From d5c30d12dfe1508328757adacc41b1b4fe7edfcd Mon Sep 17 00:00:00 2001 From: tresf Date: Thu, 12 Nov 2020 16:35:51 -0500 Subject: [PATCH] Update libffi as subtree MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit commit 8cc8f446f5aac13e107161dffbc15d1ee1a58878 Author: Mike Hommey Date: Tue Nov 10 20:41:33 2020 +0900 Allow to build with mingw-clang (#579) For some reason, compiling sysv.S with mingw-clang fails with: ``` error: invalid variant 'ffi_closure_inner@8' ``` This can be fixed (worked around?) by quoting the symbol. This works fine with mingw-gcc too. commit d817d0daa312c58548d7a081aa876027ca103766 Author: Mike Hommey Date: Tue Nov 10 20:39:25 2020 +0900 Don't use FFI_TYPE_LONGDOUBLE in the jump table in win64*.S (#580) It may have the same value as FFI_TYPE_DOUBLE per ffi.h, which possibly can make things go wrong with .org/ORG. For instance, GCC complains about "Error: attempt to move .org backwards" commit 56f7df711f70414d4f3663b34e54b122b38bab88 Author: AndreRH Date: Tue Nov 10 12:27:59 2020 +0100 aarch64: Allow FFI_WIN64 for winelib (#593) commit 8111cd06921e80d5d7192ce8d1f64733072fdbcd Author: Brandon Bergren Date: Tue Oct 27 09:07:38 2020 -0500 Add support for powerpc64le-*-freebsd*. (#581) Tests pass with no additional changes needed, tested on hardware. commit 484c1a0d7536be60002473f7bb82223efe139883 Author: Ken <21211439+kencu@users.noreply.github.com> Date: Tue Oct 27 07:07:07 2020 -0700 darwin powerpc: use go closures only if enabled (#583) SHA: 0ff9419f2e75652426469e256cb7d0748064ad58 added go closures for darwin powerpc AIX, but these have not yet been implemented for non-AIX systems use the go closures only if enabled commit 032b3cd6f7850f3ebc1269eeab7d1db3ea518d29 Author: Jeremy Huddleston Sequoia Date: Tue Oct 27 07:06:21 2020 -0700 Support building x86 and arm64 without FFI_GO_CLOSURES (#586) * x86: Support building without FFI_GO_CLOSURES Signed-off-by: Jeremy Huddleston Sequoia * arm: Support building without FFI_GO_CLOSURES Signed-off-by: Jeremy Huddleston Sequoia commit 93cf288ddd59c7cb203c1aba25de5e5b8f308da6 Author: Jeremy Huddleston Sequoia Date: Tue Oct 27 07:05:28 2020 -0700 testsuite: Add a missing include of to fix build failure in test suite (#587) Signed-off-by: Jeremy Huddleston Sequoia commit c78fcf88ab13a5cc414826a018dd46c3f886adee Author: petersn Date: Tue Oct 27 07:02:36 2020 -0700 Removing unnecessary instruction from ffi_call_unix64 (#588) unix64.S's `ffi_call_unix64` looks like it used to take six parameters, where the sixth said the number of SSE register arguments. However, currently the function only takes five parameters, and the number of SSE register arguments is encoded in the `struct register_args *` passed as the first parameter to `ffi_call_unix64`. This change removes an instruction that tries to use this missing sixth parameter as the number of SSE arguments. This fix should not change any behavior, nor fix any bugs, because a few instructions later the value moved from %r9d into %eax is overwritten by the correct value anyway. This change merely makes the code a tad less confusing, because currently the assembly moves from a register (r9) whose value is never set. commit 2d86809db6d1d626a9df40c283ec564ac98ff85d Author: jacobly0 Date: Tue Oct 27 10:00:03 2020 -0400 Fix read-only .eh_frame section test when CFLAGS contains -flto. (#590) commit f4435980fa2da335fe388db431d4a7946a99a2b9 Author: Alan Modra Date: Tue Oct 27 23:36:21 2020 +1030 Power10 libffi fixes (#585) Power10 pc-relative code doesn't use or preserve r2 as a TOC pointer. That means calling between pc-relative and TOC using code can't be done without intervening linker stubs, and a call from TOC code to pc-relative code must have a nop after the bl in order to restore r2. Now the PowerPC libffi assembly code doesn't use r2 except for the implicit use when making calls back to C, ffi_closure_helper_LINUX64 and ffi_prep_args64. So changing the assembly to interoperate with pc-relative code without stubs is easily done. * src/powerpc/linux64.S (ffi_call_LINUX64): Don't emit global entry when __PCREL__. Call using @notoc. Add nops. * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Likewise. (ffi_go_closure_linux64): Likewise. commit 407394c0aac2db4a1978a7691035889db9924a01 Author: Tim Gates Date: Sun Oct 11 15:07:40 2020 +1100 docs: fix simple typo, paramters -> parameters (#589) There is a small typo in src/x86/ffi.c. Should read `parameters` rather than `paramters`. commit ead65ca8711aa07efc0453f67136aff82f33c01f Author: Anthony Green Date: Wed Aug 19 08:34:09 2020 -0400 Re-enable s390x testing commit 8276f812a99b10d1f2c387dbd6ef2ca4f597c733 Author: Niclas Zeising Date: Fri Aug 7 21:05:23 2020 +0200 Upstream local FreeBSD patches (#567) * Add support for FreeBSD mips Add support for FreeBSD mips, this has been a local patch in the FreeBSD ports tree for quite some time. Originally submitted by sson, and committed by sbruno AT FreeBSD DOT org See https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=191909 for background details. Signed-off-by: Niclas Zeising * Add support for FreeBSD powerpcspe Add support for powerpcspe on FreeBSD This has been in the FreeBSD ports tree for some time. Originally submitted by jhibbits AT FreeBSD DOT org. Signed-off-by: Niclas Zeising * Fix abort() on FreeBSD arm related to __clear_cache() This patch has been in the FreeBSD ports tree for a number of years. Original commit by koobs AT FreeBSD DOT org Original commit message: > devel/libffi: Fix abort() on ARM related to __clear_cache() > > The current FreeBSD/ARM __clear_cache() implementation does nothing #if > __i386__ || __x86_64__ #else abort(); > > cognet@ advises this is an issue for anything !Apple that is using the > libcompiler_rt provided by Clang on ARM, and requires upstreaming. See https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=149167 for some background details. Signed-off-by: Niclas Zeising commit 392afbaf60663e530bd740da0b5a39ee9f4ba052 Author: Anthony Green Date: Wed Aug 5 21:38:51 2020 -0400 Don't test s390. Travis-ci is broken. commit e08eb1ced18598a26cd7f14fea05693e89638ed2 Author: Anthony Green Date: Wed Aug 5 17:59:59 2020 -0400 Add support for eabisim back. This was removed upstream. commit 4661ba7928b49588aec9e6976673208c8cbf0295 Merge: 45875da4b 1a4b786df Author: Tom Tromey Date: Tue Aug 4 11:08:19 2020 -0600 Merge pull request #576 from froydnj/config-updates update config.{guess,sub} commit 1a4b786df34b61af7b6738a2ce7b8d34f5377a71 Author: Nathan Froyd Date: Tue Aug 4 12:17:04 2020 -0400 update config.{guess,sub} This change is primarily for aarch64 macOS support, since that configuration is now properly supported upstream. commit 45875da4bc71578f6f45471141c14e1743239b7e Merge: fbd1de85f 8e58f88d3 Author: Tom Tromey Date: Wed Jul 1 09:32:37 2020 -0600 Merge pull request #570 from jwilk-forks/spelling aarch64: Fix typo commit 8e58f88d33eb878b0fbca1433f83bbf872d5d0f3 Author: Jakub Wilk Date: Tue Jun 30 22:59:28 2020 +0200 aarch64: Fix typo commit fbd1de85f6532f8a8f61abbd7892f8cb58b233d0 Author: Ma Jun Date: Mon Jun 29 08:02:20 2020 -0400 Add support for csky commit 0027b072627aa9d3ada71c4374561966064ecd5a Author: M. Herdiansyah Date: Thu Jun 11 19:40:24 2020 +0700 Makefile: increase compatibility with bmake (#551) commit b7af8f47d5fff2957521dc269637c00b8a4b6c75 Author: Andre Miras Date: Thu Jun 11 14:38:59 2020 +0200 Fixes macOS build on Python 3 (#563) The error was: ``` Traceback (most recent call last): File "_generate-darwin-source-and-headers.py", line 209, in generate_source_and_headers(generate_osx=not args.only_ios, generate_ios=not args.only_osx) File "_generate-darwin-source-and-headers.py", line 197, in generate_source_and_headers for header_name, tag_tuples in platform_headers.iteritems(): AttributeError: 'collections.defaultdict' object has no attribute 'iteritems' ``` commit 5c45c40c04f89cf67ade3fdfa99915ca5409e51a Author: Kentaro Hayashi Date: Thu Jun 11 21:37:41 2020 +0900 Fix Free software foundation address (#564) ref. https://www.fsf.org/about/contact/ commit 2e90bb5577ea1fe3bc1716954e790999bc3549f7 Author: Anthony Green Date: Sun Jun 7 14:31:06 2020 -0400 Add gcc bug tests back commit 5e6ca054148dfb4d78e6f79457d48b745f919def Author: Yann Sionneau Date: Fri May 1 13:59:11 2020 +0200 Add support for Kalray KVX architecture (#559) commit 4f9e20ac51ce13d46fed3c869e1deb6d9bb89444 Author: Andrew Geissler Date: Fri May 1 06:58:30 2020 -0500 ffi_powerpc.h: fix build failure with powerpc7 (#561) This is a patch pulled down from the following: https://github.com/buildroot/buildroot/blob/78926f610b1411b03464152472fd430012deb9ac/package/libffi/0004-ffi_powerpc.h-fix-build-failure-with-powerpc7.patch This issue is being hit on OpenBMC code when pulling the latest libffi tag and building on a P8 ppc64le machine. I verified this patch fixes the issue we are seeing. Below is the original commit message: Sicne commit 73dd43afc8a447ba98ea02e9aad4c6898dc77fb0, build on powerpc7 fails on: In file included from ../src/powerpc/ffi.c:33:0: ../src/powerpc/ffi_powerpc.h:61:9: error: '_Float128' is not supported on this target typedef _Float128 float128; ^~~~~~~~~ Fix this build failure by checking for __HAVE_FLOAT128 before using _Float128, as _Float128 is enabled only on specific conditions, see output/host/powerpc64-buildroot-linux-gnu/sysroot/usr/include/bits/floatn.h: /* Defined to 1 if the current compiler invocation provides a floating-point type with the IEEE 754 binary128 format, and this glibc includes corresponding *f128 interfaces for it. */ #if defined _ARCH_PWR8 && defined __LITTLE_ENDIAN__ && (_CALL_ELF == 2) \ && defined __FLOAT128__ && !defined __NO_LONG_DOUBLE_MATH # define __HAVE_FLOAT128 1 #else # define __HAVE_FLOAT128 0 #endif Fixes: - http://autobuild.buildroot.org/results/5c9dd8fb3b6a128882b6250f197c80232d8a3b53 Signed-off-by: Fabrice Fontaine Signed-off-by: Andrew Geissler Co-authored-by: Fabrice Fontaine commit 053b04c525a00abdc25ec646096a44208a1b9423 Author: Martin Storsjö Date: Sun Apr 26 05:02:14 2020 +0300 arm: Fix the clang specific version of the assembly (#556) Also fix the same error in the comment for the non-clang case. That typo there seems to have existed since the code was written in that form, in e7f15f60e86 - and when the clang specific codepath was added in e3d2812ce43, the typo in the comment made it into the actual code. commit c543849c3767cc0f0a84e3ed7b664a63724ed516 Author: Martin Storsjö Date: Sun Apr 26 05:01:03 2020 +0300 win64_armasm: Fix the spelling of ALIGN (#553) commit 15d3ea31a53bd0e5b828c90a9f7c4285c58f73e1 Author: Martin Storsjö Date: Sun Apr 26 04:59:35 2020 +0300 Use __builtin_ffs instead of ffs (#554) USE_BUILTIN_FFS is defined to 1 within __GNUC__, and the __builtin_ffs function is available since GCC 3.x at least, while the ffs function only exists on some OSes. This fixes compilation for non-x86 mingw platforms. For x86, USE_BUILTIN_FFS is explicitly disabled for windows targets - but if USE_BUILTIN_FFS is enabled based on __GNUC__, it should also use the builtin which actually is available correspondingly, not dependent on the target OS. commit c06468fa6674d3783a0edb1d0fae9afc8bc28513 Author: Martin Storsjö Date: Sun Apr 26 04:58:33 2020 +0300 Fix building for aarch64 windows with mingw toolchains (#555) * aarch64: Check _WIN32 instead of _M_ARM64 for detecting windows This fixes building for aarch64 with mingw toolchains. _M_ARM64 is predefined by MSVC, while mingw compilers predefine __aarch64__. In aarch64 specific code, change checks for _M_ARM64 into checks for _WIN32. In arch independent code, check for (defined(_M_ARM64) || defined(__aarch64__)) && defined(_WIN32) instead of just _M_ARM64. In src/closures.c, coalesce checks like defined(X86_WIN32) || defined(X86_WIN64) || defined(_M_ARM64) into plain defined(_WIN32). Technically, this enables code for ARM32 windows where it wasn't, but as far as I can see it, those codepaths should be fine for that architecture variant as well. * aarch64: Only use armasm source when building with MSVC When building for windows/arm64 with clang, the normal gas style .S source works fine. sysv.S and win64_armasm.S seem to be functionally equivalent, with only differences being due to assembler syntax. commit 8c50837f0b58ba5b2bcb1b424a2a4bfa01559fb2 Author: Fangrui Song Date: Tue Mar 10 04:36:09 2020 -0700 Improve read-write .eh_frame check (#546) llvm-objdump -h does not print BFD SEC_* constants like "READONLY", so the check will consider .eh_frame writable. clang 11 (since https://reviews.llvm.org/D73999) will error for mismatching section flags. Use readelf -S and check "WA" instead. commit 4a6414f2785a60f455c91fdd04f7745fe7fc11e0 Author: Moxie Bot Date: Mon Mar 9 21:10:53 2020 -0400 Mention ARM Pointer Authentication commit 909b37ffffe2a0bb08947b07508a3a4eed5f36a8 Author: Moxie Bot Date: Mon Mar 9 21:10:33 2020 -0400 Test on aarch64 Linux with clang commit 4c7bde32ea3af479babdf527d94f241282951cb9 Author: Ole André Vadla Ravnås Date: Tue Mar 10 02:05:42 2020 +0100 Port to iOS/arm64e (#548) commit 211e929df4388a6724b0dba4df4e3a1283dc49e9 Author: Anthony Green Date: Mon Mar 9 12:39:22 2020 -0400 Try to fix RLGL_KEY nonsense (#547) * Try to fix RLGL_KEY nonsense * Hard code policy bound API key Co-authored-by: Anthony Green commit 8eb2d2b05626b1cbbed100725bc440709499e8a6 Author: Moxie Bot Date: Mon Feb 24 10:29:20 2020 -0500 Revamp PA_LINUX and PA_HPUX target closures to use function descriptors. 2020-02-23 John David Anglin * include/ffi.h.in (FFI_CLOSURE_PTR, FFI_RESTORE_PTR): Define. * src/closures.c (ffi_closure_alloc): Convert closure pointer return by malloc to function pointer. (ffi_closure_free): Convert function pointer back to malloc pointer. * src/pa/ffi.c (ffi_closure_inner_pa32): Use union to double word align return address on stack. Adjust statements referencing return address. Convert closure argument from function pointer to standard closure pointer. (ffi_prep_closure_loc): Likewise convert closure argument back to closure pointer. Remove assembler trampolines. Setup simulated function descriptor as on ia64. src/pa/ffitarget.h (FFI_TRAMPOLINE_SIZE): Reduce to 12. src/pa/hpux32.S (ffi_closure_pa32): Retrieve closure pointer and real gp from fake gp value in register %r19. src/pa/linux.S (ffi_closure_pa32): Likewise. commit be815544c1588c6ca70120361c168b196376aa56 Author: hjl-tools Date: Sun Feb 23 07:45:42 2020 -0800 Update the ABI version to LIBFFI_BASE_8.0 (#544) Since x86 and x86-64 FFI_TRAMPOLINE_SIZE have been increased, we must bump the ABI version. This fixes: https://github.com/libffi/libffi/issues/543 commit d9abffeabe4f38bac12b864146cf974ede814411 Author: hjl-tools Date: Sat Feb 22 06:32:22 2020 -0800 x86: Fix ffi_prep_closure_loc (#542) Since FFI_TRAMPOLINE_SIZE is increased by 4 bytes to add ENDBR32, adjust jump displacement by 4 bytes. commit 624c7a35f7e1d12f917453d6c657cd5947ac57f7 Merge: 4c775d7cd 785565614 Author: Moxie Bot Date: Fri Feb 21 22:13:14 2020 -0500 Merge branch 'master' of github.com:/libffi/libffi commit 4c775d7cd6e914c6a2f66465497106cff360aeb5 Author: Moxie Bot Date: Fri Feb 21 22:12:34 2020 -0500 Update for pending 3.4 release. commit 7855656148b96c7070ec362d2a73af840025a2b7 Author: hjl-tools Date: Fri Feb 21 19:08:06 2020 -0800 x86: Add indirect branch tracking support (#540) Intel Control-flow Enforcement Technology (CET): https://software.intel.com/en-us/articles/intel-sdm contains shadow stack (SHSTK) and indirect branch tracking (IBT). When CET is enabled, ELF object files must be marked with .note.gnu.property section. When Intel CET is enabled, include in assembly codes to mark Intel CET support. Also when IBT is enabled, all indirect branch targets must start with ENDBR instruction and notrack prefix can be used to disable IBT on indirect branch. defines _CET_ENDBR which can be used in assembly codes for ENDBR instruction. If isn't included, define _CET_ENDBR as empty so that _CET_ENDBR can be used in assembly codes. Trampoline must be enlarged to add ENDBR instruction unconditionally, which is NOP on non-CET processors. This is required regardless if libffi is enabled with CET since libffi.so will be marked in legacy bitmap, but trampoline won't. Update library version for larger FFI_TRAMPOLINE_SIZE. This fixed: https://github.com/libffi/libffi/issues/474 Tested with $ CC="gcc -Wl,-z,cet-report=error -fcf-protection" CXX="g++ -Wl,-z,cet-report=error -fcf-protection" .../configure on Linux CET machines in i686, x32 and x86-64 modes. commit 4d6d2866ae43e55325e8ee96561221804602cd7a Author: Samuel Holland Date: Fri Feb 21 21:06:15 2020 -0600 Update powerpc sysv assembly for ffi_powerpc.h changes (#541) Some of the flag bits were moved when adding powerpc64 vector support. Fixes #536 commit b844a9c7f1ca792a1dfb0c09d5dae576178e6729 Author: Anthony Green Date: Fri Jan 10 10:22:10 2020 -0500 Update copyright year. commit 81b5491a4ef46103f2e1be79d368e76dba021ffa Author: Anthony Green Date: Sun Jan 5 21:37:05 2020 -0500 Fix script commit 1e08a45535d929915e14e48af38e45f8c8e4f8de Author: Anthony Green Date: Sun Jan 5 21:18:23 2020 -0500 Use rlgl API key on login commit e50b9ef8b910fa642ef158f6642e60d54d7ad740 Author: Khem Raj Date: Sat Dec 7 02:34:14 2019 -0800 powerpc64: Use memcpy to help platforms with no __int128. (#534) Signed-off-by: Khem Raj commit f9da75e157ab089363d079a781644c3e6f7db2c3 Author: Anthony Green Date: Sat Nov 30 07:37:19 2019 -0500 Remove 32-bit x86 file references to fix macosx builds commit 76c0cfea70f78586231d7994492396eebfb12d5f Author: Carl Hurd Date: Fri Nov 29 14:46:11 2019 -0500 Fixed missed #ifndef for __mips_soft_float (#442) Thank you! commit 98da256096c63406ef77f92369ada622c50ce388 Author: Anthony Green Date: Fri Nov 29 14:29:51 2019 -0500 Clean up macosx builds commit 6663047f56c2932a6b10a790f4ac6666dd181326 Author: Anthony Green Date: Fri Nov 29 07:00:35 2019 -0500 Address platforms with no __int128. commit 0069526c5adea9e6268b8a0087a6120c9c5c89fa Author: Anthony Green Date: Thu Nov 28 18:58:40 2019 -0500 Make build errors easier to debug commit 43887a9108f10124d05c8e0e67f3942cd980cb86 Author: Anthony Green Date: Thu Nov 28 17:44:51 2019 -0500 Add powerpc-eabi commit 29297445da5c2c9437de0e20f3189799a27f0301 Author: Anthony Green Date: Thu Nov 28 17:31:44 2019 -0500 Test on powerpc-eabisim commit 01a75ed76ea7e57f1b7a5c183e2b1e890e6aa0fd Author: Sergei Trofimovich Date: Thu Nov 28 12:42:41 2019 +0000 powerpc: fix build failure on power7 and older (#532) Build failure looks as: ``` libtool: compile: powerpc-unknown-linux-gnu-gcc \ -O2 -mcpu=powerpc -mtune=powerpc -pipe ... -c src/powerpc/ffi.c ... In file included from src/powerpc/ffi.c:33: src/powerpc/ffi_powerpc.h:65:9: error: '__int128' is not supported on this target 65 | typedef __int128 float128; | ^~~~~~~~ ``` The fix avoids using __int128 in favour of aligned char[16]. Closes: https://github.com/libffi/libffi/issues/531 Signed-off-by: Sergei Trofimovich commit 1ec01ea81cf18e12748082f48f7c279a3fdd505d Author: Anthony Green Date: Sun Nov 24 22:47:48 2019 -0500 Run aarch64-linux-gnu tests on travi-ci directly commit 66022e52ec0409f9170eb0884c6d0bdd048bc04e Author: Anthony Green Date: Sun Nov 24 16:52:14 2019 -0500 Fixes for ppc64le and s390x travis testing commit 932e5bb68a603885e4bd9cf1bd5146542f736ccf Author: Anthony Green Date: Sun Nov 24 16:26:14 2019 -0500 Force the use of docker commit 803db14c5a0aeb52ee521fe4fe4a7919926373de Merge: 86be66c8f fb914c366 Author: Anthony Green Date: Sun Nov 24 16:10:22 2019 -0500 Merge branch 'master' of github.com:libffi/libffi commit 86be66c8f608a84caea2ef724698093d2da2e5e2 Author: Anthony Green Date: Sun Nov 24 16:09:44 2019 -0500 Try travis-ci's new ppc64le and s390x support commit fb914c366fabb29ac373050ca0bfa4d9c1fef149 Author: Panayotis Date: Sun Nov 24 18:25:13 2019 +0200 disable obsolete 32-bit targets in macOS (#511) commit fd99c95f90f85963f5ec88630c0428a8132012db Author: Anthony Green Date: Sun Nov 24 11:11:25 2019 -0500 Minor clean-up commit 5dcb741f1544c5e18c9dbf96aeb8b61cc556a616 Author: Anthony Green Date: Sat Nov 23 10:24:58 2019 -0500 Move nested_struct3 test to closures directory commit 1aca33301a791f503d943f5e3abe2c720296eb8f Author: Anthony Green Date: Sat Nov 23 09:42:04 2019 -0500 Add missing closing brace commit d996cb283cf373936498defa902be040ec20455e Author: Anthony Green Date: Sat Nov 23 09:00:14 2019 -0500 Version 3.3 commit c72b82f465148b89f118ee73f8956dafdbac9b65 Author: Anthony Green Date: Sat Nov 23 08:48:53 2019 -0500 Remove junk file from dist commit 642d40ee6ae9a779b3c5ad930785d4b33cf94029 Author: Anthony Green Date: Sat Nov 23 07:49:58 2019 -0500 Account for moved test files commit 049da08a22b0ca6d63d2be6c209da5ac1335365c Author: Anthony Green Date: Sat Nov 23 07:44:26 2019 -0500 Add dejagnu directives accidentally removed commit 36730f5d87c2a432e1b3bf1a0129fedbf4b0a488 Author: Anthony Green Date: Fri Nov 22 19:49:38 2019 -0500 Move closure test to closure directory commit c88c0e92a9fbb4e69513394682c61aa848a035cc Author: Anthony Green Date: Fri Nov 22 19:27:34 2019 -0500 More more closure tests to the closure test directory commit a37cc1755cc276b2b9e601b9b112919f8c882a2e Merge: 332a539e8 247a5e787 Author: Anthony Green Date: Fri Nov 22 18:55:36 2019 -0500 Merge branch 'master' of github.com:libffi/libffi commit 332a539e88d26a0524c10a39283a3099d071ca10 Author: Anthony Green Date: Fri Nov 22 18:54:30 2019 -0500 Move closure tests so we can easily XFAIL them for some targets commit 1761a10645dadc0840bf1ff3a4998f9a342e8cad Author: Anthony Green Date: Fri Nov 22 18:53:09 2019 -0500 Remove gccbug detection. GCC is good now. commit 247a5e7878379b7af33338d51898d7fb9e82e686 Author: Michael Haubenwallner Date: Fri Nov 22 20:17:58 2019 +0100 Capture x86-windows host like x86_64-windows (#529) commit 8e3935fa761d3e6f19ccf5e6733fd40dee637d94 Author: Anthony Green Date: Fri Nov 22 12:42:26 2019 -0500 Manual clean-ups, and include the PDF in the source distribution. commit d01088a5194ff1a2250e7dbee04a9a167b582db7 Author: Anthony Green Date: Thu Nov 21 05:36:06 2019 -0500 Mention more major port contributors commit 970b7fce4f804df57ec582a0220c22eaeccc239a Author: Anthony Green Date: Thu Nov 21 05:04:54 2019 -0500 Update autoconf-archive m4 macros commit 34a3a661b80c0fa7817c0654ae9fd0686d3b8cd2 Author: Anthony Green Date: Thu Nov 21 04:59:39 2019 -0500 Fix formatting of README. commit 91a7fbe94e3884536cf2eb8e644bc79d6b19c80e Author: Anthony Green Date: Wed Nov 20 07:16:41 2019 -0500 Fix or1k lack-of-g++ checking in testsuite commit 31543c799a224ef446cef19a2372b054ecad3822 Author: Anthony Green Date: Wed Nov 20 06:15:55 2019 -0500 Add work-around for users who manage their own closure memory As suggested by DJ commit bd3a4687c0743076f420dee2a02075fd64fa547e Author: Anthony Green Date: Tue Nov 19 17:14:23 2019 -0500 No C++ for or1k-unknown-elf commit d6e4f96b482690e6cdaf67a70bb33d1bfe0acd15 Author: Anthony Green Date: Tue Nov 19 13:36:49 2019 -0500 No C++ for or1k commit 497018686d4df0012ea7bf1a4ab41b918b5e65bc Author: Anthony Green Date: Tue Nov 19 10:07:16 2019 -0500 Disable type warnings for or1k. commit 262cf74fd4003a928ca73e96ef8698546e535c46 Author: Anthony Green Date: Tue Nov 19 10:06:57 2019 -0500 No c++ for or1k-elf commit 54fc80dfd8d608ec718cade762c358b8fd819f38 Author: Anthony Green Date: Mon Nov 18 15:20:00 2019 -0500 Fake TRAVIS_BUILD_DIR commit 6f734f8a5ff13d106f81741c11c7cb2551706c7f Author: Anthony Green Date: Mon Nov 18 15:13:16 2019 -0500 Adapt for new old ChangeLog file commit 27d31130cabfe82683fc0e3e82247fd522f45607 Author: Anthony Green Date: Mon Nov 18 15:06:51 2019 -0500 Fix DEJAGNU variable commit da1358816e7b02e6f8431e4df336d0934bcabd4c Author: Anthony Green Date: Mon Nov 18 14:54:14 2019 -0500 Set vars commit 9a394c24951f3fc5cfbd0ff268278a7b45afc77e Author: Anthony Green Date: Mon Nov 18 14:45:29 2019 -0500 Remove verbosity commit 14bfbec0fd312001f7922ee6100cbc3155ae585b Author: Anthony Green Date: Mon Nov 18 14:40:30 2019 -0500 Pull before running commit dfa60e5eb44482d45b8c79b815e6b6789e62ce18 Author: Anthony Green Date: Mon Nov 18 14:02:33 2019 -0500 Fix cross builds for or1k commit a8223271a3d7ff1e8a1553c2a5ed9b4e1a24a774 Author: Anthony Green Date: Mon Nov 18 13:59:10 2019 -0500 More cross build debugging commit 8c9cf44cb2a2d46ddcb148508fb5350438b4a62b Author: Anthony Green Date: Mon Nov 18 13:47:42 2019 -0500 Set path to simulator. Add debugging output. commit 581048f91e9e25e3b17a2906edd8840f35999918 Author: Anthony Green Date: Mon Nov 18 13:17:27 2019 -0500 Be specific when building CI container images commit bbea6394f9f27de68aac037abb264151d6de9f64 Author: Anthony Green Date: Mon Nov 18 13:06:26 2019 -0500 Capture more or1k hosts commit 97e3b1fe1a0fbdb43bf28fe6198cb26eaa719c93 Author: Anthony Green Date: Mon Nov 18 13:02:44 2019 -0500 Add test support for or1k. Fix yaml typo. commit c88f4d2677d4dc6659a1aa6418e636f8fcf9bd9b Author: Anthony Green Date: Mon Nov 18 12:25:51 2019 -0500 Fix typo for m32r and bfin builds commit 93c203b347ea7d110355e12f832725f755e10eff Merge: d087b5956 73dd43afc Author: Anthony Green Date: Mon Nov 18 06:35:19 2019 -0500 Merge branch 'master' of github.com:libffi/libffi commit d087b595637ebb795c0cec1ff0c4409befb486b7 Author: Anthony Green Date: Mon Nov 18 06:34:42 2019 -0500 Build/test for m32r-elf and bfin-elf commit b6804d16453890b78844b1b3c1acf16af98cadb8 Author: Anthony Green Date: Sun Nov 17 09:19:14 2019 -0500 Mention binary128 long double support for Power commit 73dd43afc8a447ba98ea02e9aad4c6898dc77fb0 Author: Samuel Holland Date: Sun Nov 17 07:22:25 2019 -0600 IEEE754 binary128 long double support for PowerPC64 (#526) * powerpc: Adjust flags to make room for vector types * powerpc64 ELFv2 IEEE128 long double support commit b58bd77236e7d41fc04b4be7edd1c6728626c99b Author: Anthony Green Date: Fri Nov 15 11:18:59 2019 -0500 Re-enable mingw32 tests. commit 6f221e8269aa16f6b5eec41cfd1d7d90a9fbea0c Author: Anthony Green Date: Fri Nov 15 11:06:17 2019 -0500 Disable wine builds. commit 09dc0a71a44c95ee0d2f47de94b6b59534c2ce24 Author: Anthony Green Date: Fri Nov 15 10:19:00 2019 -0500 Stretch out timeout for test commit 28a7cc464c21b4955fba28cc55a6f095ddf5838b Author: Anthony Green Date: Fri Nov 15 06:07:51 2019 -0500 Consolidate all of the old ChangeLog files into ChangeLog.old. commit 86d3d3767a40fe1cae10c419a913dd97e7e6cd23 Author: Anthony Green Date: Wed Nov 13 10:07:00 2019 -0500 Don't build aarch64-linux-gnu commit db198a19375858879dbcc51a443c193efa0b6c0a Author: Anthony Green Date: Wed Nov 13 10:05:36 2019 -0500 Extend timeout for builds. commit 54af0256a00a40b2218950f93e5415c60f675714 Author: Anthony Green Date: Wed Nov 13 09:15:05 2019 -0500 Platform test tweaks commit 3a7580da73b7f16f275277316d00e3497cbb5a8c Author: Anthony Green Date: Wed Nov 13 07:59:18 2019 -0500 Mark java raw APIs as deprecated. commit cda60c1a8aec070a7a822c047d89d0f18cf7c5fc Author: Anthony Green Date: Wed Nov 13 06:51:44 2019 -0500 Disable mips64 tests. Fix log grabbing. commit a34254bdc1dee2857b759a22e657f92aadfa85dc Author: Anthony Green Date: Wed Nov 13 06:37:26 2019 -0500 Re-add wine based testing commit 9bd3a9742c04d2fac489f236163d7e7fbb1b7568 Author: Anthony Green Date: Wed Nov 13 06:35:40 2019 -0500 Add wine-sim.exp for wine-based testing commit de6df6dce6420d4643598c228b72da4e455a0da7 Author: Anthony Green Date: Wed Nov 13 06:32:46 2019 -0500 Fix log file URL extraction. Grab rlgl cli for OSX commit c8986e871f82b497493d28c36be1a755c2ea6b8d Author: Anthony Green Date: Wed Nov 13 06:01:43 2019 -0500 Fix log name extraction commit 6630764d244aeb3bc0797d0c3c7ccc79084061a8 Author: Anthony Green Date: Tue Nov 12 22:56:37 2019 -0500 Account for shorter ==LOGFILE== string commit aa975ec3e4ac07bd7dacec18f1c8282baeb7c398 Author: Anthony Green Date: Tue Nov 12 22:35:47 2019 -0500 Fix grep matcher commit 6dd88bc5b3f3d3d975409b7336f4127ae18da4b3 Author: Anthony Green Date: Tue Nov 12 18:44:27 2019 -0500 Debug commit 7e318cf50d8c481957b3be76e508054c70cbdf9b Author: Anthony Green Date: Tue Nov 12 17:55:35 2019 -0500 Pull test logs from cfarm commit 5123e9cf0193031bdc580cb221427cbd29ba3677 Author: Anthony Green Date: Tue Nov 12 09:21:57 2019 -0500 Use rlgl for cfarm test results commit 486f88f985affa215a9daaadcdc703c8baea9020 Author: Anthony Green Date: Tue Nov 12 09:20:21 2019 -0500 Use rlgl for cross builds. commit 042c6671b0456d17af24d61b79cbd3f9261c3f2b Author: Anthony Green Date: Tue Nov 12 09:14:20 2019 -0500 Always exit build container successfully. Don't test mingw. commit 82f7f80aa9ac3cd753bcb6d2b88cc930aee72bcc Author: Anthony Green Date: Tue Nov 12 08:58:52 2019 -0500 Run rlgl outside of the build container. commit 5b297e036e61ad1ecac9c8a3e57895737c55feb7 Author: Anthony Green Date: Tue Nov 12 05:51:09 2019 -0500 Don't bother dumping logs. Use rlgl instead. commit 2a8d88d92d56747101078c5592ab2473e6c5cb29 Author: Anthony Green Date: Mon Nov 11 21:24:58 2019 -0500 Fix rlgl usage commit 3c2b810e99a10f97d1eb6ba485eb09bfdb158084 Author: Anthony Green Date: Mon Nov 11 21:11:59 2019 -0500 Test rlgl commit 2cc11720be0c4d3bbe906be5a4aad3b2dc167072 Author: Anthony Green Date: Mon Nov 11 16:19:45 2019 -0500 Debug in-container builds commit 0de0c6a28006902c203e8dc6629cd9ef35d61e09 Author: Anthony Green Date: Sun Nov 10 08:20:46 2019 -0500 Build both iOS and Macosx commit 4e3e0586efbbc7828c15fb48e49401840beefcdd Author: Anthony Green Date: Thu Nov 7 12:35:27 2019 -0500 Fix tpyo commit f764f278160831f7e4d8ee8a48d18f5a7937f742 Author: Anthony Green Date: Thu Nov 7 12:17:29 2019 -0500 Change s390x options commit 188de63c597f23b0ccdb953b076c8cb0711fbcf0 Author: Anthony Green Date: Thu Nov 7 12:03:19 2019 -0500 Mark xfail for m68k and alpha. commit 0da221f8094ec504d91b5f88c74e98b8b9aa9eda Author: Anthony Green Date: Thu Nov 7 12:03:01 2019 -0500 Clean up. Debug s390x cpu support. commit 7e0fbf41934020c9ff76816fde5000482c7e6a6c Author: Anthony Green Date: Thu Nov 7 11:50:08 2019 -0500 Try coldfire for m68k builds commit 6a9fa770a436bbb71dbe8311e75123353e6836cc Author: Anthony Green Date: Thu Nov 7 11:37:25 2019 -0500 Add GCC_OPTIONS flag commit eebcbb1113742df7d41fc3d39eca970a026ad1d1 Author: Anthony Green Date: Thu Nov 7 11:26:53 2019 -0500 Compile m68k target for m5208 commit 08cb33ff5824b2ea24b2b62841cca1934a27f36f Author: Anthony Green Date: Thu Nov 7 11:15:13 2019 -0500 Dump supported CPUs commit 16c935000d459b1aba9e1f7b48e2c1d7dadc8cf1 Author: Anthony Green Date: Thu Nov 7 11:00:09 2019 -0500 Fix shell script commit 588e87461fe48441806be4bb5654f058e1c37649 Author: Anthony Green Date: Thu Nov 7 10:56:40 2019 -0500 Print debug output. Use podman when available. commit 04d14a66ff807a0ac2578ce886ef338aa2066b4d Author: Anthony Green Date: Thu Nov 7 09:23:10 2019 -0500 Set QEMU_CPU within the build containers commit 28d9c019dbbf6dea1ed127e1a7de829cece86639 Author: Anthony Green Date: Thu Nov 7 07:36:11 2019 -0500 Set QEMU_CPU for m68k commit 75464caf658f176848db6f48325c5067cdf94a9b Author: Anthony Green Date: Wed Nov 6 22:02:11 2019 -0500 Add alpha and m68k testing commit 259c497a2f326cc89344766e106cf73e62485a2a Author: Anthony Green Date: Wed Nov 6 11:31:17 2019 -0500 Sett CC and CXX for cross builds commit d23d3959fdb764984cfb888c5d6aed910926dd5f Author: Anthony Green Date: Wed Nov 6 11:21:46 2019 -0500 autogen before cross building commit a6dc8ef270c1f022720aedadb6b0d2678d1bc39e Author: Anthony Green Date: Wed Nov 6 11:10:59 2019 -0500 Fix typo. Don't double test. commit 136a235d3f36408628c0459d8740546acad0ca5a Author: Anthony Green Date: Wed Nov 6 10:51:33 2019 -0500 Use the right cross build container image commit 92f5309def05c66e6c5034c10d620d6fbc7a1018 Author: Anthony Green Date: Wed Nov 6 10:49:04 2019 -0500 Try s390x and sh4 cross builds/tests commit d4a28fee720c8f389b138b3f7726157fbcb7eaaf Author: Anthony Green Date: Wed Nov 6 06:47:05 2019 -0500 Revert to older test cross compilers commit 0fb9ee21e68ca3e70cf236a63ba00b5ca32031ff Author: Anthony Green Date: Wed Nov 6 06:44:19 2019 -0500 Use older test compiler. Enable s390x testing. commit 36d281ab70daf7dcf890df9c3f5bdd76f665bb68 Author: Anthony Green Date: Wed Nov 6 06:30:52 2019 -0500 Test cross-builds for hppa-linux-gnu commit c95cc4b354f24dbbf5776eb297fa036bcbfed252 Author: Anthony Green Date: Wed Nov 6 06:30:20 2019 -0500 Reference FFI_BUILDING_DLL commit 88a76477debe83ed6a75f96394853c891fb1dac4 Author: Anthony Green Date: Sun Nov 3 06:28:51 2019 -0500 rc2. hack as per: https://github.com/travis-ci/travis-ci/issues/6934 commit 45ba4aefed15be24fdbaaffec805fb02b1343e0f Author: Anthony Green Date: Thu Oct 31 19:07:04 2019 -0400 Add sparc64-linux-gnu testing back commit a26323b7e0824f0becaeaffc8d3b700dc09f741c Author: Anthony Green Date: Thu Oct 31 15:32:48 2019 -0400 Don't build sparc. Build mips. commit f325324818a6aaa9e8a75782d75c412872e6b21c Author: Anthony Green Date: Thu Oct 31 14:35:43 2019 -0400 Test mips64el-linux-gnu commit 39bad2f396de7885c9569144b5bb6e8a639ac011 Author: Anthony Green Date: Thu Oct 31 12:33:22 2019 -0400 Don't run autogen.sh all the time. commit 82719cbba06c2fc527df156d728341867d86f82c Author: Anthony Green Date: Thu Oct 31 12:23:11 2019 -0400 Test sparc64-linux-gnu on travis commit 0348d0e4fa585d9a3d9762132a3eee94f918fe3b Author: Anthony Green Date: Thu Oct 31 10:53:37 2019 -0400 Try new cfarm build for power commit 290bc3164bb0e0b3fb912c457b49e61f9a728eef Author: Anthony Green Date: Sat Oct 26 10:13:20 2019 -0400 Add make_sunver.pl to distribution. commit a8efc2f7897b439fdf2144329d35fea820498dc7 Author: Andreas Schwab Date: Sat Oct 26 14:57:05 2019 +0200 Fix FFI_STDCALL ABI (#514) Even for a stdcall function, the stack alignment is still the responsibility of the caller. Remember the original, not stack-aligned argument size, but align when setting up a stack frame. In ffi_closure_inner, return the true argument size, so that ffi_[go_]closure_STDCALL doesn't adjust too much. commit ca112537df7b9cdbccad7541aa3cb43b2a2dac9a Author: Anthony Green Date: Sat Oct 26 07:26:30 2019 -0400 Add missing build script, make_sunver.pl. commit 52b066f04fdd81c1037e1cb714a07a2a352bcf9b Author: Anthony Green Date: Thu Oct 24 07:03:43 2019 -0400 Update versions to 3.3-rc1 commit 9b10df3be2d8f81876b7cc90ee5d8833ccdf6b34 Author: Anthony Green Date: Thu Oct 24 06:41:11 2019 -0400 Remove some debugging output commit ec042885cfa5a9450bfcc87152a41529d26545de Author: Anthony Green Date: Thu Oct 24 06:39:00 2019 -0400 Don't test ppc64le with buggy qemu. commit 90fb897839bb7006a2f751d283e3d23b3bfec1ea Author: Anthony Green Date: Thu Oct 24 06:24:14 2019 -0400 Strip build status from README.md at 'make dist' time. commit 1d6059446fb986d2de933f7ccf6fe38bcdf0c76c Author: Anthony Green Date: Thu Oct 24 05:25:11 2019 -0400 Add missing dist files. commit 825b2a359468af8bf3570a5a0695e8b805c5446f Author: Anthony Green Date: Wed Oct 16 16:05:46 2019 -0400 Test on arm32v7-linux-gnu, ppc64le-linux-gnu and aarch64-linux-gnu. Use docker images and qemu to test libffi for non-x86 architectures on travis-ci. Use the LIBFFI_TEST_OPTIMIZATION environment variable to force specific optimization levels at test time. commit d2a4095af68f4530571bc3fa613dd7f5e5b815a3 Author: Anthony Green Date: Mon Oct 14 06:46:52 2019 -0400 Fix comments. commit 058aa4130445b4ef3c2f77f796c33506873031ca Author: Anthony Green Date: Mon Oct 14 05:47:07 2019 -0400 Update copyright year commit c4f61240ab19401bd86c0dfb271e243e8159abb4 Author: zhanhb <6323014+zhanhb@users.noreply.github.com> Date: Wed Oct 9 18:59:32 2019 +0800 Add long double test (#492) Required to fix build error on macos with gcc-9 commit 09f9d856112f2f105337e95e32ba9e2da63f65ae Author: pichikaudaykiran <51439150+pichikaudaykiran@users.noreply.github.com> Date: Wed Oct 9 16:26:06 2019 +0530 Making the change to correct the comment when SUN (#521) and GCC are used together commit 55c22092dc54e706a64af3a49ae9d5471a9e8317 Author: pnallan <46887249+pnallan@users.noreply.github.com> Date: Tue Oct 8 18:46:47 2019 +0530 handle compilation warnings with ftruncate API (#508) * fix me: avoid warning while handle ftruncate API Signed-off-by: Prasad Nallani * Update closures.c commit e5f0eb1552f222eb2630e40ee348b090d56412a3 Author: John Ericson Date: Tue Oct 8 06:58:52 2019 -0400 Clean up line endings (#509) The CLRF visual studio files can be kept that way, but recognized as text. The assembly file can be converted to LF. commit ea9b6639c69cbffeacd1ce0c1953c1997cf29d2e Author: Samuel Holland Date: Tue Oct 8 05:57:28 2019 -0500 PowerPC bugfixes (#520) * powerpc: Silence warnings about unused labels * powerpc: Fix a couple of comments * powerpc: Fix alignment after float structs * powerpc: Don't pad rvalues copied from FP regs * powerpc: Add missing check in struct alignment * powerpc: Support homogeneous long double structs commit 25cf9cc4a5ce4a272252de3f681d60dafd838ceb Author: Anthony Green Date: Tue Oct 8 06:24:55 2019 -0400 Use https for moxie toolchain repo commit c2a6859012d928b67a83619bd5087674a96b9254 Author: Paul Monson Date: Wed Aug 7 11:57:45 2019 -0700 fix mingw build and crashing bugs for Python Windows ARM64 (#496) * fix mingw build and crashing bugs for Python Windows ARM64 * Fix issues found in PR review commit e0b4f84fb71c6760068c9d1306e77c9382e76d8d Author: Anthony Green Date: Wed Jun 26 06:18:48 2019 -0400 Clear the apt cache commit 68668fe4f22dea7b829fb84b724f5a74091f22f2 Author: Anthony Green Date: Wed Jun 26 06:01:15 2019 -0400 More debugging output commit 1d7635d18ae4aebe4ec1cd129b0b4f71b685131e Author: Anthony Green Date: Tue Jun 25 23:39:52 2019 -0400 Debug moxie builds in travis commit d856743e6b02fcb5911491204131e277a7a4e10b Author: ossdev07 <39188636+ossdev07@users.noreply.github.com> Date: Wed Jun 26 07:31:22 2019 +0530 libffi: added ARM64 support for Windows (#486) * libffi: added ARM64 support for Windows 1. ported sysv.S to win64_armasm.S for armasm64 assembler 2. added msvc_build folder for visual studio solution 3. updated README.md for the same 4. MSVC solution created with the changes, and below test suites are tested with test script written in python. libffi.bhaible libffi.call 5. Basic functionality of above test suites are getting passed Signed-off-by: ossdev07 * Update README.md commit 80d07104c33045ea34a4d5185600495dc7461a12 Author: Anthony Green Date: Sun Apr 28 07:36:24 2019 -0400 uuencode compressed log files for travis commit fadf1eb530713fde0be9774d926bc8202c97e379 Author: Sergei Trofimovich Date: Sat Apr 27 20:53:29 2019 +0100 hppa: avoid TEXTREL in .eh_frame section (#447) Before the change hand-crafted .eh_frame section contained ABS relocation and caused TEXTREL tag to be emitted: ``` $ ./configure --host=hppa2.0-unknown-linux-gnu LDFLAGS=-Wl,-z,text $ make ... /usr/libexec/gcc/hppa2.0-unknown-linux-gnu/ld: read-only segment has dynamic relocations. ``` Link failure is caused by absolute address of FDEs encoded into .eh_frame entries. Fixed TEXTREL by using pcrel (instead of ABS) encoding for absolute addresses (__PIC__ code) by adding augmentation information ("zR" CIE type). All tests still pass on hppa2.0. The specific tests that still pass and exercise this code path: testsuite/libffi.call/unwindtest.cc testsuite/libffi.call/unwindtest_ffi_call.cc Signed-off-by: Sergei Trofimovich commit 06bf1a9deaa781b711c4d3718561e7faf303a842 Author: driver1998 Date: Sun Apr 28 03:21:44 2019 +0800 fix x86/x64 MSVC build (#487) commit db5706ff285c476aa3c0f811ff2b188319ac3ebe Author: Paul Monson Date: Fri Apr 26 04:58:58 2019 -0700 add support for 32-bit ARM on Windows (#477) * add support for 32-bit ARM on Windows * fix mismatched brace in appveyor.yml * remove arm platform from appveyor.yml for now * fix arm build * fix typo * fix assembler names * try Visual Studio 2017 * add windows arm32 to .appveyor.yml * update README.md commit d1e9b4b96cc8d237c3532cf83da0d4b99d19abb5 Merge: 05a179641 a7d6396f0 Author: Tom Tromey Date: Tue Apr 9 14:00:10 2019 -0600 Merge pull request #482 from sharkcz/aarch64 fix check for Linux/aarch64 commit a7d6396f06d5e1a726f24b746b509514f466380f Author: Dan Horák Date: Fri Mar 29 14:19:20 2019 +0100 fix check for Linux/aarch64 fixes #473 commit 05a1796419f68267250c0b8ae7138da36ab60b01 Author: Jeremy Huddleston Sequoia Date: Tue Feb 19 04:11:28 2019 -0800 Cleanup symbol exports on darwin and add architecture preprocessor checks to assist in building fat binaries (eg: i386+x86_64 on macOS or arm+aarch64 on iOS) (#450) * x86: Ensure _efi64 suffixed symbols are not exported * x86: Ensure we do not export ffi_prep_cif_machdep Signed-off-by: Jeremy Huddleston Sequoia * x86: Ensure we don't export ffi_call_win64, ffi_closure_win64, or ffi_go_closure_win64 Signed-off-by: Jeremy Huddleston Sequoia * closures: Silence a semantic warning libffi/src/closures.c:175:23: This function declaration is not a prototype Signed-off-by: Jeremy Huddleston Sequoia * aarch64: Ensure we don't export ffi_prep_cif_machdep Signed-off-by: Jeremy Huddleston Sequoia * arm: Ensure we don't export ffi_prep_cif_machdep Signed-off-by: Jeremy Huddleston Sequoia * aarch64, arm, x86: Add architecture preprocessor checks to support easier fat builds (eg: iOS) Signed-off-by: Jeremy Huddleston Sequoia * x86: Silence some static analysis warnings libffi/src/x86/ffi64.c:286:21: The left operand of '!=' is a garbage value due to array index out of bounds libffi/src/x86/ffi64.c:297:22: The left operand of '!=' is a garbage value due to array index out of bounds Signed-off-by: Jeremy Huddleston Sequoia * aarch: Use FFI_HIDDEN rather than .hidden Signed-off-by: Jeremy Huddleston Sequoia * ffi.h: Don't advertise ffi_java_rvalue_to_raw, ffi_prep_java_raw_closure, and ffi_prep_java_raw_closure_loc when FFI_NATIVE_RAW_API is 0 Signed-off-by: Jeremy Huddleston Sequoia commit 3ec522bc1c1c5a7fb66b85d4fb6a0fe80581d597 Author: dabrain34 Date: Tue Feb 19 13:09:09 2019 +0100 Fix cfi checks for old compiler (#453) cfi_sections can be unsupported when cfi_startproc and cfi_endproc are. commit e1118af50599314a2cbac3eb51a81896e8e21d0c Author: Paul Monson Date: Tue Feb 19 03:58:25 2019 -0800 changes for win32 on windows (#468) commit 44a6c28545186d78642487927952844156fc7ab5 Author: Florian Weimer Date: Tue Feb 19 12:55:11 2019 +0100 aarch64: Flush code mapping in addition to data mapping (#471) This needs a new function, ffi_data_to_code_pointer, to translate from data pointers to code pointers. Fixes issue #470. commit 042ef8c314a946ef1cd58c6e10cd74e403ef5bf9 Author: Anthony Green Date: Tue Feb 12 08:50:30 2019 -0500 Remove -Os testing. No ABI impact, and helps trim log lengths. commit 8fa88373a1d433c675b11200ccd58418e91f81e4 Merge: 737d4faa0 2c5b16428 Author: Tom Tromey Date: Wed Dec 19 14:18:40 2018 -0700 Merge pull request #461 from NativeScript/bektchiev/fix-visibility-hidden-check-in-configure fix(configure): Correctly detect visibility("hidden") support on Darwin commit 2c5b164288712cc048048d73a3cd841d845a132c Author: Martin Bektchiev Date: Wed Dec 19 18:33:08 2018 +0200 fix(configure): Correctly detect visibility("hidden") support on Darwin commit 737d4faa00d681b4c758057f67e1a02d813d01c2 Merge: a5ea7527c 4a84df4ae Author: Tom Tromey Date: Fri Nov 30 08:16:00 2018 -0700 Merge pull request #457 from NativeScript/bektchiev/fix-arm64-q3-q4-args Fix Q registers parameter passing on ARM64 commit 4a84df4ae9d33bb766fb34ce8a871d84d6e0ed9c Author: Martin Bektchiev Date: Wed Oct 31 15:53:54 2018 +0200 Fix Q registers parameter passing on ARM64 The second two quads are located at offset 32 not 16 commit a5ea7527cda8d9b7d011eb4004dfcbad54583bd2 Merge: b11b08ca8 ba73a671c Author: Tom Tromey Date: Wed Sep 19 07:29:36 2018 -0600 Merge pull request #443 from jeremyhu/master Update FFI_HIDDEN() to use .private_extern on Apple platforms and use the macro where appropriate commit b11b08ca8daacb08943ae5ea8c124771a8b82dde Merge: 980908b47 e6eac7863 Author: Tom Tromey Date: Tue Sep 18 08:26:31 2018 -0600 Merge pull request #449 from gpakosz/align-macros Prefix ALIGN_DOWN macro with FFI_ commit e6eac7863e2bf1a009ea863041b354bdb4af6b67 Author: Gregory Pakosz Date: Tue Sep 18 15:19:53 2018 +0200 Prefix ALIGN_DOWN macro with FFI_ commit 980908b47bbde09fab88ee4b2b61d8bc5d4378a7 Merge: 65da63abc 4cb776bc8 Author: Tom Tromey Date: Sat Aug 11 09:53:15 2018 -0600 Merge pull request #445 from andreas-schwab/master RISC-V go closures commit 4cb776bc8075332d2f3e59f51785d621fcda48f6 Author: Andreas Schwab Date: Thu Aug 9 12:12:29 2018 +0200 RISC-V go closures This implements go closures for RISC-V. It has been tested on riscv64-suse-linux and against the libgo testsuite. commit ba73a671cb49f8e2c4417723a9dc233e328926ae Author: Jeremy Huddleston Sequoia Date: Fri Jul 20 09:37:43 2018 -0700 Update FFI_HIDDEN() to use .private_extern on Apple platforms and use the macro where appropriate Fix issue #439 Signed-off-by: Jeremy Huddleston Sequoia commit 65da63abc843fe448aaa86015d094cf016f325ba Author: Jeremy Huddleston Sequoia Date: Mon Jun 25 04:38:58 2018 -0700 Add compact unwind for darwin/i386 (#440) * x86: Add implementation of compact unwind for ffi_call_i386, et al. Signed-off-by: Jeremy Huddleston Sequoia * x86: Use __text as the section name to avoid deprecated section name warnings. Signed-off-by: Jeremy Huddleston Sequoia * darwin: Add missing regular,debug attributes for compact unwind sections Signed-off-by: Jeremy Huddleston Sequoia commit 2309b58448b0c8349156d687d7fa8709dfb68992 Author: Shoaib Meenai Date: Sun Jun 17 17:04:24 2018 -0700 Mark sysv.S as SafeSEH compatible (#438) It contains no exception handler, so we can just emit the special @feat.00 symbol to indicate that it's trivially SafeSEH compatible. SafeSEH only applies to x86 and not x86-64, hence its inclusion in the x86-specific block. See [1] for details. [1] https://msdn.microsoft.com/en-us/library/windows/desktop/ms680547(v=vs.85).aspx#the_.sxdata_section_ commit 1d704051b2da207d715351b8613e74437d8e2eb7 Author: Jeremy Huddleston Sequoia Date: Sun Jun 17 17:01:50 2018 -0700 i386: Fix missing break; in case statement leading to incorrectly returned FFI_BAD_ABI (#437) * i386: Add missing break triggering dead store static analyzer checks. Register calling sequence is being reported as bad ABI instead of working as intended. Found-by: Clang Static Analysis Signed-off-by: Jeremy Huddleston Sequoia * Mark ffi arm sysv entry points as private_extern. Signed-off-by: Jeremy Huddleston Sequoia * x86_64: Add implementation of compact unwind for ffi_call_unix64. Signed-off-by: Jeremy Huddleston Sequoia commit b55baf0b500ccc7636a8a55e0506d9da787ad2dd Author: Anthony Green Date: Wed May 9 13:21:02 2018 -0400 Handle FFI_GNUW64 on non-Windows systems (EFI) commit 8206253fdf01226173c36f087226d8ea53584566 Author: Anthony Green Date: Wed May 9 10:50:46 2018 -0400 Mark some cases as xfail due to GCC bug commit b5ee395710e2db830749c3c95558c91b66ac111f Author: Anthony Green Date: Sat May 5 07:41:53 2018 -0400 Revert "Remove some symbol exports and cleanup newline warnings (#433)" This reverts commit a5a0f3cf36dfb4d64316414a872288c3170e6c1d. commit a5a0f3cf36dfb4d64316414a872288c3170e6c1d Author: Jeremy Huddleston Sequoia Date: Sat May 5 03:44:33 2018 -0700 Remove some symbol exports and cleanup newline warnings (#433) * build: Ensure darwin generated sources end with a new line Signed-off-by: Jeremy Huddleston Sequoia * build: Use .private_extern where missing to prevent exporting symbols that are not API Signed-off-by: Jeremy Huddleston Sequoia commit d3c54cf3a2b2bb2e889173b6a0a959517b42c47f Author: hjl-tools Date: Wed May 2 06:19:58 2018 -0700 Re-enable msabi testing (#436) * Revert "disable msabi testing for now" This reverts commit 7b7638eb0eac2adfa72f7ec9f254ba287c9947e2. * x86: Correct testing for 64-bit x86-64 Since passing -m32 to x86-64 compiler will generate i386 code, we need to check both __ILP32__ and __i386__ for 64-bit x86-64. * x86: Check __i386__ instead of i?86-*-* targets Since passing -m32 to x86-64 compiler will generate i386 code, we need to check __i386__ instead of i?86-*-* targets for i386 targets. * i386: Properly passing integer parameters in registers For thiscall and fastcall, if the paramter passed as 64-bit integer or struct, all following integer paramters will be passed on stack. * test: Add ABI_ATTR to callback_code Add ABI_ATTR to callback_code to properly test different ABIs. commit ebf241663cc0fd0d76e4db8101e842e1630a2c78 Author: Alan Modra Date: Wed May 2 13:55:29 2018 +0930 PowerPC64 ELFv1 fp arg fixes The ELFv1 ABI says: "Single precision floating point values are mapped to the second word in a single doubleword" and also "Floating point registers f1 through f13 are used consecutively to pass up to 13 floating point values, one member aggregates passed by value containing a floating point value, and to pass complex floating point values". libffi wasn't expecting float args in the second word, and wasn't passing one member aggregates in fp registers. This patch fixes those problems, making use of the existing ELFv2 homogeneous aggregate support since a one element fp struct is a special case of an homogeneous aggregate. I've also set a flag when returning pointers that might be used one day. This is just a tidy since the ppc64 assembly support code currently doesn't test FLAG_RETURNS_64BITS for integer types.. * src/powerpc/ffi_linux64.c (discover_homogeneous_aggregate): Compile for ELFv1 too, handling single element aggregates. (ffi_prep_cif_linux64_core): Call discover_homogeneous_aggregate for ELFv1. Set FLAG_RETURNS_64BITS for FFI_TYPE_POINTER return. (ffi_prep_args64): Call discover_homogeneous_aggregate for ELFv1, and handle single element structs containing float or double as if the element wasn't wrapped in a struct. Store floats in second word of doubleword slot when big-endian. (ffi_closure_helper_LINUX64): Similarly. commit a2c6c7af9ffb9500a944abc666ced0922d0de762 Author: Anthony Green Date: Tue May 1 06:46:30 2018 -0400 Add license for build-time tools commit 4c2206ace07f2fb4bef43cd4bfe952ccb584dcec Author: Tom Tromey Date: Sat Apr 28 04:46:10 2018 -0600 Fix two "return" issues in x86/ffi64.c (#431) Issue #70 pointed out that at least one compiler didn't like: return ffi_call_efi64(cif, fn, rvalue, avalue); ... where the return type is "void". This patch splits the statement into two. I also noticed that ffi_call_go here seems to do a double call. I suspect a "return" is missing here, so this patch adds it as well. commit ed3ed4d801a4d417ea304715d4d8ae581a6c6903 Merge: 8e250c390 b69f95d65 Author: Tom Tromey Date: Fri Apr 27 10:48:51 2018 -0600 Merge pull request #429 from yousong/m4subst build: fix subst toolexecdir, toolexeclibdir commit b69f95d6562b6a57c9be36da0cdf2a500adb94e9 Author: Yousong Zhou Date: Thu Apr 26 19:27:54 2018 +0800 build: fix subst toolexecdir, toolexeclibdir On CentOS 7.4, configure script generated by autogen.sh will output the following values when compiling GCC source code toolexecdir='NONE/$(target_alias)' toolexeclibdir='NONE/$(target_alias)/lib' and cause build error ... -o libffi.la -rpath NONE/riscv64-bs-linux-gnu/lib ... checking for shl_load in -ldld... libtool: link: only absolute run-paths are allowed commit 8e250c390a45f722387c1881ae5f99849a2d1e22 Author: Stephen Date: Wed Apr 25 13:17:32 2018 -0700 update to https for cygwin download (#428) commit 7d3cab7926d08aad9a8e54420d6878cb17efd185 Author: Lucas Pluvinage Date: Sat Apr 21 00:24:50 2018 +0200 xtensa-linux: use cache flush instruction only if it is available (#426) commit 159b94e5fd4aa2d88e1b5b389092cefd9472a741 Author: James Cowgill Date: Thu Apr 19 01:28:23 2018 +0100 Various MIPS Fixes (#425) * mips: simplify closure #defines This commit should have no visible effect. * mips: add special handling of variadic functions MIPS requires special handling of variadic functions which pass floating point arguments: * In the o32 ABI, all float arguments are passed in integer registers. * In the n32/n64 ABIs, float arguments after the ellipsis are passed in integer registers. Implement this in libffi. To support this in n32/n64 closures, we need to add a new mips_nfixedargs field to ffi_cif which will break the libffi ABI. This fixes the libffi.call/cls_longdouble_va.c test which was failing on 64-bit MIPS. * mips: align argn for all 64-bit types in o32 closure handler Ensure that argn is pre-aligned for all 64-bit argument types (including doubles) and not just integer types. This fixes closures of the form "f(float, double, )". Previously the first integer argument would be read from a2 which is garbage at this point (the float arguments have already "consumed" a0-a3). After this commit, argn is correctly padded between the "float" and "double" arguments so that the first integer argument is read from the stack. Fixes "double f(float,double,int)" test in #371 * mips: do not read from floating point register if returning a struct In the o32 ABI, the pointer passed in a0 used to return structures indirectly is treated as the first argument for argument allocation purposes. This means that it should inhibit floating point registers the same way that other integer arguments do. Fixes "Double f(float,Double,double)" test in #371 * mips: fix pointer cast warnings Fix two pointer cast warnings when compiled on 64-bit mips by casting through uintptr_t. Fixes mips64el part of #404 commit f2afda08e85b02888350449dcf39a6c37cfb7bc4 Merge: e27f70b8c 801c1bd71 Author: Anthony Green Date: Sun Apr 8 18:25:43 2018 -0400 Merge branch 'master' of github.com:/libffi/libffi commit e27f70b8cf2a537bef84b2cb29ad8ea6209a11b8 Author: Anthony Green Date: Sun Apr 8 18:25:34 2018 -0400 Fix case where callback arg value is split across regs and stack commit 801c1bd712ff8c76675b7aa69c29948907f1eeff Author: Andreas Krebbel <38103320+Andreas-Krebbel@users.noreply.github.com> Date: Thu Apr 5 14:27:32 2018 +0200 Fix issue #421 (#422) Fantastic - thanks for digging into this. commit 8660e6935971c5abd7b528eaf54deeccd4bbaccd Author: Anthony Green Date: Mon Apr 2 08:30:17 2018 -0400 3.3 release candidate 0 commit 7b7638eb0eac2adfa72f7ec9f254ba287c9947e2 Author: Anthony Green Date: Mon Apr 2 08:24:44 2018 -0400 disable msabi testing for now commit af6773d6ab4db0577bc6b932ab5a2f98a0a8dca2 Author: fwg Date: Mon Apr 2 13:55:31 2018 +0200 Fix appveyor windows build (#420) * Fix msvcc dll build by adding dllexport decorations to all API declarations * Fix appveyor build for VS 2013 Use the new -DFFI_BUILDING_DLL for producing a working DLL. Update the msvcc.sh wrapper script to successfully compile the testsuite files. * MSVC build: suppress warnings in testsuite * fix testsuite on appveyor commit 48bdb02867edb7e9f3785ccb4bdff1087fb44246 Author: Anthony Green Date: Thu Mar 29 07:22:57 2018 -0400 Trim some optimization tests commit f98e0f0d777bf962057e18d036989b2b89fdf416 Author: Anthony Green Date: Thu Mar 29 07:22:38 2018 -0400 Simplify matrix commit fa72b054e10acaea33ec8cc395e0ea5b31c9ac9e Author: Anthony Green Date: Thu Mar 29 07:10:23 2018 -0400 Remove warning message from clang commit 746c3ce220cb155a8f3b613550a7ecad76f9fedc Author: Anthony Green Date: Thu Mar 29 07:01:14 2018 -0400 Expand ABI tests on x86. Testsuite bug fixes. commit 2eee934d0cb2beef7c0aa29767eca1ce03452ae9 Author: Anthony Green Date: Tue Mar 27 14:45:40 2018 -0400 deal with msvc warnings commit 206b2974f70f65b4c4c4d9e62b03ee7e0753ee2d Author: Anthony Green Date: Tue Mar 27 14:37:50 2018 -0400 fix quoting issue commit 499e41c1ec6b760eff2b320ffc87e22de82a9e1d Author: Anthony Green Date: Tue Mar 27 14:35:10 2018 -0400 Don't ignore a file commit 5b41e9f6d2c099add0e02db8c8054281f96b6d0e Author: Anthony Green Date: Tue Mar 27 14:34:31 2018 -0400 Try different msvc hack commit e8cf133813e7fa2c606d93abab5b68989e5f6cbc Author: Anthony Green Date: Tue Mar 27 14:12:02 2018 -0400 msvc c99 hack commit a3e20940898ed2ca96b9df87f34117d06d900e5f Author: Anthony Green Date: Tue Mar 27 11:58:42 2018 -0400 More msvc hacks commit a82b456e98b0f3fbafe76725d1a1503d458a2178 Author: Anthony Green Date: Tue Mar 27 11:49:46 2018 -0400 msvc fixes commit a2326aaf561798c3bfd5e464b16693a094bf6c45 Author: Anthony Green Date: Tue Mar 27 11:36:04 2018 -0400 Remove debug output commit 85b6b209d7984667ce7d00351fe7e79ef6c4930c Author: Anthony Green Date: Tue Mar 27 11:35:23 2018 -0400 Force literals to float (msvc warning) commit 2872c460ba8e96416b4f4ee4a0ed183461d8d6c9 Author: Anthony Green Date: Tue Mar 27 11:28:40 2018 -0400 Fix msvc linking commit b40a386c22a2bd164819b4f398b722357222b7c1 Author: Anthony Green Date: Tue Mar 27 11:20:21 2018 -0400 Fix library path handling commit d78c5f64f834c24a769bee78cfb7db4d468b6b21 Author: Anthony Green Date: Tue Mar 27 11:14:56 2018 -0400 Deal with libpath commit f318bb63c5dc623d5a69ffc4a59a330050cac567 Author: Anthony Green Date: Tue Mar 27 08:25:26 2018 -0400 Fix library path handling commit 4f0c9cb39b7599312729278380ce63247bca226b Author: Anthony Green Date: Tue Mar 27 08:10:03 2018 -0400 Use mixed cygpath paths commit d3693b4665de7ef69b61fb4e483d5686302f9c69 Author: Anthony Green Date: Tue Mar 27 08:05:42 2018 -0400 Copy - don't symlink - to support msvc builds commit 51ce3696cd0da4716d86b73a30bd89e139ec4b74 Author: Anthony Green Date: Tue Mar 27 07:35:33 2018 -0400 Use cygpath to convert windows paths commit 70b93e4d863c1ff2a6498455339e17df3d7fe784 Author: Anthony Green Date: Tue Mar 27 07:14:59 2018 -0400 try to fix windows builds commit 377134d7a946df13b60ff9c9dacc895e82c36178 Author: Anthony Green Date: Tue Mar 27 07:06:01 2018 -0400 Reduce verbosity commit d01ef5b875dedcb9f7b3039b2334cbd594fdfcd4 Author: Anthony Green Date: Tue Mar 27 06:23:29 2018 -0400 iOS builds commit d13583f55fb31a302684cbc62c300ae81eb577e9 Author: Anthony Green Date: Tue Mar 27 06:22:57 2018 -0400 Fix -L support commit d49911544bf9d2b4fab782b0c5e76a68074223df Author: Anthony Green Date: Tue Mar 27 05:59:26 2018 -0400 Fix quoting commit 8f47ec404fcf5a082fc4537ede8cdf897f4c5226 Author: Anthony Green Date: Tue Mar 27 05:35:35 2018 -0400 Don't set AM_MAKEFLAGS commit df314193dc6cd00978b6e1517866448b604199af Author: Anthony Green Date: Tue Mar 27 05:13:38 2018 -0400 Run tests with lots of debug output commit b394947518fb15887802fe3bc7f0a337c495d65d Author: Anthony Green Date: Tue Mar 27 05:11:52 2018 -0400 Add -L and -l support commit 33da6c075299317885385f19680b58a1e9feab5e Author: Anthony Green Date: Tue Mar 27 05:07:08 2018 -0400 Mention C99 and build requirements commit e6bdbd7888b5c8a90f3c91f4d6caf3c8d574f3b5 Author: Anthony Green Date: Tue Mar 27 04:46:04 2018 -0400 Remove reference to wiki commit 24eb1f88d788399dde5fbb4804767e6b6cbf1bc8 Author: Anthony Green Date: Tue Mar 27 04:09:56 2018 -0400 Refactor travis build instructions. Add iOS build commit ffc3ff162a15ed0ed0e946815c950c69ee0c8343 Author: Anthony Green Date: Tue Mar 27 04:02:06 2018 -0400 Add issue template commit 5c2ca479e726dc2feec43e25475c63361387bf20 Author: Anthony Green Date: Tue Mar 27 04:01:37 2018 -0400 Remove uninitialized warning. Fix #163. commit 0081378017c33a4b9b6fbf20efabdd9959d6a48d Author: Anthony Green Date: Sat Mar 24 09:05:14 2018 -0400 revert commit 59d44242e15d2979291fe6793ddfb2681b7480ef Author: Anthony Green Date: Sat Mar 24 08:45:59 2018 -0400 Try bhaible tests in appveyor commit 6f7c29c5a6d2850ce53b99efcc71fcc4e8f8c8cc Author: Anthony Green Date: Sat Mar 24 08:40:04 2018 -0400 Formatting fixes commit f3c2729ee4026c9f70a6656603bac8ba2ec40a5d Author: Anthony Green Date: Fri Mar 23 11:47:29 2018 -0400 Add libffi.map.in to extra dist files commit 6d2233f96a08c5ac01a64e7b84e2bd678eed0a4b Author: Anthony Green Date: Fri Mar 23 09:06:08 2018 -0400 Add sparc solaris build notes commit 99a80d2fc4ede48348af2edfb7126d60609c4b71 Author: Anthony Green Date: Sun Mar 18 21:22:55 2018 -0400 Fix formatting commit a3e87ac8e7b7e609ab84541a69c7e51873efa437 Author: Anthony Green Date: Sun Mar 18 20:36:52 2018 -0400 Fix formatting commit cec3a3a201f17a7f018f25e1a0917bd5206e5a5a Author: Anthony Green Date: Sun Mar 18 16:00:04 2018 -0400 Don't align-double for 32-bit x86 commit f34f8039c10c4d8d728bc335dcf2b2a38f6e8b50 Author: Anthony Green Date: Sun Mar 18 15:30:00 2018 -0400 Remove debug output commit f31b915314b12904b858b54cfdf670ceaeda7b2d Author: Anthony Green Date: Sun Mar 18 15:13:21 2018 -0400 Add m4/ax_require_defined.m4 commit a2dc5848b53e9e45ff68e5ef70683b5ffb82592e Author: Anthony Green Date: Sun Mar 18 13:41:15 2018 -0400 travis debug commit 81d345b2cff039c4389fad283a108ad978b609e0 Author: Anthony Green Date: Sun Mar 18 13:32:03 2018 -0400 Update autoconf archive macros commit 369ef49f71186fc9d6ab15614488ad466fac3fc1 Author: Anthony Green Date: Sun Mar 18 12:53:42 2018 -0400 Add missing FFI_GNUW64 enum commit 56655e58c7f966685c0d5635ee7215733121ae41 Merge: 43980dd10 9bc40d87e Author: Anthony Green Date: Sun Mar 18 12:34:18 2018 -0400 Merge branch 'master' of github.com:/libffi/libffi commit 43980dd10e0d5705e52f24d63988b722e31bf330 Author: Anthony Green Date: Sun Mar 18 12:32:10 2018 -0400 Add FFI_GNUW64 ABI for GNU 80-bit long double support commit 9bc40d87ea5950969b234ed56cd2c6acd883fa0e Author: Anthony Green Date: Sun Mar 18 12:32:10 2018 -0400 Add FFI_GWIN64 ABI for GNU 80-bit long double support commit d46406088d28b038a0a0f7396d9621f431482f6a Author: Ryan C. Underwood Date: Sun Mar 18 07:00:42 2018 -0700 Fully allocate file backing writable maps (#389) When ftruncate() is used on a filesystem supporting sparse files, space in the file is not actually allocated. Then, when the file is mmap'd and libffi writes to the mapping, SIGBUS is thrown to the calling application. Instead, always fully allocate the file that will back writable maps. commit 247e44b3ef653f210de614a749b71449b8c70764 Author: Anthony Green Date: Sun Mar 18 07:01:54 2018 -0400 Fix return values commit 1f99701fad1761ffe4e9adf88cfe26f74f427e7c Author: Anthony Green Date: Sat Mar 17 22:49:58 2018 -0400 Make tests compile/run standalone commit d974207cd6373ff282a64038564d99b853dc3fab Author: Anthony Green Date: Sat Mar 17 09:03:42 2018 -0400 Formatting fixes. Add 'bug fixes' note. commit a33bfa9b12993aae2edf669f62e141a84286d4de Author: Anthony Green Date: Sat Mar 17 07:17:24 2018 -0400 xfail unwind tests for moxie commit 8bf05481436ba085df3adfff2b47a879df738d67 Author: Anthony Green Date: Fri Mar 16 23:54:28 2018 -0400 Install moxie libstdc++ commit 84b383eda67ba75a6456e4b3587f17b749e6e1a1 Author: Anthony Green Date: Fri Mar 16 23:36:51 2018 -0400 Install moxie-elf-g++ commit 6a801d042ac1212fe8c0e11abb865061e515aa43 Author: Anthony Green Date: Fri Mar 16 17:53:33 2018 -0400 Fix closure case where 8-byte value is partially passed in register. Fixes cls_many_mixed_float_double test case. commit 0f0ba9e0ec52e0297c91cb63a889980a6f847219 Author: Anthony Green Date: Fri Mar 16 07:28:33 2018 -0400 Run dejagnu's runtest with -a option to prevent travis timeouts after 10min commit 725454196afab82f7da7bd719e8bddf7cd7837dd Author: Anthony Green Date: Thu Mar 15 08:17:18 2018 -0400 fix .travis.yml commit d71051c46a795a0669b2d83804030746f790edf4 Author: Anthony Green Date: Wed Mar 14 21:21:06 2018 -0400 fix .travis.yml commit e86aa54ce3418081dca403217c9216c773b82560 Author: Anthony Green Date: Wed Mar 14 21:00:55 2018 -0400 fix .travis.yml commit 5b538ed61fb13efc8f01bad014c5485231714c96 Author: Anthony Green Date: Wed Mar 14 20:32:00 2018 -0400 fix .travis.yml commit 6ed64a7953178cf07d2793650fcd64caa6b51d53 Author: Anthony Green Date: Wed Mar 14 20:22:33 2018 -0400 fix .travis.yml commit 90809e928f8b5a0cdf0787d799111bc33689a9d1 Author: Anthony Green Date: Wed Mar 14 14:09:39 2018 -0400 fix .travis.yml commit 85e4c62351bd1e45dd6d77916778e18c435f1723 Author: Anthony Green Date: Wed Mar 14 13:47:27 2018 -0400 fix .travis.yml commit aa8a5ec4a4ec4a4460ec3cd4e396e09ceaa3a2fa Author: Anthony Green Date: Wed Mar 14 13:25:04 2018 -0400 fix .travis.yml commit 7c3e5da401ab7efef83a5bd4e7f64b0df053da0c Author: Anthony Green Date: Wed Mar 14 13:07:04 2018 -0400 fix .travis.yml commit 19aec1989382fab8bd86c556aa24751a75b8246e Author: Anthony Green Date: Wed Mar 14 12:40:58 2018 -0400 Fix travis.yml commit bf337b070977fb7e8934fc91e4d0e60209f53448 Author: Anthony Green Date: Wed Mar 14 12:15:04 2018 -0400 yaml syntax error fix commit 076d5a987e016669993154b91525b02978c56727 Author: Anthony Green Date: Wed Mar 14 11:13:32 2018 -0400 markdown fixes commit d2f7e788aecf37971c9be1cd650f78a3e4061e18 Author: Anthony Green Date: Wed Mar 14 09:06:38 2018 -0400 Fix test cases with short results commit 18fe190068563bb2bf52a0993017887f438c85a4 Author: Anthony Green Date: Wed Mar 14 08:02:10 2018 -0400 Debug travis output commit 718a4177f23e9d6c5476d0118d49a3a093d72c6c Author: Anthony Green Date: Wed Mar 14 07:40:53 2018 -0400 test commit 486c34b65deb43d6d9025b2a40284606f97f4e0b Author: Anthony Green Date: Tue Mar 13 23:31:18 2018 -0400 debug commit 00464aa52022674bdc1f5aa9b2ceb575c9a04b87 Author: Anthony Green Date: Tue Mar 13 23:30:52 2018 -0400 debug commit eb0544d0d42e5aef56ee1df0861bcd0c95102f88 Author: Anthony Green Date: Tue Mar 13 23:20:15 2018 -0400 Debug env commit 9d5bd29d8dfb5f89d0927a175443240a19e93890 Author: Anthony Green Date: Tue Mar 13 23:07:10 2018 -0400 Refactor commit 52a364d7699230be5e883f32b9ac0caaea0300b1 Author: Anthony Green Date: Tue Mar 13 22:49:48 2018 -0400 fix moxie build commit 1a32b3f146f7ade701cea74e3983b16145f93393 Author: Anthony Green Date: Tue Mar 13 22:44:13 2018 -0400 fix moxie build commit e8958c3a64b5e59f0477bcaafaf9c03337d9f445 Author: Anthony Green Date: Tue Mar 13 22:36:41 2018 -0400 revert test commit 58315ae125482c44821cfbdce41e350380024d0e Author: Anthony Green Date: Tue Mar 13 22:35:50 2018 -0400 test commit 6239c28741a8abc1bd06cb47329d8c5cdb359c90 Author: Anthony Green Date: Tue Mar 13 22:13:44 2018 -0400 Add experimental moxie travis build commit 01db31d94e97e71836fe862826bc8a2a399dd619 Author: Anthony Green Date: Tue Mar 13 20:41:55 2018 -0400 Update moxie sub opcode commit e8f5a43033d53ad631ba477b78a3def61783a1dc Author: Anthony Green Date: Tue Mar 13 09:27:00 2018 -0400 Fix appveyor badge commit d1689dd3bf78f4975aa6a6a16cbffcb794283acb Author: Anthony Green Date: Tue Mar 13 09:24:37 2018 -0400 Fix travis badge. Add REAME.md to dist files commit 785351fbc2a49f05535c412067ac51f906612779 Author: Anthony Green Date: Tue Mar 13 09:21:49 2018 -0400 Preformat release notes commit 8c7fbd1616b075e3077346c7bf56befd723a9355 Author: Anthony Green Date: Tue Mar 13 09:20:44 2018 -0400 Preformat release notes commit 54f31b00562e75e6aae2b736c70e6713618640cb Author: Anthony Green Date: Tue Mar 13 09:17:57 2018 -0400 Try to fix table commit 3a29a27a86624512f405e4c9962276485517caf1 Author: Anthony Green Date: Tue Mar 13 09:14:40 2018 -0400 Clean up markdown table commit fbfc3b048fcf3f699f4495ca70b7e58461a0df4f Author: Anthony Green Date: Tue Mar 13 09:12:21 2018 -0400 Move to markdown commit a31d3d3ba6644d308b6fb8cea2a60976e0531f4a Author: Anthony Green Date: Tue Mar 13 09:11:24 2018 -0400 Add travis and appveyor build badges commit 16313cb2746ceed6bc1d25b52604642c14758e01 Author: Anthony Green Date: Tue Mar 13 08:52:30 2018 -0400 Support compiler specific warning suppression flags commit cca6d1fb549e422514df4aad68f2ca9a8313327d Author: Anthony Green Date: Tue Mar 13 08:51:34 2018 -0400 Support compiler specific warning suppression flags commit 9291f941c2df5ddf967f701b258ac2b3cc9d0820 Author: Anthony Green Date: Tue Mar 13 08:37:21 2018 -0400 Add bhaible's missing Makefile commit bede01d82ec2f1cfe2bb42233fcd408a6e3a6323 Author: Anthony Green Date: Tue Mar 13 07:53:33 2018 -0400 Remove stray directory commit 9aefbb1031c9be7c5121ab24742e025ccff236c4 Author: Anthony Green Date: Tue Mar 13 07:51:02 2018 -0400 Update version number to next pre-release commit ddf7a8f7511a038342f56c08dac718148e2f0151 Author: Anthony Green Date: Tue Mar 13 07:47:57 2018 -0400 Update test list for dist commit 6186261cb36a33abbf31430569f83ccd7685474e Author: Anthony Green Date: Mon Mar 12 21:53:18 2018 -0400 Add Bruno Haible and Bill Triggs' libffi testsuite commit 4e8f79d5587ddb8e027f4df213e90f943e286594 Author: Anthony Green Date: Sun Mar 11 18:21:46 2018 -0400 Next release will be 3.3 commit 3840d49aaa831d649b1597518a2903dfed0d57f3 Author: Stef O'Rear Date: Sun Mar 11 05:55:15 2018 -0700 New RISC-V port (#281) * Add RISC-V support This patch adds support for the RISC-V architecture (https://riscv.org). This patch has been tested using QEMU user-mode emulation and GCC 7.2.0 in the following configurations: * -march=rv32imac -mabi=ilp32 * -march=rv32g -mabi=ilp32d * -march=rv64imac -mabi=lp64 * -march=rv64g -mabi=lp64d The ABI currently can be found at https://github.com/riscv/riscv-elf-psabi-doc/blob/master/riscv-elf.md . * Add RISC-V to README * RISC-V: fix configure.host commit dca52b55bc2ac0213c20849d7e9e19fbc9202023 Merge: 83d9aba3a 11de69ddb Author: Anthony Green Date: Sun Mar 11 08:50:01 2018 -0400 Merge pull request #406 from trofi/master ia64: fix variadic function closures with FP arguments commit 83d9aba3a44dff8426052312a9445a7ef52f1db1 Merge: ab1540569 b58caef7f Author: Anthony Green Date: Sun Mar 11 08:48:42 2018 -0400 Merge pull request #407 from trofi/ia64-small-struct ia64: fix small struct return commit ab154056937782c1c7f7005c8b46c5179bc719cb Merge: 9429968b8 e66fd6782 Author: Anthony Green Date: Sun Mar 11 08:46:58 2018 -0400 Merge pull request #409 from andreas-schwab/master Revert "Fix passing struct by value on aarch64" commit e66fd678200db456a2e9860e80451773efa96fe0 Author: Andreas Schwab Date: Tue Feb 20 10:47:09 2018 +0100 Revert "Fix passing struct by value on aarch64" This reverts commit 482b37f00467325e3389bab322525099860dd9aa. That was actually a bug in python, see . commit b58caef7fd620408be9239ac24ea89d5bc84f30b Author: Sergei Trofimovich Date: Sat Feb 17 19:00:40 2018 +0000 ia64: fix small struct return This change fixes libffi.call/struct10.c failure on ia64: FAIL: libffi.call/struct10.c -W -Wall -Wno-psabi -O0 execution test .Lst_small_struct handles returns for structs less than 32 bytes (following ia64 return value ABI [1]). Subroutine does roughly the following: ``` mov [sp+0] = r8 mov [sp+8] = r9 mov [sp+16] = r10 mov [sp+24] = r11 memcpy(destination, source=sp, 12); ``` The problem: ia64 ABI guarantees that top 16 bytes of stack are scratch space for callee function. Thus it can clobber it. [1] says (7.1 Procedure Frames): """ * Scratch area. This 16-byte region is provided as scratch storage for procedures that are called by the current procedure. Leaf procedures do not need to allocate this region. A procedure may use the 16 bytes at the top of its own frame as scratch memory, but the contents of this area are not preserved by a procedure call. """ In our case 16 top bytes are clobbered by a PLT resolver when memcpy() is called for the first time. As a result memcpy implementation reads already clobbered data frop top of stack. The fix is simple: allocate 16 bytes of scrats space prior to memcpy() call. [1]: https://www.intel.com/content/dam/www/public/us/en/documents/guides/itanium-software-runtime-architecture-guide.pdf Bug: https://bugs.gentoo.org/634190 Signed-off-by: Sergei Trofimovich commit 45da2fcbcd0ecaba673275d22b04fac3e4376e22 Author: Sergei Trofimovich Date: Sat Feb 17 18:53:02 2018 +0000 new test: return small struct The bug originally was discovered in https://bugs.gentoo.org/634190 where complicated callback was returning invalid data on ia64. This change adds minimal reproducer that fails only on ia64 as: FAIL: libffi.call/struct10.c -W -Wall -Wno-psabi -O0 execution test FAIL: libffi.call/struct10.c -W -Wall -Wno-psabi -O2 execution test FAIL: libffi.call/struct10.c -W -Wall -Wno-psabi -O3 execution test FAIL: libffi.call/struct10.c -W -Wall -Wno-psabi -Os execution test Test passes on amd64. The fix is in the following commit. Bug: https://bugs.gentoo.org/634190 Signed-off-by: Sergei Trofimovich commit 11de69ddb788e4d87ef653898878384116ac16c6 Author: Sergei Trofimovich Date: Sun Feb 11 11:29:39 2018 +0000 ia64: fix variadic function closures with FP arguments libffi test framework already flagged failures as: ``` FAIL: libffi.call/cls_double_va.c -W -Wall -Wno-psabi -O0 output pattern test, is 7.0 res: 4 0.0 res: 4 ? should match 7.0 ?es: 4 ?.0 res: 4 ``` Failure happens here at ```c // testsuite/libffi.call/cls_double_va.c ... char* format = "%.1f\n"; double doubleArg = 7; ... CHECK(ffi_prep_closure_loc(pcl, &cif, cls_double_va_fn, NULL, code) == FFI_OK); res = ((int(*)(char*, ...))(code))(format, doubleArg); ``` libffi expects 'doubleArg' to be located in 'f9' (second FP argument) but gcc placed it to 'r33' (second GR). ia64 software [1] manual described argument passing ABI in "8.5.2 Register Parameters" as: """ If an actual parameter is known to correspond to a floating-point formal parameter, the following rules apply: a) The actual parameter is passed in the next available floating-point parameter register, if one is available. Floating-point parameter registers are allocated as needed from the range f8-f15, starting with f8. b) If all available floating-point parameter registers have been used, the actual parameter is passed in the appropriate general register(s). (This case can occur only as a result of homogeneous floating-point aggregates, described below.) If a floating-point actual parameter is known to correspond to a variable-argument specification in the formal parameter list, the following rule applies: c) The actual parameter is passed in the appropriate general register(s). If the compiler cannot determine, at the point of call, whether the corresponding formal parameter is a varargs parameter, it must generate code that satisfies both of the above conditions. (The compiler’s determination may be based on prototype declarations, language standard assumptions, analysis, or other user options or information.) """ We have [c] case here and gcc uses only GR for parameter passing. The change binds known variadic arguments ro GRs instead of FPs as those are always expected to be initialized for all variadic call types. This fixes all 10 failures on ia64-unknown-linux-gnu: ``` === libffi Summary === -# of expected passes 1945 -# of unexpected failures 10 + +# of expected passes 1955 ``` [1]: https://www.intel.com/content/dam/www/public/us/en/documents/guides/itanium-software-runtime-architecture-guide.pdf Signed-off-by: Sergei Trofimovich commit 9429968b828de8775cfd7f139374888943f2186a Merge: f08b2be62 28d3b61bd Author: Anthony Green Date: Sat Feb 10 23:23:33 2018 -0500 Merge pull request #403 from frida/fix/x86-sysv-pic-closure-regression Fix x86 SysV closure in PIC mode commit f08b2be6273f471863f1ea1afb0078a96fe73d74 Merge: c194448ef 482b37f00 Author: Anthony Green Date: Sat Feb 10 23:21:49 2018 -0500 Merge pull request #405 from andreas-schwab/master Fix passing struct by value on aarch64 commit 482b37f00467325e3389bab322525099860dd9aa Author: Andreas Schwab Date: Mon Sep 18 12:44:08 2017 +0200 Fix passing struct by value on aarch64 This fixes the ctypes test in the python testsuite. commit 28d3b61bdd8b9fe786cb431165f2fb95f4e275d3 Author: Ole André Vadla Ravnås Date: Sat Jan 20 23:56:17 2018 +0100 Fix x86 SysV closure in PIC mode The assembly single-line comments swallowed up the remaining assembly code of the macros due to lack of line-endings. This is a regression introduced in b7f6d7a. commit c194448ef08b4c0183c68dda95a4b8b9cfc47985 Author: Anthony Green Date: Wed Jan 10 09:38:15 2018 -0500 adjust env usage in travis commit 5314b1067145d30acdc5bba3bebb788e8acfdaae Author: Anthony Green Date: Wed Jan 10 07:22:51 2018 -0500 test static lib commit bec6135dff061a35065efc106f920940a82c28ee Merge: f6ecf10ca d15581c69 Author: Anthony Green Date: Wed Jan 10 07:20:04 2018 -0500 Merge pull request #393 from thejunkjon/master Linker error "recompile with -fPIC" for x86_64 commit f6ecf10ca021e94d496c75cc43c417161edf28b3 Merge: c50ba182f 746dbe3a6 Author: Anthony Green Date: Fri Jan 5 16:51:44 2018 -0500 Merge pull request #401 from wzssyqa/jr-r6 mips/ffi.c: fix encoding for jr on r6 commit 746dbe3a6a79a41931c03b51df2972be4d5e5028 Author: YunQiang Su Date: Wed Jan 3 10:07:41 2018 +0800 mips/ffi.c: fix encoding for jr on r6 mips/ffi.c: instruction jr has a different encoding for r6 commit c50ba182f43537e29dd92cfd9b2fa7a30549a06e Merge: 54cb3e234 af6949b7a Author: Anthony Green Date: Mon Jan 1 15:18:46 2018 -0500 Merge pull request #398 from emaste/master Enable symbol versioning when ld is LLVM's lld commit 54cb3e2345ee8e2483d4705a9fb2967b6e5a9bd1 Merge: 716bfd831 94c102aa6 Author: Anthony Green Date: Sun Dec 31 07:46:05 2017 -0500 Merge pull request #396 from wzssyqa/master mips/n32.S: disable .set mips4 on mips r6 commit af6949b7af98404729e04227165d0e32ce550ce8 Author: Ed Maste Date: Sun Dec 17 23:08:12 2017 -0500 Enable symbol versioning when ld is LLVM's lld Fixes #397 commit 94c102aa69b04337f63498e0e6551fcdce549ae5 Author: YunQiang Su Date: Sun Dec 10 14:25:01 2017 +0800 Not set mips on mips r6 MIPS release changed encodes of some instructions, include ll/sc etc. if .set mips4 on mips r6, as will generate some wrong encode of some instructions. commit d15581c6969657c79afcff4865bdc8f5ff23a2f7 Author: jon Date: Fri Dec 1 00:34:30 2017 -0800 Updating calls to ffi_closure_unix64_inner and ffi_closure_win64_inner to use PLT. Without this fix, statically linking libffi causes the linker error i.e. 'requires dynamic R_X86_64_PC32 reloc against ffi_closure_unix64_inner which may overflow at runtime; recompile with -fPIC)' commit 716bfd83177689e2244c4707bd513003cff92c68 Merge: 4fdbb0578 cd5e9f8b8 Author: Anthony Green Date: Sun Nov 5 09:02:37 2017 -0500 Merge pull request #387 from trofi/master Makefile.am: add 'src/s390/internal.h' to source tarball commit cd5e9f8b87cc387d3ffef4db81ee6af200a84183 Author: Sergei Trofimovich Date: Sun Nov 5 13:56:42 2017 +0000 Makefile.am: add 'src/s390/internal.h' to source tarball commit 2f530de168e0253ac06e044c832132c496e8788b ("s390: Reorganize assembly") introduced new header (similar to other arches) but did not add it to source tarball. As a result build from 'make dist' tarballs failed as: ``` ../src/s390/ffi.c:34:10: fatal error: internal.h: No such file or directory #include "internal.h" ^~~~~~~~~~~~ ``` To fix it the change adds file to 'Makefile.am'. Signed-off-by: Sergei Trofimovich commit 4fdbb0578e921a9da146c2b040061a3a39fe4fda Merge: b302bc3df 2bfcd2995 Author: Anthony Green Date: Fri Nov 3 07:05:31 2017 -0400 Merge pull request #320 from 0-wiz-0/master Support NetBSD with mprotect. commit b302bc3dfe47e3ea57de11610fced3170c06df28 Merge: 14eac9386 1fb788ac8 Author: Anthony Green Date: Fri Nov 3 07:03:55 2017 -0400 Merge pull request #322 from compnerd/aarch64-base aarch64: fix index base register for AArch64 commit 14eac93869eb109bea1daeb4f00609169d5488ff Merge: ad15ae762 9fc9dc535 Author: Anthony Green Date: Fri Nov 3 06:49:03 2017 -0400 Merge pull request #384 from yan12125/fix-sgidefs-checking Fix linux detection (closes #303) commit 9fc9dc535ee7af28f49f86a8ecacb7f575c46ba4 Author: Yen Chi Hsuan Date: Fri Oct 27 16:12:56 2017 +0800 Fix linux detection (closes #303) commit 1fb788ac898290d8e3044ca4c0a4fb3c3c254ce5 Author: Saleem Abdulrasool Date: Tue Oct 10 11:37:00 2017 -0700 aarch64: fix index base register for AArch64 The base is passed in `x3`, not in `x2`. This fixes the indexing base so that the right value is used. commit ad15ae762c8b3999e626ebccb9cf454455119555 Merge: dc2ff5baa 5e4fcdccb Author: Anthony Green Date: Wed Oct 25 13:16:49 2017 -0400 Merge pull request #298 from frida/fix/ios-trampoline-table-complexity Simplify iOS trampoline table allocation commit dc2ff5baabf89d473d61ae4b468638f8ea98bb04 Merge: 927da7166 79d1509cb Author: Anthony Green Date: Wed Oct 25 13:11:40 2017 -0400 Merge pull request #323 from compnerd/x86-alloca-alignment x86: align alloca to 16-byte boundary commit 927da71662f349f656fc1c5d0aaa533005e24f47 Merge: a0455c031 181fc4cc9 Author: Anthony Green Date: Wed Oct 25 13:05:53 2017 -0400 Merge pull request #379 from jlj/master Xcode build improvements commit a0455c031824c6ab3f1a7a6ef9912edad6ebd742 Merge: 0faf61452 9d9d92b47 Author: Anthony Green Date: Wed Oct 25 13:04:23 2017 -0400 Merge pull request #383 from hjl-tools/hjl/master Hjl/master commit 0faf614529ae7ac8bc929415a6e1b076ba1b8cc5 Merge: 8d26e8c6d 3c372c384 Author: Anthony Green Date: Wed Oct 25 13:03:55 2017 -0400 Merge pull request #381 from compnerd/ffi-nullptr-deref arm: fix a level of indirection issue commit 9d9d92b47f3f73644f808ae38e97d1ea8f4d22df Author: H.J. Lu Date: Wed Oct 25 04:59:31 2017 -0700 Skip WIN64/EFI64 support for x32 Since x32 doesn't support WIN64/EFI64, skip it if __ILP32__ is defined. commit b2a343ffc68f1af5368281e063c0ca5571c611cc Author: H.J. Lu Date: Wed Oct 25 04:36:49 2017 -0700 Don't include WIN64/EFI64 for x32 Since x32 doesn't support WIN64/EFI64, don't include it for x32. Also x32 has sizeof size_t == 4. But TARGET should be X86_64. commit 3c372c384a94db23fdaf9fe64a4beb86159cf6d3 Author: Saleem Abdulrasool Date: Tue Oct 24 13:53:56 2017 -0700 arm: fix a level of indirection issue Rather than relying on the stack being 0'ed out always, do it manually. The stack generally happened to be zero, and because the compiler realizes that the tests are dealing with chars truncates the read value. However, the top 3 nibbles of the value are undefined and may be non-zero. The indirection level caused a null-pointer dereference. Explicitly scribbling on the stack during the allocation causes test failures without the original zexting behaviour. commit 181fc4cc9ca211f3bc8209cecf3b2348b0674e21 Merge: 8d26e8c6d 54529ba1d Author: Jean-Luc Jumpertz Date: Mon Oct 23 15:02:29 2017 +0200 Merge branch 'master' based on ksjogo/libffi Added a tvOS target in Xcode project. Misc Xcode project cleanup. Fix macOS build target in Xcode project. # Conflicts: # src/aarch64/ffi.c # src/x86/ffi64.c commit 54529ba1db9974b5efcbd2e893c44b6b01b0b2c8 Author: Jean-Luc Jumpertz Date: Mon Oct 23 09:49:35 2017 +0200 Added a tvOS target in Xcode project. Misc Xcode project cleanup. commit 79d1509cb06ba9067f56e2c62394d7bc60fa6bf2 Author: Saleem Abdulrasool Date: Tue Oct 10 11:39:45 2017 -0700 x86: align alloca to 16-byte boundary Align the stack allocation to a 16-byte boundary. This ensures that the stack parameters are 16-byte aligned which is needed for some instructions. commit 8d26e8c6da23b10331181a4bbf837f479ce5d7d2 Merge: 1c9171ec8 ed7488c00 Author: Tom Tromey Date: Tue Oct 17 10:46:49 2017 -0600 Merge pull request #326 from trofi/master src/ia64/unix.S: unbreak small struct handling commit ed7488c003765c7ee71b7da3e8c21f431b043809 Author: Sergei Trofimovich Date: Tue Oct 17 13:00:51 2017 +0100 src/ia64/unix.S: unbreak small struct handling commit 6e8a4460833594d5af1b4539178025da0077df19 added FFI_TYPE_COMPLEX value type (comes after FFI_TYPE_POINTER) ia64 ffi_closure_unix reiles on the ordering of FFI_ enums as ia64 has ia64-specific FFI types: small struct and FPU extesions. As a result all tests handling small structs broke. The change fixes dispatch table by adding (no-op) FFI_TYPE_COMPLEX entry This has positive effect of unbreaking most tests on ia64: === libffi Summary === -# of expected passes 1595 -# of unexpected failures 295 +# of expected passes 1930 +# of unexpected failures 10 # of unsupported tests 30 Signed-off-by: Sergei Trofimovich commit 1c9171ec89943d9b4b10a54091c8b5c104eb258d Merge: 9f1f8741b 7ad0ae7f4 Author: Anthony Green Date: Tue Oct 10 23:19:57 2017 -0400 Merge pull request #324 from compnerd/arm-zext arm: zext return value parameters commit 7ad0ae7f42f3e208431ab66a9032dc9549f978d0 Author: Saleem Abdulrasool Date: Tue Oct 10 11:44:05 2017 -0700 arm: zext return value parameters The closure function (invoked as closure->fun in ffi_closure_XXX_inner) will only populate the actual number of bytes for the true return type, which may be a character. This leaves garbage on the stack when the assembly closure function (i.e. ffi_closure_XXX) reads the return value off of the stack into r0 as a 4-byte value. ffi_closure_XXX always leaves room for at least 4 bytes here, so we can safely set them to 0. Otherwise, if there is garbage in any of these bytes, these end up in r0 and in the returned value as well. commit 2bfcd29955c02b67fa10a68cc4200f6838181e0f Author: Joerg Sonnenberger Date: Mon Oct 2 15:34:03 2017 +0200 Support NetBSD with mprotect. Signed-off-by: Thomas Klausner commit 9f1f8741b2ec0c8898ecff7b93005627b915dd02 Merge: dca078f46 6cf0dea78 Author: Tom Tromey Date: Sun Oct 1 20:25:50 2017 -0600 Merge pull request #319 from angerman/patch-5 Adds `local.exp` to DISTCLEANFILES commit 6cf0dea78a5a4584eda871e264043974d0453e53 Author: Moritz Angermann Date: Mon Oct 2 10:20:16 2017 +0800 Change CLEANFILES to DISTCLEANFILES commit dca078f468ff696134195e85fed3f48260115651 Merge: 540258d13 bccba9a34 Author: Tom Tromey Date: Sun Oct 1 20:19:53 2017 -0600 Merge pull request #316 from angerman/patch-2 Add `configure.host` and `libtool-version` to the `EXTRA_DIST` files commit 540258d13e1d427254cbeddacdd2bb1ab2c307d1 Merge: 93d8e7dd1 4c6aa78d7 Author: Tom Tromey Date: Sun Oct 1 20:17:02 2017 -0600 Merge pull request #317 from angerman/patch-3 Add src/x86/asmnames.h to noinst_HEADERS commit bccba9a34e8ba0fe671c4b94671fea2efe181b01 Merge: a1e6ccb67 59d65d74a Author: Moritz Angermann Date: Sun Oct 1 12:48:17 2017 +0800 Merged commit 6b6df1a7bb377651a07663879bd4162f0059dc5c Author: Moritz Angermann Date: Sun Oct 1 12:37:53 2017 +0800 Adds `local.exp` to CLEANFILES With #315, #316, #317, #318 and this patch, running ``` AM_DISTCHECK_CONFIGURE_FLAGS=--disable-docs make distcheck ``` should complete successfully. commit 59d65d74ae72bd24fbf78f045592c5168f0b0784 Author: Moritz Angermann Date: Sun Oct 1 12:23:31 2017 +0800 Include `libtool-version` in the EXTRA_DIST files. commit 4c6aa78d75f0c426a23c27af74155aa2535021d6 Author: Moritz Angermann Date: Sun Oct 1 11:32:41 2017 +0800 Add src/x86/asmnames.h to noinst_HEADERS In eaa59755fcbb692a8cb763c7f9f24a350aadbd30, macros from `unix64.S` were extracted into `asmnames.h` to be used with `win64.S` as well. As such these are required by `unix64.S`, which fails to build without them. commit a1e6ccb67e7d7f2b235016aa6c6733c5743945a1 Author: Moritz Angermann Date: Sun Oct 1 11:28:40 2017 +0800 Add `configure.host` to the `EXTRA_DIST` files When running `make dist`, `configure.host` would not result in the distribution tarball, however `configure` would try to read it, and as such the tarball would not be buildable. commit 93d8e7dd17b08ff195af3580584ccd5c2228202f Author: Anthony Green Date: Wed Sep 27 21:51:34 2017 -0400 Fix #265 commit 02a5145abbae2e311d5aeeee606c5b5745a99b55 Merge: 10099d6ca 7d504f7a8 Author: Anthony Green Date: Wed Sep 27 21:43:03 2017 -0400 Merge pull request #263 from ksjogo/master fix ios builds commit 10099d6cad8418a10c1508235d1f458993ac51fe Merge: c8f1aa3d0 ed8488345 Author: Anthony Green Date: Wed Sep 27 20:54:09 2017 -0400 Merge pull request #271 from frida/fix/qnx-cache-flushing arm: Fix cache flushing on QNX commit c8f1aa3d0459c6f1be6884b324b4d2d785ecfd78 Merge: db4dad97e 50e408ce5 Author: Anthony Green Date: Wed Sep 27 20:53:50 2017 -0400 Merge pull request #307 from zw3rk/master Support -ios triple commit db4dad97e2589ca700d1b5b54eeb332137a152ad Merge: 8a16f2f47 9c12209d2 Author: Anthony Green Date: Wed Sep 27 20:47:08 2017 -0400 Merge pull request #312 from fjricci/fix_ub Fix misaligned memory access in ffi_call_int commit 8a16f2f47fdb6a985d201baa02a1d226562af4d8 Merge: b23091069 0ff9419f2 Author: Anthony Green Date: Wed Sep 27 20:45:15 2017 -0400 Merge pull request #308 from trex58/master This patch enables FFI Go Closure on AIX. commit a78da73956e4b0af4556f1c0a444bd0f065c3965 Author: Jean-Luc Jumpertz Date: Mon Sep 4 15:55:34 2017 +0200 Fix macOS build target in Xcode project. - Add missing files for desktop platforms in generate-darwin-source-and-headers.py, and in the Xcode project. - Add a static library target for macOS. - Fix "implicit conversion loses integer precision" warnings for iOS mad macOS targets. commit 9c12209d2eac40238eefb4255994277918e7eff1 Author: Francis Ricci Date: Thu Aug 3 10:46:28 2017 -0700 Fix misaligned memory access in ffi_call_int commit b23091069adce469dc38fbcc9fd8ac9085d3c9d7 Merge: c0cc9f1df 9c6cb5881 Author: Tom Tromey Date: Wed Jul 12 12:43:06 2017 -0600 Merge pull request #309 from rurban/oldgcc __attribute__ deprecated (msg) only since gcc 4.5 commit 9c6cb58812aa92a4c56a94b2b51bd85b2aebc2a1 Author: Reini Urban Date: Sun Jun 18 18:26:05 2017 +0200 __attribute__ deprecated (msg) only since gcc 4.5 make it work with older compilers commit 50e408ce5d9eb6f14bf82bb3894d0e74e5f19c2c Author: Moritz Angermann Date: Sun May 21 20:44:10 2017 +0800 add i?86-*-ios -- the iOS simulator (32bit) commit a08cabe6746343564418383b879d40d3b998c399 Author: Moritz Angermann Date: Sun May 21 20:25:16 2017 +0800 add x86_64-*-ios -- the iOS simulator commit 0ff9419f2e75652426469e256cb7d0748064ad58 Author: Tony Reix Date: Wed May 17 14:57:53 2017 +0200 This patch enables FFI Go Closure on AIX. commit 19ab448d84223d2992048bce4e926eac2c44f606 Author: Moritz Angermann Date: Tue May 16 08:32:06 2017 +0800 Support -ios triple Autoconf hasn’t had an update since 2014, and it doesn’t look like it will soon[1] This updates config.{guess,sub} It adds support for e.g. `-ios`, which allows to have targets like `aarch64-apple-ios`. It basically does exactly what the config.guess script says: > It is advised that you download the most up to date version of the config scripts from The configure.ac script has been updated to relax `*-apple-darwin*` to `*-apple-*`. Similarly the `.gitignore` and `autogen.sh` needed to be updated to respect the newer `config.{sub,guess}` — [1]: http://lists.gnu.org/archive/html/autoconf/2016-07/msg00017.html commit c0cc9f1df9fd4c5e758470f05d0e48123f0638ae Merge: b841ae70a bd72848c7 Author: Tom Tromey Date: Mon May 8 15:20:39 2017 -0600 Merge pull request #302 from gpakosz/align-macros Prefix ALIGN macros with FFI_ commit bd72848c7af9302df50a7a11652c77166d17caa8 Author: Gregory Pakosz Date: Thu Apr 27 13:20:36 2017 +0200 Prefix ALIGN macros with FFI_ commit 7d504f7a8e33f6da27941f5dc5d889fe60b8b9c8 Author: ksjogo Date: Thu Mar 30 13:32:58 2017 +0200 update for xcode 8.3 commit 57d8ff044cd6320d8ebacacf06455569b4aac27d Author: Ole André Vadla Ravnås Date: Wed Mar 15 01:43:11 2017 +0100 Simplify iOS trampoline table allocation By using VM_FLAGS_OVERWRITE there is no need for speculatively allocating on a page we just deallocated. This approach eliminates the race-condition and gets rid of the retry logic. commit 7f558a9ba4e74edf54a3e95e983e8a59ddb0cc7c Author: ksjogo Date: Mon Jul 25 01:18:30 2016 +0200 set ios deployment target to 7.0 commit e76fa94b6ef9a699ee9edeb85d1731d3bb7abd46 Author: ksjogo Date: Thu Jul 14 11:28:18 2016 +0200 fix ltmain.sh in xcode builds commit 22b2fd6f18eed7b32ea959f7f3e56ea09b9ac9bf Author: ksjogo Date: Wed Jul 13 14:40:07 2016 +0200 add valid architectures commit 0040694505411785ec98e660fe81faad9630f515 Author: Johannes Goslar Date: Tue Jul 12 16:08:42 2016 +0200 Update Xcodeproj Include all currently relevent files. Call autogen is build script. Fix compiler settings. Fix mach include. commit b841ae70a05a5e11de1fca1b4551189db0895cf2 Author: Anthony Green Date: Sun Mar 19 07:42:09 2017 -0400 remove stray quote commit a94c999ba0d51ed8ec88685233f5b1ae38e894ea Author: Anthony Green Date: Sun Mar 19 07:36:07 2017 -0400 Handle fastcall declaration differently for some Microsoft compilers commit 69963d39ab7d7f201b4597ed7a24cf438e0a34bf Author: Anthony Green Date: Sun Mar 19 07:33:39 2017 -0400 We don't support 32-bit builds with the Microsoft toolchain commit 073bc9d50c1ce753735944e4228d9a401e964306 Author: Anthony Green Date: Sun Mar 19 07:25:24 2017 -0400 Enable appveyor ci support commit a0b14eea2baf9f18c0d29bc5ce4495422381f917 Merge: fa5a0fe2c 1e0d107b4 Author: Anthony Green Date: Fri Mar 17 09:20:40 2017 -0400 Merge pull request #291 from ramon-garcia/visual-studio-build Build with Visual C++ (64 bits) commit fa5a0fe2c6da906cc7c114c8002e5c7092c4dbee Merge: 5b91bbd38 bfab429ee Author: Anthony Green Date: Thu Mar 16 07:50:07 2017 -0400 Merge pull request #238 from KubaKaszycki/master Update README with a new port commit 5b91bbd388213ec19044a392c291505e2c0674b9 Merge: d754c510f 982b89c01 Author: Anthony Green Date: Thu Mar 16 00:40:58 2017 -0400 Merge pull request #288 from yan12125/std-includedir Install public headers in the standard path commit d754c510f1dd03dc8c2cd7b50bb2a1668715c5cd Merge: ebe7f447a 41185b565 Author: Anthony Green Date: Thu Mar 16 00:39:07 2017 -0400 Merge pull request #299 from matijaskala/patch-1 detect other x32 hosts commit ebe7f447abcfbf1a16dbfca26538fff03faf7c6b Author: Anthony Green Date: Thu Mar 16 00:36:10 2017 -0400 Remove osx testing. gcc was symlinked to clang. dejagnu c++ tests fail with clang. commit c9c2aa24611ec996f45a495f24d690688f3ffed6 Author: Anthony Green Date: Wed Mar 15 09:58:39 2017 -0400 Revert previous two changes. clang/MacOS problem can only be solved within dejagnu. commit 5a8fca94059fbf14ac7352ce487c9250cdd67c63 Author: Anthony Green Date: Wed Mar 15 09:52:39 2017 -0400 Fix typo commit 5b10a0198188d43fc07c271bc29ff1033e2ac9d3 Author: Anthony Green Date: Wed Mar 15 09:34:01 2017 -0400 Work around dejagnu/clang problems commit b6751258072fadadd487a6adf4962bf22b44746d Author: Anthony Green Date: Wed Mar 15 08:54:52 2017 -0400 Require sudo for installing packages commit 41185b565fe17a8f107642f480849ef369ad0baa Author: Matija Skala Date: Wed Mar 15 12:50:41 2017 +0100 detect other x32 hosts primarily x86_64-pc-linux-muslx32 while at it, add x86_64-x32-pc-linux-{gnu,musl} as well commit 6e2e041b6df6a3c3a5ca8a750dedbbd465e5ca97 Merge: dabbd1456 d42ce7b95 Author: Richard Henderson Date: Tue Mar 14 19:37:45 2017 -0700 Merge pull request #297 from frida/fix/error-path Fix error path so mutex is unlocked before returning commit 5e4fcdccb3a841f2daaa5834f726eaeb42950511 Author: Ole André Vadla Ravnås Date: Wed Mar 15 01:43:11 2017 +0100 Simplify iOS trampoline table allocation By using VM_FLAGS_OVERWRITE there is no need for speculatively allocating on a page we just deallocated. This approach eliminates the race-condition and gets rid of the retry logic. commit d42ce7b95c6d5cba7d976cc0c8c85f7645fa6b8f Author: Ole André Vadla Ravnås Date: Wed Mar 15 01:23:40 2017 +0100 Fix error path so mutex is unlocked before returning In the unusual case where ffi_trampoline_table_alloc() fails. commit dabbd145665ffca6fa676eba551f12107cd81131 Merge: 60e4250a7 6caabd198 Author: Tom Tromey Date: Mon Feb 27 21:10:00 2017 -0700 Merge pull request #295 from terrillmoore/patch-1 Correct typos in libffi.texi commit 6caabd198ec24d090d85d7eca01150b44003c76b Author: Terry Moore Date: Sun Feb 26 15:06:53 2017 -0500 Correct typos in libffi.texi One "it's" should be "its", and one "is" should be "in". commit 1e0d107b4b237edd11ca3209a9cf59786f752fe5 Author: Ramón García Fernández Date: Sun Jan 8 20:12:59 2017 +0100 Modify configure.host to detect compilation with Microsoft Visual C++ and use assembly with Intel syntax in that case commit 982b89c01aca99c7bc229914fc1521f96930919b Author: Yen Chi Hsuan Date: Sun Nov 13 19:17:19 2016 +0800 Install public headers in the standard path commit 60e4250a77eb3fde500bfd68ec40519fe34b21bd Author: Anthony Green Date: Sun Sep 4 09:34:18 2016 -0400 make executable commit c3caf38895252b82febb76c6bb323c67a4e01929 Author: Anthony Green Date: Sun Sep 4 09:21:18 2016 -0400 Update travis to test linux 32/64 and osx with a mix of gcc and clang commit c5b408ee78d097a495ea12467cf082f8008c8e17 Author: Anthony Green Date: Sun Sep 4 09:17:46 2016 -0400 xfail the unwindtest for osx as per issue #279 commit 256ce51c408766d23b2707cc79b97f673750b3b9 Merge: 2ded2a4f4 52a11f6d9 Author: Richard Henderson Date: Thu Sep 1 13:54:51 2016 -0700 Merge pull request #273 from wbx-github/master m68k: support ISA-A Coldfire CPUs commit 2ded2a4f494165c93293afc14ab0be1243cf8c49 Merge: 408c9c7d5 e247b562a Author: Richard Henderson Date: Thu Sep 1 13:30:45 2016 -0700 Merge pull request #272 from yousong/mips64-soft-float Mips64 soft float commit 408c9c7d5a4cb323c77499b1643858f4bfaf4ca7 Merge: 06d7c5190 b545ff81f Author: Tom Tromey Date: Tue Aug 23 11:32:17 2016 -0600 Merge pull request #274 from dr-hannibal/ykolerov-unused ARC: Remove unused variable commit b545ff81f58233284cd6fca97a470aa3218c23e6 Author: Yuriy Kolerov Date: Tue Aug 23 20:23:37 2016 +0300 ARC: Remove unused variable Signed-off-by: Yuriy Kolerov commit 52a11f6d9152a56aa1fb6d04eb40fd76f4cce5ff Author: Waldemar Brodkorb Date: Sat Aug 20 00:52:19 2016 +0200 m68k: support ISA-A Coldfire CPUs Fix compilation for m68k/coldfire CPUs like mcf5208. Signed-off-by: Thorsten Glaser Signed-off-by: Waldemar Brodkorb commit e247b562ac85565dfac33e676e2915489af04183 Author: Yousong Zhou Date: Mon Aug 15 17:34:37 2016 +0800 doc: fix typo Name of last argument to ffi_get_struct_offsets is `offsets`, not `sizes` Signed-off-by: Yousong Zhou commit 7a0d2c83bf2b0422e2d0ed297f847fc1eeb88bb3 Author: Yousong Zhou Date: Mon Aug 15 15:00:13 2016 +0800 mips: fix MIPS softfloat build issue The patch for o32.S is taken from OpenWrt packages repo 3a7a4bf "libffi: fix MIPS softfloat build issue with current binutils" Signed-off-by: Felix Fietkau Signed-off-by: Yousong Zhou commit 06d7c5190e910356d4212d2a79056df8b70fd27e Merge: aa7ed78c0 301166b11 Author: Tom Tromey Date: Wed Aug 10 15:06:16 2016 -0600 Merge pull request #269 from frida/fix/aarch64-variadic-closures-on-ios aarch64: Fix handling of variadic closures on iOS commit aa7ed78c0e0d3fc4daa6a86c35ecbb7c43b5d067 Merge: 1daa78799 4da814b18 Author: Tom Tromey Date: Wed Aug 10 15:03:37 2016 -0600 Merge pull request #268 from frida/fix/aarch64-large-aggregates aarch64: Fix handling of aggregates larger than 16 bytes commit 4da814b18a1bf1d2013d988e37485e568d6117d1 Author: Ole André Vadla Ravnås Date: Wed Aug 10 22:48:09 2016 +0200 aarch64: Fix handling of aggregates larger than 16 bytes Instead of allocating stack space for a pointer we would allocate stack space for the actual aggregate size. commit 1daa78799942c83c905f8d7c5439952e9ffec3bc Merge: f03ba2032 5e9ac7e25 Author: Tom Tromey Date: Wed Aug 10 14:31:09 2016 -0600 Merge pull request #270 from frida/fix/aarch64-warnings Fix aarch64 warnings commit ed8488345eaae41c161ab741b0c44299ee99dc6c Author: s1341 Date: Wed Aug 10 14:57:22 2016 +0200 arm: Fix cache flushing on QNX Use `msync()` directly as `__clear_cache()` is broken in the qnx650_gcc4.8.3 toolchain. commit 5e9ac7e252dbb0b7025027c1b6e1a4f3ec48943f Author: Ole André Vadla Ravnås Date: Wed Aug 10 15:22:19 2016 +0200 aarch64: Fix warning about unused function on iOS commit 4d1f11f6a9effce05ba51578b142827834d1f699 Author: Ole André Vadla Ravnås Date: Wed Aug 10 15:21:42 2016 +0200 aarch64: Fix operand size warning reported by Clang commit 301166b11e913f85d09f67af94264268811f23a9 Author: Ole André Vadla Ravnås Date: Wed Aug 10 15:59:56 2016 +0200 aarch64: Fix handling of variadic closures on iOS commit f03ba2032850545dc6869bb495170a8c6cbf849d Merge: 40e4063ab ef8be84de Author: Tom Tromey Date: Tue Aug 2 10:23:25 2016 -0600 Merge pull request #266 from dr-hannibal/arc-pthread Make testing easier outside of build directory and especially for ARC commit ef8be84de5f9e611add396efe908b772eb1cae1d Author: Yuriy Kolerov Date: Fri Jul 29 21:01:38 2016 +0300 Do not use fabsl() in float2.c test Some targets may support long double variables but in the same time may lack support of long double functions like fabsl(). Signed-off-by: Yuriy Kolerov commit f3201733d54bf26ac40c432f4305499dd3f20215 Author: Yuriy Kolerov Date: Fri Jul 29 19:18:41 2016 +0300 Fix output expectations in cls_dbls_struct.c test This test with invalid output expectations may fail on some targets (e.g. ARC processors). Signed-off-by: Yuriy Kolerov commit f74ea2dc7994b4867f7ab45169e1b2422072cc92 Author: Yuriy Kolerov Date: Thu Jul 28 20:57:09 2016 +0300 Allow setting an arbitary value for blddirffi in testsuite It is useful when tests are executed not from build directory. So the path of the build directory may be passed through site.exp or runtest. Signed-off-by: Yuriy Kolerov commit 31362d785ed1db52516dfe02ebe8b163c53d29d0 Author: Yuriy Kolerov Date: Thu Jul 28 18:48:23 2016 +0300 ARC: Link tests with pthread library Signed-off-by: Yuriy Kolerov commit 40e4063ab4449c644bc106298805fa8c6cfadbd4 Merge: c0829e628 cf4b2a504 Author: Richard Henderson Date: Thu Jul 7 23:11:38 2016 -0700 Merge pull request #261 from tromey/fix-260 Don't dereference "ecif" before NULL check commit cf4b2a50413ecb8931eb1a94437497694f189c47 Author: Tom Tromey Date: Fri Jun 17 10:09:44 2016 +0100 Don't dereference "ecif" before NULL check Fixes #260 commit c0829e62841100c308cc37ddd6537ca69e856de7 Merge: 96d9bf4b5 794a54d4a Author: Richard Henderson Date: Sun Jun 5 15:05:21 2016 -0700 Merge pull request #259 from rth7680/master Mark win64.S with GNU-stack note commit 794a54d4a603639adf1a79d5b5bc15da437457e2 Author: Richard Henderson Date: Sun Jun 5 14:57:00 2016 -0700 Mark win64.S with GNU-stack note commit 96d9bf4b5517bc49c0129e8cfdee43098d3771a2 Merge: 02089a1b5 52fbd12dc Author: Richard Henderson Date: Fri Jun 3 23:04:56 2016 -0700 Merge pull request #253 from iains/darwin-build Some Darwin build fixes commit 52fbd12dca1bddf0b135a24e7136f10af79a06bd Author: Iain Sandoe Date: Thu May 26 09:00:00 2016 +0100 [Darwin-ppc, build] Fixes for clang build. clang is experimental on powerpc-darwin, but the changes are appropriate to cctools as as well. Use the 'official' simplified cmpwi insn, rather than the implied one accepted by cctools. Do not re-use a set value. commit 92810b4bc101fccead2234b7385d4fa5e7a3e56a Author: Iain Sandoe Date: Thu May 26 08:56:51 2016 +0100 [Darwin-x86, build] Fix up label prefixes, remove .purgem Darwin uses a label prefix of _. cctools assembler will not accept .purgem as a directive. commit 30b3440cfd378dc67111b1380f2654334d048659 Author: Iain Sandoe Date: Thu May 26 08:55:04 2016 +0100 [Darwin, configure] Allow configure to work for -*-darwin. The exec configury is appropriate for all current targets (including using older tools on them). commit 02089a1b5cf3aa1535172c62add454a2e9375066 Merge: 5d0365c60 c88455178 Author: Tom Tromey Date: Mon May 23 09:58:51 2016 -0600 Merge pull request #237 from tschwinge/libffi_feature_test Simplify/fix libffi_feature_test commit 5d0365c605c472f40d1195cf7ba04ae08e360cb7 Merge: 584a140e7 74b3f5201 Author: Richard Henderson Date: Thu May 19 09:35:38 2016 -0700 Merge pull request #251 from berkerpeksag/delete-test-define Remove unused FFI_CLOSURE_TEST commit 74b3f5201432c37e40e24386fa193c838596cec8 Author: Berker Peksag Date: Thu May 19 18:05:36 2016 +0300 Remove unused FFI_CLOSURE_TEST It was here since the first commit c6dddbd (warning: huge diff) and it wasn't defined by the configure script. It was probably used manually during development. commit 584a140e7558dff6e4ed4009f39d11dd5e3db0ae Merge: c61775497 b3496dedb Author: Tom Tromey Date: Thu May 19 06:36:38 2016 -0600 Merge pull request #248 from berkerpeksag/patch-1 Delete empty env in .travis.yml commit b3496dedbbad6c1f4e5bf4cc100e76b86f3c801e Author: Berker Peksag Date: Thu May 19 10:53:07 2016 +0300 Delete empty env in .travis.yml commit c61775497e7bdfc892d865c2eb15415d2da7493b Merge: 9edaa695c 33d9a31d3 Author: Richard Henderson Date: Wed May 18 11:10:33 2016 -0700 Merge pull request #247 from rth7680/symver Add parent to symbol version LIBFFI_BASE_7.1 commit 33d9a31d34cfba63e1294978114be1bddd2de3fd Author: Richard Henderson Date: Wed May 18 11:01:55 2016 -0700 Add parent to symbol version LIBFFI_BASE_7.1 commit 9edaa695c54daaf58d62e84032d30d33b7484fa8 Merge: 0969a1c1b 8a0a4e25e Author: Richard Henderson Date: Wed May 18 10:56:19 2016 -0700 Merge pull request #230 from rth7680/symver Use ELF symbol versioning commit 0969a1c1b339f845202fa7507b70d193220ceb01 Merge: 571b1ca62 4a677a425 Author: Tom Tromey Date: Wed May 18 10:09:28 2016 -0600 Merge pull request #232 from berkerpeksag/signcompare Fix -Wsign-compare warnings in x86/ffi64.c commit 571b1ca6277dbafc1f340a59ea1c5ee4ddb11c94 Merge: 069d2e4e1 1e82e1cda Author: Tom Tromey Date: Wed May 18 10:08:59 2016 -0600 Merge pull request #233 from berkerpeksag/mremap Define _GNU_SOURCE on Linux for mremap() commit 069d2e4e16c5570adec9249d642f2a2ae0ffaa1c Merge: a44a5b863 1b4f5ee6a Author: Tom Tromey Date: Wed May 18 10:07:31 2016 -0600 Merge pull request #246 from berkerpeksag/update-github-links Update project URLs to use libffi/libffi commit 1b4f5ee6a4989d7eda7766c5864908c389586c6f Author: Berker Peksag Date: Wed May 18 11:12:58 2016 +0300 Update project URLs to use libffi/libffi commit 1e82e1cda43dacd8b6ab2d9ac4db33523d86f5dc Author: Berker Peksag Date: Mon Mar 7 18:38:10 2016 +0200 Define _GNU_SOURCE on Linux for mremap() This was committed to CPython's libffi copy in https://bugs.python.org/issue10309 mremap() documentation says _GNU_SOURCE needs to be defined in order to use mremap(): see the synopsis section at http://linux.die.net/man/2/mremap Original commit: https://hg.python.org/cpython/rev/9986fff720a2 Original patch was written by Hallvard B Furuseth. commit 4a677a425c18eda4bc5357b2485da57f133f908d Author: Berker Peksag Date: Sat Mar 5 09:58:38 2016 +0200 Fix -Wsign-compare warnings in x86/ffi64.c This was originally reported on the Python tracker: httpa://bugs.python.org/issue23958 The original patch was written by Steve R. Hastings. I've updated it to current master of libffi. commit a44a5b863aaae98041e11aa3256907bf912ee4d6 Merge: 9443eaeda 2627fc98e Author: Tom Tromey Date: Tue May 17 17:07:27 2016 -0600 Merge pull request #234 from berkerpeksag/disable-docs Add a note about the --disable-docs switch commit 9443eaeda867cfdeae92b10bee3cc0506892a4c5 Merge: 14be9a1d0 e169ba2b8 Author: Tom Tromey Date: Tue May 17 17:04:50 2016 -0600 Merge pull request #242 from somasis/master Fix usage on musl libc commit 8a0a4e25ef99f4ec98761f2fc075c2e8409f171b Author: Richard Henderson Date: Wed May 4 07:33:41 2016 -1000 Update symbol versioning for ffi_get_struct_offsets Given that the original symbol versioning patch has shipped with gcc-6, it would be kind to keep that as a valid base ABI and make subsequent changes to version 7.1. commit c02c3414440e816aa13a2fe8446484173d3edb0e Author: Richard Henderson Date: Thu Mar 3 13:51:44 2016 -0800 Use ELF symbol versioning At the same time, we must bump the (major) ABI version. This needed to be done anyway due to ABI breakage in the AArch64 port (see 12cf89ee and the corresponding GCC PR70024). commit 14be9a1d0713deb97edb91fe9800afe971a001d5 Merge: 1cea27375 b50eabf48 Author: Richard Henderson Date: Wed May 4 06:45:05 2016 -1000 Merge pull request #245 from rth7680/tromey-ffi-prep-cif-core-is-private A rebase of #219 commit b50eabf48927ccaf0184c6224992bbb0a125e969 Author: Tom Tromey Date: Mon Feb 22 16:00:51 2016 -0700 minor comment cleanup This patch minor cleans up ffi.h.in comments in a minor way. It fixes some typos and capitalizations, adds some periods, and reformats some comments to a more GNU-ish style. It also fixes up some stale documentation. commit 6c07077a61fd0dc5ac11759ea1b4345acf486f3c Author: Tom Tromey Date: Sun Feb 21 20:08:21 2016 -0700 Change ffi.h.in so that braces match This is a tiny refactoring to make it so brace-matching works in Emacs. commit 2fbc0369b89a7e6b089e6384b98dba904ec1cf8b Author: Tom Tromey Date: Thu Nov 19 16:38:35 2015 -0700 move ffi_prep_cif_core to ffi_common.h commit 1cea273758ad862bdc2862af864e4ef5b42f31ae Merge: 48bfae1f3 6aafb5654 Author: Tom Tromey Date: Tue May 3 08:23:51 2016 -0600 Merge pull request #231 from berkerpeksag/simplfy-oserror Simplify error handling in mkdir_p() commit 48bfae1f37551b7fc894e6a0ea3b98968d2b5c9f Merge: 6d8711057 d76975dbd Author: Richard Henderson Date: Mon May 2 20:58:57 2016 -1000 Merge pull request #236 from andreas-schwab/master Define FFI_SIZEOF_JAVA_RAW for aarch64 ILP32 commit 6d8711057bc3a3befa37eed6765231ea5d244078 Merge: 9de24a3ff e5843a3a0 Author: Richard Henderson Date: Sun May 1 11:02:40 2016 -1000 Merge pull request #241 from rth7680/fix-win64 Fix win64 abi calling from unix64 commit e169ba2b83c780058fe626856cfdb5903a85cb97 Author: Kylie McClain Date: Fri Apr 29 21:04:07 2016 -0400 Fix usage on musl libc A gcc compiled on musl does not define __gnu_linux__, it defines __linux__. Only on glibc does __gnu_linux__ get defined, but both define __linux__, so we should check for that instead. With this patch, libffi works perfectly, and passes its testsuite entirely on musl libc systems. commit e5843a3a09976f9d8fa77671e9d6c188c890199d Author: Richard Henderson Date: Fri Apr 15 16:10:08 2016 -0700 x86: Fix calling convention for ffi_closure_win64_inner Also enable testing for the cross-abi calls. commit d06751979bf0e4c9caabf0bca531d74de8cb9152 Author: Richard Henderson Date: Mon Mar 7 12:14:22 2016 -0500 x86: Copy fix for clang .org from unix64.S Clang doesn't understand .org with symbolic operands. commit bfab429ee479723cc65831a7e4327e2866036420 Author: Jakub Kaszycki Date: Fri Apr 1 21:42:41 2016 +0200 Update README with a new port This is in fact not a NEW port. Apple Mac devices are generally x86-64 now, x86 rarely. If GCC exists for this CPU, it means that libffi has been built too. commit c88455178151bea14b8561ab6830724c31af255e Author: Thomas Schwinge Date: Thu Feb 25 20:10:18 2016 +0100 Simplify/fix libffi_feature_test As detailed in , the original code (originally added in commit f1301a54bb80e6ae23f7687c68f36875dae69134 as part of pull request #145) does not work (at least not for me, in the GCC environment). commit d76975dbd2854beca7acbca946c4cbccf9bf7cee Author: Andreas Schwab Date: Wed Mar 16 12:23:07 2016 +0100 Define FFI_SIZEOF_JAVA_RAW for aarch64 ILP32 Like x32, aarch64 ILP32 needs to define FFI_SIZEOF_JAVA_RAW. This fixes the java interpreter. commit 9de24a3ff31c7446de29c1adcd9276fd7ba93897 Merge: ee718066f 38a4d72c9 Author: Anthony Green Date: Mon Mar 14 13:54:53 2016 -0400 Merge pull request #212 from tromey/struct-layout add ffi_get_struct_offsets commit 2627fc98e86e906fd7c9bc16bf345b91ee26d2ba Author: Berker Peksag Date: Mon Mar 7 18:43:03 2016 +0200 Add a note about the --disable-docs switch Closes #204 commit 6aafb5654d6741927276efff1af72a5b3614da39 Author: Berker Peksag Date: Sat Mar 5 08:30:31 2016 +0200 Simplify error handling in mkdir_p() commit ee718066f763b65d1888f9cc7155d3a1880f80e9 Merge: f2f234aef e59fb55e0 Author: Anthony Green Date: Mon Feb 22 20:15:40 2016 -0500 Merge pull request #227 from tromey/fix-documentation-build Fix documentation build commit 38a4d72c95936d27cba1ac6e84e3094ffdfaa77c Author: Tom Tromey Date: Tue Nov 17 21:18:20 2015 -0700 add ffi_get_struct_offsets commit e59fb55e09fb8aa846c820a802a03fed408d05a6 Author: Tom Tromey Date: Mon Feb 22 15:51:27 2016 -0700 remove and ignore texinfo.tex This file is provided by Automake commit 6dd5fd5a643c50aff3303ba2add3b6cc7588df78 Author: Tom Tromey Date: Mon Feb 22 15:34:39 2016 -0700 fix documentation building An earlier patch added --disable-docs, but went too far, making it impossible to build the docs. It turns out that Automake seemingly has a bug preventing the conditional build of an info file. So, this patch works around the bug by putting the info_TEXINFOS rule into a new doc/Makefile.am. Tested by building with and without --disable-docs and looking for the existence of doc/libffi.info. commit f2f234aef203a5e836b83cb772f9473f7ea0d5ce Merge: 49b95edaf fa7a25711 Author: Anthony Green Date: Sat Feb 20 06:50:29 2016 -0500 Merge pull request #190 from havardgraff/master Add configure-option to disable building docs commit 49b95edafd712ce27e08b1bb4be2987f14c54ca3 Merge: 415723b40 43fc5bca4 Author: Anthony Green Date: Sat Feb 20 06:49:40 2016 -0500 Merge pull request #194 from amodra/master Correct powerpc sysv stack argument accounting commit 415723b40ba76f4a9df9b709db1cde0c65162072 Merge: 7c27fcbe0 1f6b5a91f Author: Anthony Green Date: Sat Feb 20 06:49:19 2016 -0500 Merge pull request #104 from joshtriplett/efi64 Support the Windows/EFI calling convention on all x86-64 targets commit 7c27fcbe09534d261b4098467a3ce258daec2807 Merge: 69143d06c e06124395 Author: Anthony Green Date: Sat Feb 20 06:45:00 2016 -0500 Merge pull request #199 from Pan7/master Fix configuring msvc compiler commit 69143d06c6276d09790a4e34f2eb6d37528c08c6 Merge: 755f1e642 505346e18 Author: Anthony Green Date: Sat Feb 20 06:44:28 2016 -0500 Merge pull request #197 from foxsen/mips_go_closure Mips go closure support commit 755f1e642d97892cb54358d65e06b60d1d6b72a4 Merge: 183ee6a67 4cdedc279 Author: Anthony Green Date: Sat Feb 20 06:43:48 2016 -0500 Merge pull request #193 from rth7680/fix-70 Test case from issue #70 commit 183ee6a6729db8f3ee0b1532d7bd54a6ef305443 Merge: 0bb71b52f 877ea9bf9 Author: Anthony Green Date: Sat Feb 20 06:43:28 2016 -0500 Merge pull request #206 from edolnx/master Add configure flag to disable multi-os-directory support commit 0bb71b52f5a60643ad55f801f1c652855b0e5cc5 Merge: 296a283a4 8bec5ca26 Author: Anthony Green Date: Sat Feb 20 06:41:10 2016 -0500 Merge pull request #211 from tromey/move-prep-types-out-of-ffi.h ffi_prep_types is internal-only commit 296a283a46057a6a643f1f64534324af34b7ed19 Merge: e7c713a0b 4805bf9cc Author: Anthony Green Date: Sat Feb 20 06:40:21 2016 -0500 Merge pull request #213 from tromey/remove-FFI_TYPE remove FFI_TYPE typedef commit e7c713a0b64c425b51f3d9437a8db7ca497e6c31 Merge: 79666abca f45e9e357 Author: Anthony Green Date: Sat Feb 20 06:40:03 2016 -0500 Merge pull request #214 from tromey/document-enum-support document (lack of) enum handling in libffi commit 79666abca87ac67d2d05b477a3ba71ec18b8af0c Merge: 20c7c31ca 89b760503 Author: Anthony Green Date: Sat Feb 20 06:38:40 2016 -0500 Merge pull request #205 from filbranden/tool_opts Fix dejagnu test support for --tool_opts commit 20c7c31ca04350670012b28756d4ebc9e7b67a68 Merge: 6ba2a49e2 9c14cbfc7 Author: Anthony Green Date: Sat Feb 20 06:38:01 2016 -0500 Merge pull request #225 from ehsan/safeseh Don't clobber safeseh for msvcc.sh -clang-cl in 32-bit mode commit 89b760503ed017233c98531908dfe8d8affab65e Author: Filipe Brandenburger Date: Mon Oct 19 15:33:23 2015 -0700 Fix dejagnu test support for --tool_opts Right now it concatenates it with the existing options and then appends it to that list, fix it to simply append it as is, same as it is done with the other variables. Tested by running the following command which includes gcc options: $ make check RUNTESTFLAGS="--tool_opts '-Werror'" Without this patch, all the tests fail. With it, the test succeed. Inspecting the logs shows that -Werror was indeed used when compiling the test sources. commit 9c14cbfc7ff49c2ceb9b65c548ddaced607f0785 Author: Ehsan Akhgari Date: Thu Jan 28 16:57:18 2016 -0500 Don't clobber safeseh for msvcc.sh -clang-cl in 32-bit mode The commit fb25cd0 went a bit too far and removed safeseh when -clang-cl was passed, but that's only needed in x86-64 which is already handled by the -m64 flag. I discovered this when building Firefox x86 with clang-cl. commit 6ba2a49e226b859ddde9a37c34973cf85380c525 Merge: 8005600a9 6a3a6acd9 Author: Anthony Green Date: Sat Jan 2 20:03:41 2016 -0500 Merge pull request #222 from freakboy3742/t181 Fixed #181 -- Corrected problems with ARMv7 build under iOS. commit 8005600a920f43d89ace7ef3598057392da726e1 Merge: 70594f7f8 5e2b18a65 Author: Anthony Green Date: Sat Jan 2 16:44:29 2016 -0500 Merge pull request #224 from plicease/travis update apt-get in travis commit 5e2b18a65d3f366b2fb02ee27321095c4b5f17ac Author: Graham Ollis Date: Sat Jan 2 07:56:05 2016 -0500 update apt-get in travis commit 6a3a6acd9fdbdac6f9f9294732b9627895b77715 Merge: bc4fc07aa 70594f7f8 Author: Russell Keith-Magee Date: Mon Dec 21 00:37:37 2015 +0800 Merge branch 'master' into t181 commit bc4fc07aa58df6d7fa90b57d305d58eef96a0416 Author: Russell Keith-Magee Date: Mon Dec 21 00:37:06 2015 +0800 Fixed #181 -- Corrected problems with ARMv7 build under iOS. Based on a patch from @fealebenpae, with input from @SolaWing and @rth7680, and testing from @superdump. commit 70594f7f8d7410be47ba9a80a566175d937ab626 Merge: 88deacdd4 6affad791 Author: Anthony Green Date: Fri Dec 18 10:24:38 2015 -0500 Merge pull request #221 from tromey/document-closure-ffi_arg Document closure ffi arg commit 6affad79111c45982fd29637a46bc756abb128a3 Author: Tom Tromey Date: Thu Dec 17 16:58:10 2015 -0700 correctly document closure return promotion commit 8a7cdeb0412f9ef8bbe0a7833d662dbec40b472f Author: Tom Tromey Date: Thu Dec 17 16:51:32 2015 -0700 fix formatting of ffi_prep_closure_loc arguments commit f45e9e35700d7c547ba33b02c4186e746dea204a Author: Tom Tromey Date: Thu Nov 19 15:21:14 2015 -0700 document (lack of) enum handling in libffi commit 4805bf9ccdc176687526e9e7fca0db763cef42fb Author: Tom Tromey Date: Thu Nov 19 14:17:42 2015 -0700 remove FFI_TYPE typedef commit 8bec5ca26a9977e6c05d6bfa1bbbc7f6f99af4e7 Author: Tom Tromey Date: Wed Nov 11 21:34:43 2015 -0700 ffi_prep_types is internal-only commit 88deacdd4220b66805b6ad10abd61fa3640afb67 Author: Anthony Green Date: Sun Nov 15 08:39:17 2015 -0500 speling fix commit 5973dca595f8595a40709f645e74b06990975db3 Merge: 2a988c445 473bcbcad Author: Anthony Green Date: Sun Nov 15 08:36:48 2015 -0500 Merge pull request #209 from tromey/documentation-updates documentation fixes commit 2a988c4455d3926c59997b68f7510be63e10c841 Merge: e9de7e35f ff33ddd1a Author: Anthony Green Date: Sun Nov 15 08:33:46 2015 -0500 Merge pull request #207 from tromey/deprecate-ffi-prep-closure mark ffi_prep_closure as deprecated commit 473bcbcad32e9e320a7a5c111cefd0d8ad570245 Author: Tom Tromey Date: Sat Nov 7 12:01:06 2015 -0700 documentation fixes Fixes #78. Documentation for #33 and #35, but no fix. commit ff33ddd1abe72b6d02d2769c0379b06eb7dbb931 Author: Tom Tromey Date: Sat Nov 7 11:45:04 2015 -0700 mark ffi_prep_closure as deprecated commit 877ea9bf9ac2c98cb858c12f5a6aeeec13cf978f Author: Carl Perry Date: Thu Oct 29 22:50:42 2015 -0500 Add configure flag to disable multi-os-directory support In some environments, installing the library to GCC's multi-os-directory is not perferable. This patch adds a switch to configure "--disable-multi-os-directory" which will disable that check, typically falling back to ${libdir} unless cross-compiling. Original patch was written by Stewart Brodie, and was found at https://sourceware.org/ml/libffi-discuss/2013/msg00144.html I've just updated it to work with the current version. commit 4cdedc279f96aefda24277184594c0ddd56d3402 Author: Richard Henderson Date: Sat Jul 25 14:29:15 2015 -0700 Tidy call.exp test collection Commit c952a92e20aa6013d8202d0b3fa1d87838c83054 moved all of the complex tests to libffi.complex, but failed to remove the anti-globbing from libffi.call. commit 609db2fbe41a56c1a8cd5653d248f08b88ea803f Author: Richard Henderson Date: Sat Jul 25 12:42:04 2015 -0700 Test case from Issue #70 commit e9de7e35f2339598b16cbb375f9992643ed81209 Merge: d19b6cc78 0ebde2d24 Author: Anthony Green Date: Sun Oct 11 10:54:50 2015 -0400 Merge pull request #202 from tromey/note-prep-cif-var-usage document that there must be ntotalargs types commit d19b6cc78e956e7ffd48b39fd7c441c22e633ce2 Merge: 17ffc3655 36f3fb914 Author: Anthony Green Date: Sun Oct 11 10:54:22 2015 -0400 Merge pull request #201 from tromey/fix-texinfo-buglet Add missing "@" to @var{} commit 0ebde2d24a3deb7b0cdf5cb1e625b55af2834fde Author: Tom Tromey Date: Tue Oct 6 11:42:43 2015 -0600 document that there must be ntotalargs types commit 36f3fb91412a6f10bd2c6dde01f4d7afa1f2f2c0 Author: Tom Tromey Date: Mon Oct 5 10:20:05 2015 -0600 Add missing "@" to @var{} commit e061243958d63080956887e121d275533570f7cc Author: Pan7 Date: Mon Sep 14 08:25:09 2015 +0200 Fix configuring msvc compiler commit 505346e18fe20eee7eb69a23a48f6951858c5267 Author: Zhang Fuxin Date: Wed Aug 26 09:57:10 2015 +0800 fix type error in unwind code commit 5953c66bd7510b048b332b4e7450f3fb8d687f8b Author: foxsen <2503799872@qq.com> Date: Thu Aug 20 20:28:13 2015 +0800 add unwind infor for *go_closure; reorder the labels to make thing more clear commit f0ecd5d40397d4616ba0bf3e2521da2663b5bd97 Author: Zhang Fuxin Date: Tue Aug 11 12:47:36 2015 +0800 fix O32 stack unwind code add missing 1: label commit 6f0201c803a9a1f0f9b6fd226a42fb8daa223928 Author: foxsen <2503799872@qq.com> Date: Tue Aug 4 18:25:34 2015 +0800 various fixes for go closure support. Now all n64 tests passed. commit 697dd4e8a03625a9b1448aa696978d8dcc57d438 Author: foxsen <2503799872@qq.com> Date: Tue Aug 4 12:53:33 2015 +0800 add support for go closure support on mips commit 43fc5bca48715a522b34c1124589575063362a90 Author: Alan Modra Date: Mon Aug 3 23:34:05 2015 +0930 Correct powerpc sysv stack argument accounting ppc32 starts using the stack for integer arg passing when we run out of integer arg passing registers. Similarly, we start using the stack for floating point args when we run out of floating point registers. The decision on where an integer arg goes does not depend on number of floating point args, nor does the decision on where a floating point arg goes depend on number of integer args. Alignment of stack args also simply depends on number of stack args. This patch untangles the horrible mess we had, with intarg_count being wrongly used to count both integer args and stack words. * src/powerpc/ffi_sysv.c (ffi_prep_cif_sysv_core): Count fprs, gprs, and stack words separately. (ffi_prep_args_SYSV): Similarly. commit 1f6b5a91f417ac77d2fe9b0b3eb66293db132e2e Author: Josh Triplett Date: Sun Jul 26 16:27:34 2015 -0700 Support the WIN64/EFI64 calling convention on all X86_64 platforms Add a new calling convention FFI_EFI64, alias FFI_WIN64, on all X86_64 platforms. This allows libffi compiled on a 64-bit x86 platform to call EFI functions. Compile in ffiw64.c and win64.S on all X86_64 platforms. When compiled for a platform other than X86_WIN64, ffiw64.c suffixes its functions with _efi64, to avoid conflict with the platform's actual implementations of those functions. commit 6de51f3e04e496901ea1bd8f9b44f75c9f01b599 Author: Josh Triplett Date: Sun Jul 26 16:23:55 2015 -0700 src/x86/ffiw64.c: Don't assign a "char *" to an "unsigned char *" Declare a local variable to match the type of the struct field assigned to it, rather than adding unsigned to the type. Fixes a -Wpointer-sign warning. commit eaa59755fcbb692a8cb763c7f9f24a350aadbd30 Author: Josh Triplett Date: Sun Jul 26 17:17:16 2015 -0700 src/x86/win64.S: Handle name mangling and PIC Move the macros from unix64.S into a shared header asmnames.h and use them in win64.S too. commit c8e82d9fbffd3eeaef0266a1aac64d7bd13ee9c3 Author: Josh Triplett Date: Sun Jul 26 16:18:57 2015 -0700 src/x86/win64.S: Support compiling on non-WIN64 platforms Non-WIN64 versions of the GNU assembler don't support the .seh_* directives for structured exception handling, so wrap them in a macro that compiles to nothing. Handle the registers used for the non-Windows x86-64 calling convention when on a non-Windows platform. Distinguish between cases that should refer to the native argument registers (defined as arg0, arg1, arg2, and arg3) and cases that should always refer to the Windows argument registers. commit fa7a257113e2cfc963a0be9dca5d7b4c73999dcc Author: Havard Graff Date: Tue May 12 10:45:13 2015 +0200 Add configure-option to disable building docs commit e3d2812ce43940aacae5bab2d0e965278cb1e7ea Author: Russell Keith-Magee Date: Sat Apr 25 19:03:03 2015 +0800 Modified arm/sysv.S to remove directives not allowed by clang. commit 17ffc3655a531c116e9eb9cc933e50bb1e5c47f8 Merge: 173757ce9 a5b3eaa27 Author: Anthony Green Date: Sun Mar 29 21:38:26 2015 -0400 Merge pull request #185 from freakboy3742/t184 Fixed #184 -- Corrected source file references in Darwin source generator tool. commit 173757ce973a2f12ad36ac6cda593581b225f9e3 Merge: f5ec6279a dd9e4c8d4 Author: Anthony Green Date: Sun Mar 29 21:38:12 2015 -0400 Merge pull request #186 from freakboy3742/gitignore Added generated source file to .gitignore. commit dd9e4c8d46b65104a24a9858accd0936ca44b98f Author: Russell Keith-Magee Date: Fri Mar 13 08:03:13 2015 +0800 Added generated source file to .gitignore. commit a5b3eaa279a2e54721d44eeefdfd1962cf63caaf Author: Russell Keith-Magee Date: Fri Mar 13 07:59:11 2015 +0800 Fixed #184 -- Corrected file references in Darwin source generator tool. commit f5ec6279a4698a8752093247a500cf76096ff157 Merge: 06747d318 95df37919 Author: Anthony Green Date: Wed Mar 4 13:58:04 2015 -0500 Merge pull request #178 from rth7680/aa64-ilp32 aarch64: Handle ILP32 ABI commit 95df3791935d934d92b0c852af1e655285157b70 Author: Andrew Pinski Date: Wed Feb 11 08:31:48 2015 -0800 aarch64: Handle ILP32 ABI commit 06747d318761884e70dfa433a9548219fd779d7d Merge: db1b34b7e 6c5357757 Author: Anthony Green Date: Thu Jan 29 07:22:23 2015 -0500 Merge pull request #176 from 0-wiz-0/master Handle NetBSD/powerpc the same as FreeBSD and OpenBSD. commit 6c5357757f188577fffcd2890ef7a0f53e77de20 Author: Thomas Klausner Date: Thu Jan 29 12:32:28 2015 +0100 Handle NetBSD/powerpc the same as FreeBSD and OpenBSD. Signed-off-by: Thomas Klausner commit db1b34b7e1f5e473d17557e454a29933dfecd1af Author: Anthony Green Date: Wed Jan 28 18:08:06 2015 -0500 Remove incomplete sentence commit 56036a6b58acb9e854719e315a0738f16e354c35 Merge: 035715b33 2104b2a4f Author: Anthony Green Date: Wed Jan 28 18:03:57 2015 -0500 Merge pull request #175 from rth7680/sparc sparc: Re-introduce hand-written unwind info commit 035715b3330421806df36ec4ac7ccecc86458540 Merge: 31a618530 b0e979634 Author: Anthony Green Date: Wed Jan 28 18:03:06 2015 -0500 Merge pull request #174 from rth7680/fbsd configure: Run HAVE_AS_X86_PCREL for all X86 targets commit 2104b2a4fc923b8d82647c3fbbd8a8ca20bb24c8 Author: Richard Henderson Date: Mon Jan 26 12:43:57 2015 -0800 sparc: Re-introduce hand-written unwind info Fixes the build with the Solaris assembler. commit b0e9796344562ea8e1ef28e9b04db6be26ca29fb Author: Richard Henderson Date: Tue Jan 27 22:10:12 2015 -0800 configure: Run HAVE_AS_X86_PCREL for all X86 targets commit 31a618530737cc8f1666845f8e65a3c097f6d408 Merge: d6675c16c 536366349 Author: Anthony Green Date: Wed Jan 21 05:55:47 2015 -0500 Merge pull request #170 from fealebenpae/aarch64-trampoline-table Support closures on ARM64 iOS commit d6675c16cc2b5a0e32c4538e5a4f555c91e6bb56 Merge: a25a46a7f 3ac1610aa Author: Anthony Green Date: Wed Jan 21 05:53:21 2015 -0500 Merge pull request #172 from rth7680/fixes x86: Fix cygwin32 build commit 3ac1610aa33c887ea9b14935208943925714a33e Author: Richard Henderson Date: Mon Jan 19 20:48:40 2015 +0100 x86: Fix cygwin32 build The section syntax is just that little bit different. commit a25a46a7fe4f7ae27fbd6e08a2540678899914e9 Merge: 5cd411ad5 1ad0b1717 Author: Anthony Green Date: Fri Jan 16 17:32:07 2015 -0500 Merge pull request #171 from rth7680/fixes Fixes from mainline gcc commit 1ad0b17177526be86736e245fbcc21eadaf7cb36 Author: Richard Henderson Date: Fri Jan 16 13:30:05 2015 -0800 sparc: Also mark the return address in unwind info commit d68c8aed19d459a9fef0847f316dd9473ef263ad Author: Richard Henderson Date: Fri Jan 16 11:40:33 2015 -0800 sparc: Solaris fixes, part 2 /bin/as seems to only understand single-digit labels /bin/as knows nothing about .rept/.endr commit b740ab7cc955fc2888325e6387207f34400a6c45 Author: Rainer Orth Date: Fri Jan 16 11:32:23 2015 -0800 sparc: Solaris fixes * /bin/as requires .type fn,#function instead of @function. * /bin/as doesn't support .macro/.endm. I'm using preprocessor macros instead to implement E in src/sparc/v[89].S. commit f1560b7bbeb57042a32dfde487c3d21bd9ef6a51 Author: Rainer Orth Date: Fri Jan 16 11:31:37 2015 -0800 x86: Solaris fixes * Solaris/x86 /bin/as doesn't support .org, so I've just disabled the uses in src/x86/{sysv, unix64}.S, as on Darwin. * Solaris/x86 needs to use EH_FRAME_FLAGS so manually and compiler generated .eh_frame sections match, otherwise libffi.so fails to link: * Solaris/x86 /bin/as has different COMDAT syntax; I've disabled it for the moment. commit 536366349ced9d57c873d64e219ba30992a83c06 Author: Yavor Georgiev Date: Fri Jan 16 15:19:38 2015 +0200 aarch64: implement the trampoline table workaround for ffi closures on Apple systems This is a direct copy/paste port of the ARM code, with changes because of Aarch64 pc-relative addressing restrictions. commit 05e65b4e2064f24669e8cc1c1cece1cbb0577548 Author: Yavor Georgiev Date: Fri Jan 16 15:18:04 2015 +0200 aarch64: rewrite range syntax into list to appease Clang Clang's assembler in Xcode 6 appears to choke when the operand of st4 is a range, but is happy with a list. commit 5cd411ad5d62b1dadfd02f75f99121801c8be4c9 Author: Anthony Green Date: Tue Jan 13 15:44:03 2015 -0500 New test case for old aarch64 bug commit e46842b4149e19106a9eb6066ef022d8b6a7ad5d Author: Richard Henderson Date: Tue Jan 13 07:23:48 2015 -0800 Remove extra brackets in configure.host This table is no longer in configure.ac, needing protection from m4. commit f27c4e46734fa1342236b16161e6006a9d2557e9 Author: Richard Henderson Date: Tue Jan 13 07:22:07 2015 -0800 x86: Fix thinko in ffi_raw_call Missed structure initialization for raw path. Apparently there are no tests for this outside gcc. commit c82cc159426d8d4402375fa1ae3f045b9cf82e16 Merge: 1c61e73ad e1a5ddb30 Author: Anthony Green Date: Sat Jan 10 09:25:26 2015 -0500 Merge pull request #166 from chevah/master Fix expr error in Solaris 10 when using gcc. commit 1c61e73ad8e920d6bff337fac02de71f3a66bc38 Merge: dd0b59a5c 3fa5d70cb Author: Anthony Green Date: Sat Jan 10 09:23:30 2015 -0500 Merge pull request #165 from rth7680/pcc Support PCC as producer and consumer commit dd0b59a5cf63e0f9602c76fc89a4cb62593ff6f1 Merge: 9131039c9 b7f6d7aa9 Author: Anthony Green Date: Sat Jan 10 09:22:55 2015 -0500 Merge pull request #164 from rth7680/darwin Fix build on darwin commit 9131039c93b6ecd1c3946905a3b6dafb5dc4ee40 Merge: 4ca2262ad 7282d3289 Author: Anthony Green Date: Sat Jan 10 09:22:42 2015 -0500 Merge pull request #160 from nobu/msvc-no-complex x86: MSVC does not support Complex type commit 4ca2262ad61189276b2d95aab652a23c8db2a5b9 Merge: 58bf7d65d 5f8881a59 Author: Anthony Green Date: Sat Jan 10 09:21:37 2015 -0500 Merge pull request #159 from nobu/fix-void-arith x86: Fix void pointer arithmetic commit e1a5ddb3067f83563cb8a6b8ac4d0f58d4699262 Author: Mișu Moldovan Date: Wed Jan 7 17:51:07 2015 +0200 Fix expr error in Solaris 10 when using gcc. commit 3fa5d70cbb18b39a5e44f1c7984dedf73446bf6c Author: Richard Henderson Date: Mon Jan 5 13:03:06 2015 -0800 x86: Avoid fastcall when building with pcc Apparently, PCC doesn't support the fastcall calling convention. Nor does it issue a warning or error for the attribute that it does not understand. commit a03d2310ed53bb8b3a4610af04015ef9df6ea36c Author: Richard Henderson Date: Wed Dec 24 16:03:34 2014 -0800 x86: Load structure return address into eax commit b7f6d7aa9b0d7b19eec28a945251e09a4b65b275 Author: Richard Henderson Date: Wed Dec 10 13:37:36 2014 -0800 x86: Reinstate hand-written unwind info for sysv.S commit 6cedf81ca7cbad01b3fcc2c4475d860095bfb062 Author: Richard Henderson Date: Wed Dec 10 09:43:58 2014 -0800 x86: Expand FFI_GO_CLOSURE If we're going to have to hand-write unwind info for darwin, these macros make the job harder. commit ae842a515b6e2a486d73ce8f9dbe5fee6c3fd1ff Author: Iain Sandoe Date: Tue Nov 25 11:43:40 2014 +0100 x86: More Darwin unwind fixups EHFrame{N} IIRC is a special cue to ld64 that it should treat the unwind in the object as "special/legacy" .. [these days everything is .cfi_xxxx (except, cctools-as, as you noted)] .. without that much confusion arises with ld64's atom-isation of the eh_frame section. xxxx.eh labels are not needed for darwin ld64 >= 85.2.1 (i.e. darwin9, xcode 3.1.4) to all intents and purposes, that's all that matters now, since I think that anyone trying to build on 10.4/darwin8/xcode2.5 would have to use a later ld64 (from odcctools) for other reasons. commit 8fa3c9f24b5ef6da2f24a9f071984d033dd9e00c Author: Richard Henderson Date: Tue Nov 25 09:27:54 2014 +0100 x86: Reinstate hand-written unwind info for unix64.S One more try to get default Darwin to work. commit 5f35e0ffcc05a72ce0aacf228dc06d1262754660 Author: Richard Henderson Date: Mon Nov 24 16:26:50 2014 +0100 x86: Avoid using gas local labels Which are unsupported by Darwin cctools as. Thankfully this doesn't uglify the source too much. commit ed1ca2777c35fe2f2751de255df3e16f17bdbd8d Author: Richard Henderson Date: Mon Nov 24 13:02:03 2014 +0100 x86: Remove use of .cfi_escape The unwind info isn't 100% correct at all points during the epilogue, and not annotating is just as incorrect as the annotation. This works better on systems that do not support DW_OP_call_frame_cfa. commit 1b12593d7e595c03e5c06377feaeb8ad1b813681 Author: Richard Henderson Date: Mon Nov 24 12:55:43 2014 +0100 x86: Honor alignment of arguments Darwin aligns long-double to 16, and thus all of the long double tests were failing due to not honoring that. We ought to be able to devise a test case for GCC using __attribute__((aligned)) that would have failed too. commit 042b8dafeeee82667e00660fb1edeab72fd9de47 Author: Richard Henderson Date: Mon Nov 24 11:24:02 2014 +0100 x86: Use .balign not .align The Apple assembler defaults to power of two alignment, rather than byte alignment like everyone else. Force byte alignment by using the proper directive. commit 0172bc029cd943dce99f82b9d857c51e746574e5 Author: Richard Henderson Date: Mon Nov 24 10:42:02 2014 +0100 x86: Disable .org for Darwin commit 9f112619c187e135132b765adeedef89ee354eb2 Author: Richard Henderson Date: Sat Nov 22 20:02:43 2014 +0100 x86: Best guess at update for Darwin commit 7282d3289c6f6e2cc21e2c2b554c33af00ac3474 Author: Nobuyoshi Nakada Date: Mon Dec 22 17:14:40 2014 +0900 x86: MSVC does not support Complex type commit 5f8881a5913a9554cbc2b5fed7057627b9bb9ece Author: Nobuyoshi Nakada Date: Mon Dec 22 17:08:08 2014 +0900 x86: Fix void pointer arithmetic commit 58bf7d65d8896f198624c591cc31e6e01a09cc31 Merge: 75b2199f2 7ba30b190 Author: Anthony Green Date: Sat Dec 20 10:20:40 2014 -0500 Merge pull request #158 from rth7680/s390 S390 commit 7ba30b19060dfefefeb91970a862e5b215606602 Author: Richard Henderson Date: Fri Dec 19 11:38:17 2014 -0500 s390: Inline and tidy ffi_prep_args As per discussion with Ulrich Weigand, document the restrictions on the code within ffi_call_int as we simultaneously prepare stack frames for ffi_call_SYSV and the target function. commit f69ec6f35db93163b7f33e43129d274995c8122d Author: Richard Henderson Date: Thu Dec 18 16:21:07 2014 -0500 s390: Use pc-relative insns in 31-bit mode It's silly to stick to esa/390 features when the compiler won't. Detect when brasl and larl are used by the compiler and then use them in the assembly. commit 2f530de168e0253ac06e044c832132c496e8788b Author: Richard Henderson Date: Thu Dec 18 16:01:59 2014 -0500 s390: Reorganize assembly Avoid using ffi_prep_args as a callback; do all the work setting up the frame within ffi_call_int directly. Save fewer registers in ffi_closure_SYSV. commit 97512ded052678993ffcfa3aec0035cce1585138 Author: Richard Henderson Date: Thu Dec 18 16:01:15 2014 -0500 s390: Avoid aliasing warnings commit c860ca9ac0fc7ee44124c5637dda751f55e4fa3f Author: Richard Henderson Date: Wed Dec 17 13:24:03 2014 -0500 s390: Kill trailing whitespace commit 02b7c89967307cea91714fb6e4e53048e1b5396e Author: Dominik Vogt Date: Wed Dec 17 13:20:51 2014 -0500 s390: Go closure support commit 75b2199f264f42de814528ccf4bfd2ef427665b5 Merge: 8a3a2723a 2f6524696 Author: Anthony Green Date: Fri Dec 12 10:13:56 2014 -0500 Merge pull request #157 from rth7680/x86 Two fixes for x86 commit 8a3a2723aaa615cb95ee1c798469d7a3b2d95d31 Merge: 9ae3bc870 542e00471 Author: Anthony Green Date: Fri Dec 12 10:13:27 2014 -0500 Merge pull request #156 from rth7680/sparc sparc: Define FFI_TARGET_SPECIFIC_VARIADIC for v9 commit 2f652469684d6d2b3e54705294c7e81cffe1341f Author: Richard Henderson Date: Thu Dec 11 14:16:00 2014 -0800 x86: Handle void arguments as if an empty structure Since libffi currently doesn't allow empty structures, libgo currently maps them to ffi_type_void. Given that we'll abort on this case, handle it gracefully. commit 097ccfd6a6b94a3c3b5bfc269a7afd1dc4f7a051 Author: Richard Henderson Date: Wed Dec 10 13:25:14 2014 -0800 x86: Fix some unwind errors commit 9ae3bc870ee07820b60d368f137ea3b5daca850e Author: James Greenhalgh Date: Sat Dec 6 23:58:41 2014 -0500 Fix for https://github.com/atgreen/libffi/issues/141 commit a5a40960bb6ccb029d2d9d4e18d37f6e0dacb913 Merge: 9ca43706f 590663b30 Author: Anthony Green Date: Wed Nov 19 21:13:50 2014 -0500 Merge pull request #151 from amodra/master powerpc: go closures for linux commit 590663b30b1f35b6136cf55d8870a2394ed78c21 Author: Richard Henderson Date: Tue Nov 18 12:56:58 2014 +1030 powerpc: Fix ffi_go_closure_linux64 Unlike ffi_closure_LINUX64, this entry point is called normally, so we already have the TOC in R2 and the closure in R11. * powerpc/linux64_closure.S (ffi_closure_LINUX64): Remove a register dependency chain. (ffi_go_closure_linux64): Don't load r11 or r2. commit fa1040c111b3e423bc9c7e78d9af89470c0fa2fb Author: Alan Modra Date: Mon Nov 10 09:42:31 2014 +1030 GO closures for powerpc linux Plus .cfi async unwind info, rearrangement of ffi_call_linux64 and ffi_call_SYSV function params to avoid register copies, tweaks to trampolines. * src/powerpc/ffitarget.h (FFI_GO_CLOSURES): Define. * src/powerpc/ffi.c (ffi_call_int): New function with extra closure param, and args rearranged on ffi_call_linux64 and ffi_call_SYSV calls, extracted from .. (ffi_call): ..here. (ffi_call_go, ffi_prep_go_closure): New functions. * src/powerpc/ffi_linux64.c (ffi_prep_closure_loc_linux64): Make hidden. Only flush insn part of ELFv2 trampoline. Don't shuffle ELFv1 trampoline. (ffi_closure_helper_LINUX64): Replace closure param with cif, fun, user_data params. * src/powerpc/ffi_powerpc.h (ffi_go_closure_sysv): Declare. (ffi_go_closure_linux64): Declare. (ffi_call_SYSV, fi_call_LINUX64): Update. (ffi_prep_closure_loc_sysv, ffi_prep_closure_loc_linux64): Declare. (ffi_closure_helper_SYSV, ffi_closure_helper_LINUX64): Update. * src/powerpc/ffi_sysv.c (ASM_NEEDS_REGISTERS): Increase to 6. (ffi_prep_closure_loc_sysv): Use bcl in trampoline, put data words last, flush just the insn part. (ffi_closure_helper_SYSV): Replace closure param with cif, fun and user_data params. * src/powerpc/linux64.S (ffi_call_LINUX64): Replace hand-written .eh_frame with .cfi directives. Adjust for changed param order. Pass extra "closure" param to user function in static chain. Add .cfi directives to describe epilogue. Don't provide traceback table for ELFv2 or _CALL_LINUX. * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Replace hand-written .eh_frame with .cfi directives. Adjust for changed ffi_closure_helper_LINUX64 params. Add .cfi directives to describe epilogue. Don't provide traceback table for ELFv2 or _CALL_LINUX. (ffi_go_closure_linux64): New function. * src/powerpc/sysv.S: Remove redundant .globl ffi_prep_args_SYSV. (ffi_call_SYSV): Make hidden. Replace hand-written .eh_frame with .cfi directives. Adjust for changed params. Pass extra "closure" param to user function in static chain. Add .cfi directives to describe epilogue. * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Make hidden. Replace hand-written .eh_frame with .cfi directives. Adjust for changed ffi_closure_helper_SYSV params. Add .cfi directives to describe epilogue. Don't just use nops in the dead __NO_FPRS__ epilogues. (ffi_go_closure_sysv): New function. commit d3d06f4c94847402a08f0e4e2af2e4d726ed3180 Author: Alan Modra Date: Mon Nov 10 13:50:05 2014 +1030 Fix powerpc breakage from 6e8a4460 * src/powerpc/ffitarget.h: #error on unexpected FFI_TYPE_LAST. (FFI_PPC_TYPE_LAST): Define. (FFI_TYPE_UINT128): Define in terms of FFI_PPC_TYPE_LAST. (FFI_SYSV_TYPE_SMALL_STRUCT, FFI_V2_TYPE_FLOAT_HOMOG): Likewise. (FFI_V2_TYPE_DOUBLE_HOMOG, FFI_V2_TYPE_SMALL_STRUCT): Likewise. commit 9ca43706f9e59ee3a0b288884c90fd51d8926a9b Merge: 5211c7207 a9ed0c3a0 Author: Anthony Green Date: Wed Nov 19 11:32:08 2014 -0500 Merge pull request #148 from gpakosz/dlmalloc-preprocessor-fix dlmalloc: change defined(i386) to defined(__i386__) commit 542e004710e3f1d7c137bba305a16538cd6257d6 Author: Richard Henderson Date: Tue Nov 18 05:07:00 2014 -0800 sparc: Define FFI_TARGET_SPECIFIC_VARIADIC for v9 This is a port of http://gcc.gnu.org/viewcvs?rev=207763&root=gcc&view=rev aka GCC PR libffi/60073, to the rewritten Sparc codebase. Supposedly, we should have seen failures with the existing libffi.call/cls_double_va.c testcase, but I hadn't. Perhaps a gcc newer than 4.6.3 is required to see that... commit a9ed0c3a0428f5fd4ed2166fa95a608595e11a4d Author: Gregory Pakosz Date: Tue Nov 18 09:37:07 2014 +0100 dlmalloc: change defined(i386) to defined(__i386__) When compiling with --std==c99, GCC and clang don't define i386 but __i386__ commit 5211c7207ec81dfdfa9acdcd86b49b9ff66b9d91 Merge: 3316b666e ccdd7bb85 Author: Anthony Green Date: Mon Nov 17 09:07:07 2014 -0500 Merge pull request #147 from rth7680/go/alpha testsuite: Fix alpha static chain register name commit ccdd7bb8566b2fd1da5c4b5c8eaa2db43a69e720 Author: Richard Henderson Date: Sun Nov 16 12:12:23 2014 +0100 testsuite: Fix alpha static chain register name commit 3316b666e3074fe0eebff4fe0458979961b26651 Merge: b698638d7 dea49e202 Author: Anthony Green Date: Sat Nov 15 07:31:41 2014 -0500 Merge pull request #145 from rth7680/master Configure and testsuite cleanups, v2 commit b698638d7601de994410c5c1b2f5c2c648015b7b Merge: a353d5b16 771fabc6d Author: Anthony Green Date: Sat Nov 15 07:29:58 2014 -0500 Merge pull request #144 from atrosinenko/testsuite-fp-comparison-fix Floating point number comparison fix for testsuite commit a353d5b16cfbb60f1a42bdec74b0f272701f3ecb Merge: 4aa702a0e e029c70fa Author: Anthony Green Date: Sat Nov 15 07:27:21 2014 -0500 Merge pull request #143 from adamkaplan/patch-1 Update i386 & armv7 minimum iOS version to 7.0 commit dea49e2020b2fc97ba9293200f316f7a7b3ca73b Author: Richard Henderson Date: Fri Nov 14 13:05:14 2014 +0100 x86: Fix typo in ffi_prep_go_closure Used the wrong register for THISCALL and FASTCALL. commit c9f5b6648b0f052bbca8b50615284dd975e9ed29 Author: Richard Henderson Date: Fri Nov 14 13:04:33 2014 +0100 testsuite: Add trivial tests for Go closures commit c952a92e20aa6013d8202d0b3fa1d87838c83054 Author: Richard Henderson Date: Fri Nov 14 11:00:14 2014 +0100 testsuite: Move complex tests to their own subdirectory It seems a bit silly to isolate them by globbing vs "*complex*" when we can just as easily put them in their own subdirectory. commit f1301a54bb80e6ae23f7687c68f36875dae69134 Author: Richard Henderson Date: Fri Nov 14 10:50:29 2014 +0100 testsuite: Use feature test rather than enumeration for complex commit b5ade2fb5d9ba06519484677a5474e5dad48c2e3 Author: Richard Henderson Date: Thu Nov 13 09:06:10 2014 -0800 testsuite: Detect clang Clang doesn't like the -Wno-psabi argument that we want to pass to GCC. Since clang is detected as GCC via __GNUC__, use ax_cv_c_compiler_vendor. commit 5d69d57a053c5049df29242def492159e1fadbae Author: Richard Henderson Date: Thu Nov 13 13:50:39 2014 +0100 configure: Move target source selection into configure.host This eliminates the AM_CONDITIONAL ugliness, which eliminates just a bit of extra boilerplate for a new target. At the same time, properly categorize the EXTRA_DIST files into SOURCES and HEADERS, for the generation of ctags. commit 771fabc6d5379e64e5ab9e5f1ca6eb5960617432 Author: Anatoly Trosinenko Date: Fri Nov 14 14:21:35 2014 +0300 Take a float absolute value using fabs() instead of abs(). Replace integer abs() by floating point fabs() in the approximate equality check for float values. commit 9622ede2e24067a09380c63da8d5007ec1f03dc4 Author: Anatoly Trosinenko Date: Fri Nov 14 13:18:04 2014 +0300 Fix floating point number comparisons in testsuite/libffi.call/float[123].c. Rewrite the checks for approximate equality of floating point return values to be in the form "fabs(a - b) < EPS" instead of just "a - b < EPS". commit e029c70fa7da7a489157b2a2a88349aaffb051de Author: Adam Date: Thu Nov 13 15:08:16 2014 -0500 Update i386 & armv7 minimum iOS version to 7.0 iOS 5 (and any below 7.1) is deprecated. Apple doesn't distribute the supporting libraries for this platform anymore as of Xcode 6, so it causes a linker error. commit 57f52484573613e2def21171184f50bd72209cac Author: Richard Henderson Date: Thu Nov 13 12:42:54 2014 +0100 configure: Split out configure.host Split out the host case statement to a separate file, so that we don't have to regenerate configure.in for changes therein. commit f8632815a6515a6709802ad23909585664ba1b9d Author: Richard Henderson Date: Thu Nov 13 12:32:35 2014 +0100 powerpc: Delete patch output Clearly added by mistake. commit 4aa702a0e0ffc9b8a672af380103f607d339a822 Merge: 20562ac04 8d5debc79 Author: Anthony Green Date: Wed Nov 12 08:08:57 2014 -0500 Merge branch 'rth7680-go-closure' commit 8d5debc7904aa5c4f1fbf5d5b26450ec5a5b43c6 Author: Anthony Green Date: Wed Nov 12 08:04:51 2014 -0500 Update in preparation for next release commit 56735e05fed77a3920a8cac262666c9eb0e75b66 Merge: 20562ac04 0e303c065 Author: Richard Henderson Date: Wed Nov 12 07:11:53 2014 -0500 Merge branch 'go-closure' of https://github.com/rth7680/libffi into rth7680-go-closure Conflicts: src/aarch64/ffi.c commit 20562ac0427c3578250d04c6e34fb0127d4551cf Author: Anthony Green Date: Wed Nov 12 07:00:59 2014 -0500 Fix for AArch64. Release as 3.2.1. commit 0e303c065779afb42cfdb2ea20c0e1a557dc16f0 Author: Richard Henderson Date: Wed Nov 12 03:58:58 2014 -0800 x86: Work around clang bugs http://llvm.org/bugs/show_bug.cgi?21500 http://llvm.org/bugs/show_bug.cgi?21501 http://llvm.org/bugs/show_bug.cgi?21515 commit 6eec410f1512d0584276db5f4b83711330e830d7 Author: Richard Henderson Date: Sun Oct 26 15:29:04 2014 -0700 sparc: Re-add abi compliant structure support The original code, removed in the "rewrite" patch, was incorrect for large structures, and required dynamic allocation of a trampoline on every ffi_call. Instead, allocate a 4k entry table of all possible structure returns. The table is 80k, but is read-only and dynamically paged, which ought to be better than allocating the trampoline. This is difficult to test with gcc. One can only use -O0 at present. See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63668. commit 92022496ef7a9439f48a2ef11e460c300ac863d7 Author: Richard Henderson Date: Sun Oct 26 14:48:28 2014 -0700 sparc: Add support for Go closures commit ad89c2d9e93d5a713ae4113567b4ca6c4fbbf17c Author: Richard Henderson Date: Sun Oct 26 13:16:03 2014 -0700 sparc: Add support for complex types commit 20da5b41324de524ea3ee1f84f402828b7c332dc Author: Richard Henderson Date: Sat Oct 25 16:24:41 2014 -0700 sparc: Handle more cases of structure return directly commit 0686c2e71c8e094cf003f199e3f578bcdede43ba Author: Richard Henderson Date: Sat Oct 25 14:52:45 2014 -0700 sparc: Preprocess float point struct return We can eliminate recursion and speed structure return by flattening a nested structure tree into a bitmask. commit 2b27890ba77db6a23d37fc70897109e4b2803c2d Author: Richard Henderson Date: Fri Oct 24 16:10:48 2014 -0700 sparc: Rewrite everything It's impossible to call between v8 and v9 ABIs, because of the stack bias in the v9 ABI. So let's not pretend it's just not implemented yet. Split the v9 code out to a separate file. The register windows prevent ffi_call from setting up the entire stack frame the assembly, but we needn't make an indirect call back to prep_args. commit 5d7b54450582f31d0d5d3ecca7f330b70446f195 Author: Richard Henderson Date: Fri Oct 24 11:45:53 2014 -0700 sparc: Tidy up symbols Assembly to use local labels, .type annotation, hidden annotation. I do retain the _prefix for the symbols, but given that it wasn't done consistently across all symbols, I doubt it's actually needed. commit fd4f7cc0cabae66a91e40f8645791740b9416bc1 Author: Richard Henderson Date: Fri Oct 24 11:11:53 2014 -0700 sparc: Eliminate long double ifdefs commit ab83cbb90b280195e636976098d3674f5ffc1d0a Author: Richard Henderson Date: Wed Oct 29 14:38:42 2014 -0400 arm: Add support for Go closures commit 6fa617dabc3d99e19d47ac9bc4127c56e3163008 Author: Richard Henderson Date: Tue Oct 21 11:27:11 2014 -0400 arm: Add argument space for the hidden struct return pointer This should have been failing all along, but it's only exposed by the complex_int test case. commit a529bec27b64ab846ab2643ecbeb22e01f35c7dc Author: Richard Henderson Date: Tue Oct 21 11:26:59 2014 -0400 arm: Add support for complex types commit 5e88ebe62717b8e1e2f8d96c10f78a7134caa765 Author: Richard Henderson Date: Mon Oct 20 15:10:43 2014 -0400 arm: Remove internal FFI_TYPE constants These have been replaced by the contents of internal.h. commit a4b785ea695e929b91d64e1e8b514411ae8bafe8 Author: Richard Henderson Date: Fri Oct 17 02:07:32 2014 -0400 arm: Rewrite ffi_closure Move the push of the argument registers into ffi_closure_SYSV, reducing the size of the trampoline. commit e7f15f60e86ad8931b2c4a516945d0b76b4bda20 Author: Richard Henderson Date: Fri Oct 17 01:27:16 2014 -0400 arm: Rewrite ffi_call Use the trick to allocate the stack frame for ffi_call_SYSV within ffi_call itself. commit a74a3aaddbcc730d9aa0a4787f136c9cc19b82b3 Author: Richard Henderson Date: Fri Oct 17 01:21:22 2014 -0400 arm: Rewrite vfp_type_p Do not modify the ffi_type. Rearrange the tests so that we quickly eliminate structures that cannot match. Return an encoded value of element count and base type. commit 0d39b4bb692b1a8e4f3354badcd9e123f8276bd4 Author: Richard Henderson Date: Fri Oct 17 01:02:52 2014 -0400 arm: Deref ffi_put_arg arguments commit 57b24fb3f5471262ce57e1a912774d6e7de5ada7 Author: Richard Henderson Date: Fri Oct 17 00:53:21 2014 -0400 arm: Deref ffi_align argument commit c129bea82a25c9b0f3af3bf48c35c55257092f87 Author: Richard Henderson Date: Wed Oct 15 17:28:53 2014 -0400 arm: Reindent arm/ffi.c commit 9761b7bb70b4c47dc11dc74830964a0b3b3176d3 Author: Richard Henderson Date: Fri Oct 17 21:26:52 2014 -0700 alpha: Add support for Go closures commit f41bec3b576aa5ff8915b1188446c2dc086dfe64 Author: Richard Henderson Date: Fri Oct 17 20:46:48 2014 -0700 alpha: Add support for complex types commit 5f917371afcb4d3363fd33b547844c47264f6d4e Author: Richard Henderson Date: Fri Oct 17 13:47:26 2014 -0700 alpha: Clean up conversion of float values Don't use "real" conversion to double, lest we raise exceptions when passing signalling nans. commit 32a26b75ba92665fa093b539318e83d916e3effb Author: Richard Henderson Date: Fri Oct 17 12:33:53 2014 -0700 alpha: Reorganize cif flags Unties the backend from changes to FFI_TYPE_* constants, and allows compilation to succeed after the addition of FFI_TYPE_COMPLEX. Delete the hand-written unwind info. commit c6352b664c31835b9527d747f8a29b4754cabb0e Author: Richard Henderson Date: Thu Oct 23 00:26:14 2014 -0400 aarch64: Add support for Go closures commit 0e41c73b092064e986d484270a13046479d6bda3 Author: Richard Henderson Date: Wed Oct 22 23:48:12 2014 -0400 aarch64: Move x8 out of call_context Reduces stack size. It was only used by the closure, and there are available argument registers. commit a992f8789b5202ca180e67b16b47ee834160665c Author: Richard Henderson Date: Wed Oct 22 22:58:09 2014 -0400 aarch64: Add support for complex types commit 658b2b56001e293848ea22b70068bd7f8daa7205 Author: Richard Henderson Date: Wed Oct 22 22:36:07 2014 -0400 aarch64: Remove aarch64_flags This field was useless from the start, since the normal flags field is available for backend use. commit 4a3cbcaa4f1d3834e0b74c9a841c3f51c5e0454c Author: Richard Henderson Date: Wed Oct 22 22:32:13 2014 -0400 aarch64: Unify scalar fp and hfa handling Since an HFA of a single element is exactly the same as scalar, this tidies things up a bit. commit 12cf89ee04dc01650f71bc38ce414c953c83199a Author: Richard Henderson Date: Wed Oct 22 21:53:30 2014 -0400 aarch64: Move return value handling into ffi_closure_SYSV As with the change to ffi_call_SYSV, this avoids copying data into a temporary buffer. commit 4fe1aea1211aaf228c2db867b8ac6f1620de72d4 Author: Richard Henderson Date: Wed Oct 22 17:06:19 2014 -0400 aarch64: Move return value handling into ffi_call_SYSV This lets us pass return data directly to the caller of ffi_call in most cases, rather than storing it into temporary storage first. commit 325471ea6a7bf954943485458a1bd391635dfaa8 Author: Richard Henderson Date: Wed Oct 22 13:58:59 2014 -0400 aarch64: Merge prep_args with ffi_call Use the trick to allocate the stack frame for ffi_call_SYSV within ffi_call itself. commit 8c8161cb623585d5d0c783b9d494b9b74ada6ced Author: Richard Henderson Date: Wed Oct 22 12:52:07 2014 -0400 aarch64: Tidy up abi manipulation Avoid false abstraction, like get_x_addr. Avoid recomputing data about the type being manipulated. Use NEON insns for HFA manipulation. Note that some of the inline assembly will go away in a subsequent patch. commit b55e03665ddf2423df9baee0d3172892ba781c26 Author: Richard Henderson Date: Wed Oct 22 12:33:59 2014 -0400 aarch64: Treat void return as not passed in registers This lets us do less post-processing when there's no return value. commit 2e32f9bf6ffa6782b70fafe96bdf21c5aa44f534 Author: Richard Henderson Date: Tue Oct 21 22:49:05 2014 -0400 aarch64: Use correct return registers There are fewer return registers than argument registers. commit 95a04af134431ccc8230aca1641541a5e8fcbdc9 Author: Richard Henderson Date: Tue Oct 21 22:41:07 2014 -0400 aarch64: Reduce the size of register_context We don't need to store 32 general and vector registers. Only 8 of each are used for parameter passing. commit 77c4cddca6aeb6e545e21f235e29323e05f5a3a3 Author: Richard Henderson Date: Tue Oct 21 13:30:40 2014 -0400 aarch64: Simplify AARCH64_STACK_ALIGN The iOS abi doesn't require padding between arguments, but that's not what AARCH64_STACK_ALIGN meant. The hardware will in fact trap if the SP register is not 16 byte aligned. commit b5f147d84761dc673ffe01d6af82bcde4ea47928 Author: Richard Henderson Date: Tue Oct 21 13:27:57 2014 -0400 aarch64: Always distinguish LONGDOUBLE Avoid if-deffery by forcing FFI_TYPE_LONGDOUBLE different from FFI_TYPE_DOUBLE. This will simply be unused on hosts that define them identically. commit 38b54b9c180af13a3371e70a151a1a97e105b03f Author: Richard Henderson Date: Tue Oct 21 13:17:39 2014 -0400 aarch64: Improve is_hfa The set of functions get_homogeneous_type, element_count, and is_hfa are all intertwined and recompute data. Return a compound quantity from is_hfa that contains all the data and avoids the recomputation. commit 18b74ce54afab45fcf2a7d4eb86bb2ce9db8cec8 Author: Richard Henderson Date: Tue Oct 21 13:00:34 2014 -0400 aarch64: Fix non-apple compilation commit 2650f47fe8388f3281ec00e2d1e3bc777c1d3544 Author: Richard Henderson Date: Thu Nov 6 10:57:04 2014 -0800 x86: Use win32 name mangling for fastcall functions commit f8c64e2486c956109bf5255a0c0a85f24cc86c72 Author: Richard Henderson Date: Wed Nov 5 17:04:29 2014 +0100 x86: Add support for Go closures commit 198f469e91076f84ab6ac38055ad12a5490101cd Author: Richard Henderson Date: Wed Nov 5 16:34:41 2014 +0100 x86: Add support for Complex commit b21ec1ce783f09335c0e6c6f96ccc7e89da5a1b4 Author: Richard Henderson Date: Wed Nov 5 10:15:25 2014 +0100 x86: Rewrite closures Move everything into sysv.S, removing win32.S and freebsd.S. Handle all abis with a single ffi_closure_inner function. Move complexity of the raw THISCALL trampoline into assembly instead of the trampoline itself. Only push the context for the REGISTER abi; let the rest receive it in a register. commit b9ac94f3af9b1ca1853850911b7d3bc058296c71 Author: Richard Henderson Date: Sat Nov 1 15:10:34 2014 -0700 x86: Rewrite ffi_call Decouple the assembly from FFI_TYPE_*. Merge prep_args with ffi_call, passing the frame and the stack to the assembly. Note that this patch isn't really standalone, as this breaks closures. commit 159d3788eb52b89ed6bc5dd6ebb85c02cd150232 Author: Richard Henderson Date: Fri Oct 31 12:07:02 2014 -0700 x86: Convert to gas generated unwind info commit e7b0056d60f29675072454a575eed98053fc63d6 Author: Richard Henderson Date: Thu Oct 30 13:57:39 2014 -0700 x86: Force FFI_TYPE_LONGDOUBLE different from FFI_TYPE_DOUBLE There are few abis that set double = long double. Eliminate the conditional compilation and let this code simply be unused there. commit 4b2fad8fd583f1d9c36db0044c4d46316207aa49 Author: Richard Henderson Date: Thu Oct 30 12:41:31 2014 -0700 x86: Remove some conditional compilation Removal of ifdefs made possible to due to ffi_abi unification. commit ef76205647bca77796882d31f6ab5e889f461f07 Author: Richard Henderson Date: Thu Oct 30 12:13:31 2014 -0700 x86: Tidy ffi_abi The x86_64 unix port only handles one ABI; don't define all of the other symbols. The UNIX64 symbol retains the same value. The i386 ports ought to have the same symbols, even if we can't yet unify the values without incrementing the libffi soname. commit 7cf841324f3d15c46833e0f2ddf8a98c4b24711e Author: Richard Henderson Date: Fri Nov 7 06:56:55 2014 -0800 testsuite: Add two dg-do run markers Caught by clang warning about unused -L parameter. commit fc501750f41e2ee7f4268f490f590ce97a25a818 Author: Richard Henderson Date: Wed Nov 5 16:33:44 2014 +0100 testsuite: Fix return_complex2 vs excessive precision Use the previously computed rc2 to validate, rather than recomputing a floating point result with excess precision. commit 610c90bf7131de70318dc981a529a63ae36981b8 Author: Richard Henderson Date: Tue Oct 28 11:21:50 2014 -0700 x86_64: Add support for complex types commit 32c56831636a489a1c5084e42ffbd935902b1033 Author: Richard Henderson Date: Tue Oct 28 11:17:35 2014 -0700 x86_64: Decouple return types from FFI_TYPE constants We can better support structure returns, and as prep for complex types. commit 2e9dc1655600eb8928ebab13fa246187b6205109 Author: Richard Henderson Date: Mon Oct 27 13:41:39 2014 -0700 x86_64: Fixups for x32 commit ebd82769de5fc909ec9678aafca674ece394d173 Author: Richard Henderson Date: Thu Oct 23 23:57:06 2014 -0700 win64: Remove support from ffi.c commit 99db4d42acc7ee6d2ad206848be38083d7b5e094 Author: Richard Henderson Date: Thu Oct 23 14:12:18 2014 -0700 win64: Rewrite It's way too different from the 32-bit ABIs with which it is currently associated. As seen from all of the existing XFAILs. commit 6b62fb4a26d7363548a92f1a156562cfa648726e Author: Richard Henderson Date: Fri Oct 17 11:11:58 2014 -0700 x86-64: Support go closures Dumps all of the hand-coded unwind info for gas generated. Move jump table data into .rodata. Adjust ffi_call_unix64 to load the static chain. Split out sse portions of ffi_closure_unix64 to ffi_closure_unix64_sse rather than test cif->flags at runtime. commit 89bbde8b4e802625bdbf830ea14e938beea79835 Author: Richard Henderson Date: Fri Oct 17 10:55:11 2014 -0700 Add ffi_cfi.h Have one copy of the HAVE_AS_CFI_PSEUDO_OP code to share between all backends. commit e951d64c085267a30036907d02f94a70dee4c9d9 Author: Richard Henderson Date: Fri Oct 17 10:12:25 2014 -0700 Add entry points for interacting with Go A "ffi_go_closure" is intended to be compatible with the function descriptors used by Go, and ffi_call_go sets up the static chain parameter for calling a Go function. The entry points are disabled when a backend has not been updated, much like we do for "normal" closures. commit a0bdc5250cd36f1d2578358e33299fd2408091ac Author: Anthony Green Date: Tue Nov 11 09:43:01 2014 -0500 Fix typo commit 67c0c0705318d2e567fd48e680d63dca3619a235 Author: Anthony Green Date: Tue Nov 11 07:30:49 2014 -0500 Final 3.2 changes commit ee826b933bbdeb68ee81876ffb508c3908eed371 Author: Anthony Green Date: Tue Nov 11 07:29:26 2014 -0500 Final 3.2 changes commit c81a705d53be97899d314953987936213244b1ba Author: Anthony Green Date: Tue Nov 4 13:29:58 2014 -0500 Mention OpenRISC commit 6a17ad540d7ddb7a812687f4380bd0ef2d2865bb Merge: 990eb9d4c 0f316ab7c Author: Anthony Green Date: Sun Sep 28 00:54:00 2014 -0400 Merge pull request #134 from s-macke/openrisc Add OpenRISC support commit 990eb9d4c244324771a7d878a168bb8eba1850c0 Author: Anthony Green Date: Sun Sep 28 00:50:29 2014 -0400 Only run the complex type tests on supported platforms. commit 0f316ab7c11b8315a838a6ae4645d36ff2c46f4c Author: Sebastian Macke Date: Sat Sep 27 00:19:15 2014 +0000 Add OpenRISC support This patch adds support for the OpenRISC architecture. (http://opencores.org/or1k/Main_Page) This patch has been tested under Linux with QEMU-user emulation support. - 32 Bit - big endian - delayed instructions This is the only available configuration under Linux. The description of the ABI can be found on the official website. Is passes the testsuite except of the unwindtest_ffi_call.cc testcase, which seems to be a problem of gcc and not libffi. Some testcases of the gcc testsuite still fail. Signed-off-by: Sebastian Macke commit 6695983d2f0857caab220862de12f3f94a55bcf6 Author: Anthony Green Date: Sat Sep 20 07:44:37 2014 -0400 Add complex type support. Mostly broken right now commit fbbf48fbdf9b1491af0aefc00523a1b5d8bb9349 Author: Anthony Green Date: Sat Sep 20 07:43:51 2014 -0400 Compile tests with -Wno-psabi when using GCC commit ae8d2554c5f192be54dadc9f99a1449b4df1b222 Author: Anthony Green Date: Sat Sep 20 07:08:37 2014 -0400 Update version to 3.2 commit b2ca59d39d2192021316561ea9dda3d35d8d35d9 Author: Anthony Green Date: Sat Sep 20 07:08:05 2014 -0400 More README updates for 3.2 commit f920a01ad8eafc79daaf331499ef9a97696c46a7 Author: Anthony Green Date: Sat Sep 20 06:54:06 2014 -0400 Update release notes. commit bfcbf329c5e4bebdb32134d4e53260e17e4e66cc Author: Bernd Edlinger Date: Sat Sep 20 06:51:45 2014 -0400 2014-05-11 Bernd Edlinger Fix current cygwin-64 build problems. * src/java_raw_api.c: Remove if !defined(FFI_NO_RAW_API). * src/x86/ffi.c: Add if defined(__CYGWIN__). * src/x86/win64.S (ffi_closure_win64, ffi_call_win64): Added handling for FFI_TYPE_UINT64, FFI_TYPE_POINTER and FFI_TYPE_INT. Added SEH information. Fixed formatting. commit 32cb2ce81db9457c6bd43a2587d7fa23a2edb9b1 Author: Jakub Jelinek Date: Sat Sep 20 06:39:55 2014 -0400 2014-09-10 Jakub Jelinek * src/powerpc/linux64.S: Emit .note.GNU-stack even when POWERPC64 is not defined. * src/powerpc/linux64_closure.S: Likewise. Also test _CALL_ELF == 2. commit aaf3101ba81af8f488502881648e3f687721671e Author: Matthias Klose Date: Sat Sep 20 06:37:04 2014 -0400 Fix -Werror=declaration-after-statement problem commit 6e8a4460833594d5af1b4539178025da0077df19 Author: Dominik Vogt Date: Sat Sep 20 06:21:19 2014 -0400 2014-07-22 Dominik Vogt * src/types.c (FFI_TYPEDEF, FFI_NONCONST_TYPEDEF): Merge the macros by adding another argument that controls whether the result is const or not (FFI_LDBL_CONST): Temporary macro to reduce ifdef confusion * src/prep_cif.c (ffi_prep_cif_core): Replace list of systems with new macro FFI_TARGET_SPECIFIC_STACK_SPACE_ALLOCATION * src/pa/ffitarget.h (FFI_TARGET_SPECIFIC_STACK_SPACE_ALLOCATION): Define. * src/s390/ffitarget.h (FFI_TARGET_SPECIFIC_STACK_SPACE_ALLOCATION): Define. * src/x86/ffitarget.h (FFI_TARGET_SPECIFIC_STACK_SPACE_ALLOCATION): Define. 2014-07-22 Dominik Vogt * doc/libffi.texi (Primitive Types): Document ffi_type_complex_float, ffi_type_complex_double and ffi_type_complex_longdouble (Complex Types): New subsection. (Complex Type Example): Ditto. * testsuite/libffi.call/cls_align_complex_double.c: New FFI_TYPE_COMPLEX test. * testsuite/libffi.call/cls_align_complex_float.c: Ditto. * testsuite/libffi.call/cls_align_complex_longdouble.c: Ditto. * testsuite/libffi.call/cls_complex_double.c: Ditto. * testsuite/libffi.call/cls_complex_float.c: Ditto. * testsuite/libffi.call/cls_complex_longdouble.c: Ditto. * testsuite/libffi.call/cls_complex_struct_double.c: Ditto. * testsuite/libffi.call/cls_complex_struct_float.c: Ditto. * testsuite/libffi.call/cls_complex_struct_longdouble.c: Ditto. * testsuite/libffi.call/cls_complex_va_double.c: Ditto. * testsuite/libffi.call/cls_complex_va_float.c: Ditto. * testsuite/libffi.call/cls_complex_va_longdouble.c: Ditto. * testsuite/libffi.call/complex_double.c: Ditto. * testsuite/libffi.call/complex_defs_double.c: Ditto. * testsuite/libffi.call/complex_float.c: Ditto. * testsuite/libffi.call/complex_defs_float.c: Ditto. * testsuite/libffi.call/complex_longdouble.c: Ditto. * testsuite/libffi.call/complex_defs_longdouble.c: Ditto. * testsuite/libffi.call/complex_int.c: Ditto. * testsuite/libffi.call/many_complex_double.c: Ditto. * testsuite/libffi.call/many_complex_float.c: Ditto. * testsuite/libffi.call/many_complex_longdouble.c: Ditto. * testsuite/libffi.call/return_complex1_double.c: Ditto. * testsuite/libffi.call/return_complex1_float.c: Ditto. * testsuite/libffi.call/return_complex1_longdouble.c: Ditto. * testsuite/libffi.call/return_complex2_double.c: Ditto. * testsuite/libffi.call/return_complex2_float.c: Ditto. * testsuite/libffi.call/return_complex2_longdouble.c: Ditto. * testsuite/libffi.call/return_complex_double.c: Ditto. * testsuite/libffi.call/return_complex_float.c: Ditto. * testsuite/libffi.call/return_complex_longdouble.c: Ditto. * src/raw_api.c (ffi_raw_to_ptrarray): Handle FFI_TYPE_COMPLEX (ffi_ptrarray_to_raw): Ditto. * src/prep_cif.c (ffi_prep_cif_core): Abort if FFI_TYPE_COMPLEX is not implemented in libffi for the target. * src/java_raw_api.c (ffi_java_raw_size): FFI_TYPE_COMPLEX not supported yet (abort). (ffi_java_raw_to_ptrarray): Ditto. (ffi_java_rvalue_to_raw): Ditto. (ffi_java_raw_to_rvalue): Ditto. * src/debug.c (ffi_type_test): Add debug tests for complex types. * include/ffi.h.in (FFI_TYPE_COMPLEX): Add new FFI_TYPE_COMPLEX. (FFI_TYPE_LAST): Bump. (ffi_type_complex_float): Add new ffi_type_.... (ffi_type_complex_double): Ditto. (ffi_type_complex_longdouble): Ditto. 2014-07-22 Dominik Vogt * src/s390/ffitarget.h (FFI_TARGET_HAS_COMPLEX_TYPE): Define to provide FFI_TYPE_COMPLEX support. * src/s390/ffi.c (ffi_check_struct_type): Implement FFI_TYPE_COMPLEX (ffi_prep_args): Ditto. (ffi_prep_cif_machdep): Ditto. (ffi_closure_helper_SYSV): Ditto. commit 4c5c4088aa3e4d8103ff9ca441937da64fdd849a Merge: 862f53de0 5d6340ef2 Author: Anthony Green Date: Thu Sep 18 19:10:54 2014 -0400 Merge pull request #132 from nielsAD/master Pascal and Register calling convention support on x86 commit 862f53de04ff898a128be7eec592bae60930d6df Merge: 5df384077 aebf2c302 Author: Anthony Green Date: Thu Sep 18 19:06:08 2014 -0400 Merge pull request #130 from frida/fix/darwin-aarch64-float-alignment Fix alignment of FFI_TYPE_FLOAT for Apple's ARM64 ABI commit 5df384077f3633ec8cf2e08f9199e5a44994dbad Merge: 7b7a5284d 0f4e09d27 Author: Anthony Green Date: Thu Sep 18 19:05:35 2014 -0400 Merge pull request #129 from frida/fix/darwin-aarch64-cif-prep Fix non-variadic CIF initialization for Apple/ARM64 commit 7b7a5284de7a99edfcee5ac362dbf76251698e1e Merge: 4cd5e8401 6172a995e Author: Anthony Green Date: Thu Sep 18 18:28:05 2014 -0400 Merge pull request #124 from knuesel/master Fix issue with builddir when calling configure with absolute path commit 4cd5e840122e22636b5838943372c4ed74618e35 Merge: 1cee07be4 08853e32f Author: Anthony Green Date: Thu Sep 18 18:27:15 2014 -0400 Merge pull request #123 from ehsan/clang-cl Add support for building with clang-cl commit 5d6340ef2cf81432da79ac8b0b1b1218ab391438 Author: nielsAD Date: Mon Aug 25 17:29:44 2014 +0200 Determine whether register arguments (THISCALL/FASTCALL/REGISTER) are really passed via register to closures. Use stack if not. commit 6e346487b879b4b056a847268e381ae6efec4c21 Author: nielsAD Date: Mon Aug 25 12:23:29 2014 +0200 Fixed THISCALL/FASTCALL closures and added basic support for PASCAL/REGISTER closures. commit 098dca6b31e42dd0256ed7d966729e31961f1461 Author: nielsAD Date: Sat Aug 23 00:18:47 2014 +0200 Support for calling functions with PASCAL and REGISTER calling conventions on x86 Windows/Linux. Also changed indentation to be more consistent throughout the (adjusted) files. commit 08853e32fb6b750cda664ee6d3290d340429dd4d Author: Ehsan Akhgari Date: Thu Jul 31 20:33:33 2014 -0400 Enable forcing the usage of the static CRT in libffi's msvc wrapper This is required for AddressSanitizer builds with clang-cl. commit aebf2c3023b5ff056a708dc34169f7b549b2ec4f Author: Ole André Vadla Ravnås Date: Fri Jul 25 21:40:50 2014 +0200 Fix alignment of FFI_TYPE_FLOAT for Apple's ARM64 ABI commit 0f4e09d27dc175ce3e95d070b794351663c7220e Author: Ole André Vadla Ravnås Date: Sat Jul 26 00:11:06 2014 +0200 Fix non-variadic CIF initialization for Apple/ARM64 Turns out `aarch64_nfixedargs` wasn't initialized in the non-variadic case, resulting in undefined behavior when allocating arguments. commit 6172a995e14bdbc6150aa02a361f710626a5c293 Author: Jeremie Knuesel Date: Wed Jun 25 12:28:17 2014 +0200 Fix issue with builddir when calling configure with absolute path commit fb25cd08ed5a87640b02f0feeb10a09b37cfddbe Author: Ehsan Akhgari Date: Wed Jun 11 12:07:24 2014 -0400 Add support for building with clang-cl commit 1cee07be4712411baa5eb8af22698bc8443a6d6e Author: Anthony Green Date: Thu Jun 12 06:30:59 2014 -0400 Remove compiler warning commit ad0d1d239afc7ad09203847793dcc020856d1035 Author: Samuli Suominen Date: Thu Jun 12 06:30:21 2014 -0400 Fix paths in libffi.pc.in commit df31a85103b0cc232bbc340d7e782a3853c7fad5 Merge: 69289c10b 360aa8ca1 Author: Anthony Green Date: Fri Jun 6 08:24:35 2014 -0400 Merge pull request #122 from rvandermeulen/1014976 Don't make --enable-debug imply using the debug CRT in libffi commit 360aa8ca1ead2fdaefa7b1c6f01ad001d8abae7e Author: Mike Hommey Date: Mon Jun 2 09:21:10 2014 -0400 Bug 1014976 - Don't make --enable-debug imply using the debug CRT in libffi. commit 69289c10b0acb4b52c8b5df4cf738193c618db40 Author: Anthony Green Date: Sat May 31 08:54:08 2014 -0400 Prepare for libffi 3.1.1 commit 978c9540154d320525488db1b7049277122f736d Author: Samuli Suominen Date: Sat May 31 08:53:10 2014 -0400 Add missing GNU stack markings in win32.S commit cbc5a3c0200aa6c2cf8d6798d69a21501557c83f Author: Ryan Hill Date: Sat May 31 08:26:34 2014 -0400 Fix typo commit 9d8e3018ac3d0d338373bff7eefa9f006746f4ec Author: Anthony Green Date: Sat May 24 23:28:42 2014 -0400 Update current version. commit 18d56cca5a4ee95ad48b129100adf26231518ed0 Author: Anthony Green Date: Mon May 19 18:05:33 2014 -0400 Increment libtool version number commit 629f1029c47e522e4331988f02f32c203a070e28 Merge: 0403f332b c1166d4f6 Author: Anthony Green Date: Mon May 19 18:04:28 2014 -0400 Merge pull request #120 from l0kod/tmpfile Create temporary file with O_TMPFILE and O_CLOEXEC when available commit c1166d4f653bf6d17ad8c265cf5e8da84e2e76dc Author: Mickaël Salaün Date: Mon May 12 19:44:08 2014 +0200 closures: Check for mkostemp(3) commit 8daeed9570af72eb135c8ded460d2888f05b2e68 Author: Mickaël Salaün Date: Sun May 11 22:54:58 2014 +0200 closures: Create temporary file with O_TMPFILE and O_CLOEXEC when available The open_temp_exec_file_dir function can create a temporary file without file system accessible link. If the O_TMPFILE flag is not defined (old Linux kernel or libc) the behavior is unchanged. The open_temp_exec_file_name function now need a new argument "flags" (like O_CLOEXEC) used for temporary file creation. The O_TMPFILE flag allow temporary file creation without race condition. This feature/fix prevent another process to access the (future) executable file from the file system. The O_CLOEXEC flag automatically close the temporary file for any execve. This avoid transmitting (executable) file descriptor to a child process. commit 0403f332b1f478696c30d3d8a0e2f6eef24aaf88 Author: Anthony Green Date: Mon May 19 09:41:32 2014 -0400 Update date. Annoucing 3.1 today. commit 94ac0c168ee7b115409121d88b25a4979446c8da Author: Anthony Green Date: Mon May 19 09:37:21 2014 -0400 Increment libtool library revision number commit 57465744b6e1295d7202de5a7734df589518f1c8 Author: Anthony Green Date: Sun May 11 10:30:22 2014 -0400 Update to version 3.1 commit 0c2251a42df5108b6d9ebe5fe1cf83d0bcdf660e Author: Anthony Green Date: Sun May 11 10:22:30 2014 -0400 Support versions of git older than 1.8.5 commit 70c303cb88e23aaee91c87c56b108c50ab4f3c2f Author: Anthony Green Date: Sun May 11 09:56:40 2014 -0400 Fix testsuite for GCC 4.9.0 commit 52b3457093ed19b2a7c5fcf243c4014c90ce6225 Author: Magnus Granberg Date: Sun May 11 09:55:28 2014 -0400 Check /proc/self/status for PaX status. commit 7ba4c5d72aa440a4b21fb57e999e67c5957761da Author: Dominik Vogt Date: Sun May 11 09:52:47 2014 -0400 Use to get correct dir commit 31e0d4ecff6dc2a6c75a066ee099b52a43f6ba27 Merge: 1c0e9a729 99909eb61 Author: Anthony Green Date: Wed Apr 23 19:24:47 2014 -0400 Merge pull request #119 from joshtriplett/fastcall-fastball src/x86/win32.S: Define ffi_closure_FASTCALL in the MASM section, too commit 99909eb6184b62408d88b6b4e7ab38e84e6d0bf3 Author: Josh Triplett Date: Tue Apr 22 21:17:52 2014 -0700 src/x86/win32.S: Define ffi_closure_FASTCALL in the MASM section, too commit 1c0e9a7297ced15413c2d2d5d35f6c650c4b46c9 Merge: 93a24f216 d36952273 Author: Anthony Green Date: Mon Apr 21 12:41:56 2014 -0400 Merge pull request #101 from joshtriplett/fastcall-closures Support closures for fastcall commit d36952273d4fafbda91ecc205fc0824f7cc65e70 Author: Josh Triplett Date: Sun Apr 20 12:03:25 2014 -0700 Support fastcall closures libffi on 32-bit x86 now supports closures for all supported ABIs. Thus, rewrite the last remaining duplicated-by-ABI test (closure_stdcall and closure_thiscall) to use the generic ABI_NUM/ABI_ATTR mechanism. commit 93a24f216bcdd1018b976d697179c6d49004015a Merge: dd11a0406 2349fec9a Author: Anthony Green Date: Sat Apr 12 19:38:07 2014 -0400 Merge pull request #80 from ueno/devel Fix typo in doc commit dd11a04061cb49ce1d702545693c24eb1267d648 Merge: 8fa281235 03ca88008 Author: Anthony Green Date: Sat Apr 12 19:37:21 2014 -0400 Merge pull request #86 from joshtriplett/testsuite-CC-CXX testsuite ignores CC parameter supplied to configure or make commit 8fa2812355e685a42abf9a62fbc674d616b2edee Merge: 8a58e6b78 419503f40 Author: Anthony Green Date: Sat Apr 12 19:32:08 2014 -0400 Merge pull request #116 from frida/fix/darwin-aarch64-variadic Fix handling of variadic calls on Darwin/AArch64 commit 8a58e6b7805b736def197b8baf8e465a2a3f6913 Merge: 30b77c56f a539f7ffd Author: Anthony Green Date: Sat Apr 12 19:30:18 2014 -0400 Merge pull request #115 from frida/fix/darwin-aarch64-alignment Fix alignment of AArch64 assembler functions commit 30b77c56f95c63ecd83399aafdbad7b07330f2fd Merge: dc33cb3c9 3e2b84d29 Author: Anthony Green Date: Sat Apr 12 19:29:13 2014 -0400 Merge pull request #117 from frida/fix/windows-regression Fix Windows regression commit 3e2b84d295531720917bf46afc532fc6d877e3ec Author: Ole André Vadla Ravnås Date: Sat Apr 12 01:04:04 2014 +0200 Fix Windows regression Introduced by b5fed601948237037513a9b7f967c8fc6c9ff1f6. commit 419503f409c321fe31ff59d963ef34bb913420d0 Author: Ole André Vadla Ravnås Date: Sun Apr 6 20:54:13 2014 +0200 Fix handling of variadic calls on Darwin/AArch64 commit a539f7ffd6783aa11353d13265520e453c565fb4 Author: Ole André Vadla Ravnås Date: Sun Apr 6 20:53:02 2014 +0200 Fix alignment of AArch64 assembler functions commit dc33cb3c998da521a960385c1269c3aef552f69f Merge: c860a992f b5fed6019 Author: Anthony Green Date: Sat Apr 5 23:41:22 2014 -0400 Merge pull request #114 from joshtriplett/bounce-on-a-tiny-trampoline Fix ABI on 32-bit non-Windows x86: go back to trampoline size 10 commit b5fed601948237037513a9b7f967c8fc6c9ff1f6 Author: Josh Triplett Date: Sat Apr 5 17:33:42 2014 -0700 Fix ABI on 32-bit non-Windows x86: go back to trampoline size 10 The trampoline size is part of the ABI, so it cannot change. Move the logic from the stdcall and thiscall trampolines to the functions they call, to reduce them both to 10 bytes. This drops the previously added support for raw THISCALL closures on non-Windows. (Non-raw THISCALL closures still work.) commit 03ca880081b22efab09ba72268270f83017d3d7b Author: Josh Triplett Date: Thu Mar 27 08:44:34 2014 -0700 README: Note the testsuite changes to respect $CC and $CXX commit d74df8c5d8c6722ecb908da98c86cc8e2c755b84 Author: Josh Triplett Date: Thu Mar 27 00:44:12 2014 -0700 README: Update Windows example to set both CC and CXX commit 7d698125b1f05173f3656a89755a2eb58813b002 Author: Josh Triplett Date: Wed Mar 26 23:17:56 2014 -0700 Use the proper C++ compiler to run C++ tests Running the C compiler with -shared-libgcc -lstdc++ does not work on non-GCC compilers. commit fa5e88f170cb37c7b2b9bb015c8c5b854ffd8a3e Author: Josh Triplett Date: Wed Mar 26 23:53:57 2014 -0700 .travis.yml: Make the build command more readable by splitting at && "script" can contain multiple commands to run in sequence. commit 0c3824702d3d59d37f8c177d646303f546187683 Author: Josh Triplett Date: Wed Mar 26 14:51:32 2014 -0700 Always set CC_FOR_TARGET for dejagnu, to make the testsuite respect $CC This fixes cross-compilation and compilation with CC="gcc -m32". commit 9946a92af31b30cb7760150d1f8ca6c11b01aeea Author: Josh Triplett Date: Wed Mar 26 20:18:58 2014 -0700 Stop looking for expect and runtest above top_builddir Users wishing to test hand-compiled versions of expect and runtest can easily enough put them in their path or set EXPECT and RUNTEST themselves. commit acb202325215058639234efb7af1f04c1c8a1f44 Author: Josh Triplett Date: Wed Mar 26 20:18:41 2014 -0700 Stop setting an empty AM_RUNTESTFLAGS commit c860a992fef5d7cd7bb0975b1632d17a9fafe007 Author: Anthony Green Date: Tue Mar 25 17:02:51 2014 -0400 Upgrade version to 3.1-rc1 commit 9837073e6203048a162a226798c5d252600219ed Author: Anthony Green Date: Tue Mar 25 16:24:14 2014 -0400 Update copyright date and clean up README notes. commit 18d3baa9f597b026675baa1b4e5a5eeef7577a08 Merge: afee53738 f0c8a3157 Author: Anthony Green Date: Tue Mar 25 16:12:53 2014 -0400 Merge pull request #108 from joshtriplett/freebsd [3.1 blocker] Fix FreeBSD support commit afee53738a995e23bd2f89fd0f7b30b380566106 Merge: 7d2478568 b2d610e02 Author: Anthony Green Date: Tue Mar 25 16:12:35 2014 -0400 Merge pull request #106 from joshtriplett/darwin-award [3.1 blocker] Update OS X build system to include win32.S on 32-bit commit 7d2478568ed9f03cbf57627f449a2d2cf4d1571c Merge: beab5f334 56be47f87 Author: Anthony Green Date: Tue Mar 25 16:12:17 2014 -0400 Merge pull request #110 from joshtriplett/w64 Fix 64-bit Windows support commit beab5f334d9ec5b8b91d1cc727d1029b40358e7e Merge: 28fb19707 ef5890eba Author: Anthony Green Date: Tue Mar 25 16:07:47 2014 -0400 Merge pull request #105 from joshtriplett/win32-relocations [3.1 blocker] win32.S needs to handle relocations/GOT commit f0c8a31577172104049283f0a80c723084a5bd77 Author: Josh Triplett Date: Mon Mar 24 22:14:26 2014 -0700 Compile win32.S on FreeBSD commit b2d610e028b5ce48d1ad7e5d0debc9c321d891b2 Author: Josh Triplett Date: Fri Mar 21 11:10:13 2014 -0700 Compile win32.S on 32-bit Darwin as well commit be50b87a490e794362cb4a27ada2fbaab202adb8 Author: Josh Triplett Date: Mon Mar 24 21:44:13 2014 -0700 Always use configure to detect whether global symbols need underscores 64-bit Windows already used this check; make it universal, and use it in place of an ifdef on X86_WIN32, to handle non-Windows platforms that use the underscore, such as Darwin. commit 56be47f87629e31afbcb0774aa65735f539ee972 Author: Josh Triplett Date: Mon Mar 24 21:24:53 2014 -0700 Fix a warning on 64-bit Windows When sizeof(size_t) != sizeof(unsigned), adding a size_t to cif->bytes produces a "possible loss of data" warning. However, the size_t in question refers to the size of a single parameter. Use a cast to avoid the warning. commit 48a8eda74aad8a21b6f26df5df08fe64c043d208 Author: Josh Triplett Date: Mon Mar 24 21:21:12 2014 -0700 Avoid referencing undefined ABIs on 64-bit Windows builds 64-bit Windows does not have FFI_STDCALL, FFI_THISCALL, or FFI_FASTCALL. commit f0f4138f90345d7d67dfa6783a7e1c7cc30d3c6f Author: Josh Triplett Date: Sat Mar 22 10:00:53 2014 -0700 win32.S: Add handling for position-independent code on Darwin Newer versions of Darwin generate the necessary stub functions automatically and just need a call instruction, but accomodating older versions as well requires adding the stub. commit ef5890ebafb7cd2fbf9acf161edb55fe1382871c Author: Josh Triplett Date: Fri Mar 21 11:01:39 2014 -0700 win32.S: Use shifting for multiplication rather than repeated addition The jump table code added a register to itself twice to multiply by 4; shift the register left by 2 instead. commit 4fca48901e7e4f53bf490ed22607b2d2d8f4bfcc Author: Josh Triplett Date: Fri Mar 21 11:00:41 2014 -0700 win32.S: Make the jump tables position-independent Now that non-Windows platforms include win32.S, it needs to support building as position-independent code. This fixes build failures on target platforms that do not allow text relocations. commit 2087dcf736274286f76c69d3988fb6d7cc4fd0f5 Author: Josh Triplett Date: Fri Mar 21 10:57:06 2014 -0700 win32.S: Make calls to ffi_closure_SYSV_inner position-independent Now that non-Windows platforms include win32.S, it needs to support building as position-independent code. This fixes one source of build failures on target platforms that do not allow text relocations. commit 28fb197079cf1d11da4eef7c8c243ab05590c528 Merge: c697472fc c3dd0a1a0 Author: Anthony Green Date: Tue Mar 18 12:19:36 2014 -0400 Merge pull request #107 from rvandermeulen/msvcc Various compatibility fixes and improvements to msvcc.sh. commit c3dd0a1a0245fc174361a70876e88ae24285f861 Author: Ryan VanderMeulen Date: Tue Mar 18 12:09:45 2014 -0400 Various compatibility fixes and improvements to msvcc.sh. * Don't try to mix incompatible optimization flags in debug builds. * Workaround ax_cc_maxopt.m4 not supporting MSVC and change -O3 to -O2. * Fix MSVC warning by properly passing linker flags to compiler. * Make msvcc.sh return 1 if invalid command line options are used rather than silently eating them. * Add more comments. commit c697472fccfbb5b87b007c053cda9ef014b346b9 Merge: 83fd2bce0 e48918ecf Author: Anthony Green Date: Mon Mar 17 00:32:42 2014 -0400 Merge pull request #102 from joshtriplett/test-generic Add ABIs to the test matrix; unify many bits of the testsuite commit e48918ecf876bc85d040fc50a232059c566553a8 Author: Josh Triplett Date: Sun Mar 16 20:29:27 2014 -0700 testsuite: Add ABIs to the test matrix; unify tests across ABIs This eliminates all the *_win32.c tests in favor of the tests they were branched from, and expands test coverage to run many more tests on stdcall, thiscall, and fastcall. This same mechanism also supports testing any other target that has multiple ABIs. commit 4d4d368e5a55d9443c4c53b1b70d58ab6d8c941c Author: Josh Triplett Date: Sun Mar 16 17:02:05 2014 -0700 testsuite: Replace ffitestcxx.h with ffitest.h ffitest.h contains a superset of the functionality of ffitestcxx.h; make the C++ tests include ffitest.h instead, and remove ffitestcxx.h. commit 3f97cf3413c46caf2a79f32ac9cda4620972c2d7 Author: Josh Triplett Date: Sun Mar 16 16:53:42 2014 -0700 testsuite: Unify the C and C++ testsuites These two testsuites differ only in the source file glob and a couple of additional compiler options; unify the remaining bits. commit 0d9cce8edb937bbe771a6cdd25f671edf06d2128 Author: Josh Triplett Date: Sun Mar 16 16:22:58 2014 -0700 testsuite: ffitest.h: Parenthesize the CHECK macro commit 5695ec1444c5323e48fe4314f8c8f027625e67df Author: Josh Triplett Date: Sun Mar 16 16:04:58 2014 -0700 testsuite: Factor out a function to run a matrix of tests This commons up code from libffi.call/call.exp and libffi.special/special.exp, unifies the optimization option matrix between the two, and makes it easier to add more axes to the matrix in the future. commit dfdb02cc869855d3b68571e5f7aa77ae8c9d254a Author: Josh Triplett Date: Sun Mar 16 15:26:26 2014 -0700 testsuite: Introduce a __THISCALL__ compiler-specific macro commit 83fd2bce0456224483435d4b764063f4513fd464 Merge: 3658a0700 06ff92421 Author: Anthony Green Date: Sun Mar 16 22:03:29 2014 -0400 Merge pull request #99 from joshtriplett/gitignore .gitignore: Ignore more generated files commit 3658a0700a50d37a2fdba04fd9d79ad2f706d9f5 Merge: d948d0a72 46c5d3c30 Author: Anthony Green Date: Sun Mar 16 21:37:42 2014 -0400 Merge pull request #100 from rvandermeulen/bug-756740 Change double quotes in Makefile.am to single quotes. commit 46c5d3c30fdc2b43c076ad955078d7c5f1e75b37 Author: Ryan VanderMeulen Date: Sun Mar 16 21:16:08 2014 -0400 Change double quotes in Makefile.am to single quotes. This was originally done in PR #84, except the change was made to Makefile.in instead of Makefile.am and was therefore reverted the next time the files were regenerated. commit 06ff924215a2f9739efa2c059dc595bc4ec1c851 Author: Josh Triplett Date: Sun Mar 16 16:19:46 2014 -0700 .gitignore: Ignore more generated files The build process generates doc/libffi.info and fficonfig.h.in, so add them to .gitignore. commit bad8948346e9b8813023a0cc78a3b6eb8d9c14c6 Author: Josh Triplett Date: Sun Mar 16 15:16:18 2014 -0700 testsuite: Introduce a __STDCALL__ compiler-specific macro Several tests want to use stdcall, which differs in syntax by compiler, so introduce a macro for it in ffitest.h. commit 98a793fa36a4ab3ba24d059cb80a2891cdb940e1 Author: Josh Triplett Date: Sun Mar 16 15:20:36 2014 -0700 testsuite: Common up the ifdef blocks for compiler-specific macros commit d948d0a729c934b0224749338a3ba0a2c8f51c45 Merge: b61b472bd a86bd318e Author: Anthony Green Date: Sun Mar 16 10:53:48 2014 -0400 Merge pull request #98 from joshtriplett/unconfigure.host Merge configure.host into configure.ac commit a86bd318e2424d879d784ee7b29d6536d7a17c18 Author: Josh Triplett Date: Sun Mar 16 06:58:59 2014 -0700 Merge configure.host into configure.ac configure.host only has a single entry, and shows no signs of needing more added. commit b61b472bd0647006d6685238721002017f1d119c Author: Anthony Green Date: Sun Mar 16 09:45:55 2014 -0400 Update version to 3.1-rc0. Clean up README. commit 7a64e7dbba54e6e9f69954adfb943be1856ff928 Merge: 11a5c5c39 eef2e02a1 Author: Anthony Green Date: Sun Mar 16 09:39:08 2014 -0400 Merge pull request #97 from joshtriplett/remove-more-generated-files Remove more generated files commit 11a5c5c39f5861011f6c5ddf795da3a32b5f0082 Merge: 9a62a21f5 1c68c0721 Author: Anthony Green Date: Sun Mar 16 09:38:47 2014 -0400 Merge pull request #96 from joshtriplett/sawing-changelogs Generate ChangeLog from git in make dist; remove it from version control commit eef2e02a13d7d1c8145d47a64467f654406a3548 Author: Josh Triplett Date: Sun Mar 16 06:26:03 2014 -0700 doc: Remove autogenerated info file and stamp commit 9fb403d3c5d9643e0f561cab6d4a07b1e54907ff Author: Josh Triplett Date: Sun Mar 16 06:25:52 2014 -0700 fficonfig.h.in: Remove, configure generates it commit 1c68c07217fda78a779778c1480fedef7a58d5b4 Author: Josh Triplett Date: Sun Mar 16 06:11:58 2014 -0700 Generate ChangeLog from git in make dist Archive the existing ChangeLog to ChangeLog.libffi-3.1 commit c65ed55e655711e008282edbdd82ce95d008b4f6 Author: Josh Triplett Date: Sun Mar 16 05:52:00 2014 -0700 ChangeLog.v1: Fix typo in explanatory header. commit 9a62a21f5c3a8e1da463229f3170c8ab3031d920 Author: Anthony Green Date: Sun Mar 16 09:03:57 2014 -0400 Add missing ChangeLog entry. Clean up some entries. commit 9bc704c58cb7a049d867837e3a11e2e31886ec66 Merge: 694447aa2 e892e581d Author: Anthony Green Date: Sun Mar 16 08:41:00 2014 -0400 Merge pull request #95 from joshtriplett/news README: Update news for 3.0.14 commit e892e581d1838a06c18c7ecd50ebd79915cff92b Author: Josh Triplett Date: Sun Mar 16 05:38:24 2014 -0700 README: Update news for 3.0.14 commit 694447aa29deadd571efb4e9a26ee3f68ede1493 Merge: fdc87f3b2 45a6c21ef Author: Anthony Green Date: Sun Mar 16 08:32:05 2014 -0400 Merge pull request #93 from joshtriplett/travis-dist Make Travis check "make dist" commit 45a6c21efa944b520842e631dc54919b04884744 Author: Josh Triplett Date: Sun Mar 16 05:29:08 2014 -0700 .travis.yml: Test "make dist" too. commit fdc87f3b2ea37b58a4a9ae6c35083f544909fe3c Merge: 7412b838d e1911f78d Author: Anthony Green Date: Sun Mar 16 08:05:51 2014 -0400 Merge pull request #85 from joshtriplett/stdcall stdcall support on Linux commit e1911f78df113ca58738b66089a070d4cf747de7 Author: Josh Triplett Date: Sun Mar 16 03:25:53 2014 -0700 Add support for stdcall, thiscall, and fastcall on non-Windows x86-32 Linux supports the stdcall calling convention, either via functions explicitly declared with the stdcall attribute, or via code compiled with -mrtd which effectively makes stdcall the default. This introduces FFI_STDCALL, FFI_THISCALL, and FFI_FASTCALL on non-Windows x86-32 platforms, as non-default calling conventions. commit 7412b838d543aae4fa925862bd5702d3dacbc29a Merge: c0cc5fdaa 9531d05f6 Author: Anthony Green Date: Sun Mar 16 07:58:16 2014 -0400 Merge pull request #90 from joshtriplett/win32-unifdef prep_cif.c: Remove unnecessary ifdef for X86_WIN32 commit c0cc5fdaa237b67e86f22d2f6e13f3b42d9aae33 Merge: 98b529604 b3a5da068 Author: Anthony Green Date: Sun Mar 16 07:57:59 2014 -0400 Merge pull request #89 from joshtriplett/travis32 .travis.yml: Test on both 32-bit and 64-bit commit 9531d05f64c2a674e0197158ffad68d69f177bd0 Author: Josh Triplett Date: Sun Mar 16 01:50:02 2014 -0700 prep_cif.c: Remove unnecessary ifdef for X86_WIN32 ffi_prep_cif_core had a special case for X86_WIN32, checking for FFI_THISCALL in addition to the FFI_FIRST_ABI-to-FFI_LAST_ABI range before returning FFI_BAD_ABI. However, on X86_WIN32, FFI_THISCALL already falls in that range, making the special case unnecessary. Remove it. commit b3a5da068abd2f2983d9e67adbf41b0e0f34e37f Author: Josh Triplett Date: Sat Mar 15 23:27:56 2014 -0700 .travis.yml: Test on both 32-bit and 64-bit commit 98b52960485a261399f081915f36063de3854a5f Merge: 134ce4c02 f6dd18454 Author: Anthony Green Date: Sun Mar 16 07:51:33 2014 -0400 Merge pull request #94 from joshtriplett/esp-extra-stackery-perception ChangeLog: Document testsuite changes to remove fragile stack pointer checks commit f6dd1845434dd53e22129becdfa092c082df307c Author: Josh Triplett Date: Sun Mar 16 04:49:36 2014 -0700 ChangeLog: Document testsuite changes to remove fragile stack pointer checks commit 134ce4c0266bf994f07518fc534de53f1d3c8de8 Merge: 2680e9ea9 9c279328e Author: Anthony Green Date: Sun Mar 16 07:47:17 2014 -0400 Merge pull request #91 from joshtriplett/esp-extra-stackery-perception testsuite: Remove fragile stack pointer checks commit 9c279328ee12fc869adff63ca81f1230977bd42b Author: Josh Triplett Date: Sun Mar 16 02:31:19 2014 -0700 testsuite: Remove fragile stack pointer checks testsuite/libffi.call/closure_stdcall.c and testsuite/libffi.call/closure_thiscall.c include inline assembly to save the stack pointer before and after the call, and compare the values. However, compilers can and do leave the stack in different states for these two pieces of inline assembly, such as by saving a temporary value on the stack across the call; observed with gcc -Os, and verified as spurious through careful inspection of disassembly. commit 2680e9ea9b4c87ea8042a61e551bd667493d4bd3 Merge: 071eab32a 82f8cb24a Author: Anthony Green Date: Sun Mar 16 07:44:08 2014 -0400 Merge pull request #88 from joshtriplett/such-precision-many-fail-wow testsuite/libffi.call/many.c: Avoid spurious failure due to excess precision commit 82f8cb24a1d976db35ae31a4b86cec8926da327d Author: Josh Triplett Date: Sun Mar 16 04:27:32 2014 -0700 ChangeLog: Document many.c and many_win32.c changes to avoid spurious failures commit 88d562a8b5912e99306063fe3bc289bab6ca6ebe Author: Josh Triplett Date: Sat Mar 15 22:08:19 2014 -0700 testsuite/libffi.call/many_win32.c: Avoid spurious failure due to excess precision The test case testsuite/libffi.call/many_win32.c can spuriously fail due to excess floating-point precision. Instrumenting it with some printf calls shows differences well above FLT_EPSILON. (Note when instrumenting it that multiple computations of the difference, such as one in a print and another in the conditional, may produce different results.) Rather than complicating the test suite with architecture-specific flags to avoid excess precision, just simplify the floating-point computation to avoid a dependency on potential excess precision. commit c00a49ecd165b2d06c1c9b249d212dc843fa116f Author: Josh Triplett Date: Sat Mar 15 22:08:19 2014 -0700 testsuite/libffi.call/many.c: Avoid spurious failure due to excess precision The test case testsuite/libffi.call/many.c can spuriously fail due to excess floating-point precision. Instrumenting it with some printf calls shows differences well above FLT_EPSILON. (Note when instrumenting it that multiple computations of the difference, such as one in a print and another in the conditional, may produce different results.) Rather than complicating the test suite with architecture-specific flags to avoid excess precision, just simplify the floating-point computation to avoid a dependency on potential excess precision. commit 071eab32a7f9fbbef46c0d8f37d9985bc9cceb37 Merge: 2228c7ab1 2f44952c9 Author: Anthony Green Date: Sun Mar 16 07:36:52 2014 -0400 Merge pull request #92 from joshtriplett/autogen Re-add libtool-ldflags commit 2f44952c95765c1486fad66f57235f8d459a9748 Author: Josh Triplett Date: Sun Mar 16 04:35:12 2014 -0700 Re-add libtool-ldflags commit 2228c7ab190f3c529b9018495467b841fa21cba2 Merge: 76d19d004 35634dbce Author: Anthony Green Date: Sun Mar 16 07:25:18 2014 -0400 Merge pull request #87 from joshtriplett/autogen Remove autogenerated files from the repository commit 35634dbceaac0a1544f7385addc01d21ef1ef6a8 Author: Josh Triplett Date: Sat Mar 15 18:11:16 2014 -0700 Remove autogenerated files from the repository Add an autogen.sh to regenerate them. commit 76d19d004e36e99d261ee78261e2f52cea5e4ab1 Merge: c86d9b6cc a1a6f71bf Author: Anthony Green Date: Fri Mar 14 16:54:31 2014 -0400 Ensure the linker supports @unwind sections in libffi. commit c86d9b6cc6e16ee262844a33b40441374400758c Merge: 4efb7dbfd f8cdf1146 Author: Anthony Green Date: Fri Mar 14 16:51:20 2014 -0400 Fix merge commit 4efb7dbfd9427c478a948cd0d464210123db8de8 Merge: 634a475ea 18eb81d03 Author: Anthony Green Date: Fri Mar 14 16:47:57 2014 -0400 Merge pull request #81 from rvandermeulen/bug-756740 Allow building for mipsel with Android NDK r8. commit a1a6f71bfe4199293043b2e4cfb4c8d3cb1112f9 Author: Ryan VanderMeulen Date: Mon Mar 10 15:12:47 2014 -0400 Remove stray hunk that shouldn't have been included in this patch. commit f8cdf11467181f2a9a6b7e748167569aa58e3a81 Author: Ryan VanderMeulen Date: Mon Mar 10 15:04:58 2014 -0400 Replace double quotes with single quotes in Makefile.in to improve compatibility between some versions of MSYS and gmake. From Mozilla bug 943728. https://bugzilla.mozilla.org/show_bug.cgi?id=943728 commit dfa3738732e1bc3a7f4130395ae4bab55fcebb99 Author: Ryan VanderMeulen Date: Mon Mar 10 14:53:48 2014 -0400 Ensure the linker supports @unwind sections in libffi. From Mozilla bug 756740. https://bugzilla.mozilla.org/show_bug.cgi?id=778414 Also tracked as issue #42. https://github.com/atgreen/libffi/issues/42 commit 18eb81d032f29d645d0498ba92bddfd651f009ae Author: Ryan VanderMeulen Date: Mon Mar 10 14:43:37 2014 -0400 Allow building for mipsel with Android NDK r8. From Mozilla bug 756740. https://bugzilla.mozilla.org/show_bug.cgi?id=756740 commit 2349fec9a818fb52fd2f294bcbc7b3156cd113de Author: Daiki Ueno Date: Wed Mar 5 17:53:02 2014 +0900 Fix typo in doc commit 634a475eaf1bee31c09f7d519e31c13b64cd24df Author: Anthony Green Date: Sat Mar 1 18:37:29 2014 -0500 Update Makefile for new darwin scripts commit c7b67e874bb89859f9a07d1cf9606052b6c0dcc1 Author: Anthony Green Date: Sat Mar 1 18:34:18 2014 -0500 Add README note commit a04e30ba3dc303133d459c1ac273ceefe4d49b32 Author: Anthony Green Date: Fri Feb 28 17:20:59 2014 -0500 Add missing -DFFI_DEBUG flag commit 934dc1b5c8d6a9e727bedc72342831eb7d62c35f Merge: 11d7aa9d7 67fbef3b5 Author: Anthony Green Date: Fri Feb 28 01:10:17 2014 -0500 Merge branch 'master' of github.com:/atgreen/libffi commit 11d7aa9d7a4bbe642944edc0f07cf96db9b270b6 Merge: b40aeda31 3b44d4115 Author: Anthony Green Date: Fri Feb 28 01:06:48 2014 -0500 Merge pull request #46 from makotokato/android-clang Fix build failure when using clang for Android commit 67fbef3b56ff0ef88f9b1a7fe48cb77222fa6cec Merge: b40aeda31 3b44d4115 Author: Anthony Green Date: Fri Feb 28 01:06:48 2014 -0500 Merge pull request #46 from makotokato/android-clang Fix build failure when using clang for Android commit b40aeda31a74d95a37c723b6243aabac466e67c4 Merge: 20698abc6 53ceaf14c Author: Anthony Green Date: Fri Feb 28 01:01:29 2014 -0500 Merge branch 'master' of github.com:/atgreen/libffi commit 53ceaf14c5eeb16ba09745f0ca87cca367d41a90 Merge: 860fe6646 cc9b51868 Author: Anthony Green Date: Fri Feb 28 01:01:02 2014 -0500 Merge pull request #40 from wojdyr/master Correct the -L flag in libffi.pc.in commit 20698abc6a00092fd7fd3e434a3a29dc0f048f1e Merge: 64bd06990 1a0b01e17 Author: Anthony Green Date: Fri Feb 28 00:56:27 2014 -0500 Merge pull request #66 from ppizarro/master BlackFin fixes - Fatal error when calling a function defined in a shared library from within the function called by FFI commit 860fe6646f5ae603e99a8d1d722ddddba8b75769 Merge: 64bd06990 1a0b01e17 Author: Anthony Green Date: Fri Feb 28 00:56:27 2014 -0500 Merge pull request #66 from ppizarro/master BlackFin fixes - Fatal error when calling a function defined in a shared library from within the function called by FFI commit 64bd06990a7accf72271516a2110b86cdccd8df4 Author: Anthony Green Date: Fri Feb 28 00:52:56 2014 -0500 Add ChangeLog entry for Josh's change commit edf29c5169b06fcfc241445e152e325bc3c50e0e Merge: 33c9954f2 3998d2604 Author: Anthony Green Date: Fri Feb 28 00:50:25 2014 -0500 Merge pull request #75 from joshtriplett/longdouble Fix build error on x86 without distinct long double commit 33c9954f2eec539011a0f93270aaf013318837ae Author: Anthony Green Date: Fri Feb 28 00:38:41 2014 -0500 Rebuilt with new libtool commit 926b6989fbd08488b9105943293353d45ac527e0 Merge: 5a88c85fd cc82051c7 Author: Anthony Green Date: Fri Feb 28 00:26:57 2014 -0500 Merge branch 'master' of github.com:/atgreen/libffi Conflicts: ChangeLog commit 5a88c85fde304052bed1581ed0b6452ac2c68838 Author: Anthony Green Date: Fri Feb 28 00:23:04 2014 -0500 Fix spelling errors commit cc82051c7e80cea772c4b72da026eb7e68d598fc Author: Anthony Green Date: Fri Feb 28 00:23:04 2014 -0500 Fix spelling errors commit 001aaf4b1b56349596bb6f6b5c1613dcbbd84ea8 Author: Anthony Green Date: Fri Feb 28 00:20:17 2014 -0500 When no VFP arguments are present the IP register is used uninitialized. Initialize it to the value of FP. This fixes a number of testsuite failures when configured for armv7l-unknown-linux-gnueabihf commit 49f7729c3ce697c12408c42ccb29cdf4eb66bb85 Author: Anthony Green Date: Fri Feb 28 00:17:16 2014 -0500 aarch64 fix commit 447483d51c6aa9df7116f346a73fc1cf795f4c2b Merge: 51377bda9 b4df9cf9c Author: Anthony Green Date: Thu Feb 27 15:42:41 2014 -0500 Fix ChangeLog merge commit 3998d2604b5c0d45a098ff3119a9fd9710ef429d Author: Josh Triplett Date: Mon Feb 17 11:20:33 2014 -0800 Fix build error on x86 without distinct long double src/x86/ffi64.c: In function 'classify_argument': src/x86/ffi64.c:205:5: error: duplicate case value case FFI_TYPE_LONGDOUBLE: ^ src/x86/ffi64.c:202:5: error: previously used here case FFI_TYPE_DOUBLE: ^ commit 51377bda9aed0b2c1309c818460cab9d9ab3d46e Merge: f08da5465 40927bd3e Author: Anthony Green Date: Sat Feb 15 08:06:29 2014 -0500 Merge pull request #72 from heiher/devel MIPS N32: Fix call floating point va function commit f08da5465463e60a28f5e921f23ebf2ba984c148 Merge: 3dc3f32c3 fa5f25c20 Author: Anthony Green Date: Sat Feb 15 08:06:11 2014 -0500 Merge pull request #68 from zeldin/master Linux/ppc64: Remove assumption on contents of r11 in closure commit 40927bd3e1e7c6007025ba10854fd8a0664e47df Author: Heiher Date: Tue Jan 21 23:18:27 2014 +0800 Fix call floating point va function I'm not sure floating-point arguments in GPR or FPR before calling variable number arguments function. so, load all arguments to GPR and FPR. commit b4df9cf9cc4a9a9401a53fd6bea1f3c2c283b97b Author: Zachary Waldowski Date: Wed Feb 5 14:22:52 2014 -0500 AArch64: Fix void fall-through case when assertions are enabled commit f466aad0d91a117e42571d1d6fb434fa0433c930 Author: Zachary Waldowski Date: Tue Jan 21 16:38:31 2014 -0500 AArch64: Fix missing semicolons when assertions are enabled commit 7ea677733bd98917241852b8901a6b7580728895 Author: Anthony Green Date: Sat Nov 30 20:58:31 2013 -0500 Remove build-ios from Makefile Conflicts: ChangeLog commit 6ae046cc59c12b2cd40158d6bcb96f4a59886159 Author: Anthony Green Date: Sat Nov 30 21:06:51 2013 -0500 Mention Aarch64 on iOS commit bfc06b3fdb32abe90ce0749aedfec649df85a7ef Author: Zachary Waldowski Date: Mon Dec 30 17:36:39 2013 -0500 Update ChangeLog commit 0a0f12ce1f7be81006b08a3c81a636926d283a9b Author: Zachary Waldowski Date: Thu Jan 9 13:50:17 2014 -0500 AArch64: Remove duplicitous element_count call. This inhibits an analyzer warning by Clang. commit 4330fdcd92e67c816288d64ab230237065768206 Author: Zachary Waldowski Date: Thu Jan 9 13:53:30 2014 -0500 Darwin/aarch64: Respect iOS ABI re: stack argument alignment commit 0a333d6c3973935d4fe02aae76b10e39d3c88e07 Author: Zachary Waldowski Date: Thu Jan 9 14:03:29 2014 -0500 Darwin/aarch64: Fix size_t assumptions commit 2c18e3c76aad1b426617db05a4384e7c3a920176 Author: Zachary Waldowski Date: Mon Dec 30 16:14:02 2013 -0500 Darwin/aarch64: Fix "shadows declaration" warnings commit 1b8a8e20e344f3c55495ab6eb46bd14e843d4b3e Author: Zachary Waldowski Date: Thu Jan 9 13:55:21 2014 -0500 Darwin/aarch64: Use Clang cache invalidation builtin commit 6030cdcae776f8fb5876a53168f7d1e75d28a242 Author: Zachary Waldowski Date: Mon Dec 30 15:45:51 2013 -0500 Darwin/aarch64: Account for long double being equal to double commit 5658b0892683d2e24e4d5842978c184a7ad33858 Author: Zachary Waldowski Date: Mon Dec 30 16:33:47 2013 -0500 Darwin/aarch64: Use CNAME, restrict .size like ARM commit 07175780802acec5dc49fdedd6d20a62409a6707 Author: Zachary Waldowski Date: Mon Dec 30 17:48:22 2013 -0500 Darwin/aarch64: Fix invalid reference in assembly commit 9da28b44277fea3aeb827c35dd63d609d2524a8b Author: Zachary Waldowski Date: Mon Dec 30 16:23:21 2013 -0500 Darwin/x86_64: Fix 64-bit type shortening warnings commit 821d398f08bd1d540a5b235507812ffeee49b580 Author: Zachary Waldowski Date: Thu Jan 9 13:15:06 2014 -0500 Darwin: Merge build scripts, redo project, incl. arm64 commit 6eff9ff9e72463b9783be2514f944b6f05692054 Author: Zachary Waldowski Date: Mon Dec 30 17:48:10 2013 -0500 Darwin/iOS: Improve unified syntax use for LLVM commit ba0ea99c82aadd5957386a031e3122011bd36d52 Author: Zachary Waldowski Date: Mon Dec 30 15:27:44 2013 -0500 Fix dlmalloc warnings due to set_segment_flags, sizeof(size_t) commit 994be3a5c1d0d17b19103396103e128517fd62f9 Author: Zachary Waldowski Date: Mon Dec 30 15:27:14 2013 -0500 Darwin/iOS: Fix mis-typing of vfp_reg_free commit a8e0a835ab1f62d03ad6391760e3e8b7732d24f8 Author: Zachary Waldowski Date: Mon Dec 30 15:26:20 2013 -0500 Darwin/ARM: Assert on NULL dereference This inhibits an analyzer warning by Clang on all platforms. commit 13675341581c489ed9df8ba390c8e08a732decb2 Author: Zachary Waldowski Date: Thu Jan 9 13:42:08 2014 -0500 Darwin/i386: Inhibit Clang previous prototype warnings commit 66469c381e2e2cc96e7d409266dea0ffe177eeca Author: Zachary Waldowski Date: Thu Jan 9 13:41:45 2014 -0500 Darwin/ARM: Inhibit Clang previous prototype warnings commit 5bfe62a00d2d659eec9f19b39802b6e69844fc27 Author: Zachary Waldowski Date: Thu Jan 9 13:41:27 2014 -0500 Darwin/AArch64: Inhibit Clang previous prototype warnings commit fa5f25c20f76a6ef5e950a7ccbce826672c8a620 Author: Marcus Comstedt Date: Sat Jan 4 19:00:08 2014 +0100 Linux/ppc64: Remove assumption on contents of r11 in closure commit 1a0b01e171e9c750437cef2f18917f5a6e32c498 Author: Paulo Pizarro Date: Thu Jan 2 16:17:59 2014 -0200 When the function called by the ffi called a function defined in a shared library generate a fatal error The correction was to take into consideration the GOT. commit 3dc3f32c35db5ab995a835225f6815369735ceb7 Author: Anthony Green Date: Thu Dec 5 16:23:25 2013 -0500 Undo iOS ARM64 changes. commit 356b2cbc304bfe5bdc28b8d1c68d1ff084e9ec37 Merge: 484a75842 07345a30e Author: Anthony Green Date: Sat Nov 30 22:38:13 2013 -0500 Merge branch 'master' of github.com:/atgreen/libffi commit 484a7584260e2fbb399ce90083046834271bf9ff Author: Anthony Green Date: Sat Nov 30 21:06:51 2013 -0500 Mention Aarch64 on iOS commit 07345a30ec0a2fa45a7c363d301f57723690cfa0 Author: Anthony Green Date: Sat Nov 30 21:06:51 2013 -0500 Mention Aarch64 on iOS commit d4b931c1b872378c35f12ddbb9a6d55e7f17c65e Author: Anthony Green Date: Sat Nov 30 20:58:31 2013 -0500 Remove build-ios from Makefile commit dfbf236d70fc1ec68e6ff193584a154353508e2f Merge: 852ac3bd3 bb9740e54 Author: Anthony Green Date: Sat Nov 30 20:54:54 2013 -0500 Merge branch 'master' of github.com:/atgreen/libffi Add ChangeLog entry. commit bb9740e545205f93a525c77aa6d1cbf4ca9371f3 Merge: ac7536889 4d701e03f Author: Anthony Green Date: Sat Nov 30 17:54:39 2013 -0800 Merge pull request #60 from zwaldowski/ios-redo Mac/iOS support, including aarch64 port commit 4d701e03faa475a5eb3b54b90046114a1e27b813 Author: Zachary Waldowski Date: Sat Nov 30 13:25:27 2013 -0500 Darwin: Properly export headers from Xcode project commit 022f12eb9ad2264e838fa5fb453733f5177888f4 Author: Zachary Waldowski Date: Sat Nov 30 12:21:38 2013 -0500 Darwin: Freshen gen scripts, remove old build-ios.sh commit e820fe2025d7ad3df7584407946dfaad2af69599 Author: Zachary Waldowski Date: Sat Nov 30 12:03:51 2013 -0500 Darwin/iOS: Include x86_64+aarch64 pieces in library commit 0278284e470ec91db7cdc15ac3dcd64683305848 Author: Zachary Waldowski Date: Sat Nov 30 03:03:37 2013 -0500 Darwin/aarch64: size_t assumptions commit 9775446b6441c91cd9059215c106aa3bcf949767 Author: Zachary Waldowski Date: Sat Nov 30 02:39:34 2013 -0500 Darwin/aarch64: Fix “shadows declaration” warnings commit 4260badc37705d3618e774dfe61184ac709881c1 Author: Zachary Waldowski Date: Sat Nov 30 02:08:14 2013 -0500 Darwin/aarch64: Use Clang cache invalidation builtin commit 9fa7998d5f9250908cbf12a671479852ebadf9d1 Author: Zachary Waldowski Date: Sat Nov 30 02:07:48 2013 -0500 Darwin/aarch64: Inhibit Xcode warning commit 0e832048a93830575b0976406444e134e649a4f7 Author: Zachary Waldowski Date: Sat Nov 30 02:07:34 2013 -0500 Darwin/aarch64: double == long double commit 602dc22d76931092610234cf063f9f1b8dbc1a51 Author: Zachary Waldowski Date: Sat Nov 30 02:06:00 2013 -0500 Darwin/iOS prep script: try and compile for arm64 commit b513dfe79de4725e8a717325a9e3b5b9f69f63dc Author: Zachary Waldowski Date: Sat Nov 30 02:05:22 2013 -0500 Darwin/aarch64: Restrict .size to ELF like arm32. commit bc978099bf2812de755c076b67ef9c2547607572 Author: Zachary Waldowski Date: Sat Nov 30 02:04:57 2013 -0500 Darwin/aarch64: Potentially(?) fix compile error commit d6bb9314467c6e0683156559d23ca341c43fa3c8 Author: Zachary Waldowski Date: Sat Nov 30 02:04:22 2013 -0500 Darwin/aarch64: Use CNAME refs commit 33c46ce5680eea28d3437c8771ec1d137e226b45 Author: Zachary Waldowski Date: Sat Nov 30 04:13:42 2013 -0500 Darwin/Mac: Fix 64/32 shortening warnings commit 0612081e6c161d9d820742f995975d35da2adbc2 Author: Zachary Waldowski Date: Sat Nov 30 03:03:00 2013 -0500 Darwin: Misc size_t warnings commit 6a6247d179ec3859311c2d8775841b884f309f66 Author: Zachary Waldowski Date: Sat Nov 30 02:55:48 2013 -0500 Darwin: Fix dlmalloc warnings due to sizeof(size_t) commit 4d60d9e1e32de6166ffd63bbe9ce54cf961c78fc Author: Zachary Waldowski Date: Sat Nov 30 04:09:30 2013 -0500 Darwin: Rebuild Xcode project commit cb719a5c1c2eb391d6a5f5e02484ba4aa990a51b Author: Zachary Waldowski Date: Sat Nov 30 04:09:18 2013 -0500 Darwin/iOS: Fix LLVM 3.3 warning re: memcpy. commit 21bde92c9abb378f9c456a9d95e6f9b99ef8c920 Author: Zachary Waldowski Date: Sat Nov 30 03:43:42 2013 -0500 Darwin: Clean up, modernize generator scripts commit fd54eab74cef7891e4acaaafb71e783142ecb69e Author: Zachary Waldowski Date: Sat Nov 30 03:38:02 2013 -0500 Darwin/Mac: Also exclude OS X generated source commit 953b6f14c655141f9e7d82550a312c3eeb961091 Author: Zachary Waldowski Date: Tue Apr 24 11:16:20 2012 -0400 Darwin/iOS: More unified syntax support w/ Clang. Signed-off-by: Zachary Waldowski commit c713a55379481c339877f2e0003d97cb8d9ed80e Author: Zachary Waldowski Date: Tue Apr 24 10:25:29 2012 -0400 Darwin/iOS: Simplify RETLDM arguments for LLVM 3.1 Signed-off-by: Zachary Waldowski commit 16ba1b80028db5cb71cf86e5f79f5e48317f83c8 Author: Zachary Waldowski Date: Wed Apr 11 23:26:04 2012 -0400 Darwin: Silence Clang warnings. commit 852ac3bd302d6ed97b1ef65f4cbed69c258a48df Merge: ab79d6e21 ac7536889 Author: Anthony Green Date: Thu Nov 21 21:25:44 2013 -0500 Merge branch 'master' of github.com:/atgreen/libffi Conflicts: ChangeLog commit ab79d6e21992dd86139ba07530ff888833b78a04 Author: Alan Modra Date: Thu Nov 21 06:12:35 2013 -0500 This separates the 32-bit sysv/linux/bsd code from the 64-bit linux code, and makes it possible to link code compiled with different options to those used to compile libffi. For example, a -mlong-double-128 libffi can be used with -mlong-double-64 code. Using the return value area as a place to pass parameters wasn't such a good idea, causing a failure of cls_ulonglong.c. I didn't see this when running the mainline gcc libffi testsuite because that version of the test is inferior to the upstreamm libffi test. Using NUM_FPR_ARG_REGISTERS rather than NUM_FPR_ARG_REGISTERS64 meant that a parameter save area could be allocated before it was strictly necessary. Wrong but harmless. Found when splitting apart ffi.c into 32-bit and 64-bit support. commit ac7536889334d4be50709006d7e23536364d7891 Author: Alan Modra Date: Thu Nov 21 06:12:35 2013 -0500 This separates the 32-bit sysv/linux/bsd code from the 64-bit linux code, and makes it possible to link code compiled with different options to those used to compile libffi. For example, a -mlong-double-128 libffi can be used with -mlong-double-64 code. Using the return value area as a place to pass parameters wasn't such a good idea, causing a failure of cls_ulonglong.c. I didn't see this when running the mainline gcc libffi testsuite because that version of the test is inferior to the upstreamm libffi test. Using NUM_FPR_ARG_REGISTERS rather than NUM_FPR_ARG_REGISTERS64 meant that a parameter save area could be allocated before it was strictly necessary. Wrong but harmless. Found when splitting apart ffi.c into 32-bit and 64-bit support. commit 69df91cfb4fa6bcb644350a80bff970f27478a6a Merge: 2f450822a aa1f62c0a Author: Anthony Green Date: Mon Nov 18 06:34:04 2013 -0800 Merge pull request #59 from iains/powerpc-darwin-unwind-fix Fix PowerPC Darwin FDE encodings to use pcrel correctly. Modernise the picbase labels. commit aa1f62c0a093c30325dff1d4d2b6b4b22eb96929 Author: Iain Sandoe Date: Mon Nov 18 13:11:56 2013 +0000 Fix PowerPC Darwin FDE encodings to use pcrel correctly. Modernise the picbase labels. commit 2f450822a8698ba88441c56d152c7dc8924b127f Author: Anthony Green Date: Mon Nov 18 06:52:29 2013 -0500 Clean up code to appease modern GCC compiler. commit 16d56c51aded374730920a4acde76ff3d2860ae1 Author: Alan Modra Date: Mon Nov 18 06:36:03 2013 -0500 An #endif in the wrong place would cause compile failure on powerpcle. Using bl instead of b doesn't cause runtime failures as you might think, but does mess the processor branch prediction. commit 34f878a5ef28663f6b1d7fd26fb099429ea1579e Merge: 83f65b63d 1fd045784 Author: Anthony Green Date: Sat Nov 16 06:57:54 2013 -0500 Merge branch 'master' of github.com:/atgreen/libffi Conflicts: ChangeLog src/powerpc/ffi.c commit 83f65b63d9764a9cc7688fc5cda5ee2bd23faf54 Author: Alan Modra Date: Sat Nov 16 06:53:50 2013 -0500 Finally, this adds _CALL_ELF == 2 support. ELFv1 objects can't be linked with ELFv2 objects, so this is one case where preprocessor tests in ffi.c are fine. Also, there is no need to define a new FFI_ELFv2 or somesuch value in enum ffi_abi. FFI_LINUX64 will happily serve both ABIs. commit 1fd045784cac874b5d76b7fa931f67209a8280d3 Author: Alan Modra Date: Sat Nov 16 06:53:50 2013 -0500 Finally, this adds _CALL_ELF == 2 support. ELFv1 objects can't be linked with ELFv2 objects, so this is one case where preprocessor tests in ffi.c are fine. Also, there is no need to define a new FFI_ELFv2 or somesuch value in enum ffi_abi. FFI_LINUX64 will happily serve both ABIs. commit 362851379a49ce07d3e36e82c4e5c7b6cc16a352 Author: Alan Modra Date: Sat Nov 16 06:52:43 2013 -0500 Andreas' 2013-02-08 change reverted some breakage for struct return values from 2011-11-12, but in so doing reintroduced string instructions to sysv.S that are not supported on all powerpc variants. This patch properly copies the bounce buffer to destination in C code rather than in asm. I have tested this on powerpc64-linux, powerpc-linux and powerpc-freebsd. Well, the last on powerpc-linux by lying to configure with CC="gcc -m32 -msvr4-struct-return -mlong-double-64" \ CXX="g++ -m32 -msvr4-struct-return -mlong-double-64" \ /src/libffi-current/configure --build=powerpc-freebsd and then make && make CC="gcc -m32" CXX="g++ -m32" \ RUNTESTFLAGS=--target_board=unix/-m32/-msvr4-struct-return/-mlong-double-64\ check commit 1c06515d927d9de1582438d4eb5953890e79c5c7 Author: Alan Modra Date: Sat Nov 16 06:41:36 2013 -0500 The powerpc64 ABIs align structs passed by value, a fact ignored by gcc for quite some time. Since gcc now does the correct alignment, libffi needs to follow suit. This ought to be made selectable via a new abi value, and the #ifdefs removed from ffi.c along with many other #ifdefs present there and in assembly. I'll do that with a followup patch sometime. This is a revised version of https://sourceware.org/ml/libffi-discuss/2013/msg00162.html commit a97cf1fae575d8bfd5259c5c422025ad43911326 Author: Alan Modra Date: Sat Nov 16 06:40:13 2013 -0500 This patch prepares for ELFv2, where sizes of these areas change. It also makes some minor changes to improve code efficiency. commit 164283f4ac5972ce2ab5e015cc2ab1014c23276c Author: Alan Modra Date: Sat Nov 16 06:38:55 2013 -0500 The powerpc64 support opted to pass floating point values both in the fpr area and the parameter save area, necessary when the backend doesn't know if a function argument corresponds to the ellipsis arguments of a variadic function. This patch adds powerpc support for variadic functions, and changes the code to only pass fp in the ABI mandated area. ELFv2 needs this change since the parameter save area may not exist there. This also fixes two faulty tests that used a non-variadic function cast to call a variadic function, and spuriously reasoned that this is somehow necessary for static functions.. commit 31257b3189f81a199bc2902c22bc5f2d7c54ccde Author: Andrew Haley Date: Sat Nov 16 06:35:51 2013 -0500 Fix sample closure code commit db0ace3a38496af73eae3df02ef353736d16909f Author: Andrew Haley Date: Sat Nov 16 06:29:25 2013 -0500 Fix broken test cases commit de10f5039ed7a53382ddcc95c368d03e535edb98 Merge: 58c2577a3 f3657da27 Author: Anthony Green Date: Thu Nov 14 10:56:29 2013 -0500 Merge branch 'master' of https://github.com/bivab/libffi Conflicts: ChangeLog commit f3657da278dd63afcdd8762894a9bdaea8ef028a Author: David Schneider Date: Thu Nov 14 13:02:16 2013 +0100 update Changelog commit 58c2577a3ff80e7416ef0434769e2af23365719c Author: Alan Modra Date: Wed Nov 13 16:55:36 2013 -0500 This enshrines the current testsuite practice of using ffi_arg for returned values. It would be reasonable and logical to use the actual return argument type as passed to ffi_prep_cif, but this would mean changing a large number of tests that use ffi_arg and all backends that write results to an ffi_arg. commit 8af42f9944f7ed72c81ae360aac6a84dc11f89dc Author: Anthony Green Date: Wed Nov 13 16:40:28 2013 -0500 Respect HAVE_ALLOCA_H commit cdf405d574f479b782454516366bd4f4b9b3415e Author: David Schneider Date: Wed Nov 13 15:50:21 2013 +0100 add a testcase for the double/float issue on ARMHF commit 77f823e31ffb557a466b24f7fba845fbf7831798 Author: David Schneider Date: Wed Nov 13 14:26:57 2013 +0100 stop trying to assing vfp regs once we are done with the registers commit 37067ec5036f2a6ed7a4799f83f8f53160460344 Author: David Schneider Date: Tue Nov 12 19:49:01 2013 +0100 mark all vfp registers as used when done. To avoid assigning registers the would fit, once arguments have been on the stack, we mark all registers as used once we do not find a free register for the first time. commit 2f5b7ce545473a7f6e41193edc29407cbebe82d5 Author: Anthony Green Date: Sat Nov 9 06:16:32 2013 -0500 UltraSPARC IIi fix. Update README and build configury. commit becd754434173032f426d22ffcbfe24f55b3c137 Author: Mark Kettenis Date: Wed Nov 6 06:43:49 2013 -0500 Align the stack pointer to 16-bytes. commit 05c31093409f7b3e6d795fac21d2c954313d8162 Author: Konstantin Belousov Date: Wed Nov 6 06:40:58 2013 -0500 Mark executable as not requiring executable stack. commit cf6bf9818e8394cfcdb07a40c6a5e2ee6b01d333 Author: Anthony Green Date: Sat Nov 2 17:23:59 2013 -0400 Fix up docs commit 02177176854d16fc0f1a5958aa34da2f306630ee Merge: c2422174b c265b4cac Author: Anthony Green Date: Sat Nov 2 17:11:22 2013 -0400 Merge branch 'master' of github.com:/atgreen/libffi commit c2422174b3edc0de0b148dfd6b67087bb881c4a6 Merge: f4b843f83 d918d4780 Author: Anthony Green Date: Sat Nov 2 14:08:23 2013 -0700 Merge pull request #45 from foss-for-synopsys-dwc-arc-processors/arc_support arc: Fix build error commit c265b4cacb9130f042699a85de9c7242b3f49cc3 Merge: f4b843f83 d918d4780 Author: Anthony Green Date: Sat Nov 2 14:08:23 2013 -0700 Merge pull request #45 from foss-for-synopsys-dwc-arc-processors/arc_support arc: Fix build error commit f4b843f83710ac378c48abd87fe66bb519d30d2e Author: Anthony Green Date: Sat Nov 2 17:01:15 2013 -0400 Don't align stack for win32 commit f3cd39345713db8e414cf642b6cb65a4cfe6018c Merge: 666f3e71b 6aa15900a Author: Anthony Green Date: Sat Nov 2 13:17:57 2013 -0700 Merge pull request #51 from vbudovski/for_upstream Don't use 16 byte aligned stack for WIN32 commit 666f3e71b56d92c49fcd2d7f349b8f8ebca0f8a3 Author: Anthony Green Date: Sat Oct 26 09:12:42 2013 -0400 Add more credits to README. Tidy up. commit 73ada14e756bad97fad0e6915a821a3c7e079f81 Author: Anthony Green Date: Sat Oct 26 09:09:45 2013 -0400 Update README commit d3372c54ce7117e80d389ba875dc5b6b2213c71e Author: Mark H Weaver Date: Sat Oct 26 08:30:06 2013 -0400 Fix N32 ABI issue for MIPS. commit d6716aba8118eb0513885cfe557bedebb7016e8b Author: Anthony Green Date: Tue Oct 15 15:42:49 2013 -0400 Update travis-ci build dependencies to include texinfo commit 16b93a211bcfbe4bd0efdcf94de225a71aa0ee02 Author: Sandra Loosemore Date: Tue Oct 15 15:33:59 2013 -0400 Add nios2 port. commit 2f5626ce02fce8267ab48ceb6d7d0ed7d672a75e Author: Sandra Loosemore Date: Tue Oct 15 15:32:16 2013 -0400 Fix testsuite bug commit f64e4a865557e440774436b4c2b2fd7374290e97 Author: Marcus Shawcroft Date: Tue Oct 15 15:20:14 2013 -0400 Fix many.c testcase for Aarch64 commit 128cd1d2f358f26d9fa75a27cf2b30356f5dd903 Author: Anthony Green Date: Tue Oct 8 06:45:51 2013 -0400 Fix spelling errors commit ff06269d707cafbfef2a88afb07a79c9d1480c5f Author: Anthony Green Date: Tue Oct 8 06:32:18 2013 -0400 Update README for M88K and VAX commit d2fcbcdfbea750d1f6a9f493e2e6c4d5ffa71b34 Author: Anthony Green Date: Tue Oct 8 06:27:46 2013 -0400 Add m88k and VAX support. Update some configury bits. commit 6aa15900accc0a648cdebf11ec11d11697ebfffd Author: Vitaly Budovski Date: Thu Sep 5 12:05:06 2013 +1000 Don't use 16 byte aligned stack for WIN32 This fixes a crash when accessing __stdcall functions in Python ctypes. commit 3b44d41156149af8da2a58825fefdfa23274ae7a Author: Makoto Kato Date: Wed Jul 10 15:34:53 2013 +0900 Fix build failure when using clang for Android clang for Android generates __gnu_linux__ define, but gcc for Android doesn't. So we should add check it for Android commit d918d47809c174d62283306b282749f8db93661f Author: Mischa Jonker Date: Mon Jul 8 15:51:36 2013 +0200 arc: Fix build error One part of the patch for ARC support was missing in the upstreamed version. Signed-off-by: Mischa Jonker commit d3d099b40c122550279789200263346f120f6909 Author: Anthony Green Date: Tue Jul 2 16:11:38 2013 -0400 little-endian ppc64 support commit 0f8690a84c874ec09a090c8c6adfb93c594acac6 Author: Anthony Green Date: Tue Jul 2 15:54:40 2013 -0400 Rebuild for ARC additions commit f88118b345f27c46f5445d6e4832c498ff9a6d85 Author: Anthony Green Date: Tue Jul 2 15:51:27 2013 -0400 Revert "Merge pull request #36 from abergmeier/emscripten_fix" This reverts commit 6a4d901dde7b3f87984c563505717cde3113d16e, reversing changes made to b50a13b9c07ec09af4b9697e482acdad571e6961. commit 6a4d901dde7b3f87984c563505717cde3113d16e Merge: b50a13b9c 587002c09 Author: Anthony Green Date: Tue Jul 2 12:12:34 2013 -0700 Merge pull request #36 from abergmeier/emscripten_fix Fixes for building with Emscripten commit b50a13b9c07ec09af4b9697e482acdad571e6961 Merge: 767f1f96e b082e1509 Author: Anthony Green Date: Tue Jul 2 12:10:26 2013 -0700 Merge pull request #44 from foss-for-synopsys-dwc-arc-processors/arc_support Add ARC support commit 767f1f96e5282da44d7340e6815e9820a3f78e39 Merge: c3c40e029 b8a91d81b Author: Anthony Green Date: Tue Jul 2 12:08:04 2013 -0700 Merge pull request #43 from JensTimmerman/__m128 added include for xmmintrin.h commit b8a91d81be77d479327fdb6bdd9fdae6d18e6e63 Author: Jens Timmerman Date: Tue Jul 2 10:57:37 2013 +0200 added include for xmmintrin.h commit b082e15091961373c03d10ed0251f619ebb6ed76 Author: Mischa Jonker Date: Mon Jun 10 16:19:33 2013 +0200 Add ARC support This adds support for the ARC architecture to libffi. DesignWare ARC is a family of processors from Synopsys, Inc. This patch has been tested on a little-endian system and passes the testsuite. Signed-off-by: Mischa Jonker commit cc9b518687e46b0d1acafdd4bc3f3b281c25a3d9 Author: Marcin Wojdyr Date: Tue May 14 15:01:23 2013 +0200 Update libffi.pc.in use -L${toolexeclibdir} instead of -L${libdir} to be consistent with Makefile.am commit 587002c092cffe6e7a8d7028f246c241d03b738c Author: Andreas Bergmeier Date: Fri Apr 19 17:12:24 2013 +0200 Enable disabling of libtool on platforms where it does not work (e.g. LLVM). Build libraries normally then. commit c3c40e0290377d7cf948b072eedd8317c4bf215e Merge: ede96e4eb 4750e3c66 Author: Anthony Green Date: Sat Mar 30 05:24:14 2013 -0700 Merge pull request #34 from davidsch/armhf Fix ARM hard-float support for large numbers of VFP arguments commit 4750e3c662fd9569cb3e2d28f539685fd1ca8caf Author: David Schneider Date: Thu Mar 28 16:56:36 2013 +0100 update changelog commit 9708e7cf09f1bf815f4d6485eb1f180fabb35804 Author: David Schneider Date: Wed Mar 27 19:31:04 2013 +0100 folow the ARM hard-float ABI in ffi_prep_incoming_args_VFP commit b41120981e5e49ca2da10b94b154775f50da5f36 Author: David Schneider Date: Wed Mar 27 16:38:35 2013 +0100 create separated versions of ffi_prep_incoming_args_* for SYSV and VFP ABIs. The different versions will be called depending on the value of cif->abi commit dd26f1f39c54861c5b91931f0f37a72942c2a072 Author: David Schneider Date: Thu Mar 28 15:39:01 2013 +0100 add a failing test for closures on ARM hardfloat commit 3c1608613ab3c2184222b98c5482cddedd6b559b Author: David Schneider Date: Tue Mar 26 19:24:47 2013 +0100 extend ffi_prepare_args for FFI_VFP (hard-float ABI), fixing an issue with passing VFP arguments in VFP registers and the stack, while at the same time not using all core registers. commit 0f2ff2d4c92719be8936179f9ab674f4d1a3fd14 Author: David Schneider Date: Tue Mar 26 19:22:02 2013 +0100 separate ARM ffi_prepare_args in a version implementing the simple SYSV calling convention and one for the hard-float calling convention commit 3a352b8a8252400a83de22c7c424bf1887b4a2ef Author: David Schneider Date: Tue Mar 26 14:24:04 2013 +0100 move the hardfloat specific argument copying code to the helper function commit 5df6b7944a4225b6eb329f3886be64e04e966f29 Author: David Schneider Date: Tue Mar 26 14:02:21 2013 +0100 extract setting of arguments to be passed to a helper function commit 7d1048c471bb4b1f9d67a9e9f8e95f9a1d2e6d45 Author: David Schneider Date: Tue Mar 26 11:33:33 2013 +0100 extract code to align the argument storage pointer to a helper function commit b9f013788f0f384c423ad963475aaacb55598135 Author: David Schneider Date: Mon Mar 25 13:27:36 2013 +0100 add a testcase, that on ARM hardfloat needs more than the 8 VFP argument registers to pass arguments to a call commit 2fbdb0f231cafdb77b025d3cd8afe90cda99b3ba Author: David Schneider Date: Mon Mar 25 13:26:02 2013 +0100 use the absolute value to check the test result against an epsilon commit ede96e4eb660bbf3e0fe048135efa8106f48af5d Merge: f22ab3c68 9e34992a5 Author: Anthony Green Date: Sun Mar 17 18:38:21 2013 -0400 Merge branch 'master' of github.com:/atgreen/libffi commit 9e34992a5ea2fda1dba5875bf96dc91a7230f51f Merge: 12b1886d7 d08124bed Author: Anthony Green Date: Sun Mar 17 18:34:54 2013 -0400 Merge branch 'master' of github.com:/atgreen/libffi --- native/libffi/.appveyor.yml | 66 + native/libffi/.gitattributes | 4 + native/libffi/.github/issue_template.md | 10 + native/libffi/.gitignore | 5 +- native/libffi/.travis.yml | 95 +- native/libffi/.travis/ar-lib | 270 + native/libffi/.travis/bfin-sim.exp | 58 + .../.travis/build-cross-in-container.sh | 14 + native/libffi/.travis/build-in-container.sh | 12 + native/libffi/.travis/build.sh | 142 + native/libffi/.travis/compile | 351 + native/libffi/.travis/install.sh | 71 + native/libffi/.travis/m32r-sim.exp | 58 + native/libffi/.travis/moxie-sim.exp | 60 + native/libffi/.travis/or1k-sim.exp | 58 + native/libffi/.travis/powerpc-eabisim.exp | 58 + native/libffi/.travis/site.exp | 27 + native/libffi/.travis/wine-sim.exp | 55 + native/libffi/ChangeLog.libffi | 584 -- native/libffi/ChangeLog.libffi-3.1 | 1410 ++++ native/libffi/ChangeLog.libgcj | 40 - native/libffi/ChangeLog.old | 7410 +++++++++++++++++ native/libffi/LICENSE | 4 + native/libffi/LICENSE-BUILDTOOLS | 353 + native/libffi/Makefile.am | 143 +- native/libffi/README.md | 486 ++ native/libffi/acinclude.m4 | 387 + native/libffi/autogen.sh | 6 +- native/libffi/config.guess | 1687 ++++ native/libffi/config.sub | 1851 ++++ native/libffi/configure.ac | 161 +- native/libffi/configure.host | 326 + native/libffi/doc/Makefile.am | 3 + native/libffi/doc/libffi.texi | 312 +- native/libffi/doc/version.texi | 7 + .../generate-darwin-source-and-headers.py | 64 + native/libffi/include/Makefile.am | 6 +- native/libffi/include/ffi.h.in | 187 +- native/libffi/include/ffi_cfi.h | 55 + native/libffi/include/ffi_common.h | 25 +- native/libffi/libffi.map.in | 76 + native/libffi/libffi.pc.in | 4 + .../libffi/libffi.xcodeproj/project.pbxproj | 794 ++ native/libffi/libtool-version | 4 + native/libffi/m4/asmcfi.m4 | 2 +- native/libffi/m4/ax_append_flag.m4 | 61 +- native/libffi/m4/ax_cc_maxopt.m4 | 29 +- native/libffi/m4/ax_cflags_warn_all.m4 | 8 +- native/libffi/m4/ax_check_compile_flag.m4 | 43 +- native/libffi/m4/ax_compiler_vendor.m4 | 14 +- native/libffi/m4/ax_configure_args.m4 | 37 +- native/libffi/m4/ax_enable_builddir.m4 | 15 +- native/libffi/m4/ax_gcc_archflag.m4 | 166 +- native/libffi/m4/ax_gcc_x86_cpuid.m4 | 26 +- native/libffi/m4/ax_require_defined.m4 | 37 + native/libffi/make_sunver.pl | 333 + .../msvc_build/aarch64/Ffi_staticLib.sln | 33 + .../msvc_build/aarch64/Ffi_staticLib.vcxproj | 130 + .../aarch64/Ffi_staticLib.vcxproj.filters | 57 + .../aarch64/Ffi_staticLib.vcxproj.user | 4 + .../msvc_build/aarch64/aarch64_include/ffi.h | 511 ++ .../aarch64/aarch64_include/fficonfig.h | 219 + native/libffi/msvcc.sh | 143 +- native/libffi/src/aarch64/ffi.c | 1110 ++- native/libffi/src/aarch64/ffitarget.h | 54 +- native/libffi/src/aarch64/internal.h | 68 + native/libffi/src/aarch64/sysv.S | 442 +- native/libffi/src/aarch64/win64_armasm.S | 506 ++ native/libffi/src/alpha/ffi.c | 429 +- native/libffi/src/alpha/ffitarget.h | 4 + native/libffi/src/alpha/internal.h | 23 + native/libffi/src/alpha/osf.S | 427 +- native/libffi/src/arc/ffi.c | 14 + native/libffi/src/arm/ffi.c | 1048 ++- native/libffi/src/arm/ffitarget.h | 32 +- native/libffi/src/arm/internal.h | 7 + native/libffi/src/arm/sysv.S | 451 +- native/libffi/src/arm/sysv_msvc_arm32.S | 311 + native/libffi/src/arm/trampoline.S | 4450 ---------- native/libffi/src/closures.c | 430 +- native/libffi/src/cris/ffi.c | 4 +- native/libffi/src/csky/ffi.c | 395 + native/libffi/src/csky/ffitarget.h | 63 + native/libffi/src/csky/sysv.S | 371 + native/libffi/src/dlmalloc.c | 11 +- native/libffi/src/frv/ffi.c | 4 +- native/libffi/src/ia64/ffi.c | 30 +- native/libffi/src/ia64/ffitarget.h | 3 +- native/libffi/src/ia64/unix.S | 9 +- native/libffi/src/java_raw_api.c | 6 +- native/libffi/src/kvx/asm.h | 5 + native/libffi/src/kvx/ffi.c | 273 + native/libffi/src/kvx/ffitarget.h | 75 + native/libffi/src/kvx/sysv.S | 127 + native/libffi/src/m32r/ffi.c | 2 +- native/libffi/src/m68k/ffi.c | 4 +- native/libffi/src/m68k/sysv.S | 29 +- native/libffi/src/m88k/ffi.c | 16 + native/libffi/src/metag/ffi.c | 14 +- native/libffi/src/microblaze/ffi.c | 10 +- native/libffi/src/mips/ffi.c | 154 +- native/libffi/src/mips/ffitarget.h | 23 +- native/libffi/src/mips/n32.S | 151 +- native/libffi/src/mips/o32.S | 177 +- native/libffi/src/moxie/eabi.S | 2 +- native/libffi/src/moxie/ffi.c | 27 +- native/libffi/src/nios2/ffi.c | 8 + native/libffi/src/pa/ffi.c | 137 +- native/libffi/src/pa/ffitarget.h | 7 +- native/libffi/src/pa/hpux32.S | 6 +- native/libffi/src/pa/linux.S | 31 +- native/libffi/src/powerpc/aix.S | 240 +- native/libffi/src/powerpc/aix_closure.S | 253 +- native/libffi/src/powerpc/asm.h | 4 +- native/libffi/src/powerpc/darwin_closure.S | 7 +- native/libffi/src/powerpc/ffi.c | 61 + native/libffi/src/powerpc/ffi_darwin.c | 137 +- native/libffi/src/powerpc/ffi_linux64.c | 450 + native/libffi/src/powerpc/ffi_powerpc.h | 64 + native/libffi/src/powerpc/ffi_sysv.c | 225 + native/libffi/src/powerpc/ffitarget.h | 54 +- native/libffi/src/powerpc/linux64.S | 180 +- native/libffi/src/powerpc/linux64_closure.S | 390 + native/libffi/src/powerpc/ppc_closure.S | 177 +- native/libffi/src/powerpc/sysv.S | 109 +- native/libffi/src/prep_cif.c | 58 +- native/libffi/src/raw_api.c | 10 +- native/libffi/src/riscv/ffi.c | 481 ++ native/libffi/src/riscv/ffitarget.h | 69 + native/libffi/src/riscv/sysv.S | 293 + native/libffi/src/s390/ffi.c | 415 +- native/libffi/src/s390/ffitarget.h | 1 + native/libffi/src/s390/internal.h | 11 + native/libffi/src/s390/sysv.S | 623 +- native/libffi/src/sparc/ffi.c | 861 +- native/libffi/src/sparc/ffi64.c | 608 ++ native/libffi/src/sparc/ffitarget.h | 20 +- native/libffi/src/sparc/internal.h | 26 + native/libffi/src/sparc/v8.S | 613 +- native/libffi/src/sparc/v9.S | 547 +- native/libffi/src/types.c | 10 +- native/libffi/src/vax/ffi.c | 8 + native/libffi/src/x86/asmnames.h | 30 + native/libffi/src/x86/darwin.S | 444 - native/libffi/src/x86/darwin64.S | 416 - native/libffi/src/x86/ffi.c | 816 +- native/libffi/src/x86/ffi64.c | 433 +- native/libffi/src/x86/ffitarget.h | 85 +- native/libffi/src/x86/ffiw64.c | 318 + native/libffi/src/x86/internal.h | 29 + native/libffi/src/x86/internal64.h | 22 + native/libffi/src/x86/sysv.S | 1479 +++- native/libffi/src/x86/sysv_intel.S | 995 +++ native/libffi/src/x86/unix64.S | 793 +- native/libffi/src/x86/win64.S | 241 +- native/libffi/src/x86/win64_intel.S | 238 + native/libffi/src/xtensa/ffi.c | 4 +- native/libffi/src/xtensa/sysv.S | 7 +- native/libffi/testsuite/Makefile.am | 117 + native/libffi/testsuite/lib/libffi.exp | 352 +- .../libffi/testsuite/libffi.bhaible/Makefile | 28 + native/libffi/testsuite/libffi.bhaible/README | 78 + .../libffi/testsuite/libffi.bhaible/alignof.h | 50 + .../testsuite/libffi.bhaible/bhaible.exp | 63 + .../testsuite/libffi.bhaible/test-call.c | 1745 ++++ .../testsuite/libffi.bhaible/test-callback.c | 2885 +++++++ .../testsuite/libffi.bhaible/testcases.c | 743 ++ .../testsuite/libffi.call/align_mixed.c | 20 + .../testsuite/libffi.call/align_stdcall.c | 46 + native/libffi/testsuite/libffi.call/call.exp | 28 + native/libffi/testsuite/libffi.call/ffitest.h | 8 +- native/libffi/testsuite/libffi.call/float1.c | 4 +- native/libffi/testsuite/libffi.call/float2.c | 27 +- native/libffi/testsuite/libffi.call/float3.c | 6 +- native/libffi/testsuite/libffi.call/offsets.c | 46 + .../libffi/testsuite/libffi.call/pr1172638.c | 127 + .../libffi/testsuite/libffi.call/return_ldl.c | 2 +- .../libffi/testsuite/libffi.call/struct10.c | 57 + .../testsuite/libffi.call/unwindtest.cc | 2 +- .../libffi.call/unwindtest_ffi_call.cc | 2 +- native/libffi/testsuite/libffi.call/va_1.c | 2 +- .../testsuite/libffi.closures/closure.exp | 67 + .../closure_fn0.c | 0 .../closure_fn1.c | 0 .../closure_fn2.c | 0 .../closure_fn3.c | 0 .../closure_fn4.c | 0 .../closure_fn5.c | 0 .../closure_fn6.c | 0 .../closure_loc_fn0.c | 0 .../libffi.closures/closure_simple.c | 55 + .../cls_12byte.c | 0 .../cls_16byte.c | 0 .../cls_18byte.c | 0 .../cls_19byte.c | 0 .../cls_1_1byte.c | 0 .../cls_20byte.c | 0 .../cls_20byte1.c | 0 .../cls_24byte.c | 0 .../cls_2byte.c | 0 .../cls_3_1byte.c | 0 .../cls_3byte1.c | 0 .../cls_3byte2.c | 0 .../testsuite/libffi.closures/cls_3float.c | 95 + .../cls_4_1byte.c | 0 .../cls_4byte.c | 0 .../cls_5_1_byte.c | 0 .../cls_5byte.c | 0 .../cls_64byte.c | 0 .../cls_6_1_byte.c | 0 .../cls_6byte.c | 0 .../cls_7_1_byte.c | 0 .../cls_7byte.c | 0 .../cls_8byte.c | 0 .../cls_9byte1.c | 0 .../cls_9byte2.c | 0 .../cls_align_double.c | 0 .../cls_align_float.c | 0 .../cls_align_longdouble.c | 0 .../cls_align_longdouble_split.c | 2 - .../cls_align_longdouble_split2.c | 2 - .../cls_align_pointer.c | 0 .../cls_align_sint16.c | 0 .../cls_align_sint32.c | 0 .../cls_align_sint64.c | 0 .../cls_align_uint16.c | 0 .../cls_align_uint32.c | 0 .../cls_align_uint64.c | 0 .../cls_dbls_struct.c | 4 +- .../cls_double.c | 0 .../cls_double_va.c | 0 .../cls_float.c | 0 .../cls_longdouble.c | 2 - .../cls_longdouble_va.c | 0 .../libffi.closures/cls_many_mixed_args.c | 70 + .../cls_many_mixed_float_double.c | 55 + .../cls_multi_schar.c | 0 .../cls_multi_sshort.c | 0 .../cls_multi_sshortchar.c | 0 .../cls_multi_uchar.c | 0 .../cls_multi_ushort.c | 0 .../cls_multi_ushortchar.c | 0 .../cls_pointer.c | 0 .../cls_pointer_stack.c | 0 .../cls_schar.c | 0 .../cls_sint.c | 0 .../cls_sshort.c | 0 .../cls_struct_va1.c | 0 .../cls_uchar.c | 0 .../cls_uchar_va.c | 0 .../cls_uint.c | 0 .../cls_uint_va.c | 0 .../cls_ulong_va.c | 0 .../cls_ulonglong.c | 0 .../cls_ushort.c | 0 .../cls_ushort_va.c | 0 .../err_bad_abi.c | 0 .../testsuite/libffi.closures/ffitest.h | 138 + .../huge_struct.c | 6 +- .../nested_struct.c | 0 .../nested_struct1.c | 0 .../nested_struct10.c | 1 + .../nested_struct11.c | 0 .../nested_struct2.c | 0 .../nested_struct3.c | 0 .../nested_struct4.c | 0 .../nested_struct5.c | 0 .../nested_struct6.c | 0 .../nested_struct7.c | 0 .../nested_struct8.c | 0 .../nested_struct9.c | 0 .../problem1.c | 0 .../stret_large.c | 0 .../stret_large2.c | 0 .../stret_medium.c | 0 .../stret_medium2.c | 0 .../testclosure.c | 0 .../testsuite/libffi.closures/unwindtest.cc | 117 + .../libffi.closures/unwindtest_ffi_call.cc | 54 + .../libffi.complex/cls_align_complex.inc | 91 + .../libffi.complex/cls_align_complex_double.c | 10 + .../libffi.complex/cls_align_complex_float.c | 10 + .../cls_align_complex_longdouble.c | 10 + .../testsuite/libffi.complex/cls_complex.inc | 42 + .../libffi.complex/cls_complex_double.c | 10 + .../libffi.complex/cls_complex_float.c | 10 + .../libffi.complex/cls_complex_longdouble.c | 10 + .../libffi.complex/cls_complex_struct.inc | 71 + .../cls_complex_struct_double.c | 10 + .../libffi.complex/cls_complex_struct_float.c | 10 + .../cls_complex_struct_longdouble.c | 10 + .../libffi.complex/cls_complex_va.inc | 80 + .../libffi.complex/cls_complex_va_double.c | 10 + .../libffi.complex/cls_complex_va_float.c | 16 + .../cls_complex_va_longdouble.c | 10 + .../testsuite/libffi.complex/complex.exp | 36 + .../testsuite/libffi.complex/complex.inc | 51 + .../libffi.complex/complex_defs_double.inc | 7 + .../libffi.complex/complex_defs_float.inc | 7 + .../complex_defs_longdouble.inc | 7 + .../testsuite/libffi.complex/complex_double.c | 10 + .../testsuite/libffi.complex/complex_float.c | 10 + .../testsuite/libffi.complex/complex_int.c | 86 + .../libffi.complex/complex_longdouble.c | 10 + .../libffi/testsuite/libffi.complex/ffitest.h | 1 + .../testsuite/libffi.complex/many_complex.inc | 78 + .../libffi.complex/many_complex_double.c | 10 + .../libffi.complex/many_complex_float.c | 10 + .../libffi.complex/many_complex_longdouble.c | 10 + .../libffi.complex/return_complex.inc | 37 + .../libffi.complex/return_complex1.inc | 41 + .../libffi.complex/return_complex1_double.c | 10 + .../libffi.complex/return_complex1_float.c | 10 + .../return_complex1_longdouble.c | 10 + .../libffi.complex/return_complex2.inc | 44 + .../libffi.complex/return_complex2_double.c | 10 + .../libffi.complex/return_complex2_float.c | 10 + .../return_complex2_longdouble.c | 10 + .../libffi.complex/return_complex_double.c | 10 + .../libffi.complex/return_complex_float.c | 10 + .../return_complex_longdouble.c | 10 + native/libffi/testsuite/libffi.go/aa-direct.c | 34 + native/libffi/testsuite/libffi.go/closure1.c | 28 + native/libffi/testsuite/libffi.go/ffitest.h | 1 + native/libffi/testsuite/libffi.go/go.exp | 36 + .../libffi/testsuite/libffi.go/static-chain.h | 19 + 326 files changed, 42692 insertions(+), 10489 deletions(-) create mode 100644 native/libffi/.appveyor.yml create mode 100644 native/libffi/.gitattributes create mode 100644 native/libffi/.github/issue_template.md create mode 100755 native/libffi/.travis/ar-lib create mode 100644 native/libffi/.travis/bfin-sim.exp create mode 100755 native/libffi/.travis/build-cross-in-container.sh create mode 100755 native/libffi/.travis/build-in-container.sh create mode 100755 native/libffi/.travis/build.sh create mode 100755 native/libffi/.travis/compile create mode 100755 native/libffi/.travis/install.sh create mode 100644 native/libffi/.travis/m32r-sim.exp create mode 100644 native/libffi/.travis/moxie-sim.exp create mode 100644 native/libffi/.travis/or1k-sim.exp create mode 100644 native/libffi/.travis/powerpc-eabisim.exp create mode 100644 native/libffi/.travis/site.exp create mode 100644 native/libffi/.travis/wine-sim.exp delete mode 100644 native/libffi/ChangeLog.libffi delete mode 100644 native/libffi/ChangeLog.libgcj create mode 100644 native/libffi/ChangeLog.old create mode 100644 native/libffi/LICENSE-BUILDTOOLS create mode 100644 native/libffi/README.md create mode 100644 native/libffi/config.guess create mode 100644 native/libffi/config.sub create mode 100644 native/libffi/configure.host create mode 100644 native/libffi/doc/Makefile.am create mode 100644 native/libffi/include/ffi_cfi.h create mode 100644 native/libffi/libffi.map.in create mode 100644 native/libffi/m4/ax_require_defined.m4 create mode 100644 native/libffi/make_sunver.pl create mode 100644 native/libffi/msvc_build/aarch64/Ffi_staticLib.sln create mode 100644 native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj create mode 100644 native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj.filters create mode 100644 native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj.user create mode 100644 native/libffi/msvc_build/aarch64/aarch64_include/ffi.h create mode 100644 native/libffi/msvc_build/aarch64/aarch64_include/fficonfig.h create mode 100644 native/libffi/src/aarch64/internal.h create mode 100644 native/libffi/src/aarch64/win64_armasm.S create mode 100644 native/libffi/src/alpha/internal.h create mode 100644 native/libffi/src/arm/internal.h create mode 100644 native/libffi/src/arm/sysv_msvc_arm32.S delete mode 100644 native/libffi/src/arm/trampoline.S create mode 100644 native/libffi/src/csky/ffi.c create mode 100644 native/libffi/src/csky/ffitarget.h create mode 100644 native/libffi/src/csky/sysv.S create mode 100644 native/libffi/src/kvx/asm.h create mode 100644 native/libffi/src/kvx/ffi.c create mode 100644 native/libffi/src/kvx/ffitarget.h create mode 100644 native/libffi/src/kvx/sysv.S create mode 100644 native/libffi/src/riscv/ffi.c create mode 100644 native/libffi/src/riscv/ffitarget.h create mode 100644 native/libffi/src/riscv/sysv.S create mode 100644 native/libffi/src/s390/internal.h create mode 100644 native/libffi/src/sparc/ffi64.c create mode 100644 native/libffi/src/sparc/internal.h create mode 100644 native/libffi/src/x86/asmnames.h delete mode 100644 native/libffi/src/x86/darwin.S delete mode 100644 native/libffi/src/x86/darwin64.S create mode 100644 native/libffi/src/x86/ffiw64.c create mode 100644 native/libffi/src/x86/internal.h create mode 100644 native/libffi/src/x86/internal64.h create mode 100644 native/libffi/src/x86/sysv_intel.S create mode 100644 native/libffi/src/x86/win64_intel.S create mode 100644 native/libffi/testsuite/libffi.bhaible/Makefile create mode 100644 native/libffi/testsuite/libffi.bhaible/README create mode 100644 native/libffi/testsuite/libffi.bhaible/alignof.h create mode 100644 native/libffi/testsuite/libffi.bhaible/bhaible.exp create mode 100644 native/libffi/testsuite/libffi.bhaible/test-call.c create mode 100644 native/libffi/testsuite/libffi.bhaible/test-callback.c create mode 100644 native/libffi/testsuite/libffi.bhaible/testcases.c create mode 100644 native/libffi/testsuite/libffi.call/align_stdcall.c create mode 100644 native/libffi/testsuite/libffi.call/offsets.c create mode 100644 native/libffi/testsuite/libffi.call/pr1172638.c create mode 100644 native/libffi/testsuite/libffi.call/struct10.c create mode 100644 native/libffi/testsuite/libffi.closures/closure.exp rename native/libffi/testsuite/{libffi.call => libffi.closures}/closure_fn0.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/closure_fn1.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/closure_fn2.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/closure_fn3.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/closure_fn4.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/closure_fn5.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/closure_fn6.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/closure_loc_fn0.c (100%) create mode 100644 native/libffi/testsuite/libffi.closures/closure_simple.c rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_12byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_16byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_18byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_19byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_1_1byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_20byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_20byte1.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_24byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_2byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_3_1byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_3byte1.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_3byte2.c (100%) create mode 100644 native/libffi/testsuite/libffi.closures/cls_3float.c rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_4_1byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_4byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_5_1_byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_5byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_64byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_6_1_byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_6byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_7_1_byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_7byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_8byte.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_9byte1.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_9byte2.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_double.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_float.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_longdouble.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_longdouble_split.c (95%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_longdouble_split2.c (95%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_pointer.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_sint16.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_sint32.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_sint64.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_uint16.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_uint32.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_align_uint64.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_dbls_struct.c (95%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_double.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_double_va.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_float.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_longdouble.c (94%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_longdouble_va.c (100%) create mode 100644 native/libffi/testsuite/libffi.closures/cls_many_mixed_args.c create mode 100644 native/libffi/testsuite/libffi.closures/cls_many_mixed_float_double.c rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_multi_schar.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_multi_sshort.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_multi_sshortchar.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_multi_uchar.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_multi_ushort.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_multi_ushortchar.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_pointer.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_pointer_stack.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_schar.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_sint.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_sshort.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_struct_va1.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_uchar.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_uchar_va.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_uint.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_uint_va.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_ulong_va.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_ulonglong.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_ushort.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/cls_ushort_va.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/err_bad_abi.c (100%) create mode 100644 native/libffi/testsuite/libffi.closures/ffitest.h rename native/libffi/testsuite/{libffi.call => libffi.closures}/huge_struct.c (98%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct1.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct10.c (99%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct11.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct2.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct3.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct4.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct5.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct6.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct7.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct8.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/nested_struct9.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/problem1.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/stret_large.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/stret_large2.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/stret_medium.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/stret_medium2.c (100%) rename native/libffi/testsuite/{libffi.call => libffi.closures}/testclosure.c (100%) create mode 100644 native/libffi/testsuite/libffi.closures/unwindtest.cc create mode 100644 native/libffi/testsuite/libffi.closures/unwindtest_ffi_call.cc create mode 100644 native/libffi/testsuite/libffi.complex/cls_align_complex.inc create mode 100644 native/libffi/testsuite/libffi.complex/cls_align_complex_double.c create mode 100644 native/libffi/testsuite/libffi.complex/cls_align_complex_float.c create mode 100644 native/libffi/testsuite/libffi.complex/cls_align_complex_longdouble.c create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex.inc create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex_double.c create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex_float.c create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex_longdouble.c create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex_struct.inc create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex_struct_double.c create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex_struct_float.c create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex_struct_longdouble.c create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex_va.inc create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex_va_double.c create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex_va_float.c create mode 100644 native/libffi/testsuite/libffi.complex/cls_complex_va_longdouble.c create mode 100644 native/libffi/testsuite/libffi.complex/complex.exp create mode 100644 native/libffi/testsuite/libffi.complex/complex.inc create mode 100644 native/libffi/testsuite/libffi.complex/complex_defs_double.inc create mode 100644 native/libffi/testsuite/libffi.complex/complex_defs_float.inc create mode 100644 native/libffi/testsuite/libffi.complex/complex_defs_longdouble.inc create mode 100644 native/libffi/testsuite/libffi.complex/complex_double.c create mode 100644 native/libffi/testsuite/libffi.complex/complex_float.c create mode 100644 native/libffi/testsuite/libffi.complex/complex_int.c create mode 100644 native/libffi/testsuite/libffi.complex/complex_longdouble.c create mode 100644 native/libffi/testsuite/libffi.complex/ffitest.h create mode 100644 native/libffi/testsuite/libffi.complex/many_complex.inc create mode 100644 native/libffi/testsuite/libffi.complex/many_complex_double.c create mode 100644 native/libffi/testsuite/libffi.complex/many_complex_float.c create mode 100644 native/libffi/testsuite/libffi.complex/many_complex_longdouble.c create mode 100644 native/libffi/testsuite/libffi.complex/return_complex.inc create mode 100644 native/libffi/testsuite/libffi.complex/return_complex1.inc create mode 100644 native/libffi/testsuite/libffi.complex/return_complex1_double.c create mode 100644 native/libffi/testsuite/libffi.complex/return_complex1_float.c create mode 100644 native/libffi/testsuite/libffi.complex/return_complex1_longdouble.c create mode 100644 native/libffi/testsuite/libffi.complex/return_complex2.inc create mode 100644 native/libffi/testsuite/libffi.complex/return_complex2_double.c create mode 100644 native/libffi/testsuite/libffi.complex/return_complex2_float.c create mode 100644 native/libffi/testsuite/libffi.complex/return_complex2_longdouble.c create mode 100644 native/libffi/testsuite/libffi.complex/return_complex_double.c create mode 100644 native/libffi/testsuite/libffi.complex/return_complex_float.c create mode 100644 native/libffi/testsuite/libffi.complex/return_complex_longdouble.c create mode 100644 native/libffi/testsuite/libffi.go/aa-direct.c create mode 100644 native/libffi/testsuite/libffi.go/closure1.c create mode 100644 native/libffi/testsuite/libffi.go/ffitest.h create mode 100644 native/libffi/testsuite/libffi.go/go.exp create mode 100644 native/libffi/testsuite/libffi.go/static-chain.h diff --git a/native/libffi/.appveyor.yml b/native/libffi/.appveyor.yml new file mode 100644 index 0000000000..ece8a948c9 --- /dev/null +++ b/native/libffi/.appveyor.yml @@ -0,0 +1,66 @@ +shallow_clone: true + +# We're currently only testing libffi built with Microsoft's +# tools. +# This matrix should be expanded to include at least: +# 32- and 64-bit gcc/cygwin +# 32- and 64-bit gcc/mingw +# 32- and 64-bit clang/mingw +# and perhaps more. + +image: Visual Studio 2017 +platform: + - x64 + - x86 + - arm + - arm64 + +environment: + global: + CYG_ROOT: C:/cygwin + CYG_CACHE: C:/cygwin/var/cache/setup + CYG_MIRROR: http://mirrors.kernel.org/sourceware/cygwin/ + matrix: + - VSVER: 15 + +install: + - ps: >- + If ($env:Platform -Match "x86") { + $env:VCVARS_PLATFORM="x86" + $env:BUILD="i686-pc-cygwin" + $env:HOST="i686-pc-cygwin" + $env:MSVCC="/cygdrive/c/projects/libffi/msvcc.sh" + $env:SRC_ARCHITECTURE="x86" + } ElseIf ($env:Platform -Match "arm64") { + $env:VCVARS_PLATFORM="x86_arm64" + $env:BUILD="i686-pc-cygwin" + $env:HOST="aarch64-w64-cygwin" + $env:MSVCC="/cygdrive/c/projects/libffi/msvcc.sh -marm64" + $env:SRC_ARCHITECTURE="aarch64" + } ElseIf ($env:Platform -Match "arm") { + $env:VCVARS_PLATFORM="x86_arm" + $env:BUILD="i686-pc-cygwin" + $env:HOST="arm-w32-cygwin" + $env:MSVCC="/cygdrive/c/projects/libffi/msvcc.sh -marm" + $env:SRC_ARCHITECTURE="arm" + } Else { + $env:VCVARS_PLATFORM="amd64" + $env:BUILD="x86_64-w64-cygwin" + $env:HOST="x86_64-w64-cygwin" + $env:MSVCC="/cygdrive/c/projects/libffi/msvcc.sh -m64" + $env:SRC_ARCHITECTURE="x86" + } + - 'appveyor DownloadFile https://cygwin.com/setup-x86.exe -FileName setup.exe' + - 'setup.exe -qnNdO -R "%CYG_ROOT%" -s "%CYG_MIRROR%" -l "%CYG_CACHE%" -P dejagnu >NUL' + - '%CYG_ROOT%/bin/bash -lc "cygcheck -dc cygwin"' + - echo call VsDevCmd to set VS150COMNTOOLS + - call "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\Common7\Tools\VsDevCmd.bat" + - ps: $env:VSCOMNTOOLS=(Get-Content ("env:VS" + "$env:VSVER" + "0COMNTOOLS")) + - echo "Using Visual Studio %VSVER%.0 at %VSCOMNTOOLS%" + - call "%VSCOMNTOOLS%..\..\vc\Auxiliary\Build\vcvarsall.bat" %VCVARS_PLATFORM% + +build_script: + - c:\cygwin\bin\sh -lc "(cd $OLDPWD; ./autogen.sh;)" + - c:\cygwin\bin\sh -lc "(cd $OLDPWD; ./configure CC='%MSVCC%' CXX='%MSVCC%' LD='link' CPP='cl -nologo -EP' CXXCPP='cl -nologo -EP' CPPFLAGS='-DFFI_BUILDING_DLL' AR='/cygdrive/c/projects/libffi/.travis/ar-lib lib' NM='dumpbin -symbols' STRIP=':' --build=$BUILD --host=$HOST;)" + - c:\cygwin\bin\sh -lc "(cd $OLDPWD; cp src/%SRC_ARCHITECTURE%/ffitarget.h include; make; find .;)" + - c:\cygwin\bin\sh -lc "(cd $OLDPWD; cp `find . -name 'libffi-?.dll'` $HOST/testsuite/; make check; cat `find ./ -name libffi.log`)" diff --git a/native/libffi/.gitattributes b/native/libffi/.gitattributes new file mode 100644 index 0000000000..f7d3833e78 --- /dev/null +++ b/native/libffi/.gitattributes @@ -0,0 +1,4 @@ +* text=auto + +*.sln text eol=crlf +*.vcxproj* text eol=crlf diff --git a/native/libffi/.github/issue_template.md b/native/libffi/.github/issue_template.md new file mode 100644 index 0000000000..e197e2cea7 --- /dev/null +++ b/native/libffi/.github/issue_template.md @@ -0,0 +1,10 @@ +## System Details + + + + +## Problems Description + + + + diff --git a/native/libffi/.gitignore b/native/libffi/.gitignore index 1158ecef28..0f90100eaa 100644 --- a/native/libffi/.gitignore +++ b/native/libffi/.gitignore @@ -5,10 +5,11 @@ .dirstamp *.la Makefile +!testsuite/libffi.bhaible/Makefile Makefile.in aclocal.m4 compile -config.* +!.travis/compile configure depcomp doc/libffi.info @@ -33,3 +34,5 @@ libffi.xcodeproj/xcuserdata libffi.xcodeproj/project.xcworkspace build_*/ darwin_*/ +src/arm/trampoline.S +**/texinfo.tex \ No newline at end of file diff --git a/native/libffi/.travis.yml b/native/libffi/.travis.yml index 6c14fcd0b0..e0a69fa778 100644 --- a/native/libffi/.travis.yml +++ b/native/libffi/.travis.yml @@ -1,22 +1,83 @@ +--- +sudo: required + language: cpp -compiler: - - gcc - - clang -env: - - - - HOST=i386-pc-linux-gnu + +# For qemu-powered targets, get the list of supported processors from +# travis by setting QEMU_CPU=help, then set -mcpu= for the compilers +# accordingly. + matrix: - exclude: - - compiler: clang - env: HOST=i386-pc-linux-gnu + include: + - os: linux + env: HOST=powerpc-eabisim RUNTESTFLAGS="--target_board powerpc-eabisim" DEJAGNU="/opt/.travis/site.exp" + - os: linux + env: HOST=or1k-elf RUNTESTFLAGS="--target_board or1k-sim" DEJAGNU="/opt/.travis/site.exp" + - os: linux + env: HOST=m32r-elf RUNTESTFLAGS="--target_board m32r-sim" DEJAGNU="/opt/.travis/site.exp" + - os: linux + env: HOST=bfin-elf RUNTESTFLAGS="--target_board bfin-sim" DEJAGNU="/opt/.travis/site.exp" +# This configuration is still using the native x86 toolchain? +# - os: osx +# env: HOST=aarch64-apple-darwin13 + - os: osx + env: HOST=x86_64-apple-darwin10 + - os: linux + env: HOST=x86_64-w64-mingw32 MEVAL='export CC="x86_64-w64-mingw32-gcc" && CXX="x86_64-w64-mingw32-g++" RUNTESTFLAGS="--target_board wine-sim" DEJAGNU="$TRAVIS_BUILD_DIR/.travis/site.exp" CONFIGURE_OPTIONS=--disable-shared LIBFFI_TEST_OPTIMIZATION="-O2" + - os: linux + env: HOST=sh4-linux-gnu CONFIGURE_OPTIONS=--disable-shared QEMU_LD_PREFIX=/usr/sh4-linux-gnu + - os: linux + env: HOST=alpha-linux-gnu CONFIGURE_OPTIONS=--disable-shared QEMU_LD_PREFIX=/usr/alpha-linux-gnu + - os: linux + env: HOST=m68k-linux-gnu MEVAL='export CC="m68k-linux-gnu-gcc-8 -mcpu=547x" && CXX="m68k-linux-gnu-g++-8 -mcpu=547x"' CONFIGURE_OPTIONS=--disable-shared QEMU_LD_PREFIX=/usr/m68k-linux-gnu QEMU_CPU=cfv4e + - os: linux + arch: s390x + env: HOST=s390x-linux-gnu + - os: linux + arch: ppc64le + env: HOST=ppc64le-linux-gnu + - os: linux + arch: arm64 + env: HOST=aarch64-linux-gnu + - os: linux + arch: arm64 + env: HOST=aarch64-linux-gnu + compiler: clang + - os: linux + env: HOST=arm32v7-linux-gnu LIBFFI_TEST_OPTIMIZATION="-O0" + - os: linux + env: HOST=arm32v7-linux-gnu LIBFFI_TEST_OPTIMIZATION="-O2" + - os: linux + env: HOST=arm32v7-linux-gnu LIBFFI_TEST_OPTIMIZATION="-O2 -fomit-frame-pointer" +# The sparc64 linux system in the GCC compile farm is non-responsive. +# - os: linux +# env: HOST=sparc64-linux-gnu +# The mips64 linux system in the GCC compile farm is not allowing logins +# - os: linux +# env: HOST=mips64el-linux-gnu + - os: linux + compiler: gcc + env: HOST=i386-pc-linux-gnu MEVAL='export CC="$CC -m32" && CXX="$CXX -m32"' + - os: linux + compiler: gcc + - os: linux + compiler: gcc + env: CONFIGURE_OPTIONS=--disable-shared + - os: linux + compiler: clang + - os: linux + compiler: clang + env: CONFIGURE_OPTIONS=--disable-shared + - os: linux + env: HOST=moxie-elf MEVAL='export PATH=/opt/moxielogic/bin:$PATH && CC=moxie-elf-gcc && CXX=moxie-elf-g++' LDFLAGS=-Tsim.ld RUNTESTFLAGS="--target_board moxie-sim" DEJAGNU="$TRAVIS_BUILD_DIR/.travis/site.exp" + +before_install: + - if test x"$MEVAL" != x; then eval ${MEVAL}; fi -before_script: - - sudo apt-get install dejagnu texinfo - - if [ "$HOST" = i386-pc-linux-gnu ] ; then sudo apt-get install gcc-multilib g++-multilib && CC="$CC -m32" && CXX="$CXX -m32" ; fi +install: + - travis_wait 30 ./.travis/install.sh script: - - ./autogen.sh - - ./configure ${HOST+--host=$HOST} - - make - - make dist - - make check + - if ! test x"$MEVAL" = x; then eval ${MEVAL}; fi + - travis_wait 115 sleep infinity & + - ./.travis/build.sh \ No newline at end of file diff --git a/native/libffi/.travis/ar-lib b/native/libffi/.travis/ar-lib new file mode 100755 index 0000000000..0baa4f6076 --- /dev/null +++ b/native/libffi/.travis/ar-lib @@ -0,0 +1,270 @@ +#! /bin/sh +# Wrapper for Microsoft lib.exe + +me=ar-lib +scriptversion=2012-03-01.08; # UTC + +# Copyright (C) 2010-2018 Free Software Foundation, Inc. +# Written by Peter Rosin . +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2, or (at your option) +# any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +# As a special exception to the GNU General Public License, if you +# distribute this file as part of a program that contains a +# configuration script generated by Autoconf, you may include it under +# the same distribution terms that you use for the rest of that program. + +# This file is maintained in Automake, please report +# bugs to or send patches to +# . + + +# func_error message +func_error () +{ + echo "$me: $1" 1>&2 + exit 1 +} + +file_conv= + +# func_file_conv build_file +# Convert a $build file to $host form and store it in $file +# Currently only supports Windows hosts. +func_file_conv () +{ + file=$1 + case $file in + / | /[!/]*) # absolute file, and not a UNC file + if test -z "$file_conv"; then + # lazily determine how to convert abs files + case `uname -s` in + MINGW*) + file_conv=mingw + ;; + CYGWIN*) + file_conv=cygwin + ;; + *) + file_conv=wine + ;; + esac + fi + case $file_conv in + mingw) + file=`cmd //C echo "$file " | sed -e 's/"\(.*\) " *$/\1/'` + ;; + cygwin) + file=`cygpath -m "$file" || echo "$file"` + ;; + wine) + file=`winepath -w "$file" || echo "$file"` + ;; + esac + ;; + esac +} + +# func_at_file at_file operation archive +# Iterate over all members in AT_FILE performing OPERATION on ARCHIVE +# for each of them. +# When interpreting the content of the @FILE, do NOT use func_file_conv, +# since the user would need to supply preconverted file names to +# binutils ar, at least for MinGW. +func_at_file () +{ + operation=$2 + archive=$3 + at_file_contents=`cat "$1"` + eval set x "$at_file_contents" + shift + + for member + do + $AR -NOLOGO $operation:"$member" "$archive" || exit $? + done +} + +case $1 in + '') + func_error "no command. Try '$0 --help' for more information." + ;; + -h | --h*) + cat < libffi.log + + ./rlgl l --key=${RLGL_KEY} https://rl.gl + ID=$(./rlgl start) + ./rlgl e --id=$ID --policy=https://github.com/libffi/rlgl-policy.git libffi.log + exit $? +} + +function build_linux() +{ + ./autogen.sh + ./configure ${HOST+--host=$HOST} ${CONFIGURE_OPTIONS} || cat */config.log + make + make dist + make check RUNTESTFLAGS="-a $RUNTESTFLAGS" + + ./rlgl l --key=${RLGL_KEY} https://rl.gl + ID=$(./rlgl start) + ./rlgl e --id=$ID --policy=https://github.com/libffi/rlgl-policy.git */testsuite/libffi.log + exit $? +} + +function build_foreign_linux() +{ + ${DOCKER} run --rm -t -i -v $(pwd):/opt ${SET_QEMU_CPU} -e LIBFFI_TEST_OPTIMIZATION="${LIBFFI_TEST_OPTIMIZATION}" $2 bash -c /opt/.travis/build-in-container.sh + + ./rlgl l --key=${RLGL_KEY} https://rl.gl + ID=$(./rlgl start) + ./rlgl e --id=$ID --policy=https://github.com/libffi/rlgl-policy.git */testsuite/libffi.log + exit $? +} + +function build_cross_linux() +{ + ${DOCKER} run --rm -t -i -v $(pwd):/opt ${SET_QEMU_CPU} -e HOST="${HOST}" -e CC="${HOST}-gcc-8 ${GCC_OPTIONS}" -e CXX="${HOST}-g++-8 ${GCC_OPTIONS}" -e LIBFFI_TEST_OPTIMIZATION="${LIBFFI_TEST_OPTIMIZATION}" moxielogic/cross-ci-build-container:latest bash -c /opt/.travis/build-in-container.sh + + ./rlgl l --key=${RLGL_KEY} https://rl.gl + ID=$(./rlgl start) + ./rlgl e --id=$ID --policy=https://github.com/libffi/rlgl-policy.git */testsuite/libffi.log + exit $? +} + +function build_cross() +{ + ${DOCKER} pull quay.io/moxielogic/libffi-ci-${HOST} + ${DOCKER} run --rm -t -i -v $(pwd):/opt -e HOST="${HOST}" -e CC="${HOST}-gcc ${GCC_OPTIONS}" -e CXX="${HOST}-g++ ${GCC_OPTIONS}" -e TRAVIS_BUILD_DIR=/opt -e DEJAGNU="${DEJAGNU}" -e RUNTESTFLAGS="${RUNTESTFLAGS}" -e LIBFFI_TEST_OPTIMIZATION="${LIBFFI_TEST_OPTIMIZATION}" quay.io/moxielogic/libffi-ci-${HOST} bash -c /opt/.travis/build-cross-in-container.sh + + ./rlgl l --key=${RLGL_KEY} https://rl.gl + ID=$(./rlgl start) + ./rlgl e --id=$ID --policy=https://github.com/libffi/rlgl-policy.git */testsuite/libffi.log + exit $? +} + +function build_ios() +{ + which python +# export PYTHON_BIN=/usr/local/bin/python + ./generate-darwin-source-and-headers.py --only-ios + xcodebuild -showsdks + xcodebuild -project libffi.xcodeproj -target "libffi-iOS" -configuration Release -sdk iphoneos11.4 + exit $? +} + +function build_macosx() +{ + which python +# export PYTHON_BIN=/usr/local/bin/python + ./generate-darwin-source-and-headers.py --only-osx + xcodebuild -showsdks + xcodebuild -project libffi.xcodeproj -target "libffi-Mac" -configuration Release -sdk macosx10.13 + echo "Finished build" + exit $? +} + +case "$HOST" in + arm-apple-darwin*) + ./autogen.sh + build_ios + ;; + x86_64-apple-darwin*) + ./autogen.sh + build_macosx + ;; + arm32v7-linux-gnu) + ./autogen.sh + build_foreign_linux arm moxielogic/arm32v7-ci-build-container:latest + ;; + mips64el-linux-gnu | sparc64-linux-gnu) + build_cfarm + ;; + bfin-elf ) + ./autogen.sh + GCC_OPTIONS=-msim build_cross + ;; + m32r-elf ) + ./autogen.sh + build_cross + ;; + or1k-elf ) + ./autogen.sh + build_cross + ;; + powerpc-eabisim ) + ./autogen.sh + build_cross + ;; + m68k-linux-gnu ) + ./autogen.sh + GCC_OPTIONS=-mcpu=547x build_cross_linux + ;; + alpha-linux-gnu | sh4-linux-gnu ) + ./autogen.sh + build_cross_linux + ;; + *) + ./autogen.sh + build_linux + ;; +esac diff --git a/native/libffi/.travis/compile b/native/libffi/.travis/compile new file mode 100755 index 0000000000..655932a224 --- /dev/null +++ b/native/libffi/.travis/compile @@ -0,0 +1,351 @@ +#! /bin/sh +# Wrapper for compilers which do not understand '-c -o'. + +scriptversion=2018-03-27.18; # UTC + +# Copyright (C) 1999-2018 Free Software Foundation, Inc. +# Written by Tom Tromey . +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2, or (at your option) +# any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +# As a special exception to the GNU General Public License, if you +# distribute this file as part of a program that contains a +# configuration script generated by Autoconf, you may include it under +# the same distribution terms that you use for the rest of that program. + +# This file is maintained in Automake, please report +# bugs to or send patches to +# . + +nl=' +' + +# We need space, tab and new line, in precisely that order. Quoting is +# there to prevent tools from complaining about whitespace usage. +IFS=" "" $nl" + +file_conv= + +# func_file_conv build_file lazy +# Convert a $build file to $host form and store it in $file +# Currently only supports Windows hosts. If the determined conversion +# type is listed in (the comma separated) LAZY, no conversion will +# take place. +func_file_conv () +{ + file=$1 + case $file in + / | /[!/]*) # absolute file, and not a UNC file + if test -z "$file_conv"; then + # lazily determine how to convert abs files + case `uname -s` in + MINGW*) + file_conv=mingw + ;; + CYGWIN*) + file_conv=cygwin + ;; + *) + file_conv=wine + ;; + esac + fi + case $file_conv/,$2, in + *,$file_conv,*) + ;; + mingw/*) + file=`cmd //C echo "$file " | sed -e 's/"\(.*\) " *$/\1/'` + ;; + cygwin/*) + file=`cygpath -m "$file" || echo "$file"` + ;; + wine/*) + file=`winepath -w "$file" || echo "$file"` + ;; + esac + ;; + esac +} + +# func_cl_dashL linkdir +# Make cl look for libraries in LINKDIR +func_cl_dashL () +{ + func_file_conv "$1" + if test -z "$lib_path"; then + lib_path=$file + else + lib_path="$lib_path;$file" + fi + linker_opts="$linker_opts -LIBPATH:$file" +} + +# func_cl_dashl library +# Do a library search-path lookup for cl +func_cl_dashl () +{ + lib=$1 + found=no + save_IFS=$IFS + IFS=';' + for dir in $lib_path $LIB + do + IFS=$save_IFS + if $shared && test -f "$dir/$lib.dll.lib"; then + found=yes + lib=$dir/$lib.dll.lib + break + fi + if test -f "$dir/$lib.lib"; then + found=yes + lib=$dir/$lib.lib + break + fi + if test -f "$dir/lib$lib.a"; then + found=yes + lib=$dir/lib$lib.a + break + fi + done + IFS=$save_IFS + + if test "$found" != yes; then + lib=$lib.lib + fi +} + +# func_cl_wrapper cl arg... +# Adjust compile command to suit cl +func_cl_wrapper () +{ + # Assume a capable shell + lib_path= + shared=: + linker_opts= + for arg + do + if test -n "$eat"; then + eat= + else + case $1 in + -o) + # configure might choose to run compile as 'compile cc -o foo foo.c'. + eat=1 + case $2 in + *.o | *.[oO][bB][jJ]) + func_file_conv "$2" + set x "$@" -Fo"$file" + shift + ;; + *) + func_file_conv "$2" + set x "$@" -Fe"$file" + shift + ;; + esac + ;; + -I) + eat=1 + func_file_conv "$2" mingw + set x "$@" -I"$file" + shift + ;; + -I*) + func_file_conv "${1#-I}" mingw + set x "$@" -I"$file" + shift + ;; + -l) + eat=1 + func_cl_dashl "$2" + set x "$@" "$lib" + shift + ;; + -l*) + func_cl_dashl "${1#-l}" + set x "$@" "$lib" + shift + ;; + -L) + eat=1 + func_cl_dashL "$2" + ;; + -L*) + func_cl_dashL "${1#-L}" + ;; + -static) + shared=false + ;; + -warn) + eat=1 + ;; + -Wl,*) + arg=${1#-Wl,} + save_ifs="$IFS"; IFS=',' + for flag in $arg; do + IFS="$save_ifs" + linker_opts="$linker_opts $flag" + done + IFS="$save_ifs" + ;; + -Xlinker) + eat=1 + linker_opts="$linker_opts $2" + ;; + -*) + set x "$@" "$1" + shift + ;; + *.cc | *.CC | *.cxx | *.CXX | *.[cC]++) + func_file_conv "$1" + set x "$@" -Tp"$file" + shift + ;; + *.c | *.cpp | *.CPP | *.lib | *.LIB | *.Lib | *.OBJ | *.obj | *.[oO]) + func_file_conv "$1" mingw + set x "$@" "$file" + shift + ;; + *) + set x "$@" "$1" + shift + ;; + esac + fi + shift + done + if test -n "$linker_opts"; then + linker_opts="-link$linker_opts" + fi + exec "$@" $linker_opts + exit 1 +} + +eat= + +case $1 in + '') + echo "$0: No command. Try '$0 --help' for more information." 1>&2 + exit 1; + ;; + -h | --h*) + cat <<\EOF +Usage: compile [--help] [--version] PROGRAM [ARGS] + +Wrapper for compilers which do not understand '-c -o'. +Remove '-o dest.o' from ARGS, run PROGRAM with the remaining +arguments, and rename the output as expected. + +If you are trying to build a whole package this is not the +right script to run: please start by reading the file 'INSTALL'. + +Report bugs to . +EOF + exit $? + ;; + -v | --v*) + echo "compile $scriptversion" + exit $? + ;; + cl | *[/\\]cl | cl.exe | *[/\\]cl.exe | \ + icl | *[/\\]icl | icl.exe | *[/\\]icl.exe ) + func_cl_wrapper "$@" # Doesn't return... + ;; +esac + +ofile= +cfile= + +for arg +do + if test -n "$eat"; then + eat= + else + case $1 in + -o) + # configure might choose to run compile as 'compile cc -o foo foo.c'. + # So we strip '-o arg' only if arg is an object. + eat=1 + case $2 in + *.o | *.obj) + ofile=$2 + ;; + *) + set x "$@" -o "$2" + shift + ;; + esac + ;; + *.c) + cfile=$1 + set x "$@" "$1" + shift + ;; + *) + set x "$@" "$1" + shift + ;; + esac + fi + shift +done + +if test -z "$ofile" || test -z "$cfile"; then + # If no '-o' option was seen then we might have been invoked from a + # pattern rule where we don't need one. That is ok -- this is a + # normal compilation that the losing compiler can handle. If no + # '.c' file was seen then we are probably linking. That is also + # ok. + exec "$@" +fi + +# Name of file we expect compiler to create. +cofile=`echo "$cfile" | sed 's|^.*[\\/]||; s|^[a-zA-Z]:||; s/\.c$/.o/'` + +# Create the lock directory. +# Note: use '[/\\:.-]' here to ensure that we don't use the same name +# that we are using for the .o file. Also, base the name on the expected +# object file name, since that is what matters with a parallel build. +lockdir=`echo "$cofile" | sed -e 's|[/\\:.-]|_|g'`.d +while true; do + if mkdir "$lockdir" >/dev/null 2>&1; then + break + fi + sleep 1 +done +# FIXME: race condition here if user kills between mkdir and trap. +trap "rmdir '$lockdir'; exit 1" 1 2 15 + +# Run the compile. +"$@" +ret=$? + +if test -f "$cofile"; then + test "$cofile" = "$ofile" || mv "$cofile" "$ofile" +elif test -f "${cofile}bj"; then + test "${cofile}bj" = "$ofile" || mv "${cofile}bj" "$ofile" +fi + +rmdir "$lockdir" +exit $ret + +# Local Variables: +# mode: shell-script +# sh-indentation: 2 +# eval: (add-hook 'before-save-hook 'time-stamp) +# time-stamp-start: "scriptversion=" +# time-stamp-format: "%:y-%02m-%02d.%02H" +# time-stamp-time-zone: "UTC0" +# time-stamp-end: "; # UTC" +# End: diff --git a/native/libffi/.travis/install.sh b/native/libffi/.travis/install.sh new file mode 100755 index 0000000000..2420245710 --- /dev/null +++ b/native/libffi/.travis/install.sh @@ -0,0 +1,71 @@ +#!/bin/bash +set -x + +if [[ $TRAVIS_OS_NAME != 'linux' ]]; then + brew update > brew-update.log 2>&1 + # fix an issue with libtool on travis by reinstalling it + brew uninstall libtool; + brew install libtool dejagnu; + + # Download and extract the rlgl client + wget -qO - https://rl.gl/cli/rlgl-darwin-amd64.tgz | \ + tar --strip-components=2 -xvzf - ./rlgl/rlgl; + +else + # Download and extract the rlgl client + case $HOST in + aarch64-linux-gnu) + wget -qO - https://rl.gl/cli/rlgl-linux-arm.tgz | \ + tar --strip-components=2 -xvzf - ./rlgl/rlgl; + ;; + ppc64le-linux-gnu) + wget -qO - https://rl.gl/cli/rlgl-linux-ppc64le.tgz | \ + tar --strip-components=2 -xvzf - ./rlgl/rlgl; + ;; + s390x-linux-gnu) + wget -qO - https://rl.gl/cli/rlgl-linux-s390x.tgz | \ + tar --strip-components=2 -xvzf - ./rlgl/rlgl; + ;; + *) + wget -qO - https://rl.gl/cli/rlgl-linux-amd64.tgz | \ + tar --strip-components=2 -xvzf - ./rlgl/rlgl; + ;; + esac + + sudo apt-get clean # clear the cache + sudo apt-get update + case $HOST in + mips64el-linux-gnu | sparc64-linux-gnu) + ;; + alpha-linux-gnu | arm32v7-linux-gnu | m68k-linux-gnu | sh4-linux-gnu) + sudo apt-get install qemu-user-static + ;; + hppa-linux-gnu ) + sudo apt-get install -y qemu-user-static g++-5-hppa-linux-gnu + ;; + i386-pc-linux-gnu) + sudo apt-get install gcc-multilib g++-multilib; + ;; + moxie-elf) + echo 'deb https://repos.moxielogic.org:7114/MoxieLogic moxiedev main' | sudo tee -a /etc/apt/sources.list + sudo apt-get clean # clear the cache + sudo apt-get update ## -qq + sudo apt-get update + sudo apt-get install -y --allow-unauthenticated moxielogic-moxie-elf-gcc moxielogic-moxie-elf-gcc-c++ moxielogic-moxie-elf-gcc-libstdc++ moxielogic-moxie-elf-gdb-sim + ;; + x86_64-w64-mingw32) + sudo apt-get install gcc-mingw-w64-x86-64 g++-mingw-w64-x86-64 wine; + ;; + i686-w32-mingw32) + sudo apt-get install gcc-mingw-w64-i686 g++-mingw-w64-i686 wine; + ;; + esac + case $HOST in + arm32v7-linux-gnu) + # don't install host tools + ;; + *) + sudo apt-get install dejagnu texinfo sharutils + ;; + esac +fi diff --git a/native/libffi/.travis/m32r-sim.exp b/native/libffi/.travis/m32r-sim.exp new file mode 100644 index 0000000000..c18123f28f --- /dev/null +++ b/native/libffi/.travis/m32r-sim.exp @@ -0,0 +1,58 @@ +# Copyright (C) 2010, 2019 Free Software Foundation, Inc. +# +# This file is part of DejaGnu. +# +# DejaGnu is free software; you can redistribute it and/or modify it +# under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# DejaGnu is distributed in the hope that it will be useful, but +# WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with DejaGnu; if not, write to the Free Software Foundation, +# Inc., 51 Franklin Street, Fifth Floor, MA 02110, USA. + +# This is a list of toolchains that are supported on this board. +set_board_info target_install {m32r-elf} + +# Load the generic configuration for this board. This will define a basic set +# of routines needed by the tool to communicate with the board. +load_generic_config "sim" + +# basic-sim.exp is a basic description for the standard Cygnus simulator. +load_base_board_description "basic-sim" + +# "m32r" is the name of the sim subdir in devo/sim. +setup_sim m32r + +# No multilib options needed by default. +process_multilib_options "" + +# We only support newlib on this target. We assume that all multilib +# options have been specified before we get here. + +set_board_info compiler "[find_gcc]" +set_board_info cflags "[libgloss_include_flags] [newlib_include_flags]" +set_board_info ldflags "[libgloss_link_flags] [newlib_link_flags]" + +# Configuration settings for testsuites +set_board_info noargs 1 +set_board_info gdb,nosignals 1 +set_board_info gdb,noresults 1 +set_board_info gdb,cannot_call_functions 1 +set_board_info gdb,skip_float_tests 1 +set_board_info gdb,can_reverse 1 +set_board_info gdb,use_precord 1 + +# More time is needed +set_board_info gcc,timeout 800 +set_board_info gdb,timeout 60 + +# Used by a few gcc.c-torture testcases to delimit how large the stack can +# be. +set_board_info gcc,stack_size 5000 + diff --git a/native/libffi/.travis/moxie-sim.exp b/native/libffi/.travis/moxie-sim.exp new file mode 100644 index 0000000000..32979ea6be --- /dev/null +++ b/native/libffi/.travis/moxie-sim.exp @@ -0,0 +1,60 @@ +# Copyright (C) 2010 Free Software Foundation, Inc. +# +# This file is part of DejaGnu. +# +# DejaGnu is free software; you can redistribute it and/or modify it +# under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# DejaGnu is distributed in the hope that it will be useful, but +# WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with DejaGnu; if not, write to the Free Software Foundation, +# Inc., 51 Franklin Street, Fifth Floor, MA 02110, USA. + +# This is a list of toolchains that are supported on this board. +set_board_info target_install {moxie-elf} + +# Load the generic configuration for this board. This will define a basic set +# of routines needed by the tool to communicate with the board. +load_generic_config "sim" + +# basic-sim.exp is a basic description for the standard Cygnus simulator. +load_base_board_description "basic-sim" + +# "moxie" is the name of the sim subdir in devo/sim. +setup_sim moxie + +# No multilib options needed by default. +process_multilib_options "" + +# We only support newlib on this target. We assume that all multilib +# options have been specified before we get here. + +set_board_info compiler "[find_gcc]" +set_board_info cflags "[libgloss_include_flags] [newlib_include_flags]" +set_board_info ldflags "[libgloss_link_flags] [newlib_link_flags]" +# No linker script needed. +set_board_info ldscript "-Tsim.ld" + +# Configuration settings for testsuites +set_board_info noargs 1 +set_board_info gdb,nosignals 1 +set_board_info gdb,noresults 1 +set_board_info gdb,cannot_call_functions 1 +set_board_info gdb,skip_float_tests 1 +set_board_info gdb,can_reverse 1 +set_board_info gdb,use_precord 1 + +# More time is needed +set_board_info gcc,timeout 800 +set_board_info gdb,timeout 60 + +# Used by a few gcc.c-torture testcases to delimit how large the stack can +# be. +set_board_info gcc,stack_size 5000 + diff --git a/native/libffi/.travis/or1k-sim.exp b/native/libffi/.travis/or1k-sim.exp new file mode 100644 index 0000000000..3920413ee1 --- /dev/null +++ b/native/libffi/.travis/or1k-sim.exp @@ -0,0 +1,58 @@ +# Copyright (C) 2010, 2019 Free Software Foundation, Inc. +# +# This file is part of DejaGnu. +# +# DejaGnu is free software; you can redistribute it and/or modify it +# under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# DejaGnu is distributed in the hope that it will be useful, but +# WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with DejaGnu; if not, write to the Free Software Foundation, +# Inc., 51 Franklin Street, Fifth Floor, MA 02110, USA. + +# This is a list of toolchains that are supported on this board. +set_board_info target_install {or1k-elf} + +# Load the generic configuration for this board. This will define a basic set +# of routines needed by the tool to communicate with the board. +load_generic_config "sim" + +# basic-sim.exp is a basic description for the standard Cygnus simulator. +load_base_board_description "basic-sim" + +# "or1k" is the name of the sim subdir in devo/sim. +setup_sim or1k + +# No multilib options needed by default. +process_multilib_options "" + +# We only support newlib on this target. We assume that all multilib +# options have been specified before we get here. + +set_board_info compiler "[find_gcc]" +set_board_info cflags "[libgloss_include_flags] [newlib_include_flags]" +set_board_info ldflags "[libgloss_link_flags] [newlib_link_flags]" + +# Configuration settings for testsuites +set_board_info noargs 1 +set_board_info gdb,nosignals 1 +set_board_info gdb,noresults 1 +set_board_info gdb,cannot_call_functions 1 +set_board_info gdb,skip_float_tests 1 +set_board_info gdb,can_reverse 1 +set_board_info gdb,use_precord 1 + +# More time is needed +set_board_info gcc,timeout 800 +set_board_info gdb,timeout 60 + +# Used by a few gcc.c-torture testcases to delimit how large the stack can +# be. +set_board_info gcc,stack_size 5000 + diff --git a/native/libffi/.travis/powerpc-eabisim.exp b/native/libffi/.travis/powerpc-eabisim.exp new file mode 100644 index 0000000000..285fd4f6fe --- /dev/null +++ b/native/libffi/.travis/powerpc-eabisim.exp @@ -0,0 +1,58 @@ +# Copyright (C) 2010, 2019 Free Software Foundation, Inc. +# +# This file is part of DejaGnu. +# +# DejaGnu is free software; you can redistribute it and/or modify it +# under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# DejaGnu is distributed in the hope that it will be useful, but +# WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with DejaGnu; if not, write to the Free Software Foundation, +# Inc., 51 Franklin Street, Fifth Floor, MA 02110, USA. + +# This is a list of toolchains that are supported on this board. +set_board_info target_install {powerpc-eabisim} + +# Load the generic configuration for this board. This will define a basic set +# of routines needed by the tool to communicate with the board. +load_generic_config "sim" + +# basic-sim.exp is a basic description for the standard Cygnus simulator. +load_base_board_description "basic-sim" + +# "powerpc" is the name of the sim subdir in devo/sim. +setup_sim powerpc + +# No multilib options needed by default. +process_multilib_options "" + +# We only support newlib on this target. We assume that all multilib +# options have been specified before we get here. + +set_board_info compiler "[find_gcc]" +set_board_info cflags "[libgloss_include_flags] [newlib_include_flags]" +set_board_info ldflags "[libgloss_link_flags] [newlib_link_flags]" + +# Configuration settings for testsuites +set_board_info noargs 1 +set_board_info gdb,nosignals 1 +set_board_info gdb,noresults 1 +set_board_info gdb,cannot_call_functions 1 +set_board_info gdb,skip_float_tests 1 +set_board_info gdb,can_reverse 1 +set_board_info gdb,use_precord 1 + +# More time is needed +set_board_info gcc,timeout 800 +set_board_info gdb,timeout 60 + +# Used by a few gcc.c-torture testcases to delimit how large the stack can +# be. +set_board_info gcc,stack_size 5000 + diff --git a/native/libffi/.travis/site.exp b/native/libffi/.travis/site.exp new file mode 100644 index 0000000000..644ec63185 --- /dev/null +++ b/native/libffi/.travis/site.exp @@ -0,0 +1,27 @@ +# Copyright (C) 2008, 2010, 2018, 2019 Anthony Green + +# Make sure we look in the right place for the board description files. +if ![info exists boards_dir] { + set boards_dir {} +} + +lappend boards_dir $::env(TRAVIS_BUILD_DIR)/.travis + +verbose "Global Config File: target_triplet is $target_triplet" 2 +global target_list + +case "$target_triplet" in { + { "bfin-elf" } { + set target_list "bfin-sim" + } + { "m32r-elf" } { + set target_list "m32r-sim" + } + { "moxie-elf" } { + set target_list "moxie-sim" + } + { "or1k-elf" } { + set target_list "or1k-sim" + } +} + diff --git a/native/libffi/.travis/wine-sim.exp b/native/libffi/.travis/wine-sim.exp new file mode 100644 index 0000000000..1ad603817a --- /dev/null +++ b/native/libffi/.travis/wine-sim.exp @@ -0,0 +1,55 @@ +# Copyright (C) 2010, 2019 Free Software Foundation, Inc. +# +# This file is part of DejaGnu. +# +# DejaGnu is free software; you can redistribute it and/or modify it +# under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# DejaGnu is distributed in the hope that it will be useful, but +# WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with DejaGnu; if not, write to the Free Software Foundation, +# Inc., 51 Franklin Street, Fifth Floor, MA 02110, USA. + +# This is a list of toolchains that are supported on this board. +set_board_info target_install {i686-w64-mingw32} + +# Load the generic configuration for this board. This will define a basic set +# of routines needed by the tool to communicate with the board. +load_generic_config "sim" + +set_board_info sim "wineconsole --backend=curses" +set_board_info is_simulator 1 + +# No multilib options needed by default. +process_multilib_options "" + +# We only support newlib on this target. We assume that all multilib +# options have been specified before we get here. + +set_board_info compiler "[find_gcc]" +set_board_info cflags "[libgloss_include_flags] [newlib_include_flags]" +set_board_info ldflags "[libgloss_link_flags] [newlib_link_flags]" + +# Configuration settings for testsuites +set_board_info noargs 1 +set_board_info gdb,nosignals 1 +set_board_info gdb,noresults 1 +set_board_info gdb,cannot_call_functions 1 +set_board_info gdb,skip_float_tests 1 +set_board_info gdb,can_reverse 1 +set_board_info gdb,use_precord 1 + +# More time is needed +set_board_info gcc,timeout 800 +set_board_info gdb,timeout 60 + +# Used by a few gcc.c-torture testcases to delimit how large the stack can +# be. +set_board_info gcc,stack_size 5000 + diff --git a/native/libffi/ChangeLog.libffi b/native/libffi/ChangeLog.libffi deleted file mode 100644 index 49ba8dad3a..0000000000 --- a/native/libffi/ChangeLog.libffi +++ /dev/null @@ -1,584 +0,0 @@ -2011-02-08 Andreas Tobler - - * testsuite/lib/libffi.exp: Tweak for stand-alone mode. - -2009-12-25 Samuli Suominen - - * configure.ac: Undefine _AC_ARG_VAR_PRECIOUS for autoconf 2.64. - * configure: Rebuilt. - * fficonfig.h.in: Rebuilt. - -2009-06-16 Andrew Haley - - * testsuite/libffi.call/cls_align_sint64.c, - testsuite/libffi.call/cls_align_uint64.c, - testsuite/libffi.call/cls_longdouble_va.c, - testsuite/libffi.call/cls_ulonglong.c, - testsuite/libffi.call/return_ll1.c, - testsuite/libffi.call/stret_medium2.c: Fix printf format - specifiers. - * testsuite/libffi.call/huge_struct.c: Ad x86 XFAILs. - * testsuite/libffi.call/float2.c: Fix dg-excess-errors. - * testsuite/libffi.call/ffitest.h, - testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define. - -2009-06-12 Andrew Haley - - * testsuite/libffi.call/cls_align_sint64.c, - testsuite/libffi.call/cls_align_uint64.c, - testsuite/libffi.call/cls_ulonglong.c, - testsuite/libffi.call/return_ll1.c, - testsuite/libffi.call/stret_medium2.c: Fix printf format - specifiers. - testsuite/libffi.special/unwindtest.cc: include stdint.h. - -2009-06-11 Timothy Wall - - * Makefile.am, - configure.ac, - include/ffi.h.in, - include/ffi_common.h, - src/closures.c, - src/dlmalloc.c, - src/x86/ffi.c, - src/x86/ffitarget.h, - src/x86/win64.S (new), - README: Added win64 support (mingw or MSVC) - * Makefile.in, - include/Makefile.in, - man/Makefile.in, - testsuite/Makefile.in, - configure, - aclocal.m4: Regenerated - * ltcf-c.sh: properly escape cygwin/w32 path - * man/ffi_call.3: Clarify size requirements for return value. - * src/x86/ffi64.c: Fix filename in comment. - * src/x86/win32.S: Remove unused extern. - - * testsuite/libffi.call/closure_fn0.c, - testsuite/libffi.call/closure_fn1.c, - testsuite/libffi.call/closure_fn2.c, - testsuite/libffi.call/closure_fn3.c, - testsuite/libffi.call/closure_fn4.c, - testsuite/libffi.call/closure_fn5.c, - testsuite/libffi.call/closure_fn6.c, - testsuite/libffi.call/closure_stdcall.c, - testsuite/libffi.call/cls_12byte.c, - testsuite/libffi.call/cls_16byte.c, - testsuite/libffi.call/cls_18byte.c, - testsuite/libffi.call/cls_19byte.c, - testsuite/libffi.call/cls_1_1byte.c, - testsuite/libffi.call/cls_20byte.c, - testsuite/libffi.call/cls_20byte1.c, - testsuite/libffi.call/cls_24byte.c, - testsuite/libffi.call/cls_2byte.c, - testsuite/libffi.call/cls_3_1byte.c, - testsuite/libffi.call/cls_3byte1.c, - testsuite/libffi.call/cls_3byte2.c, - testsuite/libffi.call/cls_4_1byte.c, - testsuite/libffi.call/cls_4byte.c, - testsuite/libffi.call/cls_5_1_byte.c, - testsuite/libffi.call/cls_5byte.c, - testsuite/libffi.call/cls_64byte.c, - testsuite/libffi.call/cls_6_1_byte.c, - testsuite/libffi.call/cls_6byte.c, - testsuite/libffi.call/cls_7_1_byte.c, - testsuite/libffi.call/cls_7byte.c, - testsuite/libffi.call/cls_8byte.c, - testsuite/libffi.call/cls_9byte1.c, - testsuite/libffi.call/cls_9byte2.c, - testsuite/libffi.call/cls_align_double.c, - testsuite/libffi.call/cls_align_float.c, - testsuite/libffi.call/cls_align_longdouble.c, - testsuite/libffi.call/cls_align_longdouble_split.c, - testsuite/libffi.call/cls_align_longdouble_split2.c, - testsuite/libffi.call/cls_align_pointer.c, - testsuite/libffi.call/cls_align_sint16.c, - testsuite/libffi.call/cls_align_sint32.c, - testsuite/libffi.call/cls_align_sint64.c, - testsuite/libffi.call/cls_align_uint16.c, - testsuite/libffi.call/cls_align_uint32.c, - testsuite/libffi.call/cls_align_uint64.c, - testsuite/libffi.call/cls_dbls_struct.c, - testsuite/libffi.call/cls_double.c, - testsuite/libffi.call/cls_double_va.c, - testsuite/libffi.call/cls_float.c, - testsuite/libffi.call/cls_longdouble.c, - testsuite/libffi.call/cls_longdouble_va.c, - testsuite/libffi.call/cls_multi_schar.c, - testsuite/libffi.call/cls_multi_sshort.c, - testsuite/libffi.call/cls_multi_sshortchar.c, - testsuite/libffi.call/cls_multi_uchar.c, - testsuite/libffi.call/cls_multi_ushort.c, - testsuite/libffi.call/cls_multi_ushortchar.c, - testsuite/libffi.call/cls_pointer.c, - testsuite/libffi.call/cls_pointer_stack.c, - testsuite/libffi.call/cls_schar.c, - testsuite/libffi.call/cls_sint.c, - testsuite/libffi.call/cls_sshort.c, - testsuite/libffi.call/cls_uchar.c, - testsuite/libffi.call/cls_uint.c, - testsuite/libffi.call/cls_ulonglong.c, - testsuite/libffi.call/cls_ushort.c, - testsuite/libffi.call/err_bad_abi.c, - testsuite/libffi.call/err_bad_typedef.c, - testsuite/libffi.call/float2.c, - testsuite/libffi.call/huge_struct.c, - testsuite/libffi.call/nested_struct.c, - testsuite/libffi.call/nested_struct1.c, - testsuite/libffi.call/nested_struct10.c, - testsuite/libffi.call/nested_struct2.c, - testsuite/libffi.call/nested_struct3.c, - testsuite/libffi.call/nested_struct4.c, - testsuite/libffi.call/nested_struct5.c, - testsuite/libffi.call/nested_struct6.c, - testsuite/libffi.call/nested_struct7.c, - testsuite/libffi.call/nested_struct8.c, - testsuite/libffi.call/nested_struct9.c, - testsuite/libffi.call/problem1.c, - testsuite/libffi.call/return_ldl.c, - testsuite/libffi.call/return_ll1.c, - testsuite/libffi.call/stret_large.c, - testsuite/libffi.call/stret_large2.c, - testsuite/libffi.call/stret_medium.c, - testsuite/libffi.call/stret_medium2.c, - testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead - of checking for MMAP. Use intptr_t instead of long casts. - -2009-06-04 Andrew Haley - - * src/powerpc/ffitarget.h: Fix misapplied merge from gcc. - -2009-06-04 Andrew Haley - - * src/mips/o32.S, - src/mips/n32.S: Fix licence formatting. - -2009-06-04 Andrew Haley - - * src/x86/darwin.S: Fix licence formatting. - src/x86/win32.S: Likewise. - src/sh64/sysv.S: Likewise. - src/sh/sysv.S: Likewise. - -2009-06-04 Andrew Haley - - * src/sh64/ffi.c: Remove lint directives. Was missing from merge - of Andreas Tobler's patch from 2006-04-22. - -2009-06-04 Andrew Haley - - * src/sh/ffi.c: Apply missing hunk from Alexandre Oliva's patch of - 2007-03-07. - -2008-12-26 Timothy Wall - - * testsuite/libffi.call/cls_longdouble.c, - testsuite/libffi.call/cls_longdouble_va.c, - testsuite/libffi.call/cls_align_longdouble.c, - testsuite/libffi.call/cls_align_longdouble_split.c, - testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected - failures on x86_64 cygwin/mingw. - -2008-12-22 Timothy Wall - - * testsuite/libffi.call/closure_fn0.c, - testsuite/libffi.call/closure_fn1.c, - testsuite/libffi.call/closure_fn2.c, - testsuite/libffi.call/closure_fn3.c, - testsuite/libffi.call/closure_fn4.c, - testsuite/libffi.call/closure_fn5.c, - testsuite/libffi.call/closure_fn6.c, - testsuite/libffi.call/closure_loc_fn0.c, - testsuite/libffi.call/closure_stdcall.c, - testsuite/libffi.call/cls_align_pointer.c, - testsuite/libffi.call/cls_pointer.c, - testsuite/libffi.call/cls_pointer_stack.c: use portable cast from - pointer to integer (intptr_t). - * testsuite/libffi.call/cls_longdouble.c: disable for win64. - -2008-12-19 Anthony Green - - * configure.ac: Bump version to 3.0.8. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - * libtool-version: Increment revision. - * README: Update for new release. - -2008-11-11 Anthony Green - - * configure.ac: Bump version to 3.0.7. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - * libtool-version: Increment revision. - * README: Update for new release. - -2008-08-25 Andreas Tobler - - * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and - FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum. - Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT. - Adjust copyright notice. - * src/powerpc/ffi.c: Add two new flags to indicate if we have one - register or two register to use for FFI_SYSV structs. - (ffi_prep_cif_machdep): Pass the right register flag introduced above. - (ffi_closure_helper_SYSV): Fix the return type for - FFI_SYSV_TYPE_SMALL_STRUCT. Comment. - Adjust copyright notice. - -2008-07-24 Anthony Green - - * testsuite/libffi.call/cls_dbls_struct.c, - testsuite/libffi.call/cls_double_va.c, - testsuite/libffi.call/cls_longdouble.c, - testsuite/libffi.call/cls_longdouble_va.c, - testsuite/libffi.call/cls_pointer.c, - testsuite/libffi.call/cls_pointer_stack.c, - testsuite/libffi.call/err_bad_abi.c: Clean up failures from - compiler warnings. - -2008-07-17 Anthony Green - - * configure.ac: Bump version to 3.0.6. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - * libtool-version: Increment revision. Add documentation. - * README: Update for new release. - -2008-07-16 Kaz Kojima - - * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned - int. - -2008-07-16 Kaz Kojima - - * src/sh/sysv.S: Add .note.GNU-stack on Linux. - * src/sh64/sysv.S: Likewise. - -2008-04-03 Anthony Green - - * libffi.pc.in (Libs): Add -L${libdir}. - * configure.ac: Bump version to 3.0.5. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - * libtool-version: Increment revision. - * README: Update for new release. - -2008-04-03 Anthony Green - Xerces Ranby - - * include/ffi.h.in: Wrap definition of target architecture to - protect from double definitions. - -2008-03-22 Moriyoshi Koizumi - - * src/x86/ffi.c (ffi_prep_closure_loc): Fix for bug revealed in - closure_loc_fn0.c. - * testsuite/libffi.call/closure_loc_fn0.c (closure_loc_test_fn0): - New test. - -2008-03-04 Anthony Green - Blake Chaffin - hos@tamanegi.org - - * testsuite/libffi.call/cls_align_longdouble_split2.c - testsuite/libffi.call/cls_align_longdouble_split.c - testsuite/libffi.call/cls_dbls_struct.c - testsuite/libffi.call/cls_double_va.c - testsuite/libffi.call/cls_longdouble.c - testsuite/libffi.call/cls_longdouble_va.c - testsuite/libffi.call/cls_pointer.c - testsuite/libffi.call/cls_pointer_stack.c - testsuite/libffi.call/err_bad_abi.c - testsuite/libffi.call/err_bad_typedef.c - testsuite/libffi.call/huge_struct.c - testsuite/libffi.call/stret_large2.c - testsuite/libffi.call/stret_large.c - testsuite/libffi.call/stret_medium2.c - testsuite/libffi.call/stret_medium.c: New tests from Apple. - -2008-02-26 Jakub Jelinek - Anthony Green - - * src/alpha/osf.S: Add .note.GNU-stack on Linux. - * src/s390/sysv.S: Likewise. - * src/powerpc/linux64.S: Likewise. - * src/powerpc/linux64_closure.S: Likewise. - * src/powerpc/ppc_closure.S: Likewise. - * src/powerpc/sysv.S: Likewise. - * src/x86/unix64.S: Likewise. - * src/x86/sysv.S: Likewise. - * src/sparc/v8.S: Likewise. - * src/sparc/v9.S: Likewise. - * src/m68k/sysv.S: Likewise. - * src/ia64/unix.S: Likewise. - * src/arm/sysv.S: Likewise. - -2008-02-26 Anthony Green - Thomas Heller - - * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C - comment. - -2008-02-26 Anthony Green - Thomas Heller - - * include/ffi.h.in: Change void (*)() to void (*)(void). - -2008-02-26 Anthony Green - Thomas Heller - - * src/alpha/ffi.c: Change void (*)() to void (*)(void). - src/alpha/osf.S, src/arm/ffi.c, src/frv/ffi.c, src/ia64/ffi.c, - src/ia64/unix.S, src/java_raw_api.c, src/m32r/ffi.c, - src/mips/ffi.c, src/pa/ffi.c, src/pa/hpux32.S, src/pa/linux.S, - src/powerpc/ffi.c, src/powerpc/ffi_darwin.c, src/raw_api.c, - src/s390/ffi.c, src/sh/ffi.c, src/sh64/ffi.c, src/sparc/ffi.c, - src/x86/ffi.c, src/x86/unix64.S, src/x86/darwin64.S, - src/x86/ffi64.c: Ditto. - -2008-02-24 Anthony Green - - * configure.ac: Accept openbsd*, not just openbsd. - Bump version to 3.0.4. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - * libtool-version: Increment revision. - * README: Update for new release. - -2008-02-22 Anthony Green - - * README: Clean up list of tested platforms. - -2008-02-22 Anthony Green - - * configure.ac: Bump version to 3.0.3. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - * libtool-version: Increment revision. - * README: Update for new release. Clean up test docs. - -2008-02-22 Bjoern Koenig - Andreas Tobler - - * configure.ac: Add amd64-*-freebsd* target. - * configure: Regenerate. - -2008-02-22 Thomas Heller - - * configure.ac: Add x86 OpenBSD support. - * configure: Rebuilt. - -2008-02-21 Thomas Heller - - * README: Change "make test" to "make check". - -2008-02-21 Anthony Green - - * configure.ac: Bump version to 3.0.2. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - * libtool-version: Increment revision. - * README: Update for new release. - -2008-02-21 Björn König - - * src/x86/freebsd.S: New file. - * configure.ac: Add x86 FreeBSD support. - * Makefile.am: Ditto. - -2008-02-15 Anthony Green - - * configure.ac: Bump version to 3.0.1. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - * libtool-version: Increment revision. - * README: Update for new release. - -2008-02-15 David Daney - - * src/mips/ffi.c: Remove extra '>' from include directive. - (ffi_prep_closure_loc): Use clear_location instead of tramp. - -2008-02-15 Anthony Green - - * configure.ac: Bump version to 3.0.0. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - -2008-02-15 David Daney - - * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE): - Define (conditionally), and use it to include cachectl.h. - (ffi_prep_closure_loc): Fix cache flushing. - * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define. - -2008-02-15 Anthony Green - - * man/ffi_call.3, man/ffi_prep_cif.3, man/ffi.3: - Update dates and remove all references to ffi_prep_closure. - * configure.ac: Bump version to 2.99.9. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - -2008-02-15 Anthony Green - - * man/ffi_prep_closure.3: Delete. - * man/Makefile.am (EXTRA_DIST): Remove ffi_prep_closure.3. - (man_MANS): Ditto. - * man/Makefile.in: Rebuilt. - * configure.ac: Bump version to 2.99.8. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - -2008-02-14 Anthony Green - - * configure.ac: Bump version to 2.99.7. - * configure, doc/stamp-vti, doc/version.texi: Rebuilt. - * include/ffi.h.in LICENSE src/debug.c src/closures.c - src/ffitest.c src/s390/sysv.S src/s390/ffitarget.h - src/types.c src/m68k/ffitarget.h src/raw_api.c src/frv/ffi.c - src/frv/ffitarget.h src/sh/ffi.c src/sh/sysv.S - src/sh/ffitarget.h src/powerpc/ffitarget.h src/pa/ffi.c - src/pa/ffitarget.h src/pa/linux.S src/java_raw_api.c - src/cris/ffitarget.h src/x86/ffi.c src/x86/sysv.S - src/x86/unix64.S src/x86/win32.S src/x86/ffitarget.h - src/x86/ffi64.c src/x86/darwin.S src/ia64/ffi.c - src/ia64/ffitarget.h src/ia64/ia64_flags.h src/ia64/unix.S - src/sparc/ffi.c src/sparc/v9.S src/sparc/ffitarget.h - src/sparc/v8.S src/alpha/ffi.c src/alpha/ffitarget.h - src/alpha/osf.S src/sh64/ffi.c src/sh64/sysv.S - src/sh64/ffitarget.h src/mips/ffi.c src/mips/ffitarget.h - src/mips/n32.S src/mips/o32.S src/arm/ffi.c src/arm/sysv.S - src/arm/ffitarget.h src/prep_cif.c: Update license text. - -2008-02-14 Anthony Green - - * README: Update tested platforms. - * configure.ac: Bump version to 2.99.6. - * configure: Rebuilt. - -2008-02-14 Anthony Green - - * configure.ac: Bump version to 2.99.5. - * configure: Rebuilt. - * Makefile.am (EXTRA_DIST): Add darwin64.S - * Makefile.in: Rebuilt. - * testsuite/lib/libffi-dg.exp: Remove libstdc++ bits from GCC tree. - * LICENSE: Update WARRANTY. - -2008-02-14 Anthony Green - - * libffi.pc.in (libdir): Fix libdir definition. - * configure.ac: Bump version to 2.99.4. - * configure: Rebuilt. - -2008-02-14 Anthony Green - - * README: Update. - * libffi.info: New file. - * doc/stamp-vti: New file. - * configure.ac: Bump version to 2.99.3. - * configure: Rebuilt. - -2008-02-14 Anthony Green - - * Makefile.am (SUBDIRS): Add man dir. - * Makefile.in: Rebuilt. - * configure.ac: Create Makefile. - * configure: Rebuilt. - * man/ffi_call.3 man/ffi_prep_cif.3 man/ffi_prep_closure.3 - man/Makefile.am man/Makefile.in: New files. - -2008-02-14 Tom Tromey - - * aclocal.m4, Makefile.in, configure, fficonfig.h.in: Rebuilt. - * mdate-sh, texinfo.tex: New files. - * Makefile.am (info_TEXINFOS): New variable. - * doc/libffi.texi: New file. - * doc/version.texi: Likewise. - -2008-02-14 Anthony Green - - * Makefile.am (AM_CFLAGS): Don't compile with -D$(TARGET). - (lib_LTLIBRARIES): Define. - (toolexeclib_LIBRARIES): Undefine. - * Makefile.in: Rebuilt. - * configure.ac: Reset version to 2.99.1. - * configure.in: Rebuilt. - -2008-02-14 Anthony Green - - * libffi.pc.in: Use @PACKAGE_NAME@ and @PACKAGE_VERSION@. - * configure.ac: Reset version to 2.99.1. - * configure.in: Rebuilt. - * Makefile.am (EXTRA_DIST): Add ChangeLog.libffi. - * Makefile.in: Rebuilt. - * LICENSE: Update copyright notice. - -2008-02-14 Anthony Green - - * include/Makefile.am (nodist_includes_HEADERS): Define. Don't - distribute ffitarget.h or ffi.h from the build include dir. - * Makefile.in: Rebuilt. - -2008-02-14 Anthony Green - - * include/Makefile.am (includesdir): Install headers under libdir. - (pkgconfigdir): Define. Install libffi.pc. - * include/Makefile.in: Rebuilt. - * libffi.pc.in: Create. - * libtool-version: Increment CURRENT - * configure.ac: Add libffi.pc.in - * configure: Rebuilt. - -2008-02-03 Anthony Green - - * include/Makefile.am (includesdir): Fix header install with - DESTDIR. - * include/Makefile.in: Rebuilt. - -2008-02-03 Timothy Wall - - * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return - offset based on code pointer, not data pointer. - -2008-02-01 Anthony Green - - * include/Makefile.am: Fix header installs. - * Makefile.am: Ditto. - * include/Makefile.in: Rebuilt. - * Makefile.in: Ditto. - -2008-02-01 Anthony Green - - * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL, - FFI_INIT_TRAMPOLINE): Revert my broken changes to twall's last - patch. - -2008-01-31 Anthony Green - - * Makefile.am (EXTRA_DIST): Add missing files. - * testsuite/Makefile.am: Ditto. - * Makefile.in, testsuite/Makefile.in: Rebuilt. - -2008-01-31 Timothy Wall - - * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall - closures. - * src/x86/ffitarget.h: Increase size of trampoline for stdcall - closures. - * src/x86/win32.S: Add assembly for stdcall closure. - * src/x86/ffi.c: Initialize stdcall closure trampoline. - -2008-01-30 H.J. Lu - - PR libffi/34612 - * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when - returning struct. - - * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer" - tests. - -2008-01-30 Anthony Green - - * Makefile.am, include/Makefile.am: Move headers to - libffi_la_SOURCES for new automake. - * Makefile.in, include/Makefile.in: Rebuilt. - - * testsuite/lib/wrapper.exp: Copied from gcc tree to allow for - execution outside of gcc tree. - * testsuite/lib/target-libpath.exp: Ditto. - - * testsuite/lib/libffi-dg.exp: Many changes to allow for execution - outside of gcc tree. - diff --git a/native/libffi/ChangeLog.libffi-3.1 b/native/libffi/ChangeLog.libffi-3.1 index 8f7f50d6f9..8216dc8194 100644 --- a/native/libffi/ChangeLog.libffi-3.1 +++ b/native/libffi/ChangeLog.libffi-3.1 @@ -1,3 +1,14 @@ +<<<<<<< HEAD:native/libffi/ChangeLog.libffi-3.1 +======= +Libffi change logs used to be maintained in separate ChangeLog files. +These days we generate them directly from the git commit messages. +The old ChangeLog files are saved here in order to maintain the historical +record. + +============================================================================= +From the old ChangeLog.libffi-3.1 file... + +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878:native/libffi/ChangeLog.old 2014-03-16 Josh Triplett * ChangeLog: Archive to ChangeLog.libffi-3.1 and delete. Future @@ -5998,3 +6009,1402 @@ Thu Jul 8 14:28:42 1999 Anthony Green * src/x86/ChangeLog: Removed. * ChangeLog.v1: Created. + +============================================================================= +From the old ChangeLog.libffi file.... + +2011-02-08 Andreas Tobler + + * testsuite/lib/libffi.exp: Tweak for stand-alone mode. + +2009-12-25 Samuli Suominen + + * configure.ac: Undefine _AC_ARG_VAR_PRECIOUS for autoconf 2.64. + * configure: Rebuilt. + * fficonfig.h.in: Rebuilt. + +2009-06-16 Andrew Haley + + * testsuite/libffi.call/cls_align_sint64.c, + testsuite/libffi.call/cls_align_uint64.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_ulonglong.c, + testsuite/libffi.call/return_ll1.c, + testsuite/libffi.call/stret_medium2.c: Fix printf format + specifiers. + * testsuite/libffi.call/huge_struct.c: Ad x86 XFAILs. + * testsuite/libffi.call/float2.c: Fix dg-excess-errors. + * testsuite/libffi.call/ffitest.h, + testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define. + +2009-06-12 Andrew Haley + + * testsuite/libffi.call/cls_align_sint64.c, + testsuite/libffi.call/cls_align_uint64.c, + testsuite/libffi.call/cls_ulonglong.c, + testsuite/libffi.call/return_ll1.c, + testsuite/libffi.call/stret_medium2.c: Fix printf format + specifiers. + testsuite/libffi.special/unwindtest.cc: include stdint.h. + +2009-06-11 Timothy Wall + + * Makefile.am, + configure.ac, + include/ffi.h.in, + include/ffi_common.h, + src/closures.c, + src/dlmalloc.c, + src/x86/ffi.c, + src/x86/ffitarget.h, + src/x86/win64.S (new), + README: Added win64 support (mingw or MSVC) + * Makefile.in, + include/Makefile.in, + man/Makefile.in, + testsuite/Makefile.in, + configure, + aclocal.m4: Regenerated + * ltcf-c.sh: properly escape cygwin/w32 path + * man/ffi_call.3: Clarify size requirements for return value. + * src/x86/ffi64.c: Fix filename in comment. + * src/x86/win32.S: Remove unused extern. + + * testsuite/libffi.call/closure_fn0.c, + testsuite/libffi.call/closure_fn1.c, + testsuite/libffi.call/closure_fn2.c, + testsuite/libffi.call/closure_fn3.c, + testsuite/libffi.call/closure_fn4.c, + testsuite/libffi.call/closure_fn5.c, + testsuite/libffi.call/closure_fn6.c, + testsuite/libffi.call/closure_stdcall.c, + testsuite/libffi.call/cls_12byte.c, + testsuite/libffi.call/cls_16byte.c, + testsuite/libffi.call/cls_18byte.c, + testsuite/libffi.call/cls_19byte.c, + testsuite/libffi.call/cls_1_1byte.c, + testsuite/libffi.call/cls_20byte.c, + testsuite/libffi.call/cls_20byte1.c, + testsuite/libffi.call/cls_24byte.c, + testsuite/libffi.call/cls_2byte.c, + testsuite/libffi.call/cls_3_1byte.c, + testsuite/libffi.call/cls_3byte1.c, + testsuite/libffi.call/cls_3byte2.c, + testsuite/libffi.call/cls_4_1byte.c, + testsuite/libffi.call/cls_4byte.c, + testsuite/libffi.call/cls_5_1_byte.c, + testsuite/libffi.call/cls_5byte.c, + testsuite/libffi.call/cls_64byte.c, + testsuite/libffi.call/cls_6_1_byte.c, + testsuite/libffi.call/cls_6byte.c, + testsuite/libffi.call/cls_7_1_byte.c, + testsuite/libffi.call/cls_7byte.c, + testsuite/libffi.call/cls_8byte.c, + testsuite/libffi.call/cls_9byte1.c, + testsuite/libffi.call/cls_9byte2.c, + testsuite/libffi.call/cls_align_double.c, + testsuite/libffi.call/cls_align_float.c, + testsuite/libffi.call/cls_align_longdouble.c, + testsuite/libffi.call/cls_align_longdouble_split.c, + testsuite/libffi.call/cls_align_longdouble_split2.c, + testsuite/libffi.call/cls_align_pointer.c, + testsuite/libffi.call/cls_align_sint16.c, + testsuite/libffi.call/cls_align_sint32.c, + testsuite/libffi.call/cls_align_sint64.c, + testsuite/libffi.call/cls_align_uint16.c, + testsuite/libffi.call/cls_align_uint32.c, + testsuite/libffi.call/cls_align_uint64.c, + testsuite/libffi.call/cls_dbls_struct.c, + testsuite/libffi.call/cls_double.c, + testsuite/libffi.call/cls_double_va.c, + testsuite/libffi.call/cls_float.c, + testsuite/libffi.call/cls_longdouble.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_multi_schar.c, + testsuite/libffi.call/cls_multi_sshort.c, + testsuite/libffi.call/cls_multi_sshortchar.c, + testsuite/libffi.call/cls_multi_uchar.c, + testsuite/libffi.call/cls_multi_ushort.c, + testsuite/libffi.call/cls_multi_ushortchar.c, + testsuite/libffi.call/cls_pointer.c, + testsuite/libffi.call/cls_pointer_stack.c, + testsuite/libffi.call/cls_schar.c, + testsuite/libffi.call/cls_sint.c, + testsuite/libffi.call/cls_sshort.c, + testsuite/libffi.call/cls_uchar.c, + testsuite/libffi.call/cls_uint.c, + testsuite/libffi.call/cls_ulonglong.c, + testsuite/libffi.call/cls_ushort.c, + testsuite/libffi.call/err_bad_abi.c, + testsuite/libffi.call/err_bad_typedef.c, + testsuite/libffi.call/float2.c, + testsuite/libffi.call/huge_struct.c, + testsuite/libffi.call/nested_struct.c, + testsuite/libffi.call/nested_struct1.c, + testsuite/libffi.call/nested_struct10.c, + testsuite/libffi.call/nested_struct2.c, + testsuite/libffi.call/nested_struct3.c, + testsuite/libffi.call/nested_struct4.c, + testsuite/libffi.call/nested_struct5.c, + testsuite/libffi.call/nested_struct6.c, + testsuite/libffi.call/nested_struct7.c, + testsuite/libffi.call/nested_struct8.c, + testsuite/libffi.call/nested_struct9.c, + testsuite/libffi.call/problem1.c, + testsuite/libffi.call/return_ldl.c, + testsuite/libffi.call/return_ll1.c, + testsuite/libffi.call/stret_large.c, + testsuite/libffi.call/stret_large2.c, + testsuite/libffi.call/stret_medium.c, + testsuite/libffi.call/stret_medium2.c, + testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead + of checking for MMAP. Use intptr_t instead of long casts. + +2009-06-04 Andrew Haley + + * src/powerpc/ffitarget.h: Fix misapplied merge from gcc. + +2009-06-04 Andrew Haley + + * src/mips/o32.S, + src/mips/n32.S: Fix licence formatting. + +2009-06-04 Andrew Haley + + * src/x86/darwin.S: Fix licence formatting. + src/x86/win32.S: Likewise. + src/sh64/sysv.S: Likewise. + src/sh/sysv.S: Likewise. + +2009-06-04 Andrew Haley + + * src/sh64/ffi.c: Remove lint directives. Was missing from merge + of Andreas Tobler's patch from 2006-04-22. + +2009-06-04 Andrew Haley + + * src/sh/ffi.c: Apply missing hunk from Alexandre Oliva's patch of + 2007-03-07. + +2008-12-26 Timothy Wall + + * testsuite/libffi.call/cls_longdouble.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_align_longdouble.c, + testsuite/libffi.call/cls_align_longdouble_split.c, + testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected + failures on x86_64 cygwin/mingw. + +2008-12-22 Timothy Wall + + * testsuite/libffi.call/closure_fn0.c, + testsuite/libffi.call/closure_fn1.c, + testsuite/libffi.call/closure_fn2.c, + testsuite/libffi.call/closure_fn3.c, + testsuite/libffi.call/closure_fn4.c, + testsuite/libffi.call/closure_fn5.c, + testsuite/libffi.call/closure_fn6.c, + testsuite/libffi.call/closure_loc_fn0.c, + testsuite/libffi.call/closure_stdcall.c, + testsuite/libffi.call/cls_align_pointer.c, + testsuite/libffi.call/cls_pointer.c, + testsuite/libffi.call/cls_pointer_stack.c: use portable cast from + pointer to integer (intptr_t). + * testsuite/libffi.call/cls_longdouble.c: disable for win64. + +2008-12-19 Anthony Green + + * configure.ac: Bump version to 3.0.8. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-11-11 Anthony Green + + * configure.ac: Bump version to 3.0.7. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-08-25 Andreas Tobler + + * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and + FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum. + Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT. + Adjust copyright notice. + * src/powerpc/ffi.c: Add two new flags to indicate if we have one + register or two register to use for FFI_SYSV structs. + (ffi_prep_cif_machdep): Pass the right register flag introduced above. + (ffi_closure_helper_SYSV): Fix the return type for + FFI_SYSV_TYPE_SMALL_STRUCT. Comment. + Adjust copyright notice. + +2008-07-24 Anthony Green + + * testsuite/libffi.call/cls_dbls_struct.c, + testsuite/libffi.call/cls_double_va.c, + testsuite/libffi.call/cls_longdouble.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_pointer.c, + testsuite/libffi.call/cls_pointer_stack.c, + testsuite/libffi.call/err_bad_abi.c: Clean up failures from + compiler warnings. + +2008-07-17 Anthony Green + + * configure.ac: Bump version to 3.0.6. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. Add documentation. + * README: Update for new release. + +2008-07-16 Kaz Kojima + + * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned + int. + +2008-07-16 Kaz Kojima + + * src/sh/sysv.S: Add .note.GNU-stack on Linux. + * src/sh64/sysv.S: Likewise. + +2008-04-03 Anthony Green + + * libffi.pc.in (Libs): Add -L${libdir}. + * configure.ac: Bump version to 3.0.5. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-04-03 Anthony Green + Xerces Ranby + + * include/ffi.h.in: Wrap definition of target architecture to + protect from double definitions. + +2008-03-22 Moriyoshi Koizumi + + * src/x86/ffi.c (ffi_prep_closure_loc): Fix for bug revealed in + closure_loc_fn0.c. + * testsuite/libffi.call/closure_loc_fn0.c (closure_loc_test_fn0): + New test. + +2008-03-04 Anthony Green + Blake Chaffin + hos@tamanegi.org + + * testsuite/libffi.call/cls_align_longdouble_split2.c + testsuite/libffi.call/cls_align_longdouble_split.c + testsuite/libffi.call/cls_dbls_struct.c + testsuite/libffi.call/cls_double_va.c + testsuite/libffi.call/cls_longdouble.c + testsuite/libffi.call/cls_longdouble_va.c + testsuite/libffi.call/cls_pointer.c + testsuite/libffi.call/cls_pointer_stack.c + testsuite/libffi.call/err_bad_abi.c + testsuite/libffi.call/err_bad_typedef.c + testsuite/libffi.call/huge_struct.c + testsuite/libffi.call/stret_large2.c + testsuite/libffi.call/stret_large.c + testsuite/libffi.call/stret_medium2.c + testsuite/libffi.call/stret_medium.c: New tests from Apple. + +2008-02-26 Jakub Jelinek + Anthony Green + + * src/alpha/osf.S: Add .note.GNU-stack on Linux. + * src/s390/sysv.S: Likewise. + * src/powerpc/linux64.S: Likewise. + * src/powerpc/linux64_closure.S: Likewise. + * src/powerpc/ppc_closure.S: Likewise. + * src/powerpc/sysv.S: Likewise. + * src/x86/unix64.S: Likewise. + * src/x86/sysv.S: Likewise. + * src/sparc/v8.S: Likewise. + * src/sparc/v9.S: Likewise. + * src/m68k/sysv.S: Likewise. + * src/ia64/unix.S: Likewise. + * src/arm/sysv.S: Likewise. + +2008-02-26 Anthony Green + Thomas Heller + + * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C + comment. + +2008-02-26 Anthony Green + Thomas Heller + + * include/ffi.h.in: Change void (*)() to void (*)(void). + +2008-02-26 Anthony Green + Thomas Heller + + * src/alpha/ffi.c: Change void (*)() to void (*)(void). + src/alpha/osf.S, src/arm/ffi.c, src/frv/ffi.c, src/ia64/ffi.c, + src/ia64/unix.S, src/java_raw_api.c, src/m32r/ffi.c, + src/mips/ffi.c, src/pa/ffi.c, src/pa/hpux32.S, src/pa/linux.S, + src/powerpc/ffi.c, src/powerpc/ffi_darwin.c, src/raw_api.c, + src/s390/ffi.c, src/sh/ffi.c, src/sh64/ffi.c, src/sparc/ffi.c, + src/x86/ffi.c, src/x86/unix64.S, src/x86/darwin64.S, + src/x86/ffi64.c: Ditto. + +2008-02-24 Anthony Green + + * configure.ac: Accept openbsd*, not just openbsd. + Bump version to 3.0.4. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-02-22 Anthony Green + + * README: Clean up list of tested platforms. + +2008-02-22 Anthony Green + + * configure.ac: Bump version to 3.0.3. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. Clean up test docs. + +2008-02-22 Bjoern Koenig + Andreas Tobler + + * configure.ac: Add amd64-*-freebsd* target. + * configure: Regenerate. + +2008-02-22 Thomas Heller + + * configure.ac: Add x86 OpenBSD support. + * configure: Rebuilt. + +2008-02-21 Thomas Heller + + * README: Change "make test" to "make check". + +2008-02-21 Anthony Green + + * configure.ac: Bump version to 3.0.2. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-02-21 Björn König + + * src/x86/freebsd.S: New file. + * configure.ac: Add x86 FreeBSD support. + * Makefile.am: Ditto. + +2008-02-15 Anthony Green + + * configure.ac: Bump version to 3.0.1. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-02-15 David Daney + + * src/mips/ffi.c: Remove extra '>' from include directive. + (ffi_prep_closure_loc): Use clear_location instead of tramp. + +2008-02-15 Anthony Green + + * configure.ac: Bump version to 3.0.0. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + +2008-02-15 David Daney + + * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE): + Define (conditionally), and use it to include cachectl.h. + (ffi_prep_closure_loc): Fix cache flushing. + * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define. + +2008-02-15 Anthony Green + + * man/ffi_call.3, man/ffi_prep_cif.3, man/ffi.3: + Update dates and remove all references to ffi_prep_closure. + * configure.ac: Bump version to 2.99.9. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + +2008-02-15 Anthony Green + + * man/ffi_prep_closure.3: Delete. + * man/Makefile.am (EXTRA_DIST): Remove ffi_prep_closure.3. + (man_MANS): Ditto. + * man/Makefile.in: Rebuilt. + * configure.ac: Bump version to 2.99.8. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + +2008-02-14 Anthony Green + + * configure.ac: Bump version to 2.99.7. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * include/ffi.h.in LICENSE src/debug.c src/closures.c + src/ffitest.c src/s390/sysv.S src/s390/ffitarget.h + src/types.c src/m68k/ffitarget.h src/raw_api.c src/frv/ffi.c + src/frv/ffitarget.h src/sh/ffi.c src/sh/sysv.S + src/sh/ffitarget.h src/powerpc/ffitarget.h src/pa/ffi.c + src/pa/ffitarget.h src/pa/linux.S src/java_raw_api.c + src/cris/ffitarget.h src/x86/ffi.c src/x86/sysv.S + src/x86/unix64.S src/x86/win32.S src/x86/ffitarget.h + src/x86/ffi64.c src/x86/darwin.S src/ia64/ffi.c + src/ia64/ffitarget.h src/ia64/ia64_flags.h src/ia64/unix.S + src/sparc/ffi.c src/sparc/v9.S src/sparc/ffitarget.h + src/sparc/v8.S src/alpha/ffi.c src/alpha/ffitarget.h + src/alpha/osf.S src/sh64/ffi.c src/sh64/sysv.S + src/sh64/ffitarget.h src/mips/ffi.c src/mips/ffitarget.h + src/mips/n32.S src/mips/o32.S src/arm/ffi.c src/arm/sysv.S + src/arm/ffitarget.h src/prep_cif.c: Update license text. + +2008-02-14 Anthony Green + + * README: Update tested platforms. + * configure.ac: Bump version to 2.99.6. + * configure: Rebuilt. + +2008-02-14 Anthony Green + + * configure.ac: Bump version to 2.99.5. + * configure: Rebuilt. + * Makefile.am (EXTRA_DIST): Add darwin64.S + * Makefile.in: Rebuilt. + * testsuite/lib/libffi-dg.exp: Remove libstdc++ bits from GCC tree. + * LICENSE: Update WARRANTY. + +2008-02-14 Anthony Green + + * libffi.pc.in (libdir): Fix libdir definition. + * configure.ac: Bump version to 2.99.4. + * configure: Rebuilt. + +2008-02-14 Anthony Green + + * README: Update. + * libffi.info: New file. + * doc/stamp-vti: New file. + * configure.ac: Bump version to 2.99.3. + * configure: Rebuilt. + +2008-02-14 Anthony Green + + * Makefile.am (SUBDIRS): Add man dir. + * Makefile.in: Rebuilt. + * configure.ac: Create Makefile. + * configure: Rebuilt. + * man/ffi_call.3 man/ffi_prep_cif.3 man/ffi_prep_closure.3 + man/Makefile.am man/Makefile.in: New files. + +2008-02-14 Tom Tromey + + * aclocal.m4, Makefile.in, configure, fficonfig.h.in: Rebuilt. + * mdate-sh, texinfo.tex: New files. + * Makefile.am (info_TEXINFOS): New variable. + * doc/libffi.texi: New file. + * doc/version.texi: Likewise. + +2008-02-14 Anthony Green + + * Makefile.am (AM_CFLAGS): Don't compile with -D$(TARGET). + (lib_LTLIBRARIES): Define. + (toolexeclib_LIBRARIES): Undefine. + * Makefile.in: Rebuilt. + * configure.ac: Reset version to 2.99.1. + * configure.in: Rebuilt. + +2008-02-14 Anthony Green + + * libffi.pc.in: Use @PACKAGE_NAME@ and @PACKAGE_VERSION@. + * configure.ac: Reset version to 2.99.1. + * configure.in: Rebuilt. + * Makefile.am (EXTRA_DIST): Add ChangeLog.libffi. + * Makefile.in: Rebuilt. + * LICENSE: Update copyright notice. + +2008-02-14 Anthony Green + + * include/Makefile.am (nodist_includes_HEADERS): Define. Don't + distribute ffitarget.h or ffi.h from the build include dir. + * Makefile.in: Rebuilt. + +2008-02-14 Anthony Green + + * include/Makefile.am (includesdir): Install headers under libdir. + (pkgconfigdir): Define. Install libffi.pc. + * include/Makefile.in: Rebuilt. + * libffi.pc.in: Create. + * libtool-version: Increment CURRENT + * configure.ac: Add libffi.pc.in + * configure: Rebuilt. + +2008-02-03 Anthony Green + + * include/Makefile.am (includesdir): Fix header install with + DESTDIR. + * include/Makefile.in: Rebuilt. + +2008-02-03 Timothy Wall + + * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return + offset based on code pointer, not data pointer. + +2008-02-01 Anthony Green + + * include/Makefile.am: Fix header installs. + * Makefile.am: Ditto. + * include/Makefile.in: Rebuilt. + * Makefile.in: Ditto. + +2008-02-01 Anthony Green + + * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL, + FFI_INIT_TRAMPOLINE): Revert my broken changes to twall's last + patch. + +2008-01-31 Anthony Green + + * Makefile.am (EXTRA_DIST): Add missing files. + * testsuite/Makefile.am: Ditto. + * Makefile.in, testsuite/Makefile.in: Rebuilt. + +2008-01-31 Timothy Wall + + * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall + closures. + * src/x86/ffitarget.h: Increase size of trampoline for stdcall + closures. + * src/x86/win32.S: Add assembly for stdcall closure. + * src/x86/ffi.c: Initialize stdcall closure trampoline. + +2008-01-30 H.J. Lu + + PR libffi/34612 + * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when + returning struct. + + * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer" + tests. + +2008-01-30 Anthony Green + + * Makefile.am, include/Makefile.am: Move headers to + libffi_la_SOURCES for new automake. + * Makefile.in, include/Makefile.in: Rebuilt. + + * testsuite/lib/wrapper.exp: Copied from gcc tree to allow for + execution outside of gcc tree. + * testsuite/lib/target-libpath.exp: Ditto. + + * testsuite/lib/libffi-dg.exp: Many changes to allow for execution + outside of gcc tree. + + +============================================================================= +From the old ChangeLog.libgcj file.... + +2004-01-14 Kelley Cook + + * configure.in: Add in AC_PREREQ(2.13) + +2003-02-20 Alexandre Oliva + + * configure.in: Propagate ORIGINAL_LD_FOR_MULTILIBS to + config.status. + * configure: Rebuilt. + +2002-01-27 Alexandre Oliva + + * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST. + Remove USE_LIBDIR conditional. + * Makefile.am (toolexecdir, toolexeclibdir): Don't override. + * Makefile.in, configure: Rebuilt. + +Mon Aug 9 18:33:38 1999 Rainer Orth + + * include/Makefile.in: Rebuilt. + * Makefile.in: Rebuilt + * Makefile.am (toolexeclibdir): Add $(MULTISUBDIR) even for native + builds. + Use USE_LIBDIR. + + * configure: Rebuilt. + * configure.in (USE_LIBDIR): Define for native builds. + Use lowercase in configure --help explanations. + +1999-08-08 Anthony Green + + * include/ffi.h.in (FFI_FN): Remove `...'. + +1999-08-08 Anthony Green + + * Makefile.in: Rebuilt. + * Makefile.am (AM_CFLAGS): Compile with -fexceptions. + + * src/x86/sysv.S: Add exception handling metadata. + + +============================================================================= + +The libffi version 1 ChangeLog archive. + +Version 1 of libffi had per-directory ChangeLogs. Current and future +versions have a single ChangeLog file in the root directory. The +version 1 ChangeLogs have all been concatenated into this file for +future reference only. + +--- libffi ---------------------------------------------------------------- + +Mon Oct 5 02:17:50 1998 Anthony Green + + * configure.in: Boosted rev. + * configure, Makefile.in, aclocal.m4: Rebuilt. + * README: Boosted rev and updated release notes. + +Mon Oct 5 01:03:03 1998 Anthony Green + + * configure.in: Boosted rev. + * configure, Makefile.in, aclocal.m4: Rebuilt. + * README: Boosted rev and updated release notes. + +1998-07-25 Andreas Schwab + + * m68k/ffi.c (ffi_prep_cif_machdep): Use bitmask for cif->flags. + Correctly handle small structures. + (ffi_prep_args): Also handle small structures. + (ffi_call): Pass size of return type to ffi_call_SYSV. + * m68k/sysv.S: Adjust for above changes. Correctly align small + structures in the return value. + + * types.c (uint64, sint64) [M68K]: Change alignment to 4. + +Fri Apr 17 17:26:58 1998 Anthony Green + + * configure.in: Boosted rev. + * configure,Makefile.in,aclocal.m4: Rebuilt. + * README: Boosted rev and added release notes. + +Sun Feb 22 00:50:41 1998 Geoff Keating + + * configure.in: Add PowerPC config bits. + +1998-02-14 Andreas Schwab + + * configure.in: Add m68k config bits. Change AC_CANONICAL_SYSTEM + to AC_CANONICAL_HOST, this is not a compiler. Use $host instead + of $target. Remove AC_CHECK_SIZEOF(char), we already know the + result. Fix argument of AC_ARG_ENABLE. + * configure, fficonfig.h.in: Rebuilt. + +Tue Feb 10 20:53:40 1998 Richard Henderson + + * configure.in: Add Alpha config bits. + +Tue May 13 13:39:20 1997 Anthony Green + + * README: Updated dates and reworded Irix comments. + + * configure.in: Removed AC_PROG_RANLIB. + + * Makefile.in, aclocal.m4, config.guess, config.sub, configure, + ltmain.sh, */Makefile.in: libtoolized again and rebuilt with + automake and autoconf. + +Sat May 10 18:44:50 1997 Tom Tromey + + * configure, aclocal.m4: Rebuilt. + * configure.in: Don't compute EXTRADIST; now handled in + src/Makefile.in. Removed macros implied by AM_INIT_AUTOMAKE. + Don't run AM_MAINTAINER_MODE. + +Thu May 8 14:34:05 1997 Anthony Green + + * missing, ltmain.sh, ltconfig.sh: Created. These are new files + required by automake and libtool. + + * README: Boosted rev to 1.14. Added notes. + + * acconfig.h: Moved PACKAGE and VERSION for new automake. + + * configure.in: Changes for libtool. + + * Makefile.am (check): make test now make check. Uses libtool now. + + * Makefile.in, configure.in, aclocal.h, fficonfig.h.in: Rebuilt. + +Thu May 1 16:27:07 1997 Anthony Green + + * missing: Added file required by new automake. + +Tue Nov 26 14:10:42 1996 Anthony Green + + * acconfig.h: Added USING_PURIFY flag. This is defined when + --enable-purify-safety was used at configure time. + + * configure.in (allsources): Added --enable-purify-safety switch. + (VERSION): Boosted rev to 1.13. + * configure: Rebuilt. + +Fri Nov 22 06:46:12 1996 Anthony Green + + * configure.in (VERSION): Boosted rev to 1.12. + Removed special CFLAGS hack for gcc. + * configure: Rebuilt. + + * README: Boosted rev to 1.12. Added notes. + + * Many files: Cygnus Support changed to Cygnus Solutions. + +Wed Oct 30 11:15:25 1996 Anthony Green + + * configure.in (VERSION): Boosted rev to 1.11. + * configure: Rebuilt. + + * README: Boosted rev to 1.11. Added notes about GNU make. + +Tue Oct 29 12:25:12 1996 Anthony Green + + * configure.in: Fixed -Wall trick. + (VERSION): Boosted rev. + * configure: Rebuilt + + * acconfig.h: Needed for --enable-debug configure switch. + + * README: Boosted rev to 1.09. Added more notes on building + libffi, and LCLint. + + * configure.in: Added --enable-debug switch. Boosted rev to + 1.09. + * configure: Rebuilt + +Tue Oct 15 13:11:28 1996 Anthony Green + + * configure.in (VERSION): Boosted rev to 1.08 + * configure: Rebuilt. + + * README: Added n32 bug fix notes. + + * Makefile.am: Added "make lint" production. + * Makefile.in: Rebuilt. + +Mon Oct 14 10:54:46 1996 Anthony Green + + * README: Added web page reference. + + * configure.in, README: Boosted rev to 1.05 + * configure: Rebuilt. + + * README: Fixed n32 sample code. + +Fri Oct 11 17:09:28 1996 Anthony Green + + * README: Added sparc notes. + + * configure.in, README: Boosted rev to 1.04. + * configure: Rebuilt. + +Thu Oct 10 10:31:03 1996 Anthony Green + + * configure.in, README: Boosted rev to 1.03. + * configure: Rebuilt. + + * README: Added struct notes. + + * Makefile.am (EXTRA_DIST): Added LICENSE to distribution. + * Makefile.in: Rebuilt. + + * README: Removed Linux section. No special notes now + because aggregates arg/return types work. + +Wed Oct 9 16:16:42 1996 Anthony Green + + * README, configure.in (VERSION): Boosted rev to 1.02 + * configure: Rebuilt. + +Tue Oct 8 11:56:33 1996 Anthony Green + + * README (NOTE): Added n32 notes. + + * Makefile.am: Added test production. + * Makefile: Rebuilt + + * README: spell checked! + + * configure.in (VERSION): Boosted rev to 1.01 + * configure: Rebuilt. + +Mon Oct 7 15:50:22 1996 Anthony Green + + * configure.in: Added nasty bit to support SGI tools. + * configure: Rebuilt. + + * README: Added SGI notes. Added note about automake bug. + +Mon Oct 7 11:00:28 1996 Anthony Green + + * README: Rewrote intro, and fixed examples. + +Fri Oct 4 10:19:55 1996 Anthony Green + + * configure.in: -D$TARGET is no longer used as a compiler switch. + It is now inserted into ffi.h at configure time. + * configure: Rebuilt. + + * FFI_ABI and FFI_STATUS are now ffi_abi and ffi_status. + +Thu Oct 3 13:47:34 1996 Anthony Green + + * README, LICENSE: Created. Wrote some docs. + + * configure.in: Don't barf on i586-unknown-linuxaout. + Added EXTRADIST code for "make dist". + * configure: Rebuilt. + + * */Makefile.in: Rebuilt with patched automake. + +Tue Oct 1 17:12:25 1996 Anthony Green + + * Makefile.am, aclocal.m4, config.guess, config.sub, + configure.in, fficonfig.h.in, install-sh, mkinstalldirs, + stamp-h.in: Created + * Makefile.in, configure: Generated + +--- libffi/include -------------------------------------------------------- + +Tue Feb 24 13:09:36 1998 Anthony Green + + * ffi_mips.h: Updated FFI_TYPE_STRUCT_* values based on + ffi.h.in changes. This is a work-around for SGI's "simple" + assembler. + +Sun Feb 22 00:51:55 1998 Geoff Keating + + * ffi.h.in: PowerPC support. + +1998-02-14 Andreas Schwab + + * ffi.h.in: Add m68k support. + (FFI_TYPE_LONGDOUBLE): Make it a separate value. + +Tue Feb 10 20:55:16 1998 Richard Henderson + + * ffi.h.in (SIZEOF_ARG): Use a pointer type by default. + + * ffi.h.in: Alpha support. + +Fri Nov 22 06:48:45 1996 Anthony Green + + * ffi.h.in, ffi_common.h: Cygnus Support -> Cygnus Solutions. + +Wed Nov 20 22:31:01 1996 Anthony Green + + * ffi.h.in: Added ffi_type_void definition. + +Tue Oct 29 12:22:40 1996 Anthony Green + + * Makefile.am (hack_DATA): Always install ffi_mips.h. + + * ffi.h.in: Removed FFI_DEBUG. It's now in the correct + place (acconfig.h). + Added #include for size_t definition. + +Tue Oct 15 17:23:35 1996 Anthony Green + + * ffi.h.in, ffi_common.h, ffi_mips.h: More clean up. + Commented out #define of FFI_DEBUG. + +Tue Oct 15 13:01:06 1996 Anthony Green + + * ffi_common.h: Added bool definition. + + * ffi.h.in, ffi_common.h: Clean up based on LCLint output. + Added funny /*@...@*/ comments to annotate source. + +Mon Oct 14 12:29:23 1996 Anthony Green + + * ffi.h.in: Interface changes based on feedback from Jim + Blandy. + +Fri Oct 11 16:49:35 1996 Anthony Green + + * ffi.h.in: Small change for sparc support. + +Thu Oct 10 14:53:37 1996 Anthony Green + + * ffi_mips.h: Added FFI_TYPE_STRUCT_* definitions for + special structure return types. + +Wed Oct 9 13:55:57 1996 Anthony Green + + * ffi.h.in: Added SIZEOF_ARG definition for X86 + +Tue Oct 8 11:40:36 1996 Anthony Green + + * ffi.h.in (FFI_FN): Added macro for eliminating compiler warnings. + Use it to case your function pointers to the proper type. + + * ffi_mips.h (SIZEOF_ARG): Added magic to fix type promotion bug. + + * Makefile.am (EXTRA_DIST): Added ffi_mips.h to EXTRA_DIST. + * Makefile: Rebuilt. + + * ffi_mips.h: Created. Moved all common mips definitions here. + +Mon Oct 7 10:58:12 1996 Anthony Green + + * ffi.h.in: The SGI assember is very picky about parens. Redefined + some macros to avoid problems. + + * ffi.h.in: Added FFI_DEFAULT_ABI definitions. Also added + externs for pointer, and 64bit integral ffi_types. + +Fri Oct 4 09:51:37 1996 Anthony Green + + * ffi.h.in: Added FFI_ABI member to ffi_cif and changed + function prototypes accordingly. + Added #define @TARGET@. Now programs including ffi.h don't + have to specify this themselves. + +Thu Oct 3 15:36:44 1996 Anthony Green + + * ffi.h.in: Changed ffi_prep_cif's values from void* to void** + + * Makefile.am (EXTRA_DIST): Added EXTRA_DIST for "make dist" + to work. + * Makefile.in: Regenerated. + +Wed Oct 2 10:16:59 1996 Anthony Green + + * Makefile.am: Created + * Makefile.in: Generated + + * ffi_common.h: Added rcsid comment + +Tue Oct 1 17:13:51 1996 Anthony Green + + * ffi.h.in, ffi_common.h: Created + +--- libffi/src ------------------------------------------------------------ + +Mon Oct 5 02:17:50 1998 Anthony Green + + * arm/ffi.c, arm/sysv.S: Created. + + * Makefile.am: Added arm files. + * Makefile.in: Rebuilt. + +Mon Oct 5 01:41:38 1998 Anthony Green + + * Makefile.am (libffi_la_LDFLAGS): Incremented revision. + +Sun Oct 4 16:27:17 1998 Anthony Green + + * alpha/osf.S (ffi_call_osf): Patch for DU assembler. + + * ffitest.c (main): long long and long double return values work + for x86. + +Fri Apr 17 11:50:58 1998 Anthony Green + + * Makefile.in: Rebuilt. + + * ffitest.c (main): Floating point tests not executed for systems + with broken lond double (SunOS 4 w/ GCC). + + * types.c: Fixed x86 alignment info for long long types. + +Thu Apr 16 07:15:28 1998 Anthony Green + + * ffitest.c: Added more notes about GCC bugs under Irix 6. + +Wed Apr 15 08:42:22 1998 Anthony Green + + * ffitest.c (struct5): New test function. + (main): New test with struct5. + +Thu Mar 5 10:48:11 1998 Anthony Green + + * prep_cif.c (initialize_aggregate): Fix assertion for + nested structures. + +Tue Feb 24 16:33:41 1998 Anthony Green + + * prep_cif.c (ffi_prep_cif): Added long double support for sparc. + +Sun Feb 22 00:52:18 1998 Geoff Keating + + * powerpc/asm.h: New file. + * powerpc/ffi.c: New file. + * powerpc/sysv.S: New file. + * Makefile.am: PowerPC port. + * ffitest.c (main): Allow all tests to run even in presence of gcc + bug on PowerPC. + +1998-02-17 Anthony Green + + * mips/ffi.c: Fixed comment typo. + + * x86/ffi.c (ffi_prep_cif_machdep), x86/sysv.S (retfloat): + Fixed x86 long double return handling. + + * types.c: Fixed x86 long double alignment info. + +1998-02-14 Andreas Schwab + + * types.c: Add m68k support. + + * ffitest.c (floating): Add long double parameter. + (return_ll, ldblit): New functions to test long long and long + double return value. + (main): Fix type error in assignment of ts[1-4]_type.elements. + Add tests for long long and long double arguments and return + values. + + * prep_cif.c (ffi_prep_cif) [M68K]: Don't allocate argument for + struct value pointer. + + * m68k/ffi.c, m68k/sysv.S: New files. + * Makefile.am: Add bits for m68k port. Add kludge to work around + automake deficiency. + (test): Don't require "." in $PATH. + * Makefile.in: Rebuilt. + +Wed Feb 11 07:36:50 1998 Anthony Green + + * Makefile.in: Rebuilt. + +Tue Feb 10 20:56:00 1998 Richard Henderson + + * alpha/ffi.c, alpha/osf.S: New files. + * Makefile.am: Alpha port. + +Tue Nov 18 14:12:07 1997 Anthony Green + + * mips/ffi.c (ffi_prep_cif_machdep): Initialize rstruct_flag + for n32. + +Tue Jun 3 17:18:20 1997 Anthony Green + + * ffitest.c (main): Added hack to get structure tests working + correctly. + +Sat May 10 19:06:42 1997 Tom Tromey + + * Makefile.in: Rebuilt. + * Makefile.am (EXTRA_DIST): Explicitly list all distributable + files in subdirs. + (VERSION, CC): Removed. + +Thu May 8 17:19:01 1997 Anthony Green + + * Makefile.am: Many changes for new automake and libtool. + * Makefile.in: Rebuilt. + +Fri Nov 22 06:57:56 1996 Anthony Green + + * ffitest.c (main): Fixed test case for non mips machines. + +Wed Nov 20 22:31:59 1996 Anthony Green + + * types.c: Added ffi_type_void declaration. + +Tue Oct 29 13:07:19 1996 Anthony Green + + * ffitest.c (main): Fixed character constants. + (main): Emit warning for structure test 3 failure on Sun. + + * Makefile.am (VPATH): Fixed VPATH def'n so automake won't + strip it out. + Moved distdir hack from libffi to automake. + (ffitest): Added missing -c for $(COMPILE) (change in automake). + * Makefile.in: Rebuilt. + +Tue Oct 15 13:08:20 1996 Anthony Green + + * Makefile.am: Added "make lint" production. + * Makefile.in: Rebuilt. + + * prep_cif.c (STACK_ARG_SIZE): Improved STACK_ARG_SIZE macro. + Clean up based on LCLint output. Added funny /*@...@*/ comments to + annotate source. + + * ffitest.c, debug.c: Cleaned up code. + +Mon Oct 14 12:26:56 1996 Anthony Green + + * ffitest.c: Changes based on interface changes. + + * prep_cif.c (ffi_prep_cif): Cleaned up interface based on + feedback from Jim Blandy. + +Fri Oct 11 15:53:18 1996 Anthony Green + + * ffitest.c: Reordered tests while porting to sparc. + Made changes to handle lame structure passing for sparc. + Removed calls to fflush(). + + * prep_cif.c (ffi_prep_cif): Added special case for sparc + aggregate type arguments. + +Thu Oct 10 09:56:51 1996 Anthony Green + + * ffitest.c (main): Added structure passing/returning tests. + + * prep_cif.c (ffi_prep_cif): Perform proper initialization + of structure return types if needed. + (initialize_aggregate): Bug fix + +Wed Oct 9 16:04:20 1996 Anthony Green + + * types.c: Added special definitions for x86 (double doesn't + need double word alignment). + + * ffitest.c: Added many tests + +Tue Oct 8 09:19:22 1996 Anthony Green + + * prep_cif.c (ffi_prep_cif): Fixed assertion. + + * debug.c (ffi_assert): Must return a non void now. + + * Makefile.am: Added test production. + * Makefile: Rebuilt. + + * ffitest.c (main): Created. + + * types.c: Created. Stripped common code out of */ffi.c. + + * prep_cif.c: Added missing stdlib.h include. + + * debug.c (ffi_type_test): Used "a" to eliminate compiler + warnings in non-debug builds. Included ffi_common.h. + +Mon Oct 7 15:36:42 1996 Anthony Green + + * Makefile.am: Added a rule for .s -> .o + This is required by the SGI compiler. + * Makefile: Rebuilt. + +Fri Oct 4 09:51:08 1996 Anthony Green + + * prep_cif.c (initialize_aggregate): Moved abi specification + to ffi_prep_cif(). + +Thu Oct 3 15:37:37 1996 Anthony Green + + * prep_cif.c (ffi_prep_cif): Changed values from void* to void**. + (initialize_aggregate): Fixed aggregate type initialization. + + * Makefile.am (EXTRA_DIST): Added support code for "make dist". + * Makefile.in: Regenerated. + +Wed Oct 2 11:41:57 1996 Anthony Green + + * debug.c, prep_cif: Created. + + * Makefile.am: Added debug.o and prep_cif.o to OBJ. + * Makefile.in: Regenerated. + + * Makefile.am (INCLUDES): Added missing -I../include + * Makefile.in: Regenerated. + +Tue Oct 1 17:11:51 1996 Anthony Green + + * error.c, Makefile.am: Created. + * Makefile.in: Generated. + +--- libffi/src/x86 -------------------------------------------------------- + +Sun Oct 4 16:27:17 1998 Anthony Green + + * sysv.S (retlongdouble): Fixed long long return value support. + * ffi.c (ffi_prep_cif_machdep): Ditto. + +Wed May 13 04:30:33 1998 Anthony Green + + * ffi.c (ffi_prep_cif_machdep): Fixed long double return value + support. + +Wed Apr 15 08:43:20 1998 Anthony Green + + * ffi.c (ffi_prep_args): small struct support was missing. + +Thu May 8 16:53:58 1997 Anthony Green + + * objects.mak: Removed. + +Mon Dec 2 15:12:58 1996 Tom Tromey + + * sysv.S: Use .balign, for a.out Linux boxes. + +Tue Oct 15 13:06:50 1996 Anthony Green + + * ffi.c: Clean up based on LCLint output. + Added funny /*@...@*/ comments to annotate source. + +Fri Oct 11 16:43:38 1996 Anthony Green + + * ffi.c (ffi_call): Added assertion for bad ABIs. + +Wed Oct 9 13:57:27 1996 Anthony Green + + * sysv.S (retdouble): Fixed double return problems. + + * ffi.c (ffi_call): Corrected fn arg definition. + (ffi_prep_cif_machdep): Fixed double return problems + +Tue Oct 8 12:12:49 1996 Anthony Green + + * ffi.c: Moved ffi_type definitions to types.c. + (ffi_prep_args): Fixed type promotion bug. + +Mon Oct 7 15:53:06 1996 Anthony Green + + * ffi.c (FFI_*_TYPEDEF): Removed redundant ';' + +Fri Oct 4 09:54:53 1996 Anthony Green + + * ffi.c (ffi_call): Removed FFI_ABI arg, and swapped + remaining args. + +Wed Oct 2 10:07:05 1996 Anthony Green + + * ffi.c, sysv.S, objects.mak: Created. + (ffi_prep_cif): cif->rvalue no longer initialized to NULL. + (ffi_prep_cif_machdep): Moved machine independent cif processing + to src/prep_cif.c. Introduced ffi_prep_cif_machdep(). + +--- libffi/src/mips ------------------------------------------------------- + +Tue Feb 17 17:18:07 1998 Anthony Green + + * o32.S: Fixed typo in comment. + + * ffi.c (ffi_prep_cif_machdep): Fixed argument processing. + +Thu May 8 16:53:58 1997 Anthony Green + + * o32.s, n32.s: Wrappers for SGI tool support. + + * objects.mak: Removed. + +Tue Oct 29 14:37:45 1996 Anthony Green + + * ffi.c (ffi_prep_args): Changed int z to size_t z. + +Tue Oct 15 13:17:25 1996 Anthony Green + + * n32.S: Fixed bad stack munging. + + * ffi.c: Moved prototypes for ffi_call_?32() to here from + ffi_mips.h because extended_cif is not defined in ffi_mips.h. + +Mon Oct 14 12:42:02 1996 Anthony Green + + * ffi.c: Interface changes based on feedback from Jim Blandy. + +Thu Oct 10 11:22:16 1996 Anthony Green + + * n32.S, ffi.c: Lots of changes to support passing and + returning structures with the n32 calling convention. + + * n32.S: Fixed fn pointer bug. + + * ffi.c (ffi_prep_cif_machdep): Fix for o32 structure + return values. + (ffi_prep_args): Fixed n32 structure passing when structures + partially fit in registers. + +Wed Oct 9 13:49:25 1996 Anthony Green + + * objects.mak: Added n32.o. + + * n32.S: Created. + + * ffi.c (ffi_prep_args): Added magic to support proper + n32 processing. + +Tue Oct 8 10:37:35 1996 Anthony Green + + * ffi.c: Moved ffi_type definitions to types.c. + (ffi_prep_args): Fixed type promotion bug. + + * o32.S: This code is only built for o32 compiles. + A lot of the #define cruft has moved to ffi_mips.h. + + * ffi.c (ffi_prep_cif_machdep): Fixed arg flags. Second arg + is only processed if the first is either a float or double. + +Mon Oct 7 15:33:59 1996 Anthony Green + + * o32.S: Modified to compile under each of o32, n32 and n64. + + * ffi.c (FFI_*_TYPEDEF): Removed redundant ';' + +Fri Oct 4 09:53:25 1996 Anthony Green + + * ffi.c (ffi_call): Removed FFI_ABI arg, and swapped + remaining args. + +Wed Oct 2 17:41:22 1996 Anthony Green + + * o32.S: Removed crufty definitions. + +Wed Oct 2 12:53:42 1996 Anthony Green + + * ffi.c (ffi_prep_cif): cif->rvalue no longer initialized to NULL. + (ffi_prep_cif_machdep): Moved all machine independent cif processing + to src/prep_cif.c. Introduced ffi_prep_cif_machdep. Return types + of FFI_TYPE_STRUCT are no different than FFI_TYPE_INT. + +Tue Oct 1 17:11:02 1996 Anthony Green + + * ffi.c, o32.S, object.mak: Created + +--- libffi/src/sparc ------------------------------------------------------ + +Tue Feb 24 16:33:18 1998 Anthony Green + + * ffi.c (ffi_prep_args): Added long double support. + +Thu May 8 16:53:58 1997 Anthony Green + + * objects.mak: Removed. + +Thu May 1 16:07:56 1997 Anthony Green + + * v8.S: Fixed minor portability problem reported by + Russ McManus . + +Tue Nov 26 14:12:43 1996 Anthony Green + + * v8.S: Used STACKFRAME define elsewhere. + + * ffi.c (ffi_prep_args): Zero out space when USING_PURIFY + is set. + (ffi_prep_cif_machdep): Allocate the correct stack frame + space for functions with < 6 args. + +Tue Oct 29 15:08:55 1996 Anthony Green + + * ffi.c (ffi_prep_args): int z is now size_t z. + +Mon Oct 14 13:31:24 1996 Anthony Green + + * v8.S (ffi_call_V8): Gordon rewrites this again. It looks + great now. + + * ffi.c (ffi_call): The comment about hijacked registers + is no longer valid after gordoni hacked v8.S. + + * v8.S (ffi_call_V8): Rewrote with gordoni. Much simpler. + + * v8.S, ffi.c: ffi_call() had changed to accept more than + two args, so v8.S had to change (because it hijacks incoming + arg registers). + + * ffi.c: Interface changes based on feedback from Jim Blandy. + +Thu Oct 10 17:48:16 1996 Anthony Green + + * ffi.c, v8.S, objects.mak: Created. + + diff --git a/native/libffi/ChangeLog.libgcj b/native/libffi/ChangeLog.libgcj deleted file mode 100644 index ea5d02f19b..0000000000 --- a/native/libffi/ChangeLog.libgcj +++ /dev/null @@ -1,40 +0,0 @@ -2004-01-14 Kelley Cook - - * configure.in: Add in AC_PREREQ(2.13) - -2003-02-20 Alexandre Oliva - - * configure.in: Propagate ORIGINAL_LD_FOR_MULTILIBS to - config.status. - * configure: Rebuilt. - -2002-01-27 Alexandre Oliva - - * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST. - Remove USE_LIBDIR conditional. - * Makefile.am (toolexecdir, toolexeclibdir): Don't override. - * Makefile.in, configure: Rebuilt. - -Mon Aug 9 18:33:38 1999 Rainer Orth - - * include/Makefile.in: Rebuilt. - * Makefile.in: Rebuilt - * Makefile.am (toolexeclibdir): Add $(MULTISUBDIR) even for native - builds. - Use USE_LIBDIR. - - * configure: Rebuilt. - * configure.in (USE_LIBDIR): Define for native builds. - Use lowercase in configure --help explanations. - -1999-08-08 Anthony Green - - * include/ffi.h.in (FFI_FN): Remove `...'. - -1999-08-08 Anthony Green - - * Makefile.in: Rebuilt. - * Makefile.am (AM_CFLAGS): Compile with -fexceptions. - - * src/x86/sysv.S: Add exception handling metadata. - diff --git a/native/libffi/ChangeLog.old b/native/libffi/ChangeLog.old new file mode 100644 index 0000000000..8216dc8194 --- /dev/null +++ b/native/libffi/ChangeLog.old @@ -0,0 +1,7410 @@ +<<<<<<< HEAD:native/libffi/ChangeLog.libffi-3.1 +======= +Libffi change logs used to be maintained in separate ChangeLog files. +These days we generate them directly from the git commit messages. +The old ChangeLog files are saved here in order to maintain the historical +record. + +============================================================================= +From the old ChangeLog.libffi-3.1 file... + +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878:native/libffi/ChangeLog.old +2014-03-16 Josh Triplett + + * ChangeLog: Archive to ChangeLog.libffi-3.1 and delete. Future + changelogs will come from git, with autogenerated snapshots shipped in + distributed tarballs. + +2014-03-16 Josh Triplett + + Add support for stdcall, thiscall, and fastcall on non-Windows + x86-32. + + Linux supports the stdcall calling convention, either via + functions explicitly declared with the stdcall attribute, or via + code compiled with -mrtd which effectively makes stdcall the + default. + + This introduces FFI_STDCALL, FFI_THISCALL, and FFI_FASTCALL on + non-Windows x86-32 platforms, as non-default calling conventions. + + * Makefile.am: Compile in src/x86/win32.S on non-Windows x86-32. + * src/x86/ffitarget.h: Add FFI_STDCALL, FFI_THISCALL, and + FFI_FASTCALL on non-Windows x86-32. Increase trampoline size to + accomodate these calling conventions, and unify some ifdeffery. + * src/x86/ffi.c: Add support for FFI_STDCALL, FFI_THISCALL, and + FFI_FASTCALL on non-Windows x86-32 platforms; update ifdeffery. + * src/x86/win32.S: Support compiling on non-Windows x86-32 + platforms. On those platforms, avoid redefining the SYSV symbols + already provided by src/x86/sysv.S. + * testsuite/libffi.call/closure_stdcall.c: Run on non-Windows. + #define __stdcall if needed. + * testsuite/libffi.call/closure_thiscall.c: Run on non-Windows. + #define __fastcall if needed. + * testsuite/libffi.call/fastthis1_win32.c: Run on non-Windows. + * testsuite/libffi.call/fastthis2_win32.c: Ditto. + * testsuite/libffi.call/fastthis3_win32.c: Ditto. + * testsuite/libffi.call/many2_win32.c: Ditto. + * testsuite/libffi.call/many_win32.c: Ditto. + * testsuite/libffi.call/strlen2_win32.c: Ditto. + * testsuite/libffi.call/strlen_win32.c: Ditto. + * testsuite/libffi.call/struct1_win32.c: Ditto. + * testsuite/libffi.call/struct2_win32.c: Ditto. + +2014-03-16 Josh Triplett + + * prep_cif.c: Remove unnecessary ifdef for X86_WIN32. + ffi_prep_cif_core had a special case for X86_WIN32, checking for + FFI_THISCALL in addition to the FFI_FIRST_ABI-to-FFI_LAST_ABI + range before returning FFI_BAD_ABI. However, on X86_WIN32, + FFI_THISCALL already falls in that range, making the special case + unnecessary. Remove it. + +2014-03-16 Josh Triplett + + * testsuite/libffi.call/closure_stdcall.c, + testsuite/libffi.call/closure_thiscall.c: Remove fragile stack + pointer checks. These files included inline assembly to save the + stack pointer before and after the call, and compare the values. + However, compilers can and do leave the stack in different states + for these two pieces of inline assembly, such as by saving a + temporary value on the stack across the call; observed with gcc + -Os, and verified as spurious through careful inspection of + disassembly. + +2014-03-16 Josh Triplett + + * testsuite/libffi.call/many.c: Avoid spurious failure due to + excess floating-point precision. + * testsuite/libffi.call/many_win32.c: Ditto. + +2014-03-16 Josh Triplett + + * libtool-ldflags: Re-add. + +2014-03-16 Josh Triplett + + * Makefile.in, aclocal.m4, compile, config.guess, config.sub, + configure, depcomp, include/Makefile.in, install-sh, + libtool-ldflags, ltmain.sh, m4/libtool.m4, m4/ltoptions.m4, + m4/ltsugar.m4, m4/ltversion.m4, m4/lt~obsolete.m4, + man/Makefile.in, mdate-sh, missing, testsuite/Makefile.in: Delete + autogenerated files from version control. + * .gitignore: Add autogenerated files. + * autogen.sh: New script to generate the autogenerated files. + * README: Document requirement to run autogen.sh when building + directly from version control. + * .travis.yml: Run autogen.sh + +2014-03-14 Anthony Green + + * configure, Makefile.in: Rebuilt. + +2014-03-10 Mike Hommey + + * configure.ac: Allow building for mipsel with Android NDK r8. + * Makefile.am (AM_MAKEFLAGS): Replace double quotes with single + quotes. + +2014-03-10 Landry Breuil + + * configure.ac: Ensure the linker supports @unwind sections in libffi. + +2014-03-01 Anthony Green + + * Makefile.am (EXTRA_DIST): Replace old scripts with + generate-darwin-source-and-headers.py. + * Makefile.in: Rebuilt. + +2014-02-28 Anthony Green + + * Makefile.am (AM_CFLAGS): Reintroduce missing -DFFI_DEBUG for + --enable-debug builds. + * Makefile.in: Rebuilt. + +2014-02-28 Makoto Kato + + * src/closures.c: Fix build failure when using clang for Android. + +2014-02-28 Marcin Wojdyr + + * libffi.pc.in (toolexeclibdir): use -L${toolexeclibdir} instead + of -L${libdir}. + +2014-02-28 Paulo Pizarro + + * src/bfin/sysv.S: Calling functions in shared libraries requires + considering the GOT. + +2014-02-28 Josh Triplett + + * src/x86/ffi64.c (classify_argument): Handle case where + FFI_TYPE_LONGDOUBLE == FFI_TYPE_DOUBLE. + +2014-02-28 Anthony Green + + * ltmain.sh: Generate with libtool-2.4.2.418. + * m4/libtool.m4, m4/ltoptions.m4, m4/ltversion.m4: Ditto. + * configure: Rebuilt. + +2014-02-28 Dominik Vogt + + * configure.ac (AC_ARG_ENABLE struct): Fix typo in help + message. + (AC_ARG_ENABLE raw_api): Ditto. + * configure, fficonfig.h.in: Rebuilt. + +2014-02-28 Will Newton + + * src/arm/sysv.S: Initialize IP register with FP. + +2014-02-28 Yufeng Zhang + + * src/aarch64/sysv.S (ffi_closure_SYSV): Use x29 as the + main CFA reg; update cfi_rel_offset. + +2014-02-15 Marcus Comstedt + + * src/powerpc/ffi_linux64.c, src/powerpc/linux64_closure.S: Remove + assumption on contents of r11 in closure. + +2014-02-09 Heiher + + * src/mips/n32.S: Fix call floating point va function. + +2014-01-21 Zachary Waldowski + + * src/aarch64/ffi.c: Fix missing semicolons on assertions under + debug mode. + +2013-12-30 Zachary Waldowski + + * .gitignore: Exclude darwin_* generated source and build_* trees. + * src/aarch64/ffi.c, src/arm/ffi.c, src/x86/ffi.c: Inhibit Clang + previous prototype warnings. + * src/arm/ffi.c: Prevent NULL dereference, fix short type warning + * src/dlmalloc.c: Fix warnings from set_segment_flags return type, + and the native use of size_t for malloc on platforms + * src/arm/sysv.S: Use unified syntax. Clang clean-ups for + ARM_FUNC_START. + * generate-osx-source-and-headers.py: Remove. + * build-ios.sh: Remove. + * libffi.xcodeproj/project.pbxproj: Rebuild targets. Include + x86_64+aarch64 pieces in library. Export headers properly. + * src/x86/ffi64.c: More Clang warning clean-ups. + * src/closures.c (open_temp_exec_file_dir): Use size_t. + * src/prep_cif.c (ffi_prep_cif_core): Cast ALIGN result. + * src/aarch64/sysv.S: Use CNAME for global symbols. Only use + .size for ELF targets. + * src/aarch64/ffi.c: Clean up for double == long double. Clean up + from Clang warnings. Use Clang cache invalidation builtin. Use + size_t in place of unsigned in many places. Accommodate for + differences in Apple AArch64 ABI. + +2013-12-02 Daniel Rodríguez Troitiño + + * generate-darwin-source-and-headers.py: Clean up, modernize, + merged version of previous scripts. + +2013-11-21 Anthony Green + + * configure, Makefile.in, include/Makefile.in, include/ffi.h.in, + man/Makefile.in, testsuite/Makefile.in, fficonfig.h.in: Rebuilt. + +2013-11-21 Alan Modra + + * Makefile.am (EXTRA_DIST): Add new src/powerpc files. + (nodist_libffi_la_SOURCES ): Likewise. + * configure.ac (HAVE_LONG_DOUBLE_VARIANT): Define for powerpc. + * include/ffi.h.in (ffi_prep_types): Declare. + * src/prep_cif.c (ffi_prep_cif_core): Call ffi_prep_types. + * src/types.c (FFI_NONCONST_TYPEDEF): Define and use for + HAVE_LONG_DOUBLE_VARIANT. + * src/powerpc/ffi_powerpc.h: New file. + * src/powerpc/ffi.c: Split into.. + * src/powerpc/ffi_sysv.c: ..new file, and.. + * src/powerpc/ffi_linux64.c: ..new file, rewriting parts. + * src/powerpc/ffitarget.h (enum ffi_abi): Rewrite powerpc ABI + selection as bits controlling features. + * src/powerpc/linux64.S: For consistency, use POWERPC64 rather + than __powerpc64__. + * src/powerpc/linux64_closure.S: Likewise. + * src/powerpc/ppc_closure.S: Likewise. Move .note.FNU-stack + inside guard. + * src/powerpc/sysv.S: Likewise. + * configure: Regenerate. + * fficonfig.h.in: Regenerate. + * Makefile.in: Regenerate. + +2013-11-20 Alan Modra + + * src/powerpc/ffi.c (ffi_prep_cif_machdep_core): Use + NUM_FPR_ARG_REGISTERS64 and NUM_GPR_ARG_REGISTERS64 not their + 32-bit versions for 64-bit code. + * src/powerpc/linux64_closure.S: Don't use the return value area + as a parameter save area on ELFv2. + +2013-11-18 Iain Sandoe + + * src/powerpc/darwin.S (EH): Correct use of pcrel FDE encoding. + * src/powerpc/darwin_closure.S (EH): Likewise. Modernise picbase + labels. + +2013-11-18 Anthony Green + + * src/arm/ffi.c (ffi_call): Hoist declaration of temp to top of + function. + * src/arm/ffi.c (ffi_closure_inner): Moderize function declaration + to appease compiler. + Thanks for Gregory P. Smith . + +2013-11-18 Anthony Green + + * README (tested): Mention PowerPC ELFv2. + +2013-11-16 Alan Modra + + * src/powerpc/ppc_closure.S: Move errant #endif to where it belongs. + Don't bl .Luint128. + +2013-11-16 Alan Modra + + * src/powerpc/ffi.c (ffi_prep_cif_machdep_core): Use #if _CALL_ELF + test to select parameter save sizing for ELFv2 vs. ELFv1. + * src/powerpc/ffitarget.h (FFI_V2_TYPE_FLOAT_HOMOG, + FFI_V2_TYPE_DOUBLE_HOMOG, FFI_V2_TYPE_SMALL_STRUCT): Define. + (FFI_TRAMPOLINE_SIZE): Define variant for ELFv2. + * src/powerpc/ffi.c (FLAG_ARG_NEEDS_PSAVE): Define. + (discover_homogeneous_aggregate): New function. + (ffi_prep_args64): Adjust start of param save area for ELFv2. + Handle homogenous floating point struct parms. + (ffi_prep_cif_machdep_core): Adjust space calculation for ELFv2. + Handle ELFv2 return values. Set FLAG_ARG_NEEDS_PSAVE. Handle + homogenous floating point structs. + (ffi_call): Increase size of smst_buffer for ELFv2. Handle ELFv2. + (flush_icache): Compile for ELFv2. + (ffi_prep_closure_loc): Set up ELFv2 trampoline. + (ffi_closure_helper_LINUX64): Don't return all structs directly + to caller. Handle homogenous floating point structs. Handle + ELFv2 struct return values. + * src/powerpc/linux64.S (ffi_call_LINUX64): Set up r2 for + ELFv2. Adjust toc save location. Call function pointer using + r12. Handle FLAG_RETURNS_SMST. Don't predict branches. + * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Set up r2 + for ELFv2. Define ELFv2 versions of STACKFRAME, PARMSAVE, and + RETVAL. Handle possibly missing parameter save area. Handle + ELFv2 return values. + (.note.GNU-stack): Move inside outer #ifdef. + +2013-11-16 Alan Modra + + * src/powerpc/ffi.c (ffi_prep_cif_machdep): Revert 2013-02-08 + change. Do not consume an int arg when returning a small struct + for FFI_SYSV ABI. + (ffi_call): Only use bounce buffer when FLAG_RETURNS_SMST. + Properly copy bounce buffer to destination. + * src/powerpc/sysv.S: Revert 2013-02-08 change. + * src/powerpc/ppc_closure.S: Remove stray '+'. + +2013-11-16 Alan Modra + + * src/powerpc/ffi.c (ffi_prep_args64): Align struct parameters + according to __STRUCT_PARM_ALIGN__. + (ffi_prep_cif_machdep_core): Likewise. + (ffi_closure_helper_LINUX64): Likewise. + +2013-11-16 Alan Modra + + * src/powerpc/linux64.S (ffi_call_LINUX64): Tweak restore of r28. + (.note.GNU-stack): Move inside outer #ifdef. + * src/powerpc/linux64_closure.S (STACKFRAME, PARMSAVE, + RETVAL): Define and use throughout. + (ffi_closure_LINUX64): Save fprs before buying stack. + (.note.GNU-stack): Move inside outer #ifdef. + +2013-11-16 Alan Modra + + * src/powerpc/ffitarget.h (FFI_TARGET_SPECIFIC_VARIADIC): Define. + (FFI_EXTRA_CIF_FIELDS): Define. + * src/powerpc/ffi.c (ffi_prep_args64): Save fprs as per the + ABI, not to both fpr and param save area. + (ffi_prep_cif_machdep_core): Renamed from ffi_prep_cif_machdep. + Keep initial flags. Formatting. Remove dead FFI_LINUX_SOFT_FLOAT + code. + (ffi_prep_cif_machdep, ffi_prep_cif_machdep_var): New functions. + (ffi_closure_helper_LINUX64): Pass floating point as per ABI, + not to both fpr and parameter save areas. + + * libffi/testsuite/libffi.call/cls_double_va.c (main): Correct + function cast and don't call ffi_prep_cif. + * libffi/testsuite/libffi.call/cls_longdouble_va.c (main): Likewise. + +2013-11-15 Andrew Haley + + * doc/libffi.texi (Closure Example): Fix the sample code. + * doc/libffi.info, doc/stamp-vti, doc/version.texi: Rebuilt. + +2013-11-15 Andrew Haley + + * testsuite/libffi.call/va_struct1.c (main): Fix broken test. + * testsuite/libffi.call/cls_uint_va.c (cls_ret_T_fn): Likewise + * testsuite/libffi.call/cls_struct_va1.c (test_fn): Likewise. + * testsuite/libffi.call/va_1.c (main): Likewise. + +2013-11-14 David Schneider + + * src/arm/ffi.c: Fix register allocation for mixed float and + doubles. + * testsuite/libffi.call/cls_many_mixed_float_double.c: Testcase + for many mixed float and double arguments. + +2013-11-13 Alan Modra + + * doc/libffi.texi (Simple Example): Correct example code. + * doc/libffi.info, doc/stamp-vti, doc/version.texi: Rebuilt. + +2013-11-13 Anthony Green + + * include/ffi_common.h: Respect HAVE_ALLOCA_H for GNU compiler + based build. (Thanks to tmr111116 on github) + +2013-11-09 Anthony Green + + * m4/libtool.m4: Refresh. + * configure, Makefile.in: Rebuilt. + * README: Add more notes about next release. + +2013-11-09 Shigeharu TAKENO + + * m4/ax_gcc_archflag.m4 (ax_gcc_arch): Don't recognize + UltraSPARC-IIi as ultrasparc3. + +2013-11-06 Mark Kettenis + + * src/x86/freebsd.S (ffi_call_SYSV): Align the stack pointer to + 16-bytes. + +2013-11-06 Konstantin Belousov + + * src/x86/freebsd.S (ffi_closure_raw_SYSV): Mark the assembler + source as not requiring executable stack. + +2013-11-02 Anthony Green + + * doc/libffi.texi (The Basics): Clarify return value buffer size + requirements. Also, NULL result buffer pointers are no longer + supported. + * doc/libffi.info: Rebuilt. + +2013-11-02 Mischa Jonker + + * Makefile.am (nodist_libffi_la_SOURCES): Fix build error. + * Makefile.in: Rebuilt. + +2013-11-02 David Schneider + + * src/arm/ffi.c: more robust argument handling for closures on arm hardfloat + * testsuite/libffi.call/many_mixed.c: New file. + * testsuite/libffi.call/cls_many_mixed_args.c: More tests. + +2013-11-02 Vitaly Budovski + + * src/x86/ffi.c (ffi_prep_cif_machdep): Don't align stack for win32. + +2013-10-23 Mark H Weaver + + * src/mips/ffi.c: Fix handling of uint32_t arguments on the + MIPS N32 ABI. + +2013-10-13 Sandra Loosemore + + * README: Add Nios II to table of supported platforms. + * Makefile.am (EXTRA_DIST): Add nios2 files. + (nodist_libffi_la_SOURCES): Likewise. + * Makefile.in: Regenerated. + * configure.ac (nios2*-linux*): New host. + (NIOS2): Add AM_CONDITIONAL. + * configure: Regenerated. + * src/nios2/ffi.c: New. + * src/nios2/ffitarget.h: New. + * src/nios2/sysv.S: New. + * src/prep_cif.c (initialize_aggregate): Handle extra structure + alignment via FFI_AGGREGATE_ALIGNMENT. + (ffi_prep_cif_core): Conditionalize structure return for NIOS2. + +2013-10-10 Sandra Loosemore + + * testsuite/libffi.call/cls_many_mixed_args.c (cls_ret_double_fn): + Fix uninitialized variable. + +2013-10-11 Marcus Shawcroft + + * testsuite/libffi.call/many.c (many): Replace * with +. + +2013-10-08 Ondřej Bílka + + * src/aarch64/ffi.c, src/aarch64/sysv.S, src/arm/ffi.c, + src/arm/gentramp.sh, src/bfin/sysv.S, src/closures.c, + src/dlmalloc.c, src/ia64/ffi.c, src/microblaze/ffi.c, + src/microblaze/sysv.S, src/powerpc/darwin_closure.S, + src/powerpc/ffi.c, src/powerpc/ffi_darwin.c, src/sh/ffi.c, + src/tile/tile.S, testsuite/libffi.call/nested_struct11.c: Fix + spelling errors. + +2013-10-08 Anthony Green + + * aclocal.m4, compile, config.guess, config.sub, depcomp, + install-sh, mdate-sh, missing, texinfo.tex: Update from upstream. + * configure.ac: Update version to 3.0.14-rc0. + * Makefile.in, configure, Makefile.in, include/Makefile.in, + man/Makefile.in, testsuite/Makefile.in: Rebuilt. + * README: Mention M88K and VAX. + +2013-07-15 Miod Vallat + + * Makefile.am, + configure.ac, + src/m88k/ffi.c, + src/m88k/ffitarget.h, + src/m88k/obsd.S, + src/vax/elfbsd.S, + src/vax/ffi.c, + src/vax/ffitarget.h: Add m88k and vax support. + +2013-06-24 Alan Modra + + * src/powerpc/ffi.c (ffi_prep_args_SYSV): Move var declaration + before statements. + (ffi_prep_args64): Support little-endian. + (ffi_closure_helper_SYSV, ffi_closure_helper_LINUX64): Likewise. + * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Likewise. + * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Likewise. + +2013-06-12 Mischa Jonker + + * configure.ac: Add support for ARC. + * Makefile.am: Likewise. + * README: Add ARC details. + * src/arc/arcompact.S: New. + * src/arc/ffi.c: Likewise. + * src/arc/ffitarget.h: Likewise. + +2013-03-28 David Schneider + + * src/arm/ffi.c: Fix support for ARM hard-float calling convention. + * src/arm/sysv.S: call different methods for SYSV and VFP ABIs. + * testsuite/libffi.call/cls_many_mixed_args.c: testcase for a closure with + mixed arguments, many doubles. + * testsuite/libffi.call/many_double.c: testcase for calling a function using + more than 8 doubles. + * testcase/libffi.call/many.c: use absolute value to check result against an + epsilon + +2013-03-17 Anthony Green + + * README: Update for 3.0.13. + * configure.ac: Ditto. + * configure: Rebuilt. + * doc/*: Update version. + +2013-03-17 Dave Korn + + * src/closures.c (is_emutramp_enabled + [!FFI_MMAP_EXEC_EMUTRAMP_PAX]): Move default definition outside + enclosing #if scope. + +2013-03-17 Anthony Green + + * configure.ac: Only modify toolexecdir in certain cases. + * configure: Rebuilt. + +2013-03-16 Gilles Talis + + * src/powerpc/ffi.c (ffi_prep_args_SYSV): Don't use + fparg_count,etc on __NO_FPRS__ targets. + +2013-03-16 Alan Hourihane + + * src/m68k/sysv.S (epilogue): Don't use extb instruction on + m680000 machines. + +2013-03-16 Alex Gaynor + + * src/x86/ffi.c (ffi_prep_cif_machdep): Always align stack. + +2013-03-13 Markos Chandras + + * configure.ac: Add support for Imagination Technologies Meta. + * Makefile.am: Likewise. + * README: Add Imagination Technologies Meta details. + * src/metag/ffi.c: New. + * src/metag/ffitarget.h: Likewise. + * src/metag/sysv.S: Likewise. + +2013-02-24 Andreas Schwab + + * doc/libffi.texi (Structures): Fix missing category argument of + @deftp. + +2013-02-11 Anthony Green + + * configure.ac: Update release number to 3.0.12. + * configure: Rebuilt. + * README: Update release info. + +2013-02-10 Anthony Green + + * README: Add Moxie. + * src/moxie/ffi.c: Created. + * src/moxie/eabi.S: Created. + * src/moxie/ffitarget.h: Created. + * Makefile.am (nodist_libffi_la_SOURCES): Add Moxie. + * Makefile.in: Rebuilt. + * configure.ac: Add Moxie. + * configure: Rebuilt. + * testsuite/libffi.call/huge_struct.c: Disable format string + warnings for moxie*-*-elf tests. + +2013-02-10 Anthony Green + + * Makefile.am (LTLDFLAGS): Fix reference. + * Makefile.in: Rebuilt. + +2013-02-10 Anthony Green + + * README: Update supported platforms. Update test results link. + +2013-02-09 Anthony Green + + * testsuite/libffi.call/negint.c: Remove forced -O2. + * testsuite/libffi.call/many2.c (foo): Remove GCCism. + * testsuite/libffi.call/ffitest.h: Add default PRIuPTR definition. + + * src/sparc/v8.S (ffi_closure_v8): Import ancient ulonglong + closure return type fix developed by Martin v. Löwis for cpython + fork. + +2013-02-08 Andreas Tobler + + * src/powerpc/ffi.c (ffi_prep_cif_machdep): Fix small struct + support. + * src/powerpc/sysv.S: Ditto. + +2013-02-08 Anthony Green + + * testsuite/libffi.call/cls_longdouble.c: Remove xfail for + arm*-*-*. + +2013-02-08 Anthony Green + + * src/sparc/ffi.c (ffi_prep_closure_loc): Fix cache flushing for GCC. + +2013-02-08 Matthias Klose + + * man/ffi_prep_cif.3: Clean up for debian linter. + +2013-02-08 Peter Bergner + + * src/powerpc/ffi.c (ffi_prep_args_SYSV): Account for FP args pushed + on the stack. + +2013-02-08 Anthony Green + + * Makefile.am (EXTRA_DIST): Add missing files. + * testsuite/Makefile.am (EXTRA_DIST): Ditto. + * Makefile.in: Rebuilt. + +2013-02-08 Anthony Green + + * configure.ac: Move sparc asm config checks to within functions + for compatibility with sun tools. + * configure: Rebuilt. + * src/sparc/ffi.c (ffi_prep_closure_loc): Flush cache on v9 + systems. + * src/sparc/v8.S (ffi_flush_icache): Implement a sparc v9 cache + flusher. + +2013-02-08 Nathan Rossi + + * src/microblaze/ffi.c (ffi_closure_call_SYSV): Fix handling of + small big-endian structures. + (ffi_prep_args): Ditto. + +2013-02-07 Anthony Green + + * src/sparc/v8.S (ffi_call_v8): Fix typo from last patch + (effectively hiding ffi_call_v8). + +2013-02-07 Anthony Green + + * configure.ac: Update bug reporting address. + * configure.in: Rebuild. + + * src/sparc/v8.S (ffi_flush_icache): Out-of-line cache flusher for + Sun compiler. + * src/sparc/ffi.c (ffi_call): Remove warning. + Call ffi_flush_icache for non-GCC builds. + (ffi_prep_closure_loc): Use ffi_flush_icache. + + * Makefile.am (EXTRA_DIST): Add libtool-ldflags. + * Makefile.in: Rebuilt. + * libtool-ldflags: New file. + +2013-02-07 Daniel Schepler + + * configure.ac: Correctly identify x32 systems as 64-bit. + * m4/libtool.m4: Remove libtool expr error. + * aclocal.m4, configure: Rebuilt. + +2013-02-07 Anthony Green + + * configure.ac: Fix GCC usage test. + * configure: Rebuilt. + * README: Mention LLVM/GCC x86_64 issue. + * testsuite/Makefile.in: Rebuilt. + +2013-02-07 Anthony Green + + * testsuite/libffi.call/cls_double_va.c (main): Replace // style + comments with /* */ for xlc compiler. + * testsuite/libffi.call/stret_large.c (main): Ditto. + * testsuite/libffi.call/stret_large2.c (main): Ditto. + * testsuite/libffi.call/nested_struct1.c (main): Ditto. + * testsuite/libffi.call/huge_struct.c (main): Ditto. + * testsuite/libffi.call/float_va.c (main): Ditto. + * testsuite/libffi.call/cls_struct_va1.c (main): Ditto. + * testsuite/libffi.call/cls_pointer_stack.c (main): Ditto. + * testsuite/libffi.call/cls_pointer.c (main): Ditto. + * testsuite/libffi.call/cls_longdouble_va.c (main): Ditto. + +2013-02-06 Anthony Green + + * man/ffi_prep_cif.3: Clean up for debian lintian checker. + +2013-02-06 Anthony Green + + * Makefile.am (pkgconfigdir): Add missing pkgconfig install bits. + * Makefile.in: Rebuild. + +2013-02-02 Mark H Weaver + + * src/x86/ffi64.c (ffi_call): Sign-extend integer arguments passed + via general purpose registers. + +2013-01-21 Nathan Rossi + + * README: Add MicroBlaze details. + * Makefile.am: Add MicroBlaze support. + * configure.ac: Likewise. + * src/microblaze/ffi.c: New. + * src/microblaze/ffitarget.h: Likewise. + * src/microblaze/sysv.S: Likewise. + +2013-01-21 Nathan Rossi + * testsuite/libffi.call/return_uc.c: Fixed issue. + +2013-01-21 Chris Zankel + + * README: Add Xtensa support. + * Makefile.am: Likewise. + * configure.ac: Likewise. + * Makefile.in Regenerate. + * configure: Likewise. + * src/prep_cif.c: Handle Xtensa. + * src/xtensa: New directory. + * src/xtensa/ffi.c: New file. + * src/xtensa/ffitarget.h: Ditto. + * src/xtensa/sysv.S: Ditto. + +2013-01-11 Anthony Green + + * src/powerpc/ffi_darwin.c (ffi_prep_args): Replace // style + comments with /* */ for xlc compiler. + * src/powerpc/aix.S (ffi_call_AIX): Ditto. + * testsuite/libffi.call/ffitest.h (allocate_mmap): Delete + deprecated inline function. + * testsuite/libffi.special/ffitestcxx.h: Ditto. + * README: Add update for AIX support. + +2013-01-11 Anthony Green + + * configure.ac: Robustify pc relative reloc check. + * m4/ax_cc_maxopt.m4: Don't -malign-double. This is an ABI + changing option for 32-bit x86. + * aclocal.m4, configure: Rebuilt. + * README: Update supported target list. + +2013-01-10 Anthony Green + + * README (tested): Add Compiler column to table. + +2013-01-10 Anthony Green + + * src/x86/ffi64.c (struct register_args): Make sse array and array + of unions for sunpro compiler compatibility. + +2013-01-10 Anthony Green + + * configure.ac: Test target platform size_t size. Handle both 32 + and 64-bit builds for x86_64-* and i?86-* targets (allowing for + CFLAG option to change default settings). + * configure, aclocal.m4: Rebuilt. + +2013-01-10 Anthony Green + + * testsuite/libffi.special/special.exp: Only run exception + handling tests when using GNU compiler. + + * m4/ax_compiler_vendor.m4: New file. + * configure.ac: Test for compiler vendor and don't use + AX_CFLAGS_WARN_ALL with the sun compiler. + * aclocal.m4, configure: Rebuilt. + +2013-01-10 Anthony Green + + * include/ffi_common.h: Don't use GCCisms to define types when + building with the SUNPRO compiler. + +2013-01-10 Anthony Green + + * configure.ac: Put local.exp in the right place. + * configure: Rebuilt. + + * src/x86/ffi.c: Update comment about regparm function attributes. + * src/x86/sysv.S (ffi_closure_SYSV): The SUNPRO compiler requires + that all function arguments be passed on the stack (no regparm + support). + +2013-01-08 Anthony Green + + * configure.ac: Generate local.exp. This sets CC_FOR_TARGET + when we are using the vendor compiler. + * testsuite/Makefile.am (EXTRA_DEJAGNU_SITE_CONFIG): Point to + ../local.exp. + * configure, testsuite/Makefile.in: Rebuilt. + + * testsuite/libffi.call/call.exp: Run tests with different + options, depending on whether or not we are using gcc or the + vendor compiler. + * testsuite/lib/libffi.exp (libffi-init): Set using_gcc based on + whether or not we are building/testing with gcc. + +2013-01-08 Anthony Green + + * configure.ac: Switch x86 solaris target to X86 by default. + * configure: Rebuilt. + +2013-01-08 Anthony Green + + * configure.ac: Fix test for read-only eh_frame. + * configure: Rebuilt. + +2013-01-08 Anthony Green + + * src/x86/sysv.S, src/x86/unix64.S: Only emit DWARF unwind info + when building with the GNU toolchain. + * testsuite/libffi.call/ffitest.h (CHECK): Fix for Solaris vendor + compiler. + +2013-01-07 Thorsten Glaser + + * testsuite/libffi.call/cls_uchar_va.c, + testsuite/libffi.call/cls_ushort_va.c, + testsuite/libffi.call/va_1.c: Testsuite fixes. + +2013-01-07 Thorsten Glaser + + * src/m68k/ffi.c (CIF_FLAGS_SINT8, CIF_FLAGS_SINT16): Define. + (ffi_prep_cif_machdep): Fix 8-bit and 16-bit signed calls. + * src/m68k/sysv.S (ffi_call_SYSV, ffi_closure_SYSV): Ditto. + +2013-01-04 Anthony Green + + * Makefile.am (AM_CFLAGS): Don't automatically add -fexceptions + and -Wall. This is set in the configure script after testing for + GCC. + * Makefile.in: Rebuilt. + +2013-01-02 rofl0r + + * src/powerpc/ffi.c (ffi_prep_cif_machdep): Fix build error on ppc + when long double == double. + +2013-01-02 Reini Urban + + * Makefile.am (libffi_la_LDFLAGS): Add -no-undefined to LDFLAGS + (required for shared libs on cygwin/mingw). + * Makefile.in: Rebuilt. + +2012-10-31 Alan Modra + + * src/powerpc/linux64_closure.S: Add new ABI support. + * src/powerpc/linux64.S: Likewise. + +2012-10-30 Magnus Granberg + Pavel Labushev + + * configure.ac: New options pax_emutramp + * configure, fficonfig.h.in: Regenerated + * src/closures.c: New function emutramp_enabled_check() and + checks. + +2012-10-30 Frederick Cheung + + * configure.ac: Enable FFI_MAP_EXEC_WRIT for Darwin 12 (mountain + lion) and future version. + * configure: Rebuild. + +2012-10-30 James Greenhalgh + Marcus Shawcroft + + * README: Add details of aarch64 port. + * src/aarch64/ffi.c: New. + * src/aarch64/ffitarget.h: Likewise. + * src/aarch64/sysv.S: Likewise. + * Makefile.am: Support aarch64. + * configure.ac: Support aarch64. + * Makefile.in, configure: Rebuilt. + +2012-10-30 James Greenhalgh + Marcus Shawcroft + + * testsuite/lib/libffi.exp: Add support for aarch64. + * testsuite/libffi.call/cls_struct_va1.c: New. + * testsuite/libffi.call/cls_uchar_va.c: Likewise. + * testsuite/libffi.call/cls_uint_va.c: Likewise. + * testsuite/libffi.call/cls_ulong_va.c: Likewise. + * testsuite/libffi.call/cls_ushort_va.c: Likewise. + * testsuite/libffi.call/nested_struct11.c: Likewise. + * testsuite/libffi.call/uninitialized.c: Likewise. + * testsuite/libffi.call/va_1.c: Likewise. + * testsuite/libffi.call/va_struct1.c: Likewise. + * testsuite/libffi.call/va_struct2.c: Likewise. + * testsuite/libffi.call/va_struct3.c: Likewise. + +2012-10-12 Walter Lee + + * Makefile.am: Add TILE-Gx/TILEPro support. + * configure.ac: Likewise. + * Makefile.in: Regenerate. + * configure: Likewise. + * src/prep_cif.c (ffi_prep_cif_core): Handle TILE-Gx/TILEPro. + * src/tile: New directory. + * src/tile/ffi.c: New file. + * src/tile/ffitarget.h: Ditto. + * src/tile/tile.S: Ditto. + +2012-10-12 Matthias Klose + + * generate-osx-source-and-headers.py: Normalize whitespace. + +2012-09-14 David Edelsohn + + * configure: Regenerated. + +2012-08-26 Andrew Pinski + + PR libffi/53014 + * src/mips/ffi.c (ffi_prep_closure_loc): Allow n32 with soft-float and n64 with + soft-float. + +2012-08-08 Uros Bizjak + + * src/s390/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test, + just return FFI_BAD_ABI when things are wrong. + +2012-07-18 H.J. Lu + + PR libffi/53982 + PR libffi/53973 + * src/x86/ffitarget.h: Check __ILP32__ instead of __LP64__ for x32. + (FFI_SIZEOF_JAVA_RAW): Defined to 4 for x32. + +2012-05-16 H.J. Lu + + * configure: Regenerated. + +2012-05-05 Nicolas Lelong + + * libffi.xcodeproj/project.pbxproj: Fixes. + * README: Update for iOS builds. + +2012-04-23 Alexandre Keunecke I. de Mendonca + + * configure.ac: Add Blackfin/sysv support + * Makefile.am: Add Blackfin/sysv support + * src/bfin/ffi.c: Add Blackfin/sysv support + * src/bfin/ffitarget.h: Add Blackfin/sysv support + +2012-04-11 Anthony Green + + * Makefile.am (EXTRA_DIST): Add new script. + * Makefile.in: Rebuilt. + +2012-04-11 Zachary Waldowski + + * generate-ios-source-and-headers.py, + libffi.xcodeproj/project.pbxproj: Support a Mac static library via + Xcode. Set iOS compatibility to 4.0. Move iOS trampoline + generation into an Xcode "run script" phase. Include both as + Xcode build scripts. Don't always regenerate config files. + +2012-04-10 Anthony Green + + * src/powerpc/ffi_darwin.c (ffi_prep_args): Add missing semicolon. + +2012-04-06 Anthony Green + + * Makefile.am (EXTRA_DIST): Add new iOS/xcode files. + * Makefile.in: Rebuilt. + +2012-04-06 Mike Lewis + + * generate-ios-source-and-headers.py: New file. + * libffi.xcodeproj/project.pbxproj: New file. + * README: Update instructions on building iOS binary. + * build-ios.sh: Delete. + +2012-04-06 Anthony Green + + * src/x86/ffi64.c (UINT128): Define differently for Intel and GNU + compilers, then use it. + +2012-04-06 H.J. Lu + + * m4/libtool.m4 (_LT_ENABLE_LOCK): Support x32. + +2012-04-06 Anthony Green + + * testsuite/Makefile.am (EXTRA_DIST): Add missing test cases. + * testsuite/Makefile.in: Rebuilt. + +2012-04-05 Zachary Waldowski + + * include/ffi.h.in: Add missing trampoline table fields. + * src/arm/sysv.S: Fix ENTRY definition, and wrap symbol references + in CNAME. + * src/x86/ffi.c: Wrap Windows specific code in ifdefs. + +2012-04-02 Peter Bergner + + * src/powerpc/ffi.c (ffi_prep_args_SYSV): Declare double_tmp. + Silence casting pointer to integer of different size warning. + Delete goto to previously deleted label. + (ffi_call): Silence possibly undefined warning. + (ffi_closure_helper_SYSV): Declare variable type. + +2012-04-02 Peter Rosin + + * src/x86/win32.S (ffi_call_win32): Sign/zero extend the return + value in the Intel version as is already done for the AT&T version. + (ffi_closure_SYSV): Likewise. + (ffi_closure_raw_SYSV): Likewise. + (ffi_closure_STDCALL): Likewise. + +2012-03-29 Peter Rosin + + * src/x86/win32.S (ffi_closure_raw_THISCALL): Unify the frame + generation, fix the ENDP label and remove the surplus third arg + from the 'lea' insn. + +2012-03-29 Peter Rosin + + * src/x86/win32.S (ffi_closure_raw_SYSV): Make the 'stubraw' label + visible outside the PROC, so that ffi_closure_raw_THISCALL can see + it. Also instruct the assembler to add a frame to the function. + +2012-03-23 Peter Rosin + + * Makefile.am (AM_CPPFLAGS): Add -DFFI_BUILDING. + * Makefile.in: Rebuilt. + * include/ffi.h.in [MSVC]: Add __declspec(dllimport) decorations + to all data exports, when compiling libffi clients using MSVC. + +2012-03-29 Peter Rosin + + * src/x86/ffitarget.h (ffi_abi): Add new ABI FFI_MS_CDECL and + make it the default for MSVC. + (FFI_TYPE_MS_STRUCT): New structure return convention. + * src/x86/ffi.c (ffi_prep_cif_machdep): Tweak the structure + return convention for FFI_MS_CDECL to be FFI_TYPE_MS_STRUCT + instead of an ordinary FFI_TYPE_STRUCT. + (ffi_prep_args): Treat FFI_TYPE_MS_STRUCT as FFI_TYPE_STRUCT. + (ffi_call): Likewise. + (ffi_prep_incoming_args_SYSV): Likewise. + (ffi_raw_call): Likewise. + (ffi_prep_closure_loc): Treat FFI_MS_CDECL as FFI_SYSV. + * src/x86/win32.S (ffi_closure_SYSV): For FFI_TYPE_MS_STRUCT, + return a pointer to the result structure in eax and don't pop + that pointer from the stack, the caller takes care of it. + (ffi_call_win32): Treat FFI_TYPE_MS_STRUCT as FFI_TYPE_STRUCT. + (ffi_closure_raw_SYSV): Likewise. + +2012-03-22 Peter Rosin + + * testsuite/libffi.call/closure_stdcall.c [MSVC]: Add inline + assembly version with Intel syntax. + * testsuite/libffi.call/closure_thiscall.c [MSVC]: Likewise. + +2012-03-23 Peter Rosin + + * testsuite/libffi.call/ffitest.h: Provide abstration of + __attribute__((fastcall)) in the form of a __FASTCALL__ + define. Define it to __fastcall for MSVC. + * testsuite/libffi.call/fastthis1_win32.c: Use the above. + * testsuite/libffi.call/fastthis2_win32.c: Likewise. + * testsuite/libffi.call/fastthis3_win32.c: Likewise. + * testsuite/libffi.call/strlen2_win32.c: Likewise. + * testsuite/libffi.call/struct1_win32.c: Likewise. + * testsuite/libffi.call/struct2_win32.c: Likewise. + +2012-03-22 Peter Rosin + + * src/x86/win32.S [MSVC] (ffi_closure_THISCALL): Remove the manual + frame on function entry, MASM adds one automatically. + +2012-03-22 Peter Rosin + + * testsuite/libffi.call/ffitest.h [MSVC]: Add kludge for missing + bits in the MSVC headers. + +2012-03-22 Peter Rosin + + * testsuite/libffi.call/cls_12byte.c: Adjust to the C89 style + with no declarations after statements. + * testsuite/libffi.call/cls_16byte.c: Likewise. + * testsuite/libffi.call/cls_18byte.c: Likewise. + * testsuite/libffi.call/cls_19byte.c: Likewise. + * testsuite/libffi.call/cls_1_1byte.c: Likewise. + * testsuite/libffi.call/cls_20byte.c: Likewise. + * testsuite/libffi.call/cls_20byte1.c: Likewise. + * testsuite/libffi.call/cls_24byte.c: Likewise. + * testsuite/libffi.call/cls_2byte.c: Likewise. + * testsuite/libffi.call/cls_3_1byte.c: Likewise. + * testsuite/libffi.call/cls_3byte1.c: Likewise. + * testsuite/libffi.call/cls_3byte2.c: Likewise. + * testsuite/libffi.call/cls_4_1byte.c: Likewise. + * testsuite/libffi.call/cls_4byte.c: Likewise. + * testsuite/libffi.call/cls_5_1_byte.c: Likewise. + * testsuite/libffi.call/cls_5byte.c: Likewise. + * testsuite/libffi.call/cls_64byte.c: Likewise. + * testsuite/libffi.call/cls_6_1_byte.c: Likewise. + * testsuite/libffi.call/cls_6byte.c: Likewise. + * testsuite/libffi.call/cls_7_1_byte.c: Likewise. + * testsuite/libffi.call/cls_7byte.c: Likewise. + * testsuite/libffi.call/cls_8byte.c: Likewise. + * testsuite/libffi.call/cls_9byte1.c: Likewise. + * testsuite/libffi.call/cls_9byte2.c: Likewise. + * testsuite/libffi.call/cls_align_double.c: Likewise. + * testsuite/libffi.call/cls_align_float.c: Likewise. + * testsuite/libffi.call/cls_align_longdouble.c: Likewise. + * testsuite/libffi.call/cls_align_longdouble_split.c: Likewise. + * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise. + * testsuite/libffi.call/cls_align_pointer.c: Likewise. + * testsuite/libffi.call/cls_align_sint16.c: Likewise. + * testsuite/libffi.call/cls_align_sint32.c: Likewise. + * testsuite/libffi.call/cls_align_sint64.c: Likewise. + * testsuite/libffi.call/cls_align_uint16.c: Likewise. + * testsuite/libffi.call/cls_align_uint32.c: Likewise. + * testsuite/libffi.call/cls_align_uint64.c: Likewise. + * testsuite/libffi.call/cls_dbls_struct.c: Likewise. + * testsuite/libffi.call/cls_pointer_stack.c: Likewise. + * testsuite/libffi.call/err_bad_typedef.c: Likewise. + * testsuite/libffi.call/huge_struct.c: Likewise. + * testsuite/libffi.call/nested_struct.c: Likewise. + * testsuite/libffi.call/nested_struct1.c: Likewise. + * testsuite/libffi.call/nested_struct10.c: Likewise. + * testsuite/libffi.call/nested_struct2.c: Likewise. + * testsuite/libffi.call/nested_struct3.c: Likewise. + * testsuite/libffi.call/nested_struct4.c: Likewise. + * testsuite/libffi.call/nested_struct5.c: Likewise. + * testsuite/libffi.call/nested_struct6.c: Likewise. + * testsuite/libffi.call/nested_struct7.c: Likewise. + * testsuite/libffi.call/nested_struct8.c: Likewise. + * testsuite/libffi.call/nested_struct9.c: Likewise. + * testsuite/libffi.call/stret_large.c: Likewise. + * testsuite/libffi.call/stret_large2.c: Likewise. + * testsuite/libffi.call/stret_medium.c: Likewise. + * testsuite/libffi.call/stret_medium2.c: Likewise. + * testsuite/libffi.call/struct1.c: Likewise. + * testsuite/libffi.call/struct1_win32.c: Likewise. + * testsuite/libffi.call/struct2.c: Likewise. + * testsuite/libffi.call/struct2_win32.c: Likewise. + * testsuite/libffi.call/struct3.c: Likewise. + * testsuite/libffi.call/struct4.c: Likewise. + * testsuite/libffi.call/struct5.c: Likewise. + * testsuite/libffi.call/struct6.c: Likewise. + * testsuite/libffi.call/struct7.c: Likewise. + * testsuite/libffi.call/struct8.c: Likewise. + * testsuite/libffi.call/struct9.c: Likewise. + * testsuite/libffi.call/testclosure.c: Likewise. + +2012-03-21 Peter Rosin + + * testsuite/libffi.call/float_va.c (float_va_fn): Use %f when + printing doubles (%lf is for long doubles). + (main): Likewise. + +2012-03-21 Peter Rosin + + * testsuite/lib/target-libpath.exp [*-*-cygwin*, *-*-mingw*] + (set_ld_library_path_env_vars): Add the library search dir to PATH + (and save PATH for later). + (restore_ld_library_path_env_vars): Restore PATH. + +2012-03-21 Peter Rosin + + * testsuite/lib/target-libpath.exp [*-*-cygwin*, *-*-mingw*] + (set_ld_library_path_env_vars): Add the library search dir to PATH + (and save PATH for later). + (restore_ld_library_path_env_vars): Restore PATH. + +2012-03-20 Peter Rosin + + * testsuite/libffi.call/strlen2_win32.c (main): Remove bug. + * src/x86/win32.S [MSVC] (ffi_closure_SYSV): Make the 'stub' label + visible outside the PROC, so that ffi_closure_THISCALL can see it. + +2012-03-20 Peter Rosin + + * testsuite/libffi.call/strlen2_win32.c (main): Remove bug. + * src/x86/win32.S [MSVC] (ffi_closure_SYSV): Make the 'stub' label + visible outside the PROC, so that ffi_closure_THISCALL can see it. + +2012-03-19 Alan Hourihane + + * src/m68k/ffi.c: Add MINT support. + * src/m68k/sysv.S: Ditto. + +2012-03-06 Chung-Lin Tang + + * src/arm/ffi.c (ffi_call): Add __ARM_EABI__ guard around call to + ffi_call_VFP(). + (ffi_prep_closure_loc): Add __ARM_EABI__ guard around use of + ffi_closure_VFP. + * src/arm/sysv.S: Add __ARM_EABI__ guard around VFP code. + +2012-03-19 chennam + + * src/powerpc/ffi_darwin.c (ffi_prep_closure_loc): Fix AIX closure + support. + +2012-03-13 Kaz Kojima + + * src/sh/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test, + just return FFI_BAD_ABI when things are wrong. + * src/sh64/ffi.c (ffi_prep_closure_loc): Ditto. + +2012-03-09 David Edelsohn + + * src/powerpc/aix_closure.S (ffi_closure_ASM): Adjust for Darwin64 + change to return value of ffi_closure_helper_DARWIN and load type + from return type. + +2012-03-03 H.J. Lu + + * src/x86/ffi64.c (ffi_call): Cast the return value to unsigned + long. + (ffi_prep_closure_loc): Cast to 64bit address in trampoline. + (ffi_closure_unix64_inner): Cast return pointer to unsigned long + first. + + * src/x86/ffitarget.h (FFI_SIZEOF_ARG): Defined to 8 for x32. + (ffi_arg): Set to unsigned long long for x32. + (ffi_sarg): Set to long long for x32. + +2012-03-03 H.J. Lu + + * src/prep_cif.c (ffi_prep_cif_core): Properly check bad ABI. + +2012-03-03 Andoni Morales Alastruey + + * configure.ac: Add -no-undefined for both 32- and 64-bit x86 + windows-like hosts. + * configure: Rebuilt. + +2012-02-27 Mikael Pettersson + + PR libffi/52223 + * Makefile.am (FLAGS_TO_PASS): Define. + * Makefile.in: Regenerate. + +2012-02-23 Anthony Green + + * src/*/ffitarget.h: Ensure that users never include ffitarget.h + directly. + +2012-02-23 Kai Tietz + + PR libffi/52221 + * src/x86/ffi.c (ffi_closure_raw_THISCALL): New + prototype. + (ffi_prep_raw_closure_loc): Use ffi_closure_raw_THISCALL for + thiscall-convention. + (ffi_raw_call): Use ffi_prep_args_raw. + * src/x86/win32.S (ffi_closure_raw_THISCALL): Add + implementation for stub. + +2012-02-10 Kai Tietz + + * configure.ac (AM_LTLDFLAGS): Add -no-undefine for x64 + windows target. + * configure: Regenerated. + +2012-02-08 Kai Tietz + + * src/prep_cif.c (ffi_prep_cif): Allow for X86_WIN32 + also FFI_THISCALL. + * src/x86/ffi.c (ffi_closure_THISCALL): Add prototype. + (FFI_INIT_TRAMPOLINE_THISCALL): New trampoline code. + (ffi_prep_closure_loc): Add FFI_THISCALL support. + * src/x86/ffitarget.h (FFI_TRAMPOLINE_SIZE): Adjust size. + * src/x86/win32.S (ffi_closure_THISCALL): New closure code + for thiscall-calling convention. + * testsuite/libffi.call/closure_thiscall.c: New test. + +2012-01-28 Kai Tietz + + * src/libffi/src/x86/ffi.c (ffi_call_win32): Add new + argument to prototype for specify calling-convention. + (ffi_call): Add support for stdcall/thiscall convention. + (ffi_prep_args): Likewise. + (ffi_raw_call): Likewise. + * src/x86/ffitarget.h (ffi_abi): Add FFI_THISCALL and + FFI_FASTCALL. + * src/x86/win32.S (_ffi_call_win32): Add support for + fastcall/thiscall calling-convention calls. + * testsuite/libffi.call/fastthis1_win32.c: New test. + * testsuite/libffi.call/fastthis2_win32.c: New test. + * testsuite/libffi.call/fastthis3_win32.c: New test. + * testsuite/libffi.call/strlen2_win32.c: New test. + * testsuite/libffi.call/many2_win32.c: New test. + * testsuite/libffi.call/struct1_win32.c: New test. + * testsuite/libffi.call/struct2_win32.c: New test. + +2012-01-23 Uros Bizjak + + * src/alpha/ffi.c (ffi_prep_closure_loc): Check for bad ABI. + +2012-01-23 Anthony Green + Chris Young + + * configure.ac: Add Amiga support. + * configure: Rebuilt. + +2012-01-23 Dmitry Nadezhin + + * include/ffi_common.h (LIKELY, UNLIKELY): Fix definitions. + +2012-01-23 Andreas Schwab + + * src/m68k/sysv.S (ffi_call_SYSV): Properly test for plain + mc68000. Test for __HAVE_68881__ in addition to __MC68881__. + +2012-01-19 Jakub Jelinek + + PR rtl-optimization/48496 + * src/ia64/ffi.c (ffi_call): Fix up aliasing violations. + +2012-01-09 Rainer Orth + + * configure.ac (i?86-*-*): Set TARGET to X86_64. + * configure: Regenerate. + +2011-12-07 Andrew Pinski + + PR libffi/50051 + * src/mips/n32.S: Add ".set mips4". + +2011-11-21 Andreas Tobler + + * configure: Regenerate. + +2011-11-12 David Gilbert + + * doc/libffi.texi, include/ffi.h.in, include/ffi_common.h, + man/Makefile.am, man/ffi.3, man/ffi_prep_cif.3, + man/ffi_prep_cif_var.3, src/arm/ffi.c, src/arm/ffitarget.h, + src/cris/ffi.c, src/prep_cif.c, + testsuite/libffi.call/cls_double_va.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/float_va.c: Many changes to support variadic + function calls. + +2011-11-12 Kyle Moffett + + * src/powerpc/ffi.c, src/powerpc/ffitarget.h, + src/powerpc/ppc_closure.S, src/powerpc/sysv.S: Many changes for + softfloat powerpc variants. + +2011-11-12 Petr Salinger + + * configure.ac (FFI_EXEC_TRAMPOLINE_TABLE): Fix kfreebsd support. + * configure: Rebuilt. + +2011-11-12 Timothy Wall + + * src/arm/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV): Max + alignment of 4 for wince on ARM. + +2011-11-12 Kyle Moffett + Anthony Green + + * src/ppc/sysv.S, src/ppc/ffi.c: Remove use of ppc string + instructions (not available on some cores, like the PPC440). + +2011-11-12 Kimura Wataru + + * m4/ax_enable_builddir: Change from string comparison to numeric + comparison for wc output. + * configure.ac: Enable FFI_MMAP_EXEC_WRIT for darwin11 aka Mac OS + X 10.7. + * configure: Rebuilt. + +2011-11-12 Anthony Green + + * Makefile.am (AM_CCASFLAGS): Add -g option to build assembly + files with debug info. + * Makefile.in: Rebuilt. + +2011-11-12 Jasper Lievisse Adriaanse + + * README: Update list of supported OpenBSD systems. + +2011-11-12 Anthony Green + + * libtool-version: Update. + * Makefile.am (nodist_libffi_la_SOURCES): Add src/debug.c if + FFI_DEBUG. + (libffi_la_SOURCES): Remove src/debug.c + (EXTRA_DIST): Add src/debug.c + * Makefile.in: Rebuilt. + * README: Update for 3.0.11. + +2011-11-10 Richard Henderson + + * configure.ac (GCC_AS_CFI_PSEUDO_OP): Use it instead of inline check. + * configure, aclocal.m4: Rebuild. + +2011-09-04 Iain Sandoe + + PR libffi/49594 + * src/powerpc/darwin_closure.S (stubs): Make the stub binding + helper reference track the architecture pointer size. + +2011-08-25 Andrew Haley + + * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Remove hard-coded assembly + instructions. + * src/arm/sysv.S (ffi_arm_trampoline): Put them here instead. + +2011-07-11 Andrew Haley + + * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Clear icache. + +2011-06-29 Rainer Orth + + * testsuite/libffi.call/cls_double_va.c: Move PR number to comment. + * testsuite/libffi.call/cls_longdouble_va.c: Likewise. + +2011-06-29 Rainer Orth + + PR libffi/46660 + * testsuite/libffi.call/cls_double_va.c: xfail dg-output on + mips-sgi-irix6*. + * testsuite/libffi.call/cls_longdouble_va.c: Likewise. + +2011-06-14 Rainer Orth + + * testsuite/libffi.call/huge_struct.c (test_large_fn): Use PRIu8, + PRId8 instead of %hhu, %hhd. + * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRId8, + PRIu8): Define. + [__sgi__] (PRId8, PRIu8): Define. + +2011-04-29 Rainer Orth + + * src/alpha/osf.S (UA_SI, FDE_ENCODING, FDE_ENCODE, FDE_ARANGE): + Define. + Use them to handle ELF vs. ECOFF differences. + [__osf__] (_GLOBAL__F_ffi_call_osf): Define. + +2011-03-30 Timothy Wall + + * src/powerpc/darwin.S: Fix unknown FDE encoding. + * src/powerpc/darwin_closure.S: ditto. + +2011-02-25 Anthony Green + + * src/powerpc/ffi.c (ffi_prep_closure_loc): Allow for more + 32-bit ABIs. + +2011-02-15 Anthony Green + + * m4/ax_cc_maxopt.m4: Don't -malign-double or use -ffast-math. + * configure: Rebuilt. + +2011-02-13 Ralf Wildenhues + + * configure: Regenerate. + +2011-02-13 Anthony Green + + * include/ffi_common.h (UNLIKELY, LIKELY): Define. + * src/x86/ffi64.c (UNLIKELY, LIKELY): Remove definition. + * src/prep_cif.c (UNLIKELY, LIKELY): Remove definition. + + * src/prep_cif.c (initialize_aggregate): Convert assertion into + FFI_BAD_TYPEDEF return. Initialize arg size and alignment to 0. + + * src/pa/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test, + just return FFI_BAD_ABI when things are wrong. + * src/arm/ffi.c (ffi_prep_closure_loc): Ditto. + * src/powerpc/ffi.c (ffi_prep_closure_loc): Ditto. + * src/mips/ffi.c (ffi_prep_closure_loc): Ditto. + * src/ia64/ffi.c (ffi_prep_closure_loc): Ditto. + * src/avr32/ffi.c (ffi_prep_closure_loc): Ditto. + +2011-02-11 Anthony Green + + * src/sparc/ffi.c (ffi_prep_closure_loc): Don't ASSERT ABI test, + just return FFI_BAD_ABI when things are wrong. + +2012-02-11 Eric Botcazou + + * src/sparc/v9.S (STACKFRAME): Bump to 176. + +2011-02-09 Stuart Shelton + + http://bugs.gentoo.org/show_bug.cgi?id=286911 + * src/mips/ffitarget.h: Clean up error messages. + * src/java_raw_api.c (ffi_java_translate_args): Cast raw arg to + ffi_raw*. + * include/ffi.h.in: Add pragma for SGI compiler. + +2011-02-09 Anthony Green + + * configure.ac: Add powerpc64-*-darwin* support. + +2011-02-09 Anthony Green + + * README: Mention Interix. + +2011-02-09 Jonathan Callen + + * configure.ac: Add Interix to win32/cygwin/mingw case. + * configure: Ditto. + * src/closures.c: Treat Interix like Cygwin, instead of as a + generic win32. + +2011-02-09 Anthony Green + + * testsuite/libffi.call/err_bad_typedef.c: Remove xfail. + * testsuite/libffi.call/err_bad_abi.c: Remove xfail. + * src/x86/ffi64.c (UNLIKELY, LIKELY): Define. + (ffi_prep_closure_loc): Check for bad ABI. + * src/prep_cif.c (UNLIKELY, LIKELY): Define. + (initialize_aggregate): Check for bad types. + +2011-02-09 Landon Fuller + + * Makefile.am (EXTRA_DIST): Add build-ios.sh, src/arm/gentramp.sh, + src/arm/trampoline.S. + (nodist_libffi_la_SOURCES): Add src/arc/trampoline.S. + * configure.ac (FFI_EXEC_TRAMPOLINE_TABLE): Define. + * src/arm/ffi.c (ffi_trampoline_table) + (ffi_closure_trampoline_table_page, ffi_trampoline_table_entry) + (FFI_TRAMPOLINE_CODELOC_CONFIG, FFI_TRAMPOLINE_CONFIG_PAGE_OFFSET) + (FFI_TRAMPOLINE_COUNT, ffi_trampoline_lock, ffi_trampoline_tables) + (ffi_trampoline_table_alloc, ffi_closure_alloc, ffi_closure_free): + Define for FFI_EXEC_TRAMPOLINE_TABLE case (iOS). + (ffi_prep_closure_loc): Handl FFI_EXEC_TRAMPOLINE_TABLE case + separately. + * src/arm/sysv.S: Handle Apple iOS host. + * src/closures.c: Handle FFI_EXEC_TRAMPOLINE_TABLE case. + * build-ios.sh: New file. + * fficonfig.h.in, configure, Makefile.in: Rebuilt. + * README: Mention ARM iOS. + +2011-02-08 Oren Held + + * src/dlmalloc.c (_STRUCT_MALLINFO): Define in order to avoid + redefinition of mallinfo on HP-UX. + +2011-02-08 Ginn Chen + + * src/sparc/ffi.c (ffi_call): Make compatible with Solaris Studio + aggregate return ABI. Flush cache. + (ffi_prep_closure_loc): Flush cache. + +2011-02-11 Anthony Green + + From Tom Honermann : + * src/powerpc/aix.S (ffi_call_AIX): Support for xlc toolchain on + AIX. Declare .ffi_prep_args. Insert nops after branch + instructions so that the AIX linker can insert TOC reload + instructions. + * src/powerpc/aix_closure.S: Declare .ffi_closure_helper_DARWIN. + +2011-02-08 Ed + + * src/powerpc/asm.h: Fix grammar nit in comment. + +2011-02-08 Uli Link + + * include/ffi.h.in (FFI_64_BIT_MAX): Define and use. + +2011-02-09 Rainer Orth + + PR libffi/46661 + * testsuite/libffi.call/cls_pointer.c (main): Cast void * to + uintptr_t first. + * testsuite/libffi.call/cls_pointer_stack.c (main): Likewise. + +2011-02-08 Rafael Avila de Espindola + + * configure.ac: Fix x86 test for pc related relocs. + * configure: Rebuilt. + +2011-02-07 Joel Sherrill + + * libffi/src/m68k/ffi.c: Add RTEMS support for cache flushing. + Handle case when CPU variant does not have long double support. + * libffi/src/m68k/sysv.S: Add support for mc68000, Coldfire, + and cores with soft floating point. + +2011-02-07 Joel Sherrill + + * configure.ac: Add mips*-*-rtems* support. + * configure: Regenerate. + * src/mips/ffitarget.h: Ensure needed constants are available + for targets which do not have sgidefs.h. + +2011-01-26 Dave Korn + + PR target/40125 + * configure.ac (AM_LTLDFLAGS): Add -bindir option for windows DLLs. + * configure: Regenerate. + +2010-12-18 Iain Sandoe + + PR libffi/29152 + PR libffi/42378 + * src/powerpc/darwin_closure.S: Provide Darwin64 implementation, + update comments. + * src/powerpc/ffitarget.h (POWERPC_DARWIN64): New, + (FFI_TRAMPOLINE_SIZE): Update for Darwin64. + * src/powerpc/darwin.S: Provide Darwin64 implementation, + update comments. + * src/powerpc/ffi_darwin.c: Likewise. + +2010-12-06 Rainer Orth + + * configure.ac (libffi_cv_as_ascii_pseudo_op): Use double + backslashes. + (libffi_cv_as_string_pseudo_op): Likewise. + * configure: Regenerate. + +2010-12-03 Chung-Lin Tang + + * src/arm/sysv.S (ffi_closure_SYSV): Add UNWIND to .pad directive. + (ffi_closure_VFP): Same. + (ffi_call_VFP): Move down to before ffi_closure_VFP. Add '.fpu vfp' + directive. + +2010-12-01 Rainer Orth + + * testsuite/libffi.call/ffitest.h [__sgi] (PRId64, PRIu64): Define. + (PRIuPTR): Define. + +2010-11-29 Richard Henderson + Rainer Orth + + * src/x86/sysv.S (FDE_ENCODING, FDE_ENCODE): Define. + (.eh_frame): Use FDE_ENCODING. + (.LASFDE1, .LASFDE2, LASFDE3): Simplify with FDE_ENCODE. + +2010-11-22 Jacek Caban + + * configure.ac: Check for symbol underscores on mingw-w64. + * configure: Rebuilt. + * src/x86/win64.S: Correctly access extern symbols in respect to + underscores. + +2010-11-15 Rainer Orth + + * testsuite/lib/libffi-dg.exp: Rename ... + * testsuite/lib/libffi.exp: ... to this. + * libffi/testsuite/libffi.call/call.exp: Don't load libffi-dg.exp. + * libffi/testsuite/libffi.special/special.exp: Likewise. + +2010-10-28 Chung-Lin Tang + + * src/arm/ffi.c (ffi_prep_args): Add VFP register argument handling + code, new parameter, and return value. Update comments. + (ffi_prep_cif_machdep): Add case for VFP struct return values. Add + call to layout_vfp_args(). + (ffi_call_SYSV): Update declaration. + (ffi_call_VFP): New declaration. + (ffi_call): Add VFP struct return conditions. Call ffi_call_VFP() + when ABI is FFI_VFP. + (ffi_closure_VFP): New declaration. + (ffi_closure_SYSV_inner): Add new vfp_args parameter, update call to + ffi_prep_incoming_args_SYSV(). + (ffi_prep_incoming_args_SYSV): Update parameters. Add VFP argument + case handling. + (ffi_prep_closure_loc): Pass ffi_closure_VFP to trampoline + construction under VFP hard-float. + (rec_vfp_type_p): New function. + (vfp_type_p): Same. + (place_vfp_arg): Same. + (layout_vfp_args): Same. + * src/arm/ffitarget.h (ffi_abi): Add FFI_VFP. Define FFI_DEFAULT_ABI + based on __ARM_PCS_VFP. + (FFI_EXTRA_CIF_FIELDS): Define for adding VFP hard-float specific + fields. + (FFI_TYPE_STRUCT_VFP_FLOAT): Define internally used type code. + (FFI_TYPE_STRUCT_VFP_DOUBLE): Same. + * src/arm/sysv.S (ffi_call_SYSV): Change call of ffi_prep_args() to + direct call. Move function pointer load upwards. + (ffi_call_VFP): New function. + (ffi_closure_VFP): Same. + + * testsuite/lib/libffi-dg.exp (check-flags): New function. + (dg-skip-if): New function. + * testsuite/libffi.call/cls_double_va.c: Skip if target is arm*-*-* + and compiler options include -mfloat-abi=hard. + * testsuite/libffi.call/cls_longdouble_va.c: Same. + +2010-10-01 Jakub Jelinek + + PR libffi/45677 + * src/x86/ffi64.c (ffi_prep_cif_machdep): Ensure cif->bytes is + a multiple of 8. + * testsuite/libffi.call/many2.c: New test. + +2010-08-20 Mark Wielaard + + * src/closures.c (open_temp_exec_file_mnt): Check if getmntent_r + returns NULL. + +2010-08-09 Andreas Tobler + + * configure.ac: Add target powerpc64-*-freebsd*. + * configure: Regenerate. + * testsuite/libffi.call/cls_align_longdouble_split.c: Pass + -mlong-double-128 only to linux targets. + * testsuite/libffi.call/cls_align_longdouble_split2.c: Likewise. + * testsuite/libffi.call/cls_longdouble.c: Likewise. + * testsuite/libffi.call/huge_struct.c: Likewise. + +2010-08-05 Dan Witte + + * Makefile.am: Pass FFI_DEBUG define to msvcc.sh for linking to the + debug CRT when --enable-debug is given. + * configure.ac: Define it. + * msvcc.sh: Translate -g and -DFFI_DEBUG appropriately. + +2010-08-04 Dan Witte + + * src/x86/ffitarget.h: Add X86_ANY define for all x86/x86_64 + platforms. + * src/x86/ffi.c: Remove redundant ifdef checks. + * src/prep_cif.c: Push stack space computation into src/x86/ffi.c + for X86_ANY so return value space doesn't get added twice. + +2010-08-03 Neil Rashbrooke + + * msvcc.sh: Don't pass -safeseh to ml64 because behavior is buggy. + +2010-07-22 Dan Witte + + * src/*/ffitarget.h: Make FFI_LAST_ABI one past the last valid ABI. + * src/prep_cif.c: Fix ABI assertion. + * src/cris/ffi.c: Ditto. + +2010-07-10 Evan Phoenix + + * src/closures.c (selinux_enabled_check): Fix strncmp usage bug. + +2010-07-07 Dan Horák + + * include/ffi.h.in: Protect #define with #ifndef. + * src/powerpc/ffitarget.h: Ditto. + * src/s390/ffitarget.h: Ditto. + * src/sparc/ffitarget.h: Ditto. + +2010-07-07 Neil Roberts + + * src/x86/sysv.S (ffi_call_SYSV): Align the stack pointer to + 16-bytes. + +2010-07-02 Jakub Jelinek + + * Makefile.am (AM_MAKEFLAGS): Pass also mandir to submakes. + * Makefile.in: Regenerated. + +2010-05-19 Rainer Orth + + * configure.ac (libffi_cv_as_x86_pcrel): Check for illegal in as + output, too. + (libffi_cv_as_ascii_pseudo_op): Check for .ascii. + (libffi_cv_as_string_pseudo_op): Check for .string. + * configure: Regenerate. + * fficonfig.h.in: Regenerate. + * src/x86/sysv.S (.eh_frame): Use .ascii, .string or error. + +2010-05-11 Dan Witte + + * doc/libffi.tex: Document previous change. + +2010-05-11 Makoto Kato + + * src/x86/ffi.c (ffi_call): Don't copy structs passed by value. + +2010-05-05 Michael Kohler + + * src/dlmalloc.c (dlfree): Fix spelling. + * src/ia64/ffi.c (ffi_prep_cif_machdep): Ditto. + * configure.ac: Ditto. + * configure: Rebuilt. + +2010-04-13 Dan Witte + + * msvcc.sh: Build with -W3 instead of -Wall. + * src/powerpc/ffi_darwin.c: Remove build warnings. + * src/x86/ffi.c: Ditto. + * src/x86/ffitarget.h: Ditto. + +2010-04-12 Dan Witte + Walter Meinl + + * configure.ac: Add OS/2 support. + * configure: Rebuilt. + * src/closures.c: Ditto. + * src/dlmalloc.c: Ditto. + * src/x86/win32.S: Ditto. + +2010-04-07 Jakub Jelinek + + * testsuite/libffi.call/err_bad_abi.c: Remove unused args variable. + +2010-04-02 Ralf Wildenhues + + * Makefile.in: Regenerate. + * aclocal.m4: Regenerate. + * include/Makefile.in: Regenerate. + * man/Makefile.in: Regenerate. + * testsuite/Makefile.in: Regenerate. + +2010-03-30 Dan Witte + + * msvcc.sh: Disable build warnings. + * README (tested): Clarify windows build procedure. + +2010-03-15 Rainer Orth + + * configure.ac (libffi_cv_as_x86_64_unwind_section_type): New test. + * configure: Regenerate. + * fficonfig.h.in: Regenerate. + * libffi/src/x86/unix64.S (.eh_frame) + [HAVE_AS_X86_64_UNWIND_SECTION_TYPE]: Use @unwind section type. + +2010-03-14 Matthias Klose + + * src/x86/ffi64.c: Fix typo in comment. + * src/x86/ffi.c: Use /* ... */ comment style. + +2010-02-24 Rainer Orth + + * doc/libffi.texi (The Closure API): Fix typo. + * doc/libffi.info: Remove. + +2010-02-15 Matthias Klose + + * src/arm/sysv.S (__ARM_ARCH__): Define for processor + __ARM_ARCH_7EM__. + +2010-01-15 Anthony Green + + * README: Add notes on building with Microsoft Visual C++. + +2010-01-15 Daniel Witte + + * msvcc.sh: New file. + + * src/x86/win32.S: Port assembly routines to MSVC and #ifdef. + * src/x86/ffi.c: Tweak function declaration and remove excess + parens. + * include/ffi.h.in: Add __declspec(align(8)) to typedef struct + ffi_closure. + + * src/x86/ffi.c: Merge ffi_call_SYSV and ffi_call_STDCALL into new + function ffi_call_win32 on X86_WIN32. + * src/x86/win32.S (ffi_call_SYSV): Rename to ffi_call_win32. + (ffi_call_STDCALL): Remove. + + * src/prep_cif.c (ffi_prep_cif): Move stack space allocation code + to ffi_prep_cif_machdep for x86. + * src/x86/ffi.c (ffi_prep_cif_machdep): To here. + +2010-01-15 Oliver Kiddle + + * src/x86/ffitarget.h (ffi_abi): Check for __i386 and __amd64 for + Sun Studio compiler compatibility. + +2010-01-12 Conrad Irwin + + * doc/libffi.texi: Add closure example. + +2010-01-07 Rainer Orth + + PR libffi/40701 + * testsuite/libffi.call/ffitest.h [__alpha__ && __osf__] (PRIdLL, + PRIuLL, PRId64, PRIu64, PRIuPTR): Define. + * testsuite/libffi.call/cls_align_sint64.c: Add -Wno-format on + alpha*-dec-osf*. + * testsuite/libffi.call/cls_align_uint64.c: Likewise. + * testsuite/libffi.call/cls_ulonglong.c: Likewise. + * testsuite/libffi.call/return_ll1.c: Likewise. + * testsuite/libffi.call/stret_medium2.c: Likewise. + * testsuite/libffi.special/ffitestcxx.h (allocate_mmap): Cast + MAP_FAILED to char *. + +2010-01-06 Rainer Orth + + * src/mips/n32.S: Use .abicalls and .eh_frame with __GNUC__. + +2009-12-31 Anthony Green + + * README: Update for libffi 3.0.9. + +2009-12-27 Matthias Klose + + * configure.ac (HAVE_LONG_DOUBLE): Define for mips when + appropriate. + * configure: Rebuilt. + +2009-12-26 Anthony Green + + * testsuite/libffi.call/cls_longdouble_va.c: Mark as xfail for + avr32*-*-*. + * testsuite/libffi.call/cls_double_va.c: Ditto. + +2009-12-26 Andreas Tobler + + * testsuite/libffi.call/ffitest.h: Conditionally include stdint.h + and inttypes.h. + * testsuite/libffi.special/unwindtest.cc: Ditto. + +2009-12-26 Andreas Tobler + + * configure.ac: Add amd64-*-openbsd*. + * configure: Rebuilt. + * testsuite/lib/libffi-dg.exp (libffi_target_compile): Link + openbsd programs with -lpthread. + +2009-12-26 Anthony Green + + * testsuite/libffi.call/cls_double_va.c, + testsuite/libffi.call/cls_longdouble.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_pointer.c, + testsuite/libffi.call/cls_pointer_stack.c: Remove xfail for + mips*-*-* and arm*-*-*. + * testsuite/libffi.call/cls_align_longdouble_split.c, + testsuite/libffi.call/cls_align_longdouble_split2.c, + testsuite/libffi.call/stret_medium2.c, + testsuite/libffi.call/stret_medium.c, + testsuite/libffi.call/stret_large.c, + testsuite/libffi.call/stret_large2.c: Remove xfail for arm*-*-*. + +2009-12-31 Kay Tietz + + * testsuite/libffi.call/ffitest.h, + testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRuLL): Fix + definitions. + +2009-12-31 Carlo Bramini + + * configure.ac (AM_LTLDFLAGS): Define for windows hosts. + * Makefile.am (libffi_la_LDFLAGS): Add AM_LTLDFLAGS. + * configure: Rebuilt. + * Makefile.in: Rebuilt. + +2009-12-31 Anthony Green + Blake Chaffin. + + * testsuite/libffi.call/huge_struct.c: New test case from Blake + Chaffin @ Apple. + +2009-12-28 David Edelsohn + + * src/powerpc/ffi_darwin.c (ffi_prep_args): Copy abi and nargs to + local variables. + (aix_adjust_aggregate_sizes): New function. + (ffi_prep_cif_machdep): Call it. + +2009-12-26 Andreas Tobler + + * configure.ac: Define FFI_MMAP_EXEC_WRIT for the given targets. + * configure: Regenerate. + * fficonfig.h.in: Likewise. + * src/closures.c: Remove the FFI_MMAP_EXEC_WRIT definition for + Solaris/x86. + +2009-12-26 Andreas Schwab + + * src/powerpc/ffi.c (ffi_prep_args_SYSV): Advance intarg_count + when a float arguments is passed in memory. + (ffi_closure_helper_SYSV): Mark general registers as used up when + a 64bit or soft-float long double argument is passed in memory. + +2009-12-25 Matthias Klose + + * man/ffi_call.3: Fix #include in examples. + * doc/libffi.texi: Add dircategory. + +2009-12-25 Frank Everdij + + * include/ffi.h.in: Placed '__GNUC__' ifdef around + '__attribute__((aligned(8)))' in ffi_closure, fixes compile for + IRIX MIPSPro c99. + * include/ffi_common.h: Added '__sgi' define to non + '__attribute__((__mode__()))' integer typedefs. + * src/mips/ffi.c (ffi_call, ffi_closure_mips_inner_O32, + ffi_closure_mips_inner_N32): Added 'defined(_MIPSEB)' to BE check. + (ffi_closure_mips_inner_O32, ffi_closure_mips_inner_N32): Added + FFI_LONGDOUBLE support and alignment(N32 only). + * src/mips/ffitarget.h: Corrected '#include ' for IRIX and + fixed non '__attribute__((__mode__()))' integer typedefs. + * src/mips/n32.S: Put '#ifdef linux' around '.abicalls' and '.eh_frame' + since they are Linux/GNU Assembler specific. + +2009-12-25 Bradley Smith + + * configure.ac, Makefile.am, src/avr32/ffi.c, + src/avr32/ffitarget.h, + src/avr32/sysv.S: Add AVR32 port. + * configure, Makefile.in: Rebuilt. + +2009-12-21 Andreas Tobler + + * configure.ac: Make i?86 build on FreeBSD and OpenBSD. + * configure: Regenerate. + +2009-12-15 John David Anglin + + * testsuite/libffi.call/ffitest.h: Define PRIuPTR on PA HP-UX. + +2009-12-13 John David Anglin + + * src/pa/ffi.c (ffi_closure_inner_pa32): Handle FFI_TYPE_LONGDOUBLE + type on HP-UX. + +2012-02-13 Kai Tietz + + PR libffi/52221 + * src/x86/ffi.c (ffi_prep_raw_closure_loc): Add thiscall + support for X86_WIN32. + (FFI_INIT_TRAMPOLINE_THISCALL): Fix displacement. + +2009-12-11 Eric Botcazou + + * src/sparc/ffi.c (ffi_closure_sparc_inner_v9): Properly align 'long + double' arguments. + +2009-12-11 Eric Botcazou + + * testsuite/libffi.call/ffitest.h: Define PRIuPTR on Solaris < 10. + +2009-12-10 Rainer Orth + + PR libffi/40700 + * src/closures.c [X86_64 && __sun__ && __svr4__] + (FFI_MMAP_EXEC_WRIT): Define. + +2009-12-08 David Daney + + * testsuite/libffi.call/stret_medium.c: Remove xfail for mips*-*-* + * testsuite/libffi.call/cls_align_longdouble_split2.c: Same. + * testsuite/libffi.call/stret_large.c: Same. + * testsuite/libffi.call/cls_align_longdouble_split.c: Same. + * testsuite/libffi.call/stret_large2.c: Same. + * testsuite/libffi.call/stret_medium2.c: Same. + +2009-12-07 David Edelsohn + + * src/powerpc/aix_closure.S (libffi_closure_ASM): Fix tablejump + typo. + +2009-12-05 David Edelsohn + + * src/powerpc/aix.S: Update AIX32 code to be consistent with AIX64 + code. + * src/powerpc/aix_closure.S: Same. + +2009-12-05 Ralf Wildenhues + + * Makefile.in: Regenerate. + * configure: Regenerate. + * include/Makefile.in: Regenerate. + * man/Makefile.in: Regenerate. + * testsuite/Makefile.in: Regenerate. + +2009-12-04 David Edelsohn + + * src/powerpc/aix_closure.S: Reorganize 64-bit code to match + linux64_closure.S. + +2009-12-04 Uros Bizjak + + PR libffi/41908 + * src/x86/ffi64.c (classify_argument): Update from + gcc/config/i386/i386.c. + (ffi_closure_unix64_inner): Do not use the address of two consecutive + SSE registers directly. + * testsuite/libffi.call/cls_dbls_struct.c (main): Remove xfail + for x86_64 linux targets. + +2009-12-04 David Edelsohn + + * src/powerpc/ffi_darwin.c (ffi_closure_helper_DARWIN): Increment + pfr for long double split between fpr13 and stack. + +2009-12-03 David Edelsohn + + * src/powerpc/ffi_darwin.c (ffi_prep_args): Increment next_arg and + fparg_count twice for long double. + +2009-12-03 David Edelsohn + + PR libffi/42243 + * src/powerpc/ffi_darwin.c (ffi_prep_args): Remove extra parentheses. + +2009-12-03 Uros Bizjak + + * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string. + Remove xfails for x86 linux targets. + +2009-12-02 David Edelsohn + + * src/powerpc/ffi_darwin.c (ffi_prep_args): Fix typo in INT64 + case. + +2009-12-01 David Edelsohn + + * src/powerpc/aix.S (ffi_call_AIX): Convert to more standard + register usage. Call ffi_prep_args directly. Add long double + return value support. + * src/powerpc/ffi_darwin.c (ffi_prep_args): Double arg increment + applies to FFI_TYPE_DOUBLE. Correct fpr_base increment typo. + Separate FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases. + (ffi_prep_cif_machdep): Only 16 byte stack alignment in 64 bit + mode. + (ffi_closure_helper_DARWIN): Remove nf and ng counters. Move temp + into case. + * src/powerpc/aix_closure.S: Maintain 16 byte stack alignment. + Allocate result area between params and FPRs. + +2009-11-30 David Edelsohn + + PR target/35484 + * src/powerpc/ffitarget.h (POWERPC64): Define for PPC64 Linux and + AIX64. + * src/powerpc/aix.S: Implement AIX64 version. + * src/powerpc/aix_closure.S: Implement AIX64 version. + (ffi_closure_ASM): Use extsb, lha and displament addresses. + * src/powerpc/ffi_darwin.c (ffi_prep_args): Implement AIX64 + support. + (ffi_prep_cif_machdep): Same. + (ffi_call): Same. + (ffi_closure_helper_DARWIN): Same. + +2009-11-02 Andreas Tobler + + PR libffi/41908 + * testsuite/libffi.call/testclosure.c: New test. + +2009-09-28 Kai Tietz + + * src/x86/win64.S (_ffi_call_win64 stack): Remove for gnu + assembly version use of ___chkstk. + +2009-09-23 Matthias Klose + + PR libffi/40242, PR libffi/41443 + * src/arm/sysv.S (__ARM_ARCH__): Define for processors + __ARM_ARCH_6T2__, __ARM_ARCH_6M__, __ARM_ARCH_7__, + __ARM_ARCH_7A__, __ARM_ARCH_7R__, __ARM_ARCH_7M__. + Change the conditionals to __SOFTFP__ || __ARM_EABI__ + for -mfloat-abi=softfp to work. + +2009-09-17 Loren J. Rittle + + PR testsuite/32843 (strikes again) + * src/x86/ffi.c (ffi_prep_cif_machdep): Add X86_FREEBSD to + enable proper extension on char and short. + +2009-09-15 David Daney + + * src/java_raw_api.c (ffi_java_raw_to_rvalue): Remove special + handling for FFI_TYPE_POINTER. + * src/mips/ffitarget.h (FFI_TYPE_STRUCT_D_SOFT, + FFI_TYPE_STRUCT_F_SOFT, FFI_TYPE_STRUCT_DD_SOFT, + FFI_TYPE_STRUCT_FF_SOFT, FFI_TYPE_STRUCT_FD_SOFT, + FFI_TYPE_STRUCT_DF_SOFT, FFI_TYPE_STRUCT_SOFT): New defines. + (FFI_N32_SOFT_FLOAT, FFI_N64_SOFT_FLOAT): New ffi_abi enumerations. + (enum ffi_abi): Set FFI_DEFAULT_ABI for soft-float. + * src/mips/n32.S (ffi_call_N32): Add handling for soft-float + structure and pointer returns. + (ffi_closure_N32): Add handling for pointer returns. + * src/mips/ffi.c (ffi_prep_args, calc_n32_struct_flags, + calc_n32_return_struct_flags): Handle soft-float. + (ffi_prep_cif_machdep): Handle soft-float, fix pointer handling. + (ffi_call_N32): Declare proper argument types. + (ffi_call, copy_struct_N32, ffi_closure_mips_inner_N32): Handle + soft-float. + +2009-08-24 Ralf Wildenhues + + * configure.ac (AC_PREREQ): Bump to 2.64. + +2009-08-22 Ralf Wildenhues + + * Makefile.am (install-html, install-pdf): Remove. + * Makefile.in: Regenerate. + + * Makefile.in: Regenerate. + * aclocal.m4: Regenerate. + * configure: Regenerate. + * fficonfig.h.in: Regenerate. + * include/Makefile.in: Regenerate. + * man/Makefile.in: Regenerate. + * testsuite/Makefile.in: Regenerate. + +2011-08-22 Jasper Lievisse Adriaanse + + * configure.ac: Add OpenBSD/hppa and OpenBSD/powerpc support. + * configure: Rebuilt. + +2009-07-30 Ralf Wildenhues + + * configure.ac (_AC_ARG_VAR_PRECIOUS): Use m4_rename_force. + +2009-07-24 Dave Korn + + PR libffi/40807 + * src/x86/ffi.c (ffi_prep_cif_machdep): Also use sign/zero-extending + return types for X86_WIN32. + * src/x86/win32.S (_ffi_call_SYSV): Handle omitted return types. + (_ffi_call_STDCALL, _ffi_closure_SYSV, _ffi_closure_raw_SYSV, + _ffi_closure_STDCALL): Likewise. + + * src/closures.c (is_selinux_enabled): Define to const 0 for Cygwin. + (dlmmap, dlmunmap): Also use these functions on Cygwin. + +2009-07-11 Richard Sandiford + + PR testsuite/40699 + PR testsuite/40707 + PR testsuite/40709 + * testsuite/lib/libffi-dg.exp: Revert 2009-07-02, 2009-07-01 and + 2009-06-30 commits. + +2009-07-01 Richard Sandiford + + * testsuite/lib/libffi-dg.exp (libffi-init): Set ld_library_path + to "" before adding paths. (This reinstates an assignment that + was removed by my 2009-06-30 commit, but changes the initial + value from "." to "".) + +2009-07-01 H.J. Lu + + PR testsuite/40601 + * testsuite/lib/libffi-dg.exp (libffi-init): Properly set + gccdir. Adjust ld_library_path for gcc only if gccdir isn't + empty. + +2009-06-30 Richard Sandiford + + * testsuite/lib/libffi-dg.exp (libffi-init): Don't add "." + to ld_library_path. Use add_path. Add just find_libgcc_s + to ld_library_path, not every libgcc multilib directory. + +2009-06-16 Wim Lewis + + * src/powerpc/ffi.c: Avoid clobbering cr3 and cr4, which are + supposed to be callee-saved. + * src/powerpc/sysv.S (small_struct_return_value): Fix overrun of + return buffer for odd-size structs. + +2009-06-16 Andreas Tobler + + PR libffi/40444 + * testsuite/lib/libffi-dg.exp (libffi_target_compile): Add + allow_stack_execute for Darwin. + +2009-06-16 Andrew Haley + + * configure.ac (TARGETDIR): Add missing blank lines. + * configure: Regenerate. + +2009-06-16 Andrew Haley + + * testsuite/libffi.call/cls_align_sint64.c, + testsuite/libffi.call/cls_align_uint64.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_ulonglong.c, + testsuite/libffi.call/return_ll1.c, + testsuite/libffi.call/stret_medium2.c: Fix printf format + specifiers. + * testsuite/libffi.call/ffitest.h, + testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define. + +2009-06-15 Andrew Haley + + * testsuite/libffi.call/err_bad_typedef.c: xfail everywhere. + * testsuite/libffi.call/err_bad_abi.c: Likewise. + +2009-06-12 Andrew Haley + + * Makefile.am: Remove info_TEXINFOS. + +2009-06-12 Andrew Haley + + * ChangeLog.libffi: testsuite/libffi.call/cls_align_sint64.c, + testsuite/libffi.call/cls_align_uint64.c, + testsuite/libffi.call/cls_ulonglong.c, + testsuite/libffi.call/return_ll1.c, + testsuite/libffi.call/stret_medium2.c: Fix printf format + specifiers. + testsuite/libffi.special/unwindtest.cc: include stdint.h. + +2009-06-11 Timothy Wall + + * Makefile.am, + configure.ac, + include/ffi.h.in, + include/ffi_common.h, + src/closures.c, + src/dlmalloc.c, + src/x86/ffi.c, + src/x86/ffitarget.h, + src/x86/win64.S (new), + README: Added win64 support (mingw or MSVC) + * Makefile.in, + include/Makefile.in, + man/Makefile.in, + testsuite/Makefile.in, + configure, + aclocal.m4: Regenerated + * ltcf-c.sh: properly escape cygwin/w32 path + * man/ffi_call.3: Clarify size requirements for return value. + * src/x86/ffi64.c: Fix filename in comment. + * src/x86/win32.S: Remove unused extern. + + * testsuite/libffi.call/closure_fn0.c, + testsuite/libffi.call/closure_fn1.c, + testsuite/libffi.call/closure_fn2.c, + testsuite/libffi.call/closure_fn3.c, + testsuite/libffi.call/closure_fn4.c, + testsuite/libffi.call/closure_fn5.c, + testsuite/libffi.call/closure_fn6.c, + testsuite/libffi.call/closure_stdcall.c, + testsuite/libffi.call/cls_12byte.c, + testsuite/libffi.call/cls_16byte.c, + testsuite/libffi.call/cls_18byte.c, + testsuite/libffi.call/cls_19byte.c, + testsuite/libffi.call/cls_1_1byte.c, + testsuite/libffi.call/cls_20byte.c, + testsuite/libffi.call/cls_20byte1.c, + testsuite/libffi.call/cls_24byte.c, + testsuite/libffi.call/cls_2byte.c, + testsuite/libffi.call/cls_3_1byte.c, + testsuite/libffi.call/cls_3byte1.c, + testsuite/libffi.call/cls_3byte2.c, + testsuite/libffi.call/cls_4_1byte.c, + testsuite/libffi.call/cls_4byte.c, + testsuite/libffi.call/cls_5_1_byte.c, + testsuite/libffi.call/cls_5byte.c, + testsuite/libffi.call/cls_64byte.c, + testsuite/libffi.call/cls_6_1_byte.c, + testsuite/libffi.call/cls_6byte.c, + testsuite/libffi.call/cls_7_1_byte.c, + testsuite/libffi.call/cls_7byte.c, + testsuite/libffi.call/cls_8byte.c, + testsuite/libffi.call/cls_9byte1.c, + testsuite/libffi.call/cls_9byte2.c, + testsuite/libffi.call/cls_align_double.c, + testsuite/libffi.call/cls_align_float.c, + testsuite/libffi.call/cls_align_longdouble.c, + testsuite/libffi.call/cls_align_longdouble_split.c, + testsuite/libffi.call/cls_align_longdouble_split2.c, + testsuite/libffi.call/cls_align_pointer.c, + testsuite/libffi.call/cls_align_sint16.c, + testsuite/libffi.call/cls_align_sint32.c, + testsuite/libffi.call/cls_align_sint64.c, + testsuite/libffi.call/cls_align_uint16.c, + testsuite/libffi.call/cls_align_uint32.c, + testsuite/libffi.call/cls_align_uint64.c, + testsuite/libffi.call/cls_dbls_struct.c, + testsuite/libffi.call/cls_double.c, + testsuite/libffi.call/cls_double_va.c, + testsuite/libffi.call/cls_float.c, + testsuite/libffi.call/cls_longdouble.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_multi_schar.c, + testsuite/libffi.call/cls_multi_sshort.c, + testsuite/libffi.call/cls_multi_sshortchar.c, + testsuite/libffi.call/cls_multi_uchar.c, + testsuite/libffi.call/cls_multi_ushort.c, + testsuite/libffi.call/cls_multi_ushortchar.c, + testsuite/libffi.call/cls_pointer.c, + testsuite/libffi.call/cls_pointer_stack.c, + testsuite/libffi.call/cls_schar.c, + testsuite/libffi.call/cls_sint.c, + testsuite/libffi.call/cls_sshort.c, + testsuite/libffi.call/cls_uchar.c, + testsuite/libffi.call/cls_uint.c, + testsuite/libffi.call/cls_ulonglong.c, + testsuite/libffi.call/cls_ushort.c, + testsuite/libffi.call/err_bad_abi.c, + testsuite/libffi.call/err_bad_typedef.c, + testsuite/libffi.call/float2.c, + testsuite/libffi.call/huge_struct.c, + testsuite/libffi.call/nested_struct.c, + testsuite/libffi.call/nested_struct1.c, + testsuite/libffi.call/nested_struct10.c, + testsuite/libffi.call/nested_struct2.c, + testsuite/libffi.call/nested_struct3.c, + testsuite/libffi.call/nested_struct4.c, + testsuite/libffi.call/nested_struct5.c, + testsuite/libffi.call/nested_struct6.c, + testsuite/libffi.call/nested_struct7.c, + testsuite/libffi.call/nested_struct8.c, + testsuite/libffi.call/nested_struct9.c, + testsuite/libffi.call/problem1.c, + testsuite/libffi.call/return_ldl.c, + testsuite/libffi.call/return_ll1.c, + testsuite/libffi.call/stret_large.c, + testsuite/libffi.call/stret_large2.c, + testsuite/libffi.call/stret_medium.c, + testsuite/libffi.call/stret_medium2.c, + testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead + of checking for MMAP. Use intptr_t instead of long casts. + +2009-06-11 Kaz Kojima + + * testsuite/libffi.call/cls_longdouble_va.c: Add xfail sh*-*-linux-*. + * testsuite/libffi.call/err_bad_abi.c: Add xfail sh*-*-*. + * testsuite/libffi.call/err_bad_typedef.c: Likewise. + +2009-06-09 Andrew Haley + + * src/x86/freebsd.S: Add missing file. + +2009-06-08 Andrew Haley + + Import from libffi 3.0.8: + + * doc/libffi.texi: New file. + * doc/libffi.info: Likewise. + * doc/stamp-vti: Likewise. + * man/Makefile.am: New file. + * man/ffi_call.3: New file. + + * Makefile.am (EXTRA_DIST): Add src/x86/darwin64.S, + src/dlmalloc.c. + (nodist_libffi_la_SOURCES): Add X86_FREEBSD. + + * configure.ac: Bump version to 3.0.8. + parisc*-*-linux*: Add. + i386-*-freebsd* | i386-*-openbsd*: Add. + powerpc-*-beos*: Add. + AM_CONDITIONAL X86_FREEBSD: Add. + AC_CONFIG_FILES: Add man/Makefile. + + * include/ffi.h.in (FFI_FN): Change void (*)() to void (*)(void). + +2009-06-08 Andrew Haley + + * README: Import from libffi 3.0.8. + +2009-06-08 Andrew Haley + + * testsuite/libffi.call/err_bad_abi.c: Add xfails. + * testsuite/libffi.call/cls_longdouble_va.c: Add xfails. + * testsuite/libffi.call/cls_dbls_struct.c: Add xfail x86_64-*-linux-*. + * testsuite/libffi.call/err_bad_typedef.c: Add xfails. + + * testsuite/libffi.call/stret_medium2.c: Add __UNUSED__ to args. + * testsuite/libffi.call/stret_medium.c: Likewise. + * testsuite/libffi.call/stret_large2.c: Likewise. + * testsuite/libffi.call/stret_large.c: Likewise. + +2008-12-26 Timothy Wall + + * testsuite/libffi.call/cls_longdouble.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_align_longdouble.c, + testsuite/libffi.call/cls_align_longdouble_split.c, + testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected + failures on x86_64 cygwin/mingw. + +2008-12-22 Timothy Wall + + * testsuite/libffi.call/closure_fn0.c, + testsuite/libffi.call/closure_fn1.c, + testsuite/libffi.call/closure_fn2.c, + testsuite/libffi.call/closure_fn3.c, + testsuite/libffi.call/closure_fn4.c, + testsuite/libffi.call/closure_fn5.c, + testsuite/libffi.call/closure_fn6.c, + testsuite/libffi.call/closure_loc_fn0.c, + testsuite/libffi.call/closure_stdcall.c, + testsuite/libffi.call/cls_align_pointer.c, + testsuite/libffi.call/cls_pointer.c, + testsuite/libffi.call/cls_pointer_stack.c: use portable cast from + pointer to integer (intptr_t). + * testsuite/libffi.call/cls_longdouble.c: disable for win64. + +2008-07-24 Anthony Green + + * testsuite/libffi.call/cls_dbls_struct.c, + testsuite/libffi.call/cls_double_va.c, + testsuite/libffi.call/cls_longdouble.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_pointer.c, + testsuite/libffi.call/cls_pointer_stack.c, + testsuite/libffi.call/err_bad_abi.c: Clean up failures from + compiler warnings. + +2008-03-04 Anthony Green + Blake Chaffin + hos@tamanegi.org + + * testsuite/libffi.call/cls_align_longdouble_split2.c + testsuite/libffi.call/cls_align_longdouble_split.c + testsuite/libffi.call/cls_dbls_struct.c + testsuite/libffi.call/cls_double_va.c + testsuite/libffi.call/cls_longdouble.c + testsuite/libffi.call/cls_longdouble_va.c + testsuite/libffi.call/cls_pointer.c + testsuite/libffi.call/cls_pointer_stack.c + testsuite/libffi.call/err_bad_abi.c + testsuite/libffi.call/err_bad_typedef.c + testsuite/libffi.call/stret_large2.c + testsuite/libffi.call/stret_large.c + testsuite/libffi.call/stret_medium2.c + testsuite/libffi.call/stret_medium.c: New tests from Apple. + +2009-06-05 Andrew Haley + + * src/x86/ffitarget.h, src/x86/ffi.c: Merge stdcall changes from + libffi. + +2009-06-04 Andrew Haley + + * src/x86/ffitarget.h, src/x86/win32.S, src/x86/ffi.c: Back out + stdcall changes. + +2008-02-26 Anthony Green + Thomas Heller + + * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C + comment. + +2008-02-03 Timothy Wall + + * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return + offset based on code pointer, not data pointer. + +2008-01-31 Timothy Wall + + * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall + closures. + * src/x86/ffitarget.h: Increase size of trampoline for stdcall + closures. + * src/x86/win32.S: Add assembly for stdcall closure. + * src/x86/ffi.c: Initialize stdcall closure trampoline. + +2009-06-04 Andrew Haley + + * include/ffi.h.in: Change void (*)() to void (*)(void). + * src/x86/ffi.c: Likewise. + +2009-06-04 Andrew Haley + + * src/powerpc/ppc_closure.S: Insert licence header. + * src/powerpc/linux64_closure.S: Likewise. + * src/m68k/sysv.S: Likewise. + + * src/sh64/ffi.c: Change void (*)() to void (*)(void). + * src/powerpc/ffi.c: Likewise. + * src/powerpc/ffi_darwin.c: Likewise. + * src/m32r/ffi.c: Likewise. + * src/sh64/ffi.c: Likewise. + * src/x86/ffi64.c: Likewise. + * src/alpha/ffi.c: Likewise. + * src/alpha/osf.S: Likewise. + * src/frv/ffi.c: Likewise. + * src/s390/ffi.c: Likewise. + * src/pa/ffi.c: Likewise. + * src/pa/hpux32.S: Likewise. + * src/ia64/unix.S: Likewise. + * src/ia64/ffi.c: Likewise. + * src/sparc/ffi.c: Likewise. + * src/mips/ffi.c: Likewise. + * src/sh/ffi.c: Likewise. + +2008-02-15 David Daney + + * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE): + Define (conditionally), and use it to include cachectl.h. + (ffi_prep_closure_loc): Fix cache flushing. + * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define. + +2009-06-04 Andrew Haley + + include/ffi.h.in, + src/arm/ffitarget.h, + src/arm/ffi.c, + src/arm/sysv.S, + src/powerpc/ffitarget.h, + src/closures.c, + src/sh64/ffitarget.h, + src/sh64/ffi.c, + src/sh64/sysv.S, + src/types.c, + src/x86/ffi64.c, + src/x86/ffitarget.h, + src/x86/win32.S, + src/x86/darwin.S, + src/x86/ffi.c, + src/x86/sysv.S, + src/x86/unix64.S, + src/alpha/ffitarget.h, + src/alpha/ffi.c, + src/alpha/osf.S, + src/m68k/ffitarget.h, + src/frv/ffitarget.h, + src/frv/ffi.c, + src/s390/ffitarget.h, + src/s390/sysv.S, + src/cris/ffitarget.h, + src/pa/linux.S, + src/pa/ffitarget.h, + src/pa/ffi.c, + src/raw_api.c, + src/ia64/ffitarget.h, + src/ia64/unix.S, + src/ia64/ffi.c, + src/ia64/ia64_flags.h, + src/java_raw_api.c, + src/debug.c, + src/sparc/v9.S, + src/sparc/ffitarget.h, + src/sparc/ffi.c, + src/sparc/v8.S, + src/mips/ffitarget.h, + src/mips/n32.S, + src/mips/o32.S, + src/mips/ffi.c, + src/prep_cif.c, + src/sh/ffitarget.h, + src/sh/ffi.c, + src/sh/sysv.S: Update license text. + +2009-05-22 Dave Korn + + * src/x86/win32.S (_ffi_closure_STDCALL): New function. + (.eh_frame): Add FDE for it. + +2009-05-22 Dave Korn + + * configure.ac: Also check if assembler supports pc-relative + relocs on X86_WIN32 targets. + * configure: Regenerate. + * src/x86/win32.S (ffi_prep_args): Declare extern, not global. + (_ffi_call_SYSV): Add missing function type symbol .def and + add EH markup labels. + (_ffi_call_STDCALL): Likewise. + (_ffi_closure_SYSV): Likewise. + (_ffi_closure_raw_SYSV): Likewise. + (.eh_frame): Add hand-crafted EH data. + +2009-04-09 Jakub Jelinek + + * testsuite/lib/libffi-dg.exp: Change copyright header to refer to + version 3 of the GNU General Public License and to point readers + at the COPYING3 file and the FSF's license web page. + * testsuite/libffi.call/call.exp: Likewise. + * testsuite/libffi.special/special.exp: Likewise. + +2009-03-01 Ralf Wildenhues + + * configure: Regenerate. + +2008-12-18 Rainer Orth + + PR libffi/26048 + * configure.ac (HAVE_AS_X86_PCREL): New test. + * configure: Regenerate. + * fficonfig.h.in: Regenerate. + * src/x86/sysv.S [!FFI_NO_RAW_API]: Precalculate + RAW_CLOSURE_CIF_OFFSET, RAW_CLOSURE_FUN_OFFSET, + RAW_CLOSURE_USER_DATA_OFFSET for the Solaris 10/x86 assembler. + (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL. + * src/x86/unix64.S (.Lstore_table): Move to .text section. + (.Lload_table): Likewise. + (.eh_frame): Only use SYMBOL-. iff HAVE_AS_X86_PCREL. + +2008-12-18 Ralf Wildenhues + + * configure: Regenerate. + +2008-11-21 Eric Botcazou + + * src/sparc/ffi.c (ffi_prep_cif_machdep): Add support for + signed/unsigned int8/16 return values. + * src/sparc/v8.S (ffi_call_v8): Likewise. + (ffi_closure_v8): Likewise. + +2008-09-26 Peter O'Gorman + Steve Ellcey + + * configure: Regenerate for new libtool. + * Makefile.in: Ditto. + * include/Makefile.in: Ditto. + * aclocal.m4: Ditto. + +2008-08-25 Andreas Tobler + + * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and + FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum. + Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT. + Adjust copyright notice. + * src/powerpc/ffi.c: Add two new flags to indicate if we have one + register or two register to use for FFI_SYSV structs. + (ffi_prep_cif_machdep): Pass the right register flag introduced above. + (ffi_closure_helper_SYSV): Fix the return type for + FFI_SYSV_TYPE_SMALL_STRUCT. Comment. + Adjust copyright notice. + +2008-07-16 Kaz Kojima + + * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned + int. + +2008-06-17 Ralf Wildenhues + + * configure: Regenerate. + * include/Makefile.in: Regenerate. + * testsuite/Makefile.in: Regenerate. + +2008-06-07 Joseph Myers + + * configure.ac (parisc*-*-linux*, powerpc-*-sysv*, + powerpc-*-beos*): Remove. + * configure: Regenerate. + +2008-05-09 Julian Brown + + * Makefile.am (LTLDFLAGS): New. + (libffi_la_LDFLAGS): Use above. + * Makefile.in: Regenerate. + +2008-04-18 Paolo Bonzini + + PR bootstrap/35457 + * aclocal.m4: Regenerate. + * configure: Regenerate. + +2008-03-26 Kaz Kojima + + * src/sh/sysv.S: Add .note.GNU-stack on Linux. + * src/sh64/sysv.S: Likewise. + +2008-03-26 Daniel Jacobowitz + + * src/arm/sysv.S: Fix ARM comment marker. + +2008-03-26 Jakub Jelinek + + * src/alpha/osf.S: Add .note.GNU-stack on Linux. + * src/s390/sysv.S: Likewise. + * src/powerpc/ppc_closure.S: Likewise. + * src/powerpc/sysv.S: Likewise. + * src/x86/unix64.S: Likewise. + * src/x86/sysv.S: Likewise. + * src/sparc/v8.S: Likewise. + * src/sparc/v9.S: Likewise. + * src/m68k/sysv.S: Likewise. + * src/arm/sysv.S: Likewise. + +2008-03-16 Ralf Wildenhues + + * aclocal.m4: Regenerate. + * configure: Likewise. + * Makefile.in: Likewise. + * include/Makefile.in: Likewise. + * testsuite/Makefile.in: Likewise. + +2008-02-12 Bjoern Koenig + Andreas Tobler + + * configure.ac: Add amd64-*-freebsd* target. + * configure: Regenerate. + +2008-01-30 H.J. Lu + + PR libffi/34612 + * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when + returning struct. + + * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer" + tests. + +2008-01-24 David Edelsohn + + * configure: Regenerate. + +2008-01-06 Andreas Tobler + + * src/x86/ffi.c (ffi_prep_cif_machdep): Fix thinko. + +2008-01-05 Andreas Tobler + + PR testsuite/32843 + * src/x86/ffi.c (ffi_prep_cif_machdep): Add code for + signed/unsigned int8/16 for X86_DARWIN. + Updated copyright info. + Handle one and two byte structs with special cif->flags. + * src/x86/ffitarget.h: Add special types for one and two byte structs. + Updated copyright info. + * src/x86/darwin.S (ffi_call_SYSV): Rewrite to use a jump table like + sysv.S + Remove code to pop args from the stack after call. + Special-case signed/unsigned for int8/16, one and two byte structs. + (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8, + FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32, + FFI_TYPE_SINT32. + Updated copyright info. + +2007-12-08 David Daney + + * src/mips/n32.S (ffi_call_N32): Replace dadd with ADDU, dsub with + SUBU, add with ADDU and use smaller code sequences. + +2007-12-07 David Daney + + * src/mips/ffi.c (ffi_prep_cif_machdep): Handle long double return + type. + +2007-12-06 David Daney + + * include/ffi.h.in (FFI_SIZEOF_JAVA_RAW): Define if not already + defined. + (ffi_java_raw): New typedef. + (ffi_java_raw_call, ffi_java_ptrarray_to_raw, + ffi_java_raw_to_ptrarray): Change parameter types from ffi_raw to + ffi_java_raw. + (ffi_java_raw_closure) : Same. + (ffi_prep_java_raw_closure, ffi_prep_java_raw_closure_loc): Change + parameter types. + * src/java_raw_api.c (ffi_java_raw_size): Replace FFI_SIZEOF_ARG with + FFI_SIZEOF_JAVA_RAW. + (ffi_java_raw_to_ptrarray): Change type of raw to ffi_java_raw. + Replace FFI_SIZEOF_ARG with FFI_SIZEOF_JAVA_RAW. Use + sizeof(ffi_java_raw) for alignment calculations. + (ffi_java_ptrarray_to_raw): Same. + (ffi_java_rvalue_to_raw): Add special handling for FFI_TYPE_POINTER + if FFI_SIZEOF_JAVA_RAW == 4. + (ffi_java_raw_to_rvalue): Same. + (ffi_java_raw_call): Change type of raw to ffi_java_raw. + (ffi_java_translate_args): Same. + (ffi_prep_java_raw_closure_loc, ffi_prep_java_raw_closure): Change + parameter types. + * src/mips/ffitarget.h (FFI_SIZEOF_JAVA_RAW): Define for N32 ABI. + +2007-12-06 David Daney + + * src/mips/n32.S (ffi_closure_N32): Use 64-bit add instruction on + pointer values. + +2007-12-01 Andreas Tobler + + PR libffi/31937 + * src/powerpc/ffitarget.h: Introduce new ABI FFI_LINUX_SOFT_FLOAT. + Add local FFI_TYPE_UINT128 to handle soft-float long-double-128. + * src/powerpc/ffi.c: Distinguish between __NO_FPRS__ and not and + set the NUM_FPR_ARG_REGISTERS according to. + Add support for potential soft-float support under hard-float + architecture. + (ffi_prep_args_SYSV): Set NUM_FPR_ARG_REGISTERS to 0 in case of + FFI_LINUX_SOFT_FLOAT, handle float, doubles and long-doubles according + to the FFI_LINUX_SOFT_FLOAT ABI. + (ffi_prep_cif_machdep): Likewise. + (ffi_closure_helper_SYSV): Likewise. + * src/powerpc/ppc_closure.S: Make sure not to store float/double + on archs where __NO_FPRS__ is true. + Add FFI_TYPE_UINT128 support. + * src/powerpc/sysv.S: Add support for soft-float long-double-128. + Adjust copyright notice. + +2007-11-25 Andreas Tobler + + * src/closures.c: Move defintion of MAYBE_UNUSED from here to ... + * include/ffi_common.h: ... here. + Update copyright. + +2007-11-17 Andreas Tobler + + * src/powerpc/sysv.S: Load correct cr to compare if we have long double. + * src/powerpc/linux64.S: Likewise. + * src/powerpc/ffi.c: Add a comment to show which part goes into cr6. + * testsuite/libffi.call/return_ldl.c: New test. + +2007-09-04 + + * src/arm/sysv.S (UNWIND): New. + (Whole file): Conditionally compile unwinder directives. + * src/arm/sysv.S: Add unwinder directives. + + * src/arm/ffi.c (ffi_prep_args): Align structs by at least 4 bytes. + Only treat r0 as a struct address if we're actually returning a + struct by address. + Only copy the bytes that are actually within a struct. + (ffi_prep_cif_machdep): A Composite Type not larger than 4 bytes + is returned in r0, not passed by address. + (ffi_call): Allocate a word-sized temporary for the case where + a composite is returned in r0. + (ffi_prep_incoming_args_SYSV): Align as necessary. + +2007-08-05 Steven Newbury + + * src/arm/ffi.c (FFI_INIT_TRAMPOLINE): Use __clear_cache instead of + directly using the sys_cacheflush syscall. + +2007-07-27 Andrew Haley + + * src/arm/sysv.S (ffi_closure_SYSV): Add soft-float. + +2007-09-03 Maciej W. Rozycki + + * Makefile.am: Unify MIPS_IRIX and MIPS_LINUX into MIPS. + * configure.ac: Likewise. + * Makefile.in: Regenerate. + * include/Makefile.in: Likewise. + * testsuite/Makefile.in: Likewise. + * configure: Likewise. + +2007-08-24 David Daney + + * testsuite/libffi.call/return_sl.c: New test. + +2007-08-10 David Daney + + * testsuite/libffi.call/cls_multi_ushort.c, + testsuite/libffi.call/cls_align_uint16.c, + testsuite/libffi.call/nested_struct1.c, + testsuite/libffi.call/nested_struct3.c, + testsuite/libffi.call/cls_7_1_byte.c, + testsuite/libffi.call/nested_struct5.c, + testsuite/libffi.call/cls_double.c, + testsuite/libffi.call/nested_struct7.c, + testsuite/libffi.call/cls_sint.c, + testsuite/libffi.call/nested_struct9.c, + testsuite/libffi.call/cls_20byte1.c, + testsuite/libffi.call/cls_multi_sshortchar.c, + testsuite/libffi.call/cls_align_sint64.c, + testsuite/libffi.call/cls_3byte2.c, + testsuite/libffi.call/cls_multi_schar.c, + testsuite/libffi.call/cls_multi_uchar.c, + testsuite/libffi.call/cls_19byte.c, + testsuite/libffi.call/cls_9byte1.c, + testsuite/libffi.call/cls_align_float.c, + testsuite/libffi.call/closure_fn1.c, + testsuite/libffi.call/problem1.c, + testsuite/libffi.call/closure_fn3.c, + testsuite/libffi.call/cls_sshort.c, + testsuite/libffi.call/closure_fn5.c, + testsuite/libffi.call/cls_align_double.c, + testsuite/libffi.call/nested_struct.c, + testsuite/libffi.call/cls_2byte.c, + testsuite/libffi.call/nested_struct10.c, + testsuite/libffi.call/cls_4byte.c, + testsuite/libffi.call/cls_6byte.c, + testsuite/libffi.call/cls_8byte.c, + testsuite/libffi.call/cls_multi_sshort.c, + testsuite/libffi.call/cls_align_sint16.c, + testsuite/libffi.call/cls_align_uint32.c, + testsuite/libffi.call/cls_20byte.c, + testsuite/libffi.call/cls_float.c, + testsuite/libffi.call/nested_struct2.c, + testsuite/libffi.call/cls_5_1_byte.c, + testsuite/libffi.call/nested_struct4.c, + testsuite/libffi.call/cls_24byte.c, + testsuite/libffi.call/nested_struct6.c, + testsuite/libffi.call/cls_64byte.c, + testsuite/libffi.call/nested_struct8.c, + testsuite/libffi.call/cls_uint.c, + testsuite/libffi.call/cls_multi_ushortchar.c, + testsuite/libffi.call/cls_schar.c, + testsuite/libffi.call/cls_uchar.c, + testsuite/libffi.call/cls_align_uint64.c, + testsuite/libffi.call/cls_ulonglong.c, + testsuite/libffi.call/cls_align_longdouble.c, + testsuite/libffi.call/cls_1_1byte.c, + testsuite/libffi.call/cls_12byte.c, + testsuite/libffi.call/cls_3_1byte.c, + testsuite/libffi.call/cls_3byte1.c, + testsuite/libffi.call/cls_4_1byte.c, + testsuite/libffi.call/cls_6_1_byte.c, + testsuite/libffi.call/cls_16byte.c, + testsuite/libffi.call/cls_18byte.c, + testsuite/libffi.call/closure_fn0.c, + testsuite/libffi.call/cls_9byte2.c, + testsuite/libffi.call/closure_fn2.c, + testsuite/libffi.call/closure_fn4.c, + testsuite/libffi.call/cls_ushort.c, + testsuite/libffi.call/closure_fn6.c, + testsuite/libffi.call/cls_5byte.c, + testsuite/libffi.call/cls_align_pointer.c, + testsuite/libffi.call/cls_7byte.c, + testsuite/libffi.call/cls_align_sint32.c, + testsuite/libffi.special/unwindtest_ffi_call.cc, + testsuite/libffi.special/unwindtest.cc: Remove xfail for mips64*-*-*. + +2007-08-10 David Daney + + PR libffi/28313 + * configure.ac: Don't treat mips64 as a special case. + * Makefile.am (nodist_libffi_la_SOURCES): Add n32.S. + * configure: Regenerate + * Makefile.in: Ditto. + * fficonfig.h.in: Ditto. + * src/mips/ffitarget.h (REG_L, REG_S, SUBU, ADDU, SRL, LI): Indent. + (LA, EH_FRAME_ALIGN, FDE_ADDR_BYTES): New preprocessor macros. + (FFI_DEFAULT_ABI): Set for n64 case. + (FFI_CLOSURES, FFI_TRAMPOLINE_SIZE): Define for n32 and n64 cases. + * src/mips/n32.S (ffi_call_N32): Add debug macros and labels for FDE. + (ffi_closure_N32): New function. + (.eh_frame): New section + * src/mips/o32.S: Clean up comments. + (ffi_closure_O32): Pass ffi_closure parameter in $12. + * src/mips/ffi.c: Use FFI_MIPS_N32 instead of + _MIPS_SIM == _ABIN32 throughout. + (FFI_MIPS_STOP_HERE): New, use in place of + ffi_stop_here. + (ffi_prep_args): Use unsigned long to hold pointer values. Rewrite + to support n32/n64 ABIs. + (calc_n32_struct_flags): Rewrite. + (calc_n32_return_struct_flags): Remove unused variable. Reverse + position of flag bits. + (ffi_prep_cif_machdep): Rewrite n32 portion. + (ffi_call): Enable for n64. Add special handling for small structure + return values. + (ffi_prep_closure_loc): Add n32 and n64 support. + (ffi_closure_mips_inner_O32): Add cast to silence warning. + (copy_struct_N32, ffi_closure_mips_inner_N32): New functions. + +2007-08-08 David Daney + + * testsuite/libffi.call/ffitest.h (ffi_type_mylong): Remove definition. + * testsuite/libffi.call/cls_align_uint16.c (main): Use correct type + specifiers. + * testsuite/libffi.call/nested_struct1.c (main): Ditto. + * testsuite/libffi.call/cls_sint.c (main): Ditto. + * testsuite/libffi.call/nested_struct9.c (main): Ditto. + * testsuite/libffi.call/cls_20byte1.c (main): Ditto. + * testsuite/libffi.call/cls_9byte1.c (main): Ditto. + * testsuite/libffi.call/closure_fn1.c (main): Ditto. + * testsuite/libffi.call/closure_fn3.c (main): Ditto. + * testsuite/libffi.call/return_dbl2.c (main): Ditto. + * testsuite/libffi.call/cls_sshort.c (main): Ditto. + * testsuite/libffi.call/return_fl3.c (main): Ditto. + * testsuite/libffi.call/closure_fn5.c (main): Ditto. + * testsuite/libffi.call/nested_struct.c (main): Ditto. + * testsuite/libffi.call/nested_struct10.c (main): Ditto. + * testsuite/libffi.call/return_ll1.c (main): Ditto. + * testsuite/libffi.call/cls_8byte.c (main): Ditto. + * testsuite/libffi.call/cls_align_uint32.c (main): Ditto. + * testsuite/libffi.call/cls_align_sint16.c (main): Ditto. + * testsuite/libffi.call/cls_20byte.c (main): Ditto. + * testsuite/libffi.call/nested_struct2.c (main): Ditto. + * testsuite/libffi.call/cls_24byte.c (main): Ditto. + * testsuite/libffi.call/nested_struct6.c (main): Ditto. + * testsuite/libffi.call/cls_uint.c (main): Ditto. + * testsuite/libffi.call/cls_12byte.c (main): Ditto. + * testsuite/libffi.call/cls_16byte.c (main): Ditto. + * testsuite/libffi.call/closure_fn0.c (main): Ditto. + * testsuite/libffi.call/cls_9byte2.c (main): Ditto. + * testsuite/libffi.call/closure_fn2.c (main): Ditto. + * testsuite/libffi.call/return_dbl1.c (main): Ditto. + * testsuite/libffi.call/closure_fn4.c (main): Ditto. + * testsuite/libffi.call/closure_fn6.c (main): Ditto. + * testsuite/libffi.call/cls_align_sint32.c (main): Ditto. + +2007-08-07 Andrew Haley + + * src/x86/sysv.S (ffi_closure_raw_SYSV): Fix typo in previous + checkin. + +2007-08-06 Andrew Haley + + PR testsuite/32843 + * src/x86/sysv.S (ffi_closure_raw_SYSV): Handle FFI_TYPE_UINT8, + FFI_TYPE_SINT8, FFI_TYPE_UINT16, FFI_TYPE_SINT16, FFI_TYPE_UINT32, + FFI_TYPE_SINT32. + +2007-08-02 David Daney + + * testsuite/libffi.call/return_ul.c (main): Define return type as + ffi_arg. Use proper printf conversion specifier. + +2007-07-30 Andrew Haley + + PR testsuite/32843 + * src/x86/ffi.c (ffi_prep_cif_machdep): in x86 case, add code for + signed/unsigned int8/16. + * src/x86/sysv.S (ffi_call_SYSV): Rewrite to: + Use a jump table. + Remove code to pop args from the stack after call. + Special-case signed/unsigned int8/16. + * testsuite/libffi.call/return_sc.c (main): Revert. + +2007-07-26 Richard Guenther + + PR testsuite/32843 + * testsuite/libffi.call/return_sc.c (main): Verify call + result as signed char, not ffi_arg. + +2007-07-16 Rainer Orth + + * configure.ac (i?86-*-solaris2.1[0-9]): Set TARGET to X86_64. + * configure: Regenerate. + +2007-07-11 David Daney + + * src/mips/ffi.c: Don't include sys/cachectl.h. + (ffi_prep_closure_loc): Use __builtin___clear_cache() instead of + cacheflush(). + +2007-05-18 Aurelien Jarno + + * src/arm/ffi.c (ffi_prep_closure_loc): Renamed and ajusted + from (ffi_prep_closure): ... this. + (FFI_INIT_TRAMPOLINE): Adjust. + +2005-12-31 Phil Blundell + + * src/arm/ffi.c (ffi_prep_incoming_args_SYSV, + ffi_closure_SYSV_inner, ffi_prep_closure): New, add closure support. + * src/arm/sysv.S(ffi_closure_SYSV): Likewise. + * src/arm/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise. + (FFI_CLOSURES): Enable closure support. + +2007-07-03 Andrew Haley + + * testsuite/libffi.call/cls_multi_ushort.c, + testsuite/libffi.call/cls_align_uint16.c, + testsuite/libffi.call/nested_struct1.c, + testsuite/libffi.call/nested_struct3.c, + testsuite/libffi.call/cls_7_1_byte.c, + testsuite/libffi.call/cls_double.c, + testsuite/libffi.call/nested_struct5.c, + testsuite/libffi.call/nested_struct7.c, + testsuite/libffi.call/cls_sint.c, + testsuite/libffi.call/nested_struct9.c, + testsuite/libffi.call/cls_20byte1.c, + testsuite/libffi.call/cls_multi_sshortchar.c, + testsuite/libffi.call/cls_align_sint64.c, + testsuite/libffi.call/cls_3byte2.c, + testsuite/libffi.call/cls_multi_schar.c, + testsuite/libffi.call/cls_multi_uchar.c, + testsuite/libffi.call/cls_19byte.c, + testsuite/libffi.call/cls_9byte1.c, + testsuite/libffi.call/cls_align_float.c, + testsuite/libffi.call/closure_fn1.c, + testsuite/libffi.call/problem1.c, + testsuite/libffi.call/closure_fn3.c, + testsuite/libffi.call/cls_sshort.c, + testsuite/libffi.call/closure_fn5.c, + testsuite/libffi.call/cls_align_double.c, + testsuite/libffi.call/cls_2byte.c, + testsuite/libffi.call/nested_struct.c, + testsuite/libffi.call/nested_struct10.c, + testsuite/libffi.call/cls_4byte.c, + testsuite/libffi.call/cls_6byte.c, + testsuite/libffi.call/cls_8byte.c, + testsuite/libffi.call/cls_multi_sshort.c, + testsuite/libffi.call/cls_align_uint32.c, + testsuite/libffi.call/cls_align_sint16.c, + testsuite/libffi.call/cls_float.c, + testsuite/libffi.call/cls_20byte.c, + testsuite/libffi.call/cls_5_1_byte.c, + testsuite/libffi.call/nested_struct2.c, + testsuite/libffi.call/cls_24byte.c, + testsuite/libffi.call/nested_struct4.c, + testsuite/libffi.call/nested_struct6.c, + testsuite/libffi.call/cls_64byte.c, + testsuite/libffi.call/nested_struct8.c, + testsuite/libffi.call/cls_uint.c, + testsuite/libffi.call/cls_multi_ushortchar.c, + testsuite/libffi.call/cls_schar.c, + testsuite/libffi.call/cls_uchar.c, + testsuite/libffi.call/cls_align_uint64.c, + testsuite/libffi.call/cls_ulonglong.c, + testsuite/libffi.call/cls_align_longdouble.c, + testsuite/libffi.call/cls_1_1byte.c, + testsuite/libffi.call/cls_12byte.c, + testsuite/libffi.call/cls_3_1byte.c, + testsuite/libffi.call/cls_3byte1.c, + testsuite/libffi.call/cls_4_1byte.c, + testsuite/libffi.call/cls_6_1_byte.c, + testsuite/libffi.call/cls_16byte.c, + testsuite/libffi.call/cls_18byte.c, + testsuite/libffi.call/closure_fn0.c, + testsuite/libffi.call/cls_9byte2.c, + testsuite/libffi.call/closure_fn2.c, + testsuite/libffi.call/closure_fn4.c, + testsuite/libffi.call/cls_ushort.c, + testsuite/libffi.call/closure_fn6.c, + testsuite/libffi.call/cls_5byte.c, + testsuite/libffi.call/cls_align_pointer.c, + testsuite/libffi.call/cls_7byte.c, + testsuite/libffi.call/cls_align_sint32.c, + testsuite/libffi.special/unwindtest_ffi_call.cc, + testsuite/libffi.special/unwindtest.cc: Enable for ARM. + +2007-07-05 H.J. Lu + + * aclocal.m4: Regenerated. + +2007-06-02 Paolo Bonzini + + * configure: Regenerate. + +2007-05-23 Steve Ellcey + + * Makefile.in: Regenerate. + * configure: Regenerate. + * aclocal.m4: Regenerate. + * include/Makefile.in: Regenerate. + * testsuite/Makefile.in: Regenerate. + +2007-05-10 Roman Zippel + + * src/m68k/ffi.c (ffi_prep_incoming_args_SYSV, + ffi_closure_SYSV_inner,ffi_prep_closure): New, add closure support. + * src/m68k/sysv.S(ffi_closure_SYSV,ffi_closure_struct_SYSV): Likewise. + * src/m68k/ffitarget.h (FFI_TRAMPOLINE_SIZE): Likewise. + (FFI_CLOSURES): Enable closure support. + +2007-05-10 Roman Zippel + + * configure.ac (HAVE_AS_CFI_PSEUDO_OP): New test. + * configure: Regenerate. + * fficonfig.h.in: Regenerate. + * src/m68k/sysv.S (CFI_STARTPROC,CFI_ENDPROC, + CFI_OFFSET,CFI_DEF_CFA): New macros. + (ffi_call_SYSV): Add callframe annotation. + +2007-05-10 Roman Zippel + + * src/m68k/ffi.c (ffi_prep_args,ffi_prep_cif_machdep): Fix + numerous test suite failures. + * src/m68k/sysv.S (ffi_call_SYSV): Likewise. + +2007-04-11 Paolo Bonzini + + * Makefile.am (EXTRA_DIST): Bring up to date. + * Makefile.in: Regenerate. + * src/frv/eabi.S: Remove RCS keyword. + +2007-04-06 Richard Henderson + + * configure.ac: Tidy target case. + (HAVE_LONG_DOUBLE): Allow the target to override. + * configure: Regenerate. + * include/ffi.h.in: Don't define ffi_type_foo if + LIBFFI_HIDE_BASIC_TYPES is defined. + (ffi_type_longdouble): If not HAVE_LONG_DOUBLE, define + to ffi_type_double. + * types.c (LIBFFI_HIDE_BASIC_TYPES): Define. + (FFI_TYPEDEF, ffi_type_void): Mark the data const. + (ffi_type_longdouble): Special case for Alpha. Don't define + if long double == double. + + * src/alpha/ffi.c (FFI_TYPE_LONGDOUBLE): Assert unique value. + (ffi_prep_cif_machdep): Handle it as the 128-bit type. + (ffi_call, ffi_closure_osf_inner): Likewise. + (ffi_closure_osf_inner): Likewise. Mark hidden. + (ffi_call_osf, ffi_closure_osf): Mark hidden. + * src/alpha/ffitarget.h (FFI_LAST_ABI): Tidy definition. + * src/alpha/osf.S (ffi_call_osf, ffi_closure_osf): Mark hidden. + (load_table): Handle 128-bit long double. + + * testsuite/libffi.call/float4.c: Add -mieee for alpha. + +2007-04-06 Tom Tromey + + PR libffi/31491: + * README: Fixed bug in example. + +2007-04-03 Jakub Jelinek + + * src/closures.c: Include sys/statfs.h. + (_GNU_SOURCE): Define on Linux. + (FFI_MMAP_EXEC_SELINUX): Define. + (selinux_enabled): New variable. + (selinux_enabled_check): New function. + (is_selinux_enabled): Define. + (dlmmap): Use it. + +2007-03-24 Uros Bizjak + + * testsuite/libffi.call/return_fl2.c (return_fl): Mark as static. + Use 'volatile float sum' to create sum of floats to avoid false + negative due to excess precision on ix86 targets. + (main): Ditto. + +2007-03-08 Alexandre Oliva + + * src/powerpc/ffi.c (flush_icache): Fix left-over from previous + patch. + (ffi_prep_closure_loc): Remove unneeded casts. Add needed ones. + +2007-03-07 Alexandre Oliva + + * include/ffi.h.in (ffi_closure_alloc, ffi_closure_free): New. + (ffi_prep_closure_loc): New. + (ffi_prep_raw_closure_loc): New. + (ffi_prep_java_raw_closure_loc): New. + * src/closures.c: New file. + * src/dlmalloc.c [FFI_MMAP_EXEC_WRIT] (struct malloc_segment): + Replace sflags with exec_offset. + [FFI_MMAP_EXEC_WRIT] (mmap_exec_offset, add_segment_exec_offset, + sub_segment_exec_offset): New macros. + (get_segment_flags, set_segment_flags, check_segment_merge): New + macros. + (is_mmapped_segment, is_extern_segment): Use get_segment_flags. + (add_segment, sys_alloc, create_mspace, create_mspace_with_base, + destroy_mspace): Use new macros. + (sys_alloc): Silence warning. + * Makefile.am (libffi_la_SOURCES): Add src/closures.c. + * Makefile.in: Rebuilt. + * src/prep_cif [FFI_CLOSURES] (ffi_prep_closure): Implement in + terms of ffi_prep_closure_loc. + * src/raw_api.c (ffi_prep_raw_closure_loc): Renamed and adjusted + from... + (ffi_prep_raw_closure): ... this. Re-implement in terms of the + renamed version. + * src/java_raw_api (ffi_prep_java_raw_closure_loc): Renamed and + adjusted from... + (ffi_prep_java_raw_closure): ... this. Re-implement in terms of + the renamed version. + * src/alpha/ffi.c (ffi_prep_closure_loc): Renamed from + (ffi_prep_closure): ... this. + * src/pa/ffi.c: Likewise. + * src/cris/ffi.c: Likewise. Adjust. + * src/frv/ffi.c: Likewise. + * src/ia64/ffi.c: Likewise. + * src/mips/ffi.c: Likewise. + * src/powerpc/ffi_darwin.c: Likewise. + * src/s390/ffi.c: Likewise. + * src/sh/ffi.c: Likewise. + * src/sh64/ffi.c: Likewise. + * src/sparc/ffi.c: Likewise. + * src/x86/ffi64.c: Likewise. + * src/x86/ffi.c: Likewise. + (FFI_INIT_TRAMPOLINE): Adjust. + (ffi_prep_raw_closure_loc): Renamed and adjusted from... + (ffi_prep_raw_closure): ... this. + * src/powerpc/ffi.c (ffi_prep_closure_loc): Renamed from + (ffi_prep_closure): ... this. + (flush_icache): Adjust. + +2007-03-07 Alexandre Oliva + + * src/dlmalloc.c: New file, imported version 2.8.3 of Doug + Lea's malloc. + +2007-03-01 Brooks Moses + + * Makefile.am: Add dummy install-pdf target. + * Makefile.in: Regenerate + +2007-02-13 Andreas Krebbel + + * src/s390/ffi.c (ffi_prep_args, ffi_prep_cif_machdep, + ffi_closure_helper_SYSV): Add long double handling. + +2007-02-02 Jakub Jelinek + + * src/powerpc/linux64.S (ffi_call_LINUX64): Move restore of r2 + immediately after bctrl instruction. + +2007-01-18 Alexandre Oliva + + * Makefile.am (all-recursive, install-recursive, + mostlyclean-recursive, clean-recursive, distclean-recursive, + maintainer-clean-recursive): Add missing targets. + * Makefile.in: Rebuilt. + +2006-12-14 Andreas Tobler + + * configure.ac: Add TARGET for x86_64-*-darwin*. + * Makefile.am (nodist_libffi_la_SOURCES): Add rules for 64-bit sources + for X86_DARWIN. + * src/x86/ffitarget.h: Set trampoline size for x86_64-*-darwin*. + * src/x86/darwin64.S: New file for x86_64-*-darwin* support. + * configure: Regenerate. + * Makefile.in: Regenerate. + * include/Makefile.in: Regenerate. + * testsuite/Makefile.in: Regenerate. + * testsuite/libffi.special/unwindtest_ffi_call.cc: New test case for + ffi_call only. + +2006-12-13 Andreas Tobler + + * aclocal.m4: Regenerate with aclocal -I .. as written in the + Makefile.am. + +2006-10-31 Geoffrey Keating + + * src/powerpc/ffi_darwin.c (darwin_adjust_aggregate_sizes): New. + (ffi_prep_cif_machdep): Call darwin_adjust_aggregate_sizes for + Darwin. + * testsuite/libffi.call/nested_struct4.c: Remove Darwin XFAIL. + * testsuite/libffi.call/nested_struct6.c: Remove Darwin XFAIL. + +2006-10-10 Paolo Bonzini + Sandro Tolaini + + * configure.ac [i*86-*-darwin*]: Set X86_DARWIN symbol and + conditional. + * configure: Regenerated. + * Makefile.am (nodist_libffi_la_SOURCES) [X86_DARWIN]: New case. + (EXTRA_DIST): Add src/x86/darwin.S. + * Makefile.in: Regenerated. + * include/Makefile.in: Regenerated. + * testsuite/Makefile.in: Regenerated. + + * src/x86/ffi.c (ffi_prep_cif_machdep) [X86_DARWIN]: Treat like + X86_WIN32, and additionally align stack to 16 bytes. + * src/x86/darwin.S: New, based on sysv.S. + * src/prep_cif.c (ffi_prep_cif) [X86_DARWIN]: Align > 8-byte structs. + +2006-09-12 David Daney + + PR libffi/23935 + * include/Makefile.am: Install both ffi.h and ffitarget.h in + $(libdir)/gcc/$(target_alias)/$(gcc_version)/include. + * aclocal.m4: Regenerated for automake 1.9.6. + * Makefile.in: Regenerated. + * include/Makefile.in: Regenerated. + * testsuite/Makefile.in: Regenerated. + +2006-08-17 Andreas Tobler + + * include/ffi_common.h (struct): Revert accidental commit. + +2006-08-15 Andreas Tobler + + * include/ffi_common.h: Remove lint directives. + * include/ffi.h.in: Likewise. + +2006-07-25 Torsten Schoenfeld + + * include/ffi.h.in (ffi_type_ulong, ffi_type_slong): Define correctly + for 32-bit architectures. + * testsuite/libffi.call/return_ul.c: New test case. + +2006-07-19 David Daney + + * testsuite/libffi.call/closure_fn6.c: Remove xfail for mips, + xfail remains for mips64. + +2006-05-23 Carlos O'Donell + + * Makefile.am: Add install-html target. Add install-html to .PHONY + * Makefile.in: Regenerate. + * aclocal.m4: Regenerate. + * include/Makefile.in: Regenerate. + * testsuite/Makefile.in: Regenerate. + +2006-05-18 John David Anglin + + * pa/ffi.c (ffi_prep_args_pa32): Load floating point arguments from + stack slot. + +2006-04-22 Andreas Tobler + + * README: Remove notice about 'Crazy Comments'. + * src/debug.c: Remove lint directives. Cleanup white spaces. + * src/java_raw_api.c: Likewise. + * src/prep_cif.c: Likewise. + * src/raw_api.c: Likewise. + * src/ffitest.c: Delete. No longer needed, all test cases migrated + to the testsuite. + * src/arm/ffi.c: Remove lint directives. + * src/m32r/ffi.c: Likewise. + * src/pa/ffi.c: Likewise. + * src/powerpc/ffi.c: Likewise. + * src/powerpc/ffi_darwin.c: Likewise. + * src/sh/ffi.c: Likewise. + * src/sh64/ffi.c: Likewise. + * src/x86/ffi.c: Likewise. + * testsuite/libffi.call/float2.c: Likewise. + * testsuite/libffi.call/promotion.c: Likewise. + * testsuite/libffi.call/struct1.c: Likewise. + +2006-04-13 Andreas Tobler + + * src/pa/hpux32.S: Correct unwind offset calculation for + ffi_closure_pa32. + * src/pa/linux.S: Likewise. + +2006-04-12 James E Wilson + + PR libgcj/26483 + * src/ia64/ffi.c (stf_spill, ldf_fill): Rewrite as macros. + (hfa_type_load): Call stf_spill. + (hfa_type_store): Call ldf_fill. + (ffi_call): Adjust calls to above routines. Add local temps for + macro result. + +2006-04-10 Matthias Klose + + * testsuite/lib/libffi-dg.exp (libffi-init): Recognize multilib + directory names containing underscores. + +2006-04-07 James E Wilson + + * testsuite/libffi.call/float4.c: New testcase. + +2006-04-05 John David Anglin + Andreas Tobler + + * Makefile.am: Add PA_HPUX port. + * Makefile.in: Regenerate. + * include/Makefile.in: Likewise. + * testsuite/Makefile.in: Likewise. + * configure.ac: Add PA_HPUX rules. + * configure: Regenerate. + * src/pa/ffitarget.h: Rename linux target to PA_LINUX. + Add PA_HPUX and PA64_HPUX. + Rename FFI_LINUX ABI to FFI_PA32 ABI. + (FFI_TRAMPOLINE_SIZE): Define for 32-bit HP-UX targets. + (FFI_TYPE_SMALL_STRUCT2): Define. + (FFI_TYPE_SMALL_STRUCT4): Likewise. + (FFI_TYPE_SMALL_STRUCT8): Likewise. + (FFI_TYPE_SMALL_STRUCT3): Redefine. + (FFI_TYPE_SMALL_STRUCT5): Likewise. + (FFI_TYPE_SMALL_STRUCT6): Likewise. + (FFI_TYPE_SMALL_STRUCT7): Likewise. + * src/pa/ffi.c (ROUND_DOWN): Delete. + (fldw, fstw, fldd, fstd): Use '__asm__'. + (ffi_struct_type): Add support for FFI_TYPE_SMALL_STRUCT2, + FFI_TYPE_SMALL_STRUCT4 and FFI_TYPE_SMALL_STRUCT8. + (ffi_prep_args_LINUX): Rename to ffi_prep_args_pa32. Update comment. + Simplify incrementing of stack slot variable. Change type of local + 'n' to unsigned int. + (ffi_size_stack_LINUX): Rename to ffi_size_stack_pa32. Handle long + double on PA_HPUX. + (ffi_prep_cif_machdep): Likewise. + (ffi_call): Likewise. + (ffi_closure_inner_LINUX): Rename to ffi_closure_inner_pa32. Change + return type to ffi_status. Simplify incrementing of stack slot + variable. Only copy floating point argument registers when PA_LINUX + is true. Reformat debug statement. + Add support for FFI_TYPE_SMALL_STRUCT2, FFI_TYPE_SMALL_STRUCT4 and + FFI_TYPE_SMALL_STRUCT8. + (ffi_closure_LINUX): Rename to ffi_closure_pa32. Add 'extern' to + declaration. + (ffi_prep_closure): Make linux trampoline conditional on PA_LINUX. + Add nops to cache flush. Add trampoline for PA_HPUX. + * src/pa/hpux32.S: New file. + * src/pa/linux.S (ffi_call_LINUX): Rename to ffi_call_pa32. Rename + ffi_prep_args_LINUX to ffi_prep_args_pa32. + Localize labels. Add support for 2, 4 and 8-byte small structs. Handle + unaligned destinations in 3, 5, 6 and 7-byte small structs. Order + argument type checks so that common argument types appear first. + (ffi_closure_LINUX): Rename to ffi_closure_pa32. Rename + ffi_closure_inner_LINUX to ffi_closure_inner_pa32. + +2006-03-24 Alan Modra + + * src/powerpc/ffitarget.h (enum ffi_abi): Add FFI_LINUX. Default + for 32-bit using IBM extended double format. Fix FFI_LAST_ABI. + * src/powerpc/ffi.c (ffi_prep_args_SYSV): Handle linux variant of + FFI_TYPE_LONGDOUBLE. + (ffi_prep_args64): Assert using IBM extended double. + (ffi_prep_cif_machdep): Don't munge FFI_TYPE_LONGDOUBLE type. + Handle FFI_LINUX FFI_TYPE_LONGDOUBLE return and args. + (ffi_call): Handle FFI_LINUX. + (ffi_closure_helper_SYSV): Non FFI_LINUX long double return needs + gpr3 return pointer as for struct return. Handle FFI_LINUX + FFI_TYPE_LONGDOUBLE return and args. Don't increment "nf" + unnecessarily. + * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Load both f1 and f2 + for FFI_TYPE_LONGDOUBLE. Move epilogue insns into case table. + Don't use r6 as pointer to results, instead use sp offset. Don't + make a special call to load lr with case table address, instead + use offset from previous call. + * src/powerpc/sysv.S (ffi_call_SYSV): Save long double return. + * src/powerpc/linux64.S (ffi_call_LINUX64): Simplify long double + return. + +2006-03-15 Kaz Kojima + + * src/sh64/ffi.c (ffi_prep_cif_machdep): Handle float arguments + passed with FP registers correctly. + (ffi_closure_helper_SYSV): Likewise. + * src/sh64/sysv.S: Likewise. + +2006-03-01 Andreas Tobler + + * testsuite/libffi.special/unwindtest.cc (closure_test_fn): Mark cif, + args and userdata unused. + (closure_test_fn1): Mark cif and userdata unused. + (main): Remove unused res. + +2006-02-28 Andreas Tobler + + * testsuite/libffi.call/call.exp: Adjust FSF address. Add test runs for + -O2, -O3, -Os and the warning flags -W -Wall. + * testsuite/libffi.special/special.exp: Likewise. + * testsuite/libffi.call/ffitest.h: Add an __UNUSED__ macro to mark + unused parameter unused for gcc or else do nothing. + * testsuite/libffi.special/ffitestcxx.h: Likewise. + * testsuite/libffi.call/cls_12byte.c (cls_struct_12byte_gn): Mark cif + and userdata unused. + * testsuite/libffi.call/cls_16byte.c (cls_struct_16byte_gn): Likewise. + * testsuite/libffi.call/cls_18byte.c (cls_struct_18byte_gn): Likewise. + * testsuite/libffi.call/cls_19byte.c (cls_struct_19byte_gn): Likewise. + * testsuite/libffi.call/cls_1_1byte.c (cls_struct_1_1byte_gn): Likewise. + * testsuite/libffi.call/cls_20byte.c (cls_struct_20byte_gn): Likewise. + * testsuite/libffi.call/cls_20byte1.c (cls_struct_20byte_gn): Likewise. + * testsuite/libffi.call/cls_24byte.c (cls_struct_24byte_gn): Likewise. + * testsuite/libffi.call/cls_2byte.c (cls_struct_2byte_gn): Likewise. + * testsuite/libffi.call/cls_3_1byte.c (cls_struct_3_1byte_gn): Likewise. + * testsuite/libffi.call/cls_3byte1.c (cls_struct_3byte_gn): Likewise. + * testsuite/libffi.call/cls_3byte2.c (cls_struct_3byte_gn1): Likewise. + * testsuite/libffi.call/cls_4_1byte.c (cls_struct_4_1byte_gn): Likewise. + * testsuite/libffi.call/cls_4byte.c (cls_struct_4byte_gn): Likewise. + * testsuite/libffi.call/cls_5_1_byte.c (cls_struct_5byte_gn): Likewise. + * testsuite/libffi.call/cls_5byte.c (cls_struct_5byte_gn): Likewise. + * testsuite/libffi.call/cls_64byte.c (cls_struct_64byte_gn): Likewise. + * testsuite/libffi.call/cls_6_1_byte.c (cls_struct_6byte_gn): Likewise. + * testsuite/libffi.call/cls_6byte.c (cls_struct_6byte_gn): Likewise. + * testsuite/libffi.call/cls_7_1_byte.c (cls_struct_7byte_gn): Likewise. + * testsuite/libffi.call/cls_7byte.c (cls_struct_7byte_gn): Likewise. + * testsuite/libffi.call/cls_8byte.c (cls_struct_8byte_gn): Likewise. + * testsuite/libffi.call/cls_9byte1.c (cls_struct_9byte_gn): Likewise. + * testsuite/libffi.call/cls_9byte2.c (cls_struct_9byte_gn): Likewise. + * testsuite/libffi.call/cls_align_double.c (cls_struct_align_gn): + Likewise. + * testsuite/libffi.call/cls_align_float.c (cls_struct_align_gn): + Likewise. + * testsuite/libffi.call/cls_align_longdouble.c (cls_struct_align_gn): + Likewise. + * testsuite/libffi.call/cls_align_pointer.c (cls_struct_align_fn): Cast + void* to avoid compiler warning. + (main): Likewise. + (cls_struct_align_gn): Mark cif and userdata unused. + * testsuite/libffi.call/cls_align_sint16.c (cls_struct_align_gn): + Likewise. + * testsuite/libffi.call/cls_align_sint32.c (cls_struct_align_gn): + Likewise. + * testsuite/libffi.call/cls_align_sint64.c (cls_struct_align_gn): + Likewise. + * testsuite/libffi.call/cls_align_uint16.c (cls_struct_align_gn): + Likewise. + * testsuite/libffi.call/cls_align_uint32.c (cls_struct_align_gn): + Likewise. + * testsuite/libffi.call/cls_double.c (cls_ret_double_fn): Likewise. + * testsuite/libffi.call/cls_float.c (cls_ret_float_fn): Likewise. + * testsuite/libffi.call/cls_multi_schar.c (test_func_gn): Mark cif and + data unused. + (main): Cast res_call to silence gcc. + * testsuite/libffi.call/cls_multi_sshort.c (test_func_gn): Mark cif and + data unused. + (main): Cast res_call to silence gcc. + * testsuite/libffi.call/cls_multi_sshortchar.c (test_func_gn): Mark cif + and data unused. + (main): Cast res_call to silence gcc. + * testsuite/libffi.call/cls_multi_uchar.c (test_func_gn): Mark cif and + data unused. + (main): Cast res_call to silence gcc. + * testsuite/libffi.call/cls_multi_ushort.c (test_func_gn): Mark cif and + data unused. + (main): Cast res_call to silence gcc. + * testsuite/libffi.call/cls_multi_ushortchar.c (test_func_gn): Mark cif + and data unused. + (main): Cast res_call to silence gcc. + * testsuite/libffi.call/cls_schar.c (cls_ret_schar_fn): Mark cif and + userdata unused. + (cls_ret_schar_fn): Cast printf parameter to silence gcc. + * testsuite/libffi.call/cls_sint.c (cls_ret_sint_fn): Mark cif and + userdata unused. + (cls_ret_sint_fn): Cast printf parameter to silence gcc. + * testsuite/libffi.call/cls_sshort.c (cls_ret_sshort_fn): Mark cif and + userdata unused. + (cls_ret_sshort_fn): Cast printf parameter to silence gcc. + * testsuite/libffi.call/cls_uchar.c (cls_ret_uchar_fn): Mark cif and + userdata unused. + (cls_ret_uchar_fn): Cast printf parameter to silence gcc. + * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Mark cif and + userdata unused. + (cls_ret_uint_fn): Cast printf parameter to silence gcc. + * testsuite/libffi.call/cls_ulonglong.c (cls_ret_ulonglong_fn): Mark cif + and userdata unused. + * testsuite/libffi.call/cls_ushort.c (cls_ret_ushort_fn): Mark cif and + userdata unused. + (cls_ret_ushort_fn): Cast printf parameter to silence gcc. + * testsuite/libffi.call/float.c (floating): Remove unused parameter e. + * testsuite/libffi.call/float1.c (main): Remove unused variable i. + Cleanup white spaces. + * testsuite/libffi.call/negint.c (checking): Remove unused variable i. + * testsuite/libffi.call/nested_struct.c (cls_struct_combined_gn): Mark + cif and userdata unused. + * testsuite/libffi.call/nested_struct1.c (cls_struct_combined_gn): + Likewise. + * testsuite/libffi.call/nested_struct10.c (B_gn): Likewise. + * testsuite/libffi.call/nested_struct2.c (B_fn): Adjust printf + formatters to silence gcc. + (B_gn): Mark cif and userdata unused. + * testsuite/libffi.call/nested_struct3.c (B_gn): Mark cif and userdata + unused. + * testsuite/libffi.call/nested_struct4.c: Mention related PR. + (B_gn): Mark cif and userdata unused. + * testsuite/libffi.call/nested_struct5.c (B_gn): Mark cif and userdata + unused. + * testsuite/libffi.call/nested_struct6.c: Mention related PR. + (B_gn): Mark cif and userdata unused. + * testsuite/libffi.call/nested_struct7.c (B_gn): Mark cif and userdata + unused. + * testsuite/libffi.call/nested_struct8.c (B_gn): Likewise. + * testsuite/libffi.call/nested_struct9.c (B_gn): Likewise. + * testsuite/libffi.call/problem1.c (stub): Likewise. + * testsuite/libffi.call/pyobjc-tc.c (main): Cast the result to silence + gcc. + * testsuite/libffi.call/return_fl2.c (return_fl): Add the note mentioned + in the last commit for this test case in the test case itself. + * testsuite/libffi.call/closure_fn0.c (closure_test_fn0): Mark cif as + unused. + * testsuite/libffi.call/closure_fn1.c (closure_test_fn1): Likewise. + * testsuite/libffi.call/closure_fn2.c (closure_test_fn2): Likewise. + * testsuite/libffi.call/closure_fn3.c (closure_test_fn3): Likewise. + * testsuite/libffi.call/closure_fn4.c (closure_test_fn0): Likewise. + * testsuite/libffi.call/closure_fn5.c (closure_test_fn5): Likewise. + * testsuite/libffi.call/closure_fn6.c (closure_test_fn0): Likewise. + +2006-02-22 Kaz Kojima + + * src/sh/sysv.S: Fix register numbers in the FDE for + ffi_closure_SYSV. + +2006-02-20 Andreas Tobler + + * testsuite/libffi.call/return_fl2.c (return_fl): Remove static + declaration to avoid a false negative on ix86. See PR323. + +2006-02-18 Kaz Kojima + + * src/sh/ffi.c (ffi_closure_helper_SYSV): Remove unused variable + and cast integer to void * if needed. Update the pointer to + the FP register saved area correctly. + +2006-02-17 Andreas Tobler + + * testsuite/libffi.call/nested_struct6.c: XFAIL this test until PR25630 + is fixed. + * testsuite/libffi.call/nested_struct4.c: Likewise. + +2006-02-16 Andreas Tobler + + * testsuite/libffi.call/return_dbl.c: New test case. + * testsuite/libffi.call/return_dbl1.c: Likewise. + * testsuite/libffi.call/return_dbl2.c: Likewise. + * testsuite/libffi.call/return_fl.c: Likewise. + * testsuite/libffi.call/return_fl1.c: Likewise. + * testsuite/libffi.call/return_fl2.c: Likewise. + * testsuite/libffi.call/return_fl3.c: Likewise. + * testsuite/libffi.call/closure_fn6.c: Likewise. + + * testsuite/libffi.call/nested_struct2.c: Remove ffi_type_mylong + definition. + * testsuite/libffi.call/ffitest.h: Add ffi_type_mylong definition + here to be used by other test cases too. + + * testsuite/libffi.call/nested_struct10.c: New test case. + * testsuite/libffi.call/nested_struct9.c: Likewise. + * testsuite/libffi.call/nested_struct8.c: Likewise. + * testsuite/libffi.call/nested_struct7.c: Likewise. + * testsuite/libffi.call/nested_struct6.c: Likewise. + * testsuite/libffi.call/nested_struct5.c: Likewise. + * testsuite/libffi.call/nested_struct4.c: Likewise. + +2006-01-21 Andreas Tobler + + * configure.ac: Enable libffi for sparc64-*-freebsd*. + * configure: Rebuilt. + +2006-01-18 Jakub Jelinek + + * src/powerpc/sysv.S (smst_two_register): Don't call __ashldi3, + instead do the shifting inline. + * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't compute %r5 + shift count unconditionally. Simplify load sequences for 1, 2, 3, 4 + and 8 byte structs, for the remaining struct sizes don't call + __lshrdi3, instead do the shifting inline. + +2005-12-07 Thiemo Seufer + + * src/mips/ffitarget.h: Remove obsolete sgidefs.h include. Add + missing parentheses. + * src/mips/o32.S (ffi_call_O32): Code formatting. Define + and use A3_OFF, FP_OFF, RA_OFF. Micro-optimizations. + (ffi_closure_O32): Likewise, but with newly defined A3_OFF2, + A2_OFF2, A1_OFF2, A0_OFF2, RA_OFF2, FP_OFF2, S0_OFF2, GP_OFF2, + V1_OFF2, V0_OFF2, FA_1_1_OFF2, FA_1_0_OFF2, FA_0_1_OFF2, + FA_0_0_OFF2. + * src/mips/ffi.c (ffi_prep_args): Code formatting. Fix + endianness bugs. + (ffi_prep_closure): Improve trampoline instruction scheduling. + (ffi_closure_mips_inner_O32): Fix endianness bugs. + +2005-12-03 Alan Modra + + * src/powerpc/ffi.c: Formatting. + (ffi_prep_args_SYSV): Avoid possible aliasing problems by using unions. + (ffi_prep_args64): Likewise. + +2005-09-30 Geoffrey Keating + + * testsuite/lib/libffi-dg.exp (libffi_target_compile): For + darwin, use -shared-libgcc not -lgcc_s, and explain why. + +2005-09-26 Tom Tromey + + * testsuite/libffi.call/float1.c (value_type): New typedef. + (CANARY): New define. + (main): Check for result buffer overflow. + * src/powerpc/linux64.S: Handle linux64 long double returns. + * src/powerpc/ffi.c (FLAG_RETURNS_128BITS): New constant. + (ffi_prep_cif_machdep): Handle linux64 long double returns. + +2005-08-25 Alan Modra + + PR target/23404 + * src/powerpc/ffi.c (ffi_prep_args_SYSV): Correct placement of stack + homed fp args. + (ffi_status ffi_prep_cif_machdep): Correct stack sizing for same. + +2005-08-11 Jakub Jelinek + + * configure.ac (HAVE_HIDDEN_VISIBILITY_ATTRIBUTE): New test. + (AH_BOTTOM): Add FFI_HIDDEN definition. + * configure: Rebuilt. + * fficonfig.h.in: Rebuilt. + * src/powerpc/ffi.c (hidden): Remove. + (ffi_closure_LINUX64, ffi_prep_args64, ffi_call_LINUX64, + ffi_closure_helper_LINUX64): Use FFI_HIDDEN instead of hidden. + * src/powerpc/linux64_closure.S (ffi_closure_LINUX64, + .ffi_closure_LINUX64): Use FFI_HIDDEN instead of .hidden. + * src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): Remove, + add FFI_HIDDEN to its prototype. + (ffi_closure_SYSV_inner): New. + * src/x86/sysv.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New. + * src/x86/win32.S (ffi_closure_SYSV, ffi_closure_raw_SYSV): New. + +2005-08-10 Alfred M. Szmidt + + PR libffi/21819: + * configure: Rebuilt. + * configure.ac: Handle i*86-*-gnu*. + +2005-08-09 Jakub Jelinek + + * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Use + DW_CFA_offset_extended_sf rather than + DW_CFA_GNU_negative_offset_extended. + * src/powerpc/sysv.S (ffi_call_SYSV): Likewise. + +2005-07-22 SUGIOKA Toshinobu + + * src/sh/sysv.S (ffi_call_SYSV): Stop argument popping correctly + on sh3. + (ffi_closure_SYSV): Change the stack layout for sh3 struct argument. + * src/sh/ffi.c (ffi_prep_args): Fix sh3 argument copy, when it is + partially on register. + (ffi_closure_helper_SYSV): Likewise. + (ffi_prep_cif_machdep): Don't set too many cif->flags. + +2005-07-20 Kaz Kojima + + * src/sh/ffi.c (ffi_call): Handle small structures correctly. + Remove empty line. + * src/sh64/ffi.c (simple_type): Remove. + (return_type): Handle small structures correctly. + (ffi_prep_args): Likewise. + (ffi_call): Likewise. + (ffi_closure_helper_SYSV): Likewise. + * src/sh64/sysv.S (ffi_call_SYSV): Handle 1, 2 and 4-byte return. + Emit position independent code if PIC and remove wrong datalabel + prefixes from EH data. + +2005-07-19 Andreas Tobler + + * Makefile.am (nodist_libffi_la_SOURCES): Add POWERPC_FREEBSD. + * Makefile.in: Regenerate. + * include/Makefile.in: Likewise. + * testsuite/Makefile.in: Likewise. + * configure.ac: Add POWERPC_FREEBSD rules. + * configure: Regenerate. + * src/powerpc/ffitarget.h: Add POWERPC_FREEBSD rules. + (FFI_SYSV_TYPE_SMALL_STRUCT): Define. + * src/powerpc/ffi.c: Add flags to handle small structure returns + in ffi_call_SYSV. + (ffi_prep_cif_machdep): Handle small structures for SYSV 4 ABI. + Aka FFI_SYSV. + (ffi_closure_helper_SYSV): Likewise. + * src/powerpc/ppc_closure.S: Add return types for small structures. + * src/powerpc/sysv.S: Add bits to handle small structures for + final SYSV 4 ABI. + +2005-07-10 Andreas Tobler + + * testsuite/libffi.call/cls_5_1_byte.c: New test file. + * testsuite/libffi.call/cls_6_1_byte.c: Likewise. + * testsuite/libffi.call/cls_7_1_byte.c: Likewise. + +2005-07-05 Randolph Chung + + * src/pa/ffi.c (ffi_struct_type): Rename FFI_TYPE_SMALL_STRUCT1 + as FFI_TYPE_SMALL_STRUCT3. Break out handling for 5-7 byte + structures. Kill compilation warnings. + (ffi_closure_inner_LINUX): Print return values as hex in debug + message. Rename FFI_TYPE_SMALL_STRUCT1 as FFI_TYPE_SMALL_STRUCT3. + Properly handle 5-7 byte structure returns. + * src/pa/ffitarget.h (FFI_TYPE_SMALL_STRUCT1) + (FFI_TYPE_SMALL_STRUCT2): Remove. + (FFI_TYPE_SMALL_STRUCT3, FFI_TYPE_SMALL_STRUCT5) + (FFI_TYPE_SMALL_STRUCT6, FFI_TYPE_SMALL_STRUCT7): Define. + * src/pa/linux.S: Mark source file as using PA1.1 assembly. + (checksmst1, checksmst2): Remove. + (checksmst3): Optimize handling of 3-byte struct returns. + (checksmst567): Properly handle 5-7 byte struct returns. + +2005-06-15 Rainer Orth + + PR libgcj/21943 + * src/mips/n32.S: Enforce PIC code. + * src/mips/o32.S: Likewise. + +2005-06-15 Rainer Orth + + * configure.ac: Treat i*86-*-solaris2.10 and up as X86_64. + * configure: Regenerate. + +2005-06-01 Alan Modra + + * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Don't use JUMPTARGET + to call ffi_closure_helper_SYSV. Append @local instead. + * src/powerpc/sysv.S (ffi_call_SYSV): Likewise for ffi_prep_args_SYSV. + +2005-05-17 Kelley Cook + + * configure.ac: Use AC_C_BIGENDIAN instead of AC_C_BIGENDIAN_CROSS. + Use AC_CHECK_SIZEOF instead of AC_COMPILE_CHECK_SIZEOF. + * Makefile.am (ACLOCAL_AMFLAGS): Remove -I ../config. + * aclocal.m4, configure, fficonfig.h.in, Makefile.in, + include/Makefile.in, testsuite/Makefile.in: Regenerate. + +2005-05-09 Mike Stump + + * configure: Regenerate. + +2005-05-08 Richard Henderson + + PR libffi/21285 + * src/alpha/osf.S: Update unwind into to match code. + +2005-05-04 Andreas Degert + Richard Henderson + + * src/x86/ffi64.c (ffi_prep_cif_machdep): Save sse-used flag in + bit 11 of flags. + (ffi_call): Mask return type field. Pass ssecount to ffi_call_unix64. + (ffi_prep_closure): Set carry bit if sse-used flag set. + * src/x86/unix64.S (ffi_call_unix64): Add ssecount argument. + Only load sse registers if ssecount non-zero. + (ffi_closure_unix64): Only save sse registers if carry set on entry. + +2005-04-29 Ralf Corsepius + + * configure.ac: Add i*86-*-rtems*, sparc*-*-rtems*, + powerpc-*rtems*, arm*-*-rtems*, sh-*-rtems*. + * configure: Regenerate. + +2005-04-20 Hans-Peter Nilsson + + * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): In regsub use, + have Tcl8.3-compatible intermediate variable. + +2005-04-18 Simon Posnjak + Hans-Peter Nilsson + + * Makefile.am: Add CRIS support. + * configure.ac: Likewise. + * Makefile.in, configure, testsuite/Makefile.in, + include/Makefile.in: Regenerate. + * src/cris: New directory. + * src/cris/ffi.c, src/cris/sysv.S, src/cris/ffitarget.h: New files. + * src/prep_cif.c (ffi_prep_cif): Wrap in #ifndef __CRIS__. + + * testsuite/lib/libffi-dg.exp (libffi-dg-test-1): Replace \n with + \r?\n in output tests. + +2005-04-12 Mike Stump + + * configure: Regenerate. + +2005-03-30 Hans Boehm + + * src/ia64/ffitarget.h (ffi_arg): Use long long instead of DI. + +2005-03-30 Steve Ellcey + + * src/ia64/ffitarget.h (ffi_arg) ADD DI attribute. + (ffi_sarg) Ditto. + * src/ia64/unix.S (ffi_closure_unix): Extend gp + to 64 bits in ILP32 mode. + Load 64 bits even for short data. + +2005-03-23 Mike Stump + + * src/powerpc/darwin.S: Update for -m64 multilib. + * src/powerpc/darwin_closure.S: Likewise. + +2005-03-21 Zack Weinberg + + * configure.ac: Do not invoke TL_AC_GCC_VERSION. + Do not set tool_include_dir. + * aclocal.m4, configure, Makefile.in, testsuite/Makefile.in: + Regenerate. + * include/Makefile.am: Set gcc_version and toollibffidir. + * include/Makefile.in: Regenerate. + +2005-02-22 Andrew Haley + + * src/powerpc/ffi.c (ffi_prep_cif_machdep): Bump alignment to + odd-numbered register pairs for 64-bit integer types. + +2005-02-23 Andreas Tobler + + PR libffi/20104 + * testsuite/libffi.call/return_ll1.c: New test case. + +2005-02-11 Janis Johnson + + * testsuite/libffi.call/cls_align_longdouble.c: Remove dg-options. + * testsuite/libffi.call/float.c: Ditto. + * testsuite/libffi.call/float2.c: Ditto. + * testsuite/libffi.call/float3.c: Ditto. + +2005-02-08 Andreas Tobler + + * src/frv/ffitarget.h: Remove PPC stuff which does not belong to frv. + +2005-01-12 Eric Botcazou + + * testsuite/libffi.special/special.exp (cxx_options): Add + -shared-libgcc. + +2004-12-31 Richard Henderson + + * src/types.c (FFI_AGGREGATE_TYPEDEF): Remove. + (FFI_TYPEDEF): Rename from FFI_INTEGRAL_TYPEDEF. Replace size and + offset parameters with a type parameter; deduce size and structure + alignment. Update all users. + +2004-12-31 Richard Henderson + + * src/types.c (FFI_TYPE_POINTER): Define with sizeof. + (FFI_TYPE_LONGDOUBLE): Fix for ia64. + * src/ia64/ffitarget.h (struct ffi_ia64_trampoline_struct): Move + into ffi_prep_closure. + * src/ia64/ia64_flags.h, src/ia64/ffi.c, src/ia64/unix.S: Rewrite + from scratch. + +2004-12-27 Richard Henderson + + * src/x86/unix64.S: Fix typo in unwind info. + +2004-12-25 Richard Henderson + + * src/x86/ffi64.c (struct register_args): Rename from stackLayout. + (enum x86_64_reg_class): Add X86_64_COMPLEX_X87_CLASS. + (merge_classes): Check for it. + (SSE_CLASS_P): New. + (classify_argument): Pass byte_offset by value; perform all updates + inside struct case. + (examine_argument): Add classes argument; handle + X86_64_COMPLEX_X87_CLASS. + (ffi_prep_args): Merge into ... + (ffi_call): ... here. Share stack frame with ffi_call_unix64. + (ffi_prep_cif_machdep): Setup cif->flags for proper structure return. + (ffi_fill_return_value): Remove. + (ffi_prep_closure): Remove dead assert. + (ffi_closure_unix64_inner): Rename from ffi_closure_UNIX64_inner. + Rewrite to use struct register_args instead of va_list. Create + flags for handling structure returns. + * src/x86/unix64.S: Remove dead strings. + (ffi_call_unix64): Rename from ffi_call_UNIX64. Rewrite to share + stack frame with ffi_call. Handle structure returns properly. + (float2sse, floatfloat2sse, double2sse): Remove. + (sse2float, sse2double, sse2floatfloat): Remove. + (ffi_closure_unix64): Rename from ffi_closure_UNIX64. Rewrite + to handle structure returns properly. + +2004-12-08 David Edelsohn + + * Makefile.am (AM_MAKEFLAGS): Remove duplicate LIBCFLAGS and + PICFLAG. + * Makefile.in: Regenerated. + +2004-12-02 Richard Sandiford + + * configure.ac: Use TL_AC_GCC_VERSION to set gcc_version. + * configure, aclocal.m4, Makefile.in: Regenerate. + * include/Makefile.in, testsuite/Makefile.in: Regenerate. + +2004-11-29 Kelley Cook + + * configure: Regenerate for libtool change. + +2004-11-25 Kelley Cook + + * configure: Regenerate for libtool reversion. + +2004-11-24 Kelley Cook + + * configure: Regenerate for libtool change. + +2004-11-23 John David Anglin + + * testsuite/lib/libffi-dg.exp: Use new procs in target-libpath.exp. + +2004-11-23 Richard Sandiford + + * src/mips/o32.S (ffi_call_O32, ffi_closure_O32): Use jalr instead + of jal. Use an absolute encoding for the frame information. + +2004-11-23 Kelley Cook + + * Makefile.am: Remove no-dependencies. Add ACLOCAL_AMFLAGS. + * acinclude.m4: Delete logic for sincludes. + * aclocal.m4, Makefile.in, configure: Regenerate. + * include/Makefile: Likewise. + * testsuite/Makefile: Likewise. + +2004-11-22 Eric Botcazou + + * src/sparc/ffi.c (ffi_prep_closure): Align doubles and 64-bit integers + on a 8-byte boundary. + * src/sparc/v8.S (ffi_closure_v8): Reserve frame space for arguments. + +2004-10-27 Richard Earnshaw + + * src/arm/ffi.c (ffi_prep_cif_machdep): Handle functions that return + long long values. Round stack allocation to a multiple of 8 bytes + for ATPCS compatibility. + * src/arm/sysv.S (ffi_call_SYSV): Rework to avoid use of APCS register + names. Handle returning long long types. Add Thumb and interworking + support. Improve soft-float code. + +2004-10-27 Richard Earnshaw + + * testsuite/lib/libffi-db.exp (load_gcc_lib): New function. + (libffi_exit): New function. + (libffi_init): Build the testglue wrapper if needed. + +2004-10-25 Eric Botcazou + + PR other/18138 + * testsuite/lib/libffi-dg.exp: Accept more than one multilib libgcc. + +2004-10-25 Kazuhiro Inaoka + + * src/m32r/libffitarget.h (FFI_CLOSURES): Set to 0. + +2004-10-20 Kaz Kojima + + * src/sh/sysv.S (ffi_call_SYSV): Don't align for double data. + * testsuite/libffi.call/float3.c: New test case. + +2004-10-18 Kaz Kojima + + * src/sh/ffi.c (ffi_prep_closure): Set T bit in trampoline for + the function returning a structure pointed with R2. + * src/sh/sysv.S (ffi_closure_SYSV): Use R2 as the pointer to + the structure return value if T bit set. Emit position + independent code and EH data if PIC. + +2004-10-13 Kazuhiro Inaoka + + * Makefile.am: Add m32r support. + * configure.ac: Likewise. + * Makefile.in: Regenerate. + * confiugre: Regenerate. + * src/types.c: Add m32r port to FFI_INTERNAL_TYPEDEF + (uint64, sint64, double, longdouble) + * src/m32r: New directory. + * src/m32r/ffi.c: New file. + * src/m32r/sysv.S: Likewise. + * src/m32r/ffitarget.h: Likewise. + +2004-10-02 Kaz Kojima + + * testsuite/libffi.call/negint.c: New test case. + +2004-09-14 H.J. Lu + + PR libgcj/17465 + * testsuite/lib/libffi-dg.exp: Don't use global ld_library_path. + Set up LD_LIBRARY_PATH, SHLIB_PATH, LD_LIBRARYN32_PATH, + LD_LIBRARY64_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64 and + DYLD_LIBRARY_PATH. + +2004-09-05 Andreas Tobler + + * testsuite/libffi.call/many_win32.c: Remove whitespaces. + * testsuite/libffi.call/promotion.c: Likewise. + * testsuite/libffi.call/return_ll.c: Remove unused var. Cleanup + whitespaces. + * testsuite/libffi.call/return_sc.c: Likewise. + * testsuite/libffi.call/return_uc.c: Likewise. + +2004-09-05 Andreas Tobler + + * src/powerpc/darwin.S: Fix comments and identation. + * src/powerpc/darwin_closure.S: Likewise. + +2004-09-02 Andreas Tobler + + * src/powerpc/ffi_darwin.c: Add flag for longdouble return values. + (ffi_prep_args): Handle longdouble arguments. + (ffi_prep_cif_machdep): Set flags for longdouble. Calculate space for + longdouble. + (ffi_closure_helper_DARWIN): Add closure handling for longdouble. + * src/powerpc/darwin.S (_ffi_call_DARWIN): Add handling of longdouble + values. + * src/powerpc/darwin_closure.S (_ffi_closure_ASM): Likewise. + * src/types.c: Defined longdouble size and alignment for darwin. + +2004-09-02 Andreas Tobler + + * src/powerpc/aix.S: Remove whitespaces. + * src/powerpc/aix_closure.S: Likewise. + * src/powerpc/asm.h: Likewise. + * src/powerpc/ffi.c: Likewise. + * src/powerpc/ffitarget.h: Likewise. + * src/powerpc/linux64.S: Likewise. + * src/powerpc/linux64_closure.S: Likewise. + * src/powerpc/ppc_closure.S: Likewise. + * src/powerpc/sysv.S: Likewise. + +2004-08-30 Anthony Green + + * Makefile.am: Add frv support. + * Makefile.in, testsuite/Makefile.in: Rebuilt. + * configure.ac: Read configure.host. + * configure.in: Read configure.host. + * configure.host: New file. frv-elf needs libgloss. + * include/ffi.h.in: Force ffi_closure to have a nice big (8) + alignment. This is needed to frv and shouldn't harm the others. + * include/ffi_common.h (ALIGN_DOWN): New macro. + * src/frv/ffi.c, src/frv/ffitarget.h, src/frv/eabi.S: New files. + +2004-08-24 David Daney + + * testsuite/libffi.call/closure_fn0.c: Xfail mips64* instead of mips*. + * testsuite/libffi.call/closure_fn1.c: Likewise. + * testsuite/libffi.call/closure_fn2.c Likewise. + * testsuite/libffi.call/closure_fn3.c: Likewise. + * testsuite/libffi.call/closure_fn4.c: Likewise. + * testsuite/libffi.call/closure_fn5.c: Likewise. + * testsuite/libffi.call/cls_18byte.c: Likewise. + * testsuite/libffi.call/cls_19byte.c: Likewise. + * testsuite/libffi.call/cls_1_1byte.c: Likewise. + * testsuite/libffi.call/cls_20byte.c: Likewise. + * testsuite/libffi.call/cls_20byte1.c: Likewise. + * testsuite/libffi.call/cls_24byte.c: Likewise. + * testsuite/libffi.call/cls_2byte.c: Likewise. + * testsuite/libffi.call/cls_3_1byte.c: Likewise. + * testsuite/libffi.call/cls_3byte1.c: Likewise. + * testsuite/libffi.call/cls_3byte2.c: Likewise. + * testsuite/libffi.call/cls_4_1byte.c: Likewise. + * testsuite/libffi.call/cls_4byte.c: Likewise. + * testsuite/libffi.call/cls_64byte.c: Likewise. + * testsuite/libffi.call/cls_6byte.c: Likewise. + * testsuite/libffi.call/cls_7byte.c: Likewise. + * testsuite/libffi.call/cls_8byte.c: Likewise. + * testsuite/libffi.call/cls_9byte1.c: Likewise. + * testsuite/libffi.call/cls_9byte2.c: Likewise. + * testsuite/libffi.call/cls_align_double.c: Likewise. + * testsuite/libffi.call/cls_align_float.c: Likewise. + * testsuite/libffi.call/cls_align_longdouble.c: Likewise. + * testsuite/libffi.call/cls_align_pointer.c: Likewise. + * testsuite/libffi.call/cls_align_sint16.c: Likewise. + * testsuite/libffi.call/cls_align_sint32.c: Likewise. + * testsuite/libffi.call/cls_align_sint64.c: Likewise. + * testsuite/libffi.call/cls_align_uint16.c: Likewise. + * testsuite/libffi.call/cls_align_uint32.c: Likewise. + * testsuite/libffi.call/cls_align_uint64.c: Likewise. + * testsuite/libffi.call/cls_double.c: Likewise. + * testsuite/libffi.call/cls_float.c: Likewise. + * testsuite/libffi.call/cls_multi_schar.c: Likewise. + * testsuite/libffi.call/cls_multi_sshort.c: Likewise. + * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise. + * testsuite/libffi.call/cls_multi_uchar.c: Likewise. + * testsuite/libffi.call/cls_multi_ushort.c: Likewise. + * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise. + * testsuite/libffi.call/cls_schar.c: Likewise. + * testsuite/libffi.call/cls_sint.c: Likewise. + * testsuite/libffi.call/cls_sshort.c: Likewise. + * testsuite/libffi.call/cls_uchar.c: Likewise. + * testsuite/libffi.call/cls_uint.c: Likewise. + * testsuite/libffi.call/cls_ulonglong.c: Likewise. + * testsuite/libffi.call/cls_ushort.c: Likewise. + * testsuite/libffi.call/nested_struct.c: Likewise. + * testsuite/libffi.call/nested_struct1.c: Likewise. + * testsuite/libffi.call/nested_struct2.c: Likewise. + * testsuite/libffi.call/nested_struct3.c: Likewise. + * testsuite/libffi.call/problem1.c: Likewise. + * testsuite/libffi.special/unwindtest.cc: Likewise. + * testsuite/libffi.call/cls_12byte.c: Likewise and set return value + to zero. + * testsuite/libffi.call/cls_16byte.c: Likewise. + * testsuite/libffi.call/cls_5byte.c: Likewise. + +2004-08-23 David Daney + + PR libgcj/13141 + * src/mips/ffitarget.h (FFI_O32_SOFT_FLOAT): New ABI. + * src/mips/ffi.c (ffi_prep_args): Fix alignment calculation. + (ffi_prep_cif_machdep): Handle FFI_O32_SOFT_FLOAT floating point + parameters and return types. + (ffi_call): Handle FFI_O32_SOFT_FLOAT ABI. + (ffi_prep_closure): Ditto. + (ffi_closure_mips_inner_O32): Handle FFI_O32_SOFT_FLOAT ABI, fix + alignment calculations. + * src/mips/o32.S (ffi_closure_O32): Don't use floating point + instructions if FFI_O32_SOFT_FLOAT, make stack frame ABI compliant. + +2004-08-14 Casey Marshall + + * src/mips/ffi.c (ffi_pref_cif_machdep): set `cif->flags' to + contain `FFI_TYPE_UINT64' as return type for any 64-bit + integer (O32 ABI only). + (ffi_prep_closure): new function. + (ffi_closure_mips_inner_O32): new function. + * src/mips/ffitarget.h: Define `FFI_CLOSURES' and + `FFI_TRAMPOLINE_SIZE' appropriately if the ABI is o32. + * src/mips/o32.S (ffi_call_O32): add labels for .eh_frame. Return + 64 bit integers correctly. + (ffi_closure_O32): new function. + Added DWARF-2 unwind info for both functions. + +2004-08-10 Andrew Haley + + * src/x86/ffi64.c (ffi_prep_args ): 8-align all stack arguments. + +2004-08-01 Robert Millan + + * configure.ac: Detect knetbsd-gnu and kfreebsd-gnu. + * configure: Regenerate. + +2004-07-30 Maciej W. Rozycki + + * acinclude.m4 (AC_FUNC_MMAP_BLACKLIST): Check for + and mmap() explicitly instead of relying on preset autoconf cache + variables. + * aclocal.m4: Regenerate. + * configure: Regenerate. + +2004-07-11 Ulrich Weigand + + * src/s390/ffi.c (ffi_prep_args): Fix C aliasing violation. + (ffi_check_float_struct): Remove unused prototype. + +2004-06-30 Geoffrey Keating + + * src/powerpc/ffi_darwin.c (flush_icache): ';' is a comment + character on Darwin, use '\n\t' instead. + +2004-06-26 Matthias Klose + + * libtool-version: Fix typo in revision/age. + +2004-06-17 Matthias Klose + + * libtool-version: New. + * Makefile.am (libffi_la_LDFLAGS): Use -version-info for soname. + * Makefile.in: Regenerate. + +2004-06-15 Paolo Bonzini + + * Makefile.am: Remove useless multilib rules. + * Makefile.in: Regenerate. + * aclocal.m4: Regenerate with automake 1.8.5. + * configure.ac: Remove useless multilib configury. + * configure: Regenerate. + +2004-06-15 Paolo Bonzini + + * .cvsignore: New file. + +2004-06-10 Jakub Jelinek + + * src/ia64/unix.S (ffi_call_unix): Insert group barrier break + fp_done. + (ffi_closure_UNIX): Fix f14/f15 adjustment if FLOAT_SZ is ever + changed from 8. + +2004-06-06 Sean McNeil + + * configure.ac: Add x86_64-*-freebsd* support. + * configure: Regenerate. + +2004-04-26 Joe Buck + + Bug 15093 + * configure.ac: Test for existence of mmap and sys/mman.h before + checking blacklist. Fix suggested by Jim Wilson. + * configure: Regenerate. + +2004-04-26 Matt Austern + + * src/powerpc/darwin.S: Go through a non-lazy pointer for initial + FDE location. + * src/powerpc/darwin_closure.S: Likewise. + +2004-04-24 Andreas Tobler + + * testsuite/libffi.call/cls_multi_schar.c (main): Fix initialization + error. Reported by Thomas Heller . + * testsuite/libffi.call/cls_multi_sshort.c (main): Likewise. + * testsuite/libffi.call/cls_multi_ushort.c (main): Likewise. + +2004-03-20 Matthias Klose + + * src/pa/linux.S: Fix typo. + +2004-03-19 Matthias Klose + + * Makefile.am: Update. + * Makefile.in: Regenerate. + * src/pa/ffi.h.in: Remove. + * src/pa/ffitarget.h: New file. + +2004-02-10 Randolph Chung + + * Makefile.am: Add PA support. + * Makefile.in: Regenerate. + * include/Makefile.in: Regenerate. + * configure.ac: Add PA target. + * configure: Regenerate. + * src/pa/ffi.c: New file. + * src/pa/ffi.h.in: Add PA support. + * src/pa/linux.S: New file. + * prep_cif.c: Add PA support. + +2004-03-16 Hosaka Yuji + + * src/types.c: Fix alignment size of X86_WIN32 case int64 and + double. + * src/x86/ffi.c (ffi_prep_args): Replace ecif->cif->rtype->type + with ecif->cif->flags. + (ffi_call, ffi_prep_incoming_args_SYSV): Replace cif->rtype->type + with cif->flags. + (ffi_prep_cif_machdep): Add X86_WIN32 struct case. + (ffi_closure_SYSV): Add 1 or 2-bytes struct case for X86_WIN32. + * src/x86/win32.S (retstruct1b, retstruct2b, sc_retstruct1b, + sc_retstruct2b): Add for 1 or 2-bytes struct case. + +2004-03-15 Kelley Cook + + * configure.in: Rename file to ... + * configure.ac: ... this. + * fficonfig.h.in: Regenerate. + * Makefile.in: Regenerate. + * include/Makefile.in: Regenerate. + * testsuite/Makefile.in: Regenerate. + +2004-03-12 Matt Austern + + * src/powerpc/darwin.S: Fix EH information so it corresponds to + changes in EH format resulting from addition of linkonce support. + * src/powerpc/darwin_closure.S: Likewise. + +2004-03-11 Andreas Tobler + Paolo Bonzini + + * Makefile.am (AUTOMAKE_OPTIONS): Set them. + Remove VPATH. Remove rules for object files. Remove multilib support. + (AM_CCASFLAGS): Add. + * configure.in (AC_CONFIG_HEADERS): Relace AM_CONFIG_HEADER. + (AC_PREREQ): Bump version to 2.59. + (AC_INIT): Fill with version info and bug address. + (ORIGINAL_LD_FOR_MULTILIBS): Remove. + (AM_ENABLE_MULTILIB): Use this instead of AC_ARG_ENABLE. + De-precious CC so that the right flags are passed down to multilibs. + (AC_MSG_ERROR): Replace obsolete macro AC_ERROR. + (AC_CONFIG_FILES): Replace obsolete macro AC_LINK_FILES. + (AC_OUTPUT): Reorganize the output with AC_CONFIG_COMMANDS. + * configure: Rebuilt. + * aclocal.m4: Likewise. + * Makefile.in, include/Makefile.in, testsuite/Makefile.in: Likewise. + * fficonfig.h.in: Likewise. + +2004-03-11 Andreas Schwab + + * src/ia64/ffi.c (ffi_prep_incoming_args_UNIX): Get floating point + arguments from fp registers only for the first 8 parameter slots. + Don't convert a float parameter when passed in memory. + +2004-03-09 Hans-Peter Nilsson + + * configure: Regenerate for config/accross.m4 correction. + +2004-02-25 Matt Kraai + + * src/powerpc/ffi.c (ffi_prep_args_SYSV): Change + ecif->cif->bytes to bytes. + (ffi_prep_cif_machdep): Add braces around nested if statement. + +2004-02-09 Alan Modra + + * src/types.c (pointer): POWERPC64 has 8 byte pointers. + + * src/powerpc/ffi.c (ffi_prep_args64): Correct long double handling. + (ffi_closure_helper_LINUX64): Fix typo. + * testsuite/libffi.call/cls_align_longdouble.c: Pass -mlong-double-128 + for powerpc64-*-*. + * testsuite/libffi.call/float.c: Likewise. + * testsuite/libffi.call/float2.c: Likewise. + +2004-02-08 Alan Modra + + * src/powerpc/ffi.c (ffi_prep_cif_machdep ): Correct + long double function return and long double arg handling. + (ffi_closure_helper_LINUX64): Formatting. Delete unused "ng" var. + Use "end_pfr" instead of "nf". Correct long double handling. + Localise "temp". + * src/powerpc/linux64.S (ffi_call_LINUX64): Save f2 long double + return value. + * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Allocate + space for long double return value. Adjust stack frame and offsets. + Load f2 long double return. + +2004-02-07 Alan Modra + + * src/types.c: Use 16 byte long double for POWERPC64. + +2004-01-25 Eric Botcazou + + * src/sparc/ffi.c (ffi_prep_args_v9): Shift the parameter array + when the structure return address is passed in %o0. + (ffi_V9_return_struct): Rename into ffi_v9_layout_struct. + (ffi_v9_layout_struct): Align the field following a nested structure + on a word boundary. Use memmove instead of memcpy. + (ffi_call): Update call to ffi_V9_return_struct. + (ffi_prep_closure): Define 'ctx' only for V8. + (ffi_closure_sparc_inner): Clone into ffi_closure_sparc_inner_v8 + and ffi_closure_sparc_inner_v9. + (ffi_closure_sparc_inner_v8): Return long doubles by reference. + Always skip the structure return address. For structures and long + doubles, copy the argument directly. + (ffi_closure_sparc_inner_v9): Skip the structure return address only + if required. Shift the maximum floating-point slot accordingly. For + big structures, copy the argument directly; otherwise, left-justify the + argument and call ffi_v9_layout_struct to lay out the structure on + the stack. + * src/sparc/v8.S: Undef STACKFRAME before defining it. + (ffi_closure_v8): Pass the structure return address. Update call to + ffi_closure_sparc_inner_v8. Short-circuit FFI_TYPE_INT handling. + Skip the 'unimp' insn when returning long doubles and structures. + * src/sparc/v9.S: Undef STACKFRAME before defining it. + (ffi_closure_v9): Increase the frame size by 2 words. Short-circuit + FFI_TYPE_INT handling. Load structures both in integers and + floating-point registers on return. + * README: Update status of the SPARC port. + +2004-01-24 Andreas Tobler + + * testsuite/libffi.call/pyobjc-tc.c (main): Treat result value + as of type ffi_arg. + * testsuite/libffi.call/struct3.c (main): Fix CHECK. + +2004-01-22 Ulrich Weigand + + * testsuite/libffi.call/cls_uint.c (cls_ret_uint_fn): Treat result + value as of type ffi_arg, not unsigned int. + +2004-01-21 Michael Ritzert + + * ffi64.c (ffi_prep_args): Cast the RHS of an assignment instead + of the LHS. + +2004-01-12 Andreas Tobler + + * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_32 for + Solaris. + +2004-01-08 Rainer Orth + + * testsuite/libffi.call/ffitest.h (allocate_mmap): Cast MAP_FAILED + to void *. + +2003-12-10 Richard Henderson + + * testsuite/libffi.call/cls_align_pointer.c: Cast pointers to + size_t instead of int. + +2003-12-04 Hosaka Yuji + + * testsuite/libffi.call/many_win32.c: Include . + * testsuite/libffi.call/many_win32.c (main): Replace variable + int i with unsigned long ul. + + * testsuite/libffi.call/cls_align_uint64.c: New test case. + * testsuite/libffi.call/cls_align_sint64.c: Likewise. + * testsuite/libffi.call/cls_align_uint32.c: Likewise. + * testsuite/libffi.call/cls_align_sint32.c: Likewise. + * testsuite/libffi.call/cls_align_uint16.c: Likewise. + * testsuite/libffi.call/cls_align_sint16.c: Likewise. + * testsuite/libffi.call/cls_align_float.c: Likewise. + * testsuite/libffi.call/cls_align_double.c: Likewise. + * testsuite/libffi.call/cls_align_longdouble.c: Likewise. + * testsuite/libffi.call/cls_align_pointer.c: Likewise. + +2003-12-02 Hosaka Yuji + + PR other/13221 + * src/x86/ffi.c (ffi_prep_args, ffi_prep_incoming_args_SYSV): + Align arguments to 32 bits. + +2003-12-01 Andreas Tobler + + PR other/13221 + * testsuite/libffi.call/cls_multi_sshort.c: New test case. + * testsuite/libffi.call/cls_multi_sshortchar.c: Likewise. + * testsuite/libffi.call/cls_multi_uchar.c: Likewise. + * testsuite/libffi.call/cls_multi_schar.c: Likewise. + * testsuite/libffi.call/cls_multi_ushortchar.c: Likewise. + * testsuite/libffi.call/cls_multi_ushort.c: Likewise. + + * testsuite/libffi.special/unwindtest.cc: Cosmetics. + +2003-11-26 Kaveh R. Ghazi + + * testsuite/libffi.call/ffitest.h: Include . + * testsuite/libffi.special/ffitestcxx.h: Likewise. + +2003-11-22 Andreas Tobler + + * Makefile.in: Rebuilt. + * configure: Likewise. + * testsuite/libffi.special/unwindtest.cc: Convert the mmap to + the right type. + +2003-11-21 Andreas Jaeger + Andreas Tobler + + * acinclude.m4: Add AC_FUNC_MMAP_BLACKLIST. + * configure.in: Call AC_FUNC_MMAP_BLACKLIST. + * Makefile.in: Rebuilt. + * aclocal.m4: Likewise. + * configure: Likewise. + * fficonfig.h.in: Likewise. + * testsuite/lib/libffi-dg.exp: Add include dir. + * testsuite/libffi.call/ffitest.h: Add MMAP definitions. + * testsuite/libffi.special/ffitestcxx.h: Likewise. + * testsuite/libffi.call/closure_fn0.c: Use MMAP functionality + for ffi_closure if available. + * testsuite/libffi.call/closure_fn1.c: Likewise. + * testsuite/libffi.call/closure_fn2.c: Likewise. + * testsuite/libffi.call/closure_fn3.c: Likewise. + * testsuite/libffi.call/closure_fn4.c: Likewise. + * testsuite/libffi.call/closure_fn5.c: Likewise. + * testsuite/libffi.call/cls_12byte.c: Likewise. + * testsuite/libffi.call/cls_16byte.c: Likewise. + * testsuite/libffi.call/cls_18byte.c: Likewise. + * testsuite/libffi.call/cls_19byte.c: Likewise. + * testsuite/libffi.call/cls_1_1byte.c: Likewise. + * testsuite/libffi.call/cls_20byte.c: Likewise. + * testsuite/libffi.call/cls_20byte1.c: Likewise. + * testsuite/libffi.call/cls_24byte.c: Likewise. + * testsuite/libffi.call/cls_2byte.c: Likewise. + * testsuite/libffi.call/cls_3_1byte.c: Likewise. + * testsuite/libffi.call/cls_3byte1.c: Likewise. + * testsuite/libffi.call/cls_3byte2.c: Likewise. + * testsuite/libffi.call/cls_4_1byte.c: Likewise. + * testsuite/libffi.call/cls_4byte.c: Likewise. + * testsuite/libffi.call/cls_5byte.c: Likewise. + * testsuite/libffi.call/cls_64byte.c: Likewise. + * testsuite/libffi.call/cls_6byte.c: Likewise. + * testsuite/libffi.call/cls_7byte.c: Likewise. + * testsuite/libffi.call/cls_8byte.c: Likewise. + * testsuite/libffi.call/cls_9byte1.c: Likewise. + * testsuite/libffi.call/cls_9byte2.c: Likewise. + * testsuite/libffi.call/cls_double.c: Likewise. + * testsuite/libffi.call/cls_float.c: Likewise. + * testsuite/libffi.call/cls_schar.c: Likewise. + * testsuite/libffi.call/cls_sint.c: Likewise. + * testsuite/libffi.call/cls_sshort.c: Likewise. + * testsuite/libffi.call/cls_uchar.c: Likewise. + * testsuite/libffi.call/cls_uint.c: Likewise. + * testsuite/libffi.call/cls_ulonglong.c: Likewise. + * testsuite/libffi.call/cls_ushort.c: Likewise. + * testsuite/libffi.call/nested_struct.c: Likewise. + * testsuite/libffi.call/nested_struct1.c: Likewise. + * testsuite/libffi.call/nested_struct2.c: Likewise. + * testsuite/libffi.call/nested_struct3.c: Likewise. + * testsuite/libffi.call/problem1.c: Likewise. + * testsuite/libffi.special/unwindtest.cc: Likewise. + +2003-11-20 Andreas Tobler + + * testsuite/lib/libffi-dg.exp: Make the -lgcc_s conditional. + +2003-11-19 Andreas Tobler + + * testsuite/lib/libffi-dg.exp: Add DYLD_LIBRARY_PATH for darwin. + Add -lgcc_s to additional flags. + +2003-11-12 Andreas Tobler + + * configure.in, include/Makefile.am: PR libgcj/11147, install + the ffitarget.h header file in a gcc versioned and target + dependent place. + * configure: Regenerated. + * Makefile.in, include/Makefile.in: Likewise. + * testsuite/Makefile.in: Likewise. + +2003-11-09 Andreas Tobler + + * testsuite/libffi.call/closure_fn0.c: Print result and check + with dg-output to make debugging easier. + * testsuite/libffi.call/closure_fn1.c: Likewise. + * testsuite/libffi.call/closure_fn2.c: Likewise. + * testsuite/libffi.call/closure_fn3.c: Likewise. + * testsuite/libffi.call/closure_fn4.c: Likewise. + * testsuite/libffi.call/closure_fn5.c: Likewise. + * testsuite/libffi.call/cls_12byte.c: Likewise. + * testsuite/libffi.call/cls_16byte.c: Likewise. + * testsuite/libffi.call/cls_18byte.c: Likewise. + * testsuite/libffi.call/cls_19byte.c: Likewise. + * testsuite/libffi.call/cls_1_1byte.c: Likewise. + * testsuite/libffi.call/cls_20byte.c: Likewise. + * testsuite/libffi.call/cls_20byte1.c: Likewise. + * testsuite/libffi.call/cls_24byte.c: Likewise. + * testsuite/libffi.call/cls_2byte.c: Likewise. + * testsuite/libffi.call/cls_3_1byte.c: Likewise. + * testsuite/libffi.call/cls_3byte1.c: Likewise. + * testsuite/libffi.call/cls_3byte2.c: Likewise. + * testsuite/libffi.call/cls_4_1byte.c: Likewise. + * testsuite/libffi.call/cls_4byte.c: Likewise. + * testsuite/libffi.call/cls_5byte.c: Likewise. + * testsuite/libffi.call/cls_64byte.c: Likewise. + * testsuite/libffi.call/cls_6byte.c: Likewise. + * testsuite/libffi.call/cls_7byte.c: Likewise. + * testsuite/libffi.call/cls_8byte.c: Likewise. + * testsuite/libffi.call/cls_9byte1.c: Likewise. + * testsuite/libffi.call/cls_9byte2.c: Likewise. + * testsuite/libffi.call/cls_double.c: Likewise. + * testsuite/libffi.call/cls_float.c: Likewise. + * testsuite/libffi.call/cls_schar.c: Likewise. + * testsuite/libffi.call/cls_sint.c: Likewise. + * testsuite/libffi.call/cls_sshort.c: Likewise. + * testsuite/libffi.call/cls_uchar.c: Likewise. + * testsuite/libffi.call/cls_uint.c: Likewise. + * testsuite/libffi.call/cls_ulonglong.c: Likewise. + * testsuite/libffi.call/cls_ushort.c: Likewise. + * testsuite/libffi.call/problem1.c: Likewise. + + * testsuite/libffi.special/unwindtest.cc: Make ffi_closure + static. + +2003-11-08 Andreas Tobler + + * testsuite/libffi.call/cls_9byte2.c: New test case. + * testsuite/libffi.call/cls_9byte1.c: Likewise. + * testsuite/libffi.call/cls_64byte.c: Likewise. + * testsuite/libffi.call/cls_20byte1.c: Likewise. + * testsuite/libffi.call/cls_19byte.c: Likewise. + * testsuite/libffi.call/cls_18byte.c: Likewise. + * testsuite/libffi.call/closure_fn4.c: Likewise. + * testsuite/libffi.call/closure_fn5.c: Likewise. + * testsuite/libffi.call/cls_schar.c: Likewise. + * testsuite/libffi.call/cls_sint.c: Likewise. + * testsuite/libffi.call/cls_sshort.c: Likewise. + * testsuite/libffi.call/nested_struct2.c: Likewise. + * testsuite/libffi.call/nested_struct3.c: Likewise. + +2003-11-08 Andreas Tobler + + * testsuite/libffi.call/cls_double.c: Do a check on the result. + * testsuite/libffi.call/cls_uchar.c: Likewise. + * testsuite/libffi.call/cls_uint.c: Likewise. + * testsuite/libffi.call/cls_ulonglong.c: Likewise. + * testsuite/libffi.call/cls_ushort.c: Likewise. + * testsuite/libffi.call/return_sc.c: Cleanup whitespaces. + +2003-11-06 Andreas Tobler + + * src/prep_cif.c (ffi_prep_cif): Move the validity check after + the initialization. + +2003-10-23 Andreas Tobler + + * src/java_raw_api.c (ffi_java_ptrarray_to_raw): Replace + FFI_ASSERT(FALSE) with FFI_ASSERT(0). + +2003-10-22 David Daney + + * src/mips/ffitarget.h: Replace undefined UINT32 and friends with + __attribute__((__mode__(__SI__))) and friends. + +2003-10-22 Andreas Schwab + + * src/ia64/ffi.c: Replace FALSE/TRUE with false/true. + +2003-10-21 Andreas Tobler + + * configure.in: AC_LINK_FILES(ffitarget.h). + * configure: Regenerate. + * Makefile.in: Likewise. + * include/Makefile.in: Likewise. + * testsuite/Makefile.in: Likewise. + * fficonfig.h.in: Likewise. + +2003-10-21 Paolo Bonzini + Richard Henderson + + Avoid that ffi.h includes fficonfig.h. + + * Makefile.am (EXTRA_DIST): Include ffitarget.h files + (TARGET_SRC_MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX. + (TARGET_SRC_MIPS_SGI): Removed. + (MIPS_GCC): Renamed to TARGET_SRC_MIPS_IRIX. + (MIPS_SGI): Removed. + (CLEANFILES): Removed. + (mostlyclean-am, clean-am, mostlyclean-sub, clean-sub): New + targets. + * acconfig.h: Removed. + * configure.in: Compute sizeofs only for double and long double. + Use them to define and subst HAVE_LONG_DOUBLE. Include comments + into AC_DEFINE instead of using acconfig.h. Create + include/ffitarget.h instead of include/fficonfig.h. Rename + MIPS_GCC to MIPS_IRIX, drop MIPS_SGI since we are in gcc's tree. + AC_DEFINE EH_FRAME_FLAGS. + * include/Makefile.am (DISTCLEANFILES): New automake macro. + (hack_DATA): Add ffitarget.h. + * include/ffi.h.in: Remove all system specific definitions. + Declare raw API even if it is not installed, why bother? + Use limits.h instead of SIZEOF_* to define ffi_type_*. Do + not define EH_FRAME_FLAGS, it is in fficonfig.h now. Include + ffitarget.h instead of fficonfig.h. Remove ALIGN macro. + (UINT_ARG, INT_ARG): Removed, use ffi_arg and ffi_sarg instead. + * include/ffi_common.h (bool): Do not define. + (ffi_assert): Accept failed assertion. + (ffi_type_test): Return void and accept file/line. + (FFI_ASSERT): Pass stringized failed assertion. + (FFI_ASSERT_AT): New macro. + (FFI_ASSERT_VALID_TYPE): New macro. + (UINT8, SINT8, UINT16, SINT16, UINT32, SINT32, + UINT64, SINT64): Define here with gcc's __attribute__ macro + instead of in ffi.h + (FLOAT32, ALIGN): Define here instead of in ffi.h + * include/ffi-mips.h: Removed. Its content moved to + src/mips/ffitarget.h after separating assembly and C sections. + * src/alpha/ffi.c, src/alpha/ffi.c, src/java_raw_api.c + src/prep_cif.c, src/raw_api.c, src/ia64/ffi.c, + src/mips/ffi.c, src/mips/n32.S, src/mips/o32.S, + src/mips/ffitarget.h, src/sparc/ffi.c, src/x86/ffi64.c: + SIZEOF_ARG -> FFI_SIZEOF_ARG. + * src/ia64/ffi.c: Include stdbool.h (provided by GCC 2.95+). + * src/debug.c (ffi_assert): Accept stringized failed assertion. + (ffi_type_test): Rewritten. + * src/prep-cif.c (initialize_aggregate, ffi_prep_cif): Call + FFI_ASSERT_VALID_TYPE. + * src/alpha/ffitarget.h, src/arm/ffitarget.h, + src/ia64/ffitarget.h, src/m68k/ffitarget.h, + src/mips/ffitarget.h, src/powerpc/ffitarget.h, + src/s390/ffitarget.h, src/sh/ffitarget.h, + src/sh64/ffitarget.h, src/sparc/ffitarget.h, + src/x86/ffitarget.h: New files. + * src/alpha/osf.S, src/arm/sysv.S, src/ia64/unix.S, + src/m68k/sysv.S, src/mips/n32.S, src/mips/o32.S, + src/powerpc/aix.S, src/powerpc/darwin.S, + src/powerpc/ffi_darwin.c, src/powerpc/linux64.S, + src/powerpc/linux64_closure.S, src/powerpc/ppc_closure.S, + src/powerpc/sysv.S, src/s390/sysv.S, src/sh/sysv.S, + src/sh64/sysv.S, src/sparc/v8.S, src/sparc/v9.S, + src/x86/sysv.S, src/x86/unix64.S, src/x86/win32.S: + include fficonfig.h + +2003-10-20 Rainer Orth + + * src/mips/ffi.c: Use _ABIN32, _ABIO32 instead of external + _MIPS_SIM_NABI32, _MIPS_SIM_ABI32. + +2003-10-19 Andreas Tobler + + * src/powerpc/ffi_darwin.c (ffi_prep_args): Declare bytes again. + Used when FFI_DEBUG = 1. + +2003-10-14 Alan Modra + + * src/types.c (double, longdouble): Default POWERPC64 to 8 byte size + and align. + +2003-10-06 Rainer Orth + + * include/ffi_mips.h: Define FFI_MIPS_N32 for N32/N64 ABIs, + FFI_MIPS_O32 for O32 ABI. + +2003-10-01 Andreas Tobler + + * testsuite/lib/libffi-dg.exp: Set LD_LIBRARY_PATH_64 for + SPARC64. Cleanup whitespaces. + +2003-09-19 Andreas Tobler + + * testsuite/libffi.call/closure_fn0.c: Xfail mips, arm, + strongarm, xscale. Cleanup whitespaces. + * testsuite/libffi.call/closure_fn1.c: Likewise. + * testsuite/libffi.call/closure_fn2.c: Likewise. + * testsuite/libffi.call/closure_fn3.c: Likewise. + * testsuite/libffi.call/cls_12byte.c: Likewise. + * testsuite/libffi.call/cls_16byte.c: Likewise. + * testsuite/libffi.call/cls_1_1byte.c: Likewise. + * testsuite/libffi.call/cls_20byte.c: Likewise. + * testsuite/libffi.call/cls_24byte.c: Likewise. + * testsuite/libffi.call/cls_2byte.c: Likewise. + * testsuite/libffi.call/cls_3_1byte.c: Likewise. + * testsuite/libffi.call/cls_3byte1.c: Likewise. + * testsuite/libffi.call/cls_3byte2.c: Likewise. + * testsuite/libffi.call/cls_4_1byte.c: Likewise. + * testsuite/libffi.call/cls_4byte.c: Likewise. + * testsuite/libffi.call/cls_5byte.c: Likewise. + * testsuite/libffi.call/cls_6byte.c: Likewise. + * testsuite/libffi.call/cls_7byte.c: Likewise. + * testsuite/libffi.call/cls_8byte.c: Likewise. + * testsuite/libffi.call/cls_double.c: Likewise. + * testsuite/libffi.call/cls_float.c: Likewise. + * testsuite/libffi.call/cls_uchar.c: Likewise. + * testsuite/libffi.call/cls_uint.c: Likewise. + * testsuite/libffi.call/cls_ulonglong.c: Likewise. + * testsuite/libffi.call/cls_ushort.c: Likewise. + * testsuite/libffi.call/nested_struct.c: Likewise. + * testsuite/libffi.call/nested_struct1.c: Likewise. + * testsuite/libffi.call/problem1.c: Likewise. + * testsuite/libffi.special/unwindtest.cc: Likewise. + * testsuite/libffi.call/pyobjc-tc.c: Cleanup whitespaces. + +2003-09-18 David Edelsohn + + * src/powerpc/aix.S: Cleanup whitespaces. + * src/powerpc/aix_closure.S: Likewise. + +2003-09-18 Andreas Tobler + + * src/powerpc/darwin.S: Cleanup whitespaces, comment formatting. + * src/powerpc/darwin_closure.S: Likewise. + * src/powerpc/ffi_darwin.c: Likewise. + +2003-09-18 Andreas Tobler + David Edelsohn + + * src/types.c (double): Add AIX and Darwin to the right TYPEDEF. + * src/powerpc/aix_closure.S: Remove the pointer to the outgoing + parameter stack. + * src/powerpc/darwin_closure.S: Likewise. + * src/powerpc/ffi_darwin.c (ffi_prep_args): Handle structures + according to the Darwin/AIX ABI. + (ffi_prep_cif_machdep): Likewise. + (ffi_closure_helper_DARWIN): Likewise. + Remove the outgoing parameter stack logic. Simplify the evaluation + of the different CASE types. + (ffi_prep_clousure): Avoid the casts on lvalues. Change the branch + statement in the trampoline code. + +2003-09-18 Kaz Kojima + + * src/sh/ffi.c (ffi_prep_args): Take account into the alignement + for the register size. + (ffi_closure_helper_SYSV): Handle the structure return value + address correctly. + (ffi_closure_helper_SYSV): Return the appropriate type when + the registers are used for the structure return value. + * src/sh/sysv.S (ffi_closure_SYSV): Fix the stack layout for + the 64-bit return value. Update copyright years. + +2003-09-17 Rainer Orth + + * testsuite/lib/libffi-dg.exp (libffi_target_compile): Search in + srcdir for ffi_mips.h. + +2003-09-12 Alan Modra + + * src/prep_cif.c (initialize_aggregate): Include tail padding in + structure size. + * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Correct + placement of float result. + * testsuite/libffi.special/unwindtest.cc (closure_test_fn1): Correct + cast of "resp" for big-endian 64 bit machines. + +2003-09-11 Alan Modra + + * src/types.c (double, longdouble): Merge identical SH and ARM + typedefs, and add POWERPC64. + * src/powerpc/ffi.c (ffi_prep_args64): Correct next_arg calc for + struct split over gpr and rest. + (ffi_prep_cif_machdep): Correct intarg_count for structures. + * src/powerpc/linux64.S (ffi_call_LINUX64): Fix gpr offsets. + +2003-09-09 Andreas Tobler + + * src/powerpc/ffi.c (ffi_closure_helper_SYSV) Handle struct + passing correctly. + +2003-09-09 Alan Modra + + * configure: Regenerate. + +2003-09-04 Andreas Tobler + + * Makefile.am: Remove build rules for ffitest. + * Makefile.in: Rebuilt. + +2003-09-04 Andreas Tobler + + * src/java_raw_api.c: Include to fix compiler warning + about implicit declaration of abort(). + +2003-09-04 Andreas Tobler + + * Makefile.am: Add dejagnu test framework. Fixes PR other/11411. + * Makefile.in: Rebuilt. + * configure.in: Add dejagnu test framework. + * configure: Rebuilt. + + * testsuite/Makefile.am: New file. + * testsuite/Makefile.in: Built + * testsuite/lib/libffi-dg.exp: New file. + * testsuite/config/default.exp: Likewise. + * testsuite/libffi.call/call.exp: Likewise. + * testsuite/libffi.call/ffitest.h: Likewise. + * testsuite/libffi.call/closure_fn0.c: Likewise. + * testsuite/libffi.call/closure_fn1.c: Likewise. + * testsuite/libffi.call/closure_fn2.c: Likewise. + * testsuite/libffi.call/closure_fn3.c: Likewise. + * testsuite/libffi.call/cls_1_1byte.c: Likewise. + * testsuite/libffi.call/cls_3_1byte.c: Likewise. + * testsuite/libffi.call/cls_4_1byte.c: Likewise. + * testsuite/libffi.call/cls_2byte.c: Likewise. + * testsuite/libffi.call/cls_3byte1.c: Likewise. + * testsuite/libffi.call/cls_3byte2.c: Likewise. + * testsuite/libffi.call/cls_4byte.c: Likewise. + * testsuite/libffi.call/cls_5byte.c: Likewise. + * testsuite/libffi.call/cls_6byte.c: Likewise. + * testsuite/libffi.call/cls_7byte.c: Likewise. + * testsuite/libffi.call/cls_8byte.c: Likewise. + * testsuite/libffi.call/cls_12byte.c: Likewise. + * testsuite/libffi.call/cls_16byte.c: Likewise. + * testsuite/libffi.call/cls_20byte.c: Likewise. + * testsuite/libffi.call/cls_24byte.c: Likewise. + * testsuite/libffi.call/cls_double.c: Likewise. + * testsuite/libffi.call/cls_float.c: Likewise. + * testsuite/libffi.call/cls_uchar.c: Likewise. + * testsuite/libffi.call/cls_uint.c: Likewise. + * testsuite/libffi.call/cls_ulonglong.c: Likewise. + * testsuite/libffi.call/cls_ushort.c: Likewise. + * testsuite/libffi.call/float.c: Likewise. + * testsuite/libffi.call/float1.c: Likewise. + * testsuite/libffi.call/float2.c: Likewise. + * testsuite/libffi.call/many.c: Likewise. + * testsuite/libffi.call/many_win32.c: Likewise. + * testsuite/libffi.call/nested_struct.c: Likewise. + * testsuite/libffi.call/nested_struct1.c: Likewise. + * testsuite/libffi.call/pyobjc-tc.c: Likewise. + * testsuite/libffi.call/problem1.c: Likewise. + * testsuite/libffi.call/promotion.c: Likewise. + * testsuite/libffi.call/return_ll.c: Likewise. + * testsuite/libffi.call/return_sc.c: Likewise. + * testsuite/libffi.call/return_uc.c: Likewise. + * testsuite/libffi.call/strlen.c: Likewise. + * testsuite/libffi.call/strlen_win32.c: Likewise. + * testsuite/libffi.call/struct1.c: Likewise. + * testsuite/libffi.call/struct2.c: Likewise. + * testsuite/libffi.call/struct3.c: Likewise. + * testsuite/libffi.call/struct4.c: Likewise. + * testsuite/libffi.call/struct5.c: Likewise. + * testsuite/libffi.call/struct6.c: Likewise. + * testsuite/libffi.call/struct7.c: Likewise. + * testsuite/libffi.call/struct8.c: Likewise. + * testsuite/libffi.call/struct9.c: Likewise. + * testsuite/libffi.special/special.exp: New file. + * testsuite/libffi.special/ffitestcxx.h: Likewise. + * testsuite/libffi.special/unwindtest.cc: Likewise. + + +2003-08-13 Kaz Kojima + + * src/sh/ffi.c (OFS_INT16): Set 0 for little endian case. Update + copyright years. + +2003-08-02 Alan Modra + + * src/powerpc/ffi.c (ffi_prep_args64): Modify for changed gcc + structure passing. + (ffi_closure_helper_LINUX64): Likewise. + * src/powerpc/linux64.S: Remove code writing to parm save area. + * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Use return + address in lr from ffi_closure_helper_LINUX64 call to calculate + table address. Optimize function tail. + +2003-07-28 Andreas Tobler + + * src/sparc/ffi.c: Handle all floating point registers. + * src/sparc/v9.S: Likewise. Fixes second part of PR target/11410. + +2003-07-11 Gerald Pfeifer + + * README: Note that libffi is not part of GCC. Update the project + URL and status. + +2003-06-19 Franz Sirl + + * src/powerpc/ppc_closure.S: Include ffi.h. + +2003-06-13 Rainer Orth + + * src/x86/sysv.S: Avoid gas-only .uleb128/.sleb128 directives. + Use C style comments. + +2003-06-13 Kaz Kojima + + * Makefile.am: Add SHmedia support. Fix a typo of SH support. + * Makefile.in: Regenerate. + * configure.in (sh64-*-linux*, sh5*-*-linux*): Add target. + * configure: Regenerate. + * include/ffi.h.in: Add SHmedia support. + * src/sh64/ffi.c: New file. + * src/sh64/sysv.S: New file. + +2003-05-16 Jakub Jelinek + + * configure.in (HAVE_RO_EH_FRAME): Check whether .eh_frame section + should be read-only. + * configure: Rebuilt. + * fficonfig.h.in: Rebuilt. + * include/ffi.h.in (EH_FRAME_FLAGS): Define. + * src/alpha/osf.S: Use EH_FRAME_FLAGS. + * src/powerpc/linux64.S: Likewise. + * src/powerpc/linux64_closure.S: Likewise. Include ffi.h. + * src/powerpc/sysv.S: Use EH_FRAME_FLAGS. Use pcrel encoding + if -fpic/-fPIC/-mrelocatable. + * src/powerpc/powerpc_closure.S: Likewise. + * src/sparc/v8.S: If HAVE_RO_EH_FRAME is defined, don't include + #write in .eh_frame flags. + * src/sparc/v9.S: Likewise. + * src/x86/unix64.S: Use EH_FRAME_FLAGS. + * src/x86/sysv.S: Likewise. Use pcrel encoding if -fpic/-fPIC. + * src/s390/sysv.S: Use EH_FRAME_FLAGS. Include ffi.h. + +2003-05-07 Jeff Sturm + + Fixes PR bootstrap/10656 + * configure.in (HAVE_AS_REGISTER_PSEUDO_OP): Test assembler + support for .register pseudo-op. + * src/sparc/v8.S: Use it. + * fficonfig.h.in: Rebuilt. + * configure: Rebuilt. + +2003-04-18 Jakub Jelinek + + * include/ffi.h.in (POWERPC64): Define if 64-bit. + (enum ffi_abi): Add FFI_LINUX64 on POWERPC. + Make it the default on POWERPC64. + (FFI_TRAMPOLINE_SIZE): Define to 24 on POWERPC64. + * configure.in: Change powerpc-*-linux* into powerpc*-*-linux*. + * configure: Rebuilt. + * src/powerpc/ffi.c (hidden): Define. + (ffi_prep_args_SYSV): Renamed from + ffi_prep_args. Cast pointers to unsigned long to shut up warnings. + (NUM_GPR_ARG_REGISTERS64, NUM_FPR_ARG_REGISTERS64, + ASM_NEEDS_REGISTERS64): New. + (ffi_prep_args64): New function. + (ffi_prep_cif_machdep): Handle FFI_LINUX64 ABI. + (ffi_call): Likewise. + (ffi_prep_closure): Likewise. + (flush_icache): Surround by #ifndef POWERPC64. + (ffi_dblfl): New union type. + (ffi_closure_helper_SYSV): Use it to avoid aliasing problems. + (ffi_closure_helper_LINUX64): New function. + * src/powerpc/ppc_closure.S: Surround whole file by #ifndef + __powerpc64__. + * src/powerpc/sysv.S: Likewise. + (ffi_call_SYSV): Rename ffi_prep_args to ffi_prep_args_SYSV. + * src/powerpc/linux64.S: New file. + * src/powerpc/linux64_closure.S: New file. + * Makefile.am (EXTRA_DIST): Add src/powerpc/linux64.S and + src/powerpc/linux64_closure.S. + (TARGET_SRC_POWERPC): Likewise. + + * src/ffitest.c (closure_test_fn, closure_test_fn1, closure_test_fn2, + closure_test_fn3): Fix result printing on big-endian 64-bit + machines. + (main): Print tst2_arg instead of uninitialized tst2_result. + + * src/ffitest.c (main): Hide what closure pointer really points to + from the compiler. + +2003-04-16 Richard Earnshaw + + * configure.in (arm-*-netbsdelf*): Add configuration. + (configure): Regenerated. + +2003-04-04 Loren J. Rittle + + * include/Makefile.in: Regenerate. + +2003-03-21 Zdenek Dvorak + + * libffi/include/ffi.h.in: Define X86 instead of X86_64 in 32 + bit mode. + * libffi/src/x86/ffi.c (ffi_closure_SYSV, ffi_closure_raw_SYSV): + Receive closure pointer through parameter, read args using + __builtin_dwarf_cfa. + (FFI_INIT_TRAMPOLINE): Send closure reference through eax. + +2003-03-12 Andreas Schwab + + * configure.in: Avoid trailing /. in toolexeclibdir. + * configure: Rebuilt. + +2003-03-03 Andreas Tobler + + * src/powerpc/darwin_closure.S: Recode to fit dynamic libraries. + +2003-02-06 Andreas Tobler + + * libffi/src/powerpc/darwin_closure.S: + Fix alignement bug, allocate 8 bytes for the result. + * libffi/src/powerpc/aix_closure.S: + Likewise. + * libffi/src/powerpc/ffi_darwin.c: + Update stackframe description for aix/darwin_closure.S. + +2003-02-06 Jakub Jelinek + + * src/s390/ffi.c (ffi_closure_helper_SYSV): Add hidden visibility + attribute. + +2003-01-31 Christian Cornelssen , + Andreas Schwab + + * configure.in: Adjust command to source config-ml.in to account + for changes to the libffi_basedir definition. + (libffi_basedir): Remove ${srcdir} from value and include trailing + slash if nonempty. + + * configure: Regenerate. + +2003-01-29 Franz Sirl + + * src/powerpc/ppc_closure.S: Recode to fit shared libs. + +2003-01-28 Andrew Haley + + * include/ffi.h.in: Enable FFI_CLOSURES for x86_64. + * src/x86/ffi64.c (ffi_prep_closure): New. + (ffi_closure_UNIX64_inner): New. + * src/x86/unix64.S (ffi_closure_UNIX64): New. + +2003-01-27 Alexandre Oliva + + * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST. + Remove USE_LIBDIR conditional. + * Makefile.am (toolexecdir, toolexeclibdir): Don't override. + * Makefile.in, configure: Rebuilt. + +2003-01027 David Edelsohn + + * Makefile.am (TARGET_SRC_POWERPC_AIX): Fix typo. + * Makefile.in: Regenerate. + +2003-01-22 Andrew Haley + + * src/powerpc/darwin.S (_ffi_call_AIX): Add Augmentation size to + unwind info. + +2003-01-21 Andreas Tobler + + * src/powerpc/darwin.S: Add unwind info. + * src/powerpc/darwin_closure.S: Likewise. + +2003-01-14 Andrew Haley + + * src/x86/ffi64.c (ffi_prep_args): Check for void retval. + (ffi_prep_cif_machdep): Likewise. + * src/x86/unix64.S: Add unwind info. + +2003-01-14 Andreas Jaeger + + * src/ffitest.c (main): Only use ffi_closures if those are + supported. + +2003-01-13 Andreas Tobler + + * libffi/src/ffitest.c + add closure testcases + +2003-01-13 Kevin B. Hendricks + + * libffi/src/powerpc/ffi.c + fix alignment bug for float (4 byte aligned iso 8 byte) + +2003-01-09 Geoffrey Keating + + * src/powerpc/ffi_darwin.c: Remove RCS version string. + * src/powerpc/darwin.S: Remove RCS version string. + +2003-01-03 Jeff Sturm + + * include/ffi.h.in: Add closure defines for SPARC, SPARC64. + * src/ffitest.c (main): Use static storage for closure. + * src/sparc/ffi.c (ffi_prep_closure, ffi_closure_sparc_inner): New. + * src/sparc/v8.S (ffi_closure_v8): New. + * src/sparc/v9.S (ffi_closure_v9): New. + +2002-11-10 Ranjit Mathew + + * include/ffi.h.in: Added FFI_STDCALL ffi_type + enumeration for X86_WIN32. + * src/x86/win32.S: Added ffi_call_STDCALL function + definition. + * src/x86/ffi.c (ffi_call/ffi_raw_call): Added + switch cases for recognising FFI_STDCALL and + calling ffi_call_STDCALL if target is X86_WIN32. + * src/ffitest.c (my_stdcall_strlen/stdcall_many): + stdcall versions of the "my_strlen" and "many" + test functions (for X86_WIN32). + Added test cases to test stdcall invocation using + these functions. + +2002-12-02 Kaz Kojima + + * src/sh/sysv.S: Add DWARF2 unwind info. + +2002-11-27 Ulrich Weigand + + * src/s390/sysv.S (.eh_frame section): Make section read-only. + +2002-11-26 Jim Wilson + + * src/types.c (FFI_TYPE_POINTER): Has size 8 on IA64. + +2002-11-23 H.J. Lu + + * acinclude.m4: Add dummy AM_PROG_LIBTOOL. + Include ../config/accross.m4. + * aclocal.m4; Rebuild. + * configure: Likewise. + +2002-11-15 Ulrich Weigand + + * src/s390/sysv.S (.eh_frame section): Adapt to pcrel FDE encoding. + +2002-11-11 DJ Delorie + + * configure.in: Look for common files in the right place. + +2002-10-08 Ulrich Weigand + + * src/java_raw_api.c (ffi_java_raw_to_ptrarray): Interpret + raw data as _Jv_word values, not ffi_raw. + (ffi_java_ptrarray_to_raw): Likewise. + (ffi_java_rvalue_to_raw): New function. + (ffi_java_raw_call): Call it. + (ffi_java_raw_to_rvalue): New function. + (ffi_java_translate_args): Call it. + * src/ffitest.c (closure_test_fn): Interpret return value + as ffi_arg, not int. + * src/s390/ffi.c (ffi_prep_cif_machdep): Add missing + FFI_TYPE_POINTER case. + (ffi_closure_helper_SYSV): Likewise. Also, assume return + values extended to word size. + +2002-10-02 Andreas Jaeger + + * src/x86/ffi64.c (ffi_prep_cif_machdep): Remove debug output. + +2002-10-01 Bo Thorsen + + * include/ffi.h.in: Fix i386 win32 compilation. + +2002-09-30 Ulrich Weigand + + * configure.in: Add s390x-*-linux-* target. + * configure: Regenerate. + * include/ffi.h.in: Define S390X for s390x targets. + (FFI_CLOSURES): Define for s390/s390x. + (FFI_TRAMPOLINE_SIZE): Likewise. + (FFI_NATIVE_RAW_API): Likewise. + * src/prep_cif.c (ffi_prep_cif): Do not compute stack space for s390. + * src/types.c (FFI_TYPE_POINTER): Use 8-byte pointers on s390x. + * src/s390/ffi.c: Major rework of existing code. Add support for + s390x targets. Add closure support. + * src/s390/sysv.S: Likewise. + +2002-09-29 Richard Earnshaw + + * src/arm/sysv.S: Fix typo. + +2002-09-28 Richard Earnshaw + + * src/arm/sysv.S: If we don't have machine/asm.h and the pre-processor + has defined __USER_LABEL_PREFIX__, then use it in CNAME. + (ffi_call_SYSV): Handle soft-float. + +2002-09-27 Bo Thorsen + + * include/ffi.h.in: Fix multilib x86-64 support. + +2002-09-22 Kaveh R. Ghazi + + * Makefile.am (all-multi): Fix multilib parallel build. + +2002-07-19 Kaz Kojima + + * configure.in (sh[34]*-*-linux*): Add brackets. + * configure: Regenerate. + +2002-07-18 Kaz Kojima + + * Makefile.am: Add SH support. + * Makefile.in: Regenerate. + * configure.in (sh-*-linux*, sh[34]*-*-linux*): Add target. + * configure: Regenerate. + * include/ffi.h.in: Add SH support. + * src/sh/ffi.c: New file. + * src/sh/sysv.S: New file. + * src/types.c: Add SH support. + +2002-07-16 Bo Thorsen + + * src/x86/ffi64.c: New file that adds x86-64 support. + * src/x86/unix64.S: New file that handles argument setup for + x86-64. + * src/x86/sysv.S: Don't use this on x86-64. + * src/x86/ffi.c: Don't use this on x86-64. + Remove unused vars. + * src/prep_cif.c (ffi_prep_cif): Don't do stack size calculation + for x86-64. + * src/ffitest.c (struct6): New test that tests a special case in + the x86-64 ABI. + (struct7): Likewise. + (struct8): Likewise. + (struct9): Likewise. + (closure_test_fn): Silence warning about this when it's not used. + (main): Add the new tests. + (main): Fix a couple of wrong casts and silence some compiler warnings. + * include/ffi.h.in: Add x86-64 ABI definition. + * fficonfig.h.in: Regenerate. + * Makefile.am: Add x86-64 support. + * configure.in: Likewise. + * Makefile.in: Regenerate. + * configure: Likewise. + +2002-06-24 Bo Thorsen + + * src/types.c: Merge settings for similar architectures. + Add x86-64 sizes and alignments. + +2002-06-23 Bo Thorsen + + * src/arm/ffi.c (ffi_prep_args): Remove unused vars. + * src/sparc/ffi.c (ffi_prep_args_v8): Likewise. + * src/mips/ffi.c (ffi_prep_args): Likewise. + * src/m68k/ffi.c (ffi_prep_args): Likewise. + +2002-07-18 H.J. Lu (hjl@gnu.org) + + * Makefile.am (TARGET_SRC_MIPS_LINUX): New. + (libffi_la_SOURCES): Support MIPS_LINUX. + (libffi_convenience_la_SOURCES): Likewise. + * Makefile.in: Regenerated. + + * configure.in (mips64*-*): Skip. + (mips*-*-linux*): New. + * configure: Regenerated. + + * src/mips/ffi.c: Include . + +2002-06-06 Ulrich Weigand + + * src/s390/sysv.S: Save/restore %r6. Add DWARF-2 unwind info. + +2002-05-27 Roger Sayle + + * src/x86/ffi.c (ffi_prep_args): Remove reference to avn. + +2002-05-27 Bo Thorsen + + * src/x86/ffi.c (ffi_prep_args): Remove unused variable and + fix formatting. + +2002-05-13 Andreas Tobler + + * src/powerpc/ffi_darwin.c (ffi_prep_closure): Declare fd at + beginning of function (for older apple cc). + +2002-05-08 Alexandre Oliva + + * configure.in (ORIGINAL_LD_FOR_MULTILIBS): Preserve LD at + script entry, and set LD to it when configuring multilibs. + * configure: Rebuilt. + +2002-05-05 Jason Thorpe + + * configure.in (sparc64-*-netbsd*): Add target. + (sparc-*-netbsdelf*): Likewise. + * configure: Regenerate. + +2002-04-28 David S. Miller + + * configure.in, configure: Fix SPARC test in previous change. + +2002-04-29 Gerhard Tonn + + * Makefile.am: Add Linux for S/390 support. + * Makefile.in: Regenerate. + * configure.in: Add Linux for S/390 support. + * configure: Regenerate. + * include/ffi.h.in: Add Linux for S/390 support. + * src/s390/ffi.c: New file from libffi CVS tree. + * src/s390/sysv.S: New file from libffi CVS tree. + +2002-04-28 Jakub Jelinek + + * configure.in (HAVE_AS_SPARC_UA_PCREL): Check for working + %r_disp32(). + * src/sparc/v8.S: Use it. + * src/sparc/v9.S: Likewise. + * fficonfig.h.in: Rebuilt. + * configure: Rebuilt. + +2002-04-08 Hans Boehm + + * src/java_raw_api.c (ffi_java_raw_size): Handle FFI_TYPE_DOUBLE + correctly. + * src/ia64/unix.S: Add unwind information. Fix comments. + Save sp in a way that's compatible with unwind info. + (ffi_call_unix): Correctly restore sp in all cases. + * src/ia64/ffi.c: Add, fix comments. + +2002-04-08 Jakub Jelinek + + * src/sparc/v8.S: Make .eh_frame dependent on target word size. + +2002-04-06 Jason Thorpe + + * configure.in (alpha*-*-netbsd*): Add target. + * configure: Regenerate. + +2002-04-04 Jeff Sturm + + * src/sparc/v8.S: Add unwind info. + * src/sparc/v9.S: Likewise. + +2002-03-30 Krister Walfridsson + + * configure.in: Enable i*86-*-netbsdelf*. + * configure: Rebuilt. + +2002-03-29 David Billinghurst + + PR other/2620 + * src/mips/n32.s: Delete + * src/mips/o32.s: Delete + +2002-03-21 Loren J. Rittle + + * configure.in: Enable alpha*-*-freebsd*. + * configure: Rebuilt. + +2002-03-17 Bryce McKinlay + + * Makefile.am: libfficonvenience -> libffi_convenience. + * Makefile.in: Rebuilt. + + * Makefile.am: Define ffitest_OBJECTS. + * Makefile.in: Rebuilt. + +2002-03-07 Andreas Tobler + David Edelsohn + + * Makefile.am (EXTRA_DIST): Add Darwin and AIX closure files. + (TARGET_SRC_POWERPC_AIX): Add aix_closure.S. + (TARGET_SRC_POWERPC_DARWIN): Add darwin_closure.S. + * Makefile.in: Regenerate. + * include/ffi.h.in: Add AIX and Darwin closure definitions. + * src/powerpc/ffi_darwin.c (ffi_prep_closure): New function. + (flush_icache, flush_range): New functions. + (ffi_closure_helper_DARWIN): New function. + * src/powerpc/aix_closure.S: New file. + * src/powerpc/darwin_closure.S: New file. + +2002-02-24 Jeff Sturm + + * include/ffi.h.in: Add typedef for ffi_arg. + * src/ffitest.c (main): Declare rint with ffi_arg. + +2002-02-21 Andreas Tobler + + * src/powerpc/ffi_darwin.c (ffi_prep_args): Skip appropriate + number of GPRs for floating-point arguments. + +2002-01-31 Anthony Green + + * configure: Rebuilt. + * configure.in: Replace CHECK_SIZEOF and endian tests with + cross-compiler friendly macros. + * aclocal.m4 (AC_COMPILE_CHECK_SIZEOF, AC_C_BIGENDIAN_CROSS): New + macros. + +2002-01-18 David Edelsohn + + * src/powerpc/darwin.S (_ffi_call_AIX): New. + * src/powerpc/aix.S (ffi_call_DARWIN): New. + +2002-01-17 David Edelsohn + + * Makefile.am (EXTRA_DIST): Add Darwin and AIX files. + (TARGET_SRC_POWERPC_AIX): New. + (POWERPC_AIX): New stanza. + * Makefile.in: Regenerate. + * configure.in: Add AIX case. + * configure: Regenerate. + * include/ffi.h.in (ffi_abi): Add FFI_AIX. + * src/powerpc/ffi_darwin.c (ffi_status): Use "long" to scale frame + size. Fix "long double" support. + (ffi_call): Add FFI_AIX case. + * src/powerpc/aix.S: New. + +2001-10-09 John Hornkvist + + Implement Darwin PowerPC ABI. + * configure.in: Handle powerpc-*-darwin*. + * Makefile.am: Set source files for POWERPC_DARWIN. + * configure: Rebuilt. + * Makefile.in: Rebuilt. + * include/ffi.h.in: Define FFI_DARWIN and FFI_DEFAULT_ABI for + POWERPC_DARWIN. + * src/powerpc/darwin.S: New file. + * src/powerpc/ffi_darwin.c: New file. + +2001-10-07 Joseph S. Myers + + * src/x86/ffi.c: Fix spelling error of "separate" as "seperate". + +2001-07-16 Rainer Orth + + * src/x86/sysv.S: Avoid gas-only .balign directive. + Use C style comments. + +2001-07-16 Rainer Orth + + * src/alpha/ffi.c (ffi_prep_closure): Avoid gas-only mnemonic. + Fixes PR bootstrap/3563. + +2001-06-26 Rainer Orth + + * src/alpha/osf.S (ffi_closure_osf): Use .rdata for ECOFF. + +2001-06-25 Rainer Orth + + * configure.in: Recognize sparc*-sun-* host. + * configure: Regenerate. + +2001-06-06 Andrew Haley + + * src/alpha/osf.S (__FRAME_BEGIN__): Conditionalize for ELF. + +2001-06-03 Andrew Haley + + * src/alpha/osf.S: Add unwind info. + * src/powerpc/sysv.S: Add unwind info. + * src/powerpc/ppc_closure.S: Likewise. + +2000-05-31 Jeff Sturm + + * configure.in: Fix AC_ARG_ENABLE usage. + * configure: Rebuilt. + +2001-05-06 Bryce McKinlay + + * configure.in: Remove warning about beta code. + * configure: Rebuilt. + +2001-04-25 Hans Boehm + + * src/ia64/unix.S: Restore stack pointer when returning from + ffi_closure_UNIX. + * src/ia64/ffi.c: Fix typo in comment. + +2001-04-18 Jim Wilson + + * src/ia64/unix.S: Delete unnecessary increment and decrement of loc2 + to eliminate RAW DV. + +2001-04-12 Bryce McKinlay + + * Makefile.am: Make a libtool convenience library. + * Makefile.in: Rebuilt. + +2001-03-29 Bryce McKinlay + + * configure.in: Use different syntax for subdirectory creation. + * configure: Rebuilt. + +2001-03-27 Jon Beniston + + * configure.in: Added X86_WIN32 target (Win32, CygWin, MingW). + * configure: Rebuilt. + * Makefile.am: Added X86_WIN32 target support. + * Makefile.in: Rebuilt. + + * include/ffi.h.in: Added X86_WIN32 target support. + + * src/ffitest.c: Doesn't run structure tests for X86_WIN32 targets. + * src/types.c: Added X86_WIN32 target support. + + * src/x86/win32.S: New file. Based on sysv.S, but with EH + stuff removed and made to work with CygWin's gas. + +2001-03-26 Bryce McKinlay + + * configure.in: Make target subdirectory in build dir. + * Makefile.am: Override suffix based rules to specify correct output + subdirectory. + * Makefile.in: Rebuilt. + * configure: Rebuilt. + +2001-03-23 Kevin B Hendricks + + * src/powerpc/ppc_closure.S: New file. + * src/powerpc/ffi.c (ffi_prep_args): Fixed ABI compatibility bug + involving long long and register pairs. + (ffi_prep_closure): New function. + (flush_icache): Likewise. + (ffi_closure_helper_SYSV): Likewise. + * include/ffi.h.in (FFI_CLOSURES): Define on PPC. + (FFI_TRAMPOLINE_SIZE): Likewise. + (FFI_NATIVE_RAW_API): Likewise. + * Makefile.in: Rebuilt. + * Makefile.am (EXTRA_DIST): Added src/powerpc/ppc_closure.S. + (TARGET_SRC_POWERPC): Likewise. + +2001-03-19 Tom Tromey + + * Makefile.in: Rebuilt. + * Makefile.am (ffitest_LDFLAGS): New macro. + +2001-03-02 Nick Clifton + + * include/ffi.h.in: Remove RCS ident string. + * include/ffi_mips.h: Remove RCS ident string. + * src/debug.c: Remove RCS ident string. + * src/ffitest.c: Remove RCS ident string. + * src/prep_cif.c: Remove RCS ident string. + * src/types.c: Remove RCS ident string. + * src/alpha/ffi.c: Remove RCS ident string. + * src/alpha/osf.S: Remove RCS ident string. + * src/arm/ffi.c: Remove RCS ident string. + * src/arm/sysv.S: Remove RCS ident string. + * src/mips/ffi.c: Remove RCS ident string. + * src/mips/n32.S: Remove RCS ident string. + * src/mips/o32.S: Remove RCS ident string. + * src/sparc/ffi.c: Remove RCS ident string. + * src/sparc/v8.S: Remove RCS ident string. + * src/sparc/v9.S: Remove RCS ident string. + * src/x86/ffi.c: Remove RCS ident string. + * src/x86/sysv.S: Remove RCS ident string. + +2001-02-08 Joseph S. Myers + + * include/ffi.h.in: Change sourceware.cygnus.com references to + gcc.gnu.org. + +2000-12-09 Richard Henderson + + * src/alpha/ffi.c (ffi_call): Simplify struct return test. + (ffi_closure_osf_inner): Index rather than increment avalue + and arg_types. Give ffi_closure_osf the raw return value type. + * src/alpha/osf.S (ffi_closure_osf): Handle return value type + promotion. + +2000-12-07 Richard Henderson + + * src/raw_api.c (ffi_translate_args): Fix typo. + (ffi_prep_closure): Likewise. + + * include/ffi.h.in [ALPHA]: Define FFI_CLOSURES and + FFI_TRAMPOLINE_SIZE. + * src/alpha/ffi.c (ffi_prep_cif_machdep): Adjust minimal + cif->bytes for new ffi_call_osf implementation. + (ffi_prep_args): Absorb into ... + (ffi_call): ... here. Do all stack allocation here and + avoid a callback function. + (ffi_prep_closure, ffi_closure_osf_inner): New. + * src/alpha/osf.S (ffi_call_osf): Reimplement with no callback. + (ffi_closure_osf): New. + +2000-09-10 Alexandre Oliva + + * config.guess, config.sub, install-sh: Removed. + * ltconfig, ltmain.sh, missing, mkinstalldirs: Likewise. + * Makefile.in: Rebuilt. + + * acinclude.m4: Include libtool macros from the top level. + * aclocal.m4, configure: Rebuilt. + +2000-08-22 Alexandre Oliva + + * configure.in [i*86-*-freebsd*] (TARGET, TARGETDIR): Set. + * configure: Rebuilt. + +2000-05-11 Scott Bambrough + + * libffi/src/arm/sysv.S (ffi_call_SYSV): Doubles are not saved to + memory correctly. Use conditional instructions, not branches where + possible. + +2000-05-04 Tom Tromey + + * configure: Rebuilt. + * configure.in: Match `arm*-*-linux-*'. + From Chris Dornan . + +2000-04-28 Jakub Jelinek + + * Makefile.am (SUBDIRS): Define. + (AM_MAKEFLAGS): Likewise. + (Multilib support.): Add section. + * Makefile.in: Rebuilt. + * ltconfig (extra_compiler_flags, extra_compiler_flags_value): + New variables. Set for gcc using -print-multi-lib. Export them + to libtool. + (sparc64-*-linux-gnu*): Use libsuff 64 for search paths. + * ltmain.sh (B|b|V): Don't throw away gcc's -B, -b and -V options + for -shared links. + (extra_compiler_flags_value, extra_compiler_flags): Check these + for extra compiler options which need to be passed down in + compiler_flags. + +2000-04-16 Anthony Green + + * configure: Rebuilt. + * configure.in: Change i*86-pc-linux* to i*86-*-linux*. + +2000-04-14 Jakub Jelinek + + * include/ffi.h.in (SPARC64): Define for 64bit SPARC builds. + Set SPARC FFI_DEFAULT_ABI based on SPARC64 define. + * src/sparc/ffi.c (ffi_prep_args_v8): Renamed from ffi_prep_args. + Replace all void * sizeofs with sizeof(int). + Only compare type with FFI_TYPE_LONGDOUBLE if LONGDOUBLE is + different than DOUBLE. + Remove FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases (handled elsewhere). + (ffi_prep_args_v9): New function. + (ffi_prep_cif_machdep): Handle V9 ABI and long long on V8. + (ffi_V9_return_struct): New function. + (ffi_call): Handle FFI_V9 ABI from 64bit code and FFI_V8 ABI from + 32bit code (not yet cross-arch calls). + * src/sparc/v8.S: Add struct return delay nop. + Handle long long. + * src/sparc/v9.S: New file. + * src/prep_cif.c (ffi_prep_cif): Return structure pointer + is used on sparc64 only for structures larger than 32 bytes. + Pass by reference for structures is done for structure arguments + larger than 16 bytes. + * src/ffitest.c (main): Use 64bit rint on sparc64. + Run long long tests on sparc. + * src/types.c (FFI_TYPE_POINTER): Pointer is 64bit on alpha and + sparc64. + (FFI_TYPE_LONGDOUBLE): long double is 128 bit aligned to 128 bits + on sparc64. + * configure.in (sparc-*-linux*): New supported target. + (sparc64-*-linux*): Likewise. + * configure: Rebuilt. + * Makefile.am: Add v9.S to SPARC files. + * Makefile.in: Likewise. + (LINK): Surround $(CCLD) into double quotes, so that multilib + compiles work correctly. + +2000-04-04 Alexandre Petit-Bianco + + * configure: Rebuilt. + * configure.in: (i*86-*-solaris*): New libffi target. Patch + proposed by Bryce McKinlay. + +2000-03-20 Tom Tromey + + * Makefile.in: Hand edit for java_raw_api.lo. + +2000-03-08 Bryce McKinlay + + * config.guess, config.sub: Update from the gcc tree. + Fix for PR libgcj/168. + +2000-03-03 Tom Tromey + + * Makefile.in: Fixed ia64 by hand. + + * configure: Rebuilt. + * configure.in (--enable-multilib): New option. + (libffi_basedir): New subst. + (AC_OUTPUT): Added multilib code. + +2000-03-02 Tom Tromey + + * Makefile.in: Rebuilt. + * Makefile.am (TARGET_SRC_IA64): Use `ia64', not `alpha', as + directory name. + +2000-02-25 Hans Boehm + + * src/ia64/ffi.c, src/ia64/ia64_flags.h, src/ia64/unix.S: New + files. + * src/raw_api.c (ffi_translate_args): Fixed typo in argument + list. + (ffi_prep_raw_closure): Use ffi_translate_args, not + ffi_closure_translate. + * src/java_raw_api.c: New file. + * src/ffitest.c (closure_test_fn): New function. + (main): Define `rint' as long long on IA64. Added new test when + FFI_CLOSURES is defined. + * include/ffi.h.in (ALIGN): Use size_t, not unsigned. + (ffi_abi): Recognize IA64. + (ffi_raw): Added `flt' field. + Added "Java raw API" code. + * configure.in: Recognize ia64. + * Makefile.am (TARGET_SRC_IA64): New macro. + (libffi_la_common_SOURCES): Added java_raw_api.c. + (libffi_la_SOURCES): Define in IA64 case. + +2000-01-04 Tom Tromey + + * Makefile.in: Rebuilt with newer automake. + +1999-12-31 Tom Tromey + + * Makefile.am (INCLUDES): Added -I$(top_srcdir)/src. + +1999-09-01 Tom Tromey + + * include/ffi.h.in: Removed PACKAGE and VERSION defines and + undefs. + * fficonfig.h.in: Rebuilt. + * configure: Rebuilt. + * configure.in: Pass 3rd argument to AM_INIT_AUTOMAKE. + Use AM_PROG_LIBTOOL (automake 1.4 compatibility). + * acconfig.h: Don't #undef PACKAGE or VERSION. + +1999-08-09 Anthony Green + + * include/ffi.h.in: Try to work around messy header problem + with PACKAGE and VERSION. + + * configure: Rebuilt. + * configure.in: Change version to 2.00-beta. + + * fficonfig.h.in: Rebuilt. + * acconfig.h (FFI_NO_STRUCTS, FFI_NO_RAW_API): Define. + + * src/x86/ffi.c (ffi_raw_call): Rename. + +1999-08-02 Kresten Krab Thorup + + * src/x86/ffi.c (ffi_closure_SYSV): New function. + (ffi_prep_incoming_args_SYSV): Ditto. + (ffi_prep_closure): Ditto. + (ffi_closure_raw_SYSV): Ditto. + (ffi_prep_raw_closure): More ditto. + (ffi_call_raw): Final ditto. + + * include/ffi.h.in: Add definitions for closure and raw API. + + * src/x86/ffi.c (ffi_prep_cif_machdep): Added case for + FFI_TYPE_UINT64. + + * Makefile.am (libffi_la_common_SOURCES): Added raw_api.c + + * src/raw_api.c: New file. + + * include/ffi.h.in (ffi_raw): New type. + (UINT_ARG, SINT_ARG): New defines. + (ffi_closure, ffi_raw_closure): New types. + (ffi_prep_closure, ffi_prep_raw_closure): New declarations. + + * configure.in: Add check for endianness and sizeof void*. + + * src/x86/sysv.S (ffi_call_SYSV): Call fixup routine via argument, + instead of directly. + + * configure: Rebuilt. + +Thu Jul 8 14:28:42 1999 Anthony Green + + * configure.in: Add x86 and powerpc BeOS configurations. + From Makoto Kato . + +1999-05-09 Anthony Green + + * configure.in: Add warning about this being beta code. + Remove src/Makefile.am from the picture. + * configure: Rebuilt. + + * Makefile.am: Move logic from src/Makefile.am. Add changes + to support libffi as a target library. + * Makefile.in: Rebuilt. + + * aclocal.m4, config.guess, config.sub, ltconfig, ltmain.sh: + Upgraded to new autoconf, automake, libtool. + + * README: Tweaks. + + * LICENSE: Update copyright date. + + * src/Makefile.am, src/Makefile.in: Removed. + +1998-11-29 Anthony Green + + * include/ChangeLog: Removed. + * src/ChangeLog: Removed. + * src/mips/ChangeLog: Removed. + * src/sparc/ChangeLog: Remboved. + * src/x86/ChangeLog: Removed. + + * ChangeLog.v1: Created. + +============================================================================= +From the old ChangeLog.libffi file.... + +2011-02-08 Andreas Tobler + + * testsuite/lib/libffi.exp: Tweak for stand-alone mode. + +2009-12-25 Samuli Suominen + + * configure.ac: Undefine _AC_ARG_VAR_PRECIOUS for autoconf 2.64. + * configure: Rebuilt. + * fficonfig.h.in: Rebuilt. + +2009-06-16 Andrew Haley + + * testsuite/libffi.call/cls_align_sint64.c, + testsuite/libffi.call/cls_align_uint64.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_ulonglong.c, + testsuite/libffi.call/return_ll1.c, + testsuite/libffi.call/stret_medium2.c: Fix printf format + specifiers. + * testsuite/libffi.call/huge_struct.c: Ad x86 XFAILs. + * testsuite/libffi.call/float2.c: Fix dg-excess-errors. + * testsuite/libffi.call/ffitest.h, + testsuite/libffi.special/ffitestcxx.h (PRIdLL, PRIuLL): Define. + +2009-06-12 Andrew Haley + + * testsuite/libffi.call/cls_align_sint64.c, + testsuite/libffi.call/cls_align_uint64.c, + testsuite/libffi.call/cls_ulonglong.c, + testsuite/libffi.call/return_ll1.c, + testsuite/libffi.call/stret_medium2.c: Fix printf format + specifiers. + testsuite/libffi.special/unwindtest.cc: include stdint.h. + +2009-06-11 Timothy Wall + + * Makefile.am, + configure.ac, + include/ffi.h.in, + include/ffi_common.h, + src/closures.c, + src/dlmalloc.c, + src/x86/ffi.c, + src/x86/ffitarget.h, + src/x86/win64.S (new), + README: Added win64 support (mingw or MSVC) + * Makefile.in, + include/Makefile.in, + man/Makefile.in, + testsuite/Makefile.in, + configure, + aclocal.m4: Regenerated + * ltcf-c.sh: properly escape cygwin/w32 path + * man/ffi_call.3: Clarify size requirements for return value. + * src/x86/ffi64.c: Fix filename in comment. + * src/x86/win32.S: Remove unused extern. + + * testsuite/libffi.call/closure_fn0.c, + testsuite/libffi.call/closure_fn1.c, + testsuite/libffi.call/closure_fn2.c, + testsuite/libffi.call/closure_fn3.c, + testsuite/libffi.call/closure_fn4.c, + testsuite/libffi.call/closure_fn5.c, + testsuite/libffi.call/closure_fn6.c, + testsuite/libffi.call/closure_stdcall.c, + testsuite/libffi.call/cls_12byte.c, + testsuite/libffi.call/cls_16byte.c, + testsuite/libffi.call/cls_18byte.c, + testsuite/libffi.call/cls_19byte.c, + testsuite/libffi.call/cls_1_1byte.c, + testsuite/libffi.call/cls_20byte.c, + testsuite/libffi.call/cls_20byte1.c, + testsuite/libffi.call/cls_24byte.c, + testsuite/libffi.call/cls_2byte.c, + testsuite/libffi.call/cls_3_1byte.c, + testsuite/libffi.call/cls_3byte1.c, + testsuite/libffi.call/cls_3byte2.c, + testsuite/libffi.call/cls_4_1byte.c, + testsuite/libffi.call/cls_4byte.c, + testsuite/libffi.call/cls_5_1_byte.c, + testsuite/libffi.call/cls_5byte.c, + testsuite/libffi.call/cls_64byte.c, + testsuite/libffi.call/cls_6_1_byte.c, + testsuite/libffi.call/cls_6byte.c, + testsuite/libffi.call/cls_7_1_byte.c, + testsuite/libffi.call/cls_7byte.c, + testsuite/libffi.call/cls_8byte.c, + testsuite/libffi.call/cls_9byte1.c, + testsuite/libffi.call/cls_9byte2.c, + testsuite/libffi.call/cls_align_double.c, + testsuite/libffi.call/cls_align_float.c, + testsuite/libffi.call/cls_align_longdouble.c, + testsuite/libffi.call/cls_align_longdouble_split.c, + testsuite/libffi.call/cls_align_longdouble_split2.c, + testsuite/libffi.call/cls_align_pointer.c, + testsuite/libffi.call/cls_align_sint16.c, + testsuite/libffi.call/cls_align_sint32.c, + testsuite/libffi.call/cls_align_sint64.c, + testsuite/libffi.call/cls_align_uint16.c, + testsuite/libffi.call/cls_align_uint32.c, + testsuite/libffi.call/cls_align_uint64.c, + testsuite/libffi.call/cls_dbls_struct.c, + testsuite/libffi.call/cls_double.c, + testsuite/libffi.call/cls_double_va.c, + testsuite/libffi.call/cls_float.c, + testsuite/libffi.call/cls_longdouble.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_multi_schar.c, + testsuite/libffi.call/cls_multi_sshort.c, + testsuite/libffi.call/cls_multi_sshortchar.c, + testsuite/libffi.call/cls_multi_uchar.c, + testsuite/libffi.call/cls_multi_ushort.c, + testsuite/libffi.call/cls_multi_ushortchar.c, + testsuite/libffi.call/cls_pointer.c, + testsuite/libffi.call/cls_pointer_stack.c, + testsuite/libffi.call/cls_schar.c, + testsuite/libffi.call/cls_sint.c, + testsuite/libffi.call/cls_sshort.c, + testsuite/libffi.call/cls_uchar.c, + testsuite/libffi.call/cls_uint.c, + testsuite/libffi.call/cls_ulonglong.c, + testsuite/libffi.call/cls_ushort.c, + testsuite/libffi.call/err_bad_abi.c, + testsuite/libffi.call/err_bad_typedef.c, + testsuite/libffi.call/float2.c, + testsuite/libffi.call/huge_struct.c, + testsuite/libffi.call/nested_struct.c, + testsuite/libffi.call/nested_struct1.c, + testsuite/libffi.call/nested_struct10.c, + testsuite/libffi.call/nested_struct2.c, + testsuite/libffi.call/nested_struct3.c, + testsuite/libffi.call/nested_struct4.c, + testsuite/libffi.call/nested_struct5.c, + testsuite/libffi.call/nested_struct6.c, + testsuite/libffi.call/nested_struct7.c, + testsuite/libffi.call/nested_struct8.c, + testsuite/libffi.call/nested_struct9.c, + testsuite/libffi.call/problem1.c, + testsuite/libffi.call/return_ldl.c, + testsuite/libffi.call/return_ll1.c, + testsuite/libffi.call/stret_large.c, + testsuite/libffi.call/stret_large2.c, + testsuite/libffi.call/stret_medium.c, + testsuite/libffi.call/stret_medium2.c, + testsuite/libffi.special/unwindtest.cc: use ffi_closure_alloc instead + of checking for MMAP. Use intptr_t instead of long casts. + +2009-06-04 Andrew Haley + + * src/powerpc/ffitarget.h: Fix misapplied merge from gcc. + +2009-06-04 Andrew Haley + + * src/mips/o32.S, + src/mips/n32.S: Fix licence formatting. + +2009-06-04 Andrew Haley + + * src/x86/darwin.S: Fix licence formatting. + src/x86/win32.S: Likewise. + src/sh64/sysv.S: Likewise. + src/sh/sysv.S: Likewise. + +2009-06-04 Andrew Haley + + * src/sh64/ffi.c: Remove lint directives. Was missing from merge + of Andreas Tobler's patch from 2006-04-22. + +2009-06-04 Andrew Haley + + * src/sh/ffi.c: Apply missing hunk from Alexandre Oliva's patch of + 2007-03-07. + +2008-12-26 Timothy Wall + + * testsuite/libffi.call/cls_longdouble.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_align_longdouble.c, + testsuite/libffi.call/cls_align_longdouble_split.c, + testsuite/libffi.call/cls_align_longdouble_split2.c: mark expected + failures on x86_64 cygwin/mingw. + +2008-12-22 Timothy Wall + + * testsuite/libffi.call/closure_fn0.c, + testsuite/libffi.call/closure_fn1.c, + testsuite/libffi.call/closure_fn2.c, + testsuite/libffi.call/closure_fn3.c, + testsuite/libffi.call/closure_fn4.c, + testsuite/libffi.call/closure_fn5.c, + testsuite/libffi.call/closure_fn6.c, + testsuite/libffi.call/closure_loc_fn0.c, + testsuite/libffi.call/closure_stdcall.c, + testsuite/libffi.call/cls_align_pointer.c, + testsuite/libffi.call/cls_pointer.c, + testsuite/libffi.call/cls_pointer_stack.c: use portable cast from + pointer to integer (intptr_t). + * testsuite/libffi.call/cls_longdouble.c: disable for win64. + +2008-12-19 Anthony Green + + * configure.ac: Bump version to 3.0.8. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-11-11 Anthony Green + + * configure.ac: Bump version to 3.0.7. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-08-25 Andreas Tobler + + * src/powerpc/ffitarget.h (ffi_abi): Add FFI_LINUX and + FFI_LINUX_SOFT_FLOAT to the POWERPC_FREEBSD enum. + Add note about flag bits used for FFI_SYSV_TYPE_SMALL_STRUCT. + Adjust copyright notice. + * src/powerpc/ffi.c: Add two new flags to indicate if we have one + register or two register to use for FFI_SYSV structs. + (ffi_prep_cif_machdep): Pass the right register flag introduced above. + (ffi_closure_helper_SYSV): Fix the return type for + FFI_SYSV_TYPE_SMALL_STRUCT. Comment. + Adjust copyright notice. + +2008-07-24 Anthony Green + + * testsuite/libffi.call/cls_dbls_struct.c, + testsuite/libffi.call/cls_double_va.c, + testsuite/libffi.call/cls_longdouble.c, + testsuite/libffi.call/cls_longdouble_va.c, + testsuite/libffi.call/cls_pointer.c, + testsuite/libffi.call/cls_pointer_stack.c, + testsuite/libffi.call/err_bad_abi.c: Clean up failures from + compiler warnings. + +2008-07-17 Anthony Green + + * configure.ac: Bump version to 3.0.6. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. Add documentation. + * README: Update for new release. + +2008-07-16 Kaz Kojima + + * src/sh/ffi.c (ffi_prep_closure_loc): Turn INSN into an unsigned + int. + +2008-07-16 Kaz Kojima + + * src/sh/sysv.S: Add .note.GNU-stack on Linux. + * src/sh64/sysv.S: Likewise. + +2008-04-03 Anthony Green + + * libffi.pc.in (Libs): Add -L${libdir}. + * configure.ac: Bump version to 3.0.5. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-04-03 Anthony Green + Xerces Ranby + + * include/ffi.h.in: Wrap definition of target architecture to + protect from double definitions. + +2008-03-22 Moriyoshi Koizumi + + * src/x86/ffi.c (ffi_prep_closure_loc): Fix for bug revealed in + closure_loc_fn0.c. + * testsuite/libffi.call/closure_loc_fn0.c (closure_loc_test_fn0): + New test. + +2008-03-04 Anthony Green + Blake Chaffin + hos@tamanegi.org + + * testsuite/libffi.call/cls_align_longdouble_split2.c + testsuite/libffi.call/cls_align_longdouble_split.c + testsuite/libffi.call/cls_dbls_struct.c + testsuite/libffi.call/cls_double_va.c + testsuite/libffi.call/cls_longdouble.c + testsuite/libffi.call/cls_longdouble_va.c + testsuite/libffi.call/cls_pointer.c + testsuite/libffi.call/cls_pointer_stack.c + testsuite/libffi.call/err_bad_abi.c + testsuite/libffi.call/err_bad_typedef.c + testsuite/libffi.call/huge_struct.c + testsuite/libffi.call/stret_large2.c + testsuite/libffi.call/stret_large.c + testsuite/libffi.call/stret_medium2.c + testsuite/libffi.call/stret_medium.c: New tests from Apple. + +2008-02-26 Jakub Jelinek + Anthony Green + + * src/alpha/osf.S: Add .note.GNU-stack on Linux. + * src/s390/sysv.S: Likewise. + * src/powerpc/linux64.S: Likewise. + * src/powerpc/linux64_closure.S: Likewise. + * src/powerpc/ppc_closure.S: Likewise. + * src/powerpc/sysv.S: Likewise. + * src/x86/unix64.S: Likewise. + * src/x86/sysv.S: Likewise. + * src/sparc/v8.S: Likewise. + * src/sparc/v9.S: Likewise. + * src/m68k/sysv.S: Likewise. + * src/ia64/unix.S: Likewise. + * src/arm/sysv.S: Likewise. + +2008-02-26 Anthony Green + Thomas Heller + + * src/x86/ffi.c (ffi_closure_SYSV_inner): Change C++ comment to C + comment. + +2008-02-26 Anthony Green + Thomas Heller + + * include/ffi.h.in: Change void (*)() to void (*)(void). + +2008-02-26 Anthony Green + Thomas Heller + + * src/alpha/ffi.c: Change void (*)() to void (*)(void). + src/alpha/osf.S, src/arm/ffi.c, src/frv/ffi.c, src/ia64/ffi.c, + src/ia64/unix.S, src/java_raw_api.c, src/m32r/ffi.c, + src/mips/ffi.c, src/pa/ffi.c, src/pa/hpux32.S, src/pa/linux.S, + src/powerpc/ffi.c, src/powerpc/ffi_darwin.c, src/raw_api.c, + src/s390/ffi.c, src/sh/ffi.c, src/sh64/ffi.c, src/sparc/ffi.c, + src/x86/ffi.c, src/x86/unix64.S, src/x86/darwin64.S, + src/x86/ffi64.c: Ditto. + +2008-02-24 Anthony Green + + * configure.ac: Accept openbsd*, not just openbsd. + Bump version to 3.0.4. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-02-22 Anthony Green + + * README: Clean up list of tested platforms. + +2008-02-22 Anthony Green + + * configure.ac: Bump version to 3.0.3. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. Clean up test docs. + +2008-02-22 Bjoern Koenig + Andreas Tobler + + * configure.ac: Add amd64-*-freebsd* target. + * configure: Regenerate. + +2008-02-22 Thomas Heller + + * configure.ac: Add x86 OpenBSD support. + * configure: Rebuilt. + +2008-02-21 Thomas Heller + + * README: Change "make test" to "make check". + +2008-02-21 Anthony Green + + * configure.ac: Bump version to 3.0.2. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-02-21 Björn König + + * src/x86/freebsd.S: New file. + * configure.ac: Add x86 FreeBSD support. + * Makefile.am: Ditto. + +2008-02-15 Anthony Green + + * configure.ac: Bump version to 3.0.1. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * libtool-version: Increment revision. + * README: Update for new release. + +2008-02-15 David Daney + + * src/mips/ffi.c: Remove extra '>' from include directive. + (ffi_prep_closure_loc): Use clear_location instead of tramp. + +2008-02-15 Anthony Green + + * configure.ac: Bump version to 3.0.0. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + +2008-02-15 David Daney + + * src/mips/ffi.c (USE__BUILTIN___CLEAR_CACHE): + Define (conditionally), and use it to include cachectl.h. + (ffi_prep_closure_loc): Fix cache flushing. + * src/mips/ffitarget.h (_ABIN32, _ABI64, _ABIO32): Define. + +2008-02-15 Anthony Green + + * man/ffi_call.3, man/ffi_prep_cif.3, man/ffi.3: + Update dates and remove all references to ffi_prep_closure. + * configure.ac: Bump version to 2.99.9. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + +2008-02-15 Anthony Green + + * man/ffi_prep_closure.3: Delete. + * man/Makefile.am (EXTRA_DIST): Remove ffi_prep_closure.3. + (man_MANS): Ditto. + * man/Makefile.in: Rebuilt. + * configure.ac: Bump version to 2.99.8. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + +2008-02-14 Anthony Green + + * configure.ac: Bump version to 2.99.7. + * configure, doc/stamp-vti, doc/version.texi: Rebuilt. + * include/ffi.h.in LICENSE src/debug.c src/closures.c + src/ffitest.c src/s390/sysv.S src/s390/ffitarget.h + src/types.c src/m68k/ffitarget.h src/raw_api.c src/frv/ffi.c + src/frv/ffitarget.h src/sh/ffi.c src/sh/sysv.S + src/sh/ffitarget.h src/powerpc/ffitarget.h src/pa/ffi.c + src/pa/ffitarget.h src/pa/linux.S src/java_raw_api.c + src/cris/ffitarget.h src/x86/ffi.c src/x86/sysv.S + src/x86/unix64.S src/x86/win32.S src/x86/ffitarget.h + src/x86/ffi64.c src/x86/darwin.S src/ia64/ffi.c + src/ia64/ffitarget.h src/ia64/ia64_flags.h src/ia64/unix.S + src/sparc/ffi.c src/sparc/v9.S src/sparc/ffitarget.h + src/sparc/v8.S src/alpha/ffi.c src/alpha/ffitarget.h + src/alpha/osf.S src/sh64/ffi.c src/sh64/sysv.S + src/sh64/ffitarget.h src/mips/ffi.c src/mips/ffitarget.h + src/mips/n32.S src/mips/o32.S src/arm/ffi.c src/arm/sysv.S + src/arm/ffitarget.h src/prep_cif.c: Update license text. + +2008-02-14 Anthony Green + + * README: Update tested platforms. + * configure.ac: Bump version to 2.99.6. + * configure: Rebuilt. + +2008-02-14 Anthony Green + + * configure.ac: Bump version to 2.99.5. + * configure: Rebuilt. + * Makefile.am (EXTRA_DIST): Add darwin64.S + * Makefile.in: Rebuilt. + * testsuite/lib/libffi-dg.exp: Remove libstdc++ bits from GCC tree. + * LICENSE: Update WARRANTY. + +2008-02-14 Anthony Green + + * libffi.pc.in (libdir): Fix libdir definition. + * configure.ac: Bump version to 2.99.4. + * configure: Rebuilt. + +2008-02-14 Anthony Green + + * README: Update. + * libffi.info: New file. + * doc/stamp-vti: New file. + * configure.ac: Bump version to 2.99.3. + * configure: Rebuilt. + +2008-02-14 Anthony Green + + * Makefile.am (SUBDIRS): Add man dir. + * Makefile.in: Rebuilt. + * configure.ac: Create Makefile. + * configure: Rebuilt. + * man/ffi_call.3 man/ffi_prep_cif.3 man/ffi_prep_closure.3 + man/Makefile.am man/Makefile.in: New files. + +2008-02-14 Tom Tromey + + * aclocal.m4, Makefile.in, configure, fficonfig.h.in: Rebuilt. + * mdate-sh, texinfo.tex: New files. + * Makefile.am (info_TEXINFOS): New variable. + * doc/libffi.texi: New file. + * doc/version.texi: Likewise. + +2008-02-14 Anthony Green + + * Makefile.am (AM_CFLAGS): Don't compile with -D$(TARGET). + (lib_LTLIBRARIES): Define. + (toolexeclib_LIBRARIES): Undefine. + * Makefile.in: Rebuilt. + * configure.ac: Reset version to 2.99.1. + * configure.in: Rebuilt. + +2008-02-14 Anthony Green + + * libffi.pc.in: Use @PACKAGE_NAME@ and @PACKAGE_VERSION@. + * configure.ac: Reset version to 2.99.1. + * configure.in: Rebuilt. + * Makefile.am (EXTRA_DIST): Add ChangeLog.libffi. + * Makefile.in: Rebuilt. + * LICENSE: Update copyright notice. + +2008-02-14 Anthony Green + + * include/Makefile.am (nodist_includes_HEADERS): Define. Don't + distribute ffitarget.h or ffi.h from the build include dir. + * Makefile.in: Rebuilt. + +2008-02-14 Anthony Green + + * include/Makefile.am (includesdir): Install headers under libdir. + (pkgconfigdir): Define. Install libffi.pc. + * include/Makefile.in: Rebuilt. + * libffi.pc.in: Create. + * libtool-version: Increment CURRENT + * configure.ac: Add libffi.pc.in + * configure: Rebuilt. + +2008-02-03 Anthony Green + + * include/Makefile.am (includesdir): Fix header install with + DESTDIR. + * include/Makefile.in: Rebuilt. + +2008-02-03 Timothy Wall + + * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL): Calculate jump return + offset based on code pointer, not data pointer. + +2008-02-01 Anthony Green + + * include/Makefile.am: Fix header installs. + * Makefile.am: Ditto. + * include/Makefile.in: Rebuilt. + * Makefile.in: Ditto. + +2008-02-01 Anthony Green + + * src/x86/ffi.c (FFI_INIT_TRAMPOLINE_STDCALL, + FFI_INIT_TRAMPOLINE): Revert my broken changes to twall's last + patch. + +2008-01-31 Anthony Green + + * Makefile.am (EXTRA_DIST): Add missing files. + * testsuite/Makefile.am: Ditto. + * Makefile.in, testsuite/Makefile.in: Rebuilt. + +2008-01-31 Timothy Wall + + * testsuite/libffi.call/closure_stdcall.c: Add test for stdcall + closures. + * src/x86/ffitarget.h: Increase size of trampoline for stdcall + closures. + * src/x86/win32.S: Add assembly for stdcall closure. + * src/x86/ffi.c: Initialize stdcall closure trampoline. + +2008-01-30 H.J. Lu + + PR libffi/34612 + * src/x86/sysv.S (ffi_closure_SYSV): Pop 4 byte from stack when + returning struct. + + * testsuite/libffi.call/call.exp: Add "-O2 -fomit-frame-pointer" + tests. + +2008-01-30 Anthony Green + + * Makefile.am, include/Makefile.am: Move headers to + libffi_la_SOURCES for new automake. + * Makefile.in, include/Makefile.in: Rebuilt. + + * testsuite/lib/wrapper.exp: Copied from gcc tree to allow for + execution outside of gcc tree. + * testsuite/lib/target-libpath.exp: Ditto. + + * testsuite/lib/libffi-dg.exp: Many changes to allow for execution + outside of gcc tree. + + +============================================================================= +From the old ChangeLog.libgcj file.... + +2004-01-14 Kelley Cook + + * configure.in: Add in AC_PREREQ(2.13) + +2003-02-20 Alexandre Oliva + + * configure.in: Propagate ORIGINAL_LD_FOR_MULTILIBS to + config.status. + * configure: Rebuilt. + +2002-01-27 Alexandre Oliva + + * configure.in (toolexecdir, toolexeclibdir): Set and AC_SUBST. + Remove USE_LIBDIR conditional. + * Makefile.am (toolexecdir, toolexeclibdir): Don't override. + * Makefile.in, configure: Rebuilt. + +Mon Aug 9 18:33:38 1999 Rainer Orth + + * include/Makefile.in: Rebuilt. + * Makefile.in: Rebuilt + * Makefile.am (toolexeclibdir): Add $(MULTISUBDIR) even for native + builds. + Use USE_LIBDIR. + + * configure: Rebuilt. + * configure.in (USE_LIBDIR): Define for native builds. + Use lowercase in configure --help explanations. + +1999-08-08 Anthony Green + + * include/ffi.h.in (FFI_FN): Remove `...'. + +1999-08-08 Anthony Green + + * Makefile.in: Rebuilt. + * Makefile.am (AM_CFLAGS): Compile with -fexceptions. + + * src/x86/sysv.S: Add exception handling metadata. + + +============================================================================= + +The libffi version 1 ChangeLog archive. + +Version 1 of libffi had per-directory ChangeLogs. Current and future +versions have a single ChangeLog file in the root directory. The +version 1 ChangeLogs have all been concatenated into this file for +future reference only. + +--- libffi ---------------------------------------------------------------- + +Mon Oct 5 02:17:50 1998 Anthony Green + + * configure.in: Boosted rev. + * configure, Makefile.in, aclocal.m4: Rebuilt. + * README: Boosted rev and updated release notes. + +Mon Oct 5 01:03:03 1998 Anthony Green + + * configure.in: Boosted rev. + * configure, Makefile.in, aclocal.m4: Rebuilt. + * README: Boosted rev and updated release notes. + +1998-07-25 Andreas Schwab + + * m68k/ffi.c (ffi_prep_cif_machdep): Use bitmask for cif->flags. + Correctly handle small structures. + (ffi_prep_args): Also handle small structures. + (ffi_call): Pass size of return type to ffi_call_SYSV. + * m68k/sysv.S: Adjust for above changes. Correctly align small + structures in the return value. + + * types.c (uint64, sint64) [M68K]: Change alignment to 4. + +Fri Apr 17 17:26:58 1998 Anthony Green + + * configure.in: Boosted rev. + * configure,Makefile.in,aclocal.m4: Rebuilt. + * README: Boosted rev and added release notes. + +Sun Feb 22 00:50:41 1998 Geoff Keating + + * configure.in: Add PowerPC config bits. + +1998-02-14 Andreas Schwab + + * configure.in: Add m68k config bits. Change AC_CANONICAL_SYSTEM + to AC_CANONICAL_HOST, this is not a compiler. Use $host instead + of $target. Remove AC_CHECK_SIZEOF(char), we already know the + result. Fix argument of AC_ARG_ENABLE. + * configure, fficonfig.h.in: Rebuilt. + +Tue Feb 10 20:53:40 1998 Richard Henderson + + * configure.in: Add Alpha config bits. + +Tue May 13 13:39:20 1997 Anthony Green + + * README: Updated dates and reworded Irix comments. + + * configure.in: Removed AC_PROG_RANLIB. + + * Makefile.in, aclocal.m4, config.guess, config.sub, configure, + ltmain.sh, */Makefile.in: libtoolized again and rebuilt with + automake and autoconf. + +Sat May 10 18:44:50 1997 Tom Tromey + + * configure, aclocal.m4: Rebuilt. + * configure.in: Don't compute EXTRADIST; now handled in + src/Makefile.in. Removed macros implied by AM_INIT_AUTOMAKE. + Don't run AM_MAINTAINER_MODE. + +Thu May 8 14:34:05 1997 Anthony Green + + * missing, ltmain.sh, ltconfig.sh: Created. These are new files + required by automake and libtool. + + * README: Boosted rev to 1.14. Added notes. + + * acconfig.h: Moved PACKAGE and VERSION for new automake. + + * configure.in: Changes for libtool. + + * Makefile.am (check): make test now make check. Uses libtool now. + + * Makefile.in, configure.in, aclocal.h, fficonfig.h.in: Rebuilt. + +Thu May 1 16:27:07 1997 Anthony Green + + * missing: Added file required by new automake. + +Tue Nov 26 14:10:42 1996 Anthony Green + + * acconfig.h: Added USING_PURIFY flag. This is defined when + --enable-purify-safety was used at configure time. + + * configure.in (allsources): Added --enable-purify-safety switch. + (VERSION): Boosted rev to 1.13. + * configure: Rebuilt. + +Fri Nov 22 06:46:12 1996 Anthony Green + + * configure.in (VERSION): Boosted rev to 1.12. + Removed special CFLAGS hack for gcc. + * configure: Rebuilt. + + * README: Boosted rev to 1.12. Added notes. + + * Many files: Cygnus Support changed to Cygnus Solutions. + +Wed Oct 30 11:15:25 1996 Anthony Green + + * configure.in (VERSION): Boosted rev to 1.11. + * configure: Rebuilt. + + * README: Boosted rev to 1.11. Added notes about GNU make. + +Tue Oct 29 12:25:12 1996 Anthony Green + + * configure.in: Fixed -Wall trick. + (VERSION): Boosted rev. + * configure: Rebuilt + + * acconfig.h: Needed for --enable-debug configure switch. + + * README: Boosted rev to 1.09. Added more notes on building + libffi, and LCLint. + + * configure.in: Added --enable-debug switch. Boosted rev to + 1.09. + * configure: Rebuilt + +Tue Oct 15 13:11:28 1996 Anthony Green + + * configure.in (VERSION): Boosted rev to 1.08 + * configure: Rebuilt. + + * README: Added n32 bug fix notes. + + * Makefile.am: Added "make lint" production. + * Makefile.in: Rebuilt. + +Mon Oct 14 10:54:46 1996 Anthony Green + + * README: Added web page reference. + + * configure.in, README: Boosted rev to 1.05 + * configure: Rebuilt. + + * README: Fixed n32 sample code. + +Fri Oct 11 17:09:28 1996 Anthony Green + + * README: Added sparc notes. + + * configure.in, README: Boosted rev to 1.04. + * configure: Rebuilt. + +Thu Oct 10 10:31:03 1996 Anthony Green + + * configure.in, README: Boosted rev to 1.03. + * configure: Rebuilt. + + * README: Added struct notes. + + * Makefile.am (EXTRA_DIST): Added LICENSE to distribution. + * Makefile.in: Rebuilt. + + * README: Removed Linux section. No special notes now + because aggregates arg/return types work. + +Wed Oct 9 16:16:42 1996 Anthony Green + + * README, configure.in (VERSION): Boosted rev to 1.02 + * configure: Rebuilt. + +Tue Oct 8 11:56:33 1996 Anthony Green + + * README (NOTE): Added n32 notes. + + * Makefile.am: Added test production. + * Makefile: Rebuilt + + * README: spell checked! + + * configure.in (VERSION): Boosted rev to 1.01 + * configure: Rebuilt. + +Mon Oct 7 15:50:22 1996 Anthony Green + + * configure.in: Added nasty bit to support SGI tools. + * configure: Rebuilt. + + * README: Added SGI notes. Added note about automake bug. + +Mon Oct 7 11:00:28 1996 Anthony Green + + * README: Rewrote intro, and fixed examples. + +Fri Oct 4 10:19:55 1996 Anthony Green + + * configure.in: -D$TARGET is no longer used as a compiler switch. + It is now inserted into ffi.h at configure time. + * configure: Rebuilt. + + * FFI_ABI and FFI_STATUS are now ffi_abi and ffi_status. + +Thu Oct 3 13:47:34 1996 Anthony Green + + * README, LICENSE: Created. Wrote some docs. + + * configure.in: Don't barf on i586-unknown-linuxaout. + Added EXTRADIST code for "make dist". + * configure: Rebuilt. + + * */Makefile.in: Rebuilt with patched automake. + +Tue Oct 1 17:12:25 1996 Anthony Green + + * Makefile.am, aclocal.m4, config.guess, config.sub, + configure.in, fficonfig.h.in, install-sh, mkinstalldirs, + stamp-h.in: Created + * Makefile.in, configure: Generated + +--- libffi/include -------------------------------------------------------- + +Tue Feb 24 13:09:36 1998 Anthony Green + + * ffi_mips.h: Updated FFI_TYPE_STRUCT_* values based on + ffi.h.in changes. This is a work-around for SGI's "simple" + assembler. + +Sun Feb 22 00:51:55 1998 Geoff Keating + + * ffi.h.in: PowerPC support. + +1998-02-14 Andreas Schwab + + * ffi.h.in: Add m68k support. + (FFI_TYPE_LONGDOUBLE): Make it a separate value. + +Tue Feb 10 20:55:16 1998 Richard Henderson + + * ffi.h.in (SIZEOF_ARG): Use a pointer type by default. + + * ffi.h.in: Alpha support. + +Fri Nov 22 06:48:45 1996 Anthony Green + + * ffi.h.in, ffi_common.h: Cygnus Support -> Cygnus Solutions. + +Wed Nov 20 22:31:01 1996 Anthony Green + + * ffi.h.in: Added ffi_type_void definition. + +Tue Oct 29 12:22:40 1996 Anthony Green + + * Makefile.am (hack_DATA): Always install ffi_mips.h. + + * ffi.h.in: Removed FFI_DEBUG. It's now in the correct + place (acconfig.h). + Added #include for size_t definition. + +Tue Oct 15 17:23:35 1996 Anthony Green + + * ffi.h.in, ffi_common.h, ffi_mips.h: More clean up. + Commented out #define of FFI_DEBUG. + +Tue Oct 15 13:01:06 1996 Anthony Green + + * ffi_common.h: Added bool definition. + + * ffi.h.in, ffi_common.h: Clean up based on LCLint output. + Added funny /*@...@*/ comments to annotate source. + +Mon Oct 14 12:29:23 1996 Anthony Green + + * ffi.h.in: Interface changes based on feedback from Jim + Blandy. + +Fri Oct 11 16:49:35 1996 Anthony Green + + * ffi.h.in: Small change for sparc support. + +Thu Oct 10 14:53:37 1996 Anthony Green + + * ffi_mips.h: Added FFI_TYPE_STRUCT_* definitions for + special structure return types. + +Wed Oct 9 13:55:57 1996 Anthony Green + + * ffi.h.in: Added SIZEOF_ARG definition for X86 + +Tue Oct 8 11:40:36 1996 Anthony Green + + * ffi.h.in (FFI_FN): Added macro for eliminating compiler warnings. + Use it to case your function pointers to the proper type. + + * ffi_mips.h (SIZEOF_ARG): Added magic to fix type promotion bug. + + * Makefile.am (EXTRA_DIST): Added ffi_mips.h to EXTRA_DIST. + * Makefile: Rebuilt. + + * ffi_mips.h: Created. Moved all common mips definitions here. + +Mon Oct 7 10:58:12 1996 Anthony Green + + * ffi.h.in: The SGI assember is very picky about parens. Redefined + some macros to avoid problems. + + * ffi.h.in: Added FFI_DEFAULT_ABI definitions. Also added + externs for pointer, and 64bit integral ffi_types. + +Fri Oct 4 09:51:37 1996 Anthony Green + + * ffi.h.in: Added FFI_ABI member to ffi_cif and changed + function prototypes accordingly. + Added #define @TARGET@. Now programs including ffi.h don't + have to specify this themselves. + +Thu Oct 3 15:36:44 1996 Anthony Green + + * ffi.h.in: Changed ffi_prep_cif's values from void* to void** + + * Makefile.am (EXTRA_DIST): Added EXTRA_DIST for "make dist" + to work. + * Makefile.in: Regenerated. + +Wed Oct 2 10:16:59 1996 Anthony Green + + * Makefile.am: Created + * Makefile.in: Generated + + * ffi_common.h: Added rcsid comment + +Tue Oct 1 17:13:51 1996 Anthony Green + + * ffi.h.in, ffi_common.h: Created + +--- libffi/src ------------------------------------------------------------ + +Mon Oct 5 02:17:50 1998 Anthony Green + + * arm/ffi.c, arm/sysv.S: Created. + + * Makefile.am: Added arm files. + * Makefile.in: Rebuilt. + +Mon Oct 5 01:41:38 1998 Anthony Green + + * Makefile.am (libffi_la_LDFLAGS): Incremented revision. + +Sun Oct 4 16:27:17 1998 Anthony Green + + * alpha/osf.S (ffi_call_osf): Patch for DU assembler. + + * ffitest.c (main): long long and long double return values work + for x86. + +Fri Apr 17 11:50:58 1998 Anthony Green + + * Makefile.in: Rebuilt. + + * ffitest.c (main): Floating point tests not executed for systems + with broken lond double (SunOS 4 w/ GCC). + + * types.c: Fixed x86 alignment info for long long types. + +Thu Apr 16 07:15:28 1998 Anthony Green + + * ffitest.c: Added more notes about GCC bugs under Irix 6. + +Wed Apr 15 08:42:22 1998 Anthony Green + + * ffitest.c (struct5): New test function. + (main): New test with struct5. + +Thu Mar 5 10:48:11 1998 Anthony Green + + * prep_cif.c (initialize_aggregate): Fix assertion for + nested structures. + +Tue Feb 24 16:33:41 1998 Anthony Green + + * prep_cif.c (ffi_prep_cif): Added long double support for sparc. + +Sun Feb 22 00:52:18 1998 Geoff Keating + + * powerpc/asm.h: New file. + * powerpc/ffi.c: New file. + * powerpc/sysv.S: New file. + * Makefile.am: PowerPC port. + * ffitest.c (main): Allow all tests to run even in presence of gcc + bug on PowerPC. + +1998-02-17 Anthony Green + + * mips/ffi.c: Fixed comment typo. + + * x86/ffi.c (ffi_prep_cif_machdep), x86/sysv.S (retfloat): + Fixed x86 long double return handling. + + * types.c: Fixed x86 long double alignment info. + +1998-02-14 Andreas Schwab + + * types.c: Add m68k support. + + * ffitest.c (floating): Add long double parameter. + (return_ll, ldblit): New functions to test long long and long + double return value. + (main): Fix type error in assignment of ts[1-4]_type.elements. + Add tests for long long and long double arguments and return + values. + + * prep_cif.c (ffi_prep_cif) [M68K]: Don't allocate argument for + struct value pointer. + + * m68k/ffi.c, m68k/sysv.S: New files. + * Makefile.am: Add bits for m68k port. Add kludge to work around + automake deficiency. + (test): Don't require "." in $PATH. + * Makefile.in: Rebuilt. + +Wed Feb 11 07:36:50 1998 Anthony Green + + * Makefile.in: Rebuilt. + +Tue Feb 10 20:56:00 1998 Richard Henderson + + * alpha/ffi.c, alpha/osf.S: New files. + * Makefile.am: Alpha port. + +Tue Nov 18 14:12:07 1997 Anthony Green + + * mips/ffi.c (ffi_prep_cif_machdep): Initialize rstruct_flag + for n32. + +Tue Jun 3 17:18:20 1997 Anthony Green + + * ffitest.c (main): Added hack to get structure tests working + correctly. + +Sat May 10 19:06:42 1997 Tom Tromey + + * Makefile.in: Rebuilt. + * Makefile.am (EXTRA_DIST): Explicitly list all distributable + files in subdirs. + (VERSION, CC): Removed. + +Thu May 8 17:19:01 1997 Anthony Green + + * Makefile.am: Many changes for new automake and libtool. + * Makefile.in: Rebuilt. + +Fri Nov 22 06:57:56 1996 Anthony Green + + * ffitest.c (main): Fixed test case for non mips machines. + +Wed Nov 20 22:31:59 1996 Anthony Green + + * types.c: Added ffi_type_void declaration. + +Tue Oct 29 13:07:19 1996 Anthony Green + + * ffitest.c (main): Fixed character constants. + (main): Emit warning for structure test 3 failure on Sun. + + * Makefile.am (VPATH): Fixed VPATH def'n so automake won't + strip it out. + Moved distdir hack from libffi to automake. + (ffitest): Added missing -c for $(COMPILE) (change in automake). + * Makefile.in: Rebuilt. + +Tue Oct 15 13:08:20 1996 Anthony Green + + * Makefile.am: Added "make lint" production. + * Makefile.in: Rebuilt. + + * prep_cif.c (STACK_ARG_SIZE): Improved STACK_ARG_SIZE macro. + Clean up based on LCLint output. Added funny /*@...@*/ comments to + annotate source. + + * ffitest.c, debug.c: Cleaned up code. + +Mon Oct 14 12:26:56 1996 Anthony Green + + * ffitest.c: Changes based on interface changes. + + * prep_cif.c (ffi_prep_cif): Cleaned up interface based on + feedback from Jim Blandy. + +Fri Oct 11 15:53:18 1996 Anthony Green + + * ffitest.c: Reordered tests while porting to sparc. + Made changes to handle lame structure passing for sparc. + Removed calls to fflush(). + + * prep_cif.c (ffi_prep_cif): Added special case for sparc + aggregate type arguments. + +Thu Oct 10 09:56:51 1996 Anthony Green + + * ffitest.c (main): Added structure passing/returning tests. + + * prep_cif.c (ffi_prep_cif): Perform proper initialization + of structure return types if needed. + (initialize_aggregate): Bug fix + +Wed Oct 9 16:04:20 1996 Anthony Green + + * types.c: Added special definitions for x86 (double doesn't + need double word alignment). + + * ffitest.c: Added many tests + +Tue Oct 8 09:19:22 1996 Anthony Green + + * prep_cif.c (ffi_prep_cif): Fixed assertion. + + * debug.c (ffi_assert): Must return a non void now. + + * Makefile.am: Added test production. + * Makefile: Rebuilt. + + * ffitest.c (main): Created. + + * types.c: Created. Stripped common code out of */ffi.c. + + * prep_cif.c: Added missing stdlib.h include. + + * debug.c (ffi_type_test): Used "a" to eliminate compiler + warnings in non-debug builds. Included ffi_common.h. + +Mon Oct 7 15:36:42 1996 Anthony Green + + * Makefile.am: Added a rule for .s -> .o + This is required by the SGI compiler. + * Makefile: Rebuilt. + +Fri Oct 4 09:51:08 1996 Anthony Green + + * prep_cif.c (initialize_aggregate): Moved abi specification + to ffi_prep_cif(). + +Thu Oct 3 15:37:37 1996 Anthony Green + + * prep_cif.c (ffi_prep_cif): Changed values from void* to void**. + (initialize_aggregate): Fixed aggregate type initialization. + + * Makefile.am (EXTRA_DIST): Added support code for "make dist". + * Makefile.in: Regenerated. + +Wed Oct 2 11:41:57 1996 Anthony Green + + * debug.c, prep_cif: Created. + + * Makefile.am: Added debug.o and prep_cif.o to OBJ. + * Makefile.in: Regenerated. + + * Makefile.am (INCLUDES): Added missing -I../include + * Makefile.in: Regenerated. + +Tue Oct 1 17:11:51 1996 Anthony Green + + * error.c, Makefile.am: Created. + * Makefile.in: Generated. + +--- libffi/src/x86 -------------------------------------------------------- + +Sun Oct 4 16:27:17 1998 Anthony Green + + * sysv.S (retlongdouble): Fixed long long return value support. + * ffi.c (ffi_prep_cif_machdep): Ditto. + +Wed May 13 04:30:33 1998 Anthony Green + + * ffi.c (ffi_prep_cif_machdep): Fixed long double return value + support. + +Wed Apr 15 08:43:20 1998 Anthony Green + + * ffi.c (ffi_prep_args): small struct support was missing. + +Thu May 8 16:53:58 1997 Anthony Green + + * objects.mak: Removed. + +Mon Dec 2 15:12:58 1996 Tom Tromey + + * sysv.S: Use .balign, for a.out Linux boxes. + +Tue Oct 15 13:06:50 1996 Anthony Green + + * ffi.c: Clean up based on LCLint output. + Added funny /*@...@*/ comments to annotate source. + +Fri Oct 11 16:43:38 1996 Anthony Green + + * ffi.c (ffi_call): Added assertion for bad ABIs. + +Wed Oct 9 13:57:27 1996 Anthony Green + + * sysv.S (retdouble): Fixed double return problems. + + * ffi.c (ffi_call): Corrected fn arg definition. + (ffi_prep_cif_machdep): Fixed double return problems + +Tue Oct 8 12:12:49 1996 Anthony Green + + * ffi.c: Moved ffi_type definitions to types.c. + (ffi_prep_args): Fixed type promotion bug. + +Mon Oct 7 15:53:06 1996 Anthony Green + + * ffi.c (FFI_*_TYPEDEF): Removed redundant ';' + +Fri Oct 4 09:54:53 1996 Anthony Green + + * ffi.c (ffi_call): Removed FFI_ABI arg, and swapped + remaining args. + +Wed Oct 2 10:07:05 1996 Anthony Green + + * ffi.c, sysv.S, objects.mak: Created. + (ffi_prep_cif): cif->rvalue no longer initialized to NULL. + (ffi_prep_cif_machdep): Moved machine independent cif processing + to src/prep_cif.c. Introduced ffi_prep_cif_machdep(). + +--- libffi/src/mips ------------------------------------------------------- + +Tue Feb 17 17:18:07 1998 Anthony Green + + * o32.S: Fixed typo in comment. + + * ffi.c (ffi_prep_cif_machdep): Fixed argument processing. + +Thu May 8 16:53:58 1997 Anthony Green + + * o32.s, n32.s: Wrappers for SGI tool support. + + * objects.mak: Removed. + +Tue Oct 29 14:37:45 1996 Anthony Green + + * ffi.c (ffi_prep_args): Changed int z to size_t z. + +Tue Oct 15 13:17:25 1996 Anthony Green + + * n32.S: Fixed bad stack munging. + + * ffi.c: Moved prototypes for ffi_call_?32() to here from + ffi_mips.h because extended_cif is not defined in ffi_mips.h. + +Mon Oct 14 12:42:02 1996 Anthony Green + + * ffi.c: Interface changes based on feedback from Jim Blandy. + +Thu Oct 10 11:22:16 1996 Anthony Green + + * n32.S, ffi.c: Lots of changes to support passing and + returning structures with the n32 calling convention. + + * n32.S: Fixed fn pointer bug. + + * ffi.c (ffi_prep_cif_machdep): Fix for o32 structure + return values. + (ffi_prep_args): Fixed n32 structure passing when structures + partially fit in registers. + +Wed Oct 9 13:49:25 1996 Anthony Green + + * objects.mak: Added n32.o. + + * n32.S: Created. + + * ffi.c (ffi_prep_args): Added magic to support proper + n32 processing. + +Tue Oct 8 10:37:35 1996 Anthony Green + + * ffi.c: Moved ffi_type definitions to types.c. + (ffi_prep_args): Fixed type promotion bug. + + * o32.S: This code is only built for o32 compiles. + A lot of the #define cruft has moved to ffi_mips.h. + + * ffi.c (ffi_prep_cif_machdep): Fixed arg flags. Second arg + is only processed if the first is either a float or double. + +Mon Oct 7 15:33:59 1996 Anthony Green + + * o32.S: Modified to compile under each of o32, n32 and n64. + + * ffi.c (FFI_*_TYPEDEF): Removed redundant ';' + +Fri Oct 4 09:53:25 1996 Anthony Green + + * ffi.c (ffi_call): Removed FFI_ABI arg, and swapped + remaining args. + +Wed Oct 2 17:41:22 1996 Anthony Green + + * o32.S: Removed crufty definitions. + +Wed Oct 2 12:53:42 1996 Anthony Green + + * ffi.c (ffi_prep_cif): cif->rvalue no longer initialized to NULL. + (ffi_prep_cif_machdep): Moved all machine independent cif processing + to src/prep_cif.c. Introduced ffi_prep_cif_machdep. Return types + of FFI_TYPE_STRUCT are no different than FFI_TYPE_INT. + +Tue Oct 1 17:11:02 1996 Anthony Green + + * ffi.c, o32.S, object.mak: Created + +--- libffi/src/sparc ------------------------------------------------------ + +Tue Feb 24 16:33:18 1998 Anthony Green + + * ffi.c (ffi_prep_args): Added long double support. + +Thu May 8 16:53:58 1997 Anthony Green + + * objects.mak: Removed. + +Thu May 1 16:07:56 1997 Anthony Green + + * v8.S: Fixed minor portability problem reported by + Russ McManus . + +Tue Nov 26 14:12:43 1996 Anthony Green + + * v8.S: Used STACKFRAME define elsewhere. + + * ffi.c (ffi_prep_args): Zero out space when USING_PURIFY + is set. + (ffi_prep_cif_machdep): Allocate the correct stack frame + space for functions with < 6 args. + +Tue Oct 29 15:08:55 1996 Anthony Green + + * ffi.c (ffi_prep_args): int z is now size_t z. + +Mon Oct 14 13:31:24 1996 Anthony Green + + * v8.S (ffi_call_V8): Gordon rewrites this again. It looks + great now. + + * ffi.c (ffi_call): The comment about hijacked registers + is no longer valid after gordoni hacked v8.S. + + * v8.S (ffi_call_V8): Rewrote with gordoni. Much simpler. + + * v8.S, ffi.c: ffi_call() had changed to accept more than + two args, so v8.S had to change (because it hijacks incoming + arg registers). + + * ffi.c: Interface changes based on feedback from Jim Blandy. + +Thu Oct 10 17:48:16 1996 Anthony Green + + * ffi.c, v8.S, objects.mak: Created. + + diff --git a/native/libffi/LICENSE b/native/libffi/LICENSE index a66fab4f25..c2fd979bde 100644 --- a/native/libffi/LICENSE +++ b/native/libffi/LICENSE @@ -1,4 +1,8 @@ +<<<<<<< HEAD libffi - Copyright (c) 1996-2014 Anthony Green, Red Hat, Inc and others. +======= +libffi - Copyright (c) 1996-2020 Anthony Green, Red Hat, Inc and others. +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 See source files for details. Permission is hereby granted, free of charge, to any person obtaining diff --git a/native/libffi/LICENSE-BUILDTOOLS b/native/libffi/LICENSE-BUILDTOOLS new file mode 100644 index 0000000000..d1d626e008 --- /dev/null +++ b/native/libffi/LICENSE-BUILDTOOLS @@ -0,0 +1,353 @@ +The libffi source distribution contains certain code that is not part +of libffi, and is only used as tooling to assist with the building and +testing of libffi. This includes the msvcc.sh script used to wrap the +Microsoft compiler with GNU compatible command-line options, +make_sunver.pl, and the libffi test code distributed in the +testsuite/libffi.bhaible directory. This code is distributed with +libffi for the purpose of convenience only, and libffi is in no way +derived from this code. + +msvcc.sh an testsuite/libffi.bhaible are both distributed under the +terms of the GNU GPL version 2, as below. + + + + GNU GENERAL PUBLIC LICENSE + Version 2, June 1991 + + Copyright (C) 1989, 1991 Free Software Foundation, Inc., + 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + Everyone is permitted to copy and distribute verbatim copies + of this license document, but changing it is not allowed. + + Preamble + + The licenses for most software are designed to take away your +freedom to share and change it. By contrast, the GNU General Public +License is intended to guarantee your freedom to share and change free +software--to make sure the software is free for all its users. This +General Public License applies to most of the Free Software +Foundation's software and to any other program whose authors commit to +using it. (Some other Free Software Foundation software is covered by +the GNU Lesser General Public License instead.) You can apply it to +your programs, too. + + When we speak of free software, we are referring to freedom, not +price. Our General Public Licenses are designed to make sure that you +have the freedom to distribute copies of free software (and charge for +this service if you wish), that you receive source code or can get it +if you want it, that you can change the software or use pieces of it +in new free programs; and that you know you can do these things. + + To protect your rights, we need to make restrictions that forbid +anyone to deny you these rights or to ask you to surrender the rights. +These restrictions translate to certain responsibilities for you if you +distribute copies of the software, or if you modify it. + + For example, if you distribute copies of such a program, whether +gratis or for a fee, you must give the recipients all the rights that +you have. You must make sure that they, too, receive or can get the +source code. And you must show them these terms so they know their +rights. + + We protect your rights with two steps: (1) copyright the software, and +(2) offer you this license which gives you legal permission to copy, +distribute and/or modify the software. + + Also, for each author's protection and ours, we want to make certain +that everyone understands that there is no warranty for this free +software. If the software is modified by someone else and passed on, we +want its recipients to know that what they have is not the original, so +that any problems introduced by others will not reflect on the original +authors' reputations. + + Finally, any free program is threatened constantly by software +patents. We wish to avoid the danger that redistributors of a free +program will individually obtain patent licenses, in effect making the +program proprietary. To prevent this, we have made it clear that any +patent must be licensed for everyone's free use or not licensed at all. + + The precise terms and conditions for copying, distribution and +modification follow. + + GNU GENERAL PUBLIC LICENSE + TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION + + 0. This License applies to any program or other work which contains +a notice placed by the copyright holder saying it may be distributed +under the terms of this General Public License. The "Program", below, +refers to any such program or work, and a "work based on the Program" +means either the Program or any derivative work under copyright law: +that is to say, a work containing the Program or a portion of it, +either verbatim or with modifications and/or translated into another +language. (Hereinafter, translation is included without limitation in +the term "modification".) Each licensee is addressed as "you". + +Activities other than copying, distribution and modification are not +covered by this License; they are outside its scope. The act of +running the Program is not restricted, and the output from the Program +is covered only if its contents constitute a work based on the +Program (independent of having been made by running the Program). +Whether that is true depends on what the Program does. + + 1. You may copy and distribute verbatim copies of the Program's +source code as you receive it, in any medium, provided that you +conspicuously and appropriately publish on each copy an appropriate +copyright notice and disclaimer of warranty; keep intact all the +notices that refer to this License and to the absence of any warranty; +and give any other recipients of the Program a copy of this License +along with the Program. + +You may charge a fee for the physical act of transferring a copy, and +you may at your option offer warranty protection in exchange for a fee. + + 2. You may modify your copy or copies of the Program or any portion +of it, thus forming a work based on the Program, and copy and +distribute such modifications or work under the terms of Section 1 +above, provided that you also meet all of these conditions: + + a) You must cause the modified files to carry prominent notices + stating that you changed the files and the date of any change. + + b) You must cause any work that you distribute or publish, that in + whole or in part contains or is derived from the Program or any + part thereof, to be licensed as a whole at no charge to all third + parties under the terms of this License. + + c) If the modified program normally reads commands interactively + when run, you must cause it, when started running for such + interactive use in the most ordinary way, to print or display an + announcement including an appropriate copyright notice and a + notice that there is no warranty (or else, saying that you provide + a warranty) and that users may redistribute the program under + these conditions, and telling the user how to view a copy of this + License. (Exception: if the Program itself is interactive but + does not normally print such an announcement, your work based on + the Program is not required to print an announcement.) + +These requirements apply to the modified work as a whole. If +identifiable sections of that work are not derived from the Program, +and can be reasonably considered independent and separate works in +themselves, then this License, and its terms, do not apply to those +sections when you distribute them as separate works. But when you +distribute the same sections as part of a whole which is a work based +on the Program, the distribution of the whole must be on the terms of +this License, whose permissions for other licensees extend to the +entire whole, and thus to each and every part regardless of who wrote it. + +Thus, it is not the intent of this section to claim rights or contest +your rights to work written entirely by you; rather, the intent is to +exercise the right to control the distribution of derivative or +collective works based on the Program. + +In addition, mere aggregation of another work not based on the Program +with the Program (or with a work based on the Program) on a volume of +a storage or distribution medium does not bring the other work under +the scope of this License. + + 3. You may copy and distribute the Program (or a work based on it, +under Section 2) in object code or executable form under the terms of +Sections 1 and 2 above provided that you also do one of the following: + + a) Accompany it with the complete corresponding machine-readable + source code, which must be distributed under the terms of Sections + 1 and 2 above on a medium customarily used for software interchange; or, + + b) Accompany it with a written offer, valid for at least three + years, to give any third party, for a charge no more than your + cost of physically performing source distribution, a complete + machine-readable copy of the corresponding source code, to be + distributed under the terms of Sections 1 and 2 above on a medium + customarily used for software interchange; or, + + c) Accompany it with the information you received as to the offer + to distribute corresponding source code. (This alternative is + allowed only for noncommercial distribution and only if you + received the program in object code or executable form with such + an offer, in accord with Subsection b above.) + +The source code for a work means the preferred form of the work for +making modifications to it. For an executable work, complete source +code means all the source code for all modules it contains, plus any +associated interface definition files, plus the scripts used to +control compilation and installation of the executable. However, as a +special exception, the source code distributed need not include +anything that is normally distributed (in either source or binary +form) with the major components (compiler, kernel, and so on) of the +operating system on which the executable runs, unless that component +itself accompanies the executable. + +If distribution of executable or object code is made by offering +access to copy from a designated place, then offering equivalent +access to copy the source code from the same place counts as +distribution of the source code, even though third parties are not +compelled to copy the source along with the object code. + + 4. You may not copy, modify, sublicense, or distribute the Program +except as expressly provided under this License. Any attempt +otherwise to copy, modify, sublicense or distribute the Program is +void, and will automatically terminate your rights under this License. +However, parties who have received copies, or rights, from you under +this License will not have their licenses terminated so long as such +parties remain in full compliance. + + 5. You are not required to accept this License, since you have not +signed it. However, nothing else grants you permission to modify or +distribute the Program or its derivative works. These actions are +prohibited by law if you do not accept this License. Therefore, by +modifying or distributing the Program (or any work based on the +Program), you indicate your acceptance of this License to do so, and +all its terms and conditions for copying, distributing or modifying +the Program or works based on it. + + 6. Each time you redistribute the Program (or any work based on the +Program), the recipient automatically receives a license from the +original licensor to copy, distribute or modify the Program subject to +these terms and conditions. You may not impose any further +restrictions on the recipients' exercise of the rights granted herein. +You are not responsible for enforcing compliance by third parties to +this License. + + 7. If, as a consequence of a court judgment or allegation of patent +infringement or for any other reason (not limited to patent issues), +conditions are imposed on you (whether by court order, agreement or +otherwise) that contradict the conditions of this License, they do not +excuse you from the conditions of this License. If you cannot +distribute so as to satisfy simultaneously your obligations under this +License and any other pertinent obligations, then as a consequence you +may not distribute the Program at all. For example, if a patent +license would not permit royalty-free redistribution of the Program by +all those who receive copies directly or indirectly through you, then +the only way you could satisfy both it and this License would be to +refrain entirely from distribution of the Program. + +If any portion of this section is held invalid or unenforceable under +any particular circumstance, the balance of the section is intended to +apply and the section as a whole is intended to apply in other +circumstances. + +It is not the purpose of this section to induce you to infringe any +patents or other property right claims or to contest validity of any +such claims; this section has the sole purpose of protecting the +integrity of the free software distribution system, which is +implemented by public license practices. Many people have made +generous contributions to the wide range of software distributed +through that system in reliance on consistent application of that +system; it is up to the author/donor to decide if he or she is willing +to distribute software through any other system and a licensee cannot +impose that choice. + +This section is intended to make thoroughly clear what is believed to +be a consequence of the rest of this License. + + 8. If the distribution and/or use of the Program is restricted in +certain countries either by patents or by copyrighted interfaces, the +original copyright holder who places the Program under this License +may add an explicit geographical distribution limitation excluding +those countries, so that distribution is permitted only in or among +countries not thus excluded. In such case, this License incorporates +the limitation as if written in the body of this License. + + 9. The Free Software Foundation may publish revised and/or new versions +of the General Public License from time to time. Such new versions will +be similar in spirit to the present version, but may differ in detail to +address new problems or concerns. + +Each version is given a distinguishing version number. If the Program +specifies a version number of this License which applies to it and "any +later version", you have the option of following the terms and conditions +either of that version or of any later version published by the Free +Software Foundation. If the Program does not specify a version number of +this License, you may choose any version ever published by the Free Software +Foundation. + + 10. If you wish to incorporate parts of the Program into other free +programs whose distribution conditions are different, write to the author +to ask for permission. For software which is copyrighted by the Free +Software Foundation, write to the Free Software Foundation; we sometimes +make exceptions for this. Our decision will be guided by the two goals +of preserving the free status of all derivatives of our free software and +of promoting the sharing and reuse of software generally. + + NO WARRANTY + + 11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY +FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN +OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES +PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED +OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF +MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS +TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE +PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, +REPAIR OR CORRECTION. + + 12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING +WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR +REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, +INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING +OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED +TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY +YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER +PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE +POSSIBILITY OF SUCH DAMAGES. + + END OF TERMS AND CONDITIONS + + How to Apply These Terms to Your New Programs + + If you develop a new program, and you want it to be of the greatest +possible use to the public, the best way to achieve this is to make it +free software which everyone can redistribute and change under these terms. + + To do so, attach the following notices to the program. It is safest +to attach them to the start of each source file to most effectively +convey the exclusion of warranty; and each file should have at least +the "copyright" line and a pointer to where the full notice is found. + + + Copyright (C) + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License along + with this program; if not, write to the Free Software Foundation, Inc., + 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + +Also add information on how to contact you by electronic and paper mail. + +If the program is interactive, make it output a short notice like this +when it starts in an interactive mode: + + Gnomovision version 69, Copyright (C) year name of author + Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'. + This is free software, and you are welcome to redistribute it + under certain conditions; type `show c' for details. + +The hypothetical commands `show w' and `show c' should show the appropriate +parts of the General Public License. Of course, the commands you use may +be called something other than `show w' and `show c'; they could even be +mouse-clicks or menu items--whatever suits your program. + +You should also get your employer (if you work as a programmer) or your +school, if any, to sign a "copyright disclaimer" for the program, if +necessary. Here is a sample; alter the names: + + Yoyodyne, Inc., hereby disclaims all copyright interest in the program + `Gnomovision' (which makes passes at compilers) written by James Hacker. + + , 1 April 1989 + Ty Coon, President of Vice + +This General Public License does not permit incorporating your program into +proprietary programs. If your program is a subroutine library, you may +consider it more useful to permit linking proprietary applications with the +library. If this is what you want to do, use the GNU Lesser General +Public License instead of this License. diff --git a/native/libffi/Makefile.am b/native/libffi/Makefile.am index 0e4045100a..48839d9b13 100644 --- a/native/libffi/Makefile.am +++ b/native/libffi/Makefile.am @@ -5,7 +5,16 @@ AUTOMAKE_OPTIONS = foreign subdir-objects ACLOCAL_AMFLAGS = -I m4 SUBDIRS = include testsuite man +if BUILD_DOCS +## This hack is needed because it doesn't seem possible to make a +## conditional info_TEXINFOS in Automake. At least Automake 1.14 +## either gives errors -- if this attempted in the most +## straightforward way -- or simply unconditionally tries to build the +## info file. +SUBDIRS += doc +endif +<<<<<<< HEAD EXTRA_DIST = LICENSE ChangeLog.v1 ChangeLog.libgcj \ src/aarch64/ffi.c src/aarch64/ffitarget.h src/aarch64/sysv.S \ src/alpha/ffi.c src/alpha/osf.S \ @@ -98,27 +107,107 @@ AM_MAKEFLAGS = \ 'NM=$(NM)' \ 'RANLIB=$(RANLIB)' \ 'DESTDIR=$(DESTDIR)' +======= +EXTRA_DIST = LICENSE ChangeLog.old \ + m4/libtool.m4 m4/lt~obsolete.m4 \ + m4/ltoptions.m4 m4/ltsugar.m4 m4/ltversion.m4 \ + m4/ltversion.m4 src/debug.c msvcc.sh \ + generate-darwin-source-and-headers.py \ + libffi.xcodeproj/project.pbxproj \ + libtool-ldflags libtool-version configure.host README.md \ + libffi.map.in LICENSE-BUILDTOOLS msvc_build make_sunver.pl + +# local.exp is generated by configure +DISTCLEANFILES = local.exp +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 # Subdir rules rely on $(FLAGS_TO_PASS) FLAGS_TO_PASS = $(AM_MAKEFLAGS) MAKEOVERRIDES= +pkgconfigdir = $(libdir)/pkgconfig +pkgconfig_DATA = libffi.pc + toolexeclib_LTLIBRARIES = libffi.la noinst_LTLIBRARIES = libffi_convenience.la libffi_la_SOURCES = src/prep_cif.c src/types.c \ src/raw_api.c src/java_raw_api.c src/closures.c -pkgconfigdir = $(libdir)/pkgconfig -pkgconfig_DATA = libffi.pc +if FFI_DEBUG +libffi_la_SOURCES += src/debug.c +endif + +noinst_HEADERS = src/aarch64/ffitarget.h src/aarch64/internal.h \ + src/alpha/ffitarget.h src/alpha/internal.h \ + src/arc/ffitarget.h src/arm/ffitarget.h src/arm/internal.h \ + src/avr32/ffitarget.h src/bfin/ffitarget.h \ + src/cris/ffitarget.h src/csky/ffitarget.h src/frv/ffitarget.h \ + src/ia64/ffitarget.h src/ia64/ia64_flags.h \ + src/m32r/ffitarget.h src/m68k/ffitarget.h \ + src/m88k/ffitarget.h src/metag/ffitarget.h \ + src/microblaze/ffitarget.h src/mips/ffitarget.h \ + src/moxie/ffitarget.h src/nios2/ffitarget.h \ + src/or1k/ffitarget.h src/pa/ffitarget.h \ + src/powerpc/ffitarget.h src/powerpc/asm.h \ + src/powerpc/ffi_powerpc.h src/riscv/ffitarget.h \ + src/s390/ffitarget.h src/s390/internal.h src/sh/ffitarget.h \ + src/sh64/ffitarget.h src/sparc/ffitarget.h \ + src/sparc/internal.h src/tile/ffitarget.h src/vax/ffitarget.h \ + src/x86/ffitarget.h src/x86/internal.h src/x86/internal64.h \ + src/x86/asmnames.h src/xtensa/ffitarget.h src/dlmalloc.c \ + src/kvx/ffitarget.h + +EXTRA_libffi_la_SOURCES = src/aarch64/ffi.c src/aarch64/sysv.S \ + src/aarch64/win64_armasm.S src/alpha/ffi.c src/alpha/osf.S \ + src/arc/ffi.c src/arc/arcompact.S src/arm/ffi.c \ + src/arm/sysv.S src/arm/ffi.c src/arm/sysv_msvc_arm32.S \ + src/avr32/ffi.c src/avr32/sysv.S src/bfin/ffi.c \ + src/bfin/sysv.S src/cris/ffi.c src/cris/sysv.S src/frv/ffi.c \ + src/csky/ffi.c src/csky/sysv.S src/frv/eabi.S src/ia64/ffi.c \ + src/ia64/unix.S src/m32r/ffi.c src/m32r/sysv.S src/m68k/ffi.c \ + src/m68k/sysv.S src/m88k/ffi.c src/m88k/obsd.S \ + src/metag/ffi.c src/metag/sysv.S src/microblaze/ffi.c \ + src/microblaze/sysv.S src/mips/ffi.c src/mips/o32.S \ + src/mips/n32.S src/moxie/ffi.c src/moxie/eabi.S \ + src/nios2/ffi.c src/nios2/sysv.S src/or1k/ffi.c \ + src/or1k/sysv.S src/pa/ffi.c src/pa/linux.S src/pa/hpux32.S \ + src/powerpc/ffi.c src/powerpc/ffi_sysv.c \ + src/powerpc/ffi_linux64.c src/powerpc/sysv.S \ + src/powerpc/linux64.S src/powerpc/linux64_closure.S \ + src/powerpc/ppc_closure.S src/powerpc/aix.S \ + src/powerpc/darwin.S src/powerpc/aix_closure.S \ + src/powerpc/darwin_closure.S src/powerpc/ffi_darwin.c \ + src/riscv/ffi.c src/riscv/sysv.S src/s390/ffi.c \ + src/s390/sysv.S src/sh/ffi.c src/sh/sysv.S src/sh64/ffi.c \ + src/sh64/sysv.S src/sparc/ffi.c src/sparc/ffi64.c \ + src/sparc/v8.S src/sparc/v9.S src/tile/ffi.c src/tile/tile.S \ + src/vax/ffi.c src/vax/elfbsd.S src/x86/ffi.c src/x86/sysv.S \ + src/x86/ffiw64.c src/x86/win64.S src/x86/ffi64.c \ + src/x86/unix64.S src/x86/sysv_intel.S src/x86/win64_intel.S \ + src/xtensa/ffi.c src/xtensa/sysv.S src/kvx/ffi.c \ + src/kvx/sysv.S + +TARGET_OBJ = @TARGET_OBJ@ +libffi_la_LIBADD = $(TARGET_OBJ) + +libffi_convenience_la_SOURCES = $(libffi_la_SOURCES) +EXTRA_libffi_convenience_la_SOURCES = $(EXTRA_libffi_la_SOURCES) +libffi_convenience_la_LIBADD = $(libffi_la_LIBADD) +libffi_convenience_la_DEPENDENCIES = $(libffi_la_DEPENDENCIES) +nodist_libffi_convenience_la_SOURCES = $(nodist_libffi_la_SOURCES) -nodist_libffi_la_SOURCES = +LTLDFLAGS = $(shell $(SHELL) $(top_srcdir)/libtool-ldflags $(LDFLAGS)) +AM_CFLAGS = if FFI_DEBUG -nodist_libffi_la_SOURCES += src/debug.c +# Build debug. Define FFI_DEBUG on the commandline so that, when building with +# MSVC, it can link against the debug CRT. +AM_CFLAGS += -DFFI_DEBUG endif +<<<<<<< HEAD if MIPS nodist_libffi_la_SOURCES += src/mips/ffi.c src/mips/o32.S src/mips/n32.S endif @@ -163,10 +252,24 @@ nodist_libffi_la_SOURCES += src/m88k/ffi.c src/m88k/obsd.S endif if MOXIE nodist_libffi_la_SOURCES += src/moxie/ffi.c src/moxie/eabi.S -endif -if MICROBLAZE -nodist_libffi_la_SOURCES += src/microblaze/ffi.c src/microblaze/sysv.S -endif +======= +if LIBFFI_BUILD_VERSIONED_SHLIB +if LIBFFI_BUILD_VERSIONED_SHLIB_GNU +libffi_version_script = -Wl,--version-script,libffi.map +libffi_version_dep = libffi.map +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 +endif +if LIBFFI_BUILD_VERSIONED_SHLIB_SUN +libffi_version_script = -Wl,-M,libffi.map-sun +libffi_version_dep = libffi.map-sun +libffi.map-sun : libffi.map $(top_srcdir)/make_sunver.pl \ + $(libffi_la_OBJECTS) $(libffi_la_LIBADD) + perl $(top_srcdir)/make_sunver.pl libffi.map \ + `echo $(libffi_la_OBJECTS) $(libffi_la_LIBADD) | \ + sed 's,\([^/ ]*\)\.l\([ao]\),.libs/\1.\2,g'` \ + > $@ || (rm -f $@ ; exit 1) +endif +<<<<<<< HEAD if NIOS2 nodist_libffi_la_SOURCES += src/nios2/sysv.S src/nios2/ffi.c endif @@ -236,10 +339,19 @@ endif if VAX nodist_libffi_la_SOURCES += src/vax/elfbsd.S src/vax/ffi.c endif +======= +else +libffi_version_script = +libffi_version_dep = +endif +libffi_version_info = -version-info `grep -v '^\#' $(srcdir)/libtool-version` +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 -libffi_convenience_la_SOURCES = $(libffi_la_SOURCES) -nodist_libffi_convenience_la_SOURCES = $(nodist_libffi_la_SOURCES) +libffi.map: $(top_srcdir)/libffi.map.in + $(COMPILE) -D$(TARGET) -DGENERATE_LIBFFI_MAP \ + -E -x assembler-with-cpp -o $@ $(top_srcdir)/libffi.map.in +<<<<<<< HEAD LTLDFLAGS = $(shell $(SHELL) $(top_srcdir)/libtool-ldflags $(LDFLAGS)) AM_CFLAGS = @@ -250,9 +362,20 @@ AM_CFLAGS += -DFFI_DEBUG endif libffi_la_LDFLAGS = -no-undefined -version-info `grep -v '^\#' $(srcdir)/libtool-version` $(LTLDFLAGS) $(AM_LTLDFLAGS) +======= +libffi_la_LDFLAGS = -no-undefined $(libffi_version_info) $(libffi_version_script) $(LTLDFLAGS) $(AM_LTLDFLAGS) +libffi_la_DEPENDENCIES = $(libffi_la_LIBADD) $(libffi_version_dep) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 AM_CPPFLAGS = -I. -I$(top_srcdir)/include -Iinclude -I$(top_srcdir)/src AM_CCASFLAGS = $(AM_CPPFLAGS) dist-hook: +<<<<<<< HEAD + if [ -d $(top_srcdir)/.git ] ; then (cd $(top_srcdir); git log --no-decorate) ; else echo 'See git log for history.' ; fi > $(distdir)/ChangeLog +======= + d=`(cd $(distdir); pwd)`; (cd doc; make pdf; cp *.pdf $$d/doc) if [ -d $(top_srcdir)/.git ] ; then (cd $(top_srcdir); git log --no-decorate) ; else echo 'See git log for history.' ; fi > $(distdir)/ChangeLog + s=`awk '/was released on/{ print NR; exit}' $(top_srcdir)/README.md`; tail -n +$$(($$s-1)) $(top_srcdir)/README.md > $(distdir)/README.md + +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 diff --git a/native/libffi/README.md b/native/libffi/README.md new file mode 100644 index 0000000000..4225d85345 --- /dev/null +++ b/native/libffi/README.md @@ -0,0 +1,486 @@ +Status +====== + +[![Build Status](https://travis-ci.org/libffi/libffi.svg?branch=master)](https://travis-ci.org/libffi/libffi) +[![Build status](https://ci.appveyor.com/api/projects/status/8lko9vagbx4w2kxq?svg=true)](https://ci.appveyor.com/project/atgreen/libffi) + +libffi-3.4 was released on TBD. Check the libffi web +page for updates: . + + +What is libffi? +=============== + +Compilers for high level languages generate code that follow certain +conventions. These conventions are necessary, in part, for separate +compilation to work. One such convention is the "calling +convention". The "calling convention" is essentially a set of +assumptions made by the compiler about where function arguments will +be found on entry to a function. A "calling convention" also specifies +where the return value for a function is found. + +Some programs may not know at the time of compilation what arguments +are to be passed to a function. For instance, an interpreter may be +told at run-time about the number and types of arguments used to call +a given function. Libffi can be used in such programs to provide a +bridge from the interpreter program to compiled code. + +The libffi library provides a portable, high level programming +interface to various calling conventions. This allows a programmer to +call any function specified by a call interface description at run +time. + +FFI stands for Foreign Function Interface. A foreign function +interface is the popular name for the interface that allows code +written in one language to call code written in another language. The +libffi library really only provides the lowest, machine dependent +layer of a fully featured foreign function interface. A layer must +exist above libffi that handles type conversions for values passed +between the two languages. + + +Supported Platforms +=================== + +Libffi has been ported to many different platforms. + +At the time of release, the following basic configurations have been +tested: + +| Architecture | Operating System | Compiler | +| --------------- | ---------------- | ----------------------- | +| AArch64 (ARM64) | iOS | Clang | +| AArch64 | Linux | GCC | +| AArch64 | Windows | MSVC | +| Alpha | Linux | GCC | +| Alpha | Tru64 | GCC | +| ARC | Linux | GCC | +| ARM | Linux | GCC | +| ARM | iOS | GCC | +| ARM | Windows | MSVC | +| AVR32 | Linux | GCC | +| Blackfin | uClinux | GCC | +| CSKY | Linux | GCC | +| HPPA | HPUX | GCC | +| KVX | Linux | GCC | +| IA-64 | Linux | GCC | +| M68K | FreeMiNT | GCC | +| M68K | Linux | GCC | +| M68K | RTEMS | GCC | +| M88K | OpenBSD/mvme88k | GCC | +| Meta | Linux | GCC | +| MicroBlaze | Linux | GCC | +| MIPS | IRIX | GCC | +| MIPS | Linux | GCC | +| MIPS | RTEMS | GCC | +| MIPS64 | Linux | GCC | +| Moxie | Bare metal | GCC | +| Nios II | Linux | GCC | +| OpenRISC | Linux | GCC | +| PowerPC 32-bit | AIX | IBM XL C | +| PowerPC 64-bit | AIX | IBM XL C | +| PowerPC | AMIGA | GCC | +| PowerPC | Linux | GCC | +| PowerPC | Mac OSX | GCC | +| PowerPC | FreeBSD | GCC | +| PowerPC 64-bit | FreeBSD | GCC | +| PowerPC 64-bit | Linux ELFv1 | GCC | +| PowerPC 64-bit | Linux ELFv2 | GCC | +| RISC-V 32-bit | Linux | GCC | +| RISC-V 64-bit | Linux | GCC | +| S390 | Linux | GCC | +| S390X | Linux | GCC | +| SPARC | Linux | GCC | +| SPARC | Solaris | GCC | +| SPARC | Solaris | Oracle Solaris Studio C | +| SPARC64 | Linux | GCC | +| SPARC64 | FreeBSD | GCC | +| SPARC64 | Solaris | Oracle Solaris Studio C | +| TILE-Gx/TILEPro | Linux | GCC | +| VAX | OpenBSD/vax | GCC | +| X86 | FreeBSD | GCC | +| X86 | GNU HURD | GCC | +| X86 | Interix | GCC | +| X86 | kFreeBSD | GCC | +| X86 | Linux | GCC | +| X86 | OpenBSD | GCC | +| X86 | OS/2 | GCC | +| X86 | Solaris | GCC | +| X86 | Solaris | Oracle Solaris Studio C | +| X86 | Windows/Cygwin | GCC | +| X86 | Windows/MingW | GCC | +| X86-64 | FreeBSD | GCC | +| X86-64 | Linux | GCC | +| X86-64 | Linux/x32 | GCC | +| X86-64 | OpenBSD | GCC | +| X86-64 | Solaris | Oracle Solaris Studio C | +| X86-64 | Windows/Cygwin | GCC | +| X86-64 | Windows/MingW | GCC | +| X86-64 | Mac OSX | GCC | +| Xtensa | Linux | GCC | + +Please send additional platform test results to +libffi-discuss@sourceware.org. + +Installing libffi +================= + +First you must configure the distribution for your particular +system. Go to the directory you wish to build libffi in and run the +"configure" program found in the root directory of the libffi source +distribution. Note that building libffi requires a C99 compatible +compiler. + +If you're building libffi directly from git hosted sources, configure +won't exist yet; run ./autogen.sh first. This will require that you +install autoconf, automake and libtool. + +You may want to tell configure where to install the libffi library and +header files. To do that, use the ``--prefix`` configure switch. Libffi +will install under /usr/local by default. + +If you want to enable extra run-time debugging checks use the the +``--enable-debug`` configure switch. This is useful when your program dies +mysteriously while using libffi. + +Another useful configure switch is ``--enable-purify-safety``. Using this +will add some extra code which will suppress certain warnings when you +are using Purify with libffi. Only use this switch when using +Purify, as it will slow down the library. + +If you don't want to build documentation, use the ``--disable-docs`` +configure switch. + +It's also possible to build libffi on Windows platforms with +Microsoft's Visual C++ compiler. In this case, use the msvcc.sh +wrapper script during configuration like so: + + path/to/configure CC=path/to/msvcc.sh CXX=path/to/msvcc.sh LD=link CPP="cl -nologo -EP" CPPFLAGS="-DFFI_BUILDING_DLL" + +For 64-bit Windows builds, use ``CC="path/to/msvcc.sh -m64"`` and +``CXX="path/to/msvcc.sh -m64"``. You may also need to specify +``--build`` appropriately. + +It is also possible to build libffi on Windows platforms with the LLVM +project's clang-cl compiler, like below: + + path/to/configure CC="path/to/msvcc.sh -clang-cl" CXX="path/to/msvcc.sh -clang-cl" LD=link CPP="clang-cl -EP" + +When building with MSVC under a MingW environment, you may need to +remove the line in configure that sets 'fix_srcfile_path' to a 'cygpath' +command. ('cygpath' is not present in MingW, and is not required when +using MingW-style paths.) + +To build static library for ARM64 with MSVC using visual studio solution, msvc_build folder have + aarch64/Ffi_staticLib.sln + required header files in aarch64/aarch64_include/ + + +SPARC Solaris builds require the use of the GNU assembler and linker. +Point ``AS`` and ``LD`` environment variables at those tool prior to +configuration. + +For iOS builds, the ``libffi.xcodeproj`` Xcode project is available. + +Configure has many other options. Use ``configure --help`` to see them all. + +Once configure has finished, type "make". Note that you must be using +GNU make. You can ftp GNU make from ftp.gnu.org:/pub/gnu/make . + +To ensure that libffi is working as advertised, type "make check". +This will require that you have DejaGNU installed. + +To install the library and header files, type ``make install``. + + +History +======= + +See the git log for details at http://github.com/libffi/libffi. + + 3.4 TBD + Add support for Alibaba's CSKY architecture. + Add support for Intel Control-flow Enforcement Technology (CET). + Add support for ARM Pointer Authentication (PA). + Fix 32-bit PPC regression. + Fix MIPS soft-float problem. + + 3.3 Nov-23-19 + Add RISC-V support. + New API in support of GO closures. + Add IEEE754 binary128 long double support for 64-bit Power + Default to Microsoft's 64 bit long double ABI with Visual C++. + GNU compiler uses 80 bits (128 in memory) FFI_GNUW64 ABI. + Add Windows on ARM64 (WOA) support. + Add Windows 32-bit ARM support. + Raw java (gcj) API deprecated. + Add pre-built PDF documentation to source distribution. + Many new test cases and bug fixes. + + 3.2.1 Nov-12-14 + Build fix for non-iOS AArch64 targets. + + 3.2 Nov-11-14 + Add C99 Complex Type support (currently only supported on + s390). + Add support for PASCAL and REGISTER calling conventions on x86 + Windows/Linux. + Add OpenRISC and Cygwin-64 support. + Bug fixes. + + 3.1 May-19-14 + Add AArch64 (ARM64) iOS support. + Add Nios II support. + Add m88k and DEC VAX support. + Add support for stdcall, thiscall, and fastcall on non-Windows + 32-bit x86 targets such as Linux. + Various Android, MIPS N32, x86, FreeBSD and UltraSPARC IIi + fixes. + Make the testsuite more robust: eliminate several spurious + failures, and respect the $CC and $CXX environment variables. + Archive off the manually maintained ChangeLog in favor of git + log. + + 3.0.13 Mar-17-13 + Add Meta support. + Add missing Moxie bits. + Fix stack alignment bug on 32-bit x86. + Build fix for m68000 targets. + Build fix for soft-float Power targets. + Fix the install dir location for some platforms when building + with GCC (OS X, Solaris). + Fix Cygwin regression. + + 3.0.12 Feb-11-13 + Add Moxie support. + Add AArch64 support. + Add Blackfin support. + Add TILE-Gx/TILEPro support. + Add MicroBlaze support. + Add Xtensa support. + Add support for PaX enabled kernels with MPROTECT. + Add support for native vendor compilers on + Solaris and AIX. + Work around LLVM/GCC interoperability issue on x86_64. + + 3.0.11 Apr-11-12 + Lots of build fixes. + Add support for variadic functions (ffi_prep_cif_var). + Add Linux/x32 support. + Add thiscall, fastcall and MSVC cdecl support on Windows. + Add Amiga and newer MacOS support. + Add m68k FreeMiNT support. + Integration with iOS' xcode build tools. + Fix Octeon and MC68881 support. + Fix code pessimizations. + + 3.0.10 Aug-23-11 + Add support for Apple's iOS. + Add support for ARM VFP ABI. + Add RTEMS support for MIPS and M68K. + Fix instruction cache clearing problems on + ARM and SPARC. + Fix the N64 build on mips-sgi-irix6.5. + Enable builds with Microsoft's compiler. + Enable x86 builds with Oracle's Solaris compiler. + Fix support for calling code compiled with Oracle's Sparc + Solaris compiler. + Testsuite fixes for Tru64 Unix. + Additional platform support. + + 3.0.9 Dec-31-09 + Add AVR32 and win64 ports. Add ARM softfp support. + Many fixes for AIX, Solaris, HP-UX, *BSD. + Several PowerPC and x86-64 bug fixes. + Build DLL for windows. + + 3.0.8 Dec-19-08 + Add *BSD, BeOS, and PA-Linux support. + + 3.0.7 Nov-11-08 + Fix for ppc FreeBSD. + (thanks to Andreas Tobler) + + 3.0.6 Jul-17-08 + Fix for closures on sh. + Mark the sh/sh64 stack as non-executable. + (both thanks to Kaz Kojima) + + 3.0.5 Apr-3-08 + Fix libffi.pc file. + Fix #define ARM for IcedTea users. + Fix x86 closure bug. + + 3.0.4 Feb-24-08 + Fix x86 OpenBSD configury. + + 3.0.3 Feb-22-08 + Enable x86 OpenBSD thanks to Thomas Heller, and + x86-64 FreeBSD thanks to Björn König and Andreas Tobler. + Clean up test instruction in README. + + 3.0.2 Feb-21-08 + Improved x86 FreeBSD support. + Thanks to Björn König. + + 3.0.1 Feb-15-08 + Fix instruction cache flushing bug on MIPS. + Thanks to David Daney. + + 3.0.0 Feb-15-08 + Many changes, mostly thanks to the GCC project. + Cygnus Solutions is now Red Hat. + + [10 years go by...] + + 1.20 Oct-5-98 + Raffaele Sena produces ARM port. + + 1.19 Oct-5-98 + Fixed x86 long double and long long return support. + m68k bug fixes from Andreas Schwab. + Patch for DU assembler compatibility for the Alpha from Richard + Henderson. + + 1.18 Apr-17-98 + Bug fixes and MIPS configuration changes. + + 1.17 Feb-24-98 + Bug fixes and m68k port from Andreas Schwab. PowerPC port from + Geoffrey Keating. Various bug x86, Sparc and MIPS bug fixes. + + 1.16 Feb-11-98 + Richard Henderson produces Alpha port. + + 1.15 Dec-4-97 + Fixed an n32 ABI bug. New libtool, auto* support. + + 1.14 May-13-97 + libtool is now used to generate shared and static libraries. + Fixed a minor portability problem reported by Russ McManus + . + + 1.13 Dec-2-96 + Added --enable-purify-safety to keep Purify from complaining + about certain low level code. + Sparc fix for calling functions with < 6 args. + Linux x86 a.out fix. + + 1.12 Nov-22-96 + Added missing ffi_type_void, needed for supporting void return + types. Fixed test case for non MIPS machines. Cygnus Support + is now Cygnus Solutions. + + 1.11 Oct-30-96 + Added notes about GNU make. + + 1.10 Oct-29-96 + Added configuration fix for non GNU compilers. + + 1.09 Oct-29-96 + Added --enable-debug configure switch. Clean-ups based on LCLint + feedback. ffi_mips.h is always installed. Many configuration + fixes. Fixed ffitest.c for sparc builds. + + 1.08 Oct-15-96 + Fixed n32 problem. Many clean-ups. + + 1.07 Oct-14-96 + Gordon Irlam rewrites v8.S again. Bug fixes. + + 1.06 Oct-14-96 + Gordon Irlam improved the sparc port. + + 1.05 Oct-14-96 + Interface changes based on feedback. + + 1.04 Oct-11-96 + Sparc port complete (modulo struct passing bug). + + 1.03 Oct-10-96 + Passing struct args, and returning struct values works for + all architectures/calling conventions. Expanded tests. + + 1.02 Oct-9-96 + Added SGI n32 support. Fixed bugs in both o32 and Linux support. + Added "make test". + + 1.01 Oct-8-96 + Fixed float passing bug in mips version. Restructured some + of the code. Builds cleanly with SGI tools. + + 1.00 Oct-7-96 + First release. No public announcement. + +Authors & Credits +================= + +libffi was originally written by Anthony Green . + +The developers of the GNU Compiler Collection project have made +innumerable valuable contributions. See the ChangeLog file for +details. + +Some of the ideas behind libffi were inspired by Gianni Mariani's free +gencall library for Silicon Graphics machines. + +The closure mechanism was designed and implemented by Kresten Krab +Thorup. + +Major processor architecture ports were contributed by the following +developers: + + aarch64 Marcus Shawcroft, James Greenhalgh + alpha Richard Henderson + arc Hackers at Synopsis + arm Raffaele Sena + avr32 Bradley Smith + blackfin Alexandre Keunecke I. de Mendonca + cris Simon Posnjak, Hans-Peter Nilsson + csky Ma Jun, Zhang Wenmeng + frv Anthony Green + ia64 Hans Boehm + m32r Kazuhiro Inaoka + m68k Andreas Schwab + m88k Miod Vallat + metag Hackers at Imagination Technologies + microblaze Nathan Rossi + mips Anthony Green, Casey Marshall + mips64 David Daney + moxie Anthony Green + nios ii Sandra Loosemore + openrisc Sebastian Macke + pa Randolph Chung, Dave Anglin, Andreas Tobler + powerpc Geoffrey Keating, Andreas Tobler, + David Edelsohn, John Hornkvist + powerpc64 Jakub Jelinek + riscv Michael Knyszek, Andrew Waterman, Stef O'Rear + s390 Gerhard Tonn, Ulrich Weigand + sh Kaz Kojima + sh64 Kaz Kojima + sparc Anthony Green, Gordon Irlam + tile-gx/tilepro Walter Lee + vax Miod Vallat + x86 Anthony Green, Jon Beniston + x86-64 Bo Thorsen + xtensa Chris Zankel + +Jesper Skov and Andrew Haley both did more than their fair share of +stepping through the code and tracking down bugs. + +Thanks also to Tom Tromey for bug fixes, documentation and +configuration help. + +Thanks to Jim Blandy, who provided some useful feedback on the libffi +interface. + +Andreas Tobler has done a tremendous amount of work on the testsuite. + +Alex Oliva solved the executable page problem for SElinux. + +The list above is almost certainly incomplete and inaccurate. I'm +happy to make corrections or additions upon request. + +If you have a problem, or have found a bug, please send a note to the +author at green@moxielogic.com, or the project mailing list at +libffi-discuss@sourceware.org. diff --git a/native/libffi/acinclude.m4 b/native/libffi/acinclude.m4 index 3e8f8ba570..1a70efb010 100644 --- a/native/libffi/acinclude.m4 +++ b/native/libffi/acinclude.m4 @@ -90,3 +90,390 @@ if test $ac_cv_func_mmap_anon = yes; then [Define if mmap with MAP_ANON(YMOUS) works.]) fi ]) + +dnl ---------------------------------------------------------------------- +dnl This whole bit snagged from libstdc++-v3, via libatomic. + +dnl +dnl LIBFFI_ENABLE +dnl (FEATURE, DEFAULT, HELP-ARG, HELP-STRING) +dnl (FEATURE, DEFAULT, HELP-ARG, HELP-STRING, permit a|b|c) +dnl (FEATURE, DEFAULT, HELP-ARG, HELP-STRING, SHELL-CODE-HANDLER) +dnl +dnl See docs/html/17_intro/configury.html#enable for documentation. +dnl +m4_define([LIBFFI_ENABLE],[dnl +m4_define([_g_switch],[--enable-$1])dnl +m4_define([_g_help],[AC_HELP_STRING(_g_switch$3,[$4 @<:@default=$2@:>@])])dnl + AC_ARG_ENABLE($1,_g_help, + m4_bmatch([$5], + [^permit ], + [[ + case "$enableval" in + m4_bpatsubst([$5],[permit ])) ;; + *) AC_MSG_ERROR(Unknown argument to enable/disable $1) ;; + dnl Idea for future: generate a URL pointing to + dnl "onlinedocs/configopts.html#whatever" + esac + ]], + [^$], + [[ + case "$enableval" in + yes|no) ;; + *) AC_MSG_ERROR(Argument to enable/disable $1 must be yes or no) ;; + esac + ]], + [[$5]]), + [enable_]m4_bpatsubst([$1],-,_)[=][$2]) +m4_undefine([_g_switch])dnl +m4_undefine([_g_help])dnl +]) + +dnl +dnl If GNU ld is in use, check to see if tricky linker opts can be used. If +dnl the native linker is in use, all variables will be defined to something +dnl safe (like an empty string). +dnl +dnl Defines: +dnl SECTION_LDFLAGS='-Wl,--gc-sections' if possible +dnl OPT_LDFLAGS='-Wl,-O1' if possible +dnl LD (as a side effect of testing) +dnl Sets: +dnl with_gnu_ld +dnl libat_ld_is_gold (possibly) +dnl libat_gnu_ld_version (possibly) +dnl +dnl The last will be a single integer, e.g., version 1.23.45.0.67.89 will +dnl set libat_gnu_ld_version to 12345. Zeros cause problems. +dnl +AC_DEFUN([LIBFFI_CHECK_LINKER_FEATURES], [ + # If we're not using GNU ld, then there's no point in even trying these + # tests. Check for that first. We should have already tested for gld + # by now (in libtool), but require it now just to be safe... + test -z "$SECTION_LDFLAGS" && SECTION_LDFLAGS='' + test -z "$OPT_LDFLAGS" && OPT_LDFLAGS='' + AC_REQUIRE([AC_PROG_LD]) + AC_REQUIRE([AC_PROG_AWK]) + + # The name set by libtool depends on the version of libtool. Shame on us + # for depending on an impl detail, but c'est la vie. Older versions used + # ac_cv_prog_gnu_ld, but now it's lt_cv_prog_gnu_ld, and is copied back on + # top of with_gnu_ld (which is also set by --with-gnu-ld, so that actually + # makes sense). We'll test with_gnu_ld everywhere else, so if that isn't + # set (hence we're using an older libtool), then set it. + if test x${with_gnu_ld+set} != xset; then + if test x${ac_cv_prog_gnu_ld+set} != xset; then + # We got through "ac_require(ac_prog_ld)" and still not set? Huh? + with_gnu_ld=no + else + with_gnu_ld=$ac_cv_prog_gnu_ld + fi + fi + + # Start by getting the version number. I think the libtool test already + # does some of this, but throws away the result. + libat_ld_is_gold=no + if $LD --version 2>/dev/null | grep 'GNU gold'> /dev/null 2>&1; then + libat_ld_is_gold=yes + fi + changequote(,) + ldver=`$LD --version 2>/dev/null | + sed -e 's/GNU gold /GNU ld /;s/GNU ld version /GNU ld /;s/GNU ld ([^)]*) /GNU ld /;s/GNU ld \([0-9.][0-9.]*\).*/\1/; q'` + changequote([,]) + libat_gnu_ld_version=`echo $ldver | \ + $AWK -F. '{ if (NF<3) [$]3=0; print ([$]1*100+[$]2)*100+[$]3 }'` + + # Set --gc-sections. + if test "$with_gnu_ld" = "notbroken"; then + # GNU ld it is! Joy and bunny rabbits! + + # All these tests are for C++; save the language and the compiler flags. + # Need to do this so that g++ won't try to link in libstdc++ + ac_test_CFLAGS="${CFLAGS+set}" + ac_save_CFLAGS="$CFLAGS" + CFLAGS='-x c++ -Wl,--gc-sections' + + # Check for -Wl,--gc-sections + # XXX This test is broken at the moment, as symbols required for linking + # are now in libsupc++ (not built yet). In addition, this test has + # cored on solaris in the past. In addition, --gc-sections doesn't + # really work at the moment (keeps on discarding used sections, first + # .eh_frame and now some of the glibc sections for iconv). + # Bzzzzt. Thanks for playing, maybe next time. + AC_MSG_CHECKING([for ld that supports -Wl,--gc-sections]) + AC_TRY_RUN([ + int main(void) + { + try { throw 1; } + catch (...) { }; + return 0; + } + ], [ac_sectionLDflags=yes],[ac_sectionLDflags=no], [ac_sectionLDflags=yes]) + if test "$ac_test_CFLAGS" = set; then + CFLAGS="$ac_save_CFLAGS" + else + # this is the suspicious part + CFLAGS='' + fi + if test "$ac_sectionLDflags" = "yes"; then + SECTION_LDFLAGS="-Wl,--gc-sections $SECTION_LDFLAGS" + fi + AC_MSG_RESULT($ac_sectionLDflags) + fi + + # Set linker optimization flags. + if test x"$with_gnu_ld" = x"yes"; then + OPT_LDFLAGS="-Wl,-O1 $OPT_LDFLAGS" + fi + + AC_SUBST(SECTION_LDFLAGS) + AC_SUBST(OPT_LDFLAGS) +]) + + +dnl +dnl If GNU ld is in use, check to see if tricky linker opts can be used. If +dnl the native linker is in use, all variables will be defined to something +dnl safe (like an empty string). +dnl +dnl Defines: +dnl SECTION_LDFLAGS='-Wl,--gc-sections' if possible +dnl OPT_LDFLAGS='-Wl,-O1' if possible +dnl LD (as a side effect of testing) +dnl Sets: +dnl with_gnu_ld +dnl libat_ld_is_gold (possibly) +dnl libat_gnu_ld_version (possibly) +dnl +dnl The last will be a single integer, e.g., version 1.23.45.0.67.89 will +dnl set libat_gnu_ld_version to 12345. Zeros cause problems. +dnl +AC_DEFUN([LIBFFI_CHECK_LINKER_FEATURES], [ + # If we're not using GNU ld, then there's no point in even trying these + # tests. Check for that first. We should have already tested for gld + # by now (in libtool), but require it now just to be safe... + test -z "$SECTION_LDFLAGS" && SECTION_LDFLAGS='' + test -z "$OPT_LDFLAGS" && OPT_LDFLAGS='' + AC_REQUIRE([AC_PROG_LD]) + AC_REQUIRE([AC_PROG_AWK]) + + # The name set by libtool depends on the version of libtool. Shame on us + # for depending on an impl detail, but c'est la vie. Older versions used + # ac_cv_prog_gnu_ld, but now it's lt_cv_prog_gnu_ld, and is copied back on + # top of with_gnu_ld (which is also set by --with-gnu-ld, so that actually + # makes sense). We'll test with_gnu_ld everywhere else, so if that isn't + # set (hence we're using an older libtool), then set it. + if test x${with_gnu_ld+set} != xset; then + if test x${ac_cv_prog_gnu_ld+set} != xset; then + # We got through "ac_require(ac_prog_ld)" and still not set? Huh? + with_gnu_ld=no + else + with_gnu_ld=$ac_cv_prog_gnu_ld + fi + fi + + # Start by getting the version number. I think the libtool test already + # does some of this, but throws away the result. + libat_ld_is_gold=no + if $LD --version 2>/dev/null | grep 'GNU gold'> /dev/null 2>&1; then + libat_ld_is_gold=yes + fi + libat_ld_is_lld=no + if $LD --version 2>/dev/null | grep 'LLD '> /dev/null 2>&1; then + libat_ld_is_lld=yes + fi + changequote(,) + ldver=`$LD --version 2>/dev/null | + sed -e 's/GNU gold /GNU ld /;s/GNU ld version /GNU ld /;s/GNU ld ([^)]*) /GNU ld /;s/GNU ld \([0-9.][0-9.]*\).*/\1/; q'` + changequote([,]) + libat_gnu_ld_version=`echo $ldver | \ + $AWK -F. '{ if (NF<3) [$]3=0; print ([$]1*100+[$]2)*100+[$]3 }'` + + # Set --gc-sections. + if test "$with_gnu_ld" = "notbroken"; then + # GNU ld it is! Joy and bunny rabbits! + + # All these tests are for C++; save the language and the compiler flags. + # Need to do this so that g++ won't try to link in libstdc++ + ac_test_CFLAGS="${CFLAGS+set}" + ac_save_CFLAGS="$CFLAGS" + CFLAGS='-x c++ -Wl,--gc-sections' + + # Check for -Wl,--gc-sections + # XXX This test is broken at the moment, as symbols required for linking + # are now in libsupc++ (not built yet). In addition, this test has + # cored on solaris in the past. In addition, --gc-sections doesn't + # really work at the moment (keeps on discarding used sections, first + # .eh_frame and now some of the glibc sections for iconv). + # Bzzzzt. Thanks for playing, maybe next time. + AC_MSG_CHECKING([for ld that supports -Wl,--gc-sections]) + AC_TRY_RUN([ + int main(void) + { + try { throw 1; } + catch (...) { }; + return 0; + } + ], [ac_sectionLDflags=yes],[ac_sectionLDflags=no], [ac_sectionLDflags=yes]) + if test "$ac_test_CFLAGS" = set; then + CFLAGS="$ac_save_CFLAGS" + else + # this is the suspicious part + CFLAGS='' + fi + if test "$ac_sectionLDflags" = "yes"; then + SECTION_LDFLAGS="-Wl,--gc-sections $SECTION_LDFLAGS" + fi + AC_MSG_RESULT($ac_sectionLDflags) + fi + + # Set linker optimization flags. + if test x"$with_gnu_ld" = x"yes"; then + OPT_LDFLAGS="-Wl,-O1 $OPT_LDFLAGS" + fi + + AC_SUBST(SECTION_LDFLAGS) + AC_SUBST(OPT_LDFLAGS) +]) + + +dnl +dnl Add version tags to symbols in shared library (or not), additionally +dnl marking other symbols as private/local (or not). +dnl +dnl --enable-symvers=style adds a version script to the linker call when +dnl creating the shared library. The choice of version script is +dnl controlled by 'style'. +dnl --disable-symvers does not. +dnl + Usage: LIBFFI_ENABLE_SYMVERS[(DEFAULT)] +dnl Where DEFAULT is either 'yes' or 'no'. Passing `yes' tries to +dnl choose a default style based on linker characteristics. Passing +dnl 'no' disables versioning. +dnl +AC_DEFUN([LIBFFI_ENABLE_SYMVERS], [ + +LIBFFI_ENABLE(symvers,yes,[=STYLE], + [enables symbol versioning of the shared library], + [permit yes|no|gnu*|sun]) + +# If we never went through the LIBFFI_CHECK_LINKER_FEATURES macro, then we +# don't know enough about $LD to do tricks... +AC_REQUIRE([LIBFFI_CHECK_LINKER_FEATURES]) + +# Turn a 'yes' into a suitable default. +if test x$enable_symvers = xyes ; then + # FIXME The following test is too strict, in theory. + if test $enable_shared = no || test "x$LD" = x; then + enable_symvers=no + else + if test $with_gnu_ld = yes ; then + enable_symvers=gnu + else + case ${target_os} in + # Sun symbol versioning exists since Solaris 2.5. + solaris2.[[5-9]]* | solaris2.1[[0-9]]*) + enable_symvers=sun ;; + *) + enable_symvers=no ;; + esac + fi + fi +fi + +# Check if 'sun' was requested on non-Solaris 2 platforms. +if test x$enable_symvers = xsun ; then + case ${target_os} in + solaris2*) + # All fine. + ;; + *) + # Unlikely to work. + AC_MSG_WARN([=== You have requested Sun symbol versioning, but]) + AC_MSG_WARN([=== you are not targetting Solaris 2.]) + AC_MSG_WARN([=== Symbol versioning will be disabled.]) + enable_symvers=no + ;; + esac +fi + +# Check to see if libgcc_s exists, indicating that shared libgcc is possible. +if test $enable_symvers != no; then + AC_MSG_CHECKING([for shared libgcc]) + ac_save_CFLAGS="$CFLAGS" + CFLAGS=' -lgcc_s' + AC_TRY_LINK(, [return 0;], libat_shared_libgcc=yes, libat_shared_libgcc=no) + CFLAGS="$ac_save_CFLAGS" + if test $libat_shared_libgcc = no; then + cat > conftest.c <&1 >/dev/null \ + | sed -n 's/^.* -lgcc_s\([^ ]*\) .*$/\1/p'` +changequote([,])dnl + rm -f conftest.c conftest.so + if test x${libat_libgcc_s_suffix+set} = xset; then + CFLAGS=" -lgcc_s$libat_libgcc_s_suffix" + AC_TRY_LINK(, [return 0;], libat_shared_libgcc=yes) + CFLAGS="$ac_save_CFLAGS" + fi + fi + AC_MSG_RESULT($libat_shared_libgcc) +fi + +# For GNU ld, we need at least this version. The format is described in +# LIBFFI_CHECK_LINKER_FEATURES above. +libat_min_gnu_ld_version=21400 +# XXXXXXXXXXX libat_gnu_ld_version=21390 + +# Check to see if unspecified "yes" value can win, given results above. +# Change "yes" into either "no" or a style name. +if test $enable_symvers != no && test $libat_shared_libgcc = yes; then + if test $with_gnu_ld = yes; then + if test $libat_gnu_ld_version -ge $libat_min_gnu_ld_version ; then + enable_symvers=gnu + elif test $libat_ld_is_gold = yes ; then + enable_symvers=gnu + elif test $libat_ld_is_lld = yes ; then + enable_symvers=gnu + else + # The right tools, the right setup, but too old. Fallbacks? + AC_MSG_WARN(=== Linker version $libat_gnu_ld_version is too old for) + AC_MSG_WARN(=== full symbol versioning support in this release of GCC.) + AC_MSG_WARN(=== You would need to upgrade your binutils to version) + AC_MSG_WARN(=== $libat_min_gnu_ld_version or later and rebuild GCC.) + if test $libat_gnu_ld_version -ge 21200 ; then + # Globbing fix is present, proper block support is not. + dnl AC_MSG_WARN([=== Dude, you are soooo close. Maybe we can fake it.]) + dnl enable_symvers=??? + AC_MSG_WARN([=== Symbol versioning will be disabled.]) + enable_symvers=no + else + # 2.11 or older. + AC_MSG_WARN([=== Symbol versioning will be disabled.]) + enable_symvers=no + fi + fi + elif test $enable_symvers = sun; then + : All interesting versions of Sun ld support sun style symbol versioning. + else + # just fail for now + AC_MSG_WARN([=== You have requested some kind of symbol versioning, but]) + AC_MSG_WARN([=== either you are not using a supported linker, or you are]) + AC_MSG_WARN([=== not building a shared libgcc_s (which is required).]) + AC_MSG_WARN([=== Symbol versioning will be disabled.]) + enable_symvers=no + fi +fi +if test $enable_symvers = gnu; then + AC_DEFINE(LIBFFI_GNU_SYMBOL_VERSIONING, 1, + [Define to 1 if GNU symbol versioning is used for libatomic.]) +fi + +AM_CONDITIONAL(LIBFFI_BUILD_VERSIONED_SHLIB, test $enable_symvers != no) +AM_CONDITIONAL(LIBFFI_BUILD_VERSIONED_SHLIB_GNU, test $enable_symvers = gnu) +AM_CONDITIONAL(LIBFFI_BUILD_VERSIONED_SHLIB_SUN, test $enable_symvers = sun) +AC_MSG_NOTICE(versioning on shared library symbols is $enable_symvers) +]) diff --git a/native/libffi/autogen.sh b/native/libffi/autogen.sh index 822ff97d88..f32a4c18c9 100755 --- a/native/libffi/autogen.sh +++ b/native/libffi/autogen.sh @@ -1,2 +1,6 @@ #!/bin/sh -exec autoreconf -v -f -i \ No newline at end of file +<<<<<<< HEAD +exec autoreconf -v -f -i +======= +exec autoreconf -v -i +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 diff --git a/native/libffi/config.guess b/native/libffi/config.guess new file mode 100644 index 0000000000..e94095c5fb --- /dev/null +++ b/native/libffi/config.guess @@ -0,0 +1,1687 @@ +#! /bin/sh +# Attempt to guess a canonical system name. +# Copyright 1992-2020 Free Software Foundation, Inc. + +timestamp='2020-07-12' + +# This file is free software; you can redistribute it and/or modify it +# under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, but +# WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; if not, see . +# +# As a special exception to the GNU General Public License, if you +# distribute this file as part of a program that contains a +# configuration script generated by Autoconf, you may include it under +# the same distribution terms that you use for the rest of that +# program. This Exception is an additional permission under section 7 +# of the GNU General Public License, version 3 ("GPLv3"). +# +# Originally written by Per Bothner; maintained since 2000 by Ben Elliston. +# +# You can get the latest version of this script from: +# https://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess +# +# Please send patches to . + + +me=`echo "$0" | sed -e 's,.*/,,'` + +usage="\ +Usage: $0 [OPTION] + +Output the configuration name of the system \`$me' is run on. + +Options: + -h, --help print this help, then exit + -t, --time-stamp print date of last modification, then exit + -v, --version print version number, then exit + +Report bugs and patches to ." + +version="\ +GNU config.guess ($timestamp) + +Originally written by Per Bothner. +Copyright 1992-2020 Free Software Foundation, Inc. + +This is free software; see the source for copying conditions. There is NO +warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE." + +help=" +Try \`$me --help' for more information." + +# Parse command line +while test $# -gt 0 ; do + case $1 in + --time-stamp | --time* | -t ) + echo "$timestamp" ; exit ;; + --version | -v ) + echo "$version" ; exit ;; + --help | --h* | -h ) + echo "$usage"; exit ;; + -- ) # Stop option processing + shift; break ;; + - ) # Use stdin as input. + break ;; + -* ) + echo "$me: invalid option $1$help" >&2 + exit 1 ;; + * ) + break ;; + esac +done + +if test $# != 0; then + echo "$me: too many arguments$help" >&2 + exit 1 +fi + +# CC_FOR_BUILD -- compiler used by this script. Note that the use of a +# compiler to aid in system detection is discouraged as it requires +# temporary files to be created and, as you can see below, it is a +# headache to deal with in a portable fashion. + +# Historically, `CC_FOR_BUILD' used to be named `HOST_CC'. We still +# use `HOST_CC' if defined, but it is deprecated. + +# Portable tmp directory creation inspired by the Autoconf team. + +tmp= +# shellcheck disable=SC2172 +trap 'test -z "$tmp" || rm -fr "$tmp"' 0 1 2 13 15 + +set_cc_for_build() { + # prevent multiple calls if $tmp is already set + test "$tmp" && return 0 + : "${TMPDIR=/tmp}" + # shellcheck disable=SC2039 + { tmp=`(umask 077 && mktemp -d "$TMPDIR/cgXXXXXX") 2>/dev/null` && test -n "$tmp" && test -d "$tmp" ; } || + { test -n "$RANDOM" && tmp=$TMPDIR/cg$$-$RANDOM && (umask 077 && mkdir "$tmp" 2>/dev/null) ; } || + { tmp=$TMPDIR/cg-$$ && (umask 077 && mkdir "$tmp" 2>/dev/null) && echo "Warning: creating insecure temp directory" >&2 ; } || + { echo "$me: cannot create a temporary directory in $TMPDIR" >&2 ; exit 1 ; } + dummy=$tmp/dummy + case ${CC_FOR_BUILD-},${HOST_CC-},${CC-} in + ,,) echo "int x;" > "$dummy.c" + for driver in cc gcc c89 c99 ; do + if ($driver -c -o "$dummy.o" "$dummy.c") >/dev/null 2>&1 ; then + CC_FOR_BUILD="$driver" + break + fi + done + if test x"$CC_FOR_BUILD" = x ; then + CC_FOR_BUILD=no_compiler_found + fi + ;; + ,,*) CC_FOR_BUILD=$CC ;; + ,*,*) CC_FOR_BUILD=$HOST_CC ;; + esac +} + +# This is needed to find uname on a Pyramid OSx when run in the BSD universe. +# (ghazi@noc.rutgers.edu 1994-08-24) +if test -f /.attbin/uname ; then + PATH=$PATH:/.attbin ; export PATH +fi + +UNAME_MACHINE=`(uname -m) 2>/dev/null` || UNAME_MACHINE=unknown +UNAME_RELEASE=`(uname -r) 2>/dev/null` || UNAME_RELEASE=unknown +UNAME_SYSTEM=`(uname -s) 2>/dev/null` || UNAME_SYSTEM=unknown +UNAME_VERSION=`(uname -v) 2>/dev/null` || UNAME_VERSION=unknown + +case "$UNAME_SYSTEM" in +Linux|GNU|GNU/*) + # If the system lacks a compiler, then just pick glibc. + # We could probably try harder. + LIBC=gnu + + set_cc_for_build + cat <<-EOF > "$dummy.c" + #include + #if defined(__UCLIBC__) + LIBC=uclibc + #elif defined(__dietlibc__) + LIBC=dietlibc + #else + LIBC=gnu + #endif + EOF + eval "`$CC_FOR_BUILD -E "$dummy.c" 2>/dev/null | grep '^LIBC' | sed 's, ,,g'`" + + # If ldd exists, use it to detect musl libc. + if command -v ldd >/dev/null && \ + ldd --version 2>&1 | grep -q ^musl + then + LIBC=musl + fi + ;; +esac + +# Note: order is significant - the case branches are not exclusive. + +case "$UNAME_MACHINE:$UNAME_SYSTEM:$UNAME_RELEASE:$UNAME_VERSION" in + *:NetBSD:*:*) + # NetBSD (nbsd) targets should (where applicable) match one or + # more of the tuples: *-*-netbsdelf*, *-*-netbsdaout*, + # *-*-netbsdecoff* and *-*-netbsd*. For targets that recently + # switched to ELF, *-*-netbsd* would select the old + # object file format. This provides both forward + # compatibility and a consistent mechanism for selecting the + # object file format. + # + # Note: NetBSD doesn't particularly care about the vendor + # portion of the name. We always set it to "unknown". + sysctl="sysctl -n hw.machine_arch" + UNAME_MACHINE_ARCH=`(uname -p 2>/dev/null || \ + "/sbin/$sysctl" 2>/dev/null || \ + "/usr/sbin/$sysctl" 2>/dev/null || \ + echo unknown)` + case "$UNAME_MACHINE_ARCH" in + armeb) machine=armeb-unknown ;; + arm*) machine=arm-unknown ;; + sh3el) machine=shl-unknown ;; + sh3eb) machine=sh-unknown ;; + sh5el) machine=sh5le-unknown ;; + earmv*) + arch=`echo "$UNAME_MACHINE_ARCH" | sed -e 's,^e\(armv[0-9]\).*$,\1,'` + endian=`echo "$UNAME_MACHINE_ARCH" | sed -ne 's,^.*\(eb\)$,\1,p'` + machine="${arch}${endian}"-unknown + ;; + *) machine="$UNAME_MACHINE_ARCH"-unknown ;; + esac + # The Operating System including object format, if it has switched + # to ELF recently (or will in the future) and ABI. + case "$UNAME_MACHINE_ARCH" in + earm*) + os=netbsdelf + ;; + arm*|i386|m68k|ns32k|sh3*|sparc|vax) + set_cc_for_build + if echo __ELF__ | $CC_FOR_BUILD -E - 2>/dev/null \ + | grep -q __ELF__ + then + # Once all utilities can be ECOFF (netbsdecoff) or a.out (netbsdaout). + # Return netbsd for either. FIX? + os=netbsd + else + os=netbsdelf + fi + ;; + *) + os=netbsd + ;; + esac + # Determine ABI tags. + case "$UNAME_MACHINE_ARCH" in + earm*) + expr='s/^earmv[0-9]/-eabi/;s/eb$//' + abi=`echo "$UNAME_MACHINE_ARCH" | sed -e "$expr"` + ;; + esac + # The OS release + # Debian GNU/NetBSD machines have a different userland, and + # thus, need a distinct triplet. However, they do not need + # kernel version information, so it can be replaced with a + # suitable tag, in the style of linux-gnu. + case "$UNAME_VERSION" in + Debian*) + release='-gnu' + ;; + *) + release=`echo "$UNAME_RELEASE" | sed -e 's/[-_].*//' | cut -d. -f1,2` + ;; + esac + # Since CPU_TYPE-MANUFACTURER-KERNEL-OPERATING_SYSTEM: + # contains redundant information, the shorter form: + # CPU_TYPE-MANUFACTURER-OPERATING_SYSTEM is used. + echo "$machine-${os}${release}${abi-}" + exit ;; + *:Bitrig:*:*) + UNAME_MACHINE_ARCH=`arch | sed 's/Bitrig.//'` + echo "$UNAME_MACHINE_ARCH"-unknown-bitrig"$UNAME_RELEASE" + exit ;; + *:OpenBSD:*:*) + UNAME_MACHINE_ARCH=`arch | sed 's/OpenBSD.//'` + echo "$UNAME_MACHINE_ARCH"-unknown-openbsd"$UNAME_RELEASE" + exit ;; + *:LibertyBSD:*:*) + UNAME_MACHINE_ARCH=`arch | sed 's/^.*BSD\.//'` + echo "$UNAME_MACHINE_ARCH"-unknown-libertybsd"$UNAME_RELEASE" + exit ;; + *:MidnightBSD:*:*) + echo "$UNAME_MACHINE"-unknown-midnightbsd"$UNAME_RELEASE" + exit ;; + *:ekkoBSD:*:*) + echo "$UNAME_MACHINE"-unknown-ekkobsd"$UNAME_RELEASE" + exit ;; + *:SolidBSD:*:*) + echo "$UNAME_MACHINE"-unknown-solidbsd"$UNAME_RELEASE" + exit ;; + *:OS108:*:*) + echo "$UNAME_MACHINE"-unknown-os108_"$UNAME_RELEASE" + exit ;; + macppc:MirBSD:*:*) + echo powerpc-unknown-mirbsd"$UNAME_RELEASE" + exit ;; + *:MirBSD:*:*) + echo "$UNAME_MACHINE"-unknown-mirbsd"$UNAME_RELEASE" + exit ;; + *:Sortix:*:*) + echo "$UNAME_MACHINE"-unknown-sortix + exit ;; + *:Twizzler:*:*) + echo "$UNAME_MACHINE"-unknown-twizzler + exit ;; + *:Redox:*:*) + echo "$UNAME_MACHINE"-unknown-redox + exit ;; + mips:OSF1:*.*) + echo mips-dec-osf1 + exit ;; + alpha:OSF1:*:*) + case $UNAME_RELEASE in + *4.0) + UNAME_RELEASE=`/usr/sbin/sizer -v | awk '{print $3}'` + ;; + *5.*) + UNAME_RELEASE=`/usr/sbin/sizer -v | awk '{print $4}'` + ;; + esac + # According to Compaq, /usr/sbin/psrinfo has been available on + # OSF/1 and Tru64 systems produced since 1995. I hope that + # covers most systems running today. This code pipes the CPU + # types through head -n 1, so we only detect the type of CPU 0. + ALPHA_CPU_TYPE=`/usr/sbin/psrinfo -v | sed -n -e 's/^ The alpha \(.*\) processor.*$/\1/p' | head -n 1` + case "$ALPHA_CPU_TYPE" in + "EV4 (21064)") + UNAME_MACHINE=alpha ;; + "EV4.5 (21064)") + UNAME_MACHINE=alpha ;; + "LCA4 (21066/21068)") + UNAME_MACHINE=alpha ;; + "EV5 (21164)") + UNAME_MACHINE=alphaev5 ;; + "EV5.6 (21164A)") + UNAME_MACHINE=alphaev56 ;; + "EV5.6 (21164PC)") + UNAME_MACHINE=alphapca56 ;; + "EV5.7 (21164PC)") + UNAME_MACHINE=alphapca57 ;; + "EV6 (21264)") + UNAME_MACHINE=alphaev6 ;; + "EV6.7 (21264A)") + UNAME_MACHINE=alphaev67 ;; + "EV6.8CB (21264C)") + UNAME_MACHINE=alphaev68 ;; + "EV6.8AL (21264B)") + UNAME_MACHINE=alphaev68 ;; + "EV6.8CX (21264D)") + UNAME_MACHINE=alphaev68 ;; + "EV6.9A (21264/EV69A)") + UNAME_MACHINE=alphaev69 ;; + "EV7 (21364)") + UNAME_MACHINE=alphaev7 ;; + "EV7.9 (21364A)") + UNAME_MACHINE=alphaev79 ;; + esac + # A Pn.n version is a patched version. + # A Vn.n version is a released version. + # A Tn.n version is a released field test version. + # A Xn.n version is an unreleased experimental baselevel. + # 1.2 uses "1.2" for uname -r. + echo "$UNAME_MACHINE"-dec-osf"`echo "$UNAME_RELEASE" | sed -e 's/^[PVTX]//' | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz`" + # Reset EXIT trap before exiting to avoid spurious non-zero exit code. + exitcode=$? + trap '' 0 + exit $exitcode ;; + Amiga*:UNIX_System_V:4.0:*) + echo m68k-unknown-sysv4 + exit ;; + *:[Aa]miga[Oo][Ss]:*:*) + echo "$UNAME_MACHINE"-unknown-amigaos + exit ;; + *:[Mm]orph[Oo][Ss]:*:*) + echo "$UNAME_MACHINE"-unknown-morphos + exit ;; + *:OS/390:*:*) + echo i370-ibm-openedition + exit ;; + *:z/VM:*:*) + echo s390-ibm-zvmoe + exit ;; + *:OS400:*:*) + echo powerpc-ibm-os400 + exit ;; + arm:RISC*:1.[012]*:*|arm:riscix:1.[012]*:*) + echo arm-acorn-riscix"$UNAME_RELEASE" + exit ;; + arm*:riscos:*:*|arm*:RISCOS:*:*) + echo arm-unknown-riscos + exit ;; + SR2?01:HI-UX/MPP:*:* | SR8000:HI-UX/MPP:*:*) + echo hppa1.1-hitachi-hiuxmpp + exit ;; + Pyramid*:OSx*:*:* | MIS*:OSx*:*:* | MIS*:SMP_DC-OSx*:*:*) + # akee@wpdis03.wpafb.af.mil (Earle F. Ake) contributed MIS and NILE. + if test "`(/bin/universe) 2>/dev/null`" = att ; then + echo pyramid-pyramid-sysv3 + else + echo pyramid-pyramid-bsd + fi + exit ;; + NILE*:*:*:dcosx) + echo pyramid-pyramid-svr4 + exit ;; + DRS?6000:unix:4.0:6*) + echo sparc-icl-nx6 + exit ;; + DRS?6000:UNIX_SV:4.2*:7* | DRS?6000:isis:4.2*:7*) + case `/usr/bin/uname -p` in + sparc) echo sparc-icl-nx7; exit ;; + esac ;; + s390x:SunOS:*:*) + echo "$UNAME_MACHINE"-ibm-solaris2"`echo "$UNAME_RELEASE" | sed -e 's/[^.]*//'`" + exit ;; + sun4H:SunOS:5.*:*) + echo sparc-hal-solaris2"`echo "$UNAME_RELEASE"|sed -e 's/[^.]*//'`" + exit ;; + sun4*:SunOS:5.*:* | tadpole*:SunOS:5.*:*) + echo sparc-sun-solaris2"`echo "$UNAME_RELEASE" | sed -e 's/[^.]*//'`" + exit ;; + i86pc:AuroraUX:5.*:* | i86xen:AuroraUX:5.*:*) + echo i386-pc-auroraux"$UNAME_RELEASE" + exit ;; + i86pc:SunOS:5.*:* | i86xen:SunOS:5.*:*) + set_cc_for_build + SUN_ARCH=i386 + # If there is a compiler, see if it is configured for 64-bit objects. + # Note that the Sun cc does not turn __LP64__ into 1 like gcc does. + # This test works for both compilers. + if [ "$CC_FOR_BUILD" != no_compiler_found ]; then + if (echo '#ifdef __amd64'; echo IS_64BIT_ARCH; echo '#endif') | \ + (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) | \ + grep IS_64BIT_ARCH >/dev/null + then + SUN_ARCH=x86_64 + fi + fi + echo "$SUN_ARCH"-pc-solaris2"`echo "$UNAME_RELEASE"|sed -e 's/[^.]*//'`" + exit ;; + sun4*:SunOS:6*:*) + # According to config.sub, this is the proper way to canonicalize + # SunOS6. Hard to guess exactly what SunOS6 will be like, but + # it's likely to be more like Solaris than SunOS4. + echo sparc-sun-solaris3"`echo "$UNAME_RELEASE"|sed -e 's/[^.]*//'`" + exit ;; + sun4*:SunOS:*:*) + case "`/usr/bin/arch -k`" in + Series*|S4*) + UNAME_RELEASE=`uname -v` + ;; + esac + # Japanese Language versions have a version number like `4.1.3-JL'. + echo sparc-sun-sunos"`echo "$UNAME_RELEASE"|sed -e 's/-/_/'`" + exit ;; + sun3*:SunOS:*:*) + echo m68k-sun-sunos"$UNAME_RELEASE" + exit ;; + sun*:*:4.2BSD:*) + UNAME_RELEASE=`(sed 1q /etc/motd | awk '{print substr($5,1,3)}') 2>/dev/null` + test "x$UNAME_RELEASE" = x && UNAME_RELEASE=3 + case "`/bin/arch`" in + sun3) + echo m68k-sun-sunos"$UNAME_RELEASE" + ;; + sun4) + echo sparc-sun-sunos"$UNAME_RELEASE" + ;; + esac + exit ;; + aushp:SunOS:*:*) + echo sparc-auspex-sunos"$UNAME_RELEASE" + exit ;; + # The situation for MiNT is a little confusing. The machine name + # can be virtually everything (everything which is not + # "atarist" or "atariste" at least should have a processor + # > m68000). The system name ranges from "MiNT" over "FreeMiNT" + # to the lowercase version "mint" (or "freemint"). Finally + # the system name "TOS" denotes a system which is actually not + # MiNT. But MiNT is downward compatible to TOS, so this should + # be no problem. + atarist[e]:*MiNT:*:* | atarist[e]:*mint:*:* | atarist[e]:*TOS:*:*) + echo m68k-atari-mint"$UNAME_RELEASE" + exit ;; + atari*:*MiNT:*:* | atari*:*mint:*:* | atarist[e]:*TOS:*:*) + echo m68k-atari-mint"$UNAME_RELEASE" + exit ;; + *falcon*:*MiNT:*:* | *falcon*:*mint:*:* | *falcon*:*TOS:*:*) + echo m68k-atari-mint"$UNAME_RELEASE" + exit ;; + milan*:*MiNT:*:* | milan*:*mint:*:* | *milan*:*TOS:*:*) + echo m68k-milan-mint"$UNAME_RELEASE" + exit ;; + hades*:*MiNT:*:* | hades*:*mint:*:* | *hades*:*TOS:*:*) + echo m68k-hades-mint"$UNAME_RELEASE" + exit ;; + *:*MiNT:*:* | *:*mint:*:* | *:*TOS:*:*) + echo m68k-unknown-mint"$UNAME_RELEASE" + exit ;; + m68k:machten:*:*) + echo m68k-apple-machten"$UNAME_RELEASE" + exit ;; + powerpc:machten:*:*) + echo powerpc-apple-machten"$UNAME_RELEASE" + exit ;; + RISC*:Mach:*:*) + echo mips-dec-mach_bsd4.3 + exit ;; + RISC*:ULTRIX:*:*) + echo mips-dec-ultrix"$UNAME_RELEASE" + exit ;; + VAX*:ULTRIX*:*:*) + echo vax-dec-ultrix"$UNAME_RELEASE" + exit ;; + 2020:CLIX:*:* | 2430:CLIX:*:*) + echo clipper-intergraph-clix"$UNAME_RELEASE" + exit ;; + mips:*:*:UMIPS | mips:*:*:RISCos) + set_cc_for_build + sed 's/^ //' << EOF > "$dummy.c" +#ifdef __cplusplus +#include /* for printf() prototype */ + int main (int argc, char *argv[]) { +#else + int main (argc, argv) int argc; char *argv[]; { +#endif + #if defined (host_mips) && defined (MIPSEB) + #if defined (SYSTYPE_SYSV) + printf ("mips-mips-riscos%ssysv\\n", argv[1]); exit (0); + #endif + #if defined (SYSTYPE_SVR4) + printf ("mips-mips-riscos%ssvr4\\n", argv[1]); exit (0); + #endif + #if defined (SYSTYPE_BSD43) || defined(SYSTYPE_BSD) + printf ("mips-mips-riscos%sbsd\\n", argv[1]); exit (0); + #endif + #endif + exit (-1); + } +EOF + $CC_FOR_BUILD -o "$dummy" "$dummy.c" && + dummyarg=`echo "$UNAME_RELEASE" | sed -n 's/\([0-9]*\).*/\1/p'` && + SYSTEM_NAME=`"$dummy" "$dummyarg"` && + { echo "$SYSTEM_NAME"; exit; } + echo mips-mips-riscos"$UNAME_RELEASE" + exit ;; + Motorola:PowerMAX_OS:*:*) + echo powerpc-motorola-powermax + exit ;; + Motorola:*:4.3:PL8-*) + echo powerpc-harris-powermax + exit ;; + Night_Hawk:*:*:PowerMAX_OS | Synergy:PowerMAX_OS:*:*) + echo powerpc-harris-powermax + exit ;; + Night_Hawk:Power_UNIX:*:*) + echo powerpc-harris-powerunix + exit ;; + m88k:CX/UX:7*:*) + echo m88k-harris-cxux7 + exit ;; + m88k:*:4*:R4*) + echo m88k-motorola-sysv4 + exit ;; + m88k:*:3*:R3*) + echo m88k-motorola-sysv3 + exit ;; + AViiON:dgux:*:*) + # DG/UX returns AViiON for all architectures + UNAME_PROCESSOR=`/usr/bin/uname -p` + if [ "$UNAME_PROCESSOR" = mc88100 ] || [ "$UNAME_PROCESSOR" = mc88110 ] + then + if [ "$TARGET_BINARY_INTERFACE"x = m88kdguxelfx ] || \ + [ "$TARGET_BINARY_INTERFACE"x = x ] + then + echo m88k-dg-dgux"$UNAME_RELEASE" + else + echo m88k-dg-dguxbcs"$UNAME_RELEASE" + fi + else + echo i586-dg-dgux"$UNAME_RELEASE" + fi + exit ;; + M88*:DolphinOS:*:*) # DolphinOS (SVR3) + echo m88k-dolphin-sysv3 + exit ;; + M88*:*:R3*:*) + # Delta 88k system running SVR3 + echo m88k-motorola-sysv3 + exit ;; + XD88*:*:*:*) # Tektronix XD88 system running UTekV (SVR3) + echo m88k-tektronix-sysv3 + exit ;; + Tek43[0-9][0-9]:UTek:*:*) # Tektronix 4300 system running UTek (BSD) + echo m68k-tektronix-bsd + exit ;; + *:IRIX*:*:*) + echo mips-sgi-irix"`echo "$UNAME_RELEASE"|sed -e 's/-/_/g'`" + exit ;; + ????????:AIX?:[12].1:2) # AIX 2.2.1 or AIX 2.1.1 is RT/PC AIX. + echo romp-ibm-aix # uname -m gives an 8 hex-code CPU id + exit ;; # Note that: echo "'`uname -s`'" gives 'AIX ' + i*86:AIX:*:*) + echo i386-ibm-aix + exit ;; + ia64:AIX:*:*) + if [ -x /usr/bin/oslevel ] ; then + IBM_REV=`/usr/bin/oslevel` + else + IBM_REV="$UNAME_VERSION.$UNAME_RELEASE" + fi + echo "$UNAME_MACHINE"-ibm-aix"$IBM_REV" + exit ;; + *:AIX:2:3) + if grep bos325 /usr/include/stdio.h >/dev/null 2>&1; then + set_cc_for_build + sed 's/^ //' << EOF > "$dummy.c" + #include + + main() + { + if (!__power_pc()) + exit(1); + puts("powerpc-ibm-aix3.2.5"); + exit(0); + } +EOF + if $CC_FOR_BUILD -o "$dummy" "$dummy.c" && SYSTEM_NAME=`"$dummy"` + then + echo "$SYSTEM_NAME" + else + echo rs6000-ibm-aix3.2.5 + fi + elif grep bos324 /usr/include/stdio.h >/dev/null 2>&1; then + echo rs6000-ibm-aix3.2.4 + else + echo rs6000-ibm-aix3.2 + fi + exit ;; + *:AIX:*:[4567]) + IBM_CPU_ID=`/usr/sbin/lsdev -C -c processor -S available | sed 1q | awk '{ print $1 }'` + if /usr/sbin/lsattr -El "$IBM_CPU_ID" | grep ' POWER' >/dev/null 2>&1; then + IBM_ARCH=rs6000 + else + IBM_ARCH=powerpc + fi + if [ -x /usr/bin/lslpp ] ; then + IBM_REV=`/usr/bin/lslpp -Lqc bos.rte.libc | + awk -F: '{ print $3 }' | sed s/[0-9]*$/0/` + else + IBM_REV="$UNAME_VERSION.$UNAME_RELEASE" + fi + echo "$IBM_ARCH"-ibm-aix"$IBM_REV" + exit ;; + *:AIX:*:*) + echo rs6000-ibm-aix + exit ;; + ibmrt:4.4BSD:*|romp-ibm:4.4BSD:*) + echo romp-ibm-bsd4.4 + exit ;; + ibmrt:*BSD:*|romp-ibm:BSD:*) # covers RT/PC BSD and + echo romp-ibm-bsd"$UNAME_RELEASE" # 4.3 with uname added to + exit ;; # report: romp-ibm BSD 4.3 + *:BOSX:*:*) + echo rs6000-bull-bosx + exit ;; + DPX/2?00:B.O.S.:*:*) + echo m68k-bull-sysv3 + exit ;; + 9000/[34]??:4.3bsd:1.*:*) + echo m68k-hp-bsd + exit ;; + hp300:4.4BSD:*:* | 9000/[34]??:4.3bsd:2.*:*) + echo m68k-hp-bsd4.4 + exit ;; + 9000/[34678]??:HP-UX:*:*) + HPUX_REV=`echo "$UNAME_RELEASE"|sed -e 's/[^.]*.[0B]*//'` + case "$UNAME_MACHINE" in + 9000/31?) HP_ARCH=m68000 ;; + 9000/[34]??) HP_ARCH=m68k ;; + 9000/[678][0-9][0-9]) + if [ -x /usr/bin/getconf ]; then + sc_cpu_version=`/usr/bin/getconf SC_CPU_VERSION 2>/dev/null` + sc_kernel_bits=`/usr/bin/getconf SC_KERNEL_BITS 2>/dev/null` + case "$sc_cpu_version" in + 523) HP_ARCH=hppa1.0 ;; # CPU_PA_RISC1_0 + 528) HP_ARCH=hppa1.1 ;; # CPU_PA_RISC1_1 + 532) # CPU_PA_RISC2_0 + case "$sc_kernel_bits" in + 32) HP_ARCH=hppa2.0n ;; + 64) HP_ARCH=hppa2.0w ;; + '') HP_ARCH=hppa2.0 ;; # HP-UX 10.20 + esac ;; + esac + fi + if [ "$HP_ARCH" = "" ]; then + set_cc_for_build + sed 's/^ //' << EOF > "$dummy.c" + + #define _HPUX_SOURCE + #include + #include + + int main () + { + #if defined(_SC_KERNEL_BITS) + long bits = sysconf(_SC_KERNEL_BITS); + #endif + long cpu = sysconf (_SC_CPU_VERSION); + + switch (cpu) + { + case CPU_PA_RISC1_0: puts ("hppa1.0"); break; + case CPU_PA_RISC1_1: puts ("hppa1.1"); break; + case CPU_PA_RISC2_0: + #if defined(_SC_KERNEL_BITS) + switch (bits) + { + case 64: puts ("hppa2.0w"); break; + case 32: puts ("hppa2.0n"); break; + default: puts ("hppa2.0"); break; + } break; + #else /* !defined(_SC_KERNEL_BITS) */ + puts ("hppa2.0"); break; + #endif + default: puts ("hppa1.0"); break; + } + exit (0); + } +EOF + (CCOPTS="" $CC_FOR_BUILD -o "$dummy" "$dummy.c" 2>/dev/null) && HP_ARCH=`"$dummy"` + test -z "$HP_ARCH" && HP_ARCH=hppa + fi ;; + esac + if [ "$HP_ARCH" = hppa2.0w ] + then + set_cc_for_build + + # hppa2.0w-hp-hpux* has a 64-bit kernel and a compiler generating + # 32-bit code. hppa64-hp-hpux* has the same kernel and a compiler + # generating 64-bit code. GNU and HP use different nomenclature: + # + # $ CC_FOR_BUILD=cc ./config.guess + # => hppa2.0w-hp-hpux11.23 + # $ CC_FOR_BUILD="cc +DA2.0w" ./config.guess + # => hppa64-hp-hpux11.23 + + if echo __LP64__ | (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) | + grep -q __LP64__ + then + HP_ARCH=hppa2.0w + else + HP_ARCH=hppa64 + fi + fi + echo "$HP_ARCH"-hp-hpux"$HPUX_REV" + exit ;; + ia64:HP-UX:*:*) + HPUX_REV=`echo "$UNAME_RELEASE"|sed -e 's/[^.]*.[0B]*//'` + echo ia64-hp-hpux"$HPUX_REV" + exit ;; + 3050*:HI-UX:*:*) + set_cc_for_build + sed 's/^ //' << EOF > "$dummy.c" + #include + int + main () + { + long cpu = sysconf (_SC_CPU_VERSION); + /* The order matters, because CPU_IS_HP_MC68K erroneously returns + true for CPU_PA_RISC1_0. CPU_IS_PA_RISC returns correct + results, however. */ + if (CPU_IS_PA_RISC (cpu)) + { + switch (cpu) + { + case CPU_PA_RISC1_0: puts ("hppa1.0-hitachi-hiuxwe2"); break; + case CPU_PA_RISC1_1: puts ("hppa1.1-hitachi-hiuxwe2"); break; + case CPU_PA_RISC2_0: puts ("hppa2.0-hitachi-hiuxwe2"); break; + default: puts ("hppa-hitachi-hiuxwe2"); break; + } + } + else if (CPU_IS_HP_MC68K (cpu)) + puts ("m68k-hitachi-hiuxwe2"); + else puts ("unknown-hitachi-hiuxwe2"); + exit (0); + } +EOF + $CC_FOR_BUILD -o "$dummy" "$dummy.c" && SYSTEM_NAME=`"$dummy"` && + { echo "$SYSTEM_NAME"; exit; } + echo unknown-hitachi-hiuxwe2 + exit ;; + 9000/7??:4.3bsd:*:* | 9000/8?[79]:4.3bsd:*:*) + echo hppa1.1-hp-bsd + exit ;; + 9000/8??:4.3bsd:*:*) + echo hppa1.0-hp-bsd + exit ;; + *9??*:MPE/iX:*:* | *3000*:MPE/iX:*:*) + echo hppa1.0-hp-mpeix + exit ;; + hp7??:OSF1:*:* | hp8?[79]:OSF1:*:*) + echo hppa1.1-hp-osf + exit ;; + hp8??:OSF1:*:*) + echo hppa1.0-hp-osf + exit ;; + i*86:OSF1:*:*) + if [ -x /usr/sbin/sysversion ] ; then + echo "$UNAME_MACHINE"-unknown-osf1mk + else + echo "$UNAME_MACHINE"-unknown-osf1 + fi + exit ;; + parisc*:Lites*:*:*) + echo hppa1.1-hp-lites + exit ;; + C1*:ConvexOS:*:* | convex:ConvexOS:C1*:*) + echo c1-convex-bsd + exit ;; + C2*:ConvexOS:*:* | convex:ConvexOS:C2*:*) + if getsysinfo -f scalar_acc + then echo c32-convex-bsd + else echo c2-convex-bsd + fi + exit ;; + C34*:ConvexOS:*:* | convex:ConvexOS:C34*:*) + echo c34-convex-bsd + exit ;; + C38*:ConvexOS:*:* | convex:ConvexOS:C38*:*) + echo c38-convex-bsd + exit ;; + C4*:ConvexOS:*:* | convex:ConvexOS:C4*:*) + echo c4-convex-bsd + exit ;; + CRAY*Y-MP:*:*:*) + echo ymp-cray-unicos"$UNAME_RELEASE" | sed -e 's/\.[^.]*$/.X/' + exit ;; + CRAY*[A-Z]90:*:*:*) + echo "$UNAME_MACHINE"-cray-unicos"$UNAME_RELEASE" \ + | sed -e 's/CRAY.*\([A-Z]90\)/\1/' \ + -e y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/ \ + -e 's/\.[^.]*$/.X/' + exit ;; + CRAY*TS:*:*:*) + echo t90-cray-unicos"$UNAME_RELEASE" | sed -e 's/\.[^.]*$/.X/' + exit ;; + CRAY*T3E:*:*:*) + echo alphaev5-cray-unicosmk"$UNAME_RELEASE" | sed -e 's/\.[^.]*$/.X/' + exit ;; + CRAY*SV1:*:*:*) + echo sv1-cray-unicos"$UNAME_RELEASE" | sed -e 's/\.[^.]*$/.X/' + exit ;; + *:UNICOS/mp:*:*) + echo craynv-cray-unicosmp"$UNAME_RELEASE" | sed -e 's/\.[^.]*$/.X/' + exit ;; + F30[01]:UNIX_System_V:*:* | F700:UNIX_System_V:*:*) + FUJITSU_PROC=`uname -m | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz` + FUJITSU_SYS=`uname -p | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz | sed -e 's/\///'` + FUJITSU_REL=`echo "$UNAME_RELEASE" | sed -e 's/ /_/'` + echo "${FUJITSU_PROC}-fujitsu-${FUJITSU_SYS}${FUJITSU_REL}" + exit ;; + 5000:UNIX_System_V:4.*:*) + FUJITSU_SYS=`uname -p | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz | sed -e 's/\///'` + FUJITSU_REL=`echo "$UNAME_RELEASE" | tr ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz | sed -e 's/ /_/'` + echo "sparc-fujitsu-${FUJITSU_SYS}${FUJITSU_REL}" + exit ;; + i*86:BSD/386:*:* | i*86:BSD/OS:*:* | *:Ascend\ Embedded/OS:*:*) + echo "$UNAME_MACHINE"-pc-bsdi"$UNAME_RELEASE" + exit ;; + sparc*:BSD/OS:*:*) + echo sparc-unknown-bsdi"$UNAME_RELEASE" + exit ;; + *:BSD/OS:*:*) + echo "$UNAME_MACHINE"-unknown-bsdi"$UNAME_RELEASE" + exit ;; + arm:FreeBSD:*:*) + UNAME_PROCESSOR=`uname -p` + set_cc_for_build + if echo __ARM_PCS_VFP | $CC_FOR_BUILD -E - 2>/dev/null \ + | grep -q __ARM_PCS_VFP + then + echo "${UNAME_PROCESSOR}"-unknown-freebsd"`echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'`"-gnueabi + else + echo "${UNAME_PROCESSOR}"-unknown-freebsd"`echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'`"-gnueabihf + fi + exit ;; + *:FreeBSD:*:*) + UNAME_PROCESSOR=`/usr/bin/uname -p` + case "$UNAME_PROCESSOR" in + amd64) + UNAME_PROCESSOR=x86_64 ;; + i386) + UNAME_PROCESSOR=i586 ;; + esac + echo "$UNAME_PROCESSOR"-unknown-freebsd"`echo "$UNAME_RELEASE"|sed -e 's/[-(].*//'`" + exit ;; + i*:CYGWIN*:*) + echo "$UNAME_MACHINE"-pc-cygwin + exit ;; + *:MINGW64*:*) + echo "$UNAME_MACHINE"-pc-mingw64 + exit ;; + *:MINGW*:*) + echo "$UNAME_MACHINE"-pc-mingw32 + exit ;; + *:MSYS*:*) + echo "$UNAME_MACHINE"-pc-msys + exit ;; + i*:PW*:*) + echo "$UNAME_MACHINE"-pc-pw32 + exit ;; + *:Interix*:*) + case "$UNAME_MACHINE" in + x86) + echo i586-pc-interix"$UNAME_RELEASE" + exit ;; + authenticamd | genuineintel | EM64T) + echo x86_64-unknown-interix"$UNAME_RELEASE" + exit ;; + IA64) + echo ia64-unknown-interix"$UNAME_RELEASE" + exit ;; + esac ;; + i*:UWIN*:*) + echo "$UNAME_MACHINE"-pc-uwin + exit ;; + amd64:CYGWIN*:*:* | x86_64:CYGWIN*:*:*) + echo x86_64-pc-cygwin + exit ;; + prep*:SunOS:5.*:*) + echo powerpcle-unknown-solaris2"`echo "$UNAME_RELEASE"|sed -e 's/[^.]*//'`" + exit ;; + *:GNU:*:*) + # the GNU system + echo "`echo "$UNAME_MACHINE"|sed -e 's,[-/].*$,,'`-unknown-$LIBC`echo "$UNAME_RELEASE"|sed -e 's,/.*$,,'`" + exit ;; + *:GNU/*:*:*) + # other systems with GNU libc and userland + echo "$UNAME_MACHINE-unknown-`echo "$UNAME_SYSTEM" | sed 's,^[^/]*/,,' | tr "[:upper:]" "[:lower:]"``echo "$UNAME_RELEASE"|sed -e 's/[-(].*//'`-$LIBC" + exit ;; + *:Minix:*:*) + echo "$UNAME_MACHINE"-unknown-minix + exit ;; + aarch64:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + aarch64_be:Linux:*:*) + UNAME_MACHINE=aarch64_be + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + alpha:Linux:*:*) + case `sed -n '/^cpu model/s/^.*: \(.*\)/\1/p' /proc/cpuinfo 2>/dev/null` in + EV5) UNAME_MACHINE=alphaev5 ;; + EV56) UNAME_MACHINE=alphaev56 ;; + PCA56) UNAME_MACHINE=alphapca56 ;; + PCA57) UNAME_MACHINE=alphapca56 ;; + EV6) UNAME_MACHINE=alphaev6 ;; + EV67) UNAME_MACHINE=alphaev67 ;; + EV68*) UNAME_MACHINE=alphaev68 ;; + esac + objdump --private-headers /bin/sh | grep -q ld.so.1 + if test "$?" = 0 ; then LIBC=gnulibc1 ; fi + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + arc:Linux:*:* | arceb:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + arm*:Linux:*:*) + set_cc_for_build + if echo __ARM_EABI__ | $CC_FOR_BUILD -E - 2>/dev/null \ + | grep -q __ARM_EABI__ + then + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + else + if echo __ARM_PCS_VFP | $CC_FOR_BUILD -E - 2>/dev/null \ + | grep -q __ARM_PCS_VFP + then + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"eabi + else + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC"eabihf + fi + fi + exit ;; + avr32*:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + cris:Linux:*:*) + echo "$UNAME_MACHINE"-axis-linux-"$LIBC" + exit ;; + crisv32:Linux:*:*) + echo "$UNAME_MACHINE"-axis-linux-"$LIBC" + exit ;; + e2k:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + frv:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + hexagon:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + i*86:Linux:*:*) + echo "$UNAME_MACHINE"-pc-linux-"$LIBC" + exit ;; + ia64:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + k1om:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + m32r*:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + m68*:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + mips:Linux:*:* | mips64:Linux:*:*) + set_cc_for_build + IS_GLIBC=0 + test x"${LIBC}" = xgnu && IS_GLIBC=1 + sed 's/^ //' << EOF > "$dummy.c" + #undef CPU + #undef mips + #undef mipsel + #undef mips64 + #undef mips64el + #if ${IS_GLIBC} && defined(_ABI64) + LIBCABI=gnuabi64 + #else + #if ${IS_GLIBC} && defined(_ABIN32) + LIBCABI=gnuabin32 + #else + LIBCABI=${LIBC} + #endif + #endif + + #if ${IS_GLIBC} && defined(__mips64) && defined(__mips_isa_rev) && __mips_isa_rev>=6 + CPU=mipsisa64r6 + #else + #if ${IS_GLIBC} && !defined(__mips64) && defined(__mips_isa_rev) && __mips_isa_rev>=6 + CPU=mipsisa32r6 + #else + #if defined(__mips64) + CPU=mips64 + #else + CPU=mips + #endif + #endif + #endif + + #if defined(__MIPSEL__) || defined(__MIPSEL) || defined(_MIPSEL) || defined(MIPSEL) + MIPS_ENDIAN=el + #else + #if defined(__MIPSEB__) || defined(__MIPSEB) || defined(_MIPSEB) || defined(MIPSEB) + MIPS_ENDIAN= + #else + MIPS_ENDIAN= + #endif + #endif +EOF + eval "`$CC_FOR_BUILD -E "$dummy.c" 2>/dev/null | grep '^CPU\|^MIPS_ENDIAN\|^LIBCABI'`" + test "x$CPU" != x && { echo "$CPU${MIPS_ENDIAN}-unknown-linux-$LIBCABI"; exit; } + ;; + mips64el:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + openrisc*:Linux:*:*) + echo or1k-unknown-linux-"$LIBC" + exit ;; + or32:Linux:*:* | or1k*:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + padre:Linux:*:*) + echo sparc-unknown-linux-"$LIBC" + exit ;; + parisc64:Linux:*:* | hppa64:Linux:*:*) + echo hppa64-unknown-linux-"$LIBC" + exit ;; + parisc:Linux:*:* | hppa:Linux:*:*) + # Look for CPU level + case `grep '^cpu[^a-z]*:' /proc/cpuinfo 2>/dev/null | cut -d' ' -f2` in + PA7*) echo hppa1.1-unknown-linux-"$LIBC" ;; + PA8*) echo hppa2.0-unknown-linux-"$LIBC" ;; + *) echo hppa-unknown-linux-"$LIBC" ;; + esac + exit ;; + ppc64:Linux:*:*) + echo powerpc64-unknown-linux-"$LIBC" + exit ;; + ppc:Linux:*:*) + echo powerpc-unknown-linux-"$LIBC" + exit ;; + ppc64le:Linux:*:*) + echo powerpc64le-unknown-linux-"$LIBC" + exit ;; + ppcle:Linux:*:*) + echo powerpcle-unknown-linux-"$LIBC" + exit ;; + riscv32:Linux:*:* | riscv64:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + s390:Linux:*:* | s390x:Linux:*:*) + echo "$UNAME_MACHINE"-ibm-linux-"$LIBC" + exit ;; + sh64*:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + sh*:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + sparc:Linux:*:* | sparc64:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + tile*:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + vax:Linux:*:*) + echo "$UNAME_MACHINE"-dec-linux-"$LIBC" + exit ;; + x86_64:Linux:*:*) + set_cc_for_build + LIBCABI=$LIBC + if [ "$CC_FOR_BUILD" != no_compiler_found ]; then + if (echo '#ifdef __ILP32__'; echo IS_X32; echo '#endif') | \ + (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) | \ + grep IS_X32 >/dev/null + then + LIBCABI="$LIBC"x32 + fi + fi + echo "$UNAME_MACHINE"-pc-linux-"$LIBCABI" + exit ;; + xtensa*:Linux:*:*) + echo "$UNAME_MACHINE"-unknown-linux-"$LIBC" + exit ;; + i*86:DYNIX/ptx:4*:*) + # ptx 4.0 does uname -s correctly, with DYNIX/ptx in there. + # earlier versions are messed up and put the nodename in both + # sysname and nodename. + echo i386-sequent-sysv4 + exit ;; + i*86:UNIX_SV:4.2MP:2.*) + # Unixware is an offshoot of SVR4, but it has its own version + # number series starting with 2... + # I am not positive that other SVR4 systems won't match this, + # I just have to hope. -- rms. + # Use sysv4.2uw... so that sysv4* matches it. + echo "$UNAME_MACHINE"-pc-sysv4.2uw"$UNAME_VERSION" + exit ;; + i*86:OS/2:*:*) + # If we were able to find `uname', then EMX Unix compatibility + # is probably installed. + echo "$UNAME_MACHINE"-pc-os2-emx + exit ;; + i*86:XTS-300:*:STOP) + echo "$UNAME_MACHINE"-unknown-stop + exit ;; + i*86:atheos:*:*) + echo "$UNAME_MACHINE"-unknown-atheos + exit ;; + i*86:syllable:*:*) + echo "$UNAME_MACHINE"-pc-syllable + exit ;; + i*86:LynxOS:2.*:* | i*86:LynxOS:3.[01]*:* | i*86:LynxOS:4.[02]*:*) + echo i386-unknown-lynxos"$UNAME_RELEASE" + exit ;; + i*86:*DOS:*:*) + echo "$UNAME_MACHINE"-pc-msdosdjgpp + exit ;; + i*86:*:4.*:*) + UNAME_REL=`echo "$UNAME_RELEASE" | sed 's/\/MP$//'` + if grep Novell /usr/include/link.h >/dev/null 2>/dev/null; then + echo "$UNAME_MACHINE"-univel-sysv"$UNAME_REL" + else + echo "$UNAME_MACHINE"-pc-sysv"$UNAME_REL" + fi + exit ;; + i*86:*:5:[678]*) + # UnixWare 7.x, OpenUNIX and OpenServer 6. + case `/bin/uname -X | grep "^Machine"` in + *486*) UNAME_MACHINE=i486 ;; + *Pentium) UNAME_MACHINE=i586 ;; + *Pent*|*Celeron) UNAME_MACHINE=i686 ;; + esac + echo "$UNAME_MACHINE-unknown-sysv${UNAME_RELEASE}${UNAME_SYSTEM}${UNAME_VERSION}" + exit ;; + i*86:*:3.2:*) + if test -f /usr/options/cb.name; then + UNAME_REL=`sed -n 's/.*Version //p' /dev/null >/dev/null ; then + UNAME_REL=`(/bin/uname -X|grep Release|sed -e 's/.*= //')` + (/bin/uname -X|grep i80486 >/dev/null) && UNAME_MACHINE=i486 + (/bin/uname -X|grep '^Machine.*Pentium' >/dev/null) \ + && UNAME_MACHINE=i586 + (/bin/uname -X|grep '^Machine.*Pent *II' >/dev/null) \ + && UNAME_MACHINE=i686 + (/bin/uname -X|grep '^Machine.*Pentium Pro' >/dev/null) \ + && UNAME_MACHINE=i686 + echo "$UNAME_MACHINE"-pc-sco"$UNAME_REL" + else + echo "$UNAME_MACHINE"-pc-sysv32 + fi + exit ;; + pc:*:*:*) + # Left here for compatibility: + # uname -m prints for DJGPP always 'pc', but it prints nothing about + # the processor, so we play safe by assuming i586. + # Note: whatever this is, it MUST be the same as what config.sub + # prints for the "djgpp" host, or else GDB configure will decide that + # this is a cross-build. + echo i586-pc-msdosdjgpp + exit ;; + Intel:Mach:3*:*) + echo i386-pc-mach3 + exit ;; + paragon:*:*:*) + echo i860-intel-osf1 + exit ;; + i860:*:4.*:*) # i860-SVR4 + if grep Stardent /usr/include/sys/uadmin.h >/dev/null 2>&1 ; then + echo i860-stardent-sysv"$UNAME_RELEASE" # Stardent Vistra i860-SVR4 + else # Add other i860-SVR4 vendors below as they are discovered. + echo i860-unknown-sysv"$UNAME_RELEASE" # Unknown i860-SVR4 + fi + exit ;; + mini*:CTIX:SYS*5:*) + # "miniframe" + echo m68010-convergent-sysv + exit ;; + mc68k:UNIX:SYSTEM5:3.51m) + echo m68k-convergent-sysv + exit ;; + M680?0:D-NIX:5.3:*) + echo m68k-diab-dnix + exit ;; + M68*:*:R3V[5678]*:*) + test -r /sysV68 && { echo 'm68k-motorola-sysv'; exit; } ;; + 3[345]??:*:4.0:3.0 | 3[34]??A:*:4.0:3.0 | 3[34]??,*:*:4.0:3.0 | 3[34]??/*:*:4.0:3.0 | 4400:*:4.0:3.0 | 4850:*:4.0:3.0 | SKA40:*:4.0:3.0 | SDS2:*:4.0:3.0 | SHG2:*:4.0:3.0 | S7501*:*:4.0:3.0) + OS_REL='' + test -r /etc/.relid \ + && OS_REL=.`sed -n 's/[^ ]* [^ ]* \([0-9][0-9]\).*/\1/p' < /etc/.relid` + /bin/uname -p 2>/dev/null | grep 86 >/dev/null \ + && { echo i486-ncr-sysv4.3"$OS_REL"; exit; } + /bin/uname -p 2>/dev/null | /bin/grep entium >/dev/null \ + && { echo i586-ncr-sysv4.3"$OS_REL"; exit; } ;; + 3[34]??:*:4.0:* | 3[34]??,*:*:4.0:*) + /bin/uname -p 2>/dev/null | grep 86 >/dev/null \ + && { echo i486-ncr-sysv4; exit; } ;; + NCR*:*:4.2:* | MPRAS*:*:4.2:*) + OS_REL='.3' + test -r /etc/.relid \ + && OS_REL=.`sed -n 's/[^ ]* [^ ]* \([0-9][0-9]\).*/\1/p' < /etc/.relid` + /bin/uname -p 2>/dev/null | grep 86 >/dev/null \ + && { echo i486-ncr-sysv4.3"$OS_REL"; exit; } + /bin/uname -p 2>/dev/null | /bin/grep entium >/dev/null \ + && { echo i586-ncr-sysv4.3"$OS_REL"; exit; } + /bin/uname -p 2>/dev/null | /bin/grep pteron >/dev/null \ + && { echo i586-ncr-sysv4.3"$OS_REL"; exit; } ;; + m68*:LynxOS:2.*:* | m68*:LynxOS:3.0*:*) + echo m68k-unknown-lynxos"$UNAME_RELEASE" + exit ;; + mc68030:UNIX_System_V:4.*:*) + echo m68k-atari-sysv4 + exit ;; + TSUNAMI:LynxOS:2.*:*) + echo sparc-unknown-lynxos"$UNAME_RELEASE" + exit ;; + rs6000:LynxOS:2.*:*) + echo rs6000-unknown-lynxos"$UNAME_RELEASE" + exit ;; + PowerPC:LynxOS:2.*:* | PowerPC:LynxOS:3.[01]*:* | PowerPC:LynxOS:4.[02]*:*) + echo powerpc-unknown-lynxos"$UNAME_RELEASE" + exit ;; + SM[BE]S:UNIX_SV:*:*) + echo mips-dde-sysv"$UNAME_RELEASE" + exit ;; + RM*:ReliantUNIX-*:*:*) + echo mips-sni-sysv4 + exit ;; + RM*:SINIX-*:*:*) + echo mips-sni-sysv4 + exit ;; + *:SINIX-*:*:*) + if uname -p 2>/dev/null >/dev/null ; then + UNAME_MACHINE=`(uname -p) 2>/dev/null` + echo "$UNAME_MACHINE"-sni-sysv4 + else + echo ns32k-sni-sysv + fi + exit ;; + PENTIUM:*:4.0*:*) # Unisys `ClearPath HMP IX 4000' SVR4/MP effort + # says + echo i586-unisys-sysv4 + exit ;; + *:UNIX_System_V:4*:FTX*) + # From Gerald Hewes . + # How about differentiating between stratus architectures? -djm + echo hppa1.1-stratus-sysv4 + exit ;; + *:*:*:FTX*) + # From seanf@swdc.stratus.com. + echo i860-stratus-sysv4 + exit ;; + i*86:VOS:*:*) + # From Paul.Green@stratus.com. + echo "$UNAME_MACHINE"-stratus-vos + exit ;; + *:VOS:*:*) + # From Paul.Green@stratus.com. + echo hppa1.1-stratus-vos + exit ;; + mc68*:A/UX:*:*) + echo m68k-apple-aux"$UNAME_RELEASE" + exit ;; + news*:NEWS-OS:6*:*) + echo mips-sony-newsos6 + exit ;; + R[34]000:*System_V*:*:* | R4000:UNIX_SYSV:*:* | R*000:UNIX_SV:*:*) + if [ -d /usr/nec ]; then + echo mips-nec-sysv"$UNAME_RELEASE" + else + echo mips-unknown-sysv"$UNAME_RELEASE" + fi + exit ;; + BeBox:BeOS:*:*) # BeOS running on hardware made by Be, PPC only. + echo powerpc-be-beos + exit ;; + BeMac:BeOS:*:*) # BeOS running on Mac or Mac clone, PPC only. + echo powerpc-apple-beos + exit ;; + BePC:BeOS:*:*) # BeOS running on Intel PC compatible. + echo i586-pc-beos + exit ;; + BePC:Haiku:*:*) # Haiku running on Intel PC compatible. + echo i586-pc-haiku + exit ;; + x86_64:Haiku:*:*) + echo x86_64-unknown-haiku + exit ;; + SX-4:SUPER-UX:*:*) + echo sx4-nec-superux"$UNAME_RELEASE" + exit ;; + SX-5:SUPER-UX:*:*) + echo sx5-nec-superux"$UNAME_RELEASE" + exit ;; + SX-6:SUPER-UX:*:*) + echo sx6-nec-superux"$UNAME_RELEASE" + exit ;; + SX-7:SUPER-UX:*:*) + echo sx7-nec-superux"$UNAME_RELEASE" + exit ;; + SX-8:SUPER-UX:*:*) + echo sx8-nec-superux"$UNAME_RELEASE" + exit ;; + SX-8R:SUPER-UX:*:*) + echo sx8r-nec-superux"$UNAME_RELEASE" + exit ;; + SX-ACE:SUPER-UX:*:*) + echo sxace-nec-superux"$UNAME_RELEASE" + exit ;; + Power*:Rhapsody:*:*) + echo powerpc-apple-rhapsody"$UNAME_RELEASE" + exit ;; + *:Rhapsody:*:*) + echo "$UNAME_MACHINE"-apple-rhapsody"$UNAME_RELEASE" + exit ;; + arm64:Darwin:*:*) + echo aarch64-apple-darwin"$UNAME_RELEASE" + exit ;; + *:Darwin:*:*) + UNAME_PROCESSOR=`uname -p` + case $UNAME_PROCESSOR in + unknown) UNAME_PROCESSOR=powerpc ;; + esac + if command -v xcode-select > /dev/null 2> /dev/null && \ + ! xcode-select --print-path > /dev/null 2> /dev/null ; then + # Avoid executing cc if there is no toolchain installed as + # cc will be a stub that puts up a graphical alert + # prompting the user to install developer tools. + CC_FOR_BUILD=no_compiler_found + else + set_cc_for_build + fi + if [ "$CC_FOR_BUILD" != no_compiler_found ]; then + if (echo '#ifdef __LP64__'; echo IS_64BIT_ARCH; echo '#endif') | \ + (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) | \ + grep IS_64BIT_ARCH >/dev/null + then + case $UNAME_PROCESSOR in + i386) UNAME_PROCESSOR=x86_64 ;; + powerpc) UNAME_PROCESSOR=powerpc64 ;; + esac + fi + # On 10.4-10.6 one might compile for PowerPC via gcc -arch ppc + if (echo '#ifdef __POWERPC__'; echo IS_PPC; echo '#endif') | \ + (CCOPTS="" $CC_FOR_BUILD -E - 2>/dev/null) | \ + grep IS_PPC >/dev/null + then + UNAME_PROCESSOR=powerpc + fi + elif test "$UNAME_PROCESSOR" = i386 ; then + # uname -m returns i386 or x86_64 + UNAME_PROCESSOR=$UNAME_MACHINE + fi + echo "$UNAME_PROCESSOR"-apple-darwin"$UNAME_RELEASE" + exit ;; + *:procnto*:*:* | *:QNX:[0123456789]*:*) + UNAME_PROCESSOR=`uname -p` + if test "$UNAME_PROCESSOR" = x86; then + UNAME_PROCESSOR=i386 + UNAME_MACHINE=pc + fi + echo "$UNAME_PROCESSOR"-"$UNAME_MACHINE"-nto-qnx"$UNAME_RELEASE" + exit ;; + *:QNX:*:4*) + echo i386-pc-qnx + exit ;; + NEO-*:NONSTOP_KERNEL:*:*) + echo neo-tandem-nsk"$UNAME_RELEASE" + exit ;; + NSE-*:NONSTOP_KERNEL:*:*) + echo nse-tandem-nsk"$UNAME_RELEASE" + exit ;; + NSR-*:NONSTOP_KERNEL:*:*) + echo nsr-tandem-nsk"$UNAME_RELEASE" + exit ;; + NSV-*:NONSTOP_KERNEL:*:*) + echo nsv-tandem-nsk"$UNAME_RELEASE" + exit ;; + NSX-*:NONSTOP_KERNEL:*:*) + echo nsx-tandem-nsk"$UNAME_RELEASE" + exit ;; + *:NonStop-UX:*:*) + echo mips-compaq-nonstopux + exit ;; + BS2000:POSIX*:*:*) + echo bs2000-siemens-sysv + exit ;; + DS/*:UNIX_System_V:*:*) + echo "$UNAME_MACHINE"-"$UNAME_SYSTEM"-"$UNAME_RELEASE" + exit ;; + *:Plan9:*:*) + # "uname -m" is not consistent, so use $cputype instead. 386 + # is converted to i386 for consistency with other x86 + # operating systems. + # shellcheck disable=SC2154 + if test "$cputype" = 386; then + UNAME_MACHINE=i386 + else + UNAME_MACHINE="$cputype" + fi + echo "$UNAME_MACHINE"-unknown-plan9 + exit ;; + *:TOPS-10:*:*) + echo pdp10-unknown-tops10 + exit ;; + *:TENEX:*:*) + echo pdp10-unknown-tenex + exit ;; + KS10:TOPS-20:*:* | KL10:TOPS-20:*:* | TYPE4:TOPS-20:*:*) + echo pdp10-dec-tops20 + exit ;; + XKL-1:TOPS-20:*:* | TYPE5:TOPS-20:*:*) + echo pdp10-xkl-tops20 + exit ;; + *:TOPS-20:*:*) + echo pdp10-unknown-tops20 + exit ;; + *:ITS:*:*) + echo pdp10-unknown-its + exit ;; + SEI:*:*:SEIUX) + echo mips-sei-seiux"$UNAME_RELEASE" + exit ;; + *:DragonFly:*:*) + echo "$UNAME_MACHINE"-unknown-dragonfly"`echo "$UNAME_RELEASE"|sed -e 's/[-(].*//'`" + exit ;; + *:*VMS:*:*) + UNAME_MACHINE=`(uname -p) 2>/dev/null` + case "$UNAME_MACHINE" in + A*) echo alpha-dec-vms ; exit ;; + I*) echo ia64-dec-vms ; exit ;; + V*) echo vax-dec-vms ; exit ;; + esac ;; + *:XENIX:*:SysV) + echo i386-pc-xenix + exit ;; + i*86:skyos:*:*) + echo "$UNAME_MACHINE"-pc-skyos"`echo "$UNAME_RELEASE" | sed -e 's/ .*$//'`" + exit ;; + i*86:rdos:*:*) + echo "$UNAME_MACHINE"-pc-rdos + exit ;; + i*86:AROS:*:*) + echo "$UNAME_MACHINE"-pc-aros + exit ;; + x86_64:VMkernel:*:*) + echo "$UNAME_MACHINE"-unknown-esx + exit ;; + amd64:Isilon\ OneFS:*:*) + echo x86_64-unknown-onefs + exit ;; + *:Unleashed:*:*) + echo "$UNAME_MACHINE"-unknown-unleashed"$UNAME_RELEASE" + exit ;; +esac + +# No uname command or uname output not recognized. +set_cc_for_build +cat > "$dummy.c" < +#include +#endif +#if defined(ultrix) || defined(_ultrix) || defined(__ultrix) || defined(__ultrix__) +#if defined (vax) || defined (__vax) || defined (__vax__) || defined(mips) || defined(__mips) || defined(__mips__) || defined(MIPS) || defined(__MIPS__) +#include +#if defined(_SIZE_T_) || defined(SIGLOST) +#include +#endif +#endif +#endif +main () +{ +#if defined (sony) +#if defined (MIPSEB) + /* BFD wants "bsd" instead of "newsos". Perhaps BFD should be changed, + I don't know.... */ + printf ("mips-sony-bsd\n"); exit (0); +#else +#include + printf ("m68k-sony-newsos%s\n", +#ifdef NEWSOS4 + "4" +#else + "" +#endif + ); exit (0); +#endif +#endif + +#if defined (NeXT) +#if !defined (__ARCHITECTURE__) +#define __ARCHITECTURE__ "m68k" +#endif + int version; + version=`(hostinfo | sed -n 's/.*NeXT Mach \([0-9]*\).*/\1/p') 2>/dev/null`; + if (version < 4) + printf ("%s-next-nextstep%d\n", __ARCHITECTURE__, version); + else + printf ("%s-next-openstep%d\n", __ARCHITECTURE__, version); + exit (0); +#endif + +#if defined (MULTIMAX) || defined (n16) +#if defined (UMAXV) + printf ("ns32k-encore-sysv\n"); exit (0); +#else +#if defined (CMU) + printf ("ns32k-encore-mach\n"); exit (0); +#else + printf ("ns32k-encore-bsd\n"); exit (0); +#endif +#endif +#endif + +#if defined (__386BSD__) + printf ("i386-pc-bsd\n"); exit (0); +#endif + +#if defined (sequent) +#if defined (i386) + printf ("i386-sequent-dynix\n"); exit (0); +#endif +#if defined (ns32000) + printf ("ns32k-sequent-dynix\n"); exit (0); +#endif +#endif + +#if defined (_SEQUENT_) + struct utsname un; + + uname(&un); + if (strncmp(un.version, "V2", 2) == 0) { + printf ("i386-sequent-ptx2\n"); exit (0); + } + if (strncmp(un.version, "V1", 2) == 0) { /* XXX is V1 correct? */ + printf ("i386-sequent-ptx1\n"); exit (0); + } + printf ("i386-sequent-ptx\n"); exit (0); +#endif + +#if defined (vax) +#if !defined (ultrix) +#include +#if defined (BSD) +#if BSD == 43 + printf ("vax-dec-bsd4.3\n"); exit (0); +#else +#if BSD == 199006 + printf ("vax-dec-bsd4.3reno\n"); exit (0); +#else + printf ("vax-dec-bsd\n"); exit (0); +#endif +#endif +#else + printf ("vax-dec-bsd\n"); exit (0); +#endif +#else +#if defined(_SIZE_T_) || defined(SIGLOST) + struct utsname un; + uname (&un); + printf ("vax-dec-ultrix%s\n", un.release); exit (0); +#else + printf ("vax-dec-ultrix\n"); exit (0); +#endif +#endif +#endif +#if defined(ultrix) || defined(_ultrix) || defined(__ultrix) || defined(__ultrix__) +#if defined(mips) || defined(__mips) || defined(__mips__) || defined(MIPS) || defined(__MIPS__) +#if defined(_SIZE_T_) || defined(SIGLOST) + struct utsname *un; + uname (&un); + printf ("mips-dec-ultrix%s\n", un.release); exit (0); +#else + printf ("mips-dec-ultrix\n"); exit (0); +#endif +#endif +#endif + +#if defined (alliant) && defined (i860) + printf ("i860-alliant-bsd\n"); exit (0); +#endif + + exit (1); +} +EOF + +$CC_FOR_BUILD -o "$dummy" "$dummy.c" 2>/dev/null && SYSTEM_NAME=`$dummy` && + { echo "$SYSTEM_NAME"; exit; } + +# Apollos put the system type in the environment. +test -d /usr/apollo && { echo "$ISP-apollo-$SYSTYPE"; exit; } + +echo "$0: unable to guess system type" >&2 + +case "$UNAME_MACHINE:$UNAME_SYSTEM" in + mips:Linux | mips64:Linux) + # If we got here on MIPS GNU/Linux, output extra information. + cat >&2 <&2 <&2 </dev/null || echo unknown` +uname -r = `(uname -r) 2>/dev/null || echo unknown` +uname -s = `(uname -s) 2>/dev/null || echo unknown` +uname -v = `(uname -v) 2>/dev/null || echo unknown` + +/usr/bin/uname -p = `(/usr/bin/uname -p) 2>/dev/null` +/bin/uname -X = `(/bin/uname -X) 2>/dev/null` + +hostinfo = `(hostinfo) 2>/dev/null` +/bin/universe = `(/bin/universe) 2>/dev/null` +/usr/bin/arch -k = `(/usr/bin/arch -k) 2>/dev/null` +/bin/arch = `(/bin/arch) 2>/dev/null` +/usr/bin/oslevel = `(/usr/bin/oslevel) 2>/dev/null` +/usr/convex/getsysinfo = `(/usr/convex/getsysinfo) 2>/dev/null` + +UNAME_MACHINE = "$UNAME_MACHINE" +UNAME_RELEASE = "$UNAME_RELEASE" +UNAME_SYSTEM = "$UNAME_SYSTEM" +UNAME_VERSION = "$UNAME_VERSION" +EOF +fi + +exit 1 + +# Local variables: +# eval: (add-hook 'before-save-hook 'time-stamp) +# time-stamp-start: "timestamp='" +# time-stamp-format: "%:y-%02m-%02d" +# time-stamp-end: "'" +# End: diff --git a/native/libffi/config.sub b/native/libffi/config.sub new file mode 100644 index 0000000000..14b5150c42 --- /dev/null +++ b/native/libffi/config.sub @@ -0,0 +1,1851 @@ +#! /bin/sh +# Configuration validation subroutine script. +# Copyright 1992-2020 Free Software Foundation, Inc. + +timestamp='2020-08-05' + +# This file is free software; you can redistribute it and/or modify it +# under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, but +# WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; if not, see . +# +# As a special exception to the GNU General Public License, if you +# distribute this file as part of a program that contains a +# configuration script generated by Autoconf, you may include it under +# the same distribution terms that you use for the rest of that +# program. This Exception is an additional permission under section 7 +# of the GNU General Public License, version 3 ("GPLv3"). + + +# Please send patches to . +# +# Configuration subroutine to validate and canonicalize a configuration type. +# Supply the specified configuration type as an argument. +# If it is invalid, we print an error message on stderr and exit with code 1. +# Otherwise, we print the canonical config type on stdout and succeed. + +# You can get the latest version of this script from: +# https://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub + +# This file is supposed to be the same for all GNU packages +# and recognize all the CPU types, system types and aliases +# that are meaningful with *any* GNU software. +# Each package is responsible for reporting which valid configurations +# it does not support. The user should be able to distinguish +# a failure to support a valid configuration from a meaningless +# configuration. + +# The goal of this file is to map all the various variations of a given +# machine specification into a single specification in the form: +# CPU_TYPE-MANUFACTURER-OPERATING_SYSTEM +# or in some cases, the newer four-part form: +# CPU_TYPE-MANUFACTURER-KERNEL-OPERATING_SYSTEM +# It is wrong to echo any other type of specification. + +me=`echo "$0" | sed -e 's,.*/,,'` + +usage="\ +Usage: $0 [OPTION] CPU-MFR-OPSYS or ALIAS + +Canonicalize a configuration name. + +Options: + -h, --help print this help, then exit + -t, --time-stamp print date of last modification, then exit + -v, --version print version number, then exit + +Report bugs and patches to ." + +version="\ +GNU config.sub ($timestamp) + +Copyright 1992-2020 Free Software Foundation, Inc. + +This is free software; see the source for copying conditions. There is NO +warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE." + +help=" +Try \`$me --help' for more information." + +# Parse command line +while test $# -gt 0 ; do + case $1 in + --time-stamp | --time* | -t ) + echo "$timestamp" ; exit ;; + --version | -v ) + echo "$version" ; exit ;; + --help | --h* | -h ) + echo "$usage"; exit ;; + -- ) # Stop option processing + shift; break ;; + - ) # Use stdin as input. + break ;; + -* ) + echo "$me: invalid option $1$help" >&2 + exit 1 ;; + + *local*) + # First pass through any local machine types. + echo "$1" + exit ;; + + * ) + break ;; + esac +done + +case $# in + 0) echo "$me: missing argument$help" >&2 + exit 1;; + 1) ;; + *) echo "$me: too many arguments$help" >&2 + exit 1;; +esac + +# Split fields of configuration type +# shellcheck disable=SC2162 +IFS="-" read field1 field2 field3 field4 <&2 + exit 1 + ;; + *-*-*-*) + basic_machine=$field1-$field2 + basic_os=$field3-$field4 + ;; + *-*-*) + # Ambiguous whether COMPANY is present, or skipped and KERNEL-OS is two + # parts + maybe_os=$field2-$field3 + case $maybe_os in + nto-qnx* | linux-* | uclinux-uclibc* \ + | uclinux-gnu* | kfreebsd*-gnu* | knetbsd*-gnu* | netbsd*-gnu* \ + | netbsd*-eabi* | kopensolaris*-gnu* | cloudabi*-eabi* \ + | storm-chaos* | os2-emx* | rtmk-nova*) + basic_machine=$field1 + basic_os=$maybe_os + ;; + android-linux) + basic_machine=$field1-unknown + basic_os=linux-android + ;; + *) + basic_machine=$field1-$field2 + basic_os=$field3 + ;; + esac + ;; + *-*) + # A lone config we happen to match not fitting any pattern + case $field1-$field2 in + decstation-3100) + basic_machine=mips-dec + basic_os= + ;; + *-*) + # Second component is usually, but not always the OS + case $field2 in + # Prevent following clause from handling this valid os + sun*os*) + basic_machine=$field1 + basic_os=$field2 + ;; + # Manufacturers + dec* | mips* | sequent* | encore* | pc533* | sgi* | sony* \ + | att* | 7300* | 3300* | delta* | motorola* | sun[234]* \ + | unicom* | ibm* | next | hp | isi* | apollo | altos* \ + | convergent* | ncr* | news | 32* | 3600* | 3100* \ + | hitachi* | c[123]* | convex* | sun | crds | omron* | dg \ + | ultra | tti* | harris | dolphin | highlevel | gould \ + | cbm | ns | masscomp | apple | axis | knuth | cray \ + | microblaze* | sim | cisco \ + | oki | wec | wrs | winbond) + basic_machine=$field1-$field2 + basic_os= + ;; + *) + basic_machine=$field1 + basic_os=$field2 + ;; + esac + ;; + esac + ;; + *) + # Convert single-component short-hands not valid as part of + # multi-component configurations. + case $field1 in + 386bsd) + basic_machine=i386-pc + basic_os=bsd + ;; + a29khif) + basic_machine=a29k-amd + basic_os=udi + ;; + adobe68k) + basic_machine=m68010-adobe + basic_os=scout + ;; + alliant) + basic_machine=fx80-alliant + basic_os= + ;; + altos | altos3068) + basic_machine=m68k-altos + basic_os= + ;; + am29k) + basic_machine=a29k-none + basic_os=bsd + ;; + amdahl) + basic_machine=580-amdahl + basic_os=sysv + ;; + amiga) + basic_machine=m68k-unknown + basic_os= + ;; + amigaos | amigados) + basic_machine=m68k-unknown + basic_os=amigaos + ;; + amigaunix | amix) + basic_machine=m68k-unknown + basic_os=sysv4 + ;; + apollo68) + basic_machine=m68k-apollo + basic_os=sysv + ;; + apollo68bsd) + basic_machine=m68k-apollo + basic_os=bsd + ;; + aros) + basic_machine=i386-pc + basic_os=aros + ;; + aux) + basic_machine=m68k-apple + basic_os=aux + ;; + balance) + basic_machine=ns32k-sequent + basic_os=dynix + ;; + blackfin) + basic_machine=bfin-unknown + basic_os=linux + ;; + cegcc) + basic_machine=arm-unknown + basic_os=cegcc + ;; + convex-c1) + basic_machine=c1-convex + basic_os=bsd + ;; + convex-c2) + basic_machine=c2-convex + basic_os=bsd + ;; + convex-c32) + basic_machine=c32-convex + basic_os=bsd + ;; + convex-c34) + basic_machine=c34-convex + basic_os=bsd + ;; + convex-c38) + basic_machine=c38-convex + basic_os=bsd + ;; + cray) + basic_machine=j90-cray + basic_os=unicos + ;; + crds | unos) + basic_machine=m68k-crds + basic_os= + ;; + da30) + basic_machine=m68k-da30 + basic_os= + ;; + decstation | pmax | pmin | dec3100 | decstatn) + basic_machine=mips-dec + basic_os= + ;; + delta88) + basic_machine=m88k-motorola + basic_os=sysv3 + ;; + dicos) + basic_machine=i686-pc + basic_os=dicos + ;; + djgpp) + basic_machine=i586-pc + basic_os=msdosdjgpp + ;; + ebmon29k) + basic_machine=a29k-amd + basic_os=ebmon + ;; + es1800 | OSE68k | ose68k | ose | OSE) + basic_machine=m68k-ericsson + basic_os=ose + ;; + gmicro) + basic_machine=tron-gmicro + basic_os=sysv + ;; + go32) + basic_machine=i386-pc + basic_os=go32 + ;; + h8300hms) + basic_machine=h8300-hitachi + basic_os=hms + ;; + h8300xray) + basic_machine=h8300-hitachi + basic_os=xray + ;; + h8500hms) + basic_machine=h8500-hitachi + basic_os=hms + ;; + harris) + basic_machine=m88k-harris + basic_os=sysv3 + ;; + hp300 | hp300hpux) + basic_machine=m68k-hp + basic_os=hpux + ;; + hp300bsd) + basic_machine=m68k-hp + basic_os=bsd + ;; + hppaosf) + basic_machine=hppa1.1-hp + basic_os=osf + ;; + hppro) + basic_machine=hppa1.1-hp + basic_os=proelf + ;; + i386mach) + basic_machine=i386-mach + basic_os=mach + ;; + isi68 | isi) + basic_machine=m68k-isi + basic_os=sysv + ;; + m68knommu) + basic_machine=m68k-unknown + basic_os=linux + ;; + magnum | m3230) + basic_machine=mips-mips + basic_os=sysv + ;; + merlin) + basic_machine=ns32k-utek + basic_os=sysv + ;; + mingw64) + basic_machine=x86_64-pc + basic_os=mingw64 + ;; + mingw32) + basic_machine=i686-pc + basic_os=mingw32 + ;; + mingw32ce) + basic_machine=arm-unknown + basic_os=mingw32ce + ;; + monitor) + basic_machine=m68k-rom68k + basic_os=coff + ;; + morphos) + basic_machine=powerpc-unknown + basic_os=morphos + ;; + moxiebox) + basic_machine=moxie-unknown + basic_os=moxiebox + ;; + msdos) + basic_machine=i386-pc + basic_os=msdos + ;; + msys) + basic_machine=i686-pc + basic_os=msys + ;; + mvs) + basic_machine=i370-ibm + basic_os=mvs + ;; + nacl) + basic_machine=le32-unknown + basic_os=nacl + ;; + ncr3000) + basic_machine=i486-ncr + basic_os=sysv4 + ;; + netbsd386) + basic_machine=i386-pc + basic_os=netbsd + ;; + netwinder) + basic_machine=armv4l-rebel + basic_os=linux + ;; + news | news700 | news800 | news900) + basic_machine=m68k-sony + basic_os=newsos + ;; + news1000) + basic_machine=m68030-sony + basic_os=newsos + ;; + necv70) + basic_machine=v70-nec + basic_os=sysv + ;; + nh3000) + basic_machine=m68k-harris + basic_os=cxux + ;; + nh[45]000) + basic_machine=m88k-harris + basic_os=cxux + ;; + nindy960) + basic_machine=i960-intel + basic_os=nindy + ;; + mon960) + basic_machine=i960-intel + basic_os=mon960 + ;; + nonstopux) + basic_machine=mips-compaq + basic_os=nonstopux + ;; + os400) + basic_machine=powerpc-ibm + basic_os=os400 + ;; + OSE68000 | ose68000) + basic_machine=m68000-ericsson + basic_os=ose + ;; + os68k) + basic_machine=m68k-none + basic_os=os68k + ;; + paragon) + basic_machine=i860-intel + basic_os=osf + ;; + parisc) + basic_machine=hppa-unknown + basic_os=linux + ;; + psp) + basic_machine=mipsallegrexel-sony + basic_os=psp + ;; + pw32) + basic_machine=i586-unknown + basic_os=pw32 + ;; + rdos | rdos64) + basic_machine=x86_64-pc + basic_os=rdos + ;; + rdos32) + basic_machine=i386-pc + basic_os=rdos + ;; + rom68k) + basic_machine=m68k-rom68k + basic_os=coff + ;; + sa29200) + basic_machine=a29k-amd + basic_os=udi + ;; + sei) + basic_machine=mips-sei + basic_os=seiux + ;; + sequent) + basic_machine=i386-sequent + basic_os= + ;; + sps7) + basic_machine=m68k-bull + basic_os=sysv2 + ;; + st2000) + basic_machine=m68k-tandem + basic_os= + ;; + stratus) + basic_machine=i860-stratus + basic_os=sysv4 + ;; + sun2) + basic_machine=m68000-sun + basic_os= + ;; + sun2os3) + basic_machine=m68000-sun + basic_os=sunos3 + ;; + sun2os4) + basic_machine=m68000-sun + basic_os=sunos4 + ;; + sun3) + basic_machine=m68k-sun + basic_os= + ;; + sun3os3) + basic_machine=m68k-sun + basic_os=sunos3 + ;; + sun3os4) + basic_machine=m68k-sun + basic_os=sunos4 + ;; + sun4) + basic_machine=sparc-sun + basic_os= + ;; + sun4os3) + basic_machine=sparc-sun + basic_os=sunos3 + ;; + sun4os4) + basic_machine=sparc-sun + basic_os=sunos4 + ;; + sun4sol2) + basic_machine=sparc-sun + basic_os=solaris2 + ;; + sun386 | sun386i | roadrunner) + basic_machine=i386-sun + basic_os= + ;; + sv1) + basic_machine=sv1-cray + basic_os=unicos + ;; + symmetry) + basic_machine=i386-sequent + basic_os=dynix + ;; + t3e) + basic_machine=alphaev5-cray + basic_os=unicos + ;; + t90) + basic_machine=t90-cray + basic_os=unicos + ;; + toad1) + basic_machine=pdp10-xkl + basic_os=tops20 + ;; + tpf) + basic_machine=s390x-ibm + basic_os=tpf + ;; + udi29k) + basic_machine=a29k-amd + basic_os=udi + ;; + ultra3) + basic_machine=a29k-nyu + basic_os=sym1 + ;; + v810 | necv810) + basic_machine=v810-nec + basic_os=none + ;; + vaxv) + basic_machine=vax-dec + basic_os=sysv + ;; + vms) + basic_machine=vax-dec + basic_os=vms + ;; + vsta) + basic_machine=i386-pc + basic_os=vsta + ;; + vxworks960) + basic_machine=i960-wrs + basic_os=vxworks + ;; + vxworks68) + basic_machine=m68k-wrs + basic_os=vxworks + ;; + vxworks29k) + basic_machine=a29k-wrs + basic_os=vxworks + ;; + xbox) + basic_machine=i686-pc + basic_os=mingw32 + ;; + ymp) + basic_machine=ymp-cray + basic_os=unicos + ;; + *) + basic_machine=$1 + basic_os= + ;; + esac + ;; +esac + +# Decode 1-component or ad-hoc basic machines +case $basic_machine in + # Here we handle the default manufacturer of certain CPU types. It is in + # some cases the only manufacturer, in others, it is the most popular. + w89k) + cpu=hppa1.1 + vendor=winbond + ;; + op50n) + cpu=hppa1.1 + vendor=oki + ;; + op60c) + cpu=hppa1.1 + vendor=oki + ;; + ibm*) + cpu=i370 + vendor=ibm + ;; + orion105) + cpu=clipper + vendor=highlevel + ;; + mac | mpw | mac-mpw) + cpu=m68k + vendor=apple + ;; + pmac | pmac-mpw) + cpu=powerpc + vendor=apple + ;; + + # Recognize the various machine names and aliases which stand + # for a CPU type and a company and sometimes even an OS. + 3b1 | 7300 | 7300-att | att-7300 | pc7300 | safari | unixpc) + cpu=m68000 + vendor=att + ;; + 3b*) + cpu=we32k + vendor=att + ;; + bluegene*) + cpu=powerpc + vendor=ibm + basic_os=cnk + ;; + decsystem10* | dec10*) + cpu=pdp10 + vendor=dec + basic_os=tops10 + ;; + decsystem20* | dec20*) + cpu=pdp10 + vendor=dec + basic_os=tops20 + ;; + delta | 3300 | motorola-3300 | motorola-delta \ + | 3300-motorola | delta-motorola) + cpu=m68k + vendor=motorola + ;; + dpx2*) + cpu=m68k + vendor=bull + basic_os=sysv3 + ;; + encore | umax | mmax) + cpu=ns32k + vendor=encore + ;; + elxsi) + cpu=elxsi + vendor=elxsi + basic_os=${basic_os:-bsd} + ;; + fx2800) + cpu=i860 + vendor=alliant + ;; + genix) + cpu=ns32k + vendor=ns + ;; + h3050r* | hiux*) + cpu=hppa1.1 + vendor=hitachi + basic_os=hiuxwe2 + ;; + hp3k9[0-9][0-9] | hp9[0-9][0-9]) + cpu=hppa1.0 + vendor=hp + ;; + hp9k2[0-9][0-9] | hp9k31[0-9]) + cpu=m68000 + vendor=hp + ;; + hp9k3[2-9][0-9]) + cpu=m68k + vendor=hp + ;; + hp9k6[0-9][0-9] | hp6[0-9][0-9]) + cpu=hppa1.0 + vendor=hp + ;; + hp9k7[0-79][0-9] | hp7[0-79][0-9]) + cpu=hppa1.1 + vendor=hp + ;; + hp9k78[0-9] | hp78[0-9]) + # FIXME: really hppa2.0-hp + cpu=hppa1.1 + vendor=hp + ;; + hp9k8[67]1 | hp8[67]1 | hp9k80[24] | hp80[24] | hp9k8[78]9 | hp8[78]9 | hp9k893 | hp893) + # FIXME: really hppa2.0-hp + cpu=hppa1.1 + vendor=hp + ;; + hp9k8[0-9][13679] | hp8[0-9][13679]) + cpu=hppa1.1 + vendor=hp + ;; + hp9k8[0-9][0-9] | hp8[0-9][0-9]) + cpu=hppa1.0 + vendor=hp + ;; + i*86v32) + cpu=`echo "$1" | sed -e 's/86.*/86/'` + vendor=pc + basic_os=sysv32 + ;; + i*86v4*) + cpu=`echo "$1" | sed -e 's/86.*/86/'` + vendor=pc + basic_os=sysv4 + ;; + i*86v) + cpu=`echo "$1" | sed -e 's/86.*/86/'` + vendor=pc + basic_os=sysv + ;; + i*86sol2) + cpu=`echo "$1" | sed -e 's/86.*/86/'` + vendor=pc + basic_os=solaris2 + ;; + j90 | j90-cray) + cpu=j90 + vendor=cray + basic_os=${basic_os:-unicos} + ;; + iris | iris4d) + cpu=mips + vendor=sgi + case $basic_os in + irix*) + ;; + *) + basic_os=irix4 + ;; + esac + ;; + miniframe) + cpu=m68000 + vendor=convergent + ;; + *mint | mint[0-9]* | *MiNT | *MiNT[0-9]*) + cpu=m68k + vendor=atari + basic_os=mint + ;; + news-3600 | risc-news) + cpu=mips + vendor=sony + basic_os=newsos + ;; + next | m*-next) + cpu=m68k + vendor=next + case $basic_os in + openstep*) + ;; + nextstep*) + ;; + ns2*) + basic_os=nextstep2 + ;; + *) + basic_os=nextstep3 + ;; + esac + ;; + np1) + cpu=np1 + vendor=gould + ;; + op50n-* | op60c-*) + cpu=hppa1.1 + vendor=oki + basic_os=proelf + ;; + pa-hitachi) + cpu=hppa1.1 + vendor=hitachi + basic_os=hiuxwe2 + ;; + pbd) + cpu=sparc + vendor=tti + ;; + pbb) + cpu=m68k + vendor=tti + ;; + pc532) + cpu=ns32k + vendor=pc532 + ;; + pn) + cpu=pn + vendor=gould + ;; + power) + cpu=power + vendor=ibm + ;; + ps2) + cpu=i386 + vendor=ibm + ;; + rm[46]00) + cpu=mips + vendor=siemens + ;; + rtpc | rtpc-*) + cpu=romp + vendor=ibm + ;; + sde) + cpu=mipsisa32 + vendor=sde + basic_os=${basic_os:-elf} + ;; + simso-wrs) + cpu=sparclite + vendor=wrs + basic_os=vxworks + ;; + tower | tower-32) + cpu=m68k + vendor=ncr + ;; + vpp*|vx|vx-*) + cpu=f301 + vendor=fujitsu + ;; + w65) + cpu=w65 + vendor=wdc + ;; + w89k-*) + cpu=hppa1.1 + vendor=winbond + basic_os=proelf + ;; + none) + cpu=none + vendor=none + ;; + leon|leon[3-9]) + cpu=sparc + vendor=$basic_machine + ;; + leon-*|leon[3-9]-*) + cpu=sparc + vendor=`echo "$basic_machine" | sed 's/-.*//'` + ;; + + *-*) + # shellcheck disable=SC2162 + IFS="-" read cpu vendor <&2 + exit 1 + ;; + esac + ;; +esac + +# Here we canonicalize certain aliases for manufacturers. +case $vendor in + digital*) + vendor=dec + ;; + commodore*) + vendor=cbm + ;; + *) + ;; +esac + +# Decode manufacturer-specific aliases for certain operating systems. + +if [ x$basic_os != x ] +then + +# First recognize some ad-hoc caes, or perhaps split kernel-os, or else just +# set os. +case $basic_os in + gnu/linux*) + kernel=linux + os=`echo $basic_os | sed -e 's|gnu/linux|gnu|'` + ;; + nto-qnx*) + kernel=nto + os=`echo $basic_os | sed -e 's|nto-qnx|qnx|'` + ;; + *-*) + # shellcheck disable=SC2162 + IFS="-" read kernel os <&2 + exit 1 + ;; +esac + +# As a final step for OS-related things, validate the OS-kernel combination +# (given a valid OS), if there is a kernel. +case $kernel-$os in + linux-gnu* | linux-dietlibc* | linux-android* | linux-newlib* | linux-musl* | linux-uclibc* ) + ;; + -dietlibc* | -newlib* | -musl* | -uclibc* ) + # These are just libc implementations, not actual OSes, and thus + # require a kernel. + echo "Invalid configuration \`$1': libc \`$os' needs explicit kernel." 1>&2 + exit 1 + ;; + kfreebsd*-gnu* | kopensolaris*-gnu*) + ;; + nto-qnx*) + ;; + *-eabi* | *-gnueabi*) + ;; + -*) + # Blank kernel with real OS is always fine. + ;; + *-*) + echo "Invalid configuration \`$1': Kernel \`$kernel' not known to work with OS \`$os'." 1>&2 + exit 1 + ;; +esac + +# Here we handle the case where we know the os, and the CPU type, but not the +# manufacturer. We pick the logical manufacturer. +case $vendor in + unknown) + case $cpu-$os in + *-riscix*) + vendor=acorn + ;; + *-sunos*) + vendor=sun + ;; + *-cnk* | *-aix*) + vendor=ibm + ;; + *-beos*) + vendor=be + ;; + *-hpux*) + vendor=hp + ;; + *-mpeix*) + vendor=hp + ;; + *-hiux*) + vendor=hitachi + ;; + *-unos*) + vendor=crds + ;; + *-dgux*) + vendor=dg + ;; + *-luna*) + vendor=omron + ;; + *-genix*) + vendor=ns + ;; + *-clix*) + vendor=intergraph + ;; + *-mvs* | *-opened*) + vendor=ibm + ;; + *-os400*) + vendor=ibm + ;; + s390-* | s390x-*) + vendor=ibm + ;; + *-ptx*) + vendor=sequent + ;; + *-tpf*) + vendor=ibm + ;; + *-vxsim* | *-vxworks* | *-windiss*) + vendor=wrs + ;; + *-aux*) + vendor=apple + ;; + *-hms*) + vendor=hitachi + ;; + *-mpw* | *-macos*) + vendor=apple + ;; + *-*mint | *-mint[0-9]* | *-*MiNT | *-MiNT[0-9]*) + vendor=atari + ;; + *-vos*) + vendor=stratus + ;; + esac + ;; +esac + +echo "$cpu-$vendor-${kernel:+$kernel-}$os" +exit + +# Local variables: +# eval: (add-hook 'before-save-hook 'time-stamp) +# time-stamp-start: "timestamp='" +# time-stamp-format: "%:y-%02m-%02d" +# time-stamp-end: "'" +# End: diff --git a/native/libffi/configure.ac b/native/libffi/configure.ac index a7bf5eefae..486226eb5c 100644 --- a/native/libffi/configure.ac +++ b/native/libffi/configure.ac @@ -2,7 +2,11 @@ dnl Process this with autoconf to create configure AC_PREREQ(2.68) +<<<<<<< HEAD AC_INIT([libffi], [3.2.1], [http://github.com/atgreen/libffi/issues]) +======= +AC_INIT([libffi], [3.3], [http://github.com/libffi/libffi/issues]) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 AC_CONFIG_HEADERS([fficonfig.h]) AC_CANONICAL_SYSTEM @@ -72,6 +76,7 @@ AM_CONDITIONAL(TESTSUBDIR, test -d $srcdir/testsuite) TARGETDIR="unknown" HAVE_LONG_DOUBLE_VARIANT=0 +<<<<<<< HEAD case "$host" in aarch64*-*-*) TARGET=AARCH64; TARGETDIR=aarch64 @@ -287,16 +292,16 @@ case "$host" in xtensa*-*) TARGET=XTENSA; TARGETDIR=xtensa ;; +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 -esac +. ${srcdir}/configure.host -AC_SUBST(AM_RUNTESTFLAGS) -AC_SUBST(AM_LTLDFLAGS) - -if test $TARGETDIR = unknown; then +if test -n "${UNSUPPORTED}"; then AC_MSG_ERROR(["libffi has not been ported to $host."]) fi +<<<<<<< HEAD AM_CONDITIONAL(MIPS, test x$TARGET = xMIPS) AM_CONDITIONAL(BFIN, test x$TARGET = xBFIN) AM_CONDITIONAL(SPARC, test x$TARGET = xSPARC) @@ -337,6 +342,10 @@ AM_CONDITIONAL(PA64_HPUX, test x$TARGET = xPA64_HPUX) AM_CONDITIONAL(TILE, test x$TARGET = xTILE) AM_CONDITIONAL(VAX, test x$TARGET = xVAX) AM_CONDITIONAL(XTENSA, test x$TARGET = xXTENSA) +======= +AC_SUBST(AM_RUNTESTFLAGS) +AC_SUBST(AM_LTLDFLAGS) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 AC_HEADER_STDC AC_CHECK_FUNCS(memcpy) @@ -367,7 +376,8 @@ AC_C_BIGENDIAN GCC_AS_CFI_PSEUDO_OP -if test x$TARGET = xSPARC; then +case "$TARGET" in + SPARC) AC_CACHE_CHECK([assembler and linker support unaligned pc related relocs], libffi_cv_as_sparc_ua_pcrel, [ save_CFLAGS="$CFLAGS" @@ -396,9 +406,9 @@ if test x$TARGET = xSPARC; then AC_DEFINE(HAVE_AS_REGISTER_PSEUDO_OP, 1, [Define if your assembler supports .register.]) fi -fi + ;; -if test x$TARGET = xX86 || test x$TARGET = xX86_WIN32 || test x$TARGET = xX86_64; then + X86*) AC_CACHE_CHECK([assembler supports pc related relocs], libffi_cv_as_x86_pcrel, [ libffi_cv_as_x86_pcrel=no @@ -411,32 +421,46 @@ if test x$TARGET = xX86 || test x$TARGET = xX86_WIN32 || test x$TARGET = xX86_64 AC_DEFINE(HAVE_AS_X86_PCREL, 1, [Define if your assembler supports PC relative relocs.]) fi + ;; - AC_CACHE_CHECK([assembler .ascii pseudo-op support], - libffi_cv_as_ascii_pseudo_op, [ - libffi_cv_as_ascii_pseudo_op=unknown - # Check if we have .ascii - AC_TRY_COMPILE(,[asm (".ascii \\"string\\"");], - [libffi_cv_as_ascii_pseudo_op=yes], - [libffi_cv_as_ascii_pseudo_op=no]) - ]) - if test "x$libffi_cv_as_ascii_pseudo_op" = xyes; then - AC_DEFINE(HAVE_AS_ASCII_PSEUDO_OP, 1, - [Define if your assembler supports .ascii.]) + S390) + AC_CACHE_CHECK([compiler uses zarch features], + libffi_cv_as_s390_zarch, [ + libffi_cv_as_s390_zarch=no + echo 'void foo(void) { bar(); bar(); }' > conftest.c + if $CC $CFLAGS -S conftest.c > /dev/null 2>&1; then + if grep -q brasl conftest.s; then + libffi_cv_as_s390_zarch=yes + fi + fi + ]) + if test "x$libffi_cv_as_s390_zarch" = xyes; then + AC_DEFINE(HAVE_AS_S390_ZARCH, 1, + [Define if the compiler uses zarch features.]) fi + ;; +esac - AC_CACHE_CHECK([assembler .string pseudo-op support], - libffi_cv_as_string_pseudo_op, [ - libffi_cv_as_string_pseudo_op=unknown - # Check if we have .string - AC_TRY_COMPILE(,[asm (".string \\"string\\"");], - [libffi_cv_as_string_pseudo_op=yes], - [libffi_cv_as_string_pseudo_op=no]) - ]) - if test "x$libffi_cv_as_string_pseudo_op" = xyes; then - AC_DEFINE(HAVE_AS_STRING_PSEUDO_OP, 1, - [Define if your assembler supports .string.]) - fi +AC_CACHE_CHECK([whether compiler supports pointer authentication], + libffi_cv_as_ptrauth, [ + libffi_cv_as_ptrauth=unknown + AC_TRY_COMPILE(,[ +#ifdef __clang__ +# if __has_feature(ptrauth_calls) +# define HAVE_PTRAUTH 1 +# endif +#endif + +#ifndef HAVE_PTRAUTH +# error Pointer authentication not supported +#endif + ], + [libffi_cv_as_ptrauth=yes], + [libffi_cv_as_ptrauth=no]) +]) +if test "x$libffi_cv_as_ptrauth" = xyes; then + AC_DEFINE(HAVE_PTRAUTH, 1, + [Define if your compiler supports pointer authentication.]) fi # On PaX enable kernels that have MPROTECT enable we can't use PROT_EXEC. @@ -454,13 +478,13 @@ fi FFI_EXEC_TRAMPOLINE_TABLE=0 case "$target" in - *arm*-apple-darwin*) + *arm*-apple-* | aarch64-apple-*) FFI_EXEC_TRAMPOLINE_TABLE=1 AC_DEFINE(FFI_EXEC_TRAMPOLINE_TABLE, 1, [Cannot use PROT_EXEC on this target, so, we revert to alternative means]) ;; - *-apple-darwin1* | *-*-freebsd* | *-*-kfreebsd* | *-*-openbsd* | *-pc-solaris*) + *-apple-* | *-*-freebsd* | *-*-kfreebsd* | *-*-openbsd* | *-pc-solaris* | *-linux-android*) AC_DEFINE(FFI_MMAP_EXEC_WRIT, 1, [Cannot use malloc on this target, so, we revert to alternative means]) @@ -501,18 +525,16 @@ EOF fi if test "x$GCC" = "xyes"; then + AX_CHECK_COMPILE_FLAG(-fno-lto, libffi_cv_no_lto=-fno-lto) + AC_CACHE_CHECK([whether .eh_frame section should be read-only], libffi_cv_ro_eh_frame, [ - libffi_cv_ro_eh_frame=no + libffi_cv_ro_eh_frame=yes echo 'extern void foo (void); void bar (void) { foo (); foo (); }' > conftest.c - if $CC $CFLAGS -c -fpic -fexceptions -o conftest.o conftest.c > /dev/null 2>&1; then - objdump -h conftest.o > conftest.dump 2>&1 - libffi_eh_frame_line=`grep -n eh_frame conftest.dump | cut -d: -f 1` - libffi_test_line=`expr $libffi_eh_frame_line + 1`p - sed -n $libffi_test_line conftest.dump > conftest.line - if grep READONLY conftest.line > /dev/null; then - libffi_cv_ro_eh_frame=yes - fi + if $CC $CFLAGS -c -fpic -fexceptions $libffi_cv_no_lto -o conftest.o conftest.c > /dev/null 2>&1; then + if readelf -WS conftest.o | grep -q -n 'eh_frame .* WA'; then + libffi_cv_ro_eh_frame=no + fi fi rm -f conftest.* ]) @@ -531,7 +553,7 @@ if test "x$GCC" = "xyes"; then echo 'int __attribute__ ((visibility ("hidden"))) foo (void) { return 1 ; }' > conftest.c libffi_cv_hidden_visibility_attribute=no if AC_TRY_COMMAND(${CC-cc} -Werror -S conftest.c -o conftest.s 1>&AS_MESSAGE_LOG_FD); then - if grep '\.hidden.*foo' conftest.s >/dev/null; then + if egrep '(\.hidden|\.private_extern).*foo' conftest.s >/dev/null; then libffi_cv_hidden_visibility_attribute=yes fi fi @@ -543,10 +565,21 @@ if test "x$GCC" = "xyes"; then fi fi +AC_ARG_ENABLE(docs, + AC_HELP_STRING([--disable-docs], + [Disable building of docs (default: no)]), + [enable_docs=no], + [enable_docs=yes]) +AM_CONDITIONAL(BUILD_DOCS, [test x$enable_docs = xyes]) + AH_BOTTOM([ #ifdef HAVE_HIDDEN_VISIBILITY_ATTRIBUTE #ifdef LIBFFI_ASM +#ifdef __APPLE__ +#define FFI_HIDDEN(name) .private_extern name +#else #define FFI_HIDDEN(name) .hidden name +#endif #else #define FFI_HIDDEN __attribute__ ((visibility ("hidden"))) #endif @@ -562,6 +595,14 @@ AH_BOTTOM([ AC_SUBST(TARGET) AC_SUBST(TARGETDIR) +changequote(<,>) +TARGET_OBJ= +for i in $SOURCES; do + TARGET_OBJ="${TARGET_OBJ} src/${TARGETDIR}/"`echo $i | sed 's/[cS]$/lo/'` +done +changequote([,]) +AC_SUBST(TARGET_OBJ) + AC_SUBST(SHELL) AC_ARG_ENABLE(debug, @@ -590,11 +631,16 @@ AC_ARG_ENABLE(purify-safety, AC_DEFINE(USING_PURIFY, 1, [Define this if you are using Purify and want to suppress spurious messages.]) fi) +AC_ARG_ENABLE(multi-os-directory, +[ --disable-multi-os-directory + disable use of gcc --print-multi-os-directory to change the library installation directory]) + # These variables are only ever used when we cross-build to X86_WIN32. # And we only support this with GCC, so... if test "x$GCC" = "xyes"; then if test -n "$with_cross_host" && test x"$with_cross_host" != x"no"; then +<<<<<<< HEAD toolexecdir="${exec_prefix}"/'$(target_alias)' toolexeclibdir="${toolexecdir}"/lib else @@ -609,17 +655,40 @@ if test "x$GCC" = "xyes"; then AC_SUBST(toolexecdir) else toolexeclibdir="${libdir}" +======= + toolexecdir='${exec_prefix}'/'$(target_alias)' + toolexeclibdir='${toolexecdir}'/lib + else + toolexecdir='${libdir}'/gcc-lib/'$(target_alias)' + toolexeclibdir='${libdir}' + fi + if test x"$enable_multi_os_directory" != x"no"; then + multi_os_directory=`$CC $CFLAGS -print-multi-os-directory` + case $multi_os_directory in + .) ;; # Avoid trailing /. + ../*) toolexeclibdir=$toolexeclibdir/$multi_os_directory ;; + esac + fi + AC_SUBST(toolexecdir) +else + toolexeclibdir='${libdir}' +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 fi AC_SUBST(toolexeclibdir) +# Check linker support. +LIBFFI_ENABLE_SYMVERS + AC_CONFIG_COMMANDS(include, [test -d include || mkdir include]) AC_CONFIG_COMMANDS(src, [ test -d src || mkdir src test -d src/$TARGETDIR || mkdir src/$TARGETDIR ], [TARGETDIR="$TARGETDIR"]) -AC_CONFIG_LINKS(include/ffitarget.h:src/$TARGETDIR/ffitarget.h) - -AC_CONFIG_FILES(include/Makefile include/ffi.h Makefile testsuite/Makefile man/Makefile libffi.pc) +AC_CONFIG_FILES(include/Makefile include/ffi.h Makefile testsuite/Makefile man/Makefile doc/Makefile libffi.pc) AC_OUTPUT + +# Copy this file instead of using AC_CONFIG_LINK in order to support +# compiling with MSVC, which won't understand cygwin style symlinks. +cp ${srcdir}/src/$TARGETDIR/ffitarget.h include/ffitarget.h diff --git a/native/libffi/configure.host b/native/libffi/configure.host new file mode 100644 index 0000000000..d9623e92f6 --- /dev/null +++ b/native/libffi/configure.host @@ -0,0 +1,326 @@ +# configure.host +# +# This shell script handles all host based configuration for libffi. +# + +# THIS TABLE IS SORTED. KEEP IT THAT WAY. +# Most of the time we can define all the variables all at once... +case "${host}" in + aarch64*-*-cygwin* | aarch64*-*-mingw* | aarch64*-*-win* ) + TARGET=ARM_WIN64; TARGETDIR=aarch64 + if test "${ax_cv_c_compiler_vendor}" = "microsoft"; then + MSVC=1 + fi + ;; + + aarch64*-*-*) + TARGET=AARCH64; TARGETDIR=aarch64 + SOURCES="ffi.c sysv.S" + ;; + + alpha*-*-*) + TARGET=ALPHA; TARGETDIR=alpha; + # Support 128-bit long double, changeable via command-line switch. + HAVE_LONG_DOUBLE='defined(__LONG_DOUBLE_128__)' + SOURCES="ffi.c osf.S" + ;; + + arc*-*-*) + TARGET=ARC; TARGETDIR=arc + SOURCES="ffi.c arcompact.S" + ;; + + arm*-*-cygwin* | arm*-*-mingw* | arm*-*-win* ) + TARGET=ARM_WIN32; TARGETDIR=arm + MSVC=1 + ;; + + arm*-*-*) + TARGET=ARM; TARGETDIR=arm + SOURCES="ffi.c sysv.S" + ;; + + avr32*-*-*) + TARGET=AVR32; TARGETDIR=avr32 + SOURCES="ffi.c sysv.S" + ;; + + bfin*) + TARGET=BFIN; TARGETDIR=bfin + SOURCES="ffi.c sysv.S" + ;; + + cris-*-*) + TARGET=LIBFFI_CRIS; TARGETDIR=cris + SOURCES="ffi.c sysv.S" + ;; + + csky-*-*) + TARGET=CSKY; TARGETDIR=csky + SOURCES="ffi.c sysv.S" + ;; + + frv-*-*) + TARGET=FRV; TARGETDIR=frv + SOURCES="ffi.c eabi.S" + ;; + + hppa*-*-linux* | parisc*-*-linux* | hppa*-*-openbsd*) + TARGET=PA_LINUX; TARGETDIR=pa + SOURCES="ffi.c linux.S" + ;; + hppa*64-*-hpux*) + TARGET=PA64_HPUX; TARGETDIR=pa + ;; + hppa*-*-hpux*) + TARGET=PA_HPUX; TARGETDIR=pa + SOURCES="ffi.c hpux32.S" + ;; + + i?86-*-freebsd* | i?86-*-openbsd*) + TARGET=X86_FREEBSD; TARGETDIR=x86 + ;; + + i?86-*-cygwin* | i?86-*-mingw* | i?86-*-win* | i?86-*-os2* | i?86-*-interix* \ + | x86_64-*-cygwin* | x86_64-*-mingw* | x86_64-*-win* ) + TARGETDIR=x86 + if test $ac_cv_sizeof_size_t = 4; then + TARGET=X86_WIN32 + else + TARGET=X86_WIN64 + fi + if test "${ax_cv_c_compiler_vendor}" = "microsoft"; then + MSVC=1 + fi + # All mingw/cygwin/win32 builds require -no-undefined for sharedlib. + # We must also check with_cross_host to decide if this is a native + # or cross-build and select where to install dlls appropriately. + if test -n "$with_cross_host" && + test x"$with_cross_host" != x"no"; then + AM_LTLDFLAGS='-no-undefined -bindir "$(toolexeclibdir)"'; + else + AM_LTLDFLAGS='-no-undefined -bindir "$(bindir)"'; + fi + ;; + + i?86-*-darwin* | x86_64-*-darwin* | i?86-*-ios | x86_64-*-ios) + TARGETDIR=x86 + if test $ac_cv_sizeof_size_t = 4; then + TARGET=X86_DARWIN + else + TARGET=X86_64 + fi + ;; + + i?86-*-* | x86_64-*-* | amd64-*) + TARGETDIR=x86 + if test $ac_cv_sizeof_size_t = 4; then + case "$host" in + x86_64-*x32|x86_64-x32-*) + TARGET_X32=yes + TARGET=X86_64 + ;; + *) + echo 'int foo (void) { return __x86_64__; }' > conftest.c + if $CC $CFLAGS -Werror -S conftest.c -o conftest.s > /dev/null 2>&1; then + TARGET_X32=yes + TARGET=X86_64 + else + TARGET=X86; + fi + rm -f conftest.* + ;; + esac + else + TARGET=X86_64; + fi + ;; + + ia64*-*-*) + TARGET=IA64; TARGETDIR=ia64 + SOURCES="ffi.c unix.S" + ;; + + kvx-*-*) + TARGET=KVX; TARGETDIR=kvx + SOURCES="ffi.c sysv.S" + ;; + + m32r*-*-*) + TARGET=M32R; TARGETDIR=m32r + SOURCES="ffi.c sysv.S" + ;; + + m68k-*-*) + TARGET=M68K; TARGETDIR=m68k + SOURCES="ffi.c sysv.S" + ;; + + m88k-*-*) + TARGET=M88K; TARGETDIR=m88k + SOURCES="ffi.c obsd.S" + ;; + + microblaze*-*-*) + TARGET=MICROBLAZE; TARGETDIR=microblaze + SOURCES="ffi.c sysv.S" + ;; + + moxie-*-*) + TARGET=MOXIE; TARGETDIR=moxie + SOURCES="ffi.c eabi.S" + ;; + + metag-*-*) + TARGET=METAG; TARGETDIR=metag + SOURCES="ffi.c sysv.S" + ;; + + mips-sgi-irix5.* | mips-sgi-irix6.* | mips*-*-rtems*) + TARGET=MIPS; TARGETDIR=mips + ;; + mips*-*linux* | mips*-*-openbsd* | mips*-*-freebsd*) + # Support 128-bit long double for NewABI. + HAVE_LONG_DOUBLE='defined(__mips64)' + TARGET=MIPS; TARGETDIR=mips + ;; + + nios2*-linux*) + TARGET=NIOS2; TARGETDIR=nios2 + SOURCES="ffi.c sysv.S" + ;; + + or1k*-*-*) + TARGET=OR1K; TARGETDIR=or1k + SOURCES="ffi.c sysv.S" + ;; + + powerpc*-*-linux* | powerpc-*-sysv*) + TARGET=POWERPC; TARGETDIR=powerpc + HAVE_LONG_DOUBLE_VARIANT=1 + ;; + powerpc-*-amigaos*) + TARGET=POWERPC; TARGETDIR=powerpc + ;; + powerpc-*-eabi*) + TARGET=POWERPC; TARGETDIR=powerpc + ;; + powerpc-*-beos*) + TARGET=POWERPC; TARGETDIR=powerpc + ;; + powerpc-*-darwin* | powerpc64-*-darwin*) + TARGET=POWERPC_DARWIN; TARGETDIR=powerpc + ;; + powerpc-*-aix* | rs6000-*-aix*) + TARGET=POWERPC_AIX; TARGETDIR=powerpc + ;; + powerpc-*-freebsd* | powerpc-*-openbsd* | powerpc-*-netbsd*) + TARGET=POWERPC_FREEBSD; TARGETDIR=powerpc + HAVE_LONG_DOUBLE_VARIANT=1 + ;; + powerpcspe-*-freebsd*) + TARGET=POWERPC_FREEBSD; TARGETDIR=powerpc + CFLAGS="$CFLAGS -D__NO_FPRS__" + ;; + powerpc64-*-freebsd* | powerpc64le-*-freebsd*) + TARGET=POWERPC; TARGETDIR=powerpc + ;; + powerpc*-*-rtems*) + TARGET=POWERPC; TARGETDIR=powerpc + ;; + + riscv*-*) + TARGET=RISCV; TARGETDIR=riscv + SOURCES="ffi.c sysv.S" + ;; + + s390-*-* | s390x-*-*) + TARGET=S390; TARGETDIR=s390 + SOURCES="ffi.c sysv.S" + ;; + + sh-*-* | sh[34]*-*-*) + TARGET=SH; TARGETDIR=sh + SOURCES="ffi.c sysv.S" + ;; + sh64-*-* | sh5*-*-*) + TARGET=SH64; TARGETDIR=sh64 + SOURCES="ffi.c sysv.S" + ;; + + sparc*-*-*) + TARGET=SPARC; TARGETDIR=sparc + SOURCES="ffi.c ffi64.c v8.S v9.S" + ;; + + tile*-*) + TARGET=TILE; TARGETDIR=tile + SOURCES="ffi.c tile.S" + ;; + + vax-*-*) + TARGET=VAX; TARGETDIR=vax + SOURCES="ffi.c elfbsd.S" + ;; + + xtensa*-*) + TARGET=XTENSA; TARGETDIR=xtensa + SOURCES="ffi.c sysv.S" + ;; +esac + +# ... but some of the cases above share configury. +case "${TARGET}" in + ARM_WIN32) + SOURCES="ffi.c sysv_msvc_arm32.S" + ;; + ARM_WIN64) + if test "$MSVC" = 1; then + SOURCES="ffi.c win64_armasm.S" + else + SOURCES="ffi.c sysv.S" + fi + ;; + MIPS) + SOURCES="ffi.c o32.S n32.S" + ;; + POWERPC) + SOURCES="ffi.c ffi_sysv.c ffi_linux64.c sysv.S ppc_closure.S" + SOURCES="${SOURCES} linux64.S linux64_closure.S" + ;; + POWERPC_AIX) + SOURCES="ffi_darwin.c aix.S aix_closure.S" + ;; + POWERPC_DARWIN) + SOURCES="ffi_darwin.c darwin.S darwin_closure.S" + ;; + POWERPC_FREEBSD) + SOURCES="ffi.c ffi_sysv.c sysv.S ppc_closure.S" + ;; + X86 | X86_DARWIN | X86_FREEBSD | X86_WIN32) + if test "$MSVC" = 1; then + SOURCES="ffi.c sysv_intel.S" + else + SOURCES="ffi.c sysv.S" + fi + ;; + X86_64) + if test x"$TARGET_X32" = xyes; then + SOURCES="ffi64.c unix64.S" + else + SOURCES="ffi64.c unix64.S ffiw64.c win64.S" + fi + ;; + X86_WIN64) + if test "$MSVC" = 1; then + SOURCES="ffiw64.c win64_intel.S" + else + SOURCES="ffiw64.c win64.S" + fi + ;; +esac + +# If we failed to configure SOURCES, we can't do anything. +if test -z "${SOURCES}"; then + UNSUPPORTED=1 +fi diff --git a/native/libffi/doc/Makefile.am b/native/libffi/doc/Makefile.am new file mode 100644 index 0000000000..43b650aea7 --- /dev/null +++ b/native/libffi/doc/Makefile.am @@ -0,0 +1,3 @@ +## Process this with automake to create Makefile.in + +info_TEXINFOS = libffi.texi diff --git a/native/libffi/doc/libffi.texi b/native/libffi/doc/libffi.texi index b1c9bc367b..f52172fa6d 100644 --- a/native/libffi/doc/libffi.texi +++ b/native/libffi/doc/libffi.texi @@ -1,7 +1,8 @@ \input texinfo @c -*-texinfo-*- @c %**start of header @setfilename libffi.info -@settitle libffi +@include version.texi +@settitle libffi: the portable foreign function interface library @setchapternewpage off @c %**end of header @@ -12,32 +13,43 @@ @syncodeindex pg cp @syncodeindex tp cp -@include version.texi - @copying -This manual is for Libffi, a portable foreign-function interface +This manual is for libffi, a portable foreign function interface library. -Copyright @copyright{} 2008, 2010, 2011 Red Hat, Inc. +Copyright @copyright{} 2008--2019 Anthony Green and Red Hat, Inc. + +Permission is hereby granted, free of charge, to any person obtaining +a copy of this software and associated documentation files (the +``Software''), to deal in the Software without restriction, including +without limitation the rights to use, copy, modify, merge, publish, +distribute, sublicense, and/or sell copies of the Software, and to +permit persons to whom the Software is furnished to do so, subject to +the following conditions: -@quotation -Permission is granted to copy, distribute and/or modify this document -under the terms of the GNU General Public License as published by the -Free Software Foundation; either version 2, or (at your option) any -later version. A copy of the license is included in the -section entitled ``GNU General Public License''. +The above copyright notice and this permission notice shall be +included in all copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, +EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF +MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. +IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY +CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, +TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE +SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. -@end quotation @end copying @dircategory Development @direntry -* libffi: (libffi). Portable foreign-function interface library. +* libffi: (libffi). Portable foreign function interface library. @end direntry @titlepage -@title Libffi +@title libffi: a foreign function interface library +@subtitle For Version @value{VERSION} of libffi +@author Anthony Green @page @vskip 0pt plus 1filll @insertcopying @@ -107,6 +119,7 @@ values passed between the two languages. * Multiple ABIs:: Different passing styles on one platform. * The Closure API:: Writing a generic function. * Closure Example:: A closure example. +* Thread Safety:: Thread safety. @end menu @@ -152,16 +165,16 @@ If the function being called is variadic (varargs) then @code{ffi_prep_cif_var} must be used instead of @code{ffi_prep_cif}. @findex ffi_prep_cif_var -@defun ffi_status ffi_prep_cif_var (ffi_cif *@var{cif}, ffi_abi var{abi}, unsigned int @var{nfixedargs}, unsigned int var{ntotalargs}, ffi_type *@var{rtype}, ffi_type **@var{argtypes}) +@defun ffi_status ffi_prep_cif_var (ffi_cif *@var{cif}, ffi_abi @var{abi}, unsigned int @var{nfixedargs}, unsigned int @var{ntotalargs}, ffi_type *@var{rtype}, ffi_type **@var{argtypes}) This initializes @var{cif} according to the given parameters for -a call to a variadic function. In general it's operation is the +a call to a variadic function. In general its operation is the same as for @code{ffi_prep_cif} except that: @var{nfixedargs} is the number of fixed arguments, prior to any variadic arguments. It must be greater than zero. @var{ntotalargs} the total number of arguments, including variadic -and fixed arguments. +and fixed arguments. @var{argtypes} must have this many elements. Note that, different cif's must be prepped for calls to the same function when different numbers of arguments are passed. @@ -172,6 +185,10 @@ Also note that a call to @code{ffi_prep_cif_var} with @end defun +Note that the resulting @code{ffi_cif} holds pointers to all the +@code{ffi_type} objects that were used during initialization. You +must ensure that these type objects have a lifetime at least as long +as that of the @code{ffi_cif}. To call a function using an initialized @code{ffi_cif}, use the @code{ffi_call} function: @@ -189,6 +206,19 @@ bits), and must be suitably aligned; it is the caller's responsibility to ensure this. If @var{cif} declares that the function returns @code{void} (using @code{ffi_type_void}), then @var{rvalue} is ignored. +<<<<<<< HEAD +======= + +In most situations, @samp{libffi} will handle promotion according to +the ABI. However, for historical reasons, there is a special case +with return values that must be handled by your code. In particular, +for integral (not @code{struct}) types that are narrower than the +system register size, the return value will be widened by +@samp{libffi}. @samp{libffi} provides a type, @code{ffi_arg}, that +can be used as the return type. For example, if the CIF was defined +with a return type of @code{char}, @samp{libffi} will try to store a +full @code{ffi_arg} into the return value. +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 @var{avalues} is a vector of @code{void *} pointers that point to the memory locations holding the argument values for a call. If @var{cif} @@ -196,6 +226,13 @@ declares that the function has no arguments (i.e., @var{nargs} was 0), then @var{avalues} is ignored. Note that argument values may be modified by the callee (for instance, structs passed by value); the burden of copying pass-by-value arguments is placed on the caller. + +Note that while the return value must be register-sized, arguments +should exactly match their declared type. For example, if an argument +is a @code{short}, then the entry in @var{avalues} should point to an +object declared as @code{short}; but if the return type is +@code{short}, then @var{rvalue} should point to an object declared as +a larger type -- usually @code{ffi_arg}. @end defun @@ -246,6 +283,8 @@ int main() @menu * Primitive Types:: Built-in types. * Structures:: Structure types. +* Size and Alignment:: Size and alignment of types. +* Arrays Unions Enums:: Arrays, unions, and enumerations. * Type Example:: Structure type example. * Complex:: Complex types. * Complex Type Example:: Complex type example. @@ -370,8 +409,7 @@ when passing to @code{ffi_prep_cif}. @node Structures @subsection Structures -Although @samp{libffi} has no special support for unions or -bit-fields, it is perfectly happy passing structures back and forth. +@samp{libffi} is perfectly happy passing structures back and forth. You must first describe the structure to @samp{libffi} by creating a new @code{ffi_type} object for it. @@ -391,9 +429,166 @@ For a structure, this should be set to @code{FFI_TYPE_STRUCT}. @item ffi_type **elements This is a @samp{NULL}-terminated array of pointers to @code{ffi_type} objects. There is one element per field of the struct. + +Note that @samp{libffi} has no special support for bit-fields. You +must manage these manually. @end table @end deftp +The @code{size} and @code{alignment} fields will be filled in by +@code{ffi_prep_cif} or @code{ffi_prep_cif_var}, as needed. + +@node Size and Alignment +@subsection Size and Alignment + +@code{libffi} will set the @code{size} and @code{alignment} fields of +an @code{ffi_type} object for you. It does so using its knowledge of +the ABI. + +You might expect that you can simply read these fields for a type that +has been laid out by @code{libffi}. However, there are some caveats. + +@itemize @bullet +@item +The size or alignment of some of the built-in types may vary depending +on the chosen ABI. + +@item +The size and alignment of a new structure type will not be set by +@code{libffi} until it has been passed to @code{ffi_prep_cif} or +@code{ffi_get_struct_offsets}. + +@item +A structure type cannot be shared across ABIs. Instead each ABI needs +its own copy of the structure type. +@end itemize + +So, before examining these fields, it is safest to pass the +@code{ffi_type} object to @code{ffi_prep_cif} or +@code{ffi_get_struct_offsets} first. This function will do all the +needed setup. + +@example +ffi_type *desired_type; +ffi_abi desired_abi; +@dots{} +ffi_cif cif; +if (ffi_prep_cif (&cif, desired_abi, 0, desired_type, NULL) == FFI_OK) + @{ + size_t size = desired_type->size; + unsigned short alignment = desired_type->alignment; + @} +@end example + +@code{libffi} also provides a way to get the offsets of the members of +a structure. + +@findex ffi_get_struct_offsets +@defun ffi_status ffi_get_struct_offsets (ffi_abi abi, ffi_type *struct_type, size_t *offsets) +Compute the offset of each element of the given structure type. +@var{abi} is the ABI to use; this is needed because in some cases the +layout depends on the ABI. + +@var{offsets} is an out parameter. The caller is responsible for +providing enough space for all the results to be written -- one +element per element type in @var{struct_type}. If @var{offsets} is +@code{NULL}, then the type will be laid out but not otherwise +modified. This can be useful for accessing the type's size or layout, +as mentioned above. + +This function returns @code{FFI_OK} on success; @code{FFI_BAD_ABI} if +@var{abi} is invalid; or @code{FFI_BAD_TYPEDEF} if @var{struct_type} +is invalid in some way. Note that only @code{FFI_STRUCT} types are +valid here. +@end defun + +@node Arrays Unions Enums +@subsection Arrays, Unions, and Enumerations + +@subsubsection Arrays + +@samp{libffi} does not have direct support for arrays or unions. +However, they can be emulated using structures. + +To emulate an array, simply create an @code{ffi_type} using +@code{FFI_TYPE_STRUCT} with as many members as there are elements in +the array. + +@example +ffi_type array_type; +ffi_type **elements +int i; + +elements = malloc ((n + 1) * sizeof (ffi_type *)); +for (i = 0; i < n; ++i) + elements[i] = array_element_type; +elements[n] = NULL; + +array_type.size = array_type.alignment = 0; +array_type.type = FFI_TYPE_STRUCT; +array_type.elements = elements; +@end example + +Note that arrays cannot be passed or returned by value in C -- +structure types created like this should only be used to refer to +members of real @code{FFI_TYPE_STRUCT} objects. + +However, a phony array type like this will not cause any errors from +@samp{libffi} if you use it as an argument or return type. This may +be confusing. + +@subsubsection Unions + +A union can also be emulated using @code{FFI_TYPE_STRUCT}. In this +case, however, you must make sure that the size and alignment match +the real requirements of the union. + +One simple way to do this is to ensue that each element type is laid +out. Then, give the new structure type a single element; the size of +the largest element; and the largest alignment seen as well. + +This example uses the @code{ffi_prep_cif} trick to ensure that each +element type is laid out. + +@example +ffi_abi desired_abi; +ffi_type union_type; +ffi_type **union_elements; + +int i; +ffi_type element_types[2]; + +element_types[1] = NULL; + +union_type.size = union_type.alignment = 0; +union_type.type = FFI_TYPE_STRUCT; +union_type.elements = element_types; + +for (i = 0; union_elements[i]; ++i) + @{ + ffi_cif cif; + if (ffi_prep_cif (&cif, desired_abi, 0, union_elements[i], NULL) == FFI_OK) + @{ + if (union_elements[i]->size > union_type.size) + @{ + union_type.size = union_elements[i]; + size = union_elements[i]->size; + @} + if (union_elements[i]->alignment > union_type.alignment) + union_type.alignment = union_elements[i]->alignment; + @} + @} +@end example + +@subsubsection Enumerations + +@code{libffi} does not have any special support for C @code{enum}s. +Although any given @code{enum} is implemented using a specific +underlying integral type, exactly which type will be used cannot be +determined by @code{libffi} -- it may depend on the values in the +enumeration or on compiler flags such as @option{-fshort-enums}. +@xref{Structures unions enumerations and bit-fields implementation, , , gcc}, +for more information about how GCC handles enumerations. @node Type Example @subsection Type Example @@ -630,30 +825,47 @@ the closure function: @findex ffi_prep_closure_loc @defun ffi_status ffi_prep_closure_loc (ffi_closure *@var{closure}, ffi_cif *@var{cif}, void (*@var{fun}) (ffi_cif *@var{cif}, void *@var{ret}, void **@var{args}, void *@var{user_data}), void *@var{user_data}, void *@var{codeloc}) -Prepare a closure function. +Prepare a closure function. The arguments to +@code{ffi_prep_closure_loc} are: -@var{closure} is the address of a @code{ffi_closure} object; this is -the writable address returned by @code{ffi_closure_alloc}. +@table @var +@item closure +The address of a @code{ffi_closure} object; this is the writable +address returned by @code{ffi_closure_alloc}. + +@item cif +The @code{ffi_cif} describing the function parameters. Note that this +object, and the types to which it refers, must be kept alive until the +closure itself is freed. -@var{cif} is the @code{ffi_cif} describing the function parameters. +@item user_data +An arbitrary datum that is passed, uninterpreted, to your closure +function. -@var{user_data} is an arbitrary datum that is passed, uninterpreted, -to your closure function. +@item codeloc +The executable address returned by @code{ffi_closure_alloc}. -@var{codeloc} is the executable address returned by -@code{ffi_closure_alloc}. +@item fun +The function which will be called when the closure is invoked. It is +called with the arguments: -@var{fun} is the function which will be called when the closure is -invoked. It is called with the arguments: @table @var @item cif The @code{ffi_cif} passed to @code{ffi_prep_closure_loc}. @item ret A pointer to the memory used for the function's return value. -@var{fun} must fill this, unless the function is declared as returning -@code{void}. -@c FIXME: is this NULL for void-returning functions? + +If the function is declared as returning @code{void}, then this value +is garbage and should not be used. + +Otherwise, @var{fun} must fill the object to which this points, +following the same special promotion behavior as @code{ffi_call}. +That is, in most cases, @var{ret} points to an object of exactly the +size of the type specified when @var{cif} was constructed. However, +integral types narrower than the system register size are widened. In +these cases your program may assume that @var{ret} points to an +@code{ffi_arg} object. @item args A vector of pointers to memory holding the arguments to the function. @@ -662,10 +874,10 @@ A vector of pointers to memory holding the arguments to the function. The same @var{user_data} that was passed to @code{ffi_prep_closure_loc}. @end table +@end table @code{ffi_prep_closure_loc} will return @code{FFI_OK} if everything -went ok, and something else on error. -@c FIXME: what? +went ok, and one of the other @code{ffi_status} values on error. After calling @code{ffi_prep_closure_loc}, you can cast @var{codeloc} to the appropriate pointer-to-function type. @@ -733,6 +945,28 @@ int main() @end example +@node Thread Safety +@section Thread Safety + +@code{libffi} is not completely thread-safe. However, many parts are, +and if you follow some simple rules, you can use it safely in a +multi-threaded program. + +@itemize @bullet +@item +@code{ffi_prep_cif} may modify the @code{ffi_type} objects passed to +it. It is best to ensure that only a single thread prepares a given +@code{ffi_cif} at a time. + +@item +On some platforms, @code{ffi_prep_cif} may modify the size and +alignment of some types, depending on the chosen ABI. On these +platforms, if you switch between ABIs, you must ensure that there is +only one call to @code{ffi_prep_cif} at a time. + +Currently the only affected platform is PowerPC and the only affected +type is @code{long double}. +@end itemize @node Missing Features @chapter Missing Features @@ -747,16 +981,12 @@ Variadic closures. @item There is no support for bit fields in structures. -@item -The closure API is - -@c FIXME: ... - @item The ``raw'' API is undocumented. -@c argument promotion? -@c unions? @c anything else? + +@item +The Go API is undocumented. @end itemize Note that variadic support is very new and tested on a relatively diff --git a/native/libffi/doc/version.texi b/native/libffi/doc/version.texi index ccef70f493..92950d92a8 100644 --- a/native/libffi/doc/version.texi +++ b/native/libffi/doc/version.texi @@ -1,4 +1,11 @@ +<<<<<<< HEAD @set UPDATED 8 November 2014 @set UPDATED-MONTH November 2014 @set EDITION 3.2.1 @set VERSION 3.2.1 +======= +@set UPDATED 22 November 2019 +@set UPDATED-MONTH November 2019 +@set EDITION 3.3 +@set VERSION 3.3 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 diff --git a/native/libffi/generate-darwin-source-and-headers.py b/native/libffi/generate-darwin-source-and-headers.py index 964e861d43..00e9bf0ec3 100644 --- a/native/libffi/generate-darwin-source-and-headers.py +++ b/native/libffi/generate-darwin-source-and-headers.py @@ -14,12 +14,20 @@ class simulator_platform(Platform): sdk = 'iphonesimulator' arch = 'i386' triple = 'i386-apple-darwin11' +<<<<<<< HEAD version_min = '-miphoneos-version-min=5.1.1' +======= + version_min = '-miphoneos-version-min=7.0' +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 prefix = "#ifdef __i386__\n\n" suffix = "\n\n#endif" src_dir = 'x86' +<<<<<<< HEAD src_files = ['darwin.S', 'win32.S', 'ffi.c'] +======= + src_files = ['sysv.S', 'ffi.c', 'internal.h'] +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 class simulator64_platform(Platform): @@ -32,7 +40,11 @@ class simulator64_platform(Platform): prefix = "#ifdef __x86_64__\n\n" suffix = "\n\n#endif" src_dir = 'x86' +<<<<<<< HEAD src_files = ['darwin64.S', 'ffi64.c'] +======= + src_files = ['unix64.S', 'ffi64.c', 'ffiw64.c', 'win64.S', 'internal64.h', 'asmnames.h'] +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 class device_platform(Platform): @@ -40,12 +52,20 @@ class device_platform(Platform): sdk = 'iphoneos' arch = 'armv7' triple = 'arm-apple-darwin11' +<<<<<<< HEAD version_min = '-miphoneos-version-min=5.1.1' +======= + version_min = '-miphoneos-version-min=7.0' +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 prefix = "#ifdef __arm__\n\n" suffix = "\n\n#endif" src_dir = 'arm' +<<<<<<< HEAD src_files = ['sysv.S', 'trampoline.S', 'ffi.c'] +======= + src_files = ['sysv.S', 'ffi.c', 'internal.h'] +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 class device64_platform(Platform): @@ -58,7 +78,11 @@ class device64_platform(Platform): prefix = "#ifdef __arm64__\n\n" suffix = "\n\n#endif" src_dir = 'aarch64' +<<<<<<< HEAD src_files = ['sysv.S', 'ffi.c'] +======= + src_files = ['sysv.S', 'ffi.c', 'internal.h'] +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 class desktop32_platform(Platform): @@ -68,7 +92,11 @@ class desktop32_platform(Platform): triple = 'i386-apple-darwin10' version_min = '-mmacosx-version-min=10.6' src_dir = 'x86' +<<<<<<< HEAD src_files = ['darwin.S', 'win32.S', 'ffi.c'] +======= + src_files = ['sysv.S', 'ffi.c', 'internal.h'] +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 prefix = "#ifdef __i386__\n\n" suffix = "\n\n#endif" @@ -84,16 +112,24 @@ class desktop64_platform(Platform): prefix = "#ifdef __x86_64__\n\n" suffix = "\n\n#endif" src_dir = 'x86' +<<<<<<< HEAD src_files = ['darwin64.S', 'ffi64.c'] +======= + src_files = ['unix64.S', 'ffi64.c', 'ffiw64.c', 'win64.S', 'internal64.h', 'asmnames.h'] +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 def mkdir_p(path): try: os.makedirs(path) except OSError as exc: # Python >2.5 +<<<<<<< HEAD if exc.errno == errno.EEXIST: pass else: +======= + if exc.errno != errno.EEXIST: +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 raise @@ -102,8 +138,16 @@ def move_file(src_dir, dst_dir, filename, file_suffix=None, prefix='', suffix='' out_filename = filename if file_suffix: +<<<<<<< HEAD split_name = os.path.splitext(filename) out_filename = "%s_%s%s" % (split_name[0], file_suffix, split_name[1]) +======= + if filename in ['internal64.h', 'asmnames.h', 'internal.h']: + out_filename = filename + else: + split_name = os.path.splitext(filename) + out_filename = "%s_%s%s" % (split_name[0], file_suffix, split_name[1]) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 with open(os.path.join(src_dir, filename)) as in_file: with open(os.path.join(dst_dir, out_filename), 'w') as out_file: @@ -142,7 +186,11 @@ def xcrun_cmd(cmd): mkdir_p(build_dir) env = dict(CC=xcrun_cmd('clang'), LD=xcrun_cmd('ld'), +<<<<<<< HEAD CFLAGS='%s' % (platform.version_min)) +======= + CFLAGS='%s -fembed-bitcode' % (platform.version_min)) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 working_dir = os.getcwd() try: os.chdir(build_dir) @@ -162,24 +210,33 @@ def xcrun_cmd(cmd): platform_headers[filename].add((platform.prefix, platform.arch, platform.suffix)) +<<<<<<< HEAD def make_tramp(): with open('src/arm/trampoline.S', 'w') as tramp_out: p = subprocess.Popen(['bash', 'src/arm/gentramp.sh'], stdout=tramp_out) p.wait() +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 def generate_source_and_headers(generate_osx=True, generate_ios=True): copy_files('src', 'darwin_common/src', pattern='*.c') copy_files('include', 'darwin_common/include', pattern='*.h') if generate_ios: +<<<<<<< HEAD make_tramp() +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 copy_src_platform_files(simulator_platform) copy_src_platform_files(simulator64_platform) copy_src_platform_files(device_platform) copy_src_platform_files(device64_platform) if generate_osx: +<<<<<<< HEAD copy_src_platform_files(desktop32_platform) +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 copy_src_platform_files(desktop64_platform) platform_headers = collections.defaultdict(set) @@ -190,11 +247,18 @@ def generate_source_and_headers(generate_osx=True, generate_ios=True): build_target(device_platform, platform_headers) build_target(device64_platform, platform_headers) if generate_osx: +<<<<<<< HEAD build_target(desktop32_platform, platform_headers) build_target(desktop64_platform, platform_headers) mkdir_p('darwin_common/include') for header_name, tag_tuples in platform_headers.iteritems(): +======= + build_target(desktop64_platform, platform_headers) + + mkdir_p('darwin_common/include') + for header_name, tag_tuples in platform_headers.items(): +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 basename, suffix = os.path.splitext(header_name) with open(os.path.join('darwin_common/include', header_name), 'w') as header: for tag_tuple in tag_tuples: diff --git a/native/libffi/include/Makefile.am b/native/libffi/include/Makefile.am index fd28024973..c59df9fb3c 100644 --- a/native/libffi/include/Makefile.am +++ b/native/libffi/include/Makefile.am @@ -3,7 +3,7 @@ AUTOMAKE_OPTIONS=foreign DISTCLEANFILES=ffitarget.h -EXTRA_DIST=ffi.h.in ffi_common.h +noinst_HEADERS=ffi_common.h ffi_cfi.h +EXTRA_DIST=ffi.h.in -includesdir = $(libdir)/@PACKAGE_NAME@-@PACKAGE_VERSION@/include -nodist_includes_HEADERS = ffi.h ffitarget.h +nodist_include_HEADERS = ffi.h ffitarget.h diff --git a/native/libffi/include/ffi.h.in b/native/libffi/include/ffi.h.in index f403ae06dd..0960c1a6e7 100644 --- a/native/libffi/include/ffi.h.in +++ b/native/libffi/include/ffi.h.in @@ -1,5 +1,9 @@ /* -----------------------------------------------------------------*-C-*- +<<<<<<< HEAD libffi @VERSION@ - Copyright (c) 2011, 2014 Anthony Green +======= + libffi @VERSION@ - Copyright (c) 2011, 2014, 2019 Anthony Green +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 - Copyright (c) 1996-2003, 2007, 2008 Red Hat, Inc. Permission is hereby granted, free of charge, to any person @@ -25,23 +29,14 @@ ----------------------------------------------------------------------- */ /* ------------------------------------------------------------------- - The basic API is described in the README file. + Most of the API is documented in doc/libffi.texi. - The raw API is designed to bypass some of the argument packing - and unpacking on architectures for which it can be avoided. + The raw API is designed to bypass some of the argument packing and + unpacking on architectures for which it can be avoided. Routines + are provided to emulate the raw API if the underlying platform + doesn't allow faster implementation. - The closure API allows interpreted functions to be packaged up - inside a C function pointer, so that they can be called as C functions, - with no understanding on the client side that they are interpreted. - It can also be used in other cases in which it is necessary to package - up a user specified parameter and a function pointer as a single - function pointer. - - The closure API must be implemented in order to get its functionality, - e.g. for use by gij. Routines are provided to emulate the raw API - if the underlying platform doesn't allow faster implementation. - - More details on the raw and cloure API can be found in: + More details on the raw API can be found in: http://gcc.gnu.org/ml/java/1999-q3/msg00138.html @@ -106,8 +101,8 @@ extern "C" { # endif #endif -/* The closure code assumes that this works on pointers, i.e. a size_t */ -/* can hold a pointer. */ +/* The closure code assumes that this works on pointers, i.e. a size_t + can hold a pointer. */ typedef struct _ffi_type { @@ -117,6 +112,32 @@ typedef struct _ffi_type struct _ffi_type **elements; } ffi_type; +/* Need minimal decorations for DLLs to work on Windows. GCC has + autoimport and autoexport. Always mark externally visible symbols + as dllimport for MSVC clients, even if it means an extra indirection + when using the static version of the library. + Besides, as a workaround, they can define FFI_BUILDING if they + *know* they are going to link with the static library. */ +#if defined _MSC_VER +# if defined FFI_BUILDING_DLL /* Building libffi.DLL with msvcc.sh */ +# define FFI_API __declspec(dllexport) +# elif !defined FFI_BUILDING /* Importing libffi.DLL */ +# define FFI_API __declspec(dllimport) +# else /* Building/linking static library */ +# define FFI_API +# endif +#else +# define FFI_API +#endif + +/* The externally visible type declarations also need the MSVC DLL + decorations, or they will not be exported from the object file. */ +#if defined LIBFFI_HIDE_BASIC_TYPES +# define FFI_EXTERN FFI_API +#else +# define FFI_EXTERN extern FFI_API +#endif + #ifndef LIBFFI_HIDE_BASIC_TYPES #if SCHAR_MAX == 127 # define ffi_type_uchar ffi_type_uint8 @@ -166,21 +187,7 @@ typedef struct _ffi_type #error "long size not supported" #endif -/* Need minimal decorations for DLLs to works on Windows. */ -/* GCC has autoimport and autoexport. Rely on Libtool to */ -/* help MSVC export from a DLL, but always declare data */ -/* to be imported for MSVC clients. This costs an extra */ -/* indirection for MSVC clients using the static version */ -/* of the library, but don't worry about that. Besides, */ -/* as a workaround, they can define FFI_BUILDING if they */ -/* *know* they are going to link with the static library. */ -#if defined _MSC_VER && !defined FFI_BUILDING -#define FFI_EXTERN extern __declspec(dllimport) -#else -#define FFI_EXTERN extern -#endif - -/* These are defined in types.c */ +/* These are defined in types.c. */ FFI_EXTERN ffi_type ffi_type_void; FFI_EXTERN ffi_type ffi_type_uint8; FFI_EXTERN ffi_type ffi_type_sint8; @@ -217,8 +224,6 @@ typedef enum { FFI_BAD_ABI } ffi_status; -typedef unsigned FFI_TYPE; - typedef struct { ffi_abi abi; unsigned nargs; @@ -231,6 +236,7 @@ typedef struct { #endif } ffi_cif; +<<<<<<< HEAD #if @HAVE_LONG_DOUBLE_VARIANT@ /* Used to adjust size/alignment of ffi types. */ void ffi_prep_types (ffi_abi abi); @@ -245,6 +251,8 @@ ffi_status ffi_prep_cif_core(ffi_cif *cif, ffi_type *rtype, ffi_type **atypes); +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* ---- Definitions for the raw API -------------------------------------- */ #ifndef FFI_SIZEOF_ARG @@ -282,27 +290,34 @@ typedef ffi_raw ffi_java_raw; #endif +FFI_API void ffi_raw_call (ffi_cif *cif, void (*fn)(void), void *rvalue, ffi_raw *avalue); -void ffi_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_raw *raw); -void ffi_raw_to_ptrarray (ffi_cif *cif, ffi_raw *raw, void **args); -size_t ffi_raw_size (ffi_cif *cif); +FFI_API void ffi_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_raw *raw); +FFI_API void ffi_raw_to_ptrarray (ffi_cif *cif, ffi_raw *raw, void **args); +FFI_API size_t ffi_raw_size (ffi_cif *cif); -/* This is analogous to the raw API, except it uses Java parameter */ -/* packing, even on 64-bit machines. I.e. on 64-bit machines */ -/* longs and doubles are followed by an empty 64-bit word. */ +/* This is analogous to the raw API, except it uses Java parameter + packing, even on 64-bit machines. I.e. on 64-bit machines longs + and doubles are followed by an empty 64-bit word. */ +#if !FFI_NATIVE_RAW_API +FFI_API void ffi_java_raw_call (ffi_cif *cif, void (*fn)(void), void *rvalue, - ffi_java_raw *avalue); + ffi_java_raw *avalue) __attribute__((deprecated)); +#endif -void ffi_java_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_java_raw *raw); -void ffi_java_raw_to_ptrarray (ffi_cif *cif, ffi_java_raw *raw, void **args); -size_t ffi_java_raw_size (ffi_cif *cif); +FFI_API +void ffi_java_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_java_raw *raw) __attribute__((deprecated)); +FFI_API +void ffi_java_raw_to_ptrarray (ffi_cif *cif, ffi_java_raw *raw, void **args) __attribute__((deprecated)); +FFI_API +size_t ffi_java_raw_size (ffi_cif *cif) __attribute__((deprecated)); /* ---- Definitions for closures ----------------------------------------- */ @@ -321,25 +336,42 @@ typedef struct { ffi_cif *cif; void (*fun)(ffi_cif*,void*,void**,void*); void *user_data; +} ffi_closure #ifdef __GNUC__ -} ffi_closure __attribute__((aligned (8))); -#else -} ffi_closure; + __attribute__((aligned (8))) +#endif + ; + +#ifndef __GNUC__ # ifdef __sgi # pragma pack 0 # endif #endif -void *ffi_closure_alloc (size_t size, void **code); -void ffi_closure_free (void *); +FFI_API void *ffi_closure_alloc (size_t size, void **code); +FFI_API void ffi_closure_free (void *); + +#if defined(PA_LINUX) || defined(PA_HPUX) +#define FFI_CLOSURE_PTR(X) ((void *)((unsigned int)(X) | 2)) +#define FFI_RESTORE_PTR(X) ((void *)((unsigned int)(X) & ~3)) +#else +#define FFI_CLOSURE_PTR(X) (X) +#define FFI_RESTORE_PTR(X) (X) +#endif -ffi_status +FFI_API ffi_status ffi_prep_closure (ffi_closure*, ffi_cif *, void (*fun)(ffi_cif*,void*,void**,void*), - void *user_data); + void *user_data) +#if defined(__GNUC__) && (((__GNUC__ * 100) + __GNUC_MINOR__) >= 405) + __attribute__((deprecated ("use ffi_prep_closure_loc instead"))) +#elif defined(__GNUC__) && __GNUC__ >= 3 + __attribute__((deprecated)) +#endif + ; -ffi_status +FFI_API ffi_status ffi_prep_closure_loc (ffi_closure*, ffi_cif *, void (*fun)(ffi_cif*,void*,void**,void*), @@ -360,9 +392,9 @@ typedef struct { #if !FFI_NATIVE_RAW_API - /* if this is enabled, then a raw closure has the same layout + /* If this is enabled, then a raw closure has the same layout as a regular closure. We use this to install an intermediate - handler to do the transaltion, void** -> ffi_raw*. */ + handler to do the transaltion, void** -> ffi_raw*. */ void (*translate_args)(ffi_cif*,void*,void**,void*); void *this_closure; @@ -386,9 +418,9 @@ typedef struct { #if !FFI_NATIVE_RAW_API - /* if this is enabled, then a raw closure has the same layout + /* If this is enabled, then a raw closure has the same layout as a regular closure. We use this to install an intermediate - handler to do the transaltion, void** -> ffi_raw*. */ + handler to do the translation, void** -> ffi_raw*. */ void (*translate_args)(ffi_cif*,void*,void**,void*); void *this_closure; @@ -400,42 +432,62 @@ typedef struct { } ffi_java_raw_closure; -ffi_status +FFI_API ffi_status ffi_prep_raw_closure (ffi_raw_closure*, ffi_cif *cif, void (*fun)(ffi_cif*,void*,ffi_raw*,void*), void *user_data); -ffi_status +FFI_API ffi_status ffi_prep_raw_closure_loc (ffi_raw_closure*, ffi_cif *cif, void (*fun)(ffi_cif*,void*,ffi_raw*,void*), void *user_data, void *codeloc); -ffi_status +#if !FFI_NATIVE_RAW_API +FFI_API ffi_status ffi_prep_java_raw_closure (ffi_java_raw_closure*, ffi_cif *cif, void (*fun)(ffi_cif*,void*,ffi_java_raw*,void*), - void *user_data); + void *user_data) __attribute__((deprecated)); -ffi_status +FFI_API ffi_status ffi_prep_java_raw_closure_loc (ffi_java_raw_closure*, ffi_cif *cif, void (*fun)(ffi_cif*,void*,ffi_java_raw*,void*), void *user_data, - void *codeloc); + void *codeloc) __attribute__((deprecated)); +#endif #endif /* FFI_CLOSURES */ +#if FFI_GO_CLOSURES + +typedef struct { + void *tramp; + ffi_cif *cif; + void (*fun)(ffi_cif*,void*,void**,void*); +} ffi_go_closure; + +FFI_API ffi_status ffi_prep_go_closure (ffi_go_closure*, ffi_cif *, + void (*fun)(ffi_cif*,void*,void**,void*)); + +FFI_API void ffi_call_go (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure); + +#endif /* FFI_GO_CLOSURES */ + /* ---- Public interface definition -------------------------------------- */ +FFI_API ffi_status ffi_prep_cif(ffi_cif *cif, ffi_abi abi, unsigned int nargs, ffi_type *rtype, ffi_type **atypes); +FFI_API ffi_status ffi_prep_cif_var(ffi_cif *cif, ffi_abi abi, unsigned int nfixedargs, @@ -443,12 +495,17 @@ ffi_status ffi_prep_cif_var(ffi_cif *cif, ffi_type *rtype, ffi_type **atypes); +FFI_API void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue); -/* Useful for eliminating compiler warnings */ +FFI_API +ffi_status ffi_get_struct_offsets (ffi_abi abi, ffi_type *struct_type, + size_t *offsets); + +/* Useful for eliminating compiler warnings. */ #define FFI_FN(f) ((void (*)(void))f) /* ---- Definitions shared with assembly code ---------------------------- */ @@ -477,7 +534,11 @@ void ffi_call(ffi_cif *cif, #define FFI_TYPE_POINTER 14 #define FFI_TYPE_COMPLEX 15 +<<<<<<< HEAD /* This should always refer to the last type code (for sanity checks) */ +======= +/* This should always refer to the last type code (for sanity checks). */ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #define FFI_TYPE_LAST FFI_TYPE_COMPLEX #ifdef __cplusplus diff --git a/native/libffi/include/ffi_cfi.h b/native/libffi/include/ffi_cfi.h new file mode 100644 index 0000000000..244ce572be --- /dev/null +++ b/native/libffi/include/ffi_cfi.h @@ -0,0 +1,55 @@ +/* ----------------------------------------------------------------------- + ffi_cfi.h - Copyright (c) 2014 Red Hat, Inc. + + Conditionally assemble cfi directives. Only necessary for building libffi. + ----------------------------------------------------------------------- */ + +#ifndef FFI_CFI_H +#define FFI_CFI_H + +#ifdef HAVE_AS_CFI_PSEUDO_OP + +# define cfi_startproc .cfi_startproc +# define cfi_endproc .cfi_endproc +# define cfi_def_cfa(reg, off) .cfi_def_cfa reg, off +# define cfi_def_cfa_register(reg) .cfi_def_cfa_register reg +# define cfi_def_cfa_offset(off) .cfi_def_cfa_offset off +# define cfi_adjust_cfa_offset(off) .cfi_adjust_cfa_offset off +# define cfi_offset(reg, off) .cfi_offset reg, off +# define cfi_rel_offset(reg, off) .cfi_rel_offset reg, off +# define cfi_register(r1, r2) .cfi_register r1, r2 +# define cfi_return_column(reg) .cfi_return_column reg +# define cfi_restore(reg) .cfi_restore reg +# define cfi_same_value(reg) .cfi_same_value reg +# define cfi_undefined(reg) .cfi_undefined reg +# define cfi_remember_state .cfi_remember_state +# define cfi_restore_state .cfi_restore_state +# define cfi_window_save .cfi_window_save +# define cfi_personality(enc, exp) .cfi_personality enc, exp +# define cfi_lsda(enc, exp) .cfi_lsda enc, exp +# define cfi_escape(...) .cfi_escape __VA_ARGS__ + +#else + +# define cfi_startproc +# define cfi_endproc +# define cfi_def_cfa(reg, off) +# define cfi_def_cfa_register(reg) +# define cfi_def_cfa_offset(off) +# define cfi_adjust_cfa_offset(off) +# define cfi_offset(reg, off) +# define cfi_rel_offset(reg, off) +# define cfi_register(r1, r2) +# define cfi_return_column(reg) +# define cfi_restore(reg) +# define cfi_same_value(reg) +# define cfi_undefined(reg) +# define cfi_remember_state +# define cfi_restore_state +# define cfi_window_save +# define cfi_personality(enc, exp) +# define cfi_lsda(enc, exp) +# define cfi_escape(...) + +#endif /* HAVE_AS_CFI_PSEUDO_OP */ +#endif /* FFI_CFI_H */ diff --git a/native/libffi/include/ffi_common.h b/native/libffi/include/ffi_common.h index 37f5a9e924..76b9dd6faf 100644 --- a/native/libffi/include/ffi_common.h +++ b/native/libffi/include/ffi_common.h @@ -74,14 +74,35 @@ void ffi_type_test(ffi_type *a, char *file, int line); #define FFI_ASSERT_VALID_TYPE(x) #endif -#define ALIGN(v, a) (((((size_t) (v))-1) | ((a)-1))+1) -#define ALIGN_DOWN(v, a) (((size_t) (v)) & -a) +/* v cast to size_t and aligned up to a multiple of a */ +#define FFI_ALIGN(v, a) (((((size_t) (v))-1) | ((a)-1))+1) +/* v cast to size_t and aligned down to a multiple of a */ +#define FFI_ALIGN_DOWN(v, a) (((size_t) (v)) & -a) /* Perform machine dependent cif processing */ ffi_status ffi_prep_cif_machdep(ffi_cif *cif); ffi_status ffi_prep_cif_machdep_var(ffi_cif *cif, unsigned int nfixedargs, unsigned int ntotalargs); + +#if HAVE_LONG_DOUBLE_VARIANT +/* Used to adjust size/alignment of ffi types. */ +void ffi_prep_types (ffi_abi abi); +#endif + +/* Used internally, but overridden by some architectures */ +ffi_status ffi_prep_cif_core(ffi_cif *cif, + ffi_abi abi, + unsigned int isvariadic, + unsigned int nfixedargs, + unsigned int ntotalargs, + ffi_type *rtype, + ffi_type **atypes); + +/* Translate a data pointer to a code pointer. Needed for closures on + some targets. */ +void *ffi_data_to_code_pointer (void *data) FFI_HIDDEN; + /* Extended cif, used in callback from assembly routine */ typedef struct { diff --git a/native/libffi/libffi.map.in b/native/libffi/libffi.map.in new file mode 100644 index 0000000000..de8778ae40 --- /dev/null +++ b/native/libffi/libffi.map.in @@ -0,0 +1,76 @@ +#define LIBFFI_ASM +#define LIBFFI_H +#include +#include + +/* These version numbers correspond to the libtool-version abi numbers, + not to the libffi release numbers. */ + +LIBFFI_BASE_8.0 { + global: + /* Exported data variables. */ + ffi_type_void; + ffi_type_uint8; + ffi_type_sint8; + ffi_type_uint16; + ffi_type_sint16; + ffi_type_uint32; + ffi_type_sint32; + ffi_type_uint64; + ffi_type_sint64; + ffi_type_float; + ffi_type_double; + ffi_type_longdouble; + ffi_type_pointer; + + /* Exported functions. */ + ffi_call; + ffi_prep_cif; + ffi_prep_cif_var; + + ffi_raw_call; + ffi_ptrarray_to_raw; + ffi_raw_to_ptrarray; + ffi_raw_size; + + ffi_java_raw_call; + ffi_java_ptrarray_to_raw; + ffi_java_raw_to_ptrarray; + ffi_java_raw_size; + + ffi_get_struct_offsets; + local: + *; +}; + +#ifdef FFI_TARGET_HAS_COMPLEX_TYPE +LIBFFI_COMPLEX_8.0 { + global: + /* Exported data variables. */ + ffi_type_complex_float; + ffi_type_complex_double; + ffi_type_complex_longdouble; +} LIBFFI_BASE_8.0; +#endif + +#if FFI_CLOSURES +LIBFFI_CLOSURE_8.0 { + global: + ffi_closure_alloc; + ffi_closure_free; + ffi_prep_closure; + ffi_prep_closure_loc; + ffi_prep_raw_closure; + ffi_prep_raw_closure_loc; + ffi_prep_java_raw_closure; + ffi_prep_java_raw_closure_loc; +} LIBFFI_BASE_8.0; +#endif + +#if FFI_GO_CLOSURES +LIBFFI_GO_CLOSURE_8.0 { + global: + ffi_call_go; + ffi_prep_go_closure; +} LIBFFI_CLOSURE_8.0; +#endif diff --git a/native/libffi/libffi.pc.in b/native/libffi/libffi.pc.in index edf6fde5e2..264877ea71 100644 --- a/native/libffi/libffi.pc.in +++ b/native/libffi/libffi.pc.in @@ -2,7 +2,11 @@ prefix=@prefix@ exec_prefix=@exec_prefix@ libdir=@libdir@ toolexeclibdir=@toolexeclibdir@ +<<<<<<< HEAD includedir=${libdir}/@PACKAGE_NAME@-@PACKAGE_VERSION@/include +======= +includedir=@includedir@ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 Name: @PACKAGE_NAME@ Description: Library supporting Foreign Function Interfaces diff --git a/native/libffi/libffi.xcodeproj/project.pbxproj b/native/libffi/libffi.xcodeproj/project.pbxproj index 1cf396ffa1..cc49b3d7ed 100644 --- a/native/libffi/libffi.xcodeproj/project.pbxproj +++ b/native/libffi/libffi.xcodeproj/project.pbxproj @@ -7,6 +7,12 @@ objects = { /* Begin PBXBuildFile section */ +<<<<<<< HEAD +======= + 43B5D3F81D35473200D1E1FD /* ffiw64_x86_64.c in Sources */ = {isa = PBXBuildFile; fileRef = 43B5D3F71D35473200D1E1FD /* ffiw64_x86_64.c */; }; + 43B5D3FA1D3547CE00D1E1FD /* win64_x86_64.S in Sources */ = {isa = PBXBuildFile; fileRef = 43B5D3F91D3547CE00D1E1FD /* win64_x86_64.S */; }; + 43E9A5C81D352C1500926A8F /* unix64_x86_64.S in Sources */ = {isa = PBXBuildFile; fileRef = 43E9A5C61D352C1500926A8F /* unix64_x86_64.S */; }; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 DBFA714A187F1D8600A76262 /* ffi.h in Headers */ = {isa = PBXBuildFile; fileRef = DBFA713E187F1D8600A76262 /* ffi.h */; }; DBFA714B187F1D8600A76262 /* ffi_common.h in Headers */ = {isa = PBXBuildFile; fileRef = DBFA713F187F1D8600A76262 /* ffi_common.h */; }; DBFA714C187F1D8600A76262 /* fficonfig.h in Headers */ = {isa = PBXBuildFile; fileRef = DBFA7140187F1D8600A76262 /* fficonfig.h */; }; @@ -23,6 +29,7 @@ DBFA7178187F1D9B00A76262 /* sysv_arm64.S in Sources */ = {isa = PBXBuildFile; fileRef = DBFA716D187F1D9B00A76262 /* sysv_arm64.S */; }; DBFA7179187F1D9B00A76262 /* ffi_armv7.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA716F187F1D9B00A76262 /* ffi_armv7.c */; }; DBFA717A187F1D9B00A76262 /* sysv_armv7.S in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7170187F1D9B00A76262 /* sysv_armv7.S */; }; +<<<<<<< HEAD DBFA717B187F1D9B00A76262 /* trampoline_armv7.S in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7171187F1D9B00A76262 /* trampoline_armv7.S */; }; DBFA717C187F1D9B00A76262 /* darwin64_x86_64.S in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7173187F1D9B00A76262 /* darwin64_x86_64.S */; }; DBFA717D187F1D9B00A76262 /* darwin_i386.S in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7174187F1D9B00A76262 /* darwin_i386.S */; }; @@ -38,11 +45,60 @@ DBFA7195187F1DA100A76262 /* darwin_i386.S in Sources */ = {isa = PBXBuildFile; fileRef = DBFA718B187F1DA100A76262 /* darwin_i386.S */; }; DBFA7196187F1DA100A76262 /* ffi64_x86_64.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA718C187F1DA100A76262 /* ffi64_x86_64.c */; }; DBFA7197187F1DA100A76262 /* ffi_i386.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA718D187F1DA100A76262 /* ffi_i386.c */; }; +======= + DBFA717E187F1D9B00A76262 /* ffi64_x86_64.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7175187F1D9B00A76262 /* ffi64_x86_64.c */; }; + DBFA718F187F1DA100A76262 /* ffi_x86_64.h in Headers */ = {isa = PBXBuildFile; fileRef = DBFA7183187F1DA100A76262 /* ffi_x86_64.h */; }; + DBFA7191187F1DA100A76262 /* fficonfig_x86_64.h in Headers */ = {isa = PBXBuildFile; fileRef = DBFA7185187F1DA100A76262 /* fficonfig_x86_64.h */; }; + DBFA7193187F1DA100A76262 /* ffitarget_x86_64.h in Headers */ = {isa = PBXBuildFile; fileRef = DBFA7187187F1DA100A76262 /* ffitarget_x86_64.h */; }; + DBFA7194187F1DA100A76262 /* unix64_x86_64.S in Sources */ = {isa = PBXBuildFile; fileRef = DBFA718A187F1DA100A76262 /* unix64_x86_64.S */; }; + DBFA7196187F1DA100A76262 /* ffi64_x86_64.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA718C187F1DA100A76262 /* ffi64_x86_64.c */; }; + FDB52FB31F6144FA00AA92E6 /* unix64_x86_64.S in Sources */ = {isa = PBXBuildFile; fileRef = 43E9A5C61D352C1500926A8F /* unix64_x86_64.S */; }; + FDB52FB51F6144FA00AA92E6 /* ffi64_x86_64.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7175187F1D9B00A76262 /* ffi64_x86_64.c */; }; + FDB52FB61F6144FA00AA92E6 /* ffi_armv7.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA716F187F1D9B00A76262 /* ffi_armv7.c */; }; + FDB52FB71F6144FA00AA92E6 /* closures.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7143187F1D8600A76262 /* closures.c */; }; + FDB52FB81F6144FA00AA92E6 /* sysv_armv7.S in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7170187F1D9B00A76262 /* sysv_armv7.S */; }; + FDB52FB91F6144FA00AA92E6 /* ffiw64_x86_64.c in Sources */ = {isa = PBXBuildFile; fileRef = 43B5D3F71D35473200D1E1FD /* ffiw64_x86_64.c */; }; + FDB52FBA1F6144FA00AA92E6 /* prep_cif.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7147187F1D8600A76262 /* prep_cif.c */; }; + FDB52FBC1F6144FA00AA92E6 /* raw_api.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7148187F1D8600A76262 /* raw_api.c */; }; + FDB52FBD1F6144FA00AA92E6 /* sysv_arm64.S in Sources */ = {isa = PBXBuildFile; fileRef = DBFA716D187F1D9B00A76262 /* sysv_arm64.S */; }; + FDB52FBE1F6144FA00AA92E6 /* types.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7149187F1D8600A76262 /* types.c */; }; + FDB52FBF1F6144FA00AA92E6 /* ffi_arm64.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA716C187F1D9B00A76262 /* ffi_arm64.c */; }; + FDB52FC01F6144FA00AA92E6 /* win64_x86_64.S in Sources */ = {isa = PBXBuildFile; fileRef = 43B5D3F91D3547CE00D1E1FD /* win64_x86_64.S */; }; + FDB52FD01F614A8B00AA92E6 /* ffi.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA713E187F1D8600A76262 /* ffi.h */; }; + FDB52FD11F614AA700AA92E6 /* ffi_arm64.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA715E187F1D9B00A76262 /* ffi_arm64.h */; }; + FDB52FD21F614AAB00AA92E6 /* ffi_armv7.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA715F187F1D9B00A76262 /* ffi_armv7.h */; }; + FDB52FD41F614AB500AA92E6 /* ffi_x86_64.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7161187F1D9B00A76262 /* ffi_x86_64.h */; }; + FDB52FD51F614AE200AA92E6 /* ffi.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA713E187F1D8600A76262 /* ffi.h */; }; + FDB52FD61F614AEA00AA92E6 /* ffi_arm64.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA715E187F1D9B00A76262 /* ffi_arm64.h */; }; + FDB52FD71F614AED00AA92E6 /* ffi_x86_64.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7161187F1D9B00A76262 /* ffi_x86_64.h */; }; + FDB52FD81F614B8700AA92E6 /* ffitarget.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7141187F1D8600A76262 /* ffitarget.h */; }; + FDB52FD91F614B8E00AA92E6 /* ffitarget_arm64.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7166187F1D9B00A76262 /* ffitarget_arm64.h */; }; + FDB52FDA1F614B9300AA92E6 /* ffitarget_armv7.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7167187F1D9B00A76262 /* ffitarget_armv7.h */; }; + FDB52FDD1F614BA900AA92E6 /* ffitarget_x86_64.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7169187F1D9B00A76262 /* ffitarget_x86_64.h */; }; + FDB52FDE1F6155E300AA92E6 /* ffitarget.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7141187F1D8600A76262 /* ffitarget.h */; }; + FDB52FDF1F6155EA00AA92E6 /* ffitarget_arm64.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7166187F1D9B00A76262 /* ffitarget_arm64.h */; }; + FDB52FE01F6155EF00AA92E6 /* ffitarget_x86_64.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7169187F1D9B00A76262 /* ffitarget_x86_64.h */; }; + FDB52FE21F6156FA00AA92E6 /* ffi.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA713E187F1D8600A76262 /* ffi.h */; }; + FDB52FE31F61571A00AA92E6 /* ffi_x86_64.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7183187F1DA100A76262 /* ffi_x86_64.h */; }; + FDB52FE41F61571D00AA92E6 /* ffitarget.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7141187F1D8600A76262 /* ffitarget.h */; }; + FDB52FE61F61573100AA92E6 /* ffitarget_x86_64.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DBFA7187187F1DA100A76262 /* ffitarget_x86_64.h */; }; + FDDB2F411F5D66E200EF414E /* ffiw64_x86_64.c in Sources */ = {isa = PBXBuildFile; fileRef = FDDB2F3F1F5D666900EF414E /* ffiw64_x86_64.c */; }; + FDDB2F461F5D691E00EF414E /* win64_x86_64.S in Sources */ = {isa = PBXBuildFile; fileRef = FDDB2F441F5D68C900EF414E /* win64_x86_64.S */; }; + FDDB2F4A1F5D846400EF414E /* ffi64_x86_64.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA718C187F1DA100A76262 /* ffi64_x86_64.c */; }; + FDDB2F4C1F5D846400EF414E /* prep_cif.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7147187F1D8600A76262 /* prep_cif.c */; }; + FDDB2F4E1F5D846400EF414E /* ffiw64_x86_64.c in Sources */ = {isa = PBXBuildFile; fileRef = FDDB2F3F1F5D666900EF414E /* ffiw64_x86_64.c */; }; + FDDB2F4F1F5D846400EF414E /* types.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7149187F1D8600A76262 /* types.c */; }; + FDDB2F501F5D846400EF414E /* raw_api.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7148187F1D8600A76262 /* raw_api.c */; }; + FDDB2F511F5D846400EF414E /* closures.c in Sources */ = {isa = PBXBuildFile; fileRef = DBFA7143187F1D8600A76262 /* closures.c */; }; + FDDB2F521F5D846400EF414E /* unix64_x86_64.S in Sources */ = {isa = PBXBuildFile; fileRef = DBFA718A187F1DA100A76262 /* unix64_x86_64.S */; }; + FDDB2F531F5D846400EF414E /* win64_x86_64.S in Sources */ = {isa = PBXBuildFile; fileRef = FDDB2F441F5D68C900EF414E /* win64_x86_64.S */; }; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* End PBXBuildFile section */ /* Begin PBXCopyFilesBuildPhase section */ DB13B1641849DF1E0010F42D /* CopyFiles */ = { isa = PBXCopyFilesBuildPhase; +<<<<<<< HEAD buildActionMask = 8; dstPath = "include/$(PRODUCT_NAME)"; dstSubfolderSpec = 16; @@ -105,18 +161,61 @@ DBFA715C187F1D9B00A76262 /* darwin_ios */, DBFA7180187F1DA100A76262 /* darwin_osx */, DB13B1671849DF1E0010F42D /* Products */, +======= + buildActionMask = 12; + dstPath = "include/$(PRODUCT_NAME)"; + dstSubfolderSpec = 16; + files = ( + FDB52FD01F614A8B00AA92E6 /* ffi.h in CopyFiles */, + FDB52FD11F614AA700AA92E6 /* ffi_arm64.h in CopyFiles */, + FDB52FD21F614AAB00AA92E6 /* ffi_armv7.h in CopyFiles */, + FDB52FD41F614AB500AA92E6 /* ffi_x86_64.h in CopyFiles */, + FDB52FD81F614B8700AA92E6 /* ffitarget.h in CopyFiles */, + FDB52FD91F614B8E00AA92E6 /* ffitarget_arm64.h in CopyFiles */, + FDB52FDA1F614B9300AA92E6 /* ffitarget_armv7.h in CopyFiles */, + FDB52FDD1F614BA900AA92E6 /* ffitarget_x86_64.h in CopyFiles */, + ); + runOnlyForDeploymentPostprocessing = 0; + }; + FDB52FC11F6144FA00AA92E6 /* CopyFiles */ = { + isa = PBXCopyFilesBuildPhase; + buildActionMask = 12; + dstPath = "include/$(PRODUCT_NAME)"; + dstSubfolderSpec = 16; + files = ( + FDB52FD51F614AE200AA92E6 /* ffi.h in CopyFiles */, + FDB52FD61F614AEA00AA92E6 /* ffi_arm64.h in CopyFiles */, + FDB52FD71F614AED00AA92E6 /* ffi_x86_64.h in CopyFiles */, + FDB52FDE1F6155E300AA92E6 /* ffitarget.h in CopyFiles */, + FDB52FDF1F6155EA00AA92E6 /* ffitarget_arm64.h in CopyFiles */, + FDB52FE01F6155EF00AA92E6 /* ffitarget_x86_64.h in CopyFiles */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ); sourceTree = ""; }; +<<<<<<< HEAD DB13B1671849DF1E0010F42D /* Products */ = { isa = PBXGroup; children = ( DB13B1661849DF1E0010F42D /* libffi.a */, DB13B1911849DF510010F42D /* ffi.dylib */, +======= + FDB52FE11F6156E000AA92E6 /* CopyFiles */ = { + isa = PBXCopyFilesBuildPhase; + buildActionMask = 2147483647; + dstPath = "include/$(PRODUCT_NAME)"; + dstSubfolderSpec = 16; + files = ( + FDB52FE21F6156FA00AA92E6 /* ffi.h in CopyFiles */, + FDB52FE31F61571A00AA92E6 /* ffi_x86_64.h in CopyFiles */, + FDB52FE41F61571D00AA92E6 /* ffitarget.h in CopyFiles */, + FDB52FE61F61573100AA92E6 /* ffitarget_x86_64.h in CopyFiles */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ); name = Products; sourceTree = ""; }; +<<<<<<< HEAD DBFA713C187F1D8600A76262 /* darwin_common */ = { isa = PBXGroup; children = ( @@ -180,6 +279,139 @@ DBFA716A187F1D9B00A76262 /* src */ = { isa = PBXGroup; children = ( +======= +/* End PBXCopyFilesBuildPhase section */ + +/* Begin PBXFileReference section */ + 43B5D3F71D35473200D1E1FD /* ffiw64_x86_64.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = ffiw64_x86_64.c; sourceTree = ""; }; + 43B5D3F91D3547CE00D1E1FD /* win64_x86_64.S */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.asm; path = win64_x86_64.S; sourceTree = ""; }; + 43E9A5C61D352C1500926A8F /* unix64_x86_64.S */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.asm; path = unix64_x86_64.S; sourceTree = ""; }; + 43E9A5DA1D35373600926A8F /* internal.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = internal.h; sourceTree = ""; }; + 43E9A5DB1D35374400926A8F /* internal.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = internal.h; sourceTree = ""; }; + 43E9A5DC1D35375400926A8F /* internal.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = internal.h; sourceTree = ""; }; + 43E9A5DD1D35375400926A8F /* internal64.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = internal64.h; sourceTree = ""; }; + DB13B1661849DF1E0010F42D /* libffi.a */ = {isa = PBXFileReference; explicitFileType = archive.ar; includeInIndex = 0; path = libffi.a; sourceTree = BUILT_PRODUCTS_DIR; }; + DB13B1911849DF510010F42D /* ffi.dylib */ = {isa = PBXFileReference; explicitFileType = "compiled.mach-o.dylib"; includeInIndex = 0; path = ffi.dylib; sourceTree = BUILT_PRODUCTS_DIR; }; + DBFA713E187F1D8600A76262 /* ffi.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ffi.h; sourceTree = ""; }; + DBFA713F187F1D8600A76262 /* ffi_common.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ffi_common.h; sourceTree = ""; }; + DBFA7140187F1D8600A76262 /* fficonfig.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = fficonfig.h; sourceTree = ""; }; + DBFA7141187F1D8600A76262 /* ffitarget.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ffitarget.h; sourceTree = ""; }; + DBFA7143187F1D8600A76262 /* closures.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = closures.c; sourceTree = ""; }; + DBFA7145187F1D8600A76262 /* dlmalloc.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = dlmalloc.c; sourceTree = ""; }; + DBFA7147187F1D8600A76262 /* prep_cif.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = prep_cif.c; sourceTree = ""; }; + DBFA7148187F1D8600A76262 /* raw_api.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = raw_api.c; sourceTree = ""; }; + DBFA7149187F1D8600A76262 /* types.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = types.c; sourceTree = ""; }; + DBFA715E187F1D9B00A76262 /* ffi_arm64.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ffi_arm64.h; sourceTree = ""; }; + DBFA715F187F1D9B00A76262 /* ffi_armv7.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ffi_armv7.h; sourceTree = ""; }; + DBFA7161187F1D9B00A76262 /* ffi_x86_64.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ffi_x86_64.h; sourceTree = ""; }; + DBFA7162187F1D9B00A76262 /* fficonfig_arm64.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = fficonfig_arm64.h; sourceTree = ""; }; + DBFA7163187F1D9B00A76262 /* fficonfig_armv7.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = fficonfig_armv7.h; sourceTree = ""; }; + DBFA7165187F1D9B00A76262 /* fficonfig_x86_64.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = fficonfig_x86_64.h; sourceTree = ""; }; + DBFA7166187F1D9B00A76262 /* ffitarget_arm64.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ffitarget_arm64.h; sourceTree = ""; }; + DBFA7167187F1D9B00A76262 /* ffitarget_armv7.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ffitarget_armv7.h; sourceTree = ""; }; + DBFA7169187F1D9B00A76262 /* ffitarget_x86_64.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ffitarget_x86_64.h; sourceTree = ""; }; + DBFA716C187F1D9B00A76262 /* ffi_arm64.c */ = {isa = PBXFileReference; fileEncoding = 4; indentWidth = 2; lastKnownFileType = sourcecode.c.c; path = ffi_arm64.c; sourceTree = ""; }; + DBFA716D187F1D9B00A76262 /* sysv_arm64.S */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.asm; path = sysv_arm64.S; sourceTree = ""; }; + DBFA716F187F1D9B00A76262 /* ffi_armv7.c */ = {isa = PBXFileReference; fileEncoding = 4; indentWidth = 2; lastKnownFileType = sourcecode.c.c; path = ffi_armv7.c; sourceTree = ""; }; + DBFA7170187F1D9B00A76262 /* sysv_armv7.S */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.asm; path = sysv_armv7.S; sourceTree = ""; }; + DBFA7175187F1D9B00A76262 /* ffi64_x86_64.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = ffi64_x86_64.c; sourceTree = ""; }; + DBFA7183187F1DA100A76262 /* ffi_x86_64.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ffi_x86_64.h; sourceTree = ""; }; + DBFA7185187F1DA100A76262 /* fficonfig_x86_64.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = fficonfig_x86_64.h; sourceTree = ""; }; + DBFA7187187F1DA100A76262 /* ffitarget_x86_64.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = ffitarget_x86_64.h; sourceTree = ""; }; + DBFA718A187F1DA100A76262 /* unix64_x86_64.S */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.asm; path = unix64_x86_64.S; sourceTree = ""; }; + DBFA718C187F1DA100A76262 /* ffi64_x86_64.c */ = {isa = PBXFileReference; fileEncoding = 4; indentWidth = 2; lastKnownFileType = sourcecode.c.c; path = ffi64_x86_64.c; sourceTree = ""; }; + FDB52FC51F6144FA00AA92E6 /* libffi.a */ = {isa = PBXFileReference; explicitFileType = archive.ar; includeInIndex = 0; path = libffi.a; sourceTree = BUILT_PRODUCTS_DIR; }; + FDDB2F3E1F5D61BC00EF414E /* asmnames.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = asmnames.h; sourceTree = ""; }; + FDDB2F3F1F5D666900EF414E /* ffiw64_x86_64.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = ffiw64_x86_64.c; sourceTree = ""; }; + FDDB2F421F5D68C900EF414E /* internal64.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = internal64.h; sourceTree = ""; }; + FDDB2F431F5D68C900EF414E /* internal.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = internal.h; sourceTree = ""; }; + FDDB2F441F5D68C900EF414E /* win64_x86_64.S */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.asm; path = win64_x86_64.S; sourceTree = ""; }; + FDDB2F621F5D846400EF414E /* libffi.a */ = {isa = PBXFileReference; explicitFileType = "compiled.mach-o.dylib"; includeInIndex = 0; path = libffi.a; sourceTree = BUILT_PRODUCTS_DIR; }; +/* End PBXFileReference section */ + +/* Begin PBXGroup section */ + DB13B15B1849DEB70010F42D = { + isa = PBXGroup; + children = ( + DBFA713C187F1D8600A76262 /* darwin_common */, + DBFA715C187F1D9B00A76262 /* darwin_ios */, + DBFA7180187F1DA100A76262 /* darwin_osx */, + DB13B1671849DF1E0010F42D /* Products */, + ); + sourceTree = ""; + }; + DB13B1671849DF1E0010F42D /* Products */ = { + isa = PBXGroup; + children = ( + DB13B1661849DF1E0010F42D /* libffi.a */, + DB13B1911849DF510010F42D /* ffi.dylib */, + FDDB2F621F5D846400EF414E /* libffi.a */, + FDB52FC51F6144FA00AA92E6 /* libffi.a */, + ); + name = Products; + sourceTree = ""; + }; + DBFA713C187F1D8600A76262 /* darwin_common */ = { + isa = PBXGroup; + children = ( + DBFA713D187F1D8600A76262 /* include */, + DBFA7142187F1D8600A76262 /* src */, + ); + path = darwin_common; + sourceTree = ""; + }; + DBFA713D187F1D8600A76262 /* include */ = { + isa = PBXGroup; + children = ( + DBFA713E187F1D8600A76262 /* ffi.h */, + DBFA713F187F1D8600A76262 /* ffi_common.h */, + DBFA7140187F1D8600A76262 /* fficonfig.h */, + DBFA7141187F1D8600A76262 /* ffitarget.h */, + ); + path = include; + sourceTree = ""; + }; + DBFA7142187F1D8600A76262 /* src */ = { + isa = PBXGroup; + children = ( + DBFA7143187F1D8600A76262 /* closures.c */, + DBFA7145187F1D8600A76262 /* dlmalloc.c */, + DBFA7147187F1D8600A76262 /* prep_cif.c */, + DBFA7148187F1D8600A76262 /* raw_api.c */, + DBFA7149187F1D8600A76262 /* types.c */, + ); + path = src; + sourceTree = ""; + }; + DBFA715C187F1D9B00A76262 /* darwin_ios */ = { + isa = PBXGroup; + children = ( + DBFA715D187F1D9B00A76262 /* include */, + DBFA716A187F1D9B00A76262 /* src */, + ); + path = darwin_ios; + sourceTree = ""; + }; + DBFA715D187F1D9B00A76262 /* include */ = { + isa = PBXGroup; + children = ( + DBFA715E187F1D9B00A76262 /* ffi_arm64.h */, + DBFA715F187F1D9B00A76262 /* ffi_armv7.h */, + DBFA7161187F1D9B00A76262 /* ffi_x86_64.h */, + DBFA7162187F1D9B00A76262 /* fficonfig_arm64.h */, + DBFA7163187F1D9B00A76262 /* fficonfig_armv7.h */, + DBFA7165187F1D9B00A76262 /* fficonfig_x86_64.h */, + DBFA7166187F1D9B00A76262 /* ffitarget_arm64.h */, + DBFA7167187F1D9B00A76262 /* ffitarget_armv7.h */, + DBFA7169187F1D9B00A76262 /* ffitarget_x86_64.h */, + ); + path = include; + sourceTree = ""; + }; + DBFA716A187F1D9B00A76262 /* src */ = { + isa = PBXGroup; + children = ( +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 DBFA716B187F1D9B00A76262 /* aarch64 */, DBFA716E187F1D9B00A76262 /* arm */, DBFA7172187F1D9B00A76262 /* x86 */, @@ -190,6 +422,10 @@ DBFA716B187F1D9B00A76262 /* aarch64 */ = { isa = PBXGroup; children = ( +<<<<<<< HEAD +======= + 43E9A5DA1D35373600926A8F /* internal.h */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 DBFA716C187F1D9B00A76262 /* ffi_arm64.c */, DBFA716D187F1D9B00A76262 /* sysv_arm64.S */, ); @@ -199,9 +435,15 @@ DBFA716E187F1D9B00A76262 /* arm */ = { isa = PBXGroup; children = ( +<<<<<<< HEAD DBFA716F187F1D9B00A76262 /* ffi_armv7.c */, DBFA7170187F1D9B00A76262 /* sysv_armv7.S */, DBFA7171187F1D9B00A76262 /* trampoline_armv7.S */, +======= + 43E9A5DB1D35374400926A8F /* internal.h */, + DBFA716F187F1D9B00A76262 /* ffi_armv7.c */, + DBFA7170187F1D9B00A76262 /* sysv_armv7.S */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ); path = arm; sourceTree = ""; @@ -209,6 +451,7 @@ DBFA7172187F1D9B00A76262 /* x86 */ = { isa = PBXGroup; children = ( +<<<<<<< HEAD DBFA7173187F1D9B00A76262 /* darwin64_x86_64.S */, DBFA7174187F1D9B00A76262 /* darwin_i386.S */, DBFA7175187F1D9B00A76262 /* ffi64_x86_64.c */, @@ -234,6 +477,32 @@ DBFA7184187F1DA100A76262 /* fficonfig_i386.h */, DBFA7185187F1DA100A76262 /* fficonfig_x86_64.h */, DBFA7186187F1DA100A76262 /* ffitarget_i386.h */, +======= + 43E9A5DC1D35375400926A8F /* internal.h */, + 43E9A5DD1D35375400926A8F /* internal64.h */, + DBFA7175187F1D9B00A76262 /* ffi64_x86_64.c */, + 43B5D3F71D35473200D1E1FD /* ffiw64_x86_64.c */, + 43E9A5C61D352C1500926A8F /* unix64_x86_64.S */, + 43B5D3F91D3547CE00D1E1FD /* win64_x86_64.S */, + ); + path = x86; + sourceTree = ""; + }; + DBFA7180187F1DA100A76262 /* darwin_osx */ = { + isa = PBXGroup; + children = ( + DBFA7181187F1DA100A76262 /* include */, + DBFA7188187F1DA100A76262 /* src */, + ); + path = darwin_osx; + sourceTree = ""; + }; + DBFA7181187F1DA100A76262 /* include */ = { + isa = PBXGroup; + children = ( + DBFA7183187F1DA100A76262 /* ffi_x86_64.h */, + DBFA7185187F1DA100A76262 /* fficonfig_x86_64.h */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 DBFA7187187F1DA100A76262 /* ffitarget_x86_64.h */, ); path = include; @@ -250,10 +519,20 @@ DBFA7189187F1DA100A76262 /* x86 */ = { isa = PBXGroup; children = ( +<<<<<<< HEAD DBFA718A187F1DA100A76262 /* darwin64_x86_64.S */, DBFA718B187F1DA100A76262 /* darwin_i386.S */, DBFA718C187F1DA100A76262 /* ffi64_x86_64.c */, DBFA718D187F1DA100A76262 /* ffi_i386.c */, +======= + FDDB2F431F5D68C900EF414E /* internal.h */, + FDDB2F421F5D68C900EF414E /* internal64.h */, + FDDB2F3E1F5D61BC00EF414E /* asmnames.h */, + DBFA718C187F1DA100A76262 /* ffi64_x86_64.c */, + FDDB2F3F1F5D666900EF414E /* ffiw64_x86_64.c */, + DBFA718A187F1DA100A76262 /* unix64_x86_64.S */, + FDDB2F441F5D68C900EF414E /* win64_x86_64.S */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ); path = x86; sourceTree = ""; @@ -270,11 +549,16 @@ DBFA714A187F1D8600A76262 /* ffi.h in Headers */, DBFA718F187F1DA100A76262 /* ffi_x86_64.h in Headers */, DBFA7191187F1DA100A76262 /* fficonfig_x86_64.h in Headers */, +<<<<<<< HEAD DBFA718E187F1DA100A76262 /* ffi_i386.h in Headers */, DBFA7190187F1DA100A76262 /* fficonfig_i386.h in Headers */, DBFA714B187F1D8600A76262 /* ffi_common.h in Headers */, DBFA7193187F1DA100A76262 /* ffitarget_x86_64.h in Headers */, DBFA7192187F1DA100A76262 /* ffitarget_i386.h in Headers */, +======= + DBFA714B187F1D8600A76262 /* ffi_common.h in Headers */, + DBFA7193187F1DA100A76262 /* ffitarget_x86_64.h in Headers */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ); runOnlyForDeploymentPostprocessing = 0; }; @@ -285,7 +569,11 @@ isa = PBXNativeTarget; buildConfigurationList = DB13B18B1849DF1E0010F42D /* Build configuration list for PBXNativeTarget "libffi-iOS" */; buildPhases = ( +<<<<<<< HEAD DB13B3051849E01C0010F42D /* ShellScript */, +======= + 43B5D3FB1D35480D00D1E1FD /* Run Script */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 DB13B1621849DF1E0010F42D /* Sources */, DB13B1641849DF1E0010F42D /* CopyFiles */, ); @@ -299,6 +587,14 @@ productType = "com.apple.product-type.library.static"; }; DB13B1901849DF510010F42D /* libffi-Mac */ = { +<<<<<<< HEAD + isa = PBXNativeTarget; + buildConfigurationList = DB13B1B01849DF520010F42D /* Build configuration list for PBXNativeTarget "libffi-Mac" */; + buildPhases = ( + DB13B3061849E0490010F42D /* ShellScript */, + DB13B18D1849DF510010F42D /* Sources */, + DB13B18F1849DF510010F42D /* Headers */, +======= isa = PBXNativeTarget; buildConfigurationList = DB13B1B01849DF520010F42D /* Build configuration list for PBXNativeTarget "libffi-Mac" */; buildPhases = ( @@ -315,13 +611,59 @@ productReference = DB13B1911849DF510010F42D /* ffi.dylib */; productType = "com.apple.product-type.library.dynamic"; }; + FDB52FB01F6144FA00AA92E6 /* libffi-tvOS */ = { + isa = PBXNativeTarget; + buildConfigurationList = FDB52FC21F6144FA00AA92E6 /* Build configuration list for PBXNativeTarget "libffi-tvOS" */; + buildPhases = ( + FDB52FB11F6144FA00AA92E6 /* Run Script */, + FDB52FB21F6144FA00AA92E6 /* Sources */, + FDB52FC11F6144FA00AA92E6 /* CopyFiles */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + ); + buildRules = ( + ); + dependencies = ( + ); +<<<<<<< HEAD + name = "libffi-Mac"; + productName = ffi; + productReference = DB13B1911849DF510010F42D /* ffi.dylib */; + productType = "com.apple.product-type.library.dynamic"; +======= + name = "libffi-tvOS"; + productName = ffi; + productReference = FDB52FC51F6144FA00AA92E6 /* libffi.a */; + productType = "com.apple.product-type.library.static"; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + }; + FDDB2F471F5D846400EF414E /* libffi-static-Mac */ = { + isa = PBXNativeTarget; + buildConfigurationList = FDDB2F5F1F5D846400EF414E /* Build configuration list for PBXNativeTarget "libffi-static-Mac" */; + buildPhases = ( + FDDB2F481F5D846400EF414E /* ShellScript */, + FDDB2F491F5D846400EF414E /* Sources */, + FDB52FE11F6156E000AA92E6 /* CopyFiles */, + ); + buildRules = ( + ); + dependencies = ( + ); + name = "libffi-static-Mac"; + productName = ffi; + productReference = FDDB2F621F5D846400EF414E /* libffi.a */; + productType = "com.apple.product-type.library.dynamic"; + }; /* End PBXNativeTarget section */ /* Begin PBXProject section */ DB13B15C1849DEB70010F42D /* Project object */ = { isa = PBXProject; attributes = { +<<<<<<< HEAD LastUpgradeCheck = 0510; +======= + LastUpgradeCheck = 0830; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 }; buildConfigurationList = DB13B15F1849DEB70010F42D /* Build configuration list for PBXProject "libffi" */; compatibilityVersion = "Xcode 3.2"; @@ -336,13 +678,37 @@ projectRoot = ""; targets = ( DB13B1651849DF1E0010F42D /* libffi-iOS */, +<<<<<<< HEAD DB13B1901849DF510010F42D /* libffi-Mac */, +======= + FDB52FB01F6144FA00AA92E6 /* libffi-tvOS */, + DB13B1901849DF510010F42D /* libffi-Mac */, + FDDB2F471F5D846400EF414E /* libffi-static-Mac */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ); }; /* End PBXProject section */ /* Begin PBXShellScriptBuildPhase section */ +<<<<<<< HEAD DB13B3051849E01C0010F42D /* ShellScript */ = { +======= + 43B5D3FB1D35480D00D1E1FD /* Run Script */ = { + isa = PBXShellScriptBuildPhase; + buildActionMask = 2147483647; + files = ( + ); + inputPaths = ( + ); + name = "Run Script"; + outputPaths = ( + ); + runOnlyForDeploymentPostprocessing = 0; + shellPath = /bin/sh; + shellScript = "if [ ! -f \"./compile\" ]\nthen\nautoreconf -i -f -v\nif [ -f \"../ltmain.sh\" ]\nthen\necho \"fixing ltmain.sh for some reason\"\nmv ../ltmain.sh ./\nautoreconf -i -f -v\nfi\n/usr/bin/python generate-darwin-source-and-headers.py --only-ios\nfi"; + }; + DB13B3061849E0490010F42D /* ShellScript */ = { +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 isa = PBXShellScriptBuildPhase; buildActionMask = 2147483647; files = ( @@ -353,25 +719,53 @@ ); runOnlyForDeploymentPostprocessing = 0; shellPath = /bin/sh; +<<<<<<< HEAD shellScript = "/usr/bin/python generate-darwin-source-and-headers.py --only-ios"; }; DB13B3061849E0490010F42D /* ShellScript */ = { +======= + shellScript = "if [ ! -f \"./compile\" ]\nthen\nautoreconf -i -f -v\nif [ -f \"../ltmain.sh\" ]\nthen\necho \"fixing ltmain.sh for some reason\"\nmv ../ltmain.sh ./\nautoreconf -i -f -v\nfi\n/usr/bin/python generate-darwin-source-and-headers.py --only-osx\nfi"; + }; + FDB52FB11F6144FA00AA92E6 /* Run Script */ = { +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 isa = PBXShellScriptBuildPhase; buildActionMask = 2147483647; files = ( ); inputPaths = ( ); +<<<<<<< HEAD +======= + name = "Run Script"; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 outputPaths = ( ); runOnlyForDeploymentPostprocessing = 0; shellPath = /bin/sh; +<<<<<<< HEAD shellScript = "/usr/bin/python generate-darwin-source-and-headers.py --only-osx"; +======= + shellScript = "if [ ! -f \"./compile\" ]\nthen\nautoreconf -i -f -v\nif [ -f \"../ltmain.sh\" ]\nthen\necho \"fixing ltmain.sh for some reason\"\nmv ../ltmain.sh ./\nautoreconf -i -f -v\nfi\n/usr/bin/python generate-darwin-source-and-headers.py --only-ios\nfi"; + }; + FDDB2F481F5D846400EF414E /* ShellScript */ = { + isa = PBXShellScriptBuildPhase; + buildActionMask = 2147483647; + files = ( + ); + inputPaths = ( + ); + outputPaths = ( + ); + runOnlyForDeploymentPostprocessing = 0; + shellPath = /bin/sh; + shellScript = "if [ ! -f \"./compile\" ]\nthen\nautoreconf -i -f -v\nif [ -f \"../ltmain.sh\" ]\nthen\necho \"fixing ltmain.sh for some reason\"\nmv ../ltmain.sh ./\nautoreconf -i -f -v\nfi\n/usr/bin/python generate-darwin-source-and-headers.py --only-osx\nfi"; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 }; /* End PBXShellScriptBuildPhase section */ /* Begin PBXSourcesBuildPhase section */ DB13B1621849DF1E0010F42D /* Sources */ = { +<<<<<<< HEAD isa = PBXSourcesBuildPhase; buildActionMask = 2147483647; files = ( @@ -403,6 +797,72 @@ DBFA7159187F1D8600A76262 /* raw_api.c in Sources */, DBFA714F187F1D8600A76262 /* closures.c in Sources */, DBFA7194187F1DA100A76262 /* darwin64_x86_64.S in Sources */, +======= + isa = PBXSourcesBuildPhase; + buildActionMask = 2147483647; + files = ( + 43E9A5C81D352C1500926A8F /* unix64_x86_64.S in Sources */, + DBFA717E187F1D9B00A76262 /* ffi64_x86_64.c in Sources */, + DBFA7179187F1D9B00A76262 /* ffi_armv7.c in Sources */, + DBFA714E187F1D8600A76262 /* closures.c in Sources */, + DBFA717A187F1D9B00A76262 /* sysv_armv7.S in Sources */, + 43B5D3F81D35473200D1E1FD /* ffiw64_x86_64.c in Sources */, + DBFA7156187F1D8600A76262 /* prep_cif.c in Sources */, + DBFA7158187F1D8600A76262 /* raw_api.c in Sources */, + DBFA7178187F1D9B00A76262 /* sysv_arm64.S in Sources */, + DBFA715A187F1D8600A76262 /* types.c in Sources */, + DBFA7177187F1D9B00A76262 /* ffi_arm64.c in Sources */, + 43B5D3FA1D3547CE00D1E1FD /* win64_x86_64.S in Sources */, + ); + runOnlyForDeploymentPostprocessing = 0; + }; + DB13B18D1849DF510010F42D /* Sources */ = { + isa = PBXSourcesBuildPhase; + buildActionMask = 2147483647; + files = ( + DBFA7196187F1DA100A76262 /* ffi64_x86_64.c in Sources */, + DBFA7157187F1D8600A76262 /* prep_cif.c in Sources */, + FDDB2F411F5D66E200EF414E /* ffiw64_x86_64.c in Sources */, + DBFA715B187F1D8600A76262 /* types.c in Sources */, + DBFA7159187F1D8600A76262 /* raw_api.c in Sources */, + DBFA714F187F1D8600A76262 /* closures.c in Sources */, + DBFA7194187F1DA100A76262 /* unix64_x86_64.S in Sources */, + FDDB2F461F5D691E00EF414E /* win64_x86_64.S in Sources */, + ); + runOnlyForDeploymentPostprocessing = 0; + }; + FDB52FB21F6144FA00AA92E6 /* Sources */ = { + isa = PBXSourcesBuildPhase; + buildActionMask = 2147483647; + files = ( + FDB52FB31F6144FA00AA92E6 /* unix64_x86_64.S in Sources */, + FDB52FB51F6144FA00AA92E6 /* ffi64_x86_64.c in Sources */, + FDB52FB61F6144FA00AA92E6 /* ffi_armv7.c in Sources */, + FDB52FB71F6144FA00AA92E6 /* closures.c in Sources */, + FDB52FB81F6144FA00AA92E6 /* sysv_armv7.S in Sources */, + FDB52FB91F6144FA00AA92E6 /* ffiw64_x86_64.c in Sources */, + FDB52FBA1F6144FA00AA92E6 /* prep_cif.c in Sources */, + FDB52FBC1F6144FA00AA92E6 /* raw_api.c in Sources */, + FDB52FBD1F6144FA00AA92E6 /* sysv_arm64.S in Sources */, + FDB52FBE1F6144FA00AA92E6 /* types.c in Sources */, + FDB52FBF1F6144FA00AA92E6 /* ffi_arm64.c in Sources */, + FDB52FC01F6144FA00AA92E6 /* win64_x86_64.S in Sources */, + ); + runOnlyForDeploymentPostprocessing = 0; + }; + FDDB2F491F5D846400EF414E /* Sources */ = { + isa = PBXSourcesBuildPhase; + buildActionMask = 2147483647; + files = ( + FDDB2F4A1F5D846400EF414E /* ffi64_x86_64.c in Sources */, + FDDB2F4C1F5D846400EF414E /* prep_cif.c in Sources */, + FDDB2F4E1F5D846400EF414E /* ffiw64_x86_64.c in Sources */, + FDDB2F4F1F5D846400EF414E /* types.c in Sources */, + FDDB2F501F5D846400EF414E /* raw_api.c in Sources */, + FDDB2F511F5D846400EF414E /* closures.c in Sources */, + FDDB2F521F5D846400EF414E /* unix64_x86_64.S in Sources */, + FDDB2F531F5D846400EF414E /* win64_x86_64.S in Sources */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ); runOnlyForDeploymentPostprocessing = 0; }; @@ -410,6 +870,7 @@ /* Begin XCBuildConfiguration section */ DB13B1601849DEB70010F42D /* Debug */ = { +<<<<<<< HEAD isa = XCBuildConfiguration; buildSettings = { HEADER_SEARCH_PATHS = ( @@ -439,6 +900,66 @@ CLANG_CXX_LIBRARY = "libc++"; CLANG_ENABLE_MODULES = YES; CLANG_ENABLE_OBJC_ARC = YES; +======= + isa = XCBuildConfiguration; + buildSettings = { + CLANG_WARN_BOOL_CONVERSION = YES; + CLANG_WARN_CONSTANT_CONVERSION = YES; + CLANG_WARN_EMPTY_BODY = YES; + CLANG_WARN_ENUM_CONVERSION = YES; + CLANG_WARN_INFINITE_RECURSION = YES; + CLANG_WARN_INT_CONVERSION = YES; + CLANG_WARN_SUSPICIOUS_MOVE = YES; + CLANG_WARN_UNREACHABLE_CODE = YES; + CLANG_WARN__DUPLICATE_METHOD_MATCH = YES; + ENABLE_STRICT_OBJC_MSGSEND = YES; + ENABLE_TESTABILITY = YES; + GCC_NO_COMMON_BLOCKS = YES; + GCC_WARN_64_TO_32_BIT_CONVERSION = YES; + GCC_WARN_ABOUT_RETURN_TYPE = YES; + GCC_WARN_UNDECLARED_SELECTOR = YES; + GCC_WARN_UNINITIALIZED_AUTOS = YES; + GCC_WARN_UNUSED_FUNCTION = YES; + GCC_WARN_UNUSED_VARIABLE = YES; + HEADER_SEARCH_PATHS = ( + "$(inherited)", + darwin_common/include, + ); + ONLY_ACTIVE_ARCH = YES; + }; + name = Debug; + }; + DB13B1611849DEB70010F42D /* Release */ = { + isa = XCBuildConfiguration; + buildSettings = { + CLANG_WARN_BOOL_CONVERSION = YES; + CLANG_WARN_CONSTANT_CONVERSION = YES; + CLANG_WARN_EMPTY_BODY = YES; + CLANG_WARN_ENUM_CONVERSION = YES; + CLANG_WARN_INFINITE_RECURSION = YES; + CLANG_WARN_INT_CONVERSION = YES; + CLANG_WARN_SUSPICIOUS_MOVE = YES; + CLANG_WARN_UNREACHABLE_CODE = YES; + CLANG_WARN__DUPLICATE_METHOD_MATCH = YES; + ENABLE_STRICT_OBJC_MSGSEND = YES; + GCC_NO_COMMON_BLOCKS = YES; + GCC_WARN_64_TO_32_BIT_CONVERSION = YES; + GCC_WARN_ABOUT_RETURN_TYPE = YES; + GCC_WARN_UNDECLARED_SELECTOR = YES; + GCC_WARN_UNINITIALIZED_AUTOS = YES; + GCC_WARN_UNUSED_FUNCTION = YES; + GCC_WARN_UNUSED_VARIABLE = YES; + HEADER_SEARCH_PATHS = ( + "$(inherited)", + darwin_common/include, + ); + }; + name = Release; + }; + DB13B1871849DF1E0010F42D /* Debug */ = { + isa = XCBuildConfiguration; + buildSettings = { + ALWAYS_SEARCH_USER_PATHS = NO; CLANG_WARN_BOOL_CONVERSION = YES; CLANG_WARN_CONSTANT_CONVERSION = YES; CLANG_WARN_DIRECT_OBJC_ISA_USAGE = YES_ERROR; @@ -449,8 +970,85 @@ CLANG_WARN__DUPLICATE_METHOD_MATCH = YES; COPY_PHASE_STRIP = NO; DSTROOT = /tmp/ffi.dst; + GCC_OPTIMIZATION_LEVEL = 0; + GCC_PREPROCESSOR_DEFINITIONS = ( + "DEBUG=1", + "$(inherited)", + ); + GCC_WARN_64_TO_32_BIT_CONVERSION = YES; + GCC_WARN_ABOUT_RETURN_TYPE = YES_ERROR; + GCC_WARN_UNDECLARED_SELECTOR = YES; + GCC_WARN_UNINITIALIZED_AUTOS = YES_AGGRESSIVE; + GCC_WARN_UNUSED_FUNCTION = YES; + GCC_WARN_UNUSED_VARIABLE = YES; + HEADER_SEARCH_PATHS = ( + "$(inherited)", + darwin_ios/include, + ); + IPHONEOS_DEPLOYMENT_TARGET = 8.0; + PRODUCT_NAME = ffi; + SDKROOT = iphoneos; + SKIP_INSTALL = YES; + VALID_ARCHS = "arm64 armv7 armv7s x86_64"; + }; + name = Debug; + }; + DB13B1881849DF1E0010F42D /* Release */ = { + isa = XCBuildConfiguration; + buildSettings = { + ALWAYS_SEARCH_USER_PATHS = NO; + CLANG_WARN_BOOL_CONVERSION = YES; + CLANG_WARN_CONSTANT_CONVERSION = YES; + CLANG_WARN_DIRECT_OBJC_ISA_USAGE = YES_ERROR; + CLANG_WARN_EMPTY_BODY = YES; + CLANG_WARN_ENUM_CONVERSION = YES; + CLANG_WARN_INT_CONVERSION = YES; + CLANG_WARN_OBJC_ROOT_CLASS = YES_ERROR; + CLANG_WARN__DUPLICATE_METHOD_MATCH = YES; + COPY_PHASE_STRIP = YES; + DSTROOT = /tmp/ffi.dst; + ENABLE_NS_ASSERTIONS = NO; + GCC_WARN_64_TO_32_BIT_CONVERSION = YES; + GCC_WARN_ABOUT_RETURN_TYPE = YES_ERROR; + GCC_WARN_UNDECLARED_SELECTOR = YES; + GCC_WARN_UNINITIALIZED_AUTOS = YES_AGGRESSIVE; + GCC_WARN_UNUSED_FUNCTION = YES; + GCC_WARN_UNUSED_VARIABLE = YES; + HEADER_SEARCH_PATHS = ( + "$(inherited)", + darwin_ios/include, + ); + IPHONEOS_DEPLOYMENT_TARGET = 8.0; + PRODUCT_NAME = ffi; + SDKROOT = iphoneos; + SKIP_INSTALL = YES; + VALIDATE_PRODUCT = YES; + VALID_ARCHS = "arm64 armv7 armv7s x86_64"; + }; + name = Release; + }; + DB13B1B11849DF520010F42D /* Debug */ = { + isa = XCBuildConfiguration; + buildSettings = { + ALWAYS_SEARCH_USER_PATHS = NO; + CLANG_CXX_LANGUAGE_STANDARD = "gnu++0x"; + CLANG_CXX_LIBRARY = "libc++"; + CLANG_ENABLE_OBJC_ARC = YES; + CLANG_WARN_BOOL_CONVERSION = YES; + CLANG_WARN_CONSTANT_CONVERSION = YES; + CLANG_WARN_DIRECT_OBJC_ISA_USAGE = YES_ERROR; + CLANG_WARN_EMPTY_BODY = YES; + CLANG_WARN_ENUM_CONVERSION = YES; + CLANG_WARN_INT_CONVERSION = YES; + CLANG_WARN_OBJC_ROOT_CLASS = YES_ERROR; + CLANG_WARN__DUPLICATE_METHOD_MATCH = YES; + COMBINE_HIDPI_IMAGES = YES; + COPY_PHASE_STRIP = NO; + DYLIB_COMPATIBILITY_VERSION = 1; + DYLIB_CURRENT_VERSION = 1; GCC_C_LANGUAGE_STANDARD = gnu99; GCC_DYNAMIC_NO_PIC = NO; + GCC_ENABLE_OBJC_EXCEPTIONS = YES; GCC_OPTIMIZATION_LEVEL = 0; GCC_PREPROCESSOR_DEFINITIONS = ( "DEBUG=1", @@ -465,6 +1063,94 @@ GCC_WARN_UNUSED_VARIABLE = YES; HEADER_SEARCH_PATHS = ( "$(inherited)", + darwin_osx/include, + ); + MACOSX_DEPLOYMENT_TARGET = 10.6; + ONLY_ACTIVE_ARCH = YES; + OTHER_LDFLAGS = "-Wl,-no_compact_unwind"; + PRODUCT_NAME = ffi; + SDKROOT = macosx; + }; + name = Debug; + }; + DB13B1B21849DF520010F42D /* Release */ = { + isa = XCBuildConfiguration; + buildSettings = { + ALWAYS_SEARCH_USER_PATHS = NO; + CLANG_CXX_LANGUAGE_STANDARD = "gnu++0x"; + CLANG_CXX_LIBRARY = "libc++"; + CLANG_ENABLE_OBJC_ARC = YES; + CLANG_WARN_BOOL_CONVERSION = YES; + CLANG_WARN_CONSTANT_CONVERSION = YES; + CLANG_WARN_DIRECT_OBJC_ISA_USAGE = YES_ERROR; + CLANG_WARN_EMPTY_BODY = YES; + CLANG_WARN_ENUM_CONVERSION = YES; + CLANG_WARN_INT_CONVERSION = YES; + CLANG_WARN_OBJC_ROOT_CLASS = YES_ERROR; + CLANG_WARN__DUPLICATE_METHOD_MATCH = YES; + COMBINE_HIDPI_IMAGES = YES; + COPY_PHASE_STRIP = YES; + DEBUG_INFORMATION_FORMAT = "dwarf-with-dsym"; + DYLIB_COMPATIBILITY_VERSION = 1; + DYLIB_CURRENT_VERSION = 1; + ENABLE_NS_ASSERTIONS = NO; + GCC_C_LANGUAGE_STANDARD = gnu99; + GCC_ENABLE_OBJC_EXCEPTIONS = YES; + GCC_WARN_64_TO_32_BIT_CONVERSION = YES; + GCC_WARN_ABOUT_RETURN_TYPE = YES_ERROR; + GCC_WARN_UNDECLARED_SELECTOR = YES; + GCC_WARN_UNINITIALIZED_AUTOS = YES_AGGRESSIVE; + GCC_WARN_UNUSED_FUNCTION = YES; + GCC_WARN_UNUSED_VARIABLE = YES; + HEADER_SEARCH_PATHS = ( + "$(inherited)", + darwin_osx/include, + ); + MACOSX_DEPLOYMENT_TARGET = 10.6; + OTHER_LDFLAGS = "-Wl,-no_compact_unwind"; + PRODUCT_NAME = ffi; + SDKROOT = macosx; + }; + name = Release; + }; + FDB52FC31F6144FA00AA92E6 /* Debug */ = { + isa = XCBuildConfiguration; + buildSettings = { + ALWAYS_SEARCH_USER_PATHS = NO; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + CLANG_WARN_BOOL_CONVERSION = YES; + CLANG_WARN_CONSTANT_CONVERSION = YES; + CLANG_WARN_DIRECT_OBJC_ISA_USAGE = YES_ERROR; + CLANG_WARN_EMPTY_BODY = YES; + CLANG_WARN_ENUM_CONVERSION = YES; + CLANG_WARN_INT_CONVERSION = YES; + CLANG_WARN_OBJC_ROOT_CLASS = YES_ERROR; + CLANG_WARN__DUPLICATE_METHOD_MATCH = YES; + COPY_PHASE_STRIP = NO; +<<<<<<< HEAD + DSTROOT = /tmp/ffi.dst; + GCC_C_LANGUAGE_STANDARD = gnu99; + GCC_DYNAMIC_NO_PIC = NO; +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + GCC_OPTIMIZATION_LEVEL = 0; + GCC_PREPROCESSOR_DEFINITIONS = ( + "DEBUG=1", + "$(inherited)", + ); +<<<<<<< HEAD + GCC_SYMBOLS_PRIVATE_EXTERN = NO; +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + GCC_WARN_64_TO_32_BIT_CONVERSION = YES; + GCC_WARN_ABOUT_RETURN_TYPE = YES_ERROR; + GCC_WARN_UNDECLARED_SELECTOR = YES; + GCC_WARN_UNINITIALIZED_AUTOS = YES_AGGRESSIVE; + GCC_WARN_UNUSED_FUNCTION = YES; + GCC_WARN_UNUSED_VARIABLE = YES; + HEADER_SEARCH_PATHS = ( + "$(inherited)", +<<<<<<< HEAD "darwin_ios/include", ); IPHONEOS_DEPLOYMENT_TARGET = 5.0; @@ -485,6 +1171,21 @@ CLANG_CXX_LIBRARY = "libc++"; CLANG_ENABLE_MODULES = YES; CLANG_ENABLE_OBJC_ARC = YES; +======= + darwin_ios/include, + ); + PRODUCT_NAME = ffi; + SDKROOT = appletvos; + SKIP_INSTALL = YES; + TVOS_DEPLOYMENT_TARGET = 9.0; + }; + name = Debug; + }; + FDB52FC41F6144FA00AA92E6 /* Release */ = { + isa = XCBuildConfiguration; + buildSettings = { + ALWAYS_SEARCH_USER_PATHS = NO; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 CLANG_WARN_BOOL_CONVERSION = YES; CLANG_WARN_CONSTANT_CONVERSION = YES; CLANG_WARN_DIRECT_OBJC_ISA_USAGE = YES_ERROR; @@ -494,9 +1195,13 @@ CLANG_WARN_OBJC_ROOT_CLASS = YES_ERROR; CLANG_WARN__DUPLICATE_METHOD_MATCH = YES; COPY_PHASE_STRIP = YES; +<<<<<<< HEAD DSTROOT = /tmp/ffi.dst; ENABLE_NS_ASSERTIONS = NO; GCC_C_LANGUAGE_STANDARD = gnu99; +======= + ENABLE_NS_ASSERTIONS = NO; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 GCC_WARN_64_TO_32_BIT_CONVERSION = YES; GCC_WARN_ABOUT_RETURN_TYPE = YES_ERROR; GCC_WARN_UNDECLARED_SELECTOR = YES; @@ -505,6 +1210,7 @@ GCC_WARN_UNUSED_VARIABLE = YES; HEADER_SEARCH_PATHS = ( "$(inherited)", +<<<<<<< HEAD "darwin_ios/include", ); IPHONEOS_DEPLOYMENT_TARGET = 5.0; @@ -513,11 +1219,23 @@ PRODUCT_NAME = ffi; SDKROOT = iphoneos; SKIP_INSTALL = YES; +======= + darwin_ios/include, + ); + PRODUCT_NAME = ffi; + SDKROOT = appletvos; + SKIP_INSTALL = YES; + TVOS_DEPLOYMENT_TARGET = 9.0; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 VALIDATE_PRODUCT = YES; }; name = Release; }; +<<<<<<< HEAD DB13B1B11849DF520010F42D /* Debug */ = { +======= + FDDB2F601F5D846400EF414E /* Debug */ = { +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 isa = XCBuildConfiguration; buildSettings = { ALWAYS_SEARCH_USER_PATHS = NO; @@ -532,9 +1250,16 @@ CLANG_WARN_INT_CONVERSION = YES; CLANG_WARN_OBJC_ROOT_CLASS = YES_ERROR; CLANG_WARN__DUPLICATE_METHOD_MATCH = YES; +<<<<<<< HEAD COPY_PHASE_STRIP = NO; DYLIB_COMPATIBILITY_VERSION = 1; DYLIB_CURRENT_VERSION = 1; +======= + COMBINE_HIDPI_IMAGES = YES; + COPY_PHASE_STRIP = NO; + EXECUTABLE_EXTENSION = a; + EXECUTABLE_PREFIX = lib; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 GCC_C_LANGUAGE_STANDARD = gnu99; GCC_DYNAMIC_NO_PIC = NO; GCC_ENABLE_OBJC_EXCEPTIONS = YES; @@ -552,17 +1277,29 @@ GCC_WARN_UNUSED_VARIABLE = YES; HEADER_SEARCH_PATHS = ( "$(inherited)", +<<<<<<< HEAD "darwin_osx/include", ); MACOSX_DEPLOYMENT_TARGET = 10.6; ONLY_ACTIVE_ARCH = YES; OTHER_LDFLAGS = "-Wl,-no_compact_unwind"; +======= + darwin_osx/include, + ); + MACH_O_TYPE = staticlib; + MACOSX_DEPLOYMENT_TARGET = 10.6; + ONLY_ACTIVE_ARCH = YES; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 PRODUCT_NAME = ffi; SDKROOT = macosx; }; name = Debug; }; +<<<<<<< HEAD DB13B1B21849DF520010F42D /* Release */ = { +======= + FDDB2F611F5D846400EF414E /* Release */ = { +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 isa = XCBuildConfiguration; buildSettings = { ALWAYS_SEARCH_USER_PATHS = NO; @@ -577,11 +1314,20 @@ CLANG_WARN_INT_CONVERSION = YES; CLANG_WARN_OBJC_ROOT_CLASS = YES_ERROR; CLANG_WARN__DUPLICATE_METHOD_MATCH = YES; +<<<<<<< HEAD COPY_PHASE_STRIP = YES; DEBUG_INFORMATION_FORMAT = "dwarf-with-dsym"; DYLIB_COMPATIBILITY_VERSION = 1; DYLIB_CURRENT_VERSION = 1; ENABLE_NS_ASSERTIONS = NO; +======= + COMBINE_HIDPI_IMAGES = YES; + COPY_PHASE_STRIP = YES; + DEBUG_INFORMATION_FORMAT = "dwarf-with-dsym"; + ENABLE_NS_ASSERTIONS = NO; + EXECUTABLE_EXTENSION = a; + EXECUTABLE_PREFIX = lib; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 GCC_C_LANGUAGE_STANDARD = gnu99; GCC_ENABLE_OBJC_EXCEPTIONS = YES; GCC_WARN_64_TO_32_BIT_CONVERSION = YES; @@ -592,10 +1338,17 @@ GCC_WARN_UNUSED_VARIABLE = YES; HEADER_SEARCH_PATHS = ( "$(inherited)", +<<<<<<< HEAD "darwin_osx/include", ); MACOSX_DEPLOYMENT_TARGET = 10.6; OTHER_LDFLAGS = "-Wl,-no_compact_unwind"; +======= + darwin_osx/include, + ); + MACH_O_TYPE = staticlib; + MACOSX_DEPLOYMENT_TARGET = 10.6; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 PRODUCT_NAME = ffi; SDKROOT = macosx; }; @@ -605,6 +1358,12 @@ /* Begin XCConfigurationList section */ DB13B15F1849DEB70010F42D /* Build configuration list for PBXProject "libffi" */ = { +<<<<<<< HEAD + isa = XCConfigurationList; + buildConfigurations = ( + DB13B1601849DEB70010F42D /* Debug */, + DB13B1611849DEB70010F42D /* Release */, +======= isa = XCConfigurationList; buildConfigurations = ( DB13B1601849DEB70010F42D /* Debug */, @@ -627,6 +1386,41 @@ buildConfigurations = ( DB13B1B11849DF520010F42D /* Debug */, DB13B1B21849DF520010F42D /* Release */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + ); + defaultConfigurationIsVisible = 0; + defaultConfigurationName = Release; + }; +<<<<<<< HEAD + DB13B18B1849DF1E0010F42D /* Build configuration list for PBXNativeTarget "libffi-iOS" */ = { + isa = XCConfigurationList; + buildConfigurations = ( + DB13B1871849DF1E0010F42D /* Debug */, + DB13B1881849DF1E0010F42D /* Release */, +======= + FDB52FC21F6144FA00AA92E6 /* Build configuration list for PBXNativeTarget "libffi-tvOS" */ = { + isa = XCConfigurationList; + buildConfigurations = ( + FDB52FC31F6144FA00AA92E6 /* Debug */, + FDB52FC41F6144FA00AA92E6 /* Release */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + ); + defaultConfigurationIsVisible = 0; + defaultConfigurationName = Release; + }; +<<<<<<< HEAD + DB13B1B01849DF520010F42D /* Build configuration list for PBXNativeTarget "libffi-Mac" */ = { + isa = XCConfigurationList; + buildConfigurations = ( + DB13B1B11849DF520010F42D /* Debug */, + DB13B1B21849DF520010F42D /* Release */, +======= + FDDB2F5F1F5D846400EF414E /* Build configuration list for PBXNativeTarget "libffi-static-Mac" */ = { + isa = XCConfigurationList; + buildConfigurations = ( + FDDB2F601F5D846400EF414E /* Debug */, + FDDB2F611F5D846400EF414E /* Release */, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ); defaultConfigurationIsVisible = 0; defaultConfigurationName = Release; diff --git a/native/libffi/libtool-version b/native/libffi/libtool-version index 149a51a2ab..e365bacf94 100644 --- a/native/libffi/libtool-version +++ b/native/libffi/libtool-version @@ -26,4 +26,8 @@ # release, then set age to 0. # # CURRENT:REVISION:AGE +<<<<<<< HEAD 6:4:0 +======= +9:0:1 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 diff --git a/native/libffi/m4/asmcfi.m4 b/native/libffi/m4/asmcfi.m4 index dbf73a0b3e..3e28602274 100644 --- a/native/libffi/m4/asmcfi.m4 +++ b/native/libffi/m4/asmcfi.m4 @@ -2,7 +2,7 @@ AC_DEFUN([GCC_AS_CFI_PSEUDO_OP], [AC_CACHE_CHECK([assembler .cfi pseudo-op support], gcc_cv_as_cfi_pseudo_op, [ gcc_cv_as_cfi_pseudo_op=unknown - AC_TRY_COMPILE([asm (".cfi_startproc\n\t.cfi_endproc");],, + AC_TRY_COMPILE([asm (".cfi_sections\n\t.cfi_startproc\n\t.cfi_endproc");],, [gcc_cv_as_cfi_pseudo_op=yes], [gcc_cv_as_cfi_pseudo_op=no]) ]) diff --git a/native/libffi/m4/ax_append_flag.m4 b/native/libffi/m4/ax_append_flag.m4 index 1d38b76fb8..dd6d8b6140 100644 --- a/native/libffi/m4/ax_append_flag.m4 +++ b/native/libffi/m4/ax_append_flag.m4 @@ -1,5 +1,5 @@ # =========================================================================== -# http://www.gnu.org/software/autoconf-archive/ax_append_flag.html +# https://www.gnu.org/software/autoconf-archive/ax_append_flag.html # =========================================================================== # # SYNOPSIS @@ -23,47 +23,28 @@ # Copyright (c) 2008 Guido U. Draheim # Copyright (c) 2011 Maarten Bosmans # -# This program is free software: you can redistribute it and/or modify it -# under the terms of the GNU General Public License as published by the -# Free Software Foundation, either version 3 of the License, or (at your -# option) any later version. -# -# This program is distributed in the hope that it will be useful, but -# WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General -# Public License for more details. -# -# You should have received a copy of the GNU General Public License along -# with this program. If not, see . -# -# As a special exception, the respective Autoconf Macro's copyright owner -# gives unlimited permission to copy, distribute and modify the configure -# scripts that are the output of Autoconf when processing the Macro. You -# need not follow the terms of the GNU General Public License when using -# or distributing such scripts, even though portions of the text of the -# Macro appear in them. The GNU General Public License (GPL) does govern -# all other use of the material that constitutes the Autoconf Macro. -# -# This special exception to the GPL applies to versions of the Autoconf -# Macro released by the Autoconf Archive. When you make and distribute a -# modified version of the Autoconf Macro, you may extend this special -# exception to the GPL to apply to your modified version as well. +# Copying and distribution of this file, with or without modification, are +# permitted in any medium without royalty provided the copyright notice +# and this notice are preserved. This file is offered as-is, without any +# warranty. -#serial 2 +#serial 8 AC_DEFUN([AX_APPEND_FLAG], -[AC_PREREQ(2.59)dnl for _AC_LANG_PREFIX -AS_VAR_PUSHDEF([FLAGS], [m4_default($2,_AC_LANG_PREFIX[FLAGS])])dnl -AS_VAR_SET_IF(FLAGS, - [case " AS_VAR_GET(FLAGS) " in - *" $1 "*) - AC_RUN_LOG([: FLAGS already contains $1]) - ;; - *) - AC_RUN_LOG([: FLAGS="$FLAGS $1"]) - AS_VAR_SET(FLAGS, ["AS_VAR_GET(FLAGS) $1"]) - ;; - esac], - [AS_VAR_SET(FLAGS,["$1"])]) +[dnl +AC_PREREQ(2.64)dnl for _AC_LANG_PREFIX and AS_VAR_SET_IF +AS_VAR_PUSHDEF([FLAGS], [m4_default($2,_AC_LANG_PREFIX[FLAGS])]) +AS_VAR_SET_IF(FLAGS,[ + AS_CASE([" AS_VAR_GET(FLAGS) "], + [*" $1 "*], [AC_RUN_LOG([: FLAGS already contains $1])], + [ + AS_VAR_APPEND(FLAGS,[" $1"]) + AC_RUN_LOG([: FLAGS="$FLAGS"]) + ]) + ], + [ + AS_VAR_SET(FLAGS,[$1]) + AC_RUN_LOG([: FLAGS="$FLAGS"]) + ]) AS_VAR_POPDEF([FLAGS])dnl ])dnl AX_APPEND_FLAG diff --git a/native/libffi/m4/ax_cc_maxopt.m4 b/native/libffi/m4/ax_cc_maxopt.m4 index 62e3b5330b..9e7f1ee4a1 100644 --- a/native/libffi/m4/ax_cc_maxopt.m4 +++ b/native/libffi/m4/ax_cc_maxopt.m4 @@ -1,5 +1,5 @@ # =========================================================================== -# http://www.gnu.org/software/autoconf-archive/ax_cc_maxopt.html +# https://www.gnu.org/software/autoconf-archive/ax_cc_maxopt.html # =========================================================================== # # SYNOPSIS @@ -40,7 +40,7 @@ # Public License for more details. # # You should have received a copy of the GNU General Public License along -# with this program. If not, see . +# with this program. If not, see . # # As a special exception, the respective Autoconf Macro's copyright owner # gives unlimited permission to copy, distribute and modify the configure @@ -55,7 +55,7 @@ # modified version of the Autoconf Macro, you may extend this special # exception to the GPL to apply to your modified version as well. -#serial 13 +#serial 17 AC_DEFUN([AX_CC_MAXOPT], [ @@ -115,11 +115,19 @@ if test "$ac_test_CFLAGS" != "set"; then AX_GCC_X86_CPUID(0) AX_GCC_X86_CPUID(1) case $ax_cv_gcc_x86_cpuid_0 in # see AX_GCC_ARCHFLAG - *:756e6547:*:*) # Intel + *:756e6547:6c65746e:49656e69) # Intel case $ax_cv_gcc_x86_cpuid_1 in - *6a?:*[[234]]:*:*|*6[[789b]]?:*:*:*) icc_flags="-xK";; - *f3[[347]]:*:*:*|*f4[1347]:*:*:*) icc_flags="-xP -xN -xW -xK";; - *f??:*:*:*) icc_flags="-xN -xW -xK";; + *0?6[[78ab]]?:*:*:*|?6[[78ab]]?:*:*:*|6[[78ab]]?:*:*:*) icc_flags="-xK" ;; + *0?6[[9d]]?:*:*:*|?6[[9d]]?:*:*:*|6[[9d]]?:*:*:*|*1?65?:*:*:*) icc_flags="-xSSE2 -xB -xK" ;; + *0?6e?:*:*:*|?6e?:*:*:*|6e?:*:*:*) icc_flags="-xSSE3 -xP -xO -xB -xK" ;; + *0?6f?:*:*:*|?6f?:*:*:*|6f?:*:*:*|*1?66?:*:*:*) icc_flags="-xSSSE3 -xT -xB -xK" ;; + *1?6[[7d]]?:*:*:*) icc_flags="-xSSE4.1 -xS -xT -xB -xK" ;; + *1?6[[aef]]?:*:*:*|*2?6[[5cef]]?:*:*:*) icc_flags="-xSSE4.2 -xS -xT -xB -xK" ;; + *2?6[[ad]]?:*:*:*) icc_flags="-xAVX -SSE4.2 -xS -xT -xB -xK" ;; + *3?6[[ae]]?:*:*:*) icc_flags="-xCORE-AVX-I -xAVX -SSE4.2 -xS -xT -xB -xK" ;; + *3?6[[cf]]?:*:*:*|*4?6[[56]]?:*:*:*) icc_flags="-xCORE-AVX2 -xCORE-AVX-I -xAVX -SSE4.2 -xS -xT -xB -xK" ;; + *000?f[[346]]?:*:*:*|?f[[346]]?:*:*:*|f[[346]]?:*:*:*) icc_flags="-xSSE3 -xP -xO -xN -xW -xK" ;; + *00??f??:*:*:*|??f??:*:*:*|?f??:*:*:*|f??:*:*:*) icc_flags="-xSSE2 -xN -xW -xK" ;; esac ;; esac ;; esac @@ -141,7 +149,7 @@ if test "$ac_test_CFLAGS" != "set"; then CFLAGS="-O3 -fomit-frame-pointer" # -malign-double for x86 systems - # LIBFFI -- DON'T DO THIS - CHANGES ABI + # libffi local change -- don't align double, as it changes the ABI # AX_CHECK_COMPILE_FLAG(-malign-double, CFLAGS="$CFLAGS -malign-double") # -fstrict-aliasing for gcc-2.95+ @@ -153,6 +161,11 @@ if test "$ac_test_CFLAGS" != "set"; then AX_GCC_ARCHFLAG($acx_maxopt_portable) ;; + + microsoft) + # default optimization flags for MSVC opt builds + CFLAGS="-O2" + ;; esac if test -z "$CFLAGS"; then diff --git a/native/libffi/m4/ax_cflags_warn_all.m4 b/native/libffi/m4/ax_cflags_warn_all.m4 index 0fa3e186b4..094577e458 100644 --- a/native/libffi/m4/ax_cflags_warn_all.m4 +++ b/native/libffi/m4/ax_cflags_warn_all.m4 @@ -1,5 +1,5 @@ # =========================================================================== -# http://www.gnu.org/software/autoconf-archive/ax_cflags_warn_all.html +# https://www.gnu.org/software/autoconf-archive/ax_cflags_warn_all.html # =========================================================================== # # SYNOPSIS @@ -43,7 +43,7 @@ # Public License for more details. # # You should have received a copy of the GNU General Public License along -# with this program. If not, see . +# with this program. If not, see . # # As a special exception, the respective Autoconf Macro's copyright owner # gives unlimited permission to copy, distribute and modify the configure @@ -58,7 +58,7 @@ # modified version of the Autoconf Macro, you may extend this special # exception to the GPL to apply to your modified version as well. -#serial 14 +#serial 16 AC_DEFUN([AX_FLAGS_WARN_ALL],[dnl AS_VAR_PUSHDEF([FLAGS],[_AC_LANG_PREFIX[]FLAGS])dnl @@ -84,7 +84,7 @@ done FLAGS="$ac_save_[]FLAGS" ]) AS_VAR_POPDEF([FLAGS])dnl -AC_REQUIRE([AX_APPEND_FLAG]) +AX_REQUIRE_DEFINED([AX_APPEND_FLAG]) case ".$VAR" in .ok|.ok,*) m4_ifvaln($3,$3) ;; .|.no|.no,*) m4_default($4,[m4_ifval($2,[AX_APPEND_FLAG([$2], [$1])])]) ;; diff --git a/native/libffi/m4/ax_check_compile_flag.m4 b/native/libffi/m4/ax_check_compile_flag.m4 index c3a8d695a1..bd753b34d7 100644 --- a/native/libffi/m4/ax_check_compile_flag.m4 +++ b/native/libffi/m4/ax_check_compile_flag.m4 @@ -1,10 +1,10 @@ # =========================================================================== -# http://www.gnu.org/software/autoconf-archive/ax_check_compile_flag.html +# https://www.gnu.org/software/autoconf-archive/ax_check_compile_flag.html # =========================================================================== # # SYNOPSIS # -# AX_CHECK_COMPILE_FLAG(FLAG, [ACTION-SUCCESS], [ACTION-FAILURE], [EXTRA-FLAGS]) +# AX_CHECK_COMPILE_FLAG(FLAG, [ACTION-SUCCESS], [ACTION-FAILURE], [EXTRA-FLAGS], [INPUT]) # # DESCRIPTION # @@ -19,6 +19,8 @@ # the flags: "CFLAGS EXTRA-FLAGS FLAG". This can for example be used to # force the compiler to issue an error when a bad flag is given. # +# INPUT gives an alternative input source to AC_COMPILE_IFELSE. +# # NOTE: Implementation based on AX_CFLAGS_GCC_OPTION. Please keep this # macro in sync with AX_CHECK_{PREPROC,LINK}_FLAG. # @@ -27,45 +29,24 @@ # Copyright (c) 2008 Guido U. Draheim # Copyright (c) 2011 Maarten Bosmans # -# This program is free software: you can redistribute it and/or modify it -# under the terms of the GNU General Public License as published by the -# Free Software Foundation, either version 3 of the License, or (at your -# option) any later version. -# -# This program is distributed in the hope that it will be useful, but -# WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General -# Public License for more details. -# -# You should have received a copy of the GNU General Public License along -# with this program. If not, see . -# -# As a special exception, the respective Autoconf Macro's copyright owner -# gives unlimited permission to copy, distribute and modify the configure -# scripts that are the output of Autoconf when processing the Macro. You -# need not follow the terms of the GNU General Public License when using -# or distributing such scripts, even though portions of the text of the -# Macro appear in them. The GNU General Public License (GPL) does govern -# all other use of the material that constitutes the Autoconf Macro. -# -# This special exception to the GPL applies to versions of the Autoconf -# Macro released by the Autoconf Archive. When you make and distribute a -# modified version of the Autoconf Macro, you may extend this special -# exception to the GPL to apply to your modified version as well. +# Copying and distribution of this file, with or without modification, are +# permitted in any medium without royalty provided the copyright notice +# and this notice are preserved. This file is offered as-is, without any +# warranty. -#serial 2 +#serial 6 AC_DEFUN([AX_CHECK_COMPILE_FLAG], -[AC_PREREQ(2.59)dnl for _AC_LANG_PREFIX +[AC_PREREQ(2.64)dnl for _AC_LANG_PREFIX and AS_VAR_IF AS_VAR_PUSHDEF([CACHEVAR],[ax_cv_check_[]_AC_LANG_ABBREV[]flags_$4_$1])dnl AC_CACHE_CHECK([whether _AC_LANG compiler accepts $1], CACHEVAR, [ ax_check_save_flags=$[]_AC_LANG_PREFIX[]FLAGS _AC_LANG_PREFIX[]FLAGS="$[]_AC_LANG_PREFIX[]FLAGS $4 $1" - AC_COMPILE_IFELSE([AC_LANG_PROGRAM()], + AC_COMPILE_IFELSE([m4_default([$5],[AC_LANG_PROGRAM()])], [AS_VAR_SET(CACHEVAR,[yes])], [AS_VAR_SET(CACHEVAR,[no])]) _AC_LANG_PREFIX[]FLAGS=$ax_check_save_flags]) -AS_IF([test x"AS_VAR_GET(CACHEVAR)" = xyes], +AS_VAR_IF(CACHEVAR,yes, [m4_default([$2], :)], [m4_default([$3], :)]) AS_VAR_POPDEF([CACHEVAR])dnl diff --git a/native/libffi/m4/ax_compiler_vendor.m4 b/native/libffi/m4/ax_compiler_vendor.m4 index 73e32ea9ef..73efdb0055 100644 --- a/native/libffi/m4/ax_compiler_vendor.m4 +++ b/native/libffi/m4/ax_compiler_vendor.m4 @@ -1,5 +1,5 @@ # =========================================================================== -# http://www.gnu.org/software/autoconf-archive/ax_compiler_vendor.html +# https://www.gnu.org/software/autoconf-archive/ax_compiler_vendor.html # =========================================================================== # # SYNOPSIS @@ -29,7 +29,7 @@ # Public License for more details. # # You should have received a copy of the GNU General Public License along -# with this program. If not, see . +# with this program. If not, see . # # As a special exception, the respective Autoconf Macro's copyright owner # gives unlimited permission to copy, distribute and modify the configure @@ -44,22 +44,25 @@ # modified version of the Autoconf Macro, you may extend this special # exception to the GPL to apply to your modified version as well. -#serial 11 +#serial 17 AC_DEFUN([AX_COMPILER_VENDOR], [AC_CACHE_CHECK([for _AC_LANG compiler vendor], ax_cv_[]_AC_LANG_ABBREV[]_compiler_vendor, + dnl Please add if possible support to ax_compiler_version.m4 [# note: don't check for gcc first since some other compilers define __GNUC__ vendors="intel: __ICC,__ECC,__INTEL_COMPILER ibm: __xlc__,__xlC__,__IBMC__,__IBMCPP__ pathscale: __PATHCC__,__PATHSCALE__ clang: __clang__ + cray: _CRAYC + fujitsu: __FUJITSU + sdcc: SDCC, __SDCC gnu: __GNUC__ sun: __SUNPRO_C,__SUNPRO_CC hp: __HP_cc,__HP_aCC dec: __DECC,__DECCXX,__DECC_VER,__DECCXX_VER - borland: __BORLANDC__,__TURBOC__ + borland: __BORLANDC__,__CODEGEARC__,__TURBOC__ comeau: __COMO__ - cray: _CRAYC kai: __KCC lcc: __LCC__ sgi: __sgi,sgi @@ -67,6 +70,7 @@ AC_DEFUN([AX_COMPILER_VENDOR], metrowerks: __MWERKS__ watcom: __WATCOMC__ portland: __PGI + tcc: __TINYC__ unknown: UNKNOWN" for ventest in $vendors; do case $ventest in diff --git a/native/libffi/m4/ax_configure_args.m4 b/native/libffi/m4/ax_configure_args.m4 index 0726b1bc8a..9237efea7a 100644 --- a/native/libffi/m4/ax_configure_args.m4 +++ b/native/libffi/m4/ax_configure_args.m4 @@ -1,5 +1,5 @@ # =========================================================================== -# http://www.gnu.org/software/autoconf-archive/ax_configure_args.html +# https://www.gnu.org/software/autoconf-archive/ax_configure_args.html # =========================================================================== # # SYNOPSIS @@ -15,7 +15,7 @@ # to rely on eval'ing $ac_configure_args however some old autoconf # versions do not provide that. To ensure maximum portability of autoconf # extension macros this helper can be AC_REQUIRE'd so that -# $ac_configure_args will alsways be present. +# $ac_configure_args will always be present. # # Sadly, the traditional "exec $SHELL" of the enable_builddir macros is # spoiled now and must be replaced by "eval + exit $?". @@ -31,36 +31,15 @@ # # Copyright (c) 2008 Guido U. Draheim # -# This program is free software; you can redistribute it and/or modify it -# under the terms of the GNU General Public License as published by the -# Free Software Foundation; either version 3 of the License, or (at your -# option) any later version. -# -# This program is distributed in the hope that it will be useful, but -# WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General -# Public License for more details. -# -# You should have received a copy of the GNU General Public License along -# with this program. If not, see . -# -# As a special exception, the respective Autoconf Macro's copyright owner -# gives unlimited permission to copy, distribute and modify the configure -# scripts that are the output of Autoconf when processing the Macro. You -# need not follow the terms of the GNU General Public License when using -# or distributing such scripts, even though portions of the text of the -# Macro appear in them. The GNU General Public License (GPL) does govern -# all other use of the material that constitutes the Autoconf Macro. -# -# This special exception to the GPL applies to versions of the Autoconf -# Macro released by the Autoconf Archive. When you make and distribute a -# modified version of the Autoconf Macro, you may extend this special -# exception to the GPL to apply to your modified version as well. +# Copying and distribution of this file, with or without modification, are +# permitted in any medium without royalty provided the copyright notice +# and this notice are preserved. This file is offered as-is, without any +# warranty. -#serial 9 +#serial 14 AC_DEFUN([AX_CONFIGURE_ARGS],[ - # [$]@ is unsable in 2.60+ but earlier autoconf had no ac_configure_args + # [$]@ is unusable in 2.60+ but earlier autoconf had no ac_configure_args if test "${ac_configure_args+set}" != "set" ; then ac_configure_args= for ac_arg in ${1+"[$]@"}; do diff --git a/native/libffi/m4/ax_enable_builddir.m4 b/native/libffi/m4/ax_enable_builddir.m4 index 3cb2093718..d0106f6e0c 100644 --- a/native/libffi/m4/ax_enable_builddir.m4 +++ b/native/libffi/m4/ax_enable_builddir.m4 @@ -1,5 +1,5 @@ # =========================================================================== -# http://www.gnu.org/software/autoconf-archive/ax_enable_builddir.html +# https://www.gnu.org/software/autoconf-archive/ax_enable_builddir.html # =========================================================================== # # SYNOPSIS @@ -24,7 +24,7 @@ # toplevel builddir Makefile. It just copies the variables and # rule-targets, each extended with a default rule-execution that recurses # into the build directory of the current "HOST". You can override the -# auto-dection through `config.guess` and build-time of course, as in +# auto-detection through `config.guess` and build-time of course, as in # # make HOST=i386-mingw-cross # @@ -42,7 +42,7 @@ # into. Usually, the last one is the only one used. However, almost all # targets have an additional "*-all" rule which makes the script to # recurse into _all_ variants of the current HOST (!!) setting. The "-all" -# suffix can be overriden for the macro as well. +# suffix can be overridden for the macro as well. # # a special rule is only given for things like "dist" that will copy the # tarball from the builddir to the sourcedir (or $(PUB)) for reason of @@ -64,7 +64,7 @@ # Public License for more details. # # You should have received a copy of the GNU General Public License along -# with this program. If not, see . +# with this program. If not, see . # # As a special exception, the respective Autoconf Macro's copyright owner # gives unlimited permission to copy, distribute and modify the configure @@ -79,10 +79,11 @@ # modified version of the Autoconf Macro, you may extend this special # exception to the GPL to apply to your modified version as well. -#serial 23 +#serial 30 AC_DEFUN([AX_ENABLE_BUILDDIR],[ AC_REQUIRE([AC_CANONICAL_HOST])[]dnl +AC_REQUIRE([AC_CANONICAL_TARGET])[]dnl AC_REQUIRE([AX_CONFIGURE_ARGS])[]dnl AC_REQUIRE([AM_AUX_DIR_EXPAND])[]dnl AC_BEFORE([$0],[AM_INIT_AUTOMAKE])dnl @@ -121,7 +122,11 @@ if test ".$srcdir" = ".." ; then test -f $srcdir/$cache_file && mv $srcdir/$cache_file . AC_MSG_RESULT(....exec $SHELL $srcdir/[$]0 "--srcdir=$srcdir" "--enable-builddir=$SUB" ${1+"[$]@"}) case "[$]0" in # restart +<<<<<<< HEAD [[\\/]]* | ?:[[\\/]]*) # Asbolute name +======= + [[\\/]]* | ?:[[\\/]]*) # Absolute name +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 eval $SHELL "'[$]0'" "'--srcdir=$srcdir'" "'--enable-builddir=$SUB'" $ac_configure_args ;; *) eval $SHELL "'$srcdir/[$]0'" "'--srcdir=$srcdir'" "'--enable-builddir=$SUB'" $ac_configure_args ;; esac ; exit $? diff --git a/native/libffi/m4/ax_gcc_archflag.m4 b/native/libffi/m4/ax_gcc_archflag.m4 index aab2661c30..cc2b6927e3 100644 --- a/native/libffi/m4/ax_gcc_archflag.m4 +++ b/native/libffi/m4/ax_gcc_archflag.m4 @@ -1,5 +1,5 @@ # =========================================================================== -# http://www.gnu.org/software/autoconf-archive/ax_gcc_archflag.html +# https://www.gnu.org/software/autoconf-archive/ax_gcc_archflag.html # =========================================================================== # # SYNOPSIS @@ -36,7 +36,8 @@ # # Copyright (c) 2008 Steven G. Johnson # Copyright (c) 2008 Matteo Frigo -# Copyright (c) 2012 Tsukasa Oi +# Copyright (c) 2014 Tsukasa Oi +# Copyright (c) 2017-2018 Alexey Kopytov # # This program is free software: you can redistribute it and/or modify it # under the terms of the GNU General Public License as published by the @@ -49,7 +50,7 @@ # Public License for more details. # # You should have received a copy of the GNU General Public License along -# with this program. If not, see . +# with this program. If not, see . # # As a special exception, the respective Autoconf Macro's copyright owner # gives unlimited permission to copy, distribute and modify the configure @@ -64,11 +65,13 @@ # modified version of the Autoconf Macro, you may extend this special # exception to the GPL to apply to your modified version as well. -#serial 11 +#serial 22 AC_DEFUN([AX_GCC_ARCHFLAG], [AC_REQUIRE([AC_PROG_CC]) AC_REQUIRE([AC_CANONICAL_HOST]) +AC_REQUIRE([AC_PROG_SED]) +AC_REQUIRE([AX_COMPILER_VENDOR]) AC_ARG_WITH(gcc-arch, [AS_HELP_STRING([--with-gcc-arch=], [use architecture for gcc -march/-mtune, instead of guessing])], ax_gcc_arch=$withval, ax_gcc_arch=yes) @@ -85,63 +88,70 @@ if test "x$ax_gcc_arch" = xyes; then ax_gcc_arch="" if test "$cross_compiling" = no; then case $host_cpu in - i[[3456]]86*|x86_64*) # use cpuid codes + i[[3456]]86*|x86_64*|amd64*) # use cpuid codes AX_GCC_X86_CPUID(0) AX_GCC_X86_CPUID(1) case $ax_cv_gcc_x86_cpuid_0 in - *:756e6547:*:*) # Intel + *:756e6547:6c65746e:49656e69) # Intel case $ax_cv_gcc_x86_cpuid_1 in - *5[[48]]?:*:*:*) ax_gcc_arch="pentium-mmx pentium" ;; - *5??:*:*:*) ax_gcc_arch=pentium ;; - *0?6[[3456]]?:*:*:*) ax_gcc_arch="pentium2 pentiumpro" ;; - *0?6a?:*[[01]]:*:*) ax_gcc_arch="pentium2 pentiumpro" ;; - *0?6a?:*[[234]]:*:*) ax_gcc_arch="pentium3 pentiumpro" ;; - *0?6[[9de]]?:*:*:*) ax_gcc_arch="pentium-m pentium3 pentiumpro" ;; - *0?6[[78b]]?:*:*:*) ax_gcc_arch="pentium3 pentiumpro" ;; - *0?6f?:*:*:*|*1?66?:*:*:*) ax_gcc_arch="core2 pentium-m pentium3 pentiumpro" ;; + *5[[4578]]?:*:*:*) ax_gcc_arch="pentium-mmx pentium" ;; + *5[[123]]?:*:*:*) ax_gcc_arch=pentium ;; + *0?61?:*:*:*|?61?:*:*:*|61?:*:*:*) ax_gcc_arch=pentiumpro ;; + *0?6[[356]]?:*:*:*|?6[[356]]?:*:*:*|6[[356]]?:*:*:*) ax_gcc_arch="pentium2 pentiumpro" ;; + *0?6[[78ab]]?:*:*:*|?6[[78ab]]?:*:*:*|6[[78ab]]?:*:*:*) ax_gcc_arch="pentium3 pentiumpro" ;; + *0?6[[9d]]?:*:*:*|?6[[9d]]?:*:*:*|6[[9d]]?:*:*:*|*1?65?:*:*:*) ax_gcc_arch="pentium-m pentium3 pentiumpro" ;; + *0?6e?:*:*:*|?6e?:*:*:*|6e?:*:*:*) ax_gcc_arch="yonah pentium-m pentium3 pentiumpro" ;; + *0?6f?:*:*:*|?6f?:*:*:*|6f?:*:*:*|*1?66?:*:*:*) ax_gcc_arch="core2 pentium-m pentium3 pentiumpro" ;; *1?6[[7d]]?:*:*:*) ax_gcc_arch="penryn core2 pentium-m pentium3 pentiumpro" ;; - *1?6[[aef]]?:*:*:*|*2?6[[5cef]]?:*:*:*) ax_gcc_arch="corei7 core2 pentium-m pentium3 pentiumpro" ;; - *1?6c?:*:*:*|*[[23]]?66?:*:*:*) ax_gcc_arch="atom core2 pentium-m pentium3 pentiumpro" ;; - *2?6[[ad]]?:*:*:*) ax_gcc_arch="corei7-avx corei7 core2 pentium-m pentium3 pentiumpro" ;; - *0?6??:*:*:*) ax_gcc_arch=pentiumpro ;; - *6??:*:*:*) ax_gcc_arch="core2 pentiumpro" ;; - ?000?f3[[347]]:*:*:*|?000?f4[1347]:*:*:*|?000?f6?:*:*:*) - case $host_cpu in - x86_64*) ax_gcc_arch="nocona pentium4 pentiumpro" ;; - *) ax_gcc_arch="prescott pentium4 pentiumpro" ;; - esac ;; - ?000?f??:*:*:*) ax_gcc_arch="pentium4 pentiumpro";; + *1?6[[aef]]?:*:*:*|*2?6e?:*:*:*) ax_gcc_arch="nehalem corei7 core2 pentium-m pentium3 pentiumpro" ;; + *2?6[[5cf]]?:*:*:*) ax_gcc_arch="westmere corei7 core2 pentium-m pentium3 pentiumpro" ;; + *2?6[[ad]]?:*:*:*) ax_gcc_arch="sandybridge corei7-avx corei7 core2 pentium-m pentium3 pentiumpro" ;; + *3?6[[ae]]?:*:*:*) ax_gcc_arch="ivybridge core-avx-i corei7-avx corei7 core2 pentium-m pentium3 pentiumpro" ;; + *3?6[[cf]]?:*:*:*|*4?6[[56]]?:*:*:*) ax_gcc_arch="haswell core-avx2 core-avx-i corei7-avx corei7 core2 pentium-m pentium3 pentiumpro" ;; + *3?6d?:*:*:*|*4?6[[7f]]?:*:*:*|*5?66?:*:*:*) ax_gcc_arch="broadwell core-avx2 core-avx-i corei7-avx corei7 core2 pentium-m pentium3 pentiumpro" ;; + *1?6c?:*:*:*|*2?6[[67]]?:*:*:*|*3?6[[56]]?:*:*:*) ax_gcc_arch="bonnell atom core2 pentium-m pentium3 pentiumpro" ;; + *3?67?:*:*:*|*[[45]]?6[[ad]]?:*:*:*) ax_gcc_arch="silvermont atom core2 pentium-m pentium3 pentiumpro" ;; + *000?f[[012]]?:*:*:*|?f[[012]]?:*:*:*|f[[012]]?:*:*:*) ax_gcc_arch="pentium4 pentiumpro" ;; + *000?f[[346]]?:*:*:*|?f[[346]]?:*:*:*|f[[346]]?:*:*:*) ax_gcc_arch="nocona prescott pentium4 pentiumpro" ;; + # fallback + *5??:*:*:*) ax_gcc_arch=pentium ;; + *??6??:*:*:*) ax_gcc_arch="core2 pentiumpro" ;; + *6??:*:*:*) ax_gcc_arch=pentiumpro ;; + *00??f??:*:*:*|??f??:*:*:*|?f??:*:*:*|f??:*:*:*) ax_gcc_arch="pentium4 pentiumpro" ;; esac ;; - *:68747541:*:*) # AMD + *:68747541:444d4163:69746e65) # AMD case $ax_cv_gcc_x86_cpuid_1 in *5[[67]]?:*:*:*) ax_gcc_arch=k6 ;; - *5[[8d]]?:*:*:*) ax_gcc_arch="k6-2 k6" ;; - *5[[9]]?:*:*:*) ax_gcc_arch="k6-3 k6" ;; - *60?:*:*:*) ax_gcc_arch=k7 ;; + *5[[8]]?:*:*:*) ax_gcc_arch="k6-2 k6" ;; + *5[[9d]]?:*:*:*) ax_gcc_arch="k6-3 k6" ;; *6[[12]]?:*:*:*) ax_gcc_arch="athlon k7" ;; *6[[34]]?:*:*:*) ax_gcc_arch="athlon-tbird k7" ;; - *67?:*:*:*) ax_gcc_arch="athlon-4 athlon k7" ;; - *6[[68a]]?:*:*:*) - AX_GCC_X86_CPUID(0x80000006) # L2 cache size - case $ax_cv_gcc_x86_cpuid_0x80000006 in - *:*:*[[1-9a-f]]??????:*) # (L2 = ecx >> 16) >= 256 - ax_gcc_arch="athlon-xp athlon-4 athlon k7" ;; - *) ax_gcc_arch="athlon-4 athlon k7" ;; - esac ;; - ?00??f[[4cef8b]]?:*:*:*) ax_gcc_arch="athlon64 k8" ;; - ?00??f5?:*:*:*) ax_gcc_arch="opteron k8" ;; - ?00??f7?:*:*:*) ax_gcc_arch="athlon-fx opteron k8" ;; - ?00??f??:*:*:*) ax_gcc_arch="k8" ;; - ?05??f??:*:*:*) ax_gcc_arch="btver1 amdfam10 k8" ;; - ?06??f??:*:*:*) ax_gcc_arch="bdver1 amdfam10 k8" ;; - *f??:*:*:*) ax_gcc_arch="amdfam10 k8" ;; + *6[[678a]]?:*:*:*) ax_gcc_arch="athlon-xp athlon-4 athlon k7" ;; + *000?f[[4578bcef]]?:*:*:*|?f[[4578bcef]]?:*:*:*|f[[4578bcef]]?:*:*:*|*001?f[[4578bcf]]?:*:*:*|1?f[[4578bcf]]?:*:*:*) ax_gcc_arch="athlon64 k8" ;; + *002?f[[13457bcf]]?:*:*:*|2?f[[13457bcf]]?:*:*:*|*004?f[[138bcf]]?:*:*:*|4?f[[138bcf]]?:*:*:*|*005?f[[df]]?:*:*:*|5?f[[df]]?:*:*:*|*006?f[[8bcf]]?:*:*:*|6?f[[8bcf]]?:*:*:*|*007?f[[cf]]?:*:*:*|7?f[[cf]]?:*:*:*|*00c?f1?:*:*:*|c?f1?:*:*:*|*020?f3?:*:*:*|20?f3?:*:*:*) ax_gcc_arch="athlon64-sse3 k8-sse3 athlon64 k8" ;; + *010?f[[245689a]]?:*:*:*|10?f[[245689a]]?:*:*:*|*030?f1?:*:*:*|30?f1?:*:*:*) ax_gcc_arch="barcelona amdfam10 k8" ;; + *050?f[[12]]?:*:*:*|50?f[[12]]?:*:*:*) ax_gcc_arch="btver1 amdfam10 k8" ;; + *060?f1?:*:*:*|60?f1?:*:*:*) ax_gcc_arch="bdver1 amdfam10 k8" ;; + *060?f2?:*:*:*|60?f2?:*:*:*|*061?f[[03]]?:*:*:*|61?f[[03]]?:*:*:*) ax_gcc_arch="bdver2 bdver1 amdfam10 k8" ;; + *063?f0?:*:*:*|63?f0?:*:*:*) ax_gcc_arch="bdver3 bdver2 bdver1 amdfam10 k8" ;; + *07[[03]]?f0?:*:*:*|7[[03]]?f0?:*:*:*) ax_gcc_arch="btver2 btver1 amdfam10 k8" ;; + # fallback + *0[[13]]??f??:*:*:*|[[13]]??f??:*:*:*) ax_gcc_arch="barcelona amdfam10 k8" ;; + *020?f??:*:*:*|20?f??:*:*:*) ax_gcc_arch="athlon64-sse3 k8-sse3 athlon64 k8" ;; + *05??f??:*:*:*|5??f??:*:*:*) ax_gcc_arch="btver1 amdfam10 k8" ;; + *060?f??:*:*:*|60?f??:*:*:*) ax_gcc_arch="bdver1 amdfam10 k8" ;; + *061?f??:*:*:*|61?f??:*:*:*) ax_gcc_arch="bdver2 bdver1 amdfam10 k8" ;; + *06??f??:*:*:*|6??f??:*:*:*) ax_gcc_arch="bdver3 bdver2 bdver1 amdfam10 k8" ;; + *070?f??:*:*:*|70?f??:*:*:*) ax_gcc_arch="btver2 btver1 amdfam10 k8" ;; + *???f??:*:*:*) ax_gcc_arch="amdfam10 k8" ;; esac ;; - *:746e6543:*:*) # IDT + *:746e6543:736c7561:48727561) # IDT / VIA (Centaur) case $ax_cv_gcc_x86_cpuid_1 in *54?:*:*:*) ax_gcc_arch=winchip-c6 ;; - *58?:*:*:*) ax_gcc_arch=winchip2 ;; + *5[[89]]?:*:*:*) ax_gcc_arch=winchip2 ;; + *66?:*:*:*) ax_gcc_arch=winchip2 ;; *6[[78]]?:*:*:*) ax_gcc_arch=c3 ;; - *69?:*:*:*) ax_gcc_arch="c3-2 c3" ;; + *6[[9adf]]?:*:*:*) ax_gcc_arch="c3-2 c3" ;; esac ;; esac if test x"$ax_gcc_arch" = x; then # fallback @@ -155,7 +165,11 @@ case $host_cpu in sparc*) AC_PATH_PROG([PRTDIAG], [prtdiag], [prtdiag], [$PATH:/usr/platform/`uname -i`/sbin/:/usr/platform/`uname -m`/sbin/]) cputype=`(((grep cpu /proc/cpuinfo | cut -d: -f2) ; ($PRTDIAG -v |grep -i sparc) ; grep -i cpu /var/run/dmesg.boot ) | head -n 1) 2> /dev/null` +<<<<<<< HEAD cputype=`echo "$cputype" | tr -d ' -' | sed 's/SPARCIIi/SPARCII/' | tr $as_cr_LETTERS $as_cr_letters` +======= + cputype=`echo "$cputype" | tr -d ' -' | $SED 's/SPARCIIi/SPARCII/' |tr $as_cr_LETTERS $as_cr_letters` +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 case $cputype in *ultrasparciv*) ax_gcc_arch="ultrasparc4 ultrasparc3 ultrasparc v9" ;; *ultrasparciii*) ax_gcc_arch="ultrasparc3 ultrasparc v9" ;; @@ -177,8 +191,8 @@ case $host_cpu in alphaev79) ax_gcc_arch="ev79 ev7 ev69 ev68 ev67" ;; powerpc*) - cputype=`((grep cpu /proc/cpuinfo | head -n 1 | cut -d: -f2 | cut -d, -f1 | sed 's/ //g') ; /usr/bin/machine ; /bin/machine; grep CPU /var/run/dmesg.boot | head -n 1 | cut -d" " -f2) 2> /dev/null` - cputype=`echo $cputype | sed -e 's/ppc//g;s/ *//g'` + cputype=`((grep cpu /proc/cpuinfo | head -n 1 | cut -d: -f2 | cut -d, -f1 | $SED 's/ //g') ; /usr/bin/machine ; /bin/machine; grep CPU /var/run/dmesg.boot | head -n 1 | cut -d" " -f2) 2> /dev/null` + cputype=`echo $cputype | $SED -e 's/ppc//g;s/ *//g'` case $cputype in *750*) ax_gcc_arch="750 G3" ;; *740[[0-9]]*) ax_gcc_arch="$cputype 7400 G4" ;; @@ -188,26 +202,58 @@ case $host_cpu in *POWER4*|*power4*|*gq*) ax_gcc_arch="power4 970";; *POWER5*|*power5*|*gr*|*gs*) ax_gcc_arch="power5 power4 970";; 603ev|8240) ax_gcc_arch="$cputype 603e 603";; + *POWER7*) ax_gcc_arch="power7";; + *POWER8*) ax_gcc_arch="power8";; + *POWER9*) ax_gcc_arch="power9";; + *POWER10*) ax_gcc_arch="power10";; *) ax_gcc_arch=$cputype ;; esac ax_gcc_arch="$ax_gcc_arch powerpc" ;; + aarch64) + cpuimpl=`grep 'CPU implementer' /proc/cpuinfo 2> /dev/null | cut -d: -f2 | tr -d " " | head -n 1` + cpuarch=`grep 'CPU architecture' /proc/cpuinfo 2> /dev/null | cut -d: -f2 | tr -d " " | head -n 1` + cpuvar=`grep 'CPU variant' /proc/cpuinfo 2> /dev/null | cut -d: -f2 | tr -d " " | head -n 1` + case $cpuimpl in + 0x42) case $cpuarch in + 8) case $cpuvar in + 0x0) ax_gcc_arch="thunderx2t99 vulcan armv8.1-a armv8-a+lse armv8-a native" ;; + esac + ;; + esac + ;; + 0x43) case $cpuarch in + 8) case $cpuvar in + 0x0) ax_gcc_arch="thunderx armv8-a native" ;; + 0x1) ax_gcc_arch="thunderx+lse armv8.1-a armv8-a+lse armv8-a native" ;; + esac + ;; + esac + ;; + esac + ;; esac fi # not cross-compiling fi # guess arch if test "x$ax_gcc_arch" != x -a "x$ax_gcc_arch" != xno; then -for arch in $ax_gcc_arch; do - if test "x[]m4_default([$1],yes)" = xyes; then # if we require portable code - flags="-mtune=$arch" - # -mcpu=$arch and m$arch generate nonportable code on every arch except - # x86. And some other arches (e.g. Alpha) don't accept -mtune. Grrr. - case $host_cpu in i*86|x86_64*) flags="$flags -mcpu=$arch -m$arch";; esac - else - flags="-march=$arch -mcpu=$arch -m$arch" - fi - for flag in $flags; do - AX_CHECK_COMPILE_FLAG($flag, [ax_cv_gcc_archflag=$flag; break]) +if test "x[]m4_default([$1],yes)" = xyes; then # if we require portable code + flag_prefixes="-mtune=" + if test "x$ax_cv_[]_AC_LANG_ABBREV[]_compiler_vendor" = xclang; then flag_prefixes="-march="; fi + # -mcpu=$arch and m$arch generate nonportable code on every arch except + # x86. And some other arches (e.g. Alpha) don't accept -mtune. Grrr. + case $host_cpu in i*86|x86_64*|amd64*) flag_prefixes="$flag_prefixes -mcpu= -m";; esac +else + flag_prefixes="-march= -mcpu= -m" +fi +for flag_prefix in $flag_prefixes; do + for arch in $ax_gcc_arch; do + flag="$flag_prefix$arch" + AX_CHECK_COMPILE_FLAG($flag, [if test "x$ax_cv_[]_AC_LANG_ABBREV[]_compiler_vendor" = xclang; then + if test "x[]m4_default([$1],yes)" = xyes; then + if test "x$flag" = "x-march=$arch"; then flag=-mtune=$arch; fi + fi + fi; ax_cv_gcc_archflag=$flag; break]) done test "x$ax_cv_gcc_archflag" = xunknown || break done diff --git a/native/libffi/m4/ax_gcc_x86_cpuid.m4 b/native/libffi/m4/ax_gcc_x86_cpuid.m4 index 7d46fee021..df954658ee 100644 --- a/native/libffi/m4/ax_gcc_x86_cpuid.m4 +++ b/native/libffi/m4/ax_gcc_x86_cpuid.m4 @@ -1,17 +1,19 @@ # =========================================================================== -# http://www.gnu.org/software/autoconf-archive/ax_gcc_x86_cpuid.html +# https://www.gnu.org/software/autoconf-archive/ax_gcc_x86_cpuid.html # =========================================================================== # # SYNOPSIS # # AX_GCC_X86_CPUID(OP) +# AX_GCC_X86_CPUID_COUNT(OP, COUNT) # # DESCRIPTION # # On Pentium and later x86 processors, with gcc or a compiler that has a # compatible syntax for inline assembly instructions, run a small program # that executes the cpuid instruction with input OP. This can be used to -# detect the CPU type. +# detect the CPU type. AX_GCC_X86_CPUID_COUNT takes an additional COUNT +# parameter that gets passed into register ECX before calling cpuid. # # On output, the values of the eax, ebx, ecx, and edx registers are stored # as hexadecimal strings as "eax:ebx:ecx:edx" in the cache variable @@ -28,6 +30,7 @@ # # Copyright (c) 2008 Steven G. Johnson # Copyright (c) 2008 Matteo Frigo +# Copyright (c) 2015 Michael Petch # # This program is free software: you can redistribute it and/or modify it # under the terms of the GNU General Public License as published by the @@ -40,7 +43,7 @@ # Public License for more details. # # You should have received a copy of the GNU General Public License along -# with this program. If not, see . +# with this program. If not, see . # # As a special exception, the respective Autoconf Macro's copyright owner # gives unlimited permission to copy, distribute and modify the configure @@ -55,18 +58,25 @@ # modified version of the Autoconf Macro, you may extend this special # exception to the GPL to apply to your modified version as well. -#serial 7 +#serial 10 AC_DEFUN([AX_GCC_X86_CPUID], +[AX_GCC_X86_CPUID_COUNT($1, 0) +]) + +AC_DEFUN([AX_GCC_X86_CPUID_COUNT], [AC_REQUIRE([AC_PROG_CC]) AC_LANG_PUSH([C]) AC_CACHE_CHECK(for x86 cpuid $1 output, ax_cv_gcc_x86_cpuid_$1, [AC_RUN_IFELSE([AC_LANG_PROGRAM([#include ], [ - int op = $1, eax, ebx, ecx, edx; + int op = $1, level = $2, eax, ebx, ecx, edx; FILE *f; - __asm__("cpuid" - : "=a" (eax), "=b" (ebx), "=c" (ecx), "=d" (edx) - : "a" (op)); + __asm__ __volatile__ ("xchg %%ebx, %1\n" + "cpuid\n" + "xchg %%ebx, %1\n" + : "=a" (eax), "=r" (ebx), "=c" (ecx), "=d" (edx) + : "a" (op), "2" (level)); + f = fopen("conftest_cpuid", "w"); if (!f) return 1; fprintf(f, "%x:%x:%x:%x\n", eax, ebx, ecx, edx); fclose(f); diff --git a/native/libffi/m4/ax_require_defined.m4 b/native/libffi/m4/ax_require_defined.m4 new file mode 100644 index 0000000000..17c3eab7da --- /dev/null +++ b/native/libffi/m4/ax_require_defined.m4 @@ -0,0 +1,37 @@ +# =========================================================================== +# https://www.gnu.org/software/autoconf-archive/ax_require_defined.html +# =========================================================================== +# +# SYNOPSIS +# +# AX_REQUIRE_DEFINED(MACRO) +# +# DESCRIPTION +# +# AX_REQUIRE_DEFINED is a simple helper for making sure other macros have +# been defined and thus are available for use. This avoids random issues +# where a macro isn't expanded. Instead the configure script emits a +# non-fatal: +# +# ./configure: line 1673: AX_CFLAGS_WARN_ALL: command not found +# +# It's like AC_REQUIRE except it doesn't expand the required macro. +# +# Here's an example: +# +# AX_REQUIRE_DEFINED([AX_CHECK_LINK_FLAG]) +# +# LICENSE +# +# Copyright (c) 2014 Mike Frysinger +# +# Copying and distribution of this file, with or without modification, are +# permitted in any medium without royalty provided the copyright notice +# and this notice are preserved. This file is offered as-is, without any +# warranty. + +#serial 2 + +AC_DEFUN([AX_REQUIRE_DEFINED], [dnl + m4_ifndef([$1], [m4_fatal([macro ]$1[ is not defined; is a m4 file missing?])]) +])dnl AX_REQUIRE_DEFINED diff --git a/native/libffi/make_sunver.pl b/native/libffi/make_sunver.pl new file mode 100644 index 0000000000..8a90b1fea0 --- /dev/null +++ b/native/libffi/make_sunver.pl @@ -0,0 +1,333 @@ +#!/usr/bin/perl -w + +# make_sunver.pl +# +# This script takes at least two arguments, a GNU style version script and +# a list of object and archive files, and generates a corresponding Sun +# style version script as follows: +# +# Each glob pattern, C++ mangled pattern or literal in the input script is +# matched against all global symbols in the input objects, emitting those +# that matched (or nothing if no match was found). +# A comment with the original pattern and its type is left in the output +# file to make it easy to understand the matches. +# +# It uses elfdump when present (native), GNU readelf otherwise. +# It depends on the GNU version of c++filt, since it must understand the +# GNU mangling style. + +use FileHandle; +use IPC::Open2; + +# Enforce C locale. +$ENV{'LC_ALL'} = "C"; +$ENV{'LANG'} = "C"; + +# Input version script, GNU style. +my $symvers = shift; + +########## +# Get all the symbols from the library, match them, and add them to a hash. + +my %sym_hash = (); + +# List of objects and archives to process. +my @OBJECTS = (); + +# List of shared objects to omit from processing. +my @SHAREDOBJS = (); + +# Filter out those input archives that have corresponding shared objects to +# avoid adding all symbols matched in the archive to the output map. +foreach $file (@ARGV) { + if (($so = $file) =~ s/\.a$/.so/ && -e $so) { + printf STDERR "omitted $file -> $so\n"; + push (@SHAREDOBJS, $so); + } else { + push (@OBJECTS, $file); + } +} + +# We need to detect and ignore hidden symbols. Solaris nm can only detect +# this in the harder to parse default output format, and GNU nm not at all, +# so use elfdump -s in the native case and GNU readelf -s otherwise. +# GNU objdump -t cannot be used since it produces a variable number of +# columns. + +# The path to elfdump. +my $elfdump = "/usr/ccs/bin/elfdump"; + +if (-f $elfdump) { + open ELFDUMP,$elfdump.' -s '.(join ' ',@OBJECTS).'|' or die $!; + my $skip_arsym = 0; + + while () { + chomp; + + # Ignore empty lines. + if (/^$/) { + # End of archive symbol table, stop skipping. + $skip_arsym = 0 if $skip_arsym; + next; + } + + # Keep skipping until end of archive symbol table. + next if ($skip_arsym); + + # Ignore object name header for individual objects and archives. + next if (/:$/); + + # Ignore table header lines. + next if (/^Symbol Table Section:/); + next if (/index.*value.*size/); + + # Start of archive symbol table: start skipping. + if (/^Symbol Table: \(archive/) { + $skip_arsym = 1; + next; + } + + # Split table. + (undef, undef, undef, undef, $bind, $oth, undef, $shndx, $name) = split; + + # Error out for unknown input. + die "unknown input line:\n$_" unless defined($bind); + + # Ignore local symbols. + next if ($bind eq "LOCL"); + # Ignore hidden symbols. + next if ($oth eq "H"); + # Ignore undefined symbols. + next if ($shndx eq "UNDEF"); + # Error out for unhandled cases. + if ($bind !~ /^(GLOB|WEAK)/ or $oth ne "D") { + die "unhandled symbol:\n$_"; + } + + # Remember symbol. + $sym_hash{$name}++; + } + close ELFDUMP or die "$elfdump error"; +} else { + open READELF, 'readelf -s -W '.(join ' ',@OBJECTS).'|' or die $!; + # Process each symbol. + while () { + chomp; + + # Ignore empty lines. + next if (/^$/); + + # Ignore object name header. + next if (/^File: .*$/); + + # Ignore table header lines. + next if (/^Symbol table.*contains.*:/); + next if (/Num:.*Value.*Size/); + + # Split table. + (undef, undef, undef, undef, $bind, $vis, $ndx, $name) = split; + + # Error out for unknown input. + die "unknown input line:\n$_" unless defined($bind); + + # Ignore local symbols. + next if ($bind eq "LOCAL"); + # Ignore hidden symbols. + next if ($vis eq "HIDDEN"); + # Ignore undefined symbols. + next if ($ndx eq "UND"); + # Error out for unhandled cases. + if ($bind !~ /^(GLOBAL|WEAK)/ or $vis ne "DEFAULT") { + die "unhandled symbol:\n$_"; + } + + # Remember symbol. + $sym_hash{$name}++; + } + close READELF or die "readelf error"; +} + +########## +# The various types of glob patterns. +# +# A glob pattern that is to be applied to the demangled name: 'cxx'. +# A glob patterns that applies directly to the name in the .o files: 'glob'. +# This pattern is ignored; used for local variables (usually just '*'): 'ign'. + +# The type of the current pattern. +my $glob = 'glob'; + +# We're currently inside `extern "C++"', which Sun ld doesn't understand. +my $in_extern = 0; + +# The c++filt command to use. This *must* be GNU c++filt; the Sun Studio +# c++filt doesn't handle the GNU mangling style. +my $cxxfilt = $ENV{'CXXFILT'} || "c++filt"; + +# The current version name. +my $current_version = ""; + +# Was there any attempt to match a symbol to this version? +my $matches_attempted; + +# The number of versions which matched this symbol. +my $matched_symbols; + +open F,$symvers or die $!; + +# Print information about generating this file +print "# This file was generated by make_sunver.pl. DO NOT EDIT!\n"; +print "# It was generated by:\n"; +printf "# %s %s %s\n", $0, $symvers, (join ' ',@ARGV); +printf "# Omitted archives with corresponding shared libraries: %s\n", + (join ' ', @SHAREDOBJS) if $#SHAREDOBJS >= 0; +print "#\n\n"; + +while () { + # Lines of the form '};' + if (/^([ \t]*)(\}[ \t]*;[ \t]*)$/) { + $glob = 'glob'; + if ($in_extern) { + $in_extern--; + print "$1##$2\n"; + } else { + print; + } + next; + } + + # Lines of the form '} SOME_VERSION_NAME_1.0;' + if (/^[ \t]*\}[ \tA-Z0-9_.a-z]+;[ \t]*$/) { + $glob = 'glob'; + # We tried to match symbols agains this version, but none matched. + # Emit dummy hidden symbol to avoid marking this version WEAK. + if ($matches_attempted && $matched_symbols == 0) { + print " hidden:\n"; + print " .force_WEAK_off_$current_version = DATA S0x0 V0x0;\n"; + } + print; next; + } + + # Comment and blank lines + if (/^[ \t]*\#/) { print; next; } + if (/^[ \t]*$/) { print; next; } + + # Lines of the form '{' + if (/^([ \t]*){$/) { + if ($in_extern) { + print "$1##{\n"; + } else { + print; + } + next; + } + + # Lines of the form 'SOME_VERSION_NAME_1.1 {' + if (/^([A-Z0-9_.]+)[ \t]+{$/) { + # Record version name. + $current_version = $1; + # Reset match attempts, #matched symbols for this version. + $matches_attempted = 0; + $matched_symbols = 0; + print; + next; + } + + # Ignore 'global:' + if (/^[ \t]*global:$/) { print; next; } + + # After 'local:', globs should be ignored, they won't be exported. + if (/^[ \t]*local:$/) { + $glob = 'ign'; + print; + next; + } + + # After 'extern "C++"', globs are C++ patterns + if (/^([ \t]*)(extern \"C\+\+\"[ \t]*)$/) { + $in_extern++; + $glob = 'cxx'; + # Need to comment, Sun ld cannot handle this. + print "$1##$2\n"; next; + } + + # Chomp newline now we're done with passing through the input file. + chomp; + + # Catch globs. Note that '{}' is not allowed in globs by this script, + # so only '*' and '[]' are available. + if (/^([ \t]*)([^ \t;{}#]+);?[ \t]*$/) { + my $ws = $1; + my $ptn = $2; + # Turn the glob into a regex by replacing '*' with '.*', '?' with '.'. + # Keep $ptn so we can still print the original form. + ($pattern = $ptn) =~ s/\*/\.\*/g; + $pattern =~ s/\?/\./g; + + if ($glob eq 'ign') { + # We're in a local: * section; just continue. + print "$_\n"; + next; + } + + # Print the glob commented for human readers. + print "$ws##$ptn ($glob)\n"; + # We tried to match a symbol to this version. + $matches_attempted++; + + if ($glob eq 'glob') { + my %ptn_syms = (); + + # Match ptn against symbols in %sym_hash. + foreach my $sym (keys %sym_hash) { + # Maybe it matches one of the patterns based on the symbol in + # the .o file. + $ptn_syms{$sym}++ if ($sym =~ /^$pattern$/); + } + + foreach my $sym (sort keys(%ptn_syms)) { + $matched_symbols++; + print "$ws$sym;\n"; + } + } elsif ($glob eq 'cxx') { + my %dem_syms = (); + + # Verify that we're actually using GNU c++filt. Other versions + # most likely cannot handle GNU style symbol mangling. + my $cxxout = `$cxxfilt --version 2>&1`; + $cxxout =~ m/GNU/ or die "$0 requires GNU c++filt to function"; + + # Talk to c++filt through a pair of file descriptors. + # Need to start a fresh instance per pattern, otherwise the + # process grows to 500+ MB. + my $pid = open2(*FILTIN, *FILTOUT, $cxxfilt) or die $!; + + # Match ptn against symbols in %sym_hash. + foreach my $sym (keys %sym_hash) { + # No? Well, maybe its demangled form matches one of those + # patterns. + printf FILTOUT "%s\n",$sym; + my $dem = ; + chomp $dem; + $dem_syms{$sym}++ if ($dem =~ /^$pattern$/); + } + + close FILTOUT or die "c++filt error"; + close FILTIN or die "c++filt error"; + # Need to wait for the c++filt process to avoid lots of zombies. + waitpid $pid, 0; + + foreach my $sym (sort keys(%dem_syms)) { + $matched_symbols++; + print "$ws$sym;\n"; + } + } else { + # No? Well, then ignore it. + } + next; + } + # Important sanity check. This script can't handle lots of formats + # that GNU ld can, so be sure to error out if one is seen! + die "strange line `$_'"; +} +close F; diff --git a/native/libffi/msvc_build/aarch64/Ffi_staticLib.sln b/native/libffi/msvc_build/aarch64/Ffi_staticLib.sln new file mode 100644 index 0000000000..d9119dfac2 --- /dev/null +++ b/native/libffi/msvc_build/aarch64/Ffi_staticLib.sln @@ -0,0 +1,33 @@ + +Microsoft Visual Studio Solution File, Format Version 12.00 +# Visual Studio 15 +VisualStudioVersion = 15.0.28302.56 +MinimumVisualStudioVersion = 10.0.40219.1 +Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "Ffi_staticLib_arm64", "Ffi_staticLib.vcxproj", "{115502C0-BE05-4767-BF19-5C87D805FAD6}" +EndProject +Global + GlobalSection(SolutionConfigurationPlatforms) = preSolution + Debug|ARM64 = Debug|ARM64 + Debug|x64 = Debug|x64 + Debug|x86 = Debug|x86 + Release|ARM64 = Release|ARM64 + Release|x64 = Release|x64 + Release|x86 = Release|x86 + EndGlobalSection + GlobalSection(ProjectConfigurationPlatforms) = postSolution + {115502C0-BE05-4767-BF19-5C87D805FAD6}.Debug|ARM64.ActiveCfg = Debug|ARM64 + {115502C0-BE05-4767-BF19-5C87D805FAD6}.Debug|ARM64.Build.0 = Debug|ARM64 + {115502C0-BE05-4767-BF19-5C87D805FAD6}.Debug|x64.ActiveCfg = Debug|ARM64 + {115502C0-BE05-4767-BF19-5C87D805FAD6}.Debug|x86.ActiveCfg = Debug|ARM64 + {115502C0-BE05-4767-BF19-5C87D805FAD6}.Release|ARM64.ActiveCfg = Release|ARM64 + {115502C0-BE05-4767-BF19-5C87D805FAD6}.Release|ARM64.Build.0 = Release|ARM64 + {115502C0-BE05-4767-BF19-5C87D805FAD6}.Release|x64.ActiveCfg = Release|ARM64 + {115502C0-BE05-4767-BF19-5C87D805FAD6}.Release|x86.ActiveCfg = Release|ARM64 + EndGlobalSection + GlobalSection(SolutionProperties) = preSolution + HideSolutionNode = FALSE + EndGlobalSection + GlobalSection(ExtensibilityGlobals) = postSolution + SolutionGuid = {241C54C7-20DD-4897-9376-E6B6D1B43BD5} + EndGlobalSection +EndGlobal diff --git a/native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj b/native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj new file mode 100644 index 0000000000..3187699ca1 --- /dev/null +++ b/native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj @@ -0,0 +1,130 @@ + + + + + Debug + ARM64 + + + Release + ARM64 + + + + 15.0 + {115502C0-BE05-4767-BF19-5C87D805FAD6} + Win32Proj + FfistaticLib + 10.0.17763.0 + Ffi_staticLib_arm64 + + + + StaticLibrary + true + v141 + Unicode + + + StaticLibrary + false + v141 + true + Unicode + + + + + + + + + + + + + + + true + + + false + + + + NotUsing + Level3 + Disabled + true + FFI_BUILDING_DLL;_DEBUG;_LIB;USE_DL_PREFIX;ARM64;_M_ARM64;NDEBUG;%(PreprocessorDefinitions) + true + ..\..\include;.\aarch64_include;..\..\src\aarch64;%(AdditionalIncludeDirectories) + false + true + + + false + + + Windows + true + + + + + NotUsing + Level3 + MaxSpeed + true + true + true + FFI_BUILDING_DLL;USE_DL_PREFIX;ARM64;NDEBUG;_LIB;%(PreprocessorDefinitions) + true + ..\..\include;.\aarch64_include;..\..\src\aarch64;%(AdditionalIncludeDirectories) + true + Speed + true + ..\..\src;..\..\src\aarch64;%(AdditionalUsingDirectories) + + + Windows + true + true + true + + + true + + + + + + + + + + + + + + + + + + + + + + + cl /FA /EP /nologo /I"..\..\include" /I".\aarch64_include" /I"..\..\src\aarch64" "%(FullPath)" > $(IntDir)win64_armasm.i + armasm64 $(IntDir)win64_armasm.i /I"src\" /I"..\..\include" /I"..\..\src\aarch64" -o "$(IntDir)win64_armasm.obj" + + win64_armasm.obj;%(Outputs) + + + + + + \ No newline at end of file diff --git a/native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj.filters b/native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj.filters new file mode 100644 index 0000000000..1f8c6e13ef --- /dev/null +++ b/native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj.filters @@ -0,0 +1,57 @@ + + + + + {4FC737F1-C7A5-4376-A066-2A32D752A2FF} + cpp;c;cc;cxx;def;odl;idl;hpj;bat;asm;asmx + + + {93995380-89BD-4b04-88EB-625FBE52EBFB} + h;hh;hpp;hxx;hm;inl;inc;ipp;xsd + + + {67DA6AB6-F800-4c08-8B7A-83BB121AAD01} + rc;ico;cur;bmp;dlg;rc2;rct;bin;rgs;gif;jpg;jpeg;jpe;resx;tiff;tif;png;wav;mfcribbon-ms + + + + + Header Files + + + Header Files + + + Header Files + + + Header Files + + + Header Files + + + Header Files + + + + + Source Files + + + Source Files + + + Source Files + + + Source Files + + + Source Files + + + + + + \ No newline at end of file diff --git a/native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj.user b/native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj.user new file mode 100644 index 0000000000..be25078707 --- /dev/null +++ b/native/libffi/msvc_build/aarch64/Ffi_staticLib.vcxproj.user @@ -0,0 +1,4 @@ + + + + \ No newline at end of file diff --git a/native/libffi/msvc_build/aarch64/aarch64_include/ffi.h b/native/libffi/msvc_build/aarch64/aarch64_include/ffi.h new file mode 100644 index 0000000000..02f26a2f0f --- /dev/null +++ b/native/libffi/msvc_build/aarch64/aarch64_include/ffi.h @@ -0,0 +1,511 @@ +/* -----------------------------------------------------------------*-C-*- + libffi 3.3-rc0 - Copyright (c) 2011, 2014 Anthony Green + - Copyright (c) 1996-2003, 2007, 2008 Red Hat, Inc. + + Permission is hereby granted, free of charge, to any person + obtaining a copy of this software and associated documentation + files (the ``Software''), to deal in the Software without + restriction, including without limitation the rights to use, copy, + modify, merge, publish, distribute, sublicense, and/or sell copies + of the Software, and to permit persons to whom the Software is + furnished to do so, subject to the following conditions: + + The above copyright notice and this permission notice shall be + included in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + + ----------------------------------------------------------------------- */ + +/* ------------------------------------------------------------------- + Most of the API is documented in doc/libffi.texi. + + The raw API is designed to bypass some of the argument packing and + unpacking on architectures for which it can be avoided. Routines + are provided to emulate the raw API if the underlying platform + doesn't allow faster implementation. + + More details on the raw API can be found in: + + http://gcc.gnu.org/ml/java/1999-q3/msg00138.html + + and + + http://gcc.gnu.org/ml/java/1999-q3/msg00174.html + -------------------------------------------------------------------- */ + +#ifndef LIBFFI_H +#define LIBFFI_H + +#ifdef __cplusplus +extern "C" { +#endif + +/* Specify which architecture libffi is configured for. */ +#ifndef AARCH64 +#define AARCH64 +#endif + +/* ---- System configuration information --------------------------------- */ + +#include + +#ifndef LIBFFI_ASM + +#if defined(_MSC_VER) && !defined(__clang__) +#define __attribute__(X) +#endif + +#include +#include + +/* LONG_LONG_MAX is not always defined (not if STRICT_ANSI, for example). + But we can find it either under the correct ANSI name, or under GNU + C's internal name. */ + +#define FFI_64_BIT_MAX 9223372036854775807 + +#ifdef LONG_LONG_MAX +# define FFI_LONG_LONG_MAX LONG_LONG_MAX +#else +# ifdef LLONG_MAX +# define FFI_LONG_LONG_MAX LLONG_MAX +# ifdef _AIX52 /* or newer has C99 LLONG_MAX */ +# undef FFI_64_BIT_MAX +# define FFI_64_BIT_MAX 9223372036854775807LL +# endif /* _AIX52 or newer */ +# else +# ifdef __GNUC__ +# define FFI_LONG_LONG_MAX __LONG_LONG_MAX__ +# endif +# ifdef _AIX /* AIX 5.1 and earlier have LONGLONG_MAX */ +# ifndef __PPC64__ +# if defined (__IBMC__) || defined (__IBMCPP__) +# define FFI_LONG_LONG_MAX LONGLONG_MAX +# endif +# endif /* __PPC64__ */ +# undef FFI_64_BIT_MAX +# define FFI_64_BIT_MAX 9223372036854775807LL +# endif +# endif +#endif + +/* The closure code assumes that this works on pointers, i.e. a size_t + can hold a pointer. */ + +typedef struct _ffi_type +{ + size_t size; + unsigned short alignment; + unsigned short type; + struct _ffi_type **elements; +} ffi_type; + +/* Need minimal decorations for DLLs to work on Windows. GCC has + autoimport and autoexport. Always mark externally visible symbols + as dllimport for MSVC clients, even if it means an extra indirection + when using the static version of the library. + Besides, as a workaround, they can define FFI_BUILDING if they + *know* they are going to link with the static library. */ +#if defined _MSC_VER +# if defined FFI_BUILDING_DLL /* Building libffi.DLL with msvcc.sh */ +# define FFI_API __declspec(dllexport) +# elif !defined FFI_BUILDING /* Importing libffi.DLL */ +# define FFI_API __declspec(dllimport) +# else /* Building/linking static library */ +# define FFI_API +# endif +#else +# define FFI_API +#endif + +/* The externally visible type declarations also need the MSVC DLL + decorations, or they will not be exported from the object file. */ +#if defined LIBFFI_HIDE_BASIC_TYPES +# define FFI_EXTERN FFI_API +#else +# define FFI_EXTERN extern FFI_API +#endif + +#ifndef LIBFFI_HIDE_BASIC_TYPES +#if SCHAR_MAX == 127 +# define ffi_type_uchar ffi_type_uint8 +# define ffi_type_schar ffi_type_sint8 +#else + #error "char size not supported" +#endif + +#if SHRT_MAX == 32767 +# define ffi_type_ushort ffi_type_uint16 +# define ffi_type_sshort ffi_type_sint16 +#elif SHRT_MAX == 2147483647 +# define ffi_type_ushort ffi_type_uint32 +# define ffi_type_sshort ffi_type_sint32 +#else + #error "short size not supported" +#endif + +#if INT_MAX == 32767 +# define ffi_type_uint ffi_type_uint16 +# define ffi_type_sint ffi_type_sint16 +#elif INT_MAX == 2147483647 +# define ffi_type_uint ffi_type_uint32 +# define ffi_type_sint ffi_type_sint32 +#elif INT_MAX == 9223372036854775807 +# define ffi_type_uint ffi_type_uint64 +# define ffi_type_sint ffi_type_sint64 +#else + #error "int size not supported" +#endif + +#if LONG_MAX == 2147483647 +# if FFI_LONG_LONG_MAX != FFI_64_BIT_MAX + #error "no 64-bit data type supported" +# endif +#elif LONG_MAX != FFI_64_BIT_MAX + #error "long size not supported" +#endif + +#if LONG_MAX == 2147483647 +# define ffi_type_ulong ffi_type_uint32 +# define ffi_type_slong ffi_type_sint32 +#elif LONG_MAX == FFI_64_BIT_MAX +# define ffi_type_ulong ffi_type_uint64 +# define ffi_type_slong ffi_type_sint64 +#else + #error "long size not supported" +#endif + +/* These are defined in types.c. */ +FFI_EXTERN ffi_type ffi_type_void; +FFI_EXTERN ffi_type ffi_type_uint8; +FFI_EXTERN ffi_type ffi_type_sint8; +FFI_EXTERN ffi_type ffi_type_uint16; +FFI_EXTERN ffi_type ffi_type_sint16; +FFI_EXTERN ffi_type ffi_type_uint32; +FFI_EXTERN ffi_type ffi_type_sint32; +FFI_EXTERN ffi_type ffi_type_uint64; +FFI_EXTERN ffi_type ffi_type_sint64; +FFI_EXTERN ffi_type ffi_type_float; +FFI_EXTERN ffi_type ffi_type_double; +FFI_EXTERN ffi_type ffi_type_pointer; + +#ifndef _M_ARM64 +FFI_EXTERN ffi_type ffi_type_longdouble; +#else +#define ffi_type_longdouble ffi_type_double +#endif + +#ifdef FFI_TARGET_HAS_COMPLEX_TYPE +FFI_EXTERN ffi_type ffi_type_complex_float; +FFI_EXTERN ffi_type ffi_type_complex_double; +#if 1 +FFI_EXTERN ffi_type ffi_type_complex_longdouble; +#else +#define ffi_type_complex_longdouble ffi_type_complex_double +#endif +#endif +#endif /* LIBFFI_HIDE_BASIC_TYPES */ + +typedef enum { + FFI_OK = 0, + FFI_BAD_TYPEDEF, + FFI_BAD_ABI +} ffi_status; + +typedef struct { + ffi_abi abi; + unsigned nargs; + ffi_type **arg_types; + ffi_type *rtype; + unsigned bytes; + unsigned flags; +#ifdef FFI_EXTRA_CIF_FIELDS + FFI_EXTRA_CIF_FIELDS; +#endif +} ffi_cif; + +/* ---- Definitions for the raw API -------------------------------------- */ + +#ifndef FFI_SIZEOF_ARG +# if LONG_MAX == 2147483647 +# define FFI_SIZEOF_ARG 4 +# elif LONG_MAX == FFI_64_BIT_MAX +# define FFI_SIZEOF_ARG 8 +# endif +#endif + +#ifndef FFI_SIZEOF_JAVA_RAW +# define FFI_SIZEOF_JAVA_RAW FFI_SIZEOF_ARG +#endif + +typedef union { + ffi_sarg sint; + ffi_arg uint; + float flt; + char data[FFI_SIZEOF_ARG]; + void* ptr; +} ffi_raw; + +#if FFI_SIZEOF_JAVA_RAW == 4 && FFI_SIZEOF_ARG == 8 +/* This is a special case for mips64/n32 ABI (and perhaps others) where + sizeof(void *) is 4 and FFI_SIZEOF_ARG is 8. */ +typedef union { + signed int sint; + unsigned int uint; + float flt; + char data[FFI_SIZEOF_JAVA_RAW]; + void* ptr; +} ffi_java_raw; +#else +typedef ffi_raw ffi_java_raw; +#endif + + +FFI_API +void ffi_raw_call (ffi_cif *cif, + void (*fn)(void), + void *rvalue, + ffi_raw *avalue); + +FFI_API void ffi_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_raw *raw); +FFI_API void ffi_raw_to_ptrarray (ffi_cif *cif, ffi_raw *raw, void **args); +FFI_API size_t ffi_raw_size (ffi_cif *cif); + +/* This is analogous to the raw API, except it uses Java parameter + packing, even on 64-bit machines. I.e. on 64-bit machines longs + and doubles are followed by an empty 64-bit word. */ + +FFI_API +void ffi_java_raw_call (ffi_cif *cif, + void (*fn)(void), + void *rvalue, + ffi_java_raw *avalue); + +FFI_API +void ffi_java_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_java_raw *raw); +FFI_API +void ffi_java_raw_to_ptrarray (ffi_cif *cif, ffi_java_raw *raw, void **args); +FFI_API +size_t ffi_java_raw_size (ffi_cif *cif); + +/* ---- Definitions for closures ----------------------------------------- */ + +#if FFI_CLOSURES + +#ifdef _MSC_VER +__declspec(align(8)) +#endif +typedef struct { +#if 0 + void *trampoline_table; + void *trampoline_table_entry; +#else + char tramp[FFI_TRAMPOLINE_SIZE]; +#endif + ffi_cif *cif; + void (*fun)(ffi_cif*,void*,void**,void*); + void *user_data; +} ffi_closure +#ifdef __GNUC__ + __attribute__((aligned (8))) +#endif + ; + +#ifndef __GNUC__ +# ifdef __sgi +# pragma pack 0 +# endif +#endif + +FFI_API void *ffi_closure_alloc (size_t size, void **code); +FFI_API void ffi_closure_free (void *); + +FFI_API ffi_status +ffi_prep_closure (ffi_closure*, + ffi_cif *, + void (*fun)(ffi_cif*,void*,void**,void*), + void *user_data) +#if defined(__GNUC__) && (((__GNUC__ * 100) + __GNUC_MINOR__) >= 405) + __attribute__((deprecated ("use ffi_prep_closure_loc instead"))) +#elif defined(__GNUC__) && __GNUC__ >= 3 + __attribute__((deprecated)) +#endif + ; + +FFI_API ffi_status +ffi_prep_closure_loc (ffi_closure*, + ffi_cif *, + void (*fun)(ffi_cif*,void*,void**,void*), + void *user_data, + void*codeloc); + +#ifdef __sgi +# pragma pack 8 +#endif +typedef struct { +#if 0 + void *trampoline_table; + void *trampoline_table_entry; +#else + char tramp[FFI_TRAMPOLINE_SIZE]; +#endif + ffi_cif *cif; + +#if !FFI_NATIVE_RAW_API + + /* If this is enabled, then a raw closure has the same layout + as a regular closure. We use this to install an intermediate + handler to do the transaltion, void** -> ffi_raw*. */ + + void (*translate_args)(ffi_cif*,void*,void**,void*); + void *this_closure; + +#endif + + void (*fun)(ffi_cif*,void*,ffi_raw*,void*); + void *user_data; + +} ffi_raw_closure; + +typedef struct { +#if 0 + void *trampoline_table; + void *trampoline_table_entry; +#else + char tramp[FFI_TRAMPOLINE_SIZE]; +#endif + + ffi_cif *cif; + +#if !FFI_NATIVE_RAW_API + + /* If this is enabled, then a raw closure has the same layout + as a regular closure. We use this to install an intermediate + handler to do the translation, void** -> ffi_raw*. */ + + void (*translate_args)(ffi_cif*,void*,void**,void*); + void *this_closure; + +#endif + + void (*fun)(ffi_cif*,void*,ffi_java_raw*,void*); + void *user_data; + +} ffi_java_raw_closure; + +FFI_API ffi_status +ffi_prep_raw_closure (ffi_raw_closure*, + ffi_cif *cif, + void (*fun)(ffi_cif*,void*,ffi_raw*,void*), + void *user_data); + +FFI_API ffi_status +ffi_prep_raw_closure_loc (ffi_raw_closure*, + ffi_cif *cif, + void (*fun)(ffi_cif*,void*,ffi_raw*,void*), + void *user_data, + void *codeloc); + +FFI_API ffi_status +ffi_prep_java_raw_closure (ffi_java_raw_closure*, + ffi_cif *cif, + void (*fun)(ffi_cif*,void*,ffi_java_raw*,void*), + void *user_data); + +FFI_API ffi_status +ffi_prep_java_raw_closure_loc (ffi_java_raw_closure*, + ffi_cif *cif, + void (*fun)(ffi_cif*,void*,ffi_java_raw*,void*), + void *user_data, + void *codeloc); + +#endif /* FFI_CLOSURES */ + +#if FFI_GO_CLOSURES + +typedef struct { + void *tramp; + ffi_cif *cif; + void (*fun)(ffi_cif*,void*,void**,void*); +} ffi_go_closure; + +FFI_API ffi_status ffi_prep_go_closure (ffi_go_closure*, ffi_cif *, + void (*fun)(ffi_cif*,void*,void**,void*)); + +FFI_API void ffi_call_go (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure); + +#endif /* FFI_GO_CLOSURES */ + +/* ---- Public interface definition -------------------------------------- */ + +FFI_API +ffi_status ffi_prep_cif(ffi_cif *cif, + ffi_abi abi, + unsigned int nargs, + ffi_type *rtype, + ffi_type **atypes); + +FFI_API +ffi_status ffi_prep_cif_var(ffi_cif *cif, + ffi_abi abi, + unsigned int nfixedargs, + unsigned int ntotalargs, + ffi_type *rtype, + ffi_type **atypes); + +FFI_API +void ffi_call(ffi_cif *cif, + void (*fn)(void), + void *rvalue, + void **avalue); + +FFI_API +ffi_status ffi_get_struct_offsets (ffi_abi abi, ffi_type *struct_type, + size_t *offsets); + +/* Useful for eliminating compiler warnings. */ +#define FFI_FN(f) ((void (*)(void))f) + +/* ---- Definitions shared with assembly code ---------------------------- */ + +#endif + +/* If these change, update src/mips/ffitarget.h. */ +#define FFI_TYPE_VOID 0 +#define FFI_TYPE_INT 1 +#define FFI_TYPE_FLOAT 2 +#define FFI_TYPE_DOUBLE 3 +#ifndef _M_ARM64 +#define FFI_TYPE_LONGDOUBLE 4 +#else +#define FFI_TYPE_LONGDOUBLE FFI_TYPE_DOUBLE +#endif +#define FFI_TYPE_UINT8 5 +#define FFI_TYPE_SINT8 6 +#define FFI_TYPE_UINT16 7 +#define FFI_TYPE_SINT16 8 +#define FFI_TYPE_UINT32 9 +#define FFI_TYPE_SINT32 10 +#define FFI_TYPE_UINT64 11 +#define FFI_TYPE_SINT64 12 +#define FFI_TYPE_STRUCT 13 +#define FFI_TYPE_POINTER 14 +#define FFI_TYPE_COMPLEX 15 +/* This should always refer to the last type code (for sanity checks). */ +#define FFI_TYPE_LAST FFI_TYPE_COMPLEX + + +#ifdef __cplusplus +} +#endif + +#endif diff --git a/native/libffi/msvc_build/aarch64/aarch64_include/fficonfig.h b/native/libffi/msvc_build/aarch64/aarch64_include/fficonfig.h new file mode 100644 index 0000000000..b3d89b03ec --- /dev/null +++ b/native/libffi/msvc_build/aarch64/aarch64_include/fficonfig.h @@ -0,0 +1,219 @@ +/* fficonfig.h. Generated from fficonfig.h.in by configure. */ +/* fficonfig.h.in. Generated from configure.ac by autoheader. */ + +/* Define if building universal (internal helper macro) */ +/* #undef AC_APPLE_UNIVERSAL_BUILD */ + +/* Define to one of `_getb67', `GETB67', `getb67' for Cray-2 and Cray-YMP + systems. This function is required for `alloca.c' support on those systems. + */ +/* #undef CRAY_STACKSEG_END */ + +/* Define to 1 if using `alloca.c'. */ +/* #undef C_ALLOCA */ + +/* Define to the flags needed for the .section .eh_frame directive. */ +#define EH_FRAME_FLAGS "a" + +/* Define this if you want extra debugging. */ +/* #undef FFI_DEBUG */ + +/* Cannot use PROT_EXEC on this target, so, we revert to alternative means */ +/* #undef FFI_EXEC_TRAMPOLINE_TABLE */ + +/* Define this if you want to enable pax emulated trampolines */ +/* #undef FFI_MMAP_EXEC_EMUTRAMP_PAX */ + +/* Cannot use malloc on this target, so, we revert to alternative means */ +/* #undef FFI_MMAP_EXEC_WRIT */ + +/* Define this if you do not want support for the raw API. */ +/* #undef FFI_NO_RAW_API */ + +/* Define this if you do not want support for aggregate types. */ +/* #undef FFI_NO_STRUCTS */ + +/* Define to 1 if you have `alloca', as a function or macro. */ +#define HAVE_ALLOCA 1 + +/* Define to 1 if you have and it should be used (not on Ultrix). + */ +/*#define HAVE_ALLOCA_H 1 */ + +/* Define if your assembler supports .cfi_* directives. */ +#define HAVE_AS_CFI_PSEUDO_OP 1 + +/* Define if your assembler supports .register. */ +/* #undef HAVE_AS_REGISTER_PSEUDO_OP */ + +/* Define if the compiler uses zarch features. */ +/* #undef HAVE_AS_S390_ZARCH */ + +/* Define if your assembler and linker support unaligned PC relative relocs. + */ +/* #undef HAVE_AS_SPARC_UA_PCREL */ + +/* Define if your assembler supports unwind section type. */ +/* #undef HAVE_AS_X86_64_UNWIND_SECTION_TYPE */ + +/* Define if your assembler supports PC relative relocs. */ +/* #undef HAVE_AS_X86_PCREL */ + +/* Define to 1 if you have the header file. */ +#define HAVE_DLFCN_H 1 + +/* Define if __attribute__((visibility("hidden"))) is supported. */ +#define HAVE_HIDDEN_VISIBILITY_ATTRIBUTE 1 + +/* Define to 1 if you have the header file. */ +#define HAVE_INTTYPES_H 1 + +/* Define if you have the long double type and it is bigger than a double */ +#define HAVE_LONG_DOUBLE 1 + +/* Define if you support more than one size of the long double type */ +/* #undef HAVE_LONG_DOUBLE_VARIANT */ + +/* Define to 1 if you have the `memcpy' function. */ +#define HAVE_MEMCPY 1 + +/* Define to 1 if you have the header file. */ +#define HAVE_MEMORY_H 1 + +/* Define to 1 if you have the `mkostemp' function. */ +#define HAVE_MKOSTEMP 1 + +/* Define to 1 if you have the `mmap' function. */ +#define HAVE_MMAP 1 + +/* Define if mmap with MAP_ANON(YMOUS) works. */ +#define HAVE_MMAP_ANON 1 + +/* Define if mmap of /dev/zero works. */ +#define HAVE_MMAP_DEV_ZERO 1 + +/* Define if read-only mmap of a plain file works. */ +#define HAVE_MMAP_FILE 1 + +/* Define if .eh_frame sections should be read-only. */ +#define HAVE_RO_EH_FRAME 1 + +/* Define to 1 if you have the header file. */ +#define HAVE_STDINT_H 1 + +/* Define to 1 if you have the header file. */ +//#define HAVE_STDLIB_H 0 +#define LACKS_STDLIB_H 1 +/* Define to 1 if you have the header file. */ +#define HAVE_STRINGS_H 1 + +/* Define to 1 if you have the header file. */ +#define HAVE_STRING_H 1 + +/* Define to 1 if you have the header file. */ +#define HAVE_SYS_MMAN_H 1 + +/* Define to 1 if you have the header file. */ +#define HAVE_SYS_STAT_H 1 + +/* Define to 1 if you have the header file. */ +#define HAVE_SYS_TYPES_H 1 + +/* Define to 1 if you have the header file. */ +#define HAVE_UNISTD_H 1 + +/* Define to 1 if GNU symbol versioning is used for libatomic. */ +#define LIBFFI_GNU_SYMBOL_VERSIONING 1 + +/* Define to the sub-directory in which libtool stores uninstalled libraries. + */ +#define LT_OBJDIR ".libs/" + +/* Define to 1 if your C compiler doesn't accept -c and -o together. */ +/* #undef NO_MINUS_C_MINUS_O */ + +/* Name of package */ +#define PACKAGE "libffi" + +/* Define to the address where bug reports for this package should be sent. */ +#define PACKAGE_BUGREPORT "http://github.com/libffi/libffi/issues" + +/* Define to the full name of this package. */ +#define PACKAGE_NAME "libffi" + +/* Define to the full name and version of this package. */ +#define PACKAGE_STRING "libffi 3.3-rc0" + +/* Define to the one symbol short name of this package. */ +#define PACKAGE_TARNAME "libffi" + +/* Define to the home page for this package. */ +#define PACKAGE_URL "" + +/* Define to the version of this package. */ +#define PACKAGE_VERSION "3.3-rc0" + +/* The size of `double', as computed by sizeof. */ +#define SIZEOF_DOUBLE 8 + +/* The size of `long double', as computed by sizeof. */ +#define SIZEOF_LONG_DOUBLE 8 + +/* The size of `size_t', as computed by sizeof. */ +#define SIZEOF_SIZE_T 8 + +/* If using the C implementation of alloca, define if you know the + direction of stack growth for your system; otherwise it will be + automatically deduced at runtime. + STACK_DIRECTION > 0 => grows toward higher addresses + STACK_DIRECTION < 0 => grows toward lower addresses + STACK_DIRECTION = 0 => direction of growth unknown */ +/* #undef STACK_DIRECTION */ + +/* Define to 1 if you have the ANSI C header files. */ +#define STDC_HEADERS 1 + +/* Define if symbols are underscored. */ +/* #undef SYMBOL_UNDERSCORE */ + +/* Define this if you are using Purify and want to suppress spurious messages. + */ +/* #undef USING_PURIFY */ + +/* Version number of package */ +#define VERSION "3.3-rc0" + +/* Define WORDS_BIGENDIAN to 1 if your processor stores words with the most + significant byte first (like Motorola and SPARC, unlike Intel). */ +#if defined AC_APPLE_UNIVERSAL_BUILD +# if defined __BIG_ENDIAN__ +# define WORDS_BIGENDIAN 1 +# endif +#else +# ifndef WORDS_BIGENDIAN +/* # undef WORDS_BIGENDIAN */ +# endif +#endif + +/* Define to `unsigned int' if does not define. */ +/* #undef size_t */ + + +#ifdef HAVE_HIDDEN_VISIBILITY_ATTRIBUTE +#ifdef LIBFFI_ASM +#ifdef __APPLE__ +#define FFI_HIDDEN(name) .private_extern name +#else +#define FFI_HIDDEN(name) .hidden name +#endif +#else +#define FFI_HIDDEN __attribute__ ((visibility ("hidden"))) +#endif +#else +#ifdef LIBFFI_ASM +#define FFI_HIDDEN(name) +#else +#define FFI_HIDDEN +#endif +#endif + diff --git a/native/libffi/msvcc.sh b/native/libffi/msvcc.sh index 984722773c..556351ba89 100755 --- a/native/libffi/msvcc.sh +++ b/native/libffi/msvcc.sh @@ -44,17 +44,32 @@ args_orig=$@ args="-nologo -W3" +<<<<<<< HEAD +======= +linkargs= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 static_crt= debug_crt= cl="cl" ml="ml" safeseh="-safeseh" output= +libpaths= +libversion=7 +verbose= while [ $# -gt 0 ] do case $1 in + --verbose) + verbose=1 + shift 1 + ;; + --version) + args="-help" + shift 1 + ;; -fexceptions) # Don't enable exceptions for now. #args="$args -EHac" @@ -68,9 +83,24 @@ do safeseh= shift 1 ;; +<<<<<<< HEAD -clang-cl) cl="clang-cl" safeseh= +======= + -marm) + ml='armasm' + safeseh= + shift 1 + ;; + -marm64) + ml='armasm64' + safeseh= + shift 1 + ;; + -clang-cl) + cl="clang-cl" +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 shift 1 ;; -O0) @@ -144,6 +174,7 @@ do shift 1 ;; -I) +<<<<<<< HEAD args="$args -I\"$2\"" includes="$includes -I\"$2\"" shift 2 @@ -151,6 +182,46 @@ do -I*) args="$args -I\"$(echo $1|sed 's/^-I//g')\"" includes="$includes -I\"$(echo $1|sed 's/^-I//g')\"" +======= + p=$(cygpath -m $2) + args="$args -I$p" + includes="$includes -I$p" + shift 2 + ;; + -I*) + p=$(cygpath -m ${1#-I}) + args="$args -I$p" + includes="$includes -I$p" + shift 1 + ;; + -L) + p=$(cygpath -m $2) + linkargs="$linkargs -LIBPATH:$p" + shift 2 + ;; + -L*) + p=$(cygpath -m ${1#-L}) + linkargs="$linkargs -LIBPATH:$p" + shift 1 + ;; + -link) + # add next argument verbatim to linker args + linkargs="$linkargs $2" + shift 2 + ;; + -l*) + case $1 + in + -lffi) + linkargs="$linkargs lib${1#-l}-${libversion}.lib" + ;; + *) + # ignore other libraries like -lm, hope they are + # covered by MSVCRT + # linkargs="$linkargs ${1#-l}.lib" + ;; + esac +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 shift 1 ;; -W|-Wextra) @@ -166,6 +237,18 @@ do # libffi tests -pedantic with -Wall, so drop it also. shift 1 ;; +<<<<<<< HEAD +======= + -warn) + # ignore -warn all from libtool as well. + if test "$2" = "all"; then + shift 2 + else + args="$args -warn" + shift 1 + fi + ;; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 -Werror) args="$args -WX" shift 1 @@ -190,6 +273,7 @@ do else output="-Fe$2" fi + armasm_output="-o $2" if [ -n "$assembly" ]; then args="$args $output" else @@ -214,11 +298,24 @@ do esac done +<<<<<<< HEAD # If -Zi is specified, certain optimizations are implicitly disabled # by MSVC. Add back those optimizations if this is an optimized build. # NOTE: These arguments must come after all others. if [ -n "$opt" ]; then args="$args -link -OPT:REF -OPT:ICF -INCREMENTAL:NO" +======= +if [ -n "$linkargs" ]; then + + # If -Zi is specified, certain optimizations are implicitly disabled + # by MSVC. Add back those optimizations if this is an optimized build. + # NOTE: These arguments must come after all others. + if [ -n "$opt" ]; then + linkargs="$linkargs -OPT:REF -OPT:ICF -INCREMENTAL:NO" + fi + + args="$args -link $linkargs" +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 fi if [ -n "$static_crt" ]; then @@ -236,12 +333,38 @@ if [ -n "$assembly" ]; then outdir="." fi ppsrc="$outdir/$(basename $src|sed 's/.S$/.asm/g')" +<<<<<<< HEAD echo "$cl -nologo -EP $includes $defines $src > $ppsrc" eval "\"$cl\" -nologo -EP $includes $defines $src > $ppsrc || exit $?" +======= + + if [ $ml = "armasm" ]; then + defines="$defines -D_M_ARM" + fi + + if [ $ml = "armasm64" ]; then + defines="$defines -D_M_ARM64" + fi + + if test -n "$verbose"; then + echo "$cl -nologo -EP $includes $defines $src > $ppsrc" + fi + + "$cl" -nologo -EP $includes $defines $src > $ppsrc || exit $? +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 output="$(echo $output | sed 's%/F[dpa][^ ]*%%g')" - args="-nologo $safeseh $single $output $ppsrc" + if [ $ml = "armasm" ]; then + args="-nologo -g -oldit $armasm_output $ppsrc -errorReport:prompt" + elif [ $ml = "armasm64" ]; then + args="-nologo -g $armasm_output $ppsrc -errorReport:prompt" + else + args="-nologo $safeseh $single $output $ppsrc" + fi + + if test -n "$verbose"; then + echo "$ml $args" + fi - echo "$ml $args" eval "\"$ml\" $args" result=$? @@ -249,13 +372,29 @@ if [ -n "$assembly" ]; then #mv *.obj $outdir else args="$md $args" +<<<<<<< HEAD echo "$cl $args" # Return an error code of 1 if an invalid command line parameter is passed # instead of just ignoring it. eval "(\"$cl\" $args 2>&1 1>&3 | \ awk '{print \$0} /D9002/ {error=1} END{exit error}' >&2) 3>&1" +======= + + if test -n "$verbose"; then + echo "$cl $args" + fi + # Return an error code of 1 if an invalid command line parameter is passed + # instead of just ignoring it. Any output that is not a warning or an + # error is filtered so this command behaves more like gcc. cl.exe prints + # the name of the compiled file otherwise, which breaks the dejagnu checks + # for excess warnings and errors. + eval "(\"$cl\" $args 2>&1 1>&3 | \ + awk '{print \$0} /D9002/ {error=1} END{exit error}' >&2) 3>&1 | \ + awk '/warning|error/'" +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 result=$? fi exit $result +# vim: noai:ts=4:sw=4 diff --git a/native/libffi/src/aarch64/ffi.c b/native/libffi/src/aarch64/ffi.c index cdb7816e72..bfc21f9486 100644 --- a/native/libffi/src/aarch64/ffi.c +++ b/native/libffi/src/aarch64/ffi.c @@ -19,11 +19,19 @@ CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ +#if defined(__aarch64__) || defined(__arm64__)|| defined (_M_ARM64) #include - +#include +#include +#include #include #include +#include "internal.h" +#ifdef _WIN32 +#include /* FlushInstructionCache */ +#endif +<<<<<<< HEAD #include /* Stack alignment requirement in bytes */ @@ -37,6 +45,18 @@ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ #define N_V_ARG_REG 8 #define AARCH64_FFI_WITH_V (1 << AARCH64_FFI_WITH_V_BIT) +======= +/* Force FFI_TYPE_LONGDOUBLE to be different than FFI_TYPE_DOUBLE; + all further uses in this file will refer to the 128-bit type. */ +#if FFI_TYPE_DOUBLE != FFI_TYPE_LONGDOUBLE +# if FFI_TYPE_LONGDOUBLE != 4 +# error FFI_TYPE_LONGDOUBLE out of date +# endif +#else +# undef FFI_TYPE_LONGDOUBLE +# define FFI_TYPE_LONGDOUBLE 4 +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 union _d { @@ -44,15 +64,12 @@ union _d UINT32 s[2]; }; -struct call_context +struct _v { - UINT64 x [AARCH64_N_XREG]; - struct - { - union _d d[2]; - } v [AARCH64_N_VREG]; + union _d d[2] __attribute__((aligned(16))); }; +<<<<<<< HEAD #if defined (__clang__) && defined (__APPLE__) extern void sys_icache_invalidate (void *start, size_t len); @@ -72,33 +89,33 @@ ffi_clear_cache (void *start, void *end) static void * get_x_addr (struct call_context *context, unsigned n) +======= +struct call_context +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 { - return &context->x[n]; -} + struct _v v[N_V_ARG_REG]; + UINT64 x[N_X_ARG_REG]; +}; -static void * -get_s_addr (struct call_context *context, unsigned n) -{ -#if defined __AARCH64EB__ - return &context->v[n].d[1].s[1]; -#else - return &context->v[n].d[0].s[0]; +#if FFI_EXEC_TRAMPOLINE_TABLE + +#ifdef __MACH__ +#ifdef HAVE_PTRAUTH +#include +#endif +#include #endif -} -static void * -get_d_addr (struct call_context *context, unsigned n) -{ -#if defined __AARCH64EB__ - return &context->v[n].d[1]; #else - return &context->v[n].d[0]; + +#if defined (__clang__) && defined (__APPLE__) +extern void sys_icache_invalidate (void *start, size_t len); #endif -} -static void * -get_v_addr (struct call_context *context, unsigned n) +static inline void +ffi_clear_cache (void *start, void *end) { +<<<<<<< HEAD return &context->v[n]; } @@ -181,12 +198,29 @@ get_basic_type_alignment (unsigned short type) return 0; } } +======= +#if defined (__clang__) && defined (__APPLE__) + sys_icache_invalidate (start, (char *)end - (char *)start); +#elif defined (__GNUC__) + __builtin___clear_cache (start, end); +#elif defined (_WIN32) + FlushInstructionCache(GetCurrentProcess(), start, (char*)end - (char*)start); +#else +#error "Missing builtin to flush instruction cache" +#endif +} -/* Return the size in bytes for each of the basic types. */ +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 -static size_t -get_basic_type_size (unsigned short type) +/* A subroutine of is_vfp_type. Given a structure type, return the type code + of the first non-structure element. Recurse for structure elements. + Return -1 if the structure is in fact empty, i.e. no nested elements. */ + +static int +is_hfa0 (const ffi_type *ty) { +<<<<<<< HEAD switch (type) { case FFI_TYPE_FLOAT: @@ -234,118 +268,122 @@ extern void ffi_closure_SYSV (ffi_closure *); /* Test for an FFI floating point representation. */ +======= + ffi_type **elements = ty->elements; + int i, ret = -1; -static unsigned -is_floating_type (unsigned short type) -{ - return (type == FFI_TYPE_FLOAT || type == FFI_TYPE_DOUBLE - || type == FFI_TYPE_LONGDOUBLE); + if (elements != NULL) + for (i = 0; elements[i]; ++i) + { + ret = elements[i]->type; + if (ret == FFI_TYPE_STRUCT || ret == FFI_TYPE_COMPLEX) + { + ret = is_hfa0 (elements[i]); + if (ret < 0) + continue; + } + break; + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + + return ret; } -/* Test for a homogeneous structure. */ +/* A subroutine of is_vfp_type. Given a structure type, return true if all + of the non-structure elements are the same as CANDIDATE. */ -static unsigned short -get_homogeneous_type (ffi_type *ty) +static int +is_hfa1 (const ffi_type *ty, int candidate) { - if (ty->type == FFI_TYPE_STRUCT && ty->elements) - { - unsigned i; - unsigned short candidate_type - = get_homogeneous_type (ty->elements[0]); - for (i =1; ty->elements[i]; i++) - { - unsigned short iteration_type = 0; - /* If we have a nested struct, we must find its homogeneous type. - If that fits with our candidate type, we are still - homogeneous. */ - if (ty->elements[i]->type == FFI_TYPE_STRUCT - && ty->elements[i]->elements) - { - iteration_type = get_homogeneous_type (ty->elements[i]); - } - else - { - iteration_type = ty->elements[i]->type; - } + ffi_type **elements = ty->elements; + int i; - /* If we are not homogeneous, return FFI_TYPE_STRUCT. */ - if (candidate_type != iteration_type) - return FFI_TYPE_STRUCT; - } - return candidate_type; - } + if (elements != NULL) + for (i = 0; elements[i]; ++i) + { + int t = elements[i]->type; + if (t == FFI_TYPE_STRUCT || t == FFI_TYPE_COMPLEX) + { + if (!is_hfa1 (elements[i], candidate)) + return 0; + } + else if (t != candidate) + return 0; + } - /* Base case, we have no more levels of nesting, so we - are a basic type, and so, trivially homogeneous in that type. */ - return ty->type; + return 1; } -/* Determine the number of elements within a STRUCT. +/* Determine if TY may be allocated to the FP registers. This is both an + fp scalar type as well as an homogenous floating point aggregate (HFA). + That is, a structure consisting of 1 to 4 members of all the same type, + where that type is an fp scalar. - Note, we must handle nested structs. + Returns non-zero iff TY is an HFA. The result is the AARCH64_RET_* + constant for the type. */ - If ty is not a STRUCT this function will return 0. */ - -static unsigned -element_count (ffi_type *ty) +static int +is_vfp_type (const ffi_type *ty) { - if (ty->type == FFI_TYPE_STRUCT && ty->elements) + ffi_type **elements; + int candidate, i; + size_t size, ele_count; + + /* Quickest tests first. */ + candidate = ty->type; + switch (candidate) { - unsigned n; - unsigned elems = 0; - for (n = 0; ty->elements[n]; n++) + default: + return 0; + case FFI_TYPE_FLOAT: + case FFI_TYPE_DOUBLE: + case FFI_TYPE_LONGDOUBLE: + ele_count = 1; + goto done; + case FFI_TYPE_COMPLEX: + candidate = ty->elements[0]->type; + switch (candidate) { - if (ty->elements[n]->type == FFI_TYPE_STRUCT - && ty->elements[n]->elements) - elems += element_count (ty->elements[n]); - else - elems++; + case FFI_TYPE_FLOAT: + case FFI_TYPE_DOUBLE: + case FFI_TYPE_LONGDOUBLE: + ele_count = 2; + goto done; } - return elems; + return 0; + case FFI_TYPE_STRUCT: + break; } - return 0; -} -/* Test for a homogeneous floating point aggregate. + /* No HFA types are smaller than 4 bytes, or larger than 64 bytes. */ + size = ty->size; + if (size < 4 || size > 64) + return 0; - A homogeneous floating point aggregate is a homogeneous aggregate of - a half- single- or double- precision floating point type with one - to four elements. Note that this includes nested structs of the - basic type. */ - -static int -is_hfa (ffi_type *ty) -{ - if (ty->type == FFI_TYPE_STRUCT - && ty->elements[0] - && is_floating_type (get_homogeneous_type (ty))) + /* Find the type of the first non-structure member. */ + elements = ty->elements; + candidate = elements[0]->type; + if (candidate == FFI_TYPE_STRUCT || candidate == FFI_TYPE_COMPLEX) { - unsigned n = element_count (ty); - return n >= 1 && n <= 4; + for (i = 0; ; ++i) + { + candidate = is_hfa0 (elements[i]); + if (candidate >= 0) + break; + } } - return 0; -} - -/* Test if an ffi_type is a candidate for passing in a register. - - This test does not check that sufficient registers of the - appropriate class are actually available, merely that IFF - sufficient registers are available then the argument will be passed - in register(s). - - Note that an ffi_type that is deemed to be a register candidate - will always be returned in registers. - Returns 1 if a register candidate else 0. */ - -static int -is_register_candidate (ffi_type *ty) -{ - switch (ty->type) + /* If the first member is not a floating point type, it's not an HFA. + Also quickly re-check the size of the structure. */ + switch (candidate) { - case FFI_TYPE_VOID: case FFI_TYPE_FLOAT: + ele_count = size / sizeof(float); + if (size != ele_count * sizeof(float)) + return 0; + break; case FFI_TYPE_DOUBLE: +<<<<<<< HEAD #if FFI_TYPE_DOUBLE != FFI_TYPE_LONGDOUBLE case FFI_TYPE_LONGDOUBLE: #endif @@ -380,24 +418,39 @@ is_register_candidate (ffi_type *ty) registers required. */ return (ty->size + 7) / 8 < N_X_ARG_REG; } +======= + ele_count = size / sizeof(double); + if (size != ele_count * sizeof(double)) + return 0; break; - - default: - FFI_ASSERT (0); + case FFI_TYPE_LONGDOUBLE: + ele_count = size / sizeof(long double); + if (size != ele_count * sizeof(long double)) + return 0; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 break; + default: + return 0; } + if (ele_count > 4) + return 0; - return 0; -} - -/* Test if an ffi_type argument or result is a candidate for a vector - register. */ + /* Finally, make sure that all scalar elements are the same type. */ + for (i = 0; elements[i]; ++i) + { + int t = elements[i]->type; + if (t == FFI_TYPE_STRUCT || t == FFI_TYPE_COMPLEX) + { + if (!is_hfa1 (elements[i], candidate)) + return 0; + } + else if (t != candidate) + return 0; + } -static int -is_v_register_candidate (ffi_type *ty) -{ - return is_floating_type (ty->type) - || (ty->type == FFI_TYPE_STRUCT && is_hfa (ty)); + /* All tests succeeded. Encode the result. */ + done: + return candidate * 4 + (4 - (int)ele_count); } /* Representation of the procedure call argument marshalling @@ -419,11 +472,16 @@ struct arg_state /* Initialize a procedure call argument marshalling state. */ static void +<<<<<<< HEAD arg_init (struct arg_state *state, size_t call_frame_size) +======= +arg_init (struct arg_state *state) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 { state->ngrn = 0; state->nsrn = 0; state->nsaa = 0; +<<<<<<< HEAD #if defined (__APPLE__) state->allocating_variadic = 0; @@ -474,17 +532,28 @@ allocate_to_v (struct call_context *context, struct arg_state *state) { FFI_ASSERT (state->nsrn < N_V_ARG_REG); return get_v_addr (context, (state->nsrn)++); +======= +#if defined (__APPLE__) + state->allocating_variadic = 0; +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } /* Allocate an aligned slot on the stack and return a pointer to it. */ static void * +<<<<<<< HEAD allocate_to_stack (struct arg_state *state, void *stack, size_t alignment, size_t size) +======= +allocate_to_stack (struct arg_state *state, void *stack, + size_t alignment, size_t size) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 { - void *allocation; + size_t nsaa = state->nsaa; /* Round up the NSAA to the larger of 8 or the natural alignment of the argument's type. */ +<<<<<<< HEAD state->nsaa = ALIGN (state->nsaa, alignment); state->nsaa = ALIGN (state->nsaa, alignment); #if defined (__APPLE__) @@ -495,14 +564,26 @@ allocate_to_stack (struct arg_state *state, void *stack, size_t alignment, #endif allocation = stack + state->nsaa; +======= +#if defined (__APPLE__) + if (state->allocating_variadic && alignment < 8) + alignment = 8; +#else + if (alignment < 8) + alignment = 8; +#endif + + nsaa = FFI_ALIGN (nsaa, alignment); + state->nsaa = nsaa + size; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 - state->nsaa += size; - return allocation; + return (char *)stack + nsaa; } -static void -copy_basic_type (void *dest, void *source, unsigned short type) +static ffi_arg +extend_integer_type (void *source, int type) { +<<<<<<< HEAD /* This is necessary to ensure that basic types are copied sign extended to 64-bits as libffi expects. */ switch (type) @@ -518,85 +599,203 @@ copy_basic_type (void *dest, void *source, unsigned short type) *(long double *) dest = *(long double *) source; break; #endif +======= + switch (type) + { +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 case FFI_TYPE_UINT8: - *(ffi_arg *) dest = *(UINT8 *) source; - break; + return *(UINT8 *) source; case FFI_TYPE_SINT8: - *(ffi_sarg *) dest = *(SINT8 *) source; - break; + return *(SINT8 *) source; case FFI_TYPE_UINT16: - *(ffi_arg *) dest = *(UINT16 *) source; - break; + return *(UINT16 *) source; case FFI_TYPE_SINT16: - *(ffi_sarg *) dest = *(SINT16 *) source; - break; + return *(SINT16 *) source; case FFI_TYPE_UINT32: - *(ffi_arg *) dest = *(UINT32 *) source; - break; + return *(UINT32 *) source; case FFI_TYPE_INT: case FFI_TYPE_SINT32: - *(ffi_sarg *) dest = *(SINT32 *) source; - break; - case FFI_TYPE_POINTER: + return *(SINT32 *) source; case FFI_TYPE_UINT64: - *(ffi_arg *) dest = *(UINT64 *) source; - break; case FFI_TYPE_SINT64: - *(ffi_sarg *) dest = *(SINT64 *) source; + return *(UINT64 *) source; break; +<<<<<<< HEAD case FFI_TYPE_VOID: break; +======= + case FFI_TYPE_POINTER: + return *(uintptr_t *) source; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 default: - FFI_ASSERT (0); + abort(); } } +#if defined(_MSC_VER) +void extend_hfa_type (void *dest, void *src, int h); +#else static void -copy_hfa_to_reg_or_stack (void *memory, - ffi_type *ty, - struct call_context *context, - unsigned char *stack, - struct arg_state *state) +extend_hfa_type (void *dest, void *src, int h) { - unsigned elems = element_count (ty); - if (available_v (state) < elems) - { - /* There are insufficient V registers. Further V register allocations - are prevented, the NSAA is adjusted (by allocate_to_stack ()) - and the argument is copied to memory at the adjusted NSAA. */ - state->nsrn = N_V_ARG_REG; - memcpy (allocate_to_stack (state, stack, ty->alignment, ty->size), - memory, - ty->size); - } - else + ssize_t f = h - AARCH64_RET_S4; + void *x0; + + asm volatile ( + "adr %0, 0f\n" +" add %0, %0, %1\n" +" br %0\n" +"0: ldp s16, s17, [%3]\n" /* S4 */ +" ldp s18, s19, [%3, #8]\n" +" b 4f\n" +" ldp s16, s17, [%3]\n" /* S3 */ +" ldr s18, [%3, #8]\n" +" b 3f\n" +" ldp s16, s17, [%3]\n" /* S2 */ +" b 2f\n" +" nop\n" +" ldr s16, [%3]\n" /* S1 */ +" b 1f\n" +" nop\n" +" ldp d16, d17, [%3]\n" /* D4 */ +" ldp d18, d19, [%3, #16]\n" +" b 4f\n" +" ldp d16, d17, [%3]\n" /* D3 */ +" ldr d18, [%3, #16]\n" +" b 3f\n" +" ldp d16, d17, [%3]\n" /* D2 */ +" b 2f\n" +" nop\n" +" ldr d16, [%3]\n" /* D1 */ +" b 1f\n" +" nop\n" +" ldp q16, q17, [%3]\n" /* Q4 */ +" ldp q18, q19, [%3, #32]\n" +" b 4f\n" +" ldp q16, q17, [%3]\n" /* Q3 */ +" ldr q18, [%3, #32]\n" +" b 3f\n" +" ldp q16, q17, [%3]\n" /* Q2 */ +" b 2f\n" +" nop\n" +" ldr q16, [%3]\n" /* Q1 */ +" b 1f\n" +"4: str q19, [%2, #48]\n" +"3: str q18, [%2, #32]\n" +"2: str q17, [%2, #16]\n" +"1: str q16, [%2]" + : "=&r"(x0) + : "r"(f * 12), "r"(dest), "r"(src) + : "memory", "v16", "v17", "v18", "v19"); +} +#endif + +#if defined(_MSC_VER) +void* compress_hfa_type (void *dest, void *src, int h); +#else +static void * +compress_hfa_type (void *dest, void *reg, int h) +{ + switch (h) { +<<<<<<< HEAD int i; unsigned short type = get_homogeneous_type (ty); for (i = 0; i < elems; i++) +======= + case AARCH64_RET_S1: + if (dest == reg) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + { +#ifdef __AARCH64EB__ + dest += 12; +#endif + } + else + *(float *)dest = *(float *)reg; + break; + case AARCH64_RET_S2: + asm ("ldp q16, q17, [%1]\n\t" + "st2 { v16.s, v17.s }[0], [%0]" + : : "r"(dest), "r"(reg) : "memory", "v16", "v17"); + break; + case AARCH64_RET_S3: + asm ("ldp q16, q17, [%1]\n\t" + "ldr q18, [%1, #32]\n\t" + "st3 { v16.s, v17.s, v18.s }[0], [%0]" + : : "r"(dest), "r"(reg) : "memory", "v16", "v17", "v18"); + break; + case AARCH64_RET_S4: + asm ("ldp q16, q17, [%1]\n\t" + "ldp q18, q19, [%1, #32]\n\t" + "st4 { v16.s, v17.s, v18.s, v19.s }[0], [%0]" + : : "r"(dest), "r"(reg) : "memory", "v16", "v17", "v18", "v19"); + break; + + case AARCH64_RET_D1: + if (dest == reg) { - void *reg = allocate_to_v (context, state); - copy_basic_type (reg, memory, type); - memory += get_basic_type_size (type); +#ifdef __AARCH64EB__ + dest += 8; +#endif } + else + *(double *)dest = *(double *)reg; + break; + case AARCH64_RET_D2: + asm ("ldp q16, q17, [%1]\n\t" + "st2 { v16.d, v17.d }[0], [%0]" + : : "r"(dest), "r"(reg) : "memory", "v16", "v17"); + break; + case AARCH64_RET_D3: + asm ("ldp q16, q17, [%1]\n\t" + "ldr q18, [%1, #32]\n\t" + "st3 { v16.d, v17.d, v18.d }[0], [%0]" + : : "r"(dest), "r"(reg) : "memory", "v16", "v17", "v18"); + break; + case AARCH64_RET_D4: + asm ("ldp q16, q17, [%1]\n\t" + "ldp q18, q19, [%1, #32]\n\t" + "st4 { v16.d, v17.d, v18.d, v19.d }[0], [%0]" + : : "r"(dest), "r"(reg) : "memory", "v16", "v17", "v18", "v19"); + break; + + default: + if (dest != reg) + return memcpy (dest, reg, 16 * (4 - (h & 3))); + break; } + return dest; } +#endif /* Either allocate an appropriate register for the argument type, or if none are available, allocate a stack slot and return a pointer to the allocated space. */ static void * -allocate_to_register_or_stack (struct call_context *context, - unsigned char *stack, - struct arg_state *state, - unsigned short type) +allocate_int_to_reg_or_stack (struct call_context *context, + struct arg_state *state, + void *stack, size_t size) { - size_t alignment = get_basic_type_alignment (type); - size_t size = alignment; - switch (type) + if (state->ngrn < N_X_ARG_REG) + return &context->x[state->ngrn++]; + + state->ngrn = N_X_ARG_REG; + return allocate_to_stack (state, stack, size, size); +} + +ffi_status FFI_HIDDEN +ffi_prep_cif_machdep (ffi_cif *cif) +{ + ffi_type *rtype = cif->rtype; + size_t bytes = cif->bytes; + int flags, i, n; + + switch (rtype->type) { +<<<<<<< HEAD case FFI_TYPE_FLOAT: /* This is the only case for which the allocated stack size should not match the alignment of the type. */ @@ -612,61 +811,161 @@ allocate_to_register_or_stack (struct call_context *context, if (state->nsrn < N_V_ARG_REG) return allocate_to_v (context, state); state->nsrn = N_V_ARG_REG; +======= + case FFI_TYPE_VOID: + flags = AARCH64_RET_VOID; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 break; #endif case FFI_TYPE_UINT8: - case FFI_TYPE_SINT8: + flags = AARCH64_RET_UINT8; + break; case FFI_TYPE_UINT16: - case FFI_TYPE_SINT16: + flags = AARCH64_RET_UINT16; + break; case FFI_TYPE_UINT32: - case FFI_TYPE_SINT32: + flags = AARCH64_RET_UINT32; + break; + case FFI_TYPE_SINT8: + flags = AARCH64_RET_SINT8; + break; + case FFI_TYPE_SINT16: + flags = AARCH64_RET_SINT16; + break; case FFI_TYPE_INT: - case FFI_TYPE_POINTER: - case FFI_TYPE_UINT64: + case FFI_TYPE_SINT32: + flags = AARCH64_RET_SINT32; + break; case FFI_TYPE_SINT64: - if (state->ngrn < N_X_ARG_REG) - return allocate_to_x (context, state); - state->ngrn = N_X_ARG_REG; + case FFI_TYPE_UINT64: + flags = AARCH64_RET_INT64; break; + case FFI_TYPE_POINTER: + flags = (sizeof(void *) == 4 ? AARCH64_RET_UINT32 : AARCH64_RET_INT64); + break; + + case FFI_TYPE_FLOAT: + case FFI_TYPE_DOUBLE: + case FFI_TYPE_LONGDOUBLE: + case FFI_TYPE_STRUCT: + case FFI_TYPE_COMPLEX: + flags = is_vfp_type (rtype); + if (flags == 0) + { + size_t s = rtype->size; + if (s > 16) + { + flags = AARCH64_RET_VOID | AARCH64_RET_IN_MEM; + bytes += 8; + } + else if (s == 16) + flags = AARCH64_RET_INT128; + else if (s == 8) + flags = AARCH64_RET_INT64; + else + flags = AARCH64_RET_INT128 | AARCH64_RET_NEED_COPY; + } + break; + default: - FFI_ASSERT (0); + abort(); } - return allocate_to_stack (state, stack, alignment, size); -} + for (i = 0, n = cif->nargs; i < n; i++) + if (is_vfp_type (cif->arg_types[i])) + { + flags |= AARCH64_FLAG_ARG_V; + break; + } -/* Copy a value to an appropriate register, or if none are - available, to the stack. */ + /* Round the stack up to a multiple of the stack alignment requirement. */ + cif->bytes = (unsigned) FFI_ALIGN(bytes, 16); + cif->flags = flags; +#if defined (__APPLE__) + cif->aarch64_nfixedargs = 0; +#endif -static void -copy_to_register_or_stack (struct call_context *context, - unsigned char *stack, - struct arg_state *state, - void *value, - unsigned short type) + return FFI_OK; +} + +#if defined (__APPLE__) +/* Perform Apple-specific cif processing for variadic calls */ +ffi_status FFI_HIDDEN +ffi_prep_cif_machdep_var(ffi_cif *cif, unsigned int nfixedargs, + unsigned int ntotalargs) +{ + ffi_status status = ffi_prep_cif_machdep (cif); + cif->aarch64_nfixedargs = nfixedargs; + return status; +} +#else +ffi_status FFI_HIDDEN +ffi_prep_cif_machdep_var(ffi_cif *cif, unsigned int nfixedargs, unsigned int ntotalargs) { - copy_basic_type ( - allocate_to_register_or_stack (context, stack, state, type), - value, - type); + ffi_status status = ffi_prep_cif_machdep (cif); + cif->flags |= AARCH64_FLAG_VARARG; + return status; } +#endif /* __APPLE__ */ -/* Marshall the arguments from FFI representation to procedure call - context and stack. */ +extern void ffi_call_SYSV (struct call_context *context, void *frame, + void (*fn)(void), void *rvalue, int flags, + void *closure) FFI_HIDDEN; -static unsigned -aarch64_prep_args (struct call_context *context, unsigned char *stack, - extended_cif *ecif) +/* Call a function with the provided arguments and capture the return + value. */ +static void +ffi_call_int (ffi_cif *cif, void (*fn)(void), void *orig_rvalue, + void **avalue, void *closure) { - int i; + struct call_context *context; + void *stack, *frame, *rvalue; struct arg_state state; + size_t stack_bytes, rtype_size, rsize; + int i, nargs, flags, isvariadic = 0; + ffi_type *rtype; - arg_init (&state, ALIGN(ecif->cif->bytes, 16)); + flags = cif->flags; + rtype = cif->rtype; + rtype_size = rtype->size; + stack_bytes = cif->bytes; - for (i = 0; i < ecif->cif->nargs; i++) + if (flags & AARCH64_FLAG_VARARG) + { + isvariadic = 1; + flags &= ~AARCH64_FLAG_VARARG; + } + + /* If the target function returns a structure via hidden pointer, + then we cannot allow a null rvalue. Otherwise, mash a null + rvalue to void return type. */ + rsize = 0; + if (flags & AARCH64_RET_IN_MEM) { - ffi_type *ty = ecif->cif->arg_types[i]; - switch (ty->type) + if (orig_rvalue == NULL) + rsize = rtype_size; + } + else if (orig_rvalue == NULL) + flags &= AARCH64_FLAG_ARG_V; + else if (flags & AARCH64_RET_NEED_COPY) + rsize = 16; + + /* Allocate consectutive stack for everything we'll need. */ + context = alloca (sizeof(struct call_context) + stack_bytes + 32 + rsize); + stack = context + 1; + frame = (void*)((uintptr_t)stack + (uintptr_t)stack_bytes); + rvalue = (rsize ? (void*)((uintptr_t)frame + 32) : orig_rvalue); + + arg_init (&state); + for (i = 0, nargs = cif->nargs; i < nargs; i++) + { + ffi_type *ty = cif->arg_types[i]; + size_t s = ty->size; + void *a = avalue[i]; + int h, t; + + t = ty->type; + switch (t) { case FFI_TYPE_VOID: FFI_ASSERT (0); @@ -674,71 +973,130 @@ aarch64_prep_args (struct call_context *context, unsigned char *stack, /* If the argument is a basic type the argument is allocated to an appropriate register, or if none are available, to the stack. */ +<<<<<<< HEAD case FFI_TYPE_FLOAT: case FFI_TYPE_DOUBLE: #if FFI_TYPE_DOUBLE != FFI_TYPE_LONGDOUBLE case FFI_TYPE_LONGDOUBLE: #endif +======= + case FFI_TYPE_INT: +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 case FFI_TYPE_UINT8: case FFI_TYPE_SINT8: case FFI_TYPE_UINT16: case FFI_TYPE_SINT16: case FFI_TYPE_UINT32: - case FFI_TYPE_INT: case FFI_TYPE_SINT32: - case FFI_TYPE_POINTER: case FFI_TYPE_UINT64: case FFI_TYPE_SINT64: - copy_to_register_or_stack (context, stack, &state, - ecif->avalue[i], ty->type); + case FFI_TYPE_POINTER: + do_pointer: + { + ffi_arg ext = extend_integer_type (a, t); + if (state.ngrn < N_X_ARG_REG) + context->x[state.ngrn++] = ext; + else + { + void *d = allocate_to_stack (&state, stack, ty->alignment, s); + state.ngrn = N_X_ARG_REG; + /* Note that the default abi extends each argument + to a full 64-bit slot, while the iOS abi allocates + only enough space. */ +#ifdef __APPLE__ + memcpy(d, a, s); +#else + *(ffi_arg *)d = ext; +#endif + } + } break; + case FFI_TYPE_FLOAT: + case FFI_TYPE_DOUBLE: + case FFI_TYPE_LONGDOUBLE: case FFI_TYPE_STRUCT: - if (is_hfa (ty)) - { - copy_hfa_to_reg_or_stack (ecif->avalue[i], ty, context, - stack, &state); - } - else if (ty->size > 16) - { - /* If the argument is a composite type that is larger than 16 - bytes, then the argument has been copied to memory, and - the argument is replaced by a pointer to the copy. */ + case FFI_TYPE_COMPLEX: + { + void *dest; - copy_to_register_or_stack (context, stack, &state, - &(ecif->avalue[i]), FFI_TYPE_POINTER); - } - else if (available_x (&state) >= (ty->size + 7) / 8) - { - /* If the argument is a composite type and the size in - double-words is not more than the number of available - X registers, then the argument is copied into consecutive - X registers. */ - int j; - for (j = 0; j < (ty->size + 7) / 8; j++) - { - memcpy (allocate_to_x (context, &state), - &(((UINT64 *) ecif->avalue[i])[j]), - sizeof (UINT64)); + h = is_vfp_type (ty); + if (h) + { + int elems = 4 - (h & 3); + if (cif->abi == FFI_WIN64 && isvariadic) + { + if (state.ngrn + elems <= N_X_ARG_REG) + { + dest = &context->x[state.ngrn]; + state.ngrn += elems; + extend_hfa_type(dest, a, h); + break; + } + state.nsrn = N_X_ARG_REG; + dest = allocate_to_stack(&state, stack, ty->alignment, s); + } + else + { + if (state.nsrn + elems <= N_V_ARG_REG) + { + dest = &context->v[state.nsrn]; + state.nsrn += elems; + extend_hfa_type (dest, a, h); + break; + } + state.nsrn = N_V_ARG_REG; + dest = allocate_to_stack (&state, stack, ty->alignment, s); + } + } + else if (s > 16) + { + /* If the argument is a composite type that is larger than 16 + bytes, then the argument has been copied to memory, and + the argument is replaced by a pointer to the copy. */ + a = &avalue[i]; + t = FFI_TYPE_POINTER; + s = sizeof (void *); + goto do_pointer; + } + else + { + size_t n = (s + 7) / 8; + if (state.ngrn + n <= N_X_ARG_REG) + { + /* If the argument is a composite type and the size in + double-words is not more than the number of available + X registers, then the argument is copied into + consecutive X registers. */ + dest = &context->x[state.ngrn]; + state.ngrn += (unsigned int)n; + } + else + { + /* Otherwise, there are insufficient X registers. Further + X register allocations are prevented, the NSAA is + adjusted and the argument is copied to memory at the + adjusted NSAA. */ + state.ngrn = N_X_ARG_REG; + dest = allocate_to_stack (&state, stack, ty->alignment, s); + } } - } - else - { - /* Otherwise, there are insufficient X registers. Further X - register allocations are prevented, the NSAA is adjusted - (by allocate_to_stack ()) and the argument is copied to - memory at the adjusted NSAA. */ - state.ngrn = N_X_ARG_REG; - - memcpy (allocate_to_stack (&state, stack, ty->alignment, - ty->size), ecif->avalue + i, ty->size); + memcpy (dest, a, s); } break; default: - FFI_ASSERT (0); - break; + abort(); + } + +#if defined (__APPLE__) + if (i + 1 == cif->aarch64_nfixedargs) + { + state.ngrn = N_X_ARG_REG; + state.nsrn = N_V_ARG_REG; + state.allocating_variadic = 1; } +<<<<<<< HEAD #if defined (__APPLE__) if (i + 1 == ecif->cif->aarch64_nfixedargs) @@ -748,15 +1106,21 @@ aarch64_prep_args (struct call_context *context, unsigned char *stack, state.allocating_variadic = 1; } +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #endif } - return ecif->cif->aarch64_flags; + ffi_call_SYSV (context, frame, fn, rvalue, flags, closure); + + if (flags & AARCH64_RET_NEED_COPY) + memcpy (orig_rvalue, rvalue, rtype_size); } -ffi_status -ffi_prep_cif_machdep (ffi_cif *cif) +void +ffi_call (ffi_cif *cif, void (*fn) (void), void *rvalue, void **avalue) { +<<<<<<< HEAD /* Round the stack up to a multiple of the stack alignment requirement. */ cif->bytes = (cif->bytes + (AARCH64_STACK_ALIGN - 1)) & ~ (AARCH64_STACK_ALIGN - 1); @@ -809,9 +1173,17 @@ ffi_status ffi_prep_cif_machdep_var(ffi_cif *cif, /* Call a function with the provided arguments and capture the return value. */ +======= + ffi_call_int (cif, fn, rvalue, avalue, NULL); +} + +#ifdef FFI_GO_CLOSURES +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 void -ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +ffi_call_go (ffi_cif *cif, void (*fn) (void), void *rvalue, + void **avalue, void *closure) { +<<<<<<< HEAD extended_cif ecif; ecif.cif = cif; @@ -902,16 +1274,15 @@ ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) FFI_ASSERT (0); break; } +======= + ffi_call_int (cif, fn, rvalue, avalue, closure); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } - -static unsigned char trampoline [] = -{ 0x70, 0x00, 0x00, 0x58, /* ldr x16, 1f */ - 0x91, 0x00, 0x00, 0x10, /* adr x17, 2f */ - 0x00, 0x02, 0x1f, 0xd6 /* br x16 */ -}; +#endif /* FFI_GO_CLOSURES */ /* Build a trampoline. */ +<<<<<<< HEAD #define FFI_INIT_TRAMPOLINE(TRAMP,FUN,CTX,FLAGS) \ ({unsigned char *__tramp = (unsigned char*)(TRAMP); \ UINT64 __fun = (UINT64)(FUN); \ @@ -923,9 +1294,13 @@ static unsigned char trampoline [] = memcpy (__tramp + 28, &__flags, sizeof (__flags)); \ ffi_clear_cache(__tramp, __tramp + FFI_TRAMPOLINE_SIZE); \ }) +======= +extern void ffi_closure_SYSV (void) FFI_HIDDEN; +extern void ffi_closure_SYSV_V (void) FFI_HIDDEN; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ffi_status -ffi_prep_closure_loc (ffi_closure* closure, +ffi_prep_closure_loc (ffi_closure *closure, ffi_cif* cif, void (*fun)(ffi_cif*,void*,void**,void*), void *user_data, @@ -934,16 +1309,80 @@ ffi_prep_closure_loc (ffi_closure* closure, if (cif->abi != FFI_SYSV) return FFI_BAD_ABI; - FFI_INIT_TRAMPOLINE (&closure->tramp[0], &ffi_closure_SYSV, codeloc, - cif->aarch64_flags); + void (*start)(void); + + if (cif->flags & AARCH64_FLAG_ARG_V) + start = ffi_closure_SYSV_V; + else + start = ffi_closure_SYSV; + +#if FFI_EXEC_TRAMPOLINE_TABLE +#ifdef __MACH__ +#ifdef HAVE_PTRAUTH + codeloc = ptrauth_strip (codeloc, ptrauth_key_asia); +#endif + void **config = (void **)((uint8_t *)codeloc - PAGE_MAX_SIZE); + config[0] = closure; + config[1] = start; +#endif +#else + static const unsigned char trampoline[16] = { + 0x90, 0x00, 0x00, 0x58, /* ldr x16, tramp+16 */ + 0xf1, 0xff, 0xff, 0x10, /* adr x17, tramp+0 */ + 0x00, 0x02, 0x1f, 0xd6 /* br x16 */ + }; + char *tramp = closure->tramp; + + memcpy (tramp, trampoline, sizeof(trampoline)); + + *(UINT64 *)(tramp + 16) = (uintptr_t)start; + + ffi_clear_cache(tramp, tramp + FFI_TRAMPOLINE_SIZE); + + /* Also flush the cache for code mapping. */ +#ifdef _WIN32 + // Not using dlmalloc.c for Windows ARM64 builds + // so calling ffi_data_to_code_pointer() isn't necessary + unsigned char *tramp_code = tramp; + #else + unsigned char *tramp_code = ffi_data_to_code_pointer (tramp); + #endif + ffi_clear_cache (tramp_code, tramp_code + FFI_TRAMPOLINE_SIZE); +#endif - closure->cif = cif; + closure->cif = cif; + closure->fun = fun; closure->user_data = user_data; - closure->fun = fun; return FFI_OK; } +#ifdef FFI_GO_CLOSURES +extern void ffi_go_closure_SYSV (void) FFI_HIDDEN; +extern void ffi_go_closure_SYSV_V (void) FFI_HIDDEN; + +ffi_status +ffi_prep_go_closure (ffi_go_closure *closure, ffi_cif* cif, + void (*fun)(ffi_cif*,void*,void**,void*)) +{ + void (*start)(void); + + if (cif->abi != FFI_SYSV) + return FFI_BAD_ABI; + + if (cif->flags & AARCH64_FLAG_ARG_V) + start = ffi_go_closure_SYSV_V; + else + start = ffi_go_closure_SYSV; + + closure->tramp = start; + closure->cif = cif; + closure->fun = fun; + + return FFI_OK; +} +#endif /* FFI_GO_CLOSURES */ + /* Primary handler to setup and invoke a function within a closure. A closure when invoked enters via the assembler wrapper @@ -960,48 +1399,65 @@ ffi_prep_closure_loc (ffi_closure* closure, descriptors, invokes the wrapped function, then marshalls the return value back into the call context. */ +<<<<<<< HEAD void FFI_HIDDEN ffi_closure_SYSV_inner (ffi_closure *closure, struct call_context *context, void *stack) +======= +int FFI_HIDDEN +ffi_closure_SYSV_inner (ffi_cif *cif, + void (*fun)(ffi_cif*,void*,void**,void*), + void *user_data, + struct call_context *context, + void *stack, void *rvalue, void *struct_rvalue) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 { - ffi_cif *cif = closure->cif; void **avalue = (void**) alloca (cif->nargs * sizeof (void*)); - void *rvalue = NULL; - int i; + int i, h, nargs, flags; struct arg_state state; - arg_init (&state, ALIGN(cif->bytes, 16)); + arg_init (&state); - for (i = 0; i < cif->nargs; i++) + for (i = 0, nargs = cif->nargs; i < nargs; i++) { ffi_type *ty = cif->arg_types[i]; + int t = ty->type; + size_t n, s = ty->size; - switch (ty->type) + switch (t) { case FFI_TYPE_VOID: FFI_ASSERT (0); break; + case FFI_TYPE_INT: case FFI_TYPE_UINT8: case FFI_TYPE_SINT8: case FFI_TYPE_UINT16: case FFI_TYPE_SINT16: case FFI_TYPE_UINT32: case FFI_TYPE_SINT32: - case FFI_TYPE_INT: - case FFI_TYPE_POINTER: case FFI_TYPE_UINT64: case FFI_TYPE_SINT64: +<<<<<<< HEAD case FFI_TYPE_FLOAT: case FFI_TYPE_DOUBLE: #if FFI_TYPE_DOUBLE != FFI_TYPE_LONGDOUBLE case FFI_TYPE_LONGDOUBLE: avalue[i] = allocate_to_register_or_stack (context, stack, &state, ty->type); +======= + case FFI_TYPE_POINTER: + avalue[i] = allocate_int_to_reg_or_stack (context, &state, stack, s); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 break; #endif + case FFI_TYPE_FLOAT: + case FFI_TYPE_DOUBLE: + case FFI_TYPE_LONGDOUBLE: case FFI_TYPE_STRUCT: +<<<<<<< HEAD if (is_hfa (ty)) { unsigned n = element_count (ty); @@ -1159,21 +1615,103 @@ ffi_closure_SYSV_inner (ffi_closure *closure, struct call_context *context, { size_t size = ALIGN (cif->rtype->size, sizeof (UINT64)) ; memcpy (get_x_addr (context, 0), rvalue, size); +======= + case FFI_TYPE_COMPLEX: + h = is_vfp_type (ty); + if (h) + { + n = 4 - (h & 3); +#ifdef _WIN32 /* for handling armasm calling convention */ + if (cif->is_variadic) + { + if (state.ngrn + n <= N_X_ARG_REG) + { + void *reg = &context->x[state.ngrn]; + state.ngrn += (unsigned int)n; + + /* Eeek! We need a pointer to the structure, however the + homogeneous float elements are being passed in individual + registers, therefore for float and double the structure + is not represented as a contiguous sequence of bytes in + our saved register context. We don't need the original + contents of the register storage, so we reformat the + structure into the same memory. */ + avalue[i] = compress_hfa_type(reg, reg, h); + } + else + { + state.ngrn = N_X_ARG_REG; + state.nsrn = N_V_ARG_REG; + avalue[i] = allocate_to_stack(&state, stack, + ty->alignment, s); + } + } + else + { +#endif /* for handling armasm calling convention */ + if (state.nsrn + n <= N_V_ARG_REG) + { + void *reg = &context->v[state.nsrn]; + state.nsrn += (unsigned int)n; + avalue[i] = compress_hfa_type(reg, reg, h); + } + else + { + state.nsrn = N_V_ARG_REG; + avalue[i] = allocate_to_stack(&state, stack, + ty->alignment, s); + } +#ifdef _WIN32 /* for handling armasm calling convention */ + } +#endif /* for handling armasm calling convention */ + } + else if (s > 16) + { + /* Replace Composite type of size greater than 16 with a + pointer. */ + avalue[i] = *(void **) + allocate_int_to_reg_or_stack (context, &state, stack, + sizeof (void *)); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } else { - FFI_ASSERT (0); + n = (s + 7) / 8; + if (state.ngrn + n <= N_X_ARG_REG) + { + avalue[i] = &context->x[state.ngrn]; + state.ngrn += (unsigned int)n; + } + else + { + state.ngrn = N_X_ARG_REG; + avalue[i] = allocate_to_stack(&state, stack, + ty->alignment, s); + } } break; + default: - FFI_ASSERT (0); - break; - } - } - else - { - memcpy (&rvalue, get_x_addr (context, 8), sizeof (UINT64)); - (closure->fun) (cif, rvalue, avalue, closure->user_data); + abort(); + } + +#if defined (__APPLE__) + if (i + 1 == cif->aarch64_nfixedargs) + { + state.ngrn = N_X_ARG_REG; + state.nsrn = N_V_ARG_REG; + state.allocating_variadic = 1; + } +#endif } + + flags = cif->flags; + if (flags & AARCH64_RET_IN_MEM) + rvalue = struct_rvalue; + + fun (cif, rvalue, avalue, user_data); + + return flags; } +#endif /* (__aarch64__) || defined(__arm64__)|| defined (_M_ARM64)*/ diff --git a/native/libffi/src/aarch64/ffitarget.h b/native/libffi/src/aarch64/ffitarget.h index 4bbced26f0..9f4d41ee07 100644 --- a/native/libffi/src/aarch64/ffitarget.h +++ b/native/libffi/src/aarch64/ffitarget.h @@ -27,37 +27,79 @@ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ #endif #ifndef LIBFFI_ASM +#ifdef __ILP32__ +#define FFI_SIZEOF_ARG 8 +#define FFI_SIZEOF_JAVA_RAW 4 +typedef unsigned long long ffi_arg; +typedef signed long long ffi_sarg; +#elif defined(_WIN32) +#define FFI_SIZEOF_ARG 8 +typedef unsigned long long ffi_arg; +typedef signed long long ffi_sarg; +#else typedef unsigned long ffi_arg; typedef signed long ffi_sarg; +#endif typedef enum ffi_abi { FFI_FIRST_ABI = 0, FFI_SYSV, + FFI_WIN64, FFI_LAST_ABI, +#if defined(_WIN32) + FFI_DEFAULT_ABI = FFI_WIN64 +#else FFI_DEFAULT_ABI = FFI_SYSV +#endif } ffi_abi; #endif /* ---- Definitions for closures ----------------------------------------- */ #define FFI_CLOSURES 1 -#define FFI_TRAMPOLINE_SIZE 36 #define FFI_NATIVE_RAW_API 0 -/* ---- Internal ---- */ +#if defined (FFI_EXEC_TRAMPOLINE_TABLE) && FFI_EXEC_TRAMPOLINE_TABLE +<<<<<<< HEAD #if defined (__APPLE__) #define FFI_TARGET_SPECIFIC_VARIADIC #define FFI_EXTRA_CIF_FIELDS unsigned aarch64_flags; unsigned aarch64_nfixedargs #else #define FFI_EXTRA_CIF_FIELDS unsigned aarch64_flags +======= +#ifdef __MACH__ +#define FFI_TRAMPOLINE_SIZE 16 +#define FFI_TRAMPOLINE_CLOSURE_OFFSET 16 +#else +#error "No trampoline table implementation" #endif -#define AARCH64_FFI_WITH_V_BIT 0 +#else +#define FFI_TRAMPOLINE_SIZE 24 +#define FFI_TRAMPOLINE_CLOSURE_OFFSET FFI_TRAMPOLINE_SIZE +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 +#endif -#define AARCH64_N_XREG 32 -#define AARCH64_N_VREG 32 -#define AARCH64_CALL_CONTEXT_SIZE (AARCH64_N_XREG * 8 + AARCH64_N_VREG * 16) +#ifdef _WIN32 +#define FFI_EXTRA_CIF_FIELDS unsigned is_variadic +#endif +#define FFI_TARGET_SPECIFIC_VARIADIC + +/* ---- Internal ---- */ + +#if defined (__APPLE__) +#define FFI_EXTRA_CIF_FIELDS unsigned aarch64_nfixedargs +#elif !defined(_WIN32) +/* iOS and Windows reserve x18 for the system. Disable Go closures until + a new static chain is chosen. */ +#define FFI_GO_CLOSURES 1 +#endif + +#ifndef _WIN32 +/* No complex type on Windows */ +#define FFI_TARGET_HAS_COMPLEX_TYPE +#endif #endif diff --git a/native/libffi/src/aarch64/internal.h b/native/libffi/src/aarch64/internal.h new file mode 100644 index 0000000000..3d4d0355ba --- /dev/null +++ b/native/libffi/src/aarch64/internal.h @@ -0,0 +1,68 @@ +/* +Permission is hereby granted, free of charge, to any person obtaining +a copy of this software and associated documentation files (the +``Software''), to deal in the Software without restriction, including +without limitation the rights to use, copy, modify, merge, publish, +distribute, sublicense, and/or sell copies of the Software, and to +permit persons to whom the Software is furnished to do so, subject to +the following conditions: + +The above copyright notice and this permission notice shall be +included in all copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, +EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF +MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. +IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY +CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, +TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE +SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ + +#define AARCH64_RET_VOID 0 +#define AARCH64_RET_INT64 1 +#define AARCH64_RET_INT128 2 + +#define AARCH64_RET_UNUSED3 3 +#define AARCH64_RET_UNUSED4 4 +#define AARCH64_RET_UNUSED5 5 +#define AARCH64_RET_UNUSED6 6 +#define AARCH64_RET_UNUSED7 7 + +/* Note that FFI_TYPE_FLOAT == 2, _DOUBLE == 3, _LONGDOUBLE == 4, + so _S4 through _Q1 are layed out as (TYPE * 4) + (4 - COUNT). */ +#define AARCH64_RET_S4 8 +#define AARCH64_RET_S3 9 +#define AARCH64_RET_S2 10 +#define AARCH64_RET_S1 11 + +#define AARCH64_RET_D4 12 +#define AARCH64_RET_D3 13 +#define AARCH64_RET_D2 14 +#define AARCH64_RET_D1 15 + +#define AARCH64_RET_Q4 16 +#define AARCH64_RET_Q3 17 +#define AARCH64_RET_Q2 18 +#define AARCH64_RET_Q1 19 + +/* Note that each of the sub-64-bit integers gets two entries. */ +#define AARCH64_RET_UINT8 20 +#define AARCH64_RET_UINT16 22 +#define AARCH64_RET_UINT32 24 + +#define AARCH64_RET_SINT8 26 +#define AARCH64_RET_SINT16 28 +#define AARCH64_RET_SINT32 30 + +#define AARCH64_RET_MASK 31 + +#define AARCH64_RET_IN_MEM (1 << 5) +#define AARCH64_RET_NEED_COPY (1 << 6) + +#define AARCH64_FLAG_ARG_V_BIT 7 +#define AARCH64_FLAG_ARG_V (1 << AARCH64_FLAG_ARG_V_BIT) +#define AARCH64_FLAG_VARARG (1 << 8) + +#define N_X_ARG_REG 8 +#define N_V_ARG_REG 8 +#define CALL_CONTEXT_SIZE (N_V_ARG_REG * 16 + N_X_ARG_REG * 8) diff --git a/native/libffi/src/aarch64/sysv.S b/native/libffi/src/aarch64/sysv.S index 169eab804e..ad4d11bdb2 100644 --- a/native/libffi/src/aarch64/sysv.S +++ b/native/libffi/src/aarch64/sysv.S @@ -19,9 +19,15 @@ CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ +#if defined(__aarch64__) || defined(__arm64__) #define LIBFFI_ASM #include #include +<<<<<<< HEAD +======= +#include +#include "internal.h" +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #ifdef HAVE_MACHINE_ASM_H #include @@ -37,6 +43,7 @@ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ #endif #endif +<<<<<<< HEAD #define cfi_adjust_cfa_offset(off) .cfi_adjust_cfa_offset off #define cfi_rel_offset(reg, off) .cfi_rel_offset reg, off #define cfi_restore(reg) .cfi_restore reg @@ -149,80 +156,205 @@ CNAME(ffi_call_SYSV): ldp q2, q3, [x21, #8*32 + 32] ldp q4, q5, [x21, #8*32 + 64] ldp q6, q7, [x21, #8*32 + 96] -1: - /* Load the core argument passing registers. */ - ldp x0, x1, [x21, #0] - ldp x2, x3, [x21, #16] - ldp x4, x5, [x21, #32] - ldp x6, x7, [x21, #48] +======= +#ifdef __AARCH64EB__ +# define BE(X) X +#else +# define BE(X) 0 +#endif - /* Don't forget x8 which may be holding the address of a return buffer. - */ - ldr x8, [x21, #8*8] +#ifdef __ILP32__ +#define PTR_REG(n) w##n +#else +#define PTR_REG(n) x##n +#endif - blr x24 +#ifdef __ILP32__ +#define PTR_SIZE 4 +#else +#define PTR_SIZE 8 +#endif - /* Save the core argument passing registers. */ - stp x0, x1, [x21, #0] - stp x2, x3, [x21, #16] - stp x4, x5, [x21, #32] - stp x6, x7, [x21, #48] +#if FFI_EXEC_TRAMPOLINE_TABLE && defined(__MACH__) && defined(HAVE_PTRAUTH) +# define BR(r) braaz r +# define BLR(r) blraaz r +#else +# define BR(r) br r +# define BLR(r) blr r +#endif - /* Note nothing useful ever comes back in x8! */ + .text + .align 4 - /* Figure out if we should touch the vector registers. */ - tbz x23, #AARCH64_FFI_WITH_V_BIT, 1f +/* ffi_call_SYSV + extern void ffi_call_SYSV (void *stack, void *frame, + void (*fn)(void), void *rvalue, + int flags, void *closure); - /* Save the vector argument passing registers. */ - stp q0, q1, [x21, #8*32 + 0] - stp q2, q3, [x21, #8*32 + 32] - stp q4, q5, [x21, #8*32 + 64] - stp q6, q7, [x21, #8*32 + 96] -1: - /* All done, unwind our stack frame. */ - ldp x21, x22, [x29, # - ffi_call_SYSV_FS] - cfi_restore (x21) - cfi_restore (x22) + Therefore on entry we have: - ldp x23, x24, [x29, # - ffi_call_SYSV_FS + 16] - cfi_restore (x23) - cfi_restore (x24) + x0 stack + x1 frame + x2 fn + x3 rvalue + x4 flags + x5 closure +*/ - mov sp, x29 - cfi_def_cfa_register (sp) + cfi_startproc +CNAME(ffi_call_SYSV): + /* Use a stack frame allocated by our caller. */ + cfi_def_cfa(x1, 32); + stp x29, x30, [x1] + mov x29, x1 + mov sp, x0 + cfi_def_cfa_register(x29) + cfi_rel_offset (x29, 0) + cfi_rel_offset (x30, 8) + + mov x9, x2 /* save fn */ + mov x8, x3 /* install structure return */ +#ifdef FFI_GO_CLOSURES + mov x18, x5 /* install static chain */ +#endif + stp x3, x4, [x29, #16] /* save rvalue and flags */ + + /* Load the vector argument passing registers, if necessary. */ + tbz w4, #AARCH64_FLAG_ARG_V_BIT, 1f + ldp q0, q1, [sp, #0] + ldp q2, q3, [sp, #32] + ldp q4, q5, [sp, #64] + ldp q6, q7, [sp, #96] +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 +1: + /* Load the core argument passing registers, including + the structure return pointer. */ + ldp x0, x1, [sp, #16*N_V_ARG_REG + 0] + ldp x2, x3, [sp, #16*N_V_ARG_REG + 16] + ldp x4, x5, [sp, #16*N_V_ARG_REG + 32] + ldp x6, x7, [sp, #16*N_V_ARG_REG + 48] - ldp x29, x30, [sp], #16 - cfi_adjust_cfa_offset (-16) - cfi_restore (x29) - cfi_restore (x30) + /* Deallocate the context, leaving the stacked arguments. */ + add sp, sp, #CALL_CONTEXT_SIZE - ret + BLR(x9) /* call fn */ +<<<<<<< HEAD .cfi_endproc #ifdef __ELF__ .size CNAME(ffi_call_SYSV), .-CNAME(ffi_call_SYSV) #endif +======= + ldp x3, x4, [x29, #16] /* reload rvalue and flags */ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 -#define ffi_closure_SYSV_FS (8 * 2 + AARCH64_CALL_CONTEXT_SIZE) + /* Partially deconstruct the stack frame. */ + mov sp, x29 + cfi_def_cfa_register (sp) + ldp x29, x30, [x29] + + /* Save the return value as directed. */ + adr x5, 0f + and w4, w4, #AARCH64_RET_MASK + add x5, x5, x4, lsl #3 + br x5 + + /* Note that each table entry is 2 insns, and thus 8 bytes. + For integer data, note that we're storing into ffi_arg + and therefore we want to extend to 64 bits; these types + have two consecutive entries allocated for them. */ + .align 4 +0: ret /* VOID */ + nop +1: str x0, [x3] /* INT64 */ + ret +2: stp x0, x1, [x3] /* INT128 */ + ret +3: brk #1000 /* UNUSED */ + ret +4: brk #1000 /* UNUSED */ + ret +5: brk #1000 /* UNUSED */ + ret +6: brk #1000 /* UNUSED */ + ret +7: brk #1000 /* UNUSED */ + ret +8: st4 { v0.s, v1.s, v2.s, v3.s }[0], [x3] /* S4 */ + ret +9: st3 { v0.s, v1.s, v2.s }[0], [x3] /* S3 */ + ret +10: stp s0, s1, [x3] /* S2 */ + ret +11: str s0, [x3] /* S1 */ + ret +12: st4 { v0.d, v1.d, v2.d, v3.d }[0], [x3] /* D4 */ + ret +13: st3 { v0.d, v1.d, v2.d }[0], [x3] /* D3 */ + ret +14: stp d0, d1, [x3] /* D2 */ + ret +15: str d0, [x3] /* D1 */ + ret +16: str q3, [x3, #48] /* Q4 */ + nop +17: str q2, [x3, #32] /* Q3 */ + nop +18: stp q0, q1, [x3] /* Q2 */ + ret +19: str q0, [x3] /* Q1 */ + ret +20: uxtb w0, w0 /* UINT8 */ + str x0, [x3] +21: ret /* reserved */ + nop +22: uxth w0, w0 /* UINT16 */ + str x0, [x3] +23: ret /* reserved */ + nop +24: mov w0, w0 /* UINT32 */ + str x0, [x3] +25: ret /* reserved */ + nop +26: sxtb x0, w0 /* SINT8 */ + str x0, [x3] +27: ret /* reserved */ + nop +28: sxth x0, w0 /* SINT16 */ + str x0, [x3] +29: ret /* reserved */ + nop +30: sxtw x0, w0 /* SINT32 */ + str x0, [x3] +31: ret /* reserved */ + nop + + cfi_endproc + + .globl CNAME(ffi_call_SYSV) + FFI_HIDDEN(CNAME(ffi_call_SYSV)) +#ifdef __ELF__ + .type CNAME(ffi_call_SYSV), #function + .size CNAME(ffi_call_SYSV), .-CNAME(ffi_call_SYSV) +#endif /* ffi_closure_SYSV Closure invocation glue. This is the low level code invoked directly by the closure trampoline to setup and call a closure. - On entry x17 points to a struct trampoline_data, x16 has been clobbered + On entry x17 points to a struct ffi_closure, x16 has been clobbered all other registers are preserved. We allocate a call context and save the argument passing registers, then invoked the generic C ffi_closure_SYSV_inner() function to do all the real work, on return we load the result passing registers back from the call context. +*/ - On entry - - extern void - ffi_closure_SYSV (struct trampoline_data *); +#define ffi_closure_SYSV_FS (8*2 + CALL_CONTEXT_SIZE + 64) +<<<<<<< HEAD struct trampoline_data { UINT64 *ffi_closure; @@ -331,3 +463,225 @@ CNAME(ffi_closure_SYSV): #ifdef __ELF__ .size CNAME(ffi_closure_SYSV), .-CNAME(ffi_closure_SYSV) #endif +======= + .align 4 +CNAME(ffi_closure_SYSV_V): + cfi_startproc + stp x29, x30, [sp, #-ffi_closure_SYSV_FS]! + cfi_adjust_cfa_offset (ffi_closure_SYSV_FS) + cfi_rel_offset (x29, 0) + cfi_rel_offset (x30, 8) + + /* Save the argument passing vector registers. */ + stp q0, q1, [sp, #16 + 0] + stp q2, q3, [sp, #16 + 32] + stp q4, q5, [sp, #16 + 64] + stp q6, q7, [sp, #16 + 96] + b 0f + cfi_endproc + + .globl CNAME(ffi_closure_SYSV_V) + FFI_HIDDEN(CNAME(ffi_closure_SYSV_V)) +#ifdef __ELF__ + .type CNAME(ffi_closure_SYSV_V), #function + .size CNAME(ffi_closure_SYSV_V), . - CNAME(ffi_closure_SYSV_V) +#endif + + .align 4 + cfi_startproc +CNAME(ffi_closure_SYSV): + stp x29, x30, [sp, #-ffi_closure_SYSV_FS]! + cfi_adjust_cfa_offset (ffi_closure_SYSV_FS) + cfi_rel_offset (x29, 0) + cfi_rel_offset (x30, 8) +0: + mov x29, sp + + /* Save the argument passing core registers. */ + stp x0, x1, [sp, #16 + 16*N_V_ARG_REG + 0] + stp x2, x3, [sp, #16 + 16*N_V_ARG_REG + 16] + stp x4, x5, [sp, #16 + 16*N_V_ARG_REG + 32] + stp x6, x7, [sp, #16 + 16*N_V_ARG_REG + 48] + + /* Load ffi_closure_inner arguments. */ + ldp PTR_REG(0), PTR_REG(1), [x17, #FFI_TRAMPOLINE_CLOSURE_OFFSET] /* load cif, fn */ + ldr PTR_REG(2), [x17, #FFI_TRAMPOLINE_CLOSURE_OFFSET+PTR_SIZE*2] /* load user_data */ +.Ldo_closure: + add x3, sp, #16 /* load context */ + add x4, sp, #ffi_closure_SYSV_FS /* load stack */ + add x5, sp, #16+CALL_CONTEXT_SIZE /* load rvalue */ + mov x6, x8 /* load struct_rval */ + bl CNAME(ffi_closure_SYSV_inner) + + /* Load the return value as directed. */ +#if FFI_EXEC_TRAMPOLINE_TABLE && defined(__MACH__) && defined(HAVE_PTRAUTH) + autiza x1 +#endif + adr x1, 0f + and w0, w0, #AARCH64_RET_MASK + add x1, x1, x0, lsl #3 + add x3, sp, #16+CALL_CONTEXT_SIZE + br x1 + + /* Note that each table entry is 2 insns, and thus 8 bytes. */ + .align 4 +0: b 99f /* VOID */ + nop +1: ldr x0, [x3] /* INT64 */ + b 99f +2: ldp x0, x1, [x3] /* INT128 */ + b 99f +3: brk #1000 /* UNUSED */ + nop +4: brk #1000 /* UNUSED */ + nop +5: brk #1000 /* UNUSED */ + nop +6: brk #1000 /* UNUSED */ + nop +7: brk #1000 /* UNUSED */ + nop +8: ldr s3, [x3, #12] /* S4 */ + nop +9: ldr s2, [x3, #8] /* S3 */ + nop +10: ldp s0, s1, [x3] /* S2 */ + b 99f +11: ldr s0, [x3] /* S1 */ + b 99f +12: ldr d3, [x3, #24] /* D4 */ + nop +13: ldr d2, [x3, #16] /* D3 */ + nop +14: ldp d0, d1, [x3] /* D2 */ + b 99f +15: ldr d0, [x3] /* D1 */ + b 99f +16: ldr q3, [x3, #48] /* Q4 */ + nop +17: ldr q2, [x3, #32] /* Q3 */ + nop +18: ldp q0, q1, [x3] /* Q2 */ + b 99f +19: ldr q0, [x3] /* Q1 */ + b 99f +20: ldrb w0, [x3, #BE(7)] /* UINT8 */ + b 99f +21: brk #1000 /* reserved */ + nop +22: ldrh w0, [x3, #BE(6)] /* UINT16 */ + b 99f +23: brk #1000 /* reserved */ + nop +24: ldr w0, [x3, #BE(4)] /* UINT32 */ + b 99f +25: brk #1000 /* reserved */ + nop +26: ldrsb x0, [x3, #BE(7)] /* SINT8 */ + b 99f +27: brk #1000 /* reserved */ + nop +28: ldrsh x0, [x3, #BE(6)] /* SINT16 */ + b 99f +29: brk #1000 /* reserved */ + nop +30: ldrsw x0, [x3, #BE(4)] /* SINT32 */ + nop +31: /* reserved */ +99: ldp x29, x30, [sp], #ffi_closure_SYSV_FS + cfi_adjust_cfa_offset (-ffi_closure_SYSV_FS) + cfi_restore (x29) + cfi_restore (x30) + ret + cfi_endproc + + .globl CNAME(ffi_closure_SYSV) + FFI_HIDDEN(CNAME(ffi_closure_SYSV)) +#ifdef __ELF__ + .type CNAME(ffi_closure_SYSV), #function + .size CNAME(ffi_closure_SYSV), . - CNAME(ffi_closure_SYSV) +#endif + +#if FFI_EXEC_TRAMPOLINE_TABLE + +#ifdef __MACH__ +#include + .align PAGE_MAX_SHIFT +CNAME(ffi_closure_trampoline_table_page): + .rept PAGE_MAX_SIZE / FFI_TRAMPOLINE_SIZE + adr x16, -PAGE_MAX_SIZE + ldp x17, x16, [x16] + BR(x16) + nop /* each entry in the trampoline config page is 2*sizeof(void*) so the trampoline itself cannot be smaller than 16 bytes */ + .endr + + .globl CNAME(ffi_closure_trampoline_table_page) + FFI_HIDDEN(CNAME(ffi_closure_trampoline_table_page)) + #ifdef __ELF__ + .type CNAME(ffi_closure_trampoline_table_page), #function + .size CNAME(ffi_closure_trampoline_table_page), . - CNAME(ffi_closure_trampoline_table_page) + #endif +#endif + +#endif /* FFI_EXEC_TRAMPOLINE_TABLE */ + +#ifdef FFI_GO_CLOSURES + .align 4 +CNAME(ffi_go_closure_SYSV_V): + cfi_startproc + stp x29, x30, [sp, #-ffi_closure_SYSV_FS]! + cfi_adjust_cfa_offset (ffi_closure_SYSV_FS) + cfi_rel_offset (x29, 0) + cfi_rel_offset (x30, 8) + + /* Save the argument passing vector registers. */ + stp q0, q1, [sp, #16 + 0] + stp q2, q3, [sp, #16 + 32] + stp q4, q5, [sp, #16 + 64] + stp q6, q7, [sp, #16 + 96] + b 0f + cfi_endproc + + .globl CNAME(ffi_go_closure_SYSV_V) + FFI_HIDDEN(CNAME(ffi_go_closure_SYSV_V)) +#ifdef __ELF__ + .type CNAME(ffi_go_closure_SYSV_V), #function + .size CNAME(ffi_go_closure_SYSV_V), . - CNAME(ffi_go_closure_SYSV_V) +#endif + + .align 4 + cfi_startproc +CNAME(ffi_go_closure_SYSV): + stp x29, x30, [sp, #-ffi_closure_SYSV_FS]! + cfi_adjust_cfa_offset (ffi_closure_SYSV_FS) + cfi_rel_offset (x29, 0) + cfi_rel_offset (x30, 8) +0: + mov x29, sp + + /* Save the argument passing core registers. */ + stp x0, x1, [sp, #16 + 16*N_V_ARG_REG + 0] + stp x2, x3, [sp, #16 + 16*N_V_ARG_REG + 16] + stp x4, x5, [sp, #16 + 16*N_V_ARG_REG + 32] + stp x6, x7, [sp, #16 + 16*N_V_ARG_REG + 48] + + /* Load ffi_closure_inner arguments. */ + ldp PTR_REG(0), PTR_REG(1), [x18, #PTR_SIZE]/* load cif, fn */ + mov x2, x18 /* load user_data */ + b .Ldo_closure + cfi_endproc + + .globl CNAME(ffi_go_closure_SYSV) + FFI_HIDDEN(CNAME(ffi_go_closure_SYSV)) +#ifdef __ELF__ + .type CNAME(ffi_go_closure_SYSV), #function + .size CNAME(ffi_go_closure_SYSV), . - CNAME(ffi_go_closure_SYSV) +#endif +#endif /* FFI_GO_CLOSURES */ +#endif /* __arm64__ */ + +#if defined __ELF__ && defined __linux__ + .section .note.GNU-stack,"",%progbits +#endif + +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 diff --git a/native/libffi/src/aarch64/win64_armasm.S b/native/libffi/src/aarch64/win64_armasm.S new file mode 100644 index 0000000000..7fc185b3d8 --- /dev/null +++ b/native/libffi/src/aarch64/win64_armasm.S @@ -0,0 +1,506 @@ +/* Copyright (c) 2009, 2010, 2011, 2012 ARM Ltd. +Permission is hereby granted, free of charge, to any person obtaining +a copy of this software and associated documentation files (the +``Software''), to deal in the Software without restriction, including +without limitation the rights to use, copy, modify, merge, publish, +distribute, sublicense, and/or sell copies of the Software, and to +permit persons to whom the Software is furnished to do so, subject to +the following conditions: +The above copyright notice and this permission notice shall be +included in all copies or substantial portions of the Software. +THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, +EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF +MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. +IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY +CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, +TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE +SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ + +#define LIBFFI_ASM +#include +#include +#include +#include "internal.h" + + OPT 2 /*disable listing */ +/* For some macros to add unwind information */ +#include "ksarm64.h" + OPT 1 /*re-enable listing */ + +#define BE(X) 0 +#define PTR_REG(n) x##n +#define PTR_SIZE 8 + + IMPORT ffi_closure_SYSV_inner + EXPORT ffi_call_SYSV + EXPORT ffi_closure_SYSV_V + EXPORT ffi_closure_SYSV + EXPORT extend_hfa_type + EXPORT compress_hfa_type +#ifdef FFI_GO_CLOSURES + EXPORT ffi_go_closure_SYSV_V + EXPORT ffi_go_closure_SYSV +#endif + + TEXTAREA, ALIGN=8 + +/* ffi_call_SYSV + extern void ffi_call_SYSV (void *stack, void *frame, + void (*fn)(void), void *rvalue, + int flags, void *closure); + Therefore on entry we have: + x0 stack + x1 frame + x2 fn + x3 rvalue + x4 flags + x5 closure +*/ + + NESTED_ENTRY ffi_call_SYSV_fake + + /* For unwind information, Windows has to store fp and lr */ + PROLOG_SAVE_REG_PAIR x29, x30, #-32! + + ALTERNATE_ENTRY ffi_call_SYSV + /* Use a stack frame allocated by our caller. */ + stp x29, x30, [x1] + mov x29, x1 + mov sp, x0 + + mov x9, x2 /* save fn */ + mov x8, x3 /* install structure return */ +#ifdef FFI_GO_CLOSURES + /*mov x18, x5 install static chain */ +#endif + stp x3, x4, [x29, #16] /* save rvalue and flags */ + + /* Load the vector argument passing registers, if necessary. */ + tbz x4, #AARCH64_FLAG_ARG_V_BIT, ffi_call_SYSV_L1 + ldp q0, q1, [sp, #0] + ldp q2, q3, [sp, #32] + ldp q4, q5, [sp, #64] + ldp q6, q7, [sp, #96] + +ffi_call_SYSV_L1 + /* Load the core argument passing registers, including + the structure return pointer. */ + ldp x0, x1, [sp, #16*N_V_ARG_REG + 0] + ldp x2, x3, [sp, #16*N_V_ARG_REG + 16] + ldp x4, x5, [sp, #16*N_V_ARG_REG + 32] + ldp x6, x7, [sp, #16*N_V_ARG_REG + 48] + + /* Deallocate the context, leaving the stacked arguments. */ + add sp, sp, #CALL_CONTEXT_SIZE + + blr x9 /* call fn */ + + ldp x3, x4, [x29, #16] /* reload rvalue and flags */ + + /* Partially deconstruct the stack frame. */ + mov sp, x29 + ldp x29, x30, [x29] + + /* Save the return value as directed. */ + adr x5, ffi_call_SYSV_return + and w4, w4, #AARCH64_RET_MASK + add x5, x5, x4, lsl #3 + br x5 + + /* Note that each table entry is 2 insns, and thus 8 bytes. + For integer data, note that we're storing into ffi_arg + and therefore we want to extend to 64 bits; these types + have two consecutive entries allocated for them. */ + ALIGN 4 +ffi_call_SYSV_return + ret /* VOID */ + nop + str x0, [x3] /* INT64 */ + ret + stp x0, x1, [x3] /* INT128 */ + ret + brk #1000 /* UNUSED */ + ret + brk #1000 /* UNUSED */ + ret + brk #1000 /* UNUSED */ + ret + brk #1000 /* UNUSED */ + ret + brk #1000 /* UNUSED */ + ret + st4 { v0.s, v1.s, v2.s, v3.s }[0], [x3] /* S4 */ + ret + st3 { v0.s, v1.s, v2.s }[0], [x3] /* S3 */ + ret + stp s0, s1, [x3] /* S2 */ + ret + str s0, [x3] /* S1 */ + ret + st4 { v0.d, v1.d, v2.d, v3.d }[0], [x3] /* D4 */ + ret + st3 { v0.d, v1.d, v2.d }[0], [x3] /* D3 */ + ret + stp d0, d1, [x3] /* D2 */ + ret + str d0, [x3] /* D1 */ + ret + str q3, [x3, #48] /* Q4 */ + nop + str q2, [x3, #32] /* Q3 */ + nop + stp q0, q1, [x3] /* Q2 */ + ret + str q0, [x3] /* Q1 */ + ret + uxtb w0, w0 /* UINT8 */ + str x0, [x3] + ret /* reserved */ + nop + uxth w0, w0 /* UINT16 */ + str x0, [x3] + ret /* reserved */ + nop + mov w0, w0 /* UINT32 */ + str x0, [x3] + ret /* reserved */ + nop + sxtb x0, w0 /* SINT8 */ + str x0, [x3] + ret /* reserved */ + nop + sxth x0, w0 /* SINT16 */ + str x0, [x3] + ret /* reserved */ + nop + sxtw x0, w0 /* SINT32 */ + str x0, [x3] + ret /* reserved */ + nop + + + NESTED_END ffi_call_SYSV_fake + + +/* ffi_closure_SYSV + Closure invocation glue. This is the low level code invoked directly by + the closure trampoline to setup and call a closure. + On entry x17 points to a struct ffi_closure, x16 has been clobbered + all other registers are preserved. + We allocate a call context and save the argument passing registers, + then invoked the generic C ffi_closure_SYSV_inner() function to do all + the real work, on return we load the result passing registers back from + the call context. +*/ + +#define ffi_closure_SYSV_FS (8*2 + CALL_CONTEXT_SIZE + 64) + + NESTED_ENTRY ffi_closure_SYSV_V + PROLOG_SAVE_REG_PAIR x29, x30, #-ffi_closure_SYSV_FS! + + /* Save the argument passing vector registers. */ + stp q0, q1, [sp, #16 + 0] + stp q2, q3, [sp, #16 + 32] + stp q4, q5, [sp, #16 + 64] + stp q6, q7, [sp, #16 + 96] + + b ffi_closure_SYSV_save_argument + NESTED_END ffi_closure_SYSV_V + + NESTED_ENTRY ffi_closure_SYSV + PROLOG_SAVE_REG_PAIR x29, x30, #-ffi_closure_SYSV_FS! + +ffi_closure_SYSV_save_argument + /* Save the argument passing core registers. */ + stp x0, x1, [sp, #16 + 16*N_V_ARG_REG + 0] + stp x2, x3, [sp, #16 + 16*N_V_ARG_REG + 16] + stp x4, x5, [sp, #16 + 16*N_V_ARG_REG + 32] + stp x6, x7, [sp, #16 + 16*N_V_ARG_REG + 48] + + /* Load ffi_closure_inner arguments. */ + ldp PTR_REG(0), PTR_REG(1), [x17, #FFI_TRAMPOLINE_CLOSURE_OFFSET] /* load cif, fn */ + ldr PTR_REG(2), [x17, #FFI_TRAMPOLINE_CLOSURE_OFFSET+PTR_SIZE*2] /* load user_data */ + +do_closure + add x3, sp, #16 /* load context */ + add x4, sp, #ffi_closure_SYSV_FS /* load stack */ + add x5, sp, #16+CALL_CONTEXT_SIZE /* load rvalue */ + mov x6, x8 /* load struct_rval */ + + bl ffi_closure_SYSV_inner + + /* Load the return value as directed. */ + adr x1, ffi_closure_SYSV_return_base + and w0, w0, #AARCH64_RET_MASK + add x1, x1, x0, lsl #3 + add x3, sp, #16+CALL_CONTEXT_SIZE + br x1 + + /* Note that each table entry is 2 insns, and thus 8 bytes. */ + ALIGN 8 +ffi_closure_SYSV_return_base + b ffi_closure_SYSV_epilog /* VOID */ + nop + ldr x0, [x3] /* INT64 */ + b ffi_closure_SYSV_epilog + ldp x0, x1, [x3] /* INT128 */ + b ffi_closure_SYSV_epilog + brk #1000 /* UNUSED */ + nop + brk #1000 /* UNUSED */ + nop + brk #1000 /* UNUSED */ + nop + brk #1000 /* UNUSED */ + nop + brk #1000 /* UNUSED */ + nop + ldr s3, [x3, #12] /* S4 */ + nop + ldr s2, [x3, #8] /* S3 */ + nop + ldp s0, s1, [x3] /* S2 */ + b ffi_closure_SYSV_epilog + ldr s0, [x3] /* S1 */ + b ffi_closure_SYSV_epilog + ldr d3, [x3, #24] /* D4 */ + nop + ldr d2, [x3, #16] /* D3 */ + nop + ldp d0, d1, [x3] /* D2 */ + b ffi_closure_SYSV_epilog + ldr d0, [x3] /* D1 */ + b ffi_closure_SYSV_epilog + ldr q3, [x3, #48] /* Q4 */ + nop + ldr q2, [x3, #32] /* Q3 */ + nop + ldp q0, q1, [x3] /* Q2 */ + b ffi_closure_SYSV_epilog + ldr q0, [x3] /* Q1 */ + b ffi_closure_SYSV_epilog + ldrb w0, [x3, #BE(7)] /* UINT8 */ + b ffi_closure_SYSV_epilog + brk #1000 /* reserved */ + nop + ldrh w0, [x3, #BE(6)] /* UINT16 */ + b ffi_closure_SYSV_epilog + brk #1000 /* reserved */ + nop + ldr w0, [x3, #BE(4)] /* UINT32 */ + b ffi_closure_SYSV_epilog + brk #1000 /* reserved */ + nop + ldrsb x0, [x3, #BE(7)] /* SINT8 */ + b ffi_closure_SYSV_epilog + brk #1000 /* reserved */ + nop + ldrsh x0, [x3, #BE(6)] /* SINT16 */ + b ffi_closure_SYSV_epilog + brk #1000 /* reserved */ + nop + ldrsw x0, [x3, #BE(4)] /* SINT32 */ + nop + /* reserved */ + +ffi_closure_SYSV_epilog + EPILOG_RESTORE_REG_PAIR x29, x30, #ffi_closure_SYSV_FS! + EPILOG_RETURN + NESTED_END ffi_closure_SYSV + + +#ifdef FFI_GO_CLOSURES + NESTED_ENTRY ffi_go_closure_SYSV_V + PROLOG_SAVE_REG_PAIR x29, x30, #-ffi_closure_SYSV_FS! + + /* Save the argument passing vector registers. */ + stp q0, q1, [sp, #16 + 0] + stp q2, q3, [sp, #16 + 32] + stp q4, q5, [sp, #16 + 64] + stp q6, q7, [sp, #16 + 96] + b ffi_go_closure_SYSV_save_argument + NESTED_END ffi_go_closure_SYSV_V + + NESTED_ENTRY ffi_go_closure_SYSV + PROLOG_SAVE_REG_PAIR x29, x30, #-ffi_closure_SYSV_FS! + +ffi_go_closure_SYSV_save_argument + /* Save the argument passing core registers. */ + stp x0, x1, [sp, #16 + 16*N_V_ARG_REG + 0] + stp x2, x3, [sp, #16 + 16*N_V_ARG_REG + 16] + stp x4, x5, [sp, #16 + 16*N_V_ARG_REG + 32] + stp x6, x7, [sp, #16 + 16*N_V_ARG_REG + 48] + + /* Load ffi_closure_inner arguments. */ + ldp PTR_REG(0), PTR_REG(1), [x18, #PTR_SIZE]/* load cif, fn */ + mov x2, x18 /* load user_data */ + b do_closure + NESTED_END ffi_go_closure_SYSV + +#endif /* FFI_GO_CLOSURES */ + + +/* void extend_hfa_type (void *dest, void *src, int h) */ + + LEAF_ENTRY extend_hfa_type + + adr x3, extend_hfa_type_jump_base + and w2, w2, #AARCH64_RET_MASK + sub x2, x2, #AARCH64_RET_S4 + add x3, x3, x2, lsl #4 + br x3 + + ALIGN 4 +extend_hfa_type_jump_base + ldp s16, s17, [x1] /* S4 */ + ldp s18, s19, [x1, #8] + b extend_hfa_type_store_4 + nop + + ldp s16, s17, [x1] /* S3 */ + ldr s18, [x1, #8] + b extend_hfa_type_store_3 + nop + + ldp s16, s17, [x1] /* S2 */ + b extend_hfa_type_store_2 + nop + nop + + ldr s16, [x1] /* S1 */ + b extend_hfa_type_store_1 + nop + nop + + ldp d16, d17, [x1] /* D4 */ + ldp d18, d19, [x1, #16] + b extend_hfa_type_store_4 + nop + + ldp d16, d17, [x1] /* D3 */ + ldr d18, [x1, #16] + b extend_hfa_type_store_3 + nop + + ldp d16, d17, [x1] /* D2 */ + b extend_hfa_type_store_2 + nop + nop + + ldr d16, [x1] /* D1 */ + b extend_hfa_type_store_1 + nop + nop + + ldp q16, q17, [x1] /* Q4 */ + ldp q18, q19, [x1, #16] + b extend_hfa_type_store_4 + nop + + ldp q16, q17, [x1] /* Q3 */ + ldr q18, [x1, #16] + b extend_hfa_type_store_3 + nop + + ldp q16, q17, [x1] /* Q2 */ + b extend_hfa_type_store_2 + nop + nop + + ldr q16, [x1] /* Q1 */ + b extend_hfa_type_store_1 + +extend_hfa_type_store_4 + str q19, [x0, #48] +extend_hfa_type_store_3 + str q18, [x0, #32] +extend_hfa_type_store_2 + str q17, [x0, #16] +extend_hfa_type_store_1 + str q16, [x0] + ret + + LEAF_END extend_hfa_type + + +/* void compress_hfa_type (void *dest, void *reg, int h) */ + + LEAF_ENTRY compress_hfa_type + + adr x3, compress_hfa_type_jump_base + and w2, w2, #AARCH64_RET_MASK + sub x2, x2, #AARCH64_RET_S4 + add x3, x3, x2, lsl #4 + br x3 + + ALIGN 4 +compress_hfa_type_jump_base + ldp q16, q17, [x1] /* S4 */ + ldp q18, q19, [x1, #32] + st4 { v16.s, v17.s, v18.s, v19.s }[0], [x0] + ret + + ldp q16, q17, [x1] /* S3 */ + ldr q18, [x1, #32] + st3 { v16.s, v17.s, v18.s }[0], [x0] + ret + + ldp q16, q17, [x1] /* S2 */ + st2 { v16.s, v17.s }[0], [x0] + ret + nop + + ldr q16, [x1] /* S1 */ + st1 { v16.s }[0], [x0] + ret + nop + + ldp q16, q17, [x1] /* D4 */ + ldp q18, q19, [x1, #32] + st4 { v16.d, v17.d, v18.d, v19.d }[0], [x0] + ret + + ldp q16, q17, [x1] /* D3 */ + ldr q18, [x1, #32] + st3 { v16.d, v17.d, v18.d }[0], [x0] + ret + + ldp q16, q17, [x1] /* D2 */ + st2 { v16.d, v17.d }[0], [x0] + ret + nop + + ldr q16, [x1] /* D1 */ + st1 { v16.d }[0], [x0] + ret + nop + + ldp q16, q17, [x1] /* Q4 */ + ldp q18, q19, [x1, #32] + b compress_hfa_type_store_q4 + nop + + ldp q16, q17, [x1] /* Q3 */ + ldr q18, [x1, #32] + b compress_hfa_type_store_q3 + nop + + ldp q16, q17, [x1] /* Q2 */ + stp q16, q17, [x0] + ret + nop + + ldr q16, [x1] /* Q1 */ + str q16, [x0] + ret + +compress_hfa_type_store_q4 + str q19, [x0, #48] +compress_hfa_type_store_q3 + str q18, [x0, #32] + stp q16, q17, [x0] + ret + + LEAF_END compress_hfa_type + + END \ No newline at end of file diff --git a/native/libffi/src/alpha/ffi.c b/native/libffi/src/alpha/ffi.c index 192f691c4a..7a95e9707c 100644 --- a/native/libffi/src/alpha/ffi.c +++ b/native/libffi/src/alpha/ffi.c @@ -1,8 +1,8 @@ /* ----------------------------------------------------------------------- ffi.c - Copyright (c) 2012 Anthony Green - Copyright (c) 1998, 2001, 2007, 2008 Red Hat, Inc. - - Alpha Foreign Function Interface + Copyright (c) 1998, 2001, 2007, 2008 Red Hat, Inc. + + Alpha Foreign Function Interface Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the @@ -28,6 +28,7 @@ #include #include #include +#include "internal.h" /* Force FFI_TYPE_LONGDOUBLE to be different than FFI_TYPE_DOUBLE; all further uses in this file will refer to the 128-bit type. */ @@ -40,135 +41,286 @@ # define FFI_TYPE_LONGDOUBLE 4 #endif -extern void ffi_call_osf(void *, unsigned long, unsigned, void *, void (*)(void)) - FFI_HIDDEN; +extern void ffi_call_osf(void *stack, void *frame, unsigned flags, + void *raddr, void (*fn)(void), void *closure) + FFI_HIDDEN; extern void ffi_closure_osf(void) FFI_HIDDEN; +extern void ffi_go_closure_osf(void) FFI_HIDDEN; +/* Promote a float value to its in-register double representation. + Unlike actually casting to double, this does not trap on NaN. */ +static inline UINT64 lds(void *ptr) +{ + UINT64 ret; + asm("lds %0,%1" : "=f"(ret) : "m"(*(UINT32 *)ptr)); + return ret; +} -ffi_status +/* And the reverse. */ +static inline void sts(void *ptr, UINT64 val) +{ + asm("sts %1,%0" : "=m"(*(UINT32 *)ptr) : "f"(val)); +} + +ffi_status FFI_HIDDEN ffi_prep_cif_machdep(ffi_cif *cif) { - /* Adjust cif->bytes to represent a minimum 6 words for the temporary - register argument loading area. */ - if (cif->bytes < 6*FFI_SIZEOF_ARG) - cif->bytes = 6*FFI_SIZEOF_ARG; + size_t bytes = 0; + int flags, i, avn; + ffi_type *rtype, *itype; + + if (cif->abi != FFI_OSF) + return FFI_BAD_ABI; + + /* Compute the size of the argument area. */ + for (i = 0, avn = cif->nargs; i < avn; i++) + { + itype = cif->arg_types[i]; + switch (itype->type) + { + case FFI_TYPE_INT: + case FFI_TYPE_SINT8: + case FFI_TYPE_UINT8: + case FFI_TYPE_SINT16: + case FFI_TYPE_UINT16: + case FFI_TYPE_SINT32: + case FFI_TYPE_UINT32: + case FFI_TYPE_SINT64: + case FFI_TYPE_UINT64: + case FFI_TYPE_POINTER: + case FFI_TYPE_FLOAT: + case FFI_TYPE_DOUBLE: + case FFI_TYPE_LONGDOUBLE: + /* All take one 8 byte slot. */ + bytes += 8; + break; + + case FFI_TYPE_VOID: + case FFI_TYPE_STRUCT: + /* Passed by value in N slots. */ + bytes += FFI_ALIGN(itype->size, FFI_SIZEOF_ARG); + break; + + case FFI_TYPE_COMPLEX: + /* _Complex long double passed by reference; others in 2 slots. */ + if (itype->elements[0]->type == FFI_TYPE_LONGDOUBLE) + bytes += 8; + else + bytes += 16; + break; + + default: + abort(); + } + } /* Set the return type flag */ - switch (cif->rtype->type) + rtype = cif->rtype; + switch (rtype->type) { - case FFI_TYPE_STRUCT: + case FFI_TYPE_VOID: + flags = ALPHA_FLAGS(ALPHA_ST_VOID, ALPHA_LD_VOID); + break; + case FFI_TYPE_INT: + case FFI_TYPE_UINT32: + case FFI_TYPE_SINT32: + flags = ALPHA_FLAGS(ALPHA_ST_INT, ALPHA_LD_INT32); + break; case FFI_TYPE_FLOAT: + flags = ALPHA_FLAGS(ALPHA_ST_FLOAT, ALPHA_LD_FLOAT); + break; case FFI_TYPE_DOUBLE: - cif->flags = cif->rtype->type; + flags = ALPHA_FLAGS(ALPHA_ST_DOUBLE, ALPHA_LD_DOUBLE); + break; + case FFI_TYPE_UINT8: + flags = ALPHA_FLAGS(ALPHA_ST_INT, ALPHA_LD_UINT8); + break; + case FFI_TYPE_SINT8: + flags = ALPHA_FLAGS(ALPHA_ST_INT, ALPHA_LD_SINT8); + break; + case FFI_TYPE_UINT16: + flags = ALPHA_FLAGS(ALPHA_ST_INT, ALPHA_LD_UINT16); + break; + case FFI_TYPE_SINT16: + flags = ALPHA_FLAGS(ALPHA_ST_INT, ALPHA_LD_SINT16); + break; + case FFI_TYPE_UINT64: + case FFI_TYPE_SINT64: + case FFI_TYPE_POINTER: + flags = ALPHA_FLAGS(ALPHA_ST_INT, ALPHA_LD_INT64); break; - case FFI_TYPE_LONGDOUBLE: - /* 128-bit long double is returned in memory, like a struct. */ - cif->flags = FFI_TYPE_STRUCT; + case FFI_TYPE_STRUCT: + /* Passed in memory, with a hidden pointer. */ + flags = ALPHA_RET_IN_MEM; break; - - default: - cif->flags = FFI_TYPE_INT; + case FFI_TYPE_COMPLEX: + itype = rtype->elements[0]; + switch (itype->type) + { + case FFI_TYPE_FLOAT: + flags = ALPHA_FLAGS(ALPHA_ST_CPLXF, ALPHA_LD_CPLXF); + break; + case FFI_TYPE_DOUBLE: + flags = ALPHA_FLAGS(ALPHA_ST_CPLXD, ALPHA_LD_CPLXD); + break; + default: + if (rtype->size <= 8) + flags = ALPHA_FLAGS(ALPHA_ST_INT, ALPHA_LD_INT64); + else + flags = ALPHA_RET_IN_MEM; + break; + } break; + default: + abort(); } - + cif->flags = flags; + + /* Include the hidden structure pointer in args requirement. */ + if (flags == ALPHA_RET_IN_MEM) + bytes += 8; + /* Minimum size is 6 slots, so that ffi_call_osf can pop them. */ + if (bytes < 6*8) + bytes = 6*8; + cif->bytes = bytes; + return FFI_OK; } +static unsigned long +extend_basic_type(void *valp, int type, int argn) +{ + switch (type) + { + case FFI_TYPE_SINT8: + return *(SINT8 *)valp; + case FFI_TYPE_UINT8: + return *(UINT8 *)valp; + case FFI_TYPE_SINT16: + return *(SINT16 *)valp; + case FFI_TYPE_UINT16: + return *(UINT16 *)valp; + + case FFI_TYPE_FLOAT: + if (argn < 6) + return lds(valp); + /* FALLTHRU */ + + case FFI_TYPE_INT: + case FFI_TYPE_SINT32: + case FFI_TYPE_UINT32: + /* Note that unsigned 32-bit quantities are sign extended. */ + return *(SINT32 *)valp; + + case FFI_TYPE_SINT64: + case FFI_TYPE_UINT64: + case FFI_TYPE_POINTER: + case FFI_TYPE_DOUBLE: + return *(UINT64 *)valp; -void -ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) + default: + abort(); + } +} + +static void +ffi_call_int (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) { - unsigned long *stack, *argp; - long i, avn; + unsigned long *argp; + long i, avn, argn, flags = cif->flags; ffi_type **arg_types; - + void *frame; + /* If the return value is a struct and we don't have a return value address then we need to make one. */ - if (rvalue == NULL && cif->flags == FFI_TYPE_STRUCT) + if (rvalue == NULL && flags == ALPHA_RET_IN_MEM) rvalue = alloca(cif->rtype->size); /* Allocate the space for the arguments, plus 4 words of temp space for ffi_call_osf. */ - argp = stack = alloca(cif->bytes + 4*FFI_SIZEOF_ARG); + argp = frame = alloca(cif->bytes + 4*FFI_SIZEOF_ARG); + frame += cif->bytes; - if (cif->flags == FFI_TYPE_STRUCT) - *(void **) argp++ = rvalue; + argn = 0; + if (flags == ALPHA_RET_IN_MEM) + argp[argn++] = (unsigned long)rvalue; - i = 0; avn = cif->nargs; arg_types = cif->arg_types; - while (i < avn) + for (i = 0, avn = cif->nargs; i < avn; i++) { - size_t size = (*arg_types)->size; + ffi_type *ty = arg_types[i]; + void *valp = avalue[i]; + int type = ty->type; + size_t size; - switch ((*arg_types)->type) + switch (type) { + case FFI_TYPE_INT: case FFI_TYPE_SINT8: - *(SINT64 *) argp = *(SINT8 *)(* avalue); - break; - case FFI_TYPE_UINT8: - *(SINT64 *) argp = *(UINT8 *)(* avalue); - break; - case FFI_TYPE_SINT16: - *(SINT64 *) argp = *(SINT16 *)(* avalue); - break; - case FFI_TYPE_UINT16: - *(SINT64 *) argp = *(UINT16 *)(* avalue); - break; - case FFI_TYPE_SINT32: case FFI_TYPE_UINT32: - /* Note that unsigned 32-bit quantities are sign extended. */ - *(SINT64 *) argp = *(SINT32 *)(* avalue); - break; - case FFI_TYPE_SINT64: case FFI_TYPE_UINT64: case FFI_TYPE_POINTER: - *(UINT64 *) argp = *(UINT64 *)(* avalue); - break; - case FFI_TYPE_FLOAT: - if (argp - stack < 6) - { - /* Note the conversion -- all the fp regs are loaded as - doubles. The in-register format is the same. */ - *(double *) argp = *(float *)(* avalue); - } - else - *(float *) argp = *(float *)(* avalue); - break; - case FFI_TYPE_DOUBLE: - *(double *) argp = *(double *)(* avalue); + argp[argn] = extend_basic_type(valp, type, argn); + argn++; break; case FFI_TYPE_LONGDOUBLE: - /* 128-bit long double is passed by reference. */ - *(long double **) argp = (long double *)(* avalue); - size = sizeof (long double *); + by_reference: + /* Note that 128-bit long double is passed by reference. */ + argp[argn++] = (unsigned long)valp; break; + case FFI_TYPE_VOID: case FFI_TYPE_STRUCT: - memcpy(argp, *avalue, (*arg_types)->size); + size = ty->size; + memcpy(argp + argn, valp, size); + argn += FFI_ALIGN(size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; + break; + + case FFI_TYPE_COMPLEX: + type = ty->elements[0]->type; + if (type == FFI_TYPE_LONGDOUBLE) + goto by_reference; + + /* Most complex types passed as two separate arguments. */ + size = ty->elements[0]->size; + argp[argn] = extend_basic_type(valp, type, argn); + argp[argn + 1] = extend_basic_type(valp + size, type, argn + 1); + argn += 2; break; default: - FFI_ASSERT(0); + abort(); } - - argp += ALIGN(size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; - i++, arg_types++, avalue++; } - ffi_call_osf(stack, cif->bytes, cif->flags, rvalue, fn); + flags = (flags >> ALPHA_ST_SHIFT) & 0xff; + ffi_call_osf(argp, frame, flags, rvalue, fn, closure); +} + +void +ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +{ + ffi_call_int(cif, fn, rvalue, avalue, NULL); } +void +ffi_call_go (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) +{ + ffi_call_int(cif, fn, rvalue, avalue, closure); +} ffi_status ffi_prep_closure_loc (ffi_closure* closure, @@ -204,39 +356,56 @@ ffi_prep_closure_loc (ffi_closure* closure, return FFI_OK; } +ffi_status +ffi_prep_go_closure (ffi_go_closure* closure, + ffi_cif* cif, + void (*fun)(ffi_cif*, void*, void**, void*)) +{ + if (cif->abi != FFI_OSF) + return FFI_BAD_ABI; + + closure->tramp = (void *)ffi_go_closure_osf; + closure->cif = cif; + closure->fun = fun; + + return FFI_OK; +} long FFI_HIDDEN -ffi_closure_osf_inner(ffi_closure *closure, void *rvalue, unsigned long *argp) +ffi_closure_osf_inner (ffi_cif *cif, + void (*fun)(ffi_cif*, void*, void**, void*), + void *user_data, + void *rvalue, unsigned long *argp) { - ffi_cif *cif; void **avalue; ffi_type **arg_types; - long i, avn, argn; + long i, avn, argn, flags; - cif = closure->cif; avalue = alloca(cif->nargs * sizeof(void *)); - + flags = cif->flags; argn = 0; /* Copy the caller's structure return address to that the closure returns the data directly to the caller. */ - if (cif->flags == FFI_TYPE_STRUCT) + if (flags == ALPHA_RET_IN_MEM) { rvalue = (void *) argp[0]; argn = 1; } - i = 0; - avn = cif->nargs; arg_types = cif->arg_types; - + /* Grab the addresses of the arguments from the stack frame. */ - while (i < avn) + for (i = 0, avn = cif->nargs; i < avn; i++) { - size_t size = arg_types[i]->size; + ffi_type *ty = arg_types[i]; + int type = ty->type; + void *valp = &argp[argn]; + size_t size; - switch (arg_types[i]->type) + switch (type) { + case FFI_TYPE_INT: case FFI_TYPE_SINT8: case FFI_TYPE_UINT8: case FFI_TYPE_SINT16: @@ -246,43 +415,107 @@ ffi_closure_osf_inner(ffi_closure *closure, void *rvalue, unsigned long *argp) case FFI_TYPE_SINT64: case FFI_TYPE_UINT64: case FFI_TYPE_POINTER: + argn += 1; + break; + + case FFI_TYPE_VOID: case FFI_TYPE_STRUCT: - avalue[i] = &argp[argn]; + size = ty->size; + argn += FFI_ALIGN(size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; break; case FFI_TYPE_FLOAT: + /* Floats coming from registers need conversion from double + back to float format. */ if (argn < 6) { - /* Floats coming from registers need conversion from double - back to float format. */ - *(float *)&argp[argn - 6] = *(double *)&argp[argn - 6]; - avalue[i] = &argp[argn - 6]; + valp = &argp[argn - 6]; + sts(valp, argp[argn - 6]); } - else - avalue[i] = &argp[argn]; + argn += 1; break; case FFI_TYPE_DOUBLE: - avalue[i] = &argp[argn - (argn < 6 ? 6 : 0)]; + if (argn < 6) + valp = &argp[argn - 6]; + argn += 1; break; case FFI_TYPE_LONGDOUBLE: + by_reference: /* 128-bit long double is passed by reference. */ - avalue[i] = (long double *) argp[argn]; - size = sizeof (long double *); + valp = (void *)argp[argn]; + argn += 1; + break; + + case FFI_TYPE_COMPLEX: + type = ty->elements[0]->type; + switch (type) + { + case FFI_TYPE_SINT64: + case FFI_TYPE_UINT64: + /* Passed as separate arguments, but they wind up sequential. */ + break; + + case FFI_TYPE_INT: + case FFI_TYPE_SINT8: + case FFI_TYPE_UINT8: + case FFI_TYPE_SINT16: + case FFI_TYPE_UINT16: + case FFI_TYPE_SINT32: + case FFI_TYPE_UINT32: + /* Passed as separate arguments. Disjoint, but there's room + enough in one slot to hold the pair. */ + size = ty->elements[0]->size; + memcpy(valp + size, valp + 8, size); + break; + + case FFI_TYPE_FLOAT: + /* Passed as separate arguments. Disjoint, and each piece + may need conversion back to float. */ + if (argn < 6) + { + valp = &argp[argn - 6]; + sts(valp, argp[argn - 6]); + } + if (argn + 1 < 6) + sts(valp + 4, argp[argn + 1 - 6]); + else + *(UINT32 *)(valp + 4) = argp[argn + 1]; + break; + + case FFI_TYPE_DOUBLE: + /* Passed as separate arguments. Only disjoint if one part + is in fp regs and the other is on the stack. */ + if (argn < 5) + valp = &argp[argn - 6]; + else if (argn == 5) + { + valp = alloca(16); + ((UINT64 *)valp)[0] = argp[5 - 6]; + ((UINT64 *)valp)[1] = argp[6]; + } + break; + + case FFI_TYPE_LONGDOUBLE: + goto by_reference; + + default: + abort(); + } + argn += 2; break; default: abort (); } - argn += ALIGN(size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; - i++; + avalue[i] = valp; } /* Invoke the closure. */ - closure->fun (cif, rvalue, avalue, closure->user_data); + fun (cif, rvalue, avalue, user_data); /* Tell ffi_closure_osf how to perform return type promotions. */ - return cif->rtype->type; + return (flags >> ALPHA_LD_SHIFT) & 0xff; } diff --git a/native/libffi/src/alpha/ffitarget.h b/native/libffi/src/alpha/ffitarget.h index af145bce83..a02dbd04f6 100644 --- a/native/libffi/src/alpha/ffitarget.h +++ b/native/libffi/src/alpha/ffitarget.h @@ -44,9 +44,13 @@ typedef enum ffi_abi { } ffi_abi; #endif +#define FFI_TARGET_SPECIFIC_STACK_SPACE_ALLOCATION +#define FFI_TARGET_HAS_COMPLEX_TYPE + /* ---- Definitions for closures ----------------------------------------- */ #define FFI_CLOSURES 1 +#define FFI_GO_CLOSURES 1 #define FFI_TRAMPOLINE_SIZE 24 #define FFI_NATIVE_RAW_API 0 diff --git a/native/libffi/src/alpha/internal.h b/native/libffi/src/alpha/internal.h new file mode 100644 index 0000000000..44da1922bb --- /dev/null +++ b/native/libffi/src/alpha/internal.h @@ -0,0 +1,23 @@ +#define ALPHA_ST_VOID 0 +#define ALPHA_ST_INT 1 +#define ALPHA_ST_FLOAT 2 +#define ALPHA_ST_DOUBLE 3 +#define ALPHA_ST_CPLXF 4 +#define ALPHA_ST_CPLXD 5 + +#define ALPHA_LD_VOID 0 +#define ALPHA_LD_INT64 1 +#define ALPHA_LD_INT32 2 +#define ALPHA_LD_UINT16 3 +#define ALPHA_LD_SINT16 4 +#define ALPHA_LD_UINT8 5 +#define ALPHA_LD_SINT8 6 +#define ALPHA_LD_FLOAT 7 +#define ALPHA_LD_DOUBLE 8 +#define ALPHA_LD_CPLXF 9 +#define ALPHA_LD_CPLXD 10 + +#define ALPHA_ST_SHIFT 0 +#define ALPHA_LD_SHIFT 8 +#define ALPHA_RET_IN_MEM 0x10000 +#define ALPHA_FLAGS(S, L) (((L) << ALPHA_LD_SHIFT) | (S)) diff --git a/native/libffi/src/alpha/osf.S b/native/libffi/src/alpha/osf.S index 6b9f4dfa03..b0318282a0 100644 --- a/native/libffi/src/alpha/osf.S +++ b/native/libffi/src/alpha/osf.S @@ -1,7 +1,7 @@ /* ----------------------------------------------------------------------- - osf.S - Copyright (c) 1998, 2001, 2007, 2008, 2011 Red Hat - - Alpha/OSF Foreign Function Interface + osf.S - Copyright (c) 1998, 2001, 2007, 2008, 2011, 2014 Red Hat + + Alpha/OSF Foreign Function Interface Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the @@ -24,40 +24,49 @@ DEALINGS IN THE SOFTWARE. ----------------------------------------------------------------------- */ -#define LIBFFI_ASM +#define LIBFFI_ASM #include #include +#include +#include "internal.h" .arch ev6 .text -/* ffi_call_osf (void *args, unsigned long bytes, unsigned flags, - void *raddr, void (*fnaddr)(void)); +/* Aid in building a direct addressed jump table, 4 insns per entry. */ +.macro E index + .align 4 + .org 99b + \index * 16 +.endm + +/* ffi_call_osf (void *stack, void *frame, unsigned flags, + void *raddr, void (*fnaddr)(void), void *closure) - Bit o trickiness here -- ARGS+BYTES is the base of the stack frame + Bit o trickiness here -- FRAME is the base of the stack frame for this function. This has been allocated by ffi_call. We also deallocate some of the stack that has been alloca'd. */ - .align 3 + .align 4 .globl ffi_call_osf .ent ffi_call_osf FFI_HIDDEN(ffi_call_osf) ffi_call_osf: - .frame $15, 32, $26, 0 - .mask 0x4008000, -32 -$LFB1: - addq $16,$17,$1 + cfi_startproc + cfi_def_cfa($17, 32) mov $16, $30 - stq $26, 0($1) - stq $15, 8($1) - stq $18, 16($1) - mov $1, $15 -$LCFI1: + stq $26, 0($17) + stq $15, 8($17) + mov $17, $15 .prologue 0 + cfi_def_cfa_register($15) + cfi_rel_offset($26, 0) + cfi_rel_offset($15, 8) - stq $19, 24($1) - mov $20, $27 + stq $18, 16($17) # save flags into frame + stq $19, 24($17) # save rvalue into frame + mov $20, $27 # fn into place for call + mov $21, $1 # closure into static chain # Load up all of the (potential) argument registers. ldq $16, 0($30) @@ -77,311 +86,197 @@ $LCFI1: lda $30, 48($30) jsr $26, ($27), 0 - ldgp $29, 0($26) - - # If the return value pointer is NULL, assume no return value. - ldq $19, 24($15) - ldq $18, 16($15) +0: + ldah $29, 0($26) !gpdisp!1 + ldq $2, 24($15) # reload rvalue + lda $29, 0($29) !gpdisp!1 + ldq $3, 16($15) # reload flags + lda $1, 99f-0b($26) ldq $26, 0($15) -$LCFI2: - beq $19, $noretval - - # Store the return value out in the proper type. - cmpeq $18, FFI_TYPE_INT, $1 - bne $1, $retint - cmpeq $18, FFI_TYPE_FLOAT, $2 - bne $2, $retfloat - cmpeq $18, FFI_TYPE_DOUBLE, $3 - bne $3, $retdouble - - .align 3 -$noretval: ldq $15, 8($15) - ret + cfi_restore($26) + cfi_restore($15) + cfi_def_cfa($sp, 0) + cmoveq $2, ALPHA_ST_VOID, $3 # mash null rvalue to void + addq $3, $3, $3 + s8addq $3, $1, $1 # 99f + stcode * 16 + jmp $31, ($1), $st_int .align 4 -$retint: - stq $0, 0($19) - nop - ldq $15, 8($15) +99: +E ALPHA_ST_VOID ret - - .align 4 -$retfloat: - sts $f0, 0($19) - nop - ldq $15, 8($15) +E ALPHA_ST_INT +$st_int: + stq $0, 0($2) ret - - .align 4 -$retdouble: - stt $f0, 0($19) - nop - ldq $15, 8($15) +E ALPHA_ST_FLOAT + sts $f0, 0($2) + ret +E ALPHA_ST_DOUBLE + stt $f0, 0($2) + ret +E ALPHA_ST_CPLXF + sts $f0, 0($2) + sts $f1, 4($2) + ret +E ALPHA_ST_CPLXD + stt $f0, 0($2) + stt $f1, 8($2) ret -$LFE1: + cfi_endproc .end ffi_call_osf /* ffi_closure_osf(...) Receives the closure argument in $1. */ - .align 3 +#define CLOSURE_FS (16*8) + + .align 4 + .globl ffi_go_closure_osf + .ent ffi_go_closure_osf + FFI_HIDDEN(ffi_go_closure_osf) + +ffi_go_closure_osf: + cfi_startproc + ldgp $29, 0($27) + subq $30, CLOSURE_FS, $30 + cfi_adjust_cfa_offset(CLOSURE_FS) + stq $26, 0($30) + .prologue 1 + cfi_rel_offset($26, 0) + + stq $16, 10*8($30) + stq $17, 11*8($30) + stq $18, 12*8($30) + + ldq $16, 8($1) # load cif + ldq $17, 16($1) # load fun + mov $1, $18 # closure is user_data + br $do_closure + + cfi_endproc + .end ffi_go_closure_osf + + .align 4 .globl ffi_closure_osf .ent ffi_closure_osf FFI_HIDDEN(ffi_closure_osf) ffi_closure_osf: - .frame $30, 16*8, $26, 0 - .mask 0x4000000, -16*8 -$LFB2: + cfi_startproc ldgp $29, 0($27) - subq $30, 16*8, $30 -$LCFI5: + subq $30, CLOSURE_FS, $30 + cfi_adjust_cfa_offset(CLOSURE_FS) stq $26, 0($30) -$LCFI6: .prologue 1 + cfi_rel_offset($26, 0) # Store all of the potential argument registers in va_list format. - stt $f16, 4*8($30) - stt $f17, 5*8($30) - stt $f18, 6*8($30) - stt $f19, 7*8($30) - stt $f20, 8*8($30) - stt $f21, 9*8($30) stq $16, 10*8($30) stq $17, 11*8($30) stq $18, 12*8($30) + + ldq $16, 24($1) # load cif + ldq $17, 32($1) # load fun + ldq $18, 40($1) # load user_data + +$do_closure: stq $19, 13*8($30) stq $20, 14*8($30) stq $21, 15*8($30) + stt $f16, 4*8($30) + stt $f17, 5*8($30) + stt $f18, 6*8($30) + stt $f19, 7*8($30) + stt $f20, 8*8($30) + stt $f21, 9*8($30) # Call ffi_closure_osf_inner to do the bulk of the work. - mov $1, $16 - lda $17, 2*8($30) - lda $18, 10*8($30) + lda $19, 2*8($30) + lda $20, 10*8($30) jsr $26, ffi_closure_osf_inner - ldgp $29, 0($26) +0: + ldah $29, 0($26) !gpdisp!2 + lda $2, 99f-0b($26) + s4addq $0, 0, $1 # ldcode * 4 + ldq $0, 16($30) # preload return value + s4addq $1, $2, $1 # 99f + ldcode * 16 + lda $29, 0($29) !gpdisp!2 ldq $26, 0($30) - - # Load up the return value in the proper type. - lda $1, $load_table - s4addq $0, $1, $1 - ldl $1, 0($1) - addq $1, $29, $1 + cfi_restore($26) jmp $31, ($1), $load_32 - .align 4 -$load_none: - addq $30, 16*8, $30 +.macro epilogue + addq $30, CLOSURE_FS, $30 + cfi_adjust_cfa_offset(-CLOSURE_FS) ret + .align 4 + cfi_adjust_cfa_offset(CLOSURE_FS) +.endm .align 4 -$load_float: - lds $f0, 16($30) - nop - addq $30, 16*8, $30 - ret +99: +E ALPHA_LD_VOID + epilogue - .align 4 -$load_double: - ldt $f0, 16($30) - nop - addq $30, 16*8, $30 - ret +E ALPHA_LD_INT64 + epilogue - .align 4 -$load_u8: +E ALPHA_LD_INT32 +$load_32: + sextl $0, $0 + epilogue + +E ALPHA_LD_UINT16 + zapnot $0, 3, $0 + epilogue + +E ALPHA_LD_SINT16 #ifdef __alpha_bwx__ - ldbu $0, 16($30) - nop + sextw $0, $0 #else - ldq $0, 16($30) - and $0, 255, $0 + sll $0, 48, $0 + sra $0, 48, $0 #endif - addq $30, 16*8, $30 - ret + epilogue - .align 4 -$load_s8: +E ALPHA_LD_UINT8 + and $0, 0xff, $0 + epilogue + +E ALPHA_LD_SINT8 #ifdef __alpha_bwx__ - ldbu $0, 16($30) sextb $0, $0 #else - ldq $0, 16($30) sll $0, 56, $0 sra $0, 56, $0 #endif - addq $30, 16*8, $30 - ret + epilogue - .align 4 -$load_u16: -#ifdef __alpha_bwx__ - ldwu $0, 16($30) - nop -#else - ldq $0, 16($30) - zapnot $0, 3, $0 -#endif - addq $30, 16*8, $30 - ret +E ALPHA_LD_FLOAT + lds $f0, 16($sp) + epilogue - .align 4 -$load_s16: -#ifdef __alpha_bwx__ - ldwu $0, 16($30) - sextw $0, $0 -#else - ldq $0, 16($30) - sll $0, 48, $0 - sra $0, 48, $0 -#endif - addq $30, 16*8, $30 - ret +E ALPHA_LD_DOUBLE + ldt $f0, 16($sp) + epilogue - .align 4 -$load_32: - ldl $0, 16($30) - nop - addq $30, 16*8, $30 - ret +E ALPHA_LD_CPLXF + lds $f0, 16($sp) + lds $f1, 20($sp) + epilogue - .align 4 -$load_64: - ldq $0, 16($30) - nop - addq $30, 16*8, $30 - ret -$LFE2: +E ALPHA_LD_CPLXD + ldt $f0, 16($sp) + ldt $f1, 24($sp) + epilogue + cfi_endproc .end ffi_closure_osf -#ifdef __ELF__ -.section .rodata -#else -.rdata -#endif -$load_table: - .gprel32 $load_none # FFI_TYPE_VOID - .gprel32 $load_32 # FFI_TYPE_INT - .gprel32 $load_float # FFI_TYPE_FLOAT - .gprel32 $load_double # FFI_TYPE_DOUBLE - .gprel32 $load_none # FFI_TYPE_LONGDOUBLE - .gprel32 $load_u8 # FFI_TYPE_UINT8 - .gprel32 $load_s8 # FFI_TYPE_SINT8 - .gprel32 $load_u16 # FFI_TYPE_UINT16 - .gprel32 $load_s16 # FFI_TYPE_SINT16 - .gprel32 $load_32 # FFI_TYPE_UINT32 - .gprel32 $load_32 # FFI_TYPE_SINT32 - .gprel32 $load_64 # FFI_TYPE_UINT64 - .gprel32 $load_64 # FFI_TYPE_SINT64 - .gprel32 $load_none # FFI_TYPE_STRUCT - .gprel32 $load_64 # FFI_TYPE_POINTER - -/* Assert that the table above is in sync with ffi.h. */ - -#if FFI_TYPE_FLOAT != 2 \ - || FFI_TYPE_DOUBLE != 3 \ - || FFI_TYPE_UINT8 != 5 \ - || FFI_TYPE_SINT8 != 6 \ - || FFI_TYPE_UINT16 != 7 \ - || FFI_TYPE_SINT16 != 8 \ - || FFI_TYPE_UINT32 != 9 \ - || FFI_TYPE_SINT32 != 10 \ - || FFI_TYPE_UINT64 != 11 \ - || FFI_TYPE_SINT64 != 12 \ - || FFI_TYPE_STRUCT != 13 \ - || FFI_TYPE_POINTER != 14 \ - || FFI_TYPE_LAST != 14 -#error "osf.S out of sync with ffi.h" -#endif - -#ifdef __ELF__ -# define UA_SI .4byte -# define FDE_ENCODING 0x1b /* pcrel sdata4 */ -# define FDE_ENCODE(X) .4byte X-. -# define FDE_ARANGE(X) .4byte X -#elif defined __osf__ -# define UA_SI .align 0; .long -# define FDE_ENCODING 0x50 /* aligned absolute */ -# define FDE_ENCODE(X) .align 3; .quad X -# define FDE_ARANGE(X) .align 0; .quad X -#endif - -#ifdef __ELF__ - .section .eh_frame,EH_FRAME_FLAGS,@progbits -#elif defined __osf__ - .data - .align 3 - .globl _GLOBAL__F_ffi_call_osf -_GLOBAL__F_ffi_call_osf: -#endif -__FRAME_BEGIN__: - UA_SI $LECIE1-$LSCIE1 # Length of Common Information Entry -$LSCIE1: - UA_SI 0x0 # CIE Identifier Tag - .byte 0x1 # CIE Version - .ascii "zR\0" # CIE Augmentation - .byte 0x1 # uleb128 0x1; CIE Code Alignment Factor - .byte 0x78 # sleb128 -8; CIE Data Alignment Factor - .byte 26 # CIE RA Column - .byte 0x1 # uleb128 0x1; Augmentation size - .byte FDE_ENCODING # FDE Encoding - .byte 0xc # DW_CFA_def_cfa - .byte 30 # uleb128 column 30 - .byte 0 # uleb128 offset 0 - .align 3 -$LECIE1: -$LSFDE1: - UA_SI $LEFDE1-$LASFDE1 # FDE Length -$LASFDE1: - UA_SI $LASFDE1-__FRAME_BEGIN__ # FDE CIE offset - FDE_ENCODE($LFB1) # FDE initial location - FDE_ARANGE($LFE1-$LFB1) # FDE address range - .byte 0x0 # uleb128 0x0; Augmentation size - - .byte 0x4 # DW_CFA_advance_loc4 - UA_SI $LCFI1-$LFB1 - .byte 0x9a # DW_CFA_offset, column 26 - .byte 4 # uleb128 4*-8 - .byte 0x8f # DW_CFA_offset, column 15 - .byte 0x3 # uleb128 3*-8 - .byte 0xc # DW_CFA_def_cfa - .byte 15 # uleb128 column 15 - .byte 32 # uleb128 offset 32 - - .byte 0x4 # DW_CFA_advance_loc4 - UA_SI $LCFI2-$LCFI1 - .byte 0xda # DW_CFA_restore, column 26 - .align 3 -$LEFDE1: - -$LSFDE3: - UA_SI $LEFDE3-$LASFDE3 # FDE Length -$LASFDE3: - UA_SI $LASFDE3-__FRAME_BEGIN__ # FDE CIE offset - FDE_ENCODE($LFB2) # FDE initial location - FDE_ARANGE($LFE2-$LFB2) # FDE address range - .byte 0x0 # uleb128 0x0; Augmentation size - - .byte 0x4 # DW_CFA_advance_loc4 - UA_SI $LCFI5-$LFB2 - .byte 0xe # DW_CFA_def_cfa_offset - .byte 0x80,0x1 # uleb128 128 - - .byte 0x4 # DW_CFA_advance_loc4 - UA_SI $LCFI6-$LCFI5 - .byte 0x9a # DW_CFA_offset, column 26 - .byte 16 # uleb128 offset 16*-8 - .align 3 -$LEFDE3: -#if defined __osf__ - .align 0 - .long 0 # End of Table -#endif - #if defined __ELF__ && defined __linux__ .section .note.GNU-stack,"",@progbits #endif diff --git a/native/libffi/src/arc/ffi.c b/native/libffi/src/arc/ffi.c index 32f82a7d5b..fa55a24455 100644 --- a/native/libffi/src/arc/ffi.c +++ b/native/libffi/src/arc/ffi.c @@ -46,12 +46,18 @@ void ffi_prep_args (char *stack, extended_cif * ecif) { unsigned int i; +<<<<<<< HEAD int tmp; +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 void **p_argv; char *argp; ffi_type **p_arg; +<<<<<<< HEAD tmp = 0; +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 argp = stack; if (ecif->cif->rtype->type == FFI_TYPE_STRUCT) @@ -73,7 +79,11 @@ ffi_prep_args (char *stack, extended_cif * ecif) /* Align if necessary. */ if ((alignment - 1) & (unsigned) argp) +<<<<<<< HEAD argp = (char *) ALIGN (argp, alignment); +======= + argp = (char *) FFI_ALIGN (argp, alignment); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 z = (*p_arg)->size; if (z < sizeof (int)) @@ -225,7 +235,11 @@ ffi_closure_inner_ARCompact (ffi_closure * closure, void *rvalue, /* Align if necessary. */ if ((alignment - 1) & (unsigned) argp) +<<<<<<< HEAD argp = (char *) ALIGN (argp, alignment); +======= + argp = (char *) FFI_ALIGN (argp, alignment); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 z = (*p_argt)->size; *p_argv = (void *) argp; diff --git a/native/libffi/src/arm/ffi.c b/native/libffi/src/arm/ffi.c index 6691ab57da..88dc16a41b 100644 --- a/native/libffi/src/arm/ffi.c +++ b/native/libffi/src/arm/ffi.c @@ -28,15 +28,43 @@ DEALINGS IN THE SOFTWARE. ----------------------------------------------------------------------- */ +#if defined(__arm__) || defined(_M_ARM) +#include #include #include - +#include #include +#include "internal.h" + +#if defined(_MSC_VER) && defined(_M_ARM) +#define WIN32_LEAN_AND_MEAN +#include +#endif + +#if FFI_EXEC_TRAMPOLINE_TABLE + +#ifdef __MACH__ +#include +#endif + +#else +#ifndef _M_ARM +extern unsigned int ffi_arm_trampoline[2] FFI_HIDDEN; +#else +extern unsigned int ffi_arm_trampoline[3] FFI_HIDDEN; +#endif +#endif + +#if defined(__FreeBSD__) && defined(__arm__) +#include +#include +#endif /* Forward declares. */ -static int vfp_type_p (ffi_type *); +static int vfp_type_p (const ffi_type *); static void layout_vfp_args (ffi_cif *); +<<<<<<< HEAD int ffi_prep_args_SYSV(char *stack, extended_cif *ecif, float *vfp_space); int ffi_prep_args_VFP(char *stack, extended_cif *ecif, float *vfp_space); @@ -121,10 +149,87 @@ static size_t ffi_put_arg(ffi_type **arg_type, void **arg, char *stack) /* ffi_prep_args is called by the assembly routine once stack space has been allocated for the function's arguments +======= +static void * +ffi_align (ffi_type *ty, void *p) +{ + /* Align if necessary */ + size_t alignment; +#ifdef _WIN32_WCE + alignment = 4; +#else + alignment = ty->alignment; + if (alignment < 4) + alignment = 4; +#endif + return (void *) FFI_ALIGN (p, alignment); +} + +static size_t +ffi_put_arg (ffi_type *ty, void *src, void *dst) +{ + size_t z = ty->size; + + switch (ty->type) + { + case FFI_TYPE_SINT8: + *(UINT32 *)dst = *(SINT8 *)src; + break; + case FFI_TYPE_UINT8: + *(UINT32 *)dst = *(UINT8 *)src; + break; + case FFI_TYPE_SINT16: + *(UINT32 *)dst = *(SINT16 *)src; + break; + case FFI_TYPE_UINT16: + *(UINT32 *)dst = *(UINT16 *)src; + break; + + case FFI_TYPE_INT: + case FFI_TYPE_SINT32: + case FFI_TYPE_UINT32: + case FFI_TYPE_POINTER: +#ifndef _MSC_VER + case FFI_TYPE_FLOAT: +#endif + *(UINT32 *)dst = *(UINT32 *)src; + break; + +#ifdef _MSC_VER + // casting a float* to a UINT32* doesn't work on Windows + case FFI_TYPE_FLOAT: + *(uintptr_t *)dst = 0; + *(float *)dst = *(float *)src; + break; +#endif + + case FFI_TYPE_SINT64: + case FFI_TYPE_UINT64: + case FFI_TYPE_DOUBLE: + *(UINT64 *)dst = *(UINT64 *)src; + break; + + case FFI_TYPE_STRUCT: + case FFI_TYPE_COMPLEX: + memcpy (dst, src, z); + break; + + default: + abort(); + } + + return FFI_ALIGN (z, 4); +} + +/* ffi_prep_args is called once stack space has been allocated + for the function's arguments. + +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 The vfp_space parameter is the load area for VFP regs, the return value is cif->vfp_used (word bitset of VFP regs used for passing arguments). These are only used for the VFP hard-float ABI. */ +<<<<<<< HEAD int ffi_prep_args_SYSV(char *stack, extended_cif *ecif, float *vfp_space) { register unsigned int i; @@ -133,14 +238,53 @@ int ffi_prep_args_SYSV(char *stack, extended_cif *ecif, float *vfp_space) register ffi_type **p_arg; argp = stack; +======= +static void +ffi_prep_args_SYSV (ffi_cif *cif, int flags, void *rvalue, + void **avalue, char *argp) +{ + ffi_type **arg_types = cif->arg_types; + int i, n; - if ( ecif->cif->flags == FFI_TYPE_STRUCT ) { - *(void **) argp = ecif->rvalue; - argp += 4; - } + if (flags == ARM_TYPE_STRUCT) + { + *(void **) argp = rvalue; + argp += 4; + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 - p_argv = ecif->avalue; + for (i = 0, n = cif->nargs; i < n; i++) + { + ffi_type *ty = arg_types[i]; + argp = ffi_align (ty, argp); + argp += ffi_put_arg (ty, avalue[i], argp); + } +} +static void +ffi_prep_args_VFP (ffi_cif *cif, int flags, void *rvalue, + void **avalue, char *stack, char *vfp_space) +{ + ffi_type **arg_types = cif->arg_types; + int i, n, vi = 0; + char *argp, *regp, *eo_regp; + char stack_used = 0; + char done_with_regs = 0; + + /* The first 4 words on the stack are used for values + passed in core registers. */ + regp = stack; + eo_regp = argp = regp + 16; + + /* If the function returns an FFI_TYPE_STRUCT in memory, + that address is passed in r0 to the function. */ + if (flags == ARM_TYPE_STRUCT) + { + *(void **) regp = rvalue; + regp += 4; + } + +<<<<<<< HEAD for (i = ecif->cif->nargs, p_arg = ecif->cif->arg_types; (i != 0); i--, p_arg++, p_argv++) @@ -228,81 +372,164 @@ int ffi_prep_args_VFP(char *stack, extended_cif *ecif, float *vfp_space) } /* Indicate the VFP registers used. */ return ecif->cif->vfp_used; +======= + for (i = 0, n = cif->nargs; i < n; i++) + { + ffi_type *ty = arg_types[i]; + void *a = avalue[i]; + int is_vfp_type = vfp_type_p (ty); + + /* Allocated in VFP registers. */ + if (vi < cif->vfp_nargs && is_vfp_type) + { + char *vfp_slot = vfp_space + cif->vfp_args[vi++] * 4; + ffi_put_arg (ty, a, vfp_slot); + continue; + } + /* Try allocating in core registers. */ + else if (!done_with_regs && !is_vfp_type) + { + char *tregp = ffi_align (ty, regp); + size_t size = ty->size; + size = (size < 4) ? 4 : size; // pad + /* Check if there is space left in the aligned register + area to place the argument. */ + if (tregp + size <= eo_regp) + { + regp = tregp + ffi_put_arg (ty, a, tregp); + done_with_regs = (regp == argp); + // ensure we did not write into the stack area + FFI_ASSERT (regp <= argp); + continue; + } + /* In case there are no arguments in the stack area yet, + the argument is passed in the remaining core registers + and on the stack. */ + else if (!stack_used) + { + stack_used = 1; + done_with_regs = 1; + argp = tregp + ffi_put_arg (ty, a, tregp); + FFI_ASSERT (eo_regp < argp); + continue; + } + } + /* Base case, arguments are passed on the stack */ + stack_used = 1; + argp = ffi_align (ty, argp); + argp += ffi_put_arg (ty, a, argp); + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } /* Perform machine dependent cif processing */ -ffi_status ffi_prep_cif_machdep(ffi_cif *cif) +ffi_status FFI_HIDDEN +ffi_prep_cif_machdep (ffi_cif *cif) { - int type_code; - /* Round the stack up to a multiple of 8 bytes. This isn't needed - everywhere, but it is on some platforms, and it doesn't harm anything - when it isn't needed. */ - cif->bytes = (cif->bytes + 7) & ~7; + int flags = 0, cabi = cif->abi; + size_t bytes = cif->bytes; + + /* Map out the register placements of VFP register args. The VFP + hard-float calling conventions are slightly more sophisticated + than the base calling conventions, so we do it here instead of + in ffi_prep_args(). */ + if (cabi == FFI_VFP) + layout_vfp_args (cif); /* Set the return type flag */ switch (cif->rtype->type) { case FFI_TYPE_VOID: - case FFI_TYPE_FLOAT: - case FFI_TYPE_DOUBLE: - cif->flags = (unsigned) cif->rtype->type; + flags = ARM_TYPE_VOID; + break; + + case FFI_TYPE_INT: + case FFI_TYPE_UINT8: + case FFI_TYPE_SINT8: + case FFI_TYPE_UINT16: + case FFI_TYPE_SINT16: + case FFI_TYPE_UINT32: + case FFI_TYPE_SINT32: + case FFI_TYPE_POINTER: + flags = ARM_TYPE_INT; break; case FFI_TYPE_SINT64: case FFI_TYPE_UINT64: - cif->flags = (unsigned) FFI_TYPE_SINT64; + flags = ARM_TYPE_INT64; + break; + + case FFI_TYPE_FLOAT: + flags = (cabi == FFI_VFP ? ARM_TYPE_VFP_S : ARM_TYPE_INT); + break; + case FFI_TYPE_DOUBLE: + flags = (cabi == FFI_VFP ? ARM_TYPE_VFP_D : ARM_TYPE_INT64); break; case FFI_TYPE_STRUCT: - if (cif->abi == FFI_VFP - && (type_code = vfp_type_p (cif->rtype)) != 0) + case FFI_TYPE_COMPLEX: + if (cabi == FFI_VFP) { - /* A Composite Type passed in VFP registers, either - FFI_TYPE_STRUCT_VFP_FLOAT or FFI_TYPE_STRUCT_VFP_DOUBLE. */ - cif->flags = (unsigned) type_code; + int h = vfp_type_p (cif->rtype); + + flags = ARM_TYPE_VFP_N; + if (h == 0x100 + FFI_TYPE_FLOAT) + flags = ARM_TYPE_VFP_S; + if (h == 0x100 + FFI_TYPE_DOUBLE) + flags = ARM_TYPE_VFP_D; + if (h != 0) + break; } - else if (cif->rtype->size <= 4) - /* A Composite Type not larger than 4 bytes is returned in r0. */ - cif->flags = (unsigned)FFI_TYPE_INT; + + /* A Composite Type not larger than 4 bytes is returned in r0. + A Composite Type larger than 4 bytes, or whose size cannot + be determined statically ... is stored in memory at an + address passed [in r0]. */ + if (cif->rtype->size <= 4) + flags = ARM_TYPE_INT; else - /* A Composite Type larger than 4 bytes, or whose size cannot - be determined statically ... is stored in memory at an - address passed [in r0]. */ - cif->flags = (unsigned)FFI_TYPE_STRUCT; + { + flags = ARM_TYPE_STRUCT; + bytes += 4; + } break; default: - cif->flags = FFI_TYPE_INT; - break; + abort(); } - /* Map out the register placements of VFP register args. - The VFP hard-float calling conventions are slightly more sophisticated than - the base calling conventions, so we do it here instead of in ffi_prep_args(). */ - if (cif->abi == FFI_VFP) - layout_vfp_args (cif); + /* Round the stack up to a multiple of 8 bytes. This isn't needed + everywhere, but it is on some platforms, and it doesn't harm anything + when it isn't needed. */ + bytes = FFI_ALIGN (bytes, 8); + + /* Minimum stack space is the 4 register arguments that we pop. */ + if (bytes < 4*4) + bytes = 4*4; + + cif->bytes = bytes; + cif->flags = flags; return FFI_OK; } /* Perform machine dependent cif processing for variadic calls */ -ffi_status ffi_prep_cif_machdep_var(ffi_cif *cif, - unsigned int nfixedargs, - unsigned int ntotalargs) +ffi_status FFI_HIDDEN +ffi_prep_cif_machdep_var (ffi_cif * cif, + unsigned int nfixedargs, unsigned int ntotalargs) { /* VFP variadic calls actually use the SYSV ABI */ if (cif->abi == FFI_VFP) - cif->abi = FFI_SYSV; + cif->abi = FFI_SYSV; - return ffi_prep_cif_machdep(cif); + return ffi_prep_cif_machdep (cif); } -/* Prototypes for assembly functions, in sysv.S */ -extern void ffi_call_SYSV (void (*fn)(void), extended_cif *, unsigned, unsigned, unsigned *); -extern void ffi_call_VFP (void (*fn)(void), extended_cif *, unsigned, unsigned, unsigned *); +/* Prototypes for assembly functions, in sysv.S. */ -void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +struct call_frame { +<<<<<<< HEAD extended_cif ecif; int small_struct = (cif->flags == FFI_TYPE_INT @@ -317,38 +544,62 @@ void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) /* If the return value is a struct and we don't have a return */ /* value address then we need to make one */ +======= + void *fp; + void *lr; + void *rvalue; + int flags; + void *closure; +}; + +extern void ffi_call_SYSV (void *stack, struct call_frame *, + void (*fn) (void)) FFI_HIDDEN; +extern void ffi_call_VFP (void *vfp_space, struct call_frame *, + void (*fn) (void), unsigned vfp_used) FFI_HIDDEN; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 - if ((rvalue == NULL) && - (cif->flags == FFI_TYPE_STRUCT)) +static void +ffi_call_int (ffi_cif * cif, void (*fn) (void), void *rvalue, + void **avalue, void *closure) +{ + int flags = cif->flags; + ffi_type *rtype = cif->rtype; + size_t bytes, rsize, vfp_size; + char *stack, *vfp_space, *new_rvalue; + struct call_frame *frame; + + rsize = 0; + if (rvalue == NULL) { - ecif.rvalue = alloca(cif->rtype->size); + /* If the return value is a struct and we don't have a return + value address then we need to make one. Otherwise the return + value is in registers and we can ignore them. */ + if (flags == ARM_TYPE_STRUCT) + rsize = rtype->size; + else + flags = ARM_TYPE_VOID; } - else if (small_struct) - ecif.rvalue = &temp; - else if (vfp_struct) + else if (flags == ARM_TYPE_VFP_N) { /* Largest case is double x 4. */ - ecif.rvalue = alloca(32); + rsize = 32; } - else - ecif.rvalue = rvalue; + else if (flags == ARM_TYPE_INT && rtype->type == FFI_TYPE_STRUCT) + rsize = 4; - switch (cif->abi) - { - case FFI_SYSV: - ffi_call_SYSV (fn, &ecif, cif->bytes, cif->flags, ecif.rvalue); - break; + /* Largest case. */ + vfp_size = (cif->abi == FFI_VFP && cif->vfp_used ? 8*8: 0); - case FFI_VFP: -#ifdef __ARM_EABI__ - ffi_call_VFP (fn, &ecif, cif->bytes, cif->flags, ecif.rvalue); - break; -#endif + bytes = cif->bytes; + stack = alloca (vfp_size + bytes + sizeof(struct call_frame) + rsize); - default: - FFI_ASSERT(0); - break; + vfp_space = NULL; + if (vfp_size) + { + vfp_space = stack; + stack += vfp_size; } +<<<<<<< HEAD if (small_struct) { FFI_ASSERT(rvalue != NULL); @@ -362,32 +613,70 @@ void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) } } +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 -/** private members **/ + frame = (struct call_frame *)(stack + bytes); -static void ffi_prep_incoming_args_SYSV (char *stack, void **ret, - void** args, ffi_cif* cif, float *vfp_stack); + new_rvalue = rvalue; + if (rsize) + new_rvalue = (void *)(frame + 1); +<<<<<<< HEAD static void ffi_prep_incoming_args_VFP (char *stack, void **ret, void** args, ffi_cif* cif, float *vfp_stack); void ffi_closure_SYSV (ffi_closure *); +======= + frame->rvalue = new_rvalue; + frame->flags = flags; + frame->closure = closure; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 -void ffi_closure_VFP (ffi_closure *); + if (vfp_space) + { + ffi_prep_args_VFP (cif, flags, new_rvalue, avalue, stack, vfp_space); + ffi_call_VFP (vfp_space, frame, fn, cif->vfp_used); + } + else + { + ffi_prep_args_SYSV (cif, flags, new_rvalue, avalue, stack); + ffi_call_SYSV (stack, frame, fn); + } -/* This function is jumped to by the trampoline */ + if (rvalue && rvalue != new_rvalue) + memcpy (rvalue, new_rvalue, rtype->size); +} +<<<<<<< HEAD unsigned int FFI_HIDDEN ffi_closure_inner (ffi_closure *closure, void **respp, void *args, void *vfp_args) +======= +void +ffi_call (ffi_cif *cif, void (*fn) (void), void *rvalue, void **avalue) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 +{ + ffi_call_int (cif, fn, rvalue, avalue, NULL); +} + +#ifdef FFI_GO_CLOSURES +void +ffi_call_go (ffi_cif *cif, void (*fn) (void), void *rvalue, + void **avalue, void *closure) { - // our various things... - ffi_cif *cif; - void **arg_area; + ffi_call_int (cif, fn, rvalue, avalue, closure); +} +#endif - cif = closure->cif; - arg_area = (void**) alloca (cif->nargs * sizeof (void*)); +static void * +ffi_prep_incoming_args_SYSV (ffi_cif *cif, void *rvalue, + char *argp, void **avalue) +{ + ffi_type **arg_types = cif->arg_types; + int i, n; +<<<<<<< HEAD /* this call will initialize ARG_AREA, such that each * element in that array points to the corresponding * value on the stack; and if the function returns @@ -397,36 +686,61 @@ ffi_closure_inner (ffi_closure *closure, ffi_prep_incoming_args_VFP(args, respp, arg_area, cif, vfp_args); else ffi_prep_incoming_args_SYSV(args, respp, arg_area, cif, vfp_args); +======= + if (cif->flags == ARM_TYPE_STRUCT) + { + rvalue = *(void **) argp; + argp += 4; + } + else + { + if (cif->rtype->size && cif->rtype->size < 4) + *(uint32_t *) rvalue = 0; + } - (closure->fun) (cif, *respp, arg_area, closure->user_data); + for (i = 0, n = cif->nargs; i < n; i++) + { + ffi_type *ty = arg_types[i]; + size_t z = ty->size; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 - return cif->flags; + argp = ffi_align (ty, argp); + avalue[i] = (void *) argp; + argp += z; + } + + return rvalue; } -/*@-exportheader@*/ -static void -ffi_prep_incoming_args_SYSV(char *stack, void **rvalue, - void **avalue, ffi_cif *cif, - /* Used only under VFP hard-float ABI. */ - float *vfp_stack) -/*@=exportheader@*/ +static void * +ffi_prep_incoming_args_VFP (ffi_cif *cif, void *rvalue, char *stack, + char *vfp_space, void **avalue) { +<<<<<<< HEAD register unsigned int i; register void **p_argv; register char *argp; register ffi_type **p_arg; +======= + ffi_type **arg_types = cif->arg_types; + int i, n, vi = 0; + char *argp, *regp, *eo_regp; + char done_with_regs = 0; + char stack_used = 0; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 - argp = stack; - - if ( cif->flags == FFI_TYPE_STRUCT ) { - *rvalue = *(void **) argp; - argp += 4; - } + regp = stack; + eo_regp = argp = regp + 16; - p_argv = avalue; + if (cif->flags == ARM_TYPE_STRUCT) + { + rvalue = *(void **) regp; + regp += 4; + } - for (i = cif->nargs, p_arg = cif->arg_types; (i != 0); i--, p_arg++) + for (i = 0, n = cif->nargs; i < n; i++) { +<<<<<<< HEAD size_t z; argp = ffi_align(p_arg, argp); @@ -434,16 +748,55 @@ ffi_prep_incoming_args_SYSV(char *stack, void **rvalue, z = (*p_arg)->size; /* because we're little endian, this is what it turns into. */ +======= + ffi_type *ty = arg_types[i]; + int is_vfp_type = vfp_type_p (ty); + size_t z = ty->size; + + if (vi < cif->vfp_nargs && is_vfp_type) + { + avalue[i] = vfp_space + cif->vfp_args[vi++] * 4; + continue; + } + else if (!done_with_regs && !is_vfp_type) + { + char *tregp = ffi_align (ty, regp); + + z = (z < 4) ? 4 : z; // pad +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 - *p_argv = (void*) argp; + /* If the arguments either fits into the registers or uses registers + and stack, while we haven't read other things from the stack */ + if (tregp + z <= eo_regp || !stack_used) + { + /* Because we're little endian, this is what it turns into. */ + avalue[i] = (void *) tregp; + regp = tregp + z; + + /* If we read past the last core register, make sure we + have not read from the stack before and continue + reading after regp. */ + if (regp > eo_regp) + { + FFI_ASSERT (!stack_used); + argp = regp; + } + if (regp >= eo_regp) + { + done_with_regs = 1; + stack_used = 1; + } + continue; + } + } - p_argv++; + stack_used = 1; + argp = ffi_align (ty, argp); + avalue[i] = (void *) argp; argp += z; } - - return; -} +<<<<<<< HEAD /*@-exportheader@*/ static void ffi_prep_incoming_args_VFP(char *stack, void **rvalue, @@ -562,220 +915,183 @@ struct ffi_trampoline_table { ffi_trampoline_table *prev; ffi_trampoline_table *next; }; +======= + return rvalue; +} +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 -struct ffi_trampoline_table_entry { - void *(*trampoline)(); - ffi_trampoline_table_entry *next; +struct closure_frame +{ + char vfp_space[8*8] __attribute__((aligned(8))); + char result[8*4]; + char argp[]; }; -/* Override the standard architecture trampoline size */ -// XXX TODO - Fix -#undef FFI_TRAMPOLINE_SIZE -#define FFI_TRAMPOLINE_SIZE 12 - -/* The trampoline configuration is placed at 4080 bytes prior to the trampoline's entry point */ -#define FFI_TRAMPOLINE_CODELOC_CONFIG(codeloc) ((void **) (((uint8_t *) codeloc) - 4080)); - -/* The first 16 bytes of the config page are unused, as they are unaddressable from the trampoline page. */ -#define FFI_TRAMPOLINE_CONFIG_PAGE_OFFSET 16 - -/* Total number of trampolines that fit in one trampoline table */ -#define FFI_TRAMPOLINE_COUNT ((PAGE_SIZE - FFI_TRAMPOLINE_CONFIG_PAGE_OFFSET) / FFI_TRAMPOLINE_SIZE) - -static pthread_mutex_t ffi_trampoline_lock = PTHREAD_MUTEX_INITIALIZER; -static ffi_trampoline_table *ffi_trampoline_tables = NULL; - -static ffi_trampoline_table * -ffi_trampoline_table_alloc () +int FFI_HIDDEN +ffi_closure_inner_SYSV (ffi_cif *cif, + void (*fun) (ffi_cif *, void *, void **, void *), + void *user_data, + struct closure_frame *frame) { - ffi_trampoline_table *table = NULL; + void **avalue = (void **) alloca (cif->nargs * sizeof (void *)); + void *rvalue = ffi_prep_incoming_args_SYSV (cif, frame->result, + frame->argp, avalue); + fun (cif, rvalue, avalue, user_data); + return cif->flags; +} +<<<<<<< HEAD /* Loop until we can allocate two contiguous pages */ while (table == NULL) { vm_address_t config_page = 0x0; kern_return_t kt; +======= +int FFI_HIDDEN +ffi_closure_inner_VFP (ffi_cif *cif, + void (*fun) (ffi_cif *, void *, void **, void *), + void *user_data, + struct closure_frame *frame) +{ + void **avalue = (void **) alloca (cif->nargs * sizeof (void *)); + void *rvalue = ffi_prep_incoming_args_VFP (cif, frame->result, frame->argp, + frame->vfp_space, avalue); + fun (cif, rvalue, avalue, user_data); + return cif->flags; +} +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 - /* Try to allocate two pages */ - kt = vm_allocate (mach_task_self (), &config_page, PAGE_SIZE*2, VM_FLAGS_ANYWHERE); - if (kt != KERN_SUCCESS) { - fprintf(stderr, "vm_allocate() failure: %d at %s:%d\n", kt, __FILE__, __LINE__); - break; - } +void ffi_closure_SYSV (void) FFI_HIDDEN; +void ffi_closure_VFP (void) FFI_HIDDEN; - /* Now drop the second half of the allocation to make room for the trampoline table */ - vm_address_t trampoline_page = config_page+PAGE_SIZE; - kt = vm_deallocate (mach_task_self (), trampoline_page, PAGE_SIZE); - if (kt != KERN_SUCCESS) { - fprintf(stderr, "vm_deallocate() failure: %d at %s:%d\n", kt, __FILE__, __LINE__); - break; - } +#ifdef FFI_GO_CLOSURES +void ffi_go_closure_SYSV (void) FFI_HIDDEN; +void ffi_go_closure_VFP (void) FFI_HIDDEN; +#endif - /* Remap the trampoline table to directly follow the config page */ - vm_prot_t cur_prot; - vm_prot_t max_prot; +/* the cif must already be prep'ed */ - kt = vm_remap (mach_task_self (), &trampoline_page, PAGE_SIZE, 0x0, FALSE, mach_task_self (), (vm_address_t) &ffi_closure_trampoline_table_page, FALSE, &cur_prot, &max_prot, VM_INHERIT_SHARE); +#if defined(__FreeBSD__) && defined(__arm__) +#define __clear_cache(start, end) do { \ + struct arm_sync_icache_args ua; \ + \ + ua.addr = (uintptr_t)(start); \ + ua.len = (char *)(end) - (char *)start; \ + sysarch(ARM_SYNC_ICACHE, &ua); \ + } while (0); +#endif - /* If we lost access to the destination trampoline page, drop our config allocation mapping and retry */ - if (kt != KERN_SUCCESS) { - /* Log unexpected failures */ - if (kt != KERN_NO_SPACE) { - fprintf(stderr, "vm_remap() failure: %d at %s:%d\n", kt, __FILE__, __LINE__); - } +ffi_status +ffi_prep_closure_loc (ffi_closure * closure, + ffi_cif * cif, + void (*fun) (ffi_cif *, void *, void **, void *), + void *user_data, void *codeloc) +{ + void (*closure_func) (void) = ffi_closure_SYSV; - vm_deallocate (mach_task_self (), config_page, PAGE_SIZE); - continue; + if (cif->abi == FFI_VFP) + { + /* We only need take the vfp path if there are vfp arguments. */ + if (cif->vfp_used) + closure_func = ffi_closure_VFP; } + else if (cif->abi != FFI_SYSV) + return FFI_BAD_ABI; - /* We have valid trampoline and config pages */ - table = calloc (1, sizeof(ffi_trampoline_table)); - table->free_count = FFI_TRAMPOLINE_COUNT; - table->config_page = config_page; - table->trampoline_page = trampoline_page; - - /* Create and initialize the free list */ - table->free_list_pool = calloc(FFI_TRAMPOLINE_COUNT, sizeof(ffi_trampoline_table_entry)); +#if FFI_EXEC_TRAMPOLINE_TABLE + void **config = (void **)((uint8_t *)codeloc - PAGE_MAX_SIZE); + config[0] = closure; + config[1] = closure_func; +#else - uint16_t i; - for (i = 0; i < table->free_count; i++) { - ffi_trampoline_table_entry *entry = &table->free_list_pool[i]; - entry->trampoline = (void *) (table->trampoline_page + (i * FFI_TRAMPOLINE_SIZE)); +#ifndef _M_ARM + memcpy(closure->tramp, ffi_arm_trampoline, 8); +#else + // cast away function type so MSVC doesn't set the lower bit of the function pointer + memcpy(closure->tramp, (void*)((uintptr_t)ffi_arm_trampoline & 0xFFFFFFFE), FFI_TRAMPOLINE_CLOSURE_OFFSET); +#endif - if (i < table->free_count - 1) - entry->next = &table->free_list_pool[i+1]; - } +#if defined (__QNX__) + msync(closure->tramp, 8, 0x1000000); /* clear data map */ + msync(codeloc, 8, 0x1000000); /* clear insn map */ +#elif defined(_MSC_VER) + FlushInstructionCache(GetCurrentProcess(), closure->tramp, FFI_TRAMPOLINE_SIZE); +#else + __clear_cache(closure->tramp, closure->tramp + 8); /* clear data map */ + __clear_cache(codeloc, codeloc + 8); /* clear insn map */ +#endif +#ifdef _M_ARM + *(void(**)(void))(closure->tramp + FFI_TRAMPOLINE_CLOSURE_FUNCTION) = closure_func; +#else + *(void (**)(void))(closure->tramp + 8) = closure_func; +#endif +#endif - table->free_list = table->free_list_pool; - } + closure->cif = cif; + closure->fun = fun; + closure->user_data = user_data; - return table; + return FFI_OK; } -void * -ffi_closure_alloc (size_t size, void **code) +#ifdef FFI_GO_CLOSURES +ffi_status +ffi_prep_go_closure (ffi_go_closure *closure, ffi_cif *cif, + void (*fun) (ffi_cif *, void *, void **, void *)) { - /* Create the closure */ - ffi_closure *closure = malloc(size); - if (closure == NULL) - return NULL; + void (*closure_func) (void) = ffi_go_closure_SYSV; - pthread_mutex_lock(&ffi_trampoline_lock); - - /* Check for an active trampoline table with available entries. */ - ffi_trampoline_table *table = ffi_trampoline_tables; - if (table == NULL || table->free_list == NULL) { - table = ffi_trampoline_table_alloc (); - if (table == NULL) { - free(closure); - return NULL; + if (cif->abi == FFI_VFP) + { + /* We only need take the vfp path if there are vfp arguments. */ + if (cif->vfp_used) + closure_func = ffi_go_closure_VFP; } + else if (cif->abi != FFI_SYSV) + return FFI_BAD_ABI; - /* Insert the new table at the top of the list */ - table->next = ffi_trampoline_tables; - if (table->next != NULL) - table->next->prev = table; - - ffi_trampoline_tables = table; - } - - /* Claim the free entry */ - ffi_trampoline_table_entry *entry = ffi_trampoline_tables->free_list; - ffi_trampoline_tables->free_list = entry->next; - ffi_trampoline_tables->free_count--; - entry->next = NULL; - - pthread_mutex_unlock(&ffi_trampoline_lock); - - /* Initialize the return values */ - *code = entry->trampoline; - closure->trampoline_table = table; - closure->trampoline_table_entry = entry; + closure->tramp = closure_func; + closure->cif = cif; + closure->fun = fun; - return closure; + return FFI_OK; } +#endif -void -ffi_closure_free (void *ptr) -{ - ffi_closure *closure = ptr; - - pthread_mutex_lock(&ffi_trampoline_lock); - - /* Fetch the table and entry references */ - ffi_trampoline_table *table = closure->trampoline_table; - ffi_trampoline_table_entry *entry = closure->trampoline_table_entry; - - /* Return the entry to the free list */ - entry->next = table->free_list; - table->free_list = entry; - table->free_count++; - - /* If all trampolines within this table are free, and at least one other table exists, deallocate - * the table */ - if (table->free_count == FFI_TRAMPOLINE_COUNT && ffi_trampoline_tables != table) { - /* Remove from the list */ - if (table->prev != NULL) - table->prev->next = table->next; +/* Below are routines for VFP hard-float support. */ - if (table->next != NULL) - table->next->prev = table->prev; +/* A subroutine of vfp_type_p. Given a structure type, return the type code + of the first non-structure element. Recurse for structure elements. + Return -1 if the structure is in fact empty, i.e. no nested elements. */ - /* Deallocate pages */ - kern_return_t kt; - kt = vm_deallocate (mach_task_self (), table->config_page, PAGE_SIZE); - if (kt != KERN_SUCCESS) - fprintf(stderr, "vm_deallocate() failure: %d at %s:%d\n", kt, __FILE__, __LINE__); - - kt = vm_deallocate (mach_task_self (), table->trampoline_page, PAGE_SIZE); - if (kt != KERN_SUCCESS) - fprintf(stderr, "vm_deallocate() failure: %d at %s:%d\n", kt, __FILE__, __LINE__); - - /* Deallocate free list */ - free (table->free_list_pool); - free (table); - } else if (ffi_trampoline_tables != table) { - /* Otherwise, bump this table to the top of the list */ - table->prev = NULL; - table->next = ffi_trampoline_tables; - if (ffi_trampoline_tables != NULL) - ffi_trampoline_tables->prev = table; - - ffi_trampoline_tables = table; - } +static int +is_hfa0 (const ffi_type *ty) +{ + ffi_type **elements = ty->elements; + int i, ret = -1; - pthread_mutex_unlock (&ffi_trampoline_lock); + if (elements != NULL) + for (i = 0; elements[i]; ++i) + { + ret = elements[i]->type; + if (ret == FFI_TYPE_STRUCT || ret == FFI_TYPE_COMPLEX) + { + ret = is_hfa0 (elements[i]); + if (ret < 0) + continue; + } + break; + } - /* Free the closure */ - free (closure); + return ret; } -#else - -#define FFI_INIT_TRAMPOLINE(TRAMP,FUN,CTX) \ -({ unsigned char *__tramp = (unsigned char*)(TRAMP); \ - unsigned int __fun = (unsigned int)(FUN); \ - unsigned int __ctx = (unsigned int)(CTX); \ - unsigned char *insns = (unsigned char *)(CTX); \ - memcpy (__tramp, ffi_arm_trampoline, sizeof ffi_arm_trampoline); \ - *(unsigned int*) &__tramp[12] = __ctx; \ - *(unsigned int*) &__tramp[16] = __fun; \ - __clear_cache((&__tramp[0]), (&__tramp[19])); /* Clear data mapping. */ \ - __clear_cache(insns, insns + 3 * sizeof (unsigned int)); \ - /* Clear instruction \ - mapping. */ \ - }) - -#endif +/* A subroutine of vfp_type_p. Given a structure type, return true if all + of the non-structure elements are the same as CANDIDATE. */ -/* the cif must already be prep'ed */ - -ffi_status -ffi_prep_closure_loc (ffi_closure* closure, - ffi_cif* cif, - void (*fun)(ffi_cif*,void*,void**,void*), - void *user_data, - void *codeloc) +static int +is_hfa1 (const ffi_type *ty, int candidate) { +<<<<<<< HEAD void (*closure_func)(ffi_closure*) = NULL; if (cif->abi == FFI_SYSV) @@ -796,88 +1112,139 @@ ffi_prep_closure_loc (ffi_closure* closure, closure_func, \ codeloc); #endif +======= + ffi_type **elements = ty->elements; + int i; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 - closure->cif = cif; - closure->user_data = user_data; - closure->fun = fun; + if (elements != NULL) + for (i = 0; elements[i]; ++i) + { + int t = elements[i]->type; + if (t == FFI_TYPE_STRUCT || t == FFI_TYPE_COMPLEX) + { + if (!is_hfa1 (elements[i], candidate)) + return 0; + } + else if (t != candidate) + return 0; + } - return FFI_OK; + return 1; } -/* Below are routines for VFP hard-float support. */ +/* Determine if TY is an homogenous floating point aggregate (HFA). + That is, a structure consisting of 1 to 4 members of all the same type, + where that type is a floating point scalar. -static int rec_vfp_type_p (ffi_type *t, int *elt, int *elnum) + Returns non-zero iff TY is an HFA. The result is an encoded value where + bits 0-7 contain the type code, and bits 8-10 contain the element count. */ + +static int +vfp_type_p (const ffi_type *ty) { - switch (t->type) + ffi_type **elements; + int candidate, i; + size_t size, ele_count; + + /* Quickest tests first. */ + candidate = ty->type; + switch (ty->type) { + default: + return 0; case FFI_TYPE_FLOAT: case FFI_TYPE_DOUBLE: - *elt = (int) t->type; - *elnum = 1; - return 1; + ele_count = 1; + goto done; + case FFI_TYPE_COMPLEX: + candidate = ty->elements[0]->type; + if (candidate != FFI_TYPE_FLOAT && candidate != FFI_TYPE_DOUBLE) + return 0; + ele_count = 2; + goto done; + case FFI_TYPE_STRUCT: + break; + } - case FFI_TYPE_STRUCT_VFP_FLOAT: - *elt = FFI_TYPE_FLOAT; - *elnum = t->size / sizeof (float); - return 1; + /* No HFA types are smaller than 4 bytes, or larger than 32 bytes. */ + size = ty->size; + if (size < 4 || size > 32) + return 0; - case FFI_TYPE_STRUCT_VFP_DOUBLE: - *elt = FFI_TYPE_DOUBLE; - *elnum = t->size / sizeof (double); - return 1; + /* Find the type of the first non-structure member. */ + elements = ty->elements; + candidate = elements[0]->type; + if (candidate == FFI_TYPE_STRUCT || candidate == FFI_TYPE_COMPLEX) + { + for (i = 0; ; ++i) + { + candidate = is_hfa0 (elements[i]); + if (candidate >= 0) + break; + } + } - case FFI_TYPE_STRUCT:; - { - int base_elt = 0, total_elnum = 0; - ffi_type **el = t->elements; - while (*el) - { - int el_elt = 0, el_elnum = 0; - if (! rec_vfp_type_p (*el, &el_elt, &el_elnum) - || (base_elt && base_elt != el_elt) - || total_elnum + el_elnum > 4) - return 0; - base_elt = el_elt; - total_elnum += el_elnum; - el++; - } - *elnum = total_elnum; - *elt = base_elt; - return 1; - } - default: ; + /* If the first member is not a floating point type, it's not an HFA. + Also quickly re-check the size of the structure. */ + switch (candidate) + { + case FFI_TYPE_FLOAT: + ele_count = size / sizeof(float); + if (size != ele_count * sizeof(float)) + return 0; + break; + case FFI_TYPE_DOUBLE: + ele_count = size / sizeof(double); + if (size != ele_count * sizeof(double)) + return 0; + break; + default: + return 0; } - return 0; -} + if (ele_count > 4) + return 0; -static int vfp_type_p (ffi_type *t) -{ - int elt, elnum; - if (rec_vfp_type_p (t, &elt, &elnum)) + /* Finally, make sure that all scalar elements are the same type. */ + for (i = 0; elements[i]; ++i) { - if (t->type == FFI_TYPE_STRUCT) - { - if (elnum == 1) - t->type = elt; - else - t->type = (elt == FFI_TYPE_FLOAT - ? FFI_TYPE_STRUCT_VFP_FLOAT - : FFI_TYPE_STRUCT_VFP_DOUBLE); - } - return (int) t->type; + int t = elements[i]->type; + if (t == FFI_TYPE_STRUCT || t == FFI_TYPE_COMPLEX) + { + if (!is_hfa1 (elements[i], candidate)) + return 0; + } + else if (t != candidate) + return 0; } - return 0; + + /* All tests succeeded. Encode the result. */ + done: + return (ele_count << 8) | candidate; } +<<<<<<< HEAD static int place_vfp_arg (ffi_cif *cif, ffi_type *t) { short reg = cif->vfp_reg_free; int nregs = t->size / sizeof (float); int align = ((t->type == FFI_TYPE_STRUCT_VFP_FLOAT || t->type == FFI_TYPE_FLOAT) ? 1 : 2); +======= +static int +place_vfp_arg (ffi_cif *cif, int h) +{ + unsigned short reg = cif->vfp_reg_free; + int align = 1, nregs = h >> 8; + + if ((h & 0xff) == FFI_TYPE_DOUBLE) + align = 2, nregs *= 2; + +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* Align register number. */ if ((reg & 1) && align == 2) reg++; + while (reg + nregs <= 16) { int s, new_used = 0; @@ -892,7 +1259,7 @@ static int place_vfp_arg (ffi_cif *cif, ffi_type *t) } /* Found regs to allocate. */ cif->vfp_used |= new_used; - cif->vfp_args[cif->vfp_nargs++] = reg; + cif->vfp_args[cif->vfp_nargs++] = (signed char)reg; /* Update vfp_reg_free. */ if (cif->vfp_used & (1 << cif->vfp_reg_free)) @@ -903,7 +1270,11 @@ static int place_vfp_arg (ffi_cif *cif, ffi_type *t) cif->vfp_reg_free = reg; } return 0; +<<<<<<< HEAD next_reg: ; +======= + next_reg:; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } // done, mark all regs as used cif->vfp_reg_free = 16; @@ -911,21 +1282,30 @@ static int place_vfp_arg (ffi_cif *cif, ffi_type *t) return 1; } -static void layout_vfp_args (ffi_cif *cif) +static void +layout_vfp_args (ffi_cif * cif) { - int i; + unsigned int i; /* Init VFP fields */ cif->vfp_used = 0; cif->vfp_nargs = 0; cif->vfp_reg_free = 0; - memset (cif->vfp_args, -1, 16); /* Init to -1. */ + memset (cif->vfp_args, -1, 16); /* Init to -1. */ for (i = 0; i < cif->nargs; i++) { +<<<<<<< HEAD ffi_type *t = cif->arg_types[i]; if (vfp_type_p (t) && place_vfp_arg (cif, t) == 1) { break; } +======= + int h = vfp_type_p (cif->arg_types[i]); + if (h && place_vfp_arg (cif, h) == 1) + break; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } } + +#endif /* __arm__ or _M_ARM */ diff --git a/native/libffi/src/arm/ffitarget.h b/native/libffi/src/arm/ffitarget.h index 26d494d616..cb57b84880 100644 --- a/native/libffi/src/arm/ffitarget.h +++ b/native/libffi/src/arm/ffitarget.h @@ -43,7 +43,7 @@ typedef enum ffi_abi { FFI_SYSV, FFI_VFP, FFI_LAST_ABI, -#ifdef __ARM_PCS_VFP +#if defined(__ARM_PCS_VFP) || defined(_M_ARM) FFI_DEFAULT_ABI = FFI_VFP, #else FFI_DEFAULT_ABI = FFI_SYSV, @@ -53,19 +53,37 @@ typedef enum ffi_abi { #define FFI_EXTRA_CIF_FIELDS \ int vfp_used; \ - short vfp_reg_free, vfp_nargs; \ + unsigned short vfp_reg_free, vfp_nargs; \ signed char vfp_args[16] \ -/* Internally used. */ -#define FFI_TYPE_STRUCT_VFP_FLOAT (FFI_TYPE_LAST + 1) -#define FFI_TYPE_STRUCT_VFP_DOUBLE (FFI_TYPE_LAST + 2) - #define FFI_TARGET_SPECIFIC_VARIADIC +#ifndef _M_ARM +#define FFI_TARGET_HAS_COMPLEX_TYPE +#endif /* ---- Definitions for closures ----------------------------------------- */ #define FFI_CLOSURES 1 -#define FFI_TRAMPOLINE_SIZE 20 +#define FFI_GO_CLOSURES 1 #define FFI_NATIVE_RAW_API 0 +#if defined (FFI_EXEC_TRAMPOLINE_TABLE) && FFI_EXEC_TRAMPOLINE_TABLE + +#ifdef __MACH__ +#define FFI_TRAMPOLINE_SIZE 12 +#define FFI_TRAMPOLINE_CLOSURE_OFFSET 8 +#else +#error "No trampoline table implementation" +#endif + +#else +#ifdef _MSC_VER +#define FFI_TRAMPOLINE_SIZE 16 +#define FFI_TRAMPOLINE_CLOSURE_FUNCTION 12 +#else +#define FFI_TRAMPOLINE_SIZE 12 +#endif +#define FFI_TRAMPOLINE_CLOSURE_OFFSET FFI_TRAMPOLINE_SIZE +#endif + #endif diff --git a/native/libffi/src/arm/internal.h b/native/libffi/src/arm/internal.h new file mode 100644 index 0000000000..6cf0b2ae5d --- /dev/null +++ b/native/libffi/src/arm/internal.h @@ -0,0 +1,7 @@ +#define ARM_TYPE_VFP_S 0 +#define ARM_TYPE_VFP_D 1 +#define ARM_TYPE_VFP_N 2 +#define ARM_TYPE_INT64 3 +#define ARM_TYPE_INT 4 +#define ARM_TYPE_VOID 5 +#define ARM_TYPE_STRUCT 6 diff --git a/native/libffi/src/arm/sysv.S b/native/libffi/src/arm/sysv.S index 9b074156c7..71280ece82 100644 --- a/native/libffi/src/arm/sysv.S +++ b/native/libffi/src/arm/sysv.S @@ -1,8 +1,8 @@ /* ----------------------------------------------------------------------- sysv.S - Copyright (c) 1998, 2008, 2011 Red Hat, Inc. Copyright (c) 2011 Plausible Labs Cooperative, Inc. - - ARM Foreign Function Interface + + ARM Foreign Function Interface Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the @@ -25,9 +25,11 @@ DEALINGS IN THE SOFTWARE. ----------------------------------------------------------------------- */ -#define LIBFFI_ASM +#ifdef __arm__ +#define LIBFFI_ASM #include #include +<<<<<<< HEAD #ifdef HAVE_MACHINE_ASM_H #include #else @@ -77,38 +79,39 @@ #endif #if defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) \ +======= +#include +#include "internal.h" + +/* GCC 4.8 provides __ARM_ARCH; construct it otherwise. */ +#ifndef __ARM_ARCH +# if defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) \ + || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) \ + || defined(__ARM_ARCH_7EM__) +# define __ARM_ARCH 7 +# elif defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) \ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 || defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6Z__) \ || defined(__ARM_ARCH_6ZK__) || defined(__ARM_ARCH_6T2__) \ || defined(__ARM_ARCH_6M__) -# undef __ARM_ARCH__ -# define __ARM_ARCH__ 6 -#endif - -#if defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) \ - || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) \ - || defined(__ARM_ARCH_7EM__) -# undef __ARM_ARCH__ -# define __ARM_ARCH__ 7 -#endif - -#if __ARM_ARCH__ >= 5 -# define call_reg(x) blx x -#elif defined (__ARM_ARCH_4T__) -# define call_reg(x) mov lr, pc ; bx x -# if defined(__thumb__) || defined(__THUMB_INTERWORK__) -# define __INTERWORKING__ +# define __ARM_ARCH 6 +# elif defined(__ARM_ARCH_5__) || defined(__ARM_ARCH_5T__) \ + || defined(__ARM_ARCH_5E__) || defined(__ARM_ARCH_5TE__) \ + || defined(__ARM_ARCH_5TEJ__) +# define __ARM_ARCH 5 +# else +# define __ARM_ARCH 4 # endif -#else -# define call_reg(x) mov lr, pc ; mov pc, x #endif /* Conditionally compile unwinder directives. */ #ifdef __ARM_EABI__ -#define UNWIND +# define UNWIND(...) __VA_ARGS__ #else -#define UNWIND @ -#endif +# define UNWIND(...) +#endif +<<<<<<< HEAD .syntax unified #if defined(__thumb__) && !defined(__THUMB_INTERWORK__) @@ -130,23 +133,52 @@ _L__##name: .arm; \ ENTRY(name); \ UNWIND .fnstart +======= +#if defined(HAVE_AS_CFI_PSEUDO_OP) && defined(__ARM_EABI__) + .cfi_sections .debug_frame #endif -.macro RETLDM regs=, cond=, dirn=ia -#if defined (__INTERWORKING__) - .ifc "\regs","" - ldr\cond lr, [sp], #4 - .else - ldm\cond\dirn sp!, {\regs, lr} - .endif - bx\cond lr +#define CONCAT(a, b) CONCAT2(a, b) +#define CONCAT2(a, b) a ## b + +#ifdef __USER_LABEL_PREFIX__ +# define CNAME(X) CONCAT (__USER_LABEL_PREFIX__, X) #else - .ifc "\regs","" - ldr\cond pc, [sp], #4 - .else - ldm\cond\dirn sp!, {\regs, pc} +# define CNAME(X) X +#endif +#ifdef __ELF__ +# define SIZE(X) .size CNAME(X), . - CNAME(X) +# define TYPE(X, Y) .type CNAME(X), Y +#else +# define SIZE(X) +# define TYPE(X, Y) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 +#endif + +#define ARM_FUNC_START_LOCAL(name) \ + .align 3; \ + TYPE(CNAME(name), %function); \ + CNAME(name): + +#define ARM_FUNC_START(name) \ + .globl CNAME(name); \ + FFI_HIDDEN(CNAME(name)); \ + ARM_FUNC_START_LOCAL(name) + +#define ARM_FUNC_END(name) \ + SIZE(name) + +/* Aid in defining a jump table with 8 bytes between entries. */ +/* ??? The clang assembler doesn't handle .if with symbolic expressions. */ +#ifdef __clang__ +# define E(index) +#else +# define E(index) \ + .if . - 0b - 8*index; \ + .error "type table out of sync"; \ .endif #endif +<<<<<<< HEAD .endm @ r0: ffi_prep_args @@ -189,61 +221,121 @@ ARM_FUNC_START(ffi_call_SYSV) @ Remove the space we pushed for the args mov sp, fp +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 - @ Load r2 with the pointer to storage for the return value - ldr r2, [sp, #24] + .text + .syntax unified + .arm - @ Load r3 with the return type code - ldr r3, [sp, #12] +#ifndef __clang__ + /* We require interworking on LDM, which implies ARMv5T, + which implies the existance of BLX. */ + .arch armv5t +#endif - @ If the return value pointer is NULL, assume no return value. - cmp r2, #0 - beq LSYM(Lepilogue) + /* Note that we use STC and LDC to encode VFP instructions, + so that we do not need ".fpu vfp", nor get that added to + the object file attributes. These will not be executed + unless the FFI_VFP abi is used. */ -@ return INT - cmp r3, #FFI_TYPE_INT -#if defined(__SOFTFP__) || defined(__ARM_EABI__) - cmpne r3, #FFI_TYPE_FLOAT -#endif - streq r0, [r2] - beq LSYM(Lepilogue) + @ r0: stack + @ r1: frame + @ r2: fn + @ r3: vfp_used - @ return INT64 - cmp r3, #FFI_TYPE_SINT64 -#if defined(__SOFTFP__) || defined(__ARM_EABI__) - cmpne r3, #FFI_TYPE_DOUBLE +ARM_FUNC_START(ffi_call_VFP) + UNWIND(.fnstart) + cfi_startproc + + cmp r3, #3 @ load only d0 if possible +#ifdef __clang__ + vldrle d0, [r0] + vldmgt r0, {d0-d7} +#else + ldcle p11, cr0, [r0] @ vldrle d0, [r0] + ldcgt p11, cr0, [r0], {16} @ vldmgt r0, {d0-d7} #endif +<<<<<<< HEAD stmiaeq r2, {r0, r1} +======= + add r0, r0, #64 @ discard the vfp register args + /* FALLTHRU */ +ARM_FUNC_END(ffi_call_VFP) -#if !defined(__SOFTFP__) && !defined(__ARM_EABI__) - beq LSYM(Lepilogue) +ARM_FUNC_START(ffi_call_SYSV) + stm r1, {fp, lr} + mov fp, r1 + + @ This is a bit of a lie wrt the origin of the unwind info, but + @ now we've got the usual frame pointer and two saved registers. + UNWIND(.save {fp,lr}) + UNWIND(.setfp fp, sp) + cfi_def_cfa(fp, 8) + cfi_rel_offset(fp, 0) + cfi_rel_offset(lr, 4) + + mov sp, r0 @ install the stack pointer + mov lr, r2 @ move the fn pointer out of the way + ldr ip, [fp, #16] @ install the static chain + ldmia sp!, {r0-r3} @ move first 4 parameters in registers. + blx lr @ call fn +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 -@ return FLOAT - cmp r3, #FFI_TYPE_FLOAT - stfeqs f0, [r2] - beq LSYM(Lepilogue) + @ Load r2 with the pointer to storage for the return value + @ Load r3 with the return type code + ldr r2, [fp, #8] + ldr r3, [fp, #12] -@ return DOUBLE or LONGDOUBLE - cmp r3, #FFI_TYPE_DOUBLE - stfeqd f0, [r2] + @ Deallocate the stack with the arguments. + mov sp, fp + cfi_def_cfa_register(sp) + + @ Store values stored in registers. + .align 3 + add pc, pc, r3, lsl #3 + nop +0: +E(ARM_TYPE_VFP_S) +#ifdef __clang__ + vstr s0, [r2] +#else + stc p10, cr0, [r2] @ vstr s0, [r2] #endif - -LSYM(Lepilogue): -#if defined (__INTERWORKING__) - ldmia sp!, {r0-r3,fp, lr} - bx lr + pop {fp,pc} +E(ARM_TYPE_VFP_D) +#ifdef __clang__ + vstr d0, [r2] #else - ldmia sp!, {r0-r3,fp, pc} + stc p11, cr0, [r2] @ vstr d0, [r2] #endif - -.ffi_call_SYSV_end: - UNWIND .fnend -#ifdef __ELF__ - .size CNAME(ffi_call_SYSV),.ffi_call_SYSV_end-CNAME(ffi_call_SYSV) + pop {fp,pc} +E(ARM_TYPE_VFP_N) +#ifdef __clang__ + vstm r2, {d0-d3} +#else + stc p11, cr0, [r2], {8} @ vstm r2, {d0-d3} #endif + pop {fp,pc} +E(ARM_TYPE_INT64) + str r1, [r2, #4] + nop +E(ARM_TYPE_INT) + str r0, [r2] + pop {fp,pc} +E(ARM_TYPE_VOID) + pop {fp,pc} + nop +E(ARM_TYPE_STRUCT) + pop {fp,pc} + + cfi_endproc + UNWIND(.fnend) +ARM_FUNC_END(ffi_call_SYSV) /* +<<<<<<< HEAD unsigned int FFI_HIDDEN ffi_closure_inner (closure, respp, args) ffi_closure *closure; @@ -268,56 +360,170 @@ ARM_FUNC_START(ffi_closure_SYSV) cmp r0, #FFI_TYPE_FLOAT #if defined(__SOFTFP__) || defined(__ARM_EABI__) beq .Lretint -#else - beq .Lretfloat +======= + int ffi_closure_inner_* (cif, fun, user_data, frame) +*/ + +ARM_FUNC_START(ffi_go_closure_SYSV) + cfi_startproc + stmdb sp!, {r0-r3} @ save argument regs + cfi_adjust_cfa_offset(16) + ldr r0, [ip, #4] @ load cif + ldr r1, [ip, #8] @ load fun + mov r2, ip @ load user_data + b 0f + cfi_endproc +ARM_FUNC_END(ffi_go_closure_SYSV) + +ARM_FUNC_START(ffi_closure_SYSV) + UNWIND(.fnstart) + cfi_startproc + stmdb sp!, {r0-r3} @ save argument regs + cfi_adjust_cfa_offset(16) + +#if FFI_EXEC_TRAMPOLINE_TABLE + ldr ip, [ip] @ ip points to the config page, dereference to get the ffi_closure* #endif + ldr r0, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET] @ load cif + ldr r1, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET+4] @ load fun + ldr r2, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET+8] @ load user_data +0: + add ip, sp, #16 @ compute entry sp + sub sp, sp, #64+32 @ allocate frame + cfi_adjust_cfa_offset(64+32) + stmdb sp!, {ip,lr} + + /* Remember that EABI unwind info only applies at call sites. + We need do nothing except note the save of the stack pointer + and the link registers. */ + UNWIND(.save {sp,lr}) + cfi_adjust_cfa_offset(8) + cfi_rel_offset(lr, 4) + + add r3, sp, #8 @ load frame + bl CNAME(ffi_closure_inner_SYSV) + + @ Load values returned in registers. + add r2, sp, #8+64 @ load result + adr r3, CNAME(ffi_closure_ret) + add pc, r3, r0, lsl #3 + cfi_endproc + UNWIND(.fnend) +ARM_FUNC_END(ffi_closure_SYSV) + +ARM_FUNC_START(ffi_go_closure_VFP) + cfi_startproc + stmdb sp!, {r0-r3} @ save argument regs + cfi_adjust_cfa_offset(16) + ldr r0, [ip, #4] @ load cif + ldr r1, [ip, #8] @ load fun + mov r2, ip @ load user_data + b 0f + cfi_endproc +ARM_FUNC_END(ffi_go_closure_VFP) - cmp r0, #FFI_TYPE_DOUBLE -#if defined(__SOFTFP__) || defined(__ARM_EABI__) - beq .Lretlonglong +ARM_FUNC_START(ffi_closure_VFP) + UNWIND(.fnstart) + cfi_startproc + stmdb sp!, {r0-r3} @ save argument regs + cfi_adjust_cfa_offset(16) + +#if FFI_EXEC_TRAMPOLINE_TABLE + ldr ip, [ip] @ ip points to the config page, dereference to get the ffi_closure* +#endif + ldr r0, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET] @ load cif + ldr r1, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET+4] @ load fun + ldr r2, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET+8] @ load user_data +0: + add ip, sp, #16 + sub sp, sp, #64+32 @ allocate frame + cfi_adjust_cfa_offset(64+32) +#ifdef __clang__ + vstm sp, {d0-d7} +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #else - beq .Lretdouble + stc p11, cr0, [sp], {16} @ vstm sp, {d0-d7} #endif - - cmp r0, #FFI_TYPE_LONGDOUBLE -#if defined(__SOFTFP__) || defined(__ARM_EABI__) - beq .Lretlonglong + stmdb sp!, {ip,lr} + + /* See above. */ + UNWIND(.save {sp,lr}) + cfi_adjust_cfa_offset(8) + cfi_rel_offset(lr, 4) + + add r3, sp, #8 @ load frame + bl CNAME(ffi_closure_inner_VFP) + + @ Load values returned in registers. + add r2, sp, #8+64 @ load result + adr r3, CNAME(ffi_closure_ret) + add pc, r3, r0, lsl #3 + cfi_endproc + UNWIND(.fnend) +ARM_FUNC_END(ffi_closure_VFP) + +/* Load values returned in registers for both closure entry points. + Note that we use LDM with SP in the register set. This is deprecated + by ARM, but not yet unpredictable. */ + +ARM_FUNC_START_LOCAL(ffi_closure_ret) + cfi_startproc + cfi_rel_offset(sp, 0) + cfi_rel_offset(lr, 4) +0: +E(ARM_TYPE_VFP_S) +#ifdef __clang__ + vldr s0, [r2] #else - beq .Lretlongdouble + ldc p10, cr0, [r2] @ vldr s0, [r2] #endif - - cmp r0, #FFI_TYPE_SINT64 - beq .Lretlonglong -.Lclosure_epilogue: - add sp, sp, #16 - ldmfd sp, {sp, pc} -.Lretint: - ldr r0, [sp] - b .Lclosure_epilogue -.Lretlonglong: - ldr r0, [sp] - ldr r1, [sp, #4] - b .Lclosure_epilogue - -#if !defined(__SOFTFP__) && !defined(__ARM_EABI__) -.Lretfloat: - ldfs f0, [sp] - b .Lclosure_epilogue -.Lretdouble: - ldfd f0, [sp] - b .Lclosure_epilogue -.Lretlongdouble: - ldfd f0, [sp] - b .Lclosure_epilogue + ldm sp, {sp,pc} +E(ARM_TYPE_VFP_D) +#ifdef __clang__ + vldr d0, [r2] +#else + ldc p11, cr0, [r2] @ vldr d0, [r2] #endif - -.ffi_closure_SYSV_end: - UNWIND .fnend -#ifdef __ELF__ - .size CNAME(ffi_closure_SYSV),.ffi_closure_SYSV_end-CNAME(ffi_closure_SYSV) + ldm sp, {sp,pc} +E(ARM_TYPE_VFP_N) +#ifdef __clang__ + vldm r2, {d0-d3} +#else + ldc p11, cr0, [r2], {8} @ vldm r2, {d0-d3} +#endif + ldm sp, {sp,pc} +E(ARM_TYPE_INT64) + ldr r1, [r2, #4] + nop +E(ARM_TYPE_INT) + ldr r0, [r2] + ldm sp, {sp,pc} +E(ARM_TYPE_VOID) + ldm sp, {sp,pc} + nop +E(ARM_TYPE_STRUCT) + ldm sp, {sp,pc} + cfi_endproc +ARM_FUNC_END(ffi_closure_ret) + +#if FFI_EXEC_TRAMPOLINE_TABLE + +#ifdef __MACH__ +#include + +.align PAGE_MAX_SHIFT +ARM_FUNC_START(ffi_closure_trampoline_table_page) +.rept PAGE_MAX_SIZE / FFI_TRAMPOLINE_SIZE + adr ip, #-PAGE_MAX_SIZE @ the config page is PAGE_MAX_SIZE behind the trampoline page + sub ip, #8 @ account for pc bias + ldr pc, [ip, #4] @ jump to ffi_closure_SYSV or ffi_closure_VFP +.endr +ARM_FUNC_END(ffi_closure_trampoline_table_page) #endif +#else +<<<<<<< HEAD /* Below are VFP hard-float ABI call and closure implementations. Add VFP FPU directive here. This is only compiled into the library under EABI. */ @@ -484,11 +690,16 @@ ARM_FUNC_START(ffi_closure_VFP) .size CNAME(ffi_closure_VFP),.ffi_closure_VFP_end-CNAME(ffi_closure_VFP) #endif #endif - -ENTRY(ffi_arm_trampoline) - stmfd sp!, {r0-r3} - ldr r0, [pc] - ldr pc, [pc] +======= +ARM_FUNC_START(ffi_arm_trampoline) +0: adr ip, 0b + ldr pc, 1f +1: .long 0 +ARM_FUNC_END(ffi_arm_trampoline) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + +#endif /* FFI_EXEC_TRAMPOLINE_TABLE */ +#endif /* __arm__ */ #if defined __ELF__ && defined __linux__ .section .note.GNU-stack,"",%progbits diff --git a/native/libffi/src/arm/sysv_msvc_arm32.S b/native/libffi/src/arm/sysv_msvc_arm32.S new file mode 100644 index 0000000000..5c99d0207a --- /dev/null +++ b/native/libffi/src/arm/sysv_msvc_arm32.S @@ -0,0 +1,311 @@ +/* ----------------------------------------------------------------------- + sysv.S - Copyright (c) 1998, 2008, 2011 Red Hat, Inc. + Copyright (c) 2011 Plausible Labs Cooperative, Inc. + Copyright (c) 2019 Microsoft Corporation. + + ARM Foreign Function Interface + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + ``Software''), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice shall be included + in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + ----------------------------------------------------------------------- */ + +#define LIBFFI_ASM +#include +#include +#include +#include "internal.h" +#include "ksarm.h" + + + ; 8 byte aligned AREA to support 8 byte aligned jump tables + MACRO + NESTED_ENTRY_FFI $FuncName, $AreaName, $ExceptHandler + + ; compute the function's labels + __DeriveFunctionLabels $FuncName + + ; determine the area we will put the function into +__FuncArea SETS "|.text|" + IF "$AreaName" != "" +__FuncArea SETS "$AreaName" + ENDIF + + ; set up the exception handler itself +__FuncExceptionHandler SETS "" + IF "$ExceptHandler" != "" +__FuncExceptionHandler SETS "|$ExceptHandler|" + ENDIF + + ; switch to the specified area, jump tables require 8 byte alignment + AREA $__FuncArea,CODE,CODEALIGN,ALIGN=3,READONLY + + ; export the function name + __ExportProc $FuncName + + ; flush any pending literal pool stuff + ROUT + + ; reset the state of the unwind code tracking + __ResetUnwindState + + MEND + +; MACRO +; TABLE_ENTRY $Type, $Table +;$Type_$Table +; MEND + +#define E(index,table) return_##index##_##table + + ; r0: stack + ; r1: frame + ; r2: fn + ; r3: vfp_used + + ; fake entry point exists only to generate exists only to + ; generate .pdata for exception unwinding + NESTED_ENTRY_FFI ffi_call_VFP_fake + PROLOG_PUSH {r11, lr} ; save fp and lr for unwind + + ALTERNATE_ENTRY ffi_call_VFP + cmp r3, #3 ; load only d0 if possible + vldrle d0, [r0] + vldmgt r0, {d0-d7} + add r0, r0, #64 ; discard the vfp register args + b ffi_call_SYSV + NESTED_END ffi_call_VFP_fake + + ; fake entry point exists only to generate exists only to + ; generate .pdata for exception unwinding + NESTED_ENTRY_FFI ffi_call_SYSV_fake + PROLOG_PUSH {r11, lr} ; save fp and lr for unwind + + ALTERNATE_ENTRY ffi_call_SYSV + stm r1, {fp, lr} + mov fp, r1 + + mov sp, r0 ; install the stack pointer + mov lr, r2 ; move the fn pointer out of the way + ldr ip, [fp, #16] ; install the static chain + ldmia sp!, {r0-r3} ; move first 4 parameters in registers. + blx lr ; call fn + + ; Load r2 with the pointer to storage for the return value + ; Load r3 with the return type code + ldr r2, [fp, #8] + ldr r3, [fp, #12] + + ; Deallocate the stack with the arguments. + mov sp, fp + + ; Store values stored in registers. + ALIGN 8 + lsl r3, #3 + add r3, r3, pc + add r3, #8 + mov pc, r3 + + +E(ARM_TYPE_VFP_S, ffi_call) + ALIGN 8 + vstr s0, [r2] + pop {fp,pc} +E(ARM_TYPE_VFP_D, ffi_call) + ALIGN 8 + vstr d0, [r2] + pop {fp,pc} +E(ARM_TYPE_VFP_N, ffi_call) + ALIGN 8 + vstm r2, {d0-d3} + pop {fp,pc} +E(ARM_TYPE_INT64, ffi_call) + ALIGN 8 + str r1, [r2, #4] + nop +E(ARM_TYPE_INT, ffi_call) + ALIGN 8 + str r0, [r2] + pop {fp,pc} +E(ARM_TYPE_VOID, ffi_call) + ALIGN 8 + pop {fp,pc} + nop +E(ARM_TYPE_STRUCT, ffi_call) + ALIGN 8 + cmp r3, #ARM_TYPE_STRUCT + pop {fp,pc} + NESTED_END ffi_call_SYSV_fake + + IMPORT |ffi_closure_inner_SYSV| + /* + int ffi_closure_inner_SYSV + ( + cif, ; r0 + fun, ; r1 + user_data, ; r2 + frame ; r3 + ) + */ + + NESTED_ENTRY_FFI ffi_go_closure_SYSV + stmdb sp!, {r0-r3} ; save argument regs + ldr r0, [ip, #4] ; load cif + ldr r1, [ip, #8] ; load fun + mov r2, ip ; load user_data + b ffi_go_closure_SYSV_0 + NESTED_END ffi_go_closure_SYSV + + ; r3: ffi_closure + + ; fake entry point exists only to generate exists only to + ; generate .pdata for exception unwinding + NESTED_ENTRY_FFI ffi_closure_SYSV_fake + PROLOG_PUSH {r11, lr} ; save fp and lr for unwind + ALTERNATE_ENTRY ffi_closure_SYSV + ldmfd sp!, {ip,r0} ; restore fp (r0 is used for stack alignment) + stmdb sp!, {r0-r3} ; save argument regs + + ldr r0, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET] ; ffi_closure->cif + ldr r1, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET+4] ; ffi_closure->fun + ldr r2, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET+8] ; ffi_closure->user_data + + ALTERNATE_ENTRY ffi_go_closure_SYSV_0 + add ip, sp, #16 ; compute entry sp + + sub sp, sp, #64+32 ; allocate frame parameter (sizeof(vfp_space) = 64, sizeof(result) = 32) + mov r3, sp ; set frame parameter + stmdb sp!, {ip,lr} + + bl ffi_closure_inner_SYSV ; call the Python closure + + ; Load values returned in registers. + add r2, sp, #64+8 ; address of closure_frame->result + bl ffi_closure_ret ; move result to correct register or memory for type + + ldmfd sp!, {ip,lr} + mov sp, ip ; restore stack pointer + mov pc, lr + NESTED_END ffi_closure_SYSV_fake + + IMPORT |ffi_closure_inner_VFP| + /* + int ffi_closure_inner_VFP + ( + cif, ; r0 + fun, ; r1 + user_data, ; r2 + frame ; r3 + ) + */ + + NESTED_ENTRY_FFI ffi_go_closure_VFP + stmdb sp!, {r0-r3} ; save argument regs + ldr r0, [ip, #4] ; load cif + ldr r1, [ip, #8] ; load fun + mov r2, ip ; load user_data + b ffi_go_closure_VFP_0 + NESTED_END ffi_go_closure_VFP + + ; fake entry point exists only to generate exists only to + ; generate .pdata for exception unwinding + ; r3: closure + NESTED_ENTRY_FFI ffi_closure_VFP_fake + PROLOG_PUSH {r11, lr} ; save fp and lr for unwind + + ALTERNATE_ENTRY ffi_closure_VFP + ldmfd sp!, {ip,r0} ; restore fp (r0 is used for stack alignment) + stmdb sp!, {r0-r3} ; save argument regs + + ldr r0, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET] ; load cif + ldr r1, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET+4] ; load fun + ldr r2, [ip, #FFI_TRAMPOLINE_CLOSURE_OFFSET+8] ; load user_data + + ALTERNATE_ENTRY ffi_go_closure_VFP_0 + add ip, sp, #16 ; compute entry sp + sub sp, sp, #32 ; save space for closure_frame->result + vstmdb sp!, {d0-d7} ; push closure_frame->vfp_space + + mov r3, sp ; save closure_frame + stmdb sp!, {ip,lr} + + bl ffi_closure_inner_VFP + + ; Load values returned in registers. + add r2, sp, #64+8 ; load result + bl ffi_closure_ret + ldmfd sp!, {ip,lr} + mov sp, ip ; restore stack pointer + mov pc, lr + NESTED_END ffi_closure_VFP_fake + +/* Load values returned in registers for both closure entry points. + Note that we use LDM with SP in the register set. This is deprecated + by ARM, but not yet unpredictable. */ + + NESTED_ENTRY_FFI ffi_closure_ret + stmdb sp!, {fp,lr} + + ALIGN 8 + lsl r0, #3 + add r0, r0, pc + add r0, #8 + mov pc, r0 + +E(ARM_TYPE_VFP_S, ffi_closure) + ALIGN 8 + vldr s0, [r2] + b call_epilogue +E(ARM_TYPE_VFP_D, ffi_closure) + ALIGN 8 + vldr d0, [r2] + b call_epilogue +E(ARM_TYPE_VFP_N, ffi_closure) + ALIGN 8 + vldm r2, {d0-d3} + b call_epilogue +E(ARM_TYPE_INT64, ffi_closure) + ALIGN 8 + ldr r1, [r2, #4] + nop +E(ARM_TYPE_INT, ffi_closure) + ALIGN 8 + ldr r0, [r2] + b call_epilogue +E(ARM_TYPE_VOID, ffi_closure) + ALIGN 8 + b call_epilogue + nop +E(ARM_TYPE_STRUCT, ffi_closure) + ALIGN 8 + b call_epilogue +call_epilogue + ldmfd sp!, {fp,pc} + NESTED_END ffi_closure_ret + + AREA |.trampoline|, DATA, THUMB, READONLY + EXPORT |ffi_arm_trampoline| +|ffi_arm_trampoline| DATA +thisproc adr ip, thisproc + stmdb sp!, {ip, r0} + ldr pc, [pc, #0] + DCD 0 + ;ENDP + + END \ No newline at end of file diff --git a/native/libffi/src/arm/trampoline.S b/native/libffi/src/arm/trampoline.S deleted file mode 100644 index 935e8de179..0000000000 --- a/native/libffi/src/arm/trampoline.S +++ /dev/null @@ -1,4450 +0,0 @@ -# GENERATED CODE - DO NOT EDIT -# This file was generated by src/arm/gentramp.sh - -# Copyright (c) 2010, Plausible Labs Cooperative, Inc. -# -# Permission is hereby granted, free of charge, to any person obtaining -# a copy of this software and associated documentation files (the -# Software''), to deal in the Software without restriction, including -# without limitation the rights to use, copy, modify, merge, publish, -# distribute, sublicense, and/or sell copies of the Software, and to -# permit persons to whom the Software is furnished to do so, subject to -# the following conditions: -# -# The above copyright notice and this permission notice shall be included -# in all copies or substantial portions of the Software. -# -# THE SOFTWARE IS PROVIDED AS IS'', WITHOUT WARRANTY OF ANY KIND, -# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF -# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND -# NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT -# HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, -# WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, -# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER -# DEALINGS IN THE SOFTWARE. -# ----------------------------------------------------------------------- - -.text -.align 12 -.globl _ffi_closure_trampoline_table_page -_ffi_closure_trampoline_table_page: - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - - - // trampoline - // Save to stack - stmfd sp!, {r0-r3} - - // Load the context argument from the config page. - // This places the first usable config value at _ffi_closure_trampoline_table-4080 - // This accounts for the above 4-byte stmfd instruction, plus 8 bytes constant when loading from pc. - ldr r0, [pc, #-4092] - - // Load the jump address from the config page. - ldr pc, [pc, #-4092] - diff --git a/native/libffi/src/closures.c b/native/libffi/src/closures.c index 721ff00ea4..49b8a12dfd 100644 --- a/native/libffi/src/closures.c +++ b/native/libffi/src/closures.c @@ -1,5 +1,6 @@ /* ----------------------------------------------------------------------- - closures.c - Copyright (c) 2007, 2009, 2010 Red Hat, Inc. + closures.c - Copyright (c) 2019 Anthony Green + Copyright (c) 2007, 2009, 2010 Red Hat, Inc. Copyright (C) 2007, 2009, 2010 Free Software Foundation, Inc Copyright (c) 2011 Plausible Labs Cooperative, Inc. @@ -30,11 +31,92 @@ #define _GNU_SOURCE 1 #endif +#include #include #include +#ifdef __NetBSD__ +#include +#endif + +#if __NetBSD_Version__ - 0 >= 799007200 +/* NetBSD with PROT_MPROTECT */ +#include + +#include +#include + +static const size_t overhead = + (sizeof(max_align_t) > sizeof(void *) + sizeof(size_t)) ? + sizeof(max_align_t) + : sizeof(void *) + sizeof(size_t); + +#define ADD_TO_POINTER(p, d) ((void *)((uintptr_t)(p) + (d))) + +void * +ffi_closure_alloc (size_t size, void **code) +{ + static size_t page_size; + size_t rounded_size; + void *codeseg, *dataseg; + int prot; + + /* Expect that PAX mprotect is active and a separate code mapping is necessary. */ + if (!code) + return NULL; + + /* Obtain system page size. */ + if (!page_size) + page_size = sysconf(_SC_PAGESIZE); + + /* Round allocation size up to the next page, keeping in mind the size field and pointer to code map. */ + rounded_size = (size + overhead + page_size - 1) & ~(page_size - 1); + + /* Primary mapping is RW, but request permission to switch to PROT_EXEC later. */ + prot = PROT_READ | PROT_WRITE | PROT_MPROTECT(PROT_EXEC); + dataseg = mmap(NULL, rounded_size, prot, MAP_ANON | MAP_PRIVATE, -1, 0); + if (dataseg == MAP_FAILED) + return NULL; + + /* Create secondary mapping and switch it to RX. */ + codeseg = mremap(dataseg, rounded_size, NULL, rounded_size, MAP_REMAPDUP); + if (codeseg == MAP_FAILED) { + munmap(dataseg, rounded_size); + return NULL; + } + if (mprotect(codeseg, rounded_size, PROT_READ | PROT_EXEC) == -1) { + munmap(codeseg, rounded_size); + munmap(dataseg, rounded_size); + return NULL; + } + + /* Remember allocation size and location of the secondary mapping for ffi_closure_free. */ + memcpy(dataseg, &rounded_size, sizeof(rounded_size)); + memcpy(ADD_TO_POINTER(dataseg, sizeof(size_t)), &codeseg, sizeof(void *)); + *code = ADD_TO_POINTER(codeseg, overhead); + return ADD_TO_POINTER(dataseg, overhead); +} + +void +ffi_closure_free (void *ptr) +{ + void *codeseg, *dataseg; + size_t rounded_size; + + dataseg = ADD_TO_POINTER(ptr, -overhead); + memcpy(&rounded_size, dataseg, sizeof(rounded_size)); + memcpy(&codeseg, ADD_TO_POINTER(dataseg, sizeof(size_t)), sizeof(void *)); + munmap(dataseg, rounded_size); + munmap(codeseg, rounded_size); +} +#else /* !NetBSD with PROT_MPROTECT */ + #if !FFI_MMAP_EXEC_WRIT && !FFI_EXEC_TRAMPOLINE_TABLE +<<<<<<< HEAD # if __gnu_linux__ && !defined(__ANDROID__) +======= +# if __linux__ && !defined(__ANDROID__) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* This macro indicates it may be forbidden to map anonymous memory with both write and execute permission. Code compiled when this option is defined will attempt to map such pages once, but if it @@ -45,7 +127,7 @@ # define FFI_MMAP_EXEC_WRIT 1 # define HAVE_MNTENT 1 # endif -# if defined(X86_WIN32) || defined(X86_WIN64) || defined(__OS2__) +# if defined(_WIN32) || defined(__OS2__) /* Windows systems may have Data Execution Protection (DEP) enabled, which requires the use of VirtualMalloc/VirtualFree to alloc/free executable memory. */ @@ -54,7 +136,7 @@ #endif #if FFI_MMAP_EXEC_WRIT && !defined FFI_MMAP_EXEC_SELINUX -# ifdef __linux__ +# if defined(__linux__) && !defined(__ANDROID__) /* When defined to 1 check for SELinux and if SELinux is active, don't attempt PROT_EXEC|PROT_WRITE mapping at all, as that might cause audit messages. */ @@ -64,11 +146,222 @@ #if FFI_CLOSURES -# if FFI_EXEC_TRAMPOLINE_TABLE +#if FFI_EXEC_TRAMPOLINE_TABLE + +#ifdef __MACH__ + +#include +#include +#ifdef HAVE_PTRAUTH +#include +#endif +#include +#include + +extern void *ffi_closure_trampoline_table_page; + +typedef struct ffi_trampoline_table ffi_trampoline_table; +typedef struct ffi_trampoline_table_entry ffi_trampoline_table_entry; + +struct ffi_trampoline_table +{ + /* contiguous writable and executable pages */ + vm_address_t config_page; + vm_address_t trampoline_page; + + /* free list tracking */ + uint16_t free_count; + ffi_trampoline_table_entry *free_list; + ffi_trampoline_table_entry *free_list_pool; + + ffi_trampoline_table *prev; + ffi_trampoline_table *next; +}; + +struct ffi_trampoline_table_entry +{ + void *(*trampoline) (void); + ffi_trampoline_table_entry *next; +}; + +/* Total number of trampolines that fit in one trampoline table */ +#define FFI_TRAMPOLINE_COUNT (PAGE_MAX_SIZE / FFI_TRAMPOLINE_SIZE) + +static pthread_mutex_t ffi_trampoline_lock = PTHREAD_MUTEX_INITIALIZER; +static ffi_trampoline_table *ffi_trampoline_tables = NULL; + +static ffi_trampoline_table * +ffi_trampoline_table_alloc (void) +{ + ffi_trampoline_table *table; + vm_address_t config_page; + vm_address_t trampoline_page; + vm_address_t trampoline_page_template; + vm_prot_t cur_prot; + vm_prot_t max_prot; + kern_return_t kt; + uint16_t i; + + /* Allocate two pages -- a config page and a placeholder page */ + config_page = 0x0; + kt = vm_allocate (mach_task_self (), &config_page, PAGE_MAX_SIZE * 2, + VM_FLAGS_ANYWHERE); + if (kt != KERN_SUCCESS) + return NULL; + + /* Remap the trampoline table on top of the placeholder page */ + trampoline_page = config_page + PAGE_MAX_SIZE; + trampoline_page_template = (vm_address_t)&ffi_closure_trampoline_table_page; +#ifdef __arm__ + /* ffi_closure_trampoline_table_page can be thumb-biased on some ARM archs */ + trampoline_page_template &= ~1UL; +#endif + kt = vm_remap (mach_task_self (), &trampoline_page, PAGE_MAX_SIZE, 0x0, + VM_FLAGS_OVERWRITE, mach_task_self (), trampoline_page_template, + FALSE, &cur_prot, &max_prot, VM_INHERIT_SHARE); + if (kt != KERN_SUCCESS) + { + vm_deallocate (mach_task_self (), config_page, PAGE_MAX_SIZE * 2); + return NULL; + } + + /* We have valid trampoline and config pages */ + table = calloc (1, sizeof (ffi_trampoline_table)); + table->free_count = FFI_TRAMPOLINE_COUNT; + table->config_page = config_page; + table->trampoline_page = trampoline_page; + + /* Create and initialize the free list */ + table->free_list_pool = + calloc (FFI_TRAMPOLINE_COUNT, sizeof (ffi_trampoline_table_entry)); + + for (i = 0; i < table->free_count; i++) + { + ffi_trampoline_table_entry *entry = &table->free_list_pool[i]; + entry->trampoline = + (void *) (table->trampoline_page + (i * FFI_TRAMPOLINE_SIZE)); + + if (i < table->free_count - 1) + entry->next = &table->free_list_pool[i + 1]; + } + + table->free_list = table->free_list_pool; + + return table; +} + +static void +ffi_trampoline_table_free (ffi_trampoline_table *table) +{ + /* Remove from the list */ + if (table->prev != NULL) + table->prev->next = table->next; + + if (table->next != NULL) + table->next->prev = table->prev; + + /* Deallocate pages */ + vm_deallocate (mach_task_self (), table->config_page, PAGE_MAX_SIZE * 2); + + /* Deallocate free list */ + free (table->free_list_pool); + free (table); +} + +void * +ffi_closure_alloc (size_t size, void **code) +{ + /* Create the closure */ + ffi_closure *closure = malloc (size); + if (closure == NULL) + return NULL; + + pthread_mutex_lock (&ffi_trampoline_lock); + + /* Check for an active trampoline table with available entries. */ + ffi_trampoline_table *table = ffi_trampoline_tables; + if (table == NULL || table->free_list == NULL) + { + table = ffi_trampoline_table_alloc (); + if (table == NULL) + { + pthread_mutex_unlock (&ffi_trampoline_lock); + free (closure); + return NULL; + } + + /* Insert the new table at the top of the list */ + table->next = ffi_trampoline_tables; + if (table->next != NULL) + table->next->prev = table; + + ffi_trampoline_tables = table; + } + + /* Claim the free entry */ + ffi_trampoline_table_entry *entry = ffi_trampoline_tables->free_list; + ffi_trampoline_tables->free_list = entry->next; + ffi_trampoline_tables->free_count--; + entry->next = NULL; + + pthread_mutex_unlock (&ffi_trampoline_lock); + + /* Initialize the return values */ + *code = entry->trampoline; +#ifdef HAVE_PTRAUTH + *code = ptrauth_sign_unauthenticated (*code, ptrauth_key_asia, 0); +#endif + closure->trampoline_table = table; + closure->trampoline_table_entry = entry; + + return closure; +} + +void +ffi_closure_free (void *ptr) +{ + ffi_closure *closure = ptr; + + pthread_mutex_lock (&ffi_trampoline_lock); + + /* Fetch the table and entry references */ + ffi_trampoline_table *table = closure->trampoline_table; + ffi_trampoline_table_entry *entry = closure->trampoline_table_entry; + + /* Return the entry to the free list */ + entry->next = table->free_list; + table->free_list = entry; + table->free_count++; + + /* If all trampolines within this table are free, and at least one other table exists, deallocate + * the table */ + if (table->free_count == FFI_TRAMPOLINE_COUNT + && ffi_trampoline_tables != table) + { + ffi_trampoline_table_free (table); + } + else if (ffi_trampoline_tables != table) + { + /* Otherwise, bump this table to the top of the list */ + table->prev = NULL; + table->next = ffi_trampoline_tables; + if (ffi_trampoline_tables != NULL) + ffi_trampoline_tables->prev = table; + + ffi_trampoline_tables = table; + } + + pthread_mutex_unlock (&ffi_trampoline_lock); + + /* Free the closure */ + free (closure); +} + +#endif // Per-target implementation; It's unclear what can reasonable be shared between two OS/architecture implementations. -# elif FFI_MMAP_EXEC_WRIT /* !FFI_EXEC_TRAMPOLINE_TABLE */ +#elif FFI_MMAP_EXEC_WRIT /* !FFI_EXEC_TRAMPOLINE_TABLE */ #define USE_LOCKS 1 #define USE_DL_PREFIX 1 @@ -94,14 +387,6 @@ /* Don't allocate more than a page unless needed. */ #define DEFAULT_GRANULARITY ((size_t)malloc_getpagesize) -#if FFI_CLOSURE_TEST -/* Don't release single pages, to avoid a worst-case scenario of - continuously allocating and releasing single pages, but release - pairs of pages, which should do just as well given that allocations - are likely to be small. */ -#define DEFAULT_TRIM_THRESHOLD ((size_t)malloc_getpagesize) -#endif - #include #include #include @@ -111,7 +396,7 @@ #endif #include #include -#if !defined(X86_WIN32) && !defined(X86_WIN64) +#if !defined(_WIN32) #ifdef HAVE_MNTENT #include #endif /* HAVE_MNTENT */ @@ -237,11 +522,11 @@ static int dlmalloc_trim(size_t) MAYBE_UNUSED; static size_t dlmalloc_usable_size(void*) MAYBE_UNUSED; static void dlmalloc_stats(void) MAYBE_UNUSED; -#if !(defined(X86_WIN32) || defined(X86_WIN64) || defined(__OS2__)) || defined (__CYGWIN__) || defined(__INTERIX) +#if !(defined(_WIN32) || defined(__OS2__)) || defined (__CYGWIN__) || defined(__INTERIX) /* Use these for mmap and munmap within dlmalloc.c. */ static void *dlmmap(void *, size_t, int, int, int, off_t); static int dlmunmap(void *, size_t); -#endif /* !(defined(X86_WIN32) || defined(X86_WIN64) || defined(__OS2__)) || defined (__CYGWIN__) || defined(__INTERIX) */ +#endif /* !(defined(_WIN32) || defined(__OS2__)) || defined (__CYGWIN__) || defined(__INTERIX) */ #define mmap dlmmap #define munmap dlmunmap @@ -251,7 +536,7 @@ static int dlmunmap(void *, size_t); #undef mmap #undef munmap -#if !(defined(X86_WIN32) || defined(X86_WIN64) || defined(__OS2__)) || defined (__CYGWIN__) || defined(__INTERIX) +#if !(defined(_WIN32) || defined(__OS2__)) || defined (__CYGWIN__) || defined(__INTERIX) /* A mutex used to synchronize access to *exec* variables in this file. */ static pthread_mutex_t open_temp_exec_file_mutex = PTHREAD_MUTEX_INITIALIZER; @@ -308,7 +593,11 @@ open_temp_exec_file_dir (const char *dir) } #endif +<<<<<<< HEAD lendir = strlen (dir); +======= + lendir = (int) strlen (dir); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 tempname = __builtin_alloca (lendir + sizeof (suffix)); if (!tempname) @@ -449,6 +738,36 @@ open_temp_exec_file (void) return fd; } +/* We need to allocate space in a file that will be backing a writable + mapping. Several problems exist with the usual approaches: + - fallocate() is Linux-only + - posix_fallocate() is not available on all platforms + - ftruncate() does not allocate space on filesystems with sparse files + Failure to allocate the space will cause SIGBUS to be thrown when + the mapping is subsequently written to. */ +static int +allocate_space (int fd, off_t offset, off_t len) +{ + static size_t page_size; + + /* Obtain system page size. */ + if (!page_size) + page_size = sysconf(_SC_PAGESIZE); + + unsigned char buf[page_size]; + memset (buf, 0, page_size); + + while (len > 0) + { + off_t to_write = (len < page_size) ? len : page_size; + if (write (fd, buf, to_write) < to_write) + return -1; + len -= to_write; + } + + return 0; +} + /* Map in a chunk of memory from the temporary exec file into separate locations in the virtual memory address space, one writable and one executable. Returns the address of the writable portion, after @@ -470,7 +789,7 @@ dlmmap_locked (void *start, size_t length, int prot, int flags, off_t offset) offset = execsize; - if (ftruncate (execfd, offset + length)) + if (allocate_space (execfd, offset, length)) return MFAIL; flags &= ~(MAP_PRIVATE | MAP_ANONYMOUS); @@ -485,7 +804,13 @@ dlmmap_locked (void *start, size_t length, int prot, int flags, off_t offset) close (execfd); goto retry_open; } - ftruncate (execfd, offset); + if (ftruncate (execfd, offset) != 0) + { + /* Fixme : Error logs can be added here. Returning an error for + * ftruncte() will not add any advantage as it is being + * validating in the error case. */ + } + return MFAIL; } else if (!offset @@ -497,7 +822,12 @@ dlmmap_locked (void *start, size_t length, int prot, int flags, off_t offset) if (start == MFAIL) { munmap (ptr, length); - ftruncate (execfd, offset); + if (ftruncate (execfd, offset) != 0) + { + /* Fixme : Error logs can be added here. Returning an error for + * ftruncte() will not add any advantage as it is being + * validating in the error case. */ + } return start; } @@ -521,10 +851,6 @@ dlmmap (void *start, size_t length, int prot, && flags == (MAP_PRIVATE | MAP_ANONYMOUS) && fd == -1 && offset == 0); -#if FFI_CLOSURE_TEST - printf ("mapping in %zi\n", length); -#endif - if (execfd == -1 && is_emutramp_enabled ()) { ptr = mmap (start, length, prot & ~PROT_EXEC, flags, fd, offset); @@ -570,10 +896,6 @@ dlmunmap (void *start, size_t length) msegmentptr seg = segment_holding (gm, start); void *code; -#if FFI_CLOSURE_TEST - printf ("unmapping %zi\n", length); -#endif - if (seg && (code = add_segment_exec_offset (start, seg)) != start) { int ret = munmap (code, length); @@ -600,7 +922,7 @@ segment_holding_code (mstate m, char* addr) } #endif -#endif /* !(defined(X86_WIN32) || defined(X86_WIN64) || defined(__OS2__)) || defined (__CYGWIN__) || defined(__INTERIX) */ +#endif /* !(defined(_WIN32) || defined(__OS2__)) || defined (__CYGWIN__) || defined(__INTERIX) */ /* Allocate a chunk of memory with the given size. Returns a pointer to the writable address, and sets *CODE to the executable @@ -613,7 +935,7 @@ ffi_closure_alloc (size_t size, void **code) if (!code) return NULL; - ptr = dlmalloc (size); + ptr = FFI_CLOSURE_PTR (dlmalloc (size)); if (ptr) { @@ -625,6 +947,20 @@ ffi_closure_alloc (size_t size, void **code) return ptr; } +void * +ffi_data_to_code_pointer (void *data) +{ + msegmentptr seg = segment_holding (gm, data); + /* We expect closures to be allocated with ffi_closure_alloc(), in + which case seg will be non-NULL. However, some users take on the + burden of managing this memory themselves, in which case this + we'll just return data. */ + if (seg) + return add_segment_exec_offset (data, seg); + else + return data; +} + /* Release a chunk of memory allocated with ffi_closure_alloc. If FFI_CLOSURE_FREE_CODE is nonzero, the given address can be the writable or the executable address given. Otherwise, only the @@ -639,29 +975,9 @@ ffi_closure_free (void *ptr) ptr = sub_segment_exec_offset (ptr, seg); #endif - dlfree (ptr); + dlfree (FFI_RESTORE_PTR (ptr)); } - -#if FFI_CLOSURE_TEST -/* Do some internal sanity testing to make sure allocation and - deallocation of pages are working as intended. */ -int main () -{ - void *p[3]; -#define GET(idx, len) do { p[idx] = dlmalloc (len); printf ("allocated %zi for p[%i]\n", (len), (idx)); } while (0) -#define PUT(idx) do { printf ("freeing p[%i]\n", (idx)); dlfree (p[idx]); } while (0) - GET (0, malloc_getpagesize / 2); - GET (1, 2 * malloc_getpagesize - 64 * sizeof (void*)); - PUT (1); - GET (1, 2 * malloc_getpagesize); - GET (2, malloc_getpagesize / 2); - PUT (1); - PUT (0); - PUT (2); - return 0; -} -#endif /* FFI_CLOSURE_TEST */ # else /* ! FFI_MMAP_EXEC_WRIT */ /* On many systems, memory returned by malloc is writable and @@ -675,14 +991,22 @@ ffi_closure_alloc (size_t size, void **code) if (!code) return NULL; - return *code = malloc (size); + return *code = FFI_CLOSURE_PTR (malloc (size)); } void ffi_closure_free (void *ptr) { - free (ptr); + free (FFI_RESTORE_PTR (ptr)); +} + +void * +ffi_data_to_code_pointer (void *data) +{ + return data; } # endif /* ! FFI_MMAP_EXEC_WRIT */ #endif /* FFI_CLOSURES */ + +#endif /* NetBSD with PROT_MPROTECT */ diff --git a/native/libffi/src/cris/ffi.c b/native/libffi/src/cris/ffi.c index aaca5b1cba..9011fdec03 100644 --- a/native/libffi/src/cris/ffi.c +++ b/native/libffi/src/cris/ffi.c @@ -29,7 +29,7 @@ #include #include -#define STACK_ARG_SIZE(x) ALIGN(x, FFI_SIZEOF_ARG) +#define STACK_ARG_SIZE(x) FFI_ALIGN(x, FFI_SIZEOF_ARG) static ffi_status initialize_aggregate_packed_struct (ffi_type * arg) @@ -190,7 +190,7 @@ ffi_prep_cif_core (ffi_cif * cif, FFI_ASSERT_VALID_TYPE (*ptr); if (((*ptr)->alignment - 1) & bytes) - bytes = ALIGN (bytes, (*ptr)->alignment); + bytes = FFI_ALIGN (bytes, (*ptr)->alignment); if ((*ptr)->type == FFI_TYPE_STRUCT) { if ((*ptr)->size > 8) diff --git a/native/libffi/src/csky/ffi.c b/native/libffi/src/csky/ffi.c new file mode 100644 index 0000000000..af50b7c648 --- /dev/null +++ b/native/libffi/src/csky/ffi.c @@ -0,0 +1,395 @@ +/* ----------------------------------------------------------------------- + ffi.c + + CSKY Foreign Function Interface + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + ``Software''), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice shall be included + in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + ----------------------------------------------------------------------- */ + +#include +#include + +#include + +/* ffi_prep_args is called by the assembly routine once stack space + has been allocated for the function's arguments +*/ +void ffi_prep_args(char *stack, extended_cif *ecif) +{ + register unsigned int i; + register void **p_argv; + register char *argp; + register ffi_type **p_arg; + + argp = stack; + + if ( ecif->cif->flags == FFI_TYPE_STRUCT ) { + *(void **) argp = ecif->rvalue; + argp += 4; + } + + p_argv = ecif->avalue; + + for (i = ecif->cif->nargs, p_arg = ecif->cif->arg_types; + (i != 0); + i--, p_arg++) + { + size_t z; + size_t alignment; + + /* Align if necessary */ + alignment = (*p_arg)->alignment; +#ifdef __CSKYABIV1__ + /* + * Adapt ABIV1 bug. + * If struct's size is larger than 8 bytes, then it always alignment as 4 bytes. + */ + if (((*p_arg)->type == FFI_TYPE_STRUCT) && ((*p_arg)->size > 8) && (alignment == 8)) { + alignment = 4; + } +#endif + + if ((alignment - 1) & (unsigned) argp) { + argp = (char *) FFI_ALIGN(argp, alignment); + } + + if ((*p_arg)->type == FFI_TYPE_STRUCT) + argp = (char *) FFI_ALIGN(argp, 4); + + z = (*p_arg)->size; + if (z < sizeof(int)) + { + z = sizeof(int); + switch ((*p_arg)->type) + { + case FFI_TYPE_SINT8: + *(signed int *) argp = (signed int)*(SINT8 *)(* p_argv); + break; + + case FFI_TYPE_UINT8: + *(unsigned int *) argp = (unsigned int)*(UINT8 *)(* p_argv); + break; + + case FFI_TYPE_SINT16: + *(signed int *) argp = (signed int)*(SINT16 *)(* p_argv); + break; + + case FFI_TYPE_UINT16: + *(unsigned int *) argp = (unsigned int)*(UINT16 *)(* p_argv); + break; + + case FFI_TYPE_STRUCT: +#ifdef __CSKYBE__ + memcpy((argp + 4 - (*p_arg)->size), *p_argv, (*p_arg)->size); +#else + memcpy(argp, *p_argv, (*p_arg)->size); +#endif + break; + + default: + FFI_ASSERT(0); + } + } + else if (z == sizeof(int)) + { + *(unsigned int *) argp = (unsigned int)*(UINT32 *)(* p_argv); + } + else + { + memcpy(argp, *p_argv, z); + } + p_argv++; + argp += z; + } + + return; +} + +/* Perform machine dependent cif processing */ +ffi_status ffi_prep_cif_machdep(ffi_cif *cif) +{ + /* Round the stack up to a multiple of 8 bytes. This isn't needed + everywhere, but it is on some platforms, and it doesn't hcsky anything + when it isn't needed. */ + cif->bytes = (cif->bytes + 7) & ~7; + + /* Set the return type flag */ + switch (cif->rtype->type) + { + + case FFI_TYPE_DOUBLE: + case FFI_TYPE_SINT64: + case FFI_TYPE_UINT64: + cif->flags = (unsigned) FFI_TYPE_SINT64; + break; + + case FFI_TYPE_STRUCT: + if (cif->rtype->size <= 4) + /* A Composite Type not larger than 4 bytes is returned in r0. */ + cif->flags = (unsigned)FFI_TYPE_INT; + else if (cif->rtype->size <= 8) + /* A Composite Type not larger than 8 bytes is returned in r0, r1. */ + cif->flags = (unsigned)FFI_TYPE_SINT64; + else + /* A Composite Type larger than 8 bytes, or whose size cannot + be determined statically ... is stored in memory at an + address passed [in r0]. */ + cif->flags = (unsigned)FFI_TYPE_STRUCT; + break; + + default: + cif->flags = FFI_TYPE_INT; + break; + } + + return FFI_OK; +} + +/* Perform machine dependent cif processing for variadic calls */ +ffi_status ffi_prep_cif_machdep_var(ffi_cif *cif, + unsigned int nfixedargs, + unsigned int ntotalargs) +{ + return ffi_prep_cif_machdep(cif); +} + +/* Prototypes for assembly functions, in sysv.S */ +extern void ffi_call_SYSV (void (*fn)(void), extended_cif *, unsigned, unsigned, unsigned *); + +void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +{ + extended_cif ecif; + + int small_struct = (cif->flags == FFI_TYPE_INT + && cif->rtype->type == FFI_TYPE_STRUCT); + + ecif.cif = cif; + ecif.avalue = avalue; + + unsigned int temp; + + /* If the return value is a struct and we don't have a return */ + /* value address then we need to make one */ + + if ((rvalue == NULL) && + (cif->flags == FFI_TYPE_STRUCT)) + { + ecif.rvalue = alloca(cif->rtype->size); + } + else if (small_struct) + ecif.rvalue = &temp; + else + ecif.rvalue = rvalue; + + switch (cif->abi) + { + case FFI_SYSV: + ffi_call_SYSV (fn, &ecif, cif->bytes, cif->flags, ecif.rvalue); + break; + + default: + FFI_ASSERT(0); + break; + } + if (small_struct) +#ifdef __CSKYBE__ + memcpy (rvalue, ((unsigned char *)&temp + (4 - cif->rtype->size)), cif->rtype->size); +#else + memcpy (rvalue, &temp, cif->rtype->size); +#endif +} + +/** private members **/ + +static void ffi_prep_incoming_args_SYSV (char *stack, void **ret, + void** args, ffi_cif* cif); + +void ffi_closure_SYSV (ffi_closure *); + +/* This function is jumped to by the trampoline */ + +unsigned int +ffi_closure_SYSV_inner (closure, respp, args) + ffi_closure *closure; + void **respp; + void *args; +{ + // our various things... + ffi_cif *cif; + void **arg_area; + + cif = closure->cif; + arg_area = (void**) alloca (cif->nargs * sizeof (void*)); + + /* this call will initialize ARG_AREA, such that each + * element in that array points to the corresponding + * value on the stack; and if the function returns + * a structure, it will re-set RESP to point to the + * structure return address. */ + + ffi_prep_incoming_args_SYSV(args, respp, arg_area, cif); + + (closure->fun) (cif, *respp, arg_area, closure->user_data); + +#ifdef __CSKYBE__ + if (cif->flags == FFI_TYPE_INT && cif->rtype->type == FFI_TYPE_STRUCT) { + unsigned int tmp = 0; + tmp = *(unsigned int *)(*respp); + *(unsigned int *)(*respp) = (tmp >> ((4 - cif->rtype->size) * 8)); + } +#endif + + return cif->flags; +} + + +static void +ffi_prep_incoming_args_SYSV(char *stack, void **rvalue, + void **avalue, ffi_cif *cif) +{ + register unsigned int i; + register void **p_argv; + register char *argp; + register ffi_type **p_arg; + + argp = stack; + + if ( cif->flags == FFI_TYPE_STRUCT ) { + *rvalue = *(void **) argp; + argp += 4; + } + + p_argv = avalue; + + for (i = cif->nargs, p_arg = cif->arg_types; (i != 0); i--, p_arg++) + { + size_t z; + size_t alignment; + + alignment = (*p_arg)->alignment; + if (alignment < 4) + alignment = 4; + +#ifdef __CSKYABIV1__ + /* + * Adapt ABIV1 bug. + * If struct's size is larger than 8 bytes, then it always alignment as 4 bytes. + */ + if (((*p_arg)->type == FFI_TYPE_STRUCT) && ((*p_arg)->size > 8) && (alignment == 8)) { + alignment = 4; + } +#endif + + /* Align if necessary */ + if ((alignment - 1) & (unsigned) argp) { + argp = (char *) FFI_ALIGN(argp, alignment); + } + + z = (*p_arg)->size; + +#ifdef __CSKYBE__ + unsigned int tmp = 0; + if ((*p_arg)->size < 4) { + tmp = *(unsigned int *)argp; + memcpy(argp, ((unsigned char *)&tmp + (4 - (*p_arg)->size)), (*p_arg)->size); + } +#else + /* because we're little endian, this is what it turns into. */ +#endif + *p_argv = (void*) argp; + + p_argv++; + argp += z; + } + + return; +} + +/* How to make a trampoline. */ + +extern unsigned char ffi_csky_trampoline[TRAMPOLINE_SIZE]; + +/* + * Since there is no __clear_cache in libgcc in csky toolchain. + * define ffi_csky_cacheflush in sysv.S. + * void ffi_csky_cacheflush(uint32 start_addr, uint32 size, int cache) + */ +#define CACHEFLUSH_IN_FFI 1 +#if CACHEFLUSH_IN_FFI +extern void ffi_csky_cacheflush(unsigned char *__tramp, unsigned int k, + int i); +#define FFI_INIT_TRAMPOLINE(TRAMP,FUN,CTX) \ +({ unsigned char *__tramp = (unsigned char*)(TRAMP); \ + unsigned int __fun = (unsigned int)(FUN); \ + unsigned int __ctx = (unsigned int)(CTX); \ + unsigned char *insns = (unsigned char *)(CTX); \ + memcpy (__tramp, ffi_csky_trampoline, TRAMPOLINE_SIZE); \ + *(unsigned int*) &__tramp[TRAMPOLINE_SIZE] = __ctx; \ + *(unsigned int*) &__tramp[TRAMPOLINE_SIZE + 4] = __fun; \ + ffi_csky_cacheflush(&__tramp[0], TRAMPOLINE_SIZE, 3); /* Clear data mapping. */ \ + ffi_csky_cacheflush(insns, TRAMPOLINE_SIZE, 3); \ + /* Clear instruction \ + mapping. */ \ + }) +#else +#define FFI_INIT_TRAMPOLINE(TRAMP,FUN,CTX) \ +({ unsigned char *__tramp = (unsigned char*)(TRAMP); \ + unsigned int __fun = (unsigned int)(FUN); \ + unsigned int __ctx = (unsigned int)(CTX); \ + unsigned char *insns = (unsigned char *)(CTX); \ + memcpy (__tramp, ffi_csky_trampoline, TRAMPOLINE_SIZE); \ + *(unsigned int*) &__tramp[TRAMPOLINE_SIZE] = __ctx; \ + *(unsigned int*) &__tramp[TRAMPOLINE_SIZE + 4] = __fun; \ + __clear_cache((&__tramp[0]), (&__tramp[TRAMPOLINE_SIZE-1])); /* Clear data mapping. */ \ + __clear_cache(insns, insns + TRAMPOLINE_SIZE); \ + /* Clear instruction \ + mapping. */ \ + }) +#endif + +/* the cif must already be prep'ed */ + +ffi_status +ffi_prep_closure_loc (ffi_closure* closure, + ffi_cif* cif, + void (*fun)(ffi_cif*,void*,void**,void*), + void *user_data, + void *codeloc) +{ + void (*closure_func)(ffi_closure*) = NULL; + + if (cif->abi == FFI_SYSV) + closure_func = &ffi_closure_SYSV; + else + return FFI_BAD_ABI; + + FFI_INIT_TRAMPOLINE (&closure->tramp[0], \ + closure_func, \ + codeloc); + + closure->cif = cif; + closure->user_data = user_data; + closure->fun = fun; + + return FFI_OK; +} + + diff --git a/native/libffi/src/csky/ffitarget.h b/native/libffi/src/csky/ffitarget.h new file mode 100644 index 0000000000..f770aac66f --- /dev/null +++ b/native/libffi/src/csky/ffitarget.h @@ -0,0 +1,63 @@ +/* -----------------------------------------------------------------*-C-*- + ffitarget.h - Copyright (c) 2012 Anthony Green + Copyright (c) 2010 CodeSourcery + Copyright (c) 1996-2003 Red Hat, Inc. + + Target configuration macros for CSKY. + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + ``Software''), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice shall be included + in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + + ----------------------------------------------------------------------- */ + +#ifndef LIBFFI_TARGET_H +#define LIBFFI_TARGET_H + +#ifndef LIBFFI_H +#error "Please do not include ffitarget.h directly into your source. Use ffi.h instead." +#endif + +#ifndef LIBFFI_ASM +typedef unsigned long ffi_arg; +typedef signed long ffi_sarg; + +typedef enum ffi_abi { + FFI_FIRST_ABI = 0, + FFI_SYSV, + FFI_LAST_ABI, + FFI_DEFAULT_ABI = FFI_SYSV, +} ffi_abi; +#endif + +#ifdef __CSKYABIV2__ +#define FFI_ASM_ARGREG_SIZE 16 +#define TRAMPOLINE_SIZE 16 +#define FFI_TRAMPOLINE_SIZE 24 +#else +#define FFI_ASM_ARGREG_SIZE 24 +#define TRAMPOLINE_SIZE 20 +#define FFI_TRAMPOLINE_SIZE 28 +#endif + +/* ---- Definitions for closures ----------------------------------------- */ + +#define FFI_CLOSURES 1 +#define FFI_NATIVE_RAW_API 0 +#endif diff --git a/native/libffi/src/csky/sysv.S b/native/libffi/src/csky/sysv.S new file mode 100644 index 0000000000..21670bfd3d --- /dev/null +++ b/native/libffi/src/csky/sysv.S @@ -0,0 +1,371 @@ +/* ----------------------------------------------------------------------- + sysv.S + + CSKY Foreign Function Interface + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + ``Software''), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice shall be included + in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + ----------------------------------------------------------------------- */ + +#define LIBFFI_ASM +#include +#include + +.macro CSKY_FUNC_START name + .text + .align 2 + .globl \name + .type \name, @function + \name: +.endm + +#ifdef __CSKYABIV2__ + + /* + * a0: fn + * a1: &ecif + * a2: cif->bytes + * a3: fig->flags + * sp+0: ecif.rvalue + */ +CSKY_FUNC_START ffi_call_SYSV + /* Save registers */ + .cfi_startproc + subi sp, 28 + .cfi_def_cfa_offset 28 + stw a0, (sp, 0x0) + .cfi_offset 0, -28 + stw a1, (sp, 0x4) + .cfi_offset 1, -24 + stw a2, (sp, 0x8) + .cfi_offset 2, -20 + stw a3, (sp, 0xC) + .cfi_offset 3, -16 + stw l0, (sp, 0x10) + .cfi_offset 4, -12 + stw l1, (sp, 0x14) + .cfi_offset 5, -8 + stw lr, (sp, 0x18) + .cfi_offset 15, -4 + + mov l0, sp + .cfi_def_cfa_register 4 + + /* Make room for all of the new args. */ + subu sp, sp, a2 + + /* Place all of the ffi_prep_args in position */ + mov a0, sp + /* a1 already set */ + + /* Call ffi_prep_args(stack, &ecif) */ + jsri ffi_prep_args + + /* move first 4 parameters in registers */ + ldw a0, (sp, 0x0) + ldw a1, (sp, 0x4) + ldw a2, (sp, 0x8) + ldw a3, (sp, 0xC) + + /* and adjust stack */ + subu lr, l0, sp /* cif->bytes == l0 - sp */ + cmphsi lr, 16 + movi l1, 16 + movt lr, l1 + addu sp, sp, lr + + ldw l1, (l0, 0) /* load fn() in advance */ + + /* call (fn) (...) */ + jsr l1 + + /* Remove the space we pushed for the args */ + mov sp, l0 + + /* Load r2 with the pointer to storage for the return value */ + ldw a2, (sp, 0x1C) + + /* Load r3 with the return type code */ + ldw a3, (sp, 0xC) + + /* If the return value pointer is NULL, assume no return value. */ + cmpnei a2, 0 + bf .Lepilogue + + cmpnei a3, FFI_TYPE_STRUCT + bf .Lepilogue + + /* return INT64 */ + cmpnei a3, FFI_TYPE_SINT64 + bt .Lretint + /* stw a0, (a2, 0x0) at .Lretint */ + stw a1, (a2, 0x4) + +.Lretint: + /* return INT */ + stw a0, (a2, 0x0) + +.Lepilogue: + ldw a0, (sp, 0x0) + ldw a1, (sp, 0x4) + ldw a2, (sp, 0x8) + ldw a3, (sp, 0xC) + ldw l0, (sp, 0x10) + ldw l1, (sp, 0x14) + ldw lr, (sp, 0x18) + addi sp, sp, 28 + rts + .cfi_endproc + .size ffi_call_SYSV, .-ffi_call_SYSV + + + /* + * unsigned int FFI_HIDDEN + * ffi_closure_SYSV_inner (closure, respp, args) + * ffi_closure *closure; + * void **respp; + * void *args; + */ +CSKY_FUNC_START ffi_closure_SYSV + .cfi_startproc + mov a2, sp + addi a1, sp, 16 + subi sp, sp, 24 + .cfi_def_cfa_offset 40 + stw a1, (sp, 0x10) + .cfi_offset 1, -24 + stw lr, (sp, 0x14) + .cfi_offset 15, -20 + stw sp, (sp, 0x8) + addi a1, sp, 8 + jsri ffi_closure_SYSV_inner + ldw a0, (sp, 0x0) + /* + * if FFI_TYPE_SINT64, need a1. + * if FFI_TYPE_INT, ignore a1. + */ + ldw a1, (sp, 0x4) + + ldw lr, (sp, 0x14) + addi sp, sp, 40 + rts + .cfi_endproc + .size ffi_closure_SYSV, .-ffi_closure_SYSV + +CSKY_FUNC_START ffi_csky_trampoline + subi sp, sp, 16 + stw a0, (sp, 0x0) + stw a1, (sp, 0x4) + stw a2, (sp, 0x8) + stw a3, (sp, 0xC) + lrw a0, [.Lctx] + lrw a1, [.Lfun] + jmp a1 +.Lctx: + mov a0, a0 + mov a0, a0 +.Lfun: + + .size ffi_csky_trampoline, .-ffi_csky_trampoline + +CSKY_FUNC_START ffi_csky_cacheflush + mov t0, r7 + movi r7, 123 + trap 0 + mov r7, t0 + rts + + .size ffi_csky_cacheflush, .-ffi_csky_cacheflush + +#else /* !__CSKYABIV2__ */ + + /* + * a0: fn + * a1: &ecif + * a2: cif->bytes + * a3: fig->flags + * a4: ecif.rvalue + */ +CSKY_FUNC_START ffi_call_SYSV + /* Save registers */ + .cfi_startproc + subi sp, 32 + subi sp, 8 + .cfi_def_cfa_offset 40 + stw a0, (sp, 0x0) + .cfi_offset 2, -40 + stw a1, (sp, 0x4) + .cfi_offset 3, -36 + stw a2, (sp, 0x8) + .cfi_offset 4, -32 + stw a3, (sp, 0xC) + .cfi_offset 5, -28 + stw a4, (sp, 0x10) + .cfi_offset 6, -24 + stw a5, (sp, 0x14) + .cfi_offset 7, -20 + stw l0, (sp, 0x18) + .cfi_offset 8, -16 + stw l1, (sp, 0x1C) + .cfi_offset 9, -12 + stw lr, (sp, 0x20) + .cfi_offset 15, -8 + + mov l0, sp + .cfi_def_cfa_register 8 + + /* Make room for all of the new args. */ + subu sp, sp, a2 + + /* Place all of the ffi_prep_args in position */ + mov a0, sp + /* a1 already set */ + + /* Call ffi_prep_args(stack, &ecif) */ + jsri ffi_prep_args + + /* move first 4 parameters in registers */ + ldw a0, (sp, 0x0) + ldw a1, (sp, 0x4) + ldw a2, (sp, 0x8) + ldw a3, (sp, 0xC) + ldw a4, (sp, 0x10) + ldw a5, (sp, 0x14) + + /* and adjust stack */ + mov lr, l0 + subu lr, sp /* cif->bytes == l0 - sp */ + movi l1, 24 + cmphs lr, l1 + movt lr, l1 + addu sp, sp, lr + + ldw l1, (l0, 0) /* load fn() in advance */ + + /* call (fn) (...) */ + jsr l1 + + /* Remove the space we pushed for the args */ + mov sp, l0 + + /* Load r2 with the pointer to storage for the return value */ + ldw a2, (sp, 0x10) + + /* Load r3 with the return type code */ + ldw a3, (sp, 0xC) + + /* If the return value pointer is NULL, assume no return value. */ + cmpnei a2, 0 + bf .Lepilogue + + cmpnei a3, FFI_TYPE_STRUCT + bf .Lepilogue + + /* return INT64 */ + cmpnei a3, FFI_TYPE_SINT64 + bt .Lretint + /* stw a0, (a2, 0x0) at .Lretint */ + stw a1, (a2, 0x4) + +.Lretint: + /* return INT */ + stw a0, (a2, 0x0) + +.Lepilogue: + ldw a0, (sp, 0x0) + ldw a1, (sp, 0x4) + ldw a2, (sp, 0x8) + ldw a3, (sp, 0xC) + ldw a4, (sp, 0x10) + ldw a5, (sp, 0x14) + ldw l0, (sp, 0x18) + ldw l1, (sp, 0x1C) + ldw lr, (sp, 0x20) + addi sp, sp, 32 + addi sp, sp, 8 + rts + .cfi_endproc + + .size ffi_call_SYSV, .-ffi_call_SYSV + + + /* + * unsigned int FFI_HIDDEN + * ffi_closure_SYSV_inner (closure, respp, args) + * ffi_closure *closure; + * void **respp; + * void *args; + */ +CSKY_FUNC_START ffi_closure_SYSV + .cfi_startproc + mov a2, sp + mov a1, sp + addi a1, 24 + subi sp, sp, 24 + .cfi_def_cfa_offset 48 + stw a1, (sp, 0x10) + .cfi_offset 3, -32 + stw lr, (sp, 0x14) + .cfi_offset 15, -28 + stw sp, (sp, 0x8) + mov a1, sp + addi a1, 8 + jsri ffi_closure_SYSV_inner + ldw a0, (sp, 0x0) + /* + * if FFI_TYPE_SINT64, need a1. + * if FFI_TYPE_INT, ignore a1. + */ + ldw a1, (sp, 0x4) + + ldw lr, (sp, 0x14) + addi sp, sp, 24 + addi sp, sp, 24 + rts + .cfi_endproc + + .size ffi_closure_SYSV, .-ffi_closure_SYSV + +CSKY_FUNC_START ffi_csky_trampoline + subi sp, 24 + stw a0, (sp, 0x0) + stw a1, (sp, 0x4) + stw a2, (sp, 0x8) + stw a3, (sp, 0xC) + stw a4, (sp, 0x10) + stw a5, (sp, 0x14) + lrw a0, [.Lctx] + lrw a1, [.Lfun] + jmp a1 +.Lctx: + mov a0, a0 + mov a0, a0 +.Lfun: + + .size ffi_csky_trampoline, .-ffi_csky_trampoline + +CSKY_FUNC_START ffi_csky_cacheflush + lrw r1, 123 + trap 0 + rts + + .size ffi_csky_cacheflush, .-ffi_csky_cacheflush + +#endif /* __CSKYABIV2__ */ diff --git a/native/libffi/src/dlmalloc.c b/native/libffi/src/dlmalloc.c index 3f8f1ee7ff..31d03d3be7 100644 --- a/native/libffi/src/dlmalloc.c +++ b/native/libffi/src/dlmalloc.c @@ -438,6 +438,11 @@ DEFAULT_MMAP_THRESHOLD default: 256K */ +#if defined __linux__ && !defined _GNU_SOURCE +/* mremap() on Linux requires this via sys/mman.h */ +#define _GNU_SOURCE 1 +#endif + #ifndef WIN32 #ifdef _WIN32 #define WIN32 1 @@ -2291,7 +2296,7 @@ static size_t traverse_and_check(mstate m); #define treebin_at(M,i) (&((M)->treebins[i])) /* assign tree index for size S to variable I */ -#if defined(__GNUC__) && defined(i386) +#if defined(__GNUC__) && defined(__i386__) #define compute_tree_index(S, I)\ {\ size_t X = S >> TREEBIN_SHIFT;\ @@ -2356,7 +2361,7 @@ static size_t traverse_and_check(mstate m); /* index corresponding to given bit */ -#if defined(__GNUC__) && defined(i386) +#if defined(__GNUC__) && defined(__i386__) #define compute_bit2idx(X, I)\ {\ unsigned int J;\ @@ -2366,7 +2371,7 @@ static size_t traverse_and_check(mstate m); #else /* GNUC */ #if USE_BUILTIN_FFS -#define compute_bit2idx(X, I) I = ffs(X)-1 +#define compute_bit2idx(X, I) I = __builtin_ffs(X)-1 #else /* USE_BUILTIN_FFS */ #define compute_bit2idx(X, I)\ diff --git a/native/libffi/src/frv/ffi.c b/native/libffi/src/frv/ffi.c index 5698c89c35..ed1c65a1a3 100644 --- a/native/libffi/src/frv/ffi.c +++ b/native/libffi/src/frv/ffi.c @@ -107,7 +107,7 @@ void *ffi_prep_args(char *stack, extended_cif *ecif) count += z; } - return (stack + ((count > 24) ? 24 : ALIGN_DOWN(count, 8))); + return (stack + ((count > 24) ? 24 : FFI_ALIGN_DOWN(count, 8))); } /* Perform machine dependent cif processing */ @@ -118,7 +118,7 @@ ffi_status ffi_prep_cif_machdep(ffi_cif *cif) else cif->flags = cif->rtype->size; - cif->bytes = ALIGN (cif->bytes, 8); + cif->bytes = FFI_ALIGN (cif->bytes, 8); return FFI_OK; } diff --git a/native/libffi/src/ia64/ffi.c b/native/libffi/src/ia64/ffi.c index b77a836ddc..b1d04c3e74 100644 --- a/native/libffi/src/ia64/ffi.c +++ b/native/libffi/src/ia64/ffi.c @@ -220,8 +220,8 @@ hfa_element_type (ffi_type *type, int nested) /* Perform machine dependent cif processing. */ -ffi_status -ffi_prep_cif_machdep(ffi_cif *cif) +static ffi_status +ffi_prep_cif_machdep_core(ffi_cif *cif) { int flags; @@ -271,6 +271,22 @@ ffi_prep_cif_machdep(ffi_cif *cif) return FFI_OK; } +ffi_status +ffi_prep_cif_machdep(ffi_cif *cif) +{ + cif->nfixedargs = cif->nargs; + return ffi_prep_cif_machdep_core(cif); +} + +ffi_status +ffi_prep_cif_machdep_var(ffi_cif *cif, + unsigned int nfixedargs, + unsigned int ntotalargs MAYBE_UNUSED) +{ + cif->nfixedargs = nfixedargs; + return ffi_prep_cif_machdep_core(cif); +} + extern int ffi_call_unix (struct ia64_args *, PTR64, void (*)(void), UINT64); void @@ -454,10 +470,11 @@ ffi_closure_unix_inner (ffi_closure *closure, struct ia64_args *stack, ffi_cif *cif; void **avalue; ffi_type **p_arg; - long i, avn, gpcount, fpcount; + long i, avn, gpcount, fpcount, nfixedargs; cif = closure->cif; avn = cif->nargs; + nfixedargs = cif->nfixedargs; avalue = alloca (avn * sizeof (void *)); /* If the structure return value is passed in memory get that location @@ -468,6 +485,7 @@ ffi_closure_unix_inner (ffi_closure *closure, struct ia64_args *stack, gpcount = fpcount = 0; for (i = 0, p_arg = cif->arg_types; i < avn; i++, p_arg++) { + int named = i < nfixedargs; switch ((*p_arg)->type) { case FFI_TYPE_SINT8: @@ -491,7 +509,7 @@ ffi_closure_unix_inner (ffi_closure *closure, struct ia64_args *stack, break; case FFI_TYPE_FLOAT: - if (gpcount < 8 && fpcount < 8) + if (named && gpcount < 8 && fpcount < 8) { fpreg *addr = &stack->fp_regs[fpcount++]; float result; @@ -505,7 +523,7 @@ ffi_closure_unix_inner (ffi_closure *closure, struct ia64_args *stack, break; case FFI_TYPE_DOUBLE: - if (gpcount < 8 && fpcount < 8) + if (named && gpcount < 8 && fpcount < 8) { fpreg *addr = &stack->fp_regs[fpcount++]; double result; @@ -521,7 +539,7 @@ ffi_closure_unix_inner (ffi_closure *closure, struct ia64_args *stack, case FFI_TYPE_LONGDOUBLE: if (gpcount & 1) gpcount++; - if (LDBL_MANT_DIG == 64 && gpcount < 8 && fpcount < 8) + if (LDBL_MANT_DIG == 64 && named && gpcount < 8 && fpcount < 8) { fpreg *addr = &stack->fp_regs[fpcount++]; __float80 result; diff --git a/native/libffi/src/ia64/ffitarget.h b/native/libffi/src/ia64/ffitarget.h index e68cea6154..fd5b9a0ee1 100644 --- a/native/libffi/src/ia64/ffitarget.h +++ b/native/libffi/src/ia64/ffitarget.h @@ -50,6 +50,7 @@ typedef enum ffi_abi { #define FFI_TRAMPOLINE_SIZE 24 /* Really the following struct, which */ /* can be interpreted as a C function */ /* descriptor: */ +#define FFI_TARGET_SPECIFIC_VARIADIC 1 +#define FFI_EXTRA_CIF_FIELDS unsigned nfixedargs #endif - diff --git a/native/libffi/src/ia64/unix.S b/native/libffi/src/ia64/unix.S index 4d2a86d421..e2547e02b3 100644 --- a/native/libffi/src/ia64/unix.S +++ b/native/libffi/src/ia64/unix.S @@ -175,7 +175,6 @@ ffi_call_unix: ;; .Lst_small_struct: - add sp = -16, sp cmp.lt p6, p0 = 8, in3 cmp.lt p7, p0 = 16, in3 cmp.lt p8, p0 = 24, in3 @@ -191,6 +190,12 @@ ffi_call_unix: (p8) st8 [r18] = r11 mov out1 = sp mov out2 = in3 + ;; + // ia64 software calling convention requires + // top 16 bytes of stack to be scratch space + // PLT resolver uses that scratch space at + // 'memcpy' symbol reolution time + add sp = -16, sp br.call.sptk.many b0 = memcpy# ;; mov ar.pfs = loc0 @@ -529,6 +534,7 @@ ffi_closure_unix: data8 @pcrel(.Lst_int64) // FFI_TYPE_SINT64 data8 @pcrel(.Lst_void) // FFI_TYPE_STRUCT data8 @pcrel(.Lst_int64) // FFI_TYPE_POINTER + data8 @pcrel(.Lst_void) // FFI_TYPE_COMPLEX (not implemented) data8 @pcrel(.Lst_small_struct) // FFI_IA64_TYPE_SMALL_STRUCT data8 @pcrel(.Lst_hfa_float) // FFI_IA64_TYPE_HFA_FLOAT data8 @pcrel(.Lst_hfa_double) // FFI_IA64_TYPE_HFA_DOUBLE @@ -550,6 +556,7 @@ ffi_closure_unix: data8 @pcrel(.Lld_int) // FFI_TYPE_SINT64 data8 @pcrel(.Lld_void) // FFI_TYPE_STRUCT data8 @pcrel(.Lld_int) // FFI_TYPE_POINTER + data8 @pcrel(.Lld_void) // FFI_TYPE_COMPLEX (not implemented) data8 @pcrel(.Lld_small_struct) // FFI_IA64_TYPE_SMALL_STRUCT data8 @pcrel(.Lld_hfa_float) // FFI_IA64_TYPE_HFA_FLOAT data8 @pcrel(.Lld_hfa_double) // FFI_IA64_TYPE_HFA_DOUBLE diff --git a/native/libffi/src/java_raw_api.c b/native/libffi/src/java_raw_api.c index 127123d5bc..114d3e47fc 100644 --- a/native/libffi/src/java_raw_api.c +++ b/native/libffi/src/java_raw_api.c @@ -114,7 +114,7 @@ ffi_java_raw_to_ptrarray (ffi_cif *cif, ffi_java_raw *raw, void **args) default: *args = raw; raw += - ALIGN ((*tp)->size, sizeof(ffi_java_raw)) / sizeof(ffi_java_raw); + FFI_ALIGN ((*tp)->size, sizeof(ffi_java_raw)) / sizeof(ffi_java_raw); } } @@ -142,7 +142,7 @@ ffi_java_raw_to_ptrarray (ffi_cif *cif, ffi_java_raw *raw, void **args) #else /* FFI_SIZEOF_JAVA_RAW != 8 */ *args = (void*) raw; raw += - ALIGN ((*tp)->size, sizeof(ffi_java_raw)) / sizeof(ffi_java_raw); + FFI_ALIGN ((*tp)->size, sizeof(ffi_java_raw)) / sizeof(ffi_java_raw); #endif /* FFI_SIZEOF_JAVA_RAW == 8 */ } @@ -234,7 +234,7 @@ ffi_java_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_java_raw *raw) #else memcpy ((void*) raw->data, (void*)*args, (*tp)->size); raw += - ALIGN ((*tp)->size, sizeof(ffi_java_raw)) / sizeof(ffi_java_raw); + FFI_ALIGN ((*tp)->size, sizeof(ffi_java_raw)) / sizeof(ffi_java_raw); #endif } } diff --git a/native/libffi/src/kvx/asm.h b/native/libffi/src/kvx/asm.h new file mode 100644 index 0000000000..4edba410d8 --- /dev/null +++ b/native/libffi/src/kvx/asm.h @@ -0,0 +1,5 @@ +/* args are passed on registers from r0 up to r11 => 12*8 bytes */ +#define REG_ARGS_SIZE (12*8) +#define KVX_REGISTER_SIZE (8) +#define KVX_ABI_SLOT_SIZE (KVX_REGISTER_SIZE) +#define KVX_ABI_MAX_AGGREGATE_IN_REG_SIZE (4*KVX_ABI_SLOT_SIZE) diff --git a/native/libffi/src/kvx/ffi.c b/native/libffi/src/kvx/ffi.c new file mode 100644 index 0000000000..58f6aef5ca --- /dev/null +++ b/native/libffi/src/kvx/ffi.c @@ -0,0 +1,273 @@ +/* Copyright (c) 2020 Kalray + +Permission is hereby granted, free of charge, to any person obtaining +a copy of this software and associated documentation files (the +``Software''), to deal in the Software without restriction, including +without limitation the rights to use, copy, modify, merge, publish, +distribute, sublicense, and/or sell copies of the Software, and to +permit persons to whom the Software is furnished to do so, subject to +the following conditions: + +The above copyright notice and this permission notice shall be +included in all copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, +EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF +MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. +IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY +CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, +TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE +SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ + +#if defined(__kvx__) +#include +#include +#include +#include +#include +#include "ffi_common.h" +#include "asm.h" + +#define ALIGN(x, a) ALIGN_MASK(x, (typeof(x))(a) - 1) +#define ALIGN_MASK(x, mask) (((x) + (mask)) & ~(mask)) +#define KVX_ABI_STACK_ALIGNMENT (32) +#define KVX_ABI_STACK_ARG_ALIGNMENT (8) +#define max(a,b) ((a) > (b) ? (a) : (b)) + +#ifdef FFI_DEBUG +#define DEBUG_PRINT(...) do{ fprintf( stderr, __VA_ARGS__ ); } while(0) +#else +#define DEBUG_PRINT(...) +#endif + +struct ret_value { + unsigned long int r0; + unsigned long int r1; + unsigned long int r2; + unsigned long int r3; +}; + +extern struct ret_value ffi_call_SYSV(unsigned total_size, + unsigned size, + extended_cif *ecif, + unsigned *rvalue_addr, + void *fn, + unsigned int_ext_method); + +/* Perform machine dependent cif processing */ +ffi_status ffi_prep_cif_machdep(ffi_cif *cif) +{ + cif->flags = cif->rtype->size; + return FFI_OK; +} + +/* ffi_prep_args is called by the assembly routine once stack space + has been allocated for the function's arguments */ + +void *ffi_prep_args(char *stack, unsigned int arg_slots_size, extended_cif *ecif) +{ + char *stacktemp = stack; + char *current_arg_passed_by_value = stack + arg_slots_size; + int i, s; + ffi_type **arg; + int count = 0; + ffi_cif *cif = ecif->cif; + void **argv = ecif->avalue; + + arg = cif->arg_types; + + DEBUG_PRINT("stack: %p\n", stack); + DEBUG_PRINT("arg_slots_size: %u\n", arg_slots_size); + DEBUG_PRINT("current_arg_passed_by_value: %p\n", current_arg_passed_by_value); + DEBUG_PRINT("ecif: %p\n", ecif); + DEBUG_PRINT("ecif->avalue: %p\n", ecif->avalue); + + for (i = 0; i < cif->nargs; i++) { + + s = KVX_ABI_SLOT_SIZE; + switch((*arg)->type) { + case FFI_TYPE_SINT8: + case FFI_TYPE_UINT8: + case FFI_TYPE_SINT16: + case FFI_TYPE_UINT16: + case FFI_TYPE_SINT32: + case FFI_TYPE_UINT32: + case FFI_TYPE_FLOAT: + case FFI_TYPE_DOUBLE: + case FFI_TYPE_UINT64: + case FFI_TYPE_SINT64: + case FFI_TYPE_POINTER: + DEBUG_PRINT("INT64/32/16/8/FLOAT/DOUBLE or POINTER @%p\n", stack); + *(uint64_t *) stack = *(uint64_t *)(* argv); + break; + + case FFI_TYPE_COMPLEX: + if ((*arg)->size == 8) + *(_Complex float *) stack = *(_Complex float *)(* argv); + else if ((*arg)->size == 16) { + *(_Complex double *) stack = *(_Complex double *)(* argv); + s = 16; + } else + abort(); + break; + case FFI_TYPE_STRUCT: { + char *value; + unsigned int written_size = 0; + DEBUG_PRINT("struct by value @%p\n", stack); + if ((*arg)->size > KVX_ABI_MAX_AGGREGATE_IN_REG_SIZE) { + DEBUG_PRINT("big struct\n"); + *(uint64_t *) stack = (uintptr_t)current_arg_passed_by_value; + value = current_arg_passed_by_value; + current_arg_passed_by_value += (*arg)->size; + written_size = KVX_ABI_SLOT_SIZE; + } else { + value = stack; + written_size = (*arg)->size; + } + memcpy(value, *argv, (*arg)->size); + s = ALIGN(written_size, KVX_ABI_STACK_ARG_ALIGNMENT); + break; + } + default: + printf("Error: unsupported arg type %d\n", (*arg)->type); + abort(); + break; + + } + stack += s; + count += s; + argv++; + arg++; + } +#ifdef FFI_DEBUG + FFI_ASSERT(((intptr_t)(stacktemp + REG_ARGS_SIZE) & (KVX_ABI_STACK_ALIGNMENT-1)) == 0); +#endif + return stacktemp + REG_ARGS_SIZE; +} + +/* Perform machine dependent cif processing when we have a variadic function */ + +ffi_status ffi_prep_cif_machdep_var(ffi_cif *cif, unsigned int nfixedargs, + unsigned int ntotalargs) +{ + cif->flags = cif->rtype->size; + return FFI_OK; +} + +static unsigned long handle_small_int_ext(kvx_intext_method *int_ext_method, + const ffi_type *rtype) +{ + switch (rtype->type) { + case FFI_TYPE_SINT8: + *int_ext_method = KVX_RET_SXBD; + return KVX_REGISTER_SIZE; + + case FFI_TYPE_SINT16: + *int_ext_method = KVX_RET_SXHD; + return KVX_REGISTER_SIZE; + + case FFI_TYPE_SINT32: + *int_ext_method = KVX_RET_SXWD; + return KVX_REGISTER_SIZE; + + case FFI_TYPE_UINT8: + *int_ext_method = KVX_RET_ZXBD; + return KVX_REGISTER_SIZE; + + case FFI_TYPE_UINT16: + *int_ext_method = KVX_RET_ZXHD; + return KVX_REGISTER_SIZE; + + case FFI_TYPE_UINT32: + *int_ext_method = KVX_RET_ZXWD; + return KVX_REGISTER_SIZE; + + default: + *int_ext_method = KVX_RET_NONE; + return rtype->size; + } +} + +void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +{ + int i; + unsigned long int slot_fitting_args_size = 0; + unsigned long int total_size = 0; + unsigned long int big_struct_size = 0; + kvx_intext_method int_extension_method; + ffi_type **arg; + struct ret_value local_rvalue = {0}; + size_t wb_size; + + + /* Calculate size to allocate on stack */ + for (i = 0, arg = cif->arg_types; i < cif->nargs; i++, arg++) { + DEBUG_PRINT("argument %d, type %d, size %lu\n", i, (*arg)->type, (*arg)->size); + if (((*arg)->type == FFI_TYPE_STRUCT) || ((*arg)->type == FFI_TYPE_COMPLEX)) { + if ((*arg)->size <= KVX_ABI_MAX_AGGREGATE_IN_REG_SIZE) { + slot_fitting_args_size += ALIGN((*arg)->size, KVX_ABI_SLOT_SIZE); + } else { + slot_fitting_args_size += KVX_ABI_SLOT_SIZE; /* aggregate passed by reference */ + big_struct_size += ALIGN((*arg)->size, KVX_ABI_SLOT_SIZE); + } + } else if ((*arg)->size <= KVX_ABI_SLOT_SIZE) { + slot_fitting_args_size += KVX_ABI_SLOT_SIZE; + } else { + printf("Error: unsupported arg size %ld arg type %d\n", (*arg)->size, (*arg)->type); + abort(); /* should never happen? */ + } + } + + extended_cif ecif; + ecif.cif = cif; + ecif.avalue = avalue; + ecif.rvalue = rvalue; + + /* This implementation allocates anyway for all register based args */ + slot_fitting_args_size = max(slot_fitting_args_size, REG_ARGS_SIZE); + total_size = slot_fitting_args_size + big_struct_size; + total_size = ALIGN(total_size, KVX_ABI_STACK_ALIGNMENT); + + /* wb_size: write back size, the size we will need to write back to user + * provided buffer. In theory it should always be cif->flags which is + * cif->rtype->size. But libffi API mandates that for integral types + * of size <= system register size, then we *MUST* write back + * the size of system register size. + * in our case, if size <= 8 bytes we must write back 8 bytes. + * floats, complex and structs are not affected, only integrals. + */ + wb_size = handle_small_int_ext(&int_extension_method, cif->rtype); + + switch (cif->abi) { + case FFI_SYSV: + DEBUG_PRINT("total_size: %lu\n", total_size); + DEBUG_PRINT("slot fitting args size: %lu\n", slot_fitting_args_size); + DEBUG_PRINT("rvalue: %p\n", rvalue); + DEBUG_PRINT("fn: %p\n", fn); + DEBUG_PRINT("rsize: %u\n", cif->flags); + DEBUG_PRINT("wb_size: %u\n", wb_size); + DEBUG_PRINT("int_extension_method: %u\n", int_extension_method); + local_rvalue = ffi_call_SYSV(total_size, slot_fitting_args_size, + &ecif, rvalue, fn, int_extension_method); + if ((cif->flags <= KVX_ABI_MAX_AGGREGATE_IN_REG_SIZE) + && (cif->rtype->type != FFI_TYPE_VOID)) + memcpy(rvalue, &local_rvalue, wb_size); + break; + default: + abort(); + break; + } +} + +/* Closures not supported yet */ +ffi_status +ffi_prep_closure_loc (ffi_closure* closure, + ffi_cif* cif, + void (*fun)(ffi_cif*,void*,void**,void*), + void *user_data, + void *codeloc) +{ + return FFI_BAD_ABI; +} + +#endif /* (__kvx__) */ diff --git a/native/libffi/src/kvx/ffitarget.h b/native/libffi/src/kvx/ffitarget.h new file mode 100644 index 0000000000..8df8735e3f --- /dev/null +++ b/native/libffi/src/kvx/ffitarget.h @@ -0,0 +1,75 @@ +/* ----------------------------------------------------------------------- + ffitarget.h - Copyright (c) 2020 Kalray + + KVX Target configuration macros + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + ``Software''), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice shall be included + in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + ----------------------------------------------------------------------- */ + +#ifndef LIBFFI_TARGET_H +#define LIBFFI_TARGET_H + +#ifndef LIBFFI_H +#error "Please do not include ffitarget.h directly into your source. Use ffi.h instead." +#endif + +/* ---- System specific configurations ----------------------------------- */ + +#ifndef LIBFFI_ASM +typedef unsigned long ffi_arg; +typedef signed long ffi_sarg; + +typedef enum ffi_abi { + FFI_FIRST_ABI = 0, + FFI_SYSV, + FFI_LAST_ABI, + FFI_DEFAULT_ABI = FFI_SYSV +} ffi_abi; + +/* Those values are set depending on return type + * they are used in the assembly code in sysv.S + */ +typedef enum kvx_intext_method { + KVX_RET_NONE = 0, + KVX_RET_SXBD = 1, + KVX_RET_SXHD = 2, + KVX_RET_SXWD = 3, + KVX_RET_ZXBD = 4, + KVX_RET_ZXHD = 5, + KVX_RET_ZXWD = 6 +} kvx_intext_method; + +#endif + +/* ---- Definitions for closures ----------------------------------------- */ + +/* This is only to allow Python to compile + * but closures are not supported yet + */ +#define FFI_CLOSURES 1 +#define FFI_TRAMPOLINE_SIZE 0 + +#define FFI_NATIVE_RAW_API 0 +#define FFI_TARGET_SPECIFIC_VARIADIC 1 +#define FFI_TARGET_HAS_COMPLEX_TYPE + +#endif + diff --git a/native/libffi/src/kvx/sysv.S b/native/libffi/src/kvx/sysv.S new file mode 100644 index 0000000000..952afc7063 --- /dev/null +++ b/native/libffi/src/kvx/sysv.S @@ -0,0 +1,127 @@ +/* Copyright (c) 2020 Kalray + +Permission is hereby granted, free of charge, to any person obtaining +a copy of this software and associated documentation files (the +``Software''), to deal in the Software without restriction, including +without limitation the rights to use, copy, modify, merge, publish, +distribute, sublicense, and/or sell copies of the Software, and to +permit persons to whom the Software is furnished to do so, subject to +the following conditions: + +The above copyright notice and this permission notice shall be +included in all copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, +EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF +MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. +IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY +CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, +TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE +SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */ + +#if defined(__kvx__) +#define LIBFFI_ASM +#include +#include +#include +#include + +.text +.global ffi_call_SYSV +.type ffi_call_SYSV, @function +.type ffi_prep_args, @function +.align 8 + +/* ffi_call_SYSV + + r0: total size to allocate on stack + r1: size of arg slots + r2: extended cif structure, DO NOT REMOVE: it is used by ffi_prep_args() + r3: return value address + r4: function to call + r5: integer sign extension method to be used +*/ +ffi_call_SYSV: + addd $r12 = $r12, -64 + so (-32)[$r12] = $r20r21r22r23 + ;; + sd (0)[$r12] = $r24 + ;; + get $r23 = $ra + copyd $r20 = $r12 + sbfd $r12 = $r0, $r12 + ;; + copyd $r0 = $r12 + copyd $r21 = $r3 + copyd $r22 = $r4 + copyd $r24 = $r5 + call ffi_prep_args + ;; + lo $r8r9r10r11 = (64)[$r12] + ;; + lo $r4r5r6r7 = (32)[$r12] + ;; + lo $r0r1r2r3 = (0)[$r12] + copyd $r12 = $r0 + /* $r15 is the register used by the ABI to return big (>32 bytes) + * structs by value. + * It is also referred to as the "struct register" in the ABI. + */ + copyd $r15 = $r21 + icall $r22 + ;; + pcrel $r4 = @pcrel(.Ltable) + cb.deqz $r24 ? .Lend + ;; + addx8d $r24 = $r24, $r4 + ;; + igoto $r24 + ;; +.Ltable: +0: /* we should never arrive here */ + goto .Lerror + nop + ;; +1: /* Sign extend byte to double */ + sxbd $r0 = $r0 + goto .Lend + ;; +2: /* Sign extend half to double */ + sxhd $r0 = $r0 + goto .Lend + ;; +3: /* Sign extend word to double */ + sxwd $r0 = $r0 + goto .Lend + ;; +4: /* Zero extend byte to double */ + zxbd $r0 = $r0 + goto .Lend + ;; +5: /* Zero extend half to double */ + zxhd $r0 = $r0 + goto .Lend + ;; +6: /* Zero extend word to double */ + zxwd $r0 = $r0 + /* Fallthrough to .Lend */ + ;; +.Lend: + ld $r24 = (0)[$r12] + ;; + set $ra = $r23 + lo $r20r21r22r23 = (32)[$r20] + addd $r12 = $r20, 64 + ;; + ret + ;; +.Lerror: + errop + ;; + +#endif /* __kvx__ */ + +#if defined __ELF__ && defined __linux__ + .section .note.GNU-stack,"",%progbits +#endif + diff --git a/native/libffi/src/m32r/ffi.c b/native/libffi/src/m32r/ffi.c index 300006349b..ab8fc4e48a 100644 --- a/native/libffi/src/m32r/ffi.c +++ b/native/libffi/src/m32r/ffi.c @@ -61,7 +61,7 @@ void ffi_prep_args(char *stack, extended_cif *ecif) /* Align if necessary. */ if (((*p_arg)->alignment - 1) & (unsigned) argp) - argp = (char *) ALIGN (argp, (*p_arg)->alignment); + argp = (char *) FFI_ALIGN (argp, (*p_arg)->alignment); if (avn != 0) { diff --git a/native/libffi/src/m68k/ffi.c b/native/libffi/src/m68k/ffi.c index 0dee9383a8..033018480c 100644 --- a/native/libffi/src/m68k/ffi.c +++ b/native/libffi/src/m68k/ffi.c @@ -105,7 +105,7 @@ ffi_prep_args (void *stack, extended_cif *ecif) /* Align if necessary. */ if ((sizeof(int) - 1) & z) - z = ALIGN(z, sizeof(int)); + z = FFI_ALIGN(z, sizeof(int)); } p_argv++; @@ -297,7 +297,7 @@ ffi_prep_incoming_args_SYSV (char *stack, void **avalue, ffi_cif *cif) /* Align if necessary */ if ((sizeof(int) - 1) & z) - z = ALIGN(z, sizeof(int)); + z = FFI_ALIGN(z, sizeof(int)); } p_argv++; diff --git a/native/libffi/src/m68k/sysv.S b/native/libffi/src/m68k/sysv.S index ec2b14f3d9..ea40f1100b 100644 --- a/native/libffi/src/m68k/sysv.S +++ b/native/libffi/src/m68k/sysv.S @@ -3,7 +3,7 @@ sysv.S - Copyright (c) 2012 Alan Hourihane Copyright (c) 1998, 2012 Andreas Schwab Copyright (c) 2008 Red Hat, Inc. - Copyright (c) 2012 Thorsten Glaser + Copyright (c) 2012, 2016 Thorsten Glaser m68k Foreign Function Interface @@ -72,6 +72,15 @@ CALLFUNC(ffi_call_SYSV): pea 4(%sp) #if !defined __PIC__ jsr CALLFUNC(ffi_prep_args) +#elif defined(__uClinux__) && defined(__ID_SHARED_LIBRARY__) + move.l _current_shared_library_a5_offset_(%a5),%a0 + move.l CALLFUNC(ffi_prep_args@GOT)(%a0),%a0 + jsr (%a0) +#elif defined(__mcoldfire__) && !defined(__mcfisab__) && !defined(__mcfisac__) + move.l #_GLOBAL_OFFSET_TABLE_@GOTPC,%a0 + lea (-6,%pc,%a0),%a0 + move.l CALLFUNC(ffi_prep_args@GOT)(%a0),%a0 + jsr (%a0) #else bsr.l CALLFUNC(ffi_prep_args@PLTPC) #endif @@ -215,6 +224,15 @@ CALLFUNC(ffi_closure_SYSV): move.l %a0,-(%sp) #if !defined __PIC__ jsr CALLFUNC(ffi_closure_SYSV_inner) +#elif defined(__uClinux__) && defined(__ID_SHARED_LIBRARY__) + move.l _current_shared_library_a5_offset_(%a5),%a0 + move.l CALLFUNC(ffi_closure_SYSV_inner@GOT)(%a0),%a0 + jsr (%a0) +#elif defined(__mcoldfire__) && !defined(__mcfisab__) && !defined(__mcfisac__) + move.l #_GLOBAL_OFFSET_TABLE_@GOTPC,%a0 + lea (-6,%pc,%a0),%a0 + move.l CALLFUNC(ffi_closure_SYSV_inner@GOT)(%a0),%a0 + jsr (%a0) #else bsr.l CALLFUNC(ffi_closure_SYSV_inner@PLTPC) #endif @@ -317,6 +335,15 @@ CALLFUNC(ffi_closure_struct_SYSV): move.l %a0,-(%sp) #if !defined __PIC__ jsr CALLFUNC(ffi_closure_SYSV_inner) +#elif defined(__uClinux__) && defined(__ID_SHARED_LIBRARY__) + move.l _current_shared_library_a5_offset_(%a5),%a0 + move.l CALLFUNC(ffi_closure_SYSV_inner@GOT)(%a0),%a0 + jsr (%a0) +#elif defined(__mcoldfire__) && !defined(__mcfisab__) && !defined(__mcfisac__) + move.l #_GLOBAL_OFFSET_TABLE_@GOTPC,%a0 + lea (-6,%pc,%a0),%a0 + move.l CALLFUNC(ffi_closure_SYSV_inner@GOT)(%a0),%a0 + jsr (%a0) #else bsr.l CALLFUNC(ffi_closure_SYSV_inner@PLTPC) #endif diff --git a/native/libffi/src/m88k/ffi.c b/native/libffi/src/m88k/ffi.c index 68df494955..a14fb20453 100644 --- a/native/libffi/src/m88k/ffi.c +++ b/native/libffi/src/m88k/ffi.c @@ -134,7 +134,11 @@ ffi_prep_args (void *stack, extended_cif *ecif) /* Enforce proper stack alignment of 64-bit types */ if (argp == stackp && a > sizeof (int)) { +<<<<<<< HEAD stackp = (char *) ALIGN(stackp, a); +======= + stackp = (char *) FFI_ALIGN(stackp, a); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 argp = stackp; } @@ -177,7 +181,11 @@ ffi_prep_args (void *stack, extended_cif *ecif) /* Align if necessary. */ if ((sizeof (int) - 1) & z) +<<<<<<< HEAD z = ALIGN(z, sizeof (int)); +======= + z = FFI_ALIGN(z, sizeof (int)); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 p_argv++; @@ -320,7 +328,11 @@ ffi_prep_closure_args_OBSD (ffi_cif *cif, void **avalue, unsigned int *regp, /* Enforce proper stack alignment of 64-bit types */ if (argp == stackp && a > sizeof (int)) { +<<<<<<< HEAD stackp = (char *) ALIGN(stackp, a); +======= + stackp = (char *) FFI_ALIGN(stackp, a); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 argp = stackp; } @@ -331,7 +343,11 @@ ffi_prep_closure_args_OBSD (ffi_cif *cif, void **avalue, unsigned int *regp, /* Align if necessary */ if ((sizeof (int) - 1) & z) +<<<<<<< HEAD z = ALIGN(z, sizeof (int)); +======= + z = FFI_ALIGN(z, sizeof (int)); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 p_argv++; diff --git a/native/libffi/src/metag/ffi.c b/native/libffi/src/metag/ffi.c index 46b383e7a3..3aecb0b811 100644 --- a/native/libffi/src/metag/ffi.c +++ b/native/libffi/src/metag/ffi.c @@ -61,7 +61,7 @@ unsigned int ffi_prep_args(char *stack, extended_cif *ecif) argp -= z; /* Align if necessary */ - argp = (char *) ALIGN_DOWN(ALIGN_DOWN(argp, (*p_arg)->alignment), 4); + argp = (char *) FFI_ALIGN_DOWN(FFI_ALIGN_DOWN(argp, (*p_arg)->alignment), 4); if (z < sizeof(int)) { z = sizeof(int); @@ -93,7 +93,7 @@ unsigned int ffi_prep_args(char *stack, extended_cif *ecif) /* return the size of the arguments to be passed in registers, padded to an 8 byte boundary to preserve stack alignment */ - return ALIGN(MIN(stack - argp, 6*4), 8); + return FFI_ALIGN(MIN(stack - argp, 6*4), 8); } /* Perform machine dependent cif processing */ @@ -112,20 +112,20 @@ ffi_status ffi_prep_cif_machdep(ffi_cif *cif) /* Add any padding if necessary */ if (((*ptr)->alignment - 1) & bytes) - bytes = ALIGN(bytes, (*ptr)->alignment); + bytes = FFI_ALIGN(bytes, (*ptr)->alignment); - bytes += ALIGN((*ptr)->size, 4); + bytes += FFI_ALIGN((*ptr)->size, 4); } /* Ensure arg space is aligned to an 8-byte boundary */ - bytes = ALIGN(bytes, 8); + bytes = FFI_ALIGN(bytes, 8); /* Make space for the return structure pointer */ if (cif->rtype->type == FFI_TYPE_STRUCT) { bytes += sizeof(void*); /* Ensure stack is aligned to an 8-byte boundary */ - bytes = ALIGN(bytes, 8); + bytes = FFI_ALIGN(bytes, 8); } cif->bytes = bytes; @@ -319,7 +319,7 @@ static void ffi_prep_incoming_args_SYSV(char *stack, void **rvalue, if (alignment < 4) alignment = 4; if ((alignment - 1) & (unsigned)argp) - argp = (char *) ALIGN(argp, alignment); + argp = (char *) FFI_ALIGN(argp, alignment); z = (*p_arg)->size; *p_argv = (void*) argp; diff --git a/native/libffi/src/microblaze/ffi.c b/native/libffi/src/microblaze/ffi.c index ea962ea483..df6e33c97d 100644 --- a/native/libffi/src/microblaze/ffi.c +++ b/native/libffi/src/microblaze/ffi.c @@ -35,7 +35,7 @@ extern void ffi_closure_SYSV(void); #define WORD_SIZE sizeof(unsigned int) #define ARGS_REGISTER_SIZE (WORD_SIZE * 6) -#define WORD_ALIGN(x) ALIGN(x, WORD_SIZE) +#define WORD_FFI_ALIGN(x) FFI_ALIGN(x, WORD_SIZE) /* ffi_prep_args is called by the assembly routine once stack space has been allocated for the function's arguments */ @@ -46,12 +46,12 @@ void ffi_prep_args(void* stack, extended_cif* ecif) void** p_argv; void* stack_args_p = stack; - p_argv = ecif->avalue; - if (ecif == NULL || ecif->cif == NULL) { return; /* no description to prepare */ } + p_argv = ecif->avalue; + if ((ecif->cif->rtype != NULL) && (ecif->cif->rtype->type == FFI_TYPE_STRUCT)) { @@ -74,7 +74,7 @@ void ffi_prep_args(void* stack, extended_cif* ecif) int type = (*p_arg)->type; void* value = p_argv[i]; char* addr = stack_args_p; - int aligned_size = WORD_ALIGN(size); + int aligned_size = WORD_FFI_ALIGN(size); /* force word alignment on the stack */ stack_args_p += aligned_size; @@ -259,7 +259,7 @@ void ffi_closure_call_SYSV(void* register_args, void* stack_args, avalue[i] = ptr; break; } - ptr += WORD_ALIGN(arg_types[i]->size); + ptr += WORD_FFI_ALIGN(arg_types[i]->size); } /* set the return type info passed back to the wrapper */ diff --git a/native/libffi/src/mips/ffi.c b/native/libffi/src/mips/ffi.c index 5d0dd70cb3..979ca49f1b 100644 --- a/native/libffi/src/mips/ffi.c +++ b/native/libffi/src/mips/ffi.c @@ -29,6 +29,7 @@ #include #include +#include #include #ifdef __GNUC__ @@ -38,7 +39,9 @@ #endif #ifndef USE__BUILTIN___CLEAR_CACHE -# if defined(__OpenBSD__) +# if defined(__FreeBSD__) +# include +# elif defined(__OpenBSD__) # include # else # include @@ -116,7 +119,7 @@ static void ffi_prep_args(char *stack, if ((a - 1) & (unsigned long) argp) { - argp = (char *) ALIGN(argp, a); + argp = (char *) FFI_ALIGN(argp, a); FIX_ARGP; } @@ -247,7 +250,7 @@ calc_n32_struct_flags(int soft_float, ffi_type *arg, while ((e = arg->elements[index])) { /* Align this object. */ - *loc = ALIGN(*loc, e->alignment); + *loc = FFI_ALIGN(*loc, e->alignment); if (e->type == FFI_TYPE_DOUBLE) { /* Already aligned to FFI_SIZEOF_ARG. */ @@ -262,7 +265,7 @@ calc_n32_struct_flags(int soft_float, ffi_type *arg, index++; } /* Next Argument register at alignment of FFI_SIZEOF_ARG. */ - *arg_reg = ALIGN(*loc, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; + *arg_reg = FFI_ALIGN(*loc, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; return flags; } @@ -322,9 +325,10 @@ calc_n32_return_struct_flags(int soft_float, ffi_type *arg) #endif /* Perform machine dependent cif processing */ -ffi_status ffi_prep_cif_machdep(ffi_cif *cif) +static ffi_status ffi_prep_cif_machdep_int(ffi_cif *cif, unsigned nfixedargs) { cif->flags = 0; + cif->mips_nfixedargs = nfixedargs; #ifdef FFI_MIPS_O32 /* Set the flags necessary for O32 processing. FFI_O32_SOFT_FLOAT @@ -333,7 +337,7 @@ ffi_status ffi_prep_cif_machdep(ffi_cif *cif) if (cif->rtype->type != FFI_TYPE_STRUCT && cif->abi == FFI_O32) { - if (cif->nargs > 0) + if (cif->nargs > 0 && cif->nargs == nfixedargs) { switch ((cif->arg_types)[0]->type) { @@ -450,7 +454,9 @@ ffi_status ffi_prep_cif_machdep(ffi_cif *cif) while (count-- > 0 && arg_reg < 8) { type = (cif->arg_types)[index]->type; - if (soft_float) + + // Pass variadic arguments in integer registers even if they're floats + if (soft_float || index >= nfixedargs) { switch (type) { @@ -474,9 +480,9 @@ ffi_status ffi_prep_cif_machdep(ffi_cif *cif) break; case FFI_TYPE_LONGDOUBLE: /* Align it. */ - arg_reg = ALIGN(arg_reg, 2); + arg_reg = FFI_ALIGN(arg_reg, 2); /* Treat it as two adjacent doubles. */ - if (soft_float) + if (soft_float || index >= nfixedargs) { arg_reg += 2; } @@ -493,7 +499,7 @@ ffi_status ffi_prep_cif_machdep(ffi_cif *cif) case FFI_TYPE_STRUCT: loc = arg_reg * FFI_SIZEOF_ARG; - cif->flags += calc_n32_struct_flags(soft_float, + cif->flags += calc_n32_struct_flags(soft_float || index >= nfixedargs, (cif->arg_types)[index], &loc, &arg_reg); break; @@ -578,17 +584,30 @@ ffi_status ffi_prep_cif_machdep(ffi_cif *cif) return FFI_OK; } +ffi_status ffi_prep_cif_machdep(ffi_cif *cif) +{ + return ffi_prep_cif_machdep_int(cif, cif->nargs); +} + +ffi_status ffi_prep_cif_machdep_var(ffi_cif *cif, + unsigned nfixedargs, + unsigned ntotalargs MAYBE_UNUSED) +{ + return ffi_prep_cif_machdep_int(cif, nfixedargs); +} + /* Low level routine for calling O32 functions */ extern int ffi_call_O32(void (*)(char *, extended_cif *, int, int), extended_cif *, unsigned, - unsigned, unsigned *, void (*)(void)); + unsigned, unsigned *, void (*)(void), void *closure); /* Low level routine for calling N32 functions */ extern int ffi_call_N32(void (*)(char *, extended_cif *, int, int), extended_cif *, unsigned, - unsigned, void *, void (*)(void)); + unsigned, void *, void (*)(void), void *closure); -void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +void ffi_call_int(ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) { extended_cif ecif; @@ -610,7 +629,7 @@ void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) case FFI_O32: case FFI_O32_SOFT_FLOAT: ffi_call_O32(ffi_prep_args, &ecif, cif->bytes, - cif->flags, ecif.rvalue, fn); + cif->flags, ecif.rvalue, fn, closure); break; #endif @@ -642,7 +661,7 @@ void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) #endif } ffi_call_N32(ffi_prep_args, &ecif, cif->bytes, - cif->flags, rvalue_copy, fn); + cif->flags, rvalue_copy, fn, closure); if (copy_rvalue) memcpy(ecif.rvalue, rvalue_copy + copy_offset, cif->rtype->size); } @@ -655,11 +674,27 @@ void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) } } +void +ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +{ + ffi_call_int (cif, fn, rvalue, avalue, NULL); +} + +void +ffi_call_go (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) +{ + ffi_call_int (cif, fn, rvalue, avalue, closure); +} + + #if FFI_CLOSURES #if defined(FFI_MIPS_O32) extern void ffi_closure_O32(void); +extern void ffi_go_closure_O32(void); #else extern void ffi_closure_N32(void); +extern void ffi_go_closure_N32(void); #endif /* FFI_MIPS_O32 */ ffi_status @@ -698,7 +733,11 @@ ffi_prep_closure_loc (ffi_closure *closure, /* lui $12,high(codeloc) */ tramp[2] = 0x3c0c0000 | ((unsigned)codeloc >> 16); /* jr $25 */ +#if !defined(__mips_isa_rev) || (__mips_isa_rev<6) tramp[3] = 0x03200008; +#else + tramp[3] = 0x03200009; +#endif /* ori $12,low(codeloc) */ tramp[4] = 0x358c0000 | ((unsigned)codeloc & 0xffff); #else @@ -726,7 +765,11 @@ ffi_prep_closure_loc (ffi_closure *closure, /* ori $25,low(fn) */ tramp[10] = 0x37390000 | ((unsigned long)fn & 0xffff); /* jr $25 */ +#if !defined(__mips_isa_rev) || (__mips_isa_rev<6) tramp[11] = 0x03200008; +#else + tramp[11] = 0x03200009; +#endif /* ori $12,low(codeloc) */ tramp[12] = 0x358c0000 | ((unsigned long)codeloc & 0xffff); @@ -736,11 +779,13 @@ ffi_prep_closure_loc (ffi_closure *closure, closure->fun = fun; closure->user_data = user_data; +#if !defined(__FreeBSD__) #ifdef USE__BUILTIN___CLEAR_CACHE __builtin___clear_cache(clear_location, clear_location + FFI_TRAMPOLINE_SIZE); #else cacheflush (clear_location, FFI_TRAMPOLINE_SIZE, ICACHE); #endif +#endif /* ! __FreeBSD__ */ return FFI_OK; } @@ -762,27 +807,28 @@ ffi_prep_closure_loc (ffi_closure *closure, * Based on the similar routine for sparc. */ int -ffi_closure_mips_inner_O32 (ffi_closure *closure, +ffi_closure_mips_inner_O32 (ffi_cif *cif, + void (*fun)(ffi_cif*, void*, void**, void*), + void *user_data, void *rvalue, ffi_arg *ar, double *fpr) { - ffi_cif *cif; void **avaluep; ffi_arg *avalue; ffi_type **arg_types; int i, avn, argn, seen_int; - cif = closure->cif; avalue = alloca (cif->nargs * sizeof (ffi_arg)); avaluep = alloca (cif->nargs * sizeof (ffi_arg)); - seen_int = (cif->abi == FFI_O32_SOFT_FLOAT); + seen_int = (cif->abi == FFI_O32_SOFT_FLOAT) || (cif->mips_nfixedargs != cif->nargs); argn = 0; if ((cif->flags >> (FFI_FLAG_BITS * 2)) == FFI_TYPE_STRUCT) { - rvalue = (void *)(UINT32)ar[0]; + rvalue = (void *)(uintptr_t)ar[0]; argn = 1; + seen_int = 1; } i = 0; @@ -791,6 +837,8 @@ ffi_closure_mips_inner_O32 (ffi_closure *closure, while (i < avn) { + if (arg_types[i]->alignment == 8 && (argn & 0x1)) + argn++; if (i < 2 && !seen_int && (arg_types[i]->type == FFI_TYPE_FLOAT || arg_types[i]->type == FFI_TYPE_DOUBLE || @@ -805,8 +853,6 @@ ffi_closure_mips_inner_O32 (ffi_closure *closure, } else { - if (arg_types[i]->alignment == 8 && (argn & 0x1)) - argn++; switch (arg_types[i]->type) { case FFI_TYPE_SINT8: @@ -835,12 +881,12 @@ ffi_closure_mips_inner_O32 (ffi_closure *closure, } seen_int = 1; } - argn += ALIGN(arg_types[i]->size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; + argn += FFI_ALIGN(arg_types[i]->size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; i++; } /* Invoke the closure. */ - (closure->fun) (cif, rvalue, avaluep, closure->user_data); + fun(cif, rvalue, avaluep, user_data); if (cif->abi == FFI_O32_SOFT_FLOAT) { @@ -876,7 +922,7 @@ copy_struct_N32(char *target, unsigned offset, ffi_abi abi, ffi_type *type, char *argp; char *fpp; - o = ALIGN(offset, elt_type->alignment); + o = FFI_ALIGN(offset, elt_type->alignment); arg_offset += o - offset; offset = o; argn += arg_offset / sizeof(ffi_arg); @@ -916,11 +962,12 @@ copy_struct_N32(char *target, unsigned offset, ffi_abi abi, ffi_type *type, * */ int -ffi_closure_mips_inner_N32 (ffi_closure *closure, +ffi_closure_mips_inner_N32 (ffi_cif *cif, + void (*fun)(ffi_cif*, void*, void**, void*), + void *user_data, void *rvalue, ffi_arg *ar, ffi_arg *fpr) { - ffi_cif *cif; void **avaluep; ffi_arg *avalue; ffi_type **arg_types; @@ -928,7 +975,6 @@ ffi_closure_mips_inner_N32 (ffi_closure *closure, int soft_float; ffi_arg *argp; - cif = closure->cif; soft_float = cif->abi == FFI_N64_SOFT_FLOAT || cif->abi == FFI_N32_SOFT_FLOAT; avalue = alloca (cif->nargs * sizeof (ffi_arg)); @@ -956,10 +1002,10 @@ ffi_closure_mips_inner_N32 (ffi_closure *closure, || arg_types[i]->type == FFI_TYPE_DOUBLE || arg_types[i]->type == FFI_TYPE_LONGDOUBLE) { - argp = (argn >= 8 || soft_float) ? ar + argn : fpr + argn; - if ((arg_types[i]->type == FFI_TYPE_LONGDOUBLE) && ((unsigned)argp & (arg_types[i]->alignment-1))) + argp = (argn >= 8 || i >= cif->mips_nfixedargs || soft_float) ? ar + argn : fpr + argn; + if ((arg_types[i]->type == FFI_TYPE_LONGDOUBLE) && ((uintptr_t)argp & (arg_types[i]->alignment-1))) { - argp=(ffi_arg*)ALIGN(argp,arg_types[i]->alignment); + argp=(ffi_arg*)FFI_ALIGN(argp,arg_types[i]->alignment); argn++; } #if defined(__MIPSEB__) || defined(_MIPSEB) @@ -974,7 +1020,7 @@ ffi_closure_mips_inner_N32 (ffi_closure *closure, unsigned type = arg_types[i]->type; if (arg_types[i]->alignment > sizeof(ffi_arg)) - argn = ALIGN(argn, arg_types[i]->alignment / sizeof(ffi_arg)); + argn = FFI_ALIGN(argn, arg_types[i]->alignment / sizeof(ffi_arg)); argp = ar + argn; @@ -1025,7 +1071,7 @@ ffi_closure_mips_inner_N32 (ffi_closure *closure, it was passed in registers. */ avaluep[i] = alloca(arg_types[i]->size); copy_struct_N32(avaluep[i], 0, cif->abi, arg_types[i], - argn, 0, ar, fpr, soft_float); + argn, 0, ar, fpr, i >= cif->mips_nfixedargs || soft_float); break; } @@ -1035,16 +1081,54 @@ ffi_closure_mips_inner_N32 (ffi_closure *closure, break; } } - argn += ALIGN(arg_types[i]->size, sizeof(ffi_arg)) / sizeof(ffi_arg); + argn += FFI_ALIGN(arg_types[i]->size, sizeof(ffi_arg)) / sizeof(ffi_arg); i++; } /* Invoke the closure. */ - (closure->fun) (cif, rvalue, avaluep, closure->user_data); + fun (cif, rvalue, avaluep, user_data); return cif->flags >> (FFI_FLAG_BITS * 8); } #endif /* FFI_MIPS_N32 */ +#if defined(FFI_MIPS_O32) +extern void ffi_closure_O32(void); +extern void ffi_go_closure_O32(void); +#else +extern void ffi_closure_N32(void); +extern void ffi_go_closure_N32(void); +#endif /* FFI_MIPS_O32 */ + +ffi_status +ffi_prep_go_closure (ffi_go_closure* closure, ffi_cif* cif, + void (*fun)(ffi_cif*,void*,void**,void*)) +{ + void * fn; + +#if defined(FFI_MIPS_O32) + if (cif->abi != FFI_O32 && cif->abi != FFI_O32_SOFT_FLOAT) + return FFI_BAD_ABI; + fn = ffi_go_closure_O32; +#else +#if _MIPS_SIM ==_ABIN32 + if (cif->abi != FFI_N32 + && cif->abi != FFI_N32_SOFT_FLOAT) + return FFI_BAD_ABI; +#else + if (cif->abi != FFI_N64 + && cif->abi != FFI_N64_SOFT_FLOAT) + return FFI_BAD_ABI; +#endif + fn = ffi_go_closure_N32; +#endif /* FFI_MIPS_O32 */ + + closure->tramp = (void *)fn; + closure->cif = cif; + closure->fun = fun; + + return FFI_OK; +} + #endif /* FFI_CLOSURES */ diff --git a/native/libffi/src/mips/ffitarget.h b/native/libffi/src/mips/ffitarget.h index 717d65951c..fdd5ca90ac 100644 --- a/native/libffi/src/mips/ffitarget.h +++ b/native/libffi/src/mips/ffitarget.h @@ -32,7 +32,7 @@ #error "Please do not include ffitarget.h directly into your source. Use ffi.h instead." #endif -#ifdef linux +#ifdef __linux__ # include #elif defined(__rtems__) /* @@ -41,7 +41,7 @@ #define _MIPS_SIM_ABI32 1 #define _MIPS_SIM_NABI32 2 #define _MIPS_SIM_ABI64 3 -#elif !defined(__OpenBSD__) +#elif !defined(__OpenBSD__) && !defined(__FreeBSD__) # include #endif @@ -224,24 +224,21 @@ typedef enum ffi_abi { #endif } ffi_abi; -#define FFI_EXTRA_CIF_FIELDS unsigned rstruct_flag +#define FFI_EXTRA_CIF_FIELDS unsigned rstruct_flag; unsigned mips_nfixedargs +#define FFI_TARGET_SPECIFIC_VARIADIC #endif /* !LIBFFI_ASM */ /* ---- Definitions for closures ----------------------------------------- */ -#if defined(FFI_MIPS_O32) #define FFI_CLOSURES 1 -#define FFI_TRAMPOLINE_SIZE 20 -#else -/* N32/N64. */ -# define FFI_CLOSURES 1 -#if _MIPS_SIM==_ABI64 -#define FFI_TRAMPOLINE_SIZE 52 +#define FFI_GO_CLOSURES 1 +#define FFI_NATIVE_RAW_API 0 + +#if defined(FFI_MIPS_O32) || (_MIPS_SIM ==_ABIN32) +# define FFI_TRAMPOLINE_SIZE 20 #else -#define FFI_TRAMPOLINE_SIZE 20 +# define FFI_TRAMPOLINE_SIZE 56 #endif -#endif /* FFI_MIPS_O32 */ -#define FFI_NATIVE_RAW_API 0 #endif diff --git a/native/libffi/src/mips/n32.S b/native/libffi/src/mips/n32.S index c6985d30a6..1a940b6d2a 100644 --- a/native/libffi/src/mips/n32.S +++ b/native/libffi/src/mips/n32.S @@ -37,36 +37,43 @@ #define flags a3 #define raddr a4 #define fn a5 +#define closure a6 -#define SIZEOF_FRAME ( 8 * FFI_SIZEOF_ARG ) +/* Note: to keep stack 16 byte aligned we need even number slots + used 9 slots here +*/ +#define SIZEOF_FRAME ( 10 * FFI_SIZEOF_ARG ) #ifdef __GNUC__ .abicalls #endif +#if !defined(__mips_isa_rev) || (__mips_isa_rev<6) .set mips4 +#endif .text .align 2 .globl ffi_call_N32 .ent ffi_call_N32 ffi_call_N32: -.LFB3: +.LFB0: .frame $fp, SIZEOF_FRAME, ra .mask 0xc0000000,-FFI_SIZEOF_ARG .fmask 0x00000000,0 # Prologue SUBU $sp, SIZEOF_FRAME # Frame size -.LCFI0: +.LCFI00: REG_S $fp, SIZEOF_FRAME - 2*FFI_SIZEOF_ARG($sp) # Save frame pointer REG_S ra, SIZEOF_FRAME - 1*FFI_SIZEOF_ARG($sp) # Save return address -.LCFI1: +.LCFI01: move $fp, $sp -.LCFI3: +.LCFI02: move t9, callback # callback function pointer REG_S bytes, 2*FFI_SIZEOF_ARG($fp) # bytes REG_S flags, 3*FFI_SIZEOF_ARG($fp) # flags REG_S raddr, 4*FFI_SIZEOF_ARG($fp) # raddr REG_S fn, 5*FFI_SIZEOF_ARG($fp) # fn + REG_S closure, 6*FFI_SIZEOF_ARG($fp) # closure # Allocate at least 4 words in the argstack move v0, bytes @@ -107,6 +114,16 @@ loadregs: REG_L t6, 3*FFI_SIZEOF_ARG($fp) # load the flags word into t6. +#ifdef __mips_soft_float + REG_L a0, 0*FFI_SIZEOF_ARG(t9) + REG_L a1, 1*FFI_SIZEOF_ARG(t9) + REG_L a2, 2*FFI_SIZEOF_ARG(t9) + REG_L a3, 3*FFI_SIZEOF_ARG(t9) + REG_L a4, 4*FFI_SIZEOF_ARG(t9) + REG_L a5, 5*FFI_SIZEOF_ARG(t9) + REG_L a6, 6*FFI_SIZEOF_ARG(t9) + REG_L a7, 7*FFI_SIZEOF_ARG(t9) +#else and t4, t6, ((1< 24) ? 24 : ALIGN_DOWN(count, 8))); + return (stack + ((count > 24) ? 24 : FFI_ALIGN_DOWN(count, 8))); } /* Perform machine dependent cif processing */ @@ -111,7 +111,7 @@ ffi_status ffi_prep_cif_machdep(ffi_cif *cif) else cif->flags = cif->rtype->size; - cif->bytes = ALIGN (cif->bytes, 8); + cif->bytes = FFI_ALIGN (cif->bytes, 8); return FFI_OK; } @@ -159,7 +159,7 @@ void ffi_closure_eabi (unsigned arg1, unsigned arg2, unsigned arg3, unsigned arg4, unsigned arg5, unsigned arg6) { /* This function is called by a trampoline. The trampoline stows a - pointer to the ffi_closure object in $r7. We must save this + pointer to the ffi_closure object in $r12. We must save this pointer in a place that will persist while we do our work. */ register ffi_closure *creg __asm__ ("$r12"); ffi_closure *closure = creg; @@ -215,7 +215,18 @@ void ffi_closure_eabi (unsigned arg1, unsigned arg2, unsigned arg3, break; default: /* This is an 8-byte value. */ - avalue[i] = ptr; + if (ptr == (char *) ®ister_args[5]) + { + /* The value is split across two locations */ + unsigned *ip = alloca(8); + avalue[i] = ip; + ip[0] = *(unsigned *) ptr; + ip[1] = *(unsigned *) stack_args; + } + else + { + avalue[i] = ptr; + } ptr += 4; break; } @@ -223,8 +234,10 @@ void ffi_closure_eabi (unsigned arg1, unsigned arg2, unsigned arg3, /* If we've handled more arguments than fit in registers, start looking at the those passed on the stack. */ - if (ptr == ®ister_args[6]) + if (ptr == (char *) ®ister_args[6]) ptr = stack_args; + else if (ptr == (char *) ®ister_args[7]) + ptr = stack_args + 4; } /* Invoke the closure. */ @@ -257,7 +270,7 @@ ffi_prep_closure_loc (ffi_closure* closure, fn = (unsigned long) ffi_closure_eabi; - tramp[0] = 0x01e0; /* ldi.l $r7, .... */ + tramp[0] = 0x01e0; /* ldi.l $r12, .... */ tramp[1] = cls >> 16; tramp[2] = cls & 0xffff; tramp[3] = 0x1a00; /* jmpa .... */ diff --git a/native/libffi/src/nios2/ffi.c b/native/libffi/src/nios2/ffi.c index 2efa033f9c..36e0d27158 100644 --- a/native/libffi/src/nios2/ffi.c +++ b/native/libffi/src/nios2/ffi.c @@ -101,7 +101,11 @@ void ffi_prep_args (char *stack, extended_cif *ecif) /* Align argp as appropriate for the argument type. */ if ((alignment - 1) & (unsigned) argp) +<<<<<<< HEAD argp = (char *) ALIGN (argp, alignment); +======= + argp = (char *) FFI_ALIGN (argp, alignment); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* Copy the argument, promoting integral types smaller than a word to word size. */ @@ -230,7 +234,11 @@ ffi_closure_helper (unsigned char *args, /* Align argp as appropriate for the argument type. */ if ((alignment - 1) & (unsigned) argp) +<<<<<<< HEAD argp = (char *) ALIGN (argp, alignment); +======= + argp = (char *) FFI_ALIGN (argp, alignment); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* Arguments smaller than an int are promoted to int. */ if (size < sizeof (int)) diff --git a/native/libffi/src/pa/ffi.c b/native/libffi/src/pa/ffi.c index 4ce2bc6f0e..95e669473e 100644 --- a/native/libffi/src/pa/ffi.c +++ b/native/libffi/src/pa/ffi.c @@ -421,12 +421,15 @@ ffi_status ffi_closure_inner_pa32(ffi_closure *closure, UINT32 *stack) ffi_cif *cif; void **avalue; void *rvalue; - UINT32 ret[2]; /* function can return up to 64-bits in registers */ + /* Functions can return up to 64-bits in registers. Return address + must be double word aligned. */ + union { double rd; UINT32 ret[2]; } u; ffi_type **p_arg; char *tmp; int i, avn; unsigned int slot = FIRST_ARG_SLOT; register UINT32 r28 asm("r28"); + ffi_closure *c = (ffi_closure *)FFI_RESTORE_PTR (closure); cif = closure->cif; @@ -434,7 +437,7 @@ ffi_status ffi_closure_inner_pa32(ffi_closure *closure, UINT32 *stack) if (cif->flags == FFI_TYPE_STRUCT) rvalue = (void *)r28; else - rvalue = &ret[0]; + rvalue = &u; avalue = (void **)alloca(cif->nargs * FFI_SIZEOF_ARG); avn = cif->nargs; @@ -529,35 +532,35 @@ ffi_status ffi_closure_inner_pa32(ffi_closure *closure, UINT32 *stack) } /* Invoke the closure. */ - (closure->fun) (cif, rvalue, avalue, closure->user_data); + (c->fun) (cif, rvalue, avalue, c->user_data); - debug(3, "after calling function, ret[0] = %08x, ret[1] = %08x\n", ret[0], - ret[1]); + debug(3, "after calling function, ret[0] = %08x, ret[1] = %08x\n", u.ret[0], + u.ret[1]); /* Store the result using the lower 2 bytes of the flags. */ switch (cif->flags) { case FFI_TYPE_UINT8: - *(stack - FIRST_ARG_SLOT) = (UINT8)(ret[0] >> 24); + *(stack - FIRST_ARG_SLOT) = (UINT8)(u.ret[0] >> 24); break; case FFI_TYPE_SINT8: - *(stack - FIRST_ARG_SLOT) = (SINT8)(ret[0] >> 24); + *(stack - FIRST_ARG_SLOT) = (SINT8)(u.ret[0] >> 24); break; case FFI_TYPE_UINT16: - *(stack - FIRST_ARG_SLOT) = (UINT16)(ret[0] >> 16); + *(stack - FIRST_ARG_SLOT) = (UINT16)(u.ret[0] >> 16); break; case FFI_TYPE_SINT16: - *(stack - FIRST_ARG_SLOT) = (SINT16)(ret[0] >> 16); + *(stack - FIRST_ARG_SLOT) = (SINT16)(u.ret[0] >> 16); break; case FFI_TYPE_INT: case FFI_TYPE_SINT32: case FFI_TYPE_UINT32: - *(stack - FIRST_ARG_SLOT) = ret[0]; + *(stack - FIRST_ARG_SLOT) = u.ret[0]; break; case FFI_TYPE_SINT64: case FFI_TYPE_UINT64: - *(stack - FIRST_ARG_SLOT) = ret[0]; - *(stack - FIRST_ARG_SLOT - 1) = ret[1]; + *(stack - FIRST_ARG_SLOT) = u.ret[0]; + *(stack - FIRST_ARG_SLOT - 1) = u.ret[1]; break; case FFI_TYPE_DOUBLE: @@ -577,7 +580,7 @@ ffi_status ffi_closure_inner_pa32(ffi_closure *closure, UINT32 *stack) case FFI_TYPE_SMALL_STRUCT4: tmp = (void*)(stack - FIRST_ARG_SLOT); tmp += 4 - cif->rtype->size; - memcpy((void*)tmp, &ret[0], cif->rtype->size); + memcpy((void*)tmp, &u, cif->rtype->size); break; case FFI_TYPE_SMALL_STRUCT5: @@ -598,7 +601,7 @@ ffi_status ffi_closure_inner_pa32(ffi_closure *closure, UINT32 *stack) } memset (ret2, 0, sizeof (ret2)); - memcpy ((char *)ret2 + off, ret, 8 - off); + memcpy ((char *)ret2 + off, &u, 8 - off); *(stack - FIRST_ARG_SLOT) = ret2[0]; *(stack - FIRST_ARG_SLOT - 1) = ret2[1]; @@ -630,89 +633,41 @@ ffi_prep_closure_loc (ffi_closure* closure, void *user_data, void *codeloc) { - UINT32 *tramp = (UINT32 *)(closure->tramp); -#ifdef PA_HPUX - UINT32 *tmp; -#endif + ffi_closure *c = (ffi_closure *)FFI_RESTORE_PTR (closure); + + /* The layout of a function descriptor. A function pointer with the PLABEL + bit set points to a function descriptor. */ + struct pa32_fd + { + UINT32 code_pointer; + UINT32 gp; + }; + + struct ffi_pa32_trampoline_struct + { + UINT32 code_pointer; /* Pointer to ffi_closure_unix. */ + UINT32 fake_gp; /* Pointer to closure, installed as gp. */ + UINT32 real_gp; /* Real gp value. */ + }; + + struct ffi_pa32_trampoline_struct *tramp; + struct pa32_fd *fd; if (cif->abi != FFI_PA32) return FFI_BAD_ABI; - /* Make a small trampoline that will branch to our - handler function. Use PC-relative addressing. */ - -#ifdef PA_LINUX - tramp[0] = 0xeaa00000; /* b,l .+8,%r21 ; %r21 <- pc+8 */ - tramp[1] = 0xd6a01c1e; /* depi 0,31,2,%r21 ; mask priv bits */ - tramp[2] = 0x4aa10028; /* ldw 20(%r21),%r1 ; load plabel */ - tramp[3] = 0x36b53ff1; /* ldo -8(%r21),%r21 ; get closure addr */ - tramp[4] = 0x0c201096; /* ldw 0(%r1),%r22 ; address of handler */ - tramp[5] = 0xeac0c000; /* bv%r0(%r22) ; branch to handler */ - tramp[6] = 0x0c281093; /* ldw 4(%r1),%r19 ; GP of handler */ - tramp[7] = ((UINT32)(ffi_closure_pa32) & ~2); - - /* Flush d/icache -- have to flush up 2 two lines because of - alignment. */ - __asm__ volatile( - "fdc 0(%0)\n\t" - "fdc %1(%0)\n\t" - "fic 0(%%sr4, %0)\n\t" - "fic %1(%%sr4, %0)\n\t" - "sync\n\t" - "nop\n\t" - "nop\n\t" - "nop\n\t" - "nop\n\t" - "nop\n\t" - "nop\n\t" - "nop\n" - : - : "r"((unsigned long)tramp & ~31), - "r"(32 /* stride */) - : "memory"); -#endif + /* Get function descriptor address for ffi_closure_pa32. */ + fd = (struct pa32_fd *)((UINT32)ffi_closure_pa32 & ~3); -#ifdef PA_HPUX - tramp[0] = 0xeaa00000; /* b,l .+8,%r21 ; %r21 <- pc+8 */ - tramp[1] = 0xd6a01c1e; /* depi 0,31,2,%r21 ; mask priv bits */ - tramp[2] = 0x4aa10038; /* ldw 28(%r21),%r1 ; load plabel */ - tramp[3] = 0x36b53ff1; /* ldo -8(%r21),%r21 ; get closure addr */ - tramp[4] = 0x0c201096; /* ldw 0(%r1),%r22 ; address of handler */ - tramp[5] = 0x02c010b4; /* ldsid (%r22),%r20 ; load space id */ - tramp[6] = 0x00141820; /* mtsp %r20,%sr0 ; into %sr0 */ - tramp[7] = 0xe2c00000; /* be 0(%sr0,%r22) ; branch to handler */ - tramp[8] = 0x0c281093; /* ldw 4(%r1),%r19 ; GP of handler */ - tramp[9] = ((UINT32)(ffi_closure_pa32) & ~2); - - /* Flush d/icache -- have to flush three lines because of alignment. */ - __asm__ volatile( - "copy %1,%0\n\t" - "fdc,m %2(%0)\n\t" - "fdc,m %2(%0)\n\t" - "fdc,m %2(%0)\n\t" - "ldsid (%1),%0\n\t" - "mtsp %0,%%sr0\n\t" - "copy %1,%0\n\t" - "fic,m %2(%%sr0,%0)\n\t" - "fic,m %2(%%sr0,%0)\n\t" - "fic,m %2(%%sr0,%0)\n\t" - "sync\n\t" - "nop\n\t" - "nop\n\t" - "nop\n\t" - "nop\n\t" - "nop\n\t" - "nop\n\t" - "nop\n" - : "=&r" ((unsigned long)tmp) - : "r" ((unsigned long)tramp & ~31), - "r" (32/* stride */) - : "memory"); -#endif + /* Setup trampoline. */ + tramp = (struct ffi_pa32_trampoline_struct *)c->tramp; + tramp->code_pointer = fd->code_pointer; + tramp->fake_gp = (UINT32)codeloc & ~3; + tramp->real_gp = fd->gp; - closure->cif = cif; - closure->user_data = user_data; - closure->fun = fun; + c->cif = cif; + c->user_data = user_data; + c->fun = fun; return FFI_OK; } diff --git a/native/libffi/src/pa/ffitarget.h b/native/libffi/src/pa/ffitarget.h index fff4c6b382..df1209eb16 100644 --- a/native/libffi/src/pa/ffitarget.h +++ b/native/libffi/src/pa/ffitarget.h @@ -68,12 +68,7 @@ typedef enum ffi_abi { #define FFI_CLOSURES 1 #define FFI_NATIVE_RAW_API 0 - -#ifdef PA_LINUX -#define FFI_TRAMPOLINE_SIZE 32 -#else -#define FFI_TRAMPOLINE_SIZE 40 -#endif +#define FFI_TRAMPOLINE_SIZE 12 #define FFI_TYPE_SMALL_STRUCT2 -1 #define FFI_TYPE_SMALL_STRUCT3 -2 diff --git a/native/libffi/src/pa/hpux32.S b/native/libffi/src/pa/hpux32.S index 40528bad75..d0e5f69529 100644 --- a/native/libffi/src/pa/hpux32.S +++ b/native/libffi/src/pa/hpux32.S @@ -259,7 +259,7 @@ L$done L$FE1 /* void ffi_closure_pa32(void); - Called with closure argument in %r21 */ + Called with closure argument in %r19 */ .SPACE $TEXT$ .SUBSPA $CODE$ @@ -285,7 +285,9 @@ L$CFI22 stw %arg2, -44(%r3) stw %arg3, -48(%r3) - copy %r21, %arg0 + /* Retrieve closure pointer and real gp. */ + copy %r19, %arg0 + ldw 8(%r19), %r19 bl ffi_closure_inner_pa32, %r2 copy %r3, %arg1 ldwm -64(%sp), %r3 diff --git a/native/libffi/src/pa/linux.S b/native/libffi/src/pa/linux.S index f11ae76807..33ef0b137a 100644 --- a/native/libffi/src/pa/linux.S +++ b/native/libffi/src/pa/linux.S @@ -252,7 +252,7 @@ ffi_call_pa32: .LFE1: /* void ffi_closure_pa32(void); - Called with closure argument in %r21 */ + Called with closure argument in %r19 */ .export ffi_closure_pa32,code .import ffi_closure_inner_pa32,code @@ -277,7 +277,9 @@ ffi_closure_pa32: stw %arg2, -44(%r3) stw %arg3, -48(%r3) - copy %r21, %arg0 + /* Retrieve closure pointer and real gp. */ + copy %r19, %arg0 + ldw 8(%r19), %r19 bl ffi_closure_inner_pa32, %r2 copy %r3, %arg1 @@ -297,10 +299,18 @@ ffi_closure_pa32: .LSCIE1: .word 0x0 ;# CIE Identifier Tag .byte 0x1 ;# CIE Version +#ifdef __PIC__ + .ascii "zR\0" ;# CIE Augmentation: 'z' - data, 'R' - DW_EH_PE_... data +#else .ascii "\0" ;# CIE Augmentation +#endif .uleb128 0x1 ;# CIE Code Alignment Factor .sleb128 4 ;# CIE Data Alignment Factor .byte 0x2 ;# CIE RA Column +#ifdef __PIC__ + .uleb128 0x1 ;# Augmentation size + .byte 0x1b ;# FDE Encoding (DW_EH_PE_pcrel|DW_EH_PE_sdata4) +#endif .byte 0xc ;# DW_CFA_def_cfa .uleb128 0x1e .uleb128 0x0 @@ -310,9 +320,15 @@ ffi_closure_pa32: .word .LEFDE1-.LASFDE1 ;# FDE Length .LASFDE1: .word .LASFDE1-.Lframe1 ;# FDE CIE offset - .word .LFB1 ;# FDE initial location +#ifdef __PIC__ + .word .LFB1-. ;# FDE initial location +#else + .word .LFB1 ;# FDE initial location +#endif .word .LFE1-.LFB1 ;# FDE address range - +#ifdef __PIC__ + .uleb128 0x0 ;# Augmentation size: no data +#endif .byte 0x4 ;# DW_CFA_advance_loc4 .word .LCFI11-.LFB1 .byte 0x83 ;# DW_CFA_offset, column 0x3 @@ -338,8 +354,15 @@ ffi_closure_pa32: .word .LEFDE2-.LASFDE2 ;# FDE Length .LASFDE2: .word .LASFDE2-.Lframe1 ;# FDE CIE offset +#ifdef __PIC__ + .word .LFB2-. ;# FDE initial location +#else .word .LFB2 ;# FDE initial location +#endif .word .LFE2-.LFB2 ;# FDE address range +#ifdef __PIC__ + .uleb128 0x0 ;# Augmentation size: no data +#endif .byte 0x4 ;# DW_CFA_advance_loc4 .word .LCFI21-.LFB2 .byte 0x83 ;# DW_CFA_offset, column 0x3 diff --git a/native/libffi/src/powerpc/aix.S b/native/libffi/src/powerpc/aix.S index 349e78c266..7ba541595f 100644 --- a/native/libffi/src/powerpc/aix.S +++ b/native/libffi/src/powerpc/aix.S @@ -106,6 +106,10 @@ ffi_call_AIX: .llong .ffi_call_AIX, TOC[tc0], 0 .csect .text[PR] .ffi_call_AIX: + .function .ffi_call_AIX,.ffi_call_AIX,16,044,LFE..0-LFB..0 + .bf __LINE__ + .line 1 +LFB..0: /* Save registers we use. */ mflr r0 @@ -115,8 +119,10 @@ ffi_call_AIX: std r31, -8(r1) std r0, 16(r1) +LCFI..0: mr r28, r1 /* our AP. */ stdux r1, r1, r4 +LCFI..1: /* Save arguments over call... */ mr r31, r5 /* flags, */ @@ -202,12 +208,16 @@ L(fp_return_value): L(float_return_value): stfs f1, 0(r30) b L(done_return_value) - +LFE..0: #else /* ! __64BIT__ */ .long .ffi_call_AIX, TOC[tc0], 0 .csect .text[PR] .ffi_call_AIX: + .function .ffi_call_AIX,.ffi_call_AIX,16,044,LFE..0-LFB..0 + .bf __LINE__ + .line 1 +LFB..0: /* Save registers we use. */ mflr r0 @@ -217,8 +227,10 @@ L(float_return_value): stw r31, -4(r1) stw r0, 8(r1) +LCFI..0: mr r28, r1 /* out AP. */ stwux r1, r1, r4 +LCFI..1: /* Save arguments over call... */ mr r31, r5 /* flags, */ @@ -304,11 +316,144 @@ L(fp_return_value): L(float_return_value): stfs f1, 0(r30) b L(done_return_value) +LFE..0: #endif + .ef __LINE__ .long 0 .byte 0,0,0,1,128,4,0,0 /* END(ffi_call_AIX) */ + /* void ffi_call_go_AIX(extended_cif *ecif, unsigned long bytes, + * unsigned int flags, unsigned int *rvalue, + * void (*fn)(), + * void (*prep_args)(extended_cif*, unsigned *const), + * void *closure); + * r3=ecif, r4=bytes, r5=flags, r6=rvalue, r7=fn, r8=prep_args, r9=closure + */ + +.csect .text[PR] + .align 2 + .globl ffi_call_go_AIX + .globl .ffi_call_go_AIX +.csect ffi_call_go_AIX[DS] +ffi_call_go_AIX: +#ifdef __64BIT__ + .llong .ffi_call_go_AIX, TOC[tc0], 0 + .csect .text[PR] +.ffi_call_go_AIX: + .function .ffi_call_go_AIX,.ffi_call_go_AIX,16,044,LFE..1-LFB..1 + .bf __LINE__ + .line 1 +LFB..1: + /* Save registers we use. */ + mflr r0 + + std r28,-32(r1) + std r29,-24(r1) + std r30,-16(r1) + std r31, -8(r1) + + std r9, 8(r1) /* closure, saved in cr field. */ + std r0, 16(r1) +LCFI..2: + mr r28, r1 /* our AP. */ + stdux r1, r1, r4 +LCFI..3: + + /* Save arguments over call... */ + mr r31, r5 /* flags, */ + mr r30, r6 /* rvalue, */ + mr r29, r7 /* function address, */ + std r2, 40(r1) + + /* Call ffi_prep_args. */ + mr r4, r1 + bl .ffi_prep_args + nop + + /* Now do the call. */ + ld r0, 0(r29) + ld r2, 8(r29) + ld r11, 8(r28) /* closure */ + /* Set up cr1 with bits 4-7 of the flags. */ + mtcrf 0x40, r31 + mtctr r0 + /* Load all those argument registers. */ + /* We have set up a nice stack frame, just load it into registers. */ + ld r3, 40+(1*8)(r1) + ld r4, 40+(2*8)(r1) + ld r5, 40+(3*8)(r1) + ld r6, 40+(4*8)(r1) + nop + ld r7, 40+(5*8)(r1) + ld r8, 40+(6*8)(r1) + ld r9, 40+(7*8)(r1) + ld r10,40+(8*8)(r1) + + b L1 +LFE..1: +#else /* ! __64BIT__ */ + + .long .ffi_call_go_AIX, TOC[tc0], 0 + .csect .text[PR] +.ffi_call_go_AIX: + .function .ffi_call_go_AIX,.ffi_call_go_AIX,16,044,LFE..1-LFB..1 + .bf __LINE__ + .line 1 + /* Save registers we use. */ +LFB..1: + mflr r0 + + stw r28,-16(r1) + stw r29,-12(r1) + stw r30, -8(r1) + stw r31, -4(r1) + + stw r9, 4(r1) /* closure, saved in cr field. */ + stw r0, 8(r1) +LCFI..2: + mr r28, r1 /* out AP. */ + stwux r1, r1, r4 +LCFI..3: + + /* Save arguments over call... */ + mr r31, r5 /* flags, */ + mr r30, r6 /* rvalue, */ + mr r29, r7 /* function address, */ + stw r2, 20(r1) + + /* Call ffi_prep_args. */ + mr r4, r1 + bl .ffi_prep_args + nop + + /* Now do the call. */ + lwz r0, 0(r29) + lwz r2, 4(r29) + lwz r11, 4(r28) /* closure */ + /* Set up cr1 with bits 4-7 of the flags. */ + mtcrf 0x40, r31 + mtctr r0 + /* Load all those argument registers. */ + /* We have set up a nice stack frame, just load it into registers. */ + lwz r3, 20+(1*4)(r1) + lwz r4, 20+(2*4)(r1) + lwz r5, 20+(3*4)(r1) + lwz r6, 20+(4*4)(r1) + nop + lwz r7, 20+(5*4)(r1) + lwz r8, 20+(6*4)(r1) + lwz r9, 20+(7*4)(r1) + lwz r10,20+(8*4)(r1) + + b L1 +LFE..1: +#endif + .ef __LINE__ + .long 0 + .byte 0,0,0,1,128,4,0,0 +/* END(ffi_call_go_AIX) */ + .csect .text[PR] .align 2 .globl ffi_call_DARWIN @@ -326,3 +471,96 @@ ffi_call_DARWIN: .long 0 .byte 0,0,0,0,0,0,0,0 /* END(ffi_call_DARWIN) */ + +/* EH frame stuff. */ + +#define LR_REGNO 0x41 /* Link Register (65), see rs6000.md */ +#ifdef __64BIT__ +#define PTRSIZE 8 +#define LOG2_PTRSIZE 3 +#define FDE_ENCODING 0x1c /* DW_EH_PE_pcrel|DW_EH_PE_sdata8 */ +#define EH_DATA_ALIGN_FACT 0x78 /* LEB128 -8 */ +#else +#define PTRSIZE 4 +#define LOG2_PTRSIZE 2 +#define FDE_ENCODING 0x1b /* DW_EH_PE_pcrel|DW_EH_PE_sdata4 */ +#define EH_DATA_ALIGN_FACT 0x7c /* LEB128 -4 */ +#endif + .csect _unwind.ro_[RO],4 + .align LOG2_PTRSIZE + .globl _GLOBAL__F_libffi_src_powerpc_aix +_GLOBAL__F_libffi_src_powerpc_aix: +Lframe..1: + .vbyte 4,LECIE..1-LSCIE..1 /* CIE Length */ +LSCIE..1: + .vbyte 4,0 /* CIE Identifier Tag */ + .byte 0x3 /* CIE Version */ + .byte "zR" /* CIE Augmentation */ + .byte 0 + .byte 0x1 /* uleb128 0x1; CIE Code Alignment Factor */ + .byte EH_DATA_ALIGN_FACT /* leb128 -4/-8; CIE Data Alignment Factor */ + .byte 0x41 /* CIE RA Column */ + .byte 0x1 /* uleb128 0x1; Augmentation size */ + .byte FDE_ENCODING /* FDE Encoding (pcrel|sdata4/8) */ + .byte 0xc /* DW_CFA_def_cfa */ + .byte 0x1 /* uleb128 0x1; Register r1 */ + .byte 0 /* uleb128 0x0; Offset 0 */ + .align LOG2_PTRSIZE +LECIE..1: +LSFDE..1: + .vbyte 4,LEFDE..1-LASFDE..1 /* FDE Length */ +LASFDE..1: + .vbyte 4,LASFDE..1-Lframe..1 /* FDE CIE offset */ + .vbyte PTRSIZE,LFB..0-$ /* FDE initial location */ + .vbyte PTRSIZE,LFE..0-LFB..0 /* FDE address range */ + .byte 0 /* uleb128 0x0; Augmentation size */ + .byte 0x4 /* DW_CFA_advance_loc4 */ + .vbyte 4,LCFI..0-LFB..0 + .byte 0x11 /* DW_CFA_def_offset_extended_sf */ + .byte LR_REGNO /* uleb128 LR_REGNO; Register LR */ + .byte 0x7e /* leb128 -2; Offset -2 (8/16) */ + .byte 0x9f /* DW_CFA_offset Register r31 */ + .byte 0x1 /* uleb128 0x1; Offset 1 (-4/-8) */ + .byte 0x9e /* DW_CFA_offset Register r30 */ + .byte 0x2 /* uleb128 0x2; Offset 2 (-8/-16) */ + .byte 0x9d /* DW_CFA_offset Register r29 */ + .byte 0x3 /* uleb128 0x3; Offset 3 (-12/-24) */ + .byte 0x9c /* DW_CFA_offset Register r28 */ + .byte 0x4 /* uleb128 0x4; Offset 4 (-16/-32) */ + .byte 0x4 /* DW_CFA_advance_loc4 */ + .vbyte 4,LCFI..1-LCFI..0 + .byte 0xd /* DW_CFA_def_cfa_register */ + .byte 0x1c /* uleb128 28; Register r28 */ + .align LOG2_PTRSIZE +LEFDE..1: +LSFDE..2: + .vbyte 4,LEFDE..2-LASFDE..2 /* FDE Length */ +LASFDE..2: + .vbyte 4,LASFDE..2-Lframe..1 /* FDE CIE offset */ + .vbyte PTRSIZE,LFB..1-$ /* FDE initial location */ + .vbyte PTRSIZE,LFE..1-LFB..1 /* FDE address range */ + .byte 0 /* uleb128 0x0; Augmentation size */ + .byte 0x4 /* DW_CFA_advance_loc4 */ + .vbyte 4,LCFI..2-LFB..1 + .byte 0x11 /* DW_CFA_def_offset_extended_sf */ + .byte LR_REGNO /* uleb128 LR_REGNO; Register LR */ + .byte 0x7e /* leb128 -2; Offset -2 (8/16) */ + .byte 0x9f /* DW_CFA_offset Register r31 */ + .byte 0x1 /* uleb128 0x1; Offset 1 (-4/-8) */ + .byte 0x9e /* DW_CFA_offset Register r30 */ + .byte 0x2 /* uleb128 0x2; Offset 2 (-8/-16) */ + .byte 0x9d /* DW_CFA_offset Register r29 */ + .byte 0x3 /* uleb128 0x3; Offset 3 (-12/-24) */ + .byte 0x9c /* DW_CFA_offset Register r28 */ + .byte 0x4 /* uleb128 0x4; Offset 4 (-16/-32) */ + .byte 0x4 /* DW_CFA_advance_loc4 */ + .vbyte 4,LCFI..3-LCFI..2 + .byte 0xd /* DW_CFA_def_cfa_register */ + .byte 0x1c /* uleb128 28; Register r28 */ + .align LOG2_PTRSIZE +LEFDE..2: + .vbyte 4,0 /* End of FDEs */ + + .csect .text[PR] + .ref _GLOBAL__F_libffi_src_powerpc_aix /* Prevents garbage collection by AIX linker */ + diff --git a/native/libffi/src/powerpc/aix_closure.S b/native/libffi/src/powerpc/aix_closure.S index aabd3c3c1e..132c785edd 100644 --- a/native/libffi/src/powerpc/aix_closure.S +++ b/native/libffi/src/powerpc/aix_closure.S @@ -80,6 +80,7 @@ .set f21,21 .extern .ffi_closure_helper_DARWIN + .extern .ffi_go_closure_helper_DARWIN #define LIBFFI_ASM #define JUMPTARGET(name) name @@ -101,6 +102,10 @@ ffi_closure_ASM: .llong .ffi_closure_ASM, TOC[tc0], 0 .csect .text[PR] .ffi_closure_ASM: + .function .ffi_closure_ASM,.ffi_closure_ASM,16,044,LFE..0-LFB..0 + .bf __LINE__ + .line 1 +LFB..0: /* we want to build up an area for the parameters passed */ /* in registers (both floating point and integer) */ @@ -117,8 +122,7 @@ ffi_closure_ASM: std r9, 48+(6*8)(r1) std r10, 48+(7*8)(r1) std r0, 16(r1) /* save the return address */ - - +LCFI..0: /* 48 Bytes (Linkage Area) */ /* 64 Bytes (params) */ /* 16 Bytes (result) */ @@ -128,6 +132,7 @@ ffi_closure_ASM: stdu r1, -240(r1) /* skip over caller save area keep stack aligned to 16 */ +LCFI..1: /* next save fpr 1 to fpr 13 (aligned to 8) */ stfd f1, 128+(0*8)(r1) @@ -161,6 +166,8 @@ ffi_closure_ASM: bl .ffi_closure_helper_DARWIN nop +.Ldoneclosure: + /* now r3 contains the return type */ /* so use it to look up in a table */ /* so we know how to deal with each type */ @@ -270,12 +277,17 @@ L..finish: mtlr r0 addi r1, r1, 240 blr +LFE..0: #else /* ! __64BIT__ */ .long .ffi_closure_ASM, TOC[tc0], 0 .csect .text[PR] .ffi_closure_ASM: + .function .ffi_closure_ASM,.ffi_closure_ASM,16,044,LFE..0-LFB..0 + .bf __LINE__ + .line 1 +LFB..0: /* we want to build up an area for the parameters passed */ /* in registers (both floating point and integer) */ @@ -292,7 +304,7 @@ L..finish: stw r9, 24+(6*4)(r1) stw r10, 24+(7*4)(r1) stw r0, 8(r1) - +LCFI..0: /* 24 Bytes (Linkage Area) */ /* 32 Bytes (params) */ /* 16 Bytes (result) */ @@ -301,6 +313,7 @@ L..finish: stwu r1, -176(r1) /* skip over caller save area keep stack aligned to 16 */ +LCFI..1: /* next save fpr 1 to fpr 13 (aligned to 8) */ stfd f1, 72+(0*8)(r1) @@ -334,6 +347,8 @@ L..finish: bl .ffi_closure_helper_DARWIN nop +.Ldoneclosure: + /* now r3 contains the return type */ /* so use it to look up in a table */ /* so we know how to deal with each type */ @@ -443,5 +458,237 @@ L..60: L..finish: addi r1, r1, 176 blr +LFE..0: #endif + .ef __LINE__ /* END(ffi_closure_ASM) */ + + +.csect .text[PR] + .align 2 + .globl ffi_go_closure_ASM + .globl .ffi_go_closure_ASM +.csect ffi_go_closure_ASM[DS] +ffi_go_closure_ASM: +#ifdef __64BIT__ + .llong .ffi_go_closure_ASM, TOC[tc0], 0 + .csect .text[PR] +.ffi_go_closure_ASM: + .function .ffi_go_closure_ASM,.ffi_go_closure_ASM,16,044,LFE..1-LFB..1 + .bf __LINE__ + .line 1 +LFB..1: +/* we want to build up an area for the parameters passed */ +/* in registers (both floating point and integer) */ + + /* we store gpr 3 to gpr 10 (aligned to 4) + in the parents outgoing area */ + std r3, 48+(0*8)(r1) + std r4, 48+(1*8)(r1) + std r5, 48+(2*8)(r1) + std r6, 48+(3*8)(r1) + mflr r0 + + std r7, 48+(4*8)(r1) + std r8, 48+(5*8)(r1) + std r9, 48+(6*8)(r1) + std r10, 48+(7*8)(r1) + std r0, 16(r1) /* save the return address */ +LCFI..2: + /* 48 Bytes (Linkage Area) */ + /* 64 Bytes (params) */ + /* 16 Bytes (result) */ + /* 104 Bytes (13*8 from FPR) */ + /* 8 Bytes (alignment) */ + /* 240 Bytes */ + + stdu r1, -240(r1) /* skip over caller save area + keep stack aligned to 16 */ +LCFI..3: + + /* next save fpr 1 to fpr 13 (aligned to 8) */ + stfd f1, 128+(0*8)(r1) + stfd f2, 128+(1*8)(r1) + stfd f3, 128+(2*8)(r1) + stfd f4, 128+(3*8)(r1) + stfd f5, 128+(4*8)(r1) + stfd f6, 128+(5*8)(r1) + stfd f7, 128+(6*8)(r1) + stfd f8, 128+(7*8)(r1) + stfd f9, 128+(8*8)(r1) + stfd f10, 128+(9*8)(r1) + stfd f11, 128+(10*8)(r1) + stfd f12, 128+(11*8)(r1) + stfd f13, 128+(12*8)(r1) + + /* set up registers for the routine that actually does the work */ + mr r3, r11 /* go closure */ + + /* now load up the pointer to the result storage */ + addi r4, r1, 112 + + /* now load up the pointer to the saved gpr registers */ + addi r5, r1, 288 + + /* now load up the pointer to the saved fpr registers */ + addi r6, r1, 128 + + /* make the call */ + bl .ffi_go_closure_helper_DARWIN + nop + + b .Ldoneclosure +LFE..1: + +#else /* ! __64BIT__ */ + + .long .ffi_go_closure_ASM, TOC[tc0], 0 + .csect .text[PR] +.ffi_go_closure_ASM: + .function .ffi_go_closure_ASM,.ffi_go_closure_ASM,16,044,LFE..1-LFB..1 + .bf __LINE__ + .line 1 +LFB..1: +/* we want to build up an area for the parameters passed */ +/* in registers (both floating point and integer) */ + + /* we store gpr 3 to gpr 10 (aligned to 4) + in the parents outgoing area */ + stw r3, 24+(0*4)(r1) + stw r4, 24+(1*4)(r1) + stw r5, 24+(2*4)(r1) + stw r6, 24+(3*4)(r1) + mflr r0 + + stw r7, 24+(4*4)(r1) + stw r8, 24+(5*4)(r1) + stw r9, 24+(6*4)(r1) + stw r10, 24+(7*4)(r1) + stw r0, 8(r1) +LCFI..2: + /* 24 Bytes (Linkage Area) */ + /* 32 Bytes (params) */ + /* 16 Bytes (result) */ + /* 104 Bytes (13*8 from FPR) */ + /* 176 Bytes */ + + stwu r1, -176(r1) /* skip over caller save area + keep stack aligned to 16 */ +LCFI..3: + + /* next save fpr 1 to fpr 13 (aligned to 8) */ + stfd f1, 72+(0*8)(r1) + stfd f2, 72+(1*8)(r1) + stfd f3, 72+(2*8)(r1) + stfd f4, 72+(3*8)(r1) + stfd f5, 72+(4*8)(r1) + stfd f6, 72+(5*8)(r1) + stfd f7, 72+(6*8)(r1) + stfd f8, 72+(7*8)(r1) + stfd f9, 72+(8*8)(r1) + stfd f10, 72+(9*8)(r1) + stfd f11, 72+(10*8)(r1) + stfd f12, 72+(11*8)(r1) + stfd f13, 72+(12*8)(r1) + + /* set up registers for the routine that actually does the work */ + mr r3, 11 /* go closure */ + + /* now load up the pointer to the result storage */ + addi r4, r1, 56 + + /* now load up the pointer to the saved gpr registers */ + addi r5, r1, 200 + + /* now load up the pointer to the saved fpr registers */ + addi r6, r1, 72 + + /* make the call */ + bl .ffi_go_closure_helper_DARWIN + nop + + b .Ldoneclosure +LFE..1: +#endif + .ef __LINE__ +/* END(ffi_go_closure_ASM) */ + +/* EH frame stuff. */ + +#define LR_REGNO 0x41 /* Link Register (65), see rs6000.md */ +#ifdef __64BIT__ +#define PTRSIZE 8 +#define LOG2_PTRSIZE 3 +#define CFA_OFFSET 0xf0,0x01 /* LEB128 240 */ +#define FDE_ENCODING 0x1c /* DW_EH_PE_pcrel|DW_EH_PE_sdata8 */ +#define EH_DATA_ALIGN_FACT 0x78 /* LEB128 -8 */ +#else +#define PTRSIZE 4 +#define LOG2_PTRSIZE 2 +#define CFA_OFFSET 0xb0,0x01 /* LEB128 176 */ +#define FDE_ENCODING 0x1b /* DW_EH_PE_pcrel|DW_EH_PE_sdata4 */ +#define EH_DATA_ALIGN_FACT 0x7c /* LEB128 -4 */ +#endif + + .csect _unwind.ro_[RO],4 + .align LOG2_PTRSIZE + .globl _GLOBAL__F_libffi_src_powerpc_aix_closure +_GLOBAL__F_libffi_src_powerpc_aix_closure: +Lframe..1: + .vbyte 4,LECIE..1-LSCIE..1 /* CIE Length */ +LSCIE..1: + .vbyte 4,0 /* CIE Identifier Tag */ + .byte 0x3 /* CIE Version */ + .byte "zR" /* CIE Augmentation */ + .byte 0 + .byte 0x1 /* uleb128 0x1; CIE Code Alignment Factor */ + .byte EH_DATA_ALIGN_FACT /* leb128 -4/-8; CIE Data Alignment Factor */ + .byte LR_REGNO /* CIE RA Column */ + .byte 0x1 /* uleb128 0x1; Augmentation size */ + .byte FDE_ENCODING /* FDE Encoding (pcrel|sdata4/8) */ + .byte 0xc /* DW_CFA_def_cfa */ + .byte 0x1 /* uleb128 0x1; Register r1 */ + .byte 0 /* uleb128 0x0; Offset 0 */ + .align LOG2_PTRSIZE +LECIE..1: +LSFDE..1: + .vbyte 4,LEFDE..1-LASFDE..1 /* FDE Length */ +LASFDE..1: + .vbyte 4,LASFDE..1-Lframe..1 /* FDE CIE offset */ + .vbyte PTRSIZE,LFB..0-$ /* FDE initial location */ + .vbyte PTRSIZE,LFE..0-LFB..0 /* FDE address range */ + .byte 0 /* uleb128 0x0; Augmentation size */ + .byte 0x4 /* DW_CFA_advance_loc4 */ + .vbyte 4,LCFI..1-LCFI..0 + .byte 0xe /* DW_CFA_def_cfa_offset */ + .byte CFA_OFFSET /* uleb128 176/240 */ + .byte 0x4 /* DW_CFA_advance_loc4 */ + .vbyte 4,LCFI..0-LFB..0 + .byte 0x11 /* DW_CFA_offset_extended_sf */ + .byte LR_REGNO /* uleb128 LR_REGNO; Register LR */ + .byte 0x7e /* leb128 -2; Offset -2 (8/16) */ + .align LOG2_PTRSIZE +LEFDE..1: +LSFDE..2: + .vbyte 4,LEFDE..2-LASFDE..2 /* FDE Length */ +LASFDE..2: + .vbyte 4,LASFDE..2-Lframe..1 /* FDE CIE offset */ + .vbyte PTRSIZE,LFB..1-$ /* FDE initial location */ + .vbyte PTRSIZE,LFE..1-LFB..1 /* FDE address range */ + .byte 0 /* uleb128 0x0; Augmentation size */ + .byte 0x4 /* DW_CFA_advance_loc4 */ + .vbyte 4,LCFI..3-LCFI..2 + .byte 0xe /* DW_CFA_def_cfa_offset */ + .byte CFA_OFFSET /* uleb128 176/240 */ + .byte 0x4 /* DW_CFA_advance_loc4 */ + .vbyte 4,LCFI..2-LFB..1 + .byte 0x11 /* DW_CFA_offset_extended_sf */ + .byte LR_REGNO /* uleb128 LR_REGNO; Register LR */ + .byte 0x7e /* leb128 -2; Offset -2 (8/16) */ + .align LOG2_PTRSIZE +LEFDE..2: + .vbyte 4,0 /* End of FDEs */ + + .csect .text[PR] + .ref _GLOBAL__F_libffi_src_powerpc_aix_closure /* Prevents garbage collection by AIX linker */ + diff --git a/native/libffi/src/powerpc/asm.h b/native/libffi/src/powerpc/asm.h index 994f62d079..27b22f670a 100644 --- a/native/libffi/src/powerpc/asm.h +++ b/native/libffi/src/powerpc/asm.h @@ -93,7 +93,7 @@ /* EALIGN is like ENTRY, but does alignment to 'words'*4 bytes past a 2^align boundary. */ #ifdef PROF -#define EALIGN(name, alignt, words) \ +#define EFFI_ALIGN(name, alignt, words) \ ASM_GLOBAL_DIRECTIVE C_SYMBOL_NAME(name); \ ASM_TYPE_DIRECTIVE (C_SYMBOL_NAME(name),@function) \ .align ALIGNARG(2); \ @@ -104,7 +104,7 @@ EALIGN_W_##words; \ 0: #else /* PROF */ -#define EALIGN(name, alignt, words) \ +#define EFFI_ALIGN(name, alignt, words) \ ASM_GLOBAL_DIRECTIVE C_SYMBOL_NAME(name); \ ASM_TYPE_DIRECTIVE (C_SYMBOL_NAME(name),@function) \ .align ALIGNARG(alignt); \ diff --git a/native/libffi/src/powerpc/darwin_closure.S b/native/libffi/src/powerpc/darwin_closure.S index c7734d4198..17456058c1 100644 --- a/native/libffi/src/powerpc/darwin_closure.S +++ b/native/libffi/src/powerpc/darwin_closure.S @@ -353,7 +353,7 @@ Lret_type13: bgt Lstructend ; not a special small case b Lsmallstruct ; see if we need more. #else - cmpi 0,r0,4 + cmpwi 0,r0,4 bgt Lfinish ; not by value lg r3,0(r5) b Lfinish @@ -494,8 +494,13 @@ LSFDE1: LASFDE1: .long LASFDE1-EH_frame1 ; FDE CIE offset .g_long Lstartcode-. ; FDE initial location +<<<<<<< HEAD .set L$set$3,LFE1-Lstartcode .g_long L$set$3 ; FDE address range +======= + .set L$set$2,LFE1-Lstartcode + .g_long L$set$2 ; FDE address range +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 .byte 0x0 ; uleb128 0x0; Augmentation size .byte 0x4 ; DW_CFA_advance_loc4 .set L$set$3,LCFI1-LCFI0 diff --git a/native/libffi/src/powerpc/ffi.c b/native/libffi/src/powerpc/ffi.c index efb441bbfc..d56ba68b8a 100644 --- a/native/libffi/src/powerpc/ffi.c +++ b/native/libffi/src/powerpc/ffi.c @@ -70,8 +70,17 @@ ffi_prep_cif_machdep_var (ffi_cif *cif, #endif } +<<<<<<< HEAD void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +======= +static void +ffi_call_int (ffi_cif *cif, + void (*fn) (void), + void *rvalue, + void **avalue, + void *closure) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 { /* The final SYSV ABI says that structures smaller or equal 8 bytes are returned in r3/r4. A draft ABI used by linux instead returns @@ -81,8 +90,14 @@ ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) can write r3 and r4 to memory without worrying about struct size. For ELFv2 ABI, use a bounce buffer for homogeneous structs too, +<<<<<<< HEAD for similar reasons. */ unsigned long smst_buffer[8]; +======= + for similar reasons. This bounce buffer must be aligned to 16 + bytes for use with homogeneous structs of vectors (float128). */ + float128 smst_buffer[8]; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 extended_cif ecif; ecif.cif = cif; @@ -97,9 +112,16 @@ ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) ecif.rvalue = alloca (cif->rtype->size); #ifdef POWERPC64 +<<<<<<< HEAD ffi_call_LINUX64 (&ecif, -(long) cif->bytes, cif->flags, ecif.rvalue, fn); #else ffi_call_SYSV (&ecif, -cif->bytes, cif->flags, ecif.rvalue, fn); +======= + ffi_call_LINUX64 (&ecif, fn, ecif.rvalue, cif->flags, closure, + -(long) cif->bytes); +#else + ffi_call_SYSV (&ecif, fn, ecif.rvalue, cif->flags, closure, -cif->bytes); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #endif /* Check for a bounce-buffered return value */ @@ -116,8 +138,14 @@ ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) # endif /* The SYSV ABI returns a structure of up to 8 bytes in size left-padded in r3/r4, and the ELFv2 ABI similarly returns a +<<<<<<< HEAD structure of up to 8 bytes in size left-padded in r3. */ if (rsize <= 8) +======= + structure of up to 8 bytes in size left-padded in r3. But + note that a structure of a single float is not paddded. */ + if (rsize <= 8 && (cif->flags & FLAG_RETURNS_FP) == 0) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 memcpy (rvalue, (char *) smst_buffer + 8 - rsize, rsize); else #endif @@ -125,7 +153,22 @@ ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) } } +void +ffi_call (ffi_cif *cif, void (*fn) (void), void *rvalue, void **avalue) +{ + ffi_call_int (cif, fn, rvalue, avalue, NULL); +} + +<<<<<<< HEAD +======= +void +ffi_call_go (ffi_cif *cif, void (*fn) (void), void *rvalue, void **avalue, + void *closure) +{ + ffi_call_int (cif, fn, rvalue, avalue, closure); +} +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ffi_status ffi_prep_closure_loc (ffi_closure *closure, ffi_cif *cif, @@ -138,4 +181,22 @@ ffi_prep_closure_loc (ffi_closure *closure, #else return ffi_prep_closure_loc_sysv (closure, cif, fun, user_data, codeloc); #endif +<<<<<<< HEAD +======= +} + +ffi_status +ffi_prep_go_closure (ffi_go_closure *closure, + ffi_cif *cif, + void (*fun) (ffi_cif *, void *, void **, void *)) +{ +#ifdef POWERPC64 + closure->tramp = ffi_go_closure_linux64; +#else + closure->tramp = ffi_go_closure_sysv; +#endif + closure->cif = cif; + closure->fun = fun; + return FFI_OK; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } diff --git a/native/libffi/src/powerpc/ffi_darwin.c b/native/libffi/src/powerpc/ffi_darwin.c index cf6fb6d4b7..64bb94dfae 100644 --- a/native/libffi/src/powerpc/ffi_darwin.c +++ b/native/libffi/src/powerpc/ffi_darwin.c @@ -34,6 +34,10 @@ extern void ffi_closure_ASM (void); +#if defined (FFI_GO_CLOSURES) +extern void ffi_go_closure_ASM (void); +#endif + enum { /* The assembly depends on these exact flags. For Darwin64 (when FLAG_RETURNS_STRUCT is set): @@ -255,7 +259,7 @@ ffi_prep_args (extended_cif *ecif, unsigned long *const stack) case FFI_TYPE_STRUCT: size_al = (*ptr)->size; #if defined(POWERPC_DARWIN64) - next_arg = (unsigned long *)ALIGN((char *)next_arg, (*ptr)->alignment); + next_arg = (unsigned long *)FFI_ALIGN((char *)next_arg, (*ptr)->alignment); darwin64_pass_struct_by_value (*ptr, (char *) *p_argv, (unsigned) size_al, (unsigned int *) &fparg_count, @@ -266,7 +270,7 @@ ffi_prep_args (extended_cif *ecif, unsigned long *const stack) /* If the first member of the struct is a double, then include enough padding in the struct size to align it to double-word. */ if ((*ptr)->elements[0]->type == FFI_TYPE_DOUBLE) - size_al = ALIGN((*ptr)->size, 8); + size_al = FFI_ALIGN((*ptr)->size, 8); # if defined(POWERPC64) FFI_ASSERT (abi != FFI_DARWIN); @@ -352,7 +356,7 @@ darwin64_struct_size_exceeds_gprs_p (ffi_type *s, char *src, unsigned *nfpr) ffi_type *p = s->elements[i]; /* Find the start of this item (0 for the first one). */ if (i > 0) - struct_offset = ALIGN(struct_offset, p->alignment); + struct_offset = FFI_ALIGN(struct_offset, p->alignment); item_base = src + struct_offset; @@ -436,7 +440,7 @@ darwin64_pass_struct_floats (ffi_type *s, char *src, ffi_type *p = s->elements[i]; /* Find the start of this item (0 for the first one). */ if (i > 0) - struct_offset = ALIGN(struct_offset, p->alignment); + struct_offset = FFI_ALIGN(struct_offset, p->alignment); item_base = src + struct_offset; switch (p->type) @@ -527,7 +531,7 @@ darwin64_struct_floats_to_mem (ffi_type *s, char *dest, double *fprs, unsigned * ffi_type *p = s->elements[i]; /* Find the start of this item (0 for the first one). */ if (i > 0) - struct_offset = ALIGN(struct_offset, p->alignment); + struct_offset = FFI_ALIGN(struct_offset, p->alignment); item_base = dest + struct_offset; switch (p->type) @@ -604,10 +608,10 @@ darwin_adjust_aggregate_sizes (ffi_type *s) align = 4; #endif /* Pad, if necessary, before adding the current item. */ - s->size = ALIGN(s->size, align) + p->size; + s->size = FFI_ALIGN(s->size, align) + p->size; } - s->size = ALIGN(s->size, s->alignment); + s->size = FFI_ALIGN(s->size, s->alignment); /* This should not be necessary on m64, but harmless. */ if (s->elements[0]->type == FFI_TYPE_UINT64 @@ -640,10 +644,10 @@ aix_adjust_aggregate_sizes (ffi_type *s) align = p->alignment; if (i != 0 && p->type == FFI_TYPE_DOUBLE) align = 4; - s->size = ALIGN(s->size, align) + p->size; + s->size = FFI_ALIGN(s->size, align) + p->size; } - s->size = ALIGN(s->size, s->alignment); + s->size = FFI_ALIGN(s->size, s->alignment); if (s->elements[0]->type == FFI_TYPE_UINT64 || s->elements[0]->type == FFI_TYPE_SINT64 @@ -809,9 +813,9 @@ ffi_prep_cif_machdep (ffi_cif *cif) 16-byte-aligned. */ if (fparg_count >= NUM_FPR_ARG_REGISTERS) #if defined (POWERPC64) - intarg_count = ALIGN(intarg_count, 2); + intarg_count = FFI_ALIGN(intarg_count, 2); #else - intarg_count = ALIGN(intarg_count, 4); + intarg_count = FFI_ALIGN(intarg_count, 4); #endif break; #endif @@ -838,7 +842,7 @@ ffi_prep_cif_machdep (ffi_cif *cif) #if defined(POWERPC_DARWIN64) align_words = (*ptr)->alignment >> 3; if (align_words) - intarg_count = ALIGN(intarg_count, align_words); + intarg_count = FFI_ALIGN(intarg_count, align_words); /* Base size of the struct. */ intarg_count += (size_al + 7) / 8; /* If 16 bytes then don't worry about floats. */ @@ -848,11 +852,11 @@ ffi_prep_cif_machdep (ffi_cif *cif) #else align_words = (*ptr)->alignment >> 2; if (align_words) - intarg_count = ALIGN(intarg_count, align_words); + intarg_count = FFI_ALIGN(intarg_count, align_words); /* If the first member of the struct is a double, then align the struct to double-word. if ((*ptr)->elements[0]->type == FFI_TYPE_DOUBLE) - size_al = ALIGN((*ptr)->size, 8); */ + size_al = FFI_ALIGN((*ptr)->size, 8); */ # ifdef POWERPC64 intarg_count += (size_al + 7) / 8; # else @@ -897,7 +901,7 @@ ffi_prep_cif_machdep (ffi_cif *cif) bytes += NUM_GPR_ARG_REGISTERS * sizeof(long); /* The stack space allocated needs to be a multiple of 16 bytes. */ - bytes = ALIGN(bytes, 16) ; + bytes = FFI_ALIGN(bytes, 16) ; cif->flags = flags; cif->bytes = bytes; @@ -908,6 +912,11 @@ ffi_prep_cif_machdep (ffi_cif *cif) extern void ffi_call_AIX(extended_cif *, long, unsigned, unsigned *, void (*fn)(void), void (*fn2)(void)); +#if defined (FFI_GO_CLOSURES) +extern void ffi_call_go_AIX(extended_cif *, long, unsigned, unsigned *, + void (*fn)(void), void (*fn2)(void), void *closure); +#endif + extern void ffi_call_DARWIN(extended_cif *, long, unsigned, unsigned *, void (*fn)(void), void (*fn2)(void), ffi_type*); @@ -946,6 +955,40 @@ ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) } } +#if defined (FFI_GO_CLOSURES) +void +ffi_call_go (ffi_cif *cif, void (*fn) (void), void *rvalue, void **avalue, + void *closure) +{ + extended_cif ecif; + + ecif.cif = cif; + ecif.avalue = avalue; + + /* If the return value is a struct and we don't have a return + value address then we need to make one. */ + + if ((rvalue == NULL) && + (cif->rtype->type == FFI_TYPE_STRUCT)) + { + ecif.rvalue = alloca (cif->rtype->size); + } + else + ecif.rvalue = rvalue; + + switch (cif->abi) + { + case FFI_AIX: + ffi_call_go_AIX(&ecif, -(long)cif->bytes, cif->flags, ecif.rvalue, fn, + FFI_FN(ffi_prep_args), closure); + break; + default: + FFI_ASSERT(0); + break; + } +} +#endif + static void flush_icache(char *); static void flush_range(char *, int); @@ -1074,6 +1117,32 @@ ffi_prep_closure_loc (ffi_closure* closure, return FFI_OK; } +#if defined (FFI_GO_CLOSURES) +ffi_status +ffi_prep_go_closure (ffi_go_closure* closure, + ffi_cif* cif, + void (*fun)(ffi_cif*, void*, void**, void*)) +{ + switch (cif->abi) + { + case FFI_AIX: + + FFI_ASSERT (cif->abi == FFI_AIX); + + closure->tramp = (void *)ffi_go_closure_ASM; + closure->cif = cif; + closure->fun = fun; + return FFI_OK; + + // For now, ffi_prep_go_closure is only implemented for AIX, not for Darwin + default: + return FFI_BAD_ABI; + break; + } + return FFI_OK; +} +#endif + static void flush_icache(char *addr) { @@ -1108,6 +1177,12 @@ ffi_type * ffi_closure_helper_DARWIN (ffi_closure *, void *, unsigned long *, ffi_dblfl *); +#if defined (FFI_GO_CLOSURES) +ffi_type * +ffi_go_closure_helper_DARWIN (ffi_go_closure*, void *, + unsigned long *, ffi_dblfl *); +#endif + /* Basically the trampoline invokes ffi_closure_ASM, and on entry, r11 holds the address of the closure. After storing the registers that could possibly contain @@ -1115,8 +1190,10 @@ ffi_closure_helper_DARWIN (ffi_closure *, void *, up space for a return value, ffi_closure_ASM invokes the following helper function to do most of the work. */ -ffi_type * -ffi_closure_helper_DARWIN (ffi_closure *closure, void *rvalue, +static ffi_type * +ffi_closure_helper_common (ffi_cif* cif, + void (*fun)(ffi_cif*, void*, void**, void*), + void *user_data, void *rvalue, unsigned long *pgr, ffi_dblfl *pfr) { /* rvalue is the pointer to space for return value in closure assembly @@ -1134,14 +1211,12 @@ ffi_closure_helper_DARWIN (ffi_closure *closure, void *rvalue, void ** avalue; ffi_type ** arg_types; long i, avn; - ffi_cif * cif; ffi_dblfl * end_pfr = pfr + NUM_FPR_ARG_REGISTERS; unsigned size_al; #if defined(POWERPC_DARWIN64) unsigned fpsused = 0; #endif - cif = closure->cif; avalue = alloca (cif->nargs * sizeof(void *)); if (cif->rtype->type == FFI_TYPE_STRUCT) @@ -1208,7 +1283,7 @@ ffi_closure_helper_DARWIN (ffi_closure *closure, void *rvalue, case FFI_TYPE_STRUCT: size_al = arg_types[i]->size; #if defined(POWERPC_DARWIN64) - pgr = (unsigned long *)ALIGN((char *)pgr, arg_types[i]->alignment); + pgr = (unsigned long *)FFI_ALIGN((char *)pgr, arg_types[i]->alignment); if (size_al < 3 || size_al == 4) { avalue[i] = ((char *)pgr)+8-size_al; @@ -1233,7 +1308,7 @@ ffi_closure_helper_DARWIN (ffi_closure *closure, void *rvalue, /* If the first member of the struct is a double, then align the struct to double-word. */ if (arg_types[i]->elements[0]->type == FFI_TYPE_DOUBLE) - size_al = ALIGN(arg_types[i]->size, 8); + size_al = FFI_ALIGN(arg_types[i]->size, 8); # if defined(POWERPC64) FFI_ASSERT (cif->abi != FFI_DARWIN); avalue[i] = pgr; @@ -1352,8 +1427,26 @@ ffi_closure_helper_DARWIN (ffi_closure *closure, void *rvalue, i++; } - (closure->fun) (cif, rvalue, avalue, closure->user_data); + (fun) (cif, rvalue, avalue, user_data); /* Tell ffi_closure_ASM to perform return type promotions. */ return cif->rtype; } + +ffi_type * +ffi_closure_helper_DARWIN (ffi_closure *closure, void *rvalue, + unsigned long *pgr, ffi_dblfl *pfr) +{ + return ffi_closure_helper_common (closure->cif, closure->fun, + closure->user_data, rvalue, pgr, pfr); +} + +#if defined (FFI_GO_CLOSURES) +ffi_type * +ffi_go_closure_helper_DARWIN (ffi_go_closure *closure, void *rvalue, + unsigned long *pgr, ffi_dblfl *pfr) +{ + return ffi_closure_helper_common (closure->cif, closure->fun, + closure, rvalue, pgr, pfr); +} +#endif diff --git a/native/libffi/src/powerpc/ffi_linux64.c b/native/libffi/src/powerpc/ffi_linux64.c index b087af8c60..d97eef16de 100644 --- a/native/libffi/src/powerpc/ffi_linux64.c +++ b/native/libffi/src/powerpc/ffi_linux64.c @@ -38,7 +38,12 @@ /* About the LINUX64 ABI. */ enum { NUM_GPR_ARG_REGISTERS64 = 8, +<<<<<<< HEAD NUM_FPR_ARG_REGISTERS64 = 13 +======= + NUM_FPR_ARG_REGISTERS64 = 13, + NUM_VEC_ARG_REGISTERS64 = 12, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 }; enum { ASM_NEEDS_REGISTERS64 = 4 }; @@ -62,12 +67,41 @@ ffi_prep_types_linux64 (ffi_abi abi) #endif +<<<<<<< HEAD #if _CALL_ELF == 2 static unsigned int discover_homogeneous_aggregate (const ffi_type *t, unsigned int *elnum) { switch (t->type) { +======= +static unsigned int +discover_homogeneous_aggregate (ffi_abi abi, + const ffi_type *t, + unsigned int *elnum) +{ + switch (t->type) + { +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE + case FFI_TYPE_LONGDOUBLE: + /* 64-bit long doubles are equivalent to doubles. */ + if ((abi & FFI_LINUX_LONG_DOUBLE_128) == 0) + { + *elnum = 1; + return FFI_TYPE_DOUBLE; + } + /* IBM extended precision values use unaligned pairs + of FPRs, but according to the ABI must be considered + distinct from doubles. They are also limited to a + maximum of four members in a homogeneous aggregate. */ + else if ((abi & FFI_LINUX_LONG_DOUBLE_IEEE128) == 0) + { + *elnum = 2; + return FFI_TYPE_LONGDOUBLE; + } + /* Fall through. */ +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 case FFI_TYPE_FLOAT: case FFI_TYPE_DOUBLE: *elnum = 1; @@ -80,14 +114,28 @@ discover_homogeneous_aggregate (const ffi_type *t, unsigned int *elnum) while (*el) { unsigned int el_elt, el_elnum = 0; +<<<<<<< HEAD el_elt = discover_homogeneous_aggregate (*el, &el_elnum); +======= + el_elt = discover_homogeneous_aggregate (abi, *el, &el_elnum); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 if (el_elt == 0 || (base_elt && base_elt != el_elt)) return 0; base_elt = el_elt; total_elnum += el_elnum; +<<<<<<< HEAD + if (total_elnum > 8) + return 0; +======= +#if _CALL_ELF == 2 if (total_elnum > 8) return 0; +#else + if (total_elnum > 1) + return 0; +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 el++; } *elnum = total_elnum; @@ -98,7 +146,10 @@ discover_homogeneous_aggregate (const ffi_type *t, unsigned int *elnum) return 0; } } +<<<<<<< HEAD #endif +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* Perform machine dependent cif processing */ @@ -107,6 +158,7 @@ ffi_prep_cif_linux64_core (ffi_cif *cif) { ffi_type **ptr; unsigned bytes; +<<<<<<< HEAD unsigned i, fparg_count = 0, intarg_count = 0; unsigned flags = cif->flags; #if _CALL_ELF == 2 @@ -116,6 +168,25 @@ ffi_prep_cif_linux64_core (ffi_cif *cif) #if FFI_TYPE_LONGDOUBLE == FFI_TYPE_DOUBLE /* If compiled without long double support.. */ if ((cif->abi & FFI_LINUX_LONG_DOUBLE_128) != 0) +======= + unsigned i, fparg_count = 0, intarg_count = 0, vecarg_count = 0; + unsigned flags = cif->flags; + unsigned elt, elnum, rtype; + +#if FFI_TYPE_LONGDOUBLE == FFI_TYPE_DOUBLE + /* If compiled without long double support... */ + if ((cif->abi & FFI_LINUX_LONG_DOUBLE_128) != 0 || + (cif->abi & FFI_LINUX_LONG_DOUBLE_IEEE128) != 0) + return FFI_BAD_ABI; +#elif !defined(__VEC__) + /* If compiled without vector register support (used by assembly)... */ + if ((cif->abi & FFI_LINUX_LONG_DOUBLE_IEEE128) != 0) + return FFI_BAD_ABI; +#else + /* If the IEEE128 flag is set, but long double is only 64 bits wide... */ + if ((cif->abi & FFI_LINUX_LONG_DOUBLE_128) == 0 && + (cif->abi & FFI_LINUX_LONG_DOUBLE_IEEE128) != 0) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 return FFI_BAD_ABI; #endif @@ -137,10 +208,26 @@ ffi_prep_cif_linux64_core (ffi_cif *cif) #endif /* Return value handling. */ +<<<<<<< HEAD switch (cif->rtype->type) { #if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE case FFI_TYPE_LONGDOUBLE: +======= + rtype = cif->rtype->type; +#if _CALL_ELF == 2 +homogeneous: +#endif + switch (rtype) + { +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE + case FFI_TYPE_LONGDOUBLE: + if ((cif->abi & FFI_LINUX_LONG_DOUBLE_IEEE128) != 0) + { + flags |= FLAG_RETURNS_VEC; + break; + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 if ((cif->abi & FFI_LINUX_LONG_DOUBLE_128) != 0) flags |= FLAG_RETURNS_128BITS; /* Fall through. */ @@ -157,11 +244,16 @@ ffi_prep_cif_linux64_core (ffi_cif *cif) /* Fall through. */ case FFI_TYPE_UINT64: case FFI_TYPE_SINT64: +<<<<<<< HEAD +======= + case FFI_TYPE_POINTER: +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 flags |= FLAG_RETURNS_64BITS; break; case FFI_TYPE_STRUCT: #if _CALL_ELF == 2 +<<<<<<< HEAD elt = discover_homogeneous_aggregate (cif->rtype, &elnum); if (elt) { @@ -175,6 +267,20 @@ ffi_prep_cif_linux64_core (ffi_cif *cif) flags |= FLAG_RETURNS_SMST; break; } +======= + elt = discover_homogeneous_aggregate (cif->abi, cif->rtype, &elnum); + if (elt) + { + flags |= FLAG_RETURNS_SMST; + rtype = elt; + goto homogeneous; + } + if (cif->rtype->size <= 16) + { + flags |= FLAG_RETURNS_SMST; + break; + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #endif intarg_count++; flags |= FLAG_RETVAL_REFERENCE; @@ -196,6 +302,18 @@ ffi_prep_cif_linux64_core (ffi_cif *cif) { #if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE case FFI_TYPE_LONGDOUBLE: +<<<<<<< HEAD +======= + if ((cif->abi & FFI_LINUX_LONG_DOUBLE_IEEE128) != 0) + { + vecarg_count++; + /* Align to 16 bytes, plus the 16-byte argument. */ + intarg_count = (intarg_count + 3) & ~0x1; + if (vecarg_count > NUM_VEC_ARG_REGISTERS64) + flags |= FLAG_ARG_NEEDS_PSAVE; + break; + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 if ((cif->abi & FFI_LINUX_LONG_DOUBLE_128) != 0) { fparg_count++; @@ -219,11 +337,29 @@ ffi_prep_cif_linux64_core (ffi_cif *cif) align = 16; align = align / 8; if (align > 1) +<<<<<<< HEAD intarg_count = ALIGN (intarg_count, align); } intarg_count += ((*ptr)->size + 7) / 8; #if _CALL_ELF == 2 elt = discover_homogeneous_aggregate (*ptr, &elnum); +======= + intarg_count = FFI_ALIGN (intarg_count, align); + } + intarg_count += ((*ptr)->size + 7) / 8; + elt = discover_homogeneous_aggregate (cif->abi, *ptr, &elnum); +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE + if (elt == FFI_TYPE_LONGDOUBLE && + (cif->abi & FFI_LINUX_LONG_DOUBLE_IEEE128) != 0) + { + vecarg_count += elnum; + if (vecarg_count > NUM_VEC_ARG_REGISTERS64) + flags |= FLAG_ARG_NEEDS_PSAVE; + break; + } + else +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 if (elt) { fparg_count += elnum; @@ -231,7 +367,10 @@ ffi_prep_cif_linux64_core (ffi_cif *cif) flags |= FLAG_ARG_NEEDS_PSAVE; } else +<<<<<<< HEAD #endif +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 { if (intarg_count > NUM_GPR_ARG_REGISTERS64) flags |= FLAG_ARG_NEEDS_PSAVE; @@ -263,10 +402,23 @@ ffi_prep_cif_linux64_core (ffi_cif *cif) flags |= FLAG_FP_ARGUMENTS; if (intarg_count > 4) flags |= FLAG_4_GPR_ARGUMENTS; +<<<<<<< HEAD +======= + if (vecarg_count != 0) + flags |= FLAG_VEC_ARGUMENTS; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* Space for the FPR registers, if needed. */ if (fparg_count != 0) bytes += NUM_FPR_ARG_REGISTERS64 * sizeof (double); +<<<<<<< HEAD +======= + /* Space for the vector registers, if needed, aligned to 16 bytes. */ + if (vecarg_count != 0) { + bytes = (bytes + 15) & ~0xF; + bytes += NUM_VEC_ARG_REGISTERS64 * sizeof (float128); + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* Stack space. */ #if _CALL_ELF == 2 @@ -349,6 +501,11 @@ ffi_prep_cif_linux64_var (ffi_cif *cif, |--------------------------------------------| | | FPR registers f1-f13 (optional) 13*8 | | |--------------------------------------------| | +<<<<<<< HEAD +======= + | VEC registers v2-v13 (optional) 12*16 | | + |--------------------------------------------| | +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 | Parameter save area | | |--------------------------------------------| | | TOC save area 8 | | @@ -378,6 +535,10 @@ ffi_prep_args64 (extended_cif *ecif, unsigned long *const stack) unsigned long *ul; float *f; double *d; +<<<<<<< HEAD +======= + float128 *f128; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 size_t p; } valp; @@ -391,11 +552,23 @@ ffi_prep_args64 (extended_cif *ecif, unsigned long *const stack) valp rest; valp next_arg; +<<<<<<< HEAD /* 'fpr_base' points at the space for fpr3, and grows upwards as +======= + /* 'fpr_base' points at the space for f1, and grows upwards as +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 we use FPR registers. */ valp fpr_base; unsigned int fparg_count; +<<<<<<< HEAD +======= + /* 'vec_base' points at the space for v2, and grows upwards as + we use vector registers. */ + valp vec_base; + unsigned int vecarg_count; + +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 unsigned int i, words, nargs, nfixedargs; ffi_type **ptr; double double_tmp; @@ -412,6 +585,10 @@ ffi_prep_args64 (extended_cif *ecif, unsigned long *const stack) unsigned long **ul; float **f; double **d; +<<<<<<< HEAD +======= + float128 **f128; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } p_argv; unsigned long gprvalue; unsigned long align; @@ -426,11 +603,27 @@ ffi_prep_args64 (extended_cif *ecif, unsigned long *const stack) #endif fpr_base.d = gpr_base.d - NUM_FPR_ARG_REGISTERS64; fparg_count = 0; +<<<<<<< HEAD +======= + /* Place the vector args below the FPRs, if used, else the GPRs. */ + if (ecif->cif->flags & FLAG_FP_ARGUMENTS) + vec_base.p = fpr_base.p & ~0xF; + else + vec_base.p = gpr_base.p; + vec_base.f128 -= NUM_VEC_ARG_REGISTERS64; + vecarg_count = 0; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 next_arg.ul = gpr_base.ul; /* Check that everything starts aligned properly. */ FFI_ASSERT (((unsigned long) (char *) stack & 0xF) == 0); FFI_ASSERT (((unsigned long) stacktop.c & 0xF) == 0); +<<<<<<< HEAD +======= + FFI_ASSERT (((unsigned long) gpr_base.c & 0xF) == 0); + FFI_ASSERT (((unsigned long) gpr_end.c & 0xF) == 0); + FFI_ASSERT (((unsigned long) vec_base.c & 0xF) == 0); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 FFI_ASSERT ((bytes & 0xF) == 0); /* Deal with return values that are actually pass-by-reference. */ @@ -449,14 +642,37 @@ ffi_prep_args64 (extended_cif *ecif, unsigned long *const stack) i < nargs; i++, ptr++, p_argv.v++) { +<<<<<<< HEAD #if _CALL_ELF == 2 unsigned int elt, elnum; #endif +======= + unsigned int elt, elnum; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 switch ((*ptr)->type) { #if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE case FFI_TYPE_LONGDOUBLE: +<<<<<<< HEAD +======= + if ((ecif->cif->abi & FFI_LINUX_LONG_DOUBLE_IEEE128) != 0) + { + next_arg.p = FFI_ALIGN (next_arg.p, 16); + if (next_arg.ul == gpr_end.ul) + next_arg.ul = rest.ul; + if (vecarg_count < NUM_VEC_ARG_REGISTERS64 && i < nfixedargs) + memcpy (vec_base.f128++, *p_argv.f128, sizeof (float128)); + else + memcpy (next_arg.f128, *p_argv.f128, sizeof (float128)); + if (++next_arg.f128 == gpr_end.f128) + next_arg.f128 = rest.f128; + vecarg_count++; + FFI_ASSERT (__LDBL_MANT_DIG__ == 113); + FFI_ASSERT (flags & FLAG_VEC_ARGUMENTS); + break; + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 if ((ecif->cif->abi & FFI_LINUX_LONG_DOUBLE_128) != 0) { double_tmp = (*p_argv.d)[0]; @@ -494,6 +710,12 @@ ffi_prep_args64 (extended_cif *ecif, unsigned long *const stack) /* Fall through. */ #endif case FFI_TYPE_DOUBLE: +<<<<<<< HEAD +======= +#if _CALL_ELF != 2 + do_double: +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 double_tmp = **p_argv.d; if (fparg_count < NUM_FPR_ARG_REGISTERS64 && i < nfixedargs) { @@ -512,17 +734,43 @@ ffi_prep_args64 (extended_cif *ecif, unsigned long *const stack) break; case FFI_TYPE_FLOAT: +<<<<<<< HEAD +======= +#if _CALL_ELF != 2 + do_float: +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 double_tmp = **p_argv.f; if (fparg_count < NUM_FPR_ARG_REGISTERS64 && i < nfixedargs) { *fpr_base.d++ = double_tmp; #if _CALL_ELF != 2 if ((flags & FLAG_COMPAT) != 0) +<<<<<<< HEAD *next_arg.f = (float) double_tmp; #endif } else *next_arg.f = (float) double_tmp; +======= + { +# ifndef __LITTLE_ENDIAN__ + next_arg.f[1] = (float) double_tmp; +# else + next_arg.f[0] = (float) double_tmp; +# endif + } +#endif + } + else + { +# ifndef __LITTLE_ENDIAN__ + next_arg.f[1] = (float) double_tmp; +# else + next_arg.f[0] = (float) double_tmp; +# endif + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 if (++next_arg.ul == gpr_end.ul) next_arg.ul = rest.ul; fparg_count++; @@ -536,19 +784,58 @@ ffi_prep_args64 (extended_cif *ecif, unsigned long *const stack) if (align > 16) align = 16; if (align > 1) +<<<<<<< HEAD next_arg.p = ALIGN (next_arg.p, align); } #if _CALL_ELF == 2 elt = discover_homogeneous_aggregate (*ptr, &elnum); if (elt) { +======= + { + next_arg.p = FFI_ALIGN (next_arg.p, align); + if (next_arg.ul == gpr_end.ul) + next_arg.ul = rest.ul; + } + } + elt = discover_homogeneous_aggregate (ecif->cif->abi, *ptr, &elnum); + if (elt) + { +#if _CALL_ELF == 2 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 union { void *v; float *f; double *d; +<<<<<<< HEAD } arg; arg.v = *p_argv.v; +======= + float128 *f128; + } arg; + + arg.v = *p_argv.v; +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE + if (elt == FFI_TYPE_LONGDOUBLE && + (ecif->cif->abi & FFI_LINUX_LONG_DOUBLE_IEEE128) != 0) + { + do + { + if (vecarg_count < NUM_VEC_ARG_REGISTERS64 + && i < nfixedargs) + memcpy (vec_base.f128++, arg.f128, sizeof (float128)); + else + memcpy (next_arg.f128, arg.f128++, sizeof (float128)); + if (++next_arg.f128 == gpr_end.f128) + next_arg.f128 = rest.f128; + vecarg_count++; + } + while (--elnum != 0); + } + else +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 if (elt == FFI_TYPE_FLOAT) { do @@ -564,11 +851,17 @@ ffi_prep_args64 (extended_cif *ecif, unsigned long *const stack) fparg_count++; } while (--elnum != 0); +<<<<<<< HEAD if ((next_arg.p & 3) != 0) { if (++next_arg.f == gpr_end.f) next_arg.f = rest.f; } +======= + if ((next_arg.p & 7) != 0) + if (++next_arg.f == gpr_end.f) + next_arg.f = rest.f; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } else do @@ -583,9 +876,20 @@ ffi_prep_args64 (extended_cif *ecif, unsigned long *const stack) fparg_count++; } while (--elnum != 0); +<<<<<<< HEAD } else #endif +======= +#else + if (elt == FFI_TYPE_FLOAT) + goto do_float; + else + goto do_double; +#endif + } + else +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 { words = ((*ptr)->size + 7) / 8; if (next_arg.ul >= gpr_base.ul && next_arg.ul + words > gpr_end.ul) @@ -667,7 +971,12 @@ flush_icache (char *wraddr, char *xaddr, int size) } #endif +<<<<<<< HEAD ffi_status +======= + +ffi_status FFI_HIDDEN +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ffi_prep_closure_loc_linux64 (ffi_closure *closure, ffi_cif *cif, void (*fun) (ffi_cif *, void *, void **, void *), @@ -688,17 +997,28 @@ ffi_prep_closure_loc_linux64 (ffi_closure *closure, /* 2: .quad context */ *(void **) &tramp[4] = (void *) ffi_closure_LINUX64; *(void **) &tramp[6] = codeloc; +<<<<<<< HEAD flush_icache ((char *)tramp, (char *)codeloc, FFI_TRAMPOLINE_SIZE); +======= + flush_icache ((char *) tramp, (char *) codeloc, 4 * 4); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #else void **tramp = (void **) &closure->tramp[0]; if (cif->abi < FFI_LINUX || cif->abi >= FFI_LAST_ABI) return FFI_BAD_ABI; +<<<<<<< HEAD /* Copy function address and TOC from ffi_closure_LINUX64. */ memcpy (tramp, (char *) ffi_closure_LINUX64, 16); tramp[2] = tramp[1]; tramp[1] = codeloc; +======= + /* Copy function address and TOC from ffi_closure_LINUX64 OPD. */ + memcpy (&tramp[0], (void **) ffi_closure_LINUX64, sizeof (void *)); + tramp[1] = codeloc; + memcpy (&tramp[2], (void **) ffi_closure_LINUX64 + 1, sizeof (void *)); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #endif closure->cif = cif; @@ -710,22 +1030,43 @@ ffi_prep_closure_loc_linux64 (ffi_closure *closure, int FFI_HIDDEN +<<<<<<< HEAD ffi_closure_helper_LINUX64 (ffi_closure *closure, void *rvalue, unsigned long *pst, ffi_dblfl *pfr) +======= +ffi_closure_helper_LINUX64 (ffi_cif *cif, + void (*fun) (ffi_cif *, void *, void **, void *), + void *user_data, + void *rvalue, + unsigned long *pst, + ffi_dblfl *pfr, + float128 *pvec) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 { /* rvalue is the pointer to space for return value in closure assembly */ /* pst is the pointer to parameter save area (r3-r10 are stored into its first 8 slots by ffi_closure_LINUX64) */ /* pfr is the pointer to where f1-f13 are stored in ffi_closure_LINUX64 */ +<<<<<<< HEAD +======= + /* pvec is the pointer to where v2-v13 are stored in ffi_closure_LINUX64 */ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 void **avalue; ffi_type **arg_types; unsigned long i, avn, nfixedargs; +<<<<<<< HEAD ffi_cif *cif; ffi_dblfl *end_pfr = pfr + NUM_FPR_ARG_REGISTERS64; unsigned long align; cif = closure->cif; +======= + ffi_dblfl *end_pfr = pfr + NUM_FPR_ARG_REGISTERS64; + float128 *end_pvec = pvec + NUM_VEC_ARG_REGISTERS64; + unsigned long align; + +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 avalue = alloca (cif->nargs * sizeof (void *)); /* Copy the caller's structure return value address so that the @@ -791,6 +1132,7 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure, void *rvalue, if (align > 16) align = 16; if (align > 1) +<<<<<<< HEAD pst = (unsigned long *) ALIGN ((size_t) pst, align); } elt = 0; @@ -799,11 +1141,23 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure, void *rvalue, #endif if (elt) { +======= + pst = (unsigned long *) FFI_ALIGN ((size_t) pst, align); + } + elt = discover_homogeneous_aggregate (cif->abi, arg_types[i], &elnum); + if (elt) + { +#if _CALL_ELF == 2 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 union { void *v; unsigned long *ul; float *f; double *d; +<<<<<<< HEAD +======= + float128 *f128; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 size_t p; } to, from; @@ -811,6 +1165,20 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure, void *rvalue, aggregate size is not greater than the space taken by the registers so store back to the register/parameter save arrays. */ +<<<<<<< HEAD +======= +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE + if (elt == FFI_TYPE_LONGDOUBLE && + (cif->abi & FFI_LINUX_LONG_DOUBLE_IEEE128) != 0) + { + if (pvec + elnum <= end_pvec) + to.v = pvec; + else + to.v = pst; + } + else +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 if (pfr + elnum <= end_pfr) to.v = pfr; else @@ -818,6 +1186,26 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure, void *rvalue, avalue[i] = to.v; from.ul = pst; +<<<<<<< HEAD +======= +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE + if (elt == FFI_TYPE_LONGDOUBLE && + (cif->abi & FFI_LINUX_LONG_DOUBLE_IEEE128) != 0) + { + do + { + if (pvec < end_pvec && i < nfixedargs) + memcpy (to.f128, pvec++, sizeof (float128)); + else + memcpy (to.f128, from.f128, sizeof (float128)); + to.f128++; + from.f128++; + } + while (--elnum != 0); + } + else +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 if (elt == FFI_TYPE_FLOAT) { do @@ -850,6 +1238,15 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure, void *rvalue, } while (--elnum != 0); } +<<<<<<< HEAD +======= +#else + if (elt == FFI_TYPE_FLOAT) + goto do_float; + else + goto do_double; +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } else { @@ -867,7 +1264,22 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure, void *rvalue, #if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE case FFI_TYPE_LONGDOUBLE: +<<<<<<< HEAD if ((cif->abi & FFI_LINUX_LONG_DOUBLE_128) != 0) +======= + if ((cif->abi & FFI_LINUX_LONG_DOUBLE_IEEE128) != 0) + { + if (((unsigned long) pst & 0xF) != 0) + ++pst; + if (pvec < end_pvec && i < nfixedargs) + avalue[i] = pvec++; + else + avalue[i] = pst; + pst += 2; + break; + } + else if ((cif->abi & FFI_LINUX_LONG_DOUBLE_128) != 0) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 { if (pfr + 1 < end_pfr && i + 1 < nfixedargs) { @@ -891,6 +1303,12 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure, void *rvalue, /* Fall through. */ #endif case FFI_TYPE_DOUBLE: +<<<<<<< HEAD +======= +#if _CALL_ELF != 2 + do_double: +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* On the outgoing stack all values are aligned to 8 */ /* there are 13 64bit floating point registers */ @@ -905,6 +1323,12 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure, void *rvalue, break; case FFI_TYPE_FLOAT: +<<<<<<< HEAD +======= +#if _CALL_ELF != 2 + do_float: +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 if (pfr < end_pfr && i < nfixedargs) { /* Float values are stored as doubles in the @@ -914,7 +1338,17 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure, void *rvalue, pfr++; } else +<<<<<<< HEAD avalue[i] = pst; +======= + { +#ifndef __LITTLE_ENDIAN__ + avalue[i] = (char *) pst + 4; +#else + avalue[i] = pst; +#endif + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 pst++; break; @@ -925,19 +1359,35 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure, void *rvalue, i++; } +<<<<<<< HEAD (closure->fun) (cif, rvalue, avalue, closure->user_data); +======= + (*fun) (cif, rvalue, avalue, user_data); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* Tell ffi_closure_LINUX64 how to perform return type promotions. */ if ((cif->flags & FLAG_RETURNS_SMST) != 0) { +<<<<<<< HEAD if ((cif->flags & FLAG_RETURNS_FP) == 0) return FFI_V2_TYPE_SMALL_STRUCT + cif->rtype->size - 1; +======= + if ((cif->flags & (FLAG_RETURNS_FP | FLAG_RETURNS_VEC)) == 0) + return FFI_V2_TYPE_SMALL_STRUCT + cif->rtype->size - 1; + else if ((cif->flags & FLAG_RETURNS_VEC) != 0) + return FFI_V2_TYPE_VECTOR_HOMOG; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 else if ((cif->flags & FLAG_RETURNS_64BITS) != 0) return FFI_V2_TYPE_DOUBLE_HOMOG; else return FFI_V2_TYPE_FLOAT_HOMOG; } +<<<<<<< HEAD +======= + if ((cif->flags & FLAG_RETURNS_VEC) != 0) + return FFI_V2_TYPE_VECTOR; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 return cif->rtype->type; } #endif diff --git a/native/libffi/src/powerpc/ffi_powerpc.h b/native/libffi/src/powerpc/ffi_powerpc.h index 2e61653d1a..96dd6d3f56 100644 --- a/native/libffi/src/powerpc/ffi_powerpc.h +++ b/native/libffi/src/powerpc/ffi_powerpc.h @@ -31,6 +31,7 @@ enum { /* The assembly depends on these exact flags. */ /* These go in cr7 */ +<<<<<<< HEAD FLAG_RETURNS_SMST = 1 << (31-31), /* Used for FFI_SYSV small structs. */ FLAG_RETURNS_NOTHING = 1 << (31-30), FLAG_RETURNS_FP = 1 << (31-29), @@ -40,13 +41,30 @@ enum { FLAG_RETURNS_128BITS = 1 << (31-27), FLAG_COMPAT = 1 << (31- 8), /* Not used by assembly */ +======= + FLAG_RETURNS_SMST = 1 << (31-31), /* Used for FFI_SYSV small structs. */ + FLAG_RETURNS_NOTHING = 1 << (31-30), + FLAG_RETURNS_FP = 1 << (31-29), + FLAG_RETURNS_VEC = 1 << (31-28), + + /* These go in cr6 */ + FLAG_RETURNS_64BITS = 1 << (31-27), + FLAG_RETURNS_128BITS = 1 << (31-26), + + FLAG_COMPAT = 1 << (31- 8), /* Not used by assembly */ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* These go in cr1 */ FLAG_ARG_NEEDS_COPY = 1 << (31- 7), /* Used by sysv code */ FLAG_ARG_NEEDS_PSAVE = FLAG_ARG_NEEDS_COPY, /* Used by linux64 code */ FLAG_FP_ARGUMENTS = 1 << (31- 6), /* cr1.eq; specified by ABI */ FLAG_4_GPR_ARGUMENTS = 1 << (31- 5), +<<<<<<< HEAD FLAG_RETVAL_REFERENCE = 1 << (31- 4) +======= + FLAG_RETVAL_REFERENCE = 1 << (31- 4), + FLAG_VEC_ARGUMENTS = 1 << (31- 3), +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 }; typedef union @@ -55,6 +73,7 @@ typedef union double d; } ffi_dblfl; +<<<<<<< HEAD void FFI_HIDDEN ffi_closure_SYSV (void); void FFI_HIDDEN ffi_call_SYSV(extended_cif *, unsigned, unsigned, unsigned *, void (*)(void)); @@ -67,11 +86,56 @@ int FFI_HIDDEN ffi_closure_helper_SYSV (ffi_closure *, void *, unsigned long *, void FFI_HIDDEN ffi_call_LINUX64(extended_cif *, unsigned long, unsigned long, unsigned long *, void (*)(void)); void FFI_HIDDEN ffi_closure_LINUX64 (void); +======= +#if defined(__FLOAT128_TYPE__) && defined(__HAVE_FLOAT128) +typedef _Float128 float128; +#elif defined(__FLOAT128__) +typedef __float128 float128; +#else +typedef char float128[16] __attribute__((aligned(16))); +#endif + +void FFI_HIDDEN ffi_closure_SYSV (void); +void FFI_HIDDEN ffi_go_closure_sysv (void); +void FFI_HIDDEN ffi_call_SYSV(extended_cif *, void (*)(void), void *, + unsigned, void *, int); + +void FFI_HIDDEN ffi_prep_types_sysv (ffi_abi); +ffi_status FFI_HIDDEN ffi_prep_cif_sysv (ffi_cif *); +ffi_status FFI_HIDDEN ffi_prep_closure_loc_sysv (ffi_closure *, + ffi_cif *, + void (*) (ffi_cif *, void *, + void **, void *), + void *, void *); +int FFI_HIDDEN ffi_closure_helper_SYSV (ffi_cif *, + void (*) (ffi_cif *, void *, + void **, void *), + void *, void *, unsigned long *, + ffi_dblfl *, unsigned long *); + +void FFI_HIDDEN ffi_call_LINUX64(extended_cif *, void (*) (void), void *, + unsigned long, void *, long); +void FFI_HIDDEN ffi_closure_LINUX64 (void); +void FFI_HIDDEN ffi_go_closure_linux64 (void); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 void FFI_HIDDEN ffi_prep_types_linux64 (ffi_abi); ffi_status FFI_HIDDEN ffi_prep_cif_linux64 (ffi_cif *); ffi_status FFI_HIDDEN ffi_prep_cif_linux64_var (ffi_cif *, unsigned int, unsigned int); void FFI_HIDDEN ffi_prep_args64 (extended_cif *, unsigned long *const); +<<<<<<< HEAD int FFI_HIDDEN ffi_closure_helper_LINUX64 (ffi_closure *, void *, unsigned long *, ffi_dblfl *); +======= +ffi_status FFI_HIDDEN ffi_prep_closure_loc_linux64 (ffi_closure *, ffi_cif *, + void (*) (ffi_cif *, void *, + void **, void *), + void *, void *); +int FFI_HIDDEN ffi_closure_helper_LINUX64 (ffi_cif *, + void (*) (ffi_cif *, void *, + void **, void *), + void *, void *, + unsigned long *, ffi_dblfl *, + float128 *); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 diff --git a/native/libffi/src/powerpc/ffi_sysv.c b/native/libffi/src/powerpc/ffi_sysv.c index fbe85fe914..38e269acc7 100644 --- a/native/libffi/src/powerpc/ffi_sysv.c +++ b/native/libffi/src/powerpc/ffi_sysv.c @@ -36,7 +36,11 @@ /* About the SYSV ABI. */ +<<<<<<< HEAD #define ASM_NEEDS_REGISTERS 4 +======= +#define ASM_NEEDS_REGISTERS 6 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #define NUM_GPR_ARG_REGISTERS 8 #define NUM_FPR_ARG_REGISTERS 8 @@ -93,7 +97,11 @@ ffi_prep_cif_sysv_core (ffi_cif *cif) { ffi_type **ptr; unsigned bytes; +<<<<<<< HEAD unsigned i, fparg_count = 0, intarg_count = 0; +======= + unsigned i, fpr_count = 0, gpr_count = 0, stack_count = 0; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 unsigned flags = cif->flags; unsigned struct_copy_size = 0; unsigned type = cif->rtype->type; @@ -155,7 +163,11 @@ ffi_prep_cif_sysv_core (ffi_cif *cif) flags |= FLAG_RETURNS_SMST; break; } +<<<<<<< HEAD intarg_count++; +======= + gpr_count++; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 flags |= FLAG_RETVAL_REFERENCE; /* Fall through. */ case FFI_TYPE_VOID: @@ -182,6 +194,7 @@ ffi_prep_cif_sysv_core (ffi_cif *cif) { #if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE case FFI_TYPE_LONGDOUBLE: +<<<<<<< HEAD fparg_count++; /* Fall thru */ #endif @@ -193,13 +206,47 @@ ffi_prep_cif_sysv_core (ffi_cif *cif) && intarg_count >= NUM_GPR_ARG_REGISTERS && intarg_count % 2 != 0) intarg_count++; +======= + if (fpr_count >= NUM_FPR_ARG_REGISTERS - 1) + { + fpr_count = NUM_FPR_ARG_REGISTERS; + /* 8-byte align long doubles. */ + stack_count += stack_count & 1; + stack_count += 4; + } + else + fpr_count += 2; +#ifdef __NO_FPRS__ + return FFI_BAD_ABI; +#endif + break; +#endif + + case FFI_TYPE_DOUBLE: + if (fpr_count >= NUM_FPR_ARG_REGISTERS) + { + /* 8-byte align doubles. */ + stack_count += stack_count & 1; + stack_count += 2; + } + else + fpr_count += 1; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #ifdef __NO_FPRS__ return FFI_BAD_ABI; #endif break; case FFI_TYPE_FLOAT: +<<<<<<< HEAD fparg_count++; +======= + if (fpr_count >= NUM_FPR_ARG_REGISTERS) + /* Yes, we don't follow the ABI, but neither does gcc. */ + stack_count += 1; + else + fpr_count += 1; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #ifdef __NO_FPRS__ return FFI_BAD_ABI; #endif @@ -208,11 +255,21 @@ ffi_prep_cif_sysv_core (ffi_cif *cif) case FFI_TYPE_UINT128: /* A long double in FFI_LINUX_SOFT_FLOAT can use only a set of four consecutive gprs. If we do not have enough, we +<<<<<<< HEAD have to adjust the intarg_count value. */ if (intarg_count >= NUM_GPR_ARG_REGISTERS - 3 && intarg_count < NUM_GPR_ARG_REGISTERS) intarg_count = NUM_GPR_ARG_REGISTERS; intarg_count += 4; +======= + have to adjust the gpr_count value. */ + if (gpr_count >= NUM_GPR_ARG_REGISTERS - 3) + gpr_count = NUM_GPR_ARG_REGISTERS; + if (gpr_count >= NUM_GPR_ARG_REGISTERS) + stack_count += 4; + else + gpr_count += 4; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 break; case FFI_TYPE_UINT64: @@ -225,10 +282,21 @@ ffi_prep_cif_sysv_core (ffi_cif *cif) Also, only certain register pairs can be used for passing long long int -- specifically (r3,r4), (r5,r6), (r7,r8), (r9,r10). */ +<<<<<<< HEAD if (intarg_count == NUM_GPR_ARG_REGISTERS-1 || intarg_count % 2 != 0) intarg_count++; intarg_count += 2; +======= + gpr_count += gpr_count & 1; + if (gpr_count >= NUM_GPR_ARG_REGISTERS) + { + stack_count += stack_count & 1; + stack_count += 2; + } + else + gpr_count += 2; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 break; case FFI_TYPE_STRUCT: @@ -249,7 +317,14 @@ ffi_prep_cif_sysv_core (ffi_cif *cif) case FFI_TYPE_SINT8: /* Everything else is passed as a 4-byte word in a GPR, either the object itself or a pointer to it. */ +<<<<<<< HEAD intarg_count++; +======= + if (gpr_count >= NUM_GPR_ARG_REGISTERS) + stack_count += 1; + else + gpr_count += 1; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 break; default: @@ -257,14 +332,21 @@ ffi_prep_cif_sysv_core (ffi_cif *cif) } } +<<<<<<< HEAD if (fparg_count != 0) flags |= FLAG_FP_ARGUMENTS; if (intarg_count > 4) +======= + if (fpr_count != 0) + flags |= FLAG_FP_ARGUMENTS; + if (gpr_count > 4) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 flags |= FLAG_4_GPR_ARGUMENTS; if (struct_copy_size != 0) flags |= FLAG_ARG_NEEDS_COPY; /* Space for the FPR registers, if needed. */ +<<<<<<< HEAD if (fparg_count != 0) bytes += NUM_FPR_ARG_REGISTERS * sizeof (double); @@ -273,6 +355,13 @@ ffi_prep_cif_sysv_core (ffi_cif *cif) bytes += (intarg_count - NUM_GPR_ARG_REGISTERS) * sizeof (int); if (fparg_count > NUM_FPR_ARG_REGISTERS) bytes += (fparg_count - NUM_FPR_ARG_REGISTERS) * sizeof (double); +======= + if (fpr_count != 0) + bytes += NUM_FPR_ARG_REGISTERS * sizeof (double); + + /* Stack space. */ + bytes += stack_count * sizeof (int); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* The stack space allocated needs to be a multiple of 16 bytes. */ bytes = (bytes + 15) & ~0xF; @@ -367,13 +456,21 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) /* 'gpr_base' points at the space for gpr3, and grows upwards as we use GPR registers. */ valp gpr_base; +<<<<<<< HEAD int intarg_count; +======= + valp gpr_end; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #ifndef __NO_FPRS__ /* 'fpr_base' points at the space for fpr1, and grows upwards as we use FPR registers. */ valp fpr_base; +<<<<<<< HEAD int fparg_count; +======= + valp fpr_end; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #endif /* 'copy_space' grows down as we put structures in it. It should @@ -405,11 +502,19 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) unsigned gprvalue; stacktop.c = (char *) stack + bytes; +<<<<<<< HEAD gpr_base.u = stacktop.u - ASM_NEEDS_REGISTERS - NUM_GPR_ARG_REGISTERS; intarg_count = 0; #ifndef __NO_FPRS__ fpr_base.d = gpr_base.d - NUM_FPR_ARG_REGISTERS; fparg_count = 0; +======= + gpr_end.u = stacktop.u - ASM_NEEDS_REGISTERS; + gpr_base.u = gpr_end.u - NUM_GPR_ARG_REGISTERS; +#ifndef __NO_FPRS__ + fpr_end.d = gpr_base.d; + fpr_base.d = fpr_end.d - NUM_FPR_ARG_REGISTERS; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 copy_space.c = ((flags & FLAG_FP_ARGUMENTS) ? fpr_base.c : gpr_base.c); #else copy_space.c = gpr_base.c; @@ -425,10 +530,14 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) /* Deal with return values that are actually pass-by-reference. */ if (flags & FLAG_RETVAL_REFERENCE) +<<<<<<< HEAD { *gpr_base.u++ = (unsigned long) (char *) ecif->rvalue; intarg_count++; } +======= + *gpr_base.u++ = (unsigned) (char *) ecif->rvalue; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* Now for the arguments. */ p_argv.v = ecif->avalue; @@ -448,6 +557,7 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) case FFI_TYPE_LONGDOUBLE: double_tmp = (*p_argv.d)[0]; +<<<<<<< HEAD if (fparg_count >= NUM_FPR_ARG_REGISTERS - 1) { if (intarg_count >= NUM_GPR_ARG_REGISTERS @@ -456,6 +566,13 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) intarg_count++; next_arg.u++; } +======= + if (fpr_base.d >= fpr_end.d - 1) + { + fpr_base.d = fpr_end.d; + if (((next_arg.u - stack) & 1) != 0) + next_arg.u += 1; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 *next_arg.d = double_tmp; next_arg.u += 2; double_tmp = (*p_argv.d)[1]; @@ -468,14 +585,18 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) double_tmp = (*p_argv.d)[1]; *fpr_base.d++ = double_tmp; } +<<<<<<< HEAD fparg_count += 2; +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 FFI_ASSERT (flags & FLAG_FP_ARGUMENTS); break; # endif case FFI_TYPE_DOUBLE: double_tmp = **p_argv.d; +<<<<<<< HEAD if (fparg_count >= NUM_FPR_ARG_REGISTERS) { if (intarg_count >= NUM_GPR_ARG_REGISTERS @@ -484,17 +605,27 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) intarg_count++; next_arg.u++; } +======= + if (fpr_base.d >= fpr_end.d) + { + if (((next_arg.u - stack) & 1) != 0) + next_arg.u += 1; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 *next_arg.d = double_tmp; next_arg.u += 2; } else *fpr_base.d++ = double_tmp; +<<<<<<< HEAD fparg_count++; +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 FFI_ASSERT (flags & FLAG_FP_ARGUMENTS); break; case FFI_TYPE_FLOAT: double_tmp = **p_argv.f; +<<<<<<< HEAD if (fparg_count >= NUM_FPR_ARG_REGISTERS) { *next_arg.f = (float) double_tmp; @@ -504,6 +635,15 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) else *fpr_base.d++ = double_tmp; fparg_count++; +======= + if (fpr_base.d >= fpr_end.d) + { + *next_arg.f = (float) double_tmp; + next_arg.u += 1; + } + else + *fpr_base.d++ = double_tmp; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 FFI_ASSERT (flags & FLAG_FP_ARGUMENTS); break; #endif /* have FPRs */ @@ -513,6 +653,7 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) is passed in four consecutive GPRs if available. A maximum of 2 long doubles can be passed in gprs. If we do not have 4 GPRs left, the long double is passed on the stack, 4-byte aligned. */ +<<<<<<< HEAD { unsigned int int_tmp; unsigned int ii; @@ -549,6 +690,36 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) intarg_count++; next_arg.u++; } +======= + if (gpr_base.u >= gpr_end.u - 3) + { + unsigned int ii; + gpr_base.u = gpr_end.u; + for (ii = 0; ii < 4; ii++) + { + unsigned int int_tmp = (*p_argv.ui)[ii]; + *next_arg.u++ = int_tmp; + } + } + else + { + unsigned int ii; + for (ii = 0; ii < 4; ii++) + { + unsigned int int_tmp = (*p_argv.ui)[ii]; + *gpr_base.u++ = int_tmp; + } + } + break; + + case FFI_TYPE_UINT64: + case FFI_TYPE_SINT64: + if (gpr_base.u >= gpr_end.u - 1) + { + gpr_base.u = gpr_end.u; + if (((next_arg.u - stack) & 1) != 0) + next_arg.u++; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 *next_arg.ll = **p_argv.ll; next_arg.u += 2; } @@ -559,6 +730,7 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) (r5,r6), (r7,r8), (r9,r10). If next arg is long long but not correct starting register of pair then skip until the proper starting register. */ +<<<<<<< HEAD if (intarg_count % 2 != 0) { intarg_count ++; @@ -567,6 +739,12 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) *gpr_base.ll++ = **p_argv.ll; } intarg_count += 2; +======= + if (((gpr_end.u - gpr_base.u) & 1) != 0) + gpr_base.u++; + *gpr_base.ll++ = **p_argv.ll; + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 break; case FFI_TYPE_STRUCT: @@ -601,17 +779,25 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) gprvalue = **p_argv.ui; putgpr: +<<<<<<< HEAD if (intarg_count >= NUM_GPR_ARG_REGISTERS) *next_arg.u++ = gprvalue; else *gpr_base.u++ = gprvalue; intarg_count++; +======= + if (gpr_base.u >= gpr_end.u) + *next_arg.u++ = gprvalue; + else + *gpr_base.u++ = gprvalue; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 break; } } /* Check that we didn't overrun the stack... */ FFI_ASSERT (copy_space.c >= next_arg.c); +<<<<<<< HEAD FFI_ASSERT (gpr_base.u <= stacktop.u - ASM_NEEDS_REGISTERS); /* The assert below is testing that the number of integer arguments agrees with the number found in ffi_prep_cif_machdep(). However, intarg_count @@ -624,6 +810,14 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *const stack) <= stacktop.u - ASM_NEEDS_REGISTERS - NUM_GPR_ARG_REGISTERS); #endif FFI_ASSERT (flags & FLAG_4_GPR_ARGUMENTS || intarg_count <= 4); +======= + FFI_ASSERT (gpr_base.u <= gpr_end.u); +#ifndef __NO_FPRS__ + FFI_ASSERT (fpr_base.u <= fpr_end.u); +#endif + FFI_ASSERT (((flags & FLAG_4_GPR_ARGUMENTS) != 0) + == (gpr_end.u - gpr_base.u < 4)); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } #define MIN_CACHE_LINE_SIZE 8 @@ -654,6 +848,7 @@ ffi_prep_closure_loc_sysv (ffi_closure *closure, tramp = (unsigned int *) &closure->tramp[0]; tramp[0] = 0x7c0802a6; /* mflr r0 */ +<<<<<<< HEAD tramp[1] = 0x4800000d; /* bl 10 */ tramp[4] = 0x7d6802a6; /* mflr r11 */ tramp[5] = 0x7c0803a6; /* mtlr r0 */ @@ -666,6 +861,20 @@ ffi_prep_closure_loc_sysv (ffi_closure *closure, /* Flush the icache. */ flush_icache ((char *)tramp, (char *)codeloc, FFI_TRAMPOLINE_SIZE); +======= + tramp[1] = 0x429f0005; /* bcl 20,31,.+4 */ + tramp[2] = 0x7d6802a6; /* mflr r11 */ + tramp[3] = 0x7c0803a6; /* mtlr r0 */ + tramp[4] = 0x800b0018; /* lwz r0,24(r11) */ + tramp[5] = 0x816b001c; /* lwz r11,28(r11) */ + tramp[6] = 0x7c0903a6; /* mtctr r0 */ + tramp[7] = 0x4e800420; /* bctr */ + *(void **) &tramp[8] = (void *) ffi_closure_SYSV; /* function */ + *(void **) &tramp[9] = codeloc; /* context */ + + /* Flush the icache. */ + flush_icache ((char *)tramp, (char *)codeloc, 8 * 4); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 closure->cif = cif; closure->fun = fun; @@ -682,8 +891,17 @@ ffi_prep_closure_loc_sysv (ffi_closure *closure, following helper function to do most of the work. */ int +<<<<<<< HEAD ffi_closure_helper_SYSV (ffi_closure *closure, void *rvalue, unsigned long *pgr, ffi_dblfl *pfr, +======= +ffi_closure_helper_SYSV (ffi_cif *cif, + void (*fun) (ffi_cif *, void *, void **, void *), + void *user_data, + void *rvalue, + unsigned long *pgr, + ffi_dblfl *pfr, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 unsigned long *pst) { /* rvalue is the pointer to space for return value in closure assembly */ @@ -699,7 +917,10 @@ ffi_closure_helper_SYSV (ffi_closure *closure, void *rvalue, #endif long ng = 0; /* number of general registers already used */ +<<<<<<< HEAD ffi_cif *cif = closure->cif; +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 unsigned size = cif->rtype->size; unsigned short rtypenum = cif->rtype->type; @@ -915,7 +1136,11 @@ ffi_closure_helper_SYSV (ffi_closure *closure, void *rvalue, i++; } +<<<<<<< HEAD (closure->fun) (cif, rvalue, avalue, closure->user_data); +======= + (*fun) (cif, rvalue, avalue, user_data); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* Tell ffi_closure_SYSV how to perform return type promotions. Because the FFI_SYSV ABI returns the structures <= 8 bytes in diff --git a/native/libffi/src/powerpc/ffitarget.h b/native/libffi/src/powerpc/ffitarget.h index b47b0f5d3a..4f71b2cc25 100644 --- a/native/libffi/src/powerpc/ffitarget.h +++ b/native/libffi/src/powerpc/ffitarget.h @@ -91,15 +91,28 @@ typedef enum ffi_abi { /* This and following bits can reuse FFI_COMPAT values. */ FFI_LINUX_STRUCT_ALIGN = 1, FFI_LINUX_LONG_DOUBLE_128 = 2, +<<<<<<< HEAD +======= + FFI_LINUX_LONG_DOUBLE_IEEE128 = 4, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 FFI_DEFAULT_ABI = (FFI_LINUX # ifdef __STRUCT_PARM_ALIGN__ | FFI_LINUX_STRUCT_ALIGN # endif # ifdef __LONG_DOUBLE_128__ | FFI_LINUX_LONG_DOUBLE_128 +<<<<<<< HEAD # endif ), FFI_LAST_ABI = 12 +======= +# ifdef __LONG_DOUBLE_IEEE128__ + | FFI_LINUX_LONG_DOUBLE_IEEE128 +# endif +# endif + ), + FFI_LAST_ABI = 16 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 # else /* This bit, always set in new code, must not be set in any of the @@ -138,16 +151,29 @@ typedef enum ffi_abi { #define FFI_CLOSURES 1 #define FFI_NATIVE_RAW_API 0 #if defined (POWERPC) || defined (POWERPC_FREEBSD) +<<<<<<< HEAD +# define FFI_TARGET_SPECIFIC_VARIADIC 1 +# define FFI_EXTRA_CIF_FIELDS unsigned nfixedargs +#endif +======= +# define FFI_GO_CLOSURES 1 # define FFI_TARGET_SPECIFIC_VARIADIC 1 # define FFI_EXTRA_CIF_FIELDS unsigned nfixedargs #endif +#if defined (POWERPC_AIX) +# define FFI_GO_CLOSURES 1 +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 -/* For additional types like the below, take care about the order in - ppc_closures.S. They must follow after the FFI_TYPE_LAST. */ +/* ppc_closure.S and linux64_closure.S expect this. */ +#define FFI_PPC_TYPE_LAST FFI_TYPE_POINTER -/* Needed for soft-float long-double-128 support. */ -#define FFI_TYPE_UINT128 (FFI_TYPE_LAST + 1) +/* We define additional types below. If generic types are added that + must be supported by powerpc libffi then it is likely that + FFI_PPC_TYPE_LAST needs increasing *and* the jump tables in + ppc_closure.S and linux64_closure.S be extended. */ +<<<<<<< HEAD /* Needed for FFI_SYSV small structure returns. */ #define FFI_SYSV_TYPE_SMALL_STRUCT (FFI_TYPE_LAST + 2) @@ -155,6 +181,26 @@ typedef enum ffi_abi { #define FFI_V2_TYPE_FLOAT_HOMOG (FFI_TYPE_LAST + 1) #define FFI_V2_TYPE_DOUBLE_HOMOG (FFI_TYPE_LAST + 2) #define FFI_V2_TYPE_SMALL_STRUCT (FFI_TYPE_LAST + 3) +======= +#if !(FFI_TYPE_LAST == FFI_PPC_TYPE_LAST \ + || (FFI_TYPE_LAST == FFI_TYPE_COMPLEX \ + && !defined FFI_TARGET_HAS_COMPLEX_TYPE)) +# error "You likely have a broken powerpc libffi" +#endif + +/* Needed for soft-float long-double-128 support. */ +#define FFI_TYPE_UINT128 (FFI_PPC_TYPE_LAST + 1) + +/* Needed for FFI_SYSV small structure returns. */ +#define FFI_SYSV_TYPE_SMALL_STRUCT (FFI_PPC_TYPE_LAST + 2) + +/* Used by ELFv2 for homogenous structure returns. */ +#define FFI_V2_TYPE_VECTOR (FFI_PPC_TYPE_LAST + 1) +#define FFI_V2_TYPE_VECTOR_HOMOG (FFI_PPC_TYPE_LAST + 2) +#define FFI_V2_TYPE_FLOAT_HOMOG (FFI_PPC_TYPE_LAST + 3) +#define FFI_V2_TYPE_DOUBLE_HOMOG (FFI_PPC_TYPE_LAST + 4) +#define FFI_V2_TYPE_SMALL_STRUCT (FFI_PPC_TYPE_LAST + 5) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #if _CALL_ELF == 2 # define FFI_TRAMPOLINE_SIZE 32 diff --git a/native/libffi/src/powerpc/linux64.S b/native/libffi/src/powerpc/linux64.S index d2acb7007b..3e529160c5 100644 --- a/native/libffi/src/powerpc/linux64.S +++ b/native/libffi/src/powerpc/linux64.S @@ -32,11 +32,22 @@ #ifdef POWERPC64 .hidden ffi_call_LINUX64 .globl ffi_call_LINUX64 +<<<<<<< HEAD # if _CALL_ELF == 2 .text ffi_call_LINUX64: addis %r2, %r12, .TOC.-ffi_call_LINUX64@ha addi %r2, %r2, .TOC.-ffi_call_LINUX64@l +======= + .text + .cfi_startproc +# if _CALL_ELF == 2 +ffi_call_LINUX64: +# ifndef __PCREL__ + addis %r2, %r12, .TOC.-ffi_call_LINUX64@ha + addi %r2, %r2, .TOC.-ffi_call_LINUX64@l +# endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 .localentry ffi_call_LINUX64, . - ffi_call_LINUX64 # else .section ".opd","aw" @@ -57,20 +68,38 @@ ffi_call_LINUX64: .ffi_call_LINUX64: # endif # endif +<<<<<<< HEAD .LFB1: +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 mflr %r0 std %r28, -32(%r1) std %r29, -24(%r1) std %r30, -16(%r1) std %r31, -8(%r1) + std %r7, 8(%r1) /* closure, saved in cr field. */ std %r0, 16(%r1) mr %r28, %r1 /* our AP. */ +<<<<<<< HEAD .LCFI0: stdux %r1, %r1, %r4 mr %r31, %r5 /* flags, */ mr %r30, %r6 /* rvalue, */ mr %r29, %r7 /* function address. */ +======= + .cfi_def_cfa_register 28 + .cfi_offset 65, 16 + .cfi_offset 31, -8 + .cfi_offset 30, -16 + .cfi_offset 29, -24 + .cfi_offset 28, -32 + + stdux %r1, %r1, %r8 + mr %r31, %r6 /* flags, */ + mr %r30, %r5 /* rvalue, */ + mr %r29, %r4 /* function address. */ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /* Save toc pointer, not for the ffi_prep_args64 call, but for the later bctrl function call. */ # if _CALL_ELF == 2 @@ -82,9 +111,21 @@ ffi_call_LINUX64: /* Call ffi_prep_args64. */ mr %r4, %r1 # if defined _CALL_LINUX || _CALL_ELF == 2 +<<<<<<< HEAD bl ffi_prep_args64 # else bl .ffi_prep_args64 +======= +# ifdef __PCREL__ + bl ffi_prep_args64@notoc +# else + bl ffi_prep_args64 + nop +# endif +# else + bl .ffi_prep_args64 + nop +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 # endif # if _CALL_ELF == 2 @@ -92,44 +133,78 @@ ffi_call_LINUX64: # else ld %r12, 0(%r29) ld %r2, 8(%r29) +<<<<<<< HEAD ld %r11, 16(%r29) +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 # endif /* Now do the call. */ - /* Set up cr1 with bits 4-7 of the flags. */ - mtcrf 0x40, %r31 + /* Set up cr1 with bits 3-7 of the flags. */ + mtcrf 0xc0, %r31 /* Get the address to call into CTR. */ mtctr %r12 /* Load all those argument registers. */ - ld %r3, -32-(8*8)(%r28) - ld %r4, -32-(7*8)(%r28) - ld %r5, -32-(6*8)(%r28) - ld %r6, -32-(5*8)(%r28) + addi %r29, %r28, -32-(8*8) + ld %r3, (0*8)(%r29) + ld %r4, (1*8)(%r29) + ld %r5, (2*8)(%r29) + ld %r6, (3*8)(%r29) bf- 5, 1f - ld %r7, -32-(4*8)(%r28) - ld %r8, -32-(3*8)(%r28) - ld %r9, -32-(2*8)(%r28) - ld %r10, -32-(1*8)(%r28) + ld %r7, (4*8)(%r29) + ld %r8, (5*8)(%r29) + ld %r9, (6*8)(%r29) + ld %r10, (7*8)(%r29) 1: /* Load all the FP registers. */ bf- 6, 2f - lfd %f1, -32-(21*8)(%r28) - lfd %f2, -32-(20*8)(%r28) - lfd %f3, -32-(19*8)(%r28) - lfd %f4, -32-(18*8)(%r28) - lfd %f5, -32-(17*8)(%r28) - lfd %f6, -32-(16*8)(%r28) - lfd %f7, -32-(15*8)(%r28) - lfd %f8, -32-(14*8)(%r28) - lfd %f9, -32-(13*8)(%r28) - lfd %f10, -32-(12*8)(%r28) - lfd %f11, -32-(11*8)(%r28) - lfd %f12, -32-(10*8)(%r28) - lfd %f13, -32-(9*8)(%r28) + addi %r29, %r29, -(14*8) + lfd %f1, ( 1*8)(%r29) + lfd %f2, ( 2*8)(%r29) + lfd %f3, ( 3*8)(%r29) + lfd %f4, ( 4*8)(%r29) + lfd %f5, ( 5*8)(%r29) + lfd %f6, ( 6*8)(%r29) + lfd %f7, ( 7*8)(%r29) + lfd %f8, ( 8*8)(%r29) + lfd %f9, ( 9*8)(%r29) + lfd %f10, (10*8)(%r29) + lfd %f11, (11*8)(%r29) + lfd %f12, (12*8)(%r29) + lfd %f13, (13*8)(%r29) 2: + /* Load all the vector registers. */ + bf- 3, 3f + addi %r29, %r29, -16 + lvx %v13, 0, %r29 + addi %r29, %r29, -16 + lvx %v12, 0, %r29 + addi %r29, %r29, -16 + lvx %v11, 0, %r29 + addi %r29, %r29, -16 + lvx %v10, 0, %r29 + addi %r29, %r29, -16 + lvx %v9, 0, %r29 + addi %r29, %r29, -16 + lvx %v8, 0, %r29 + addi %r29, %r29, -16 + lvx %v7, 0, %r29 + addi %r29, %r29, -16 + lvx %v6, 0, %r29 + addi %r29, %r29, -16 + lvx %v5, 0, %r29 + addi %r29, %r29, -16 + lvx %v4, 0, %r29 + addi %r29, %r29, -16 + lvx %v3, 0, %r29 + addi %r29, %r29, -16 + lvx %v2, 0, %r29 +3: + /* Make the call. */ + ld %r11, 8(%r28) bctrl /* This must follow the call immediately, the unwinder @@ -145,12 +220,17 @@ ffi_call_LINUX64: bt 31, .Lstruct_return_value bt 30, .Ldone_return_value bt 29, .Lfp_return_value +<<<<<<< HEAD +======= + bt 28, .Lvec_return_value +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 std %r3, 0(%r30) /* Fall through... */ .Ldone_return_value: /* Restore the registers we used and return. */ mr %r1, %r28 + .cfi_def_cfa_register 1 ld %r0, 16(%r28) ld %r28, -32(%r28) mtlr %r0 @@ -159,11 +239,16 @@ ffi_call_LINUX64: ld %r31, -8(%r1) blr +.Lvec_return_value: + stvx %v2, 0, %r30 + b .Ldone_return_value + .Lfp_return_value: - bf 28, .Lfloat_return_value - stfd %f1, 0(%r30) + .cfi_def_cfa_register 28 mtcrf 0x02, %r31 /* cr6 */ - bf 27, .Ldone_return_value + bf 27, .Lfloat_return_value + stfd %f1, 0(%r30) + bf 26, .Ldone_return_value stfd %f2, 8(%r30) b .Ldone_return_value .Lfloat_return_value: @@ -171,8 +256,14 @@ ffi_call_LINUX64: b .Ldone_return_value .Lstruct_return_value: +<<<<<<< HEAD bf 29, .Lsmall_struct bf 28, .Lfloat_homog_return_value +======= + bf 29, .Lvec_homog_or_small_struct + mtcrf 0x02, %r31 /* cr6 */ + bf 27, .Lfloat_homog_return_value +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 stfd %f1, 0(%r30) stfd %f2, 8(%r30) stfd %f3, 16(%r30) @@ -194,11 +285,34 @@ ffi_call_LINUX64: stfs %f8, 28(%r30) b .Ldone_return_value +<<<<<<< HEAD +======= +.Lvec_homog_or_small_struct: + bf 28, .Lsmall_struct + stvx %v2, 0, %r30 + addi %r30, %r30, 16 + stvx %v3, 0, %r30 + addi %r30, %r30, 16 + stvx %v4, 0, %r30 + addi %r30, %r30, 16 + stvx %v5, 0, %r30 + addi %r30, %r30, 16 + stvx %v6, 0, %r30 + addi %r30, %r30, 16 + stvx %v7, 0, %r30 + addi %r30, %r30, 16 + stvx %v8, 0, %r30 + addi %r30, %r30, 16 + stvx %v9, 0, %r30 + b .Ldone_return_value + +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 .Lsmall_struct: std %r3, 0(%r30) std %r4, 8(%r30) b .Ldone_return_value +<<<<<<< HEAD .LFE1: .long 0 .byte 0,12,0,1,128,4,0,0 @@ -208,10 +322,22 @@ ffi_call_LINUX64: # ifdef _CALL_LINUX .size ffi_call_LINUX64,.-.L.ffi_call_LINUX64 # else +======= + .cfi_endproc +# if _CALL_ELF == 2 + .size ffi_call_LINUX64,.-ffi_call_LINUX64 +# else +# ifdef _CALL_LINUX + .size ffi_call_LINUX64,.-.L.ffi_call_LINUX64 +# else + .long 0 + .byte 0,12,0,1,128,4,0,0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 .size .ffi_call_LINUX64,.-.ffi_call_LINUX64 # endif # endif +<<<<<<< HEAD .section .eh_frame,EH_FRAME_FLAGS,@progbits .Lframe1: .4byte .LECIE1-.LSCIE1 # Length of Common Information Entry @@ -254,6 +380,8 @@ ffi_call_LINUX64: .align 3 .LEFDE1: +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #endif #if (defined __ELF__ && defined __linux__) || _CALL_ELF == 2 diff --git a/native/libffi/src/powerpc/linux64_closure.S b/native/libffi/src/powerpc/linux64_closure.S index 97421a43c2..7f3b4a08a8 100644 --- a/native/libffi/src/powerpc/linux64_closure.S +++ b/native/libffi/src/powerpc/linux64_closure.S @@ -33,11 +33,22 @@ #ifdef POWERPC64 FFI_HIDDEN (ffi_closure_LINUX64) .globl ffi_closure_LINUX64 +<<<<<<< HEAD # if _CALL_ELF == 2 .text ffi_closure_LINUX64: addis %r2, %r12, .TOC.-ffi_closure_LINUX64@ha addi %r2, %r2, .TOC.-ffi_closure_LINUX64@l +======= + .text + .cfi_startproc +# if _CALL_ELF == 2 +ffi_closure_LINUX64: +# ifndef __PCREL__ + addis %r2, %r12, .TOC.-ffi_closure_LINUX64@ha + addi %r2, %r2, .TOC.-ffi_closure_LINUX64@l +# endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 .localentry ffi_closure_LINUX64, . - ffi_closure_LINUX64 # else .section ".opd","aw" @@ -60,9 +71,21 @@ ffi_closure_LINUX64: # endif # if _CALL_ELF == 2 +<<<<<<< HEAD # 32 byte special reg save area + 64 byte parm save area # + 64 byte retval area + 13*8 fpr save area + round to 16 # define STACKFRAME 272 +======= +# ifdef __VEC__ +# 32 byte special reg save area + 64 byte parm save area +# + 128 byte retval area + 13*8 fpr save area + 12*16 vec save area + round to 16 +# define STACKFRAME 528 +# else +# 32 byte special reg save area + 64 byte parm save area +# + 64 byte retval area + 13*8 fpr save area + round to 16 +# define STACKFRAME 272 +# endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 # define PARMSAVE 32 # define RETVAL PARMSAVE+64 # else @@ -73,6 +96,7 @@ ffi_closure_LINUX64: # define RETVAL PARMSAVE+64 # endif +<<<<<<< HEAD .LFB1: # if _CALL_ELF == 2 ld %r12, FFI_TRAMPOLINE_SIZE(%r11) # closure->cif @@ -121,7 +145,62 @@ ffi_closure_LINUX64: # load up the pointer to the parm save area addi %r5, %r1, PARMSAVE # endif +======= +# if _CALL_ELF == 2 + ld %r12, FFI_TRAMPOLINE_SIZE(%r11) # closure->cif + mflr %r0 + lwz %r12, 28(%r12) # cif->flags + mtcrf 0x40, %r12 + addi %r12, %r1, PARMSAVE + bt 7, 0f + # Our caller has not allocated a parameter save area. + # We need to allocate one here and use it to pass gprs to + # ffi_closure_helper_LINUX64. + addi %r12, %r1, -STACKFRAME+PARMSAVE +0: + # Save general regs into parm save area + std %r3, 0(%r12) + std %r4, 8(%r12) + std %r5, 16(%r12) + std %r6, 24(%r12) + std %r7, 32(%r12) + std %r8, 40(%r12) + std %r9, 48(%r12) + std %r10, 56(%r12) + + # load up the pointer to the parm save area + mr %r7, %r12 +# else + # copy r2 to r11 and load TOC into r2 + mr %r11, %r2 + ld %r2, 16(%r2) + mflr %r0 + # Save general regs into parm save area + # This is the parameter save area set up by our caller. + std %r3, PARMSAVE+0(%r1) + std %r4, PARMSAVE+8(%r1) + std %r5, PARMSAVE+16(%r1) + std %r6, PARMSAVE+24(%r1) + std %r7, PARMSAVE+32(%r1) + std %r8, PARMSAVE+40(%r1) + std %r9, PARMSAVE+48(%r1) + std %r10, PARMSAVE+56(%r1) + + # load up the pointer to the parm save area + addi %r7, %r1, PARMSAVE +# endif + std %r0, 16(%r1) + + # closure->cif + ld %r3, FFI_TRAMPOLINE_SIZE(%r11) + # closure->fun + ld %r4, FFI_TRAMPOLINE_SIZE+8(%r11) + # closure->user_data + ld %r5, FFI_TRAMPOLINE_SIZE+16(%r11) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + +.Ldoclosure: # next save fpr 1 to fpr 13 stfd %f1, -104+(0*8)(%r1) stfd %f2, -104+(1*8)(%r1) @@ -137,6 +216,7 @@ ffi_closure_LINUX64: stfd %f12, -104+(11*8)(%r1) stfd %f13, -104+(12*8)(%r1) +<<<<<<< HEAD # load up the pointer to the saved fpr registers */ addi %r6, %r1, -104 @@ -155,7 +235,62 @@ ffi_closure_LINUX64: # else bl .ffi_closure_helper_LINUX64 # endif +======= + # load up the pointer to the saved fpr registers + addi %r8, %r1, -104 + +# ifdef __VEC__ + # load up the pointer to the saved vector registers + # 8 bytes padding for 16-byte alignment at -112(%r1) + addi %r9, %r8, -24 + stvx %v13, 0, %r9 + addi %r9, %r9, -16 + stvx %v12, 0, %r9 + addi %r9, %r9, -16 + stvx %v11, 0, %r9 + addi %r9, %r9, -16 + stvx %v10, 0, %r9 + addi %r9, %r9, -16 + stvx %v9, 0, %r9 + addi %r9, %r9, -16 + stvx %v8, 0, %r9 + addi %r9, %r9, -16 + stvx %v7, 0, %r9 + addi %r9, %r9, -16 + stvx %v6, 0, %r9 + addi %r9, %r9, -16 + stvx %v5, 0, %r9 + addi %r9, %r9, -16 + stvx %v4, 0, %r9 + addi %r9, %r9, -16 + stvx %v3, 0, %r9 + addi %r9, %r9, -16 + stvx %v2, 0, %r9 +# endif + + # load up the pointer to the result storage + addi %r6, %r1, -STACKFRAME+RETVAL + + stdu %r1, -STACKFRAME(%r1) + .cfi_def_cfa_offset STACKFRAME + .cfi_offset 65, 16 + + # make the call +# if defined _CALL_LINUX || _CALL_ELF == 2 +# ifdef __PCREL__ + bl ffi_closure_helper_LINUX64@notoc .Lret: +# else + bl ffi_closure_helper_LINUX64 +.Lret: + nop +# endif +# else + bl .ffi_closure_helper_LINUX64 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 +.Lret: + nop +# endif # now r3 contains the return type # so use it to look up in a table @@ -182,7 +317,12 @@ ffi_closure_LINUX64: # case FFI_TYPE_VOID mtlr %r0 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME nop # case FFI_TYPE_INT # ifdef __LITTLE_ENDIAN__ @@ -192,17 +332,32 @@ ffi_closure_LINUX64: # endif mtlr %r0 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME # case FFI_TYPE_FLOAT lfs %f1, RETVAL+0(%r1) mtlr %r0 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME # case FFI_TYPE_DOUBLE lfd %f1, RETVAL+0(%r1) mtlr %r0 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME # case FFI_TYPE_LONGDOUBLE lfd %f1, RETVAL+0(%r1) mtlr %r0 @@ -216,7 +371,12 @@ ffi_closure_LINUX64: # endif mtlr %r0 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME # case FFI_TYPE_SINT8 # ifdef __LITTLE_ENDIAN__ lbz %r3, RETVAL+0(%r1) @@ -235,7 +395,12 @@ ffi_closure_LINUX64: mtlr %r0 .Lfinish: addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME # case FFI_TYPE_SINT16 # ifdef __LITTLE_ENDIAN__ lha %r3, RETVAL+0(%r1) @@ -244,7 +409,12 @@ ffi_closure_LINUX64: # endif mtlr %r0 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME # case FFI_TYPE_UINT32 # ifdef __LITTLE_ENDIAN__ lwz %r3, RETVAL+0(%r1) @@ -253,7 +423,12 @@ ffi_closure_LINUX64: # endif mtlr %r0 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME # case FFI_TYPE_SINT32 # ifdef __LITTLE_ENDIAN__ lwa %r3, RETVAL+0(%r1) @@ -262,27 +437,97 @@ ffi_closure_LINUX64: # endif mtlr %r0 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME # case FFI_TYPE_UINT64 ld %r3, RETVAL+0(%r1) mtlr %r0 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME # case FFI_TYPE_SINT64 ld %r3, RETVAL+0(%r1) mtlr %r0 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME # case FFI_TYPE_STRUCT mtlr %r0 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 blr + .cfi_def_cfa_offset STACKFRAME nop # case FFI_TYPE_POINTER ld %r3, RETVAL+0(%r1) +<<<<<<< HEAD + mtlr %r0 + addi %r1, %r1, STACKFRAME + blr +# case FFI_V2_TYPE_FLOAT_HOMOG + lfs %f1, RETVAL+0(%r1) + lfs %f2, RETVAL+4(%r1) + lfs %f3, RETVAL+8(%r1) + b .Lmorefloat +# case FFI_V2_TYPE_DOUBLE_HOMOG + lfd %f1, RETVAL+0(%r1) + lfd %f2, RETVAL+8(%r1) + lfd %f3, RETVAL+16(%r1) + lfd %f4, RETVAL+24(%r1) + mtlr %r0 + lfd %f5, RETVAL+32(%r1) + lfd %f6, RETVAL+40(%r1) + lfd %f7, RETVAL+48(%r1) + lfd %f8, RETVAL+56(%r1) + addi %r1, %r1, STACKFRAME + blr +.Lmorefloat: + lfs %f4, RETVAL+12(%r1) + mtlr %r0 + lfs %f5, RETVAL+16(%r1) + lfs %f6, RETVAL+20(%r1) + lfs %f7, RETVAL+24(%r1) + lfs %f8, RETVAL+28(%r1) + addi %r1, %r1, STACKFRAME + blr +.Lsmall: +# ifdef __LITTLE_ENDIAN__ + ld %r3,RETVAL+0(%r1) mtlr %r0 + ld %r4,RETVAL+8(%r1) addi %r1, %r1, STACKFRAME blr +======= + mtlr %r0 + addi %r1, %r1, STACKFRAME + .cfi_def_cfa_offset 0 + blr + .cfi_def_cfa_offset STACKFRAME +# case FFI_V2_TYPE_VECTOR + addi %r3, %r1, RETVAL + lvx %v2, 0, %r3 + mtlr %r0 + b .Lfinish +# case FFI_V2_TYPE_VECTOR_HOMOG + addi %r3, %r1, RETVAL + lvx %v2, 0, %r3 + addi %r3, %r3, 16 + b .Lmorevector # case FFI_V2_TYPE_FLOAT_HOMOG lfs %f1, RETVAL+0(%r1) lfs %f2, RETVAL+4(%r1) @@ -299,7 +544,28 @@ ffi_closure_LINUX64: lfd %f7, RETVAL+48(%r1) lfd %f8, RETVAL+56(%r1) addi %r1, %r1, STACKFRAME + .cfi_def_cfa_offset 0 + blr + .cfi_def_cfa_offset STACKFRAME +.Lmorevector: + lvx %v3, 0, %r3 + addi %r3, %r3, 16 + lvx %v4, 0, %r3 + addi %r3, %r3, 16 + lvx %v5, 0, %r3 + mtlr %r0 + addi %r3, %r3, 16 + lvx %v6, 0, %r3 + addi %r3, %r3, 16 + lvx %v7, 0, %r3 + addi %r3, %r3, 16 + lvx %v8, 0, %r3 + addi %r3, %r3, 16 + lvx %v9, 0, %r3 + addi %r1, %r1, STACKFRAME + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset STACKFRAME .Lmorefloat: lfs %f4, RETVAL+12(%r1) mtlr %r0 @@ -308,14 +574,18 @@ ffi_closure_LINUX64: lfs %f7, RETVAL+24(%r1) lfs %f8, RETVAL+28(%r1) addi %r1, %r1, STACKFRAME + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset STACKFRAME .Lsmall: # ifdef __LITTLE_ENDIAN__ ld %r3,RETVAL+0(%r1) mtlr %r0 ld %r4,RETVAL+8(%r1) addi %r1, %r1, STACKFRAME + .cfi_def_cfa_offset 0 blr +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 # else # A struct smaller than a dword is returned in the low bits of r3 # ie. right justified. Larger structs are passed left justified @@ -328,16 +598,27 @@ ffi_closure_LINUX64: mtlr %r0 ld %r4,RETVAL+8(%r1) addi %r1, %r1, STACKFRAME +<<<<<<< HEAD blr +======= + .cfi_def_cfa_offset 0 + blr + .cfi_def_cfa_offset STACKFRAME +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 .Lsmalldown: addi %r5, %r5, FFI_V2_TYPE_SMALL_STRUCT + 7 mtlr %r0 sldi %r5, %r5, 3 addi %r1, %r1, STACKFRAME +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 0 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 srd %r3, %r3, %r5 blr # endif +<<<<<<< HEAD .LFE1: .long 0 .byte 0,12,0,1,128,0,0,0 @@ -385,6 +666,115 @@ ffi_closure_LINUX64: .align 3 .LEFDE1: +======= + .cfi_endproc +# if _CALL_ELF == 2 + .size ffi_closure_LINUX64,.-ffi_closure_LINUX64 +# else +# ifdef _CALL_LINUX + .size ffi_closure_LINUX64,.-.L.ffi_closure_LINUX64 +# else + .long 0 + .byte 0,12,0,1,128,0,0,0 + .size .ffi_closure_LINUX64,.-.ffi_closure_LINUX64 +# endif +# endif + + + FFI_HIDDEN (ffi_go_closure_linux64) + .globl ffi_go_closure_linux64 + .text + .cfi_startproc +# if _CALL_ELF == 2 +ffi_go_closure_linux64: +# ifndef __PCREL__ + addis %r2, %r12, .TOC.-ffi_go_closure_linux64@ha + addi %r2, %r2, .TOC.-ffi_go_closure_linux64@l +# endif + .localentry ffi_go_closure_linux64, . - ffi_go_closure_linux64 +# else + .section ".opd","aw" + .align 3 +ffi_go_closure_linux64: +# ifdef _CALL_LINUX + .quad .L.ffi_go_closure_linux64,.TOC.@tocbase,0 + .type ffi_go_closure_linux64,@function + .text +.L.ffi_go_closure_linux64: +# else + FFI_HIDDEN (.ffi_go_closure_linux64) + .globl .ffi_go_closure_linux64 + .quad .ffi_go_closure_linux64,.TOC.@tocbase,0 + .size ffi_go_closure_linux64,24 + .type .ffi_go_closure_linux64,@function + .text +.ffi_go_closure_linux64: +# endif +# endif + +# if _CALL_ELF == 2 + ld %r12, 8(%r11) # closure->cif + mflr %r0 + lwz %r12, 28(%r12) # cif->flags + mtcrf 0x40, %r12 + addi %r12, %r1, PARMSAVE + bt 7, 0f + # Our caller has not allocated a parameter save area. + # We need to allocate one here and use it to pass gprs to + # ffi_closure_helper_LINUX64. + addi %r12, %r1, -STACKFRAME+PARMSAVE +0: + # Save general regs into parm save area + std %r3, 0(%r12) + std %r4, 8(%r12) + std %r5, 16(%r12) + std %r6, 24(%r12) + std %r7, 32(%r12) + std %r8, 40(%r12) + std %r9, 48(%r12) + std %r10, 56(%r12) + + # load up the pointer to the parm save area + mr %r7, %r12 +# else + mflr %r0 + # Save general regs into parm save area + # This is the parameter save area set up by our caller. + std %r3, PARMSAVE+0(%r1) + std %r4, PARMSAVE+8(%r1) + std %r5, PARMSAVE+16(%r1) + std %r6, PARMSAVE+24(%r1) + std %r7, PARMSAVE+32(%r1) + std %r8, PARMSAVE+40(%r1) + std %r9, PARMSAVE+48(%r1) + std %r10, PARMSAVE+56(%r1) + + # load up the pointer to the parm save area + addi %r7, %r1, PARMSAVE +# endif + std %r0, 16(%r1) + + # closure->cif + ld %r3, 8(%r11) + # closure->fun + ld %r4, 16(%r11) + # user_data + mr %r5, %r11 + b .Ldoclosure + + .cfi_endproc +# if _CALL_ELF == 2 + .size ffi_go_closure_linux64,.-ffi_go_closure_linux64 +# else +# ifdef _CALL_LINUX + .size ffi_go_closure_linux64,.-.L.ffi_go_closure_linux64 +# else + .long 0 + .byte 0,12,0,1,128,0,0,0 + .size .ffi_go_closure_linux64,.-.ffi_go_closure_linux64 +# endif +# endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #endif #if (defined __ELF__ && defined __linux__) || _CALL_ELF == 2 diff --git a/native/libffi/src/powerpc/ppc_closure.S b/native/libffi/src/powerpc/ppc_closure.S index 075922cbb0..38c006beb3 100644 --- a/native/libffi/src/powerpc/ppc_closure.S +++ b/native/libffi/src/powerpc/ppc_closure.S @@ -33,13 +33,14 @@ #ifndef POWERPC64 +FFI_HIDDEN(ffi_closure_SYSV) ENTRY(ffi_closure_SYSV) -.LFB1: + .cfi_startproc stwu %r1,-144(%r1) -.LCFI0: + .cfi_def_cfa_offset 144 mflr %r0 -.LCFI1: stw %r0,148(%r1) + .cfi_offset 65, 4 # we want to build up an areas for the parameters passed # in registers (both floating point and integer) @@ -48,6 +49,17 @@ ENTRY(ffi_closure_SYSV) stw %r3, 16(%r1) stw %r4, 20(%r1) stw %r5, 24(%r1) + + # set up registers for the routine that does the work + + # closure->cif + lwz %r3,FFI_TRAMPOLINE_SIZE(%r11) + # closure->fun + lwz %r4,FFI_TRAMPOLINE_SIZE+4(%r11) + # closure->user_data + lwz %r5,FFI_TRAMPOLINE_SIZE+8(%r11) + +.Ldoclosure: stw %r6, 28(%r1) stw %r7, 32(%r1) stw %r8, 36(%r1) @@ -66,23 +78,18 @@ ENTRY(ffi_closure_SYSV) stfd %f8, 104(%r1) #endif - # set up registers for the routine that actually does the work - # get the context pointer from the trampoline - mr %r3,%r11 + # pointer to the result storage + addi %r6,%r1,112 - # now load up the pointer to the result storage - addi %r4,%r1,112 + # pointer to the saved gpr registers + addi %r7,%r1,16 - # now load up the pointer to the saved gpr registers - addi %r5,%r1,16 + # pointer to the saved fpr registers + addi %r8,%r1,48 - # now load up the pointer to the saved fpr registers */ - addi %r6,%r1,48 - - # now load up the pointer to the outgoing parameter - # stack in the previous frame + # pointer to the outgoing parameter save area in the previous frame # i.e. the previous frame pointer + 8 - addi %r7,%r1,152 + addi %r9,%r1,152 # make the call bl ffi_closure_helper_SYSV@local @@ -101,7 +108,6 @@ ENTRY(ffi_closure_SYSV) add %r3,%r3,%r4 # add contents of table to table address mtctr %r3 bctr # jump to it -.LFE1: # Each of the ret_typeX code fragments has to be exactly 16 bytes long # (4 instructions). For cache effectiveness we align to a 16 byte boundary @@ -111,7 +117,9 @@ ENTRY(ffi_closure_SYSV) .Lret_type0: mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 nop # case FFI_TYPE_INT @@ -119,31 +127,33 @@ ENTRY(ffi_closure_SYSV) mtlr %r0 .Lfinish: addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_TYPE_FLOAT #ifndef __NO_FPRS__ lfs %f1,112+0(%r1) - mtlr %r0 - addi %r1,%r1,144 #else nop - nop - nop #endif + mtlr %r0 + addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_TYPE_DOUBLE #ifndef __NO_FPRS__ lfd %f1,112+0(%r1) - mtlr %r0 - addi %r1,%r1,144 #else nop - nop - nop #endif + mtlr %r0 + addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_TYPE_LONGDOUBLE #ifndef __NO_FPRS__ @@ -152,10 +162,12 @@ ENTRY(ffi_closure_SYSV) mtlr %r0 b .Lfinish #else - nop - nop - nop + mtlr %r0 + addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 + nop #endif # case FFI_TYPE_UINT8 @@ -166,7 +178,9 @@ ENTRY(ffi_closure_SYSV) #endif mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_TYPE_SINT8 #ifdef __LITTLE_ENDIAN__ @@ -186,7 +200,9 @@ ENTRY(ffi_closure_SYSV) #endif mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_TYPE_SINT16 #ifdef __LITTLE_ENDIAN__ @@ -196,19 +212,25 @@ ENTRY(ffi_closure_SYSV) #endif mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_TYPE_UINT32 lwz %r3,112+0(%r1) mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_TYPE_SINT32 lwz %r3,112+0(%r1) mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_TYPE_UINT64 lwz %r3,112+0(%r1) @@ -225,14 +247,18 @@ ENTRY(ffi_closure_SYSV) # case FFI_TYPE_STRUCT mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 nop # case FFI_TYPE_POINTER lwz %r3,112+0(%r1) mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_TYPE_UINT128 lwz %r3,112+0(%r1) @@ -245,20 +271,30 @@ ENTRY(ffi_closure_SYSV) lbz %r3,112+0(%r1) mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_SYSV_TYPE_SMALL_STRUCT + 2. Two byte struct. lhz %r3,112+0(%r1) mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_SYSV_TYPE_SMALL_STRUCT + 3. Three byte struct. lwz %r3,112+0(%r1) #ifdef __LITTLE_ENDIAN__ mtlr %r0 addi %r1,%r1,144 +<<<<<<< HEAD blr +======= + .cfi_def_cfa_offset 0 + blr + .cfi_def_cfa_offset 144 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #else srwi %r3,%r3,8 mtlr %r0 @@ -269,7 +305,9 @@ ENTRY(ffi_closure_SYSV) lwz %r3,112+0(%r1) mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr + .cfi_def_cfa_offset 144 # case FFI_SYSV_TYPE_SMALL_STRUCT + 5. Five byte struct. lwz %r3,112+0(%r1) @@ -319,65 +357,54 @@ ENTRY(ffi_closure_SYSV) or %r4,%r6,%r4 mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr +<<<<<<< HEAD +======= + .cfi_def_cfa_offset 144 +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #endif .Luint128: lwz %r6,112+12(%r1) mtlr %r0 addi %r1,%r1,144 + .cfi_def_cfa_offset 0 blr +<<<<<<< HEAD +======= + .cfi_endproc +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 END(ffi_closure_SYSV) - .section ".eh_frame",EH_FRAME_FLAGS,@progbits -.Lframe1: - .4byte .LECIE1-.LSCIE1 # Length of Common Information Entry -.LSCIE1: - .4byte 0x0 # CIE Identifier Tag - .byte 0x1 # CIE Version -#if defined _RELOCATABLE || defined __PIC__ - .ascii "zR\0" # CIE Augmentation -#else - .ascii "\0" # CIE Augmentation -#endif - .uleb128 0x1 # CIE Code Alignment Factor - .sleb128 -4 # CIE Data Alignment Factor - .byte 0x41 # CIE RA Column -#if defined _RELOCATABLE || defined __PIC__ - .uleb128 0x1 # Augmentation size - .byte 0x1b # FDE Encoding (pcrel sdata4) -#endif - .byte 0xc # DW_CFA_def_cfa - .uleb128 0x1 - .uleb128 0x0 - .align 2 -.LECIE1: -.LSFDE1: - .4byte .LEFDE1-.LASFDE1 # FDE Length -.LASFDE1: - .4byte .LASFDE1-.Lframe1 # FDE CIE offset -#if defined _RELOCATABLE || defined __PIC__ - .4byte .LFB1-. # FDE initial location -#else - .4byte .LFB1 # FDE initial location -#endif - .4byte .LFE1-.LFB1 # FDE address range -#if defined _RELOCATABLE || defined __PIC__ - .uleb128 0x0 # Augmentation size -#endif - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI0-.LFB1 - .byte 0xe # DW_CFA_def_cfa_offset - .uleb128 144 - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI1-.LCFI0 - .byte 0x11 # DW_CFA_offset_extended_sf - .uleb128 0x41 - .sleb128 -1 - .align 2 -.LEFDE1: +<<<<<<< HEAD +======= +FFI_HIDDEN(ffi_go_closure_sysv) +ENTRY(ffi_go_closure_sysv) + .cfi_startproc + stwu %r1,-144(%r1) + .cfi_def_cfa_offset 144 + mflr %r0 + stw %r0,148(%r1) + .cfi_offset 65, 4 + + stw %r3, 16(%r1) + stw %r4, 20(%r1) + stw %r5, 24(%r1) + + # closure->cif + lwz %r3,4(%r11) + # closure->fun + lwz %r4,8(%r11) + # user_data + mr %r5,%r11 + b .Ldoclosure + .cfi_endproc +END(ffi_go_closure_sysv) + +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #if defined __ELF__ && defined __linux__ .section .note.GNU-stack,"",@progbits #endif diff --git a/native/libffi/src/powerpc/sysv.S b/native/libffi/src/powerpc/sysv.S index fed2380c9f..b05334046d 100644 --- a/native/libffi/src/powerpc/sysv.S +++ b/native/libffi/src/powerpc/sysv.S @@ -31,34 +31,39 @@ #include #ifndef POWERPC64 +<<<<<<< HEAD .globl ffi_prep_args_SYSV +======= +FFI_HIDDEN(ffi_call_SYSV) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ENTRY(ffi_call_SYSV) -.LFB1: + .cfi_startproc /* Save the old stack pointer as AP. */ - mr %r8,%r1 + mr %r10,%r1 + .cfi_def_cfa_register 10 -.LCFI0: /* Allocate the stack space we need. */ - stwux %r1,%r1,%r4 + stwux %r1,%r1,%r8 /* Save registers we use. */ mflr %r9 - stw %r28,-16(%r8) -.LCFI1: - stw %r29,-12(%r8) -.LCFI2: - stw %r30, -8(%r8) -.LCFI3: - stw %r31, -4(%r8) -.LCFI4: - stw %r9, 4(%r8) -.LCFI5: + stw %r28,-16(%r10) + stw %r29,-12(%r10) + stw %r30, -8(%r10) + stw %r31, -4(%r10) + stw %r9, 4(%r10) + .cfi_offset 65, 4 + .cfi_offset 31, -4 + .cfi_offset 30, -8 + .cfi_offset 29, -12 + .cfi_offset 28, -16 /* Save arguments over call... */ - mr %r31,%r5 /* flags, */ - mr %r30,%r6 /* rvalue, */ - mr %r29,%r7 /* function address, */ - mr %r28,%r8 /* our AP. */ -.LCFI6: + stw %r7, -20(%r10) /* closure, */ + mr %r31,%r6 /* flags, */ + mr %r30,%r5 /* rvalue, */ + mr %r29,%r4 /* function address, */ + mr %r28,%r10 /* our AP. */ + .cfi_def_cfa_register 28 /* Call ffi_prep_args_SYSV. */ mr %r4,%r1 @@ -70,49 +75,49 @@ ENTRY(ffi_call_SYSV) /* Get the address to call into CTR. */ mtctr %r29 /* Load all those argument registers. */ - lwz %r3,-16-(8*4)(%r28) - lwz %r4,-16-(7*4)(%r28) - lwz %r5,-16-(6*4)(%r28) - lwz %r6,-16-(5*4)(%r28) + lwz %r3,-24-(8*4)(%r28) + lwz %r4,-24-(7*4)(%r28) + lwz %r5,-24-(6*4)(%r28) + lwz %r6,-24-(5*4)(%r28) bf- 5,1f nop - lwz %r7,-16-(4*4)(%r28) - lwz %r8,-16-(3*4)(%r28) - lwz %r9,-16-(2*4)(%r28) - lwz %r10,-16-(1*4)(%r28) + lwz %r7,-24-(4*4)(%r28) + lwz %r8,-24-(3*4)(%r28) + lwz %r9,-24-(2*4)(%r28) + lwz %r10,-24-(1*4)(%r28) nop 1: #ifndef __NO_FPRS__ /* Load all the FP registers. */ bf- 6,2f - lfd %f1,-16-(8*4)-(8*8)(%r28) - lfd %f2,-16-(8*4)-(7*8)(%r28) - lfd %f3,-16-(8*4)-(6*8)(%r28) - lfd %f4,-16-(8*4)-(5*8)(%r28) + lfd %f1,-24-(8*4)-(8*8)(%r28) + lfd %f2,-24-(8*4)-(7*8)(%r28) + lfd %f3,-24-(8*4)-(6*8)(%r28) + lfd %f4,-24-(8*4)-(5*8)(%r28) nop - lfd %f5,-16-(8*4)-(4*8)(%r28) - lfd %f6,-16-(8*4)-(3*8)(%r28) - lfd %f7,-16-(8*4)-(2*8)(%r28) - lfd %f8,-16-(8*4)-(1*8)(%r28) + lfd %f5,-24-(8*4)-(4*8)(%r28) + lfd %f6,-24-(8*4)-(3*8)(%r28) + lfd %f7,-24-(8*4)-(2*8)(%r28) + lfd %f8,-24-(8*4)-(1*8)(%r28) #endif 2: /* Make the call. */ + lwz %r11, -20(%r28) bctrl /* Now, deal with the return value. */ - mtcrf 0x01,%r31 /* cr7 */ + mtcrf 0x03,%r31 /* cr6-cr7 */ bt- 31,L(small_struct_return_value) bt- 30,L(done_return_value) #ifndef __NO_FPRS__ bt- 29,L(fp_return_value) #endif stw %r3,0(%r30) - bf+ 28,L(done_return_value) + bf+ 27,L(done_return_value) stw %r4,4(%r30) - mtcrf 0x02,%r31 /* cr6 */ - bf 27,L(done_return_value) + bf 26,L(done_return_value) stw %r5,8(%r30) stw %r6,12(%r30) /* Fall through... */ @@ -125,15 +130,27 @@ L(done_return_value): lwz %r30, -8(%r28) lwz %r29,-12(%r28) lwz %r28,-16(%r28) + .cfi_remember_state + /* At this point we don't have a cfa register. Say all our + saved regs have been restored. */ + .cfi_same_value 65 + .cfi_same_value 31 + .cfi_same_value 30 + .cfi_same_value 29 + .cfi_same_value 28 + /* Hopefully this works.. */ + .cfi_def_cfa_register 1 + .cfi_offset 1, 0 lwz %r1,0(%r1) + .cfi_same_value 1 blr #ifndef __NO_FPRS__ L(fp_return_value): - bf 28,L(float_return_value) + .cfi_restore_state + bf 27,L(float_return_value) stfd %f1,0(%r30) - mtcrf 0x02,%r31 /* cr6 */ - bf 27,L(done_return_value) + bf 26,L(done_return_value) stfd %f2,8(%r30) b L(done_return_value) L(float_return_value): @@ -150,10 +167,14 @@ L(small_struct_return_value): stw %r3, 0(%r30) stw %r4, 4(%r30) b L(done_return_value) +<<<<<<< HEAD +======= + .cfi_endproc +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 -.LFE1: END(ffi_call_SYSV) +<<<<<<< HEAD .section ".eh_frame",EH_FRAME_FLAGS,@progbits .Lframe1: .4byte .LECIE1-.LSCIE1 /* Length of Common Information Entry */ @@ -214,6 +235,8 @@ END(ffi_call_SYSV) .align 2 .LEFDE1: +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #if defined __ELF__ && defined __linux__ .section .note.GNU-stack,"",@progbits #endif diff --git a/native/libffi/src/prep_cif.c b/native/libffi/src/prep_cif.c index be5eae379e..a95f733c94 100644 --- a/native/libffi/src/prep_cif.c +++ b/native/libffi/src/prep_cif.c @@ -29,12 +29,12 @@ /* Round up to FFI_SIZEOF_ARG. */ -#define STACK_ARG_SIZE(x) ALIGN(x, FFI_SIZEOF_ARG) +#define STACK_ARG_SIZE(x) FFI_ALIGN(x, FFI_SIZEOF_ARG) /* Perform machine independent initialization of aggregate type specifications. */ -static ffi_status initialize_aggregate(ffi_type *arg) +static ffi_status initialize_aggregate(ffi_type *arg, size_t *offsets) { ffi_type **ptr; @@ -52,13 +52,15 @@ static ffi_status initialize_aggregate(ffi_type *arg) while ((*ptr) != NULL) { if (UNLIKELY(((*ptr)->size == 0) - && (initialize_aggregate((*ptr)) != FFI_OK))) + && (initialize_aggregate((*ptr), NULL) != FFI_OK))) return FFI_BAD_TYPEDEF; /* Perform a sanity check on the argument type */ FFI_ASSERT_VALID_TYPE(*ptr); - arg->size = ALIGN(arg->size, (*ptr)->alignment); + arg->size = FFI_ALIGN(arg->size, (*ptr)->alignment); + if (offsets) + *offsets++ = arg->size; arg->size += (*ptr)->size; arg->alignment = (arg->alignment > (*ptr)->alignment) ? @@ -74,7 +76,14 @@ static ffi_status initialize_aggregate(ffi_type *arg) struct A { long a; char b; }; struct B { struct A x; char y; }; should find y at an offset of 2*sizeof(long) and result in a total size of 3*sizeof(long). */ - arg->size = ALIGN (arg->size, arg->alignment); + arg->size = FFI_ALIGN (arg->size, arg->alignment); + + /* On some targets, the ABI defines that structures have an additional + alignment beyond the "natural" one based on their elements. */ +#ifdef FFI_AGGREGATE_ALIGNMENT + if (FFI_AGGREGATE_ALIGNMENT > arg->alignment) + arg->alignment = FFI_AGGREGATE_ALIGNMENT; +#endif /* On some targets, the ABI defines that structures have an additional alignment beyond the "natural" one based on their elements. */ @@ -127,13 +136,20 @@ ffi_status FFI_HIDDEN ffi_prep_cif_core(ffi_cif *cif, ffi_abi abi, cif->rtype = rtype; cif->flags = 0; +#if (defined(_M_ARM64) || defined(__aarch64__)) && defined(_WIN32) + cif->is_variadic = isvariadic; +#endif +#if HAVE_LONG_DOUBLE_VARIANT + ffi_prep_types (abi); +#endif #if HAVE_LONG_DOUBLE_VARIANT ffi_prep_types (abi); #endif /* Initialize the return type if necessary */ - if ((cif->rtype->size == 0) && (initialize_aggregate(cif->rtype) != FFI_OK)) + if ((cif->rtype->size == 0) + && (initialize_aggregate(cif->rtype, NULL) != FFI_OK)) return FFI_BAD_TYPEDEF; #ifndef FFI_TARGET_HAS_COMPLEX_TYPE @@ -147,9 +163,6 @@ ffi_status FFI_HIDDEN ffi_prep_cif_core(ffi_cif *cif, ffi_abi abi, #if !defined FFI_TARGET_SPECIFIC_STACK_SPACE_ALLOCATION /* Make space for the return structure pointer */ if (cif->rtype->type == FFI_TYPE_STRUCT -#ifdef SPARC - && (cif->abi != FFI_V9 || cif->rtype->size > 32) -#endif #ifdef TILE && (cif->rtype->size > 10 * FFI_SIZEOF_ARG) #endif @@ -167,7 +180,8 @@ ffi_status FFI_HIDDEN ffi_prep_cif_core(ffi_cif *cif, ffi_abi abi, { /* Initialize any uninitialized aggregate type definitions */ - if (((*ptr)->size == 0) && (initialize_aggregate((*ptr)) != FFI_OK)) + if (((*ptr)->size == 0) + && (initialize_aggregate((*ptr), NULL) != FFI_OK)) return FFI_BAD_TYPEDEF; #ifndef FFI_TARGET_HAS_COMPLEX_TYPE @@ -179,6 +193,7 @@ ffi_status FFI_HIDDEN ffi_prep_cif_core(ffi_cif *cif, ffi_abi abi, FFI_ASSERT_VALID_TYPE(*ptr); #if !defined FFI_TARGET_SPECIFIC_STACK_SPACE_ALLOCATION +<<<<<<< HEAD #ifdef SPARC if (((*ptr)->type == FFI_TYPE_STRUCT && ((*ptr)->size > 16 || cif->abi != FFI_V9)) @@ -191,6 +206,12 @@ ffi_status FFI_HIDDEN ffi_prep_cif_core(ffi_cif *cif, ffi_abi abi, /* Add any padding if necessary */ if (((*ptr)->alignment - 1) & bytes) bytes = (unsigned)ALIGN(bytes, (*ptr)->alignment); +======= + { + /* Add any padding if necessary */ + if (((*ptr)->alignment - 1) & bytes) + bytes = (unsigned)FFI_ALIGN(bytes, (*ptr)->alignment); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #ifdef TILE if (bytes < 10 * FFI_SIZEOF_ARG && @@ -206,7 +227,7 @@ ffi_status FFI_HIDDEN ffi_prep_cif_core(ffi_cif *cif, ffi_abi abi, bytes = 6*4; #endif - bytes += STACK_ARG_SIZE((*ptr)->size); + bytes += (unsigned int)STACK_ARG_SIZE((*ptr)->size); } #endif } @@ -251,3 +272,18 @@ ffi_prep_closure (ffi_closure* closure, } #endif + +ffi_status +ffi_get_struct_offsets (ffi_abi abi, ffi_type *struct_type, size_t *offsets) +{ + if (! (abi > FFI_FIRST_ABI && abi < FFI_LAST_ABI)) + return FFI_BAD_ABI; + if (struct_type->type != FFI_TYPE_STRUCT) + return FFI_BAD_TYPEDEF; + +#if HAVE_LONG_DOUBLE_VARIANT + ffi_prep_types (abi); +#endif + + return initialize_aggregate(struct_type, offsets); +} diff --git a/native/libffi/src/raw_api.c b/native/libffi/src/raw_api.c index 276cb22807..be156116cb 100644 --- a/native/libffi/src/raw_api.c +++ b/native/libffi/src/raw_api.c @@ -43,10 +43,10 @@ ffi_raw_size (ffi_cif *cif) { #if !FFI_NO_STRUCTS if ((*at)->type == FFI_TYPE_STRUCT) - result += ALIGN (sizeof (void*), FFI_SIZEOF_ARG); + result += FFI_ALIGN (sizeof (void*), FFI_SIZEOF_ARG); else #endif - result += ALIGN ((*at)->size, FFI_SIZEOF_ARG); + result += FFI_ALIGN ((*at)->size, FFI_SIZEOF_ARG); } return result; @@ -98,7 +98,7 @@ ffi_raw_to_ptrarray (ffi_cif *cif, ffi_raw *raw, void **args) default: *args = raw; - raw += ALIGN ((*tp)->size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; + raw += FFI_ALIGN ((*tp)->size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; } } @@ -123,7 +123,7 @@ ffi_raw_to_ptrarray (ffi_cif *cif, ffi_raw *raw, void **args) else { *args = (void*) raw; - raw += ALIGN ((*tp)->size, sizeof (void*)) / sizeof (void*); + raw += FFI_ALIGN ((*tp)->size, sizeof (void*)) / sizeof (void*); } } @@ -186,7 +186,7 @@ ffi_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_raw *raw) default: memcpy ((void*) raw->data, (void*)*args, (*tp)->size); - raw += ALIGN ((*tp)->size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; + raw += FFI_ALIGN ((*tp)->size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; } } } diff --git a/native/libffi/src/riscv/ffi.c b/native/libffi/src/riscv/ffi.c new file mode 100644 index 0000000000..c910858896 --- /dev/null +++ b/native/libffi/src/riscv/ffi.c @@ -0,0 +1,481 @@ +/* ----------------------------------------------------------------------- + ffi.c - Copyright (c) 2015 Michael Knyszek + 2015 Andrew Waterman + 2018 Stef O'Rear + Based on MIPS N32/64 port + + RISC-V Foreign Function Interface + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + ``Software''), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice shall be included + in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + ----------------------------------------------------------------------- */ + +#include +#include + +#include +#include + +#if __riscv_float_abi_double +#define ABI_FLEN 64 +#define ABI_FLOAT double +#elif __riscv_float_abi_single +#define ABI_FLEN 32 +#define ABI_FLOAT float +#endif + +#define NARGREG 8 +#define STKALIGN 16 +#define MAXCOPYARG (2 * sizeof(double)) + +typedef struct call_context +{ +#if ABI_FLEN + ABI_FLOAT fa[8]; +#endif + size_t a[8]; + /* used by the assembly code to in-place construct its own stack frame */ + char frame[16]; +} call_context; + +typedef struct call_builder +{ + call_context *aregs; + int used_integer; + int used_float; + size_t *used_stack; +} call_builder; + +/* integer (not pointer) less than ABI XLEN */ +/* FFI_TYPE_INT does not appear to be used */ +#if __SIZEOF_POINTER__ == 8 +#define IS_INT(type) ((type) >= FFI_TYPE_UINT8 && (type) <= FFI_TYPE_SINT64) +#else +#define IS_INT(type) ((type) >= FFI_TYPE_UINT8 && (type) <= FFI_TYPE_SINT32) +#endif + +#if ABI_FLEN +typedef struct { + char as_elements, type1, offset2, type2; +} float_struct_info; + +#if ABI_FLEN >= 64 +#define IS_FLOAT(type) ((type) >= FFI_TYPE_FLOAT && (type) <= FFI_TYPE_DOUBLE) +#else +#define IS_FLOAT(type) ((type) == FFI_TYPE_FLOAT) +#endif + +static ffi_type **flatten_struct(ffi_type *in, ffi_type **out, ffi_type **out_end) { + int i; + if (out == out_end) return out; + if (in->type != FFI_TYPE_STRUCT) { + *(out++) = in; + } else { + for (i = 0; in->elements[i]; i++) + out = flatten_struct(in->elements[i], out, out_end); + } + return out; +} + +/* Structs with at most two fields after flattening, one of which is of + floating point type, are passed in multiple registers if sufficient + registers are available. */ +static float_struct_info struct_passed_as_elements(call_builder *cb, ffi_type *top) { + float_struct_info ret = {0, 0, 0, 0}; + ffi_type *fields[3]; + int num_floats, num_ints; + int num_fields = flatten_struct(top, fields, fields + 3) - fields; + + if (num_fields == 1) { + if (IS_FLOAT(fields[0]->type)) { + ret.as_elements = 1; + ret.type1 = fields[0]->type; + } + } else if (num_fields == 2) { + num_floats = IS_FLOAT(fields[0]->type) + IS_FLOAT(fields[1]->type); + num_ints = IS_INT(fields[0]->type) + IS_INT(fields[1]->type); + if (num_floats == 0 || num_floats + num_ints != 2) + return ret; + if (cb->used_float + num_floats > NARGREG || cb->used_integer + (2 - num_floats) > NARGREG) + return ret; + if (!IS_FLOAT(fields[0]->type) && !IS_FLOAT(fields[1]->type)) + return ret; + + ret.type1 = fields[0]->type; + ret.type2 = fields[1]->type; + ret.offset2 = FFI_ALIGN(fields[0]->size, fields[1]->alignment); + ret.as_elements = 1; + } + + return ret; +} +#endif + +/* allocates a single register, float register, or XLEN-sized stack slot to a datum */ +static void marshal_atom(call_builder *cb, int type, void *data) { + size_t value = 0; + switch (type) { + case FFI_TYPE_UINT8: value = *(uint8_t *)data; break; + case FFI_TYPE_SINT8: value = *(int8_t *)data; break; + case FFI_TYPE_UINT16: value = *(uint16_t *)data; break; + case FFI_TYPE_SINT16: value = *(int16_t *)data; break; + /* 32-bit quantities are always sign-extended in the ABI */ + case FFI_TYPE_UINT32: value = *(int32_t *)data; break; + case FFI_TYPE_SINT32: value = *(int32_t *)data; break; +#if __SIZEOF_POINTER__ == 8 + case FFI_TYPE_UINT64: value = *(uint64_t *)data; break; + case FFI_TYPE_SINT64: value = *(int64_t *)data; break; +#endif + case FFI_TYPE_POINTER: value = *(size_t *)data; break; + + /* float values may be recoded in an implementation-defined way + by hardware conforming to 2.1 or earlier, so use asm to + reinterpret floats as doubles */ +#if ABI_FLEN >= 32 + case FFI_TYPE_FLOAT: + asm("" : "=f"(cb->aregs->fa[cb->used_float++]) : "0"(*(float *)data)); + return; +#endif +#if ABI_FLEN >= 64 + case FFI_TYPE_DOUBLE: + asm("" : "=f"(cb->aregs->fa[cb->used_float++]) : "0"(*(double *)data)); + return; +#endif + default: FFI_ASSERT(0); break; + } + + if (cb->used_integer == NARGREG) { + *cb->used_stack++ = value; + } else { + cb->aregs->a[cb->used_integer++] = value; + } +} + +static void unmarshal_atom(call_builder *cb, int type, void *data) { + size_t value; + switch (type) { +#if ABI_FLEN >= 32 + case FFI_TYPE_FLOAT: + asm("" : "=f"(*(float *)data) : "0"(cb->aregs->fa[cb->used_float++])); + return; +#endif +#if ABI_FLEN >= 64 + case FFI_TYPE_DOUBLE: + asm("" : "=f"(*(double *)data) : "0"(cb->aregs->fa[cb->used_float++])); + return; +#endif + } + + if (cb->used_integer == NARGREG) { + value = *cb->used_stack++; + } else { + value = cb->aregs->a[cb->used_integer++]; + } + + switch (type) { + case FFI_TYPE_UINT8: *(uint8_t *)data = value; break; + case FFI_TYPE_SINT8: *(uint8_t *)data = value; break; + case FFI_TYPE_UINT16: *(uint16_t *)data = value; break; + case FFI_TYPE_SINT16: *(uint16_t *)data = value; break; + case FFI_TYPE_UINT32: *(uint32_t *)data = value; break; + case FFI_TYPE_SINT32: *(uint32_t *)data = value; break; +#if __SIZEOF_POINTER__ == 8 + case FFI_TYPE_UINT64: *(uint64_t *)data = value; break; + case FFI_TYPE_SINT64: *(uint64_t *)data = value; break; +#endif + case FFI_TYPE_POINTER: *(size_t *)data = value; break; + default: FFI_ASSERT(0); break; + } +} + +/* adds an argument to a call, or a not by reference return value */ +static void marshal(call_builder *cb, ffi_type *type, int var, void *data) { + size_t realign[2]; + +#if ABI_FLEN + if (!var && type->type == FFI_TYPE_STRUCT) { + float_struct_info fsi = struct_passed_as_elements(cb, type); + if (fsi.as_elements) { + marshal_atom(cb, fsi.type1, data); + if (fsi.offset2) + marshal_atom(cb, fsi.type2, ((char*)data) + fsi.offset2); + return; + } + } + + if (!var && cb->used_float < NARGREG && IS_FLOAT(type->type)) { + marshal_atom(cb, type->type, data); + return; + } +#endif + + if (type->size > 2 * __SIZEOF_POINTER__) { + /* pass by reference */ + marshal_atom(cb, FFI_TYPE_POINTER, &data); + } else if (IS_INT(type->type) || type->type == FFI_TYPE_POINTER) { + marshal_atom(cb, type->type, data); + } else { + /* overlong integers, soft-float floats, and structs without special + float handling are treated identically from this point on */ + + /* variadics are aligned even in registers */ + if (type->alignment > __SIZEOF_POINTER__) { + if (var) + cb->used_integer = FFI_ALIGN(cb->used_integer, 2); + cb->used_stack = (size_t *)FFI_ALIGN(cb->used_stack, 2*__SIZEOF_POINTER__); + } + + memcpy(realign, data, type->size); + if (type->size > 0) + marshal_atom(cb, FFI_TYPE_POINTER, realign); + if (type->size > __SIZEOF_POINTER__) + marshal_atom(cb, FFI_TYPE_POINTER, realign + 1); + } +} + +/* for arguments passed by reference returns the pointer, otherwise the arg is copied (up to MAXCOPYARG bytes) */ +static void *unmarshal(call_builder *cb, ffi_type *type, int var, void *data) { + size_t realign[2]; + void *pointer; + +#if ABI_FLEN + if (!var && type->type == FFI_TYPE_STRUCT) { + float_struct_info fsi = struct_passed_as_elements(cb, type); + if (fsi.as_elements) { + unmarshal_atom(cb, fsi.type1, data); + if (fsi.offset2) + unmarshal_atom(cb, fsi.type2, ((char*)data) + fsi.offset2); + return data; + } + } + + if (!var && cb->used_float < NARGREG && IS_FLOAT(type->type)) { + unmarshal_atom(cb, type->type, data); + return data; + } +#endif + + if (type->size > 2 * __SIZEOF_POINTER__) { + /* pass by reference */ + unmarshal_atom(cb, FFI_TYPE_POINTER, (char*)&pointer); + return pointer; + } else if (IS_INT(type->type) || type->type == FFI_TYPE_POINTER) { + unmarshal_atom(cb, type->type, data); + return data; + } else { + /* overlong integers, soft-float floats, and structs without special + float handling are treated identically from this point on */ + + /* variadics are aligned even in registers */ + if (type->alignment > __SIZEOF_POINTER__) { + if (var) + cb->used_integer = FFI_ALIGN(cb->used_integer, 2); + cb->used_stack = (size_t *)FFI_ALIGN(cb->used_stack, 2*__SIZEOF_POINTER__); + } + + if (type->size > 0) + unmarshal_atom(cb, FFI_TYPE_POINTER, realign); + if (type->size > __SIZEOF_POINTER__) + unmarshal_atom(cb, FFI_TYPE_POINTER, realign + 1); + memcpy(data, realign, type->size); + return data; + } +} + +static int passed_by_ref(call_builder *cb, ffi_type *type, int var) { +#if ABI_FLEN + if (!var && type->type == FFI_TYPE_STRUCT) { + float_struct_info fsi = struct_passed_as_elements(cb, type); + if (fsi.as_elements) return 0; + } +#endif + + return type->size > 2 * __SIZEOF_POINTER__; +} + +/* Perform machine dependent cif processing */ +ffi_status ffi_prep_cif_machdep(ffi_cif *cif) { + cif->riscv_nfixedargs = cif->nargs; + return FFI_OK; +} + +/* Perform machine dependent cif processing when we have a variadic function */ + +ffi_status ffi_prep_cif_machdep_var(ffi_cif *cif, unsigned int nfixedargs, unsigned int ntotalargs) { + cif->riscv_nfixedargs = nfixedargs; + return FFI_OK; +} + +/* Low level routine for calling functions */ +extern void ffi_call_asm (void *stack, struct call_context *regs, + void (*fn) (void), void *closure) FFI_HIDDEN; + +static void +ffi_call_int (ffi_cif *cif, void (*fn) (void), void *rvalue, void **avalue, + void *closure) +{ + /* this is a conservative estimate, assuming a complex return value and + that all remaining arguments are long long / __int128 */ + size_t arg_bytes = cif->nargs <= 3 ? 0 : + FFI_ALIGN(2 * sizeof(size_t) * (cif->nargs - 3), STKALIGN); + size_t rval_bytes = 0; + if (rvalue == NULL && cif->rtype->size > 2*__SIZEOF_POINTER__) + rval_bytes = FFI_ALIGN(cif->rtype->size, STKALIGN); + size_t alloc_size = arg_bytes + rval_bytes + sizeof(call_context); + + /* the assembly code will deallocate all stack data at lower addresses + than the argument region, so we need to allocate the frame and the + return value after the arguments in a single allocation */ + size_t alloc_base; + /* Argument region must be 16-byte aligned */ + if (_Alignof(max_align_t) >= STKALIGN) { + /* since sizeof long double is normally 16, the compiler will + guarantee alloca alignment to at least that much */ + alloc_base = (size_t)alloca(alloc_size); + } else { + alloc_base = FFI_ALIGN(alloca(alloc_size + STKALIGN - 1), STKALIGN); + } + + if (rval_bytes) + rvalue = (void*)(alloc_base + arg_bytes); + + call_builder cb; + cb.used_float = cb.used_integer = 0; + cb.aregs = (call_context*)(alloc_base + arg_bytes + rval_bytes); + cb.used_stack = (void*)alloc_base; + + int return_by_ref = passed_by_ref(&cb, cif->rtype, 0); + if (return_by_ref) + marshal(&cb, &ffi_type_pointer, 0, &rvalue); + + int i; + for (i = 0; i < cif->nargs; i++) + marshal(&cb, cif->arg_types[i], i >= cif->riscv_nfixedargs, avalue[i]); + + ffi_call_asm ((void *) alloc_base, cb.aregs, fn, closure); + + cb.used_float = cb.used_integer = 0; + if (!return_by_ref && rvalue) + unmarshal(&cb, cif->rtype, 0, rvalue); +} + +void +ffi_call (ffi_cif *cif, void (*fn) (void), void *rvalue, void **avalue) +{ + ffi_call_int(cif, fn, rvalue, avalue, NULL); +} + +void +ffi_call_go (ffi_cif *cif, void (*fn) (void), void *rvalue, + void **avalue, void *closure) +{ + ffi_call_int(cif, fn, rvalue, avalue, closure); +} + +extern void ffi_closure_asm(void) FFI_HIDDEN; + +ffi_status ffi_prep_closure_loc(ffi_closure *closure, ffi_cif *cif, void (*fun)(ffi_cif*,void*,void**,void*), void *user_data, void *codeloc) +{ + uint32_t *tramp = (uint32_t *) &closure->tramp[0]; + uint64_t fn = (uint64_t) (uintptr_t) ffi_closure_asm; + + if (cif->abi <= FFI_FIRST_ABI || cif->abi >= FFI_LAST_ABI) + return FFI_BAD_ABI; + + /* we will call ffi_closure_inner with codeloc, not closure, but as long + as the memory is readable it should work */ + + tramp[0] = 0x00000317; /* auipc t1, 0 (i.e. t0 <- codeloc) */ +#if __SIZEOF_POINTER__ == 8 + tramp[1] = 0x01033383; /* ld t2, 16(t1) */ +#else + tramp[1] = 0x01032383; /* lw t2, 16(t1) */ +#endif + tramp[2] = 0x00038067; /* jr t2 */ + tramp[3] = 0x00000013; /* nop */ + tramp[4] = fn; + tramp[5] = fn >> 32; + + closure->cif = cif; + closure->fun = fun; + closure->user_data = user_data; + + __builtin___clear_cache(codeloc, codeloc + FFI_TRAMPOLINE_SIZE); + + return FFI_OK; +} + +extern void ffi_go_closure_asm (void) FFI_HIDDEN; + +ffi_status +ffi_prep_go_closure (ffi_go_closure *closure, ffi_cif *cif, + void (*fun) (ffi_cif *, void *, void **, void *)) +{ + if (cif->abi <= FFI_FIRST_ABI || cif->abi >= FFI_LAST_ABI) + return FFI_BAD_ABI; + + closure->tramp = (void *) ffi_go_closure_asm; + closure->cif = cif; + closure->fun = fun; + + return FFI_OK; +} + +/* Called by the assembly code with aregs pointing to saved argument registers + and stack pointing to the stacked arguments. Return values passed in + registers will be reloaded from aregs. */ +void FFI_HIDDEN +ffi_closure_inner (ffi_cif *cif, + void (*fun) (ffi_cif *, void *, void **, void *), + void *user_data, + size_t *stack, call_context *aregs) +{ + void **avalue = alloca(cif->nargs * sizeof(void*)); + /* storage for arguments which will be copied by unmarshal(). We could + theoretically avoid the copies in many cases and use at most 128 bytes + of memory, but allocating disjoint storage for each argument is + simpler. */ + char *astorage = alloca(cif->nargs * MAXCOPYARG); + void *rvalue; + call_builder cb; + int return_by_ref; + int i; + + cb.aregs = aregs; + cb.used_integer = cb.used_float = 0; + cb.used_stack = stack; + + return_by_ref = passed_by_ref(&cb, cif->rtype, 0); + if (return_by_ref) + unmarshal(&cb, &ffi_type_pointer, 0, &rvalue); + else + rvalue = alloca(cif->rtype->size); + + for (i = 0; i < cif->nargs; i++) + avalue[i] = unmarshal(&cb, cif->arg_types[i], + i >= cif->riscv_nfixedargs, astorage + i*MAXCOPYARG); + + fun (cif, rvalue, avalue, user_data); + + if (!return_by_ref && cif->rtype->type != FFI_TYPE_VOID) { + cb.used_integer = cb.used_float = 0; + marshal(&cb, cif->rtype, 0, rvalue); + } +} diff --git a/native/libffi/src/riscv/ffitarget.h b/native/libffi/src/riscv/ffitarget.h new file mode 100644 index 0000000000..75e6462f5b --- /dev/null +++ b/native/libffi/src/riscv/ffitarget.h @@ -0,0 +1,69 @@ +/* -----------------------------------------------------------------*-C-*- + ffitarget.h - 2014 Michael Knyszek + + Target configuration macros for RISC-V. + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + ``Software''), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice shall be included + in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + + ----------------------------------------------------------------------- */ + +#ifndef LIBFFI_TARGET_H +#define LIBFFI_TARGET_H + +#ifndef LIBFFI_H +#error "Please do not include ffitarget.h directly into your source. Use ffi.h instead." +#endif + +#ifndef __riscv +#error "libffi was configured for a RISC-V target but this does not appear to be a RISC-V compiler." +#endif + +#ifndef LIBFFI_ASM + +typedef unsigned long ffi_arg; +typedef signed long ffi_sarg; + +/* FFI_UNUSED_NN and riscv_unused are to maintain ABI compatibility with a + distributed Berkeley patch from 2014, and can be removed at SONAME bump */ +typedef enum ffi_abi { + FFI_FIRST_ABI = 0, + FFI_SYSV, + FFI_UNUSED_1, + FFI_UNUSED_2, + FFI_UNUSED_3, + FFI_LAST_ABI, + + FFI_DEFAULT_ABI = FFI_SYSV +} ffi_abi; + +#endif /* LIBFFI_ASM */ + +/* ---- Definitions for closures ----------------------------------------- */ + +#define FFI_CLOSURES 1 +#define FFI_GO_CLOSURES 1 +#define FFI_TRAMPOLINE_SIZE 24 +#define FFI_NATIVE_RAW_API 0 +#define FFI_EXTRA_CIF_FIELDS unsigned riscv_nfixedargs; unsigned riscv_unused; +#define FFI_TARGET_SPECIFIC_VARIADIC + +#endif + diff --git a/native/libffi/src/riscv/sysv.S b/native/libffi/src/riscv/sysv.S new file mode 100644 index 0000000000..522d0b0055 --- /dev/null +++ b/native/libffi/src/riscv/sysv.S @@ -0,0 +1,293 @@ +/* ----------------------------------------------------------------------- + ffi.c - Copyright (c) 2015 Michael Knyszek + 2015 Andrew Waterman + 2018 Stef O'Rear + + RISC-V Foreign Function Interface + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + ``Software''), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice shall be included + in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + ----------------------------------------------------------------------- */ + +#define LIBFFI_ASM +#include +#include + +/* Define aliases so that we can handle all ABIs uniformly */ + +#if __SIZEOF_POINTER__ == 8 +#define PTRS 8 +#define LARG ld +#define SARG sd +#else +#define PTRS 4 +#define LARG lw +#define SARG sw +#endif + +#if __riscv_float_abi_double +#define FLTS 8 +#define FLARG fld +#define FSARG fsd +#elif __riscv_float_abi_single +#define FLTS 4 +#define FLARG flw +#define FSARG fsw +#else +#define FLTS 0 +#endif + +#define fp s0 + + .text + .globl ffi_call_asm + .type ffi_call_asm, @function + .hidden ffi_call_asm +/* + struct call_context { + floatreg fa[8]; + intreg a[8]; + intreg pad[rv32 ? 2 : 0]; + intreg save_fp, save_ra; + } + void ffi_call_asm (size_t *stackargs, struct call_context *regargs, + void (*fn) (void), void *closure); +*/ + +#define FRAME_LEN (8 * FLTS + 8 * PTRS + 16) + +ffi_call_asm: + .cfi_startproc + + /* + We are NOT going to set up an ordinary stack frame. In order to pass + the stacked args to the called function, we adjust our stack pointer to + a0, which is in the _caller's_ alloca area. We establish our own stack + frame at the end of the call_context. + + Anything below the arguments will be freed at this point, although we + preserve the call_context so that it can be read back in the caller. + */ + + .cfi_def_cfa 11, FRAME_LEN # interim CFA based on a1 + SARG fp, FRAME_LEN - 2*PTRS(a1) + .cfi_offset 8, -2*PTRS + SARG ra, FRAME_LEN - 1*PTRS(a1) + .cfi_offset 1, -1*PTRS + + addi fp, a1, FRAME_LEN + mv sp, a0 + .cfi_def_cfa 8, 0 # our frame is fully set up + + # Load arguments + mv t1, a2 + mv t2, a3 + +#if FLTS + FLARG fa0, -FRAME_LEN+0*FLTS(fp) + FLARG fa1, -FRAME_LEN+1*FLTS(fp) + FLARG fa2, -FRAME_LEN+2*FLTS(fp) + FLARG fa3, -FRAME_LEN+3*FLTS(fp) + FLARG fa4, -FRAME_LEN+4*FLTS(fp) + FLARG fa5, -FRAME_LEN+5*FLTS(fp) + FLARG fa6, -FRAME_LEN+6*FLTS(fp) + FLARG fa7, -FRAME_LEN+7*FLTS(fp) +#endif + + LARG a0, -FRAME_LEN+8*FLTS+0*PTRS(fp) + LARG a1, -FRAME_LEN+8*FLTS+1*PTRS(fp) + LARG a2, -FRAME_LEN+8*FLTS+2*PTRS(fp) + LARG a3, -FRAME_LEN+8*FLTS+3*PTRS(fp) + LARG a4, -FRAME_LEN+8*FLTS+4*PTRS(fp) + LARG a5, -FRAME_LEN+8*FLTS+5*PTRS(fp) + LARG a6, -FRAME_LEN+8*FLTS+6*PTRS(fp) + LARG a7, -FRAME_LEN+8*FLTS+7*PTRS(fp) + + /* Call */ + jalr t1 + + /* Save return values - only a0/a1 (fa0/fa1) are used */ +#if FLTS + FSARG fa0, -FRAME_LEN+0*FLTS(fp) + FSARG fa1, -FRAME_LEN+1*FLTS(fp) +#endif + + SARG a0, -FRAME_LEN+8*FLTS+0*PTRS(fp) + SARG a1, -FRAME_LEN+8*FLTS+1*PTRS(fp) + + /* Restore and return */ + addi sp, fp, -FRAME_LEN + .cfi_def_cfa 2, FRAME_LEN + LARG ra, -1*PTRS(fp) + .cfi_restore 1 + LARG fp, -2*PTRS(fp) + .cfi_restore 8 + ret + .cfi_endproc + .size ffi_call_asm, .-ffi_call_asm + + +/* + ffi_closure_asm. Expects address of the passed-in ffi_closure in t1. + void ffi_closure_inner (ffi_cif *cif, + void (*fun) (ffi_cif *, void *, void **, void *), + void *user_data, + size_t *stackargs, struct call_context *regargs) +*/ + + .globl ffi_closure_asm + .hidden ffi_closure_asm + .type ffi_closure_asm, @function +ffi_closure_asm: + .cfi_startproc + + addi sp, sp, -FRAME_LEN + .cfi_def_cfa_offset FRAME_LEN + + /* make a frame */ + SARG fp, FRAME_LEN - 2*PTRS(sp) + .cfi_offset 8, -2*PTRS + SARG ra, FRAME_LEN - 1*PTRS(sp) + .cfi_offset 1, -1*PTRS + addi fp, sp, FRAME_LEN + + /* save arguments */ +#if FLTS + FSARG fa0, 0*FLTS(sp) + FSARG fa1, 1*FLTS(sp) + FSARG fa2, 2*FLTS(sp) + FSARG fa3, 3*FLTS(sp) + FSARG fa4, 4*FLTS(sp) + FSARG fa5, 5*FLTS(sp) + FSARG fa6, 6*FLTS(sp) + FSARG fa7, 7*FLTS(sp) +#endif + + SARG a0, 8*FLTS+0*PTRS(sp) + SARG a1, 8*FLTS+1*PTRS(sp) + SARG a2, 8*FLTS+2*PTRS(sp) + SARG a3, 8*FLTS+3*PTRS(sp) + SARG a4, 8*FLTS+4*PTRS(sp) + SARG a5, 8*FLTS+5*PTRS(sp) + SARG a6, 8*FLTS+6*PTRS(sp) + SARG a7, 8*FLTS+7*PTRS(sp) + + /* enter C */ + LARG a0, FFI_TRAMPOLINE_SIZE+0*PTRS(t1) + LARG a1, FFI_TRAMPOLINE_SIZE+1*PTRS(t1) + LARG a2, FFI_TRAMPOLINE_SIZE+2*PTRS(t1) + addi a3, sp, FRAME_LEN + mv a4, sp + + call ffi_closure_inner + + /* return values */ +#if FLTS + FLARG fa0, 0*FLTS(sp) + FLARG fa1, 1*FLTS(sp) +#endif + + LARG a0, 8*FLTS+0*PTRS(sp) + LARG a1, 8*FLTS+1*PTRS(sp) + + /* restore and return */ + LARG ra, FRAME_LEN-1*PTRS(sp) + .cfi_restore 1 + LARG fp, FRAME_LEN-2*PTRS(sp) + .cfi_restore 8 + addi sp, sp, FRAME_LEN + .cfi_def_cfa_offset 0 + ret + .cfi_endproc + .size ffi_closure_asm, .-ffi_closure_asm + +/* + ffi_go_closure_asm. Expects address of the passed-in ffi_go_closure in t2. + void ffi_closure_inner (ffi_cif *cif, + void (*fun) (ffi_cif *, void *, void **, void *), + void *user_data, + size_t *stackargs, struct call_context *regargs) +*/ + + .globl ffi_go_closure_asm + .hidden ffi_go_closure_asm + .type ffi_go_closure_asm, @function +ffi_go_closure_asm: + .cfi_startproc + + addi sp, sp, -FRAME_LEN + .cfi_def_cfa_offset FRAME_LEN + + /* make a frame */ + SARG fp, FRAME_LEN - 2*PTRS(sp) + .cfi_offset 8, -2*PTRS + SARG ra, FRAME_LEN - 1*PTRS(sp) + .cfi_offset 1, -1*PTRS + addi fp, sp, FRAME_LEN + + /* save arguments */ +#if FLTS + FSARG fa0, 0*FLTS(sp) + FSARG fa1, 1*FLTS(sp) + FSARG fa2, 2*FLTS(sp) + FSARG fa3, 3*FLTS(sp) + FSARG fa4, 4*FLTS(sp) + FSARG fa5, 5*FLTS(sp) + FSARG fa6, 6*FLTS(sp) + FSARG fa7, 7*FLTS(sp) +#endif + + SARG a0, 8*FLTS+0*PTRS(sp) + SARG a1, 8*FLTS+1*PTRS(sp) + SARG a2, 8*FLTS+2*PTRS(sp) + SARG a3, 8*FLTS+3*PTRS(sp) + SARG a4, 8*FLTS+4*PTRS(sp) + SARG a5, 8*FLTS+5*PTRS(sp) + SARG a6, 8*FLTS+6*PTRS(sp) + SARG a7, 8*FLTS+7*PTRS(sp) + + /* enter C */ + LARG a0, 1*PTRS(t2) + LARG a1, 2*PTRS(t2) + mv a2, t2 + addi a3, sp, FRAME_LEN + mv a4, sp + + call ffi_closure_inner + + /* return values */ +#if FLTS + FLARG fa0, 0*FLTS(sp) + FLARG fa1, 1*FLTS(sp) +#endif + + LARG a0, 8*FLTS+0*PTRS(sp) + LARG a1, 8*FLTS+1*PTRS(sp) + + /* restore and return */ + LARG ra, FRAME_LEN-1*PTRS(sp) + .cfi_restore 1 + LARG fp, FRAME_LEN-2*PTRS(sp) + .cfi_restore 8 + addi sp, sp, FRAME_LEN + .cfi_def_cfa_offset 0 + ret + .cfi_endproc + .size ffi_go_closure_asm, .-ffi_go_closure_asm diff --git a/native/libffi/src/s390/ffi.c b/native/libffi/src/s390/ffi.c index 520ec7c61f..c75c894d31 100644 --- a/native/libffi/src/s390/ffi.c +++ b/native/libffi/src/s390/ffi.c @@ -1,9 +1,9 @@ /* ----------------------------------------------------------------------- ffi.c - Copyright (c) 2000, 2007 Software AG Copyright (c) 2008 Red Hat, Inc - + S390 Foreign Function Interface - + Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the ``Software''), to deal in the Software without restriction, including @@ -11,10 +11,10 @@ distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: - + The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. - + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. @@ -27,24 +27,23 @@ /* Includes */ /* -------- */ /*====================================================================*/ - + #include #include - -#include -#include - +#include +#include "internal.h" + /*====================== End of Includes =============================*/ - + /*====================================================================*/ /* Defines */ /* ------- */ /*====================================================================*/ -/* Maximum number of GPRs available for argument passing. */ +/* Maximum number of GPRs available for argument passing. */ #define MAX_GPRARGS 5 -/* Maximum number of FPRs available for argument passing. */ +/* Maximum number of FPRs available for argument passing. */ #ifdef __s390x__ #define MAX_FPRARGS 4 #else @@ -54,47 +53,30 @@ /* Round to multiple of 16. */ #define ROUND_SIZE(size) (((size) + 15) & ~15) -/* If these values change, sysv.S must be adapted! */ -#define FFI390_RET_VOID 0 -#define FFI390_RET_STRUCT 1 -#define FFI390_RET_FLOAT 2 -#define FFI390_RET_DOUBLE 3 -#define FFI390_RET_INT32 4 -#define FFI390_RET_INT64 5 - /*===================== End of Defines ===============================*/ - -/*====================================================================*/ -/* Prototypes */ -/* ---------- */ -/*====================================================================*/ - -static void ffi_prep_args (unsigned char *, extended_cif *); -void -#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ > 2) -__attribute__ ((visibility ("hidden"))) -#endif -ffi_closure_helper_SYSV (ffi_closure *, unsigned long *, - unsigned long long *, unsigned long *); -/*====================== End of Prototypes ===========================*/ - /*====================================================================*/ /* Externals */ /* --------- */ /*====================================================================*/ - -extern void ffi_call_SYSV(unsigned, - extended_cif *, - void (*)(unsigned char *, extended_cif *), - unsigned, - void *, - void (*fn)(void)); + +struct call_frame +{ + void *back_chain; + void *eos; + unsigned long gpr_args[5]; + unsigned long gpr_save[9]; + unsigned long long fpr_args[4]; +}; + +extern void FFI_HIDDEN ffi_call_SYSV(struct call_frame *, unsigned, void *, + void (*fn)(void), void *); extern void ffi_closure_SYSV(void); - +extern void ffi_go_closure_SYSV(void); + /*====================== End of Externals ============================*/ - + /*====================================================================*/ /* */ /* Name - ffi_check_struct_type. */ @@ -103,7 +85,7 @@ extern void ffi_closure_SYSV(void); /* general purpose or floating point register. */ /* */ /*====================================================================*/ - + static int ffi_check_struct_type (ffi_type *arg) { @@ -111,7 +93,7 @@ ffi_check_struct_type (ffi_type *arg) /* If the struct has just one element, look at that element to find out whether to consider the struct as floating point. */ - while (arg->type == FFI_TYPE_STRUCT + while (arg->type == FFI_TYPE_STRUCT && arg->elements[0] && !arg->elements[1]) arg = arg->elements[0]; @@ -144,6 +126,7 @@ ffi_check_struct_type (ffi_type *arg) /* Other structs are passed via a pointer to the data. */ return FFI_TYPE_POINTER; } +<<<<<<< HEAD /*======================== End of Routine ============================*/ @@ -328,9 +311,11 @@ ffi_prep_args (unsigned char *stack, extended_cif *ecif) } } } +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 /*======================== End of Routine ============================*/ - + /*====================================================================*/ /* */ /* Name - ffi_prep_cif_machdep. */ @@ -338,8 +323,8 @@ ffi_prep_args (unsigned char *stack, extended_cif *ecif) /* Function - Perform machine dependent CIF processing. */ /* */ /*====================================================================*/ - -ffi_status + +ffi_status FFI_HIDDEN ffi_prep_cif_machdep(ffi_cif *cif) { size_t struct_size = 0; @@ -350,7 +335,7 @@ ffi_prep_cif_machdep(ffi_cif *cif) ffi_type **ptr; int i; - /* Determine return value handling. */ + /* Determine return value handling. */ switch (cif->rtype->type) { @@ -364,7 +349,7 @@ ffi_prep_cif_machdep(ffi_cif *cif) case FFI_TYPE_COMPLEX: cif->flags = FFI390_RET_STRUCT; n_gpr++; /* We need one GPR to pass the pointer. */ - break; + break; /* Floating point values are returned in fpr 0. */ case FFI_TYPE_FLOAT: @@ -403,14 +388,14 @@ ffi_prep_cif_machdep(ffi_cif *cif) cif->flags = FFI390_RET_INT32; #endif break; - + default: FFI_ASSERT (0); break; } /* Now for the arguments. */ - + for (ptr = cif->arg_types, i = cif->nargs; i > 0; i--, ptr++) @@ -438,7 +423,7 @@ ffi_prep_cif_machdep(ffi_cif *cif) } /* Now handle all primitive int/float data types. */ - switch (type) + switch (type) { /* The first MAX_FPRARGS floating point arguments go in FPRs, the rest overflow to the stack. */ @@ -449,7 +434,7 @@ ffi_prep_cif_machdep(ffi_cif *cif) else n_ov += sizeof (double) / sizeof (long); break; - + case FFI_TYPE_FLOAT: if (n_fpr < MAX_FPRARGS) n_fpr++; @@ -459,9 +444,9 @@ ffi_prep_cif_machdep(ffi_cif *cif) /* On 31-bit machines, 64-bit integers are passed in GPR pairs, if one is still available, or else on the stack. If only one - register is free, skip the register (it won't be used for any + register is free, skip the register (it won't be used for any subsequent argument either). */ - + #ifndef __s390x__ case FFI_TYPE_UINT64: case FFI_TYPE_SINT64: @@ -477,7 +462,7 @@ ffi_prep_cif_machdep(ffi_cif *cif) /* Everything else is passed in GPRs (until MAX_GPRARGS have been used) or overflows to the stack. */ - default: + default: if (n_gpr < MAX_GPRARGS) n_gpr++; else @@ -490,12 +475,12 @@ ffi_prep_cif_machdep(ffi_cif *cif) and temporary structure copies. */ cif->bytes = ROUND_SIZE (n_ov * sizeof (long)) + struct_size; - + return FFI_OK; } - + /*======================== End of Routine ============================*/ - + /*====================================================================*/ /* */ /* Name - ffi_call. */ @@ -503,42 +488,195 @@ ffi_prep_cif_machdep(ffi_cif *cif) /* Function - Call the FFI routine. */ /* */ /*====================================================================*/ - -void -ffi_call(ffi_cif *cif, - void (*fn)(void), - void *rvalue, - void **avalue) + +static void +ffi_call_int(ffi_cif *cif, + void (*fn)(void), + void *rvalue, + void **avalue, + void *closure) { int ret_type = cif->flags; - extended_cif ecif; - - ecif.cif = cif; - ecif.avalue = avalue; - ecif.rvalue = rvalue; + size_t rsize = 0, bytes = cif->bytes; + unsigned char *stack, *p_struct; + struct call_frame *frame; + unsigned long *p_ov, *p_gpr; + unsigned long long *p_fpr; + int n_fpr, n_gpr, n_ov, i, n; + ffi_type **arg_types; + + FFI_ASSERT (cif->abi == FFI_SYSV); /* If we don't have a return value, we need to fake one. */ if (rvalue == NULL) { - if (ret_type == FFI390_RET_STRUCT) - ecif.rvalue = alloca (cif->rtype->size); + if (ret_type & FFI390_RET_IN_MEM) + rsize = cif->rtype->size; else ret_type = FFI390_RET_VOID; - } + } + + /* The stack space will be filled with those areas: + + dummy structure return (highest addresses) + FPR argument register save area + GPR argument register save area + stack frame for ffi_call_SYSV + temporary struct copies + overflow argument area (lowest addresses) + + We set up the following pointers: + + p_fpr: bottom of the FPR area (growing upwards) + p_gpr: bottom of the GPR area (growing upwards) + p_ov: bottom of the overflow area (growing upwards) + p_struct: top of the struct copy area (growing downwards) + + All areas are kept aligned to twice the word size. + + Note that we're going to create the stack frame for both + ffi_call_SYSV _and_ the target function right here. This + works because we don't make any function calls with more + than 5 arguments (indeed only memcpy and ffi_call_SYSV), + and thus we don't have any stacked outgoing parameters. */ - switch (cif->abi) + stack = alloca (bytes + sizeof(struct call_frame) + rsize); + frame = (struct call_frame *)(stack + bytes); + if (rsize) + rvalue = frame + 1; + + /* Link the new frame back to the one from this function. */ + frame->back_chain = __builtin_frame_address (0); + + /* Fill in all of the argument stuff. */ + p_ov = (unsigned long *)stack; + p_struct = (unsigned char *)frame; + p_gpr = frame->gpr_args; + p_fpr = frame->fpr_args; + n_fpr = n_gpr = n_ov = 0; + + /* If we returning a structure then we set the first parameter register + to the address of where we are returning this structure. */ + if (cif->flags & FFI390_RET_IN_MEM) + p_gpr[n_gpr++] = (uintptr_t) rvalue; + + /* Now for the arguments. */ + arg_types = cif->arg_types; + for (i = 0, n = cif->nargs; i < n; ++i) { - case FFI_SYSV: - ffi_call_SYSV (cif->bytes, &ecif, ffi_prep_args, - ret_type, ecif.rvalue, fn); - break; - - default: - FFI_ASSERT (0); - break; + ffi_type *ty = arg_types[i]; + void *arg = avalue[i]; + int type = ty->type; + ffi_arg val; + + restart: + switch (type) + { + case FFI_TYPE_SINT8: + val = *(SINT8 *)arg; + goto do_int; + case FFI_TYPE_UINT8: + val = *(UINT8 *)arg; + goto do_int; + case FFI_TYPE_SINT16: + val = *(SINT16 *)arg; + goto do_int; + case FFI_TYPE_UINT16: + val = *(UINT16 *)arg; + goto do_int; + case FFI_TYPE_INT: + case FFI_TYPE_SINT32: + val = *(SINT32 *)arg; + goto do_int; + case FFI_TYPE_UINT32: + val = *(UINT32 *)arg; + goto do_int; + case FFI_TYPE_POINTER: + val = *(uintptr_t *)arg; + do_int: + *(n_gpr < MAX_GPRARGS ? p_gpr + n_gpr++ : p_ov + n_ov++) = val; + break; + + case FFI_TYPE_UINT64: + case FFI_TYPE_SINT64: +#ifdef __s390x__ + val = *(UINT64 *)arg; + goto do_int; +#else + if (n_gpr == MAX_GPRARGS-1) + n_gpr = MAX_GPRARGS; + if (n_gpr < MAX_GPRARGS) + p_gpr[n_gpr++] = ((UINT32 *) arg)[0], + p_gpr[n_gpr++] = ((UINT32 *) arg)[1]; + else + p_ov[n_ov++] = ((UINT32 *) arg)[0], + p_ov[n_ov++] = ((UINT32 *) arg)[1]; +#endif + break; + + case FFI_TYPE_DOUBLE: + if (n_fpr < MAX_FPRARGS) + p_fpr[n_fpr++] = *(UINT64 *) arg; + else + { +#ifdef __s390x__ + p_ov[n_ov++] = *(UINT64 *) arg; +#else + p_ov[n_ov++] = ((UINT32 *) arg)[0], + p_ov[n_ov++] = ((UINT32 *) arg)[1]; +#endif + } + break; + + case FFI_TYPE_FLOAT: + val = *(UINT32 *)arg; + if (n_fpr < MAX_FPRARGS) + p_fpr[n_fpr++] = (UINT64)val << 32; + else + p_ov[n_ov++] = val; + break; + + case FFI_TYPE_STRUCT: + /* Check how a structure type is passed. */ + type = ffi_check_struct_type (ty); + /* Some structures are passed via a type they contain. */ + if (type != FFI_TYPE_POINTER) + goto restart; + /* ... otherwise, passed by reference. fallthru. */ + +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE + case FFI_TYPE_LONGDOUBLE: + /* 16-byte long double is passed via reference. */ +#endif + case FFI_TYPE_COMPLEX: + /* Complex types are passed via reference. */ + p_struct -= ROUND_SIZE (ty->size); + memcpy (p_struct, arg, ty->size); + val = (uintptr_t)p_struct; + goto do_int; + + default: + FFI_ASSERT (0); + break; + } } + + ffi_call_SYSV (frame, ret_type & FFI360_RET_MASK, rvalue, fn, closure); } - + +void +ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +{ + ffi_call_int(cif, fn, rvalue, avalue, NULL); +} + +void +ffi_call_go (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) +{ + ffi_call_int(cif, fn, rvalue, avalue, closure); +} + /*======================== End of Routine ============================*/ /*====================================================================*/ @@ -548,9 +686,11 @@ ffi_call(ffi_cif *cif, /* Function - Call a FFI closure target function. */ /* */ /*====================================================================*/ - -void -ffi_closure_helper_SYSV (ffi_closure *closure, + +void FFI_HIDDEN +ffi_closure_helper_SYSV (ffi_cif *cif, + void (*fun)(ffi_cif*,void*,void**,void*), + void *user_data, unsigned long *p_gpr, unsigned long long *p_fpr, unsigned long *p_ov) @@ -569,21 +709,16 @@ ffi_closure_helper_SYSV (ffi_closure *closure, int i; /* Allocate buffer for argument list pointers. */ + p_arg = avalue = alloca (cif->nargs * sizeof (void *)); - p_arg = avalue = alloca (closure->cif->nargs * sizeof (void *)); - - /* If we returning a structure, pass the structure address - directly to the target function. Otherwise, have the target + /* If we returning a structure, pass the structure address + directly to the target function. Otherwise, have the target function store the return value to the GPR save area. */ - - if (closure->cif->flags == FFI390_RET_STRUCT) + if (cif->flags & FFI390_RET_IN_MEM) rvalue = (void *) p_gpr[n_gpr++]; /* Now for the arguments. */ - - for (ptr = closure->cif->arg_types, i = closure->cif->nargs; - i > 0; - i--, p_arg++, ptr++) + for (ptr = cif->arg_types, i = cif->nargs; i > 0; i--, p_arg++, ptr++) { int deref_struct_pointer = 0; int type = (*ptr)->type; @@ -602,7 +737,7 @@ ffi_closure_helper_SYSV (ffi_closure *closure, else type = ffi_check_struct_type (*ptr); - /* If we pass the struct via pointer, remember to + /* If we pass the struct via pointer, remember to retrieve the pointer later. */ if (type == FFI_TYPE_POINTER) deref_struct_pointer = 1; @@ -610,30 +745,32 @@ ffi_closure_helper_SYSV (ffi_closure *closure, /* Pointers are passed like UINTs of the same size. */ if (type == FFI_TYPE_POINTER) + { #ifdef __s390x__ - type = FFI_TYPE_UINT64; + type = FFI_TYPE_UINT64; #else - type = FFI_TYPE_UINT32; + type = FFI_TYPE_UINT32; #endif + } /* Now handle all primitive int/float data types. */ - switch (type) + switch (type) { case FFI_TYPE_DOUBLE: if (n_fpr < MAX_FPRARGS) *p_arg = &p_fpr[n_fpr++]; else - *p_arg = &p_ov[n_ov], + *p_arg = &p_ov[n_ov], n_ov += sizeof (double) / sizeof (long); break; - + case FFI_TYPE_FLOAT: if (n_fpr < MAX_FPRARGS) *p_arg = &p_fpr[n_fpr++]; else *p_arg = (char *)&p_ov[n_ov++] + sizeof (long) - 4; break; - + case FFI_TYPE_UINT64: case FFI_TYPE_SINT64: #ifdef __s390x__ @@ -650,7 +787,7 @@ ffi_closure_helper_SYSV (ffi_closure *closure, *p_arg = &p_ov[n_ov], n_ov += 2; #endif break; - + case FFI_TYPE_INT: case FFI_TYPE_UINT32: case FFI_TYPE_SINT32: @@ -659,7 +796,7 @@ ffi_closure_helper_SYSV (ffi_closure *closure, else *p_arg = (char *)&p_ov[n_ov++] + sizeof (long) - 4; break; - + case FFI_TYPE_UINT16: case FFI_TYPE_SINT16: if (n_gpr < MAX_GPRARGS) @@ -675,7 +812,7 @@ ffi_closure_helper_SYSV (ffi_closure *closure, else *p_arg = (char *)&p_ov[n_ov++] + sizeof (long) - 1; break; - + default: FFI_ASSERT (0); break; @@ -689,10 +826,10 @@ ffi_closure_helper_SYSV (ffi_closure *closure, /* Call the target function. */ - (closure->fun) (closure->cif, rvalue, avalue, closure->user_data); + (fun) (cif, rvalue, avalue, user_data); /* Convert the return value. */ - switch (closure->cif->rtype->type) + switch (cif->rtype->type) { /* Void is easy, and so is struct. */ case FFI_TYPE_VOID: @@ -743,7 +880,7 @@ ffi_closure_helper_SYSV (ffi_closure *closure, break; } } - + /*======================== End of Routine ============================*/ /*====================================================================*/ @@ -753,7 +890,7 @@ ffi_closure_helper_SYSV (ffi_closure *closure, /* Function - Prepare a FFI closure. */ /* */ /*====================================================================*/ - + ffi_status ffi_prep_closure_loc (ffi_closure *closure, ffi_cif *cif, @@ -761,32 +898,46 @@ ffi_prep_closure_loc (ffi_closure *closure, void *user_data, void *codeloc) { + static unsigned short const template[] = { + 0x0d10, /* basr %r1,0 */ +#ifndef __s390x__ + 0x9801, 0x1006, /* lm %r0,%r1,6(%r1) */ +#else + 0xeb01, 0x100e, 0x0004, /* lmg %r0,%r1,14(%r1) */ +#endif + 0x07f1 /* br %r1 */ + }; + + unsigned long *tramp = (unsigned long *)&closure->tramp; + if (cif->abi != FFI_SYSV) return FFI_BAD_ABI; -#ifndef __s390x__ - *(short *)&closure->tramp [0] = 0x0d10; /* basr %r1,0 */ - *(short *)&closure->tramp [2] = 0x9801; /* lm %r0,%r1,6(%r1) */ - *(short *)&closure->tramp [4] = 0x1006; - *(short *)&closure->tramp [6] = 0x07f1; /* br %r1 */ - *(long *)&closure->tramp [8] = (long)codeloc; - *(long *)&closure->tramp[12] = (long)&ffi_closure_SYSV; -#else - *(short *)&closure->tramp [0] = 0x0d10; /* basr %r1,0 */ - *(short *)&closure->tramp [2] = 0xeb01; /* lmg %r0,%r1,14(%r1) */ - *(short *)&closure->tramp [4] = 0x100e; - *(short *)&closure->tramp [6] = 0x0004; - *(short *)&closure->tramp [8] = 0x07f1; /* br %r1 */ - *(long *)&closure->tramp[16] = (long)codeloc; - *(long *)&closure->tramp[24] = (long)&ffi_closure_SYSV; -#endif - + memcpy (tramp, template, sizeof(template)); + tramp[2] = (unsigned long)codeloc; + tramp[3] = (unsigned long)&ffi_closure_SYSV; + closure->cif = cif; - closure->user_data = user_data; closure->fun = fun; - + closure->user_data = user_data; + return FFI_OK; } /*======================== End of Routine ============================*/ - + +/* Build a Go language closure. */ + +ffi_status +ffi_prep_go_closure (ffi_go_closure *closure, ffi_cif *cif, + void (*fun)(ffi_cif*,void*,void**,void*)) +{ + if (cif->abi != FFI_SYSV) + return FFI_BAD_ABI; + + closure->tramp = ffi_go_closure_SYSV; + closure->cif = cif; + closure->fun = fun; + + return FFI_OK; +} diff --git a/native/libffi/src/s390/ffitarget.h b/native/libffi/src/s390/ffitarget.h index 0e4868a1c4..d8a4ee4bf1 100644 --- a/native/libffi/src/s390/ffitarget.h +++ b/native/libffi/src/s390/ffitarget.h @@ -58,6 +58,7 @@ typedef enum ffi_abi { /* ---- Definitions for closures ----------------------------------------- */ #define FFI_CLOSURES 1 +#define FFI_GO_CLOSURES 1 #ifdef S390X #define FFI_TRAMPOLINE_SIZE 32 #else diff --git a/native/libffi/src/s390/internal.h b/native/libffi/src/s390/internal.h new file mode 100644 index 0000000000..b8755786f9 --- /dev/null +++ b/native/libffi/src/s390/internal.h @@ -0,0 +1,11 @@ +/* If these values change, sysv.S must be adapted! */ +#define FFI390_RET_DOUBLE 0 +#define FFI390_RET_FLOAT 1 +#define FFI390_RET_INT64 2 +#define FFI390_RET_INT32 3 +#define FFI390_RET_VOID 4 + +#define FFI360_RET_MASK 7 +#define FFI390_RET_IN_MEM 8 + +#define FFI390_RET_STRUCT (FFI390_RET_VOID | FFI390_RET_IN_MEM) diff --git a/native/libffi/src/s390/sysv.S b/native/libffi/src/s390/sysv.S index 4731a3177b..c4b5006aed 100644 --- a/native/libffi/src/s390/sysv.S +++ b/native/libffi/src/s390/sysv.S @@ -1,9 +1,9 @@ /* ----------------------------------------------------------------------- sysv.S - Copyright (c) 2000 Software AG Copyright (c) 2008 Red Hat, Inc. - + S390 Foreign Function Interface - + Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the ``Software''), to deal in the Software without restriction, including @@ -11,10 +11,10 @@ distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: - + The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. - + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND @@ -29,405 +29,296 @@ #include #include + .text + #ifndef __s390x__ - -.text - - # r2: cif->bytes - # r3: &ecif - # r4: ffi_prep_args - # r5: ret_type - # r6: ecif.rvalue - # ov: fn - + + # r2: frame + # r3: ret_type + # r4: ret_addr + # r5: fun + # r6: closure + # This assumes we are using gas. + .balign 8 .globl ffi_call_SYSV + FFI_HIDDEN(ffi_call_SYSV) .type ffi_call_SYSV,%function ffi_call_SYSV: -.LFB1: - stm %r6,%r15,24(%r15) # Save registers -.LCFI0: - basr %r13,0 # Set up base register -.Lbase: - lr %r11,%r15 # Set up frame pointer -.LCFI1: - sr %r15,%r2 - ahi %r15,-96-48 # Allocate stack - lr %r8,%r6 # Save ecif.rvalue - sr %r9,%r9 - ic %r9,.Ltable-.Lbase(%r13,%r5) # Load epilog address - l %r7,96(%r11) # Load function address - st %r11,0(%r15) # Set up back chain - ahi %r11,-48 # Register save area -.LCFI2: - - la %r2,96(%r15) # Save area - # r3 already holds &ecif - basr %r14,%r4 # Call ffi_prep_args - - lm %r2,%r6,0(%r11) # Load arguments - ld %f0,32(%r11) - ld %f2,40(%r11) - la %r14,0(%r13,%r9) # Set return address - br %r7 # ... and call function - -.LretNone: # Return void - l %r4,48+56(%r11) - lm %r6,%r15,48+24(%r11) - br %r4 - -.LretFloat: - l %r4,48+56(%r11) - ste %f0,0(%r8) # Return float - lm %r6,%r15,48+24(%r11) - br %r4 - -.LretDouble: - l %r4,48+56(%r11) - std %f0,0(%r8) # Return double - lm %r6,%r15,48+24(%r11) - br %r4 - -.LretInt32: - l %r4,48+56(%r11) - st %r2,0(%r8) # Return int - lm %r6,%r15,48+24(%r11) - br %r4 - -.LretInt64: - l %r4,48+56(%r11) - stm %r2,%r3,0(%r8) # Return long long - lm %r6,%r15,48+24(%r11) - br %r4 - + .cfi_startproc + st %r6,44(%r2) # Save registers + stm %r12,%r14,48(%r2) + lr %r13,%r2 # Install frame pointer + .cfi_rel_offset r6, 44 + .cfi_rel_offset r12, 48 + .cfi_rel_offset r13, 52 + .cfi_rel_offset r14, 56 + .cfi_def_cfa_register r13 + st %r2,0(%r15) # Set up back chain + sla %r3,3 # ret_type *= 8 + lr %r12,%r4 # Save ret_addr + lr %r1,%r5 # Save fun + lr %r0,%r6 # Install static chain + + # Set return address, so that there is only one indirect jump. +#ifdef HAVE_AS_S390_ZARCH + larl %r14,.Ltable + ar %r14,%r3 +#else + basr %r14,0 +0: la %r14,.Ltable-0b(%r14,%r3) +#endif + + lm %r2,%r6,8(%r13) # Load arguments + ld %f0,64(%r13) + ld %f2,72(%r13) + br %r1 # ... and call function + + .balign 8 .Ltable: - .byte .LretNone-.Lbase # FFI390_RET_VOID - .byte .LretNone-.Lbase # FFI390_RET_STRUCT - .byte .LretFloat-.Lbase # FFI390_RET_FLOAT - .byte .LretDouble-.Lbase # FFI390_RET_DOUBLE - .byte .LretInt32-.Lbase # FFI390_RET_INT32 - .byte .LretInt64-.Lbase # FFI390_RET_INT64 +# FFI390_RET_DOUBLE + std %f0,0(%r12) + j .Ldone + + .balign 8 +# FFI390_RET_FLOAT + ste %f0,0(%r12) + j .Ldone + + .balign 8 +# FFI390_RET_INT64 + st %r3,4(%r12) + nop + # fallthru + + .balign 8 +# FFI390_RET_INT32 + st %r2,0(%r12) + nop + # fallthru + + .balign 8 +# FFI390_RET_VOID +.Ldone: + l %r14,56(%r13) + l %r12,48(%r13) + l %r6,44(%r13) + l %r13,52(%r13) + .cfi_restore 14 + .cfi_restore 13 + .cfi_restore 12 + .cfi_restore 6 + .cfi_def_cfa r15, 96 + br %r14 + .cfi_endproc + .size ffi_call_SYSV,.-ffi_call_SYSV -.LFE1: -.ffi_call_SYSV_end: - .size ffi_call_SYSV,.ffi_call_SYSV_end-ffi_call_SYSV + .balign 8 + .globl ffi_go_closure_SYSV + FFI_HIDDEN(ffi_go_closure_SYSV) + .type ffi_go_closure_SYSV,%function +ffi_go_closure_SYSV: + .cfi_startproc + stm %r2,%r6,8(%r15) # Save arguments + lr %r4,%r0 # Load closure -> user_data + l %r2,4(%r4) # ->cif + l %r3,8(%r4) # ->fun + j .Ldoclosure + .cfi_endproc + .balign 8 .globl ffi_closure_SYSV + FFI_HIDDEN(ffi_closure_SYSV) .type ffi_closure_SYSV,%function ffi_closure_SYSV: -.LFB2: + .cfi_startproc + stm %r2,%r6,8(%r15) # Save arguments + lr %r4,%r0 # Closure + l %r2,16(%r4) # ->cif + l %r3,20(%r4) # ->fun + l %r4,24(%r4) # ->user_data +.Ldoclosure: stm %r12,%r15,48(%r15) # Save registers -.LCFI10: + lr %r12,%r15 + .cfi_def_cfa_register r12 + .cfi_rel_offset r6, 24 + .cfi_rel_offset r12, 48 + .cfi_rel_offset r13, 52 + .cfi_rel_offset r14, 56 + .cfi_rel_offset r15, 60 +#ifndef HAVE_AS_S390_ZARCH basr %r13,0 # Set up base register .Lcbase: - stm %r2,%r6,8(%r15) # Save arguments - std %f0,64(%r15) - std %f2,72(%r15) - lr %r1,%r15 # Set up stack frame - ahi %r15,-96 -.LCFI11: - l %r12,.Lchelper-.Lcbase(%r13) # Get helper function - lr %r2,%r0 # Closure - la %r3,8(%r1) # GPRs - la %r4,64(%r1) # FPRs - la %r5,96(%r1) # Overflow - st %r1,0(%r15) # Set up back chain - - bas %r14,0(%r12,%r13) # Call helper - - l %r4,96+56(%r15) - ld %f0,96+64(%r15) # Load return registers - lm %r2,%r3,96+8(%r15) - lm %r12,%r15,96+48(%r15) - br %r4 + l %r1,.Lchelper-.Lcbase(%r13) # Get helper function +#endif + ahi %r15,-96-8 # Set up stack frame + st %r12,0(%r15) # Set up back chain + + std %f0,64(%r12) # Save fp arguments + std %f2,72(%r12) + + la %r5,96(%r12) # Overflow + st %r5,96(%r15) + la %r6,64(%r12) # FPRs + la %r5,8(%r12) # GPRs +#ifdef HAVE_AS_S390_ZARCH + brasl %r14,ffi_closure_helper_SYSV +#else + bas %r14,0(%r1,%r13) # Call helper +#endif + lr %r15,%r12 + .cfi_def_cfa_register r15 + lm %r12,%r14,48(%r12) # Restore saved registers + l %r6,24(%r15) + ld %f0,64(%r15) # Load return registers + lm %r2,%r3,8(%r15) + br %r14 + .cfi_endproc + +#ifndef HAVE_AS_S390_ZARCH .align 4 .Lchelper: .long ffi_closure_helper_SYSV-.Lcbase +#endif -.LFE2: - -.ffi_closure_SYSV_end: - .size ffi_closure_SYSV,.ffi_closure_SYSV_end-ffi_closure_SYSV - - - .section .eh_frame,EH_FRAME_FLAGS,@progbits -.Lframe1: - .4byte .LECIE1-.LSCIE1 # Length of Common Information Entry -.LSCIE1: - .4byte 0x0 # CIE Identifier Tag - .byte 0x1 # CIE Version - .ascii "zR\0" # CIE Augmentation - .uleb128 0x1 # CIE Code Alignment Factor - .sleb128 -4 # CIE Data Alignment Factor - .byte 0xe # CIE RA Column - .uleb128 0x1 # Augmentation size - .byte 0x1b # FDE Encoding (pcrel sdata4) - .byte 0xc # DW_CFA_def_cfa - .uleb128 0xf - .uleb128 0x60 - .align 4 -.LECIE1: -.LSFDE1: - .4byte .LEFDE1-.LASFDE1 # FDE Length -.LASFDE1: - .4byte .LASFDE1-.Lframe1 # FDE CIE offset - .4byte .LFB1-. # FDE initial location - .4byte .LFE1-.LFB1 # FDE address range - .uleb128 0x0 # Augmentation size - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI0-.LFB1 - .byte 0x8f # DW_CFA_offset, column 0xf - .uleb128 0x9 - .byte 0x8e # DW_CFA_offset, column 0xe - .uleb128 0xa - .byte 0x8d # DW_CFA_offset, column 0xd - .uleb128 0xb - .byte 0x8c # DW_CFA_offset, column 0xc - .uleb128 0xc - .byte 0x8b # DW_CFA_offset, column 0xb - .uleb128 0xd - .byte 0x8a # DW_CFA_offset, column 0xa - .uleb128 0xe - .byte 0x89 # DW_CFA_offset, column 0x9 - .uleb128 0xf - .byte 0x88 # DW_CFA_offset, column 0x8 - .uleb128 0x10 - .byte 0x87 # DW_CFA_offset, column 0x7 - .uleb128 0x11 - .byte 0x86 # DW_CFA_offset, column 0x6 - .uleb128 0x12 - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI1-.LCFI0 - .byte 0xd # DW_CFA_def_cfa_register - .uleb128 0xb - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI2-.LCFI1 - .byte 0xe # DW_CFA_def_cfa_offset - .uleb128 0x90 - .align 4 -.LEFDE1: -.LSFDE2: - .4byte .LEFDE2-.LASFDE2 # FDE Length -.LASFDE2: - .4byte .LASFDE2-.Lframe1 # FDE CIE offset - .4byte .LFB2-. # FDE initial location - .4byte .LFE2-.LFB2 # FDE address range - .uleb128 0x0 # Augmentation size - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI10-.LFB2 - .byte 0x8f # DW_CFA_offset, column 0xf - .uleb128 0x9 - .byte 0x8e # DW_CFA_offset, column 0xe - .uleb128 0xa - .byte 0x8d # DW_CFA_offset, column 0xd - .uleb128 0xb - .byte 0x8c # DW_CFA_offset, column 0xc - .uleb128 0xc - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI11-.LCFI10 - .byte 0xe # DW_CFA_def_cfa_offset - .uleb128 0xc0 - .align 4 -.LEFDE2: + .size ffi_closure_SYSV,.-ffi_closure_SYSV #else - -.text - - # r2: cif->bytes - # r3: &ecif - # r4: ffi_prep_args - # r5: ret_type - # r6: ecif.rvalue - # ov: fn - + + # r2: frame + # r3: ret_type + # r4: ret_addr + # r5: fun + # r6: closure + # This assumes we are using gas. + .balign 8 .globl ffi_call_SYSV + FFI_HIDDEN(ffi_call_SYSV) .type ffi_call_SYSV,%function ffi_call_SYSV: -.LFB1: - stmg %r6,%r15,48(%r15) # Save registers -.LCFI0: - larl %r13,.Lbase # Set up base register - lgr %r11,%r15 # Set up frame pointer -.LCFI1: - sgr %r15,%r2 - aghi %r15,-160-80 # Allocate stack - lgr %r8,%r6 # Save ecif.rvalue - llgc %r9,.Ltable-.Lbase(%r13,%r5) # Load epilog address - lg %r7,160(%r11) # Load function address - stg %r11,0(%r15) # Set up back chain - aghi %r11,-80 # Register save area -.LCFI2: - - la %r2,160(%r15) # Save area - # r3 already holds &ecif - basr %r14,%r4 # Call ffi_prep_args - - lmg %r2,%r6,0(%r11) # Load arguments - ld %f0,48(%r11) - ld %f2,56(%r11) - ld %f4,64(%r11) - ld %f6,72(%r11) - la %r14,0(%r13,%r9) # Set return address - br %r7 # ... and call function - -.Lbase: -.LretNone: # Return void - lg %r4,80+112(%r11) - lmg %r6,%r15,80+48(%r11) - br %r4 - -.LretFloat: - lg %r4,80+112(%r11) - ste %f0,0(%r8) # Return float - lmg %r6,%r15,80+48(%r11) - br %r4 - -.LretDouble: - lg %r4,80+112(%r11) - std %f0,0(%r8) # Return double - lmg %r6,%r15,80+48(%r11) - br %r4 - -.LretInt32: - lg %r4,80+112(%r11) - st %r2,0(%r8) # Return int - lmg %r6,%r15,80+48(%r11) - br %r4 - -.LretInt64: - lg %r4,80+112(%r11) - stg %r2,0(%r8) # Return long - lmg %r6,%r15,80+48(%r11) - br %r4 - + .cfi_startproc + stg %r6,88(%r2) # Save registers + stmg %r12,%r14,96(%r2) + lgr %r13,%r2 # Install frame pointer + .cfi_rel_offset r6, 88 + .cfi_rel_offset r12, 96 + .cfi_rel_offset r13, 104 + .cfi_rel_offset r14, 112 + .cfi_def_cfa_register r13 + stg %r2,0(%r15) # Set up back chain + larl %r14,.Ltable # Set up return address + slag %r3,%r3,3 # ret_type *= 8 + lgr %r12,%r4 # Save ret_addr + lgr %r1,%r5 # Save fun + lgr %r0,%r6 # Install static chain + agr %r14,%r3 + lmg %r2,%r6,16(%r13) # Load arguments + ld %f0,128(%r13) + ld %f2,136(%r13) + ld %f4,144(%r13) + ld %f6,152(%r13) + br %r1 # ... and call function + + .balign 8 .Ltable: - .byte .LretNone-.Lbase # FFI390_RET_VOID - .byte .LretNone-.Lbase # FFI390_RET_STRUCT - .byte .LretFloat-.Lbase # FFI390_RET_FLOAT - .byte .LretDouble-.Lbase # FFI390_RET_DOUBLE - .byte .LretInt32-.Lbase # FFI390_RET_INT32 - .byte .LretInt64-.Lbase # FFI390_RET_INT64 +# FFI390_RET_DOUBLE + std %f0,0(%r12) + j .Ldone + + .balign 8 +# FFI390_RET_DOUBLE + ste %f0,0(%r12) + j .Ldone + + .balign 8 +# FFI390_RET_INT64 + stg %r2,0(%r12) + + .balign 8 +# FFI390_RET_INT32 + # Never used, as we always store type ffi_arg. + # But the stg above is 6 bytes and we cannot + # jump around this case, so fall through. + nop + nop + + .balign 8 +# FFI390_RET_VOID +.Ldone: + lg %r14,112(%r13) + lg %r12,96(%r13) + lg %r6,88(%r13) + lg %r13,104(%r13) + .cfi_restore r14 + .cfi_restore r13 + .cfi_restore r12 + .cfi_restore r6 + .cfi_def_cfa r15, 160 + br %r14 + .cfi_endproc + .size ffi_call_SYSV,.-ffi_call_SYSV -.LFE1: -.ffi_call_SYSV_end: - .size ffi_call_SYSV,.ffi_call_SYSV_end-ffi_call_SYSV + + .balign 8 + .globl ffi_go_closure_SYSV + FFI_HIDDEN(ffi_go_closure_SYSV) + .type ffi_go_closure_SYSV,%function +ffi_go_closure_SYSV: + .cfi_startproc + stmg %r2,%r6,16(%r15) # Save arguments + lgr %r4,%r0 # Load closure -> user_data + lg %r2,8(%r4) # ->cif + lg %r3,16(%r4) # ->fun + j .Ldoclosure + .cfi_endproc + .size ffi_go_closure_SYSV,.-ffi_go_closure_SYSV + .balign 8 .globl ffi_closure_SYSV + FFI_HIDDEN(ffi_closure_SYSV) .type ffi_closure_SYSV,%function ffi_closure_SYSV: -.LFB2: - stmg %r14,%r15,112(%r15) # Save registers -.LCFI10: + .cfi_startproc stmg %r2,%r6,16(%r15) # Save arguments - std %f0,128(%r15) - std %f2,136(%r15) - std %f4,144(%r15) - std %f6,152(%r15) - lgr %r1,%r15 # Set up stack frame - aghi %r15,-160 -.LCFI11: - lgr %r2,%r0 # Closure - la %r3,16(%r1) # GPRs - la %r4,128(%r1) # FPRs - la %r5,160(%r1) # Overflow - stg %r1,0(%r15) # Set up back chain - + lgr %r4,%r0 # Load closure + lg %r2,32(%r4) # ->cif + lg %r3,40(%r4) # ->fun + lg %r4,48(%r4) # ->user_data +.Ldoclosure: + stmg %r13,%r15,104(%r15) # Save registers + lgr %r13,%r15 + .cfi_def_cfa_register r13 + .cfi_rel_offset r6, 48 + .cfi_rel_offset r13, 104 + .cfi_rel_offset r14, 112 + .cfi_rel_offset r15, 120 + aghi %r15,-160-16 # Set up stack frame + stg %r13,0(%r15) # Set up back chain + + std %f0,128(%r13) # Save fp arguments + std %f2,136(%r13) + std %f4,144(%r13) + std %f6,152(%r13) + la %r5,160(%r13) # Overflow + stg %r5,160(%r15) + la %r6,128(%r13) # FPRs + la %r5,16(%r13) # GPRs brasl %r14,ffi_closure_helper_SYSV # Call helper - lg %r14,160+112(%r15) - ld %f0,160+128(%r15) # Load return registers - lg %r2,160+16(%r15) - la %r15,160(%r15) + lgr %r15,%r13 + .cfi_def_cfa_register r15 + lmg %r13,%r14,104(%r13) # Restore saved registers + lg %r6,48(%r15) + ld %f0,128(%r15) # Load return registers + lg %r2,16(%r15) br %r14 -.LFE2: - -.ffi_closure_SYSV_end: - .size ffi_closure_SYSV,.ffi_closure_SYSV_end-ffi_closure_SYSV - - - - .section .eh_frame,EH_FRAME_FLAGS,@progbits -.Lframe1: - .4byte .LECIE1-.LSCIE1 # Length of Common Information Entry -.LSCIE1: - .4byte 0x0 # CIE Identifier Tag - .byte 0x1 # CIE Version - .ascii "zR\0" # CIE Augmentation - .uleb128 0x1 # CIE Code Alignment Factor - .sleb128 -8 # CIE Data Alignment Factor - .byte 0xe # CIE RA Column - .uleb128 0x1 # Augmentation size - .byte 0x1b # FDE Encoding (pcrel sdata4) - .byte 0xc # DW_CFA_def_cfa - .uleb128 0xf - .uleb128 0xa0 - .align 8 -.LECIE1: -.LSFDE1: - .4byte .LEFDE1-.LASFDE1 # FDE Length -.LASFDE1: - .4byte .LASFDE1-.Lframe1 # FDE CIE offset - .4byte .LFB1-. # FDE initial location - .4byte .LFE1-.LFB1 # FDE address range - .uleb128 0x0 # Augmentation size - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI0-.LFB1 - .byte 0x8f # DW_CFA_offset, column 0xf - .uleb128 0x5 - .byte 0x8e # DW_CFA_offset, column 0xe - .uleb128 0x6 - .byte 0x8d # DW_CFA_offset, column 0xd - .uleb128 0x7 - .byte 0x8c # DW_CFA_offset, column 0xc - .uleb128 0x8 - .byte 0x8b # DW_CFA_offset, column 0xb - .uleb128 0x9 - .byte 0x8a # DW_CFA_offset, column 0xa - .uleb128 0xa - .byte 0x89 # DW_CFA_offset, column 0x9 - .uleb128 0xb - .byte 0x88 # DW_CFA_offset, column 0x8 - .uleb128 0xc - .byte 0x87 # DW_CFA_offset, column 0x7 - .uleb128 0xd - .byte 0x86 # DW_CFA_offset, column 0x6 - .uleb128 0xe - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI1-.LCFI0 - .byte 0xd # DW_CFA_def_cfa_register - .uleb128 0xb - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI2-.LCFI1 - .byte 0xe # DW_CFA_def_cfa_offset - .uleb128 0xf0 - .align 8 -.LEFDE1: -.LSFDE2: - .4byte .LEFDE2-.LASFDE2 # FDE Length -.LASFDE2: - .4byte .LASFDE2-.Lframe1 # FDE CIE offset - .4byte .LFB2-. # FDE initial location - .4byte .LFE2-.LFB2 # FDE address range - .uleb128 0x0 # Augmentation size - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI10-.LFB2 - .byte 0x8f # DW_CFA_offset, column 0xf - .uleb128 0x5 - .byte 0x8e # DW_CFA_offset, column 0xe - .uleb128 0x6 - .byte 0x4 # DW_CFA_advance_loc4 - .4byte .LCFI11-.LCFI10 - .byte 0xe # DW_CFA_def_cfa_offset - .uleb128 0x140 - .align 8 -.LEFDE2: - -#endif + .cfi_endproc + .size ffi_closure_SYSV,.-ffi_closure_SYSV +#endif /* !s390x */ #if defined __ELF__ && defined __linux__ .section .note.GNU-stack,"",@progbits diff --git a/native/libffi/src/sparc/ffi.c b/native/libffi/src/sparc/ffi.c index 9f0fded489..9e406d0af6 100644 --- a/native/libffi/src/sparc/ffi.c +++ b/native/libffi/src/sparc/ffi.c @@ -27,655 +27,442 @@ #include #include - #include +#include "internal.h" +#ifndef SPARC64 -/* ffi_prep_args is called by the assembly routine once stack space - has been allocated for the function's arguments */ - -void ffi_prep_args_v8(char *stack, extended_cif *ecif) -{ - int i; - void **p_argv; - char *argp; - ffi_type **p_arg; - - /* Skip 16 words for the window save area */ - argp = stack + 16*sizeof(int); - - /* This should only really be done when we are returning a structure, - however, it's faster just to do it all the time... - - if ( ecif->cif->rtype->type == FFI_TYPE_STRUCT ) */ - *(int *) argp = (long)ecif->rvalue; - - /* And 1 word for the structure return value. */ - argp += sizeof(int); - -#ifdef USING_PURIFY - /* Purify will probably complain in our assembly routine, unless we - zero out this memory. */ - - ((int*)argp)[0] = 0; - ((int*)argp)[1] = 0; - ((int*)argp)[2] = 0; - ((int*)argp)[3] = 0; - ((int*)argp)[4] = 0; - ((int*)argp)[5] = 0; -#endif - - p_argv = ecif->avalue; - - for (i = ecif->cif->nargs, p_arg = ecif->cif->arg_types; i; i--, p_arg++) - { - size_t z; - - if ((*p_arg)->type == FFI_TYPE_STRUCT -#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE - || (*p_arg)->type == FFI_TYPE_LONGDOUBLE -#endif - ) - { - *(unsigned int *) argp = (unsigned long)(* p_argv); - z = sizeof(int); - } - else - { - z = (*p_arg)->size; - if (z < sizeof(int)) - { - z = sizeof(int); - switch ((*p_arg)->type) - { - case FFI_TYPE_SINT8: - *(signed int *) argp = *(SINT8 *)(* p_argv); - break; - - case FFI_TYPE_UINT8: - *(unsigned int *) argp = *(UINT8 *)(* p_argv); - break; - - case FFI_TYPE_SINT16: - *(signed int *) argp = *(SINT16 *)(* p_argv); - break; - - case FFI_TYPE_UINT16: - *(unsigned int *) argp = *(UINT16 *)(* p_argv); - break; - - default: - FFI_ASSERT(0); - } - } - else - { - memcpy(argp, *p_argv, z); - } - } - p_argv++; - argp += z; - } - - return; -} - -int ffi_prep_args_v9(char *stack, extended_cif *ecif) -{ - int i, ret = 0; - int tmp; - void **p_argv; - char *argp; - ffi_type **p_arg; - - tmp = 0; - - /* Skip 16 words for the window save area */ - argp = stack + 16*sizeof(long long); - -#ifdef USING_PURIFY - /* Purify will probably complain in our assembly routine, unless we - zero out this memory. */ - - ((long long*)argp)[0] = 0; - ((long long*)argp)[1] = 0; - ((long long*)argp)[2] = 0; - ((long long*)argp)[3] = 0; - ((long long*)argp)[4] = 0; - ((long long*)argp)[5] = 0; -#endif - - p_argv = ecif->avalue; - - if (ecif->cif->rtype->type == FFI_TYPE_STRUCT && - ecif->cif->rtype->size > 32) - { - *(unsigned long long *) argp = (unsigned long)ecif->rvalue; - argp += sizeof(long long); - tmp = 1; - } - - for (i = 0, p_arg = ecif->cif->arg_types; i < ecif->cif->nargs; - i++, p_arg++) - { - size_t z; - - z = (*p_arg)->size; - switch ((*p_arg)->type) - { - case FFI_TYPE_STRUCT: - if (z > 16) - { - /* For structures larger than 16 bytes we pass reference. */ - *(unsigned long long *) argp = (unsigned long)* p_argv; - argp += sizeof(long long); - tmp++; - p_argv++; - continue; - } - /* FALLTHROUGH */ - case FFI_TYPE_FLOAT: - case FFI_TYPE_DOUBLE: +/* Force FFI_TYPE_LONGDOUBLE to be different than FFI_TYPE_DOUBLE; + all further uses in this file will refer to the 128-bit type. */ #if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE - case FFI_TYPE_LONGDOUBLE: +# if FFI_TYPE_LONGDOUBLE != 4 +# error FFI_TYPE_LONGDOUBLE out of date +# endif +#else +# undef FFI_TYPE_LONGDOUBLE +# define FFI_TYPE_LONGDOUBLE 4 #endif - ret = 1; /* We should promote into FP regs as well as integer. */ - break; - } - if (z < sizeof(long long)) - { - switch ((*p_arg)->type) - { - case FFI_TYPE_SINT8: - *(signed long long *) argp = *(SINT8 *)(* p_argv); - break; - - case FFI_TYPE_UINT8: - *(unsigned long long *) argp = *(UINT8 *)(* p_argv); - break; - - case FFI_TYPE_SINT16: - *(signed long long *) argp = *(SINT16 *)(* p_argv); - break; - - case FFI_TYPE_UINT16: - *(unsigned long long *) argp = *(UINT16 *)(* p_argv); - break; - - case FFI_TYPE_SINT32: - *(signed long long *) argp = *(SINT32 *)(* p_argv); - break; - - case FFI_TYPE_UINT32: - *(unsigned long long *) argp = *(UINT32 *)(* p_argv); - break; - - case FFI_TYPE_FLOAT: - *(float *) (argp + 4) = *(FLOAT32 *)(* p_argv); /* Right justify */ - break; - - case FFI_TYPE_STRUCT: - memcpy(argp, *p_argv, z); - break; - - default: - FFI_ASSERT(0); - } - z = sizeof(long long); - tmp++; - } - else if (z == sizeof(long long)) - { - memcpy(argp, *p_argv, z); - z = sizeof(long long); - tmp++; - } - else - { - if ((tmp & 1) && (*p_arg)->alignment > 8) - { - tmp++; - argp += sizeof(long long); - } - memcpy(argp, *p_argv, z); - z = 2 * sizeof(long long); - tmp += 2; - } - p_argv++; - argp += z; - } - - return ret; -} /* Perform machine dependent cif processing */ -ffi_status ffi_prep_cif_machdep(ffi_cif *cif) +ffi_status FFI_HIDDEN +ffi_prep_cif_machdep(ffi_cif *cif) { - int wordsize; - - if (cif->abi != FFI_V9) - { - wordsize = 4; - - /* If we are returning a struct, this will already have been added. - Otherwise we need to add it because it's always got to be there! */ - - if (cif->rtype->type != FFI_TYPE_STRUCT) - cif->bytes += wordsize; - - /* sparc call frames require that space is allocated for 6 args, - even if they aren't used. Make that space if necessary. */ - - if (cif->bytes < 4*6+4) - cif->bytes = 4*6+4; - } - else - { - wordsize = 8; - - /* sparc call frames require that space is allocated for 6 args, - even if they aren't used. Make that space if necessary. */ - - if (cif->bytes < 8*6) - cif->bytes = 8*6; - } - - /* Adjust cif->bytes. to include 16 words for the window save area, - and maybe the struct/union return pointer area, */ - - cif->bytes += 16 * wordsize; - - /* The stack must be 2 word aligned, so round bytes up - appropriately. */ - - cif->bytes = ALIGN(cif->bytes, 2 * wordsize); + ffi_type *rtype = cif->rtype; + int rtt = rtype->type; + size_t bytes; + int i, n, flags; /* Set the return type flag */ - switch (cif->rtype->type) + switch (rtt) { case FFI_TYPE_VOID: + flags = SPARC_RET_VOID; + break; case FFI_TYPE_FLOAT: + flags = SPARC_RET_F_1; + break; case FFI_TYPE_DOUBLE: -#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE - case FFI_TYPE_LONGDOUBLE: -#endif - cif->flags = cif->rtype->type; + flags = SPARC_RET_F_2; break; - + case FFI_TYPE_LONGDOUBLE: case FFI_TYPE_STRUCT: - if (cif->abi == FFI_V9 && cif->rtype->size > 32) - cif->flags = FFI_TYPE_VOID; - else - cif->flags = FFI_TYPE_STRUCT; + flags = (rtype->size & 0xfff) << SPARC_SIZEMASK_SHIFT; + flags |= SPARC_RET_STRUCT; break; - case FFI_TYPE_SINT8: + flags = SPARC_RET_SINT8; + break; case FFI_TYPE_UINT8: + flags = SPARC_RET_UINT8; + break; case FFI_TYPE_SINT16: + flags = SPARC_RET_SINT16; + break; case FFI_TYPE_UINT16: - if (cif->abi == FFI_V9) - cif->flags = FFI_TYPE_INT; - else - cif->flags = cif->rtype->type; + flags = SPARC_RET_UINT16; + break; + case FFI_TYPE_INT: + case FFI_TYPE_SINT32: + case FFI_TYPE_UINT32: + case FFI_TYPE_POINTER: + flags = SPARC_RET_UINT32; break; - case FFI_TYPE_SINT64: case FFI_TYPE_UINT64: - if (cif->abi == FFI_V9) - cif->flags = FFI_TYPE_INT; - else - cif->flags = FFI_TYPE_SINT64; + flags = SPARC_RET_INT64; break; - - default: - cif->flags = FFI_TYPE_INT; + case FFI_TYPE_COMPLEX: + rtt = rtype->elements[0]->type; + switch (rtt) + { + case FFI_TYPE_FLOAT: + flags = SPARC_RET_F_2; + break; + case FFI_TYPE_DOUBLE: + flags = SPARC_RET_F_4; + break; + case FFI_TYPE_LONGDOUBLE: + flags = SPARC_RET_F_8; + break; + case FFI_TYPE_SINT64: + case FFI_TYPE_UINT64: + flags = SPARC_RET_INT128; + break; + case FFI_TYPE_INT: + case FFI_TYPE_SINT32: + case FFI_TYPE_UINT32: + flags = SPARC_RET_INT64; + break; + case FFI_TYPE_SINT16: + case FFI_TYPE_UINT16: + flags = SP_V8_RET_CPLX16; + break; + case FFI_TYPE_SINT8: + case FFI_TYPE_UINT8: + flags = SP_V8_RET_CPLX8; + break; + default: + abort(); + } break; + default: + abort(); } - return FFI_OK; -} - -int ffi_v9_layout_struct(ffi_type *arg, int off, char *ret, char *intg, char *flt) -{ - ffi_type **ptr = &arg->elements[0]; + cif->flags = flags; - while (*ptr != NULL) + bytes = 0; + for (i = 0, n = cif->nargs; i < n; ++i) { - if (off & ((*ptr)->alignment - 1)) - off = ALIGN(off, (*ptr)->alignment); + ffi_type *ty = cif->arg_types[i]; + size_t z = ty->size; + int tt = ty->type; - switch ((*ptr)->type) + switch (tt) { case FFI_TYPE_STRUCT: - off = ffi_v9_layout_struct(*ptr, off, ret, intg, flt); - off = ALIGN(off, FFI_SIZEOF_ARG); - break; - case FFI_TYPE_FLOAT: - case FFI_TYPE_DOUBLE: -#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE case FFI_TYPE_LONGDOUBLE: -#endif - memmove(ret + off, flt + off, (*ptr)->size); - off += (*ptr)->size; + by_reference: + /* Passed by reference. */ + z = 4; break; + + case FFI_TYPE_COMPLEX: + tt = ty->elements[0]->type; + if (tt == FFI_TYPE_FLOAT || z > 8) + goto by_reference; + /* FALLTHRU */ + default: - memmove(ret + off, intg + off, (*ptr)->size); - off += (*ptr)->size; - break; + z = FFI_ALIGN(z, 4); } - ptr++; + bytes += z; } - return off; -} + /* Sparc call frames require that space is allocated for 6 args, + even if they aren't used. Make that space if necessary. */ + if (bytes < 6 * 4) + bytes = 6 * 4; -#ifdef SPARC64 -extern int ffi_call_v9(void *, extended_cif *, unsigned, - unsigned, unsigned *, void (*fn)(void)); -#else -extern int ffi_call_v8(void *, extended_cif *, unsigned, - unsigned, unsigned *, void (*fn)(void)); -#endif + /* The ABI always requires space for the struct return pointer. */ + bytes += 4; -#ifndef __GNUC__ -void ffi_flush_icache (void *, size_t); -#endif + /* The stack must be 2 word aligned, so round bytes up appropriately. */ + bytes = FFI_ALIGN(bytes, 2 * 4); -void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) -{ - extended_cif ecif; - void *rval = rvalue; + /* Include the call frame to prep_args. */ + bytes += 4*16 + 4*8; + cif->bytes = bytes; - ecif.cif = cif; - ecif.avalue = avalue; + return FFI_OK; +} + +extern void ffi_call_v8(ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, size_t bytes, void *closure) FFI_HIDDEN; - /* If the return value is a struct and we don't have a return */ - /* value address then we need to make one */ +int FFI_HIDDEN +ffi_prep_args_v8(ffi_cif *cif, unsigned long *argp, void *rvalue, void **avalue) +{ + ffi_type **p_arg; + int flags = cif->flags; + int i, nargs; - ecif.rvalue = rvalue; - if (cif->rtype->type == FFI_TYPE_STRUCT) + if (rvalue == NULL) { - if (cif->rtype->size <= 32) - rval = alloca(64); + if ((flags & SPARC_FLAG_RET_MASK) == SPARC_RET_STRUCT) + { + /* Since we pass the pointer to the callee, we need a value. + We allowed for this space in ffi_call, before ffi_call_v8 + alloca'd the space. */ + rvalue = (char *)argp + cif->bytes; + } else { - rval = NULL; - if (rvalue == NULL) - ecif.rvalue = alloca(cif->rtype->size); + /* Otherwise, we can ignore the return value. */ + flags = SPARC_RET_VOID; } } - switch (cif->abi) - { - case FFI_V8: -#ifdef SPARC64 - /* We don't yet support calling 32bit code from 64bit */ - FFI_ASSERT(0); -#else - if (rvalue && (cif->rtype->type == FFI_TYPE_STRUCT -#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE - || cif->flags == FFI_TYPE_LONGDOUBLE + /* This could only really be done when we are returning a structure. + However, the space is reserved so we can do it unconditionally. */ + *argp++ = (unsigned long)rvalue; + +#ifdef USING_PURIFY + /* Purify will probably complain in our assembly routine, + unless we zero out this memory. */ + memset(argp, 0, 6*4); #endif - )) + + p_arg = cif->arg_types; + for (i = 0, nargs = cif->nargs; i < nargs; i++) + { + ffi_type *ty = p_arg[i]; + void *a = avalue[i]; + int tt = ty->type; + size_t z; + + switch (tt) { - /* For v8, we need an "unimp" with size of returning struct */ - /* behind "call", so we alloc some executable space for it. */ - /* l7 is used, we need to make sure v8.S doesn't use %l7. */ - unsigned int *call_struct = NULL; - ffi_closure_alloc(32, (void **)&call_struct); - if (call_struct) + case FFI_TYPE_STRUCT: + case FFI_TYPE_LONGDOUBLE: + by_reference: + *argp++ = (unsigned long)a; + break; + + case FFI_TYPE_DOUBLE: + case FFI_TYPE_UINT64: + case FFI_TYPE_SINT64: + memcpy(argp, a, 8); + argp += 2; + break; + + case FFI_TYPE_INT: + case FFI_TYPE_FLOAT: + case FFI_TYPE_UINT32: + case FFI_TYPE_SINT32: + case FFI_TYPE_POINTER: + *argp++ = *(unsigned *)a; + break; + + case FFI_TYPE_UINT8: + *argp++ = *(UINT8 *)a; + break; + case FFI_TYPE_SINT8: + *argp++ = *(SINT8 *)a; + break; + case FFI_TYPE_UINT16: + *argp++ = *(UINT16 *)a; + break; + case FFI_TYPE_SINT16: + *argp++ = *(SINT16 *)a; + break; + + case FFI_TYPE_COMPLEX: + tt = ty->elements[0]->type; + z = ty->size; + if (tt == FFI_TYPE_FLOAT || z > 8) + goto by_reference; + if (z < 4) { - unsigned long f = (unsigned long)fn; - call_struct[0] = 0xae10001f; /* mov %i7, %l7 */ - call_struct[1] = 0xbe10000f; /* mov %o7, %i7 */ - call_struct[2] = 0x03000000 | f >> 10; /* sethi %hi(fn), %g1 */ - call_struct[3] = 0x9fc06000 | (f & 0x3ff); /* jmp %g1+%lo(fn), %o7 */ - call_struct[4] = 0x01000000; /* nop */ - if (cif->rtype->size < 0x7f) - call_struct[5] = cif->rtype->size; /* unimp */ - else - call_struct[5] = 0x01000000; /* nop */ - call_struct[6] = 0x81c7e008; /* ret */ - call_struct[7] = 0xbe100017; /* mov %l7, %i7 */ -#ifdef __GNUC__ - asm volatile ("iflush %0; iflush %0+8; iflush %0+16; iflush %0+24" : : - "r" (call_struct) : "memory"); - /* SPARC v8 requires 5 instructions for flush to be visible */ - asm volatile ("nop; nop; nop; nop; nop"); -#else - ffi_flush_icache (call_struct, 32); -#endif - ffi_call_v8(ffi_prep_args_v8, &ecif, cif->bytes, - cif->flags, rvalue, call_struct); - ffi_closure_free(call_struct); + memcpy((char *)argp + 4 - z, a, z); + argp++; } else { - ffi_call_v8(ffi_prep_args_v8, &ecif, cif->bytes, - cif->flags, rvalue, fn); + memcpy(argp, a, z); + argp += z / 4; } + break; + + default: + abort(); } - else - { - ffi_call_v8(ffi_prep_args_v8, &ecif, cif->bytes, - cif->flags, rvalue, fn); - } -#endif - break; - case FFI_V9: -#ifdef SPARC64 - ffi_call_v9(ffi_prep_args_v9, &ecif, cif->bytes, - cif->flags, rval, fn); - if (rvalue && rval && cif->rtype->type == FFI_TYPE_STRUCT) - ffi_v9_layout_struct(cif->rtype, 0, (char *)rvalue, (char *)rval, ((char *)rval)+32); -#else - /* And vice versa */ - FFI_ASSERT(0); -#endif - break; - default: - FFI_ASSERT(0); - break; } + + return flags; +} + +static void +ffi_call_int (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) +{ + size_t bytes = cif->bytes; + + FFI_ASSERT (cif->abi == FFI_V8); + + /* If we've not got a return value, we need to create one if we've + got to pass the return value to the callee. Otherwise ignore it. */ + if (rvalue == NULL + && (cif->flags & SPARC_FLAG_RET_MASK) == SPARC_RET_STRUCT) + bytes += FFI_ALIGN (cif->rtype->size, 8); + + ffi_call_v8(cif, fn, rvalue, avalue, -bytes, closure); } +void +ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +{ + ffi_call_int (cif, fn, rvalue, avalue, NULL); +} -#ifdef SPARC64 -extern void ffi_closure_v9(void); +void +ffi_call_go (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) +{ + ffi_call_int (cif, fn, rvalue, avalue, closure); +} + +#ifdef __GNUC__ +static inline void +ffi_flush_icache (void *p) +{ + /* SPARC v8 requires 5 instructions for flush to be visible */ + asm volatile ("iflush %0; iflush %0+8; nop; nop; nop; nop; nop" + : : "r" (p) : "memory"); +} #else -extern void ffi_closure_v8(void); +extern void ffi_flush_icache (void *) FFI_HIDDEN; #endif +extern void ffi_closure_v8(void) FFI_HIDDEN; +extern void ffi_go_closure_v8(void) FFI_HIDDEN; + ffi_status -ffi_prep_closure_loc (ffi_closure* closure, - ffi_cif* cif, +ffi_prep_closure_loc (ffi_closure *closure, + ffi_cif *cif, void (*fun)(ffi_cif*, void*, void**, void*), void *user_data, void *codeloc) { unsigned int *tramp = (unsigned int *) &closure->tramp[0]; - unsigned long fn; -#ifdef SPARC64 - /* Trampoline address is equal to the closure address. We take advantage - of that to reduce the trampoline size by 8 bytes. */ - if (cif->abi != FFI_V9) - return FFI_BAD_ABI; - fn = (unsigned long) ffi_closure_v9; - tramp[0] = 0x83414000; /* rd %pc, %g1 */ - tramp[1] = 0xca586010; /* ldx [%g1+16], %g5 */ - tramp[2] = 0x81c14000; /* jmp %g5 */ - tramp[3] = 0x01000000; /* nop */ - *((unsigned long *) &tramp[4]) = fn; -#else - unsigned long ctx = (unsigned long) codeloc; + unsigned long ctx = (unsigned long) closure; + unsigned long fn = (unsigned long) ffi_closure_v8; + if (cif->abi != FFI_V8) return FFI_BAD_ABI; - fn = (unsigned long) ffi_closure_v8; + tramp[0] = 0x03000000 | fn >> 10; /* sethi %hi(fn), %g1 */ tramp[1] = 0x05000000 | ctx >> 10; /* sethi %hi(ctx), %g2 */ tramp[2] = 0x81c06000 | (fn & 0x3ff); /* jmp %g1+%lo(fn) */ tramp[3] = 0x8410a000 | (ctx & 0x3ff);/* or %g2, %lo(ctx) */ -#endif closure->cif = cif; closure->fun = fun; closure->user_data = user_data; - /* Flush the Icache. closure is 8 bytes aligned. */ -#ifdef __GNUC__ -#ifdef SPARC64 - asm volatile ("flush %0; flush %0+8" : : "r" (closure) : "memory"); -#else - asm volatile ("iflush %0; iflush %0+8" : : "r" (closure) : "memory"); - /* SPARC v8 requires 5 instructions for flush to be visible */ - asm volatile ("nop; nop; nop; nop; nop"); -#endif -#else - ffi_flush_icache (closure, 16); -#endif + ffi_flush_icache (closure); return FFI_OK; } -int -ffi_closure_sparc_inner_v8(ffi_closure *closure, - void *rvalue, unsigned long *gpr, unsigned long *scratch) +ffi_status +ffi_prep_go_closure (ffi_go_closure *closure, ffi_cif *cif, + void (*fun)(ffi_cif*, void*, void**, void*)) { - ffi_cif *cif; - ffi_type **arg_types; - void **avalue; - int i, argn; - - cif = closure->cif; - arg_types = cif->arg_types; - avalue = alloca(cif->nargs * sizeof(void *)); - - /* Copy the caller's structure return address so that the closure - returns the data directly to the caller. */ - if (cif->flags == FFI_TYPE_STRUCT -#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE - || cif->flags == FFI_TYPE_LONGDOUBLE -#endif - ) - rvalue = (void *) gpr[0]; - - /* Always skip the structure return address. */ - argn = 1; - - /* Grab the addresses of the arguments from the stack frame. */ - for (i = 0; i < cif->nargs; i++) - { - if (arg_types[i]->type == FFI_TYPE_STRUCT -#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE - || arg_types[i]->type == FFI_TYPE_LONGDOUBLE -#endif - ) - { - /* Straight copy of invisible reference. */ - avalue[i] = (void *)gpr[argn++]; - } - else if ((arg_types[i]->type == FFI_TYPE_DOUBLE - || arg_types[i]->type == FFI_TYPE_SINT64 - || arg_types[i]->type == FFI_TYPE_UINT64) - /* gpr is 8-byte aligned. */ - && (argn % 2) != 0) - { - /* Align on a 8-byte boundary. */ - scratch[0] = gpr[argn]; - scratch[1] = gpr[argn+1]; - avalue[i] = scratch; - scratch -= 2; - argn += 2; - } - else - { - /* Always right-justify. */ - argn += ALIGN(arg_types[i]->size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; - avalue[i] = ((char *) &gpr[argn]) - arg_types[i]->size; - } - } + if (cif->abi != FFI_V8) + return FFI_BAD_ABI; - /* Invoke the closure. */ - (closure->fun) (cif, rvalue, avalue, closure->user_data); + closure->tramp = ffi_go_closure_v8; + closure->cif = cif; + closure->fun = fun; - /* Tell ffi_closure_sparc how to perform return type promotions. */ - return cif->rtype->type; + return FFI_OK; } -int -ffi_closure_sparc_inner_v9(ffi_closure *closure, - void *rvalue, unsigned long *gpr, double *fpr) +int FFI_HIDDEN +ffi_closure_sparc_inner_v8(ffi_cif *cif, + void (*fun)(ffi_cif*, void*, void**, void*), + void *user_data, void *rvalue, + unsigned long *argp) { - ffi_cif *cif; ffi_type **arg_types; void **avalue; - int i, argn, fp_slot_max; + int i, nargs, flags; - cif = closure->cif; arg_types = cif->arg_types; - avalue = alloca(cif->nargs * sizeof(void *)); + nargs = cif->nargs; + flags = cif->flags; + avalue = alloca(nargs * sizeof(void *)); /* Copy the caller's structure return address so that the closure - returns the data directly to the caller. */ - if (cif->flags == FFI_TYPE_VOID - && cif->rtype->type == FFI_TYPE_STRUCT) + returns the data directly to the caller. Also install it so we + can return the address in %o0. */ + if ((flags & SPARC_FLAG_RET_MASK) == SPARC_RET_STRUCT) { - rvalue = (void *) gpr[0]; - /* Skip the structure return address. */ - argn = 1; + void *new_rvalue = (void *)*argp; + *(void **)rvalue = new_rvalue; + rvalue = new_rvalue; } - else - argn = 0; - fp_slot_max = 16 - argn; + /* Always skip the structure return address. */ + argp++; /* Grab the addresses of the arguments from the stack frame. */ - for (i = 0; i < cif->nargs; i++) + for (i = 0; i < nargs; i++) { - if (arg_types[i]->type == FFI_TYPE_STRUCT) + ffi_type *ty = arg_types[i]; + int tt = ty->type; + void *a = argp; + size_t z; + + switch (tt) { - if (arg_types[i]->size > 16) - { - /* Straight copy of invisible reference. */ - avalue[i] = (void *)gpr[argn++]; - } - else + case FFI_TYPE_STRUCT: + case FFI_TYPE_LONGDOUBLE: + by_reference: + /* Straight copy of invisible reference. */ + a = (void *)*argp; + break; + + case FFI_TYPE_DOUBLE: + case FFI_TYPE_SINT64: + case FFI_TYPE_UINT64: + if ((unsigned long)a & 7) { - /* Left-justify. */ - ffi_v9_layout_struct(arg_types[i], - 0, - (char *) &gpr[argn], - (char *) &gpr[argn], - (char *) &fpr[argn]); - avalue[i] = &gpr[argn]; - argn += ALIGN(arg_types[i]->size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; + /* Align on a 8-byte boundary. */ + UINT64 *tmp = alloca(8); + *tmp = ((UINT64)argp[0] << 32) | argp[1]; + a = tmp; } - } - else - { - /* Right-justify. */ - argn += ALIGN(arg_types[i]->size, FFI_SIZEOF_ARG) / FFI_SIZEOF_ARG; + argp++; + break; - /* Align on a 16-byte boundary. */ -#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE - if (arg_types[i]->type == FFI_TYPE_LONGDOUBLE && (argn % 2) != 0) - argn++; -#endif - if (i < fp_slot_max - && (arg_types[i]->type == FFI_TYPE_FLOAT - || arg_types[i]->type == FFI_TYPE_DOUBLE -#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE - || arg_types[i]->type == FFI_TYPE_LONGDOUBLE -#endif - )) - avalue[i] = ((char *) &fpr[argn]) - arg_types[i]->size; - else - avalue[i] = ((char *) &gpr[argn]) - arg_types[i]->size; + case FFI_TYPE_INT: + case FFI_TYPE_FLOAT: + case FFI_TYPE_UINT32: + case FFI_TYPE_SINT32: + case FFI_TYPE_POINTER: + break; + case FFI_TYPE_UINT16: + case FFI_TYPE_SINT16: + a += 2; + break; + case FFI_TYPE_UINT8: + case FFI_TYPE_SINT8: + a += 3; + break; + + case FFI_TYPE_COMPLEX: + tt = ty->elements[0]->type; + z = ty->size; + if (tt == FFI_TYPE_FLOAT || z > 8) + goto by_reference; + if (z < 4) + a += 4 - z; + else if (z > 4) + argp++; + break; + + default: + abort(); } + argp++; + avalue[i] = a; } /* Invoke the closure. */ - (closure->fun) (cif, rvalue, avalue, closure->user_data); + fun (cif, rvalue, avalue, user_data); /* Tell ffi_closure_sparc how to perform return type promotions. */ - return cif->rtype->type; + return flags; } +#endif /* !SPARC64 */ diff --git a/native/libffi/src/sparc/ffi64.c b/native/libffi/src/sparc/ffi64.c new file mode 100644 index 0000000000..9e04061a42 --- /dev/null +++ b/native/libffi/src/sparc/ffi64.c @@ -0,0 +1,608 @@ +/* ----------------------------------------------------------------------- + ffi.c - Copyright (c) 2011, 2013 Anthony Green + Copyright (c) 1996, 2003-2004, 2007-2008 Red Hat, Inc. + + SPARC Foreign Function Interface + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + ``Software''), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice shall be included + in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + ----------------------------------------------------------------------- */ + +#include +#include +#include +#include "internal.h" + +/* Force FFI_TYPE_LONGDOUBLE to be different than FFI_TYPE_DOUBLE; + all further uses in this file will refer to the 128-bit type. */ +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE +# if FFI_TYPE_LONGDOUBLE != 4 +# error FFI_TYPE_LONGDOUBLE out of date +# endif +#else +# undef FFI_TYPE_LONGDOUBLE +# define FFI_TYPE_LONGDOUBLE 4 +#endif + +#ifdef SPARC64 + +/* Flatten the contents of a structure to the parts that are passed in + floating point registers. The return is a bit mask wherein bit N + set means bytes [4*n, 4*n+3] are passed in %fN. + + We encode both the (running) size (maximum 32) and mask (maxumum 255) + into one integer. The size is placed in the low byte, so that align + and addition work correctly. The mask is placed in the second byte. */ + +static int +ffi_struct_float_mask (ffi_type *outer_type, int size_mask) +{ + ffi_type **elts; + ffi_type *t; + + if (outer_type->type == FFI_TYPE_COMPLEX) + { + int m = 0, tt = outer_type->elements[0]->type; + size_t z = outer_type->size; + + if (tt == FFI_TYPE_FLOAT + || tt == FFI_TYPE_DOUBLE + || tt == FFI_TYPE_LONGDOUBLE) + m = (1 << (z / 4)) - 1; + return (m << 8) | z; + } + FFI_ASSERT (outer_type->type == FFI_TYPE_STRUCT); + + for (elts = outer_type->elements; (t = *elts) != NULL; elts++) + { + size_t z = t->size; + int o, m, tt; + + size_mask = FFI_ALIGN(size_mask, t->alignment); + switch (t->type) + { + case FFI_TYPE_STRUCT: + size_mask = ffi_struct_float_mask (t, size_mask); + continue; + case FFI_TYPE_COMPLEX: + tt = t->elements[0]->type; + if (tt != FFI_TYPE_FLOAT + && tt != FFI_TYPE_DOUBLE + && tt != FFI_TYPE_LONGDOUBLE) + break; + /* FALLTHRU */ + case FFI_TYPE_FLOAT: + case FFI_TYPE_DOUBLE: + case FFI_TYPE_LONGDOUBLE: + m = (1 << (z / 4)) - 1; /* compute mask for type */ + o = (size_mask >> 2) & 0x3f; /* extract word offset */ + size_mask |= m << (o + 8); /* insert mask into place */ + break; + } + size_mask += z; + } + + size_mask = FFI_ALIGN(size_mask, outer_type->alignment); + FFI_ASSERT ((size_mask & 0xff) == outer_type->size); + + return size_mask; +} + +/* Merge floating point data into integer data. If the structure is + entirely floating point, simply return a pointer to the fp data. */ + +static void * +ffi_struct_float_merge (int size_mask, void *vi, void *vf) +{ + int size = size_mask & 0xff; + int mask = size_mask >> 8; + int n = size >> 2; + + if (mask == 0) + return vi; + else if (mask == (1 << n) - 1) + return vf; + else + { + unsigned int *wi = vi, *wf = vf; + int i; + + for (i = 0; i < n; ++i) + if ((mask >> i) & 1) + wi[i] = wf[i]; + + return vi; + } +} + +/* Similar, but place the data into VD in the end. */ + +void FFI_HIDDEN +ffi_struct_float_copy (int size_mask, void *vd, void *vi, void *vf) +{ + int size = size_mask & 0xff; + int mask = size_mask >> 8; + int n = size >> 2; + + if (mask == 0) + ; + else if (mask == (1 << n) - 1) + vi = vf; + else + { + unsigned int *wd = vd, *wi = vi, *wf = vf; + int i; + + for (i = 0; i < n; ++i) + wd[i] = ((mask >> i) & 1 ? wf : wi)[i]; + return; + } + memcpy (vd, vi, size); +} + +/* Perform machine dependent cif processing */ + +static ffi_status +ffi_prep_cif_machdep_core(ffi_cif *cif) +{ + ffi_type *rtype = cif->rtype; + int rtt = rtype->type; + size_t bytes = 0; + int i, n, flags; + + /* Set the return type flag */ + switch (rtt) + { + case FFI_TYPE_VOID: + flags = SPARC_RET_VOID; + break; + case FFI_TYPE_FLOAT: + flags = SPARC_RET_F_1; + break; + case FFI_TYPE_DOUBLE: + flags = SPARC_RET_F_2; + break; + case FFI_TYPE_LONGDOUBLE: + flags = SPARC_RET_F_4; + break; + + case FFI_TYPE_COMPLEX: + case FFI_TYPE_STRUCT: + if (rtype->size > 32) + { + flags = SPARC_RET_VOID | SPARC_FLAG_RET_IN_MEM; + bytes = 8; + } + else + { + int size_mask = ffi_struct_float_mask (rtype, 0); + int word_size = (size_mask >> 2) & 0x3f; + int all_mask = (1 << word_size) - 1; + int fp_mask = size_mask >> 8; + + flags = (size_mask << SPARC_SIZEMASK_SHIFT) | SPARC_RET_STRUCT; + + /* For special cases of all-int or all-fp, we can return + the value directly without popping through a struct copy. */ + if (fp_mask == 0) + { + if (rtype->alignment >= 8) + { + if (rtype->size == 8) + flags = SPARC_RET_INT64; + else if (rtype->size == 16) + flags = SPARC_RET_INT128; + } + } + else if (fp_mask == all_mask) + switch (word_size) + { + case 1: flags = SPARC_RET_F_1; break; + case 2: flags = SPARC_RET_F_2; break; + case 3: flags = SP_V9_RET_F_3; break; + case 4: flags = SPARC_RET_F_4; break; + /* 5 word structures skipped; handled via RET_STRUCT. */ + case 6: flags = SPARC_RET_F_6; break; + /* 7 word structures skipped; handled via RET_STRUCT. */ + case 8: flags = SPARC_RET_F_8; break; + } + } + break; + + case FFI_TYPE_SINT8: + flags = SPARC_RET_SINT8; + break; + case FFI_TYPE_UINT8: + flags = SPARC_RET_UINT8; + break; + case FFI_TYPE_SINT16: + flags = SPARC_RET_SINT16; + break; + case FFI_TYPE_UINT16: + flags = SPARC_RET_UINT16; + break; + case FFI_TYPE_INT: + case FFI_TYPE_SINT32: + flags = SP_V9_RET_SINT32; + break; + case FFI_TYPE_UINT32: + flags = SPARC_RET_UINT32; + break; + case FFI_TYPE_SINT64: + case FFI_TYPE_UINT64: + case FFI_TYPE_POINTER: + flags = SPARC_RET_INT64; + break; + + default: + abort(); + } + + bytes = 0; + for (i = 0, n = cif->nargs; i < n; ++i) + { + ffi_type *ty = cif->arg_types[i]; + size_t z = ty->size; + size_t a = ty->alignment; + + switch (ty->type) + { + case FFI_TYPE_COMPLEX: + case FFI_TYPE_STRUCT: + /* Large structs passed by reference. */ + if (z > 16) + { + a = z = 8; + break; + } + /* Small structs may be passed in integer or fp regs or both. */ + if (bytes >= 16*8) + break; + if ((ffi_struct_float_mask (ty, 0) & 0xff00) == 0) + break; + /* FALLTHRU */ + case FFI_TYPE_FLOAT: + case FFI_TYPE_DOUBLE: + case FFI_TYPE_LONGDOUBLE: + flags |= SPARC_FLAG_FP_ARGS; + break; + } + bytes = FFI_ALIGN(bytes, a); + bytes += FFI_ALIGN(z, 8); + } + + /* Sparc call frames require that space is allocated for 6 args, + even if they aren't used. Make that space if necessary. */ + if (bytes < 6 * 8) + bytes = 6 * 8; + + /* The stack must be 2 word aligned, so round bytes up appropriately. */ + bytes = FFI_ALIGN(bytes, 16); + + /* Include the call frame to prep_args. */ + bytes += 8*16 + 8*8; + + cif->bytes = bytes; + cif->flags = flags; + return FFI_OK; +} + +ffi_status FFI_HIDDEN +ffi_prep_cif_machdep(ffi_cif *cif) +{ + cif->nfixedargs = cif->nargs; + return ffi_prep_cif_machdep_core(cif); +} + +ffi_status FFI_HIDDEN +ffi_prep_cif_machdep_var(ffi_cif *cif, unsigned nfixedargs, unsigned ntotalargs) +{ + cif->nfixedargs = nfixedargs; + return ffi_prep_cif_machdep_core(cif); +} + +extern void ffi_call_v9(ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, size_t bytes, void *closure) FFI_HIDDEN; + +/* ffi_prep_args is called by the assembly routine once stack space + has been allocated for the function's arguments */ + +int FFI_HIDDEN +ffi_prep_args_v9(ffi_cif *cif, unsigned long *argp, void *rvalue, void **avalue) +{ + ffi_type **p_arg; + int flags = cif->flags; + int i, nargs; + + if (rvalue == NULL) + { + if (flags & SPARC_FLAG_RET_IN_MEM) + { + /* Since we pass the pointer to the callee, we need a value. + We allowed for this space in ffi_call, before ffi_call_v8 + alloca'd the space. */ + rvalue = (char *)argp + cif->bytes; + } + else + { + /* Otherwise, we can ignore the return value. */ + flags = SPARC_RET_VOID; + } + } + +#ifdef USING_PURIFY + /* Purify will probably complain in our assembly routine, + unless we zero out this memory. */ + memset(argp, 0, 6*8); +#endif + + if (flags & SPARC_FLAG_RET_IN_MEM) + *argp++ = (unsigned long)rvalue; + + p_arg = cif->arg_types; + for (i = 0, nargs = cif->nargs; i < nargs; i++) + { + ffi_type *ty = p_arg[i]; + void *a = avalue[i]; + size_t z; + + switch (ty->type) + { + case FFI_TYPE_SINT8: + *argp++ = *(SINT8 *)a; + break; + case FFI_TYPE_UINT8: + *argp++ = *(UINT8 *)a; + break; + case FFI_TYPE_SINT16: + *argp++ = *(SINT16 *)a; + break; + case FFI_TYPE_UINT16: + *argp++ = *(UINT16 *)a; + break; + case FFI_TYPE_INT: + case FFI_TYPE_SINT32: + *argp++ = *(SINT32 *)a; + break; + case FFI_TYPE_UINT32: + case FFI_TYPE_FLOAT: + *argp++ = *(UINT32 *)a; + break; + case FFI_TYPE_SINT64: + case FFI_TYPE_UINT64: + case FFI_TYPE_POINTER: + case FFI_TYPE_DOUBLE: + *argp++ = *(UINT64 *)a; + break; + + case FFI_TYPE_LONGDOUBLE: + case FFI_TYPE_COMPLEX: + case FFI_TYPE_STRUCT: + z = ty->size; + if (z > 16) + { + /* For structures larger than 16 bytes we pass reference. */ + *argp++ = (unsigned long)a; + break; + } + if (((unsigned long)argp & 15) && ty->alignment > 8) + argp++; + memcpy(argp, a, z); + argp += FFI_ALIGN(z, 8) / 8; + break; + + default: + abort(); + } + } + + return flags; +} + +static void +ffi_call_int(ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) +{ + size_t bytes = cif->bytes; + + FFI_ASSERT (cif->abi == FFI_V9); + + if (rvalue == NULL && (cif->flags & SPARC_FLAG_RET_IN_MEM)) + bytes += FFI_ALIGN (cif->rtype->size, 16); + + ffi_call_v9(cif, fn, rvalue, avalue, -bytes, closure); +} + +void +ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +{ + ffi_call_int(cif, fn, rvalue, avalue, NULL); +} + +void +ffi_call_go(ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) +{ + ffi_call_int(cif, fn, rvalue, avalue, closure); +} + +#ifdef __GNUC__ +static inline void +ffi_flush_icache (void *p) +{ + asm volatile ("flush %0; flush %0+8" : : "r" (p) : "memory"); +} +#else +extern void ffi_flush_icache (void *) FFI_HIDDEN; +#endif + +extern void ffi_closure_v9(void) FFI_HIDDEN; +extern void ffi_go_closure_v9(void) FFI_HIDDEN; + +ffi_status +ffi_prep_closure_loc (ffi_closure* closure, + ffi_cif* cif, + void (*fun)(ffi_cif*, void*, void**, void*), + void *user_data, + void *codeloc) +{ + unsigned int *tramp = (unsigned int *) &closure->tramp[0]; + unsigned long fn; + + if (cif->abi != FFI_V9) + return FFI_BAD_ABI; + + /* Trampoline address is equal to the closure address. We take advantage + of that to reduce the trampoline size by 8 bytes. */ + fn = (unsigned long) ffi_closure_v9; + tramp[0] = 0x83414000; /* rd %pc, %g1 */ + tramp[1] = 0xca586010; /* ldx [%g1+16], %g5 */ + tramp[2] = 0x81c14000; /* jmp %g5 */ + tramp[3] = 0x01000000; /* nop */ + *((unsigned long *) &tramp[4]) = fn; + + closure->cif = cif; + closure->fun = fun; + closure->user_data = user_data; + + ffi_flush_icache (closure); + + return FFI_OK; +} + +ffi_status +ffi_prep_go_closure (ffi_go_closure* closure, ffi_cif* cif, + void (*fun)(ffi_cif*, void*, void**, void*)) +{ + if (cif->abi != FFI_V9) + return FFI_BAD_ABI; + + closure->tramp = ffi_go_closure_v9; + closure->cif = cif; + closure->fun = fun; + + return FFI_OK; +} + +int FFI_HIDDEN +ffi_closure_sparc_inner_v9(ffi_cif *cif, + void (*fun)(ffi_cif*, void*, void**, void*), + void *user_data, void *rvalue, + unsigned long *gpr, unsigned long *fpr) +{ + ffi_type **arg_types; + void **avalue; + int i, argn, argx, nargs, flags, nfixedargs; + + arg_types = cif->arg_types; + nargs = cif->nargs; + flags = cif->flags; + nfixedargs = cif->nfixedargs; + + avalue = alloca(nargs * sizeof(void *)); + + /* Copy the caller's structure return address so that the closure + returns the data directly to the caller. */ + if (flags & SPARC_FLAG_RET_IN_MEM) + { + rvalue = (void *) gpr[0]; + /* Skip the structure return address. */ + argn = 1; + } + else + argn = 0; + + /* Grab the addresses of the arguments from the stack frame. */ + for (i = 0; i < nargs; i++, argn = argx) + { + int named = i < nfixedargs; + ffi_type *ty = arg_types[i]; + void *a = &gpr[argn]; + size_t z; + + argx = argn + 1; + switch (ty->type) + { + case FFI_TYPE_COMPLEX: + case FFI_TYPE_STRUCT: + z = ty->size; + if (z > 16) + a = *(void **)a; + else + { + argx = argn + FFI_ALIGN (z, 8) / 8; + if (named && argn < 16) + { + int size_mask = ffi_struct_float_mask (ty, 0); + int argn_mask = (0xffff00 >> argn) & 0xff00; + + /* Eliminate fp registers off the end. */ + size_mask = (size_mask & 0xff) | (size_mask & argn_mask); + a = ffi_struct_float_merge (size_mask, gpr+argn, fpr+argn); + } + } + break; + + case FFI_TYPE_LONGDOUBLE: + argn = FFI_ALIGN (argn, 2); + a = (named && argn < 16 ? fpr : gpr) + argn; + argx = argn + 2; + break; + case FFI_TYPE_DOUBLE: + if (named && argn < 16) + a = fpr + argn; + break; + case FFI_TYPE_FLOAT: + if (named && argn < 16) + a = fpr + argn; + a += 4; + break; + + case FFI_TYPE_UINT64: + case FFI_TYPE_SINT64: + case FFI_TYPE_POINTER: + break; + case FFI_TYPE_INT: + case FFI_TYPE_UINT32: + case FFI_TYPE_SINT32: + a += 4; + break; + case FFI_TYPE_UINT16: + case FFI_TYPE_SINT16: + a += 6; + break; + case FFI_TYPE_UINT8: + case FFI_TYPE_SINT8: + a += 7; + break; + + default: + abort(); + } + avalue[i] = a; + } + + /* Invoke the closure. */ + fun (cif, rvalue, avalue, user_data); + + /* Tell ffi_closure_sparc how to perform return type promotions. */ + return flags; +} +#endif /* SPARC64 */ diff --git a/native/libffi/src/sparc/ffitarget.h b/native/libffi/src/sparc/ffitarget.h index d89f7877a8..2f4cd9a7a9 100644 --- a/native/libffi/src/sparc/ffitarget.h +++ b/native/libffi/src/sparc/ffitarget.h @@ -46,21 +46,29 @@ typedef signed long ffi_sarg; typedef enum ffi_abi { FFI_FIRST_ABI = 0, - FFI_V8, - FFI_V8PLUS, - FFI_V9, - FFI_LAST_ABI, #ifdef SPARC64 - FFI_DEFAULT_ABI = FFI_V9 + FFI_V9, + FFI_DEFAULT_ABI = FFI_V9, #else - FFI_DEFAULT_ABI = FFI_V8 + FFI_V8, + FFI_DEFAULT_ABI = FFI_V8, #endif + FFI_LAST_ABI } ffi_abi; #endif +#define FFI_TARGET_SPECIFIC_STACK_SPACE_ALLOCATION 1 +#define FFI_TARGET_HAS_COMPLEX_TYPE 1 + +#ifdef SPARC64 +# define FFI_TARGET_SPECIFIC_VARIADIC 1 +# define FFI_EXTRA_CIF_FIELDS unsigned int nfixedargs +#endif + /* ---- Definitions for closures ----------------------------------------- */ #define FFI_CLOSURES 1 +#define FFI_GO_CLOSURES 1 #define FFI_NATIVE_RAW_API 0 #ifdef SPARC64 diff --git a/native/libffi/src/sparc/internal.h b/native/libffi/src/sparc/internal.h new file mode 100644 index 0000000000..0a66472bad --- /dev/null +++ b/native/libffi/src/sparc/internal.h @@ -0,0 +1,26 @@ +#define SPARC_RET_VOID 0 +#define SPARC_RET_STRUCT 1 +#define SPARC_RET_UINT8 2 +#define SPARC_RET_SINT8 3 +#define SPARC_RET_UINT16 4 +#define SPARC_RET_SINT16 5 +#define SPARC_RET_UINT32 6 +#define SP_V9_RET_SINT32 7 /* v9 only */ +#define SP_V8_RET_CPLX16 7 /* v8 only */ +#define SPARC_RET_INT64 8 +#define SPARC_RET_INT128 9 + +/* Note that F_7 is missing, and is handled by SPARC_RET_STRUCT. */ +#define SPARC_RET_F_8 10 +#define SPARC_RET_F_6 11 +#define SPARC_RET_F_4 12 +#define SPARC_RET_F_2 13 +#define SP_V9_RET_F_3 14 /* v9 only */ +#define SP_V8_RET_CPLX8 14 /* v8 only */ +#define SPARC_RET_F_1 15 + +#define SPARC_FLAG_RET_MASK 15 +#define SPARC_FLAG_RET_IN_MEM 32 +#define SPARC_FLAG_FP_ARGS 64 + +#define SPARC_SIZEMASK_SHIFT 8 diff --git a/native/libffi/src/sparc/v8.S b/native/libffi/src/sparc/v8.S index 6bf7ac056e..a2e4908fd4 100644 --- a/native/libffi/src/sparc/v8.S +++ b/native/libffi/src/sparc/v8.S @@ -1,8 +1,8 @@ /* ----------------------------------------------------------------------- v8.S - Copyright (c) 2013 The Written Word, Inc. Copyright (c) 1996, 1997, 2003, 2004, 2008 Red Hat, Inc. - - SPARC Foreign Function Interface + + SPARC Foreign Function Interface Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the @@ -25,179 +25,253 @@ DEALINGS IN THE SOFTWARE. ----------------------------------------------------------------------- */ -#define LIBFFI_ASM +#define LIBFFI_ASM #include #include +#include "internal.h" + +#ifndef SPARC64 -#define STACKFRAME 96 /* Minimum stack framesize for SPARC */ -#define ARGS (64+4) /* Offset of register area in frame */ +#define C2(X, Y) X ## Y +#define C1(X, Y) C2(X, Y) + +#ifdef __USER_LABEL_PREFIX__ +# define C(Y) C1(__USER_LABEL_PREFIX__, Y) +#else +# define C(Y) Y +#endif +#define L(Y) C1(.L, Y) -#ifndef __GNUC__ .text + +#ifndef __GNUC__ .align 8 -.globl ffi_flush_icache -.globl _ffi_flush_icache - -ffi_flush_icache: -_ffi_flush_icache: - add %o0, %o1, %o2 -#ifdef SPARC64 -1: flush %o0 -#else + .globl C(ffi_flush_icache) + .type C(ffi_flush_icache),#function + FFI_HIDDEN(C(ffi_flush_icache)) + +C(ffi_flush_icache): 1: iflush %o0 -#endif - add %o0, 8, %o0 - cmp %o0, %o2 - blt 1b + iflush %o+8 nop nop nop nop nop retl - nop -.ffi_flush_icache_end: - .size ffi_flush_icache,.ffi_flush_icache_end-ffi_flush_icache + nop + .size C(ffi_flush_icache), . - C(ffi_flush_icache) +#endif + +#if defined(__sun__) && defined(__svr4__) +# define E(INDEX) .align 16 +#else +# define E(INDEX) .align 16; .org 2b + INDEX * 16 #endif - .text .align 8 -.globl ffi_call_v8 -.globl _ffi_call_v8 - -ffi_call_v8: -_ffi_call_v8: -.LLFB1: - save %sp, -STACKFRAME, %sp -.LLCFI0: - - sub %sp, %i2, %sp ! alloca() space in stack for frame to set up - add %sp, STACKFRAME, %l0 ! %l0 has start of - ! frame to set up - - mov %l0, %o0 ! call routine to set up frame - call %i0 - mov %i1, %o1 ! (delay) - - ld [%l0+ARGS], %o0 ! call foreign function - ld [%l0+ARGS+4], %o1 - ld [%l0+ARGS+8], %o2 - ld [%l0+ARGS+12], %o3 - ld [%l0+ARGS+16], %o4 - ld [%l0+ARGS+20], %o5 - call %i5 - mov %l0, %sp ! (delay) switch to frame - nop ! STRUCT returning functions skip 12 instead of 8 bytes - - ! If the return value pointer is NULL, assume no return value. - tst %i4 - bz done + .globl C(ffi_call_v8) + .type C(ffi_call_v8),#function + FFI_HIDDEN(C(ffi_call_v8)) + +C(ffi_call_v8): +.LUW0: + ! Allocate a stack frame sized by ffi_call. + save %sp, %o4, %sp +.LUW1: + mov %i0, %o0 ! copy cif + add %sp, 64+32, %o1 ! load args area + mov %i2, %o2 ! copy rvalue + call C(ffi_prep_args_v8) + mov %i3, %o3 ! copy avalue + + add %sp, 32, %sp ! deallocate prep frame + and %o0, SPARC_FLAG_RET_MASK, %l0 ! save return type + srl %o0, SPARC_SIZEMASK_SHIFT, %l1 ! save return size + ld [%sp+64+4], %o0 ! load all argument registers + ld [%sp+64+8], %o1 + ld [%sp+64+12], %o2 + ld [%sp+64+16], %o3 + cmp %l0, SPARC_RET_STRUCT ! struct return needs an unimp 4 + ld [%sp+64+20], %o4 + be 8f + ld [%sp+64+24], %o5 + + ! Call foreign function + call %i1 + mov %i5, %g2 ! load static chain + +0: call 1f ! load pc in %o7 + sll %l0, 4, %l0 +1: add %o7, %l0, %o7 ! o7 = 0b + ret_type*16 + jmp %o7+(2f-0b) + nop + + ! Note that each entry is 4 insns, enforced by the E macro. + .align 16 +2: +E(SPARC_RET_VOID) + ret + restore +E(SPARC_RET_STRUCT) + unimp +E(SPARC_RET_UINT8) + and %o0, 0xff, %o0 + st %o0, [%i2] + ret + restore +E(SPARC_RET_SINT8) + sll %o0, 24, %o0 + b 7f + sra %o0, 24, %o0 +E(SPARC_RET_UINT16) + sll %o0, 16, %o0 + b 7f + srl %o0, 16, %o0 +E(SPARC_RET_SINT16) + sll %o0, 16, %o0 + b 7f + sra %o0, 16, %o0 +E(SPARC_RET_UINT32) +7: st %o0, [%i2] + ret + restore +E(SP_V8_RET_CPLX16) + sth %o0, [%i2+2] + b 9f + srl %o0, 16, %o0 +E(SPARC_RET_INT64) + st %o0, [%i2] + st %o1, [%i2+4] + ret + restore +E(SPARC_RET_INT128) + std %o0, [%i2] + std %o2, [%i2+8] + ret + restore +E(SPARC_RET_F_8) + st %f7, [%i2+7*4] nop - - cmp %i3, FFI_TYPE_INT - be,a done - st %o0, [%i4] ! (delay) - - cmp %i3, FFI_TYPE_FLOAT - be,a done - st %f0, [%i4+0] ! (delay) - - cmp %i3, FFI_TYPE_DOUBLE - be,a double - st %f0, [%i4+0] ! (delay) - - cmp %i3, FFI_TYPE_SINT8 - be,a sint8 - sll %o0, 24, %o0 ! (delay) - - cmp %i3, FFI_TYPE_UINT8 - be,a uint8 - sll %o0, 24, %o0 ! (delay) - - cmp %i3, FFI_TYPE_SINT16 - be,a sint16 - sll %o0, 16, %o0 ! (delay) - - cmp %i3, FFI_TYPE_UINT16 - be,a uint16 - sll %o0, 16, %o0 ! (delay) - - cmp %i3, FFI_TYPE_SINT64 - be,a longlong - st %o0, [%i4+0] ! (delay) -done: + st %f6, [%i2+6*4] + nop +E(SPARC_RET_F_6) + st %f5, [%i2+5*4] + nop + st %f4, [%i2+4*4] + nop +E(SPARC_RET_F_4) + st %f3, [%i2+3*4] + nop + st %f2, [%i2+2*4] + nop +E(SPARC_RET_F_2) + st %f1, [%i2+4] + st %f0, [%i2] ret - restore - -double: - st %f1, [%i4+4] + restore +E(SP_V8_RET_CPLX8) + stb %o0, [%i2+1] + b 0f + srl %o0, 8, %o0 +E(SPARC_RET_F_1) + st %f0, [%i2] ret - restore + restore -sint8: - sra %o0, 24, %o0 - st %o0, [%i4+0] + .align 8 +9: sth %o0, [%i2] ret - restore - -uint8: - srl %o0, 24, %o0 - st %o0, [%i4+0] + restore + .align 8 +0: stb %o0, [%i2] ret - restore + restore -sint16: - sra %o0, 16, %o0 - st %o0, [%i4+0] - ret - restore + ! Struct returning functions expect and skip the unimp here. + ! To make it worse, conforming callees examine the unimp and + ! make sure the low 12 bits of the unimp match the size of + ! the struct being returned. + .align 8 +8: call 1f ! load pc in %o7 + sll %l1, 2, %l0 ! size * 4 +1: sll %l1, 4, %l1 ! size * 16 + add %l0, %l1, %l0 ! size * 20 + add %o7, %l0, %o7 ! o7 = 8b + size*20 + jmp %o7+(2f-8b) + mov %i5, %g2 ! load static chain +2: + +/* The Sun assembler doesn't understand .rept 0x1000. */ +#define rept1 \ + call %i1; \ + nop; \ + unimp (. - 2b) / 20; \ + ret; \ + restore -uint16: - srl %o0, 16, %o0 - st %o0, [%i4+0] - ret - restore +#define rept16 \ + rept1; rept1; rept1; rept1; \ + rept1; rept1; rept1; rept1; \ + rept1; rept1; rept1; rept1; \ + rept1; rept1; rept1; rept1 -longlong: - st %o1, [%i4+4] - ret - restore -.LLFE1: +#define rept256 \ + rept16; rept16; rept16; rept16; \ + rept16; rept16; rept16; rept16; \ + rept16; rept16; rept16; rept16; \ + rept16; rept16; rept16; rept16 -.ffi_call_v8_end: - .size ffi_call_v8,.ffi_call_v8_end-ffi_call_v8 + rept256; rept256; rept256; rept256 + rept256; rept256; rept256; rept256 + rept256; rept256; rept256; rept256 + rept256; rept256; rept256; rept256 +.LUW2: + .size C(ffi_call_v8),. - C(ffi_call_v8) -#undef STACKFRAME -#define STACKFRAME 104 /* 16*4 register window + - 1*4 struct return + - 6*4 args backing store + - 3*4 locals */ + +/* 16*4 register window + 1*4 struct return + 6*4 args backing store + + 8*4 return storage + 1*4 alignment. */ +#define STACKFRAME (16*4 + 4 + 6*4 + 8*4 + 4) /* ffi_closure_v8(...) Receives the closure argument in %g2. */ - .text - .align 8 - .globl ffi_closure_v8 - -ffi_closure_v8: #ifdef HAVE_AS_REGISTER_PSEUDO_OP - .register %g2, #scratch + .register %g2, #scratch #endif -.LLFB2: - ! Reserve frame space for all arguments in case - ! we need to align them on a 8-byte boundary. - ld [%g2+FFI_TRAMPOLINE_SIZE], %g1 - ld [%g1+4], %g1 - sll %g1, 3, %g1 - add %g1, STACKFRAME, %g1 - ! %g1 == STACKFRAME + 8*nargs - neg %g1 - save %sp, %g1, %sp -.LLCFI1: + .align 8 + .globl C(ffi_go_closure_v8) + .type C(ffi_go_closure_v8),#function + FFI_HIDDEN(C(ffi_go_closure_v8)) + +C(ffi_go_closure_v8): +.LUW3: + save %sp, -STACKFRAME, %sp +.LUW4: + ld [%g2+4], %o0 ! load cif + ld [%g2+8], %o1 ! load fun + b 0f + mov %g2, %o2 ! load user_data +.LUW5: + .size C(ffi_go_closure_v8), . - C(ffi_go_closure_v8) + + .align 8 + .globl C(ffi_closure_v8) + .type C(ffi_closure_v8),#function + FFI_HIDDEN(C(ffi_closure_v8)) + +C(ffi_closure_v8): +.LUW6: + save %sp, -STACKFRAME, %sp +.LUW7: + ld [%g2+FFI_TRAMPOLINE_SIZE], %o0 ! load cif + ld [%g2+FFI_TRAMPOLINE_SIZE+4], %o1 ! load fun + ld [%g2+FFI_TRAMPOLINE_SIZE+8], %o2 ! load user_data +0: ! Store all of the potential argument registers in va_list format. st %i0, [%fp+68+0] st %i1, [%fp+68+4] @@ -207,140 +281,163 @@ ffi_closure_v8: st %i5, [%fp+68+20] ! Call ffi_closure_sparc_inner to do the bulk of the work. - mov %g2, %o0 - add %fp, -8, %o1 - add %fp, 64, %o2 + add %fp, -8*4, %o3 call ffi_closure_sparc_inner_v8 - add %fp, -16, %o3 - - ! Load up the return value in the proper type. - ! See ffi_prep_cif_machdep for the list of cases. - cmp %o0, FFI_TYPE_VOID - be done1 - - cmp %o0, FFI_TYPE_INT - be done1 - ld [%fp-8], %i0 - - cmp %o0, FFI_TYPE_FLOAT - be,a done1 - ld [%fp-8], %f0 - - cmp %o0, FFI_TYPE_DOUBLE - be,a done1 - ldd [%fp-8], %f0 - -#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE - cmp %o0, FFI_TYPE_LONGDOUBLE - be done2 -#endif - - cmp %o0, FFI_TYPE_STRUCT - be done2 - - cmp %o0, FFI_TYPE_SINT64 - be,a done1 - ldd [%fp-8], %i0 - - cmp %o0, FFI_TYPE_UINT64 - be,a done1 - ldd [%fp-8], %i0 - - ld [%fp-8], %i0 -done1: - jmp %i7+8 + add %fp, 64, %o4 + +0: call 1f + and %o0, SPARC_FLAG_RET_MASK, %o0 +1: sll %o0, 4, %o0 ! o0 = o0 * 16 + add %o7, %o0, %o7 ! o7 = 0b + o0*16 + jmp %o7+(2f-0b) + add %fp, -8*4, %i2 + + ! Note that each entry is 4 insns, enforced by the E macro. + .align 16 +2: +E(SPARC_RET_VOID) + ret restore -done2: - ! Skip 'unimp'. +E(SPARC_RET_STRUCT) + ld [%i2], %i0 jmp %i7+12 restore -.LLFE2: - -.ffi_closure_v8_end: - .size ffi_closure_v8,.ffi_closure_v8_end-ffi_closure_v8 +E(SPARC_RET_UINT8) + ldub [%i2+3], %i0 + ret + restore +E(SPARC_RET_SINT8) + ldsb [%i2+3], %i0 + ret + restore +E(SPARC_RET_UINT16) + lduh [%i2+2], %i0 + ret + restore +E(SPARC_RET_SINT16) + ldsh [%i2+2], %i0 + ret + restore +E(SPARC_RET_UINT32) + ld [%i2], %i0 + ret + restore +E(SP_V8_RET_CPLX16) + ld [%i2], %i0 + ret + restore +E(SPARC_RET_INT64) + ldd [%i2], %i0 + ret + restore +E(SPARC_RET_INT128) + ldd [%i2], %i0 + ldd [%i2+8], %i2 + ret + restore +E(SPARC_RET_F_8) + ld [%i2+7*4], %f7 + nop + ld [%i2+6*4], %f6 + nop +E(SPARC_RET_F_6) + ld [%i2+5*4], %f5 + nop + ld [%i2+4*4], %f4 + nop +E(SPARC_RET_F_4) + ld [%i2+3*4], %f3 + nop + ld [%i2+2*4], %f2 + nop +E(SPARC_RET_F_2) + ldd [%i2], %f0 + ret + restore +E(SP_V8_RET_CPLX8) + lduh [%i2], %i0 + ret + restore +E(SPARC_RET_F_1) + ld [%i2], %f0 + ret + restore -#ifdef SPARC64 -#define WS 8 -#define nword xword -#define uanword uaxword -#else -#define WS 4 -#define nword long -#define uanword uaword -#endif +.LUW8: + .size C(ffi_closure_v8), . - C(ffi_closure_v8) #ifdef HAVE_RO_EH_FRAME - .section ".eh_frame",#alloc + .section ".eh_frame",#alloc #else - .section ".eh_frame",#alloc,#write + .section ".eh_frame",#alloc,#write #endif -.LLframe1: - .uaword .LLECIE1-.LLSCIE1 ! Length of Common Information Entry -.LLSCIE1: - .uaword 0x0 ! CIE Identifier Tag - .byte 0x1 ! CIE Version - .ascii "zR\0" ! CIE Augmentation - .byte 0x1 ! uleb128 0x1; CIE Code Alignment Factor - .byte 0x80-WS ! sleb128 -WS; CIE Data Alignment Factor - .byte 0xf ! CIE RA Column - .byte 0x1 ! uleb128 0x1; Augmentation size -#ifdef HAVE_AS_SPARC_UA_PCREL - .byte 0x1b ! FDE Encoding (pcrel sdata4) -#else - .byte 0x50 ! FDE Encoding (aligned absolute) -#endif - .byte 0xc ! DW_CFA_def_cfa - .byte 0xe ! uleb128 0xe - .byte 0x0 ! uleb128 0x0 - .align WS -.LLECIE1: -.LLSFDE1: - .uaword .LLEFDE1-.LLASFDE1 ! FDE Length -.LLASFDE1: - .uaword .LLASFDE1-.LLframe1 ! FDE CIE offset + #ifdef HAVE_AS_SPARC_UA_PCREL - .uaword %r_disp32(.LLFB1) - .uaword .LLFE1-.LLFB1 ! FDE address range +# define FDE_ADDR(X) %r_disp32(X) #else - .align WS - .nword .LLFB1 - .uanword .LLFE1-.LLFB1 ! FDE address range +# define FDE_ADDR(X) X #endif - .byte 0x0 ! uleb128 0x0; Augmentation size - .byte 0x4 ! DW_CFA_advance_loc4 - .uaword .LLCFI0-.LLFB1 - .byte 0xd ! DW_CFA_def_cfa_register - .byte 0x1e ! uleb128 0x1e - .byte 0x2d ! DW_CFA_GNU_window_save - .byte 0x9 ! DW_CFA_register - .byte 0xf ! uleb128 0xf - .byte 0x1f ! uleb128 0x1f - .align WS -.LLEFDE1: -.LLSFDE2: - .uaword .LLEFDE2-.LLASFDE2 ! FDE Length -.LLASFDE2: - .uaword .LLASFDE2-.LLframe1 ! FDE CIE offset + + .align 4 +.LCIE: + .long .LECIE - .LSCIE ! CIE Length +.LSCIE: + .long 0 ! CIE Identifier Tag + .byte 1 ! CIE Version + .ascii "zR\0" ! CIE Augmentation + .byte 4 ! CIE Code Alignment Factor + .byte 0x7c ! CIE Data Alignment Factor + .byte 15 ! CIE RA Column + .byte 1 ! Augmentation size #ifdef HAVE_AS_SPARC_UA_PCREL - .uaword %r_disp32(.LLFB2) - .uaword .LLFE2-.LLFB2 ! FDE address range + .byte 0x1b ! FDE Encoding (pcrel sdata4) #else - .align WS - .nword .LLFB2 - .uanword .LLFE2-.LLFB2 ! FDE address range + .byte 0x50 ! FDE Encoding (aligned absolute) #endif - .byte 0x0 ! uleb128 0x0; Augmentation size - .byte 0x4 ! DW_CFA_advance_loc4 - .uaword .LLCFI1-.LLFB2 - .byte 0xd ! DW_CFA_def_cfa_register - .byte 0x1e ! uleb128 0x1e - .byte 0x2d ! DW_CFA_GNU_window_save - .byte 0x9 ! DW_CFA_register - .byte 0xf ! uleb128 0xf - .byte 0x1f ! uleb128 0x1f - .align WS -.LLEFDE2: - + .byte 0xc, 14, 0 ! DW_CFA_def_cfa, %o6, offset 0 + .align 4 +.LECIE: + + .long .LEFDE1 - .LSFDE1 ! FDE Length +.LSFDE1: + .long .LSFDE1 - .LCIE ! FDE CIE offset + .long FDE_ADDR(.LUW0) ! Initial location + .long .LUW2 - .LUW0 ! Address range + .byte 0 ! Augmentation size + .byte 0x40+1 ! DW_CFA_advance_loc 4 + .byte 0xd, 30 ! DW_CFA_def_cfa_register, %i6 + .byte 0x2d ! DW_CFA_GNU_window_save + .byte 0x9, 15, 31 ! DW_CFA_register, %o7, %i7 + .align 4 +.LEFDE1: + + .long .LEFDE2 - .LSFDE2 ! FDE Length +.LSFDE2: + .long .LSFDE2 - .LCIE ! FDE CIE offset + .long FDE_ADDR(.LUW3) ! Initial location + .long .LUW5 - .LUW3 ! Address range + .byte 0 ! Augmentation size + .byte 0x40+1 ! DW_CFA_advance_loc 4 + .byte 0xd, 30 ! DW_CFA_def_cfa_register, %i6 + .byte 0x2d ! DW_CFA_GNU_window_save + .byte 0x9, 15, 31 ! DW_CFA_register, %o7, %i7 + .align 4 +.LEFDE2: + + .long .LEFDE3 - .LSFDE3 ! FDE Length +.LSFDE3: + .long .LSFDE3 - .LCIE ! FDE CIE offset + .long FDE_ADDR(.LUW6) ! Initial location + .long .LUW8 - .LUW6 ! Address range + .byte 0 ! Augmentation size + .byte 0x40+1 ! DW_CFA_advance_loc 4 + .byte 0xd, 30 ! DW_CFA_def_cfa_register, %i6 + .byte 0x2d ! DW_CFA_GNU_window_save + .byte 0x9, 15, 31 ! DW_CFA_register, %o7, %i7 + .align 4 +.LEFDE3: + +#endif /* !SPARC64 */ #if defined __ELF__ && defined __linux__ .section .note.GNU-stack,"",@progbits #endif diff --git a/native/libffi/src/sparc/v9.S b/native/libffi/src/sparc/v9.S index bf31a2b511..55f8f4324c 100644 --- a/native/libffi/src/sparc/v9.S +++ b/native/libffi/src/sparc/v9.S @@ -27,105 +27,176 @@ #define LIBFFI_ASM #include #include +#include "internal.h" #ifdef SPARC64 -/* Only compile this in for 64bit builds, because otherwise the object file - will have inproper architecture due to used instructions. */ -#define STACKFRAME 176 /* Minimum stack framesize for SPARC 64-bit */ -#define STACK_BIAS 2047 -#define ARGS (128) /* Offset of register area in frame */ - -.text - .align 8 -.globl ffi_call_v9 -.globl _ffi_call_v9 - -ffi_call_v9: -_ffi_call_v9: -.LLFB1: - save %sp, -STACKFRAME, %sp -.LLCFI0: - - sub %sp, %i2, %sp ! alloca() space in stack for frame to set up - add %sp, STACKFRAME+STACK_BIAS, %l0 ! %l0 has start of - ! frame to set up - - mov %l0, %o0 ! call routine to set up frame - call %i0 - mov %i1, %o1 ! (delay) - brz,pt %o0, 1f - ldx [%l0+ARGS], %o0 ! call foreign function - - ldd [%l0+ARGS], %f0 - ldd [%l0+ARGS+8], %f2 - ldd [%l0+ARGS+16], %f4 - ldd [%l0+ARGS+24], %f6 - ldd [%l0+ARGS+32], %f8 - ldd [%l0+ARGS+40], %f10 - ldd [%l0+ARGS+48], %f12 - ldd [%l0+ARGS+56], %f14 - ldd [%l0+ARGS+64], %f16 - ldd [%l0+ARGS+72], %f18 - ldd [%l0+ARGS+80], %f20 - ldd [%l0+ARGS+88], %f22 - ldd [%l0+ARGS+96], %f24 - ldd [%l0+ARGS+104], %f26 - ldd [%l0+ARGS+112], %f28 - ldd [%l0+ARGS+120], %f30 - -1: ldx [%l0+ARGS+8], %o1 - ldx [%l0+ARGS+16], %o2 - ldx [%l0+ARGS+24], %o3 - ldx [%l0+ARGS+32], %o4 - ldx [%l0+ARGS+40], %o5 - call %i5 - sub %l0, STACK_BIAS, %sp ! (delay) switch to frame - - ! If the return value pointer is NULL, assume no return value. - brz,pn %i4, done - nop +#define C2(X, Y) X ## Y +#define C1(X, Y) C2(X, Y) - cmp %i3, FFI_TYPE_INT - be,a,pt %icc, done - stx %o0, [%i4+0] ! (delay) +#ifdef __USER_LABEL_PREFIX__ +# define C(Y) C1(__USER_LABEL_PREFIX__, Y) +#else +# define C(Y) Y +#endif +#define L(Y) C1(.L, Y) - cmp %i3, FFI_TYPE_FLOAT - be,a,pn %icc, done - st %f0, [%i4+0] ! (delay) +#if defined(__sun__) && defined(__svr4__) +# define E(INDEX) .align 16 +#else +# define E(INDEX) .align 16; .org 2b + INDEX * 16 +#endif - cmp %i3, FFI_TYPE_DOUBLE - be,a,pn %icc, done - std %f0, [%i4+0] ! (delay) +#define STACK_BIAS 2047 - cmp %i3, FFI_TYPE_STRUCT - be,pn %icc, dostruct + .text + .align 8 + .globl C(ffi_call_v9) + .type C(ffi_call_v9),#function + FFI_HIDDEN(C(ffi_call_v9)) + +C(ffi_call_v9): +.LUW0: + save %sp, %o4, %sp +.LUW1: + mov %i0, %o0 ! copy cif + add %sp, STACK_BIAS+128+48, %o1 ! load args area + mov %i2, %o2 ! copy rvalue + call C(ffi_prep_args_v9) + mov %i3, %o3 ! copy avalue + + andcc %o0, SPARC_FLAG_FP_ARGS, %g0 ! need fp regs? + add %sp, 48, %sp ! deallocate prep frame + be,pt %xcc, 1f + mov %o0, %l0 ! save flags + + ldd [%sp+STACK_BIAS+128], %f0 ! load all fp arg regs + ldd [%sp+STACK_BIAS+128+8], %f2 + ldd [%sp+STACK_BIAS+128+16], %f4 + ldd [%sp+STACK_BIAS+128+24], %f6 + ldd [%sp+STACK_BIAS+128+32], %f8 + ldd [%sp+STACK_BIAS+128+40], %f10 + ldd [%sp+STACK_BIAS+128+48], %f12 + ldd [%sp+STACK_BIAS+128+56], %f14 + ldd [%sp+STACK_BIAS+128+64], %f16 + ldd [%sp+STACK_BIAS+128+72], %f18 + ldd [%sp+STACK_BIAS+128+80], %f20 + ldd [%sp+STACK_BIAS+128+88], %f22 + ldd [%sp+STACK_BIAS+128+96], %f24 + ldd [%sp+STACK_BIAS+128+104], %f26 + ldd [%sp+STACK_BIAS+128+112], %f28 + ldd [%sp+STACK_BIAS+128+120], %f30 + +1: ldx [%sp+STACK_BIAS+128], %o0 ! load all int arg regs + ldx [%sp+STACK_BIAS+128+8], %o1 + ldx [%sp+STACK_BIAS+128+16], %o2 + ldx [%sp+STACK_BIAS+128+24], %o3 + ldx [%sp+STACK_BIAS+128+32], %o4 + ldx [%sp+STACK_BIAS+128+40], %o5 + call %i1 + mov %i5, %g5 ! load static chain + +0: call 1f ! load pc in %o7 + and %l0, SPARC_FLAG_RET_MASK, %l1 +1: sll %l1, 4, %l1 + add %o7, %l1, %o7 ! o7 = 0b + ret_type*16 + jmp %o7+(2f-0b) + nop - cmp %i3, FFI_TYPE_LONGDOUBLE - bne,pt %icc, done + .align 16 +2: +E(SPARC_RET_VOID) + return %i7+8 + nop +E(SPARC_RET_STRUCT) + add %sp, STACK_BIAS-64+128+48, %l2 + sub %sp, 64, %sp + b 8f + stx %o0, [%l2] +E(SPARC_RET_UINT8) + and %o0, 0xff, %i0 + return %i7+8 + stx %o0, [%o2] +E(SPARC_RET_SINT8) + sll %o0, 24, %o0 + sra %o0, 24, %i0 + return %i7+8 + stx %o0, [%o2] +E(SPARC_RET_UINT16) + sll %o0, 16, %o0 + srl %o0, 16, %i0 + return %i7+8 + stx %o0, [%o2] +E(SPARC_RET_SINT16) + sll %o0, 16, %o0 + sra %o0, 16, %i0 + return %i7+8 + stx %o0, [%o2] +E(SPARC_RET_UINT32) + srl %o0, 0, %i0 + return %i7+8 + stx %o0, [%o2] +E(SP_V9_RET_SINT32) + sra %o0, 0, %i0 + return %i7+8 + stx %o0, [%o2] +E(SPARC_RET_INT64) + stx %o0, [%i2] + return %i7+8 + nop +E(SPARC_RET_INT128) + stx %o0, [%i2] + stx %o1, [%i2+8] + return %i7+8 nop - std %f0, [%i4+0] - std %f2, [%i4+8] - -done: ret - restore - -dostruct: - /* This will not work correctly for unions. */ - stx %o0, [%i4+0] - stx %o1, [%i4+8] - stx %o2, [%i4+16] - stx %o3, [%i4+24] - std %f0, [%i4+32] - std %f2, [%i4+40] - std %f4, [%i4+48] - std %f6, [%i4+56] - ret - restore -.LLFE1: - -.ffi_call_v9_end: - .size ffi_call_v9,.ffi_call_v9_end-ffi_call_v9 +E(SPARC_RET_F_8) + st %f7, [%i2+7*4] + nop + st %f6, [%i2+6*4] + nop +E(SPARC_RET_F_6) + st %f5, [%i2+5*4] + nop + st %f4, [%i2+4*4] + nop +E(SPARC_RET_F_4) + std %f2, [%i2+2*4] + return %i7+8 + std %f0, [%o2] +E(SPARC_RET_F_2) + return %i7+8 + std %f0, [%o2] +E(SP_V9_RET_F_3) + st %f2, [%i2+2*4] + nop + st %f1, [%i2+1*4] + nop +E(SPARC_RET_F_1) + return %i7+8 + st %f0, [%o2] + + ! Finish the SPARC_RET_STRUCT sequence. + .align 8 +8: stx %o1, [%l2+8] + stx %o2, [%l2+16] + stx %o3, [%l2+24] + std %f0, [%l2+32] + std %f2, [%l2+40] + std %f4, [%l2+48] + std %f6, [%l2+56] + + ! Copy the structure into place. + srl %l0, SPARC_SIZEMASK_SHIFT, %o0 ! load size_mask + mov %i2, %o1 ! load dst + mov %l2, %o2 ! load src_gp + call C(ffi_struct_float_copy) + add %l2, 32, %o3 ! load src_fp + + return %i7+8 + nop + +.LUW2: + .size C(ffi_call_v9), . - C(ffi_call_v9) #undef STACKFRAME @@ -138,15 +209,36 @@ dostruct: Receives the closure argument in %g1. */ - .text .align 8 - .globl ffi_closure_v9 + .globl C(ffi_go_closure_v9) + .type C(ffi_go_closure_v9),#function + FFI_HIDDEN(C(ffi_go_closure_v9)) -ffi_closure_v9: -.LLFB2: +C(ffi_go_closure_v9): +.LUW3: save %sp, -STACKFRAME, %sp -.LLCFI1: +.LUW4: + ldx [%g5+8], %o0 + ldx [%g5+16], %o1 + b 0f + mov %g5, %o2 + +.LUW5: + .size C(ffi_go_closure_v9), . - C(ffi_go_closure_v9) + .align 8 + .globl C(ffi_closure_v9) + .type C(ffi_closure_v9),#function + FFI_HIDDEN(C(ffi_closure_v9)) + +C(ffi_closure_v9): +.LUW6: + save %sp, -STACKFRAME, %sp +.LUW7: + ldx [%g1+FFI_TRAMPOLINE_SIZE], %o0 + ldx [%g1+FFI_TRAMPOLINE_SIZE+8], %o1 + ldx [%g1+FFI_TRAMPOLINE_SIZE+16], %o2 +0: ! Store all of the potential argument registers in va_list format. stx %i0, [FP+128+0] stx %i1, [FP+128+8] @@ -174,134 +266,175 @@ ffi_closure_v9: std %f30, [FP-8] ! Call ffi_closure_sparc_inner to do the bulk of the work. - mov %g1, %o0 - add %fp, STACK_BIAS-160, %o1 - add %fp, STACK_BIAS+128, %o2 - call ffi_closure_sparc_inner_v9 - add %fp, STACK_BIAS-128, %o3 - - ! Load up the return value in the proper type. - ! See ffi_prep_cif_machdep for the list of cases. - cmp %o0, FFI_TYPE_VOID - be,pn %icc, done1 - - cmp %o0, FFI_TYPE_INT - be,pn %icc, integer - - cmp %o0, FFI_TYPE_FLOAT - be,a,pn %icc, done1 - ld [FP-160], %f0 - - cmp %o0, FFI_TYPE_DOUBLE - be,a,pn %icc, done1 - ldd [FP-160], %f0 - -#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE - cmp %o0, FFI_TYPE_LONGDOUBLE - be,a,pn %icc, longdouble1 - ldd [FP-160], %f0 -#endif + add %fp, STACK_BIAS-160, %o3 + add %fp, STACK_BIAS+128, %o4 + call C(ffi_closure_sparc_inner_v9) + add %fp, STACK_BIAS-128, %o5 + +0: call 1f ! load pc in %o7 + and %o0, SPARC_FLAG_RET_MASK, %o0 +1: sll %o0, 4, %o0 ! o2 = i2 * 16 + add %o7, %o0, %o7 ! o7 = 0b + i2*16 + jmp %o7+(2f-0b) + nop - ! FFI_TYPE_STRUCT - ldx [FP-152], %i1 - ldx [FP-144], %i2 - ldx [FP-136], %i3 + ! Note that we cannot load the data in the delay slot of + ! the return insn because the data is in the stack frame + ! that is deallocated by the return. + .align 16 +2: +E(SPARC_RET_VOID) + return %i7+8 + nop +E(SPARC_RET_STRUCT) + ldx [FP-160], %i0 ldd [FP-160], %f0 - ldd [FP-152], %f2 - ldd [FP-144], %f4 - ldd [FP-136], %f6 - -integer: + b 8f + ldx [FP-152], %i1 +E(SPARC_RET_UINT8) + ldub [FP-160+7], %i0 + return %i7+8 + nop +E(SPARC_RET_SINT8) + ldsb [FP-160+7], %i0 + return %i7+8 + nop +E(SPARC_RET_UINT16) + lduh [FP-160+6], %i0 + return %i7+8 + nop +E(SPARC_RET_SINT16) + ldsh [FP-160+6], %i0 + return %i7+8 + nop +E(SPARC_RET_UINT32) + lduw [FP-160+4], %i0 + return %i7+8 + nop +E(SP_V9_RET_SINT32) + ldsw [FP-160+4], %i0 + return %i7+8 + nop +E(SPARC_RET_INT64) ldx [FP-160], %i0 + return %i7+8 + nop +E(SPARC_RET_INT128) + ldx [FP-160], %i0 + ldx [FP-160+8], %i1 + return %i7+8 + nop +E(SPARC_RET_F_8) + ld [FP-160+7*4], %f7 + nop + ld [FP-160+6*4], %f6 + nop +E(SPARC_RET_F_6) + ld [FP-160+5*4], %f5 + nop + ld [FP-160+4*4], %f4 + nop +E(SPARC_RET_F_4) + ldd [FP-160], %f0 + ldd [FP-160+8], %f2 + return %i7+8 + nop +E(SPARC_RET_F_2) + ldd [FP-160], %f0 + return %i7+8 + nop +E(SP_V9_RET_F_3) + ld [FP-160+2*4], %f2 + nop + ld [FP-160+1*4], %f1 + nop +E(SPARC_RET_F_1) + ld [FP-160], %f0 + return %i7+8 + nop -done1: - ret - restore - -#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE -longdouble1: - ldd [FP-152], %f2 - ret - restore -#endif -.LLFE2: + ! Finish the SPARC_RET_STRUCT sequence. + .align 8 +8: ldd [FP-152], %f2 + ldx [FP-144], %i2 + ldd [FP-144], %f4 + ldx [FP-136], %i3 + ldd [FP-136], %f6 + return %i7+8 + nop -.ffi_closure_v9_end: - .size ffi_closure_v9,.ffi_closure_v9_end-ffi_closure_v9 +.LUW8: + .size C(ffi_closure_v9), . - C(ffi_closure_v9) #ifdef HAVE_RO_EH_FRAME - .section ".eh_frame",#alloc -#else - .section ".eh_frame",#alloc,#write -#endif -.LLframe1: - .uaword .LLECIE1-.LLSCIE1 ! Length of Common Information Entry -.LLSCIE1: - .uaword 0x0 ! CIE Identifier Tag - .byte 0x1 ! CIE Version - .ascii "zR\0" ! CIE Augmentation - .byte 0x1 ! uleb128 0x1; CIE Code Alignment Factor - .byte 0x78 ! sleb128 -8; CIE Data Alignment Factor - .byte 0xf ! CIE RA Column - .byte 0x1 ! uleb128 0x1; Augmentation size -#ifdef HAVE_AS_SPARC_UA_PCREL - .byte 0x1b ! FDE Encoding (pcrel sdata4) + .section ".eh_frame",#alloc #else - .byte 0x50 ! FDE Encoding (aligned absolute) + .section ".eh_frame",#alloc,#write #endif - .byte 0xc ! DW_CFA_def_cfa - .byte 0xe ! uleb128 0xe - .byte 0xff,0xf ! uleb128 0x7ff - .align 8 -.LLECIE1: -.LLSFDE1: - .uaword .LLEFDE1-.LLASFDE1 ! FDE Length -.LLASFDE1: - .uaword .LLASFDE1-.LLframe1 ! FDE CIE offset + #ifdef HAVE_AS_SPARC_UA_PCREL - .uaword %r_disp32(.LLFB1) - .uaword .LLFE1-.LLFB1 ! FDE address range +# define FDE_RANGE(B, E) .long %r_disp32(B), E - B #else - .align 8 - .xword .LLFB1 - .uaxword .LLFE1-.LLFB1 ! FDE address range +# define FDE_RANGE(B, E) .align 8; .xword B, E - B #endif - .byte 0x0 ! uleb128 0x0; Augmentation size - .byte 0x4 ! DW_CFA_advance_loc4 - .uaword .LLCFI0-.LLFB1 - .byte 0xd ! DW_CFA_def_cfa_register - .byte 0x1e ! uleb128 0x1e - .byte 0x2d ! DW_CFA_GNU_window_save - .byte 0x9 ! DW_CFA_register - .byte 0xf ! uleb128 0xf - .byte 0x1f ! uleb128 0x1f + .align 8 -.LLEFDE1: -.LLSFDE2: - .uaword .LLEFDE2-.LLASFDE2 ! FDE Length -.LLASFDE2: - .uaword .LLASFDE2-.LLframe1 ! FDE CIE offset +.LCIE: + .long .LECIE - .LSCIE ! CIE Length +.LSCIE: + .long 0 ! CIE Identifier Tag + .byte 1 ! CIE Version + .ascii "zR\0" ! CIE Augmentation + .byte 4 ! CIE Code Alignment Factor + .byte 0x78 ! CIE Data Alignment Factor + .byte 15 ! CIE RA Column + .byte 1 ! Augmentation size #ifdef HAVE_AS_SPARC_UA_PCREL - .uaword %r_disp32(.LLFB2) - .uaword .LLFE2-.LLFB2 ! FDE address range + .byte 0x1b ! FDE Encoding (pcrel sdata4) #else - .align 8 - .xword .LLFB2 - .uaxword .LLFE2-.LLFB2 ! FDE address range -#endif - .byte 0x0 ! uleb128 0x0; Augmentation size - .byte 0x4 ! DW_CFA_advance_loc4 - .uaword .LLCFI1-.LLFB2 - .byte 0xd ! DW_CFA_def_cfa_register - .byte 0x1e ! uleb128 0x1e - .byte 0x2d ! DW_CFA_GNU_window_save - .byte 0x9 ! DW_CFA_register - .byte 0xf ! uleb128 0xf - .byte 0x1f ! uleb128 0x1f - .align 8 -.LLEFDE2: + .byte 0x50 ! FDE Encoding (aligned absolute) #endif - + .byte 0xc, 14, 0xff, 0xf ! DW_CFA_def_cfa, %o6, offset 0x7ff + .align 8 +.LECIE: + + .long .LEFDE1 - .LSFDE1 ! FDE Length +.LSFDE1: + .long .LSFDE1 - .LCIE ! FDE CIE offset + FDE_RANGE(.LUW0, .LUW2) + .byte 0 ! Augmentation size + .byte 0x40+1 ! DW_CFA_advance_loc 4 + .byte 0xd, 30 ! DW_CFA_def_cfa_register, %i6 + .byte 0x2d ! DW_CFA_GNU_window_save + .byte 0x9, 15, 31 ! DW_CFA_register, %o7, %i7 + .align 8 +.LEFDE1: + + .long .LEFDE2 - .LSFDE2 ! FDE Length +.LSFDE2: + .long .LSFDE2 - .LCIE ! FDE CIE offset + FDE_RANGE(.LUW3, .LUW5) + .byte 0 ! Augmentation size + .byte 0x40+1 ! DW_CFA_advance_loc 4 + .byte 0xd, 30 ! DW_CFA_def_cfa_register, %i6 + .byte 0x2d ! DW_CFA_GNU_window_save + .byte 0x9, 15, 31 ! DW_CFA_register, %o7, %i7 + .align 8 +.LEFDE2: + + .long .LEFDE3 - .LSFDE3 ! FDE Length +.LSFDE3: + .long .LSFDE3 - .LCIE ! FDE CIE offset + FDE_RANGE(.LUW6, .LUW8) + .byte 0 ! Augmentation size + .byte 0x40+1 ! DW_CFA_advance_loc 4 + .byte 0xd, 30 ! DW_CFA_def_cfa_register, %i6 + .byte 0x2d ! DW_CFA_GNU_window_save + .byte 0x9, 15, 31 ! DW_CFA_register, %o7, %i7 + .align 8 +.LEFDE3: + +#endif /* SPARC64 */ #ifdef __linux__ .section .note.GNU-stack,"",@progbits #endif diff --git a/native/libffi/src/types.c b/native/libffi/src/types.c index 7e80aec6eb..eb18c78696 100644 --- a/native/libffi/src/types.c +++ b/native/libffi/src/types.c @@ -38,6 +38,10 @@ struct struct_align_##name { \ char c; \ type x; \ }; \ +<<<<<<< HEAD +======= +FFI_EXTERN \ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 maybe_const ffi_type ffi_type_##name = { \ sizeof(type), \ offsetof(struct struct_align_##name, x), \ @@ -52,6 +56,10 @@ struct struct_align_complex_##name { \ char c; \ _Complex type x; \ }; \ +<<<<<<< HEAD +======= +FFI_EXTERN \ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 maybe_const ffi_type ffi_type_complex_##name = { \ sizeof(_Complex type), \ offsetof(struct struct_align_complex_##name, x), \ @@ -60,7 +68,7 @@ maybe_const ffi_type ffi_type_complex_##name = { \ } /* Size and alignment are fake here. They must not be 0. */ -const ffi_type ffi_type_void = { +FFI_EXTERN const ffi_type ffi_type_void = { 1, 1, FFI_TYPE_VOID, NULL }; diff --git a/native/libffi/src/vax/ffi.c b/native/libffi/src/vax/ffi.c index f4d6bbb4f4..c1544fdf53 100644 --- a/native/libffi/src/vax/ffi.c +++ b/native/libffi/src/vax/ffi.c @@ -108,7 +108,11 @@ ffi_prep_args (extended_cif *ecif, void *stack) /* Align if necessary. */ if ((sizeof(int) - 1) & z) +<<<<<<< HEAD z = ALIGN(z, sizeof(int)); +======= + z = FFI_ALIGN(z, sizeof(int)); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } p_argv++; @@ -215,7 +219,11 @@ ffi_prep_closure_elfbsd (ffi_cif *cif, void **avalue, char *stackp) /* Align if necessary */ if ((sizeof (int) - 1) & z) +<<<<<<< HEAD z = ALIGN(z, sizeof (int)); +======= + z = FFI_ALIGN(z, sizeof (int)); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 p_argv++; stackp += z; diff --git a/native/libffi/src/x86/asmnames.h b/native/libffi/src/x86/asmnames.h new file mode 100644 index 0000000000..7551021e17 --- /dev/null +++ b/native/libffi/src/x86/asmnames.h @@ -0,0 +1,30 @@ +#ifndef ASMNAMES_H +#define ASMNAMES_H + +#define C2(X, Y) X ## Y +#define C1(X, Y) C2(X, Y) +#ifdef __USER_LABEL_PREFIX__ +# define C(X) C1(__USER_LABEL_PREFIX__, X) +#else +# define C(X) X +#endif + +#ifdef __APPLE__ +# define L(X) C1(L, X) +#else +# define L(X) C1(.L, X) +#endif + +#if defined(__ELF__) && defined(__PIC__) +# define PLT(X) X@PLT +#else +# define PLT(X) X +#endif + +#ifdef __ELF__ +# define ENDF(X) .type X,@function; .size X, . - X +#else +# define ENDF(X) +#endif + +#endif /* ASMNAMES_H */ diff --git a/native/libffi/src/x86/darwin.S b/native/libffi/src/x86/darwin.S deleted file mode 100644 index 8f0f0707aa..0000000000 --- a/native/libffi/src/x86/darwin.S +++ /dev/null @@ -1,444 +0,0 @@ -/* ----------------------------------------------------------------------- - darwin.S - Copyright (c) 1996, 1998, 2001, 2002, 2003, 2005 Red Hat, Inc. - Copyright (C) 2008 Free Software Foundation, Inc. - - X86 Foreign Function Interface - - Permission is hereby granted, free of charge, to any person obtaining - a copy of this software and associated documentation files (the - ``Software''), to deal in the Software without restriction, including - without limitation the rights to use, copy, modify, merge, publish, - distribute, sublicense, and/or sell copies of the Software, and to - permit persons to whom the Software is furnished to do so, subject to - the following conditions: - - The above copyright notice and this permission notice shall be included - in all copies or substantial portions of the Software. - - THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, - EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND - NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT - HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, - WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, - OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - DEALINGS IN THE SOFTWARE. - ----------------------------------------------------------------------- - */ - -#ifndef __x86_64__ - -#define LIBFFI_ASM -#include -#include - -.text - -.globl _ffi_prep_args - - .align 4 -.globl _ffi_call_SYSV - -_ffi_call_SYSV: -.LFB1: - pushl %ebp -.LCFI0: - movl %esp,%ebp -.LCFI1: - subl $8,%esp - /* Make room for all of the new args. */ - movl 16(%ebp),%ecx - subl %ecx,%esp - - movl %esp,%eax - - /* Place all of the ffi_prep_args in position */ - subl $8,%esp - pushl 12(%ebp) - pushl %eax - call *8(%ebp) - - /* Return stack to previous state and call the function */ - addl $16,%esp - - call *28(%ebp) - - /* Load %ecx with the return type code */ - movl 20(%ebp),%ecx - - /* Protect %esi. We're going to pop it in the epilogue. */ - pushl %esi - - /* If the return value pointer is NULL, assume no return value. */ - cmpl $0,24(%ebp) - jne 0f - - /* Even if there is no space for the return value, we are - obliged to handle floating-point values. */ - cmpl $FFI_TYPE_FLOAT,%ecx - jne noretval - fstp %st(0) - - jmp epilogue -0: - .align 4 - call 1f -.Lstore_table: - .long noretval-.Lstore_table /* FFI_TYPE_VOID */ - .long retint-.Lstore_table /* FFI_TYPE_INT */ - .long retfloat-.Lstore_table /* FFI_TYPE_FLOAT */ - .long retdouble-.Lstore_table /* FFI_TYPE_DOUBLE */ - .long retlongdouble-.Lstore_table /* FFI_TYPE_LONGDOUBLE */ - .long retuint8-.Lstore_table /* FFI_TYPE_UINT8 */ - .long retsint8-.Lstore_table /* FFI_TYPE_SINT8 */ - .long retuint16-.Lstore_table /* FFI_TYPE_UINT16 */ - .long retsint16-.Lstore_table /* FFI_TYPE_SINT16 */ - .long retint-.Lstore_table /* FFI_TYPE_UINT32 */ - .long retint-.Lstore_table /* FFI_TYPE_SINT32 */ - .long retint64-.Lstore_table /* FFI_TYPE_UINT64 */ - .long retint64-.Lstore_table /* FFI_TYPE_SINT64 */ - .long retstruct-.Lstore_table /* FFI_TYPE_STRUCT */ - .long retint-.Lstore_table /* FFI_TYPE_POINTER */ - .long retstruct1b-.Lstore_table /* FFI_TYPE_SMALL_STRUCT_1B */ - .long retstruct2b-.Lstore_table /* FFI_TYPE_SMALL_STRUCT_2B */ -1: - pop %esi - add (%esi, %ecx, 4), %esi - jmp *%esi - - /* Sign/zero extend as appropriate. */ -retsint8: - movsbl %al, %eax - jmp retint - -retsint16: - movswl %ax, %eax - jmp retint - -retuint8: - movzbl %al, %eax - jmp retint - -retuint16: - movzwl %ax, %eax - jmp retint - -retfloat: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - fstps (%ecx) - jmp epilogue - -retdouble: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - fstpl (%ecx) - jmp epilogue - -retlongdouble: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - fstpt (%ecx) - jmp epilogue - -retint64: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - movl %eax,0(%ecx) - movl %edx,4(%ecx) - jmp epilogue - -retstruct1b: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - movb %al,0(%ecx) - jmp epilogue - -retstruct2b: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - movw %ax,0(%ecx) - jmp epilogue - -retint: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - movl %eax,0(%ecx) - -retstruct: - /* Nothing to do! */ - -noretval: -epilogue: - popl %esi - movl %ebp,%esp - popl %ebp - ret - -.LFE1: -.ffi_call_SYSV_end: - - .align 4 -FFI_HIDDEN (ffi_closure_SYSV) -.globl _ffi_closure_SYSV - -_ffi_closure_SYSV: -.LFB2: - pushl %ebp -.LCFI2: - movl %esp, %ebp -.LCFI3: - subl $40, %esp - leal -24(%ebp), %edx - movl %edx, -12(%ebp) /* resp */ - leal 8(%ebp), %edx - movl %edx, 4(%esp) /* args = __builtin_dwarf_cfa () */ - leal -12(%ebp), %edx - movl %edx, (%esp) /* &resp */ - movl %ebx, 8(%esp) -.LCFI7: - call L_ffi_closure_SYSV_inner$stub - movl 8(%esp), %ebx - movl -12(%ebp), %ecx - cmpl $FFI_TYPE_INT, %eax - je .Lcls_retint - - /* Handle FFI_TYPE_UINT8, FFI_TYPE_SINT8, FFI_TYPE_UINT16, - FFI_TYPE_SINT16, FFI_TYPE_UINT32, FFI_TYPE_SINT32. */ - cmpl $FFI_TYPE_UINT64, %eax - jge 0f - cmpl $FFI_TYPE_UINT8, %eax - jge .Lcls_retint - -0: cmpl $FFI_TYPE_FLOAT, %eax - je .Lcls_retfloat - cmpl $FFI_TYPE_DOUBLE, %eax - je .Lcls_retdouble - cmpl $FFI_TYPE_LONGDOUBLE, %eax - je .Lcls_retldouble - cmpl $FFI_TYPE_SINT64, %eax - je .Lcls_retllong - cmpl $FFI_TYPE_SMALL_STRUCT_1B, %eax - je .Lcls_retstruct1b - cmpl $FFI_TYPE_SMALL_STRUCT_2B, %eax - je .Lcls_retstruct2b - cmpl $FFI_TYPE_STRUCT, %eax - je .Lcls_retstruct -.Lcls_epilogue: - movl %ebp, %esp - popl %ebp - ret -.Lcls_retint: - movl (%ecx), %eax - jmp .Lcls_epilogue -.Lcls_retfloat: - flds (%ecx) - jmp .Lcls_epilogue -.Lcls_retdouble: - fldl (%ecx) - jmp .Lcls_epilogue -.Lcls_retldouble: - fldt (%ecx) - jmp .Lcls_epilogue -.Lcls_retllong: - movl (%ecx), %eax - movl 4(%ecx), %edx - jmp .Lcls_epilogue -.Lcls_retstruct1b: - movsbl (%ecx), %eax - jmp .Lcls_epilogue -.Lcls_retstruct2b: - movswl (%ecx), %eax - jmp .Lcls_epilogue -.Lcls_retstruct: - lea -8(%ebp),%esp - movl %ebp, %esp - popl %ebp - ret $4 -.LFE2: - -#if !FFI_NO_RAW_API - -#define RAW_CLOSURE_CIF_OFFSET ((FFI_TRAMPOLINE_SIZE + 3) & ~3) -#define RAW_CLOSURE_FUN_OFFSET (RAW_CLOSURE_CIF_OFFSET + 4) -#define RAW_CLOSURE_USER_DATA_OFFSET (RAW_CLOSURE_FUN_OFFSET + 4) -#define CIF_FLAGS_OFFSET 20 - - .align 4 -FFI_HIDDEN (ffi_closure_raw_SYSV) -.globl _ffi_closure_raw_SYSV - -_ffi_closure_raw_SYSV: -.LFB3: - pushl %ebp -.LCFI4: - movl %esp, %ebp -.LCFI5: - pushl %esi -.LCFI6: - subl $36, %esp - movl RAW_CLOSURE_CIF_OFFSET(%eax), %esi /* closure->cif */ - movl RAW_CLOSURE_USER_DATA_OFFSET(%eax), %edx /* closure->user_data */ - movl %edx, 12(%esp) /* user_data */ - leal 8(%ebp), %edx /* __builtin_dwarf_cfa () */ - movl %edx, 8(%esp) /* raw_args */ - leal -24(%ebp), %edx - movl %edx, 4(%esp) /* &res */ - movl %esi, (%esp) /* cif */ - call *RAW_CLOSURE_FUN_OFFSET(%eax) /* closure->fun */ - movl CIF_FLAGS_OFFSET(%esi), %eax /* rtype */ - cmpl $FFI_TYPE_INT, %eax - je .Lrcls_retint - - /* Handle FFI_TYPE_UINT8, FFI_TYPE_SINT8, FFI_TYPE_UINT16, - FFI_TYPE_SINT16, FFI_TYPE_UINT32, FFI_TYPE_SINT32. */ - cmpl $FFI_TYPE_UINT64, %eax - jge 0f - cmpl $FFI_TYPE_UINT8, %eax - jge .Lrcls_retint -0: - cmpl $FFI_TYPE_FLOAT, %eax - je .Lrcls_retfloat - cmpl $FFI_TYPE_DOUBLE, %eax - je .Lrcls_retdouble - cmpl $FFI_TYPE_LONGDOUBLE, %eax - je .Lrcls_retldouble - cmpl $FFI_TYPE_SINT64, %eax - je .Lrcls_retllong -.Lrcls_epilogue: - addl $36, %esp - popl %esi - popl %ebp - ret -.Lrcls_retint: - movl -24(%ebp), %eax - jmp .Lrcls_epilogue -.Lrcls_retfloat: - flds -24(%ebp) - jmp .Lrcls_epilogue -.Lrcls_retdouble: - fldl -24(%ebp) - jmp .Lrcls_epilogue -.Lrcls_retldouble: - fldt -24(%ebp) - jmp .Lrcls_epilogue -.Lrcls_retllong: - movl -24(%ebp), %eax - movl -20(%ebp), %edx - jmp .Lrcls_epilogue -.LFE3: -#endif - -.section __IMPORT,__jump_table,symbol_stubs,self_modifying_code+pure_instructions,5 -L_ffi_closure_SYSV_inner$stub: - .indirect_symbol _ffi_closure_SYSV_inner - hlt ; hlt ; hlt ; hlt ; hlt - - -.section __TEXT,__eh_frame,coalesced,no_toc+strip_static_syms+live_support -EH_frame1: - .set L$set$0,LECIE1-LSCIE1 - .long L$set$0 -LSCIE1: - .long 0x0 - .byte 0x1 - .ascii "zR\0" - .byte 0x1 - .byte 0x7c - .byte 0x8 - .byte 0x1 - .byte 0x10 - .byte 0xc - .byte 0x5 - .byte 0x4 - .byte 0x88 - .byte 0x1 - .align 2 -LECIE1: -.globl _ffi_call_SYSV.eh -_ffi_call_SYSV.eh: -LSFDE1: - .set L$set$1,LEFDE1-LASFDE1 - .long L$set$1 -LASFDE1: - .long LASFDE1-EH_frame1 - .long .LFB1-. - .set L$set$2,.LFE1-.LFB1 - .long L$set$2 - .byte 0x0 - .byte 0x4 - .set L$set$3,.LCFI0-.LFB1 - .long L$set$3 - .byte 0xe - .byte 0x8 - .byte 0x84 - .byte 0x2 - .byte 0x4 - .set L$set$4,.LCFI1-.LCFI0 - .long L$set$4 - .byte 0xd - .byte 0x4 - .align 2 -LEFDE1: -.globl _ffi_closure_SYSV.eh -_ffi_closure_SYSV.eh: -LSFDE2: - .set L$set$5,LEFDE2-LASFDE2 - .long L$set$5 -LASFDE2: - .long LASFDE2-EH_frame1 - .long .LFB2-. - .set L$set$6,.LFE2-.LFB2 - .long L$set$6 - .byte 0x0 - .byte 0x4 - .set L$set$7,.LCFI2-.LFB2 - .long L$set$7 - .byte 0xe - .byte 0x8 - .byte 0x84 - .byte 0x2 - .byte 0x4 - .set L$set$8,.LCFI3-.LCFI2 - .long L$set$8 - .byte 0xd - .byte 0x4 - .align 2 -LEFDE2: - -#if !FFI_NO_RAW_API - -.globl _ffi_closure_raw_SYSV.eh -_ffi_closure_raw_SYSV.eh: -LSFDE3: - .set L$set$10,LEFDE3-LASFDE3 - .long L$set$10 -LASFDE3: - .long LASFDE3-EH_frame1 - .long .LFB3-. - .set L$set$11,.LFE3-.LFB3 - .long L$set$11 - .byte 0x0 - .byte 0x4 - .set L$set$12,.LCFI4-.LFB3 - .long L$set$12 - .byte 0xe - .byte 0x8 - .byte 0x84 - .byte 0x2 - .byte 0x4 - .set L$set$13,.LCFI5-.LCFI4 - .long L$set$13 - .byte 0xd - .byte 0x4 - .byte 0x4 - .set L$set$14,.LCFI6-.LCFI5 - .long L$set$14 - .byte 0x85 - .byte 0x3 - .align 2 -LEFDE3: - -#endif - -#endif /* ifndef __x86_64__ */ diff --git a/native/libffi/src/x86/darwin64.S b/native/libffi/src/x86/darwin64.S deleted file mode 100644 index 2f7394ef4b..0000000000 --- a/native/libffi/src/x86/darwin64.S +++ /dev/null @@ -1,416 +0,0 @@ -/* ----------------------------------------------------------------------- - darwin64.S - Copyright (c) 2006 Free Software Foundation, Inc. - Copyright (c) 2008 Red Hat, Inc. - derived from unix64.S - - x86-64 Foreign Function Interface for Darwin. - - Permission is hereby granted, free of charge, to any person obtaining - a copy of this software and associated documentation files (the - ``Software''), to deal in the Software without restriction, including - without limitation the rights to use, copy, modify, merge, publish, - distribute, sublicense, and/or sell copies of the Software, and to - permit persons to whom the Software is furnished to do so, subject to - the following conditions: - - The above copyright notice and this permission notice shall be included - in all copies or substantial portions of the Software. - - THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, EXPRESS - OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. - IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY CLAIM, DAMAGES OR - OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - OTHER DEALINGS IN THE SOFTWARE. - ----------------------------------------------------------------------- */ - -#ifdef __x86_64__ -#define LIBFFI_ASM -#include -#include - - .file "darwin64.S" -.text - -/* ffi_call_unix64 (void *args, unsigned long bytes, unsigned flags, - void *raddr, void (*fnaddr)(void)); - - Bit o trickiness here -- ARGS+BYTES is the base of the stack frame - for this function. This has been allocated by ffi_call. We also - deallocate some of the stack that has been alloca'd. */ - - .align 3 - .globl _ffi_call_unix64 - -_ffi_call_unix64: -LUW0: - movq (%rsp), %r10 /* Load return address. */ - leaq (%rdi, %rsi), %rax /* Find local stack base. */ - movq %rdx, (%rax) /* Save flags. */ - movq %rcx, 8(%rax) /* Save raddr. */ - movq %rbp, 16(%rax) /* Save old frame pointer. */ - movq %r10, 24(%rax) /* Relocate return address. */ - movq %rax, %rbp /* Finalize local stack frame. */ -LUW1: - movq %rdi, %r10 /* Save a copy of the register area. */ - movq %r8, %r11 /* Save a copy of the target fn. */ - movl %r9d, %eax /* Set number of SSE registers. */ - - /* Load up all argument registers. */ - movq (%r10), %rdi - movq 8(%r10), %rsi - movq 16(%r10), %rdx - movq 24(%r10), %rcx - movq 32(%r10), %r8 - movq 40(%r10), %r9 - testl %eax, %eax - jnz Lload_sse -Lret_from_load_sse: - - /* Deallocate the reg arg area. */ - leaq 176(%r10), %rsp - - /* Call the user function. */ - call *%r11 - - /* Deallocate stack arg area; local stack frame in redzone. */ - leaq 24(%rbp), %rsp - - movq 0(%rbp), %rcx /* Reload flags. */ - movq 8(%rbp), %rdi /* Reload raddr. */ - movq 16(%rbp), %rbp /* Reload old frame pointer. */ -LUW2: - - /* The first byte of the flags contains the FFI_TYPE. */ - movzbl %cl, %r10d - leaq Lstore_table(%rip), %r11 - movslq (%r11, %r10, 4), %r10 - addq %r11, %r10 - jmp *%r10 - -Lstore_table: - .long Lst_void-Lstore_table /* FFI_TYPE_VOID */ - .long Lst_sint32-Lstore_table /* FFI_TYPE_INT */ - .long Lst_float-Lstore_table /* FFI_TYPE_FLOAT */ - .long Lst_double-Lstore_table /* FFI_TYPE_DOUBLE */ - .long Lst_ldouble-Lstore_table /* FFI_TYPE_LONGDOUBLE */ - .long Lst_uint8-Lstore_table /* FFI_TYPE_UINT8 */ - .long Lst_sint8-Lstore_table /* FFI_TYPE_SINT8 */ - .long Lst_uint16-Lstore_table /* FFI_TYPE_UINT16 */ - .long Lst_sint16-Lstore_table /* FFI_TYPE_SINT16 */ - .long Lst_uint32-Lstore_table /* FFI_TYPE_UINT32 */ - .long Lst_sint32-Lstore_table /* FFI_TYPE_SINT32 */ - .long Lst_int64-Lstore_table /* FFI_TYPE_UINT64 */ - .long Lst_int64-Lstore_table /* FFI_TYPE_SINT64 */ - .long Lst_struct-Lstore_table /* FFI_TYPE_STRUCT */ - .long Lst_int64-Lstore_table /* FFI_TYPE_POINTER */ - - .text - .align 3 -Lst_void: - ret - .align 3 -Lst_uint8: - movzbq %al, %rax - movq %rax, (%rdi) - ret - .align 3 -Lst_sint8: - movsbq %al, %rax - movq %rax, (%rdi) - ret - .align 3 -Lst_uint16: - movzwq %ax, %rax - movq %rax, (%rdi) - .align 3 -Lst_sint16: - movswq %ax, %rax - movq %rax, (%rdi) - ret - .align 3 -Lst_uint32: - movl %eax, %eax - movq %rax, (%rdi) - .align 3 -Lst_sint32: - cltq - movq %rax, (%rdi) - ret - .align 3 -Lst_int64: - movq %rax, (%rdi) - ret - .align 3 -Lst_float: - movss %xmm0, (%rdi) - ret - .align 3 -Lst_double: - movsd %xmm0, (%rdi) - ret -Lst_ldouble: - fstpt (%rdi) - ret - .align 3 -Lst_struct: - leaq -20(%rsp), %rsi /* Scratch area in redzone. */ - - /* We have to locate the values now, and since we don't want to - write too much data into the user's return value, we spill the - value to a 16 byte scratch area first. Bits 8, 9, and 10 - control where the values are located. Only one of the three - bits will be set; see ffi_prep_cif_machdep for the pattern. */ - movd %xmm0, %r10 - movd %xmm1, %r11 - testl $0x100, %ecx - cmovnz %rax, %rdx - cmovnz %r10, %rax - testl $0x200, %ecx - cmovnz %r10, %rdx - testl $0x400, %ecx - cmovnz %r10, %rax - cmovnz %r11, %rdx - movq %rax, (%rsi) - movq %rdx, 8(%rsi) - - /* Bits 12-31 contain the true size of the structure. Copy from - the scratch area to the true destination. */ - shrl $12, %ecx - rep movsb - ret - - /* Many times we can avoid loading any SSE registers at all. - It's not worth an indirect jump to load the exact set of - SSE registers needed; zero or all is a good compromise. */ - .align 3 -LUW3: -Lload_sse: - movdqa 48(%r10), %xmm0 - movdqa 64(%r10), %xmm1 - movdqa 80(%r10), %xmm2 - movdqa 96(%r10), %xmm3 - movdqa 112(%r10), %xmm4 - movdqa 128(%r10), %xmm5 - movdqa 144(%r10), %xmm6 - movdqa 160(%r10), %xmm7 - jmp Lret_from_load_sse - -LUW4: - .align 3 - .globl _ffi_closure_unix64 - -_ffi_closure_unix64: -LUW5: - /* The carry flag is set by the trampoline iff SSE registers - are used. Don't clobber it before the branch instruction. */ - leaq -200(%rsp), %rsp -LUW6: - movq %rdi, (%rsp) - movq %rsi, 8(%rsp) - movq %rdx, 16(%rsp) - movq %rcx, 24(%rsp) - movq %r8, 32(%rsp) - movq %r9, 40(%rsp) - jc Lsave_sse -Lret_from_save_sse: - - movq %r10, %rdi - leaq 176(%rsp), %rsi - movq %rsp, %rdx - leaq 208(%rsp), %rcx - call _ffi_closure_unix64_inner - - /* Deallocate stack frame early; return value is now in redzone. */ - addq $200, %rsp -LUW7: - - /* The first byte of the return value contains the FFI_TYPE. */ - movzbl %al, %r10d - leaq Lload_table(%rip), %r11 - movslq (%r11, %r10, 4), %r10 - addq %r11, %r10 - jmp *%r10 - -Lload_table: - .long Lld_void-Lload_table /* FFI_TYPE_VOID */ - .long Lld_int32-Lload_table /* FFI_TYPE_INT */ - .long Lld_float-Lload_table /* FFI_TYPE_FLOAT */ - .long Lld_double-Lload_table /* FFI_TYPE_DOUBLE */ - .long Lld_ldouble-Lload_table /* FFI_TYPE_LONGDOUBLE */ - .long Lld_int8-Lload_table /* FFI_TYPE_UINT8 */ - .long Lld_int8-Lload_table /* FFI_TYPE_SINT8 */ - .long Lld_int16-Lload_table /* FFI_TYPE_UINT16 */ - .long Lld_int16-Lload_table /* FFI_TYPE_SINT16 */ - .long Lld_int32-Lload_table /* FFI_TYPE_UINT32 */ - .long Lld_int32-Lload_table /* FFI_TYPE_SINT32 */ - .long Lld_int64-Lload_table /* FFI_TYPE_UINT64 */ - .long Lld_int64-Lload_table /* FFI_TYPE_SINT64 */ - .long Lld_struct-Lload_table /* FFI_TYPE_STRUCT */ - .long Lld_int64-Lload_table /* FFI_TYPE_POINTER */ - - .text - .align 3 -Lld_void: - ret - .align 3 -Lld_int8: - movzbl -24(%rsp), %eax - ret - .align 3 -Lld_int16: - movzwl -24(%rsp), %eax - ret - .align 3 -Lld_int32: - movl -24(%rsp), %eax - ret - .align 3 -Lld_int64: - movq -24(%rsp), %rax - ret - .align 3 -Lld_float: - movss -24(%rsp), %xmm0 - ret - .align 3 -Lld_double: - movsd -24(%rsp), %xmm0 - ret - .align 3 -Lld_ldouble: - fldt -24(%rsp) - ret - .align 3 -Lld_struct: - /* There are four possibilities here, %rax/%rdx, %xmm0/%rax, - %rax/%xmm0, %xmm0/%xmm1. We collapse two by always loading - both rdx and xmm1 with the second word. For the remaining, - bit 8 set means xmm0 gets the second word, and bit 9 means - that rax gets the second word. */ - movq -24(%rsp), %rcx - movq -16(%rsp), %rdx - movq -16(%rsp), %xmm1 - testl $0x100, %eax - cmovnz %rdx, %rcx - movd %rcx, %xmm0 - testl $0x200, %eax - movq -24(%rsp), %rax - cmovnz %rdx, %rax - ret - - /* See the comment above Lload_sse; the same logic applies here. */ - .align 3 -LUW8: -Lsave_sse: - movdqa %xmm0, 48(%rsp) - movdqa %xmm1, 64(%rsp) - movdqa %xmm2, 80(%rsp) - movdqa %xmm3, 96(%rsp) - movdqa %xmm4, 112(%rsp) - movdqa %xmm5, 128(%rsp) - movdqa %xmm6, 144(%rsp) - movdqa %xmm7, 160(%rsp) - jmp Lret_from_save_sse - -LUW9: -.section __TEXT,__eh_frame,coalesced,no_toc+strip_static_syms+live_support -EH_frame1: - .set L$set$0,LECIE1-LSCIE1 /* CIE Length */ - .long L$set$0 -LSCIE1: - .long 0x0 /* CIE Identifier Tag */ - .byte 0x1 /* CIE Version */ - .ascii "zR\0" /* CIE Augmentation */ - .byte 0x1 /* uleb128 0x1; CIE Code Alignment Factor */ - .byte 0x78 /* sleb128 -8; CIE Data Alignment Factor */ - .byte 0x10 /* CIE RA Column */ - .byte 0x1 /* uleb128 0x1; Augmentation size */ - .byte 0x10 /* FDE Encoding (pcrel sdata4) */ - .byte 0xc /* DW_CFA_def_cfa, %rsp offset 8 */ - .byte 0x7 /* uleb128 0x7 */ - .byte 0x8 /* uleb128 0x8 */ - .byte 0x90 /* DW_CFA_offset, column 0x10 */ - .byte 0x1 - .align 3 -LECIE1: - .globl _ffi_call_unix64.eh -_ffi_call_unix64.eh: -LSFDE1: - .set L$set$1,LEFDE1-LASFDE1 /* FDE Length */ - .long L$set$1 -LASFDE1: - .long LASFDE1-EH_frame1 /* FDE CIE offset */ - .quad LUW0-. /* FDE initial location */ - .set L$set$2,LUW4-LUW0 /* FDE address range */ - .quad L$set$2 - .byte 0x0 /* Augmentation size */ - .byte 0x4 /* DW_CFA_advance_loc4 */ - .set L$set$3,LUW1-LUW0 - .long L$set$3 - - /* New stack frame based off rbp. This is a itty bit of unwind - trickery in that the CFA *has* changed. There is no easy way - to describe it correctly on entry to the function. Fortunately, - it doesn't matter too much since at all points we can correctly - unwind back to ffi_call. Note that the location to which we - moved the return address is (the new) CFA-8, so from the - perspective of the unwind info, it hasn't moved. */ - .byte 0xc /* DW_CFA_def_cfa, %rbp offset 32 */ - .byte 0x6 - .byte 0x20 - .byte 0x80+6 /* DW_CFA_offset, %rbp offset 2*-8 */ - .byte 0x2 - .byte 0xa /* DW_CFA_remember_state */ - - .byte 0x4 /* DW_CFA_advance_loc4 */ - .set L$set$4,LUW2-LUW1 - .long L$set$4 - .byte 0xc /* DW_CFA_def_cfa, %rsp offset 8 */ - .byte 0x7 - .byte 0x8 - .byte 0xc0+6 /* DW_CFA_restore, %rbp */ - - .byte 0x4 /* DW_CFA_advance_loc4 */ - .set L$set$5,LUW3-LUW2 - .long L$set$5 - .byte 0xb /* DW_CFA_restore_state */ - - .align 3 -LEFDE1: - .globl _ffi_closure_unix64.eh -_ffi_closure_unix64.eh: -LSFDE3: - .set L$set$6,LEFDE3-LASFDE3 /* FDE Length */ - .long L$set$6 -LASFDE3: - .long LASFDE3-EH_frame1 /* FDE CIE offset */ - .quad LUW5-. /* FDE initial location */ - .set L$set$7,LUW9-LUW5 /* FDE address range */ - .quad L$set$7 - .byte 0x0 /* Augmentation size */ - - .byte 0x4 /* DW_CFA_advance_loc4 */ - .set L$set$8,LUW6-LUW5 - .long L$set$8 - .byte 0xe /* DW_CFA_def_cfa_offset */ - .byte 208,1 /* uleb128 208 */ - .byte 0xa /* DW_CFA_remember_state */ - - .byte 0x4 /* DW_CFA_advance_loc4 */ - .set L$set$9,LUW7-LUW6 - .long L$set$9 - .byte 0xe /* DW_CFA_def_cfa_offset */ - .byte 0x8 - - .byte 0x4 /* DW_CFA_advance_loc4 */ - .set L$set$10,LUW8-LUW7 - .long L$set$10 - .byte 0xb /* DW_CFA_restore_state */ - - .align 3 -LEFDE3: - .subsections_via_symbols - -#endif /* __x86_64__ */ diff --git a/native/libffi/src/x86/ffi.c b/native/libffi/src/x86/ffi.c index 34f80b3bf0..383c74d91d 100644 --- a/native/libffi/src/x86/ffi.c +++ b/native/libffi/src/x86/ffi.c @@ -1,5 +1,6 @@ /* ----------------------------------------------------------------------- - ffi.c - Copyright (c) 1996, 1998, 1999, 2001, 2007, 2008 Red Hat, Inc. + ffi.c - Copyright (c) 2017 Anthony Green + Copyright (c) 1996, 1998, 1999, 2001, 2007, 2008 Red Hat, Inc. Copyright (c) 2002 Ranjit Mathew Copyright (c) 2002 Bo Thorsen Copyright (c) 2002 Roger Sayle @@ -27,6 +28,7 @@ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. ----------------------------------------------------------------------- */ +<<<<<<< HEAD #include #if !defined(__x86_64__) || defined(_WIN64) || defined(__CYGWIN__) @@ -34,10 +36,15 @@ #include #endif +======= + +#if defined(__i386__) || defined(_M_IX86) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #include #include - +#include #include +<<<<<<< HEAD /* ffi_prep_args is called by the assembly routine once stack space @@ -195,11 +202,33 @@ unsigned int ffi_prep_args(char *stack, extended_cif *ecif) #ifdef X86_WIN64 argp += (z + sizeof(void*) - 1) & ~(sizeof(void*) - 1); +======= +#include "internal.h" + +/* Force FFI_TYPE_LONGDOUBLE to be different than FFI_TYPE_DOUBLE; + all further uses in this file will refer to the 80-bit type. */ +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE +# if FFI_TYPE_LONGDOUBLE != 4 +# error FFI_TYPE_LONGDOUBLE out of date +# endif #else - argp += z; +# undef FFI_TYPE_LONGDOUBLE +# define FFI_TYPE_LONGDOUBLE 4 +#endif + +#if defined(__GNUC__) && !defined(__declspec) +# define __declspec(x) __attribute__((x)) +#endif + +#if defined(_MSC_VER) && defined(_M_IX86) +/* Stack is not 16-byte aligned on Windows. */ +#define STACK_ALIGN(bytes) (bytes) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 +#else +#define STACK_ALIGN(bytes) FFI_ALIGN (bytes, 16) #endif - } +<<<<<<< HEAD #ifndef X86_WIN64 /* We need to move the register-passed arguments for thiscall/fastcall/register on top of stack, so that those can be moved to registers by call-handler. */ @@ -231,101 +260,146 @@ unsigned int ffi_prep_args(char *stack, extended_cif *ecif) #endif return 0; } - -/* Perform machine dependent cif processing */ -ffi_status ffi_prep_cif_machdep(ffi_cif *cif) +======= +/* Perform machine dependent cif processing. */ +ffi_status FFI_HIDDEN +ffi_prep_cif_machdep(ffi_cif *cif) { - unsigned int i; - ffi_type **ptr; + size_t bytes = 0; + int i, n, flags, cabi = cif->abi; + + switch (cabi) + { + case FFI_SYSV: + case FFI_STDCALL: + case FFI_THISCALL: + case FFI_FASTCALL: + case FFI_MS_CDECL: + case FFI_PASCAL: + case FFI_REGISTER: + break; + default: + return FFI_BAD_ABI; + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 - /* Set the return type flag */ switch (cif->rtype->type) { case FFI_TYPE_VOID: + flags = X86_RET_VOID; + break; + case FFI_TYPE_FLOAT: + flags = X86_RET_FLOAT; + break; + case FFI_TYPE_DOUBLE: + flags = X86_RET_DOUBLE; + break; + case FFI_TYPE_LONGDOUBLE: + flags = X86_RET_LDOUBLE; + break; case FFI_TYPE_UINT8: + flags = X86_RET_UINT8; + break; case FFI_TYPE_UINT16: + flags = X86_RET_UINT16; + break; case FFI_TYPE_SINT8: + flags = X86_RET_SINT8; + break; case FFI_TYPE_SINT16: -#ifdef X86_WIN64 - case FFI_TYPE_UINT32: + flags = X86_RET_SINT16; + break; + case FFI_TYPE_INT: case FFI_TYPE_SINT32: -#endif - case FFI_TYPE_SINT64: - case FFI_TYPE_FLOAT: - case FFI_TYPE_DOUBLE: -#ifndef X86_WIN64 -#if FFI_TYPE_DOUBLE != FFI_TYPE_LONGDOUBLE - case FFI_TYPE_LONGDOUBLE: -#endif -#endif - cif->flags = (unsigned) cif->rtype->type; + case FFI_TYPE_UINT32: + case FFI_TYPE_POINTER: + flags = X86_RET_INT32; break; - + case FFI_TYPE_SINT64: case FFI_TYPE_UINT64: -#ifdef X86_WIN64 - case FFI_TYPE_POINTER: -#endif - cif->flags = FFI_TYPE_SINT64; + flags = X86_RET_INT64; break; - case FFI_TYPE_STRUCT: #ifndef X86 + /* ??? This should be a different ABI rather than an ifdef. */ if (cif->rtype->size == 1) - { - cif->flags = FFI_TYPE_SMALL_STRUCT_1B; /* same as char size */ - } + flags = X86_RET_STRUCT_1B; else if (cif->rtype->size == 2) - { - cif->flags = FFI_TYPE_SMALL_STRUCT_2B; /* same as short size */ - } + flags = X86_RET_STRUCT_2B; else if (cif->rtype->size == 4) - { -#ifdef X86_WIN64 - cif->flags = FFI_TYPE_SMALL_STRUCT_4B; -#else - cif->flags = FFI_TYPE_INT; /* same as int type */ -#endif - } + flags = X86_RET_INT32; else if (cif->rtype->size == 8) - { - cif->flags = FFI_TYPE_SINT64; /* same as int64 type */ - } + flags = X86_RET_INT64; else #endif - { -#ifdef X86_WIN32 - if (cif->abi == FFI_MS_CDECL) - cif->flags = FFI_TYPE_MS_STRUCT; - else -#endif - cif->flags = FFI_TYPE_STRUCT; - /* allocate space for return value pointer */ - cif->bytes += ALIGN(sizeof(void*), FFI_SIZEOF_ARG); - } - break; - - default: -#ifdef X86_WIN64 - cif->flags = FFI_TYPE_SINT64; + { + do_struct: + switch (cabi) + { + case FFI_THISCALL: + case FFI_FASTCALL: + case FFI_STDCALL: + case FFI_MS_CDECL: + flags = X86_RET_STRUCTARG; + break; + default: + flags = X86_RET_STRUCTPOP; + break; + } + /* Allocate space for return value pointer. */ + bytes += FFI_ALIGN (sizeof(void*), FFI_SIZEOF_ARG); + } break; - case FFI_TYPE_INT: - cif->flags = FFI_TYPE_SINT32; -#else - cif->flags = FFI_TYPE_INT; -#endif + case FFI_TYPE_COMPLEX: + switch (cif->rtype->elements[0]->type) + { + case FFI_TYPE_DOUBLE: + case FFI_TYPE_LONGDOUBLE: + case FFI_TYPE_SINT64: + case FFI_TYPE_UINT64: + goto do_struct; + case FFI_TYPE_FLOAT: + case FFI_TYPE_INT: + case FFI_TYPE_SINT32: + case FFI_TYPE_UINT32: + flags = X86_RET_INT64; + break; + case FFI_TYPE_SINT16: + case FFI_TYPE_UINT16: + flags = X86_RET_INT32; + break; + case FFI_TYPE_SINT8: + case FFI_TYPE_UINT8: + flags = X86_RET_STRUCT_2B; + break; + default: + return FFI_BAD_TYPEDEF; + } break; + default: + return FFI_BAD_TYPEDEF; } + cif->flags = flags; - for (ptr = cif->arg_types, i = cif->nargs; i > 0; i--, ptr++) + for (i = 0, n = cif->nargs; i < n; i++) { +<<<<<<< HEAD #ifndef X86_WIN64 if (cif->abi != FFI_STDCALL && cif->abi != FFI_FASTCALL && cif->abi != FFI_THISCALL) #endif if (((*ptr)->alignment - 1) & cif->bytes) cif->bytes = ALIGN(cif->bytes, (*ptr)->alignment); cif->bytes += (unsigned)ALIGN((*ptr)->size, FFI_SIZEOF_ARG); +======= + ffi_type *t = cif->arg_types[i]; + + bytes = FFI_ALIGN (bytes, t->alignment); + bytes += FFI_ALIGN (t->size, FFI_SIZEOF_ARG); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } + cif->bytes = bytes; +<<<<<<< HEAD #ifdef X86_WIN64 /* ensure space for storing four registers */ cif->bytes += 4 * FFI_SIZEOF_ARG; @@ -337,10 +411,37 @@ ffi_status ffi_prep_cif_machdep(ffi_cif *cif) #endif cif->bytes = (cif->bytes + 15) & ~0xF; #endif - +======= return FFI_OK; } +static ffi_arg +extend_basic_type(void *arg, int type) +{ + switch (type) + { + case FFI_TYPE_SINT8: + return *(SINT8 *)arg; + case FFI_TYPE_UINT8: + return *(UINT8 *)arg; + case FFI_TYPE_SINT16: + return *(SINT16 *)arg; + case FFI_TYPE_UINT16: + return *(UINT16 *)arg; + + case FFI_TYPE_SINT32: + case FFI_TYPE_UINT32: + case FFI_TYPE_POINTER: + case FFI_TYPE_FLOAT: + return *(UINT32 *)arg; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + + default: + abort(); + } +} + +<<<<<<< HEAD #ifdef X86_WIN64 extern int ffi_call_win64(unsigned int (*)(char *, extended_cif *), extended_cif *, @@ -351,10 +452,54 @@ ffi_call_win32(unsigned int (*)(char *, extended_cif *), extended_cif *, unsigned, unsigned, unsigned, unsigned *, void (*fn)(void)); extern void ffi_call_SYSV(void (*)(char *, extended_cif *), extended_cif *, unsigned, unsigned, unsigned *, void (*fn)(void)); +======= +struct call_frame +{ + void *ebp; /* 0 */ + void *retaddr; /* 4 */ + void (*fn)(void); /* 8 */ + int flags; /* 12 */ + void *rvalue; /* 16 */ + unsigned regs[3]; /* 20-28 */ +}; + +struct abi_params +{ + int dir; /* parameter growth direction */ + int static_chain; /* the static chain register used by gcc */ + int nregs; /* number of register parameters */ + int regs[3]; +}; + +static const struct abi_params abi_params[FFI_LAST_ABI] = { + [FFI_SYSV] = { 1, R_ECX, 0 }, + [FFI_THISCALL] = { 1, R_EAX, 1, { R_ECX } }, + [FFI_FASTCALL] = { 1, R_EAX, 2, { R_ECX, R_EDX } }, + [FFI_STDCALL] = { 1, R_ECX, 0 }, + [FFI_PASCAL] = { -1, R_ECX, 0 }, + /* ??? No defined static chain; gcc does not support REGISTER. */ + [FFI_REGISTER] = { -1, R_ECX, 3, { R_EAX, R_EDX, R_ECX } }, + [FFI_MS_CDECL] = { 1, R_ECX, 0 } +}; + +#ifdef HAVE_FASTCALL + #ifdef _MSC_VER + #define FFI_DECLARE_FASTCALL __fastcall + #else + #define FFI_DECLARE_FASTCALL __declspec(fastcall) + #endif +#else + #define FFI_DECLARE_FASTCALL +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #endif -void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +extern void FFI_DECLARE_FASTCALL ffi_call_i386(struct call_frame *, char *) FFI_HIDDEN; + +static void +ffi_call_int (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) { +<<<<<<< HEAD extended_cif ecif; ecif.cif = cif; @@ -370,23 +515,57 @@ void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) && cif->rtype->size != 2 && cif->rtype->size != 4 && cif->rtype->size != 8))) +======= + size_t rsize, bytes; + struct call_frame *frame; + char *stack, *argp; + ffi_type **arg_types; + int flags, cabi, i, n, dir, narg_reg; + const struct abi_params *pabi; + + flags = cif->flags; + cabi = cif->abi; + pabi = &abi_params[cabi]; + dir = pabi->dir; + + rsize = 0; + if (rvalue == NULL) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 { - ecif.rvalue = alloca((cif->rtype->size + 0xF) & ~0xF); + switch (flags) + { + case X86_RET_FLOAT: + case X86_RET_DOUBLE: + case X86_RET_LDOUBLE: + case X86_RET_STRUCTPOP: + case X86_RET_STRUCTARG: + /* The float cases need to pop the 387 stack. + The struct cases need to pass a valid pointer to the callee. */ + rsize = cif->rtype->size; + break; + default: + /* We can pretend that the callee returns nothing. */ + flags = X86_RET_VOID; + break; + } } -#else - if (rvalue == NULL - && (cif->flags == FFI_TYPE_STRUCT - || cif->flags == FFI_TYPE_MS_STRUCT)) - { - ecif.rvalue = alloca(cif->rtype->size); - } -#endif - else - ecif.rvalue = rvalue; - - - switch (cif->abi) + + bytes = STACK_ALIGN (cif->bytes); + stack = alloca(bytes + sizeof(*frame) + rsize); + argp = (dir < 0 ? stack + bytes : stack); + frame = (struct call_frame *)(stack + bytes); + if (rsize) + rvalue = frame + 1; + + frame->fn = fn; + frame->flags = flags; + frame->rvalue = rvalue; + frame->regs[pabi->static_chain] = (unsigned)closure; + + narg_reg = 0; + switch (flags) { +<<<<<<< HEAD #ifdef X86_WIN64 case FFI_WIN64: ffi_call_win64(ffi_prep_args, &ecif, cif->bytes, @@ -413,11 +592,32 @@ void ffi_call(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) #endif default: FFI_ASSERT(0); +======= + case X86_RET_STRUCTARG: + /* The pointer is passed as the first argument. */ + if (pabi->nregs > 0) + { + frame->regs[pabi->regs[0]] = (unsigned)rvalue; + narg_reg = 1; + break; + } + /* fallthru */ + case X86_RET_STRUCTPOP: + *(void **)argp = rvalue; + argp += sizeof(void *); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 break; } -} + arg_types = cif->arg_types; + for (i = 0, n = cif->nargs; i < n; i++) + { + ffi_type *ty = arg_types[i]; + void *valp = avalue[i]; + size_t z = ty->size; + int t = ty->type; +<<<<<<< HEAD /** private members **/ /* The following __attribute__((regparm(1))) decorations will have no effect @@ -474,12 +674,73 @@ ffi_closure_win64_inner (ffi_closure *closure, void *args) { TODO: structure sizes of 3 5 6 7 are returned by reference, too!!! */ return cif->rtype->size > sizeof(void *) ? resp : *(void **)resp; +======= + if (z <= FFI_SIZEOF_ARG && t != FFI_TYPE_STRUCT) + { + ffi_arg val = extend_basic_type (valp, t); + + if (t != FFI_TYPE_FLOAT && narg_reg < pabi->nregs) + frame->regs[pabi->regs[narg_reg++]] = val; + else if (dir < 0) + { + argp -= 4; + *(ffi_arg *)argp = val; + } + else + { + *(ffi_arg *)argp = val; + argp += 4; + } + } + else + { + size_t za = FFI_ALIGN (z, FFI_SIZEOF_ARG); + size_t align = FFI_SIZEOF_ARG; + + /* Issue 434: For thiscall and fastcall, if the paramter passed + as 64-bit integer or struct, all following integer parameters + will be passed on stack. */ + if ((cabi == FFI_THISCALL || cabi == FFI_FASTCALL) + && (t == FFI_TYPE_SINT64 + || t == FFI_TYPE_UINT64 + || t == FFI_TYPE_STRUCT)) + narg_reg = 2; + + /* Alignment rules for arguments are quite complex. Vectors and + structures with 16 byte alignment get it. Note that long double + on Darwin does have 16 byte alignment, and does not get this + alignment if passed directly; a structure with a long double + inside, however, would get 16 byte alignment. Since libffi does + not support vectors, we need non concern ourselves with other + cases. */ + if (t == FFI_TYPE_STRUCT && ty->alignment >= 16) + align = 16; + + if (dir < 0) + { + /* ??? These reverse argument ABIs are probably too old + to have cared about alignment. Someone should check. */ + argp -= za; + memcpy (argp, valp, z); + } + else + { + argp = (char *)FFI_ALIGN (argp, align); + memcpy (argp, valp, z); + argp += za; + } + } + } + FFI_ASSERT (dir > 0 || argp == stack); + + ffi_call_i386 (frame, stack); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } -#else -unsigned int FFI_HIDDEN __attribute__ ((regparm(1))) -ffi_closure_SYSV_inner (ffi_closure *closure, void **respp, void *args) +void +ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) { +<<<<<<< HEAD /* our various things... */ ffi_cif *cif; void **arg_area; @@ -728,6 +989,141 @@ ffi_prep_incoming_args(char *stack, void **rvalue, void **avalue, } /* the cif must already be prep'ed */ +======= + ffi_call_int (cif, fn, rvalue, avalue, NULL); +} + +#ifdef FFI_GO_CLOSURES +void +ffi_call_go (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) +{ + ffi_call_int (cif, fn, rvalue, avalue, closure); +} +#endif + +/** private members **/ + +void FFI_HIDDEN ffi_closure_i386(void); +void FFI_HIDDEN ffi_closure_STDCALL(void); +void FFI_HIDDEN ffi_closure_REGISTER(void); + +struct closure_frame +{ + unsigned rettemp[4]; /* 0 */ + unsigned regs[3]; /* 16-24 */ + ffi_cif *cif; /* 28 */ + void (*fun)(ffi_cif*,void*,void**,void*); /* 32 */ + void *user_data; /* 36 */ +}; + +int FFI_HIDDEN FFI_DECLARE_FASTCALL +ffi_closure_inner (struct closure_frame *frame, char *stack) +{ + ffi_cif *cif = frame->cif; + int cabi, i, n, flags, dir, narg_reg; + const struct abi_params *pabi; + ffi_type **arg_types; + char *argp; + void *rvalue; + void **avalue; + + cabi = cif->abi; + flags = cif->flags; + narg_reg = 0; + rvalue = frame->rettemp; + pabi = &abi_params[cabi]; + dir = pabi->dir; + argp = (dir < 0 ? stack + STACK_ALIGN (cif->bytes) : stack); + + switch (flags) + { + case X86_RET_STRUCTARG: + if (pabi->nregs > 0) + { + rvalue = (void *)frame->regs[pabi->regs[0]]; + narg_reg = 1; + frame->rettemp[0] = (unsigned)rvalue; + break; + } + /* fallthru */ + case X86_RET_STRUCTPOP: + rvalue = *(void **)argp; + argp += sizeof(void *); + frame->rettemp[0] = (unsigned)rvalue; + break; + } + + n = cif->nargs; + avalue = alloca(sizeof(void *) * n); + + arg_types = cif->arg_types; + for (i = 0; i < n; ++i) + { + ffi_type *ty = arg_types[i]; + size_t z = ty->size; + int t = ty->type; + void *valp; + + if (z <= FFI_SIZEOF_ARG && t != FFI_TYPE_STRUCT) + { + if (t != FFI_TYPE_FLOAT && narg_reg < pabi->nregs) + valp = &frame->regs[pabi->regs[narg_reg++]]; + else if (dir < 0) + { + argp -= 4; + valp = argp; + } + else + { + valp = argp; + argp += 4; + } + } + else + { + size_t za = FFI_ALIGN (z, FFI_SIZEOF_ARG); + size_t align = FFI_SIZEOF_ARG; + + /* See the comment in ffi_call_int. */ + if (t == FFI_TYPE_STRUCT && ty->alignment >= 16) + align = 16; + + /* Issue 434: For thiscall and fastcall, if the paramter passed + as 64-bit integer or struct, all following integer parameters + will be passed on stack. */ + if ((cabi == FFI_THISCALL || cabi == FFI_FASTCALL) + && (t == FFI_TYPE_SINT64 + || t == FFI_TYPE_UINT64 + || t == FFI_TYPE_STRUCT)) + narg_reg = 2; + + if (dir < 0) + { + /* ??? These reverse argument ABIs are probably too old + to have cared about alignment. Someone should check. */ + argp -= za; + valp = argp; + } + else + { + argp = (char *)FFI_ALIGN (argp, align); + valp = argp; + argp += za; + } + } + + avalue[i] = valp; + } + + frame->fun (cif, rvalue, avalue, frame->user_data); + + if (cabi == FFI_STDCALL) + return flags + (cif->bytes << X86_RET_POP_SHIFT); + else + return flags; +} +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 ffi_status ffi_prep_closure_loc (ffi_closure* closure, @@ -736,6 +1132,7 @@ ffi_prep_closure_loc (ffi_closure* closure, void *user_data, void *codeloc) { +<<<<<<< HEAD #ifdef X86_WIN64 #define ISFLOAT(IDX) (cif->arg_types[IDX]->type == FFI_TYPE_FLOAT || cif->arg_types[IDX]->type == FFI_TYPE_DOUBLE) #define FLAG(IDX) (cif->nargs>(IDX)&&ISFLOAT(IDX)?(1<<(IDX)):0) @@ -780,38 +1177,109 @@ ffi_prep_closure_loc (ffi_closure* closure, } #ifdef X86_WIN32 else if (cif->abi == FFI_MS_CDECL) +======= + char *tramp = closure->tramp; + void (*dest)(void); + int op = 0xb8; /* movl imm, %eax */ + + switch (cif->abi) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 { - FFI_INIT_TRAMPOLINE (&closure->tramp[0], - &ffi_closure_SYSV, - (void*)codeloc); + case FFI_SYSV: + case FFI_THISCALL: + case FFI_FASTCALL: + case FFI_MS_CDECL: + dest = ffi_closure_i386; + break; + case FFI_STDCALL: + case FFI_PASCAL: + dest = ffi_closure_STDCALL; + break; + case FFI_REGISTER: + dest = ffi_closure_REGISTER; + op = 0x68; /* pushl imm */ + break; + default: + return FFI_BAD_ABI; } -#endif /* X86_WIN32 */ -#endif /* !X86_WIN64 */ - else + + /* endbr32. */ + *(UINT32 *) tramp = 0xfb1e0ff3; + + /* movl or pushl immediate. */ + tramp[4] = op; + *(void **)(tramp + 5) = codeloc; + + /* jmp dest */ + tramp[9] = 0xe9; + *(unsigned *)(tramp + 10) = (unsigned)dest - ((unsigned)codeloc + 14); + + closure->cif = cif; + closure->fun = fun; + closure->user_data = user_data; + + return FFI_OK; +} + +#ifdef FFI_GO_CLOSURES + +void FFI_HIDDEN ffi_go_closure_EAX(void); +void FFI_HIDDEN ffi_go_closure_ECX(void); +void FFI_HIDDEN ffi_go_closure_STDCALL(void); + +ffi_status +ffi_prep_go_closure (ffi_go_closure* closure, ffi_cif* cif, + void (*fun)(ffi_cif*,void*,void**,void*)) +{ + void (*dest)(void); + + switch (cif->abi) { + case FFI_SYSV: + case FFI_MS_CDECL: + dest = ffi_go_closure_ECX; + break; + case FFI_THISCALL: + case FFI_FASTCALL: + dest = ffi_go_closure_EAX; + break; + case FFI_STDCALL: + case FFI_PASCAL: + dest = ffi_go_closure_STDCALL; + break; + case FFI_REGISTER: + default: return FFI_BAD_ABI; } - - closure->cif = cif; - closure->user_data = user_data; - closure->fun = fun; + + closure->tramp = dest; + closure->cif = cif; + closure->fun = fun; return FFI_OK; } +#endif /* FFI_GO_CLOSURES */ + /* ------- Native raw API support -------------------------------- */ #if !FFI_NO_RAW_API +void FFI_HIDDEN ffi_closure_raw_SYSV(void); +void FFI_HIDDEN ffi_closure_raw_THISCALL(void); + ffi_status -ffi_prep_raw_closure_loc (ffi_raw_closure* closure, - ffi_cif* cif, +ffi_prep_raw_closure_loc (ffi_raw_closure *closure, + ffi_cif *cif, void (*fun)(ffi_cif*,void*,ffi_raw*,void*), void *user_data, void *codeloc) { + char *tramp = closure->tramp; + void (*dest)(void); int i; +<<<<<<< HEAD if (cif->abi != FFI_SYSV #ifdef X86_WIN32 && cif->abi != FFI_THISCALL @@ -820,12 +1288,23 @@ ffi_prep_raw_closure_loc (ffi_raw_closure* closure, return FFI_BAD_ABI; /* we currently don't support certain kinds of arguments for raw +======= + /* We currently don't support certain kinds of arguments for raw +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 closures. This should be implemented by a separate assembly language routine, since it would require argument processing, something we don't do now for performance. */ - for (i = cif->nargs-1; i >= 0; i--) + switch (cif->arg_types[i]->type) + { + case FFI_TYPE_STRUCT: + case FFI_TYPE_LONGDOUBLE: + return FFI_BAD_TYPEDEF; + } + + switch (cif->abi) { +<<<<<<< HEAD FFI_ASSERT (cif->arg_types[i]->type != FFI_TYPE_STRUCT); FFI_ASSERT (cif->arg_types[i]->type != FFI_TYPE_LONGDOUBLE); } @@ -841,15 +1320,34 @@ ffi_prep_raw_closure_loc (ffi_raw_closure* closure, else if (cif->abi == FFI_THISCALL) { FFI_INIT_TRAMPOLINE_RAW_THISCALL (&closure->tramp[0], &ffi_closure_raw_THISCALL, codeloc, cif->bytes); +======= + case FFI_THISCALL: + dest = ffi_closure_raw_THISCALL; + break; + case FFI_SYSV: + dest = ffi_closure_raw_SYSV; + break; + default: + return FFI_BAD_ABI; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } -#endif - closure->cif = cif; + + /* movl imm, %eax. */ + tramp[0] = 0xb8; + *(void **)(tramp + 1) = codeloc; + + /* jmp dest */ + tramp[5] = 0xe9; + *(unsigned *)(tramp + 6) = (unsigned)dest - ((unsigned)codeloc + 10); + + closure->cif = cif; + closure->fun = fun; closure->user_data = user_data; - closure->fun = fun; return FFI_OK; } +<<<<<<< HEAD static unsigned int ffi_prep_args_raw(char *stack, extended_cif *ecif) { @@ -890,30 +1388,58 @@ ffi_prep_args_raw(char *stack, extended_cif *ecif) * libffi-1.20, this is not the case.) */ +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 void -ffi_raw_call(ffi_cif *cif, void (*fn)(void), void *rvalue, ffi_raw *fake_avalue) +ffi_raw_call(ffi_cif *cif, void (*fn)(void), void *rvalue, ffi_raw *avalue) { - extended_cif ecif; - void **avalue = (void **)fake_avalue; - - ecif.cif = cif; - ecif.avalue = avalue; - - /* If the return value is a struct and we don't have a return */ - /* value address then we need to make one */ - - if (rvalue == NULL - && (cif->flags == FFI_TYPE_STRUCT - || cif->flags == FFI_TYPE_MS_STRUCT)) + size_t rsize, bytes; + struct call_frame *frame; + char *stack, *argp; + ffi_type **arg_types; + int flags, cabi, i, n, narg_reg; + const struct abi_params *pabi; + + flags = cif->flags; + cabi = cif->abi; + pabi = &abi_params[cabi]; + + rsize = 0; + if (rvalue == NULL) { - ecif.rvalue = alloca(cif->rtype->size); + switch (flags) + { + case X86_RET_FLOAT: + case X86_RET_DOUBLE: + case X86_RET_LDOUBLE: + case X86_RET_STRUCTPOP: + case X86_RET_STRUCTARG: + /* The float cases need to pop the 387 stack. + The struct cases need to pass a valid pointer to the callee. */ + rsize = cif->rtype->size; + break; + default: + /* We can pretend that the callee returns nothing. */ + flags = X86_RET_VOID; + break; + } } - else - ecif.rvalue = rvalue; - - - switch (cif->abi) + + bytes = STACK_ALIGN (cif->bytes); + argp = stack = + (void *)((uintptr_t)alloca(bytes + sizeof(*frame) + rsize + 15) & ~16); + frame = (struct call_frame *)(stack + bytes); + if (rsize) + rvalue = frame + 1; + + frame->fn = fn; + frame->flags = flags; + frame->rvalue = rvalue; + + narg_reg = 0; + switch (flags) { +<<<<<<< HEAD #ifndef X86_WIN32 case FFI_SYSV: ffi_call_SYSV(ffi_prep_args_raw, &ecif, cif->bytes, cif->flags, @@ -935,13 +1461,55 @@ ffi_raw_call(ffi_cif *cif, void (*fn)(void), void *rvalue, ffi_raw *fake_avalue) #endif default: FFI_ASSERT(0); +======= + case X86_RET_STRUCTARG: + /* The pointer is passed as the first argument. */ + if (pabi->nregs > 0) + { + frame->regs[pabi->regs[0]] = (unsigned)rvalue; + narg_reg = 1; + break; + } + /* fallthru */ + case X86_RET_STRUCTPOP: + *(void **)argp = rvalue; + argp += sizeof(void *); + bytes -= sizeof(void *); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 break; } -} - -#endif -#endif /* !__x86_64__ || X86_WIN64 */ + arg_types = cif->arg_types; + for (i = 0, n = cif->nargs; narg_reg < pabi->nregs && i < n; i++) + { + ffi_type *ty = arg_types[i]; + size_t z = ty->size; + int t = ty->type; + + if (z <= FFI_SIZEOF_ARG && t != FFI_TYPE_STRUCT && t != FFI_TYPE_FLOAT) + { + ffi_arg val = extend_basic_type (avalue, t); + frame->regs[pabi->regs[narg_reg++]] = val; + z = FFI_SIZEOF_ARG; + } + else + { + memcpy (argp, avalue, z); + z = FFI_ALIGN (z, FFI_SIZEOF_ARG); + argp += z; + } + avalue += z; + bytes -= z; + } + if (i < n) + memcpy (argp, avalue, bytes); +<<<<<<< HEAD +======= + ffi_call_i386 (frame, stack); +} +#endif /* !FFI_NO_RAW_API */ +#endif /* __i386__ */ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 diff --git a/native/libffi/src/x86/ffi64.c b/native/libffi/src/x86/ffi64.c index 5a5e043835..8f513e4a12 100644 --- a/native/libffi/src/x86/ffi64.c +++ b/native/libffi/src/x86/ffi64.c @@ -1,6 +1,6 @@ /* ----------------------------------------------------------------------- - ffi64.c - Copyright (c) 2013 The Written Word, Inc. - Copyright (c) 2011 Anthony Green + ffi64.c - Copyright (c) 2011, 2018 Anthony Green + Copyright (c) 2013 The Written Word, Inc. Copyright (c) 2008, 2010 Red Hat, Inc. Copyright (c) 2002, 2007 Bo Thorsen @@ -32,6 +32,8 @@ #include #include +#include +#include "internal64.h" #ifdef __x86_64__ @@ -62,10 +64,15 @@ struct register_args /* Registers for argument passing. */ UINT64 gpr[MAX_GPR_REGS]; union big_int_union sse[MAX_SSE_REGS]; +<<<<<<< HEAD +======= + UINT64 rax; /* ssecount */ + UINT64 r10; /* static chain */ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 }; extern void ffi_call_unix64 (void *args, unsigned long bytes, unsigned flags, - void *raddr, void (*fnaddr)(void), unsigned ssecount); + void *raddr, void (*fnaddr)(void)) FFI_HIDDEN; /* All reference to register classes here is identical to the code in gcc/config/i386/i386.c. Do *not* change one without the other. */ @@ -167,6 +174,7 @@ classify_argument (ffi_type *type, enum x86_64_reg_class classes[], case FFI_TYPE_UINT64: case FFI_TYPE_SINT64: case FFI_TYPE_POINTER: + do_integer: { size_t size = byte_offset + type->size; @@ -188,7 +196,7 @@ classify_argument (ffi_type *type, enum x86_64_reg_class classes[], } else if (size <= 16) { - classes[0] = classes[1] = X86_64_INTEGERSI_CLASS; + classes[0] = classes[1] = X86_64_INTEGER_CLASS; return 2; } else @@ -214,7 +222,11 @@ classify_argument (ffi_type *type, enum x86_64_reg_class classes[], const size_t UNITS_PER_WORD = 8; size_t words = (type->size + UNITS_PER_WORD - 1) / UNITS_PER_WORD; ffi_type **ptr; +<<<<<<< HEAD int i; +======= + unsigned int i; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 enum x86_64_reg_class subclasses[MAX_CLASSES]; /* If the struct is larger than 32 bytes, pass it on the stack. */ @@ -228,6 +240,7 @@ classify_argument (ffi_type *type, enum x86_64_reg_class classes[], signalize memory class, so handle it as special case. */ if (!words) { + case FFI_TYPE_VOID: classes[0] = X86_64_NO_CLASS; return 1; } @@ -237,7 +250,7 @@ classify_argument (ffi_type *type, enum x86_64_reg_class classes[], { size_t num; - byte_offset = ALIGN (byte_offset, (*ptr)->alignment); + byte_offset = FFI_ALIGN (byte_offset, (*ptr)->alignment); num = classify_argument (*ptr, subclasses, byte_offset % 8); if (num == 0) @@ -276,7 +289,7 @@ classify_argument (ffi_type *type, enum x86_64_reg_class classes[], /* The X86_64_SSEUP_CLASS should be always preceded by X86_64_SSE_CLASS or X86_64_SSEUP_CLASS. */ - if (classes[i] == X86_64_SSEUP_CLASS + if (i > 1 && classes[i] == X86_64_SSEUP_CLASS && classes[i - 1] != X86_64_SSE_CLASS && classes[i - 1] != X86_64_SSEUP_CLASS) { @@ -287,7 +300,7 @@ classify_argument (ffi_type *type, enum x86_64_reg_class classes[], /* If X86_64_X87UP_CLASS isn't preceded by X86_64_X87_CLASS, everything should be passed in memory. */ - if (classes[i] == X86_64_X87UP_CLASS + if (i > 1 && classes[i] == X86_64_X87UP_CLASS && (classes[i - 1] != X86_64_X87_CLASS)) { /* The first one should never be X86_64_X87UP_CLASS. */ @@ -297,11 +310,42 @@ classify_argument (ffi_type *type, enum x86_64_reg_class classes[], } return words; } - - default: - FFI_ASSERT(0); + case FFI_TYPE_COMPLEX: + { + ffi_type *inner = type->elements[0]; + switch (inner->type) + { + case FFI_TYPE_INT: + case FFI_TYPE_UINT8: + case FFI_TYPE_SINT8: + case FFI_TYPE_UINT16: + case FFI_TYPE_SINT16: + case FFI_TYPE_UINT32: + case FFI_TYPE_SINT32: + case FFI_TYPE_UINT64: + case FFI_TYPE_SINT64: + goto do_integer; + + case FFI_TYPE_FLOAT: + classes[0] = X86_64_SSE_CLASS; + if (byte_offset % 8) + { + classes[1] = X86_64_SSESF_CLASS; + return 2; + } + return 1; + case FFI_TYPE_DOUBLE: + classes[0] = classes[1] = X86_64_SSEDF_CLASS; + return 2; +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE + case FFI_TYPE_LONGDOUBLE: + classes[0] = X86_64_COMPLEX_X87_CLASS; + return 1; +#endif + } + } } - return 0; /* Never reached. */ + abort(); } /* Examine the argument and return set number of register required in each @@ -313,7 +357,12 @@ examine_argument (ffi_type *type, enum x86_64_reg_class classes[MAX_CLASSES], _Bool in_return, int *pngpr, int *pnsse) { size_t n; +<<<<<<< HEAD int i, ngpr, nsse; +======= + unsigned int i; + int ngpr, nsse; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 n = classify_argument (type, classes, 0); if (n == 0) @@ -351,18 +400,80 @@ examine_argument (ffi_type *type, enum x86_64_reg_class classes[MAX_CLASSES], /* Perform machine dependent cif processing. */ -ffi_status +#ifndef __ILP32__ +extern ffi_status +ffi_prep_cif_machdep_efi64(ffi_cif *cif); +#endif + +ffi_status FFI_HIDDEN ffi_prep_cif_machdep (ffi_cif *cif) { +<<<<<<< HEAD int gprcount, ssecount, i, avn, ngpr, nsse, flags; enum x86_64_reg_class classes[MAX_CLASSES]; size_t bytes, n; +======= + int gprcount, ssecount, i, avn, ngpr, nsse; + unsigned flags; + enum x86_64_reg_class classes[MAX_CLASSES]; + size_t bytes, n, rtype_size; + ffi_type *rtype; + +#ifndef __ILP32__ + if (cif->abi == FFI_EFI64 || cif->abi == FFI_GNUW64) + return ffi_prep_cif_machdep_efi64(cif); +#endif + if (cif->abi != FFI_UNIX64) + return FFI_BAD_ABI; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 gprcount = ssecount = 0; - flags = cif->rtype->type; - if (flags != FFI_TYPE_VOID) + rtype = cif->rtype; + rtype_size = rtype->size; + switch (rtype->type) { + case FFI_TYPE_VOID: + flags = UNIX64_RET_VOID; + break; + case FFI_TYPE_UINT8: + flags = UNIX64_RET_UINT8; + break; + case FFI_TYPE_SINT8: + flags = UNIX64_RET_SINT8; + break; + case FFI_TYPE_UINT16: + flags = UNIX64_RET_UINT16; + break; + case FFI_TYPE_SINT16: + flags = UNIX64_RET_SINT16; + break; + case FFI_TYPE_UINT32: + flags = UNIX64_RET_UINT32; + break; + case FFI_TYPE_INT: + case FFI_TYPE_SINT32: + flags = UNIX64_RET_SINT32; + break; + case FFI_TYPE_UINT64: + case FFI_TYPE_SINT64: + flags = UNIX64_RET_INT64; + break; + case FFI_TYPE_POINTER: + flags = (sizeof(void *) == 4 ? UNIX64_RET_UINT32 : UNIX64_RET_INT64); + break; + case FFI_TYPE_FLOAT: + flags = UNIX64_RET_XMM32; + break; + case FFI_TYPE_DOUBLE: + flags = UNIX64_RET_XMM64; + break; +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE + case FFI_TYPE_LONGDOUBLE: + flags = UNIX64_RET_X87; + break; +#endif + case FFI_TYPE_STRUCT: n = examine_argument (cif->rtype, classes, 1, &ngpr, &nsse); if (n == 0) { @@ -370,22 +481,62 @@ ffi_prep_cif_machdep (ffi_cif *cif) memory is the first argument. Allocate a register for it. */ gprcount++; /* We don't have to do anything in asm for the return. */ - flags = FFI_TYPE_VOID; + flags = UNIX64_RET_VOID | UNIX64_FLAG_RET_IN_MEM; } - else if (flags == FFI_TYPE_STRUCT) + else { - /* Mark which registers the result appears in. */ _Bool sse0 = SSE_CLASS_P (classes[0]); - _Bool sse1 = n == 2 && SSE_CLASS_P (classes[1]); - if (sse0 && !sse1) - flags |= 1 << 8; - else if (!sse0 && sse1) - flags |= 1 << 9; - else if (sse0 && sse1) - flags |= 1 << 10; - /* Mark the true size of the structure. */ - flags |= cif->rtype->size << 12; + + if (rtype_size == 4 && sse0) + flags = UNIX64_RET_XMM32; + else if (rtype_size == 8) + flags = sse0 ? UNIX64_RET_XMM64 : UNIX64_RET_INT64; + else + { + _Bool sse1 = n == 2 && SSE_CLASS_P (classes[1]); + if (sse0 && sse1) + flags = UNIX64_RET_ST_XMM0_XMM1; + else if (sse0) + flags = UNIX64_RET_ST_XMM0_RAX; + else if (sse1) + flags = UNIX64_RET_ST_RAX_XMM0; + else + flags = UNIX64_RET_ST_RAX_RDX; + flags |= rtype_size << UNIX64_SIZE_SHIFT; + } + } + break; + case FFI_TYPE_COMPLEX: + switch (rtype->elements[0]->type) + { + case FFI_TYPE_UINT8: + case FFI_TYPE_SINT8: + case FFI_TYPE_UINT16: + case FFI_TYPE_SINT16: + case FFI_TYPE_INT: + case FFI_TYPE_UINT32: + case FFI_TYPE_SINT32: + case FFI_TYPE_UINT64: + case FFI_TYPE_SINT64: + flags = UNIX64_RET_ST_RAX_RDX | ((unsigned) rtype_size << UNIX64_SIZE_SHIFT); + break; + case FFI_TYPE_FLOAT: + flags = UNIX64_RET_XMM64; + break; + case FFI_TYPE_DOUBLE: + flags = UNIX64_RET_ST_XMM0_XMM1 | (16 << UNIX64_SIZE_SHIFT); + break; +#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE + case FFI_TYPE_LONGDOUBLE: + flags = UNIX64_RET_X87_2; + break; +#endif + default: + return FFI_BAD_TYPEDEF; } + break; + default: + return FFI_BAD_TYPEDEF; } /* Go over all arguments and determine the way they should be passed. @@ -402,7 +553,7 @@ ffi_prep_cif_machdep (ffi_cif *cif) if (align < 8) align = 8; - bytes = ALIGN (bytes, align); + bytes = FFI_ALIGN (bytes, align); bytes += cif->arg_types[i]->size; } else @@ -412,44 +563,54 @@ ffi_prep_cif_machdep (ffi_cif *cif) } } if (ssecount) - flags |= 1 << 11; + flags |= UNIX64_FLAG_XMM_ARGS; + cif->flags = flags; +<<<<<<< HEAD cif->bytes = (unsigned)ALIGN (bytes, 8); +======= + cif->bytes = (unsigned) FFI_ALIGN (bytes, 8); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 return FFI_OK; } -void -ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +static void +ffi_call_int (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) { enum x86_64_reg_class classes[MAX_CLASSES]; char *stack, *argp; ffi_type **arg_types; - int gprcount, ssecount, ngpr, nsse, i, avn; - _Bool ret_in_memory; + int gprcount, ssecount, ngpr, nsse, i, avn, flags; struct register_args *reg_args; /* Can't call 32-bit mode from 64-bit mode. */ FFI_ASSERT (cif->abi == FFI_UNIX64); /* If the return value is a struct and we don't have a return value - address then we need to make one. Note the setting of flags to - VOID above in ffi_prep_cif_machdep. */ - ret_in_memory = (cif->rtype->type == FFI_TYPE_STRUCT - && (cif->flags & 0xff) == FFI_TYPE_VOID); - if (rvalue == NULL && ret_in_memory) - rvalue = alloca (cif->rtype->size); + address then we need to make one. Otherwise we can ignore it. */ + flags = cif->flags; + if (rvalue == NULL) + { + if (flags & UNIX64_FLAG_RET_IN_MEM) + rvalue = alloca (cif->rtype->size); + else + flags = UNIX64_RET_VOID; + } /* Allocate the space for the arguments, plus 4 words of temp space. */ stack = alloca (sizeof (struct register_args) + cif->bytes + 4*8); reg_args = (struct register_args *) stack; argp = stack + sizeof (struct register_args); + reg_args->r10 = (uintptr_t) closure; + gprcount = ssecount = 0; /* If the return value is passed in memory, add the pointer as the first integer argument. */ - if (ret_in_memory) + if (flags & UNIX64_FLAG_RET_IN_MEM) reg_args->gpr[gprcount++] = (unsigned long) rvalue; avn = cif->nargs; @@ -471,7 +632,7 @@ ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) align = 8; /* Pass this argument in memory. */ - argp = (void *) ALIGN (argp, align); + argp = (void *) FFI_ALIGN (argp, align); memcpy (argp, avalue[i], size); argp += size; } @@ -479,12 +640,15 @@ ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) { /* The argument is passed entirely in registers. */ char *a = (char *) avalue[i]; - int j; + unsigned int j; for (j = 0; j < n; j++, a += 8, size -= 8) { switch (classes[j]) { + case X86_64_NO_CLASS: + case X86_64_SSEUP_CLASS: + break; case X86_64_INTEGER_CLASS: case X86_64_INTEGERSI_CLASS: /* Sign-extend integer arguments passed in general @@ -494,26 +658,26 @@ ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) switch (arg_types[i]->type) { case FFI_TYPE_SINT8: - *(SINT64 *)®_args->gpr[gprcount] = (SINT64) *((SINT8 *) a); + reg_args->gpr[gprcount] = (SINT64) *((SINT8 *) a); break; case FFI_TYPE_SINT16: - *(SINT64 *)®_args->gpr[gprcount] = (SINT64) *((SINT16 *) a); + reg_args->gpr[gprcount] = (SINT64) *((SINT16 *) a); break; case FFI_TYPE_SINT32: - *(SINT64 *)®_args->gpr[gprcount] = (SINT64) *((SINT32 *) a); + reg_args->gpr[gprcount] = (SINT64) *((SINT32 *) a); break; default: reg_args->gpr[gprcount] = 0; - memcpy (®_args->gpr[gprcount], a, size < 8 ? size : 8); + memcpy (®_args->gpr[gprcount], a, size); } gprcount++; break; case X86_64_SSE_CLASS: case X86_64_SSEDF_CLASS: - reg_args->sse[ssecount++].i64 = *(UINT64 *) a; + memcpy (®_args->sse[ssecount++].i64, a, sizeof(UINT64)); break; case X86_64_SSESF_CLASS: - reg_args->sse[ssecount++].i32 = *(UINT32 *) a; + memcpy (®_args->sse[ssecount++].i32, a, sizeof(UINT32)); break; default: abort(); @@ -521,13 +685,65 @@ ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) } } } + reg_args->rax = ssecount; ffi_call_unix64 (stack, cif->bytes + sizeof (struct register_args), - cif->flags, rvalue, fn, ssecount); + flags, rvalue, fn); +} + +#ifndef __ILP32__ +extern void +ffi_call_efi64(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue); +#endif + +void +ffi_call (ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +{ +#ifndef __ILP32__ + if (cif->abi == FFI_EFI64 || cif->abi == FFI_GNUW64) + { + ffi_call_efi64(cif, fn, rvalue, avalue); + return; + } +#endif + ffi_call_int (cif, fn, rvalue, avalue, NULL); +} + +#ifdef FFI_GO_CLOSURES + +#ifndef __ILP32__ +extern void +ffi_call_go_efi64(ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure); +#endif + +void +ffi_call_go (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) +{ +#ifndef __ILP32__ + if (cif->abi == FFI_EFI64 || cif->abi == FFI_GNUW64) + { + ffi_call_go_efi64(cif, fn, rvalue, avalue, closure); + return; + } +#endif + ffi_call_int (cif, fn, rvalue, avalue, closure); } +#endif /* FFI_GO_CLOSURES */ -extern void ffi_closure_unix64(void); +extern void ffi_closure_unix64(void) FFI_HIDDEN; +extern void ffi_closure_unix64_sse(void) FFI_HIDDEN; + +#ifndef __ILP32__ +extern ffi_status +ffi_prep_closure_loc_efi64(ffi_closure* closure, + ffi_cif* cif, + void (*fun)(ffi_cif*, void*, void**, void*), + void *user_data, + void *codeloc); +#endif ffi_status ffi_prep_closure_loc (ffi_closure* closure, @@ -536,29 +752,33 @@ ffi_prep_closure_loc (ffi_closure* closure, void *user_data, void *codeloc) { - volatile unsigned short *tramp; - - /* Sanity check on the cif ABI. */ - { - int abi = cif->abi; - if (UNLIKELY (! (abi > FFI_FIRST_ABI && abi < FFI_LAST_ABI))) - return FFI_BAD_ABI; - } - - tramp = (volatile unsigned short *) &closure->tramp[0]; + static const unsigned char trampoline[24] = { + /* endbr64 */ + 0xf3, 0x0f, 0x1e, 0xfa, + /* leaq -0xb(%rip),%r10 # 0x0 */ + 0x4c, 0x8d, 0x15, 0xf5, 0xff, 0xff, 0xff, + /* jmpq *0x7(%rip) # 0x18 */ + 0xff, 0x25, 0x07, 0x00, 0x00, 0x00, + /* nopl 0(%rax) */ + 0x0f, 0x1f, 0x80, 0x00, 0x00, 0x00, 0x00 + }; + void (*dest)(void); + char *tramp = closure->tramp; - tramp[0] = 0xbb49; /* mov , %r11 */ - *((unsigned long long * volatile) &tramp[1]) - = (unsigned long) ffi_closure_unix64; - tramp[5] = 0xba49; /* mov , %r10 */ - *((unsigned long long * volatile) &tramp[6]) - = (unsigned long) codeloc; +#ifndef __ILP32__ + if (cif->abi == FFI_EFI64 || cif->abi == FFI_GNUW64) + return ffi_prep_closure_loc_efi64(closure, cif, fun, user_data, codeloc); +#endif + if (cif->abi != FFI_UNIX64) + return FFI_BAD_ABI; - /* Set the carry bit iff the function uses any sse registers. - This is clc or stc, together with the first byte of the jmp. */ - tramp[10] = cif->flags & (1 << 11) ? 0x49f9 : 0x49f8; + if (cif->flags & UNIX64_FLAG_XMM_ARGS) + dest = ffi_closure_unix64_sse; + else + dest = ffi_closure_unix64; - tramp[11] = 0xe3ff; /* jmp *%r11 */ + memcpy (tramp, trampoline, sizeof(trampoline)); + *(UINT64 *)(tramp + sizeof (trampoline)) = (uintptr_t)dest; closure->cif = cif; closure->fun = fun; @@ -567,24 +787,28 @@ ffi_prep_closure_loc (ffi_closure* closure, return FFI_OK; } -int -ffi_closure_unix64_inner(ffi_closure *closure, void *rvalue, - struct register_args *reg_args, char *argp) +int FFI_HIDDEN +ffi_closure_unix64_inner(ffi_cif *cif, + void (*fun)(ffi_cif*, void*, void**, void*), + void *user_data, + void *rvalue, + struct register_args *reg_args, + char *argp) { - ffi_cif *cif; void **avalue; ffi_type **arg_types; long i, avn; int gprcount, ssecount, ngpr, nsse; - int ret; + int flags; - cif = closure->cif; - avalue = alloca(cif->nargs * sizeof(void *)); + avn = cif->nargs; + flags = cif->flags; + avalue = alloca(avn * sizeof(void *)); gprcount = ssecount = 0; - ret = cif->rtype->type; - if (ret != FFI_TYPE_VOID) + if (flags & UNIX64_FLAG_RET_IN_MEM) { +<<<<<<< HEAD enum x86_64_reg_class classes[MAX_CLASSES]; size_t n = examine_argument (cif->rtype, classes, 1, &ngpr, &nsse); if (n == 0) @@ -605,11 +829,21 @@ ffi_closure_unix64_inner(ffi_closure *closure, void *rvalue, else if (sse0 && !sse1) ret |= 1 << 9; } +======= + /* On return, %rax will contain the address that was passed + by the caller in %rdi. */ + void *r = (void *)(uintptr_t)reg_args->gpr[gprcount++]; + *(void **)rvalue = r; + rvalue = r; + flags = (sizeof(void *) == 4 ? UNIX64_RET_UINT32 : UNIX64_RET_INT64); +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } - avn = cif->nargs; arg_types = cif->arg_types; +<<<<<<< HEAD +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 for (i = 0; i < avn; ++i) { enum x86_64_reg_class classes[MAX_CLASSES]; @@ -627,7 +861,7 @@ ffi_closure_unix64_inner(ffi_closure *closure, void *rvalue, align = 8; /* Pass this argument in memory. */ - argp = (void *) ALIGN (argp, align); + argp = (void *) FFI_ALIGN (argp, align); avalue[i] = argp; argp += arg_types[i]->size; } @@ -653,7 +887,7 @@ ffi_closure_unix64_inner(ffi_closure *closure, void *rvalue, else { char *a = alloca (16); - int j; + unsigned int j; avalue[i] = a; for (j = 0; j < n; j++, a += 8) @@ -667,10 +901,43 @@ ffi_closure_unix64_inner(ffi_closure *closure, void *rvalue, } /* Invoke the closure. */ - closure->fun (cif, rvalue, avalue, closure->user_data); + fun (cif, rvalue, avalue, user_data); /* Tell assembly how to perform return type promotions. */ - return ret; + return flags; +} + +#ifdef FFI_GO_CLOSURES + +extern void ffi_go_closure_unix64(void) FFI_HIDDEN; +extern void ffi_go_closure_unix64_sse(void) FFI_HIDDEN; + +#ifndef __ILP32__ +extern ffi_status +ffi_prep_go_closure_efi64(ffi_go_closure* closure, ffi_cif* cif, + void (*fun)(ffi_cif*, void*, void**, void*)); +#endif + +ffi_status +ffi_prep_go_closure (ffi_go_closure* closure, ffi_cif* cif, + void (*fun)(ffi_cif*, void*, void**, void*)) +{ +#ifndef __ILP32__ + if (cif->abi == FFI_EFI64 || cif->abi == FFI_GNUW64) + return ffi_prep_go_closure_efi64(closure, cif, fun); +#endif + if (cif->abi != FFI_UNIX64) + return FFI_BAD_ABI; + + closure->tramp = (cif->flags & UNIX64_FLAG_XMM_ARGS + ? ffi_go_closure_unix64_sse + : ffi_go_closure_unix64); + closure->cif = cif; + closure->fun = fun; + + return FFI_OK; } +#endif /* FFI_GO_CLOSURES */ + #endif /* __x86_64__ */ diff --git a/native/libffi/src/x86/ffitarget.h b/native/libffi/src/x86/ffitarget.h index ff0f71857d..9f71e1dde7 100644 --- a/native/libffi/src/x86/ffitarget.h +++ b/native/libffi/src/x86/ffitarget.h @@ -1,5 +1,9 @@ /* -----------------------------------------------------------------*-C-*- +<<<<<<< HEAD ffitarget.h - Copyright (c) 2012, 2014 Anthony Green +======= + ffitarget.h - Copyright (c) 2012, 2014, 2018 Anthony Green +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 Copyright (c) 1996-2003, 2010 Red Hat, Inc. Copyright (C) 2008 Free Software Foundation, Inc. @@ -78,7 +82,9 @@ typedef signed long ffi_sarg; #endif typedef enum ffi_abi { +#if defined(X86_WIN64) FFI_FIRST_ABI = 0, +<<<<<<< HEAD /* ---- Intel x86 Win32 ---------- */ #ifdef X86_WIN32 @@ -89,19 +95,39 @@ typedef enum ffi_abi { FFI_MS_CDECL, FFI_PASCAL, FFI_REGISTER, +======= + FFI_WIN64, /* sizeof(long double) == 8 - microsoft compilers */ + FFI_GNUW64, /* sizeof(long double) == 16 - GNU compilers */ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 FFI_LAST_ABI, -#ifdef _MSC_VER - FFI_DEFAULT_ABI = FFI_MS_CDECL -#else - FFI_DEFAULT_ABI = FFI_SYSV -#endif +#ifdef __GNUC__ + FFI_DEFAULT_ABI = FFI_GNUW64 +#else + FFI_DEFAULT_ABI = FFI_WIN64 +#endif -#elif defined(X86_WIN64) +#elif defined(X86_64) || (defined (__x86_64__) && defined (X86_DARWIN)) + FFI_FIRST_ABI = 1, + FFI_UNIX64, FFI_WIN64, + FFI_EFI64 = FFI_WIN64, + FFI_GNUW64, FFI_LAST_ABI, - FFI_DEFAULT_ABI = FFI_WIN64 + FFI_DEFAULT_ABI = FFI_UNIX64 +#elif defined(X86_WIN32) + FFI_FIRST_ABI = 0, + FFI_SYSV = 1, + FFI_STDCALL = 2, + FFI_THISCALL = 3, + FFI_FASTCALL = 4, + FFI_MS_CDECL = 5, + FFI_PASCAL = 6, + FFI_REGISTER = 7, + FFI_LAST_ABI, + FFI_DEFAULT_ABI = FFI_MS_CDECL #else +<<<<<<< HEAD /* ---- Intel x86 and AMD x86-64 - */ FFI_SYSV, FFI_UNIX64, /* Unix variants all use the same ABI for x86-64 */ @@ -110,12 +136,18 @@ typedef enum ffi_abi { FFI_STDCALL, FFI_PASCAL, FFI_REGISTER, +======= + FFI_FIRST_ABI = 0, + FFI_SYSV = 1, + FFI_THISCALL = 3, + FFI_FASTCALL = 4, + FFI_STDCALL = 5, + FFI_PASCAL = 6, + FFI_REGISTER = 7, + FFI_MS_CDECL = 8, +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 FFI_LAST_ABI, -#if defined(__i386__) || defined(__i386) FFI_DEFAULT_ABI = FFI_SYSV -#else - FFI_DEFAULT_ABI = FFI_UNIX64 -#endif #endif } ffi_abi; #endif @@ -123,18 +155,21 @@ typedef enum ffi_abi { /* ---- Definitions for closures ----------------------------------------- */ #define FFI_CLOSURES 1 +#define FFI_GO_CLOSURES 1 + #define FFI_TYPE_SMALL_STRUCT_1B (FFI_TYPE_LAST + 1) #define FFI_TYPE_SMALL_STRUCT_2B (FFI_TYPE_LAST + 2) #define FFI_TYPE_SMALL_STRUCT_4B (FFI_TYPE_LAST + 3) #define FFI_TYPE_MS_STRUCT (FFI_TYPE_LAST + 4) -#if defined (X86_64) || (defined (__x86_64__) && defined (X86_DARWIN)) -#define FFI_TRAMPOLINE_SIZE 24 -#define FFI_NATIVE_RAW_API 0 -#else -#ifdef X86_WIN32 -#define FFI_TRAMPOLINE_SIZE 52 +#if defined (X86_64) || defined(X86_WIN64) \ + || (defined (__x86_64__) && defined (X86_DARWIN)) +/* 4 bytes of ENDBR64 + 7 bytes of LEA + 6 bytes of JMP + 7 bytes of NOP + + 8 bytes of pointer. */ +# define FFI_TRAMPOLINE_SIZE 32 +# define FFI_NATIVE_RAW_API 0 #else +<<<<<<< HEAD #ifdef X86_WIN64 #define FFI_TRAMPOLINE_SIZE 29 #define FFI_NATIVE_RAW_API 0 @@ -145,7 +180,21 @@ typedef enum ffi_abi { #endif #ifndef X86_WIN64 #define FFI_NATIVE_RAW_API 1 /* x86 has native raw api support */ -#endif +======= +/* 4 bytes of ENDBR32 + 5 bytes of MOV + 5 bytes of JMP + 2 unused + bytes. */ +# define FFI_TRAMPOLINE_SIZE 16 +# define FFI_NATIVE_RAW_API 1 /* x86 has native raw api support */ +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 +#endif + +#if !defined(GENERATE_LIBFFI_MAP) && defined(__ASSEMBLER__) \ + && defined(__CET__) +# include +# define _CET_NOTRACK notrack +#else +# define _CET_ENDBR +# define _CET_NOTRACK #endif #endif diff --git a/native/libffi/src/x86/ffiw64.c b/native/libffi/src/x86/ffiw64.c new file mode 100644 index 0000000000..a43a9ebe94 --- /dev/null +++ b/native/libffi/src/x86/ffiw64.c @@ -0,0 +1,318 @@ +/* ----------------------------------------------------------------------- + ffiw64.c - Copyright (c) 2018 Anthony Green + Copyright (c) 2014 Red Hat, Inc. + + x86 win64 Foreign Function Interface + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + ``Software''), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice shall be included + in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + ----------------------------------------------------------------------- */ + +#if defined(__x86_64__) || defined(_M_AMD64) +#include +#include +#include +#include + +#ifdef X86_WIN64 +#define EFI64(name) name +#else +#define EFI64(name) FFI_HIDDEN name##_efi64 +#endif + +struct win64_call_frame +{ + UINT64 rbp; /* 0 */ + UINT64 retaddr; /* 8 */ + UINT64 fn; /* 16 */ + UINT64 flags; /* 24 */ + UINT64 rvalue; /* 32 */ +}; + +extern void ffi_call_win64 (void *stack, struct win64_call_frame *, + void *closure) FFI_HIDDEN; + +ffi_status FFI_HIDDEN +EFI64(ffi_prep_cif_machdep)(ffi_cif *cif) +{ + int flags, n; + + switch (cif->abi) + { + case FFI_WIN64: + case FFI_GNUW64: + break; + default: + return FFI_BAD_ABI; + } + + flags = cif->rtype->type; + switch (flags) + { + default: + break; + case FFI_TYPE_LONGDOUBLE: + /* GCC returns long double values by reference, like a struct */ + if (cif->abi == FFI_GNUW64) + flags = FFI_TYPE_STRUCT; + break; + case FFI_TYPE_COMPLEX: + flags = FFI_TYPE_STRUCT; + /* FALLTHRU */ + case FFI_TYPE_STRUCT: + switch (cif->rtype->size) + { + case 8: + flags = FFI_TYPE_UINT64; + break; + case 4: + flags = FFI_TYPE_SMALL_STRUCT_4B; + break; + case 2: + flags = FFI_TYPE_SMALL_STRUCT_2B; + break; + case 1: + flags = FFI_TYPE_SMALL_STRUCT_1B; + break; + } + break; + } + cif->flags = flags; + + /* Each argument either fits in a register, an 8 byte slot, or is + passed by reference with the pointer in the 8 byte slot. */ + n = cif->nargs; + n += (flags == FFI_TYPE_STRUCT); + if (n < 4) + n = 4; + cif->bytes = n * 8; + + return FFI_OK; +} + +static void +ffi_call_int (ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) +{ + int i, j, n, flags; + UINT64 *stack; + size_t rsize; + struct win64_call_frame *frame; + + FFI_ASSERT(cif->abi == FFI_GNUW64 || cif->abi == FFI_WIN64); + + flags = cif->flags; + rsize = 0; + + /* If we have no return value for a structure, we need to create one. + Otherwise we can ignore the return type entirely. */ + if (rvalue == NULL) + { + if (flags == FFI_TYPE_STRUCT) + rsize = cif->rtype->size; + else + flags = FFI_TYPE_VOID; + } + + stack = alloca(cif->bytes + sizeof(struct win64_call_frame) + rsize); + frame = (struct win64_call_frame *)((char *)stack + cif->bytes); + if (rsize) + rvalue = frame + 1; + + frame->fn = (uintptr_t)fn; + frame->flags = flags; + frame->rvalue = (uintptr_t)rvalue; + + j = 0; + if (flags == FFI_TYPE_STRUCT) + { + stack[0] = (uintptr_t)rvalue; + j = 1; + } + + for (i = 0, n = cif->nargs; i < n; ++i, ++j) + { + switch (cif->arg_types[i]->size) + { + case 8: + stack[j] = *(UINT64 *)avalue[i]; + break; + case 4: + stack[j] = *(UINT32 *)avalue[i]; + break; + case 2: + stack[j] = *(UINT16 *)avalue[i]; + break; + case 1: + stack[j] = *(UINT8 *)avalue[i]; + break; + default: + stack[j] = (uintptr_t)avalue[i]; + break; + } + } + + ffi_call_win64 (stack, frame, closure); +} + +void +EFI64(ffi_call)(ffi_cif *cif, void (*fn)(void), void *rvalue, void **avalue) +{ + ffi_call_int (cif, fn, rvalue, avalue, NULL); +} + +void +EFI64(ffi_call_go)(ffi_cif *cif, void (*fn)(void), void *rvalue, + void **avalue, void *closure) +{ + ffi_call_int (cif, fn, rvalue, avalue, closure); +} + + +extern void ffi_closure_win64(void) FFI_HIDDEN; + +#ifdef FFI_GO_CLOSURES +extern void ffi_go_closure_win64(void) FFI_HIDDEN; +#endif + +ffi_status +EFI64(ffi_prep_closure_loc)(ffi_closure* closure, + ffi_cif* cif, + void (*fun)(ffi_cif*, void*, void**, void*), + void *user_data, + void *codeloc) +{ + static const unsigned char trampoline[FFI_TRAMPOLINE_SIZE - 8] = { + /* endbr64 */ + 0xf3, 0x0f, 0x1e, 0xfa, + /* leaq -0xb(%rip),%r10 # 0x0 */ + 0x4c, 0x8d, 0x15, 0xf5, 0xff, 0xff, 0xff, + /* jmpq *0x7(%rip) # 0x18 */ + 0xff, 0x25, 0x07, 0x00, 0x00, 0x00, + /* nopl 0(%rax) */ + 0x0f, 0x1f, 0x80, 0x00, 0x00, 0x00, 0x00 + }; + char *tramp = closure->tramp; + + switch (cif->abi) + { + case FFI_WIN64: + case FFI_GNUW64: + break; + default: + return FFI_BAD_ABI; + } + + memcpy (tramp, trampoline, sizeof(trampoline)); + *(UINT64 *)(tramp + sizeof (trampoline)) = (uintptr_t)ffi_closure_win64; + + closure->cif = cif; + closure->fun = fun; + closure->user_data = user_data; + + return FFI_OK; +} + +#ifdef FFI_GO_CLOSURES +ffi_status +EFI64(ffi_prep_go_closure)(ffi_go_closure* closure, ffi_cif* cif, + void (*fun)(ffi_cif*, void*, void**, void*)) +{ + switch (cif->abi) + { + case FFI_WIN64: + case FFI_GNUW64: + break; + default: + return FFI_BAD_ABI; + } + + closure->tramp = ffi_go_closure_win64; + closure->cif = cif; + closure->fun = fun; + + return FFI_OK; +} +#endif + +struct win64_closure_frame +{ + UINT64 rvalue[2]; + UINT64 fargs[4]; + UINT64 retaddr; + UINT64 args[]; +}; + +/* Force the inner function to use the MS ABI. When compiling on win64 + this is a nop. When compiling on unix, this simplifies the assembly, + and places the burden of saving the extra call-saved registers on + the compiler. */ +int FFI_HIDDEN __attribute__((ms_abi)) +ffi_closure_win64_inner(ffi_cif *cif, + void (*fun)(ffi_cif*, void*, void**, void*), + void *user_data, + struct win64_closure_frame *frame) +{ + void **avalue; + void *rvalue; + int i, n, nreg, flags; + + avalue = alloca(cif->nargs * sizeof(void *)); + rvalue = frame->rvalue; + nreg = 0; + + /* When returning a structure, the address is in the first argument. + We must also be prepared to return the same address in eax, so + install that address in the frame and pretend we return a pointer. */ + flags = cif->flags; + if (flags == FFI_TYPE_STRUCT) + { + rvalue = (void *)(uintptr_t)frame->args[0]; + frame->rvalue[0] = frame->args[0]; + nreg = 1; + } + + for (i = 0, n = cif->nargs; i < n; ++i, ++nreg) + { + size_t size = cif->arg_types[i]->size; + size_t type = cif->arg_types[i]->type; + void *a; + + if (type == FFI_TYPE_DOUBLE || type == FFI_TYPE_FLOAT) + { + if (nreg < 4) + a = &frame->fargs[nreg]; + else + a = &frame->args[nreg]; + } + else if (size == 1 || size == 2 || size == 4 || size == 8) + a = &frame->args[nreg]; + else + a = (void *)(uintptr_t)frame->args[nreg]; + + avalue[i] = a; + } + + /* Invoke the closure. */ + fun (cif, rvalue, avalue, user_data); + return flags; +} + +#endif /* __x86_64__ */ diff --git a/native/libffi/src/x86/internal.h b/native/libffi/src/x86/internal.h new file mode 100644 index 0000000000..09771ba8cf --- /dev/null +++ b/native/libffi/src/x86/internal.h @@ -0,0 +1,29 @@ +#define X86_RET_FLOAT 0 +#define X86_RET_DOUBLE 1 +#define X86_RET_LDOUBLE 2 +#define X86_RET_SINT8 3 +#define X86_RET_SINT16 4 +#define X86_RET_UINT8 5 +#define X86_RET_UINT16 6 +#define X86_RET_INT64 7 +#define X86_RET_INT32 8 +#define X86_RET_VOID 9 +#define X86_RET_STRUCTPOP 10 +#define X86_RET_STRUCTARG 11 +#define X86_RET_STRUCT_1B 12 +#define X86_RET_STRUCT_2B 13 +#define X86_RET_UNUSED14 14 +#define X86_RET_UNUSED15 15 + +#define X86_RET_TYPE_MASK 15 +#define X86_RET_POP_SHIFT 4 + +#define R_EAX 0 +#define R_EDX 1 +#define R_ECX 2 + +#ifdef __PCC__ +# define HAVE_FASTCALL 0 +#else +# define HAVE_FASTCALL 1 +#endif diff --git a/native/libffi/src/x86/internal64.h b/native/libffi/src/x86/internal64.h new file mode 100644 index 0000000000..512e95523e --- /dev/null +++ b/native/libffi/src/x86/internal64.h @@ -0,0 +1,22 @@ +#define UNIX64_RET_VOID 0 +#define UNIX64_RET_UINT8 1 +#define UNIX64_RET_UINT16 2 +#define UNIX64_RET_UINT32 3 +#define UNIX64_RET_SINT8 4 +#define UNIX64_RET_SINT16 5 +#define UNIX64_RET_SINT32 6 +#define UNIX64_RET_INT64 7 +#define UNIX64_RET_XMM32 8 +#define UNIX64_RET_XMM64 9 +#define UNIX64_RET_X87 10 +#define UNIX64_RET_X87_2 11 +#define UNIX64_RET_ST_XMM0_RAX 12 +#define UNIX64_RET_ST_RAX_XMM0 13 +#define UNIX64_RET_ST_XMM0_XMM1 14 +#define UNIX64_RET_ST_RAX_RDX 15 + +#define UNIX64_RET_LAST 15 + +#define UNIX64_FLAG_RET_IN_MEM (1 << 10) +#define UNIX64_FLAG_XMM_ARGS (1 << 11) +#define UNIX64_SIZE_SHIFT 12 diff --git a/native/libffi/src/x86/sysv.S b/native/libffi/src/x86/sysv.S index 3bd5477e47..d8ab4b0264 100644 --- a/native/libffi/src/x86/sysv.S +++ b/native/libffi/src/x86/sysv.S @@ -1,6 +1,7 @@ /* ----------------------------------------------------------------------- - sysv.S - Copyright (c) 2013 The Written Word, Inc. - - Copyright (c) 1996,1998,2001-2003,2005,2008,2010 Red Hat, Inc. + sysv.S - Copyright (c) 2017 Anthony Green + - Copyright (c) 2013 The Written Word, Inc. + - Copyright (c) 1996,1998,2001-2003,2005,2008,2010 Red Hat, Inc. X86 Foreign Function Interface @@ -25,458 +26,1112 @@ DEALINGS IN THE SOFTWARE. ----------------------------------------------------------------------- */ -#ifndef __x86_64__ +#ifdef __i386__ +#ifndef _MSC_VER #define LIBFFI_ASM #include #include +#include "internal.h" -.text - -.globl ffi_prep_args - - .align 4 -.globl ffi_call_SYSV - .type ffi_call_SYSV,@function - -ffi_call_SYSV: -.LFB1: - pushl %ebp -.LCFI0: - movl %esp,%ebp -.LCFI1: - /* Make room for all of the new args. */ - movl 16(%ebp),%ecx - subl %ecx,%esp - - /* Align the stack pointer to 16-bytes */ - andl $0xfffffff0, %esp - - movl %esp,%eax - - /* Place all of the ffi_prep_args in position */ - pushl 12(%ebp) - pushl %eax - call *8(%ebp) - - /* Return stack to previous state and call the function */ - addl $8,%esp - - call *28(%ebp) - - /* Load %ecx with the return type code */ - movl 20(%ebp),%ecx - - /* Protect %esi. We're going to pop it in the epilogue. */ - pushl %esi - - /* If the return value pointer is NULL, assume no return value. */ - cmpl $0,24(%ebp) - jne 0f - - /* Even if there is no space for the return value, we are - obliged to handle floating-point values. */ - cmpl $FFI_TYPE_FLOAT,%ecx - jne noretval - fstp %st(0) - - jmp epilogue - -0: - call 1f - -.Lstore_table: - .long noretval-.Lstore_table /* FFI_TYPE_VOID */ - .long retint-.Lstore_table /* FFI_TYPE_INT */ - .long retfloat-.Lstore_table /* FFI_TYPE_FLOAT */ - .long retdouble-.Lstore_table /* FFI_TYPE_DOUBLE */ - .long retlongdouble-.Lstore_table /* FFI_TYPE_LONGDOUBLE */ - .long retuint8-.Lstore_table /* FFI_TYPE_UINT8 */ - .long retsint8-.Lstore_table /* FFI_TYPE_SINT8 */ - .long retuint16-.Lstore_table /* FFI_TYPE_UINT16 */ - .long retsint16-.Lstore_table /* FFI_TYPE_SINT16 */ - .long retint-.Lstore_table /* FFI_TYPE_UINT32 */ - .long retint-.Lstore_table /* FFI_TYPE_SINT32 */ - .long retint64-.Lstore_table /* FFI_TYPE_UINT64 */ - .long retint64-.Lstore_table /* FFI_TYPE_SINT64 */ - .long retstruct-.Lstore_table /* FFI_TYPE_STRUCT */ - .long retint-.Lstore_table /* FFI_TYPE_POINTER */ - -1: - pop %esi - add (%esi, %ecx, 4), %esi - jmp *%esi - - /* Sign/zero extend as appropriate. */ -retsint8: - movsbl %al, %eax - jmp retint +#define C2(X, Y) X ## Y +#define C1(X, Y) C2(X, Y) +#ifdef __USER_LABEL_PREFIX__ +# define C(X) C1(__USER_LABEL_PREFIX__, X) +#else +# define C(X) X +#endif -retsint16: - movswl %ax, %eax - jmp retint +#ifdef X86_DARWIN +# define L(X) C1(L, X) +#else +# define L(X) C1(.L, X) +#endif -retuint8: - movzbl %al, %eax - jmp retint +#ifdef __ELF__ +# define ENDF(X) .type X,@function; .size X, . - X +#else +# define ENDF(X) +#endif -retuint16: - movzwl %ax, %eax - jmp retint - -retfloat: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - fstps (%ecx) - jmp epilogue - -retdouble: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - fstpl (%ecx) - jmp epilogue - -retlongdouble: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - fstpt (%ecx) - jmp epilogue - -retint64: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - movl %eax,0(%ecx) - movl %edx,4(%ecx) - jmp epilogue - -retint: - /* Load %ecx with the pointer to storage for the return value */ - movl 24(%ebp),%ecx - movl %eax,0(%ecx) - -retstruct: - /* Nothing to do! */ - -noretval: -epilogue: - popl %esi - movl %ebp,%esp - popl %ebp - ret -.LFE1: -.ffi_call_SYSV_end: - .size ffi_call_SYSV,.ffi_call_SYSV_end-ffi_call_SYSV - - .align 4 -FFI_HIDDEN (ffi_closure_SYSV) -.globl ffi_closure_SYSV - .type ffi_closure_SYSV, @function - -ffi_closure_SYSV: -.LFB2: - pushl %ebp -.LCFI2: - movl %esp, %ebp -.LCFI3: - subl $40, %esp - leal -24(%ebp), %edx - movl %edx, -12(%ebp) /* resp */ - leal 8(%ebp), %edx -#ifdef __SUNPRO_C - /* The SUNPRO compiler doesn't support GCC's regparm function - attribute, so we have to pass all three arguments to - ffi_closure_SYSV_inner on the stack. */ - movl %edx, 8(%esp) /* args = __builtin_dwarf_cfa () */ - leal -12(%ebp), %edx - movl %edx, 4(%esp) /* &resp */ - movl %eax, (%esp) /* closure */ +/* Handle win32 fastcall name mangling. */ +#ifdef X86_WIN32 +# define ffi_call_i386 "@ffi_call_i386@8" +# define ffi_closure_inner "@ffi_closure_inner@8" #else - movl %edx, 4(%esp) /* args = __builtin_dwarf_cfa () */ - leal -12(%ebp), %edx - movl %edx, (%esp) /* &resp */ +# define ffi_call_i386 C(ffi_call_i386) +# define ffi_closure_inner C(ffi_closure_inner) #endif -#if defined HAVE_HIDDEN_VISIBILITY_ATTRIBUTE || !defined __PIC__ - call ffi_closure_SYSV_inner + +/* This macro allows the safe creation of jump tables without an + actual table. The entry points into the table are all 8 bytes. + The use of ORG asserts that we're at the correct location. */ +/* ??? The clang assembler doesn't handle .org with symbolic expressions. */ +#if defined(__clang__) || defined(__APPLE__) || (defined (__sun__) && defined(__svr4__)) +# define E(BASE, X) .balign 8 #else - movl %ebx, 8(%esp) -.LCFI7: - call 1f -1: popl %ebx - addl $_GLOBAL_OFFSET_TABLE_+[.-1b], %ebx - call ffi_closure_SYSV_inner@PLT - movl 8(%esp), %ebx +# define E(BASE, X) .balign 8; .org BASE + X * 8 #endif - movl -12(%ebp), %ecx - cmpl $FFI_TYPE_INT, %eax - je .Lcls_retint - - /* Handle FFI_TYPE_UINT8, FFI_TYPE_SINT8, FFI_TYPE_UINT16, - FFI_TYPE_SINT16, FFI_TYPE_UINT32, FFI_TYPE_SINT32. */ - cmpl $FFI_TYPE_UINT64, %eax - jge 0f - cmpl $FFI_TYPE_UINT8, %eax - jge .Lcls_retint - -0: cmpl $FFI_TYPE_FLOAT, %eax - je .Lcls_retfloat - cmpl $FFI_TYPE_DOUBLE, %eax - je .Lcls_retdouble - cmpl $FFI_TYPE_LONGDOUBLE, %eax - je .Lcls_retldouble - cmpl $FFI_TYPE_SINT64, %eax - je .Lcls_retllong - cmpl $FFI_TYPE_STRUCT, %eax - je .Lcls_retstruct -.Lcls_epilogue: - movl %ebp, %esp - popl %ebp - ret -.Lcls_retint: - movl (%ecx), %eax - jmp .Lcls_epilogue -.Lcls_retfloat: - flds (%ecx) - jmp .Lcls_epilogue -.Lcls_retdouble: - fldl (%ecx) - jmp .Lcls_epilogue -.Lcls_retldouble: - fldt (%ecx) - jmp .Lcls_epilogue -.Lcls_retllong: - movl (%ecx), %eax - movl 4(%ecx), %edx - jmp .Lcls_epilogue -.Lcls_retstruct: - movl %ebp, %esp - popl %ebp - ret $4 -.LFE2: - .size ffi_closure_SYSV, .-ffi_closure_SYSV -#if !FFI_NO_RAW_API + .text + .balign 16 + .globl ffi_call_i386 + FFI_HIDDEN(ffi_call_i386) + +/* This is declared as -/* Precalculate for e.g. the Solaris 10/x86 assembler. */ -#if FFI_TRAMPOLINE_SIZE == 10 -#define RAW_CLOSURE_CIF_OFFSET 12 -#define RAW_CLOSURE_FUN_OFFSET 16 -#define RAW_CLOSURE_USER_DATA_OFFSET 20 -#elif FFI_TRAMPOLINE_SIZE == 24 -#define RAW_CLOSURE_CIF_OFFSET 24 -#define RAW_CLOSURE_FUN_OFFSET 28 -#define RAW_CLOSURE_USER_DATA_OFFSET 32 + void ffi_call_i386(struct call_frame *frame, char *argp) + __attribute__((fastcall)); + + Thus the arguments are present in + + ecx: frame + edx: argp +*/ + +ffi_call_i386: +L(UW0): + # cfi_startproc + _CET_ENDBR +#if !HAVE_FASTCALL + movl 4(%esp), %ecx + movl 8(%esp), %edx +#endif + movl (%esp), %eax /* move the return address */ + movl %ebp, (%ecx) /* store %ebp into local frame */ + movl %eax, 4(%ecx) /* store retaddr into local frame */ + + /* New stack frame based off ebp. This is a itty bit of unwind + trickery in that the CFA *has* changed. There is no easy way + to describe it correctly on entry to the function. Fortunately, + it doesn't matter too much since at all points we can correctly + unwind back to ffi_call. Note that the location to which we + moved the return address is (the new) CFA-4, so from the + perspective of the unwind info, it hasn't moved. */ + movl %ecx, %ebp +L(UW1): + # cfi_def_cfa(%ebp, 8) + # cfi_rel_offset(%ebp, 0) + + movl %edx, %esp /* set outgoing argument stack */ + movl 20+R_EAX*4(%ebp), %eax /* set register arguments */ + movl 20+R_EDX*4(%ebp), %edx + movl 20+R_ECX*4(%ebp), %ecx + + call *8(%ebp) + + movl 12(%ebp), %ecx /* load return type code */ + movl %ebx, 8(%ebp) /* preserve %ebx */ +L(UW2): + # cfi_rel_offset(%ebx, 8) + + andl $X86_RET_TYPE_MASK, %ecx +#ifdef __PIC__ + call C(__x86.get_pc_thunk.bx) +L(pc1): + leal L(store_table)-L(pc1)(%ebx, %ecx, 8), %ebx #else -#define RAW_CLOSURE_CIF_OFFSET ((FFI_TRAMPOLINE_SIZE + 3) & ~3) -#define RAW_CLOSURE_FUN_OFFSET (RAW_CLOSURE_CIF_OFFSET + 4) -#define RAW_CLOSURE_USER_DATA_OFFSET (RAW_CLOSURE_FUN_OFFSET + 4) + leal L(store_table)(,%ecx, 8), %ebx #endif -#define CIF_FLAGS_OFFSET 20 - - .align 4 -FFI_HIDDEN (ffi_closure_raw_SYSV) -.globl ffi_closure_raw_SYSV - .type ffi_closure_raw_SYSV, @function - -ffi_closure_raw_SYSV: -.LFB3: - pushl %ebp -.LCFI4: - movl %esp, %ebp -.LCFI5: - pushl %esi -.LCFI6: - subl $36, %esp - movl RAW_CLOSURE_CIF_OFFSET(%eax), %esi /* closure->cif */ - movl RAW_CLOSURE_USER_DATA_OFFSET(%eax), %edx /* closure->user_data */ - movl %edx, 12(%esp) /* user_data */ - leal 8(%ebp), %edx /* __builtin_dwarf_cfa () */ - movl %edx, 8(%esp) /* raw_args */ - leal -24(%ebp), %edx - movl %edx, 4(%esp) /* &res */ - movl %esi, (%esp) /* cif */ - call *RAW_CLOSURE_FUN_OFFSET(%eax) /* closure->fun */ - movl CIF_FLAGS_OFFSET(%esi), %eax /* rtype */ - cmpl $FFI_TYPE_INT, %eax - je .Lrcls_retint - - /* Handle FFI_TYPE_UINT8, FFI_TYPE_SINT8, FFI_TYPE_UINT16, - FFI_TYPE_SINT16, FFI_TYPE_UINT32, FFI_TYPE_SINT32. */ - cmpl $FFI_TYPE_UINT64, %eax - jge 0f - cmpl $FFI_TYPE_UINT8, %eax - jge .Lrcls_retint -0: - cmpl $FFI_TYPE_FLOAT, %eax - je .Lrcls_retfloat - cmpl $FFI_TYPE_DOUBLE, %eax - je .Lrcls_retdouble - cmpl $FFI_TYPE_LONGDOUBLE, %eax - je .Lrcls_retldouble - cmpl $FFI_TYPE_SINT64, %eax - je .Lrcls_retllong -.Lrcls_epilogue: - addl $36, %esp - popl %esi + movl 16(%ebp), %ecx /* load result address */ + _CET_NOTRACK jmp *%ebx + + .balign 8 +L(store_table): +E(L(store_table), X86_RET_FLOAT) + fstps (%ecx) + jmp L(e1) +E(L(store_table), X86_RET_DOUBLE) + fstpl (%ecx) + jmp L(e1) +E(L(store_table), X86_RET_LDOUBLE) + fstpt (%ecx) + jmp L(e1) +E(L(store_table), X86_RET_SINT8) + movsbl %al, %eax + mov %eax, (%ecx) + jmp L(e1) +E(L(store_table), X86_RET_SINT16) + movswl %ax, %eax + mov %eax, (%ecx) + jmp L(e1) +E(L(store_table), X86_RET_UINT8) + movzbl %al, %eax + mov %eax, (%ecx) + jmp L(e1) +E(L(store_table), X86_RET_UINT16) + movzwl %ax, %eax + mov %eax, (%ecx) + jmp L(e1) +E(L(store_table), X86_RET_INT64) + movl %edx, 4(%ecx) + /* fallthru */ +E(L(store_table), X86_RET_INT32) + movl %eax, (%ecx) + /* fallthru */ +E(L(store_table), X86_RET_VOID) +L(e1): + movl 8(%ebp), %ebx + movl %ebp, %esp popl %ebp +L(UW3): + # cfi_remember_state + # cfi_def_cfa(%esp, 4) + # cfi_restore(%ebx) + # cfi_restore(%ebp) ret -.Lrcls_retint: - movl -24(%ebp), %eax - jmp .Lrcls_epilogue -.Lrcls_retfloat: - flds -24(%ebp) - jmp .Lrcls_epilogue -.Lrcls_retdouble: - fldl -24(%ebp) - jmp .Lrcls_epilogue -.Lrcls_retldouble: - fldt -24(%ebp) - jmp .Lrcls_epilogue -.Lrcls_retllong: - movl -24(%ebp), %eax - movl -20(%ebp), %edx - jmp .Lrcls_epilogue -.LFE3: - .size ffi_closure_raw_SYSV, .-ffi_closure_raw_SYSV +L(UW4): + # cfi_restore_state + +E(L(store_table), X86_RET_STRUCTPOP) + jmp L(e1) +E(L(store_table), X86_RET_STRUCTARG) + jmp L(e1) +E(L(store_table), X86_RET_STRUCT_1B) + movb %al, (%ecx) + jmp L(e1) +E(L(store_table), X86_RET_STRUCT_2B) + movw %ax, (%ecx) + jmp L(e1) + + /* Fill out the table so that bad values are predictable. */ +E(L(store_table), X86_RET_UNUSED14) + ud2 +E(L(store_table), X86_RET_UNUSED15) + ud2 + +L(UW5): + # cfi_endproc +ENDF(ffi_call_i386) + +/* The inner helper is declared as + + void ffi_closure_inner(struct closure_frame *frame, char *argp) + __attribute_((fastcall)) + + Thus the arguments are placed in + + ecx: frame + edx: argp +*/ + +/* Macros to help setting up the closure_data structure. */ + +#if HAVE_FASTCALL +# define closure_FS (40 + 4) +# define closure_CF 0 +#else +# define closure_FS (8 + 40 + 12) +# define closure_CF 8 #endif -#if defined __GNUC__ -/* Only emit dwarf unwind info when building with GNU toolchain. */ +#define FFI_CLOSURE_SAVE_REGS \ + movl %eax, closure_CF+16+R_EAX*4(%esp); \ + movl %edx, closure_CF+16+R_EDX*4(%esp); \ + movl %ecx, closure_CF+16+R_ECX*4(%esp) -#if defined __PIC__ -# if defined __sun__ && defined __svr4__ -/* 32-bit Solaris 2/x86 uses datarel encoding for PIC. GNU ld before 2.22 - doesn't correctly sort .eh_frame_hdr with mixed encodings, so match this. */ -# define FDE_ENCODING 0x30 /* datarel */ -# define FDE_ENCODE(X) X@GOTOFF -# else -# define FDE_ENCODING 0x1b /* pcrel sdata4 */ -# if defined HAVE_AS_X86_PCREL -# define FDE_ENCODE(X) X-. -# else -# define FDE_ENCODE(X) X@rel -# endif -# endif +#define FFI_CLOSURE_COPY_TRAMP_DATA \ + movl FFI_TRAMPOLINE_SIZE(%eax), %edx; /* copy cif */ \ + movl FFI_TRAMPOLINE_SIZE+4(%eax), %ecx; /* copy fun */ \ + movl FFI_TRAMPOLINE_SIZE+8(%eax), %eax; /* copy user_data */ \ + movl %edx, closure_CF+28(%esp); \ + movl %ecx, closure_CF+32(%esp); \ + movl %eax, closure_CF+36(%esp) + +#if HAVE_FASTCALL +# define FFI_CLOSURE_PREP_CALL \ + movl %esp, %ecx; /* load closure_data */ \ + leal closure_FS+4(%esp), %edx; /* load incoming stack */ #else -# define FDE_ENCODING 0 /* absolute */ -# define FDE_ENCODE(X) X +# define FFI_CLOSURE_PREP_CALL \ + leal closure_CF(%esp), %ecx; /* load closure_data */ \ + leal closure_FS+4(%esp), %edx; /* load incoming stack */ \ + movl %ecx, (%esp); \ + movl %edx, 4(%esp) #endif - .section .eh_frame,EH_FRAME_FLAGS,@progbits -.Lframe1: - .long .LECIE1-.LSCIE1 /* Length of Common Information Entry */ -.LSCIE1: - .long 0x0 /* CIE Identifier Tag */ - .byte 0x1 /* CIE Version */ -#ifdef HAVE_AS_ASCII_PSEUDO_OP +#define FFI_CLOSURE_CALL_INNER(UWN) \ + call ffi_closure_inner + +#define FFI_CLOSURE_MASK_AND_JUMP(N, UW) \ + andl $X86_RET_TYPE_MASK, %eax; \ + leal L(C1(load_table,N))(, %eax, 8), %edx; \ + movl closure_CF(%esp), %eax; /* optimiztic load */ \ + _CET_NOTRACK jmp *%edx + #ifdef __PIC__ - .ascii "zR\0" /* CIE Augmentation */ +# if defined X86_DARWIN || defined HAVE_HIDDEN_VISIBILITY_ATTRIBUTE +# undef FFI_CLOSURE_MASK_AND_JUMP +# define FFI_CLOSURE_MASK_AND_JUMP(N, UW) \ + andl $X86_RET_TYPE_MASK, %eax; \ + call C(__x86.get_pc_thunk.dx); \ +L(C1(pc,N)): \ + leal L(C1(load_table,N))-L(C1(pc,N))(%edx, %eax, 8), %edx; \ + movl closure_CF(%esp), %eax; /* optimiztic load */ \ + _CET_NOTRACK jmp *%edx +# else +# define FFI_CLOSURE_CALL_INNER_SAVE_EBX +# undef FFI_CLOSURE_CALL_INNER +# define FFI_CLOSURE_CALL_INNER(UWN) \ + movl %ebx, 40(%esp); /* save ebx */ \ +L(C1(UW,UWN)): \ + /* cfi_rel_offset(%ebx, 40); */ \ + call C(__x86.get_pc_thunk.bx); /* load got register */ \ + addl $C(_GLOBAL_OFFSET_TABLE_), %ebx; \ + call ffi_closure_inner@PLT +# undef FFI_CLOSURE_MASK_AND_JUMP +# define FFI_CLOSURE_MASK_AND_JUMP(N, UWN) \ + andl $X86_RET_TYPE_MASK, %eax; \ + leal L(C1(load_table,N))@GOTOFF(%ebx, %eax, 8), %edx; \ + movl 40(%esp), %ebx; /* restore ebx */ \ +L(C1(UW,UWN)): \ + /* cfi_restore(%ebx); */ \ + movl closure_CF(%esp), %eax; /* optimiztic load */ \ + _CET_NOTRACK jmp *%edx +# endif /* DARWIN || HIDDEN */ +#endif /* __PIC__ */ + + .balign 16 + .globl C(ffi_go_closure_EAX) + FFI_HIDDEN(C(ffi_go_closure_EAX)) +C(ffi_go_closure_EAX): +L(UW6): + # cfi_startproc + _CET_ENDBR + subl $closure_FS, %esp +L(UW7): + # cfi_def_cfa_offset(closure_FS + 4) + FFI_CLOSURE_SAVE_REGS + movl 4(%eax), %edx /* copy cif */ + movl 8(%eax), %ecx /* copy fun */ + movl %edx, closure_CF+28(%esp) + movl %ecx, closure_CF+32(%esp) + movl %eax, closure_CF+36(%esp) /* closure is user_data */ + jmp L(do_closure_i386) +L(UW8): + # cfi_endproc +ENDF(C(ffi_go_closure_EAX)) + + .balign 16 + .globl C(ffi_go_closure_ECX) + FFI_HIDDEN(C(ffi_go_closure_ECX)) +C(ffi_go_closure_ECX): +L(UW9): + # cfi_startproc + _CET_ENDBR + subl $closure_FS, %esp +L(UW10): + # cfi_def_cfa_offset(closure_FS + 4) + FFI_CLOSURE_SAVE_REGS + movl 4(%ecx), %edx /* copy cif */ + movl 8(%ecx), %eax /* copy fun */ + movl %edx, closure_CF+28(%esp) + movl %eax, closure_CF+32(%esp) + movl %ecx, closure_CF+36(%esp) /* closure is user_data */ + jmp L(do_closure_i386) +L(UW11): + # cfi_endproc +ENDF(C(ffi_go_closure_ECX)) + +/* The closure entry points are reached from the ffi_closure trampoline. + On entry, %eax contains the address of the ffi_closure. */ + + .balign 16 + .globl C(ffi_closure_i386) + FFI_HIDDEN(C(ffi_closure_i386)) + +C(ffi_closure_i386): +L(UW12): + # cfi_startproc + _CET_ENDBR + subl $closure_FS, %esp +L(UW13): + # cfi_def_cfa_offset(closure_FS + 4) + + FFI_CLOSURE_SAVE_REGS + FFI_CLOSURE_COPY_TRAMP_DATA + + /* Entry point from preceeding Go closures. */ +L(do_closure_i386): + + FFI_CLOSURE_PREP_CALL + FFI_CLOSURE_CALL_INNER(14) + FFI_CLOSURE_MASK_AND_JUMP(2, 15) + + .balign 8 +L(load_table2): +E(L(load_table2), X86_RET_FLOAT) + flds closure_CF(%esp) + jmp L(e2) +E(L(load_table2), X86_RET_DOUBLE) + fldl closure_CF(%esp) + jmp L(e2) +E(L(load_table2), X86_RET_LDOUBLE) + fldt closure_CF(%esp) + jmp L(e2) +E(L(load_table2), X86_RET_SINT8) + movsbl %al, %eax + jmp L(e2) +E(L(load_table2), X86_RET_SINT16) + movswl %ax, %eax + jmp L(e2) +E(L(load_table2), X86_RET_UINT8) + movzbl %al, %eax + jmp L(e2) +E(L(load_table2), X86_RET_UINT16) + movzwl %ax, %eax + jmp L(e2) +E(L(load_table2), X86_RET_INT64) + movl closure_CF+4(%esp), %edx + jmp L(e2) +E(L(load_table2), X86_RET_INT32) + nop + /* fallthru */ +E(L(load_table2), X86_RET_VOID) +L(e2): + addl $closure_FS, %esp +L(UW16): + # cfi_adjust_cfa_offset(-closure_FS) + ret +L(UW17): + # cfi_adjust_cfa_offset(closure_FS) +E(L(load_table2), X86_RET_STRUCTPOP) + addl $closure_FS, %esp +L(UW18): + # cfi_adjust_cfa_offset(-closure_FS) + ret $4 +L(UW19): + # cfi_adjust_cfa_offset(closure_FS) +E(L(load_table2), X86_RET_STRUCTARG) + jmp L(e2) +E(L(load_table2), X86_RET_STRUCT_1B) + movzbl %al, %eax + jmp L(e2) +E(L(load_table2), X86_RET_STRUCT_2B) + movzwl %ax, %eax + jmp L(e2) + + /* Fill out the table so that bad values are predictable. */ +E(L(load_table2), X86_RET_UNUSED14) + ud2 +E(L(load_table2), X86_RET_UNUSED15) + ud2 + +L(UW20): + # cfi_endproc +ENDF(C(ffi_closure_i386)) + + .balign 16 + .globl C(ffi_go_closure_STDCALL) + FFI_HIDDEN(C(ffi_go_closure_STDCALL)) +C(ffi_go_closure_STDCALL): +L(UW21): + # cfi_startproc + _CET_ENDBR + subl $closure_FS, %esp +L(UW22): + # cfi_def_cfa_offset(closure_FS + 4) + FFI_CLOSURE_SAVE_REGS + movl 4(%ecx), %edx /* copy cif */ + movl 8(%ecx), %eax /* copy fun */ + movl %edx, closure_CF+28(%esp) + movl %eax, closure_CF+32(%esp) + movl %ecx, closure_CF+36(%esp) /* closure is user_data */ + jmp L(do_closure_STDCALL) +L(UW23): + # cfi_endproc +ENDF(C(ffi_go_closure_STDCALL)) + +/* For REGISTER, we have no available parameter registers, and so we + enter here having pushed the closure onto the stack. */ + + .balign 16 + .globl C(ffi_closure_REGISTER) + FFI_HIDDEN(C(ffi_closure_REGISTER)) +C(ffi_closure_REGISTER): +L(UW24): + # cfi_startproc + # cfi_def_cfa(%esp, 8) + # cfi_offset(%eip, -8) + _CET_ENDBR + subl $closure_FS-4, %esp +L(UW25): + # cfi_def_cfa_offset(closure_FS + 4) + FFI_CLOSURE_SAVE_REGS + movl closure_FS-4(%esp), %ecx /* load retaddr */ + movl closure_FS(%esp), %eax /* load closure */ + movl %ecx, closure_FS(%esp) /* move retaddr */ + jmp L(do_closure_REGISTER) +L(UW26): + # cfi_endproc +ENDF(C(ffi_closure_REGISTER)) + +/* For STDCALL (and others), we need to pop N bytes of arguments off + the stack following the closure. The amount needing to be popped + is returned to us from ffi_closure_inner. */ + + .balign 16 + .globl C(ffi_closure_STDCALL) + FFI_HIDDEN(C(ffi_closure_STDCALL)) +C(ffi_closure_STDCALL): +L(UW27): + # cfi_startproc + _CET_ENDBR + subl $closure_FS, %esp +L(UW28): + # cfi_def_cfa_offset(closure_FS + 4) + + FFI_CLOSURE_SAVE_REGS + + /* Entry point from ffi_closure_REGISTER. */ +L(do_closure_REGISTER): + + FFI_CLOSURE_COPY_TRAMP_DATA + + /* Entry point from preceeding Go closure. */ +L(do_closure_STDCALL): + + FFI_CLOSURE_PREP_CALL + FFI_CLOSURE_CALL_INNER(29) + + movl %eax, %ecx + shrl $X86_RET_POP_SHIFT, %ecx /* isolate pop count */ + leal closure_FS(%esp, %ecx), %ecx /* compute popped esp */ + movl closure_FS(%esp), %edx /* move return address */ + movl %edx, (%ecx) + + /* From this point on, the value of %esp upon return is %ecx+4, + and we've copied the return address to %ecx to make return easy. + There's no point in representing this in the unwind info, as + there is always a window between the mov and the ret which + will be wrong from one point of view or another. */ + + FFI_CLOSURE_MASK_AND_JUMP(3, 30) + + .balign 8 +L(load_table3): +E(L(load_table3), X86_RET_FLOAT) + flds closure_CF(%esp) + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_DOUBLE) + fldl closure_CF(%esp) + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_LDOUBLE) + fldt closure_CF(%esp) + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_SINT8) + movsbl %al, %eax + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_SINT16) + movswl %ax, %eax + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_UINT8) + movzbl %al, %eax + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_UINT16) + movzwl %ax, %eax + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_INT64) + movl closure_CF+4(%esp), %edx + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_INT32) + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_VOID) + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_STRUCTPOP) + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_STRUCTARG) + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_STRUCT_1B) + movzbl %al, %eax + movl %ecx, %esp + ret +E(L(load_table3), X86_RET_STRUCT_2B) + movzwl %ax, %eax + movl %ecx, %esp + ret + + /* Fill out the table so that bad values are predictable. */ +E(L(load_table3), X86_RET_UNUSED14) + ud2 +E(L(load_table3), X86_RET_UNUSED15) + ud2 + +L(UW31): + # cfi_endproc +ENDF(C(ffi_closure_STDCALL)) + +#if !FFI_NO_RAW_API + +#define raw_closure_S_FS (16+16+12) + + .balign 16 + .globl C(ffi_closure_raw_SYSV) + FFI_HIDDEN(C(ffi_closure_raw_SYSV)) +C(ffi_closure_raw_SYSV): +L(UW32): + # cfi_startproc + _CET_ENDBR + subl $raw_closure_S_FS, %esp +L(UW33): + # cfi_def_cfa_offset(raw_closure_S_FS + 4) + movl %ebx, raw_closure_S_FS-4(%esp) +L(UW34): + # cfi_rel_offset(%ebx, raw_closure_S_FS-4) + + movl FFI_TRAMPOLINE_SIZE+8(%eax), %edx /* load cl->user_data */ + movl %edx, 12(%esp) + leal raw_closure_S_FS+4(%esp), %edx /* load raw_args */ + movl %edx, 8(%esp) + leal 16(%esp), %edx /* load &res */ + movl %edx, 4(%esp) + movl FFI_TRAMPOLINE_SIZE(%eax), %ebx /* load cl->cif */ + movl %ebx, (%esp) + call *FFI_TRAMPOLINE_SIZE+4(%eax) /* call cl->fun */ + + movl 20(%ebx), %eax /* load cif->flags */ + andl $X86_RET_TYPE_MASK, %eax +#ifdef __PIC__ + call C(__x86.get_pc_thunk.bx) +L(pc4): + leal L(load_table4)-L(pc4)(%ebx, %eax, 8), %ecx #else - .ascii "\0" /* CIE Augmentation */ + leal L(load_table4)(,%eax, 8), %ecx #endif -#elif defined HAVE_AS_STRING_PSEUDO_OP + movl raw_closure_S_FS-4(%esp), %ebx +L(UW35): + # cfi_restore(%ebx) + movl 16(%esp), %eax /* Optimistic load */ + jmp *%ecx + + .balign 8 +L(load_table4): +E(L(load_table4), X86_RET_FLOAT) + flds 16(%esp) + jmp L(e4) +E(L(load_table4), X86_RET_DOUBLE) + fldl 16(%esp) + jmp L(e4) +E(L(load_table4), X86_RET_LDOUBLE) + fldt 16(%esp) + jmp L(e4) +E(L(load_table4), X86_RET_SINT8) + movsbl %al, %eax + jmp L(e4) +E(L(load_table4), X86_RET_SINT16) + movswl %ax, %eax + jmp L(e4) +E(L(load_table4), X86_RET_UINT8) + movzbl %al, %eax + jmp L(e4) +E(L(load_table4), X86_RET_UINT16) + movzwl %ax, %eax + jmp L(e4) +E(L(load_table4), X86_RET_INT64) + movl 16+4(%esp), %edx + jmp L(e4) +E(L(load_table4), X86_RET_INT32) + nop + /* fallthru */ +E(L(load_table4), X86_RET_VOID) +L(e4): + addl $raw_closure_S_FS, %esp +L(UW36): + # cfi_adjust_cfa_offset(-raw_closure_S_FS) + ret +L(UW37): + # cfi_adjust_cfa_offset(raw_closure_S_FS) +E(L(load_table4), X86_RET_STRUCTPOP) + addl $raw_closure_S_FS, %esp +L(UW38): + # cfi_adjust_cfa_offset(-raw_closure_S_FS) + ret $4 +L(UW39): + # cfi_adjust_cfa_offset(raw_closure_S_FS) +E(L(load_table4), X86_RET_STRUCTARG) + jmp L(e4) +E(L(load_table4), X86_RET_STRUCT_1B) + movzbl %al, %eax + jmp L(e4) +E(L(load_table4), X86_RET_STRUCT_2B) + movzwl %ax, %eax + jmp L(e4) + + /* Fill out the table so that bad values are predictable. */ +E(L(load_table4), X86_RET_UNUSED14) + ud2 +E(L(load_table4), X86_RET_UNUSED15) + ud2 + +L(UW40): + # cfi_endproc +ENDF(C(ffi_closure_raw_SYSV)) + +#define raw_closure_T_FS (16+16+8) + + .balign 16 + .globl C(ffi_closure_raw_THISCALL) + FFI_HIDDEN(C(ffi_closure_raw_THISCALL)) +C(ffi_closure_raw_THISCALL): +L(UW41): + # cfi_startproc + _CET_ENDBR + /* Rearrange the stack such that %ecx is the first argument. + This means moving the return address. */ + popl %edx +L(UW42): + # cfi_def_cfa_offset(0) + # cfi_register(%eip, %edx) + pushl %ecx +L(UW43): + # cfi_adjust_cfa_offset(4) + pushl %edx +L(UW44): + # cfi_adjust_cfa_offset(4) + # cfi_rel_offset(%eip, 0) + subl $raw_closure_T_FS, %esp +L(UW45): + # cfi_adjust_cfa_offset(raw_closure_T_FS) + movl %ebx, raw_closure_T_FS-4(%esp) +L(UW46): + # cfi_rel_offset(%ebx, raw_closure_T_FS-4) + + movl FFI_TRAMPOLINE_SIZE+8(%eax), %edx /* load cl->user_data */ + movl %edx, 12(%esp) + leal raw_closure_T_FS+4(%esp), %edx /* load raw_args */ + movl %edx, 8(%esp) + leal 16(%esp), %edx /* load &res */ + movl %edx, 4(%esp) + movl FFI_TRAMPOLINE_SIZE(%eax), %ebx /* load cl->cif */ + movl %ebx, (%esp) + call *FFI_TRAMPOLINE_SIZE+4(%eax) /* call cl->fun */ + + movl 20(%ebx), %eax /* load cif->flags */ + andl $X86_RET_TYPE_MASK, %eax #ifdef __PIC__ - .string "zR" /* CIE Augmentation */ + call C(__x86.get_pc_thunk.bx) +L(pc5): + leal L(load_table5)-L(pc5)(%ebx, %eax, 8), %ecx #else - .string "" /* CIE Augmentation */ + leal L(load_table5)(,%eax, 8), %ecx #endif + movl raw_closure_T_FS-4(%esp), %ebx +L(UW47): + # cfi_restore(%ebx) + movl 16(%esp), %eax /* Optimistic load */ + jmp *%ecx + + .balign 8 +L(load_table5): +E(L(load_table5), X86_RET_FLOAT) + flds 16(%esp) + jmp L(e5) +E(L(load_table5), X86_RET_DOUBLE) + fldl 16(%esp) + jmp L(e5) +E(L(load_table5), X86_RET_LDOUBLE) + fldt 16(%esp) + jmp L(e5) +E(L(load_table5), X86_RET_SINT8) + movsbl %al, %eax + jmp L(e5) +E(L(load_table5), X86_RET_SINT16) + movswl %ax, %eax + jmp L(e5) +E(L(load_table5), X86_RET_UINT8) + movzbl %al, %eax + jmp L(e5) +E(L(load_table5), X86_RET_UINT16) + movzwl %ax, %eax + jmp L(e5) +E(L(load_table5), X86_RET_INT64) + movl 16+4(%esp), %edx + jmp L(e5) +E(L(load_table5), X86_RET_INT32) + nop + /* fallthru */ +E(L(load_table5), X86_RET_VOID) +L(e5): + addl $raw_closure_T_FS, %esp +L(UW48): + # cfi_adjust_cfa_offset(-raw_closure_T_FS) + /* Remove the extra %ecx argument we pushed. */ + ret $4 +L(UW49): + # cfi_adjust_cfa_offset(raw_closure_T_FS) +E(L(load_table5), X86_RET_STRUCTPOP) + addl $raw_closure_T_FS, %esp +L(UW50): + # cfi_adjust_cfa_offset(-raw_closure_T_FS) + ret $8 +L(UW51): + # cfi_adjust_cfa_offset(raw_closure_T_FS) +E(L(load_table5), X86_RET_STRUCTARG) + jmp L(e5) +E(L(load_table5), X86_RET_STRUCT_1B) + movzbl %al, %eax + jmp L(e5) +E(L(load_table5), X86_RET_STRUCT_2B) + movzwl %ax, %eax + jmp L(e5) + + /* Fill out the table so that bad values are predictable. */ +E(L(load_table5), X86_RET_UNUSED14) + ud2 +E(L(load_table5), X86_RET_UNUSED15) + ud2 + +L(UW52): + # cfi_endproc +ENDF(C(ffi_closure_raw_THISCALL)) + +#endif /* !FFI_NO_RAW_API */ + +#ifdef X86_DARWIN +# define COMDAT(X) \ + .section __TEXT,__text,coalesced,pure_instructions; \ + .weak_definition X; \ + FFI_HIDDEN(X) +#elif defined __ELF__ && !(defined(__sun__) && defined(__svr4__)) +# define COMDAT(X) \ + .section .text.X,"axG",@progbits,X,comdat; \ + .globl X; \ + FFI_HIDDEN(X) #else -#error missing .ascii/.string +# define COMDAT(X) #endif - .byte 0x1 /* .uleb128 0x1; CIE Code Alignment Factor */ - .byte 0x7c /* .sleb128 -4; CIE Data Alignment Factor */ - .byte 0x8 /* CIE RA Column */ -#ifdef __PIC__ - .byte 0x1 /* .uleb128 0x1; Augmentation size */ - .byte FDE_ENCODING + +#if defined(__PIC__) + COMDAT(C(__x86.get_pc_thunk.bx)) +C(__x86.get_pc_thunk.bx): + movl (%esp), %ebx + ret +ENDF(C(__x86.get_pc_thunk.bx)) +# if defined X86_DARWIN || defined HAVE_HIDDEN_VISIBILITY_ATTRIBUTE + COMDAT(C(__x86.get_pc_thunk.dx)) +C(__x86.get_pc_thunk.dx): + movl (%esp), %edx + ret +ENDF(C(__x86.get_pc_thunk.dx)) +#endif /* DARWIN || HIDDEN */ +#endif /* __PIC__ */ + +/* Sadly, OSX cctools-as doesn't understand .cfi directives at all. */ + +#ifdef __APPLE__ +.section __TEXT,__eh_frame,coalesced,no_toc+strip_static_syms+live_support +EHFrame0: +#elif defined(X86_WIN32) +.section .eh_frame,"r" +#elif defined(HAVE_AS_X86_64_UNWIND_SECTION_TYPE) +.section .eh_frame,EH_FRAME_FLAGS,@unwind +#else +.section .eh_frame,EH_FRAME_FLAGS,@progbits #endif - .byte 0xc /* DW_CFA_def_cfa */ - .byte 0x4 /* .uleb128 0x4 */ - .byte 0x4 /* .uleb128 0x4 */ - .byte 0x88 /* DW_CFA_offset, column 0x8 */ - .byte 0x1 /* .uleb128 0x1 */ - .align 4 -.LECIE1: -.LSFDE1: - .long .LEFDE1-.LASFDE1 /* FDE Length */ -.LASFDE1: - .long .LASFDE1-.Lframe1 /* FDE CIE offset */ - .long FDE_ENCODE(.LFB1) /* FDE initial location */ - .long .LFE1-.LFB1 /* FDE address range */ -#ifdef __PIC__ - .byte 0x0 /* .uleb128 0x0; Augmentation size */ + +#ifdef HAVE_AS_X86_PCREL +# define PCREL(X) X - . +#else +# define PCREL(X) X@rel #endif - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LCFI0-.LFB1 - .byte 0xe /* DW_CFA_def_cfa_offset */ - .byte 0x8 /* .uleb128 0x8 */ - .byte 0x85 /* DW_CFA_offset, column 0x5 */ - .byte 0x2 /* .uleb128 0x2 */ - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LCFI1-.LCFI0 - .byte 0xd /* DW_CFA_def_cfa_register */ - .byte 0x5 /* .uleb128 0x5 */ - .align 4 -.LEFDE1: -.LSFDE2: - .long .LEFDE2-.LASFDE2 /* FDE Length */ -.LASFDE2: - .long .LASFDE2-.Lframe1 /* FDE CIE offset */ - .long FDE_ENCODE(.LFB2) /* FDE initial location */ - .long .LFE2-.LFB2 /* FDE address range */ -#ifdef __PIC__ - .byte 0x0 /* .uleb128 0x0; Augmentation size */ + +/* Simplify advancing between labels. Assume DW_CFA_advance_loc1 fits. */ +#define ADV(N, P) .byte 2, L(N)-L(P) + + .balign 4 +L(CIE): + .set L(set0),L(ECIE)-L(SCIE) + .long L(set0) /* CIE Length */ +L(SCIE): + .long 0 /* CIE Identifier Tag */ + .byte 1 /* CIE Version */ + .ascii "zR\0" /* CIE Augmentation */ + .byte 1 /* CIE Code Alignment Factor */ + .byte 0x7c /* CIE Data Alignment Factor */ + .byte 0x8 /* CIE RA Column */ + .byte 1 /* Augmentation size */ + .byte 0x1b /* FDE Encoding (pcrel sdata4) */ + .byte 0xc, 4, 4 /* DW_CFA_def_cfa, %esp offset 4 */ + .byte 0x80+8, 1 /* DW_CFA_offset, %eip offset 1*-4 */ + .balign 4 +L(ECIE): + + .set L(set1),L(EFDE1)-L(SFDE1) + .long L(set1) /* FDE Length */ +L(SFDE1): + .long L(SFDE1)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW0)) /* Initial location */ + .long L(UW5)-L(UW0) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW1, UW0) + .byte 0xc, 5, 8 /* DW_CFA_def_cfa, %ebp 8 */ + .byte 0x80+5, 2 /* DW_CFA_offset, %ebp 2*-4 */ + ADV(UW2, UW1) + .byte 0x80+3, 0 /* DW_CFA_offset, %ebx 0*-4 */ + ADV(UW3, UW2) + .byte 0xa /* DW_CFA_remember_state */ + .byte 0xc, 4, 4 /* DW_CFA_def_cfa, %esp 4 */ + .byte 0xc0+3 /* DW_CFA_restore, %ebx */ + .byte 0xc0+5 /* DW_CFA_restore, %ebp */ + ADV(UW4, UW3) + .byte 0xb /* DW_CFA_restore_state */ + .balign 4 +L(EFDE1): + + .set L(set2),L(EFDE2)-L(SFDE2) + .long L(set2) /* FDE Length */ +L(SFDE2): + .long L(SFDE2)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW6)) /* Initial location */ + .long L(UW8)-L(UW6) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW7, UW6) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE2): + + .set L(set3),L(EFDE3)-L(SFDE3) + .long L(set3) /* FDE Length */ +L(SFDE3): + .long L(SFDE3)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW9)) /* Initial location */ + .long L(UW11)-L(UW9) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW10, UW9) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE3): + + .set L(set4),L(EFDE4)-L(SFDE4) + .long L(set4) /* FDE Length */ +L(SFDE4): + .long L(SFDE4)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW12)) /* Initial location */ + .long L(UW20)-L(UW12) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW13, UW12) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ +#ifdef FFI_CLOSURE_CALL_INNER_SAVE_EBX + ADV(UW14, UW13) + .byte 0x80+3, (40-(closure_FS+4))/-4 /* DW_CFA_offset %ebx */ + ADV(UW15, UW14) + .byte 0xc0+3 /* DW_CFA_restore %ebx */ + ADV(UW16, UW15) +#else + ADV(UW16, UW13) #endif - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LCFI2-.LFB2 - .byte 0xe /* DW_CFA_def_cfa_offset */ - .byte 0x8 /* .uleb128 0x8 */ - .byte 0x85 /* DW_CFA_offset, column 0x5 */ - .byte 0x2 /* .uleb128 0x2 */ - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LCFI3-.LCFI2 - .byte 0xd /* DW_CFA_def_cfa_register */ - .byte 0x5 /* .uleb128 0x5 */ -#if !defined HAVE_HIDDEN_VISIBILITY_ATTRIBUTE && defined __PIC__ - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LCFI7-.LCFI3 - .byte 0x83 /* DW_CFA_offset, column 0x3 */ - .byte 0xa /* .uleb128 0xa */ + .byte 0xe, 4 /* DW_CFA_def_cfa_offset */ + ADV(UW17, UW16) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + ADV(UW18, UW17) + .byte 0xe, 4 /* DW_CFA_def_cfa_offset */ + ADV(UW19, UW18) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE4): + + .set L(set5),L(EFDE5)-L(SFDE5) + .long L(set5) /* FDE Length */ +L(SFDE5): + .long L(SFDE5)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW21)) /* Initial location */ + .long L(UW23)-L(UW21) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW22, UW21) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE5): + + .set L(set6),L(EFDE6)-L(SFDE6) + .long L(set6) /* FDE Length */ +L(SFDE6): + .long L(SFDE6)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW24)) /* Initial location */ + .long L(UW26)-L(UW24) /* Address range */ + .byte 0 /* Augmentation size */ + .byte 0xe, 8 /* DW_CFA_def_cfa_offset */ + .byte 0x80+8, 2 /* DW_CFA_offset %eip, 2*-4 */ + ADV(UW25, UW24) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE6): + + .set L(set7),L(EFDE7)-L(SFDE7) + .long L(set7) /* FDE Length */ +L(SFDE7): + .long L(SFDE7)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW27)) /* Initial location */ + .long L(UW31)-L(UW27) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW28, UW27) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ +#ifdef FFI_CLOSURE_CALL_INNER_SAVE_EBX + ADV(UW29, UW28) + .byte 0x80+3, (40-(closure_FS+4))/-4 /* DW_CFA_offset %ebx */ + ADV(UW30, UW29) + .byte 0xc0+3 /* DW_CFA_restore %ebx */ #endif - .align 4 -.LEFDE2: + .balign 4 +L(EFDE7): #if !FFI_NO_RAW_API + .set L(set8),L(EFDE8)-L(SFDE8) + .long L(set8) /* FDE Length */ +L(SFDE8): + .long L(SFDE8)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW32)) /* Initial location */ + .long L(UW40)-L(UW32) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW33, UW32) + .byte 0xe, raw_closure_S_FS+4 /* DW_CFA_def_cfa_offset */ + ADV(UW34, UW33) + .byte 0x80+3, 2 /* DW_CFA_offset %ebx 2*-4 */ + ADV(UW35, UW34) + .byte 0xc0+3 /* DW_CFA_restore %ebx */ + ADV(UW36, UW35) + .byte 0xe, 4 /* DW_CFA_def_cfa_offset */ + ADV(UW37, UW36) + .byte 0xe, raw_closure_S_FS+4 /* DW_CFA_def_cfa_offset */ + ADV(UW38, UW37) + .byte 0xe, 4 /* DW_CFA_def_cfa_offset */ + ADV(UW39, UW38) + .byte 0xe, raw_closure_S_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE8): -.LSFDE3: - .long .LEFDE3-.LASFDE3 /* FDE Length */ -.LASFDE3: - .long .LASFDE3-.Lframe1 /* FDE CIE offset */ - .long FDE_ENCODE(.LFB3) /* FDE initial location */ - .long .LFE3-.LFB3 /* FDE address range */ -#ifdef __PIC__ - .byte 0x0 /* .uleb128 0x0; Augmentation size */ -#endif - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LCFI4-.LFB3 - .byte 0xe /* DW_CFA_def_cfa_offset */ - .byte 0x8 /* .uleb128 0x8 */ - .byte 0x85 /* DW_CFA_offset, column 0x5 */ - .byte 0x2 /* .uleb128 0x2 */ - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LCFI5-.LCFI4 - .byte 0xd /* DW_CFA_def_cfa_register */ - .byte 0x5 /* .uleb128 0x5 */ - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LCFI6-.LCFI5 - .byte 0x86 /* DW_CFA_offset, column 0x6 */ - .byte 0x3 /* .uleb128 0x3 */ - .align 4 -.LEFDE3: + .set L(set9),L(EFDE9)-L(SFDE9) + .long L(set9) /* FDE Length */ +L(SFDE9): + .long L(SFDE9)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW41)) /* Initial location */ + .long L(UW52)-L(UW41) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW42, UW41) + .byte 0xe, 0 /* DW_CFA_def_cfa_offset */ + .byte 0x9, 8, 2 /* DW_CFA_register %eip, %edx */ + ADV(UW43, UW42) + .byte 0xe, 4 /* DW_CFA_def_cfa_offset */ + ADV(UW44, UW43) + .byte 0xe, 8 /* DW_CFA_def_cfa_offset */ + .byte 0x80+8, 2 /* DW_CFA_offset %eip 2*-4 */ + ADV(UW45, UW44) + .byte 0xe, raw_closure_T_FS+8 /* DW_CFA_def_cfa_offset */ + ADV(UW46, UW45) + .byte 0x80+3, 3 /* DW_CFA_offset %ebx 3*-4 */ + ADV(UW47, UW46) + .byte 0xc0+3 /* DW_CFA_restore %ebx */ + ADV(UW48, UW47) + .byte 0xe, 8 /* DW_CFA_def_cfa_offset */ + ADV(UW49, UW48) + .byte 0xe, raw_closure_T_FS+8 /* DW_CFA_def_cfa_offset */ + ADV(UW50, UW49) + .byte 0xe, 8 /* DW_CFA_def_cfa_offset */ + ADV(UW51, UW50) + .byte 0xe, raw_closure_T_FS+8 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE9): +#endif /* !FFI_NO_RAW_API */ +#ifdef _WIN32 + .def @feat.00; + .scl 3; + .type 0; + .endef + .globl @feat.00 +@feat.00 = 1 #endif -#endif -#endif /* ifndef __x86_64__ */ +#ifdef __APPLE__ + .subsections_via_symbols + .section __LD,__compact_unwind,regular,debug + + /* compact unwind for ffi_call_i386 */ + .long C(ffi_call_i386) + .set L1,L(UW5)-L(UW0) + .long L1 + .long 0x04000000 /* use dwarf unwind info */ + .long 0 + .long 0 + + /* compact unwind for ffi_go_closure_EAX */ + .long C(ffi_go_closure_EAX) + .set L2,L(UW8)-L(UW6) + .long L2 + .long 0x04000000 /* use dwarf unwind info */ + .long 0 + .long 0 + + /* compact unwind for ffi_go_closure_ECX */ + .long C(ffi_go_closure_ECX) + .set L3,L(UW11)-L(UW9) + .long L3 + .long 0x04000000 /* use dwarf unwind info */ + .long 0 + .long 0 + + /* compact unwind for ffi_closure_i386 */ + .long C(ffi_closure_i386) + .set L4,L(UW20)-L(UW12) + .long L4 + .long 0x04000000 /* use dwarf unwind info */ + .long 0 + .long 0 + + /* compact unwind for ffi_go_closure_STDCALL */ + .long C(ffi_go_closure_STDCALL) + .set L5,L(UW23)-L(UW21) + .long L5 + .long 0x04000000 /* use dwarf unwind info */ + .long 0 + .long 0 + + /* compact unwind for ffi_closure_REGISTER */ + .long C(ffi_closure_REGISTER) + .set L6,L(UW26)-L(UW24) + .long L6 + .long 0x04000000 /* use dwarf unwind info */ + .long 0 + .long 0 + + /* compact unwind for ffi_closure_STDCALL */ + .long C(ffi_closure_STDCALL) + .set L7,L(UW31)-L(UW27) + .long L7 + .long 0x04000000 /* use dwarf unwind info */ + .long 0 + .long 0 + + /* compact unwind for ffi_closure_raw_SYSV */ + .long C(ffi_closure_raw_SYSV) + .set L8,L(UW40)-L(UW32) + .long L8 + .long 0x04000000 /* use dwarf unwind info */ + .long 0 + .long 0 + + /* compact unwind for ffi_closure_raw_THISCALL */ + .long C(ffi_closure_raw_THISCALL) + .set L9,L(UW52)-L(UW41) + .long L9 + .long 0x04000000 /* use dwarf unwind info */ + .long 0 + .long 0 +#endif /* __APPLE__ */ + +#endif /* ifndef _MSC_VER */ +#endif /* ifdef __i386__ */ #if defined __ELF__ && defined __linux__ .section .note.GNU-stack,"",@progbits diff --git a/native/libffi/src/x86/sysv_intel.S b/native/libffi/src/x86/sysv_intel.S new file mode 100644 index 0000000000..3cafd71ce1 --- /dev/null +++ b/native/libffi/src/x86/sysv_intel.S @@ -0,0 +1,995 @@ +/* ----------------------------------------------------------------------- + sysv.S - Copyright (c) 2017 Anthony Green + - Copyright (c) 2013 The Written Word, Inc. + - Copyright (c) 1996,1998,2001-2003,2005,2008,2010 Red Hat, Inc. + + X86 Foreign Function Interface + + Permission is hereby granted, free of charge, to any person obtaining + a copy of this software and associated documentation files (the + ``Software''), to deal in the Software without restriction, including + without limitation the rights to use, copy, modify, merge, publish, + distribute, sublicense, and/or sell copies of the Software, and to + permit persons to whom the Software is furnished to do so, subject to + the following conditions: + + The above copyright notice and this permission notice shall be included + in all copies or substantial portions of the Software. + + THE SOFTWARE IS PROVIDED ``AS IS'', WITHOUT WARRANTY OF ANY KIND, + EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + DEALINGS IN THE SOFTWARE. + ----------------------------------------------------------------------- */ + +#ifndef __x86_64__ +#ifdef _MSC_VER + +#define LIBFFI_ASM +#include +#include +#include +#include "internal.h" + +#define C2(X, Y) X ## Y +#define C1(X, Y) C2(X, Y) +#define L(X) C1(L, X) +# define ENDF(X) X ENDP + +/* This macro allows the safe creation of jump tables without an + actual table. The entry points into the table are all 8 bytes. + The use of ORG asserts that we're at the correct location. */ +/* ??? The clang assembler doesn't handle .org with symbolic expressions. */ +#if defined(__clang__) || defined(__APPLE__) || (defined (__sun__) && defined(__svr4__)) +# define E(BASE, X) ALIGN 8 +#else +# define E(BASE, X) ALIGN 8; ORG BASE + X * 8 +#endif + + .686P + .MODEL FLAT + +EXTRN @ffi_closure_inner@8:PROC +_TEXT SEGMENT + +/* This is declared as + + void ffi_call_i386(struct call_frame *frame, char *argp) + __attribute__((fastcall)); + + Thus the arguments are present in + + ecx: frame + edx: argp +*/ + +ALIGN 16 +PUBLIC @ffi_call_i386@8 +@ffi_call_i386@8 PROC +L(UW0): + cfi_startproc + #if !HAVE_FASTCALL + mov ecx, [esp+4] + mov edx, [esp+8] + #endif + mov eax, [esp] /* move the return address */ + mov [ecx], ebp /* store ebp into local frame */ + mov [ecx+4], eax /* store retaddr into local frame */ + + /* New stack frame based off ebp. This is a itty bit of unwind + trickery in that the CFA *has* changed. There is no easy way + to describe it correctly on entry to the function. Fortunately, + it doesn't matter too much since at all points we can correctly + unwind back to ffi_call. Note that the location to which we + moved the return address is (the new) CFA-4, so from the + perspective of the unwind info, it hasn't moved. */ + mov ebp, ecx +L(UW1): + // cfi_def_cfa(%ebp, 8) + // cfi_rel_offset(%ebp, 0) + + mov esp, edx /* set outgoing argument stack */ + mov eax, [20+R_EAX*4+ebp] /* set register arguments */ + mov edx, [20+R_EDX*4+ebp] + mov ecx, [20+R_ECX*4+ebp] + + call dword ptr [ebp+8] + + mov ecx, [12+ebp] /* load return type code */ + mov [ebp+8], ebx /* preserve %ebx */ +L(UW2): + // cfi_rel_offset(%ebx, 8) + + and ecx, X86_RET_TYPE_MASK + lea ebx, [L(store_table) + ecx * 8] + mov ecx, [ebp+16] /* load result address */ + jmp ebx + + ALIGN 8 +L(store_table): +E(L(store_table), X86_RET_FLOAT) + fstp DWORD PTR [ecx] + jmp L(e1) +E(L(store_table), X86_RET_DOUBLE) + fstp QWORD PTR [ecx] + jmp L(e1) +E(L(store_table), X86_RET_LDOUBLE) + fstp QWORD PTR [ecx] + jmp L(e1) +E(L(store_table), X86_RET_SINT8) + movsx eax, al + mov [ecx], eax + jmp L(e1) +E(L(store_table), X86_RET_SINT16) + movsx eax, ax + mov [ecx], eax + jmp L(e1) +E(L(store_table), X86_RET_UINT8) + movzx eax, al + mov [ecx], eax + jmp L(e1) +E(L(store_table), X86_RET_UINT16) + movzx eax, ax + mov [ecx], eax + jmp L(e1) +E(L(store_table), X86_RET_INT64) + mov [ecx+4], edx + /* fallthru */ +E(L(store_table), X86_RET_int 32) + mov [ecx], eax + /* fallthru */ +E(L(store_table), X86_RET_VOID) +L(e1): + mov ebx, [ebp+8] + mov esp, ebp + pop ebp +L(UW3): + // cfi_remember_state + // cfi_def_cfa(%esp, 4) + // cfi_restore(%ebx) + // cfi_restore(%ebp) + ret +L(UW4): + // cfi_restore_state + +E(L(store_table), X86_RET_STRUCTPOP) + jmp L(e1) +E(L(store_table), X86_RET_STRUCTARG) + jmp L(e1) +E(L(store_table), X86_RET_STRUCT_1B) + mov [ecx], al + jmp L(e1) +E(L(store_table), X86_RET_STRUCT_2B) + mov [ecx], ax + jmp L(e1) + + /* Fill out the table so that bad values are predictable. */ +E(L(store_table), X86_RET_UNUSED14) + int 3 +E(L(store_table), X86_RET_UNUSED15) + int 3 + +L(UW5): + // cfi_endproc +ENDF(@ffi_call_i386@8) + +/* The inner helper is declared as + + void ffi_closure_inner(struct closure_frame *frame, char *argp) + __attribute_((fastcall)) + + Thus the arguments are placed in + + ecx: frame + edx: argp +*/ + +/* Macros to help setting up the closure_data structure. */ + +#if HAVE_FASTCALL +# define closure_FS (40 + 4) +# define closure_CF 0 +#else +# define closure_FS (8 + 40 + 12) +# define closure_CF 8 +#endif + +FFI_CLOSURE_SAVE_REGS MACRO + mov [esp + closure_CF+16+R_EAX*4], eax + mov [esp + closure_CF+16+R_EDX*4], edx + mov [esp + closure_CF+16+R_ECX*4], ecx +ENDM + +FFI_CLOSURE_COPY_TRAMP_DATA MACRO + mov edx, [eax+FFI_TRAMPOLINE_SIZE] /* copy cif */ + mov ecx, [eax+FFI_TRAMPOLINE_SIZE+4] /* copy fun */ + mov eax, [eax+FFI_TRAMPOLINE_SIZE+8]; /* copy user_data */ + mov [esp+closure_CF+28], edx + mov [esp+closure_CF+32], ecx + mov [esp+closure_CF+36], eax +ENDM + +#if HAVE_FASTCALL +FFI_CLOSURE_PREP_CALL MACRO + mov ecx, esp /* load closure_data */ + lea edx, [esp+closure_FS+4] /* load incoming stack */ +ENDM +#else +FFI_CLOSURE_PREP_CALL MACRO + lea ecx, [esp+closure_CF] /* load closure_data */ + lea edx, [esp+closure_FS+4] /* load incoming stack */ + mov [esp], ecx + mov [esp+4], edx +ENDM +#endif + +FFI_CLOSURE_CALL_INNER MACRO UWN + call @ffi_closure_inner@8 +ENDM + +FFI_CLOSURE_MASK_AND_JUMP MACRO LABEL + and eax, X86_RET_TYPE_MASK + lea edx, [LABEL+eax*8] + mov eax, [esp+closure_CF] /* optimiztic load */ + jmp edx +ENDM + +ALIGN 16 +PUBLIC ffi_go_closure_EAX +ffi_go_closure_EAX PROC C +L(UW6): + // cfi_startproc + sub esp, closure_FS +L(UW7): + // cfi_def_cfa_offset(closure_FS + 4) + FFI_CLOSURE_SAVE_REGS + mov edx, [eax+4] /* copy cif */ + mov ecx, [eax +8] /* copy fun */ + mov [esp+closure_CF+28], edx + mov [esp+closure_CF+32], ecx + mov [esp+closure_CF+36], eax /* closure is user_data */ + jmp L(do_closure_i386) +L(UW8): + // cfi_endproc +ENDF(ffi_go_closure_EAX) + +ALIGN 16 +PUBLIC ffi_go_closure_ECX +ffi_go_closure_ECX PROC C +L(UW9): + // cfi_startproc + sub esp, closure_FS +L(UW10): + // cfi_def_cfa_offset(closure_FS + 4) + FFI_CLOSURE_SAVE_REGS + mov edx, [ecx+4] /* copy cif */ + mov eax, [ecx+8] /* copy fun */ + mov [esp+closure_CF+28], edx + mov [esp+closure_CF+32], eax + mov [esp+closure_CF+36], ecx /* closure is user_data */ + jmp L(do_closure_i386) +L(UW11): + // cfi_endproc +ENDF(ffi_go_closure_ECX) + +/* The closure entry points are reached from the ffi_closure trampoline. + On entry, %eax contains the address of the ffi_closure. */ + +ALIGN 16 +PUBLIC ffi_closure_i386 +ffi_closure_i386 PROC C +L(UW12): + // cfi_startproc + sub esp, closure_FS +L(UW13): + // cfi_def_cfa_offset(closure_FS + 4) + + FFI_CLOSURE_SAVE_REGS + FFI_CLOSURE_COPY_TRAMP_DATA + + /* Entry point from preceeding Go closures. */ +L(do_closure_i386):: + + FFI_CLOSURE_PREP_CALL + FFI_CLOSURE_CALL_INNER(14) + FFI_CLOSURE_MASK_AND_JUMP L(C1(load_table,2)) + + ALIGN 8 +L(load_table2): +E(L(load_table2), X86_RET_FLOAT) + fld dword ptr [esp+closure_CF] + jmp L(e2) +E(L(load_table2), X86_RET_DOUBLE) + fld qword ptr [esp+closure_CF] + jmp L(e2) +E(L(load_table2), X86_RET_LDOUBLE) + fld qword ptr [esp+closure_CF] + jmp L(e2) +E(L(load_table2), X86_RET_SINT8) + movsx eax, al + jmp L(e2) +E(L(load_table2), X86_RET_SINT16) + movsx eax, ax + jmp L(e2) +E(L(load_table2), X86_RET_UINT8) + movzx eax, al + jmp L(e2) +E(L(load_table2), X86_RET_UINT16) + movzx eax, ax + jmp L(e2) +E(L(load_table2), X86_RET_INT64) + mov edx, [esp+closure_CF+4] + jmp L(e2) +E(L(load_table2), X86_RET_INT32) + nop + /* fallthru */ +E(L(load_table2), X86_RET_VOID) +L(e2): + add esp, closure_FS +L(UW16): + // cfi_adjust_cfa_offset(-closure_FS) + ret +L(UW17): + // cfi_adjust_cfa_offset(closure_FS) +E(L(load_table2), X86_RET_STRUCTPOP) + add esp, closure_FS +L(UW18): + // cfi_adjust_cfa_offset(-closure_FS) + ret 4 +L(UW19): + // cfi_adjust_cfa_offset(closure_FS) +E(L(load_table2), X86_RET_STRUCTARG) + jmp L(e2) +E(L(load_table2), X86_RET_STRUCT_1B) + movzx eax, al + jmp L(e2) +E(L(load_table2), X86_RET_STRUCT_2B) + movzx eax, ax + jmp L(e2) + + /* Fill out the table so that bad values are predictable. */ +E(L(load_table2), X86_RET_UNUSED14) + int 3 +E(L(load_table2), X86_RET_UNUSED15) + int 3 + +L(UW20): + // cfi_endproc +ENDF(ffi_closure_i386) + +ALIGN 16 +PUBLIC ffi_go_closure_STDCALL +ffi_go_closure_STDCALL PROC C +L(UW21): + // cfi_startproc + sub esp, closure_FS +L(UW22): + // cfi_def_cfa_offset(closure_FS + 4) + FFI_CLOSURE_SAVE_REGS + mov edx, [ecx+4] /* copy cif */ + mov eax, [ecx+8] /* copy fun */ + mov [esp+closure_CF+28], edx + mov [esp+closure_CF+32], eax + mov [esp+closure_CF+36], ecx /* closure is user_data */ + jmp L(do_closure_STDCALL) +L(UW23): + // cfi_endproc +ENDF(ffi_go_closure_STDCALL) + +/* For REGISTER, we have no available parameter registers, and so we + enter here having pushed the closure onto the stack. */ + +ALIGN 16 +PUBLIC ffi_closure_REGISTER +ffi_closure_REGISTER PROC C +L(UW24): + // cfi_startproc + // cfi_def_cfa(%esp, 8) + // cfi_offset(%eip, -8) + sub esp, closure_FS-4 +L(UW25): + // cfi_def_cfa_offset(closure_FS + 4) + FFI_CLOSURE_SAVE_REGS + mov ecx, [esp+closure_FS-4] /* load retaddr */ + mov eax, [esp+closure_FS] /* load closure */ + mov [esp+closure_FS], ecx /* move retaddr */ + jmp L(do_closure_REGISTER) +L(UW26): + // cfi_endproc +ENDF(ffi_closure_REGISTER) + +/* For STDCALL (and others), we need to pop N bytes of arguments off + the stack following the closure. The amount needing to be popped + is returned to us from ffi_closure_inner. */ + +ALIGN 16 +PUBLIC ffi_closure_STDCALL +ffi_closure_STDCALL PROC C +L(UW27): + // cfi_startproc + sub esp, closure_FS +L(UW28): + // cfi_def_cfa_offset(closure_FS + 4) + + FFI_CLOSURE_SAVE_REGS + + /* Entry point from ffi_closure_REGISTER. */ +L(do_closure_REGISTER):: + + FFI_CLOSURE_COPY_TRAMP_DATA + + /* Entry point from preceeding Go closure. */ +L(do_closure_STDCALL):: + + FFI_CLOSURE_PREP_CALL + FFI_CLOSURE_CALL_INNER(29) + + mov ecx, eax + shr ecx, X86_RET_POP_SHIFT /* isolate pop count */ + lea ecx, [esp+closure_FS+ecx] /* compute popped esp */ + mov edx, [esp+closure_FS] /* move return address */ + mov [ecx], edx + + /* From this point on, the value of %esp upon return is %ecx+4, + and we've copied the return address to %ecx to make return easy. + There's no point in representing this in the unwind info, as + there is always a window between the mov and the ret which + will be wrong from one point of view or another. */ + + FFI_CLOSURE_MASK_AND_JUMP L(C1(load_table,3)) + + ALIGN 8 +L(load_table3): +E(L(load_table3), X86_RET_FLOAT) + fld DWORD PTR [esp+closure_CF] + mov esp, ecx + ret +E(L(load_table3), X86_RET_DOUBLE) + fld QWORD PTR [esp+closure_CF] + mov esp, ecx + ret +E(L(load_table3), X86_RET_LDOUBLE) + fld QWORD PTR [esp+closure_CF] + mov esp, ecx + ret +E(L(load_table3), X86_RET_SINT8) + movsx eax, al + mov esp, ecx + ret +E(L(load_table3), X86_RET_SINT16) + movsx eax, ax + mov esp, ecx + ret +E(L(load_table3), X86_RET_UINT8) + movzx eax, al + mov esp, ecx + ret +E(L(load_table3), X86_RET_UINT16) + movzx eax, ax + mov esp, ecx + ret +E(L(load_table3), X86_RET_INT64) + mov edx, [esp+closure_CF+4] + mov esp, ecx + ret +E(L(load_table3), X86_RET_int 32) + mov esp, ecx + ret +E(L(load_table3), X86_RET_VOID) + mov esp, ecx + ret +E(L(load_table3), X86_RET_STRUCTPOP) + mov esp, ecx + ret +E(L(load_table3), X86_RET_STRUCTARG) + mov esp, ecx + ret +E(L(load_table3), X86_RET_STRUCT_1B) + movzx eax, al + mov esp, ecx + ret +E(L(load_table3), X86_RET_STRUCT_2B) + movzx eax, ax + mov esp, ecx + ret + + /* Fill out the table so that bad values are predictable. */ +E(L(load_table3), X86_RET_UNUSED14) + int 3 +E(L(load_table3), X86_RET_UNUSED15) + int 3 + +L(UW31): + // cfi_endproc +ENDF(ffi_closure_STDCALL) + +#if !FFI_NO_RAW_API + +#define raw_closure_S_FS (16+16+12) + +ALIGN 16 +PUBLIC ffi_closure_raw_SYSV +ffi_closure_raw_SYSV PROC C +L(UW32): + // cfi_startproc + sub esp, raw_closure_S_FS +L(UW33): + // cfi_def_cfa_offset(raw_closure_S_FS + 4) + mov [esp+raw_closure_S_FS-4], ebx +L(UW34): + // cfi_rel_offset(%ebx, raw_closure_S_FS-4) + + mov edx, [eax+FFI_TRAMPOLINE_SIZE+8] /* load cl->user_data */ + mov [esp+12], edx + lea edx, [esp+raw_closure_S_FS+4] /* load raw_args */ + mov [esp+8], edx + lea edx, [esp+16] /* load &res */ + mov [esp+4], edx + mov ebx, [eax+FFI_TRAMPOLINE_SIZE] /* load cl->cif */ + mov [esp], ebx + call DWORD PTR [eax+FFI_TRAMPOLINE_SIZE+4] /* call cl->fun */ + + mov eax, [ebx+20] /* load cif->flags */ + and eax, X86_RET_TYPE_MASK +// #ifdef __PIC__ +// call __x86.get_pc_thunk.bx +// L(pc4): +// lea ecx, L(load_table4)-L(pc4)(%ebx, %eax, 8), %ecx +// #else + lea ecx, [L(load_table4)+eax+8] +// #endif + mov ebx, [esp+raw_closure_S_FS-4] +L(UW35): + // cfi_restore(%ebx) + mov eax, [esp+16] /* Optimistic load */ + jmp dword ptr [ecx] + + ALIGN 8 +L(load_table4): +E(L(load_table4), X86_RET_FLOAT) + fld DWORD PTR [esp +16] + jmp L(e4) +E(L(load_table4), X86_RET_DOUBLE) + fld QWORD PTR [esp +16] + jmp L(e4) +E(L(load_table4), X86_RET_LDOUBLE) + fld QWORD PTR [esp +16] + jmp L(e4) +E(L(load_table4), X86_RET_SINT8) + movsx eax, al + jmp L(e4) +E(L(load_table4), X86_RET_SINT16) + movsx eax, ax + jmp L(e4) +E(L(load_table4), X86_RET_UINT8) + movzx eax, al + jmp L(e4) +E(L(load_table4), X86_RET_UINT16) + movzx eax, ax + jmp L(e4) +E(L(load_table4), X86_RET_INT64) + mov edx, [esp+16+4] + jmp L(e4) +E(L(load_table4), X86_RET_int 32) + nop + /* fallthru */ +E(L(load_table4), X86_RET_VOID) +L(e4): + add esp, raw_closure_S_FS +L(UW36): + // cfi_adjust_cfa_offset(-raw_closure_S_FS) + ret +L(UW37): + // cfi_adjust_cfa_offset(raw_closure_S_FS) +E(L(load_table4), X86_RET_STRUCTPOP) + add esp, raw_closure_S_FS +L(UW38): + // cfi_adjust_cfa_offset(-raw_closure_S_FS) + ret 4 +L(UW39): + // cfi_adjust_cfa_offset(raw_closure_S_FS) +E(L(load_table4), X86_RET_STRUCTARG) + jmp L(e4) +E(L(load_table4), X86_RET_STRUCT_1B) + movzx eax, al + jmp L(e4) +E(L(load_table4), X86_RET_STRUCT_2B) + movzx eax, ax + jmp L(e4) + + /* Fill out the table so that bad values are predictable. */ +E(L(load_table4), X86_RET_UNUSED14) + int 3 +E(L(load_table4), X86_RET_UNUSED15) + int 3 + +L(UW40): + // cfi_endproc +ENDF(ffi_closure_raw_SYSV) + +#define raw_closure_T_FS (16+16+8) + +ALIGN 16 +PUBLIC ffi_closure_raw_THISCALL +ffi_closure_raw_THISCALL PROC C +L(UW41): + // cfi_startproc + /* Rearrange the stack such that %ecx is the first argument. + This means moving the return address. */ + pop edx +L(UW42): + // cfi_def_cfa_offset(0) + // cfi_register(%eip, %edx) + push ecx +L(UW43): + // cfi_adjust_cfa_offset(4) + push edx +L(UW44): + // cfi_adjust_cfa_offset(4) + // cfi_rel_offset(%eip, 0) + sub esp, raw_closure_T_FS +L(UW45): + // cfi_adjust_cfa_offset(raw_closure_T_FS) + mov [esp+raw_closure_T_FS-4], ebx +L(UW46): + // cfi_rel_offset(%ebx, raw_closure_T_FS-4) + + mov edx, [eax+FFI_TRAMPOLINE_SIZE+8] /* load cl->user_data */ + mov [esp+12], edx + lea edx, [esp+raw_closure_T_FS+4] /* load raw_args */ + mov [esp+8], edx + lea edx, [esp+16] /* load &res */ + mov [esp+4], edx + mov ebx, [eax+FFI_TRAMPOLINE_SIZE] /* load cl->cif */ + mov [esp], ebx + call DWORD PTR [eax+FFI_TRAMPOLINE_SIZE+4] /* call cl->fun */ + + mov eax, [ebx+20] /* load cif->flags */ + and eax, X86_RET_TYPE_MASK +// #ifdef __PIC__ +// call __x86.get_pc_thunk.bx +// L(pc5): +// leal L(load_table5)-L(pc5)(%ebx, %eax, 8), %ecx +// #else + lea ecx, [L(load_table5)+eax*8] +//#endif + mov ebx, [esp+raw_closure_T_FS-4] +L(UW47): + // cfi_restore(%ebx) + mov eax, [esp+16] /* Optimistic load */ + jmp DWORD PTR [ecx] + + AlIGN 4 +L(load_table5): +E(L(load_table5), X86_RET_FLOAT) + fld DWORD PTR [esp +16] + jmp L(e5) +E(L(load_table5), X86_RET_DOUBLE) + fld QWORD PTR [esp +16] + jmp L(e5) +E(L(load_table5), X86_RET_LDOUBLE) + fld QWORD PTR [esp+16] + jmp L(e5) +E(L(load_table5), X86_RET_SINT8) + movsx eax, al + jmp L(e5) +E(L(load_table5), X86_RET_SINT16) + movsx eax, ax + jmp L(e5) +E(L(load_table5), X86_RET_UINT8) + movzx eax, al + jmp L(e5) +E(L(load_table5), X86_RET_UINT16) + movzx eax, ax + jmp L(e5) +E(L(load_table5), X86_RET_INT64) + mov edx, [esp+16+4] + jmp L(e5) +E(L(load_table5), X86_RET_int 32) + nop + /* fallthru */ +E(L(load_table5), X86_RET_VOID) +L(e5): + add esp, raw_closure_T_FS +L(UW48): + // cfi_adjust_cfa_offset(-raw_closure_T_FS) + /* Remove the extra %ecx argument we pushed. */ + ret 4 +L(UW49): + // cfi_adjust_cfa_offset(raw_closure_T_FS) +E(L(load_table5), X86_RET_STRUCTPOP) + add esp, raw_closure_T_FS +L(UW50): + // cfi_adjust_cfa_offset(-raw_closure_T_FS) + ret 8 +L(UW51): + // cfi_adjust_cfa_offset(raw_closure_T_FS) +E(L(load_table5), X86_RET_STRUCTARG) + jmp L(e5) +E(L(load_table5), X86_RET_STRUCT_1B) + movzx eax, al + jmp L(e5) +E(L(load_table5), X86_RET_STRUCT_2B) + movzx eax, ax + jmp L(e5) + + /* Fill out the table so that bad values are predictable. */ +E(L(load_table5), X86_RET_UNUSED14) + int 3 +E(L(load_table5), X86_RET_UNUSED15) + int 3 + +L(UW52): + // cfi_endproc +ENDF(ffi_closure_raw_THISCALL) + +#endif /* !FFI_NO_RAW_API */ + +#ifdef X86_DARWIN +# define COMDAT(X) \ + .section __TEXT,__text,coalesced,pure_instructions; \ + .weak_definition X; \ + FFI_HIDDEN(X) +#elif defined __ELF__ && !(defined(__sun__) && defined(__svr4__)) +# define COMDAT(X) \ + .section .text.X,"axG",@progbits,X,comdat; \ + PUBLIC X; \ + FFI_HIDDEN(X) +#else +# define COMDAT(X) +#endif + +// #if defined(__PIC__) +// COMDAT(C(__x86.get_pc_thunk.bx)) +// C(__x86.get_pc_thunk.bx): +// movl (%esp), %ebx +// ret +// ENDF(C(__x86.get_pc_thunk.bx)) +// # if defined X86_DARWIN || defined HAVE_HIDDEN_VISIBILITY_ATTRIBUTE +// COMDAT(C(__x86.get_pc_thunk.dx)) +// C(__x86.get_pc_thunk.dx): +// movl (%esp), %edx +// ret +// ENDF(C(__x86.get_pc_thunk.dx)) +// #endif /* DARWIN || HIDDEN */ +// #endif /* __PIC__ */ + +#if 0 +/* Sadly, OSX cctools-as doesn't understand .cfi directives at all. */ + +#ifdef __APPLE__ +.section __TEXT,__eh_frame,coalesced,no_toc+strip_static_syms+live_support +EHFrame0: +#elif defined(X86_WIN32) +.section .eh_frame,"r" +#elif defined(HAVE_AS_X86_64_UNWIND_SECTION_TYPE) +.section .eh_frame,EH_FRAME_FLAGS,@unwind +#else +.section .eh_frame,EH_FRAME_FLAGS,@progbits +#endif + +#ifdef HAVE_AS_X86_PCREL +# define PCREL(X) X - . +#else +# define PCREL(X) X@rel +#endif + +/* Simplify advancing between labels. Assume DW_CFA_advance_loc1 fits. */ +#define ADV(N, P) .byte 2, L(N)-L(P) + + .balign 4 +L(CIE): + .set L(set0),L(ECIE)-L(SCIE) + .long L(set0) /* CIE Length */ +L(SCIE): + .long 0 /* CIE Identifier Tag */ + .byte 1 /* CIE Version */ + .ascii "zR\0" /* CIE Augmentation */ + .byte 1 /* CIE Code Alignment Factor */ + .byte 0x7c /* CIE Data Alignment Factor */ + .byte 0x8 /* CIE RA Column */ + .byte 1 /* Augmentation size */ + .byte 0x1b /* FDE Encoding (pcrel sdata4) */ + .byte 0xc, 4, 4 /* DW_CFA_def_cfa, %esp offset 4 */ + .byte 0x80+8, 1 /* DW_CFA_offset, %eip offset 1*-4 */ + .balign 4 +L(ECIE): + + .set L(set1),L(EFDE1)-L(SFDE1) + .long L(set1) /* FDE Length */ +L(SFDE1): + .long L(SFDE1)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW0)) /* Initial location */ + .long L(UW5)-L(UW0) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW1, UW0) + .byte 0xc, 5, 8 /* DW_CFA_def_cfa, %ebp 8 */ + .byte 0x80+5, 2 /* DW_CFA_offset, %ebp 2*-4 */ + ADV(UW2, UW1) + .byte 0x80+3, 0 /* DW_CFA_offset, %ebx 0*-4 */ + ADV(UW3, UW2) + .byte 0xa /* DW_CFA_remember_state */ + .byte 0xc, 4, 4 /* DW_CFA_def_cfa, %esp 4 */ + .byte 0xc0+3 /* DW_CFA_restore, %ebx */ + .byte 0xc0+5 /* DW_CFA_restore, %ebp */ + ADV(UW4, UW3) + .byte 0xb /* DW_CFA_restore_state */ + .balign 4 +L(EFDE1): + + .set L(set2),L(EFDE2)-L(SFDE2) + .long L(set2) /* FDE Length */ +L(SFDE2): + .long L(SFDE2)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW6)) /* Initial location */ + .long L(UW8)-L(UW6) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW7, UW6) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE2): + + .set L(set3),L(EFDE3)-L(SFDE3) + .long L(set3) /* FDE Length */ +L(SFDE3): + .long L(SFDE3)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW9)) /* Initial location */ + .long L(UW11)-L(UW9) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW10, UW9) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE3): + + .set L(set4),L(EFDE4)-L(SFDE4) + .long L(set4) /* FDE Length */ +L(SFDE4): + .long L(SFDE4)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW12)) /* Initial location */ + .long L(UW20)-L(UW12) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW13, UW12) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ +#ifdef FFI_CLOSURE_CALL_INNER_SAVE_EBX + ADV(UW14, UW13) + .byte 0x80+3, (40-(closure_FS+4))/-4 /* DW_CFA_offset %ebx */ + ADV(UW15, UW14) + .byte 0xc0+3 /* DW_CFA_restore %ebx */ + ADV(UW16, UW15) +#else + ADV(UW16, UW13) +#endif + .byte 0xe, 4 /* DW_CFA_def_cfa_offset */ + ADV(UW17, UW16) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + ADV(UW18, UW17) + .byte 0xe, 4 /* DW_CFA_def_cfa_offset */ + ADV(UW19, UW18) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE4): + + .set L(set5),L(EFDE5)-L(SFDE5) + .long L(set5) /* FDE Length */ +L(SFDE5): + .long L(SFDE5)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW21)) /* Initial location */ + .long L(UW23)-L(UW21) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW22, UW21) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE5): + + .set L(set6),L(EFDE6)-L(SFDE6) + .long L(set6) /* FDE Length */ +L(SFDE6): + .long L(SFDE6)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW24)) /* Initial location */ + .long L(UW26)-L(UW24) /* Address range */ + .byte 0 /* Augmentation size */ + .byte 0xe, 8 /* DW_CFA_def_cfa_offset */ + .byte 0x80+8, 2 /* DW_CFA_offset %eip, 2*-4 */ + ADV(UW25, UW24) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE6): + + .set L(set7),L(EFDE7)-L(SFDE7) + .long L(set7) /* FDE Length */ +L(SFDE7): + .long L(SFDE7)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW27)) /* Initial location */ + .long L(UW31)-L(UW27) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW28, UW27) + .byte 0xe, closure_FS+4 /* DW_CFA_def_cfa_offset */ +#ifdef FFI_CLOSURE_CALL_INNER_SAVE_EBX + ADV(UW29, UW28) + .byte 0x80+3, (40-(closure_FS+4))/-4 /* DW_CFA_offset %ebx */ + ADV(UW30, UW29) + .byte 0xc0+3 /* DW_CFA_restore %ebx */ +#endif + .balign 4 +L(EFDE7): + +#if !FFI_NO_RAW_API + .set L(set8),L(EFDE8)-L(SFDE8) + .long L(set8) /* FDE Length */ +L(SFDE8): + .long L(SFDE8)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW32)) /* Initial location */ + .long L(UW40)-L(UW32) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW33, UW32) + .byte 0xe, raw_closure_S_FS+4 /* DW_CFA_def_cfa_offset */ + ADV(UW34, UW33) + .byte 0x80+3, 2 /* DW_CFA_offset %ebx 2*-4 */ + ADV(UW35, UW34) + .byte 0xc0+3 /* DW_CFA_restore %ebx */ + ADV(UW36, UW35) + .byte 0xe, 4 /* DW_CFA_def_cfa_offset */ + ADV(UW37, UW36) + .byte 0xe, raw_closure_S_FS+4 /* DW_CFA_def_cfa_offset */ + ADV(UW38, UW37) + .byte 0xe, 4 /* DW_CFA_def_cfa_offset */ + ADV(UW39, UW38) + .byte 0xe, raw_closure_S_FS+4 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE8): + + .set L(set9),L(EFDE9)-L(SFDE9) + .long L(set9) /* FDE Length */ +L(SFDE9): + .long L(SFDE9)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW41)) /* Initial location */ + .long L(UW52)-L(UW41) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW42, UW41) + .byte 0xe, 0 /* DW_CFA_def_cfa_offset */ + .byte 0x9, 8, 2 /* DW_CFA_register %eip, %edx */ + ADV(UW43, UW42) + .byte 0xe, 4 /* DW_CFA_def_cfa_offset */ + ADV(UW44, UW43) + .byte 0xe, 8 /* DW_CFA_def_cfa_offset */ + .byte 0x80+8, 2 /* DW_CFA_offset %eip 2*-4 */ + ADV(UW45, UW44) + .byte 0xe, raw_closure_T_FS+8 /* DW_CFA_def_cfa_offset */ + ADV(UW46, UW45) + .byte 0x80+3, 3 /* DW_CFA_offset %ebx 3*-4 */ + ADV(UW47, UW46) + .byte 0xc0+3 /* DW_CFA_restore %ebx */ + ADV(UW48, UW47) + .byte 0xe, 8 /* DW_CFA_def_cfa_offset */ + ADV(UW49, UW48) + .byte 0xe, raw_closure_T_FS+8 /* DW_CFA_def_cfa_offset */ + ADV(UW50, UW49) + .byte 0xe, 8 /* DW_CFA_def_cfa_offset */ + ADV(UW51, UW50) + .byte 0xe, raw_closure_T_FS+8 /* DW_CFA_def_cfa_offset */ + .balign 4 +L(EFDE9): +#endif /* !FFI_NO_RAW_API */ + +#ifdef _WIN32 + .def @feat.00; + .scl 3; + .type 0; + .endef + PUBLIC @feat.00 +@feat.00 = 1 +#endif + +#endif /* ifndef _MSC_VER */ +#endif /* ifndef __x86_64__ */ + +#if defined __ELF__ && defined __linux__ + .section .note.GNU-stack,"",@progbits +#endif +#endif + +END \ No newline at end of file diff --git a/native/libffi/src/x86/unix64.S b/native/libffi/src/x86/unix64.S index dcd6bc71ea..89d7db11fe 100644 --- a/native/libffi/src/x86/unix64.S +++ b/native/libffi/src/x86/unix64.S @@ -30,8 +30,24 @@ #define LIBFFI_ASM #include #include +#include "internal64.h" +#include "asmnames.h" -.text + .text + +/* This macro allows the safe creation of jump tables without an + actual table. The entry points into the table are all 8 bytes. + The use of ORG asserts that we're at the correct location. */ +/* ??? The clang assembler doesn't handle .org with symbolic expressions. */ +#if defined(__clang__) || defined(__APPLE__) || (defined (__sun__) && defined(__svr4__)) +# define E(BASE, X) .balign 8 +#else +# ifdef __CET__ +# define E(BASE, X) .balign 8; .org BASE + X * 16 +# else +# define E(BASE, X) .balign 8; .org BASE + X * 8 +# endif +#endif /* ffi_call_unix64 (void *args, unsigned long bytes, unsigned flags, void *raddr, void (*fnaddr)(void)); @@ -40,12 +56,13 @@ for this function. This has been allocated by ffi_call. We also deallocate some of the stack that has been alloca'd. */ - .align 2 - .globl ffi_call_unix64 - .type ffi_call_unix64,@function + .balign 8 + .globl C(ffi_call_unix64) + FFI_HIDDEN(C(ffi_call_unix64)) -ffi_call_unix64: -.LUW0: +C(ffi_call_unix64): +L(UW0): + _CET_ENDBR movq (%rsp), %r10 /* Load return address. */ leaq (%rdi, %rsi), %rax /* Find local stack base. */ movq %rdx, (%rax) /* Save flags. */ @@ -53,24 +70,36 @@ ffi_call_unix64: movq %rbp, 16(%rax) /* Save old frame pointer. */ movq %r10, 24(%rax) /* Relocate return address. */ movq %rax, %rbp /* Finalize local stack frame. */ -.LUW1: + + /* New stack frame based off rbp. This is a itty bit of unwind + trickery in that the CFA *has* changed. There is no easy way + to describe it correctly on entry to the function. Fortunately, + it doesn't matter too much since at all points we can correctly + unwind back to ffi_call. Note that the location to which we + moved the return address is (the new) CFA-8, so from the + perspective of the unwind info, it hasn't moved. */ +L(UW1): + /* cfi_def_cfa(%rbp, 32) */ + /* cfi_rel_offset(%rbp, 16) */ + movq %rdi, %r10 /* Save a copy of the register area. */ movq %r8, %r11 /* Save a copy of the target fn. */ - movl %r9d, %eax /* Set number of SSE registers. */ /* Load up all argument registers. */ movq (%r10), %rdi - movq 8(%r10), %rsi - movq 16(%r10), %rdx - movq 24(%r10), %rcx - movq 32(%r10), %r8 - movq 40(%r10), %r9 + movq 0x08(%r10), %rsi + movq 0x10(%r10), %rdx + movq 0x18(%r10), %rcx + movq 0x20(%r10), %r8 + movq 0x28(%r10), %r9 + movl 0xb0(%r10), %eax /* Set number of SSE registers. */ testl %eax, %eax - jnz .Lload_sse -.Lret_from_load_sse: + jnz L(load_sse) +L(ret_from_load_sse): - /* Deallocate the reg arg area. */ - leaq 176(%r10), %rsp + /* Deallocate the reg arg area, except for r10, then load via pop. */ + leaq 0xb8(%r10), %rsp + popq %r10 /* Call the user function. */ call *%r11 @@ -81,352 +110,512 @@ ffi_call_unix64: movq 0(%rbp), %rcx /* Reload flags. */ movq 8(%rbp), %rdi /* Reload raddr. */ movq 16(%rbp), %rbp /* Reload old frame pointer. */ -.LUW2: +L(UW2): + /* cfi_remember_state */ + /* cfi_def_cfa(%rsp, 8) */ + /* cfi_restore(%rbp) */ /* The first byte of the flags contains the FFI_TYPE. */ + cmpb $UNIX64_RET_LAST, %cl movzbl %cl, %r10d - leaq .Lstore_table(%rip), %r11 - movslq (%r11, %r10, 4), %r10 - addq %r11, %r10 + leaq L(store_table)(%rip), %r11 + ja L(sa) +#ifdef __CET__ + /* NB: Originally, each slot is 8 byte. 4 bytes of ENDBR64 + + 4 bytes NOP padding double slot size to 16 bytes. */ + addl %r10d, %r10d +#endif + leaq (%r11, %r10, 8), %r10 + + /* Prep for the structure cases: scratch area in redzone. */ + leaq -20(%rsp), %rsi jmp *%r10 -.Lstore_table: - .long .Lst_void-.Lstore_table /* FFI_TYPE_VOID */ - .long .Lst_sint32-.Lstore_table /* FFI_TYPE_INT */ - .long .Lst_float-.Lstore_table /* FFI_TYPE_FLOAT */ - .long .Lst_double-.Lstore_table /* FFI_TYPE_DOUBLE */ - .long .Lst_ldouble-.Lstore_table /* FFI_TYPE_LONGDOUBLE */ - .long .Lst_uint8-.Lstore_table /* FFI_TYPE_UINT8 */ - .long .Lst_sint8-.Lstore_table /* FFI_TYPE_SINT8 */ - .long .Lst_uint16-.Lstore_table /* FFI_TYPE_UINT16 */ - .long .Lst_sint16-.Lstore_table /* FFI_TYPE_SINT16 */ - .long .Lst_uint32-.Lstore_table /* FFI_TYPE_UINT32 */ - .long .Lst_sint32-.Lstore_table /* FFI_TYPE_SINT32 */ - .long .Lst_int64-.Lstore_table /* FFI_TYPE_UINT64 */ - .long .Lst_int64-.Lstore_table /* FFI_TYPE_SINT64 */ - .long .Lst_struct-.Lstore_table /* FFI_TYPE_STRUCT */ - .long .Lst_int64-.Lstore_table /* FFI_TYPE_POINTER */ - - .align 2 -.Lst_void: + .balign 8 +L(store_table): +E(L(store_table), UNIX64_RET_VOID) + _CET_ENDBR ret - .align 2 - -.Lst_uint8: - movzbq %al, %rax +E(L(store_table), UNIX64_RET_UINT8) + _CET_ENDBR + movzbl %al, %eax movq %rax, (%rdi) ret - .align 2 -.Lst_sint8: - movsbq %al, %rax +E(L(store_table), UNIX64_RET_UINT16) + _CET_ENDBR + movzwl %ax, %eax movq %rax, (%rdi) ret - .align 2 -.Lst_uint16: - movzwq %ax, %rax +E(L(store_table), UNIX64_RET_UINT32) + _CET_ENDBR + movl %eax, %eax movq %rax, (%rdi) - .align 2 -.Lst_sint16: - movswq %ax, %rax + ret +E(L(store_table), UNIX64_RET_SINT8) + _CET_ENDBR + movsbq %al, %rax movq %rax, (%rdi) ret - .align 2 -.Lst_uint32: - movl %eax, %eax +E(L(store_table), UNIX64_RET_SINT16) + _CET_ENDBR + movswq %ax, %rax movq %rax, (%rdi) - .align 2 -.Lst_sint32: + ret +E(L(store_table), UNIX64_RET_SINT32) + _CET_ENDBR cltq movq %rax, (%rdi) ret - .align 2 -.Lst_int64: +E(L(store_table), UNIX64_RET_INT64) + _CET_ENDBR movq %rax, (%rdi) ret - - .align 2 -.Lst_float: - movss %xmm0, (%rdi) +E(L(store_table), UNIX64_RET_XMM32) + _CET_ENDBR + movd %xmm0, (%rdi) ret - .align 2 -.Lst_double: - movsd %xmm0, (%rdi) +E(L(store_table), UNIX64_RET_XMM64) + _CET_ENDBR + movq %xmm0, (%rdi) ret -.Lst_ldouble: +E(L(store_table), UNIX64_RET_X87) + _CET_ENDBR fstpt (%rdi) ret - - .align 2 -.Lst_struct: - leaq -20(%rsp), %rsi /* Scratch area in redzone. */ - - /* We have to locate the values now, and since we don't want to - write too much data into the user's return value, we spill the - value to a 16 byte scratch area first. Bits 8, 9, and 10 - control where the values are located. Only one of the three - bits will be set; see ffi_prep_cif_machdep for the pattern. */ - movd %xmm0, %r10 - movd %xmm1, %r11 - testl $0x100, %ecx - cmovnz %rax, %rdx - cmovnz %r10, %rax - testl $0x200, %ecx - cmovnz %r10, %rdx - testl $0x400, %ecx - cmovnz %r10, %rax - cmovnz %r11, %rdx - movq %rax, (%rsi) +E(L(store_table), UNIX64_RET_X87_2) + _CET_ENDBR + fstpt (%rdi) + fstpt 16(%rdi) + ret +E(L(store_table), UNIX64_RET_ST_XMM0_RAX) + _CET_ENDBR + movq %rax, 8(%rsi) + jmp L(s3) +E(L(store_table), UNIX64_RET_ST_RAX_XMM0) + _CET_ENDBR + movq %xmm0, 8(%rsi) + jmp L(s2) +E(L(store_table), UNIX64_RET_ST_XMM0_XMM1) + _CET_ENDBR + movq %xmm1, 8(%rsi) + jmp L(s3) +E(L(store_table), UNIX64_RET_ST_RAX_RDX) + _CET_ENDBR movq %rdx, 8(%rsi) - - /* Bits 12-31 contain the true size of the structure. Copy from - the scratch area to the true destination. */ - shrl $12, %ecx +L(s2): + movq %rax, (%rsi) + shrl $UNIX64_SIZE_SHIFT, %ecx rep movsb ret + .balign 8 +L(s3): + movq %xmm0, (%rsi) + shrl $UNIX64_SIZE_SHIFT, %ecx + rep movsb + ret + +L(sa): call PLT(C(abort)) /* Many times we can avoid loading any SSE registers at all. It's not worth an indirect jump to load the exact set of SSE registers needed; zero or all is a good compromise. */ - .align 2 -.LUW3: -.Lload_sse: - movdqa 48(%r10), %xmm0 - movdqa 64(%r10), %xmm1 - movdqa 80(%r10), %xmm2 - movdqa 96(%r10), %xmm3 - movdqa 112(%r10), %xmm4 - movdqa 128(%r10), %xmm5 - movdqa 144(%r10), %xmm6 - movdqa 160(%r10), %xmm7 - jmp .Lret_from_load_sse - -.LUW4: - .size ffi_call_unix64,.-ffi_call_unix64 - - .align 2 - .globl ffi_closure_unix64 - .type ffi_closure_unix64,@function - -ffi_closure_unix64: -.LUW5: - /* The carry flag is set by the trampoline iff SSE registers - are used. Don't clobber it before the branch instruction. */ - leaq -200(%rsp), %rsp -.LUW6: - movq %rdi, (%rsp) - movq %rsi, 8(%rsp) - movq %rdx, 16(%rsp) - movq %rcx, 24(%rsp) - movq %r8, 32(%rsp) - movq %r9, 40(%rsp) - jc .Lsave_sse -.Lret_from_save_sse: - - movq %r10, %rdi - leaq 176(%rsp), %rsi - movq %rsp, %rdx - leaq 208(%rsp), %rcx - call ffi_closure_unix64_inner@PLT + .balign 2 +L(UW3): + /* cfi_restore_state */ +L(load_sse): + movdqa 0x30(%r10), %xmm0 + movdqa 0x40(%r10), %xmm1 + movdqa 0x50(%r10), %xmm2 + movdqa 0x60(%r10), %xmm3 + movdqa 0x70(%r10), %xmm4 + movdqa 0x80(%r10), %xmm5 + movdqa 0x90(%r10), %xmm6 + movdqa 0xa0(%r10), %xmm7 + jmp L(ret_from_load_sse) + +L(UW4): +ENDF(C(ffi_call_unix64)) + +/* 6 general registers, 8 vector registers, + 32 bytes of rvalue, 8 bytes of alignment. */ +#define ffi_closure_OFS_G 0 +#define ffi_closure_OFS_V (6*8) +#define ffi_closure_OFS_RVALUE (ffi_closure_OFS_V + 8*16) +#define ffi_closure_FS (ffi_closure_OFS_RVALUE + 32 + 8) + +/* The location of rvalue within the red zone after deallocating the frame. */ +#define ffi_closure_RED_RVALUE (ffi_closure_OFS_RVALUE - ffi_closure_FS) + + .balign 2 + .globl C(ffi_closure_unix64_sse) + FFI_HIDDEN(C(ffi_closure_unix64_sse)) + +C(ffi_closure_unix64_sse): +L(UW5): + _CET_ENDBR + subq $ffi_closure_FS, %rsp +L(UW6): + /* cfi_adjust_cfa_offset(ffi_closure_FS) */ + + movdqa %xmm0, ffi_closure_OFS_V+0x00(%rsp) + movdqa %xmm1, ffi_closure_OFS_V+0x10(%rsp) + movdqa %xmm2, ffi_closure_OFS_V+0x20(%rsp) + movdqa %xmm3, ffi_closure_OFS_V+0x30(%rsp) + movdqa %xmm4, ffi_closure_OFS_V+0x40(%rsp) + movdqa %xmm5, ffi_closure_OFS_V+0x50(%rsp) + movdqa %xmm6, ffi_closure_OFS_V+0x60(%rsp) + movdqa %xmm7, ffi_closure_OFS_V+0x70(%rsp) + jmp L(sse_entry1) + +L(UW7): +ENDF(C(ffi_closure_unix64_sse)) + + .balign 2 + .globl C(ffi_closure_unix64) + FFI_HIDDEN(C(ffi_closure_unix64)) + +C(ffi_closure_unix64): +L(UW8): + _CET_ENDBR + subq $ffi_closure_FS, %rsp +L(UW9): + /* cfi_adjust_cfa_offset(ffi_closure_FS) */ +L(sse_entry1): + movq %rdi, ffi_closure_OFS_G+0x00(%rsp) + movq %rsi, ffi_closure_OFS_G+0x08(%rsp) + movq %rdx, ffi_closure_OFS_G+0x10(%rsp) + movq %rcx, ffi_closure_OFS_G+0x18(%rsp) + movq %r8, ffi_closure_OFS_G+0x20(%rsp) + movq %r9, ffi_closure_OFS_G+0x28(%rsp) + +#ifdef __ILP32__ + movl FFI_TRAMPOLINE_SIZE(%r10), %edi /* Load cif */ + movl FFI_TRAMPOLINE_SIZE+4(%r10), %esi /* Load fun */ + movl FFI_TRAMPOLINE_SIZE+8(%r10), %edx /* Load user_data */ +#else + movq FFI_TRAMPOLINE_SIZE(%r10), %rdi /* Load cif */ + movq FFI_TRAMPOLINE_SIZE+8(%r10), %rsi /* Load fun */ + movq FFI_TRAMPOLINE_SIZE+16(%r10), %rdx /* Load user_data */ +#endif +L(do_closure): + leaq ffi_closure_OFS_RVALUE(%rsp), %rcx /* Load rvalue */ + movq %rsp, %r8 /* Load reg_args */ + leaq ffi_closure_FS+8(%rsp), %r9 /* Load argp */ + call PLT(C(ffi_closure_unix64_inner)) /* Deallocate stack frame early; return value is now in redzone. */ - addq $200, %rsp -.LUW7: + addq $ffi_closure_FS, %rsp +L(UW10): + /* cfi_adjust_cfa_offset(-ffi_closure_FS) */ /* The first byte of the return value contains the FFI_TYPE. */ + cmpb $UNIX64_RET_LAST, %al movzbl %al, %r10d - leaq .Lload_table(%rip), %r11 - movslq (%r11, %r10, 4), %r10 - addq %r11, %r10 + leaq L(load_table)(%rip), %r11 + ja L(la) +#ifdef __CET__ + /* NB: Originally, each slot is 8 byte. 4 bytes of ENDBR64 + + 4 bytes NOP padding double slot size to 16 bytes. */ + addl %r10d, %r10d +#endif + leaq (%r11, %r10, 8), %r10 + leaq ffi_closure_RED_RVALUE(%rsp), %rsi jmp *%r10 -.Lload_table: - .long .Lld_void-.Lload_table /* FFI_TYPE_VOID */ - .long .Lld_int32-.Lload_table /* FFI_TYPE_INT */ - .long .Lld_float-.Lload_table /* FFI_TYPE_FLOAT */ - .long .Lld_double-.Lload_table /* FFI_TYPE_DOUBLE */ - .long .Lld_ldouble-.Lload_table /* FFI_TYPE_LONGDOUBLE */ - .long .Lld_int8-.Lload_table /* FFI_TYPE_UINT8 */ - .long .Lld_int8-.Lload_table /* FFI_TYPE_SINT8 */ - .long .Lld_int16-.Lload_table /* FFI_TYPE_UINT16 */ - .long .Lld_int16-.Lload_table /* FFI_TYPE_SINT16 */ - .long .Lld_int32-.Lload_table /* FFI_TYPE_UINT32 */ - .long .Lld_int32-.Lload_table /* FFI_TYPE_SINT32 */ - .long .Lld_int64-.Lload_table /* FFI_TYPE_UINT64 */ - .long .Lld_int64-.Lload_table /* FFI_TYPE_SINT64 */ - .long .Lld_struct-.Lload_table /* FFI_TYPE_STRUCT */ - .long .Lld_int64-.Lload_table /* FFI_TYPE_POINTER */ - - .align 2 -.Lld_void: + .balign 8 +L(load_table): +E(L(load_table), UNIX64_RET_VOID) + _CET_ENDBR ret - - .align 2 -.Lld_int8: - movzbl -24(%rsp), %eax +E(L(load_table), UNIX64_RET_UINT8) + _CET_ENDBR + movzbl (%rsi), %eax ret - .align 2 -.Lld_int16: - movzwl -24(%rsp), %eax +E(L(load_table), UNIX64_RET_UINT16) + _CET_ENDBR + movzwl (%rsi), %eax ret - .align 2 -.Lld_int32: - movl -24(%rsp), %eax +E(L(load_table), UNIX64_RET_UINT32) + _CET_ENDBR + movl (%rsi), %eax ret - .align 2 -.Lld_int64: - movq -24(%rsp), %rax +E(L(load_table), UNIX64_RET_SINT8) + _CET_ENDBR + movsbl (%rsi), %eax ret - - .align 2 -.Lld_float: - movss -24(%rsp), %xmm0 +E(L(load_table), UNIX64_RET_SINT16) + _CET_ENDBR + movswl (%rsi), %eax ret - .align 2 -.Lld_double: - movsd -24(%rsp), %xmm0 +E(L(load_table), UNIX64_RET_SINT32) + _CET_ENDBR + movl (%rsi), %eax ret - .align 2 -.Lld_ldouble: - fldt -24(%rsp) +E(L(load_table), UNIX64_RET_INT64) + _CET_ENDBR + movq (%rsi), %rax ret - - .align 2 -.Lld_struct: - /* There are four possibilities here, %rax/%rdx, %xmm0/%rax, - %rax/%xmm0, %xmm0/%xmm1. We collapse two by always loading - both rdx and xmm1 with the second word. For the remaining, - bit 8 set means xmm0 gets the second word, and bit 9 means - that rax gets the second word. */ - movq -24(%rsp), %rcx - movq -16(%rsp), %rdx - movq -16(%rsp), %xmm1 - testl $0x100, %eax - cmovnz %rdx, %rcx - movd %rcx, %xmm0 - testl $0x200, %eax - movq -24(%rsp), %rax - cmovnz %rdx, %rax +E(L(load_table), UNIX64_RET_XMM32) + _CET_ENDBR + movd (%rsi), %xmm0 + ret +E(L(load_table), UNIX64_RET_XMM64) + _CET_ENDBR + movq (%rsi), %xmm0 + ret +E(L(load_table), UNIX64_RET_X87) + _CET_ENDBR + fldt (%rsi) + ret +E(L(load_table), UNIX64_RET_X87_2) + _CET_ENDBR + fldt 16(%rsi) + fldt (%rsi) + ret +E(L(load_table), UNIX64_RET_ST_XMM0_RAX) + _CET_ENDBR + movq 8(%rsi), %rax + jmp L(l3) +E(L(load_table), UNIX64_RET_ST_RAX_XMM0) + _CET_ENDBR + movq 8(%rsi), %xmm0 + jmp L(l2) +E(L(load_table), UNIX64_RET_ST_XMM0_XMM1) + _CET_ENDBR + movq 8(%rsi), %xmm1 + jmp L(l3) +E(L(load_table), UNIX64_RET_ST_RAX_RDX) + _CET_ENDBR + movq 8(%rsi), %rdx +L(l2): + movq (%rsi), %rax + ret + .balign 8 +L(l3): + movq (%rsi), %xmm0 ret - /* See the comment above .Lload_sse; the same logic applies here. */ - .align 2 -.LUW8: -.Lsave_sse: - movdqa %xmm0, 48(%rsp) - movdqa %xmm1, 64(%rsp) - movdqa %xmm2, 80(%rsp) - movdqa %xmm3, 96(%rsp) - movdqa %xmm4, 112(%rsp) - movdqa %xmm5, 128(%rsp) - movdqa %xmm6, 144(%rsp) - movdqa %xmm7, 160(%rsp) - jmp .Lret_from_save_sse - -.LUW9: - .size ffi_closure_unix64,.-ffi_closure_unix64 - -#ifdef __GNUC__ -/* Only emit DWARF unwind info when building with the GNU toolchain. */ - -#ifdef HAVE_AS_X86_64_UNWIND_SECTION_TYPE - .section .eh_frame,"a",@unwind -#else - .section .eh_frame,"a",@progbits -#endif -.Lframe1: - .long .LECIE1-.LSCIE1 /* CIE Length */ -.LSCIE1: - .long 0 /* CIE Identifier Tag */ - .byte 1 /* CIE Version */ - .ascii "zR\0" /* CIE Augmentation */ - .uleb128 1 /* CIE Code Alignment Factor */ - .sleb128 -8 /* CIE Data Alignment Factor */ - .byte 0x10 /* CIE RA Column */ - .uleb128 1 /* Augmentation size */ - .byte 0x1b /* FDE Encoding (pcrel sdata4) */ - .byte 0xc /* DW_CFA_def_cfa, %rsp offset 8 */ - .uleb128 7 - .uleb128 8 - .byte 0x80+16 /* DW_CFA_offset, %rip offset 1*-8 */ - .uleb128 1 - .align 8 -.LECIE1: -.LSFDE1: - .long .LEFDE1-.LASFDE1 /* FDE Length */ -.LASFDE1: - .long .LASFDE1-.Lframe1 /* FDE CIE offset */ -#if HAVE_AS_X86_PCREL - .long .LUW0-. /* FDE initial location */ +L(la): call PLT(C(abort)) + +L(UW11): +ENDF(C(ffi_closure_unix64)) + + .balign 2 + .globl C(ffi_go_closure_unix64_sse) + FFI_HIDDEN(C(ffi_go_closure_unix64_sse)) + +C(ffi_go_closure_unix64_sse): +L(UW12): + _CET_ENDBR + subq $ffi_closure_FS, %rsp +L(UW13): + /* cfi_adjust_cfa_offset(ffi_closure_FS) */ + + movdqa %xmm0, ffi_closure_OFS_V+0x00(%rsp) + movdqa %xmm1, ffi_closure_OFS_V+0x10(%rsp) + movdqa %xmm2, ffi_closure_OFS_V+0x20(%rsp) + movdqa %xmm3, ffi_closure_OFS_V+0x30(%rsp) + movdqa %xmm4, ffi_closure_OFS_V+0x40(%rsp) + movdqa %xmm5, ffi_closure_OFS_V+0x50(%rsp) + movdqa %xmm6, ffi_closure_OFS_V+0x60(%rsp) + movdqa %xmm7, ffi_closure_OFS_V+0x70(%rsp) + jmp L(sse_entry2) + +L(UW14): +ENDF(C(ffi_go_closure_unix64_sse)) + + .balign 2 + .globl C(ffi_go_closure_unix64) + FFI_HIDDEN(C(ffi_go_closure_unix64)) + +C(ffi_go_closure_unix64): +L(UW15): + _CET_ENDBR + subq $ffi_closure_FS, %rsp +L(UW16): + /* cfi_adjust_cfa_offset(ffi_closure_FS) */ +L(sse_entry2): + movq %rdi, ffi_closure_OFS_G+0x00(%rsp) + movq %rsi, ffi_closure_OFS_G+0x08(%rsp) + movq %rdx, ffi_closure_OFS_G+0x10(%rsp) + movq %rcx, ffi_closure_OFS_G+0x18(%rsp) + movq %r8, ffi_closure_OFS_G+0x20(%rsp) + movq %r9, ffi_closure_OFS_G+0x28(%rsp) + +#ifdef __ILP32__ + movl 4(%r10), %edi /* Load cif */ + movl 8(%r10), %esi /* Load fun */ + movl %r10d, %edx /* Load closure (user_data) */ #else - .long .LUW0@rel + movq 8(%r10), %rdi /* Load cif */ + movq 16(%r10), %rsi /* Load fun */ + movq %r10, %rdx /* Load closure (user_data) */ #endif - .long .LUW4-.LUW0 /* FDE address range */ - .uleb128 0x0 /* Augmentation size */ + jmp L(do_closure) - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LUW1-.LUW0 +L(UW17): +ENDF(C(ffi_go_closure_unix64)) - /* New stack frame based off rbp. This is a itty bit of unwind - trickery in that the CFA *has* changed. There is no easy way - to describe it correctly on entry to the function. Fortunately, - it doesn't matter too much since at all points we can correctly - unwind back to ffi_call. Note that the location to which we - moved the return address is (the new) CFA-8, so from the - perspective of the unwind info, it hasn't moved. */ - .byte 0xc /* DW_CFA_def_cfa, %rbp offset 32 */ - .uleb128 6 - .uleb128 32 - .byte 0x80+6 /* DW_CFA_offset, %rbp offset 2*-8 */ - .uleb128 2 - .byte 0xa /* DW_CFA_remember_state */ +/* Sadly, OSX cctools-as doesn't understand .cfi directives at all. */ - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LUW2-.LUW1 - .byte 0xc /* DW_CFA_def_cfa, %rsp offset 8 */ - .uleb128 7 - .uleb128 8 - .byte 0xc0+6 /* DW_CFA_restore, %rbp */ +#ifdef __APPLE__ +.section __TEXT,__eh_frame,coalesced,no_toc+strip_static_syms+live_support +EHFrame0: +#elif defined(HAVE_AS_X86_64_UNWIND_SECTION_TYPE) +.section .eh_frame,"a",@unwind +#else +.section .eh_frame,"a",@progbits +#endif - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LUW3-.LUW2 - .byte 0xb /* DW_CFA_restore_state */ +#ifdef HAVE_AS_X86_PCREL +# define PCREL(X) X - . +#else +# define PCREL(X) X@rel +#endif - .align 8 -.LEFDE1: -.LSFDE3: - .long .LEFDE3-.LASFDE3 /* FDE Length */ -.LASFDE3: - .long .LASFDE3-.Lframe1 /* FDE CIE offset */ -#if HAVE_AS_X86_PCREL - .long .LUW5-. /* FDE initial location */ +/* Simplify advancing between labels. Assume DW_CFA_advance_loc1 fits. */ +#ifdef __CET__ +/* Use DW_CFA_advance_loc2 when IBT is enabled. */ +# define ADV(N, P) .byte 3; .2byte L(N)-L(P) #else - .long .LUW5@rel +# define ADV(N, P) .byte 2, L(N)-L(P) #endif - .long .LUW9-.LUW5 /* FDE address range */ - .uleb128 0x0 /* Augmentation size */ - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LUW6-.LUW5 - .byte 0xe /* DW_CFA_def_cfa_offset */ - .uleb128 208 + .balign 8 +L(CIE): + .set L(set0),L(ECIE)-L(SCIE) + .long L(set0) /* CIE Length */ +L(SCIE): + .long 0 /* CIE Identifier Tag */ + .byte 1 /* CIE Version */ + .ascii "zR\0" /* CIE Augmentation */ + .byte 1 /* CIE Code Alignment Factor */ + .byte 0x78 /* CIE Data Alignment Factor */ + .byte 0x10 /* CIE RA Column */ + .byte 1 /* Augmentation size */ + .byte 0x1b /* FDE Encoding (pcrel sdata4) */ + .byte 0xc, 7, 8 /* DW_CFA_def_cfa, %rsp offset 8 */ + .byte 0x80+16, 1 /* DW_CFA_offset, %rip offset 1*-8 */ + .balign 8 +L(ECIE): + + .set L(set1),L(EFDE1)-L(SFDE1) + .long L(set1) /* FDE Length */ +L(SFDE1): + .long L(SFDE1)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW0)) /* Initial location */ + .long L(UW4)-L(UW0) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW1, UW0) + .byte 0xc, 6, 32 /* DW_CFA_def_cfa, %rbp 32 */ + .byte 0x80+6, 2 /* DW_CFA_offset, %rbp 2*-8 */ + ADV(UW2, UW1) .byte 0xa /* DW_CFA_remember_state */ - - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LUW7-.LUW6 - .byte 0xe /* DW_CFA_def_cfa_offset */ - .uleb128 8 - - .byte 0x4 /* DW_CFA_advance_loc4 */ - .long .LUW8-.LUW7 + .byte 0xc, 7, 8 /* DW_CFA_def_cfa, %rsp 8 */ + .byte 0xc0+6 /* DW_CFA_restore, %rbp */ + ADV(UW3, UW2) .byte 0xb /* DW_CFA_restore_state */ + .balign 8 +L(EFDE1): + + .set L(set2),L(EFDE2)-L(SFDE2) + .long L(set2) /* FDE Length */ +L(SFDE2): + .long L(SFDE2)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW5)) /* Initial location */ + .long L(UW7)-L(UW5) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW6, UW5) + .byte 0xe /* DW_CFA_def_cfa_offset */ + .byte ffi_closure_FS + 8, 1 /* uleb128, assuming 128 <= FS < 255 */ + .balign 8 +L(EFDE2): + + .set L(set3),L(EFDE3)-L(SFDE3) + .long L(set3) /* FDE Length */ +L(SFDE3): + .long L(SFDE3)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW8)) /* Initial location */ + .long L(UW11)-L(UW8) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW9, UW8) + .byte 0xe /* DW_CFA_def_cfa_offset */ + .byte ffi_closure_FS + 8, 1 /* uleb128, assuming 128 <= FS < 255 */ + ADV(UW10, UW9) + .byte 0xe, 8 /* DW_CFA_def_cfa_offset 8 */ +L(EFDE3): + + .set L(set4),L(EFDE4)-L(SFDE4) + .long L(set4) /* FDE Length */ +L(SFDE4): + .long L(SFDE4)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW12)) /* Initial location */ + .long L(UW14)-L(UW12) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW13, UW12) + .byte 0xe /* DW_CFA_def_cfa_offset */ + .byte ffi_closure_FS + 8, 1 /* uleb128, assuming 128 <= FS < 255 */ + .balign 8 +L(EFDE4): + + .set L(set5),L(EFDE5)-L(SFDE5) + .long L(set5) /* FDE Length */ +L(SFDE5): + .long L(SFDE5)-L(CIE) /* FDE CIE offset */ + .long PCREL(L(UW15)) /* Initial location */ + .long L(UW17)-L(UW15) /* Address range */ + .byte 0 /* Augmentation size */ + ADV(UW16, UW15) + .byte 0xe /* DW_CFA_def_cfa_offset */ + .byte ffi_closure_FS + 8, 1 /* uleb128, assuming 128 <= FS < 255 */ + .balign 8 +L(EFDE5): +#ifdef __APPLE__ + .subsections_via_symbols + .section __LD,__compact_unwind,regular,debug + + /* compact unwind for ffi_call_unix64 */ + .quad C(ffi_call_unix64) + .set L1,L(UW4)-L(UW0) + .long L1 + .long 0x04000000 /* use dwarf unwind info */ + .quad 0 + .quad 0 + + /* compact unwind for ffi_closure_unix64_sse */ + .quad C(ffi_closure_unix64_sse) + .set L2,L(UW7)-L(UW5) + .long L2 + .long 0x04000000 /* use dwarf unwind info */ + .quad 0 + .quad 0 + + /* compact unwind for ffi_closure_unix64 */ + .quad C(ffi_closure_unix64) + .set L3,L(UW11)-L(UW8) + .long L3 + .long 0x04000000 /* use dwarf unwind info */ + .quad 0 + .quad 0 + + /* compact unwind for ffi_go_closure_unix64_sse */ + .quad C(ffi_go_closure_unix64_sse) + .set L4,L(UW14)-L(UW12) + .long L4 + .long 0x04000000 /* use dwarf unwind info */ + .quad 0 + .quad 0 + + /* compact unwind for ffi_go_closure_unix64 */ + .quad C(ffi_go_closure_unix64) + .set L5,L(UW17)-L(UW15) + .long L5 + .long 0x04000000 /* use dwarf unwind info */ + .quad 0 + .quad 0 +#endif - .align 8 -.LEFDE3: - -#endif /* __GNUC__ */ - #endif /* __x86_64__ */ - #if defined __ELF__ && defined __linux__ .section .note.GNU-stack,"",@progbits #endif diff --git a/native/libffi/src/x86/win64.S b/native/libffi/src/x86/win64.S index 3f4981c4df..8b18330ed9 100644 --- a/native/libffi/src/x86/win64.S +++ b/native/libffi/src/x86/win64.S @@ -1,6 +1,8 @@ +#ifdef __x86_64__ #define LIBFFI_ASM #include #include +<<<<<<< HEAD /* Constants for ffi_call_win64 */ #define STACK 0 @@ -257,15 +259,40 @@ ret_void$: ffi_call_win64 ENDP _TEXT ENDS END +======= +#include +#include "asmnames.h" +#if defined(HAVE_AS_CFI_PSEUDO_OP) + .cfi_sections .debug_frame +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + +#ifdef X86_WIN64 +#define SEH(...) __VA_ARGS__ +#define arg0 %rcx +#define arg1 %rdx +#define arg2 %r8 +#define arg3 %r9 #else +#define SEH(...) +#define arg0 %rdi +#define arg1 %rsi +#define arg2 %rdx +#define arg3 %rcx +#endif -#ifdef SYMBOL_UNDERSCORE -#define SYMBOL_NAME(name) _##name +/* This macro allows the safe creation of jump tables without an + actual table. The entry points into the table are all 8 bytes. + The use of ORG asserts that we're at the correct location. */ +/* ??? The clang assembler doesn't handle .org with symbolic expressions. */ +#if defined(__clang__) || defined(__APPLE__) || (defined (__sun__) && defined(__svr4__)) +# define E(BASE, X) .balign 8 #else -#define SYMBOL_NAME(name) name +# define E(BASE, X) .balign 8; .org BASE + (X) * 8 #endif +<<<<<<< HEAD .text .extern SYMBOL_NAME(ffi_closure_win64_inner) @@ -518,3 +545,211 @@ Lch_done: .seh_endproc #endif /* !_MSC_VER */ +======= + .text + +/* ffi_call_win64 (void *stack, struct win64_call_frame *frame, void *r10) + + Bit o trickiness here -- FRAME is the base of the stack frame + for this function. This has been allocated by ffi_call. We also + deallocate some of the stack that has been alloca'd. */ + + .align 8 + .globl C(ffi_call_win64) + FFI_HIDDEN(C(ffi_call_win64)) + + SEH(.seh_proc ffi_call_win64) +C(ffi_call_win64): + cfi_startproc + _CET_ENDBR + /* Set up the local stack frame and install it in rbp/rsp. */ + movq (%rsp), %rax + movq %rbp, (arg1) + movq %rax, 8(arg1) + movq arg1, %rbp + cfi_def_cfa(%rbp, 16) + cfi_rel_offset(%rbp, 0) + SEH(.seh_pushreg %rbp) + SEH(.seh_setframe %rbp, 0) + SEH(.seh_endprologue) + movq arg0, %rsp + + movq arg2, %r10 + + /* Load all slots into both general and xmm registers. */ + movq (%rsp), %rcx + movsd (%rsp), %xmm0 + movq 8(%rsp), %rdx + movsd 8(%rsp), %xmm1 + movq 16(%rsp), %r8 + movsd 16(%rsp), %xmm2 + movq 24(%rsp), %r9 + movsd 24(%rsp), %xmm3 + + call *16(%rbp) + + movl 24(%rbp), %ecx + movq 32(%rbp), %r8 + leaq 0f(%rip), %r10 + cmpl $FFI_TYPE_SMALL_STRUCT_4B, %ecx + leaq (%r10, %rcx, 8), %r10 + ja 99f + _CET_NOTRACK jmp *%r10 + +/* Below, we're space constrained most of the time. Thus we eschew the + modern "mov, pop, ret" sequence (5 bytes) for "leave, ret" (2 bytes). */ +.macro epilogue + leaveq + cfi_remember_state + cfi_def_cfa(%rsp, 8) + cfi_restore(%rbp) + ret + cfi_restore_state +.endm + + .align 8 +0: +E(0b, FFI_TYPE_VOID) + epilogue +E(0b, FFI_TYPE_INT) + movslq %eax, %rax + movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_FLOAT) + movss %xmm0, (%r8) + epilogue +E(0b, FFI_TYPE_DOUBLE) + movsd %xmm0, (%r8) + epilogue +// FFI_TYPE_LONGDOUBLE may be FFI_TYPE_DOUBLE but we need a different value here. +E(0b, FFI_TYPE_DOUBLE + 1) + call PLT(C(abort)) +E(0b, FFI_TYPE_UINT8) + movzbl %al, %eax + movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_SINT8) + movsbq %al, %rax + jmp 98f +E(0b, FFI_TYPE_UINT16) + movzwl %ax, %eax + movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_SINT16) + movswq %ax, %rax + jmp 98f +E(0b, FFI_TYPE_UINT32) + movl %eax, %eax + movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_SINT32) + movslq %eax, %rax + movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_UINT64) +98: movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_SINT64) + movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_STRUCT) + epilogue +E(0b, FFI_TYPE_POINTER) + movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_COMPLEX) + call PLT(C(abort)) +E(0b, FFI_TYPE_SMALL_STRUCT_1B) + movb %al, (%r8) + epilogue +E(0b, FFI_TYPE_SMALL_STRUCT_2B) + movw %ax, (%r8) + epilogue +E(0b, FFI_TYPE_SMALL_STRUCT_4B) + movl %eax, (%r8) + epilogue + + .align 8 +99: call PLT(C(abort)) + + epilogue + + cfi_endproc + SEH(.seh_endproc) + + +/* 32 bytes of outgoing register stack space, 8 bytes of alignment, + 16 bytes of result, 32 bytes of xmm registers. */ +#define ffi_clo_FS (32+8+16+32) +#define ffi_clo_OFF_R (32+8) +#define ffi_clo_OFF_X (32+8+16) + + .align 8 + .globl C(ffi_go_closure_win64) + FFI_HIDDEN(C(ffi_go_closure_win64)) + + SEH(.seh_proc ffi_go_closure_win64) +C(ffi_go_closure_win64): + cfi_startproc + _CET_ENDBR + /* Save all integer arguments into the incoming reg stack space. */ + movq %rcx, 8(%rsp) + movq %rdx, 16(%rsp) + movq %r8, 24(%rsp) + movq %r9, 32(%rsp) + + movq 8(%r10), %rcx /* load cif */ + movq 16(%r10), %rdx /* load fun */ + movq %r10, %r8 /* closure is user_data */ + jmp 0f + cfi_endproc + SEH(.seh_endproc) + + .align 8 + .globl C(ffi_closure_win64) + FFI_HIDDEN(C(ffi_closure_win64)) + + SEH(.seh_proc ffi_closure_win64) +C(ffi_closure_win64): + cfi_startproc + _CET_ENDBR + /* Save all integer arguments into the incoming reg stack space. */ + movq %rcx, 8(%rsp) + movq %rdx, 16(%rsp) + movq %r8, 24(%rsp) + movq %r9, 32(%rsp) + + movq FFI_TRAMPOLINE_SIZE(%r10), %rcx /* load cif */ + movq FFI_TRAMPOLINE_SIZE+8(%r10), %rdx /* load fun */ + movq FFI_TRAMPOLINE_SIZE+16(%r10), %r8 /* load user_data */ +0: + subq $ffi_clo_FS, %rsp + cfi_adjust_cfa_offset(ffi_clo_FS) + SEH(.seh_stackalloc ffi_clo_FS) + SEH(.seh_endprologue) + + /* Save all sse arguments into the stack frame. */ + movsd %xmm0, ffi_clo_OFF_X(%rsp) + movsd %xmm1, ffi_clo_OFF_X+8(%rsp) + movsd %xmm2, ffi_clo_OFF_X+16(%rsp) + movsd %xmm3, ffi_clo_OFF_X+24(%rsp) + + leaq ffi_clo_OFF_R(%rsp), %r9 + call PLT(C(ffi_closure_win64_inner)) + + /* Load the result into both possible result registers. */ + movq ffi_clo_OFF_R(%rsp), %rax + movsd ffi_clo_OFF_R(%rsp), %xmm0 + + addq $ffi_clo_FS, %rsp + cfi_adjust_cfa_offset(-ffi_clo_FS) + ret + + cfi_endproc + SEH(.seh_endproc) +#endif /* __x86_64__ */ + +#if defined __ELF__ && defined __linux__ + .section .note.GNU-stack,"",@progbits +#endif +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 diff --git a/native/libffi/src/x86/win64_intel.S b/native/libffi/src/x86/win64_intel.S new file mode 100644 index 0000000000..970a4f997a --- /dev/null +++ b/native/libffi/src/x86/win64_intel.S @@ -0,0 +1,238 @@ +#define LIBFFI_ASM +#include +#include +#include +#include "asmnames.h" + +#if defined(HAVE_AS_CFI_PSEUDO_OP) + .cfi_sections .debug_frame +#endif + +#ifdef X86_WIN64 +#define SEH(...) __VA_ARGS__ +#define arg0 rcx +#define arg1 rdx +#define arg2 r8 +#define arg3 r9 +#else +#define SEH(...) +#define arg0 rdi +#define arg1 rsi +#define arg2 rdx +#define arg3 rcx +#endif + +/* This macro allows the safe creation of jump tables without an + actual table. The entry points into the table are all 8 bytes. + The use of ORG asserts that we're at the correct location. */ +/* ??? The clang assembler doesn't handle .org with symbolic expressions. */ +#if defined(__clang__) || defined(__APPLE__) || (defined (__sun__) && defined(__svr4__)) +# define E(BASE, X) ALIGN 8 +#else +# define E(BASE, X) ALIGN 8; ORG BASE + (X) * 8 +#endif + + .CODE + extern PLT(C(abort)):near + extern C(ffi_closure_win64_inner):near + +/* ffi_call_win64 (void *stack, struct win64_call_frame *frame, void *r10) + + Bit o trickiness here -- FRAME is the base of the stack frame + for this function. This has been allocated by ffi_call. We also + deallocate some of the stack that has been alloca'd. */ + + ALIGN 8 + PUBLIC C(ffi_call_win64) + + ; SEH(.safesh ffi_call_win64) +C(ffi_call_win64) proc SEH(frame) + cfi_startproc + /* Set up the local stack frame and install it in rbp/rsp. */ + mov RAX, [RSP] ; movq (%rsp), %rax + mov [arg1], RBP ; movq %rbp, (arg1) + mov [arg1 + 8], RAX; movq %rax, 8(arg1) + mov RBP, arg1; movq arg1, %rbp + cfi_def_cfa(rbp, 16) + cfi_rel_offset(rbp, 0) + SEH(.pushreg rbp) + SEH(.setframe rbp, 0) + SEH(.endprolog) + mov RSP, arg0 ; movq arg0, %rsp + + mov R10, arg2 ; movq arg2, %r10 + + /* Load all slots into both general and xmm registers. */ + mov RCX, [RSP] ; movq (%rsp), %rcx + movsd XMM0, qword ptr [RSP] ; movsd (%rsp), %xmm0 + mov RDX, [RSP + 8] ;movq 8(%rsp), %rdx + movsd XMM1, qword ptr [RSP + 8]; movsd 8(%rsp), %xmm1 + mov R8, [RSP + 16] ; movq 16(%rsp), %r8 + movsd XMM2, qword ptr [RSP + 16] ; movsd 16(%rsp), %xmm2 + mov R9, [RSP + 24] ; movq 24(%rsp), %r9 + movsd XMM3, qword ptr [RSP + 24] ;movsd 24(%rsp), %xmm3 + + CALL qword ptr [RBP + 16] ; call *16(%rbp) + + mov ECX, [RBP + 24] ; movl 24(%rbp), %ecx + mov R8, [RBP + 32] ; movq 32(%rbp), %r8 + LEA R10, ffi_call_win64_tab ; leaq 0f(%rip), %r10 + CMP ECX, FFI_TYPE_SMALL_STRUCT_4B ; cmpl $FFI_TYPE_SMALL_STRUCT_4B, %ecx + LEA R10, [R10 + RCX*8] ; leaq (%r10, %rcx, 8), %r10 + JA L99 ; ja 99f + JMP R10 ; jmp *%r10 + +/* Below, we're space constrained most of the time. Thus we eschew the + modern "mov, pop, ret" sequence (5 bytes) for "leave, ret" (2 bytes). */ +epilogue macro + LEAVE + cfi_remember_state + cfi_def_cfa(rsp, 8) + cfi_restore(rbp) + RET + cfi_restore_state +endm + + ALIGN 8 +ffi_call_win64_tab LABEL NEAR +E(0b, FFI_TYPE_VOID) + epilogue +E(0b, FFI_TYPE_INT) + movsxd rax, eax ; movslq %eax, %rax + mov qword ptr [r8], rax; movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_FLOAT) + movss dword ptr [r8], xmm0 ; movss %xmm0, (%r8) + epilogue +E(0b, FFI_TYPE_DOUBLE) + movsd qword ptr[r8], xmm0; movsd %xmm0, (%r8) + epilogue +// FFI_TYPE_LONGDOUBLE may be FFI_TYPE_DOUBLE but we need a different value here. +E(0b, FFI_TYPE_DOUBLE + 1) + call PLT(C(abort)) +E(0b, FFI_TYPE_UINT8) + movzx eax, al ;movzbl %al, %eax + mov qword ptr[r8], rax; movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_SINT8) + movsx rax, al ; movsbq %al, %rax + jmp L98 +E(0b, FFI_TYPE_UINT16) + movzx eax, ax ; movzwl %ax, %eax + mov qword ptr[r8], rax; movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_SINT16) + movsx rax, ax; movswq %ax, %rax + jmp L98 +E(0b, FFI_TYPE_UINT32) + mov eax, eax; movl %eax, %eax + mov qword ptr[r8], rax ; movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_SINT32) + movsxd rax, eax; movslq %eax, %rax + mov qword ptr [r8], rax; movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_UINT64) +L98 LABEL near + mov qword ptr [r8], rax ; movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_SINT64) + mov qword ptr [r8], rax;movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_STRUCT) + epilogue +E(0b, FFI_TYPE_POINTER) + mov qword ptr [r8], rax ;movq %rax, (%r8) + epilogue +E(0b, FFI_TYPE_COMPLEX) + call PLT(C(abort)) +E(0b, FFI_TYPE_SMALL_STRUCT_1B) + mov byte ptr [r8], al ; movb %al, (%r8) + epilogue +E(0b, FFI_TYPE_SMALL_STRUCT_2B) + mov word ptr [r8], ax ; movw %ax, (%r8) + epilogue +E(0b, FFI_TYPE_SMALL_STRUCT_4B) + mov dword ptr [r8], eax ; movl %eax, (%r8) + epilogue + + align 8 +L99 LABEL near + call PLT(C(abort)) + + epilogue + + cfi_endproc + C(ffi_call_win64) endp + + +/* 32 bytes of outgoing register stack space, 8 bytes of alignment, + 16 bytes of result, 32 bytes of xmm registers. */ +#define ffi_clo_FS (32+8+16+32) +#define ffi_clo_OFF_R (32+8) +#define ffi_clo_OFF_X (32+8+16) + + align 8 + PUBLIC C(ffi_go_closure_win64) + +C(ffi_go_closure_win64) proc + cfi_startproc + /* Save all integer arguments into the incoming reg stack space. */ + mov qword ptr [rsp + 8], rcx; movq %rcx, 8(%rsp) + mov qword ptr [rsp + 16], rdx; movq %rdx, 16(%rsp) + mov qword ptr [rsp + 24], r8; movq %r8, 24(%rsp) + mov qword ptr [rsp + 32], r9 ;movq %r9, 32(%rsp) + + mov rcx, qword ptr [r10 + 8]; movq 8(%r10), %rcx /* load cif */ + mov rdx, qword ptr [r10 + 16]; movq 16(%r10), %rdx /* load fun */ + mov r8, r10 ; movq %r10, %r8 /* closure is user_data */ + jmp ffi_closure_win64_2 + cfi_endproc + C(ffi_go_closure_win64) endp + + align 8 + +PUBLIC C(ffi_closure_win64) +C(ffi_closure_win64) PROC FRAME + cfi_startproc + /* Save all integer arguments into the incoming reg stack space. */ + mov qword ptr [rsp + 8], rcx; movq %rcx, 8(%rsp) + mov qword ptr [rsp + 16], rdx; movq %rdx, 16(%rsp) + mov qword ptr [rsp + 24], r8; movq %r8, 24(%rsp) + mov qword ptr [rsp + 32], r9; movq %r9, 32(%rsp) + + mov rcx, qword ptr [FFI_TRAMPOLINE_SIZE + r10] ;movq FFI_TRAMPOLINE_SIZE(%r10), %rcx /* load cif */ + mov rdx, qword ptr [FFI_TRAMPOLINE_SIZE + 8 + r10] ; movq FFI_TRAMPOLINE_SIZE+8(%r10), %rdx /* load fun */ + mov r8, qword ptr [FFI_TRAMPOLINE_SIZE+16+r10] ;movq FFI_TRAMPOLINE_SIZE+16(%r10), %r8 /* load user_data */ +ffi_closure_win64_2 LABEL near + sub rsp, ffi_clo_FS ;subq $ffi_clo_FS, %rsp + cfi_adjust_cfa_offset(ffi_clo_FS) + SEH(.allocstack ffi_clo_FS) + SEH(.endprolog) + + /* Save all sse arguments into the stack frame. */ + movsd qword ptr [ffi_clo_OFF_X + rsp], xmm0 ; movsd %xmm0, ffi_clo_OFF_X(%rsp) + movsd qword ptr [ffi_clo_OFF_X+8+rsp], xmm1 ; movsd %xmm1, ffi_clo_OFF_X+8(%rsp) + movsd qword ptr [ffi_clo_OFF_X+16+rsp], xmm2 ; movsd %xmm2, ffi_clo_OFF_X+16(%rsp) + movsd qword ptr [ffi_clo_OFF_X+24+rsp], xmm3 ; movsd %xmm3, ffi_clo_OFF_X+24(%rsp) + + lea r9, [ffi_clo_OFF_R + rsp] ; leaq ffi_clo_OFF_R(%rsp), %r9 + call C(ffi_closure_win64_inner) + + /* Load the result into both possible result registers. */ + + mov rax, qword ptr [ffi_clo_OFF_R + rsp] ;movq ffi_clo_OFF_R(%rsp), %rax + movsd xmm0, qword ptr [rsp + ffi_clo_OFF_R] ;movsd ffi_clo_OFF_R(%rsp), %xmm0 + + add rsp, ffi_clo_FS ;addq $ffi_clo_FS, %rsp + cfi_adjust_cfa_offset(-ffi_clo_FS) + ret + + cfi_endproc + C(ffi_closure_win64) endp + +#if defined __ELF__ && defined __linux__ + .section .note.GNU-stack,"",@progbits +#endif +_text ends +end \ No newline at end of file diff --git a/native/libffi/src/xtensa/ffi.c b/native/libffi/src/xtensa/ffi.c index fd94dafbe3..9a0575ff90 100644 --- a/native/libffi/src/xtensa/ffi.c +++ b/native/libffi/src/xtensa/ffi.c @@ -89,7 +89,7 @@ ffi_status ffi_prep_cif_machdep(ffi_cif *cif) /* Round the stack up to a full 4 register frame, just in case (we use this size in movsp). This way, it's also a multiple of 8 bytes for 64-bit arguments. */ - cif->bytes = ALIGN(cif->bytes, 16); + cif->bytes = FFI_ALIGN(cif->bytes, 16); return FFI_OK; } @@ -205,7 +205,7 @@ void ffi_call(ffi_cif* cif, void(*fn)(void), void *rvalue, void **avalue) if (flags == FFI_TYPE_STRUCT && (rsize <= 16 || rvalue == NULL)) { - alloc = alloca(ALIGN(rsize, 4)); + alloc = alloca(FFI_ALIGN(rsize, 4)); ecif.rvalue = alloc; } else diff --git a/native/libffi/src/xtensa/sysv.S b/native/libffi/src/xtensa/sysv.S index 64e6a0918d..e9421796fd 100644 --- a/native/libffi/src/xtensa/sysv.S +++ b/native/libffi/src/xtensa/sysv.S @@ -169,8 +169,13 @@ ENTRY(ffi_cacheflush) entry a1, 16 -1: dhwbi a2, 0 +1: +#if XCHAL_DCACHE_SIZE + dhwbi a2, 0 +#endif +#if XCHAL_ICACHE_SIZE ihi a2, 0 +#endif addi a2, a2, 4 blt a2, a3, 1b diff --git a/native/libffi/testsuite/Makefile.am b/native/libffi/testsuite/Makefile.am index da10465d28..e705afc8a5 100644 --- a/native/libffi/testsuite/Makefile.am +++ b/native/libffi/testsuite/Makefile.am @@ -6,6 +6,7 @@ EXTRA_DEJAGNU_SITE_CONFIG=../local.exp CLEANFILES = *.exe core* *.log *.sum +<<<<<<< HEAD EXTRA_DIST = config/default.exp libffi.call/cls_19byte.c \ libffi.call/cls_align_longdouble_split.c \ libffi.call/closure_loc_fn0.c libffi.call/cls_schar.c \ @@ -83,3 +84,119 @@ libffi.call/va_struct3.c \ libffi.call/strlen2.c \ libffi.call/strlen3.c \ libffi.call/strlen4.c +======= +EXTRA_DIST = lib/target-libpath.exp lib/libffi.exp lib/wrapper.exp \ +libffi.call/strlen4.c libffi.call/struct10.c libffi.call/many_mixed.c \ +libffi.call/float.c libffi.call/struct5.c libffi.call/return_fl3.c \ +libffi.call/return_fl1.c libffi.call/call.exp libffi.call/pyobjc-tc.c \ +libffi.call/float_va.c libffi.call/struct8.c libffi.call/pr1172638.c \ +libffi.call/return_sc.c libffi.call/va_struct1.c \ +libffi.call/align_stdcall.c libffi.call/struct9.c libffi.call/va_1.c \ +libffi.call/va_struct2.c libffi.call/return_fl2.c \ +libffi.call/align_mixed.c libffi.call/ffitest.h libffi.call/struct4.c \ +libffi.call/return_ldl.c libffi.call/float3.c libffi.call/return_sl.c \ +libffi.call/return_dbl1.c libffi.call/err_bad_typedef.c \ +libffi.call/return_ll1.c libffi.call/return_dbl2.c \ +libffi.call/negint.c libffi.closures/nested_struct3.c \ +libffi.call/struct2.c libffi.call/struct3.c libffi.call/return_fl.c \ +libffi.call/offsets.c libffi.call/struct7.c libffi.call/va_struct3.c \ +libffi.call/float1.c libffi.call/uninitialized.c libffi.call/many2.c \ +libffi.call/struct6.c libffi.call/strlen2.c libffi.call/float2.c \ +libffi.call/return_ul.c libffi.call/struct1.c libffi.call/strlen3.c \ +libffi.call/return_dbl.c libffi.call/float4.c libffi.call/many.c \ +libffi.call/strlen.c libffi.call/return_uc.c libffi.call/many_double.c \ +libffi.call/return_ll.c libffi.call/promotion.c \ +libffi.complex/complex_defs_longdouble.inc \ +libffi.complex/cls_align_complex_float.c \ +libffi.complex/cls_complex_va_float.c \ +libffi.complex/cls_complex_struct_float.c \ +libffi.complex/return_complex2_longdouble.c \ +libffi.complex/cls_complex_float.c \ +libffi.complex/return_complex_longdouble.c \ +libffi.complex/return_complex2_float.c libffi.complex/cls_complex.inc \ +libffi.complex/cls_complex_va_longdouble.c \ +libffi.complex/return_complex_double.c \ +libffi.complex/return_complex.inc libffi.complex/many_complex.inc \ +libffi.complex/complex_float.c libffi.complex/cls_align_complex.inc \ +libffi.complex/return_complex2_double.c \ +libffi.complex/many_complex_float.c libffi.complex/ffitest.h \ +libffi.complex/return_complex1_double.c \ +libffi.complex/cls_complex_struct_longdouble.c \ +libffi.complex/complex_defs_double.inc \ +libffi.complex/cls_complex_va_double.c \ +libffi.complex/many_complex_double.c \ +libffi.complex/return_complex2.inc \ +libffi.complex/return_complex1_float.c \ +libffi.complex/complex_longdouble.c \ +libffi.complex/complex_defs_float.inc \ +libffi.complex/cls_complex_double.c \ +libffi.complex/cls_align_complex_double.c \ +libffi.complex/cls_align_complex_longdouble.c \ +libffi.complex/complex_double.c libffi.complex/cls_complex_va.inc \ +libffi.complex/many_complex_longdouble.c libffi.complex/complex.inc \ +libffi.complex/return_complex1_longdouble.c \ +libffi.complex/complex_int.c libffi.complex/cls_complex_longdouble.c \ +libffi.complex/cls_complex_struct_double.c \ +libffi.complex/return_complex1.inc libffi.complex/complex.exp \ +libffi.complex/cls_complex_struct.inc \ +libffi.complex/return_complex_float.c libffi.go/closure1.c \ +libffi.go/aa-direct.c libffi.go/ffitest.h libffi.go/go.exp \ +libffi.go/static-chain.h libffi.bhaible/bhaible.exp \ +libffi.bhaible/test-call.c libffi.bhaible/alignof.h \ +libffi.bhaible/testcases.c libffi.bhaible/test-callback.c \ +libffi.bhaible/Makefile libffi.bhaible/README config/default.exp \ +libffi.closures/cls_multi_sshort.c \ +libffi.closures/cls_align_longdouble_split2.c \ +libffi.closures/cls_1_1byte.c libffi.closures/cls_uint_va.c \ +libffi.closures/cls_3_1byte.c libffi.closures/cls_many_mixed_args.c \ +libffi.closures/cls_20byte1.c libffi.closures/cls_pointer_stack.c \ +libffi.closures/cls_align_float.c libffi.closures/cls_5_1_byte.c \ +libffi.closures/cls_9byte1.c libffi.closures/cls_align_uint32.c \ +libffi.closures/stret_medium.c libffi.closures/cls_3byte1.c \ +libffi.closures/cls_align_uint64.c libffi.closures/cls_longdouble_va.c \ +libffi.closures/cls_align_pointer.c libffi.closures/cls_19byte.c \ +libffi.closures/cls_ushort.c libffi.closures/cls_align_sint32.c \ +libffi.closures/cls_ulonglong.c libffi.closures/cls_struct_va1.c \ +libffi.closures/cls_9byte2.c libffi.closures/closure_fn5.c \ +libffi.closures/cls_5byte.c libffi.closures/cls_3float.c \ +libffi.closures/closure.exp libffi.closures/cls_schar.c \ +libffi.closures/closure_fn4.c libffi.closures/cls_uchar_va.c \ +libffi.closures/closure_fn0.c libffi.closures/huge_struct.c \ +libffi.closures/cls_ushort_va.c \ +libffi.closures/cls_64byte.c libffi.closures/cls_longdouble.c \ +libffi.closures/cls_ulong_va.c libffi.closures/cls_6_1_byte.c \ +libffi.closures/cls_align_uint16.c libffi.closures/closure_fn2.c \ +libffi.closures/unwindtest_ffi_call.cc \ +libffi.closures/cls_multi_ushortchar.c libffi.closures/cls_8byte.c \ +libffi.closures/ffitest.h libffi.closures/nested_struct8.c \ +libffi.closures/cls_pointer.c libffi.closures/nested_struct2.c \ +libffi.closures/nested_struct.c libffi.closures/cls_multi_schar.c \ +libffi.closures/cls_align_longdouble_split.c \ +libffi.closures/cls_uchar.c libffi.closures/nested_struct9.c \ +libffi.closures/cls_float.c libffi.closures/stret_medium2.c \ +libffi.closures/closure_loc_fn0.c libffi.closures/cls_6byte.c \ +libffi.closures/closure_simple.c libffi.closures/cls_align_double.c \ +libffi.closures/cls_multi_uchar.c libffi.closures/cls_4_1byte.c \ +libffi.closures/closure_fn3.c libffi.closures/cls_align_sint64.c \ +libffi.closures/nested_struct1.c libffi.closures/unwindtest.cc \ +libffi.closures/nested_struct5.c libffi.closures/cls_multi_ushort.c \ +libffi.closures/nested_struct11.c \ +libffi.closures/cls_multi_sshortchar.c \ +libffi.closures/cls_align_longdouble.c \ +libffi.closures/cls_dbls_struct.c \ +libffi.closures/cls_many_mixed_float_double.c \ +libffi.closures/stret_large.c libffi.closures/stret_large2.c \ +libffi.closures/cls_align_sint16.c libffi.closures/cls_2byte.c \ +libffi.closures/nested_struct4.c libffi.closures/problem1.c \ +libffi.closures/testclosure.c libffi.closures/nested_struct6.c \ +libffi.closures/cls_4byte.c libffi.closures/cls_24byte.c \ +libffi.closures/nested_struct10.c libffi.closures/cls_uint.c \ +libffi.closures/cls_12byte.c libffi.closures/cls_sint.c \ +libffi.closures/cls_7_1_byte.c libffi.closures/cls_sshort.c \ +libffi.closures/cls_16byte.c libffi.closures/nested_struct7.c \ +libffi.closures/cls_double_va.c libffi.closures/cls_3byte2.c \ +libffi.closures/cls_double.c libffi.closures/cls_7byte.c \ +libffi.closures/closure_fn6.c libffi.closures/closure_fn1.c \ +libffi.closures/cls_20byte.c libffi.closures/cls_18byte.c \ +libffi.closures/err_bad_abi.c +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 diff --git a/native/libffi/testsuite/lib/libffi.exp b/native/libffi/testsuite/lib/libffi.exp index 5051d3175e..0cf035ba3b 100644 --- a/native/libffi/testsuite/lib/libffi.exp +++ b/native/libffi/testsuite/lib/libffi.exp @@ -1,4 +1,8 @@ +<<<<<<< HEAD # Copyright (C) 2003, 2005, 2008, 2009, 2010, 2011, 2014 Free Software Foundation, Inc. +======= +# Copyright (C) 2003, 2005, 2008, 2009, 2010, 2011, 2014, 2019 Free Software Foundation, Inc. +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by @@ -24,6 +28,183 @@ load_lib libgloss.exp load_gcc_lib target-libpath.exp load_gcc_lib wrapper.exp +proc check_effective_target_gccbug { } { + global has_gccbug + return $has_gccbug +} + +# Return 1 if the target matches the effective target 'arg', 0 otherwise. +# This can be used with any check_* proc that takes no argument and +# returns only 1 or 0. It could be used with check_* procs that take +# arguments with keywords that pass particular arguments. + +proc is-effective-target { arg } { + global et_index + set selected 0 + if { ![info exists et_index] } { + # Initialize the effective target index that is used in some + # check_effective_target_* procs. + set et_index 0 + } + if { [info procs check_effective_target_${arg}] != [list] } { + set selected [check_effective_target_${arg}] + } else { + error "unknown effective target keyword `$arg'" + } + verbose "is-effective-target: $arg $selected" 2 + return $selected +} + +proc is-effective-target-keyword { arg } { + if { [info procs check_effective_target_${arg}] != [list] } { + return 1 + } else { + return 0 + } +} + +# Intercept the call to the DejaGnu version of dg-process-target to +# support use of an effective-target keyword in place of a list of +# target triplets to xfail or skip a test. +# +# The argument to dg-process-target is the keyword "target" or "xfail" +# followed by a selector: +# target-triplet-1 ... +# effective-target-keyword +# selector-expression +# +# For a target list the result is "S" if the target is selected, "N" otherwise. +# For an xfail list the result is "F" if the target is affected, "P" otherwise. + +# In contexts that allow either "target" or "xfail" the argument can be +# target selector1 xfail selector2 +# which returns "N" if selector1 is not selected, otherwise the result of +# "xfail selector2". +# +# A selector expression appears within curly braces and uses a single logical +# operator: !, &&, or ||. An operand is another selector expression, an +# effective-target keyword, or a list of target triplets within quotes or +# curly braces. + +if { [info procs saved-dg-process-target] == [list] } { + rename dg-process-target saved-dg-process-target + + # Evaluate an operand within a selector expression. + proc selector_opd { op } { + set selector "target" + lappend selector $op + set answer [ expr { [dg-process-target $selector] == "S" } ] + verbose "selector_opd: `$op' $answer" 2 + return $answer + } + + # Evaluate a target triplet list within a selector expression. + # Unlike other operands, this needs to be expanded from a list to + # the same string as "target". + proc selector_list { op } { + set selector "target [join $op]" + set answer [ expr { [dg-process-target $selector] == "S" } ] + verbose "selector_list: `$op' $answer" 2 + return $answer + } + + # Evaluate a selector expression. + proc selector_expression { exp } { + if { [llength $exp] == 2 } { + if [string match "!" [lindex $exp 0]] { + set op1 [lindex $exp 1] + set answer [expr { ! [selector_opd $op1] }] + } else { + # Assume it's a list of target triplets. + set answer [selector_list $exp] + } + } elseif { [llength $exp] == 3 } { + set op1 [lindex $exp 0] + set opr [lindex $exp 1] + set op2 [lindex $exp 2] + if [string match "&&" $opr] { + set answer [expr { [selector_opd $op1] && [selector_opd $op2] }] + } elseif [string match "||" $opr] { + set answer [expr { [selector_opd $op1] || [selector_opd $op2] }] + } else { + # Assume it's a list of target triplets. + set answer [selector_list $exp] + } + } else { + # Assume it's a list of target triplets. + set answer [selector_list $exp] + } + + verbose "selector_expression: `$exp' $answer" 2 + return $answer + } + + # Evaluate "target selector" or "xfail selector". + + proc dg-process-target-1 { args } { + verbose "dg-process-target-1: `$args'" 2 + + # Extract the 'what' keyword from the argument list. + set selector [string trim [lindex $args 0]] + if [regexp "^xfail " $selector] { + set what "xfail" + } elseif [regexp "^target " $selector] { + set what "target" + } else { + error "syntax error in target selector \"$selector\"" + } + + # Extract the rest of the list, which might be a keyword. + regsub "^${what}" $selector "" rest + set rest [string trim $rest] + + if [is-effective-target-keyword $rest] { + # The selector is an effective target keyword. + if [is-effective-target $rest] { + return [expr { $what == "xfail" ? "F" : "S" }] + } else { + return [expr { $what == "xfail" ? "P" : "N" }] + } + } + + if [string match "{*}" $rest] { + if [selector_expression [lindex $rest 0]] { + return [expr { $what == "xfail" ? "F" : "S" }] + } else { + return [expr { $what == "xfail" ? "P" : "N" }] + } + } + + # The selector is not an effective-target keyword, so process + # the list of target triplets. + return [saved-dg-process-target $selector] + } + + # Intercept calls to the DejaGnu function. In addition to + # processing "target selector" or "xfail selector", handle + # "target selector1 xfail selector2". + + proc dg-process-target { args } { + verbose "replacement dg-process-target: `$args'" 2 + + set selector [string trim [lindex $args 0]] + + # If the argument list contains both 'target' and 'xfail', + # process 'target' and, if that succeeds, process 'xfail'. + if [regexp "^target .* xfail .*" $selector] { + set xfail_index [string first "xfail" $selector] + set xfail_selector [string range $selector $xfail_index end] + set target_selector [string range $selector 0 [expr $xfail_index-1]] + set target_selector [string trim $target_selector] + if { [dg-process-target-1 $target_selector] == "N" } { + return "N" + } + return [dg-process-target-1 $xfail_selector] + + } + return [dg-process-target-1 $selector] + } +} # Define libffi callbacks for dg.exp. @@ -89,6 +270,12 @@ proc libffi-dg-test { prog do_what extra_tool_flags } { return [libffi-dg-test-1 target_compile $prog $do_what $extra_tool_flags] } +proc libffi-dg-prune { target_triplet text } { + # We get this with some qemu emulated systems (eg. ppc64le-linux-gnu) + regsub -all "(^|\n)\[^\n\]*unable to perform all requested operations" $text "" text + return $text +} + proc libffi-init { args } { global gluefile wrap_flags; global srcdir @@ -100,46 +287,42 @@ proc libffi-init { args } { global libffi_link_flags global tool_root_dir global ld_library_path + global compiler_vendor - global using_gcc + if ![info exists blddirffi] { + set blddirffi [pwd]/.. + } - set blddirffi [pwd]/.. verbose "libffi $blddirffi" - # Are we building with GCC? - set tmp [grep ../config.status "GCC='yes'"] - if { [string match $tmp "GCC='yes'"] } { - - set using_gcc "yes" - - set gccdir [lookfor_file $tool_root_dir gcc/libgcc.a] - if {$gccdir != ""} { - set gccdir [file dirname $gccdir] - } - verbose "gccdir $gccdir" - - set ld_library_path "." - append ld_library_path ":${gccdir}" + # Which compiler are we building with? + set tmp [grep "$blddirffi/config.log" "^ax_cv_c_compiler_vendor.*$"] + regexp -- {^[^=]*=(.*)$} $tmp nil compiler_vendor - set compiler "${gccdir}/xgcc" - if { [is_remote host] == 0 && [which $compiler] != 0 } { - foreach i "[exec $compiler --print-multi-lib]" { - set mldir "" - regexp -- "\[a-z0-9=_/\.-\]*;" $i mldir - set mldir [string trimright $mldir "\;@"] - if { "$mldir" == "." } { - continue - } - if { [llength [glob -nocomplain ${gccdir}/${mldir}/libgcc_s*.so.*]] >= 1 } { - append ld_library_path ":${gccdir}/${mldir}" + if { [string match $compiler_vendor "gnu"] } { + set gccdir [lookfor_file $tool_root_dir gcc/libgcc.a] + if {$gccdir != ""} { + set gccdir [file dirname $gccdir] + } + verbose "gccdir $gccdir" + + set ld_library_path "." + append ld_library_path ":${gccdir}" + + set compiler "${gccdir}/xgcc" + if { [is_remote host] == 0 && [which $compiler] != 0 } { + foreach i "[exec $compiler --print-multi-lib]" { + set mldir "" + regexp -- "\[a-z0-9=_/\.-\]*;" $i mldir + set mldir [string trimright $mldir "\;@"] + if { "$mldir" == "." } { + continue + } + if { [llength [glob -nocomplain ${gccdir}/${mldir}/libgcc_s*.so.*]] >= 1 } { + append ld_library_path ":${gccdir}/${mldir}" + } } - } - } - - } else { - - set using_gcc "no" - + } } # add the library path for libffi. @@ -179,7 +362,7 @@ proc libffi_target_compile { source dest type options } { global libffi_link_flags global libffi_include global target_triplet - + global compiler_vendor if { [target_info needs_status_wrapper]!="" && [info exists gluefile] } { lappend options "libs=${gluefile}" @@ -189,7 +372,7 @@ proc libffi_target_compile { source dest type options } { # TOOL_OPTIONS must come first, so that it doesn't override testcase # specific options. if [info exists TOOL_OPTIONS] { - lappend options [concat "additional_flags=$TOOL_OPTIONS" $options]; + lappend options "additional_flags=$TOOL_OPTIONS" } # search for ffi_mips.h in srcdir, too @@ -222,14 +405,43 @@ proc libffi_target_compile { source dest type options } { lappend options "libs= -lpthread" } +<<<<<<< HEAD if { [string match "*.cc" $source] } { lappend options "c++" +======= + # this may be required for g++, but just confused clang. + if { [string match "*.cc" $source] } { + lappend options "c++" + } + + if { [string match "arc*-*-linux*" $target_triplet] } { + lappend options "libs= -lpthread" +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } verbose "options: $options" return [target_compile $source $dest $type $options] } +# TEST should be a preprocessor condition. Returns true if it holds. +proc libffi_feature_test { test } { + set src "ffitest[pid].c" + + set f [open $src "w"] + puts $f "#include " + puts $f $test + puts $f "/* OK */" + puts $f "#else" + puts $f "# error Failed $test" + puts $f "#endif" + close $f + + set lines [libffi_target_compile $src /dev/null assembly ""] + file delete $src + + return [string match "" $lines] +} + # Utility routines. # @@ -278,6 +490,7 @@ proc libffi-dg-runtest { testcases default-extra-flags } { } proc run-many-tests { testcases extra_flags } { +<<<<<<< HEAD global using_gcc if { [string match $using_gcc "yes"] } { set common "-W -Wall -Wno-psabi" @@ -291,12 +504,60 @@ proc run-many-tests { testcases extra_flags } { set targetabis { "" } if [string match $using_gcc "yes"] { if [istarget "i?86-*-*"] { +======= + global compiler_vendor + global has_gccbug + global env + switch $compiler_vendor { + "clang" { + set common "-W -Wall" + if [info exists env(LIBFFI_TEST_OPTIMIZATION)] { + set optimizations [ list $env(LIBFFI_TEST_OPTIMIZATION) ] + } else { + set optimizations { "-O0" "-O2" } + } + } + "gnu" { + set common "-W -Wall -Wno-psabi" + if [info exists env(LIBFFI_TEST_OPTIMIZATION)] { + set optimizations [ list $env(LIBFFI_TEST_OPTIMIZATION) ] + } else { + set optimizations { "-O0" "-O2" } + } + } + default { + # Assume we are using the vendor compiler. + set common "" + if [info exists env(LIBFFI_TEST_OPTIMIZATION)] { + set optimizations [ list $env(LIBFFI_TEST_OPTIMIZATION) ] + } else { + set optimizations { "" } + } + } + } + + info exists env(LD_LIBRARY_PATH) + + set targetabis { "" } + if [string match $compiler_vendor "gnu"] { + if [libffi_feature_test "#ifdef __i386__"] { +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 set targetabis { "" "-DABI_NUM=FFI_STDCALL -DABI_ATTR=__STDCALL__" "-DABI_NUM=FFI_THISCALL -DABI_ATTR=__THISCALL__" "-DABI_NUM=FFI_FASTCALL -DABI_ATTR=__FASTCALL__" } +<<<<<<< HEAD +======= + } elseif { [istarget "x86_64-*-*"] \ + && [libffi_feature_test "#if !defined __ILP32__ \ + && !defined __i386__"] } { + set targetabis { + "" + "-DABI_NUM=FFI_GNUW64 -DABI_ATTR=__MSABI__" + } +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } } @@ -311,8 +572,27 @@ proc run-many-tests { testcases extra_flags } { foreach opt $optimizations { foreach abi $abis { set options [concat $common $opt $abi] +<<<<<<< HEAD verbose "Testing $testname, $options" 1 dg-test $test $options "" +======= + set has_gccbug false; + if { [string match $compiler_vendor "gnu"] \ + && [string match "*MSABI*" $abi] \ + && ( ( [string match "*DGTEST=57 *" $common] \ + && [string match "*call.c*" $testname] ) \ + || ( [string match "*DGTEST=54 *" $common] \ + && [string match "*callback*" $testname] ) \ + || [string match "*DGTEST=55 *" $common] \ + || [string match "*DGTEST=56 *" $common] ) } then { + if [libffi_feature_test "#if (__GNUC__ < 9) || ((__GNUC__ == 9) && (__GNUC_MINOR__ < 3))"] { + set has_gccbug true; + } + } + verbose "Testing $testname, $options" 1 + verbose "has_gccbug = $has_gccbug" 1 + dg-test $test $options "" +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 } } } diff --git a/native/libffi/testsuite/libffi.bhaible/Makefile b/native/libffi/testsuite/libffi.bhaible/Makefile new file mode 100644 index 0000000000..3322de94f7 --- /dev/null +++ b/native/libffi/testsuite/libffi.bhaible/Makefile @@ -0,0 +1,28 @@ +CC = gcc +CFLAGS = -O2 -Wall +prefix = +includedir = $(prefix)/include +libdir = $(prefix)/lib +CPPFLAGS = -I$(includedir) +LDFLAGS = -L$(libdir) -Wl,-rpath,$(libdir) + +all: check-call check-callback + +test-call: test-call.c testcases.c + $(CC) $(CPPFLAGS) $(CFLAGS) $(LDFLAGS) -o test-call test-call.c -lffi + +test-callback: test-callback.c testcases.c + $(CC) $(CPPFLAGS) $(CFLAGS) $(LDFLAGS) -o test-callback test-callback.c -lffi + +check-call: test-call + ./test-call > test-call.out + LC_ALL=C uniq -u < test-call.out > failed-call + test '!' -s failed-call + +check-callback: test-callback + ./test-callback > test-callback.out + LC_ALL=C uniq -u < test-callback.out > failed-callback + test '!' -s failed-callback + +clean: + rm -f test-call test-callback test-call.out test-callback.out failed-call failed-callback diff --git a/native/libffi/testsuite/libffi.bhaible/README b/native/libffi/testsuite/libffi.bhaible/README new file mode 100644 index 0000000000..be8540b64f --- /dev/null +++ b/native/libffi/testsuite/libffi.bhaible/README @@ -0,0 +1,78 @@ +This package contains a test suite for libffi. + +This test suite can be compiled with a C compiler. No need for 'expect' +or some other package that is often not installed. + +The test suite consists of 81 C functions, each with a different signature. +* test-call verifies that calling each function directly produces the same + results as calling the function indirectly through 'ffi_call'. +* test-callback verifies that calling each function directly produces the same + results as calling a function that is a callback (object build by + 'ffi_prep_closure_loc') and simulates the original function. + +Each direct or indirect invocation should produce one line of output to +stdout. A correct output consists of paired lines, such as + +void f(void): +void f(void): +int f(void):->99 +int f(void):->99 +int f(int):(1)->2 +int f(int):(1)->2 +int f(2*int):(1,2)->3 +int f(2*int):(1,2)->3 +... + +The Makefile then creates two files: +* failed-call, which consists of the non-paired lines of output of + 'test-call', +* failed-callback, which consists of the non-paired lines of output of + 'test-callback'. + +The test suite passes if both failed-call and failed-callback come out +as empty. + + +How to use the test suite +------------------------- + +1. Modify the Makefile's variables + prefix = the directory in which libffi was installed + CC = the C compiler, often with options such as "-m32" or "-m64" + that enforce a certain ABI, + CFLAGS = optimization options (need to change them only for non-GCC + compilers) +2. Run "make". If it fails already in "test-call", run also + "make check-callback". +3. If this failed, inspect the output files. + + +How to interpret the results +---------------------------- + +The failed-call and failed-callback files consist of paired lines: +The first line is the result of the direct invocation. +The second line is the result of invocation through libffi. + +For example, this output + +uchar f(uchar,ushort,uint,ulong):(97,2,3,4)->255 +uchar f(uchar,ushort,uint,ulong):(97,2,3,4)->0 + +indicates that the arguments were passed correctly, but the return +value came out wrong. + +And this output + +float f(17*float,3*int,L):(0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.1,1.2,1.3,1.4,1.5,1.6,1.7,1.8,6,7,8,561,1105,1729,2465,2821,6601)->15319.1 +float f(17*float,3*int,L):(0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.1,1.2,1.3,1.4,1.5,1.6,1.7,1.8,-140443648,10,268042216,-72537980,-140443648,-140443648,-140443648,-140443648,-140443648)->-6.47158e+08 + +indicates that integer arguments that come after 17 floating-point arguments +were not passed correctly. + + +Credits +------- + +The test suite is based on the one of GNU libffcall-2.0. +Authors: Bill Triggs, Bruno Haible diff --git a/native/libffi/testsuite/libffi.bhaible/alignof.h b/native/libffi/testsuite/libffi.bhaible/alignof.h new file mode 100644 index 0000000000..00604a513a --- /dev/null +++ b/native/libffi/testsuite/libffi.bhaible/alignof.h @@ -0,0 +1,50 @@ +/* Determine alignment of types. + Copyright (C) 2003-2004, 2006, 2009-2017 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, see . */ + +#ifndef _ALIGNOF_H +#define _ALIGNOF_H + +#include + +/* alignof_slot (TYPE) + Determine the alignment of a structure slot (field) of a given type, + at compile time. Note that the result depends on the ABI. + This is the same as alignof (TYPE) and _Alignof (TYPE), defined in + if __alignof_is_defined is 1. + Note: The result cannot be used as a value for an 'enum' constant, + due to bugs in HP-UX 10.20 cc and AIX 3.2.5 xlc. */ +#if defined __cplusplus + template struct alignof_helper { char __slot1; type __slot2; }; +# define alignof_slot(type) offsetof (alignof_helper, __slot2) +#else +# define alignof_slot(type) offsetof (struct { char __slot1; type __slot2; }, __slot2) +#endif + +/* alignof_type (TYPE) + Determine the good alignment of an object of the given type at compile time. + Note that this is not necessarily the same as alignof_slot(type). + For example, with GNU C on x86 platforms: alignof_type(double) = 8, but + - when -malign-double is not specified: alignof_slot(double) = 4, + - when -malign-double is specified: alignof_slot(double) = 8. + Note: The result cannot be used as a value for an 'enum' constant, + due to bugs in HP-UX 10.20 cc and AIX 3.2.5 xlc. */ +#if defined __GNUC__ || defined __IBM__ALIGNOF__ +# define alignof_type __alignof__ +#else +# define alignof_type alignof_slot +#endif + +#endif /* _ALIGNOF_H */ diff --git a/native/libffi/testsuite/libffi.bhaible/bhaible.exp b/native/libffi/testsuite/libffi.bhaible/bhaible.exp new file mode 100644 index 0000000000..44aebc5e06 --- /dev/null +++ b/native/libffi/testsuite/libffi.bhaible/bhaible.exp @@ -0,0 +1,63 @@ +# Copyright (C) 2003, 2006, 2009, 2010, 2014, 2018 Free Software Foundation, Inc. + +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; see the file COPYING3. If not see +# . + +dg-init +libffi-init + +global srcdir subdir +global compiler_vendor + +# The conversion of this testsuite into a dejagnu compatible testsuite +# was done in a pretty lazy fashion, and requires the use of compiler +# flags to disable warnings for now. +if { [string match $compiler_vendor "gnu"] } { + set warning_options "-Wno-unused-variable -Wno-unused-parameter -Wno-unused-but-set-variable -Wno-uninitialized"; +} +if { [string match $compiler_vendor "microsoft"] } { + # -wd4996 suggest use of vsprintf_s instead of vsprintf + # -wd4116 unnamed type definition + # -wd4101 unreferenced local variable + # -wd4244 warning about implicit double to float conversion + set warning_options "-wd4996 -wd4116 -wd4101 -wd4244"; +} +if { ![string match $compiler_vendor "microsoft"] && ![string match $compiler_vendor "gnu"] } { + set warning_options "-Wno-unused-variable -Wno-unused-parameter -Wno-uninitialized"; +} + + +set tlist [lsort [glob -nocomplain -- $srcdir/$subdir/test-call.c]] + +for {set i 1} {$i < 82} {incr i} { + run-many-tests $tlist [format "-DDGTEST=%d %s" $i $warning_options] +} + +set tlist [lsort [glob -nocomplain -- $srcdir/$subdir/test-callback.c]] + +for {set i 1} {$i < 81} {incr i} { + if { [libffi_feature_test "#if FFI_CLOSURES"] } { + run-many-tests $tlist [format "-DDGTEST=%d %s" $i $warning_options] + } else { + foreach test $tlist { + unsupported [format "%s -DDGTEST=%d %s" $test $i $warning_options] + } + } +} + +dg-finish + +# Local Variables: +# tcl-indent-level:4 +# End: diff --git a/native/libffi/testsuite/libffi.bhaible/test-call.c b/native/libffi/testsuite/libffi.bhaible/test-call.c new file mode 100644 index 0000000000..cf9219e266 --- /dev/null +++ b/native/libffi/testsuite/libffi.bhaible/test-call.c @@ -0,0 +1,1745 @@ +/** + Copyright 1993 Bill Triggs + Copyright 1995-2017 Bruno Haible + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . +**/ + +/* { dg-do run { xfail gccbug } } */ + +#include +#include +#include +#include +#include "alignof.h" +#include + +/* libffi testsuite local changes -------------------------------- */ +#ifdef DGTEST +/* Redefine exit(1) as a test failure */ +#define exit(V) (void)((V) ? (abort(), 1) : exit(0)) +int count = 0; +char rbuf1[2048]; +char rbuf2[2048]; +int _fprintf(FILE *stream, const char *format, ...) +{ + va_list args; + va_start(args, format); + + switch (count++) + { + case 0: + case 1: + vsprintf(&rbuf1[strlen(rbuf1)], format, args); + break; + case 2: + printf("%s", rbuf1); + vsprintf(rbuf2, format, args); + break; + case 3: + vsprintf(&rbuf2[strlen(rbuf2)], format, args); + printf("%s", rbuf2); + if (strcmp (rbuf1, rbuf2)) abort(); + break; + } + + va_end(args); + + return 0; +} +#define fprintf _fprintf +#endif +/* --------------------------------------------------------------- */ + +#include "testcases.c" + +#ifndef ABI_NUM +#define ABI_NUM FFI_DEFAULT_ABI +#endif + +/* Definitions that ought to be part of libffi. */ +static ffi_type ffi_type_char; +#define ffi_type_slonglong ffi_type_sint64 +#define ffi_type_ulonglong ffi_type_uint64 + +/* libffi does not support arrays inside structs. */ +#define SKIP_EXTRA_STRUCTS + +#define FFI_PREP_CIF(cif,argtypes,rettype) \ + if (ffi_prep_cif(&(cif),ABI_NUM,sizeof(argtypes)/sizeof(argtypes[0]),&rettype,argtypes) != FFI_OK) abort() +#define FFI_PREP_CIF_NOARGS(cif,rettype) \ + if (ffi_prep_cif(&(cif),ABI_NUM,0,&rettype,NULL) != FFI_OK) abort() +#define FFI_CALL(cif,fn,args,retaddr) \ + ffi_call(&(cif),(void(*)(void))(fn),retaddr,args) + +long clear_traces_i (long a, long b, long c, long d, long e, long f, long g, long h, + long i, long j, long k, long l, long m, long n, long o, long p) +{ return 0; } +float clear_traces_f (float a, float b, float c, float d, float e, float f, float g, + float h, float i, float j, float k, float l, float m, float n, + float o, float p) +{ return 0.0; } +double clear_traces_d (double a, double b, double c, double d, double e, double f, double g, + double h, double i, double j, double k, double l, double m, double n, + double o, double p) +{ return 0.0; } +J clear_traces_J (void) +{ J j; j.l1 = j.l2 = 0; return j; } +void clear_traces (void) +{ clear_traces_i(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0); + clear_traces_f(0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0); + clear_traces_d(0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0); + clear_traces_J(); +} + +void + void_tests (void) +{ +#if (!defined(DGTEST)) || DGTEST == 1 + v_v(); + clear_traces(); + { + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_void); + { + FFI_CALL(cif,v_v,NULL,NULL); + } + } +#endif + return; +} +void + int_tests (void) +{ + int ir; + ffi_arg retvalue; +#if (!defined(DGTEST)) || DGTEST == 2 + ir = i_v(); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + { + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_sint); + { + FFI_CALL(cif,i_v,NULL,&retvalue); + ir = retvalue; + } + } + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 3 + ir = i_i(i1); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_sint); + { + /*const*/ void* args[] = { &i1 }; + FFI_CALL(cif,i_i,args,&retvalue); + ir = retvalue; + } + } + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 4 + ir = i_i2(i1,i2); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_sint); + { + /*const*/ void* args[] = { &i1, &i2 }; + FFI_CALL(cif,i_i2,args,&retvalue); + ir = retvalue; + } + } + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 5 + ir = i_i4(i1,i2,i3,i4); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_sint); + { + /*const*/ void* args[] = { &i1, &i2, &i3, &i4 }; + FFI_CALL(cif,i_i4,args,&retvalue); + ir = retvalue; + } + } + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 6 + ir = i_i8(i1,i2,i3,i4,i5,i6,i7,i8); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_sint); + { + /*const*/ void* args[] = { &i1, &i2, &i3, &i4, &i5, &i6, &i7, &i8 }; + FFI_CALL(cif,i_i8,args,&retvalue); + ir = retvalue; + } + } + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 7 + ir = i_i16(i1,i2,i3,i4,i5,i6,i7,i8,i9,i10,i11,i12,i13,i14,i15,i16); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_sint); + { + /*const*/ void* args[] = { &i1, &i2, &i3, &i4, &i5, &i6, &i7, &i8, &i9, &i10, &i11, &i12, &i13, &i14, &i15, &i16 }; + FFI_CALL(cif,i_i16,args,&retvalue); + ir = retvalue; + } + } + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + + return; +} +void + float_tests (void) +{ + float fr; + +#if (!defined(DGTEST)) || DGTEST == 8 + fr = f_f(f1); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1 }; + FFI_CALL(cif,f_f,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 9 + fr = f_f2(f1,f2); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2 }; + FFI_CALL(cif,f_f2,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 10 + fr = f_f4(f1,f2,f3,f4); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &f3, &f4 }; + FFI_CALL(cif,f_f4,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 11 + fr = f_f8(f1,f2,f3,f4,f5,f6,f7,f8); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &f3, &f4, &f5, &f6, &f7, &f8 }; + FFI_CALL(cif,f_f8,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 12 + fr = f_f16(f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,f14,f15,f16); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &f3, &f4, &f5, &f6, &f7, &f8, &f9, &f10, &f11, &f12, &f13, &f14, &f15, &f16 }; + FFI_CALL(cif,f_f16,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 13 + fr = f_f24(f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,f14,f15,f16,f17,f18,f19,f20,f21,f22,f23,f24); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &f3, &f4, &f5, &f6, &f7, &f8, &f9, &f10, &f11, &f12, &f13, &f14, &f15, &f16, &f17, &f18, &f19, &f20, &f21, &f22, &f23, &f24 }; + FFI_CALL(cif,f_f24,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif +} +void + double_tests (void) +{ + double dr; + +#if (!defined(DGTEST)) || DGTEST == 14 + + dr = d_d(d1); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1 }; + FFI_CALL(cif,d_d,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 15 + dr = d_d2(d1,d2); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2 }; + FFI_CALL(cif,d_d2,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 16 + dr = d_d4(d1,d2,d3,d4); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2, &d3, &d4 }; + FFI_CALL(cif,d_d4,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 17 + dr = d_d8(d1,d2,d3,d4,d5,d6,d7,d8); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2, &d3, &d4, &d5, &d6, &d7, &d8 }; + FFI_CALL(cif,d_d8,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 18 + dr = d_d16(d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,d13,d14,d15,d16); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2, &d3, &d4, &d5, &d6, &d7, &d8, &d9, &d10, &d11, &d12, &d13, &d14, &d15, &d16 }; + FFI_CALL(cif,d_d16,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + return; +} +void + pointer_tests (void) +{ + void* vpr; + +#if (!defined(DGTEST)) || DGTEST == 19 + vpr = vp_vpdpcpsp(&uc1,&d2,str3,&I4); + fprintf(out,"->0x%p\n",vpr); + fflush(out); + vpr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_pointer, &ffi_type_pointer, &ffi_type_pointer, &ffi_type_pointer }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_pointer); + { + void* puc1 = &uc1; + void* pd2 = &d2; + void* pstr3 = str3; + void* pI4 = &I4; + /*const*/ void* args[] = { &puc1, &pd2, &pstr3, &pI4 }; + FFI_CALL(cif,vp_vpdpcpsp,args,&vpr); + } + } + fprintf(out,"->0x%p\n",vpr); + fflush(out); +#endif + return; +} +void + mixed_number_tests (void) +{ + uchar ucr; + ushort usr; + float fr; + double dr; + long long llr; + + /* Unsigned types. + */ +#if (!defined(DGTEST)) || DGTEST == 20 + ucr = uc_ucsil(uc1, us2, ui3, ul4); + fprintf(out,"->%u\n",ucr); + fflush(out); + ucr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_uchar, &ffi_type_ushort, &ffi_type_uint, &ffi_type_ulong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_uchar); + { + ffi_arg r; + /*const*/ void* args[] = { &uc1, &us2, &ui3, &ul4 }; + FFI_CALL(cif,uc_ucsil,args,&r); + ucr = (uchar) r; + } + } + fprintf(out,"->%u\n",ucr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 21 + /* Mixed int & float types. + */ + dr = d_iidd(i1,i2,d3,d4); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &i1, &i2, &d3, &d4 }; + FFI_CALL(cif,d_iidd,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 22 + dr = d_iiidi(i1,i2,i3,d4,i5); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &i1, &i2, &i3, &d4, &i5 }; + FFI_CALL(cif,d_iiidi,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 23 + dr = d_idid(i1,d2,i3,d4); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_double, &ffi_type_sint, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &i1, &d2, &i3, &d4 }; + FFI_CALL(cif,d_idid,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 24 + dr = d_fdi(f1,d2,i3); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &f1, &d2, &i3 }; + FFI_CALL(cif,d_fdi,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 25 + usr = us_cdcd(c1,d2,c3,d4); + fprintf(out,"->%u\n",usr); + fflush(out); + usr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_char, &ffi_type_double, &ffi_type_char, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_ushort); + { + ffi_arg rint; + /*const*/ void* args[] = { &c1, &d2, &c3, &d4 }; + FFI_CALL(cif,us_cdcd,args,&rint); + usr = (ushort) rint; + } + } + fprintf(out,"->%u\n",usr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 26 + /* Long long types. + */ + llr = ll_iiilli(i1,i2,i3,ll1,i13); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_slonglong, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + { + /*const*/ void* args[] = { &i1, &i2, &i3, &ll1, &i13 }; + FFI_CALL(cif,ll_iiilli,args,&llr); + } + } + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 27 + llr = ll_flli(f13,ll1,i13); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_slonglong, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + { + /*const*/ void* args[] = { &f13, &ll1, &i13 }; + FFI_CALL(cif,ll_flli,args,&llr); + } + } + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 28 + fr = f_fi(f1,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &i9 }; + FFI_CALL(cif,f_fi,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 29 + fr = f_f2i(f1,f2,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &i9 }; + FFI_CALL(cif,f_f2i,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 30 + fr = f_f3i(f1,f2,f3,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &f3, &i9 }; + FFI_CALL(cif,f_f3i,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 31 + fr = f_f4i(f1,f2,f3,f4,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &f3, &f4, &i9 }; + FFI_CALL(cif,f_f4i,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 32 + fr = f_f7i(f1,f2,f3,f4,f5,f6,f7,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &f3, &f4, &f5, &f6, &f7, &i9 }; + FFI_CALL(cif,f_f7i,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 33 + fr = f_f8i(f1,f2,f3,f4,f5,f6,f7,f8,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &f3, &f4, &f5, &f6, &f7, &f8, &i9 }; + FFI_CALL(cif,f_f8i,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 34 + fr = f_f12i(f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &f3, &f4, &f5, &f6, &f7, &f8, &f9, &f10, &f11, &f12, &i9 }; + FFI_CALL(cif,f_f12i,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 35 + fr = f_f13i(f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &f3, &f4, &f5, &f6, &f7, &f8, &f9, &f10, &f11, &f12, &f13, &i9 }; + FFI_CALL(cif,f_f13i,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 36 + dr = d_di(d1,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &i9 }; + FFI_CALL(cif,d_di,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 37 + dr = d_d2i(d1,d2,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2, &i9 }; + FFI_CALL(cif,d_d2i,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 38 + dr = d_d3i(d1,d2,d3,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2, &d3, &i9 }; + FFI_CALL(cif,d_d3i,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 39 + dr = d_d4i(d1,d2,d3,d4,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2, &d3, &d4, &i9 }; + FFI_CALL(cif,d_d4i,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 40 + dr = d_d7i(d1,d2,d3,d4,d5,d6,d7,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2, &d3, &d4, &d5, &d6, &d7, &i9 }; + FFI_CALL(cif,d_d7i,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 41 + dr = d_d8i(d1,d2,d3,d4,d5,d6,d7,d8,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2, &d3, &d4, &d5, &d6, &d7, &d8, &i9 }; + FFI_CALL(cif,d_d8i,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 42 + dr = d_d12i(d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2, &d3, &d4, &d5, &d6, &d7, &d8, &d9, &d10, &d11, &d12, &i9 }; + FFI_CALL(cif,d_d12i,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 43 + dr = d_d13i(d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,d13,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2, &d3, &d4, &d5, &d6, &d7, &d8, &d9, &d10, &d11, &d12, &d13, &i9 }; + FFI_CALL(cif,d_d13i,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + return; +} +void + small_structure_return_tests (void) +{ +#if (!defined(DGTEST)) || DGTEST == 44 + { + Size1 r = S1_v(); + fprintf(out,"->{%c}\n",r.x1); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + { + ffi_type* ffi_type_Size1_elements[] = { &ffi_type_char, NULL }; + ffi_type ffi_type_Size1; + ffi_type_Size1.type = FFI_TYPE_STRUCT; + ffi_type_Size1.size = sizeof(Size1); + ffi_type_Size1.alignment = alignof_slot(Size1); + ffi_type_Size1.elements = ffi_type_Size1_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size1); + { + FFI_CALL(cif,S1_v,NULL,&r); + } + } + fprintf(out,"->{%c}\n",r.x1); + fflush(out); + } +#endif +#if (!defined(DGTEST)) || DGTEST == 45 + { + Size2 r = S2_v(); + fprintf(out,"->{%c%c}\n",r.x1,r.x2); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + { + ffi_type* ffi_type_Size2_elements[] = { &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size2; + ffi_type_Size2.type = FFI_TYPE_STRUCT; + ffi_type_Size2.size = sizeof(Size2); + ffi_type_Size2.alignment = alignof_slot(Size2); + ffi_type_Size2.elements = ffi_type_Size2_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size2); + { + FFI_CALL(cif,S2_v,NULL,&r); + } + } + fprintf(out,"->{%c%c}\n",r.x1,r.x2); + fflush(out); + } +#endif +#if (!defined(DGTEST)) || DGTEST == 46 + { + Size3 r = S3_v(); + fprintf(out,"->{%c%c%c}\n",r.x1,r.x2,r.x3); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + { + ffi_type* ffi_type_Size3_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size3; + ffi_type_Size3.type = FFI_TYPE_STRUCT; + ffi_type_Size3.size = sizeof(Size3); + ffi_type_Size3.alignment = alignof_slot(Size3); + ffi_type_Size3.elements = ffi_type_Size3_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size3); + { + FFI_CALL(cif,S3_v,NULL,&r); + } + } + fprintf(out,"->{%c%c%c}\n",r.x1,r.x2,r.x3); + fflush(out); + } +#endif +#if (!defined(DGTEST)) || DGTEST == 47 + { + Size4 r = S4_v(); + fprintf(out,"->{%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + { + ffi_type* ffi_type_Size4_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size4; + ffi_type_Size4.type = FFI_TYPE_STRUCT; + ffi_type_Size4.size = sizeof(Size4); + ffi_type_Size4.alignment = alignof_slot(Size4); + ffi_type_Size4.elements = ffi_type_Size4_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size4); + { + FFI_CALL(cif,S4_v,NULL,&r); + } + } + fprintf(out,"->{%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4); + fflush(out); + } +#endif +#if (!defined(DGTEST)) || DGTEST == 48 + { + Size7 r = S7_v(); + fprintf(out,"->{%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + { + ffi_type* ffi_type_Size7_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size7; + ffi_type_Size7.type = FFI_TYPE_STRUCT; + ffi_type_Size7.size = sizeof(Size7); + ffi_type_Size7.alignment = alignof_slot(Size7); + ffi_type_Size7.elements = ffi_type_Size7_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size7); + { + FFI_CALL(cif,S7_v,NULL,&r); + } + } + fprintf(out,"->{%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7); + fflush(out); + } +#endif +#if (!defined(DGTEST)) || DGTEST == 49 + { + Size8 r = S8_v(); + fprintf(out,"->{%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + { + ffi_type* ffi_type_Size8_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size8; + ffi_type_Size8.type = FFI_TYPE_STRUCT; + ffi_type_Size8.size = sizeof(Size8); + ffi_type_Size8.alignment = alignof_slot(Size8); + ffi_type_Size8.elements = ffi_type_Size8_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size8); + { + FFI_CALL(cif,S8_v,NULL,&r); + } + } + fprintf(out,"->{%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8); + fflush(out); + } +#endif +#if (!defined(DGTEST)) || DGTEST == 50 + { + Size12 r = S12_v(); + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + { + ffi_type* ffi_type_Size12_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size12; + ffi_type_Size12.type = FFI_TYPE_STRUCT; + ffi_type_Size12.size = sizeof(Size12); + ffi_type_Size12.alignment = alignof_slot(Size12); + ffi_type_Size12.elements = ffi_type_Size12_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size12); + { + FFI_CALL(cif,S12_v,NULL,&r); + } + } + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12); + fflush(out); + } +#endif +#if (!defined(DGTEST)) || DGTEST == 51 + { + Size15 r = S15_v(); + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12,r.x13,r.x14,r.x15); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + { + ffi_type* ffi_type_Size15_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size15; + ffi_type_Size15.type = FFI_TYPE_STRUCT; + ffi_type_Size15.size = sizeof(Size15); + ffi_type_Size15.alignment = alignof_slot(Size15); + ffi_type_Size15.elements = ffi_type_Size15_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size15); + { + FFI_CALL(cif,S15_v,NULL,&r); + } + } + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12,r.x13,r.x14,r.x15); + fflush(out); + } +#endif +#if (!defined(DGTEST)) || DGTEST == 52 + { + Size16 r = S16_v(); + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12,r.x13,r.x14,r.x15,r.x16); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + { + ffi_type* ffi_type_Size16_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size16; + ffi_type_Size16.type = FFI_TYPE_STRUCT; + ffi_type_Size16.size = sizeof(Size16); + ffi_type_Size16.alignment = alignof_slot(Size16); + ffi_type_Size16.elements = ffi_type_Size16_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size16); + { + FFI_CALL(cif,S16_v,NULL,&r); + } + } + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12,r.x13,r.x14,r.x15,r.x16); + fflush(out); + } +#endif +} +void + structure_tests (void) +{ + Int Ir; + Char Cr; + Float Fr; + Double Dr; + J Jr; +#ifndef SKIP_EXTRA_STRUCTS + T Tr; + X Xr; +#endif + +#if (!defined(DGTEST)) || DGTEST == 53 + Ir = I_III(I1,I2,I3); + fprintf(out,"->{%d}\n",Ir.x); + fflush(out); + Ir.x = 0; clear_traces(); + { + ffi_type* ffi_type_Int_elements[] = { &ffi_type_sint, NULL }; + ffi_type ffi_type_Int; + ffi_type_Int.type = FFI_TYPE_STRUCT; + ffi_type_Int.size = sizeof(Int); + ffi_type_Int.alignment = alignof_slot(Int); + ffi_type_Int.elements = ffi_type_Int_elements; + ffi_type* argtypes[] = { &ffi_type_Int, &ffi_type_Int, &ffi_type_Int }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_Int); + { + /*const*/ void* args[] = { &I1, &I2, &I3 }; + FFI_CALL(cif,I_III,args,&Ir); + } + } + fprintf(out,"->{%d}\n",Ir.x); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 54 + Cr = C_CdC(C1,d2,C3); + fprintf(out,"->{'%c'}\n",Cr.x); + fflush(out); + Cr.x = '\0'; clear_traces(); + { + ffi_type* ffi_type_Char_elements[] = { &ffi_type_char, NULL }; + ffi_type ffi_type_Char; + ffi_type_Char.type = FFI_TYPE_STRUCT; + ffi_type_Char.size = sizeof(Char); + ffi_type_Char.alignment = alignof_slot(Char); + ffi_type_Char.elements = ffi_type_Char_elements; + ffi_type* argtypes[] = { &ffi_type_Char, &ffi_type_double, &ffi_type_Char }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_Char); + { + /*const*/ void* args[] = { &C1, &d2, &C3 }; + FFI_CALL(cif,C_CdC,args,&Cr); + } + } + fprintf(out,"->{'%c'}\n",Cr.x); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 55 + Fr = F_Ffd(F1,f2,d3); + fprintf(out,"->{%g}\n",Fr.x); + fflush(out); + Fr.x = 0.0; clear_traces(); + { + ffi_type* ffi_type_Float_elements[] = { &ffi_type_float, NULL }; + ffi_type ffi_type_Float; + ffi_type_Float.type = FFI_TYPE_STRUCT; + ffi_type_Float.size = sizeof(Float); + ffi_type_Float.alignment = alignof_slot(Float); + ffi_type_Float.elements = ffi_type_Float_elements; + ffi_type* argtypes[] = { &ffi_type_Float, &ffi_type_float, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_Float); + { + /*const*/ void* args[] = { &F1, &f2, &d3 }; + FFI_CALL(cif,F_Ffd,args,&Fr); + } + } + fprintf(out,"->{%g}\n",Fr.x); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 56 + Dr = D_fDd(f1,D2,d3); + fprintf(out,"->{%g}\n",Dr.x); + fflush(out); + Dr.x = 0.0; clear_traces(); + { + ffi_type* ffi_type_Double_elements[] = { &ffi_type_double, NULL }; + ffi_type ffi_type_Double; + ffi_type_Double.type = FFI_TYPE_STRUCT; + ffi_type_Double.size = sizeof(Double); + ffi_type_Double.alignment = alignof_slot(Double); + ffi_type_Double.elements = ffi_type_Double_elements; + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_Double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_Double); + { + /*const*/ void* args[] = { &f1, &D2, &d3 }; + FFI_CALL(cif,D_fDd,args,&Dr); + } + } + fprintf(out,"->{%g}\n",Dr.x); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 57 + Dr = D_Dfd(D1,f2,d3); + fprintf(out,"->{%g}\n",Dr.x); + fflush(out); + Dr.x = 0.0; clear_traces(); + { + ffi_type* ffi_type_Double_elements[] = { &ffi_type_double, NULL }; + ffi_type ffi_type_Double; + ffi_type_Double.type = FFI_TYPE_STRUCT; + ffi_type_Double.size = sizeof(Double); + ffi_type_Double.alignment = alignof_slot(Double); + ffi_type_Double.elements = ffi_type_Double_elements; + ffi_type* argtypes[] = { &ffi_type_Double, &ffi_type_float, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_Double); + { + /*const*/ void* args[] = { &D1, &f2, &d3 }; + FFI_CALL(cif,D_Dfd,args,&Dr); + } + } + fprintf(out,"->{%g}\n",Dr.x); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 58 + Jr = J_JiJ(J1,i2,J2); + fprintf(out,"->{%ld,%ld}\n",Jr.l1,Jr.l2); + fflush(out); + Jr.l1 = Jr.l2 = 0; clear_traces(); + { + ffi_type* ffi_type_J_elements[] = { &ffi_type_slong, &ffi_type_slong, NULL }; + ffi_type ffi_type_J; + ffi_type_J.type = FFI_TYPE_STRUCT; + ffi_type_J.size = sizeof(J); + ffi_type_J.alignment = alignof_slot(J); + ffi_type_J.elements = ffi_type_J_elements; + ffi_type* argtypes[] = { &ffi_type_J, &ffi_type_sint, &ffi_type_J }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_J); + { + /*const*/ void* args[] = { &J1, &i2, &J2 }; + FFI_CALL(cif,J_JiJ,args,&Jr); + } + } + fprintf(out,"->{%ld,%ld}\n",Jr.l1,Jr.l2); + fflush(out); +#endif +#ifndef SKIP_EXTRA_STRUCTS +#if (!defined(DGTEST)) || DGTEST == 59 + Tr = T_TcT(T1,' ',T2); + fprintf(out,"->{\"%c%c%c\"}\n",Tr.c[0],Tr.c[1],Tr.c[2]); + fflush(out); + Tr.c[0] = Tr.c[1] = Tr.c[2] = 0; clear_traces(); + { + ffi_type* ffi_type_T_elements[] = { ??, NULL }; + ffi_type ffi_type_T; + ffi_type_T.type = FFI_TYPE_STRUCT; + ffi_type_T.size = sizeof(T); + ffi_type_T.alignment = alignof_slot(T); + ffi_type_T.elements = ffi_type_T_elements; + ffi_type* argtypes[] = { &ffi_type_T, &ffi_type_char, &ffi_type_T }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_T); + { + char space = ' '; + /*const*/ void* args[] = { &T1, &space, &T2 }; + FFI_CALL(cif,T_TcT,args,&Tr); + } + } + fprintf(out,"->{\"%c%c%c\"}\n",Tr.c[0],Tr.c[1],Tr.c[2]); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 60 + Xr = X_BcdB(B1,c2,d3,B2); + fprintf(out,"->{\"%s\",'%c'}\n",Xr.c,Xr.c1); + fflush(out); + Xr.c[0]=Xr.c1='\0'; clear_traces(); + { + ffi_type* ffi_type_X_elements[] = { ??, NULL }; + ffi_type ffi_type_X; + ffi_type_X.type = FFI_TYPE_STRUCT; + ffi_type_X.size = sizeof(X); + ffi_type_X.alignment = alignof_slot(X); + ffi_type_X.elements = ffi_type_X_elements; + ffi_type* argtypes[] = { &ffi_type_X, &ffi_type_char, &ffi_type_double, &ffi_type_X }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_X); + { + /*const*/ void* args[] = { &B1, &c2, &d3, &B2 }; + FFI_CALL(cif,X_BcdB,args,&Xr); + } + } + fprintf(out,"->{\"%s\",'%c'}\n",Xr.c,Xr.c1); + fflush(out); +#endif +#endif + + return; +} + +void + gpargs_boundary_tests (void) +{ + ffi_type* ffi_type_K_elements[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, NULL }; + ffi_type ffi_type_K; + ffi_type* ffi_type_L_elements[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, NULL }; + ffi_type ffi_type_L; + long lr; + long long llr; + float fr; + double dr; + + ffi_type_K.type = FFI_TYPE_STRUCT; + ffi_type_K.size = sizeof(K); + ffi_type_K.alignment = alignof_slot(K); + ffi_type_K.elements = ffi_type_K_elements; + + ffi_type_L.type = FFI_TYPE_STRUCT; + ffi_type_L.size = sizeof(L); + ffi_type_L.alignment = alignof_slot(L); + ffi_type_L.elements = ffi_type_L_elements; + +#if (!defined(DGTEST)) || DGTEST == 61 + lr = l_l0K(K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + { + /*const*/ void* args[] = { &K1, &l9 }; + FFI_CALL(cif,l_l0K,args,&lr); + } + } + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 62 + lr = l_l1K(l1,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + { + /*const*/ void* args[] = { &l1, &K1, &l9 }; + FFI_CALL(cif,l_l1K,args,&lr); + } + } + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 63 + lr = l_l2K(l1,l2,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + { + /*const*/ void* args[] = { &l1, &l2, &K1, &l9 }; + FFI_CALL(cif,l_l2K,args,&lr); + } + } + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 64 + lr = l_l3K(l1,l2,l3,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &K1, &l9 }; + FFI_CALL(cif,l_l3K,args,&lr); + } + } + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 65 + lr = l_l4K(l1,l2,l3,l4,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &l4, &K1, &l9 }; + FFI_CALL(cif,l_l4K,args,&lr); + } + } + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 66 + lr = l_l5K(l1,l2,l3,l4,l5,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &l4, &l5, &K1, &l9 }; + FFI_CALL(cif,l_l5K,args,&lr); + } + } + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 67 + lr = l_l6K(l1,l2,l3,l4,l5,l6,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &l4, &l5, &l6, &K1, &l9 }; + FFI_CALL(cif,l_l6K,args,&lr); + } + } + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 68 + fr = f_f17l3L(f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,f14,f15,f16,f17,l6,l7,l8,L1); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_L }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + { + /*const*/ void* args[] = { &f1, &f2, &f3, &f4, &f5, &f6, &f7, &f8, &f9, &f10, &f11, &f12, &f13, &f14, &f15, &f16, &f17, &l6, &l7, &l8, &L1 }; + FFI_CALL(cif,f_f17l3L,args,&fr); + } + } + fprintf(out,"->%g\n",fr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 69 + dr = d_d17l3L(d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,d13,d14,d15,d16,d17,l6,l7,l8,L1); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_L }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &d1, &d2, &d3, &d4, &d5, &d6, &d7, &d8, &d9, &d10, &d11, &d12, &d13, &d14, &d15, &d16, &d17, &l6, &l7, &l8, &L1 }; + FFI_CALL(cif,d_d17l3L,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 70 + llr = ll_l2ll(l1,l2,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + { + /*const*/ void* args[] = { &l1, &l2, &ll1, &l9 }; + FFI_CALL(cif,ll_l2ll,args,&llr); + } + } + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 71 + llr = ll_l3ll(l1,l2,l3,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &ll1, &l9 }; + FFI_CALL(cif,ll_l3ll,args,&llr); + } + } + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 72 + llr = ll_l4ll(l1,l2,l3,l4,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &l4, &ll1, &l9 }; + FFI_CALL(cif,ll_l4ll,args,&llr); + } + } + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 73 + llr = ll_l5ll(l1,l2,l3,l4,l5,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &l4, &l5, &ll1, &l9 }; + FFI_CALL(cif,ll_l5ll,args,&llr); + } + } + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 74 + llr = ll_l6ll(l1,l2,l3,l4,l5,l6,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &l4, &l5, &l6, &ll1, &l9 }; + FFI_CALL(cif,ll_l6ll,args,&llr); + } + } + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 75 + llr = ll_l7ll(l1,l2,l3,l4,l5,l6,l7,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &l4, &l5, &l6, &l7, &ll1, &l9 }; + FFI_CALL(cif,ll_l7ll,args,&llr); + } + } + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 76 + dr = d_l2d(l1,l2,d2,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &l1, &l2, &d2, &l9 }; + FFI_CALL(cif,d_l2d,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 77 + dr = d_l3d(l1,l2,l3,d2,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &d2, &l9 }; + FFI_CALL(cif,d_l3d,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 78 + dr = d_l4d(l1,l2,l3,l4,d2,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &l4, &d2, &l9 }; + FFI_CALL(cif,d_l4d,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 79 + dr = d_l5d(l1,l2,l3,l4,l5,d2,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &l4, &l5, &d2, &l9 }; + FFI_CALL(cif,d_l5d,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 80 + dr = d_l6d(l1,l2,l3,l4,l5,l6,d2,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &l4, &l5, &l6, &d2, &l9 }; + FFI_CALL(cif,d_l6d,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif +#if (!defined(DGTEST)) || DGTEST == 81 + dr = d_l7d(l1,l2,l3,l4,l5,l6,l7,d2,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + { + /*const*/ void* args[] = { &l1, &l2, &l3, &l4, &l5, &l6, &l7, &d2, &l9 }; + FFI_CALL(cif,d_l7d,args,&dr); + } + } + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + return; +} + +int + main (void) +{ + ffi_type_char = (char)(-1) < 0 ? ffi_type_schar : ffi_type_uchar; + out = stdout; + + void_tests(); + int_tests(); + float_tests(); + double_tests(); + pointer_tests(); + mixed_number_tests(); + small_structure_return_tests(); + structure_tests(); + gpargs_boundary_tests(); + + exit(0); +} diff --git a/native/libffi/testsuite/libffi.bhaible/test-callback.c b/native/libffi/testsuite/libffi.bhaible/test-callback.c new file mode 100644 index 0000000000..0b16799d86 --- /dev/null +++ b/native/libffi/testsuite/libffi.bhaible/test-callback.c @@ -0,0 +1,2885 @@ +/* + * Copyright 1993 Bill Triggs + * Copyright 1995-2017 Bruno Haible + * + * This program is free software: you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* { dg-do run { xfail gccbug } } */ + +#include +#include +#include +#include +#include "alignof.h" +#include + +/* libffi testsuite local changes -------------------------------- */ +#ifdef DGTEST +/* Redefine exit(1) as a test failure */ +#define exit(V) (void)((V) ? (abort(), 1) : exit(0)) +int count = 0; +char rbuf1[2048]; +char rbuf2[2048]; +int _fprintf(FILE *stream, const char *format, ...) +{ + va_list args; + va_start(args, format); + + switch (count++) + { + case 0: + case 1: + vsprintf(&rbuf1[strlen(rbuf1)], format, args); + break; + case 2: + printf("%s", rbuf1); + vsprintf(rbuf2, format, args); + break; + case 3: + vsprintf(&rbuf2[strlen(rbuf2)], format, args); + printf("%s", rbuf2); + if (strcmp (rbuf1, rbuf2)) abort(); + break; + } + + va_end(args); + + return 0; +} +#define fprintf _fprintf +#endif +/* --------------------------------------------------------------- */ + +#include "testcases.c" + +#ifndef ABI_NUM +#define ABI_NUM FFI_DEFAULT_ABI +#endif + +/* Definitions that ought to be part of libffi. */ +static ffi_type ffi_type_char; +#define ffi_type_slonglong ffi_type_sint64 +#define ffi_type_ulonglong ffi_type_uint64 + +/* libffi does not support arrays inside structs. */ +#define SKIP_EXTRA_STRUCTS + +#define FFI_PREP_CIF(cif,argtypes,rettype) \ + if (ffi_prep_cif(&(cif),ABI_NUM,sizeof(argtypes)/sizeof(argtypes[0]),&rettype,argtypes) != FFI_OK) abort() +#define FFI_PREP_CIF_NOARGS(cif,rettype) \ + if (ffi_prep_cif(&(cif),ABI_NUM,0,&rettype,NULL) != FFI_OK) abort() + +#if defined(__sparc__) && defined(__sun) && defined(__SUNPRO_C) /* SUNWspro cc */ +/* SunPRO cc miscompiles the simulator function for X_BcdB: d.i[1] is + * temporarily stored in %l2 and put onto the stack from %l2, but in between + * the copy of X has used %l2 as a counter without saving and restoring its + * value. + */ +#define SKIP_X +#endif +#if defined(__mipsn32__) && !defined(__GNUC__) +/* The X test crashes for an unknown reason. */ +#define SKIP_X +#endif + + +/* These functions simulate the behaviour of the functions defined in testcases.c. */ + +/* void tests */ +void v_v_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&v_v) { fprintf(out,"wrong data for v_v\n"); exit(1); } + fprintf(out,"void f(void):\n"); + fflush(out); +} + +/* int tests */ +void i_v_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&i_v) { fprintf(out,"wrong data for i_v\n"); exit(1); } + {int r=99; + fprintf(out,"int f(void):"); + fflush(out); + *(ffi_arg*)retp = r; +}} +void i_i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&i_i) { fprintf(out,"wrong data for i_i\n"); exit(1); } + int a = *(int*)(*args++); + int r=a+1; + fprintf(out,"int f(int):(%d)",a); + fflush(out); + *(ffi_arg*)retp = r; +} +void i_i2_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&i_i2) { fprintf(out,"wrong data for i_i2\n"); exit(1); } + {int a = *(int*)(*args++); + int b = *(int*)(*args++); + int r=a+b; + fprintf(out,"int f(2*int):(%d,%d)",a,b); + fflush(out); + *(ffi_arg*)retp = r; +}} +void i_i4_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&i_i4) { fprintf(out,"wrong data for i_i4\n"); exit(1); } + {int a = *(int*)(*args++); + int b = *(int*)(*args++); + int c = *(int*)(*args++); + int d = *(int*)(*args++); + int r=a+b+c+d; + fprintf(out,"int f(4*int):(%d,%d,%d,%d)",a,b,c,d); + fflush(out); + *(ffi_arg*)retp = r; +}} +void i_i8_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&i_i8) { fprintf(out,"wrong data for i_i8\n"); exit(1); } + {int a = *(int*)(*args++); + int b = *(int*)(*args++); + int c = *(int*)(*args++); + int d = *(int*)(*args++); + int e = *(int*)(*args++); + int f = *(int*)(*args++); + int g = *(int*)(*args++); + int h = *(int*)(*args++); + int r=a+b+c+d+e+f+g+h; + fprintf(out,"int f(8*int):(%d,%d,%d,%d,%d,%d,%d,%d)",a,b,c,d,e,f,g,h); + fflush(out); + *(ffi_arg*)retp = r; +}} +void i_i16_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&i_i16) { fprintf(out,"wrong data for i_i16\n"); exit(1); } + {int a = *(int*)(*args++); + int b = *(int*)(*args++); + int c = *(int*)(*args++); + int d = *(int*)(*args++); + int e = *(int*)(*args++); + int f = *(int*)(*args++); + int g = *(int*)(*args++); + int h = *(int*)(*args++); + int i = *(int*)(*args++); + int j = *(int*)(*args++); + int k = *(int*)(*args++); + int l = *(int*)(*args++); + int m = *(int*)(*args++); + int n = *(int*)(*args++); + int o = *(int*)(*args++); + int p = *(int*)(*args++); + int r=a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p; + fprintf(out,"int f(16*int):(%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d)", + a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p); + fflush(out); + *(ffi_arg*)retp = r; +}} + +/* float tests */ +void f_f_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f) { fprintf(out,"wrong data for f_f\n"); exit(1); } + {float a = *(float*)(*args++); + float r=a+1.0; + fprintf(out,"float f(float):(%g)",a); + fflush(out); + *(float*)retp = r; +}} +void f_f2_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f2) { fprintf(out,"wrong data for f_f2\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float r=a+b; + fprintf(out,"float f(2*float):(%g,%g)",a,b); + fflush(out); + *(float*)retp = r; +}} +void f_f4_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f4) { fprintf(out,"wrong data for f_f4\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float c = *(float*)(*args++); + float d = *(float*)(*args++); + float r=a+b+c+d; + fprintf(out,"float f(4*float):(%g,%g,%g,%g)",a,b,c,d); + fflush(out); + *(float*)retp = r; +}} +void f_f8_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f8) { fprintf(out,"wrong data for f_f8\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float c = *(float*)(*args++); + float d = *(float*)(*args++); + float e = *(float*)(*args++); + float f = *(float*)(*args++); + float g = *(float*)(*args++); + float h = *(float*)(*args++); + float r=a+b+c+d+e+f+g+h; + fprintf(out,"float f(8*float):(%g,%g,%g,%g,%g,%g,%g,%g)",a,b,c,d,e,f,g,h); + fflush(out); + *(float*)retp = r; +}} +void f_f16_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f16) { fprintf(out,"wrong data for f_f16\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float c = *(float*)(*args++); + float d = *(float*)(*args++); + float e = *(float*)(*args++); + float f = *(float*)(*args++); + float g = *(float*)(*args++); + float h = *(float*)(*args++); + float i = *(float*)(*args++); + float j = *(float*)(*args++); + float k = *(float*)(*args++); + float l = *(float*)(*args++); + float m = *(float*)(*args++); + float n = *(float*)(*args++); + float o = *(float*)(*args++); + float p = *(float*)(*args++); + float r=a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p; + fprintf(out,"float f(16*float):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g)",a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p); + fflush(out); + *(float*)retp = r; +}} +void f_f24_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f24) { fprintf(out,"wrong data for f_f24\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float c = *(float*)(*args++); + float d = *(float*)(*args++); + float e = *(float*)(*args++); + float f = *(float*)(*args++); + float g = *(float*)(*args++); + float h = *(float*)(*args++); + float i = *(float*)(*args++); + float j = *(float*)(*args++); + float k = *(float*)(*args++); + float l = *(float*)(*args++); + float m = *(float*)(*args++); + float n = *(float*)(*args++); + float o = *(float*)(*args++); + float p = *(float*)(*args++); + float q = *(float*)(*args++); + float s = *(float*)(*args++); + float t = *(float*)(*args++); + float u = *(float*)(*args++); + float v = *(float*)(*args++); + float w = *(float*)(*args++); + float x = *(float*)(*args++); + float y = *(float*)(*args++); + float r=a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p+q+s+t+u+v+w+x+y; + fprintf(out,"float f(24*float):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g)",a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,s,t,u,v,w,x,y); + fflush(out); + *(float*)retp = r; +}} + +/* double tests */ +void d_d_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d) { fprintf(out,"wrong data for d_d\n"); exit(1); } + {double a = *(double*)(*args++); + double r=a+1.0; + fprintf(out,"double f(double):(%g)",a); + fflush(out); + *(double*)retp = r; +}} +void d_d2_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d2) { fprintf(out,"wrong data for d_d2\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + double r=a+b; + fprintf(out,"double f(2*double):(%g,%g)",a,b); + fflush(out); + *(double*)retp = r; +}} +void d_d4_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d4) { fprintf(out,"wrong data for d_d4\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + double c = *(double*)(*args++); + double d = *(double*)(*args++); + double r=a+b+c+d; + fprintf(out,"double f(4*double):(%g,%g,%g,%g)",a,b,c,d); + fflush(out); + *(double*)retp = r; +}} +void d_d8_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d8) { fprintf(out,"wrong data for d_d8\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + double c = *(double*)(*args++); + double d = *(double*)(*args++); + double e = *(double*)(*args++); + double f = *(double*)(*args++); + double g = *(double*)(*args++); + double h = *(double*)(*args++); + double r=a+b+c+d+e+f+g+h; + fprintf(out,"double f(8*double):(%g,%g,%g,%g,%g,%g,%g,%g)",a,b,c,d,e,f,g,h); + fflush(out); + *(double*)retp = r; +}} +void d_d16_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d16) { fprintf(out,"wrong data for d_d16\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + double c = *(double*)(*args++); + double d = *(double*)(*args++); + double e = *(double*)(*args++); + double f = *(double*)(*args++); + double g = *(double*)(*args++); + double h = *(double*)(*args++); + double i = *(double*)(*args++); + double j = *(double*)(*args++); + double k = *(double*)(*args++); + double l = *(double*)(*args++); + double m = *(double*)(*args++); + double n = *(double*)(*args++); + double o = *(double*)(*args++); + double p = *(double*)(*args++); + double r=a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p; + fprintf(out,"double f(16*double):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g)",a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p); + fflush(out); + *(double*)retp = r; +}} + +/* pointer tests */ +void vp_vpdpcpsp_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&vp_vpdpcpsp) { fprintf(out,"wrong data for vp_vpdpcpsp\n"); exit(1); } + {void* a = *(void* *)(*args++); + double* b = *(double* *)(*args++); + char* c = *(char* *)(*args++); + Int* d = *(Int* *)(*args++); + void* ret = (char*)b + 1; + fprintf(out,"void* f(void*,double*,char*,Int*):(0x%p,0x%p,0x%p,0x%p)",a,b,c,d); + fflush(out); + *(void* *)retp = ret; +}} + +/* mixed number tests */ +void uc_ucsil_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&uc_ucsil) { fprintf(out,"wrong data for uc_ucsil\n"); exit(1); } + {uchar a = *(unsigned char *)(*args++); + ushort b = *(unsigned short *)(*args++); + uint c = *(unsigned int *)(*args++); + ulong d = *(unsigned long *)(*args++); + uchar r = (uchar)-1; + fprintf(out,"uchar f(uchar,ushort,uint,ulong):(%u,%u,%u,%lu)",a,b,c,d); + fflush(out); + *(ffi_arg *)retp = r; +}} +void d_iidd_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_iidd) { fprintf(out,"wrong data for d_iidd\n"); exit(1); } + {int a = *(int*)(*args++); + int b = *(int*)(*args++); + double c = *(double*)(*args++); + double d = *(double*)(*args++); + double r=a+b+c+d; + fprintf(out,"double f(int,int,double,double):(%d,%d,%g,%g)",a,b,c,d); + fflush(out); + *(double*)retp = r; +}} +void d_iiidi_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_iiidi) { fprintf(out,"wrong data for d_iiidi\n"); exit(1); } + {int a = *(int*)(*args++); + int b = *(int*)(*args++); + int c = *(int*)(*args++); + double d = *(double*)(*args++); + int e = *(int*)(*args++); + double r=a+b+c+d+e; + fprintf(out,"double f(int,int,int,double,int):(%d,%d,%d,%g,%d)",a,b,c,d,e); + fflush(out); + *(double*)retp = r; +}} +void d_idid_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_idid) { fprintf(out,"wrong data for d_idid\n"); exit(1); } + {int a = *(int*)(*args++); + double b = *(double*)(*args++); + int c = *(int*)(*args++); + double d = *(double*)(*args++); + double r=a+b+c+d; + fprintf(out,"double f(int,double,int,double):(%d,%g,%d,%g)",a,b,c,d); + fflush(out); + *(double*)retp = r; +}} +void d_fdi_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_fdi) { fprintf(out,"wrong data for d_fdi\n"); exit(1); } + {float a = *(float*)(*args++); + double b = *(double*)(*args++); + int c = *(int*)(*args++); + double r=a+b+c; + fprintf(out,"double f(float,double,int):(%g,%g,%d)",a,b,c); + fflush(out); + *(double*)retp = r; +}} +void us_cdcd_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&us_cdcd) { fprintf(out,"wrong data for us_cdcd\n"); exit(1); } + {char a = *(char*)(*args++); + double b = *(double*)(*args++); + char c = *(char*)(*args++); + double d = *(double*)(*args++); + ushort r = (ushort)(a + b + c + d); + fprintf(out,"ushort f(char,double,char,double):('%c',%g,'%c',%g)",a,b,c,d); + fflush(out); + *(ffi_arg *)retp = r; +}} +void ll_iiilli_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&ll_iiilli) { fprintf(out,"wrong data for ll_iiilli\n"); exit(1); } + {int a = *(int*)(*args++); + int b = *(int*)(*args++); + int c = *(int*)(*args++); + long long d = *(long long *)(*args++); + int e = *(int*)(*args++); + long long r = (long long)(int)a + (long long)(int)b + (long long)(int)c + d + (long long)e; + fprintf(out,"long long f(int,int,int,long long,int):(%d,%d,%d,0x%lx%08lx,%d)",a,b,c,(long)(d>>32),(long)(d&0xffffffff),e); + fflush(out); + *(long long *)retp = r; +}} +void ll_flli_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&ll_flli) { fprintf(out,"wrong data for ll_flli\n"); exit(1); } + {float a = *(float*)(*args++); + long long b = *(long long *)(*args++); + int c = *(int*)(*args++); + long long r = (long long)(int)a + b + (long long)c; + fprintf(out,"long long f(float,long long,int):(%g,0x%lx%08lx,0x%lx)",a,(long)(b>>32),(long)(b&0xffffffff),(long)c); + fflush(out); + *(long long *)retp = r; +}} +void f_fi_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_fi) { fprintf(out,"wrong data for f_fi\n"); exit(1); } + {float a = *(float*)(*args++); + int z = *(int*)(*args++); + float r=a+z; + fprintf(out,"float f(float,int):(%g,%d)",a,z); + fflush(out); + *(float*)retp = r; +}} +void f_f2i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f2i) { fprintf(out,"wrong data for f_f2i\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + int z = *(int*)(*args++); + float r=a+b+z; + fprintf(out,"float f(2*float,int):(%g,%g,%d)",a,b,z); + fflush(out); + *(float*)retp = r; +}} +void f_f3i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f3i) { fprintf(out,"wrong data for f_f3i\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float c = *(float*)(*args++); + int z = *(int*)(*args++); + float r=a+b+c+z; + fprintf(out,"float f(3*float,int):(%g,%g,%g,%d)",a,b,c,z); + fflush(out); + *(float*)retp = r; +}} +void f_f4i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f4i) { fprintf(out,"wrong data for f_f4i\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float c = *(float*)(*args++); + float d = *(float*)(*args++); + int z = *(int*)(*args++); + float r=a+b+c+d+z; + fprintf(out,"float f(4*float,int):(%g,%g,%g,%g,%d)",a,b,c,d,z); + fflush(out); + *(float*)retp = r; +}} +void f_f7i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f7i) { fprintf(out,"wrong data for f_f7i\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float c = *(float*)(*args++); + float d = *(float*)(*args++); + float e = *(float*)(*args++); + float f = *(float*)(*args++); + float g = *(float*)(*args++); + int z = *(int*)(*args++); + float r=a+b+c+d+e+f+g+z; + fprintf(out,"float f(7*float,int):(%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,z); + fflush(out); + *(float*)retp = r; +}} +void f_f8i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f8i) { fprintf(out,"wrong data for f_f8i\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float c = *(float*)(*args++); + float d = *(float*)(*args++); + float e = *(float*)(*args++); + float f = *(float*)(*args++); + float g = *(float*)(*args++); + float h = *(float*)(*args++); + int z = *(int*)(*args++); + float r=a+b+c+d+e+f+g+h+z; + fprintf(out,"float f(8*float,int):(%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,z); + fflush(out); + *(float*)retp = r; +}} +void f_f12i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f12i) { fprintf(out,"wrong data for f_f12i\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float c = *(float*)(*args++); + float d = *(float*)(*args++); + float e = *(float*)(*args++); + float f = *(float*)(*args++); + float g = *(float*)(*args++); + float h = *(float*)(*args++); + float i = *(float*)(*args++); + float j = *(float*)(*args++); + float k = *(float*)(*args++); + float l = *(float*)(*args++); + int z = *(int*)(*args++); + float r=a+b+c+d+e+f+g+h+i+j+k+l+z; + fprintf(out,"float f(12*float,int):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,i,j,k,l,z); + fflush(out); + *(float*)retp = r; +}} +void f_f13i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f13i) { fprintf(out,"wrong data for f_f13i\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float c = *(float*)(*args++); + float d = *(float*)(*args++); + float e = *(float*)(*args++); + float f = *(float*)(*args++); + float g = *(float*)(*args++); + float h = *(float*)(*args++); + float i = *(float*)(*args++); + float j = *(float*)(*args++); + float k = *(float*)(*args++); + float l = *(float*)(*args++); + float m = *(float*)(*args++); + int z = *(int*)(*args++); + float r=a+b+c+d+e+f+g+h+i+j+k+l+m+z; + fprintf(out,"float f(13*float,int):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,i,j,k,l,m,z); + fflush(out); + *(float*)retp = r; +}} +void d_di_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_di) { fprintf(out,"wrong data for d_di\n"); exit(1); } + {double a = *(double*)(*args++); + int z = *(int*)(*args++); + double r=a+z; + fprintf(out,"double f(double,int):(%g,%d)",a,z); + fflush(out); + *(double*)retp = r; +}} +void d_d2i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d2i) { fprintf(out,"wrong data for d_d2i\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + int z = *(int*)(*args++); + double r=a+b+z; + fprintf(out,"double f(2*double,int):(%g,%g,%d)",a,b,z); + fflush(out); + *(double*)retp = r; +}} +void d_d3i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d3i) { fprintf(out,"wrong data for d_d3i\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + double c = *(double*)(*args++); + int z = *(int*)(*args++); + double r=a+b+c+z; + fprintf(out,"double f(3*double,int):(%g,%g,%g,%d)",a,b,c,z); + fflush(out); + *(double*)retp = r; +}} +void d_d4i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d4i) { fprintf(out,"wrong data for d_d4i\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + double c = *(double*)(*args++); + double d = *(double*)(*args++); + int z = *(int*)(*args++); + double r=a+b+c+d+z; + fprintf(out,"double f(4*double,int):(%g,%g,%g,%g,%d)",a,b,c,d,z); + fflush(out); + *(double*)retp = r; +}} +void d_d7i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d7i) { fprintf(out,"wrong data for d_d7i\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + double c = *(double*)(*args++); + double d = *(double*)(*args++); + double e = *(double*)(*args++); + double f = *(double*)(*args++); + double g = *(double*)(*args++); + int z = *(int*)(*args++); + double r=a+b+c+d+e+f+g+z; + fprintf(out,"double f(7*double,int):(%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,z); + fflush(out); + *(double*)retp = r; +}} +void d_d8i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d8i) { fprintf(out,"wrong data for d_d8i\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + double c = *(double*)(*args++); + double d = *(double*)(*args++); + double e = *(double*)(*args++); + double f = *(double*)(*args++); + double g = *(double*)(*args++); + double h = *(double*)(*args++); + int z = *(int*)(*args++); + double r=a+b+c+d+e+f+g+h+z; + fprintf(out,"double f(8*double,int):(%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,z); + fflush(out); + *(double*)retp = r; +}} +void d_d12i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d12i) { fprintf(out,"wrong data for d_d12i\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + double c = *(double*)(*args++); + double d = *(double*)(*args++); + double e = *(double*)(*args++); + double f = *(double*)(*args++); + double g = *(double*)(*args++); + double h = *(double*)(*args++); + double i = *(double*)(*args++); + double j = *(double*)(*args++); + double k = *(double*)(*args++); + double l = *(double*)(*args++); + int z = *(int*)(*args++); + double r=a+b+c+d+e+f+g+h+i+j+k+l+z; + fprintf(out,"double f(12*double,int):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,i,j,k,l,z); + fflush(out); + *(double*)retp = r; +}} +void d_d13i_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d13i) { fprintf(out,"wrong data for d_d13i\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + double c = *(double*)(*args++); + double d = *(double*)(*args++); + double e = *(double*)(*args++); + double f = *(double*)(*args++); + double g = *(double*)(*args++); + double h = *(double*)(*args++); + double i = *(double*)(*args++); + double j = *(double*)(*args++); + double k = *(double*)(*args++); + double l = *(double*)(*args++); + double m = *(double*)(*args++); + int z = *(int*)(*args++); + double r=a+b+c+d+e+f+g+h+i+j+k+l+m+z; + fprintf(out,"double f(13*double,int):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,i,j,k,l,m,z); + fflush(out); + *(double*)retp = r; +}} + +/* small structure return tests */ +void S1_v_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&S1_v) { fprintf(out,"wrong data for S1_v\n"); exit(1); } + {Size1 r = Size1_1; + fprintf(out,"Size1 f(void):"); + fflush(out); + *(Size1*)retp = r; +}} +void S2_v_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&S2_v) { fprintf(out,"wrong data for S2_v\n"); exit(1); } + {Size2 r = Size2_1; + fprintf(out,"Size2 f(void):"); + fflush(out); + *(Size2*)retp = r; +}} +void S3_v_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&S3_v) { fprintf(out,"wrong data for S3_v\n"); exit(1); } + {Size3 r = Size3_1; + fprintf(out,"Size3 f(void):"); + fflush(out); + *(Size3*)retp = r; +}} +void S4_v_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&S4_v) { fprintf(out,"wrong data for S4_v\n"); exit(1); } + {Size4 r = Size4_1; + fprintf(out,"Size4 f(void):"); + fflush(out); + *(Size4*)retp = r; +}} +void S7_v_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&S7_v) { fprintf(out,"wrong data for S7_v\n"); exit(1); } + {Size7 r = Size7_1; + fprintf(out,"Size7 f(void):"); + fflush(out); + *(Size7*)retp = r; +}} +void S8_v_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&S8_v) { fprintf(out,"wrong data for S8_v\n"); exit(1); } + {Size8 r = Size8_1; + fprintf(out,"Size8 f(void):"); + fflush(out); + *(Size8*)retp = r; +}} +void S12_v_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&S12_v) { fprintf(out,"wrong data for S12_v\n"); exit(1); } + {Size12 r = Size12_1; + fprintf(out,"Size12 f(void):"); + fflush(out); + *(Size12*)retp = r; +}} +void S15_v_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&S15_v) { fprintf(out,"wrong data for S15_v\n"); exit(1); } + {Size15 r = Size15_1; + fprintf(out,"Size15 f(void):"); + fflush(out); + *(Size15*)retp = r; +}} +void S16_v_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&S16_v) { fprintf(out,"wrong data for S16_v\n"); exit(1); } + {Size16 r = Size16_1; + fprintf(out,"Size16 f(void):"); + fflush(out); + *(Size16*)retp = r; +}} + +/* structure tests */ +void I_III_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&I_III) { fprintf(out,"wrong data for I_III\n"); exit(1); } + {Int a = *(Int*)(*args++); + Int b = *(Int*)(*args++); + Int c = *(Int*)(*args++); + Int r; + r.x = a.x + b.x + c.x; + fprintf(out,"Int f(Int,Int,Int):({%d},{%d},{%d})",a.x,b.x,c.x); + fflush(out); + *(Int*)retp = r; +}} +void C_CdC_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&C_CdC) { fprintf(out,"wrong data for C_CdC\n"); exit(1); } + {Char a = *(Char*)(*args++); + double b = *(double*)(*args++); + Char c = *(Char*)(*args++); + Char r; + r.x = (a.x + c.x)/2; + fprintf(out,"Char f(Char,double,Char):({'%c'},%g,{'%c'})",a.x,b,c.x); + fflush(out); + *(Char*)retp = r; +}} +void F_Ffd_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&F_Ffd) { fprintf(out,"wrong data for F_Ffd\n"); exit(1); } + {Float a = *(Float*)(*args++); + float b = *(float*)(*args++); + double c = *(double*)(*args++); + Float r; + r.x = a.x + b + c; + fprintf(out,"Float f(Float,float,double):({%g},%g,%g)",a.x,b,c); + fflush(out); + *(Float*)retp = r; +}} +void D_fDd_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&D_fDd) { fprintf(out,"wrong data for D_fDd\n"); exit(1); } + {float a = *(float*)(*args++); + Double b = *(Double*)(*args++); + double c = *(double*)(*args++); + Double r; + r.x = a + b.x + c; + fprintf(out,"Double f(float,Double,double):(%g,{%g},%g)",a,b.x,c); + fflush(out); + *(Double*)retp = r; +}} +void D_Dfd_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&D_Dfd) { fprintf(out,"wrong data for D_Dfd\n"); exit(1); } + {Double a = *(Double*)(*args++); + float b = *(float*)(*args++); + double c = *(double*)(*args++); + Double r; + r.x = a.x + b + c; + fprintf(out,"Double f(Double,float,double):({%g},%g,%g)",a.x,b,c); + fflush(out); + *(Double*)retp = r; +}} +void J_JiJ_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&J_JiJ) { fprintf(out,"wrong data for J_JiJ\n"); exit(1); } + {J a = *(J*)(*args++); + int b= *(int*)(*args++); + J c = *(J*)(*args++); + J r; + r.l1 = a.l1+c.l1; r.l2 = a.l2+b+c.l2; + fprintf(out,"J f(J,int,J):({%ld,%ld},%d,{%ld,%ld})",a.l1,a.l2,b,c.l1,c.l2); + fflush(out); + *(J*)retp = r; +}} +#ifndef SKIP_EXTRA_STRUCTS +void T_TcT_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&T_TcT) { fprintf(out,"wrong data for T_TcT\n"); exit(1); } + {T a = *(T*)(*args++); + char b = *(char*)(*args++); + T c = *(T*)(*args++); + T r; + r.c[0]='b'; r.c[1]=c.c[1]; r.c[2]=c.c[2]; + fprintf(out,"T f(T,char,T):({\"%c%c%c\"},'%c',{\"%c%c%c\"})",a.c[0],a.c[1],a.c[2],b,c.c[0],c.c[1],c.c[2]); + fflush(out); + *(T*)retp = r; +}} +void X_BcdB_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&X_BcdB) { fprintf(out,"wrong data for X_BcdB\n"); exit(1); } + {B a = *(B*)(*args++); + char b = *(char*)(*args++); + double c = *(double*)(*args++); + B d = *(B*)(*args++); + static X xr={"return val",'R'}; + X r; + r = xr; + r.c1 = b; + fprintf(out,"X f(B,char,double,B):({%g,{%d,%d,%d}},'%c',%g,{%g,{%d,%d,%d}})", + a.d,a.i[0],a.i[1],a.i[2],b,c,d.d,d.i[0],d.i[1],d.i[2]); + fflush(out); + *(X*)retp = r; +}} +#endif + +/* gpargs boundary tests */ +void l_l0K_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&l_l0K) { fprintf(out,"wrong data for l_l0K\n"); exit(1); } + {K b = *(K*)(*args++); + long c = *(long*)(*args++); + long r = b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(K,long):(%ld,%ld,%ld,%ld,%ld)",b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + *(ffi_arg*)retp = r; +}} +void l_l1K_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&l_l1K) { fprintf(out,"wrong data for l_l1K\n"); exit(1); } + {long a1 = *(long*)(*args++); + K b = *(K*)(*args++); + long c = *(long*)(*args++); + long r = a1 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld)",a1,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + *(ffi_arg*)retp = r; +}} +void l_l2K_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&l_l2K) { fprintf(out,"wrong data for l_l2K\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + K b = *(K*)(*args++); + long c = *(long*)(*args++); + long r = a1 + a2 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(2*long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a1,a2,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + *(ffi_arg*)retp = r; +}} +void l_l3K_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&l_l3K) { fprintf(out,"wrong data for l_l3K\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + K b = *(K*)(*args++); + long c = *(long*)(*args++); + long r = a1 + a2 + a3 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(3*long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a1,a2,a3,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + *(ffi_arg*)retp = r; +}} +void l_l4K_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&l_l4K) { fprintf(out,"wrong data for l_l4K\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long a4 = *(long*)(*args++); + K b = *(K*)(*args++); + long c = *(long*)(*args++); + long r = a1 + a2 + a3 + a4 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(4*long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a1,a2,a3,a4,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + *(ffi_arg*)retp = r; +}} +void l_l5K_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&l_l5K) { fprintf(out,"wrong data for l_l5K\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long a4 = *(long*)(*args++); + long a5 = *(long*)(*args++); + K b = *(K*)(*args++); + long c = *(long*)(*args++); + long r = a1 + a2 + a3 + a4 + a5 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(5*long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a1,a2,a3,a4,a5,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + *(ffi_arg*)retp = r; +}} +void l_l6K_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&l_l6K) { fprintf(out,"wrong data for l_l6K\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long a4 = *(long*)(*args++); + long a5 = *(long*)(*args++); + long a6 = *(long*)(*args++); + K b = *(K*)(*args++); + long c = *(long*)(*args++); + long r = a1 + a2 + a3 + a4 + a5 + a6 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(6*long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a1,a2,a3,a4,a5,a6,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + *(ffi_arg*)retp = r; +}} +void f_f17l3L_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&f_f17l3L) { fprintf(out,"wrong data for f_f17l3L\n"); exit(1); } + {float a = *(float*)(*args++); + float b = *(float*)(*args++); + float c = *(float*)(*args++); + float d = *(float*)(*args++); + float e = *(float*)(*args++); + float f = *(float*)(*args++); + float g = *(float*)(*args++); + float h = *(float*)(*args++); + float i = *(float*)(*args++); + float j = *(float*)(*args++); + float k = *(float*)(*args++); + float l = *(float*)(*args++); + float m = *(float*)(*args++); + float n = *(float*)(*args++); + float o = *(float*)(*args++); + float p = *(float*)(*args++); + float q = *(float*)(*args++); + long s = *(long*)(*args++); + long t = *(long*)(*args++); + long u = *(long*)(*args++); + L z = *(L*)(*args++); + float r = a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p+q+s+t+u+z.l1+z.l2+z.l3+z.l4+z.l5+z.l6; + fprintf(out,"float f(17*float,3*int,L):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,s,t,u,z.l1,z.l2,z.l3,z.l4,z.l5,z.l6); + fflush(out); + *(float*)retp = r; +}} +void d_d17l3L_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_d17l3L) { fprintf(out,"wrong data for d_d17l3L\n"); exit(1); } + {double a = *(double*)(*args++); + double b = *(double*)(*args++); + double c = *(double*)(*args++); + double d = *(double*)(*args++); + double e = *(double*)(*args++); + double f = *(double*)(*args++); + double g = *(double*)(*args++); + double h = *(double*)(*args++); + double i = *(double*)(*args++); + double j = *(double*)(*args++); + double k = *(double*)(*args++); + double l = *(double*)(*args++); + double m = *(double*)(*args++); + double n = *(double*)(*args++); + double o = *(double*)(*args++); + double p = *(double*)(*args++); + double q = *(double*)(*args++); + long s = *(long*)(*args++); + long t = *(long*)(*args++); + long u = *(long*)(*args++); + L z = *(L*)(*args++); + double r = a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p+q+s+t+u+z.l1+z.l2+z.l3+z.l4+z.l5+z.l6; + fprintf(out,"double f(17*double,3*int,L):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,s,t,u,z.l1,z.l2,z.l3,z.l4,z.l5,z.l6); + fflush(out); + *(double*)retp = r; +}} +void ll_l2ll_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&ll_l2ll) { fprintf(out,"wrong data for ll_l2ll\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long long b = *(long long *)(*args++); + long c = *(long*)(*args++); + long long r = (long long) (a1 + a2) + b + c; + fprintf(out,"long long f(2*long,long long,long):(%ld,%ld,0x%lx%08lx,%ld)",a1,a2,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + *(long long *)retp = r; +}} +void ll_l3ll_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&ll_l3ll) { fprintf(out,"wrong data for ll_l3ll\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long long b = *(long long *)(*args++); + long c = *(long*)(*args++); + long long r = (long long) (a1 + a2 + a3) + b + c; + fprintf(out,"long long f(3*long,long long,long):(%ld,%ld,%ld,0x%lx%08lx,%ld)",a1,a2,a3,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + *(long long *)retp = r; +}} +void ll_l4ll_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&ll_l4ll) { fprintf(out,"wrong data for ll_l4ll\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long a4 = *(long*)(*args++); + long long b = *(long long *)(*args++); + long c = *(long*)(*args++); + long long r = (long long) (a1 + a2 + a3 + a4) + b + c; + fprintf(out,"long long f(4*long,long long,long):(%ld,%ld,%ld,%ld,0x%lx%08lx,%ld)",a1,a2,a3,a4,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + *(long long *)retp = r; +}} +void ll_l5ll_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&ll_l5ll) { fprintf(out,"wrong data for ll_l5ll\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long a4 = *(long*)(*args++); + long a5 = *(long*)(*args++); + long long b = *(long long *)(*args++); + long c = *(long*)(*args++); + long long r = (long long) (a1 + a2 + a3 + a4 + a5) + b + c; + fprintf(out,"long long f(5*long,long long,long):(%ld,%ld,%ld,%ld,%ld,0x%lx%08lx,%ld)",a1,a2,a3,a4,a5,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + *(long long *)retp = r; +}} +void ll_l6ll_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&ll_l6ll) { fprintf(out,"wrong data for ll_l6ll\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long a4 = *(long*)(*args++); + long a5 = *(long*)(*args++); + long a6 = *(long*)(*args++); + long long b = *(long long *)(*args++); + long c = *(long*)(*args++); + long long r = (long long) (a1 + a2 + a3 + a4 + a5 + a6) + b + c; + fprintf(out,"long long f(6*long,long long,long):(%ld,%ld,%ld,%ld,%ld,%ld,0x%lx%08lx,%ld)",a1,a2,a3,a4,a5,a6,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + *(long long *)retp = r; +}} +void ll_l7ll_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&ll_l7ll) { fprintf(out,"wrong data for ll_l7ll\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long a4 = *(long*)(*args++); + long a5 = *(long*)(*args++); + long a6 = *(long*)(*args++); + long a7 = *(long*)(*args++); + long long b = *(long long *)(*args++); + long c = *(long*)(*args++); + long long r = (long long) (a1 + a2 + a3 + a4 + a5 + a6 + a7) + b + c; + fprintf(out,"long long f(7*long,long long,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,0x%lx%08lx,%ld)",a1,a2,a3,a4,a5,a6,a7,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + *(long long *)retp = r; +}} +void d_l2d_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_l2d) { fprintf(out,"wrong data for d_l2d\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + double b = *(double*)(*args++); + long c = *(long*)(*args++); + double r = (double) (a1 + a2) + b + c; + fprintf(out,"double f(2*long,double,long):(%ld,%ld,%g,%ld)",a1,a2,b,c); + fflush(out); + *(double*)retp = r; +}} +void d_l3d_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_l3d) { fprintf(out,"wrong data for d_l3d\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + double b = *(double*)(*args++); + long c = *(long*)(*args++); + double r = (double) (a1 + a2 + a3) + b + c; + fprintf(out,"double f(3*long,double,long):(%ld,%ld,%ld,%g,%ld)",a1,a2,a3,b,c); + fflush(out); + *(double*)retp = r; +}} +void d_l4d_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_l4d) { fprintf(out,"wrong data for d_l4d\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long a4 = *(long*)(*args++); + double b = *(double*)(*args++); + long c = *(long*)(*args++); + double r = (double) (a1 + a2 + a3 + a4) + b + c; + fprintf(out,"double f(4*long,double,long):(%ld,%ld,%ld,%ld,%g,%ld)",a1,a2,a3,a4,b,c); + fflush(out); + *(double*)retp = r; +}} +void d_l5d_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_l5d) { fprintf(out,"wrong data for d_l5d\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long a4 = *(long*)(*args++); + long a5 = *(long*)(*args++); + double b = *(double*)(*args++); + long c = *(long*)(*args++); + double r = (double) (a1 + a2 + a3 + a4 + a5) + b + c; + fprintf(out,"double f(5*long,double,long):(%ld,%ld,%ld,%ld,%ld,%g,%ld)",a1,a2,a3,a4,a5,b,c); + fflush(out); + *(double*)retp = r; +}} +void d_l6d_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_l6d) { fprintf(out,"wrong data for d_l6d\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long a4 = *(long*)(*args++); + long a5 = *(long*)(*args++); + long a6 = *(long*)(*args++); + double b = *(double*)(*args++); + long c = *(long*)(*args++); + double r = (double) (a1 + a2 + a3 + a4 + a5 + a6) + b + c; + fprintf(out,"double f(6*long,double,long):(%ld,%ld,%ld,%ld,%ld,%ld,%g,%ld)",a1,a2,a3,a4,a5,a6,b,c); + fflush(out); + *(double*)retp = r; +}} +void d_l7d_simulator (ffi_cif* cif, void* retp, /*const*/ void* /*const*/ *args, void* data) +{ + if (data != (void*)&d_l7d) { fprintf(out,"wrong data for d_l7d\n"); exit(1); } + {long a1 = *(long*)(*args++); + long a2 = *(long*)(*args++); + long a3 = *(long*)(*args++); + long a4 = *(long*)(*args++); + long a5 = *(long*)(*args++); + long a6 = *(long*)(*args++); + long a7 = *(long*)(*args++); + double b = *(double*)(*args++); + long c = *(long*)(*args++); + double r = (double) (a1 + a2 + a3 + a4 + a5 + a6 + a7) + b + c; + fprintf(out,"double f(7*long,double,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,%g,%ld)",a1,a2,a3,a4,a5,a6,a7,b,c); + fflush(out); + *(double*)retp = r; +}} + + +/* + * The way we run these tests - first call the function directly, then + * through vacall() - there is the danger that arguments or results seem + * to be passed correctly, but what we are seeing are in fact the vestiges + * (traces) or the previous call. This may seriously fake the test. + * Avoid this by clearing the registers between the first and the second call. + */ +long clear_traces_i (long a, long b, long c, long d, long e, long f, long g, long h, + long i, long j, long k, long l, long m, long n, long o, long p) +{ return 0; } +float clear_traces_f (float a, float b, float c, float d, float e, float f, float g, + float h, float i, float j, float k, float l, float m, float n, + float o, float p) +{ return 0.0; } +double clear_traces_d (double a, double b, double c, double d, double e, double f, double g, + double h, double i, double j, double k, double l, double m, double n, + double o, double p) +{ return 0.0; } +J clear_traces_J (void) +{ J j; j.l1 = j.l2 = 0; return j; } +void clear_traces (void) +{ clear_traces_i(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0); + clear_traces_f(0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0); + clear_traces_d(0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0); + clear_traces_J(); +} + +int main (void) +{ + void* callback_code; + void* callback_writable; +#define ALLOC_CALLBACK() \ + callback_writable = ffi_closure_alloc(sizeof(ffi_closure),&callback_code); \ + if (!callback_writable) abort() +#define PREP_CALLBACK(cif,simulator,data) \ + if (ffi_prep_closure_loc(callback_writable,&(cif),simulator,data,callback_code) != FFI_OK) abort() +#define FREE_CALLBACK() \ + ffi_closure_free(callback_writable) + + ffi_type_char = (char)(-1) < 0 ? ffi_type_schar : ffi_type_uchar; + out = stdout; + +#if (!defined(DGTEST)) || DGTEST == 1 + /* void tests */ + v_v(); + clear_traces(); + ALLOC_CALLBACK(); + { + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_void); + PREP_CALLBACK(cif,v_v_simulator,(void*)&v_v); + ((void (ABI_ATTR *) (void)) callback_code) (); + } + FREE_CALLBACK(); +#endif + + /* int tests */ + { int ir; + +#if (!defined(DGTEST)) || DGTEST == 2 + ir = i_v(); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_sint); + PREP_CALLBACK(cif,i_v_simulator,(void*)&i_v); + ir = ((int (ABI_ATTR *) (void)) callback_code) (); + } + FREE_CALLBACK(); + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 3 + ir = i_i(i1); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_sint); + PREP_CALLBACK(cif,i_i_simulator,(void*)&i_i); + ir = ((int (ABI_ATTR *) (int)) callback_code) (i1); + } + FREE_CALLBACK(); + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 4 + ir = i_i2(i1,i2); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_sint); + PREP_CALLBACK(cif,i_i2_simulator,(void*)&i_i2); + ir = ((int (ABI_ATTR *) (int,int)) callback_code) (i1,i2); + } + FREE_CALLBACK(); + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 5 + ir = i_i4(i1,i2,i3,i4); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_sint); + PREP_CALLBACK(cif,i_i4_simulator,(void*)&i_i4); + ir = ((int (ABI_ATTR *) (int,int,int,int)) callback_code) (i1,i2,i3,i4); + } + FREE_CALLBACK(); + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 6 + ir = i_i8(i1,i2,i3,i4,i5,i6,i7,i8); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_sint); + PREP_CALLBACK(cif,i_i8_simulator,(void*)&i_i8); + ir = ((int (ABI_ATTR *) (int,int,int,int,int,int,int,int)) callback_code) (i1,i2,i3,i4,i5,i6,i7,i8); + } + FREE_CALLBACK(); + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 7 + ir = i_i16(i1,i2,i3,i4,i5,i6,i7,i8,i9,i10,i11,i12,i13,i14,i15,i16); + fprintf(out,"->%d\n",ir); + fflush(out); + ir = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_sint); + PREP_CALLBACK(cif,i_i16_simulator,(void*)&i_i16); + ir = ((int (ABI_ATTR *) (int,int,int,int,int,int,int,int,int,int,int,int,int,int,int,int)) callback_code) (i1,i2,i3,i4,i5,i6,i7,i8,i9,i10,i11,i12,i13,i14,i15,i16); + } + FREE_CALLBACK(); + fprintf(out,"->%d\n",ir); + fflush(out); +#endif + } + + /* float tests */ + { float fr; + +#if (!defined(DGTEST)) || DGTEST == 8 + fr = f_f(f1); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f_simulator,(void*)&f_f); + fr = ((float (ABI_ATTR *) (float)) callback_code) (f1); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 9 + fr = f_f2(f1,f2); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f2_simulator,(void*)&f_f2); + fr = ((float (ABI_ATTR *) (float,float)) callback_code) (f1,f2); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 10 + fr = f_f4(f1,f2,f3,f4); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f4_simulator,(void*)&f_f4); + fr = ((float (ABI_ATTR *) (float,float,float,float)) callback_code) (f1,f2,f3,f4); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 11 + fr = f_f8(f1,f2,f3,f4,f5,f6,f7,f8); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f8_simulator,(void*)&f_f8); + fr = ((float (ABI_ATTR *) (float,float,float,float,float,float,float,float)) callback_code) (f1,f2,f3,f4,f5,f6,f7,f8); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 12 + fr = f_f16(f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,f14,f15,f16); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f16_simulator,(void*)&f_f16); + fr = ((float (ABI_ATTR *) (float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float)) callback_code) (f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,f14,f15,f16); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 13 + fr = f_f24(f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,f14,f15,f16,f17,f18,f19,f20,f21,f22,f23,f24); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f24_simulator,(void*)&f_f24); + fr = ((float (ABI_ATTR *) (float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float)) callback_code) (f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,f14,f15,f16,f17,f18,f19,f20,f21,f22,f23,f24); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + + } + + /* double tests */ + { double dr; + +#if (!defined(DGTEST)) || DGTEST == 14 + dr = d_d(d1); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d_simulator,(void*)&d_d); + dr = ((double (ABI_ATTR *) (double)) callback_code) (d1); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 15 + dr = d_d2(d1,d2); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d2_simulator,(void*)&d_d2); + dr = ((double (ABI_ATTR *) (double,double)) callback_code) (d1,d2); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 16 + dr = d_d4(d1,d2,d3,d4); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d4_simulator,(void*)&d_d4); + dr = ((double (ABI_ATTR *) (double,double,double,double)) callback_code) (d1,d2,d3,d4); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 17 + dr = d_d8(d1,d2,d3,d4,d5,d6,d7,d8); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d8_simulator,(void*)&d_d8); + dr = ((double (ABI_ATTR *) (double,double,double,double,double,double,double,double)) callback_code) (d1,d2,d3,d4,d5,d6,d7,d8); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 18 + dr = d_d16(d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,d13,d14,d15,d16); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d16_simulator,(void*)&d_d16); + dr = ((double (ABI_ATTR *) (double,double,double,double,double,double,double,double,double,double,double,double,double,double,double,double)) callback_code) (d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,d13,d14,d15,d16); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + } + + /* pointer tests */ + { void* vpr; + +#if (!defined(DGTEST)) || DGTEST == 19 + vpr = vp_vpdpcpsp(&uc1,&d2,str3,&I4); + fprintf(out,"->0x%p\n",vpr); + fflush(out); + vpr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_pointer, &ffi_type_pointer, &ffi_type_pointer, &ffi_type_pointer }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_pointer); + PREP_CALLBACK(cif,vp_vpdpcpsp_simulator,(void*)&vp_vpdpcpsp); + vpr = ((void* (ABI_ATTR *) (void*,double*,char*,Int*)) callback_code) (&uc1,&d2,str3,&I4); + } + FREE_CALLBACK(); + fprintf(out,"->0x%p\n",vpr); + fflush(out); +#endif + } + + /* mixed number tests */ + { uchar ucr; + ushort usr; + float fr; + double dr; + long long llr; + +#if (!defined(DGTEST)) || DGTEST == 20 + ucr = uc_ucsil(uc1,us2,ui3,ul4); + fprintf(out,"->%u\n",ucr); + fflush(out); + ucr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_uchar, &ffi_type_ushort, &ffi_type_uint, &ffi_type_ulong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_uchar); + PREP_CALLBACK(cif,uc_ucsil_simulator,(void*)&uc_ucsil); + ucr = ((uchar (ABI_ATTR *) (uchar,ushort,uint,ulong)) callback_code) (uc1,us2,ui3,ul4); + } + FREE_CALLBACK(); + fprintf(out,"->%u\n",ucr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 21 + dr = d_iidd(i1,i2,d3,d4); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_iidd_simulator,(void*)&d_iidd); + dr = ((double (ABI_ATTR *) (int,int,double,double)) callback_code) (i1,i2,d3,d4); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 22 + dr = d_iiidi(i1,i2,i3,d4,i5); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_iiidi_simulator,(void*)&d_iiidi); + dr = ((double (ABI_ATTR *) (int,int,int,double,int)) callback_code) (i1,i2,i3,d4,i5); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 23 + dr = d_idid(i1,d2,i3,d4); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_double, &ffi_type_sint, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_idid_simulator,(void*)&d_idid); + dr = ((double (ABI_ATTR *) (int,double,int,double)) callback_code) (i1,d2,i3,d4); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 24 + dr = d_fdi(f1,d2,i3); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_fdi_simulator,(void*)&d_fdi); + dr = ((double (ABI_ATTR *) (float,double,int)) callback_code) (f1,d2,i3); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 25 + usr = us_cdcd(c1,d2,c3,d4); + fprintf(out,"->%u\n",usr); + fflush(out); + usr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_char, &ffi_type_double, &ffi_type_char, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_ushort); + PREP_CALLBACK(cif,us_cdcd_simulator,(void*)&us_cdcd); + usr = ((ushort (ABI_ATTR *) (char,double,char,double)) callback_code) (c1,d2,c3,d4); + } + FREE_CALLBACK(); + fprintf(out,"->%u\n",usr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 26 + llr = ll_iiilli(i1,i2,i3,ll1,i13); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_sint, &ffi_type_sint, &ffi_type_sint, &ffi_type_slonglong, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + PREP_CALLBACK(cif,ll_iiilli_simulator,(void*)&ll_iiilli); + llr = ((long long (ABI_ATTR *) (int,int,int,long long,int)) callback_code) (i1,i2,i3,ll1,i13); + } + FREE_CALLBACK(); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 27 + llr = ll_flli(f13,ll1,i13); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_slonglong, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + PREP_CALLBACK(cif,ll_flli_simulator,(void*)&ll_flli); + llr = ((long long (ABI_ATTR *) (float,long long,int)) callback_code) (f13,ll1,i13); + } + FREE_CALLBACK(); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 28 + fr = f_fi(f1,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_fi_simulator,(void*)&f_fi); + fr = ((float (ABI_ATTR *) (float,int)) callback_code) (f1,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 29 + fr = f_f2i(f1,f2,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f2i_simulator,(void*)&f_f2i); + fr = ((float (ABI_ATTR *) (float,float,int)) callback_code) (f1,f2,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 30 + fr = f_f3i(f1,f2,f3,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f3i_simulator,(void*)&f_f3i); + fr = ((float (ABI_ATTR *) (float,float,float,int)) callback_code) (f1,f2,f3,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 31 + fr = f_f4i(f1,f2,f3,f4,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f4i_simulator,(void*)&f_f4i); + fr = ((float (ABI_ATTR *) (float,float,float,float,int)) callback_code) (f1,f2,f3,f4,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 32 + fr = f_f7i(f1,f2,f3,f4,f5,f6,f7,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f7i_simulator,(void*)&f_f7i); + fr = ((float (ABI_ATTR *) (float,float,float,float,float,float,float,int)) callback_code) (f1,f2,f3,f4,f5,f6,f7,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 33 + fr = f_f8i(f1,f2,f3,f4,f5,f6,f7,f8,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f8i_simulator,(void*)&f_f8i); + fr = ((float (ABI_ATTR *) (float,float,float,float,float,float,float,float,int)) callback_code) (f1,f2,f3,f4,f5,f6,f7,f8,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 34 + fr = f_f13i(f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,i9); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f13i_simulator,(void*)&f_f13i); + fr = ((float (ABI_ATTR *) (float,float,float,float,float,float,float,float,float,float,float,float,float,int)) callback_code) (f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 35 + dr = d_di(d1,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_di_simulator,(void*)&d_di); + dr = ((double (ABI_ATTR *) (double,int)) callback_code) (d1,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 36 + dr = d_d2i(d1,d2,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d2i_simulator,(void*)&d_d2i); + dr = ((double (ABI_ATTR *) (double,double,int)) callback_code) (d1,d2,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 37 + dr = d_d3i(d1,d2,d3,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d3i_simulator,(void*)&d_d3i); + dr = ((double (ABI_ATTR *) (double,double,double,int)) callback_code) (d1,d2,d3,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 38 + dr = d_d4i(d1,d2,d3,d4,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d4i_simulator,(void*)&d_d4i); + dr = ((double (ABI_ATTR *) (double,double,double,double,int)) callback_code) (d1,d2,d3,d4,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 39 + dr = d_d7i(d1,d2,d3,d4,d5,d6,d7,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d7i_simulator,(void*)&d_d7i); + dr = ((double (ABI_ATTR *) (double,double,double,double,double,double,double,int)) callback_code) (d1,d2,d3,d4,d5,d6,d7,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 40 + dr = d_d8i(d1,d2,d3,d4,d5,d6,d7,d8,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d8i_simulator,(void*)&d_d8i); + dr = ((double (ABI_ATTR *) (double,double,double,double,double,double,double,double,int)) callback_code) (d1,d2,d3,d4,d5,d6,d7,d8,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 41 + dr = d_d12i(d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d12i_simulator,(void*)&d_d12i); + dr = ((double (ABI_ATTR *) (double,double,double,double,double,double,double,double,double,double,double,double,int)) callback_code) (d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 42 + dr = d_d13i(d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,d13,i9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_sint }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d13i_simulator,(void*)&d_d13i); + dr = ((double (ABI_ATTR *) (double,double,double,double,double,double,double,double,double,double,double,double,double,int)) callback_code) (d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,d13,i9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + } + + /* small structure return tests */ +#if (!defined(DGTEST)) || DGTEST == 43 + { + Size1 r = S1_v(); + fprintf(out,"->{%c}\n",r.x1); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Size1_elements[] = { &ffi_type_char, NULL }; + ffi_type ffi_type_Size1; + ffi_type_Size1.type = FFI_TYPE_STRUCT; + ffi_type_Size1.size = sizeof(Size1); + ffi_type_Size1.alignment = alignof_slot(Size1); + ffi_type_Size1.elements = ffi_type_Size1_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size1); + PREP_CALLBACK(cif,S1_v_simulator,(void*)&S1_v); + r = ((Size1 (ABI_ATTR *) (void)) callback_code) (); + } + FREE_CALLBACK(); + fprintf(out,"->{%c}\n",r.x1); + fflush(out); + } +#endif + +#if (!defined(DGTEST)) || DGTEST == 44 + { + Size2 r = S2_v(); + fprintf(out,"->{%c%c}\n",r.x1,r.x2); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Size2_elements[] = { &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size2; + ffi_type_Size2.type = FFI_TYPE_STRUCT; + ffi_type_Size2.size = sizeof(Size2); + ffi_type_Size2.alignment = alignof_slot(Size2); + ffi_type_Size2.elements = ffi_type_Size2_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size2); + PREP_CALLBACK(cif,S2_v_simulator,(void*)&S2_v); + r = ((Size2 (ABI_ATTR *) (void)) callback_code) (); + } + FREE_CALLBACK(); + fprintf(out,"->{%c%c}\n",r.x1,r.x2); + fflush(out); + } +#endif + +#if (!defined(DGTEST)) || DGTEST == 45 + { + Size3 r = S3_v(); + fprintf(out,"->{%c%c%c}\n",r.x1,r.x2,r.x3); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Size3_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size3; + ffi_type_Size3.type = FFI_TYPE_STRUCT; + ffi_type_Size3.size = sizeof(Size3); + ffi_type_Size3.alignment = alignof_slot(Size3); + ffi_type_Size3.elements = ffi_type_Size3_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size3); + PREP_CALLBACK(cif,S3_v_simulator,(void*)&S3_v); + r = ((Size3 (ABI_ATTR *) (void)) callback_code) (); + } + FREE_CALLBACK(); + fprintf(out,"->{%c%c%c}\n",r.x1,r.x2,r.x3); + fflush(out); + } +#endif + +#if (!defined(DGTEST)) || DGTEST == 46 + { + Size4 r = S4_v(); + fprintf(out,"->{%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Size4_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size4; + ffi_type_Size4.type = FFI_TYPE_STRUCT; + ffi_type_Size4.size = sizeof(Size4); + ffi_type_Size4.alignment = alignof_slot(Size4); + ffi_type_Size4.elements = ffi_type_Size4_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size4); + PREP_CALLBACK(cif,S4_v_simulator,(void*)&S4_v); + r = ((Size4 (ABI_ATTR *) (void)) callback_code) (); + } + FREE_CALLBACK(); + fprintf(out,"->{%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4); + fflush(out); + } +#endif + +#if (!defined(DGTEST)) || DGTEST == 47 + { + Size7 r = S7_v(); + fprintf(out,"->{%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Size7_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size7; + ffi_type_Size7.type = FFI_TYPE_STRUCT; + ffi_type_Size7.size = sizeof(Size7); + ffi_type_Size7.alignment = alignof_slot(Size7); + ffi_type_Size7.elements = ffi_type_Size7_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size7); + PREP_CALLBACK(cif,S7_v_simulator,(void*)&S7_v); + r = ((Size7 (ABI_ATTR *) (void)) callback_code) (); + } + FREE_CALLBACK(); + fprintf(out,"->{%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7); + fflush(out); + } +#endif + +#if (!defined(DGTEST)) || DGTEST == 48 + { + Size8 r = S8_v(); + fprintf(out,"->{%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Size8_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size8; + ffi_type_Size8.type = FFI_TYPE_STRUCT; + ffi_type_Size8.size = sizeof(Size8); + ffi_type_Size8.alignment = alignof_slot(Size8); + ffi_type_Size8.elements = ffi_type_Size8_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size8); + PREP_CALLBACK(cif,S8_v_simulator,(void*)&S8_v); + r = ((Size8 (ABI_ATTR *) (void)) callback_code) (); + } + FREE_CALLBACK(); + fprintf(out,"->{%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8); + fflush(out); + } +#endif + +#if (!defined(DGTEST)) || DGTEST == 49 + { + Size12 r = S12_v(); + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Size12_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size12; + ffi_type_Size12.type = FFI_TYPE_STRUCT; + ffi_type_Size12.size = sizeof(Size12); + ffi_type_Size12.alignment = alignof_slot(Size12); + ffi_type_Size12.elements = ffi_type_Size12_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size12); + PREP_CALLBACK(cif,S12_v_simulator,(void*)&S12_v); + r = ((Size12 (ABI_ATTR *) (void)) callback_code) (); + } + FREE_CALLBACK(); + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12); + fflush(out); + } +#endif + +#if (!defined(DGTEST)) || DGTEST == 50 + { + Size15 r = S15_v(); + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12,r.x13,r.x14,r.x15); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Size15_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size15; + ffi_type_Size15.type = FFI_TYPE_STRUCT; + ffi_type_Size15.size = sizeof(Size15); + ffi_type_Size15.alignment = alignof_slot(Size15); + ffi_type_Size15.elements = ffi_type_Size15_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size15); + PREP_CALLBACK(cif,S15_v_simulator,(void*)&S15_v); + r = ((Size15 (ABI_ATTR *) (void)) callback_code) (); + } + FREE_CALLBACK(); + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12,r.x13,r.x14,r.x15); + fflush(out); + } +#endif + +#if (!defined(DGTEST)) || DGTEST == 51 + { + Size16 r = S16_v(); + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12,r.x13,r.x14,r.x15,r.x16); + fflush(out); + memset(&r,0,sizeof(r)); clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Size16_elements[] = { &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, &ffi_type_char, NULL }; + ffi_type ffi_type_Size16; + ffi_type_Size16.type = FFI_TYPE_STRUCT; + ffi_type_Size16.size = sizeof(Size16); + ffi_type_Size16.alignment = alignof_slot(Size16); + ffi_type_Size16.elements = ffi_type_Size16_elements; + ffi_cif cif; + FFI_PREP_CIF_NOARGS(cif,ffi_type_Size16); + PREP_CALLBACK(cif,S16_v_simulator,(void*)&S16_v); + r = ((Size16 (ABI_ATTR *) (void)) callback_code) (); + } + FREE_CALLBACK(); + fprintf(out,"->{%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c%c}\n",r.x1,r.x2,r.x3,r.x4,r.x5,r.x6,r.x7,r.x8,r.x9,r.x10,r.x11,r.x12,r.x13,r.x14,r.x15,r.x16); + fflush(out); + } +#endif + + + /* structure tests */ + { Int Ir; + Char Cr; + Float Fr; + Double Dr; + J Jr; +#ifndef SKIP_EXTRA_STRUCTS + T Tr; + X Xr; +#endif + +#if (!defined(DGTEST)) || DGTEST == 52 + Ir = I_III(I1,I2,I3); + fprintf(out,"->{%d}\n",Ir.x); + fflush(out); + Ir.x = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Int_elements[] = { &ffi_type_sint, NULL }; + ffi_type ffi_type_Int; + ffi_type_Int.type = FFI_TYPE_STRUCT; + ffi_type_Int.size = sizeof(Int); + ffi_type_Int.alignment = alignof_slot(Int); + ffi_type_Int.elements = ffi_type_Int_elements; + ffi_type* argtypes[] = { &ffi_type_Int, &ffi_type_Int, &ffi_type_Int }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_Int); + PREP_CALLBACK(cif,I_III_simulator,(void*)&I_III); + Ir = ((Int (ABI_ATTR *) (Int,Int,Int)) callback_code) (I1,I2,I3); + } + FREE_CALLBACK(); + fprintf(out,"->{%d}\n",Ir.x); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 53 + Cr = C_CdC(C1,d2,C3); + fprintf(out,"->{'%c'}\n",Cr.x); + fflush(out); + Cr.x = '\0'; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Char_elements[] = { &ffi_type_char, NULL }; + ffi_type ffi_type_Char; + ffi_type_Char.type = FFI_TYPE_STRUCT; + ffi_type_Char.size = sizeof(Char); + ffi_type_Char.alignment = alignof_slot(Char); + ffi_type_Char.elements = ffi_type_Char_elements; + ffi_type* argtypes[] = { &ffi_type_Char, &ffi_type_double, &ffi_type_Char }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_Char); + PREP_CALLBACK(cif,C_CdC_simulator,(void*)&C_CdC); + Cr = ((Char (ABI_ATTR *) (Char,double,Char)) callback_code) (C1,d2,C3); + } + FREE_CALLBACK(); + fprintf(out,"->{'%c'}\n",Cr.x); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 54 + Fr = F_Ffd(F1,f2,d3); + fprintf(out,"->{%g}\n",Fr.x); + fflush(out); + Fr.x = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Float_elements[] = { &ffi_type_float, NULL }; + ffi_type ffi_type_Float; + ffi_type_Float.type = FFI_TYPE_STRUCT; + ffi_type_Float.size = sizeof(Float); + ffi_type_Float.alignment = alignof_slot(Float); + ffi_type_Float.elements = ffi_type_Float_elements; + ffi_type* argtypes[] = { &ffi_type_Float, &ffi_type_float, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_Float); + PREP_CALLBACK(cif,F_Ffd_simulator,(void*)&F_Ffd); + Fr = ((Float (ABI_ATTR *) (Float,float,double)) callback_code) (F1,f2,d3); + } + FREE_CALLBACK(); + fprintf(out,"->{%g}\n",Fr.x); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 55 + Dr = D_fDd(f1,D2,d3); + fprintf(out,"->{%g}\n",Dr.x); + fflush(out); + Dr.x = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Double_elements[] = { &ffi_type_double, NULL }; + ffi_type ffi_type_Double; + ffi_type_Double.type = FFI_TYPE_STRUCT; + ffi_type_Double.size = sizeof(Double); + ffi_type_Double.alignment = alignof_slot(Double); + ffi_type_Double.elements = ffi_type_Double_elements; + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_Double, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_Double); + PREP_CALLBACK(cif,D_fDd_simulator,(void*)&D_fDd); + Dr = ((Double (ABI_ATTR *) (float,Double,double)) callback_code) (f1,D2,d3); + } + FREE_CALLBACK(); + fprintf(out,"->{%g}\n",Dr.x); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 56 + Dr = D_Dfd(D1,f2,d3); + fprintf(out,"->{%g}\n",Dr.x); + fflush(out); + Dr.x = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_Double_elements[] = { &ffi_type_double, NULL }; + ffi_type ffi_type_Double; + ffi_type_Double.type = FFI_TYPE_STRUCT; + ffi_type_Double.size = sizeof(Double); + ffi_type_Double.alignment = alignof_slot(Double); + ffi_type_Double.elements = ffi_type_Double_elements; + ffi_type* argtypes[] = { &ffi_type_Double, &ffi_type_float, &ffi_type_double }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_Double); + PREP_CALLBACK(cif,D_Dfd_simulator,(void*)&D_Dfd); + Dr = ((Double (ABI_ATTR *) (Double,float,double)) callback_code) (D1,f2,d3); + } + FREE_CALLBACK(); + fprintf(out,"->{%g}\n",Dr.x); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 57 + Jr = J_JiJ(J1,i2,J2); + fprintf(out,"->{%ld,%ld}\n",Jr.l1,Jr.l2); + fflush(out); + Jr.l1 = Jr.l2 = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_J_elements[] = { &ffi_type_slong, &ffi_type_slong, NULL }; + ffi_type ffi_type_J; + ffi_type_J.type = FFI_TYPE_STRUCT; + ffi_type_J.size = sizeof(J); + ffi_type_J.alignment = alignof_slot(J); + ffi_type_J.elements = ffi_type_J_elements; + ffi_type* argtypes[] = { &ffi_type_J, &ffi_type_sint, &ffi_type_J }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_J); + PREP_CALLBACK(cif,J_JiJ_simulator,(void*)&J_JiJ); + Jr = ((J (ABI_ATTR *) (J,int,J)) callback_code) (J1,i2,J2); + } + FREE_CALLBACK(); + fprintf(out,"->{%ld,%ld}\n",Jr.l1,Jr.l2); + fflush(out); +#endif + +#ifndef SKIP_EXTRA_STRUCTS +#if (!defined(DGTEST)) || DGTEST == 58 + Tr = T_TcT(T1,' ',T2); + fprintf(out,"->{\"%c%c%c\"}\n",Tr.c[0],Tr.c[1],Tr.c[2]); + fflush(out); + Tr.c[0] = Tr.c[1] = Tr.c[2] = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_T_elements[] = { ??, NULL }; + ffi_type ffi_type_T; + ffi_type_T.type = FFI_TYPE_STRUCT; + ffi_type_T.size = sizeof(T); + ffi_type_T.alignment = alignof_slot(T); + ffi_type_T.elements = ffi_type_T_elements; + ffi_type* argtypes[] = { &ffi_type_T, &ffi_type_char, &ffi_type_T }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_T); + PREP_CALLBACK(cif,T_TcT_simulator,(void*)&T_TcT); + Tr = ((T (ABI_ATTR *) (T,char,T)) callback_code) (T1,' ',T2); + } + FREE_CALLBACK(); + fprintf(out,"->{\"%c%c%c\"}\n",Tr.c[0],Tr.c[1],Tr.c[2]); + fflush(out); +#endif + +#ifndef SKIP_X +#if (!defined(DGTEST)) || DGTEST == 59 + Xr = X_BcdB(B1,c2,d3,B2); + fprintf(out,"->{\"%s\",'%c'}\n",Xr.c,Xr.c1); + fflush(out); + Xr.c[0]=Xr.c1='\0'; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* ffi_type_X_elements[] = { ??, NULL }; + ffi_type ffi_type_X; + ffi_type_X.type = FFI_TYPE_STRUCT; + ffi_type_X.size = sizeof(X); + ffi_type_X.alignment = alignof_slot(X); + ffi_type_X.elements = ffi_type_X_elements; + ffi_type* argtypes[] = { &ffi_type_X, &ffi_type_char, &ffi_type_double, &ffi_type_X }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_X); + PREP_CALLBACK(cif,X_BcdB_simulator,(void*)&X_BcdB); + Xr = ((X (ABI_ATTR *) (B,char,double,B)) callback_code) (B1,c2,d3,B2); + } + FREE_CALLBACK(); + fprintf(out,"->{\"%s\",'%c'}\n",Xr.c,Xr.c1); + fflush(out); +#endif +#endif +#endif + } + + + /* gpargs boundary tests */ + { + ffi_type* ffi_type_K_elements[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, NULL }; + ffi_type ffi_type_K; + ffi_type* ffi_type_L_elements[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, NULL }; + ffi_type ffi_type_L; + long lr; + long long llr; + float fr; + double dr; + + ffi_type_K.type = FFI_TYPE_STRUCT; + ffi_type_K.size = sizeof(K); + ffi_type_K.alignment = alignof_slot(K); + ffi_type_K.elements = ffi_type_K_elements; + + ffi_type_L.type = FFI_TYPE_STRUCT; + ffi_type_L.size = sizeof(L); + ffi_type_L.alignment = alignof_slot(L); + ffi_type_L.elements = ffi_type_L_elements; + +#if (!defined(DGTEST)) || DGTEST == 60 + lr = l_l0K(K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + PREP_CALLBACK(cif,l_l0K_simulator,(void*)l_l0K); + lr = ((long (ABI_ATTR *) (K,long)) callback_code) (K1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 61 + lr = l_l1K(l1,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + PREP_CALLBACK(cif,l_l1K_simulator,(void*)l_l1K); + lr = ((long (ABI_ATTR *) (long,K,long)) callback_code) (l1,K1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 62 + lr = l_l2K(l1,l2,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + PREP_CALLBACK(cif,l_l2K_simulator,(void*)l_l2K); + lr = ((long (ABI_ATTR *) (long,long,K,long)) callback_code) (l1,l2,K1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 63 + lr = l_l3K(l1,l2,l3,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + PREP_CALLBACK(cif,l_l3K_simulator,(void*)l_l3K); + lr = ((long (ABI_ATTR *) (long,long,long,K,long)) callback_code) (l1,l2,l3,K1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 64 + lr = l_l4K(l1,l2,l3,l4,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + PREP_CALLBACK(cif,l_l4K_simulator,(void*)l_l4K); + lr = ((long (ABI_ATTR *) (long,long,long,long,K,long)) callback_code) (l1,l2,l3,l4,K1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 65 + lr = l_l5K(l1,l2,l3,l4,l5,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + PREP_CALLBACK(cif,l_l5K_simulator,(void*)l_l5K); + lr = ((long (ABI_ATTR *) (long,long,long,long,long,K,long)) callback_code) (l1,l2,l3,l4,l5,K1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 66 + lr = l_l6K(l1,l2,l3,l4,l5,l6,K1,l9); + fprintf(out,"->%ld\n",lr); + fflush(out); + lr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_K, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slong); + PREP_CALLBACK(cif,l_l6K_simulator,(void*)l_l6K); + lr = ((long (ABI_ATTR *) (long,long,long,long,long,long,K,long)) callback_code) (l1,l2,l3,l4,l5,l6,K1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%ld\n",lr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 67 + fr = f_f17l3L(f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,f14,f15,f16,f17,l6,l7,l8,L1); + fprintf(out,"->%g\n",fr); + fflush(out); + fr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_float, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_L }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_float); + PREP_CALLBACK(cif,f_f17l3L_simulator,(void*)&f_f17l3L); + fr = ((float (ABI_ATTR *) (float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,float,long,long,long,L)) callback_code) (f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,f14,f15,f16,f17,l6,l7,l8,L1); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",fr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 68 + dr = d_d17l3L(d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,d13,d14,d15,d16,d17,l6,l7,l8,L1); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_double, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_L }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_d17l3L_simulator,(void*)&d_d17l3L); + dr = ((double (ABI_ATTR *) (double,double,double,double,double,double,double,double,double,double,double,double,double,double,double,double,double,long,long,long,L)) callback_code) (d1,d2,d3,d4,d5,d6,d7,d8,d9,d10,d11,d12,d13,d14,d15,d16,d17,l6,l7,l8,L1); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 69 + llr = ll_l2ll(l1,l2,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + PREP_CALLBACK(cif,ll_l2ll_simulator,(void*)ll_l2ll); + llr = ((long long (ABI_ATTR *) (long,long,long long,long)) callback_code) (l1,l2,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 70 + llr = ll_l3ll(l1,l2,l3,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + PREP_CALLBACK(cif,ll_l3ll_simulator,(void*)ll_l3ll); + llr = ((long long (ABI_ATTR *) (long,long,long,long long,long)) callback_code) (l1,l2,l3,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 71 + llr = ll_l4ll(l1,l2,l3,l4,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + PREP_CALLBACK(cif,ll_l4ll_simulator,(void*)ll_l4ll); + llr = ((long long (ABI_ATTR *) (long,long,long,long,long long,long)) callback_code) (l1,l2,l3,l4,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 72 + llr = ll_l5ll(l1,l2,l3,l4,l5,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + PREP_CALLBACK(cif,ll_l5ll_simulator,(void*)ll_l5ll); + llr = ((long long (ABI_ATTR *) (long,long,long,long,long,long long,long)) callback_code) (l1,l2,l3,l4,l5,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 73 + llr = ll_l6ll(l1,l2,l3,l4,l5,l6,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + PREP_CALLBACK(cif,ll_l6ll_simulator,(void*)ll_l6ll); + llr = ((long long (ABI_ATTR *) (long,long,long,long,long,long,long long,long)) callback_code) (l1,l2,l3,l4,l5,l6,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 74 + llr = ll_l7ll(l1,l2,l3,l4,l5,l6,l7,ll1,l9); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); + llr = 0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slonglong, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_slonglong); + PREP_CALLBACK(cif,ll_l7ll_simulator,(void*)ll_l7ll); + llr = ((long long (ABI_ATTR *) (long,long,long,long,long,long,long,long long,long)) callback_code) (l1,l2,l3,l4,l5,l6,l7,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->0x%lx%08lx\n",(long)(llr>>32),(long)(llr&0xffffffff)); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 75 + dr = d_l2d(l1,l2,ll1,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_l2d_simulator,(void*)d_l2d); + dr = ((double (ABI_ATTR *) (long,long,double,long)) callback_code) (l1,l2,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 76 + dr = d_l3d(l1,l2,l3,ll1,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_l3d_simulator,(void*)d_l3d); + dr = ((double (ABI_ATTR *) (long,long,long,double,long)) callback_code) (l1,l2,l3,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 77 + dr = d_l4d(l1,l2,l3,l4,ll1,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_l4d_simulator,(void*)d_l4d); + dr = ((double (ABI_ATTR *) (long,long,long,long,double,long)) callback_code) (l1,l2,l3,l4,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 78 + dr = d_l5d(l1,l2,l3,l4,l5,ll1,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_l5d_simulator,(void*)d_l5d); + dr = ((double (ABI_ATTR *) (long,long,long,long,long,double,long)) callback_code) (l1,l2,l3,l4,l5,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 79 + dr = d_l6d(l1,l2,l3,l4,l5,l6,ll1,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_l6d_simulator,(void*)d_l6d); + dr = ((double (ABI_ATTR *) (long,long,long,long,long,long,double,long)) callback_code) (l1,l2,l3,l4,l5,l6,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + +#if (!defined(DGTEST)) || DGTEST == 80 + dr = d_l7d(l1,l2,l3,l4,l5,l6,l7,ll1,l9); + fprintf(out,"->%g\n",dr); + fflush(out); + dr = 0.0; clear_traces(); + ALLOC_CALLBACK(); + { + ffi_type* argtypes[] = { &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_slong, &ffi_type_double, &ffi_type_slong }; + ffi_cif cif; + FFI_PREP_CIF(cif,argtypes,ffi_type_double); + PREP_CALLBACK(cif,d_l7d_simulator,(void*)d_l7d); + dr = ((double (ABI_ATTR *) (long,long,long,long,long,long,long,double,long)) callback_code) (l1,l2,l3,l4,l5,l6,l7,ll1,l9); + } + FREE_CALLBACK(); + fprintf(out,"->%g\n",dr); + fflush(out); +#endif + + } + + exit(0); +} + diff --git a/native/libffi/testsuite/libffi.bhaible/testcases.c b/native/libffi/testsuite/libffi.bhaible/testcases.c new file mode 100644 index 0000000000..d25ebf48c9 --- /dev/null +++ b/native/libffi/testsuite/libffi.bhaible/testcases.c @@ -0,0 +1,743 @@ +/* + * Copyright 1993 Bill Triggs + * Copyright 1995-2017 Bruno Haible + * + * This program is free software: you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 3 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* This file defines test functions of selected signatures, that exercise + dark corners of the various ABIs. */ + +#include + +FILE* out; + +#define uchar unsigned char +#define ushort unsigned short +#define uint unsigned int +#define ulong unsigned long + +typedef struct { char x; } Char; +typedef struct { short x; } Short; +typedef struct { int x; } Int; +typedef struct { long x; } Long; +typedef struct { float x; } Float; +typedef struct { double x; } Double; +typedef struct { char c; float f; } A; +typedef struct { double d; int i[3]; } B; +typedef struct { long l1; long l2; } J; +typedef struct { long l1; long l2; long l3; long l4; } K; +typedef struct { long l1; long l2; long l3; long l4; long l5; long l6; } L; +typedef struct { char x1; } Size1; +typedef struct { char x1; char x2; } Size2; +typedef struct { char x1; char x2; char x3; } Size3; +typedef struct { char x1; char x2; char x3; char x4; } Size4; +typedef struct { + char x1; char x2; char x3; char x4; char x5; char x6; char x7; +} Size7; +typedef struct { + char x1; char x2; char x3; char x4; char x5; char x6; char x7; char x8; +} Size8; +typedef struct { + char x1; char x2; char x3; char x4; char x5; char x6; char x7; char x8; + char x9; char x10; char x11; char x12; +} Size12; +typedef struct { + char x1; char x2; char x3; char x4; char x5; char x6; char x7; char x8; + char x9; char x10; char x11; char x12; char x13; char x14; char x15; +} Size15; +typedef struct { + char x1; char x2; char x3; char x4; char x5; char x6; char x7; char x8; + char x9; char x10; char x11; char x12; char x13; char x14; char x15; char x16; +} Size16; +typedef struct { char c[3]; } T; +typedef struct { char c[33],c1; } X; + +char c1='a', c2=127, c3=(char)128, c4=(char)255, c5=-1; +short s1=32767, s2=(short)32768, s3=3, s4=4, s5=5, s6=6, s7=7, s8=8, s9=9; +int i1=1, i2=2, i3=3, i4=4, i5=5, i6=6, i7=7, i8=8, i9=9, + i10=11, i11=12, i12=13, i13=14, i14=15, i15=16, i16=17; +long l1=1, l2=2, l3=3, l4=4, l5=5, l6=6, l7=7, l8=8, l9=9; +long long ll1 = 3875056143130689530LL; +float f1=0.1f, f2=0.2f, f3=0.3f, f4=0.4f, f5=0.5f, f6=0.6f, f7=0.7f, f8=0.8f, f9=0.9f, + f10=1.1f, f11=1.2f, f12=1.3f, f13=1.4f, f14=1.5f, f15=1.6f, f16=1.7f, f17=1.8f, + f18=1.9f, f19=2.1f, f20=2.2f, f21=2.3f, f22=2.4f, f23=2.5f, f24=2.6f; +double d1=0.1, d2=0.2, d3=0.3, d4=0.4, d5=0.5, d6=0.6, d7=0.7, d8=0.8, d9=0.9, + d10=1.1, d11=1.2, d12=1.3, d13=1.4, d14=1.5, d15=1.6, d16=1.7, d17=1.8; + +uchar uc1='a', uc2=127, uc3=128, uc4=255, uc5=(uchar)-1; +ushort us1=1, us2=2, us3=3, us4=4, us5=5, us6=6, us7=7, us8=8, us9=9; +uint ui1=1, ui2=2, ui3=3, ui4=4, ui5=5, ui6=6, ui7=7, ui8=8, ui9=9; +ulong ul1=1, ul2=2, ul3=3, ul4=4, ul5=5, ul6=6, ul7=7, ul8=8, ul9=9; + +char *str1="hello",str2[]="goodbye",*str3="still here?"; +Char C1={'A'}, C2={'B'}, C3={'C'}, C4={'\377'}, C5={(char)(-1)}; +Short S1={1}, S2={2}, S3={3}, S4={4}, S5={5}, S6={6}, S7={7}, S8={8}, S9={9}; +Int I1={1}, I2={2}, I3={3}, I4={4}, I5={5}, I6={6}, I7={7}, I8={8}, I9={9}; +Float F1={0.1f}, F2={0.2f}, F3={0.3f}, F4={0.4f}, F5={0.5f}, F6={0.6f}, F7={0.7f}, F8={0.8f}, F9={0.9f}; +Double D1={0.1}, D2={0.2}, D3={0.3}, D4={0.4}, D5={0.5}, D6={0.6}, D7={0.7}, D8={0.8}, D9={0.9}; + +A A1={'a',0.1f},A2={'b',0.2f},A3={'\377',0.3f}; +B B1={0.1,{1,2,3}},B2={0.2,{5,4,3}}; +J J1={47,11},J2={73,55}; +K K1={19,69,12,28}; +L L1={561,1105,1729,2465,2821,6601}; /* A002997 */ +Size1 Size1_1={'a'}; +Size2 Size2_1={'a','b'}; +Size3 Size3_1={'a','b','c'}; +Size4 Size4_1={'a','b','c','d'}; +Size7 Size7_1={'a','b','c','d','e','f','g'}; +Size8 Size8_1={'a','b','c','d','e','f','g','h'}; +Size12 Size12_1={'a','b','c','d','e','f','g','h','i','j','k','l'}; +Size15 Size15_1={'a','b','c','d','e','f','g','h','i','j','k','l','m','n','o'}; +Size16 Size16_1={'a','b','c','d','e','f','g','h','i','j','k','l','m','n','o','p'}; +T T1={{'t','h','e'}},T2={{'f','o','x'}}; +X X1={"abcdefghijklmnopqrstuvwxyzABCDEF",'G'}, X2={"123",'9'}, X3={"return-return-return",'R'}; + +#if defined(__GNUC__) +#define __STDCALL__ __attribute__((stdcall)) +#define __THISCALL__ __attribute__((thiscall)) +#define __FASTCALL__ __attribute__((fastcall)) +#define __MSABI__ __attribute__((ms_abi)) +#else +#define __STDCALL__ __stdcall +#define __THISCALL__ __thiscall +#define __FASTCALL__ __fastcall +#endif + +#ifndef ABI_ATTR +#define ABI_ATTR +#endif + +/* void tests */ +void ABI_ATTR v_v (void) +{ + fprintf(out,"void f(void):\n"); + fflush(out); +} + +/* int tests */ +int ABI_ATTR i_v (void) +{ + int r=99; + fprintf(out,"int f(void):"); + fflush(out); + return r; +} +int ABI_ATTR i_i (int a) +{ + int r=a+1; + fprintf(out,"int f(int):(%d)",a); + fflush(out); + return r; +} +int ABI_ATTR i_i2 (int a, int b) +{ + int r=a+b; + fprintf(out,"int f(2*int):(%d,%d)",a,b); + fflush(out); + return r; +} +int ABI_ATTR i_i4 (int a, int b, int c, int d) +{ + int r=a+b+c+d; + fprintf(out,"int f(4*int):(%d,%d,%d,%d)",a,b,c,d); + fflush(out); + return r; +} +int ABI_ATTR i_i8 (int a, int b, int c, int d, int e, int f, int g, int h) +{ + int r=a+b+c+d+e+f+g+h; + fprintf(out,"int f(8*int):(%d,%d,%d,%d,%d,%d,%d,%d)",a,b,c,d,e,f,g,h); + fflush(out); + return r; +} +int ABI_ATTR i_i16 (int a, int b, int c, int d, int e, int f, int g, int h, + int i, int j, int k, int l, int m, int n, int o, int p) +{ + int r=a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p; + fprintf(out,"int f(16*int):(%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d,%d)", + a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p); + fflush(out); + return r; +} + +/* float tests */ +float ABI_ATTR f_f (float a) +{ + float r=a+1.0f; + fprintf(out,"float f(float):(%g)",a); + fflush(out); + return r; +} +float ABI_ATTR f_f2 (float a, float b) +{ + float r=a+b; + fprintf(out,"float f(2*float):(%g,%g)",a,b); + fflush(out); + return r; +} +float ABI_ATTR f_f4 (float a, float b, float c, float d) +{ + float r=a+b+c+d; + fprintf(out,"float f(4*float):(%g,%g,%g,%g)",a,b,c,d); + fflush(out); + return r; +} +float ABI_ATTR f_f8 (float a, float b, float c, float d, float e, float f, + float g, float h) +{ + float r=a+b+c+d+e+f+g+h; + fprintf(out,"float f(8*float):(%g,%g,%g,%g,%g,%g,%g,%g)",a,b,c,d,e,f,g,h); + fflush(out); + return r; +} +float ABI_ATTR f_f16 (float a, float b, float c, float d, float e, float f, float g, float h, + float i, float j, float k, float l, float m, float n, float o, float p) +{ + float r=a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p; + fprintf(out,"float f(16*float):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g)",a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p); + fflush(out); + return r; +} +float ABI_ATTR f_f24 (float a, float b, float c, float d, float e, float f, float g, float h, + float i, float j, float k, float l, float m, float n, float o, float p, + float q, float s, float t, float u, float v, float w, float x, float y) +{ + float r=a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p+q+s+t+u+v+w+x+y; + fprintf(out,"float f(24*float):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g)",a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,s,t,u,v,w,x,y); + fflush(out); + return r; +} + +/* double tests */ +double ABI_ATTR d_d (double a) +{ + double r=a+1.0; + fprintf(out,"double f(double):(%g)",a); + fflush(out); + return r; +} +double ABI_ATTR d_d2 (double a, double b) +{ + double r=a+b; + fprintf(out,"double f(2*double):(%g,%g)",a,b); + fflush(out); + return r; +} +double ABI_ATTR d_d4 (double a, double b, double c, double d) +{ + double r=a+b+c+d; + fprintf(out,"double f(4*double):(%g,%g,%g,%g)",a,b,c,d); + fflush(out); + return r; +} +double ABI_ATTR d_d8 (double a, double b, double c, double d, double e, double f, + double g, double h) +{ + double r=a+b+c+d+e+f+g+h; + fprintf(out,"double f(8*double):(%g,%g,%g,%g,%g,%g,%g,%g)",a,b,c,d,e,f,g,h); + fflush(out); + return r; +} +double ABI_ATTR d_d16 (double a, double b, double c, double d, double e, double f, + double g, double h, double i, double j, double k, double l, + double m, double n, double o, double p) +{ + double r=a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p; + fprintf(out,"double f(16*double):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g)",a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p); + fflush(out); + return r; +} + +/* pointer tests */ +void* ABI_ATTR vp_vpdpcpsp (void* a, double* b, char* c, Int* d) +{ + void* ret = (char*)b + 1; + fprintf(out,"void* f(void*,double*,char*,Int*):(0x%p,0x%p,0x%p,0x%p)",a,b,c,d); + fflush(out); + return ret; +} + +/* mixed number tests */ +uchar ABI_ATTR uc_ucsil (uchar a, ushort b, uint c, ulong d) +{ + uchar r = (uchar)-1; + fprintf(out,"uchar f(uchar,ushort,uint,ulong):(%u,%u,%u,%lu)",a,b,c,d); + fflush(out); + return r; +} +double ABI_ATTR d_iidd (int a, int b, double c, double d) +{ + double r = a+b+c+d; + fprintf(out,"double f(int,int,double,double):(%d,%d,%g,%g)",a,b,c,d); + fflush(out); + return r; +} +double ABI_ATTR d_iiidi (int a, int b, int c, double d, int e) +{ + double r = a+b+c+d+e; + fprintf(out,"double f(int,int,int,double,int):(%d,%d,%d,%g,%d)",a,b,c,d,e); + fflush(out); + return r; +} +double ABI_ATTR d_idid (int a, double b, int c, double d) +{ + double r = a+b+c+d; + fprintf(out,"double f(int,double,int,double):(%d,%g,%d,%g)",a,b,c,d); + fflush(out); + return r; +} +double ABI_ATTR d_fdi (float a, double b, int c) +{ + double r = a+b+c; + fprintf(out,"double f(float,double,int):(%g,%g,%d)",a,b,c); + fflush(out); + return r; +} +ushort ABI_ATTR us_cdcd (char a, double b, char c, double d) +{ + ushort r = (ushort)(a + b + c + d); + fprintf(out,"ushort f(char,double,char,double):('%c',%g,'%c',%g)",a,b,c,d); + fflush(out); + return r; +} + +long long ABI_ATTR ll_iiilli (int a, int b, int c, long long d, int e) +{ + long long r = (long long)(int)a+(long long)(int)b+(long long)(int)c+d+(long long)(int)e; + fprintf(out,"long long f(int,int,int,long long,int):(%d,%d,%d,0x%lx%08lx,%d)",a,b,c,(long)(d>>32),(long)(d&0xffffffff),e); + fflush(out); + return r; +} +long long ABI_ATTR ll_flli (float a, long long b, int c) +{ + long long r = (long long)(int)a + b + (long long)c; + fprintf(out,"long long f(float,long long,int):(%g,0x%lx%08lx,0x%lx)",a,(long)(b>>32),(long)(b&0xffffffff),(long)c); + fflush(out); + return r; +} + +float ABI_ATTR f_fi (float a, int z) +{ + float r = a+z; + fprintf(out,"float f(float,int):(%g,%d)",a,z); + fflush(out); + return r; +} +float ABI_ATTR f_f2i (float a, float b, int z) +{ + float r = a+b+z; + fprintf(out,"float f(2*float,int):(%g,%g,%d)",a,b,z); + fflush(out); + return r; +} +float ABI_ATTR f_f3i (float a, float b, float c, int z) +{ + float r = a+b+c+z; + fprintf(out,"float f(3*float,int):(%g,%g,%g,%d)",a,b,c,z); + fflush(out); + return r; +} +float ABI_ATTR f_f4i (float a, float b, float c, float d, int z) +{ + float r = a+b+c+d+z; + fprintf(out,"float f(4*float,int):(%g,%g,%g,%g,%d)",a,b,c,d,z); + fflush(out); + return r; +} +float ABI_ATTR f_f7i (float a, float b, float c, float d, float e, float f, float g, + int z) +{ + float r = a+b+c+d+e+f+g+z; + fprintf(out,"float f(7*float,int):(%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,z); + fflush(out); + return r; +} +float ABI_ATTR f_f8i (float a, float b, float c, float d, float e, float f, float g, + float h, int z) +{ + float r = a+b+c+d+e+f+g+h+z; + fprintf(out,"float f(8*float,int):(%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,z); + fflush(out); + return r; +} +float ABI_ATTR f_f12i (float a, float b, float c, float d, float e, float f, float g, + float h, float i, float j, float k, float l, int z) +{ + float r = a+b+c+d+e+f+g+h+i+j+k+l+z; + fprintf(out,"float f(12*float,int):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,i,j,k,l,z); + fflush(out); + return r; +} +float ABI_ATTR f_f13i (float a, float b, float c, float d, float e, float f, float g, + float h, float i, float j, float k, float l, float m, int z) +{ + float r = a+b+c+d+e+f+g+h+i+j+k+l+m+z; + fprintf(out,"float f(13*float,int):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,i,j,k,l,m,z); + fflush(out); + return r; +} + +double ABI_ATTR d_di (double a, int z) +{ + double r = a+z; + fprintf(out,"double f(double,int):(%g,%d)",a,z); + fflush(out); + return r; +} +double ABI_ATTR d_d2i (double a, double b, int z) +{ + double r = a+b+z; + fprintf(out,"double f(2*double,int):(%g,%g,%d)",a,b,z); + fflush(out); + return r; +} +double ABI_ATTR d_d3i (double a, double b, double c, int z) +{ + double r = a+b+c+z; + fprintf(out,"double f(3*double,int):(%g,%g,%g,%d)",a,b,c,z); + fflush(out); + return r; +} +double ABI_ATTR d_d4i (double a, double b, double c, double d, int z) +{ + double r = a+b+c+d+z; + fprintf(out,"double f(4*double,int):(%g,%g,%g,%g,%d)",a,b,c,d,z); + fflush(out); + return r; +} +double ABI_ATTR d_d7i (double a, double b, double c, double d, double e, double f, + double g, int z) +{ + double r = a+b+c+d+e+f+g+z; + fprintf(out,"double f(7*double,int):(%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,z); + fflush(out); + return r; +} +double ABI_ATTR d_d8i (double a, double b, double c, double d, double e, double f, + double g, double h, int z) +{ + double r = a+b+c+d+e+f+g+h+z; + fprintf(out,"double f(8*double,int):(%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,z); + fflush(out); + return r; +} +double ABI_ATTR d_d12i (double a, double b, double c, double d, double e, double f, + double g, double h, double i, double j, double k, double l, + int z) +{ + double r = a+b+c+d+e+f+g+h+i+j+k+l+z; + fprintf(out,"double f(12*double,int):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,i,j,k,l,z); + fflush(out); + return r; +} +double ABI_ATTR d_d13i (double a, double b, double c, double d, double e, double f, + double g, double h, double i, double j, double k, double l, + double m, int z) +{ + double r = a+b+c+d+e+f+g+h+i+j+k+l+m+z; + fprintf(out,"double f(13*double,int):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%d)",a,b,c,d,e,f,g,h,i,j,k,l,m,z); + fflush(out); + return r; +} + +/* small structure return tests */ +Size1 ABI_ATTR S1_v (void) +{ + fprintf(out,"Size1 f(void):"); + fflush(out); + return Size1_1; +} +Size2 ABI_ATTR S2_v (void) +{ + fprintf(out,"Size2 f(void):"); + fflush(out); + return Size2_1; +} +Size3 ABI_ATTR S3_v (void) +{ + fprintf(out,"Size3 f(void):"); + fflush(out); + return Size3_1; +} +Size4 ABI_ATTR S4_v (void) +{ + fprintf(out,"Size4 f(void):"); + fflush(out); + return Size4_1; +} +Size7 ABI_ATTR S7_v (void) +{ + fprintf(out,"Size7 f(void):"); + fflush(out); + return Size7_1; +} +Size8 ABI_ATTR S8_v (void) +{ + fprintf(out,"Size8 f(void):"); + fflush(out); + return Size8_1; +} +Size12 ABI_ATTR S12_v (void) +{ + fprintf(out,"Size12 f(void):"); + fflush(out); + return Size12_1; +} +Size15 ABI_ATTR S15_v (void) +{ + fprintf(out,"Size15 f(void):"); + fflush(out); + return Size15_1; +} +Size16 ABI_ATTR S16_v (void) +{ + fprintf(out,"Size16 f(void):"); + fflush(out); + return Size16_1; +} + +/* structure tests */ +Int ABI_ATTR I_III (Int a, Int b, Int c) +{ + Int r; + r.x = a.x + b.x + c.x; + fprintf(out,"Int f(Int,Int,Int):({%d},{%d},{%d})",a.x,b.x,c.x); + fflush(out); + return r; +} +Char ABI_ATTR C_CdC (Char a, double b, Char c) +{ + Char r; + r.x = (a.x + c.x)/2; + fprintf(out,"Char f(Char,double,Char):({'%c'},%g,{'%c'})",a.x,b,c.x); + fflush(out); + return r; +} +Float ABI_ATTR F_Ffd (Float a, float b, double c) +{ + Float r; + r.x = (float) (a.x + b + c); + fprintf(out,"Float f(Float,float,double):({%g},%g,%g)",a.x,b,c); + fflush(out); + return r; +} +Double ABI_ATTR D_fDd (float a, Double b, double c) +{ + Double r; + r.x = a + b.x + c; + fprintf(out,"Double f(float,Double,double):(%g,{%g},%g)",a,b.x,c); + fflush(out); + return r; +} +Double ABI_ATTR D_Dfd (Double a, float b, double c) +{ + Double r; + r.x = a.x + b + c; + fprintf(out,"Double f(Double,float,double):({%g},%g,%g)",a.x,b,c); + fflush(out); + return r; +} +J ABI_ATTR J_JiJ (J a, int b, J c) +{ + J r; + r.l1 = a.l1+c.l1; r.l2 = a.l2+b+c.l2; + fprintf(out,"J f(J,int,J):({%ld,%ld},%d,{%ld,%ld})",a.l1,a.l2,b,c.l1,c.l2); + fflush(out); + return r; +} +T ABI_ATTR T_TcT (T a, char b, T c) +{ + T r; + r.c[0]='b'; r.c[1]=c.c[1]; r.c[2]=c.c[2]; + fprintf(out,"T f(T,char,T):({\"%c%c%c\"},'%c',{\"%c%c%c\"})",a.c[0],a.c[1],a.c[2],b,c.c[0],c.c[1],c.c[2]); + fflush(out); + return r; +} +X ABI_ATTR X_BcdB (B a, char b, double c, B d) +{ + static X xr={"return val",'R'}; + X r; + r = xr; + r.c1 = b; + fprintf(out,"X f(B,char,double,B):({%g,{%d,%d,%d}},'%c',%g,{%g,{%d,%d,%d}})", + a.d,a.i[0],a.i[1],a.i[2],b,c,d.d,d.i[0],d.i[1],d.i[2]); + fflush(out); + return r; +} + +/* Test for cases where some argument (especially structure, 'long long', or + 'double') may be passed partially in general-purpose argument registers + and partially on the stack. Different ABIs pass between 4 and 8 arguments + (or none) in general-purpose argument registers. */ + +long ABI_ATTR l_l0K (K b, long c) +{ + long r = b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(K,long):(%ld,%ld,%ld,%ld,%ld)",b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + return r; +} +long ABI_ATTR l_l1K (long a1, K b, long c) +{ + long r = a1 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld)",a1,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + return r; +} +long ABI_ATTR l_l2K (long a1, long a2, K b, long c) +{ + long r = a1 + a2 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(2*long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a1,a2,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + return r; +} +long ABI_ATTR l_l3K (long a1, long a2, long a3, K b, long c) +{ + long r = a1 + a2 + a3 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(3*long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a1,a2,a3,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + return r; +} +long ABI_ATTR l_l4K (long a1, long a2, long a3, long a4, K b, long c) +{ + long r = a1 + a2 + a3 + a4 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(4*long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a1,a2,a3,a4,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + return r; +} +long ABI_ATTR l_l5K (long a1, long a2, long a3, long a4, long a5, K b, long c) +{ + long r = a1 + a2 + a3 + a4 + a5 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(5*long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a1,a2,a3,a4,a5,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + return r; +} +long ABI_ATTR l_l6K (long a1, long a2, long a3, long a4, long a5, long a6, K b, long c) +{ + long r = a1 + a2 + a3 + a4 + a5 + a6 + b.l1 + b.l2 + b.l3 + b.l4 + c; + fprintf(out,"long f(6*long,K,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a1,a2,a3,a4,a5,a6,b.l1,b.l2,b.l3,b.l4,c); + fflush(out); + return r; +} +/* These tests is crafted on the knowledge that for all known ABIs: + * 17 > number of floating-point argument registers, + * 3 < number of general-purpose argument registers < 3 + 6. */ +float ABI_ATTR f_f17l3L (float a, float b, float c, float d, float e, float f, float g, + float h, float i, float j, float k, float l, float m, float n, + float o, float p, float q, + long s, long t, long u, L z) +{ + float r = a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p+q+s+t+u+z.l1+z.l2+z.l3+z.l4+z.l5+z.l6; + fprintf(out,"float f(17*float,3*int,L):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,s,t,u,z.l1,z.l2,z.l3,z.l4,z.l5,z.l6); + fflush(out); + return r; +} +double ABI_ATTR d_d17l3L (double a, double b, double c, double d, double e, double f, + double g, double h, double i, double j, double k, double l, + double m, double n, double o, double p, double q, + long s, long t, long u, L z) +{ + double r = a+b+c+d+e+f+g+h+i+j+k+l+m+n+o+p+q+s+t+u+z.l1+z.l2+z.l3+z.l4+z.l5+z.l6; + fprintf(out,"double f(17*double,3*int,L):(%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%g,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld,%ld)",a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,s,t,u,z.l1,z.l2,z.l3,z.l4,z.l5,z.l6); + fflush(out); + return r; +} + +long long ABI_ATTR ll_l2ll (long a1, long a2, long long b, long c) +{ + long long r = (long long) (a1 + a2) + b + c; + fprintf(out,"long long f(2*long,long long,long):(%ld,%ld,0x%lx%08lx,%ld)",a1,a2,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + return r; +} +long long ABI_ATTR ll_l3ll (long a1, long a2, long a3, long long b, long c) +{ + long long r = (long long) (a1 + a2 + a3) + b + c; + fprintf(out,"long long f(3*long,long long,long):(%ld,%ld,%ld,0x%lx%08lx,%ld)",a1,a2,a3,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + return r; +} +long long ABI_ATTR ll_l4ll (long a1, long a2, long a3, long a4, long long b, long c) +{ + long long r = (long long) (a1 + a2 + a3 + a4) + b + c; + fprintf(out,"long long f(4*long,long long,long):(%ld,%ld,%ld,%ld,0x%lx%08lx,%ld)",a1,a2,a3,a4,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + return r; +} +long long ABI_ATTR ll_l5ll (long a1, long a2, long a3, long a4, long a5, long long b, long c) +{ + long long r = (long long) (a1 + a2 + a3 + a4 + a5) + b + c; + fprintf(out,"long long f(5*long,long long,long):(%ld,%ld,%ld,%ld,%ld,0x%lx%08lx,%ld)",a1,a2,a3,a4,a5,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + return r; +} +long long ABI_ATTR ll_l6ll (long a1, long a2, long a3, long a4, long a5, long a6, long long b, long c) +{ + long long r = (long long) (a1 + a2 + a3 + a4 + a5 + a6) + b + c; + fprintf(out,"long long f(6*long,long long,long):(%ld,%ld,%ld,%ld,%ld,%ld,0x%lx%08lx,%ld)",a1,a2,a3,a4,a5,a6,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + return r; +} +long long ABI_ATTR ll_l7ll (long a1, long a2, long a3, long a4, long a5, long a6, long a7, long long b, long c) +{ + long long r = (long long) (a1 + a2 + a3 + a4 + a5 + a6 + a7) + b + c; + fprintf(out,"long long f(7*long,long long,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,0x%lx%08lx,%ld)",a1,a2,a3,a4,a5,a6,a7,(long)(b>>32),(long)(b&0xffffffff),c); + fflush(out); + return r; +} + +double ABI_ATTR d_l2d (long a1, long a2, double b, long c) +{ + double r = (double) (a1 + a2) + b + c; + fprintf(out,"double f(2*long,double,long):(%ld,%ld,%g,%ld)",a1,a2,b,c); + fflush(out); + return r; +} +double ABI_ATTR d_l3d (long a1, long a2, long a3, double b, long c) +{ + double r = (double) (a1 + a2 + a3) + b + c; + fprintf(out,"double f(3*long,double,long):(%ld,%ld,%ld,%g,%ld)",a1,a2,a3,b,c); + fflush(out); + return r; +} +double ABI_ATTR d_l4d (long a1, long a2, long a3, long a4, double b, long c) +{ + double r = (double) (a1 + a2 + a3 + a4) + b + c; + fprintf(out,"double f(4*long,double,long):(%ld,%ld,%ld,%ld,%g,%ld)",a1,a2,a3,a4,b,c); + fflush(out); + return r; +} +double ABI_ATTR d_l5d (long a1, long a2, long a3, long a4, long a5, double b, long c) +{ + double r = (double) (a1 + a2 + a3 + a4 + a5) + b + c; + fprintf(out,"double f(5*long,double,long):(%ld,%ld,%ld,%ld,%ld,%g,%ld)",a1,a2,a3,a4,a5,b,c); + fflush(out); + return r; +} +double ABI_ATTR d_l6d (long a1, long a2, long a3, long a4, long a5, long a6, double b, long c) +{ + double r = (double) (a1 + a2 + a3 + a4 + a5 + a6) + b + c; + fprintf(out,"double f(6*long,double,long):(%ld,%ld,%ld,%ld,%ld,%ld,%g,%ld)",a1,a2,a3,a4,a5,a6,b,c); + fflush(out); + return r; +} +double ABI_ATTR d_l7d (long a1, long a2, long a3, long a4, long a5, long a6, long a7, double b, long c) +{ + double r = (double) (a1 + a2 + a3 + a4 + a5 + a6 + a7) + b + c; + fprintf(out,"double f(7*long,double,long):(%ld,%ld,%ld,%ld,%ld,%ld,%ld,%g,%ld)",a1,a2,a3,a4,a5,a6,a7,b,c); + fflush(out); + return r; +} diff --git a/native/libffi/testsuite/libffi.call/align_mixed.c b/native/libffi/testsuite/libffi.call/align_mixed.c index fe8e254813..6f7cfeee9e 100755 --- a/native/libffi/testsuite/libffi.call/align_mixed.c +++ b/native/libffi/testsuite/libffi.call/align_mixed.c @@ -7,7 +7,10 @@ /* { dg-do run } */ #include "ffitest.h" +<<<<<<< HEAD #include +======= +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 static float ABI_ATTR align_arguments(int i1, double f2, @@ -17,11 +20,24 @@ static float ABI_ATTR align_arguments(int i1, return i1+f2+i3+f4; } +<<<<<<< HEAD int main (void) { ffi_cif cif; ffi_type *args[4] = {&ffi_type_int, &ffi_type_double, &ffi_type_int, &ffi_type_double}; float fa[2] = {1,2}; +======= +int main(void) +{ + ffi_cif cif; + ffi_type *args[4] = { + &ffi_type_sint, + &ffi_type_double, + &ffi_type_sint, + &ffi_type_double + }; + double fa[2] = {1,2}; +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 int ia[2] = {1,2}; void *values[4] = {&ia[0], &fa[0], &ia[1], &fa[1]}; float f, ff; @@ -34,7 +50,11 @@ int main (void) ffi_call(&cif, FFI_FN(align_arguments), &f, values); +<<<<<<< HEAD if (f - ff < FLT_EPSILON) +======= + if (f == ff) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 printf("align arguments tests ok!\n"); else CHECK(0); diff --git a/native/libffi/testsuite/libffi.call/align_stdcall.c b/native/libffi/testsuite/libffi.call/align_stdcall.c new file mode 100644 index 0000000000..5e5cb860cc --- /dev/null +++ b/native/libffi/testsuite/libffi.call/align_stdcall.c @@ -0,0 +1,46 @@ +/* Area: ffi_call + Purpose: Check for proper argument alignment. + Limitations: none. + PR: none. + Originator: (from many_win32.c) */ + +/* { dg-do run } */ + +#include "ffitest.h" + +static float ABI_ATTR align_arguments(int i1, + double f2, + int i3, + double f4) +{ + return i1+f2+i3+f4; +} + +int main(void) +{ + ffi_cif cif; + ffi_type *args[4] = { + &ffi_type_sint, + &ffi_type_double, + &ffi_type_sint, + &ffi_type_double + }; + double fa[2] = {1,2}; + int ia[2] = {1,2}; + void *values[4] = {&ia[0], &fa[0], &ia[1], &fa[1]}; + float f, ff; + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, ABI_NUM, 4, + &ffi_type_float, args) == FFI_OK); + + ff = align_arguments(ia[0], fa[0], ia[1], fa[1]);; + + ffi_call(&cif, FFI_FN(align_arguments), &f, values); + + if (f == ff) + printf("align arguments tests ok!\n"); + else + CHECK(0); + exit(0); +} diff --git a/native/libffi/testsuite/libffi.call/call.exp b/native/libffi/testsuite/libffi.call/call.exp index 5177f07f11..9a69f04a51 100644 --- a/native/libffi/testsuite/libffi.call/call.exp +++ b/native/libffi/testsuite/libffi.call/call.exp @@ -19,6 +19,7 @@ libffi-init global srcdir subdir +<<<<<<< HEAD set tlist [lsearch -inline -all -not -glob [lsort [glob -nocomplain -- $srcdir/$subdir/*.{c,cc}]] *complex*] set ctlist [lsearch -inline -all -glob [lsort [glob -nocomplain -- $srcdir/$subdir/*.{c,cc}]] *complex*] @@ -33,7 +34,34 @@ if { ![istarget s390*] } { } else { run-many-tests $ctlist "" +======= +if { [string match $compiler_vendor "microsoft"] } { + # -wd4005 macro redefinition + # -wd4244 implicit conversion to type of smaller size + # -wd4305 truncation to smaller type + # -wd4477 printf %lu of uintptr_t + # -wd4312 implicit conversion to type of greater size + # -wd4311 pointer truncation to unsigned long + # -EHsc C++ Exception Handling (no SEH exceptions) + set additional_options "-wd4005 -wd4244 -wd4305 -wd4477 -wd4312 -wd4311 -EHsc"; +} else { + set additional_options ""; +} + +set tlist [lsort [glob -nocomplain -- $srcdir/$subdir/*.c]] + +run-many-tests $tlist $additional_options +set tlist [lsort [glob -nocomplain -- $srcdir/$subdir/*.cc]] +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 + +# No C++ for or1k +if { [istarget "or1k-*-*"] } { + foreach test $tlist { + unsupported "$test" + } +} else { + run-many-tests $tlist $additional_options } dg-finish diff --git a/native/libffi/testsuite/libffi.call/ffitest.h b/native/libffi/testsuite/libffi.call/ffitest.h index 15d5e44123..1880cc7bf8 100644 --- a/native/libffi/testsuite/libffi.call/ffitest.h +++ b/native/libffi/testsuite/libffi.call/ffitest.h @@ -24,6 +24,10 @@ #define __STDCALL__ __attribute__((stdcall)) #define __THISCALL__ __attribute__((thiscall)) #define __FASTCALL__ __attribute__((fastcall)) +<<<<<<< HEAD +======= +#define __MSABI__ __attribute__((ms_abi)) +>>>>>>> 8cc8f446f5aac13e107161dffbc15d1ee1a58878 #else #define __UNUSED__ #define __STDCALL__ __stdcall @@ -63,7 +67,7 @@ #endif /* MinGW kludge. */ -#ifdef _WIN64 +#if defined(_WIN64) | defined(_WIN32) #define PRIdLL "I64d" #define PRIuLL "I64u" #else @@ -123,12 +127,14 @@ /* MSVC kludge. */ #if defined _MSC_VER +#if !defined(__cplusplus) || defined(__STDC_FORMAT_MACROS) #define PRIuPTR "lu" #define PRIu8 "u" #define PRId8 "d" #define PRIu64 "I64u" #define PRId64 "I64d" #endif +#endif #ifndef PRIuPTR #define PRIuPTR "u" diff --git a/native/libffi/testsuite/libffi.call/float1.c b/native/libffi/testsuite/libffi.call/float1.c index 991d059fe4..c48493c6b2 100644 --- a/native/libffi/testsuite/libffi.call/float1.c +++ b/native/libffi/testsuite/libffi.call/float1.c @@ -8,6 +8,8 @@ #include "ffitest.h" #include "float.h" +#include + typedef union { double d; @@ -47,7 +49,7 @@ int main (void) /* These are not always the same!! Check for a reasonable delta */ - CHECK(result[0].d - dblit(f) < DBL_EPSILON); + CHECK(fabs(result[0].d - dblit(f)) < DBL_EPSILON); /* Check the canary. */ for (i = 0; i < sizeof (double); ++i) diff --git a/native/libffi/testsuite/libffi.call/float2.c b/native/libffi/testsuite/libffi.call/float2.c index a0b296cf4b..57cd9e30f9 100644 --- a/native/libffi/testsuite/libffi.call/float2.c +++ b/native/libffi/testsuite/libffi.call/float2.c @@ -3,13 +3,13 @@ Limitations: none. PR: none. Originator: From the original ffitest.c */ - -/* { dg-excess-errors "fails" { target x86_64-*-mingw* x86_64-*-cygwin* } } */ -/* { dg-do run { xfail x86_64-*-mingw* x86_64-*-cygwin* } } */ +/* { dg-do run } */ #include "ffitest.h" #include "float.h" +#include + static long double ldblit(float f) { return (long double) (((long double) f)/ (long double) 3.0); @@ -22,6 +22,7 @@ int main (void) void *values[MAX_ARGS]; float f; long double ld; + long double original; args[0] = &ffi_type_float; values[0] = &f; @@ -32,24 +33,26 @@ int main (void) f = 3.14159; -#if 1 - /* This is ifdef'd out for now. long double support under SunOS/gcc - is pretty much non-existent. You'll get the odd bus error in library - routines like printf(). */ +#if defined(__sun) && defined(__GNUC__) + /* long double support under SunOS/gcc is pretty much non-existent. + You'll get the odd bus error in library routines like printf() */ +#else printf ("%Lf\n", ldblit(f)); #endif + ld = 666; ffi_call(&cif, FFI_FN(ldblit), &ld, values); -#if 1 - /* This is ifdef'd out for now. long double support under SunOS/gcc - is pretty much non-existent. You'll get the odd bus error in library - routines like printf(). */ +#if defined(__sun) && defined(__GNUC__) + /* long double support under SunOS/gcc is pretty much non-existent. + You'll get the odd bus error in library routines like printf() */ +#else printf ("%Lf, %Lf, %Lf, %Lf\n", ld, ldblit(f), ld - ldblit(f), LDBL_EPSILON); #endif /* These are not always the same!! Check for a reasonable delta */ - if (ld - ldblit(f) < LDBL_EPSILON) + original = ldblit(f); + if (((ld > original) ? (ld - original) : (original - ld)) < LDBL_EPSILON) puts("long double return value tests ok!"); else CHECK(0); diff --git a/native/libffi/testsuite/libffi.call/float3.c b/native/libffi/testsuite/libffi.call/float3.c index 76bd5f287c..bab3206a17 100644 --- a/native/libffi/testsuite/libffi.call/float3.c +++ b/native/libffi/testsuite/libffi.call/float3.c @@ -9,6 +9,8 @@ #include "ffitest.h" #include "float.h" +#include + static double floating_1(float a, double b, long double c) { return (double) a + b + (double) c; @@ -49,7 +51,7 @@ int main (void) ffi_call(&cif, FFI_FN(floating_1), &rd, values); - CHECK(rd - floating_1(f, d, ld) < DBL_EPSILON); + CHECK(fabs(rd - floating_1(f, d, ld)) < DBL_EPSILON); args[0] = &ffi_type_longdouble; values[0] = &ld; @@ -66,7 +68,7 @@ int main (void) ffi_call(&cif, FFI_FN(floating_2), &rd, values); - CHECK(rd - floating_2(ld, d, f) < DBL_EPSILON); + CHECK(fabs(rd - floating_2(ld, d, f)) < DBL_EPSILON); exit (0); } diff --git a/native/libffi/testsuite/libffi.call/offsets.c b/native/libffi/testsuite/libffi.call/offsets.c new file mode 100644 index 0000000000..23d88b35d5 --- /dev/null +++ b/native/libffi/testsuite/libffi.call/offsets.c @@ -0,0 +1,46 @@ +/* Area: Struct layout + Purpose: Test ffi_get_struct_offsets + Limitations: none. + PR: none. + Originator: Tom Tromey. */ + +/* { dg-do run } */ +#include "ffitest.h" +#include + +struct test_1 +{ + char c; + float f; + char c2; + int i; +}; + +int +main (void) +{ + ffi_type test_1_type; + ffi_type *test_1_elements[5]; + size_t test_1_offsets[4]; + + test_1_elements[0] = &ffi_type_schar; + test_1_elements[1] = &ffi_type_float; + test_1_elements[2] = &ffi_type_schar; + test_1_elements[3] = &ffi_type_sint; + test_1_elements[4] = NULL; + + test_1_type.size = 0; + test_1_type.alignment = 0; + test_1_type.type = FFI_TYPE_STRUCT; + test_1_type.elements = test_1_elements; + + CHECK (ffi_get_struct_offsets (FFI_DEFAULT_ABI, &test_1_type, test_1_offsets) + == FFI_OK); + CHECK (test_1_type.size == sizeof (struct test_1)); + CHECK (offsetof (struct test_1, c) == test_1_offsets[0]); + CHECK (offsetof (struct test_1, f) == test_1_offsets[1]); + CHECK (offsetof (struct test_1, c2) == test_1_offsets[2]); + CHECK (offsetof (struct test_1, i) == test_1_offsets[3]); + + return 0; +} diff --git a/native/libffi/testsuite/libffi.call/pr1172638.c b/native/libffi/testsuite/libffi.call/pr1172638.c new file mode 100644 index 0000000000..7da1621cd6 --- /dev/null +++ b/native/libffi/testsuite/libffi.call/pr1172638.c @@ -0,0 +1,127 @@ +/* Area: ffi_call + Purpose: Reproduce bug found in python ctypes + Limitations: none. + PR: Fedora 1174037 */ + +/* { dg-do run } */ +#include "ffitest.h" + +typedef struct { + long x; + long y; +} POINT; + +typedef struct { + long left; + long top; + long right; + long bottom; +} RECT; + +static RECT ABI_ATTR pr_test(int i __UNUSED__, RECT ar __UNUSED__, + RECT* br __UNUSED__, POINT cp __UNUSED__, + RECT dr __UNUSED__, RECT *er __UNUSED__, + POINT fp, RECT gr __UNUSED__) +{ + RECT result; + + result.left = fp.x; + result.right = fp.y; + result.top = fp.x; + result.bottom = fp.y; + + return result; +} + +int main (void) +{ + ffi_cif cif; + ffi_type *args[MAX_ARGS]; + void *values[MAX_ARGS]; + ffi_type point_type, rect_type; + ffi_type *point_type_elements[3]; + ffi_type *rect_type_elements[5]; + + int i; + POINT cp, fp; + RECT ar, br, dr, er, gr; + RECT *p1, *p2; + + /* This is a hack to get a properly aligned result buffer */ + RECT *rect_result = + (RECT *) malloc (sizeof(RECT)); + + point_type.size = 0; + point_type.alignment = 0; + point_type.type = FFI_TYPE_STRUCT; + point_type.elements = point_type_elements; + point_type_elements[0] = &ffi_type_slong; + point_type_elements[1] = &ffi_type_slong; + point_type_elements[2] = NULL; + + rect_type.size = 0; + rect_type.alignment = 0; + rect_type.type = FFI_TYPE_STRUCT; + rect_type.elements = rect_type_elements; + rect_type_elements[0] = &ffi_type_slong; + rect_type_elements[1] = &ffi_type_slong; + rect_type_elements[2] = &ffi_type_slong; + rect_type_elements[3] = &ffi_type_slong; + rect_type_elements[4] = NULL; + + args[0] = &ffi_type_sint; + args[1] = &rect_type; + args[2] = &ffi_type_pointer; + args[3] = &point_type; + args[4] = &rect_type; + args[5] = &ffi_type_pointer; + args[6] = &point_type; + args[7] = &rect_type; + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, ABI_NUM, 8, &rect_type, args) == FFI_OK); + + i = 1; + ar.left = 2; + ar.right = 3; + ar.top = 4; + ar.bottom = 5; + br.left = 6; + br.right = 7; + br.top = 8; + br.bottom = 9; + cp.x = 10; + cp.y = 11; + dr.left = 12; + dr.right = 13; + dr.top = 14; + dr.bottom = 15; + er.left = 16; + er.right = 17; + er.top = 18; + er.bottom = 19; + fp.x = 20; + fp.y = 21; + gr.left = 22; + gr.right = 23; + gr.top = 24; + gr.bottom = 25; + + values[0] = &i; + values[1] = &ar; + p1 = &br; + values[2] = &p1; + values[3] = &cp; + values[4] = &dr; + p2 = &er; + values[5] = &p2; + values[6] = &fp; + values[7] = &gr; + + ffi_call (&cif, FFI_FN(pr_test), rect_result, values); + + CHECK(rect_result->top == 20); + + free (rect_result); + exit(0); +} diff --git a/native/libffi/testsuite/libffi.call/return_ldl.c b/native/libffi/testsuite/libffi.call/return_ldl.c index 5c2fe65aec..52a92fe071 100644 --- a/native/libffi/testsuite/libffi.call/return_ldl.c +++ b/native/libffi/testsuite/libffi.call/return_ldl.c @@ -3,8 +3,8 @@ Limitations: none. PR: none. Originator: 20071113 */ +/* { dg-do run } */ -/* { dg-do run { xfail x86_64-*-mingw* x86_64-*-cygwin* } } */ #include "ffitest.h" static long double return_ldl(long double ldl) diff --git a/native/libffi/testsuite/libffi.call/struct10.c b/native/libffi/testsuite/libffi.call/struct10.c new file mode 100644 index 0000000000..17b1377424 --- /dev/null +++ b/native/libffi/testsuite/libffi.call/struct10.c @@ -0,0 +1,57 @@ +/* Area: ffi_call + Purpose: Check structures. + Limitations: none. + PR: none. + Originator: Sergei Trofimovich + + The test originally discovered in ruby's bindings + for ffi in https://bugs.gentoo.org/634190 */ + +/* { dg-do run } */ +#include "ffitest.h" + +struct s { + int s32; + float f32; + signed char s8; +}; + +struct s make_s(void) { + struct s r; + r.s32 = 0x1234; + r.f32 = 7.0; + r.s8 = 0x78; + return r; +} + +int main() { + ffi_cif cif; + struct s r; + ffi_type rtype; + ffi_type* s_fields[] = { + &ffi_type_sint, + &ffi_type_float, + &ffi_type_schar, + NULL, + }; + + rtype.size = 0; + rtype.alignment = 0, + rtype.type = FFI_TYPE_STRUCT, + rtype.elements = s_fields, + + r.s32 = 0xbad; + r.f32 = 999.999; + r.s8 = 0x51; + + // Here we emulate the following call: + //r = make_s(); + + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 0, &rtype, NULL) == FFI_OK); + ffi_call(&cif, FFI_FN(make_s), &r, NULL); + + CHECK(r.s32 == 0x1234); + CHECK(r.f32 == 7.0); + CHECK(r.s8 == 0x78); + exit(0); +} diff --git a/native/libffi/testsuite/libffi.call/unwindtest.cc b/native/libffi/testsuite/libffi.call/unwindtest.cc index 67cfefeca6..e1145657d1 100644 --- a/native/libffi/testsuite/libffi.call/unwindtest.cc +++ b/native/libffi/testsuite/libffi.call/unwindtest.cc @@ -4,7 +4,7 @@ PR: none. Originator: Jeff Sturm */ -/* { dg-do run } */ +/* { dg-do run { xfail x86_64-apple-darwin* moxie*-*-* } } */ #include "ffitest.h" diff --git a/native/libffi/testsuite/libffi.call/unwindtest_ffi_call.cc b/native/libffi/testsuite/libffi.call/unwindtest_ffi_call.cc index ec611647e3..153d240940 100644 --- a/native/libffi/testsuite/libffi.call/unwindtest_ffi_call.cc +++ b/native/libffi/testsuite/libffi.call/unwindtest_ffi_call.cc @@ -4,7 +4,7 @@ PR: none. Originator: Andreas Tobler 20061213 */ -/* { dg-do run } */ +/* { dg-do run { xfail moxie*-*-* } } */ #include "ffitest.h" diff --git a/native/libffi/testsuite/libffi.call/va_1.c b/native/libffi/testsuite/libffi.call/va_1.c index 7f96809ea9..59d085c832 100644 --- a/native/libffi/testsuite/libffi.call/va_1.c +++ b/native/libffi/testsuite/libffi.call/va_1.c @@ -5,7 +5,7 @@ Originator: ARM Ltd. */ /* { dg-do run } */ -/* { dg-output "" { xfail avr32*-*-* } } */ +/* { dg-output "" { xfail avr32*-*-* m68k-*-* alpha-*-* } } */ #include "ffitest.h" #include diff --git a/native/libffi/testsuite/libffi.closures/closure.exp b/native/libffi/testsuite/libffi.closures/closure.exp new file mode 100644 index 0000000000..ed4145ca84 --- /dev/null +++ b/native/libffi/testsuite/libffi.closures/closure.exp @@ -0,0 +1,67 @@ +# Copyright (C) 2003, 2006, 2009, 2010, 2014, 2019 Free Software Foundation, Inc. +# Copyright (C) 2019 Anthony Green + +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; see the file COPYING3. If not see +# . + +dg-init +libffi-init + +global srcdir subdir + +if { [string match $compiler_vendor "microsoft"] } { + # -wd4005 macro redefinition + # -wd4244 implicit conversion to type of smaller size + # -wd4305 truncation to smaller type + # -wd4477 printf %lu of uintptr_t + # -wd4312 implicit conversion to type of greater size + # -wd4311 pointer truncation to unsigned long + # -EHsc C++ Exception Handling (no SEH exceptions) + set additional_options "-wd4005 -wd4244 -wd4305 -wd4477 -wd4312 -wd4311 -EHsc"; +} else { + set additional_options ""; +} + +set tlist [lsort [glob -nocomplain -- $srcdir/$subdir/*.c]] + +if { [libffi_feature_test "#if FFI_CLOSURES"] } { + run-many-tests $tlist "" +} else { + foreach test $tlist { + unsupported "$test" + } +} + +set tlist [lsort [glob -nocomplain -- $srcdir/$subdir/*.cc]] + +# No C++ for or1k +if { [istarget "or1k-*-*"] } { + foreach test $tlist { + unsupported "$test" + } +} else { + if { [libffi_feature_test "#if FFI_CLOSURES"] } { + run-many-tests $tlist $additional_options + } else { + foreach test $tlist { + unsupported "$test" + } + } +} + +dg-finish + +# Local Variables: +# tcl-indent-level:4 +# End: diff --git a/native/libffi/testsuite/libffi.call/closure_fn0.c b/native/libffi/testsuite/libffi.closures/closure_fn0.c similarity index 100% rename from native/libffi/testsuite/libffi.call/closure_fn0.c rename to native/libffi/testsuite/libffi.closures/closure_fn0.c diff --git a/native/libffi/testsuite/libffi.call/closure_fn1.c b/native/libffi/testsuite/libffi.closures/closure_fn1.c similarity index 100% rename from native/libffi/testsuite/libffi.call/closure_fn1.c rename to native/libffi/testsuite/libffi.closures/closure_fn1.c diff --git a/native/libffi/testsuite/libffi.call/closure_fn2.c b/native/libffi/testsuite/libffi.closures/closure_fn2.c similarity index 100% rename from native/libffi/testsuite/libffi.call/closure_fn2.c rename to native/libffi/testsuite/libffi.closures/closure_fn2.c diff --git a/native/libffi/testsuite/libffi.call/closure_fn3.c b/native/libffi/testsuite/libffi.closures/closure_fn3.c similarity index 100% rename from native/libffi/testsuite/libffi.call/closure_fn3.c rename to native/libffi/testsuite/libffi.closures/closure_fn3.c diff --git a/native/libffi/testsuite/libffi.call/closure_fn4.c b/native/libffi/testsuite/libffi.closures/closure_fn4.c similarity index 100% rename from native/libffi/testsuite/libffi.call/closure_fn4.c rename to native/libffi/testsuite/libffi.closures/closure_fn4.c diff --git a/native/libffi/testsuite/libffi.call/closure_fn5.c b/native/libffi/testsuite/libffi.closures/closure_fn5.c similarity index 100% rename from native/libffi/testsuite/libffi.call/closure_fn5.c rename to native/libffi/testsuite/libffi.closures/closure_fn5.c diff --git a/native/libffi/testsuite/libffi.call/closure_fn6.c b/native/libffi/testsuite/libffi.closures/closure_fn6.c similarity index 100% rename from native/libffi/testsuite/libffi.call/closure_fn6.c rename to native/libffi/testsuite/libffi.closures/closure_fn6.c diff --git a/native/libffi/testsuite/libffi.call/closure_loc_fn0.c b/native/libffi/testsuite/libffi.closures/closure_loc_fn0.c similarity index 100% rename from native/libffi/testsuite/libffi.call/closure_loc_fn0.c rename to native/libffi/testsuite/libffi.closures/closure_loc_fn0.c diff --git a/native/libffi/testsuite/libffi.closures/closure_simple.c b/native/libffi/testsuite/libffi.closures/closure_simple.c new file mode 100644 index 0000000000..5a4e728d4a --- /dev/null +++ b/native/libffi/testsuite/libffi.closures/closure_simple.c @@ -0,0 +1,55 @@ +/* Area: closure_call + Purpose: Check simple closure handling with all ABIs + Limitations: none. + PR: none. + Originator: */ + +/* { dg-do run } */ +#include "ffitest.h" + +static void +closure_test(ffi_cif* cif __UNUSED__, void* resp, void** args, void* userdata) +{ + *(ffi_arg*)resp = + (int)*(int *)args[0] + (int)(*(int *)args[1]) + + (int)(*(int *)args[2]) + (int)(*(int *)args[3]) + + (int)(intptr_t)userdata; + + printf("%d %d %d %d: %d\n", + (int)*(int *)args[0], (int)(*(int *)args[1]), + (int)(*(int *)args[2]), (int)(*(int *)args[3]), + (int)*(ffi_arg *)resp); + +} + +typedef int (ABI_ATTR *closure_test_type0)(int, int, int, int); + +int main (void) +{ + ffi_cif cif; + void *code; + ffi_closure *pcl = ffi_closure_alloc(sizeof(ffi_closure), &code); + ffi_type * cl_arg_types[17]; + int res; + + cl_arg_types[0] = &ffi_type_uint; + cl_arg_types[1] = &ffi_type_uint; + cl_arg_types[2] = &ffi_type_uint; + cl_arg_types[3] = &ffi_type_uint; + cl_arg_types[4] = NULL; + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, ABI_NUM, 4, + &ffi_type_sint, cl_arg_types) == FFI_OK); + + CHECK(ffi_prep_closure_loc(pcl, &cif, closure_test, + (void *) 3 /* userdata */, code) == FFI_OK); + + res = (*(closure_test_type0)code)(0, 1, 2, 3); + /* { dg-output "0 1 2 3: 9" } */ + + printf("res: %d\n",res); + /* { dg-output "\nres: 9" } */ + + exit(0); +} diff --git a/native/libffi/testsuite/libffi.call/cls_12byte.c b/native/libffi/testsuite/libffi.closures/cls_12byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_12byte.c rename to native/libffi/testsuite/libffi.closures/cls_12byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_16byte.c b/native/libffi/testsuite/libffi.closures/cls_16byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_16byte.c rename to native/libffi/testsuite/libffi.closures/cls_16byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_18byte.c b/native/libffi/testsuite/libffi.closures/cls_18byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_18byte.c rename to native/libffi/testsuite/libffi.closures/cls_18byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_19byte.c b/native/libffi/testsuite/libffi.closures/cls_19byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_19byte.c rename to native/libffi/testsuite/libffi.closures/cls_19byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_1_1byte.c b/native/libffi/testsuite/libffi.closures/cls_1_1byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_1_1byte.c rename to native/libffi/testsuite/libffi.closures/cls_1_1byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_20byte.c b/native/libffi/testsuite/libffi.closures/cls_20byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_20byte.c rename to native/libffi/testsuite/libffi.closures/cls_20byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_20byte1.c b/native/libffi/testsuite/libffi.closures/cls_20byte1.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_20byte1.c rename to native/libffi/testsuite/libffi.closures/cls_20byte1.c diff --git a/native/libffi/testsuite/libffi.call/cls_24byte.c b/native/libffi/testsuite/libffi.closures/cls_24byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_24byte.c rename to native/libffi/testsuite/libffi.closures/cls_24byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_2byte.c b/native/libffi/testsuite/libffi.closures/cls_2byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_2byte.c rename to native/libffi/testsuite/libffi.closures/cls_2byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_3_1byte.c b/native/libffi/testsuite/libffi.closures/cls_3_1byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_3_1byte.c rename to native/libffi/testsuite/libffi.closures/cls_3_1byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_3byte1.c b/native/libffi/testsuite/libffi.closures/cls_3byte1.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_3byte1.c rename to native/libffi/testsuite/libffi.closures/cls_3byte1.c diff --git a/native/libffi/testsuite/libffi.call/cls_3byte2.c b/native/libffi/testsuite/libffi.closures/cls_3byte2.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_3byte2.c rename to native/libffi/testsuite/libffi.closures/cls_3byte2.c diff --git a/native/libffi/testsuite/libffi.closures/cls_3float.c b/native/libffi/testsuite/libffi.closures/cls_3float.c new file mode 100644 index 0000000000..48888f8405 --- /dev/null +++ b/native/libffi/testsuite/libffi.closures/cls_3float.c @@ -0,0 +1,95 @@ +/* Area: ffi_call, closure_call + Purpose: Check structure passing with different structure size. + Depending on the ABI. Check overlapping. + Limitations:>none. + PR: none. + Originator: 20171026 */ + +/* { dg-do run } */ + +#include "ffitest.h" + +typedef struct cls_struct_3float { + float f; + float g; + float h; +} cls_struct_3float; + +cls_struct_3float cls_struct_3float_fn(struct cls_struct_3float a1, + struct cls_struct_3float a2) +{ + struct cls_struct_3float result; + + result.f = a1.f + a2.f; + result.g = a1.g + a2.g; + result.h = a1.h + a2.h; + + printf("%g %g %g %g %g %g: %g %g %g\n", a1.f, a1.g, a1.h, + a2.f, a2.g, a2.h, result.f, result.g, result.h); + + return result; +} + +static void +cls_struct_3float_gn(ffi_cif *cif __UNUSED__, void* resp, void **args, + void* userdata __UNUSED__) +{ + struct cls_struct_3float a1, a2; + + a1 = *(struct cls_struct_3float*)(args[0]); + a2 = *(struct cls_struct_3float*)(args[1]); + + *(cls_struct_3float*)resp = cls_struct_3float_fn(a1, a2); +} + +int main (void) +{ + ffi_cif cif; + void *code; + ffi_closure *pcl = ffi_closure_alloc(sizeof(ffi_closure), &code); + void *args_dbl[3]; + ffi_type* cls_struct_fields[4]; + ffi_type cls_struct_type; + ffi_type* dbl_arg_types[3]; + + struct cls_struct_3float g_dbl = { 1.0f, 2.0f, 3.0f }; + struct cls_struct_3float f_dbl = { 1.0f, 2.0f, 3.0f }; + struct cls_struct_3float res_dbl; + + cls_struct_fields[0] = &ffi_type_float; + cls_struct_fields[1] = &ffi_type_float; + cls_struct_fields[2] = &ffi_type_float; + cls_struct_fields[3] = NULL; + + cls_struct_type.size = 0; + cls_struct_type.alignment = 0; + cls_struct_type.type = FFI_TYPE_STRUCT; + cls_struct_type.elements = cls_struct_fields; + + dbl_arg_types[0] = &cls_struct_type; + dbl_arg_types[1] = &cls_struct_type; + dbl_arg_types[2] = NULL; + + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 2, &cls_struct_type, + dbl_arg_types) == FFI_OK); + + args_dbl[0] = &g_dbl; + args_dbl[1] = &f_dbl; + args_dbl[2] = NULL; + + ffi_call(&cif, FFI_FN(cls_struct_3float_fn), &res_dbl, args_dbl); + /* { dg-output "1 2 3 1 2 3: 2 4 6" } */ + printf("res: %g %g %g\n", res_dbl.f, res_dbl.g, res_dbl.h); + /* { dg-output "\nres: 2 4 6" } */ + + CHECK(ffi_prep_closure_loc(pcl, &cif, cls_struct_3float_gn, NULL, code) == + FFI_OK); + + res_dbl = ((cls_struct_3float(*)(cls_struct_3float, + cls_struct_3float))(code))(g_dbl, f_dbl); + /* { dg-output "\n1 2 3 1 2 3: 2 4 6" } */ + printf("res: %g %g %g\n", res_dbl.f, res_dbl.g, res_dbl.h); + /* { dg-output "\nres: 2 4 6" } */ + + exit(0); +} diff --git a/native/libffi/testsuite/libffi.call/cls_4_1byte.c b/native/libffi/testsuite/libffi.closures/cls_4_1byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_4_1byte.c rename to native/libffi/testsuite/libffi.closures/cls_4_1byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_4byte.c b/native/libffi/testsuite/libffi.closures/cls_4byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_4byte.c rename to native/libffi/testsuite/libffi.closures/cls_4byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_5_1_byte.c b/native/libffi/testsuite/libffi.closures/cls_5_1_byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_5_1_byte.c rename to native/libffi/testsuite/libffi.closures/cls_5_1_byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_5byte.c b/native/libffi/testsuite/libffi.closures/cls_5byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_5byte.c rename to native/libffi/testsuite/libffi.closures/cls_5byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_64byte.c b/native/libffi/testsuite/libffi.closures/cls_64byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_64byte.c rename to native/libffi/testsuite/libffi.closures/cls_64byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_6_1_byte.c b/native/libffi/testsuite/libffi.closures/cls_6_1_byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_6_1_byte.c rename to native/libffi/testsuite/libffi.closures/cls_6_1_byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_6byte.c b/native/libffi/testsuite/libffi.closures/cls_6byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_6byte.c rename to native/libffi/testsuite/libffi.closures/cls_6byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_7_1_byte.c b/native/libffi/testsuite/libffi.closures/cls_7_1_byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_7_1_byte.c rename to native/libffi/testsuite/libffi.closures/cls_7_1_byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_7byte.c b/native/libffi/testsuite/libffi.closures/cls_7byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_7byte.c rename to native/libffi/testsuite/libffi.closures/cls_7byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_8byte.c b/native/libffi/testsuite/libffi.closures/cls_8byte.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_8byte.c rename to native/libffi/testsuite/libffi.closures/cls_8byte.c diff --git a/native/libffi/testsuite/libffi.call/cls_9byte1.c b/native/libffi/testsuite/libffi.closures/cls_9byte1.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_9byte1.c rename to native/libffi/testsuite/libffi.closures/cls_9byte1.c diff --git a/native/libffi/testsuite/libffi.call/cls_9byte2.c b/native/libffi/testsuite/libffi.closures/cls_9byte2.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_9byte2.c rename to native/libffi/testsuite/libffi.closures/cls_9byte2.c diff --git a/native/libffi/testsuite/libffi.call/cls_align_double.c b/native/libffi/testsuite/libffi.closures/cls_align_double.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_align_double.c rename to native/libffi/testsuite/libffi.closures/cls_align_double.c diff --git a/native/libffi/testsuite/libffi.call/cls_align_float.c b/native/libffi/testsuite/libffi.closures/cls_align_float.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_align_float.c rename to native/libffi/testsuite/libffi.closures/cls_align_float.c diff --git a/native/libffi/testsuite/libffi.call/cls_align_longdouble.c b/native/libffi/testsuite/libffi.closures/cls_align_longdouble.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_align_longdouble.c rename to native/libffi/testsuite/libffi.closures/cls_align_longdouble.c diff --git a/native/libffi/testsuite/libffi.call/cls_align_longdouble_split.c b/native/libffi/testsuite/libffi.closures/cls_align_longdouble_split.c similarity index 95% rename from native/libffi/testsuite/libffi.call/cls_align_longdouble_split.c rename to native/libffi/testsuite/libffi.closures/cls_align_longdouble_split.c index 15f9365465..cc1c43b8ca 100644 --- a/native/libffi/testsuite/libffi.call/cls_align_longdouble_split.c +++ b/native/libffi/testsuite/libffi.closures/cls_align_longdouble_split.c @@ -4,10 +4,8 @@ PR: none. Originator: 20031203 */ -/* { dg-excess-errors "no long double format" { xfail x86_64-*-mingw* x86_64-*-cygwin* } } */ /* { dg-do run { xfail strongarm*-*-* xscale*-*-* } } */ /* { dg-options -mlong-double-128 { target powerpc64*-*-linux* } } */ -/* { dg-output "" { xfail x86_64-*-mingw* x86_64-*-cygwin* } } */ #include "ffitest.h" diff --git a/native/libffi/testsuite/libffi.call/cls_align_longdouble_split2.c b/native/libffi/testsuite/libffi.closures/cls_align_longdouble_split2.c similarity index 95% rename from native/libffi/testsuite/libffi.call/cls_align_longdouble_split2.c rename to native/libffi/testsuite/libffi.closures/cls_align_longdouble_split2.c index ca1c356cbe..5d3bec0719 100644 --- a/native/libffi/testsuite/libffi.call/cls_align_longdouble_split2.c +++ b/native/libffi/testsuite/libffi.closures/cls_align_longdouble_split2.c @@ -5,10 +5,8 @@ Originator: Blake Chaffin 6/18/2007 */ -/* { dg-excess-errors "no long double format" { xfail x86_64-*-mingw* x86_64-*-cygwin* } } */ /* { dg-do run { xfail strongarm*-*-* } } */ /* { dg-options -mlong-double-128 { target powerpc64*-*-linux* } } */ -/* { dg-output "" { xfail x86_64-*-mingw* x86_64-*-cygwin* } } */ #include "ffitest.h" diff --git a/native/libffi/testsuite/libffi.call/cls_align_pointer.c b/native/libffi/testsuite/libffi.closures/cls_align_pointer.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_align_pointer.c rename to native/libffi/testsuite/libffi.closures/cls_align_pointer.c diff --git a/native/libffi/testsuite/libffi.call/cls_align_sint16.c b/native/libffi/testsuite/libffi.closures/cls_align_sint16.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_align_sint16.c rename to native/libffi/testsuite/libffi.closures/cls_align_sint16.c diff --git a/native/libffi/testsuite/libffi.call/cls_align_sint32.c b/native/libffi/testsuite/libffi.closures/cls_align_sint32.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_align_sint32.c rename to native/libffi/testsuite/libffi.closures/cls_align_sint32.c diff --git a/native/libffi/testsuite/libffi.call/cls_align_sint64.c b/native/libffi/testsuite/libffi.closures/cls_align_sint64.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_align_sint64.c rename to native/libffi/testsuite/libffi.closures/cls_align_sint64.c diff --git a/native/libffi/testsuite/libffi.call/cls_align_uint16.c b/native/libffi/testsuite/libffi.closures/cls_align_uint16.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_align_uint16.c rename to native/libffi/testsuite/libffi.closures/cls_align_uint16.c diff --git a/native/libffi/testsuite/libffi.call/cls_align_uint32.c b/native/libffi/testsuite/libffi.closures/cls_align_uint32.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_align_uint32.c rename to native/libffi/testsuite/libffi.closures/cls_align_uint32.c diff --git a/native/libffi/testsuite/libffi.call/cls_align_uint64.c b/native/libffi/testsuite/libffi.closures/cls_align_uint64.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_align_uint64.c rename to native/libffi/testsuite/libffi.closures/cls_align_uint64.c diff --git a/native/libffi/testsuite/libffi.call/cls_dbls_struct.c b/native/libffi/testsuite/libffi.closures/cls_dbls_struct.c similarity index 95% rename from native/libffi/testsuite/libffi.call/cls_dbls_struct.c rename to native/libffi/testsuite/libffi.closures/cls_dbls_struct.c index d6637911e5..e451dea51b 100644 --- a/native/libffi/testsuite/libffi.call/cls_dbls_struct.c +++ b/native/libffi/testsuite/libffi.closures/cls_dbls_struct.c @@ -57,10 +57,10 @@ int main(int argc __UNUSED__, char** argv __UNUSED__) CHECK(ffi_prep_closure_loc(pcl, &cif, closure_test_gn, NULL, code) == FFI_OK); ((void*(*)(Dbls))(code))(arg); - /* { dg-output "1.0 2.0\n" } */ + /* { dg-output "1.0 2.0" } */ closure_test_fn(arg); - /* { dg-output "1.0 2.0\n" } */ + /* { dg-output "\n1.0 2.0" } */ return 0; } diff --git a/native/libffi/testsuite/libffi.call/cls_double.c b/native/libffi/testsuite/libffi.closures/cls_double.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_double.c rename to native/libffi/testsuite/libffi.closures/cls_double.c diff --git a/native/libffi/testsuite/libffi.call/cls_double_va.c b/native/libffi/testsuite/libffi.closures/cls_double_va.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_double_va.c rename to native/libffi/testsuite/libffi.closures/cls_double_va.c diff --git a/native/libffi/testsuite/libffi.call/cls_float.c b/native/libffi/testsuite/libffi.closures/cls_float.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_float.c rename to native/libffi/testsuite/libffi.closures/cls_float.c diff --git a/native/libffi/testsuite/libffi.call/cls_longdouble.c b/native/libffi/testsuite/libffi.closures/cls_longdouble.c similarity index 94% rename from native/libffi/testsuite/libffi.call/cls_longdouble.c rename to native/libffi/testsuite/libffi.closures/cls_longdouble.c index 5dc9ac7685..d24e72e4ae 100644 --- a/native/libffi/testsuite/libffi.call/cls_longdouble.c +++ b/native/libffi/testsuite/libffi.closures/cls_longdouble.c @@ -4,12 +4,10 @@ PR: none. Originator: Blake Chaffin */ -/* { dg-excess-errors "no long double format" { xfail x86_64-*-mingw* x86_64-*-cygwin* } } */ /* This test is known to PASS on armv7l-unknown-linux-gnueabihf, so I have remove the xfail for arm*-*-* below, until we know more. */ /* { dg-do run { xfail strongarm*-*-* xscale*-*-* } } */ /* { dg-options -mlong-double-128 { target powerpc64*-*-linux* } } */ -/* { dg-output "" { xfail x86_64-*-mingw* x86_64-*-cygwin* } } */ #include "ffitest.h" diff --git a/native/libffi/testsuite/libffi.call/cls_longdouble_va.c b/native/libffi/testsuite/libffi.closures/cls_longdouble_va.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_longdouble_va.c rename to native/libffi/testsuite/libffi.closures/cls_longdouble_va.c diff --git a/native/libffi/testsuite/libffi.closures/cls_many_mixed_args.c b/native/libffi/testsuite/libffi.closures/cls_many_mixed_args.c new file mode 100644 index 0000000000..7fd6c8207a --- /dev/null +++ b/native/libffi/testsuite/libffi.closures/cls_many_mixed_args.c @@ -0,0 +1,70 @@ +/* Area: closure_call + Purpose: Check closures called with many args of mixed types + Limitations: none. + PR: none. + Originator: */ + +/* { dg-do run } */ +#include "ffitest.h" +#include +#include + +#define NARGS 16 + +static void cls_ret_double_fn(ffi_cif* cif __UNUSED__, void* resp, void** args, + void* userdata __UNUSED__) +{ + int i; + double r = 0; + double t; + for(i = 0; i < NARGS; i++) + { + if(i == 4 || i == 9 || i == 11 || i == 13 || i == 15) + { + t = *(long int *)args[i]; + CHECK(t == i+1); + } + else + { + t = *(double *)args[i]; + CHECK(fabs(t - ((i+1) * 0.1)) < FLT_EPSILON); + } + r += t; + } + *(double *)resp = r; +} +typedef double (*cls_ret_double)(double, double, double, double, long int, +double, double, double, double, long int, double, long int, double, long int, +double, long int); + +int main (void) +{ + ffi_cif cif; + void *code; + ffi_closure *pcl = ffi_closure_alloc(sizeof(ffi_closure), &code); + ffi_type * cl_arg_types[NARGS]; + double res; + int i; + double expected = 64.9; + + for(i = 0; i < NARGS; i++) + { + if(i == 4 || i == 9 || i == 11 || i == 13 || i == 15) + cl_arg_types[i] = &ffi_type_slong; + else + cl_arg_types[i] = &ffi_type_double; + } + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, NARGS, + &ffi_type_double, cl_arg_types) == FFI_OK); + + CHECK(ffi_prep_closure_loc(pcl, &cif, cls_ret_double_fn, NULL, code) == FFI_OK); + + res = (((cls_ret_double)code))(0.1, 0.2, 0.3, 0.4, 5, 0.6, 0.7, 0.8, 0.9, 10, + 1.1, 12, 1.3, 14, 1.5, 16); + if (fabs(res - expected) < FLT_EPSILON) + exit(0); + else + abort(); +} diff --git a/native/libffi/testsuite/libffi.closures/cls_many_mixed_float_double.c b/native/libffi/testsuite/libffi.closures/cls_many_mixed_float_double.c new file mode 100644 index 0000000000..62b0697ac0 --- /dev/null +++ b/native/libffi/testsuite/libffi.closures/cls_many_mixed_float_double.c @@ -0,0 +1,55 @@ +/* Area: closure_call + Purpose: Check register allocation for closure calls with many float and double arguments + Limitations: none. + PR: none. + Originator: */ + +/* { dg-do run } */ +#include "ffitest.h" +#include +#include + +#define NARGS 16 + +static void cls_mixed_float_double_fn(ffi_cif* cif , void* ret, void** args, + void* userdata __UNUSED__) +{ + double r = 0; + unsigned int i; + double t; + for(i=0; i < cif->nargs; i++) + { + if(cif->arg_types[i] == &ffi_type_double) { + t = *(((double**)(args))[i]); + } else { + t = *(((float**)(args))[i]); + } + r += t; + } + *((double*)ret) = r; +} +typedef double (*cls_mixed)(double, float, double, double, double, double, double, float, float, double, float, float); + +int main (void) +{ + ffi_cif cif; + ffi_closure *closure; + void* code; + ffi_type *argtypes[12] = {&ffi_type_double, &ffi_type_float, &ffi_type_double, + &ffi_type_double, &ffi_type_double, &ffi_type_double, + &ffi_type_double, &ffi_type_float, &ffi_type_float, + &ffi_type_double, &ffi_type_float, &ffi_type_float}; + + + closure = ffi_closure_alloc(sizeof(ffi_closure), (void**)&code); + if(closure ==NULL) + abort(); + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 12, &ffi_type_double, argtypes) == FFI_OK); + CHECK(ffi_prep_closure_loc(closure, &cif, cls_mixed_float_double_fn, NULL, code) == FFI_OK); + double ret = ((cls_mixed)code)(0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.1, 1.2); + ffi_closure_free(closure); + if(fabs(ret - 7.8) < FLT_EPSILON) + exit(0); + else + abort(); +} diff --git a/native/libffi/testsuite/libffi.call/cls_multi_schar.c b/native/libffi/testsuite/libffi.closures/cls_multi_schar.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_multi_schar.c rename to native/libffi/testsuite/libffi.closures/cls_multi_schar.c diff --git a/native/libffi/testsuite/libffi.call/cls_multi_sshort.c b/native/libffi/testsuite/libffi.closures/cls_multi_sshort.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_multi_sshort.c rename to native/libffi/testsuite/libffi.closures/cls_multi_sshort.c diff --git a/native/libffi/testsuite/libffi.call/cls_multi_sshortchar.c b/native/libffi/testsuite/libffi.closures/cls_multi_sshortchar.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_multi_sshortchar.c rename to native/libffi/testsuite/libffi.closures/cls_multi_sshortchar.c diff --git a/native/libffi/testsuite/libffi.call/cls_multi_uchar.c b/native/libffi/testsuite/libffi.closures/cls_multi_uchar.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_multi_uchar.c rename to native/libffi/testsuite/libffi.closures/cls_multi_uchar.c diff --git a/native/libffi/testsuite/libffi.call/cls_multi_ushort.c b/native/libffi/testsuite/libffi.closures/cls_multi_ushort.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_multi_ushort.c rename to native/libffi/testsuite/libffi.closures/cls_multi_ushort.c diff --git a/native/libffi/testsuite/libffi.call/cls_multi_ushortchar.c b/native/libffi/testsuite/libffi.closures/cls_multi_ushortchar.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_multi_ushortchar.c rename to native/libffi/testsuite/libffi.closures/cls_multi_ushortchar.c diff --git a/native/libffi/testsuite/libffi.call/cls_pointer.c b/native/libffi/testsuite/libffi.closures/cls_pointer.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_pointer.c rename to native/libffi/testsuite/libffi.closures/cls_pointer.c diff --git a/native/libffi/testsuite/libffi.call/cls_pointer_stack.c b/native/libffi/testsuite/libffi.closures/cls_pointer_stack.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_pointer_stack.c rename to native/libffi/testsuite/libffi.closures/cls_pointer_stack.c diff --git a/native/libffi/testsuite/libffi.call/cls_schar.c b/native/libffi/testsuite/libffi.closures/cls_schar.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_schar.c rename to native/libffi/testsuite/libffi.closures/cls_schar.c diff --git a/native/libffi/testsuite/libffi.call/cls_sint.c b/native/libffi/testsuite/libffi.closures/cls_sint.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_sint.c rename to native/libffi/testsuite/libffi.closures/cls_sint.c diff --git a/native/libffi/testsuite/libffi.call/cls_sshort.c b/native/libffi/testsuite/libffi.closures/cls_sshort.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_sshort.c rename to native/libffi/testsuite/libffi.closures/cls_sshort.c diff --git a/native/libffi/testsuite/libffi.call/cls_struct_va1.c b/native/libffi/testsuite/libffi.closures/cls_struct_va1.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_struct_va1.c rename to native/libffi/testsuite/libffi.closures/cls_struct_va1.c diff --git a/native/libffi/testsuite/libffi.call/cls_uchar.c b/native/libffi/testsuite/libffi.closures/cls_uchar.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_uchar.c rename to native/libffi/testsuite/libffi.closures/cls_uchar.c diff --git a/native/libffi/testsuite/libffi.call/cls_uchar_va.c b/native/libffi/testsuite/libffi.closures/cls_uchar_va.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_uchar_va.c rename to native/libffi/testsuite/libffi.closures/cls_uchar_va.c diff --git a/native/libffi/testsuite/libffi.call/cls_uint.c b/native/libffi/testsuite/libffi.closures/cls_uint.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_uint.c rename to native/libffi/testsuite/libffi.closures/cls_uint.c diff --git a/native/libffi/testsuite/libffi.call/cls_uint_va.c b/native/libffi/testsuite/libffi.closures/cls_uint_va.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_uint_va.c rename to native/libffi/testsuite/libffi.closures/cls_uint_va.c diff --git a/native/libffi/testsuite/libffi.call/cls_ulong_va.c b/native/libffi/testsuite/libffi.closures/cls_ulong_va.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_ulong_va.c rename to native/libffi/testsuite/libffi.closures/cls_ulong_va.c diff --git a/native/libffi/testsuite/libffi.call/cls_ulonglong.c b/native/libffi/testsuite/libffi.closures/cls_ulonglong.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_ulonglong.c rename to native/libffi/testsuite/libffi.closures/cls_ulonglong.c diff --git a/native/libffi/testsuite/libffi.call/cls_ushort.c b/native/libffi/testsuite/libffi.closures/cls_ushort.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_ushort.c rename to native/libffi/testsuite/libffi.closures/cls_ushort.c diff --git a/native/libffi/testsuite/libffi.call/cls_ushort_va.c b/native/libffi/testsuite/libffi.closures/cls_ushort_va.c similarity index 100% rename from native/libffi/testsuite/libffi.call/cls_ushort_va.c rename to native/libffi/testsuite/libffi.closures/cls_ushort_va.c diff --git a/native/libffi/testsuite/libffi.call/err_bad_abi.c b/native/libffi/testsuite/libffi.closures/err_bad_abi.c similarity index 100% rename from native/libffi/testsuite/libffi.call/err_bad_abi.c rename to native/libffi/testsuite/libffi.closures/err_bad_abi.c diff --git a/native/libffi/testsuite/libffi.closures/ffitest.h b/native/libffi/testsuite/libffi.closures/ffitest.h new file mode 100644 index 0000000000..cfce1ad589 --- /dev/null +++ b/native/libffi/testsuite/libffi.closures/ffitest.h @@ -0,0 +1,138 @@ +#include +#include +#include +#include +#include +#include "fficonfig.h" + +#if defined HAVE_STDINT_H +#include +#endif + +#if defined HAVE_INTTYPES_H +#include +#endif + +#define MAX_ARGS 256 + +#define CHECK(x) (void)(!(x) ? (abort(), 1) : 0) + +/* Define macros so that compilers other than gcc can run the tests. */ +#undef __UNUSED__ +#if defined(__GNUC__) +#define __UNUSED__ __attribute__((__unused__)) +#define __STDCALL__ __attribute__((stdcall)) +#define __THISCALL__ __attribute__((thiscall)) +#define __FASTCALL__ __attribute__((fastcall)) +#define __MSABI__ __attribute__((ms_abi)) +#else +#define __UNUSED__ +#define __STDCALL__ __stdcall +#define __THISCALL__ __thiscall +#define __FASTCALL__ __fastcall +#endif + +#ifndef ABI_NUM +#define ABI_NUM FFI_DEFAULT_ABI +#define ABI_ATTR +#endif + +/* Prefer MAP_ANON(YMOUS) to /dev/zero, since we don't need to keep a + file open. */ +#ifdef HAVE_MMAP_ANON +# undef HAVE_MMAP_DEV_ZERO + +# include +# ifndef MAP_FAILED +# define MAP_FAILED -1 +# endif +# if !defined (MAP_ANONYMOUS) && defined (MAP_ANON) +# define MAP_ANONYMOUS MAP_ANON +# endif +# define USING_MMAP + +#endif + +#ifdef HAVE_MMAP_DEV_ZERO + +# include +# ifndef MAP_FAILED +# define MAP_FAILED -1 +# endif +# define USING_MMAP + +#endif + +/* MinGW kludge. */ +#if defined(_WIN64) | defined(_WIN32) +#define PRIdLL "I64d" +#define PRIuLL "I64u" +#else +#define PRIdLL "lld" +#define PRIuLL "llu" +#endif + +/* Tru64 UNIX kludge. */ +#if defined(__alpha__) && defined(__osf__) +/* Tru64 UNIX V4.0 doesn't support %lld/%lld, but long is 64-bit. */ +#undef PRIdLL +#define PRIdLL "ld" +#undef PRIuLL +#define PRIuLL "lu" +#define PRId8 "hd" +#define PRIu8 "hu" +#define PRId64 "ld" +#define PRIu64 "lu" +#define PRIuPTR "lu" +#endif + +/* PA HP-UX kludge. */ +#if defined(__hppa__) && defined(__hpux__) && !defined(PRIuPTR) +#define PRIuPTR "lu" +#endif + +/* IRIX kludge. */ +#if defined(__sgi) +/* IRIX 6.5 provides all definitions, but only for C99 + compilations. */ +#define PRId8 "hhd" +#define PRIu8 "hhu" +#if (_MIPS_SZLONG == 32) +#define PRId64 "lld" +#define PRIu64 "llu" +#endif +/* This doesn't match , which always has "lld" here, but the + arguments are uint64_t, int64_t, which are unsigned long, long for + 64-bit in . */ +#if (_MIPS_SZLONG == 64) +#define PRId64 "ld" +#define PRIu64 "lu" +#endif +/* This doesn't match , which has "u" here, but the arguments + are uintptr_t, which is always unsigned long. */ +#define PRIuPTR "lu" +#endif + +/* Solaris < 10 kludge. */ +#if defined(__sun__) && defined(__svr4__) && !defined(PRIuPTR) +#if defined(__arch64__) || defined (__x86_64__) +#define PRIuPTR "lu" +#else +#define PRIuPTR "u" +#endif +#endif + +/* MSVC kludge. */ +#if defined _MSC_VER +#if !defined(__cplusplus) || defined(__STDC_FORMAT_MACROS) +#define PRIuPTR "lu" +#define PRIu8 "u" +#define PRId8 "d" +#define PRIu64 "I64u" +#define PRId64 "I64d" +#endif +#endif + +#ifndef PRIuPTR +#define PRIuPTR "u" +#endif diff --git a/native/libffi/testsuite/libffi.call/huge_struct.c b/native/libffi/testsuite/libffi.closures/huge_struct.c similarity index 98% rename from native/libffi/testsuite/libffi.call/huge_struct.c rename to native/libffi/testsuite/libffi.closures/huge_struct.c index 657fe54148..e8e1d8601f 100644 --- a/native/libffi/testsuite/libffi.call/huge_struct.c +++ b/native/libffi/testsuite/libffi.closures/huge_struct.c @@ -5,11 +5,11 @@ Originator: Blake Chaffin 6/18/2007 */ -/* { dg-excess-errors "" { target x86_64-*-mingw* x86_64-*-cygwin* } } */ /* { dg-do run { xfail strongarm*-*-* xscale*-*-* } } */ /* { dg-options -mlong-double-128 { target powerpc64*-*-linux* } } */ -/* { dg-options -Wformat=0 { target moxie*-*-elf } } */ -/* { dg-output "" { xfail x86_64-*-mingw* x86_64-*-cygwin* } } */ +/* { dg-options -Wformat=0 { target moxie*-*-elf or1k-*-* } } */ + +#include #include "ffitest.h" diff --git a/native/libffi/testsuite/libffi.call/nested_struct.c b/native/libffi/testsuite/libffi.closures/nested_struct.c similarity index 100% rename from native/libffi/testsuite/libffi.call/nested_struct.c rename to native/libffi/testsuite/libffi.closures/nested_struct.c diff --git a/native/libffi/testsuite/libffi.call/nested_struct1.c b/native/libffi/testsuite/libffi.closures/nested_struct1.c similarity index 100% rename from native/libffi/testsuite/libffi.call/nested_struct1.c rename to native/libffi/testsuite/libffi.closures/nested_struct1.c diff --git a/native/libffi/testsuite/libffi.call/nested_struct10.c b/native/libffi/testsuite/libffi.closures/nested_struct10.c similarity index 99% rename from native/libffi/testsuite/libffi.call/nested_struct10.c rename to native/libffi/testsuite/libffi.closures/nested_struct10.c index 34a74e7187..3cf2b44abe 100644 --- a/native/libffi/testsuite/libffi.call/nested_struct10.c +++ b/native/libffi/testsuite/libffi.closures/nested_struct10.c @@ -32,6 +32,7 @@ static B B_fn(struct A b2, struct B b3, struct C b4) result.x.a = b2.a + b3.x.a + b3.z + b4.d; result.x.b = b2.b + b3.x.b + b3.y + b4.e; result.y = b2.b + b3.x.b + b4.e; + result.z = 0; printf("%d %d %d %d %d %d %d %d: %d %d %d\n", (int)b2.a, b2.b, (int)b3.x.a, b3.x.b, b3.y, b3.z, (int)b4.d, b4.e, diff --git a/native/libffi/testsuite/libffi.call/nested_struct11.c b/native/libffi/testsuite/libffi.closures/nested_struct11.c similarity index 100% rename from native/libffi/testsuite/libffi.call/nested_struct11.c rename to native/libffi/testsuite/libffi.closures/nested_struct11.c diff --git a/native/libffi/testsuite/libffi.call/nested_struct2.c b/native/libffi/testsuite/libffi.closures/nested_struct2.c similarity index 100% rename from native/libffi/testsuite/libffi.call/nested_struct2.c rename to native/libffi/testsuite/libffi.closures/nested_struct2.c diff --git a/native/libffi/testsuite/libffi.call/nested_struct3.c b/native/libffi/testsuite/libffi.closures/nested_struct3.c similarity index 100% rename from native/libffi/testsuite/libffi.call/nested_struct3.c rename to native/libffi/testsuite/libffi.closures/nested_struct3.c diff --git a/native/libffi/testsuite/libffi.call/nested_struct4.c b/native/libffi/testsuite/libffi.closures/nested_struct4.c similarity index 100% rename from native/libffi/testsuite/libffi.call/nested_struct4.c rename to native/libffi/testsuite/libffi.closures/nested_struct4.c diff --git a/native/libffi/testsuite/libffi.call/nested_struct5.c b/native/libffi/testsuite/libffi.closures/nested_struct5.c similarity index 100% rename from native/libffi/testsuite/libffi.call/nested_struct5.c rename to native/libffi/testsuite/libffi.closures/nested_struct5.c diff --git a/native/libffi/testsuite/libffi.call/nested_struct6.c b/native/libffi/testsuite/libffi.closures/nested_struct6.c similarity index 100% rename from native/libffi/testsuite/libffi.call/nested_struct6.c rename to native/libffi/testsuite/libffi.closures/nested_struct6.c diff --git a/native/libffi/testsuite/libffi.call/nested_struct7.c b/native/libffi/testsuite/libffi.closures/nested_struct7.c similarity index 100% rename from native/libffi/testsuite/libffi.call/nested_struct7.c rename to native/libffi/testsuite/libffi.closures/nested_struct7.c diff --git a/native/libffi/testsuite/libffi.call/nested_struct8.c b/native/libffi/testsuite/libffi.closures/nested_struct8.c similarity index 100% rename from native/libffi/testsuite/libffi.call/nested_struct8.c rename to native/libffi/testsuite/libffi.closures/nested_struct8.c diff --git a/native/libffi/testsuite/libffi.call/nested_struct9.c b/native/libffi/testsuite/libffi.closures/nested_struct9.c similarity index 100% rename from native/libffi/testsuite/libffi.call/nested_struct9.c rename to native/libffi/testsuite/libffi.closures/nested_struct9.c diff --git a/native/libffi/testsuite/libffi.call/problem1.c b/native/libffi/testsuite/libffi.closures/problem1.c similarity index 100% rename from native/libffi/testsuite/libffi.call/problem1.c rename to native/libffi/testsuite/libffi.closures/problem1.c diff --git a/native/libffi/testsuite/libffi.call/stret_large.c b/native/libffi/testsuite/libffi.closures/stret_large.c similarity index 100% rename from native/libffi/testsuite/libffi.call/stret_large.c rename to native/libffi/testsuite/libffi.closures/stret_large.c diff --git a/native/libffi/testsuite/libffi.call/stret_large2.c b/native/libffi/testsuite/libffi.closures/stret_large2.c similarity index 100% rename from native/libffi/testsuite/libffi.call/stret_large2.c rename to native/libffi/testsuite/libffi.closures/stret_large2.c diff --git a/native/libffi/testsuite/libffi.call/stret_medium.c b/native/libffi/testsuite/libffi.closures/stret_medium.c similarity index 100% rename from native/libffi/testsuite/libffi.call/stret_medium.c rename to native/libffi/testsuite/libffi.closures/stret_medium.c diff --git a/native/libffi/testsuite/libffi.call/stret_medium2.c b/native/libffi/testsuite/libffi.closures/stret_medium2.c similarity index 100% rename from native/libffi/testsuite/libffi.call/stret_medium2.c rename to native/libffi/testsuite/libffi.closures/stret_medium2.c diff --git a/native/libffi/testsuite/libffi.call/testclosure.c b/native/libffi/testsuite/libffi.closures/testclosure.c similarity index 100% rename from native/libffi/testsuite/libffi.call/testclosure.c rename to native/libffi/testsuite/libffi.closures/testclosure.c diff --git a/native/libffi/testsuite/libffi.closures/unwindtest.cc b/native/libffi/testsuite/libffi.closures/unwindtest.cc new file mode 100644 index 0000000000..e1145657d1 --- /dev/null +++ b/native/libffi/testsuite/libffi.closures/unwindtest.cc @@ -0,0 +1,117 @@ +/* Area: ffi_closure, unwind info + Purpose: Check if the unwind information is passed correctly. + Limitations: none. + PR: none. + Originator: Jeff Sturm */ + +/* { dg-do run { xfail x86_64-apple-darwin* moxie*-*-* } } */ + +#include "ffitest.h" + +void ABI_ATTR +closure_test_fn(ffi_cif* cif __UNUSED__, void* resp __UNUSED__, + void** args __UNUSED__, void* userdata __UNUSED__) +{ + throw 9; +} + +typedef void (*closure_test_type)(); + +void closure_test_fn1(ffi_cif* cif __UNUSED__, void* resp, + void** args, void* userdata __UNUSED__) + { + *(ffi_arg*)resp = + (int)*(float *)args[0] +(int)(*(float *)args[1]) + + (int)(*(float *)args[2]) + (int)*(float *)args[3] + + (int)(*(signed short *)args[4]) + (int)(*(float *)args[5]) + + (int)*(float *)args[6] + (int)(*(int *)args[7]) + + (int)(*(double*)args[8]) + (int)*(int *)args[9] + + (int)(*(int *)args[10]) + (int)(*(float *)args[11]) + + (int)*(int *)args[12] + (int)(*(int *)args[13]) + + (int)(*(int *)args[14]) + *(int *)args[15] + (int)(intptr_t)userdata; + + printf("%d %d %d %d %d %d %d %d %d %d %d %d %d %d %d %d %d: %d\n", + (int)*(float *)args[0], (int)(*(float *)args[1]), + (int)(*(float *)args[2]), (int)*(float *)args[3], + (int)(*(signed short *)args[4]), (int)(*(float *)args[5]), + (int)*(float *)args[6], (int)(*(int *)args[7]), + (int)(*(double *)args[8]), (int)*(int *)args[9], + (int)(*(int *)args[10]), (int)(*(float *)args[11]), + (int)*(int *)args[12], (int)(*(int *)args[13]), + (int)(*(int *)args[14]), *(int *)args[15], + (int)(intptr_t)userdata, (int)*(ffi_arg*)resp); + + throw (int)*(ffi_arg*)resp; +} + +typedef int (*closure_test_type1)(float, float, float, float, signed short, + float, float, int, double, int, int, float, + int, int, int, int); + +int main (void) +{ + ffi_cif cif; + void *code; + ffi_closure *pcl = (ffi_closure *)ffi_closure_alloc(sizeof(ffi_closure), &code); + ffi_type * cl_arg_types[17]; + + { + cl_arg_types[1] = NULL; + + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 0, + &ffi_type_void, cl_arg_types) == FFI_OK); + CHECK(ffi_prep_closure_loc(pcl, &cif, closure_test_fn, NULL, code) == FFI_OK); + + try + { + (*((closure_test_type)(code)))(); + } catch (int exception_code) + { + CHECK(exception_code == 9); + } + + printf("part one OK\n"); + /* { dg-output "part one OK" } */ + } + + { + + cl_arg_types[0] = &ffi_type_float; + cl_arg_types[1] = &ffi_type_float; + cl_arg_types[2] = &ffi_type_float; + cl_arg_types[3] = &ffi_type_float; + cl_arg_types[4] = &ffi_type_sshort; + cl_arg_types[5] = &ffi_type_float; + cl_arg_types[6] = &ffi_type_float; + cl_arg_types[7] = &ffi_type_uint; + cl_arg_types[8] = &ffi_type_double; + cl_arg_types[9] = &ffi_type_uint; + cl_arg_types[10] = &ffi_type_uint; + cl_arg_types[11] = &ffi_type_float; + cl_arg_types[12] = &ffi_type_uint; + cl_arg_types[13] = &ffi_type_uint; + cl_arg_types[14] = &ffi_type_uint; + cl_arg_types[15] = &ffi_type_uint; + cl_arg_types[16] = NULL; + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 16, + &ffi_type_sint, cl_arg_types) == FFI_OK); + + CHECK(ffi_prep_closure_loc(pcl, &cif, closure_test_fn1, + (void *) 3 /* userdata */, code) == FFI_OK); + try + { + (*((closure_test_type1)code)) + (1.1, 2.2, 3.3, 4.4, 127, 5.5, 6.6, 8, 9, 10, 11, 12.0, 13, + 19, 21, 1); + /* { dg-output "\n1 2 3 4 127 5 6 8 9 10 11 12 13 19 21 1 3: 255" } */ + } catch (int exception_code) + { + CHECK(exception_code == 255); + } + printf("part two OK\n"); + /* { dg-output "\npart two OK" } */ + } + exit(0); +} diff --git a/native/libffi/testsuite/libffi.closures/unwindtest_ffi_call.cc b/native/libffi/testsuite/libffi.closures/unwindtest_ffi_call.cc new file mode 100644 index 0000000000..153d240940 --- /dev/null +++ b/native/libffi/testsuite/libffi.closures/unwindtest_ffi_call.cc @@ -0,0 +1,54 @@ +/* Area: ffi_call, unwind info + Purpose: Check if the unwind information is passed correctly. + Limitations: none. + PR: none. + Originator: Andreas Tobler 20061213 */ + +/* { dg-do run { xfail moxie*-*-* } } */ + +#include "ffitest.h" + +static int checking(int a __UNUSED__, short b __UNUSED__, + signed char c __UNUSED__) +{ + throw 9; +} + +int main (void) +{ + ffi_cif cif; + ffi_type *args[MAX_ARGS]; + void *values[MAX_ARGS]; + ffi_arg rint; + + signed int si; + signed short ss; + signed char sc; + + args[0] = &ffi_type_sint; + values[0] = &si; + args[1] = &ffi_type_sshort; + values[1] = &ss; + args[2] = &ffi_type_schar; + values[2] = ≻ + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 3, + &ffi_type_sint, args) == FFI_OK); + + si = -6; + ss = -12; + sc = -1; + { + try + { + ffi_call(&cif, FFI_FN(checking), &rint, values); + } catch (int exception_code) + { + CHECK(exception_code == 9); + } + printf("part one OK\n"); + /* { dg-output "part one OK" } */ + } + exit(0); +} diff --git a/native/libffi/testsuite/libffi.complex/cls_align_complex.inc b/native/libffi/testsuite/libffi.complex/cls_align_complex.inc new file mode 100644 index 0000000000..4a812edf05 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_align_complex.inc @@ -0,0 +1,91 @@ +/* -*-c-*- */ +#include "ffitest.h" +#include + +typedef struct cls_struct_align { + unsigned char a; + _Complex T_C_TYPE b; + unsigned char c; +} cls_struct_align; + +cls_struct_align cls_struct_align_fn( + struct cls_struct_align a1, struct cls_struct_align a2) +{ + struct cls_struct_align result; + + result.a = a1.a + a2.a; + result.b = a1.b + a2.b; + result.c = a1.c + a2.c; + + printf("%d %f,%fi %d %d %f,%fi %d: %d %f,%fi %d\n", + a1.a, T_CONV creal (a1.b), T_CONV cimag (a1.b), a1.c, + a2.a, T_CONV creal (a2.b), T_CONV cimag (a2.b), a2.c, + result.a, T_CONV creal (result.b), T_CONV cimag (result.b), result.c); + + return result; +} + +static void +cls_struct_align_gn(ffi_cif* cif __UNUSED__, void* resp, void** args, + void* userdata __UNUSED__) +{ + + struct cls_struct_align a1, a2; + + a1 = *(struct cls_struct_align*)(args[0]); + a2 = *(struct cls_struct_align*)(args[1]); + + *(cls_struct_align*)resp = cls_struct_align_fn(a1, a2); +} + +int main (void) +{ + ffi_cif cif; + void *code; + ffi_closure *pcl = ffi_closure_alloc(sizeof(ffi_closure), &code); + void* args_c[5]; + ffi_type* cls_struct_fields[4]; + ffi_type cls_struct_type; + ffi_type* c_arg_types[5]; + + struct cls_struct_align g_c = { 12, 4951 + 7 * I, 127 }; + struct cls_struct_align f_c = { 1, 9320 + 1 * I, 13 }; + struct cls_struct_align res_c; + + cls_struct_type.size = 0; + cls_struct_type.alignment = 0; + cls_struct_type.type = FFI_TYPE_STRUCT; + cls_struct_type.elements = cls_struct_fields; + + cls_struct_fields[0] = &ffi_type_uchar; + cls_struct_fields[1] = &T_FFI_TYPE; + cls_struct_fields[2] = &ffi_type_uchar; + cls_struct_fields[3] = NULL; + + c_arg_types[0] = &cls_struct_type; + c_arg_types[1] = &cls_struct_type; + c_arg_types[2] = NULL; + + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 2, &cls_struct_type, + c_arg_types) == FFI_OK); + + args_c[0] = &g_c; + args_c[1] = &f_c; + args_c[2] = NULL; + + ffi_call(&cif, FFI_FN(cls_struct_align_fn), &res_c, args_c); + /* { dg-output "12 4951,7i 127 1 9320,1i 13: 13 14271,8i 140" } */ + printf("res: %d %f,%fi %d\n", + res_c.a, T_CONV creal (res_c.b), T_CONV cimag (res_c.b), res_c.c); + /* { dg-output "\nres: 13 14271,8i 140" } */ + + CHECK(ffi_prep_closure_loc(pcl, &cif, cls_struct_align_gn, NULL, code) == FFI_OK); + + res_c = ((cls_struct_align(*)(cls_struct_align, cls_struct_align))(code))(g_c, f_c); + /* { dg-output "\n12 4951,7i 127 1 9320,1i 13: 13 14271,8i 140" } */ + printf("res: %d %f,%fi %d\n", + res_c.a, T_CONV creal (res_c.b), T_CONV cimag (res_c.b), res_c.c); + /* { dg-output "\nres: 13 14271,8i 140" } */ + + exit(0); +} diff --git a/native/libffi/testsuite/libffi.complex/cls_align_complex_double.c b/native/libffi/testsuite/libffi.complex/cls_align_complex_double.c new file mode 100644 index 0000000000..0dff23ae47 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_align_complex_double.c @@ -0,0 +1,10 @@ +/* Area: ffi_call, closure_call + Purpose: Check structure alignment of complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_double.inc" +#include "cls_align_complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/cls_align_complex_float.c b/native/libffi/testsuite/libffi.complex/cls_align_complex_float.c new file mode 100644 index 0000000000..0affbd07e5 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_align_complex_float.c @@ -0,0 +1,10 @@ +/* Area: ffi_call, closure_call + Purpose: Check structure alignment of complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_float.inc" +#include "cls_align_complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/cls_align_complex_longdouble.c b/native/libffi/testsuite/libffi.complex/cls_align_complex_longdouble.c new file mode 100644 index 0000000000..7889ba859b --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_align_complex_longdouble.c @@ -0,0 +1,10 @@ +/* Area: ffi_call, closure_call + Purpose: Check structure alignment of complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_longdouble.inc" +#include "cls_align_complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/cls_complex.inc b/native/libffi/testsuite/libffi.complex/cls_complex.inc new file mode 100644 index 0000000000..f9374044e8 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex.inc @@ -0,0 +1,42 @@ +/* -*-c-*- */ +#include "ffitest.h" +#include + +static void cls_ret_complex_fn(ffi_cif* cif __UNUSED__, void* resp, void** args, + void* userdata __UNUSED__) + { + _Complex T_C_TYPE *pa; + _Complex T_C_TYPE *pr; + pa = (_Complex T_C_TYPE *)args[0]; + pr = (_Complex T_C_TYPE *)resp; + *pr = *pa; + + printf("%.6f,%.6fi: %.6f,%.6fi\n", + T_CONV creal (*pa), T_CONV cimag (*pa), + T_CONV creal (*pr), T_CONV cimag (*pr)); + } +typedef _Complex T_C_TYPE (*cls_ret_complex)(_Complex T_C_TYPE); + +int main (void) +{ + ffi_cif cif; + void *code; + ffi_closure *pcl = ffi_closure_alloc(sizeof(ffi_closure), &code); + ffi_type * cl_arg_types[2]; + _Complex T_C_TYPE res; + + cl_arg_types[0] = &T_FFI_TYPE; + cl_arg_types[1] = NULL; + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1, + &T_FFI_TYPE, cl_arg_types) == FFI_OK); + + CHECK(ffi_prep_closure_loc(pcl, &cif, cls_ret_complex_fn, NULL, code) == FFI_OK); + + res = (*((cls_ret_complex)code))(0.125 + 128.0 * I); + printf("res: %.6f,%.6fi\n", T_CONV creal (res), T_CONV cimag (res)); + CHECK (res == (0.125 + 128.0 * I)); + + exit(0); +} diff --git a/native/libffi/testsuite/libffi.complex/cls_complex_double.c b/native/libffi/testsuite/libffi.complex/cls_complex_double.c new file mode 100644 index 0000000000..05e35340e0 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex_double.c @@ -0,0 +1,10 @@ +/* Area: closure_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_double.inc" +#include "cls_complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/cls_complex_float.c b/native/libffi/testsuite/libffi.complex/cls_complex_float.c new file mode 100644 index 0000000000..5df7849d4c --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex_float.c @@ -0,0 +1,10 @@ +/* Area: closure_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_float.inc" +#include "cls_complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/cls_complex_longdouble.c b/native/libffi/testsuite/libffi.complex/cls_complex_longdouble.c new file mode 100644 index 0000000000..2b1c320a20 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex_longdouble.c @@ -0,0 +1,10 @@ +/* Area: closure_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_longdouble.inc" +#include "cls_complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/cls_complex_struct.inc b/native/libffi/testsuite/libffi.complex/cls_complex_struct.inc new file mode 100644 index 0000000000..df8708d1c4 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex_struct.inc @@ -0,0 +1,71 @@ +/* -*-c-*- */ +#include "ffitest.h" +#include + +typedef struct Cs { + _Complex T_C_TYPE x; + _Complex T_C_TYPE y; +} Cs; + +Cs gc; + +void +closure_test_fn(Cs p) +{ + printf("%.1f,%.1fi %.1f,%.1fi\n", + T_CONV creal (p.x), T_CONV cimag (p.x), + T_CONV creal (p.y), T_CONV cimag (p.y)); + gc = p; +} + +void +closure_test_gn(ffi_cif* cif __UNUSED__, void* resp __UNUSED__, + void** args, void* userdata __UNUSED__) +{ + closure_test_fn(*(Cs*)args[0]); +} + +int main(int argc __UNUSED__, char** argv __UNUSED__) +{ + ffi_cif cif; + + void *code; + ffi_closure *pcl = ffi_closure_alloc(sizeof(ffi_closure), &code); + ffi_type *cl_arg_types[1]; + + ffi_type ts1_type; + ffi_type* ts1_type_elements[4]; + + Cs arg = { 1.0 + 11.0 * I, 2.0 + 22.0 * I}; + + ts1_type.size = 0; + ts1_type.alignment = 0; + ts1_type.type = FFI_TYPE_STRUCT; + ts1_type.elements = ts1_type_elements; + + ts1_type_elements[0] = &T_FFI_TYPE; + ts1_type_elements[1] = &T_FFI_TYPE; + ts1_type_elements[2] = NULL; + + cl_arg_types[0] = &ts1_type; + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1, + &ffi_type_void, cl_arg_types) == FFI_OK); + + CHECK(ffi_prep_closure_loc(pcl, &cif, closure_test_gn, NULL, code) == FFI_OK); + + gc.x = 0.0 + 0.0 * I; + gc.y = 0.0 + 0.0 * I; + ((void*(*)(Cs))(code))(arg); + /* { dg-output "1.0,11.0i 2.0,22.0i\n" } */ + CHECK (gc.x == arg.x && gc.y == arg.y); + + gc.x = 0.0 + 0.0 * I; + gc.y = 0.0 + 0.0 * I; + closure_test_fn(arg); + /* { dg-output "1.0,11.0i 2.0,22.0i\n" } */ + CHECK (gc.x == arg.x && gc.y == arg.y); + + return 0; +} diff --git a/native/libffi/testsuite/libffi.complex/cls_complex_struct_double.c b/native/libffi/testsuite/libffi.complex/cls_complex_struct_double.c new file mode 100644 index 0000000000..ec71346a3f --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex_struct_double.c @@ -0,0 +1,10 @@ +/* Area: ffi_call, closure_call + Purpose: Check complex arguments in structs. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_double.inc" +#include "cls_complex_struct.inc" diff --git a/native/libffi/testsuite/libffi.complex/cls_complex_struct_float.c b/native/libffi/testsuite/libffi.complex/cls_complex_struct_float.c new file mode 100644 index 0000000000..96fdf75043 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex_struct_float.c @@ -0,0 +1,10 @@ +/* Area: ffi_call, closure_call + Purpose: Check complex arguments in structs. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_float.inc" +#include "cls_complex_struct.inc" diff --git a/native/libffi/testsuite/libffi.complex/cls_complex_struct_longdouble.c b/native/libffi/testsuite/libffi.complex/cls_complex_struct_longdouble.c new file mode 100644 index 0000000000..005b467398 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex_struct_longdouble.c @@ -0,0 +1,10 @@ +/* Area: ffi_call, closure_call + Purpose: Check complex arguments in structs. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_longdouble.inc" +#include "cls_complex_struct.inc" diff --git a/native/libffi/testsuite/libffi.complex/cls_complex_va.inc b/native/libffi/testsuite/libffi.complex/cls_complex_va.inc new file mode 100644 index 0000000000..8a3e15f0f6 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex_va.inc @@ -0,0 +1,80 @@ +/* -*-c-*- */ +#include "ffitest.h" +#include +#include +#include +#include + +static _Complex T_C_TYPE gComplexValue1 = 1 + 2 * I; +static _Complex T_C_TYPE gComplexValue2 = 3 + 4 * I; + +static int cls_variadic(const char *format, ...) +{ + va_list ap; + _Complex T_C_TYPE p1, p2; + + va_start (ap, format); + p1 = va_arg (ap, _Complex T_C_TYPE); + p2 = va_arg (ap, _Complex T_C_TYPE); + va_end (ap); + + return printf(format, T_CONV creal (p1), T_CONV cimag (p1), + T_CONV creal (p2), T_CONV cimag (p2)); +} + +static void +cls_complex_va_fn(ffi_cif* cif __UNUSED__, void* resp, + void** args, void* userdata __UNUSED__) +{ + char* format = *(char**)args[0]; + gComplexValue1 = *(_Complex T_C_TYPE*)args[1]; + gComplexValue2 = *(_Complex T_C_TYPE*)args[2]; + + *(ffi_arg*)resp = + printf(format, + T_CONV creal (gComplexValue1), T_CONV cimag (gComplexValue1), + T_CONV creal (gComplexValue2), T_CONV cimag (gComplexValue2)); +} + +int main (void) +{ + ffi_cif cif; + void *code; + ffi_closure *pcl = ffi_closure_alloc(sizeof(ffi_closure), &code); + void* args[4]; + ffi_type* arg_types[4]; + char *format = "%.1f,%.1fi %.1f,%.1fi\n"; + + _Complex T_C_TYPE complexArg1 = 1.0 + 22.0 *I; + _Complex T_C_TYPE complexArg2 = 333.0 + 4444.0 *I; + ffi_arg res = 0; + + arg_types[0] = &ffi_type_pointer; + arg_types[1] = &T_FFI_TYPE; + arg_types[2] = &T_FFI_TYPE; + arg_types[3] = NULL; + + /* This printf call is variadic */ + CHECK(ffi_prep_cif_var(&cif, FFI_DEFAULT_ABI, 1, 3, &ffi_type_sint, + arg_types) == FFI_OK); + + args[0] = &format; + args[1] = &complexArg1; + args[2] = &complexArg2; + args[3] = NULL; + + ffi_call(&cif, FFI_FN(cls_variadic), &res, args); + printf("res: %d\n", (int) res); + CHECK (res == 24); + + CHECK(ffi_prep_closure_loc(pcl, &cif, cls_complex_va_fn, NULL, code) + == FFI_OK); + + res = ((int(*)(char *, ...))(code))(format, complexArg1, complexArg2); + CHECK (gComplexValue1 == complexArg1); + CHECK (gComplexValue2 == complexArg2); + printf("res: %d\n", (int) res); + CHECK (res == 24); + + exit(0); +} diff --git a/native/libffi/testsuite/libffi.complex/cls_complex_va_double.c b/native/libffi/testsuite/libffi.complex/cls_complex_va_double.c new file mode 100644 index 0000000000..879ccf3b81 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex_va_double.c @@ -0,0 +1,10 @@ +/* Area: ffi_call, closure_call + Purpose: Test complex' passed in variable argument lists. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_double.inc" +#include "cls_complex_va.inc" diff --git a/native/libffi/testsuite/libffi.complex/cls_complex_va_float.c b/native/libffi/testsuite/libffi.complex/cls_complex_va_float.c new file mode 100644 index 0000000000..2b17826045 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex_va_float.c @@ -0,0 +1,16 @@ +/* Area: ffi_call, closure_call + Purpose: Test complex' passed in variable argument lists. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +/* Alpha splits _Complex into two arguments. It's illegal to pass + float through varargs, so _Complex float goes badly. In sort of + gets passed as _Complex double, but the compiler doesn't agree + with itself on this issue. */ +/* { dg-do run { xfail alpha*-*-* } } */ + +#include "complex_defs_float.inc" +#include "cls_complex_va.inc" diff --git a/native/libffi/testsuite/libffi.complex/cls_complex_va_longdouble.c b/native/libffi/testsuite/libffi.complex/cls_complex_va_longdouble.c new file mode 100644 index 0000000000..6eca9656ea --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/cls_complex_va_longdouble.c @@ -0,0 +1,10 @@ +/* Area: ffi_call, closure_call + Purpose: Test complex' passed in variable argument lists. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_longdouble.inc" +#include "cls_complex_va.inc" diff --git a/native/libffi/testsuite/libffi.complex/complex.exp b/native/libffi/testsuite/libffi.complex/complex.exp new file mode 100644 index 0000000000..4631db285a --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/complex.exp @@ -0,0 +1,36 @@ +# Copyright (C) 2003, 2006, 2009, 2010, 2014 Free Software Foundation, Inc. + +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; see the file COPYING3. If not see +# . + +dg-init +libffi-init + +global srcdir subdir + +set tlist [lsort [glob -nocomplain -- $srcdir/$subdir/*.{c,cc}]] + +if { [libffi_feature_test "#ifdef FFI_TARGET_HAS_COMPLEX_TYPE"] } { + run-many-tests $tlist "" +} else { + foreach test $tlist { + unsupported "$test" + } +} + +dg-finish + +# Local Variables: +# tcl-indent-level:4 +# End: diff --git a/native/libffi/testsuite/libffi.complex/complex.inc b/native/libffi/testsuite/libffi.complex/complex.inc new file mode 100644 index 0000000000..515ae3e60d --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/complex.inc @@ -0,0 +1,51 @@ +/* -*-c-*-*/ +#include "ffitest.h" +#include + +static _Complex T_C_TYPE f_complex(_Complex T_C_TYPE c, int x, int *py) +{ + c = -(2 * creal (c)) + (cimag (c) + 1)* I; + *py += x; + + return c; +} + +int main (void) +{ + ffi_cif cif; + ffi_type *args[MAX_ARGS]; + void *values[MAX_ARGS]; + + _Complex T_C_TYPE tc_arg; + _Complex T_C_TYPE tc_result; + int tc_int_arg_x; + int tc_y; + int *tc_ptr_arg_y = &tc_y; + + args[0] = &T_FFI_TYPE; + args[1] = &ffi_type_sint; + args[2] = &ffi_type_pointer; + values[0] = &tc_arg; + values[1] = &tc_int_arg_x; + values[2] = &tc_ptr_arg_y; + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 3, + &T_FFI_TYPE, args) == FFI_OK); + + tc_arg = 1 + 7 * I; + tc_int_arg_x = 1234; + tc_y = 9876; + ffi_call(&cif, FFI_FN(f_complex), &tc_result, values); + + printf ("%f,%fi %f,%fi, x %d 1234, y %d 11110\n", + T_CONV creal (tc_result), T_CONV cimag (tc_result), + T_CONV creal (2.0), T_CONV creal (8.0), tc_int_arg_x, tc_y); + + CHECK (creal (tc_result) == -2); + CHECK (cimag (tc_result) == 8); + CHECK (tc_int_arg_x == 1234); + CHECK (*tc_ptr_arg_y == 11110); + + exit(0); +} diff --git a/native/libffi/testsuite/libffi.complex/complex_defs_double.inc b/native/libffi/testsuite/libffi.complex/complex_defs_double.inc new file mode 100644 index 0000000000..3583e166d6 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/complex_defs_double.inc @@ -0,0 +1,7 @@ +/* -*-c-*- */ +/* Complex base type. */ +#define T_FFI_TYPE ffi_type_complex_double +/* C type corresponding to the base type. */ +#define T_C_TYPE double +/* C cast for a value of type T_C_TYPE that is passed to printf. */ +#define T_CONV diff --git a/native/libffi/testsuite/libffi.complex/complex_defs_float.inc b/native/libffi/testsuite/libffi.complex/complex_defs_float.inc new file mode 100644 index 0000000000..bbd9375cbf --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/complex_defs_float.inc @@ -0,0 +1,7 @@ +/* -*-c-*- */ +/* Complex base type. */ +#define T_FFI_TYPE ffi_type_complex_float +/* C type corresponding to the base type. */ +#define T_C_TYPE float +/* C cast for a value of type T_C_TYPE that is passed to printf. */ +#define T_CONV (double) diff --git a/native/libffi/testsuite/libffi.complex/complex_defs_longdouble.inc b/native/libffi/testsuite/libffi.complex/complex_defs_longdouble.inc new file mode 100644 index 0000000000..14b9f243f4 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/complex_defs_longdouble.inc @@ -0,0 +1,7 @@ +/* -*-c-*- */ +/* Complex base type. */ +#define T_FFI_TYPE ffi_type_complex_longdouble +/* C type corresponding to the base type. */ +#define T_C_TYPE long double +/* C cast for a value of type T_C_TYPE that is passed to printf. */ +#define T_CONV diff --git a/native/libffi/testsuite/libffi.complex/complex_double.c b/native/libffi/testsuite/libffi.complex/complex_double.c new file mode 100644 index 0000000000..8a3297b2a2 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/complex_double.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check complex types. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_double.inc" +#include "complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/complex_float.c b/native/libffi/testsuite/libffi.complex/complex_float.c new file mode 100644 index 0000000000..5044ebbcaa --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/complex_float.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check complex types. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_float.inc" +#include "complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/complex_int.c b/native/libffi/testsuite/libffi.complex/complex_int.c new file mode 100644 index 0000000000..bac319081e --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/complex_int.c @@ -0,0 +1,86 @@ +/* Area: ffi_call + Purpose: Check non-standard complex types. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "ffitest.h" +#include "ffi.h" +#include + +_Complex int f_complex(_Complex int c, int x, int *py) +{ + __real__ c = -2 * __real__ c; + __imag__ c = __imag__ c + 1; + *py += x; + return c; +} + +/* + * This macro can be used to define new complex type descriptors + * in a platform independent way. + * + * name: Name of the new descriptor is ffi_type_complex_. + * type: The C base type of the complex type. + */ +#define FFI_COMPLEX_TYPEDEF(name, type, ffitype) \ + static ffi_type *ffi_elements_complex_##name [2] = { \ + (ffi_type *)(&ffitype), NULL \ + }; \ + struct struct_align_complex_##name { \ + char c; \ + _Complex type x; \ + }; \ + ffi_type ffi_type_complex_##name = { \ + sizeof(_Complex type), \ + offsetof(struct struct_align_complex_##name, x), \ + FFI_TYPE_COMPLEX, \ + (ffi_type **)ffi_elements_complex_##name \ + } + +/* Define new complex type descriptors using the macro: */ +/* ffi_type_complex_sint */ +FFI_COMPLEX_TYPEDEF(sint, int, ffi_type_sint); +/* ffi_type_complex_uchar */ +FFI_COMPLEX_TYPEDEF(uchar, unsigned char, ffi_type_uint8); + +int main (void) +{ + ffi_cif cif; + ffi_type *args[MAX_ARGS]; + void *values[MAX_ARGS]; + + _Complex int tc_arg; + _Complex int tc_result; + int tc_int_arg_x; + int tc_y; + int *tc_ptr_arg_y = &tc_y; + + args[0] = &ffi_type_complex_sint; + args[1] = &ffi_type_sint; + args[2] = &ffi_type_pointer; + values[0] = &tc_arg; + values[1] = &tc_int_arg_x; + values[2] = &tc_ptr_arg_y; + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 3, &ffi_type_complex_sint, args) + == FFI_OK); + + tc_arg = 1 + 7 * I; + tc_int_arg_x = 1234; + tc_y = 9876; + ffi_call(&cif, FFI_FN(f_complex), &tc_result, values); + + printf ("%d,%di %d,%di, x %d 1234, y %d 11110\n", + (int)tc_result, (int)(tc_result * -I), 2, 8, tc_int_arg_x, tc_y); + /* dg-output "-2,8i 2,8i, x 1234 1234, y 11110 11110" */ + CHECK (creal (tc_result) == -2); + CHECK (cimag (tc_result) == 8); + CHECK (tc_int_arg_x == 1234); + CHECK (*tc_ptr_arg_y == 11110); + + exit(0); +} diff --git a/native/libffi/testsuite/libffi.complex/complex_longdouble.c b/native/libffi/testsuite/libffi.complex/complex_longdouble.c new file mode 100644 index 0000000000..7e78366093 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/complex_longdouble.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check complex types. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_longdouble.inc" +#include "complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/ffitest.h b/native/libffi/testsuite/libffi.complex/ffitest.h new file mode 100644 index 0000000000..d27d362d6a --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/ffitest.h @@ -0,0 +1 @@ +#include "../libffi.call/ffitest.h" diff --git a/native/libffi/testsuite/libffi.complex/many_complex.inc b/native/libffi/testsuite/libffi.complex/many_complex.inc new file mode 100644 index 0000000000..e37a774397 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/many_complex.inc @@ -0,0 +1,78 @@ +/* -*-c-*- */ +#include "ffitest.h" + +#include +#include + +static _Complex T_C_TYPE many(_Complex T_C_TYPE c1, + _Complex T_C_TYPE c2, + _Complex T_C_TYPE c3, + _Complex T_C_TYPE c4, + _Complex T_C_TYPE c5, + _Complex T_C_TYPE c6, + _Complex T_C_TYPE c7, + _Complex T_C_TYPE c8, + _Complex T_C_TYPE c9, + _Complex T_C_TYPE c10, + _Complex T_C_TYPE c11, + _Complex T_C_TYPE c12, + _Complex T_C_TYPE c13) +{ + printf("0 :%f,%fi\n" + "1 :%f,%fi\n" + "2 :%f,%fi\n" + "3 :%f,%fi\n" + "4 :%f,%fi\n" + "5 :%f,%fi\n" + "6 :%f,%fi\n" + "7 :%f,%fi\n" + "8 :%f,%fi\n" + "9 :%f,%fi\n" + "10:%f,%fi\n" + "11:%f,%fi\n" + "12:%f,%fi\n", + T_CONV creal (c1), T_CONV cimag (c1), + T_CONV creal (c2), T_CONV cimag (c2), + T_CONV creal (c3), T_CONV cimag (c3), + T_CONV creal (c4), T_CONV cimag (c4), + T_CONV creal (c5), T_CONV cimag (c5), + T_CONV creal (c6), T_CONV cimag (c6), + T_CONV creal (c7), T_CONV cimag (c7), + T_CONV creal (c8), T_CONV cimag (c8), + T_CONV creal (c9), T_CONV cimag (c9), + T_CONV creal (c10), T_CONV cimag (c10), + T_CONV creal (c11), T_CONV cimag (c11), + T_CONV creal (c12), T_CONV cimag (c12), + T_CONV creal (c13), T_CONV cimag (c13)); + + return (c1+c2-c3-c4+c5+c6+c7-c8-c9-c10-c11+c12+c13); +} + +int main (void) +{ + ffi_cif cif; + ffi_type *args[13]; + void *values[13]; + _Complex T_C_TYPE ca[13]; + _Complex T_C_TYPE c, cc; + int i; + + for (i = 0; i < 13; i++) + { + args[i] = &T_FFI_TYPE; + values[i] = &ca[i]; + ca[i] = i + (-20 - i) * I; + } + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 13, &T_FFI_TYPE, args) == FFI_OK); + + ffi_call(&cif, FFI_FN(many), &c, values); + + cc = many(ca[0], ca[1], ca[2], ca[3], ca[4], ca[5], ca[6], ca[7], ca[8], + ca[9], ca[10], ca[11], ca[12]); + CHECK(creal (cc) == creal (c)); + CHECK(cimag (cc) == cimag (c)); + + exit(0); +} diff --git a/native/libffi/testsuite/libffi.complex/many_complex_double.c b/native/libffi/testsuite/libffi.complex/many_complex_double.c new file mode 100644 index 0000000000..3fd53c3354 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/many_complex_double.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex, with many arguments + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_double.inc" +#include "many_complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/many_complex_float.c b/native/libffi/testsuite/libffi.complex/many_complex_float.c new file mode 100644 index 0000000000..c43d21cd9e --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/many_complex_float.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex, with many arguments + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_float.inc" +#include "many_complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/many_complex_longdouble.c b/native/libffi/testsuite/libffi.complex/many_complex_longdouble.c new file mode 100644 index 0000000000..dbab723969 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/many_complex_longdouble.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex, with many arguments + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_longdouble.inc" +#include "many_complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/return_complex.inc b/native/libffi/testsuite/libffi.complex/return_complex.inc new file mode 100644 index 0000000000..8bf0c1fbab --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex.inc @@ -0,0 +1,37 @@ +/* -*-c-*- */ +#include "ffitest.h" +#include + +static _Complex T_C_TYPE return_c(_Complex T_C_TYPE c) +{ + printf ("%f,%fi\n", T_CONV creal (c), T_CONV cimag (c)); + return 2 * c; +} +int main (void) +{ + ffi_cif cif; + ffi_type *args[MAX_ARGS]; + void *values[MAX_ARGS]; + _Complex T_C_TYPE c, rc, rc2; + T_C_TYPE cr, ci; + + args[0] = &T_FFI_TYPE; + values[0] = &c; + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1, + &T_FFI_TYPE, args) == FFI_OK); + + for (cr = -127.0; cr < 127; cr++) + { + ci = 1000.0 - cr; + c = cr + ci * I; + ffi_call(&cif, FFI_FN(return_c), &rc, values); + rc2 = return_c(c); + printf ("%f,%fi vs %f,%fi\n", + T_CONV creal (rc), T_CONV cimag (rc), + T_CONV creal (rc2), T_CONV cimag (rc2)); + CHECK(rc == 2 * c); + } + exit(0); +} diff --git a/native/libffi/testsuite/libffi.complex/return_complex1.inc b/native/libffi/testsuite/libffi.complex/return_complex1.inc new file mode 100644 index 0000000000..7cecc0fefa --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex1.inc @@ -0,0 +1,41 @@ +/* -*-c-*- */ +#include "ffitest.h" +#include + +static _Complex T_C_TYPE return_c(_Complex T_C_TYPE c1, float fl2, unsigned int in3, _Complex T_C_TYPE c4) +{ + return c1 + fl2 + in3 + c4; +} +int main (void) +{ + ffi_cif cif; + ffi_type *args[MAX_ARGS]; + void *values[MAX_ARGS]; + _Complex T_C_TYPE c1, c4, rc, rc2; + float fl2; + unsigned int in3; + args[0] = &T_FFI_TYPE; + args[1] = &ffi_type_float; + args[2] = &ffi_type_uint; + args[3] = &T_FFI_TYPE; + values[0] = &c1; + values[1] = &fl2; + values[2] = &in3; + values[3] = &c4; + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 4, + &T_FFI_TYPE, args) == FFI_OK); + c1 = 127.0 + 255.0 * I; + fl2 = 128.0; + in3 = 255; + c4 = 512.7 + 1024.1 * I; + + ffi_call(&cif, FFI_FN(return_c), &rc, values); + rc2 = return_c(c1, fl2, in3, c4); + printf ("%f,%fi vs %f,%fi\n", + T_CONV creal (rc), T_CONV cimag (rc), + T_CONV creal (rc2), T_CONV cimag (rc2)); + CHECK(rc == rc2); + exit(0); +} diff --git a/native/libffi/testsuite/libffi.complex/return_complex1_double.c b/native/libffi/testsuite/libffi.complex/return_complex1_double.c new file mode 100644 index 0000000000..727410d563 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex1_double.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_double.inc" +#include "return_complex1.inc" diff --git a/native/libffi/testsuite/libffi.complex/return_complex1_float.c b/native/libffi/testsuite/libffi.complex/return_complex1_float.c new file mode 100644 index 0000000000..a2aeada847 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex1_float.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_float.inc" +#include "return_complex1.inc" diff --git a/native/libffi/testsuite/libffi.complex/return_complex1_longdouble.c b/native/libffi/testsuite/libffi.complex/return_complex1_longdouble.c new file mode 100644 index 0000000000..103504bf64 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex1_longdouble.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_longdouble.inc" +#include "return_complex1.inc" diff --git a/native/libffi/testsuite/libffi.complex/return_complex2.inc b/native/libffi/testsuite/libffi.complex/return_complex2.inc new file mode 100644 index 0000000000..265170bf71 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex2.inc @@ -0,0 +1,44 @@ +/* -*-c-*- */ +#include "ffitest.h" +#include + +_Complex T_C_TYPE +return_c(_Complex T_C_TYPE c1, _Complex T_C_TYPE c2, + unsigned int in3, _Complex T_C_TYPE c4) +{ + volatile _Complex T_C_TYPE r = c1 + c2 + in3 + c4; + return r; +} + +int main (void) +{ + ffi_cif cif; + ffi_type *args[MAX_ARGS]; + void *values[MAX_ARGS]; + _Complex T_C_TYPE c1, c2, c4, rc, rc2; + unsigned int in3; + args[0] = &T_FFI_TYPE; + args[1] = &T_FFI_TYPE; + args[2] = &ffi_type_uint; + args[3] = &T_FFI_TYPE; + values[0] = &c1; + values[1] = &c2; + values[2] = &in3; + values[3] = &c4; + + /* Initialize the cif */ + CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 4, + &T_FFI_TYPE, args) == FFI_OK); + c1 = 127.0 + 255.0 * I; + c2 = 128.0 + 256.0; + in3 = 255; + c4 = 512.7 + 1024.1 * I; + + ffi_call(&cif, FFI_FN(return_c), &rc, values); + rc2 = return_c(c1, c2, in3, c4); + printf ("%f,%fi vs %f,%fi\n", + T_CONV creal (rc), T_CONV cimag (rc), + T_CONV creal (rc2), T_CONV cimag (rc2)); + CHECK(rc == rc2); + exit(0); +} diff --git a/native/libffi/testsuite/libffi.complex/return_complex2_double.c b/native/libffi/testsuite/libffi.complex/return_complex2_double.c new file mode 100644 index 0000000000..ab9efacb4a --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex2_double.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_double.inc" +#include "return_complex2.inc" diff --git a/native/libffi/testsuite/libffi.complex/return_complex2_float.c b/native/libffi/testsuite/libffi.complex/return_complex2_float.c new file mode 100644 index 0000000000..d7f22c2a0c --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex2_float.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_float.inc" +#include "return_complex2.inc" diff --git a/native/libffi/testsuite/libffi.complex/return_complex2_longdouble.c b/native/libffi/testsuite/libffi.complex/return_complex2_longdouble.c new file mode 100644 index 0000000000..3edea629df --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex2_longdouble.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_longdouble.inc" +#include "return_complex2.inc" diff --git a/native/libffi/testsuite/libffi.complex/return_complex_double.c b/native/libffi/testsuite/libffi.complex/return_complex_double.c new file mode 100644 index 0000000000..e2497cc84a --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex_double.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_double.inc" +#include "return_complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/return_complex_float.c b/native/libffi/testsuite/libffi.complex/return_complex_float.c new file mode 100644 index 0000000000..a35528ff93 --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex_float.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_float.inc" +#include "return_complex.inc" diff --git a/native/libffi/testsuite/libffi.complex/return_complex_longdouble.c b/native/libffi/testsuite/libffi.complex/return_complex_longdouble.c new file mode 100644 index 0000000000..142d7becba --- /dev/null +++ b/native/libffi/testsuite/libffi.complex/return_complex_longdouble.c @@ -0,0 +1,10 @@ +/* Area: ffi_call + Purpose: Check return value complex. + Limitations: none. + PR: none. + Originator: . */ + +/* { dg-do run } */ + +#include "complex_defs_longdouble.inc" +#include "return_complex.inc" diff --git a/native/libffi/testsuite/libffi.go/aa-direct.c b/native/libffi/testsuite/libffi.go/aa-direct.c new file mode 100644 index 0000000000..b00c404ab3 --- /dev/null +++ b/native/libffi/testsuite/libffi.go/aa-direct.c @@ -0,0 +1,34 @@ +/* { dg-do run } */ + +#include "static-chain.h" + +#if defined(__GNUC__) && !defined(__clang__) && defined(STATIC_CHAIN_REG) + +#include "ffitest.h" + +/* Blatent assumption here that the prologue doesn't clobber the + static chain for trivial functions. If this is not true, don't + define STATIC_CHAIN_REG, and we'll test what we can via other tests. */ +void *doit(void) +{ + register void *chain __asm__(STATIC_CHAIN_REG); + return chain; +} + +int main() +{ + ffi_cif cif; + void *result; + + CHECK(ffi_prep_cif(&cif, ABI_NUM, 0, &ffi_type_pointer, NULL) == FFI_OK); + + ffi_call_go(&cif, FFI_FN(doit), &result, NULL, &result); + + CHECK(result == &result); + + return 0; +} + +#else /* UNSUPPORTED */ +int main() { return 0; } +#endif diff --git a/native/libffi/testsuite/libffi.go/closure1.c b/native/libffi/testsuite/libffi.go/closure1.c new file mode 100644 index 0000000000..7b34afc88c --- /dev/null +++ b/native/libffi/testsuite/libffi.go/closure1.c @@ -0,0 +1,28 @@ +/* { dg-do run } */ + +#include "ffitest.h" + +void doit(ffi_cif *cif, void *rvalue, void **avalue, void *closure) +{ + (void)cif; + (void)avalue; + *(void **)rvalue = closure; +} + +typedef void * (*FN)(void); + +int main() +{ + ffi_cif cif; + ffi_go_closure cl; + void *result; + + CHECK(ffi_prep_cif(&cif, ABI_NUM, 0, &ffi_type_pointer, NULL) == FFI_OK); + CHECK(ffi_prep_go_closure(&cl, &cif, doit) == FFI_OK); + + ffi_call_go(&cif, FFI_FN(*(FN *)&cl), &result, NULL, &cl); + + CHECK(result == &cl); + + exit(0); +} diff --git a/native/libffi/testsuite/libffi.go/ffitest.h b/native/libffi/testsuite/libffi.go/ffitest.h new file mode 100644 index 0000000000..d27d362d6a --- /dev/null +++ b/native/libffi/testsuite/libffi.go/ffitest.h @@ -0,0 +1 @@ +#include "../libffi.call/ffitest.h" diff --git a/native/libffi/testsuite/libffi.go/go.exp b/native/libffi/testsuite/libffi.go/go.exp new file mode 100644 index 0000000000..100c5e75b4 --- /dev/null +++ b/native/libffi/testsuite/libffi.go/go.exp @@ -0,0 +1,36 @@ +# Copyright (C) 2003, 2006, 2009, 2010, 2014 Free Software Foundation, Inc. + +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 3 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; see the file COPYING3. If not see +# . + +dg-init +libffi-init + +global srcdir subdir + +set tlist [lsort [glob -nocomplain -- $srcdir/$subdir/*.{c,cc}]] + +if { [libffi_feature_test "#ifdef FFI_GO_CLOSURES"] } { + run-many-tests $tlist "" +} else { + foreach test $tlist { + unsupported "$test" + } +} + +dg-finish + +# Local Variables: +# tcl-indent-level:4 +# End: diff --git a/native/libffi/testsuite/libffi.go/static-chain.h b/native/libffi/testsuite/libffi.go/static-chain.h new file mode 100644 index 0000000000..3675b40a54 --- /dev/null +++ b/native/libffi/testsuite/libffi.go/static-chain.h @@ -0,0 +1,19 @@ +#ifdef __aarch64__ +# define STATIC_CHAIN_REG "x18" +#elif defined(__alpha__) +# define STATIC_CHAIN_REG "$1" +#elif defined(__arm__) +# define STATIC_CHAIN_REG "ip" +#elif defined(__sparc__) +# if defined(__arch64__) || defined(__sparcv9) +# define STATIC_CHAIN_REG "g5" +# else +# define STATIC_CHAIN_REG "g2" +# endif +#elif defined(__x86_64__) +# define STATIC_CHAIN_REG "r10" +#elif defined(__i386__) +# ifndef ABI_NUM +# define STATIC_CHAIN_REG "ecx" /* FFI_DEFAULT_ABI only */ +# endif +#endif