[llvm-readobj,ELF] Support --decompress/-z #82594
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Created using spr 1.3.4
|
@llvm/pr-subscribers-llvm-binary-utilities Author: Fangrui Song (MaskRay) ChangesWhen a section has the SHF_COMPRESSED flag, -p/-x dump the compressed Close #82507 Full diff: https://github.com/llvm/llvm-project/pull/82594.diff 8 Files Affected:
diff --git a/llvm/docs/CommandGuide/llvm-readelf.rst b/llvm/docs/CommandGuide/llvm-readelf.rst
index 6ee4a5dfb15917..c302a6316ccc9b 100644
--- a/llvm/docs/CommandGuide/llvm-readelf.rst
+++ b/llvm/docs/CommandGuide/llvm-readelf.rst
@@ -38,6 +38,11 @@ OPTIONS
Display the contents of the basic block address map section(s), which contain the
address of each function, along with the relative offset of each basic block.
+.. option:: --decompress, -z
+
+ Dump decompressed section content when used with ``-x`` or ``-p``.
+ If the section(s) are not compressed, they are displayed as is.
+
.. option:: --demangle, -C
Display demangled symbol names in the output.
@@ -72,7 +77,7 @@ OPTIONS
.. option:: --elf-output-style=<value>
- Format ELF information in the specified style. Valid options are ``LLVM``,
+ Format ELF information in the specified style. Valid options are ``````LLVM``,
``GNU``, and ``JSON``. ``LLVM`` output is an expanded and structured format.
``GNU`` (the default) output mimics the equivalent GNU :program:`readelf`
output. ``JSON`` is JSON formatted output intended for machine consumption.
diff --git a/llvm/docs/CommandGuide/llvm-readobj.rst b/llvm/docs/CommandGuide/llvm-readobj.rst
index cb9232ef5e560a..6d78a038723445 100644
--- a/llvm/docs/CommandGuide/llvm-readobj.rst
+++ b/llvm/docs/CommandGuide/llvm-readobj.rst
@@ -56,6 +56,11 @@ file formats.
Display the address-significance table.
+.. option:: --decompress, -z
+
+ Dump decompressed section content when used with ``-x`` or ``-p``.
+ If the section(s) are not compressed, they are displayed as is.
+
.. option:: --expand-relocs
When used with :option:`--relocs`, display each relocation in an expanded
diff --git a/llvm/test/tools/llvm-readobj/ELF/decompress-unsupported-zlib.test b/llvm/test/tools/llvm-readobj/ELF/decompress-unsupported-zlib.test
new file mode 100644
index 00000000000000..f4c73de7ca6c9d
--- /dev/null
+++ b/llvm/test/tools/llvm-readobj/ELF/decompress-unsupported-zlib.test
@@ -0,0 +1,32 @@
+# UNSUPPORTED: zlib
+# RUN: yaml2obj %s -o %t
+# RUN: llvm-readobj -z -p .a -x .b %t 2>&1 | FileCheck %s -DFILE=%t
+
+# CHECK: String dump of section '.a':
+# CHECK-NEXT: warning: '[[FILE]]': LLVM was not built with LLVM_ENABLE_ZLIB or did not find zlib at build time
+# CHECK-NEXT: [ 0] .
+# CHECK-NEXT: [ 8] .
+# CHECK-NEXT: [ 10] .
+# CHECK-NEXT: [ 18] x.c.
+# CHECK-NEXT: [ 1e] .
+# CHECK-NEXT: [ 20] .
+# CHECK-NEXT: Hex dump of section '.b':
+# CHECK-NEXT: warning: '[[FILE]]': LLVM was not built with LLVM_ENABLE_ZLIB or did not find zlib at build time
+# CHECK-NEXT: 0x00000000 01000000 00000000 01000000 00000000 ................
+# CHECK-NEXT: 0x00000010 01000000 00000000 789c6304 00000200 ........x.c.....
+# CHECK-NEXT: 0x00000020 02 .
+
+--- !ELF
+FileHeader:
+ Class: ELFCLASS64
+ Data: ELFDATA2LSB
+ Type: ET_REL
+Sections:
+ - Name: .a
+ Type: SHT_PROGBITS
+ Flags: [SHF_COMPRESSED]
+ Content: 010000000000000001000000000000000100000000000000789c63040000020002
+ - Name: .b
+ Type: SHT_PROGBITS
+ Flags: [SHF_COMPRESSED]
+ Content: 010000000000000001000000000000000100000000000000789c63040000020002
diff --git a/llvm/test/tools/llvm-readobj/ELF/decompress.test b/llvm/test/tools/llvm-readobj/ELF/decompress.test
new file mode 100644
index 00000000000000..638b5cc37d3a89
--- /dev/null
+++ b/llvm/test/tools/llvm-readobj/ELF/decompress.test
@@ -0,0 +1,67 @@
+# REQUIRES: zlib
+## Test --decompress/-z
+
+# RUN: yaml2obj %s -o %t
+
+# RUN: llvm-readelf -z -x .strings -x .not_null_terminated %t | FileCheck %s --check-prefix=HEX
+# RUN: llvm-readobj --decompress -p .strings -p .not_null_terminated %t | FileCheck %s --check-prefix=STR
+
+# HEX: Hex dump of section '.strings':
+# HEX-NEXT: 0x00000000 68657265 00617265 00736f6d 65007374 here.are.some.st
+# HEX-NEXT: 0x00000010 72696e67 7300 rings.
+# HEX: Hex dump of section '.not_null_terminated':
+# HEX-NEXT: 0x00000000 6e6f006e 756c6c no.null
+
+# RUN: llvm-readobj -x .strings -p .not_null_terminated %t | FileCheck %s --check-prefix=COMPRESSED
+
+# COMPRESSED: String dump of section '.not_null_terminated':
+# COMPRESSED-NEXT: [ 0] no
+# COMPRESSED-NEXT: [ 3] null
+# COMPRESSED-NEXT: Hex dump of section '.strings':
+# COMPRESSED-NEXT: 0x00000000 01000000 00000000 16000000 00000000 ................
+# COMPRESSED-NEXT: 0x00000010 00000000 00000000 789ccb48 2d4a6548 ........x..H-JeH
+# COMPRESSED-NEXT: 0x00000020 04e2e2fc 5c205152 9499975e cc000058 ....\ QR...^...X
+# COMPRESSED-NEXT: 0x00000030 2e079b ...
+
+# STR: String dump of section '.strings':
+# STR-NEXT: [ 0] here
+# STR-NEXT: [ 5] are
+# STR-NEXT: [ 9] some
+# STR-NEXT: [ e] strings
+# STR-EMPTY:
+# STR-NEXT: String dump of section '.not_null_terminated':
+# STR-NEXT: [ 0] no
+# STR-NEXT: [ 3] null{{$}}
+# STR-NOT: {{.}}
+
+# RUN: llvm-readelf -z -p .invalid1 -x .invalid2 %t 2>&1 | FileCheck %s -DFILE=%t --check-prefix=INVALID
+
+# INVALID: String dump of section '.invalid1':
+# INVALID-NEXT: warning: '[[FILE]]': corrupted compressed section header
+# INVALID-NEXT: [ 0] .
+# INVALID-NEXT: Hex dump of section '.invalid2':
+# INVALID-NEXT: warning: '[[FILE]]': zlib error: Z_DATA_ERROR
+# INVALID-NEXT: 0x00000000 01000000 00000000 16000000 00000000 ................
+# INVALID-NEXT: 0x00000010 00000000 00000000 78 ........x
+
+--- !ELF
+FileHeader:
+ Class: ELFCLASS64
+ Data: ELFDATA2LSB
+ Type: ET_REL
+Sections:
+ - Name: .strings
+ Type: SHT_PROGBITS
+ Flags: [SHF_COMPRESSED]
+ Content: 010000000000000016000000000000000000000000000000789ccb482d4a654804e2e2fc5c2051529499975ecc0000582e079b
+ - Name: .not_null_terminated
+ Type: SHT_PROGBITS
+ Content: 6e6f006e756c6c
+ - Name: .invalid1
+ Type: SHT_PROGBITS
+ Flags: [SHF_COMPRESSED]
+ Content: 01
+ - Name: .invalid2
+ Type: SHT_PROGBITS
+ Flags: [SHF_COMPRESSED]
+ Content: 01000000000000001600000000000000000000000000000078
diff --git a/llvm/tools/llvm-readobj/ObjDumper.cpp b/llvm/tools/llvm-readobj/ObjDumper.cpp
index 59060ac217e32f..0d3fea71aafd42 100644
--- a/llvm/tools/llvm-readobj/ObjDumper.cpp
+++ b/llvm/tools/llvm-readobj/ObjDumper.cpp
@@ -14,6 +14,7 @@
#include "ObjDumper.h"
#include "llvm-readobj.h"
#include "llvm/Object/Archive.h"
+#include "llvm/Object/Decompressor.h"
#include "llvm/Object/ObjectFile.h"
#include "llvm/Support/Error.h"
#include "llvm/Support/FormatVariadic.h"
@@ -142,8 +143,23 @@ getSectionRefsByNameOrIndex(const object::ObjectFile &Obj,
return Ret;
}
+static void maybeDecompress(const object::ObjectFile &Obj,
+ StringRef SectionName, StringRef &SectionContent,
+ SmallString<0> &Out) {
+ Expected<object::Decompressor> Decompressor = object::Decompressor::create(
+ SectionName, SectionContent, Obj.isLittleEndian(), Obj.is64Bit());
+ if (!Decompressor)
+ reportWarning(Decompressor.takeError(), Obj.getFileName());
+ else if (auto Err = Decompressor->resizeAndDecompress(Out))
+ reportWarning(std::move(Err), Obj.getFileName());
+ else
+ SectionContent = Out;
+}
+
void ObjDumper::printSectionsAsString(const object::ObjectFile &Obj,
- ArrayRef<std::string> Sections) {
+ ArrayRef<std::string> Sections,
+ bool Decompress) {
+ SmallString<0> Out;
bool First = true;
for (object::SectionRef Section :
getSectionRefsByNameOrIndex(Obj, Sections)) {
@@ -156,12 +172,16 @@ void ObjDumper::printSectionsAsString(const object::ObjectFile &Obj,
StringRef SectionContent =
unwrapOrError(Obj.getFileName(), Section.getContents());
+ if (Decompress && Section.isCompressed())
+ maybeDecompress(Obj, SectionName, SectionContent, Out);
printAsStringList(SectionContent);
}
}
void ObjDumper::printSectionsAsHex(const object::ObjectFile &Obj,
- ArrayRef<std::string> Sections) {
+ ArrayRef<std::string> Sections,
+ bool Decompress) {
+ SmallString<0> Out;
bool First = true;
for (object::SectionRef Section :
getSectionRefsByNameOrIndex(Obj, Sections)) {
@@ -174,6 +194,8 @@ void ObjDumper::printSectionsAsHex(const object::ObjectFile &Obj,
StringRef SectionContent =
unwrapOrError(Obj.getFileName(), Section.getContents());
+ if (Decompress && Section.isCompressed())
+ maybeDecompress(Obj, SectionName, SectionContent, Out);
const uint8_t *SecContent = SectionContent.bytes_begin();
const uint8_t *SecEnd = SecContent + SectionContent.size();
diff --git a/llvm/tools/llvm-readobj/ObjDumper.h b/llvm/tools/llvm-readobj/ObjDumper.h
index 1d679453581bc8..3958dd3a333332 100644
--- a/llvm/tools/llvm-readobj/ObjDumper.h
+++ b/llvm/tools/llvm-readobj/ObjDumper.h
@@ -175,9 +175,9 @@ class ObjDumper {
void printAsStringList(StringRef StringContent, size_t StringDataOffset = 0);
void printSectionsAsString(const object::ObjectFile &Obj,
- ArrayRef<std::string> Sections);
+ ArrayRef<std::string> Sections, bool Decompress);
void printSectionsAsHex(const object::ObjectFile &Obj,
- ArrayRef<std::string> Sections);
+ ArrayRef<std::string> Sections, bool Decompress);
std::function<Error(const Twine &Msg)> WarningHandler;
void reportUniqueWarning(Error Err) const;
diff --git a/llvm/tools/llvm-readobj/Opts.td b/llvm/tools/llvm-readobj/Opts.td
index e2d93c6ec229e9..018facc278e891 100644
--- a/llvm/tools/llvm-readobj/Opts.td
+++ b/llvm/tools/llvm-readobj/Opts.td
@@ -20,6 +20,7 @@ def all : FF<"all", "Equivalent to setting: --file-header, --program-headers, --
def arch_specific : FF<"arch-specific", "Display architecture-specific information">;
def bb_addr_map : FF<"bb-addr-map", "Display the BB address map section">;
def cg_profile : FF<"cg-profile", "Display call graph profile section">;
+def decompress : FF<"decompress", "Dump decompressed section content when used with -x or -p">;
defm demangle : BB<"demangle", "Demangle symbol names", "Do not demangle symbol names (default)">;
def dependent_libraries : FF<"dependent-libraries", "Display the dependent libraries section">;
def dyn_relocations : FF<"dyn-relocations", "Display the dynamic relocation entries in the file">;
@@ -139,3 +140,4 @@ def : F<"u", "Alias for --unwind">, Alias<unwind>;
def : F<"X", "Alias for --extra-sym-info">, Alias<extra_sym_info>, Group<grp_elf>;
def : F<"V", "Alias for --version-info">, Alias<version_info>, Group<grp_elf>;
def : JoinedOrSeparate<["-"], "x">, Alias<hex_dump_EQ>, HelpText<"Alias for --hex-dump">, MetaVarName<"<name or index>">;
+def : F<"z", "Alias for --decompress">, Alias<decompress>;
diff --git a/llvm/tools/llvm-readobj/llvm-readobj.cpp b/llvm/tools/llvm-readobj/llvm-readobj.cpp
index f9d605d35244bf..979433d69011c3 100644
--- a/llvm/tools/llvm-readobj/llvm-readobj.cpp
+++ b/llvm/tools/llvm-readobj/llvm-readobj.cpp
@@ -97,6 +97,7 @@ static bool ArchSpecificInfo;
static bool BBAddrMap;
bool ExpandRelocs;
static bool CGProfile;
+static bool Decompress;
bool Demangle;
static bool DependentLibraries;
static bool DynRelocs;
@@ -212,6 +213,7 @@ static void parseOptions(const opt::InputArgList &Args) {
opts::ArchSpecificInfo = Args.hasArg(OPT_arch_specific);
opts::BBAddrMap = Args.hasArg(OPT_bb_addr_map);
opts::CGProfile = Args.hasArg(OPT_cg_profile);
+ opts::Decompress = Args.hasArg(OPT_decompress);
opts::Demangle = Args.hasFlag(OPT_demangle, OPT_no_demangle, false);
opts::DependentLibraries = Args.hasArg(OPT_dependent_libraries);
opts::DynRelocs = Args.hasArg(OPT_dyn_relocations);
@@ -439,9 +441,9 @@ static void dumpObject(ObjectFile &Obj, ScopedPrinter &Writer,
Dumper->printSymbols(opts::Symbols, opts::DynamicSymbols,
opts::ExtraSymInfo, SymComp);
if (!opts::StringDump.empty())
- Dumper->printSectionsAsString(Obj, opts::StringDump);
+ Dumper->printSectionsAsString(Obj, opts::StringDump, opts::Decompress);
if (!opts::HexDump.empty())
- Dumper->printSectionsAsHex(Obj, opts::HexDump);
+ Dumper->printSectionsAsHex(Obj, opts::HexDump, opts::Decompress);
if (opts::HashTable)
Dumper->printHashTable();
if (opts::GnuHashTable)
|
Created using spr 1.3.4
jh7370
approved these changes
Feb 22, 2024
| @@ -0,0 +1,67 @@ | |||
| # REQUIRES: zlib | |||
| ## Test --decompress/-z | |||
There was a problem hiding this comment.
Suggested change
| ## Test --decompress/-z | |
| ## Test --decompress/-z. |
Nit.
| @@ -0,0 +1,67 @@ | |||
| # REQUIRES: zlib | |||
There was a problem hiding this comment.
Perhaps worth naming decompress-zlib.test, to allow for a zstd version too? (Same applies for the unsupported case).
There was a problem hiding this comment.
Sounds good. Done. Added a zstd version.
There is decompress-zlib-unsupported.test but no decompress-zstd-unsupported.test. The latter probably doesn't add much coverage.
There was a problem hiding this comment.
That's fine, but one thought that did occur to me is that I suspect zlib is available on more machines than zstd, therefore by having a zstd unsupported test case, you'd be more likely to have coverage of both the supported and unsupported code paths on the same machine.
I'm okay if you still want to do it though.
There was a problem hiding this comment.
I think Windows is a popular OS that often misses zlib support. However, "often" may degrade to "sometimes". Anyway, I just added decompress-zstd-unsupported.test...
Created using spr 1.3.4
Created using spr 1.3.4
emaste
approved these changes
Feb 22, 2024
Created using spr 1.3.4
Thanks. Added |
MaskRay
deleted the
users/MaskRay/spr/llvm-readobjelf-support-decompress-z
branch
llvmbot
pushed a commit
to llvmbot/llvm-project
that referenced
this pull request
Feb 22, 2024
When a section has the SHF_COMPRESSED flag, -p/-x dump the compressed content by default. In GNU readelf, if --decompress/-z is specified, -p/-x will dump the decompressed content. This patch implements the option. Close llvm#82507 (cherry picked from commit 26d71d9)
brad0
added a commit
to brad0/llvm-project
that referenced
this pull request
Feb 23, 2024
When a section has the SHF_COMPRESSED flag, -p/-x dump the compressed content by default. In GNU readelf, if --decompress/-z is specified, -p/-x will dump the decompressed content. This patch implements the option. Close llvm#82507 (cherry picked from commit 26d71d9)
llvmbot
pushed a commit
to llvmbot/llvm-project
that referenced
this pull request
Feb 23, 2024
When a section has the SHF_COMPRESSED flag, -p/-x dump the compressed content by default. In GNU readelf, if --decompress/-z is specified, -p/-x will dump the decompressed content. This patch implements the option. Close llvm#82507 (cherry picked from commit 26d71d9)
llvmbot
pushed a commit
to llvmbot/llvm-project
that referenced
this pull request
Feb 23, 2024
When a section has the SHF_COMPRESSED flag, -p/-x dump the compressed content by default. In GNU readelf, if --decompress/-z is specified, -p/-x will dump the decompressed content. This patch implements the option. Close llvm#82507 (cherry picked from commit 26d71d9)
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When a section has the SHF_COMPRESSED flag, -p/-x dump the compressed
content by default. In GNU readelf, if --decompress/-z is specified,
-p/-x will dump the decompressed content. This patch implements the
option.
Close #82507