-
Notifications
You must be signed in to change notification settings - Fork 12.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Track test suite failures on s390x #105383
Comments
Thanks for your attention on this!
I think I caught most of these in #102379, but only by ignoring them. There's a suggestion that we might apply more fine-grained filtering on that, but I haven't gotten around to implementing that yet.
Generally speaking, that's defined by Platform Support and the related Target Tier Policy, where tier-2 (like s390x) is only cross-built in CI, and tier-1 adds test gating. Aarch64-linux is currently the only non-x86 target at tier-1. Part of the challenge is that most developers simply don't have access to s390x hardware to diagnose problems, so it would be burdensome to make that a blocking target. Perhaps you have resources that could be applied more ad-hoc? e.g. Google has their own CI building Rust with the latest LLVM main branch, from which they report issues (or usually direct fixes). If you could set up similar external CI for s390x testing, we could get a shorter feedback loop without making it a blocking tier-1 target. |
Several codegen tests are currently failing due to making assumptions that are not valid for the s390x architecture: - catch-unwind.rs: fails due to inlining differences. Already ignored on another platform for the same reason. Solution: Ignore on s390x. - remap_path_prefix/main.rs: fails due to different alignment requirement for string constants. Solution: Do not test for the alignment requirement. - repr-transparent-aggregates-1.rs: many ABI assumptions. Already ignored on many platforms for the same reason. Solution: Ignore on s390x. - repr-transparent.rs: no vector ABI by default on s390x. Already ignored on another platform for a similar reason. Solution: Ignore on s390x. - uninit-consts.rs: hard-coded little-endian constant. Solution: Match both little- and big-endian versions. Fixes part of rust-lang#105383.
This is certainly something we could set up. I guess I'm mostly concerned about endian issues, which seem to crop up regularly. Having at least one big-endian target in the CI might be useful ... But we can start with our own external CI, and see how that goes over time. |
…Simulacrum Fix failing codegen tests on s390x Several codegen tests are currently failing due to making assumptions that are not valid for the s390x architecture: - catch-unwind.rs: fails due to inlining differences. Already ignored on another platform for the same reason. Solution: Ignore on s390x. - remap_path_prefix/main.rs: fails due to different alignment requirement for string constants. Solution: Do not test for the alignment requirement. - repr-transparent-aggregates-1.rs: many ABI assumptions. Already ignored on many platforms for the same reason. Solution: Ignore on s390x. - repr-transparent.rs: no vector ABI by default on s390x. Already ignored on another platform for a similar reason. Solution: Ignore on s390x. - uninit-consts.rs: hard-coded little-endian constant. Solution: Match both little- and big-endian versions. Fixes part of rust-lang#105383.
Fix failing codegen tests on s390x Several codegen tests are currently failing due to making assumptions that are not valid for the s390x architecture: - catch-unwind.rs: fails due to inlining differences. Already ignored on another platform for the same reason. Solution: Ignore on s390x. - remap_path_prefix/main.rs: fails due to different alignment requirement for string constants. Solution: Do not test for the alignment requirement. - repr-transparent-aggregates-1.rs: many ABI assumptions. Already ignored on many platforms for the same reason. Solution: Ignore on s390x. - repr-transparent.rs: no vector ABI by default on s390x. Already ignored on another platform for a similar reason. Solution: Ignore on s390x. - uninit-consts.rs: hard-coded little-endian constant. Solution: Match both little- and big-endian versions. Fixes part of rust-lang/rust#105383.
The src/test/ui/issues/issue-74564-if-expr-stack-overflow.rs test case added to verify rust-lang#74564 still crashes with a stack overflow on s390x-ibm-linux. Symptom is a very deep recursion in compiler/rustc_lint/src/early.rs: fn visit_expr(&mut self, e: &'a ast::Expr) { self.with_lint_attrs(e.id, &e.attrs, |cx| { lint_callback!(cx, check_expr, e); ast_visit::walk_expr(cx, e); }) } (where walk_expr recursively calls back into visit_expr). The crash happens at a nesting depth of over 17000 stack frames when using the default 8 MB stack size on s390x. This patch fixes the problem by adding a ensure_sufficient_stack call to the with_lint_attrs routine (which also should take care of all the other mutually recursive visitors here). Fixes part of rust-lang#105383.
A number of tests under ui/const-ptr and ui/consts are currently failing on big-endian platforms as the binary encoding of some constants is hard-coded in the stderr test files. Fix this by providing a normalize-stderr-test rule that strips out the raw bytes hex dump, so the comparison can be done in an endianness-independent manner. Note that in most cases, this means the tests are now also independent of word size, so the 32bit and 64bit cases can be re-unified. To keep tests that verify the details of those raw bytes dumps, a new test case raw-bytes.rs performs the tests where the hex dumps were stripped out a second time, but only on little- endian platforms. In addition, src/test/ui/const-ptr/forbidden_slices.rs exposes an endian-specific difference in this diagnostic output: constructing invalid value at .<deref>[0]: encountered 0x11, but expected a boolean depending on which byte of D0 is not a boolean value (0 or 1). Fixed this by choosing a value of D0 that differs from 0 or 1 in all bytes. Fixes part of rust-lang#105383.
…r=oli-obk Fix ui constant tests for big-endian platforms A number of tests under ui/const-ptr and ui/consts are currently failing on big-endian platforms as the binary encoding of some constants is hard-coded in the stderr test files. Fix this by a combination of two types of changes: - Where possible (i.e. where the particular value of a constant does not affect the purpose of the test), choose constant values that have the same encoding on big- and little-endian platforms. - Where this is not possible, provide a normalize-stderr-test rule that transforms the printed big-endian encoding of such constants into the corresponding little-endian form. Fixes part of rust-lang#105383.
…r=oli-obk Fix ui constant tests for big-endian platforms A number of tests under ui/const-ptr and ui/consts are currently failing on big-endian platforms as the binary encoding of some constants is hard-coded in the stderr test files. Fix this by a combination of two types of changes: - Where possible (i.e. where the particular value of a constant does not affect the purpose of the test), choose constant values that have the same encoding on big- and little-endian platforms. - Where this is not possible, provide a normalize-stderr-test rule that transforms the printed big-endian encoding of such constants into the corresponding little-endian form. Fixes part of rust-lang#105383.
…r=oli-obk Fix ui constant tests for big-endian platforms A number of tests under ui/const-ptr and ui/consts are currently failing on big-endian platforms as the binary encoding of some constants is hard-coded in the stderr test files. Fix this by a combination of two types of changes: - Where possible (i.e. where the particular value of a constant does not affect the purpose of the test), choose constant values that have the same encoding on big- and little-endian platforms. - Where this is not possible, provide a normalize-stderr-test rule that transforms the printed big-endian encoding of such constants into the corresponding little-endian form. Fixes part of rust-lang#105383.
…strieb Fix stack overflow in recursive AST walk in early lint The src/test/ui/issues/issue-74564-if-expr-stack-overflow.rs test case added to verify rust-lang#74564 still crashes with a stack overflow on s390x-ibm-linux. Symptom is a very deep recursion in compiler/rustc_lint/src/early.rs: fn visit_expr(&mut self, e: &'a ast::Expr) { self.with_lint_attrs(e.id, &e.attrs, |cx| { lint_callback!(cx, check_expr, e); ast_visit::walk_expr(cx, e); }) } (where walk_expr recursively calls back into visit_expr). The crash happens at a nesting depth of over 17000 stack frames when using the default 8 MB stack size on s390x. This patch fixes the problem by adding a ensure_sufficient_stack call to the with_lint_attrs routine (which also should take care of all the other mutually recursive visitors here). Fixes part of rust-lang#105383.
The test cases src/test/mir-opt/building/custom/consts.rs and src/test/mir-opt/const_prop/mutable_variable_no_prop.rs are currently failing on big-endian platforms as the binary encoding of some constants is hard-coded in the MIR test files. Fix this by choosing constant values that have the same encoding on big- and little-endian platforms. The test case src/test/mir-opt/issues/issue_75439.rs is failing as well, but since the purpose of the test is to validate handling of big-endian integer encodings on a little-endian platform, it does not make much sense to run it on big-endian platforms in the first place - we can just ignore it there. Fixed part of rust-lang#105383.
Fix stack overflow in recursive AST walk in early lint The src/test/ui/issues/issue-74564-if-expr-stack-overflow.rs test case added to verify rust-lang/rust#74564 still crashes with a stack overflow on s390x-ibm-linux. Symptom is a very deep recursion in compiler/rustc_lint/src/early.rs: fn visit_expr(&mut self, e: &'a ast::Expr) { self.with_lint_attrs(e.id, &e.attrs, |cx| { lint_callback!(cx, check_expr, e); ast_visit::walk_expr(cx, e); }) } (where walk_expr recursively calls back into visit_expr). The crash happens at a nesting depth of over 17000 stack frames when using the default 8 MB stack size on s390x. This patch fixes the problem by adding a ensure_sufficient_stack call to the with_lint_attrs routine (which also should take care of all the other mutually recursive visitors here). Fixes part of rust-lang/rust#105383.
… r=Mark-Simulacrum Fix mir-opt tests for big-endian platforms The test cases src/test/mir-opt/building/custom/consts.rs and src/test/mir-opt/const_prop/mutable_variable_no_prop.rs are currently failing on big-endian platforms as the binary encoding of some constants is hard-coded in the MIR test files. Fix this by choosing constant values that have the same encoding on big- and little-endian platforms. The test case src/test/mir-opt/issues/issue_75439.rs is failing as well, but since the purpose of the test is to validate handling of big-endian integer encodings on a little-endian platform, it does not make much sense to run it on big-endian platforms in the first place - we can just ignore it there. Fixed part of rust-lang#105383.
… r=Mark-Simulacrum Fix mir-opt tests for big-endian platforms The test cases src/test/mir-opt/building/custom/consts.rs and src/test/mir-opt/const_prop/mutable_variable_no_prop.rs are currently failing on big-endian platforms as the binary encoding of some constants is hard-coded in the MIR test files. Fix this by choosing constant values that have the same encoding on big- and little-endian platforms. The test case src/test/mir-opt/issues/issue_75439.rs is failing as well, but since the purpose of the test is to validate handling of big-endian integer encodings on a little-endian platform, it does not make much sense to run it on big-endian platforms in the first place - we can just ignore it there. Fixed part of rust-lang#105383.
… r=Mark-Simulacrum Fix mir-opt tests for big-endian platforms The test cases src/test/mir-opt/building/custom/consts.rs and src/test/mir-opt/const_prop/mutable_variable_no_prop.rs are currently failing on big-endian platforms as the binary encoding of some constants is hard-coded in the MIR test files. Fix this by choosing constant values that have the same encoding on big- and little-endian platforms. The test case src/test/mir-opt/issues/issue_75439.rs is failing as well, but since the purpose of the test is to validate handling of big-endian integer encodings on a little-endian platform, it does not make much sense to run it on big-endian platforms in the first place - we can just ignore it there. Fixed part of rust-lang#105383.
All failures tracked here have now be resolved; the Rust test suite is currently clean on s390x. |
As of commit 7767cbb, the tests/ui/consts/const-eval/ub-int-array.rs test is failing on big-endian platforms (in particular s390x), as the stderr output contains a hex dump that depends on endianness. Since this point intentionally verifies the hex dump to check the uninitialized byte markers, I think we should not simply standardize away the hex dump as is done with some of the other tests in this directory. However, most of the test is already endian-independent. The only exception is one line of hex dump, which can also be made endian-independent by choosing appropriate constants in the source code. Since the 32bit and 64bit stderr outputs were already (and remain) identical, I've merged them and removed the stderr-per-bitwidth marker. Fixes (again) rust-lang#105383.
…RalfJung Fix ub-int-array test for big-endian platforms As of commit 7767cbb, the tests/ui/consts/const-eval/ub-int-array.rs test is failing on big-endian platforms (in particular s390x), as the stderr output contains a hex dump that depends on endianness. Since this point intentionally verifies the hex dump to check the uninitialized byte markers, I think we should not simply standardize away the hex dump as is done with some of the other tests in this directory. However, most of the test is already endian-independent. The only exception is one line of hex dump, which can also be made endian-independent by choosing appropriate constants in the source code. Since the 32bit and 64bit stderr outputs were already (and remain) identical, I've merged them and removed the stderr-per-bitwidth marker. Fixes (again) rust-lang#105383.
Fix stack overflow in recursive AST walk in early lint The src/test/ui/issues/issue-74564-if-expr-stack-overflow.rs test case added to verify rust-lang/rust#74564 still crashes with a stack overflow on s390x-ibm-linux. Symptom is a very deep recursion in compiler/rustc_lint/src/early.rs: fn visit_expr(&mut self, e: &'a ast::Expr) { self.with_lint_attrs(e.id, &e.attrs, |cx| { lint_callback!(cx, check_expr, e); ast_visit::walk_expr(cx, e); }) } (where walk_expr recursively calls back into visit_expr). The crash happens at a nesting depth of over 17000 stack frames when using the default 8 MB stack size on s390x. This patch fixes the problem by adding a ensure_sufficient_stack call to the with_lint_attrs routine (which also should take care of all the other mutually recursive visitors here). Fixes part of rust-lang/rust#105383.
Fix stack overflow in recursive AST walk in early lint The src/test/ui/issues/issue-74564-if-expr-stack-overflow.rs test case added to verify rust-lang/rust#74564 still crashes with a stack overflow on s390x-ibm-linux. Symptom is a very deep recursion in compiler/rustc_lint/src/early.rs: fn visit_expr(&mut self, e: &'a ast::Expr) { self.with_lint_attrs(e.id, &e.attrs, |cx| { lint_callback!(cx, check_expr, e); ast_visit::walk_expr(cx, e); }) } (where walk_expr recursively calls back into visit_expr). The crash happens at a nesting depth of over 17000 stack frames when using the default 8 MB stack size on s390x. This patch fixes the problem by adding a ensure_sufficient_stack call to the with_lint_attrs routine (which also should take care of all the other mutually recursive visitors here). Fixes part of rust-lang/rust#105383.
A number of mir-opt test cases are failing again on big-endian platforms, due to endian-specific binary encodings in the output files that are being checked by the test. To fix those, one of the following two strategies is used: a) Where a binary encoded value directly originates from some immediate constant in the test case source code, change that constant to an endian-invariant value (like 0x11000011). b) Where that is not easily possible or would change the essence of what the test is checking for, disable the test on big-endian platforms. Fixes (again) rust-lang#105383.
Running the
./x.py test
suite ons390x
currently results in a number of failing tests due to various issues. I've gone through and fixed all of those, resulting in a clean run with all my local fixes applied. Opening this issue to track getting all requires fixes upstream.Specifically, I'm seeing the following failures caused by actual code problems:
This is already described in run-make-fulldeps/c-link-to-rust-va-list-fn fails on s390x #84628, and causes the following failures:
src/test/ui/abi/variadic-ffi.rs
src/test/run-make-fulldeps/c-link-to-rust-va-list-fn
I've opened Lint suggestions depend on host endianness #105379 to track this specifically. This causes failure of:
src/test/rustdoc-ui/unknown-renamed-lints.rs
The test case described in Stack overflow in rustc, regression from 1.44 #74564 still fails on s390x even though the original fixes are applied. This is probably due to the larger stack requirements on the platforms. Can be fixed by adding another call to
ensure_sufficient_stack
. This causes failure of:src/test/ui/issues/issue-74564-if-expr-stack-overflow.rs
In addition, a number of test cases show (false positive) failures due to assumptions that aren't correct for the platform:
src/test/run-make/issue-71519
(Test needs to be ignored.)
src/doc/unstable-book/src/language-features/abi-efiapi.md
(Test needs to be ignored.)
src/test/codegen/catch-unwind.rs
src/test/codegen/remap_path_prefix/main.rs
src/test/codegen/repr-transparent-aggregates-1.rs
src/test/codegen/repr-transparent.rs
src/test/codegen/uninit-consts.rs
(I have fixes for these.)
src/test/ui/const-ptr/forbidden_slices.rs
src/test/ui/consts/const-eval/ub-enum.rs
src/test/ui/consts/const-eval/ub-nonnull.rs
src/test/ui/consts/const-eval/ub-ref-ptr.rs
src/test/ui/consts/const-eval/ub-uninhabit.rs
src/test/ui/consts/const-eval/ub-wide-ptr.rs
src/test/ui/consts/issue-83182.rs
src/test/ui/consts/std/alloc.rs
src/test/ui/consts/validate_never_arrays.rs
Some (but not all) of these are listed in 1.53.0 broke ui/consts/const-eval/ub-*.rs tests on s390x (big-endian, 64-bit) (regression) #89577, with various solutions being proposed. Options would be to either ignore those on big-endian hosts, provide multiple versions of the output to match against, and/or try to handle big- and little-endian outputs within the same file (e.g. via some transformation).
src/test/mir-opt/const_prop/mutable_variable_no_prop.rs
src/test/mir-opt/issues/issue_75439.rs
src/test/mir-opt/building/custom/consts.rs
Longer term, I'm wondering what the best way would be to ensure that the test suite remains clean on the platform - can this be included in CI somehow (either via a native machine somewhere or via qemu-based testing)?
CC @cuviper
The text was updated successfully, but these errors were encountered: