[ubsan] Disable sigaction.cpp on ubsan-tsan

[Caslyn]

Summary

Following https://reviews.llvm.org/D155272, restrict to FAIL to ubsan-tsan tests so non-ubsan-tsan tests may still pass, shown by https://lab.llvm.org/buildbot/#/builders/179/builds/6669.

[Caslyn]

compiler-rt/test/ubsan/TestCases/Misc/Linux/sigaction.cpp

8

Build failure examples here:

clang-linux-x64:

• UBSan-ThreadSanitizer-x86_64 :: TestCases/Misc/Linux/sigaction.cpp

clang-linux-arm64 (Not sure why these are continuing to fail after D155272):

• UBSan-AddressSanitizer-aarch64 :: TestCases/Misc/Linux/sigaction.cpp
• UBSan-MemorySanitizer-aarch64 :: TestCases/Misc/Linux/sigaction.cpp
• UBSan-Standalone-aarch64 :: TestCases/Misc/Linux/sigaction.cpp

[Caslyn]

Amend XFAIL directive

[Caslyn]

Amend XFAIL to include ubsan-tsan

[paulkirth]

compiler-rt/test/ubsan/TestCases/Misc/Linux/sigaction.cpp

8

do you intend to only disable for aarch64 w/ ubsan + tsan?

[Caslyn]

compiler-rt/test/ubsan/TestCases/Misc/Linux/sigaction.cpp

8

Added ubsan-tsan to disqualifier condition to permit passing cases in https://lab.llvm.org/buildbot/#/builders/179/builds/6669:

• UBSan-AddressSanitizer-aarch64 :: TestCases/Misc/Linux/sigaction.cpp
• UBSan-MemorySanitizer-aarch64 :: TestCases/Misc/Linux/sigaction.cpp
• UBSan-Standalone-aarch64 :: TestCases/Misc/Linux/sigaction.cpp

8

Yes, so that aarch64 buliders may at least pass. It remains unclear how to handle the Linux-64 case for now, since it's not fully apparent that it impacts upstream builders.

[paulkirth]

https://llvm.org/docs/TestingGuide.html#constraining-test-execution is helpful for figuring out the constraints, as I think the above change enables the test on more targets, instead of on all aarch64 or ubsan-tsan.

you probably want XFAIL: aarch64{{.*}, ubsan-tsan

[paulkirth]

ah, my mistake. feel free to ignore my following comment

[hctim]

Hey @Caslyn, looks like this unfortunately didn't fix the buildbot breakage caused by https://reviews.llvm.org/D155272, I still see UBSan-ThreadSanitizer-aarch64 :: TestCases/Misc/Linux/sigaction.cpp broken at head (https://lab.llvm.org/buildbot/#/builders/238/builds/4398).

[dvyukov]

Unexpectedly Passed Tests (1):
UBSan-ThreadSanitizer-aarch64 :: TestCases/Misc/Linux/sigaction.cpp

Do we just disable it instead of XFAIL?

[mysterymath]

In D155317#4505039, @dvyukov wrote:

Unexpectedly Passed Tests (1):
UBSan-ThreadSanitizer-aarch64 :: TestCases/Misc/Linux/sigaction.cpp

Do we just disable it instead of XFAIL?

I think so. We can look into reincorporating this test into that configuration once it's been determined why it fails on some builders and passes on others.

compiler-rt/test/ubsan/TestCases/Misc/Linux/sigaction.cpp

 // RUN: %clangxx -fsanitize=undefined -shared-libsan %s -o %t && %run %t 2>&1 | FileCheck %s
 
 // Ensure ubsan runtime/interceptors are lazily initialized if called early.
 
 // The test seems to segfault on aarch64 with tsan:
 // https://lab.llvm.org/buildbot/#/builders/179/builds/6662
 // Reason unknown, needs debugging.
-// XFAIL: target=aarch64{{.*}}
+// XFAIL: target=aarch64{{.*}} && ubsan-tsan
 
 #include <assert.h>
 #include <signal.h>
 #include <stdio.h>
 
 __attribute__((constructor(0))) void ctor() {
   fprintf(stderr, "INIT\n");
   struct sigaction old;