[llvm-dev] [HWASAN] Is Buildbot missing hwasan tests? (original) (raw)

Chandler Carruth via llvm-dev llvm-dev at lists.llvm.org
Wed Jan 2 13:38:49 PST 2019

+Evgenii Stepanov <eugenis at google.com>

On Wed, Jan 2, 2019 at 10:42 AM David Greene via llvm-dev < llvm-dev at lists.llvm.org> wrote:

After updating from trunk today, I am seeing this failure in hwasan:

FAIL: HWAddressSanitizer-x8664 :: TestCases/sizes.cpp (19011 of 49508) ******************** TEST 'HWAddressSanitizer-x8664 :: TestCases/sizes.cpp' FAILED ******************** Command Output (stderr): -- + : 'RUN: at line 1' + /build/./bin/clang --driver-mode=g++ -fsanitize=hwaddress -mllvm -hwasan-generate-tags-with-calls -m64 -gline-tables-only /compiler-rt/test/hwasan/TestCases/sizes.cpp -lstdc++ -o /build/projects/compiler-rt/test/hwasan/X8664/TestCases/Output/sizes.cpp.tmp /install/binutils/x8664/bin/ld.gold: warning: Cannot export local symbol _'hwasanthreadenter' /install/binutils/x8664/bin/ld.gold: warning: Cannot export local symbol _'hwasanthreadexit' /install/binutils/x8664/bin/ld.gold: warning: Cannot export local symbol _'hwasantls' /tmp/littmpGtMlYc/sizes-acae44.o:sizes.cpp:function main: error: _undefined reference to 'hwasanmemset' /tmp/littmpGtMlYc/sizes-acae44.o:sizes.cpp:function main: error: _undefined reference to 'hwasanmemset' clang-8: error: linker command failed with exit code 1 (use -v to see invocation) I looked through the buildbots and I don't see anything that runs hwasan tests. As far as I can tell, the hwasan tests are only run as part of check-all when compiler-rt is included in a top-level LLVM build or as part of check-hwasan. No clang or llvm builder include compiler-rt AFAICT, so check-all doesn't test it. I don't see check-hwasan in this builder, or any of the other sanitizer builders I checked: http://lab.llvm.org:8011/builders/sanitizer-x8664-linux/builds/18677 Is there a hole in our testing? Anyone else seeing this failure? Any idea of the cause? "hwasanmemset" doesn't appear anywhere in compiler-rt's history, but it shows up in LLVM: commit 631b5bc6125d979804e3e563935a93e2577e617c Author: Eugene Leviant <eleviant at accesssoftek.com> Date: Thu Dec 20 09:04:33 2018 +0000 [HWASAN] Add support for memory intrinsics Differential revision: https://reviews.llvm.org/D55117

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@349728 91177308-0d34-0410-b5e6-96231b3b80d8 This commit turned on memory intrinsic instrumentation by default and is the culprit, I'm guessing: commit 73121b2cc825d26b849e5c6fa04c690a6a35fb6e Author: Eugene Leviant <eleviant at accesssoftek.com> Date: Mon Dec 24 16:02:48 2018 +0000 [HWASAN] Instrument memorty intrinsics by default Differential revision: https://reviews.llvm.org/D55926 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@350055 91177308-0d34-0410-b5e6-96231b3b80d8 -David

LLVM Developers mailing list llvm-dev at lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20190102/9d462864/attachment.html>

More information about the llvm-dev mailing list