Skip to content

new asan/failures failures since compiler-rt/libcxx update to llvm-12. #14360

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

Closed
sbc100 opened this issue Jun 3, 2021 · 0 comments · Fixed by #14371
Closed

new asan/failures failures since compiler-rt/libcxx update to llvm-12. #14360

sbc100 opened this issue Jun 3, 2021 · 0 comments · Fixed by #14371
Assignees

Comments

@sbc100
Copy link
Collaborator

sbc100 commented Jun 3, 2021

From #14280

"It looks several asan/lsan tests started to break after this. Starting commit: https://ci.chromium.org/ui/p/emscripten-releases/builders/ci/linux-test-suites/b8846238787694172272/overview

@sbc100 sbc100 self-assigned this Jun 3, 2021
sbc100 added a commit that referenced this issue Jun 3, 2021
This test leaks allocates the asyncify stack during the exit handlers and
its never freed.

Interestingly enough this leak was not caught prior to #14280.
I did some investigation and found that prior to #14280 this
chunk was bring detected a reachable via the conservative stack
scan here:

 https://github.com/emscripten-core/emscripten/blob/98a8cbf5032f8d369be636a59180a59f1eb8833f/system/lib/compiler-rt/lib/lsan/lsan_common_emscripten.cpp#L141-L153

If I comment out this stack scan then this leak will be
detected even before #14280

Also, add `--profiling` to asan and lsan builds to make
error reports more useful.

See #14360
sbc100 added a commit that referenced this issue Jun 3, 2021
The logic of this macro was inverted when the llvm-12 update to
compiler-rt (#14280).  This caused asan.test_setjmp_noleak to start
failing.

Its not clear it we want to keep this long term but restoring
it at least fixes the failing test.

See #14360
sbc100 added a commit that referenced this issue Jun 3, 2021
This test leaks allocates the asyncify stack during the exit handlers and
its never freed.

Interestingly enough this leak was not caught prior to #14280.
I did some investigation and found that prior to #14280 this
chunk was bring detected a reachable via the conservative stack
scan here:

 https://github.com/emscripten-core/emscripten/blob/98a8cbf5032f8d369be636a59180a59f1eb8833f/system/lib/compiler-rt/lib/lsan/lsan_common_emscripten.cpp#L141-L153

If I comment out this stack scan then this leak will be
detected even before #14280

Also, add `--profiling` to asan and lsan builds to make
error reports more useful.

See #14360
sbc100 added a commit that referenced this issue Jun 3, 2021
The logic of this macro was inverted when the llvm-12 update to
compiler-rt (#14280).  This caused asan.test_setjmp_noleak to start
failing.

Its not clear it we want to keep this long term but restoring
it at least fixes the failing test.

See #14360
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

1 participant