[llvm-branch-commits] [compiler-rt] [LSan] Mark create_thread_leak.cpp as UNSUPPORTED: darwin. (PR #69187)

Henrik G. Olsson via llvm-branch-commits llvm-branch-commits at lists.llvm.org
Mon Oct 16 03:36:30 PDT 2023


https://github.com/hnrklssn created https://github.com/llvm/llvm-project/pull/69187

This test failed in https://green.lab.llvm.org/green/job/clang-stage1-RA-release-17.x/63/ but not in https://green.lab.llvm.org/green/job/clang-stage1-RA-release-17.x/64/, which uses the same checkout of the LLVM repo. The only difference is https://github.com/llvm/llvm-zorg/commit/0344f4509a281a2b343da19fc4d8523921d33717 which is an llvm-zorg commit and doesn’t seem related. Cherry-picking this to prevent flaky test failures on the release branch:

```
It started to fail in a flaky manner a few days ago on GreenDragon buildbots (i.e. x86_64-darwin). I didn't track down the root cause but LSan isn't actually supported on darwin anyway, so UNSUPPORTED seems appropriate.

Prior art: 3ff080b5.

(cherry picked from commit 0a3519d5a27b9400250b5f6656b50148021e7496)
```

>From b9a64e3d5f5f7dec55a5ff6fc47df4ad3fc8900a Mon Sep 17 00:00:00 2001
From: Artem Dergachev <adergachev at apple.com>
Date: Thu, 31 Aug 2023 16:36:38 -0700
Subject: [PATCH] [LSan] Mark create_thread_leak.cpp as UNSUPPORTED: darwin.

It started to fail in a flaky manner a few days ago on GreenDragon buildbots
(i.e. x86_64-darwin). I didn't track down the root cause but LSan isn't
actually supported on darwin anyway, so UNSUPPORTED seems appropriate.

Prior art: 3ff080b5.

(cherry picked from commit 0a3519d5a27b9400250b5f6656b50148021e7496)
---
 compiler-rt/test/lsan/TestCases/create_thread_leak.cpp | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/compiler-rt/test/lsan/TestCases/create_thread_leak.cpp b/compiler-rt/test/lsan/TestCases/create_thread_leak.cpp
index 92e70b2e37c3dd4..564d4b66384e8aa 100644
--- a/compiler-rt/test/lsan/TestCases/create_thread_leak.cpp
+++ b/compiler-rt/test/lsan/TestCases/create_thread_leak.cpp
@@ -6,6 +6,9 @@
 // RUN: %run not %t 10 0 0 1 2>&1 | FileCheck %s --check-prefixes=LEAK,LEAK234
 // RUN: %run %t 10 0 0 0
 
+// This test appears to be flaky on x86_64-darwin buildbots.
+// UNSUPPORTED: darwin
+
 #include <pthread.h>
 #include <stdlib.h>
 



More information about the llvm-branch-commits mailing list