diff options
| author | Johannes Doerfert <johannes@jdoerfert.de> | 2022-04-11 13:32:22 -0500 |
|---|---|---|
| committer | Johannes Doerfert <johannes@jdoerfert.de> | 2022-04-12 16:42:50 -0500 |
| commit | 9dc7da3f9cb48ed7ba6c4806f4519d997f6671b5 (patch) | |
| tree | 633f39c8babf970965a9d8429a9350635929f57c /lldb/test/API/python_api/debugger/TestDebuggerAPI.py | |
| parent | 0f070bee8254678a3501ccb9a3f37933fe517ca7 (diff) | |
[GlobalsModRef][FIX] Ensure we honor synchronizing effects of intrinsics
This is a long standing problem that resurfaces once in a while [0].
There might actually be two problems because I'm not 100% sure if the
issue underlying https://reviews.llvm.org/D115302 would be solved by
this or not. Anyway.
In 2008 we thought intrinsics do not read/write globals passed to them:
https://github.com/llvm/llvm-project/commit/d4133ac31535ce5176f97e9fc81825af8a808760
This is not correct given that intrinsics can synchronize threads and
cause effects to effectively become visible.
NOTE: I did not yet modify any tests but only tried out the reproducer
of https://github.com/llvm/llvm-project/issues/54851.
Fixes: https://github.com/llvm/llvm-project/issues/54851
[0] https://discourse.llvm.org/t/bug-gvn-memdep-bug-in-the-presence-of-intrinsics/59402
Differential Revision: https://reviews.llvm.org/D123531
Diffstat (limited to 'lldb/test/API/python_api/debugger/TestDebuggerAPI.py')
0 files changed, 0 insertions, 0 deletions
