diff options
| author | Joseph Huber <huberjn@outlook.com> | 2025-01-06 08:11:08 -0600 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-01-06 08:11:08 -0600 |
| commit | 81fae0d5e3d3378959483ccd7709a212731bff3f (patch) | |
| tree | 0d9373d6e2571502c37020e4c8c601a00ae388c2 /clang/lib/CodeGen/CodeGenFunction.cpp | |
| parent | 3f7905733820851bc4f65cb4af693c3101cbf20d (diff) | |
| download | llvm-81fae0d5e3d3378959483ccd7709a212731bff3f.zip llvm-81fae0d5e3d3378959483ccd7709a212731bff3f.tar.gz llvm-81fae0d5e3d3378959483ccd7709a212731bff3f.tar.bz2 | |
[Clang][AMDGPU] Stop defaulting to `one-as` for all atomic scopes (#120095)
Summary:
The documentation at
https://llvm.org/docs/AMDGPUUsage.html#memory-scopes states that these
'one-as' modifiers are more specific versions of the scopes that only
apply to a specific address space. This doesn't make sense for fences
which have no associated address space to use, and it's a more
restrictive version the normal scope. This should not tbe the default
behavior, but it is currently emitted in all cases except for
sequentially consistent.
Diffstat (limited to 'clang/lib/CodeGen/CodeGenFunction.cpp')
0 files changed, 0 insertions, 0 deletions