diff options
author | Harald van Dijk <harald@gigawatt.nl> | 2024-03-13 12:08:39 +0000 |
---|---|---|
committer | GitHub <noreply@github.com> | 2024-03-13 12:08:39 +0000 |
commit | ceb744eb2fa0895db1526110462745962fdf43c0 (patch) | |
tree | ceed3689eeec4ced7b35f2e28b1bad4cdee272e0 /clang/lib/CodeGen/CodeGenModule.cpp | |
parent | 7cd61f888c479da51215071336b34f6918cad3d8 (diff) | |
download | llvm-ceb744eb2fa0895db1526110462745962fdf43c0.zip llvm-ceb744eb2fa0895db1526110462745962fdf43c0.tar.gz llvm-ceb744eb2fa0895db1526110462745962fdf43c0.tar.bz2 |
[AMDGPU] Fix canonicalization of truncated values. (#83054)
We were relying on roundings to implicitly canonicalize, which is
generally safe, except with roundings that may be optimized away.
Fixes #82937.
Diffstat (limited to 'clang/lib/CodeGen/CodeGenModule.cpp')
0 files changed, 0 insertions, 0 deletions