projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: ce38c2d) | patch
AMDGPU/R600: Implement allowsMisalignedMemoryAccess
authorMatt Arsenault <Matthew.Arsenault@amd.com>
Mon, 22 Feb 2016 21:04:16 +0000 (21:04 +0000)
committerMatt Arsenault <Matthew.Arsenault@amd.com>
Mon, 22 Feb 2016 21:04:16 +0000 (21:04 +0000)
commitfa67bdbde06f330a7d0c1ced7023951f6a238b8e
treece3b82f3e0f8b2f1c4e4aadb82aaf9ead7155e76tree | snapshot
parentce38c2ddf6a575f92d8d01871061fd0d4bc2f305commit | diff
AMDGPU/R600: Implement allowsMisalignedMemoryAccess

This avoids some test regressions in a future commit
when unaligned operations are expanded when they
have custom lowering.

llvm-svn: 261570
llvm/lib/Target/AMDGPU/R600ISelLowering.cpp diff | blob | history
llvm/lib/Target/AMDGPU/R600ISelLowering.h diff | blob | history
llvm/test/CodeGen/AMDGPU/store.ll diff | blob | history
Domain: System / Toolchain;