projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: cc4b692) | patch
AMDGPU: Fix offsets for < 4-byte aggregate kernel arguments
authorMatt Arsenault <Matthew.Arsenault@amd.com>
Fri, 7 Dec 2018 22:12:17 +0000 (22:12 +0000)
committerMatt Arsenault <Matthew.Arsenault@amd.com>
Fri, 7 Dec 2018 22:12:17 +0000 (22:12 +0000)
commitb5613ecf173d5eeca82dce3be2b269feb4a75082
tree70a8e65b9ca1fe8a54cdf817eb39e29b0631d7e5tree | snapshot
parentcc4b6920b36f81ff3a4370c6646de07e268e838bcommit | diff
AMDGPU: Fix offsets for < 4-byte aggregate kernel arguments

We were still using the rounded down offset and alignment even though
they aren't handled because you can't trivially bitcast the loaded
value.

llvm-svn: 348658
llvm/lib/Target/AMDGPU/AMDGPULowerKernelArguments.cpp diff | blob | history
llvm/test/CodeGen/AMDGPU/kernel-args.ll diff | blob | history
Domain: System / Toolchain;