projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: bfce0c2) | patch
AMDGPU/GlobalISel: Assume VGPR for G_FRAME_INDEX
authorMatt Arsenault <Matthew.Arsenault@amd.com>
Wed, 2 Oct 2019 01:02:24 +0000 (01:02 +0000)
committerMatt Arsenault <Matthew.Arsenault@amd.com>
Wed, 2 Oct 2019 01:02:24 +0000 (01:02 +0000)
commitcdfe5efe9b3c2b132b221f5099892c47339d035b
tree9e6519eb1c5a51db601db44d9c18a05d4852c467tree | snapshot
parentbfce0c2664dec15f9bd957ddc5ef25e13cd52fd9commit | diff
AMDGPU/GlobalISel: Assume VGPR for G_FRAME_INDEX

In principle this should behave as any other constant. However
eliminateFrameIndex currently assumes a VALU use and uses a vector
shift. Work around this by selecting to VGPR for now until
eliminateFrameIndex is fixed.

llvm-svn: 373415
llvm/lib/Target/AMDGPU/AMDGPURegisterBankInfo.cpp diff | blob | history
llvm/test/CodeGen/AMDGPU/GlobalISel/regbankselect-frame-index.mir diff | blob | history
Domain: System / Toolchain;