]> granicus.if.org Git - llvm/commit
AMDGPU: Fix creating invalid copy when adjusting dmask
authorMatt Arsenault <Matthew.Arsenault@amd.com>
Mon, 4 Dec 2017 22:18:27 +0000 (22:18 +0000)
committerMatt Arsenault <Matthew.Arsenault@amd.com>
Mon, 4 Dec 2017 22:18:27 +0000 (22:18 +0000)
commit79f2fee592d37f7a57f82bb31bb26e74440e7305
tree9ce763ed8308b5f1d6952524378f6c72d9420990
parentaaadfb2c37e10db72a11b7f625ed765e92222719
AMDGPU: Fix creating invalid copy when adjusting dmask

Move the entire optimization to one place. Before it was possible
to adjust dmask without changing the register class of the output
instruction, since they were done in separate places. Fix all
lane sizes and move all of the optimization into the DAG folding.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@319705 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
lib/Target/AMDGPU/AMDGPUInstrInfo.cpp
lib/Target/AMDGPU/AMDGPUInstrInfo.h
lib/Target/AMDGPU/SIISelLowering.cpp
lib/Target/AMDGPU/SIISelLowering.h
lib/Target/AMDGPU/SIInstrInfo.td
test/CodeGen/AMDGPU/adjust-writemask-invalid-copy.ll [new file with mode: 0644]