[AMDGPU] Fixed dpp combine of VOP1
authorStanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com>
Wed, 9 Oct 2019 22:02:58 +0000 (22:02 +0000)
committerStanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com>
Wed, 9 Oct 2019 22:02:58 +0000 (22:02 +0000)
commitc6dec1d8288cd78ca0d6032551f826ce7a5bbc4e
tree92809e6ad36adbe7f2e23be9085b372c6cb1c722
parent47363a148f1d3003f0dd1cce234b60cdc8cc48bd
[AMDGPU] Fixed dpp combine of VOP1

If original instruction did not have source modifiers they were
not added to the new DPP instruction as well, even if needed.

Differential Revision: https://reviews.llvm.org/D68729

llvm-svn: 374241
llvm/lib/Target/AMDGPU/GCNDPPCombine.cpp
llvm/test/CodeGen/AMDGPU/dpp_combine.mir