diff options
| author | Stanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com> | 2019-10-21 13:43:01 -0700 |
|---|---|---|
| committer | Stanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com> | 2019-10-23 18:42:48 -0700 |
| commit | 61e7a61bdccfae2c55e1ab598621204030feac7a (patch) | |
| tree | 71908d01bcba9e7b43121c52dfeb4e1f6a1f5767 /llvm/lib/Target | |
| parent | e3d26b42b95c658ac5805188291f4316b47a15dd (diff) | |
| download | bcm5719-llvm-61e7a61bdccfae2c55e1ab598621204030feac7a.tar.gz bcm5719-llvm-61e7a61bdccfae2c55e1ab598621204030feac7a.zip | |
[AMDGPU] Allow folding of sgpr to vgpr copy
Potentially sgpr to sgpr copy should also be possible.
That is however trickier because we may end up with a
wrong register class at use because of xm0/xexec permutations.
Differential Revision: https://reviews.llvm.org/D69280
Diffstat (limited to 'llvm/lib/Target')
| -rw-r--r-- | llvm/lib/Target/AMDGPU/SIFoldOperands.cpp | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/llvm/lib/Target/AMDGPU/SIFoldOperands.cpp b/llvm/lib/Target/AMDGPU/SIFoldOperands.cpp index 69ac367f7ab..a0e43101a7f 100644 --- a/llvm/lib/Target/AMDGPU/SIFoldOperands.cpp +++ b/llvm/lib/Target/AMDGPU/SIFoldOperands.cpp @@ -627,10 +627,11 @@ void SIFoldOperands::foldOperand( CopiesToReplace.push_back(UseMI); } else { if (UseMI->isCopy() && OpToFold.isReg() && - Register::isVirtualRegister(UseMI->getOperand(0).getReg()) && + UseMI->getOperand(0).getReg().isVirtual() && TRI->isVectorRegister(*MRI, UseMI->getOperand(0).getReg()) && - TRI->isVectorRegister(*MRI, UseMI->getOperand(1).getReg()) && !UseMI->getOperand(1).getSubReg()) { + LLVM_DEBUG(dbgs() << "Folding " << OpToFold + << "\n into " << *UseMI << '\n'); unsigned Size = TII->getOpSize(*UseMI, 1); UseMI->getOperand(1).setReg(OpToFold.getReg()); UseMI->getOperand(1).setSubReg(OpToFold.getSubReg()); |
