llvm.org GIT mirror llvm / 03e1ccd
Merging r268287: ------------------------------------------------------------------------ r268287 | thomas.stellard | 2016-05-02 12:37:56 -0700 (Mon, 02 May 2016) | 19 lines AMDGPU/SI: Set the kill flag on temp VGPRs used to restore SGPRs from scratch Summary: When we restore an SGPR value from scratch, we first load it into a temporary VGPR and then use v_readlane_b32 to copy the value from the VGPR back into an SGPR. We weren't setting the kill flag on the VGPR in the v_readlane_b32 instruction, so the register scavenger wasn't able to re-use this temp value later. I wasn't able to create a lit test for this. Reviewers: arsenm Subscribers: arsenm, llvm-commits Differential Revision: http://reviews.llvm.org/D19744 ------------------------------------------------------------------------ git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_38@271770 91177308-0d34-0410-b5e6-96231b3b80d8 Tom Stellard 3 years ago
1 changed file(s) with 1 addition(s) and 1 deletion(s). Raw diff Collapse all Expand all
410410 .addMemOperand(MMO);
411411 BuildMI(*MBB, MI, DL,
412412 TII->getMCOpcodeFromPseudo(AMDGPU::V_READLANE_B32), SubReg)
413 .addReg(TmpReg)
413 .addReg(TmpReg, RegState::Kill)
414414 .addImm(0)
415415 .addReg(MI->getOperand(0).getReg(), RegState::ImplicitDefine);
416416 }