intel/llvm
1
0
Fork 0
mirror of https://github.com/intel/llvm.git synced 2026-01-13 11:02:04 +08:00
Code Issues Packages Projects Releases Wiki Activity

AMDGPU: Stop using the wavemask register class for SCC cross class copies (#161801)

Browse Source 
SCC should be copied to a 32-bit SGPR. Using a wave mask doesn't make
sense.
This commit is contained in:
Matt Arsenault
2025-10-07 00:44:45 +09:00
committed by GitHub
parent 6620e53511
commit 1cc9a8c127
1 changed files with 1 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp
View File

@@ -1118,9 +1118,7 @@ SIRegisterInfo::getPointerRegClass(unsigned Kind) const {
const TargetRegisterClass *
SIRegisterInfo::getCrossCopyRegClass(const TargetRegisterClass *RC) const {
if (RC == &AMDGPU::SCC_CLASSRegClass)
return getWaveMaskRegClass();
return RC;
return RC == &AMDGPU::SCC_CLASSRegClass ? &AMDGPU::SReg_32RegClass : RC;
}
static unsigned getNumSubRegsForSpillOp(const MachineInstr &MI,