[llvm] 52df4a1 - [AMDGPU] Fix typos "SPGR" / "VPGR" in comments
Jay Foad via llvm-commits
llvm-commits at lists.llvm.org
Mon Mar 2 03:50:02 PST 2026
Author: Jay Foad
Date: 2026-03-02T11:48:38Z
New Revision: 52df4a19599b9b51bc8ac7c612e979e361aded93
URL: https://github.com/llvm/llvm-project/commit/52df4a19599b9b51bc8ac7c612e979e361aded93
DIFF: https://github.com/llvm/llvm-project/commit/52df4a19599b9b51bc8ac7c612e979e361aded93.diff
LOG: [AMDGPU] Fix typos "SPGR" / "VPGR" in comments
Added:
Modified:
llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp
Removed:
################################################################################
diff --git a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
index 41608cfb3457b..7b548a43df422 100644
--- a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
@@ -7728,7 +7728,7 @@ bool SIInstrWorklist::isDeferred(MachineInstr *MI) {
}
// Legalize size mismatches between 16bit and 32bit registers in v2s copy
-// lowering (change spgr to vgpr).
+// lowering (change sgpr to vgpr).
// This is mainly caused by 16bit SALU and 16bit VALU using reg with
diff erent
// size. Need to legalize the size of the operands during the vgpr lowering
// chain. This can be removed after we have sgpr16 in place
diff --git a/llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp b/llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp
index cbd08f0fb5dff..196e551932659 100644
--- a/llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp
+++ b/llvm/lib/Target/AMDGPU/SILowerSGPRSpills.cpp
@@ -302,7 +302,7 @@ bool SILowerSGPRSpills::spillCalleeSavedRegs(
void SILowerSGPRSpills::updateLaneVGPRDomInstr(
int FI, MachineBasicBlock *MBB, MachineBasicBlock::iterator InsertPt,
DenseMap<Register, MachineBasicBlock::iterator> &LaneVGPRDomInstr) {
- // For the Def of a virtual LaneVPGR to dominate all its uses, we should
+ // For the Def of a virtual LaneVGPR to dominate all its uses, we should
// insert an IMPLICIT_DEF before the dominating spill. Switching to a
// depth first order doesn't really help since the machine function can be in
// the unstructured control flow post-SSA. For each virtual register, hence
More information about the llvm-commits
mailing list