Author: Michael Liao Date: 2020-12-14T13:08:13-05:00 New Revision: 1fd1f638b68ca4a9cf3bd071a7cba0bac0b189c6
URL: https://github.com/llvm/llvm-project/commit/1fd1f638b68ca4a9cf3bd071a7cba0bac0b189c6 DIFF: https://github.com/llvm/llvm-project/commit/1fd1f638b68ca4a9cf3bd071a7cba0bac0b189c6.diff LOG: [amdgpu] Fix a crash case when `V_CNDMASK` could be simplified. - Once an instruction is simplified, foldable candidates from it should be invalidated or skipped as the operand index is no longer valid. Differential Revision: https://reviews.llvm.org/D93174 Added: llvm/test/CodeGen/AMDGPU/fold-cndmask-wave32.mir Modified: llvm/lib/Target/AMDGPU/SIFoldOperands.cpp Removed: ################################################################################ diff --git a/llvm/lib/Target/AMDGPU/SIFoldOperands.cpp b/llvm/lib/Target/AMDGPU/SIFoldOperands.cpp index 0df893ede7b8..bfba432848d4 100644 --- a/llvm/lib/Target/AMDGPU/SIFoldOperands.cpp +++ b/llvm/lib/Target/AMDGPU/SIFoldOperands.cpp @@ -1257,8 +1257,11 @@ void SIFoldOperands::foldInstOperand(MachineInstr &MI, for (MachineInstr *Copy : CopiesToReplace) Copy->addImplicitDefUseOperands(*MF); + SmallPtrSet<MachineInstr *, 16> Folded; for (FoldCandidate &Fold : FoldList) { assert(!Fold.isReg() || Fold.OpToFold); + if (Folded.count(Fold.UseMI)) + continue; if (Fold.isReg() && Fold.OpToFold->getReg().isVirtual()) { Register Reg = Fold.OpToFold->getReg(); MachineInstr *DefMI = Fold.OpToFold->getParent(); @@ -1278,7 +1281,8 @@ void SIFoldOperands::foldInstOperand(MachineInstr &MI, LLVM_DEBUG(dbgs() << "Folded source from " << MI << " into OpNo " << static_cast<int>(Fold.UseOpNo) << " of " << *Fold.UseMI << '\n'); - tryFoldInst(TII, Fold.UseMI); + if (tryFoldInst(TII, Fold.UseMI)) + Folded.insert(Fold.UseMI); } else if (Fold.isCommuted()) { // Restoring instruction's original operand order if fold has failed. TII->commuteInstruction(*Fold.UseMI, false); diff --git a/llvm/test/CodeGen/AMDGPU/fold-cndmask-wave32.mir b/llvm/test/CodeGen/AMDGPU/fold-cndmask-wave32.mir new file mode 100644 index 000000000000..a921ce33682c --- /dev/null +++ b/llvm/test/CodeGen/AMDGPU/fold-cndmask-wave32.mir @@ -0,0 +1,20 @@ +# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py +# RUN: llc -march=amdgcn -mcpu=gfx1030 -run-pass si-fold-operands -verify-machineinstrs -o - %s | FileCheck %s + +--- +name: fold_cndmask +tracksRegLiveness: true +registers: +body: | + bb.0.entry: + ; CHECK-LABEL: name: fold_cndmask + ; CHECK: [[DEF:%[0-9]+]]:sreg_32_xm0_xexec = IMPLICIT_DEF + ; CHECK: [[S_MOV_B32_:%[0-9]+]]:sreg_32 = S_MOV_B32 0 + ; CHECK: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec + ; CHECK: [[COPY:%[0-9]+]]:vgpr_32 = COPY [[S_MOV_B32_]] + %0:sreg_32_xm0_xexec = IMPLICIT_DEF + %1:sreg_32 = S_MOV_B32 0 + %2:vgpr_32 = COPY %1:sreg_32 + %3:vgpr_32 = V_CNDMASK_B32_e64 0, %1:sreg_32, 0, %2:vgpr_32, %0:sreg_32_xm0_xexec, implicit $exec + +... _______________________________________________ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
