[AMDGPU] Prevent hang in SIFoldOperands by caching uses (#82099)

foldOperands() for REG_SEQUENCE has recursion that can trigger an infinite loop
as the method can modify the operand order, which messes up the range-based
for loop. This patch fixes the issue by caching the uses for processing beforehand,
and then iterating over the cache rather using the instruction iterator.

Author: choikwa

llvm/lib/Target/AMDGPU/SIFoldOperands.cpp

@@ -775,20 +775,23 @@ void SIFoldOperands::foldOperand(
     Register RegSeqDstReg = UseMI->getOperand(0).getReg();
     unsigned RegSeqDstSubReg = UseMI->getOperand(UseOpIdx + 1).getImm();
 
-    for (auto &RSUse : make_early_inc_range(MRI->use_nodbg_operands(RegSeqDstReg))) {
-      MachineInstr *RSUseMI = RSUse.getParent();
+    // Grab the use operands first
+    SmallVector<MachineOperand *, 4> UsesToProcess;
+    for (auto &Use : MRI->use_nodbg_operands(RegSeqDstReg))
+      UsesToProcess.push_back(&Use);
+    for (auto *RSUse : UsesToProcess) {
+      MachineInstr *RSUseMI = RSUse->getParent();
 
       if (tryToFoldACImm(UseMI->getOperand(0), RSUseMI,
-                         RSUseMI->getOperandNo(&RSUse), FoldList))
+                         RSUseMI->getOperandNo(RSUse), FoldList))
         continue;
 
-      if (RSUse.getSubReg() != RegSeqDstSubReg)
+      if (RSUse->getSubReg() != RegSeqDstSubReg)
         continue;
 
-      foldOperand(OpToFold, RSUseMI, RSUseMI->getOperandNo(&RSUse), FoldList,
+      foldOperand(OpToFold, RSUseMI, RSUseMI->getOperandNo(RSUse), FoldList,
                   CopiesToReplace);
     }
-
     return;
   }
 

llvm/test/CodeGen/AMDGPU/si-fold-reg-sequence.mir

@@ -0,0 +1,18 @@
+# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -run-pass=si-fold-operands -verify-machineinstrs -o - %s
+
+---
+name:            fold_reg_sequence
+body:             |
+  bb.0:
+    liveins: $vgpr0_vgpr1, $vgpr2
+
+    %0:sreg_32 = S_MOV_B32 0
+    %1:sreg_32 = S_MOV_B32 429
+    %2:sreg_64 = REG_SEQUENCE killed %1, %subreg.sub0, %0, %subreg.sub1
+    %3:vgpr_32 = V_MUL_HI_U32_e64 $vgpr2, %2.sub0, implicit $exec
+    %4:vgpr_32 = GLOBAL_LOAD_DWORD $vgpr0_vgpr1, 0, 0, implicit $exec :: (load (s32), addrspace 1)
+    %5:vgpr_32 = V_MUL_HI_U32_e64 %4, %2.sub0, implicit $exec
+    S_ENDPGM 0
+
+...
+