AMDGPU/SI: Remove source uses of VCCReg

llvm-svn: 244379

1 files changed, 32 insertions, 11 deletions

diff --git a/llvm/lib/Target/AMDGPU/SIShrinkInstructions.cpp b/llvm/lib/Target/AMDGPU/SIShrinkInstructions.cpp
index 5d00bdd6..720b323 100644
--- a/llvm/lib/Target/AMDGPU/SIShrinkInstructions.cpp
+++ b/llvm/lib/Target/AMDGPU/SIShrinkInstructions.cpp
@@ -187,6 +187,21 @@ static void foldImmediates(MachineInstr &MI, const SIInstrInfo *TII,
 
 }
 
+// Copy MachineOperand with all flags except setting it as implicit.
+static MachineOperand copyRegOperandAsImplicit(const MachineOperand &Orig) {
+  assert(!Orig.isImplicit());
+  return MachineOperand::CreateReg(Orig.getReg(),
+                                   Orig.isDef(),
+                                   true,
+                                   Orig.isKill(),
+                                   Orig.isDead(),
+                                   Orig.isUndef(),
+                                   Orig.isEarlyClobber(),
+                                   Orig.getSubReg(),
+                                   Orig.isDebug(),
+                                   Orig.isInternalRead());
+}
+
 bool SIShrinkInstructions::runOnMachineFunction(MachineFunction &MF) {
   MachineRegisterInfo &MRI = MF.getRegInfo();
   const SIInstrInfo *TII =
@@ -236,14 +251,10 @@ bool SIShrinkInstructions::runOnMachineFunction(MachineFunction &MF) {
       if (TII->isVOPC(Op32)) {
         unsigned DstReg = MI.getOperand(0).getReg();
         if (TargetRegisterInfo::isVirtualRegister(DstReg)) {
-          // VOPC instructions can only write to the VCC register.  We can't
-          // force them to use VCC here, because the register allocator has
-          // trouble with sequences like this, which cause the allocator to run
-          // out of registers if vreg0 and vreg1 belong to the VCCReg register
-          // class:
-          // vreg0 = VOPC;
-          // vreg1 = VOPC;
-          // S_AND_B64 vreg0, vreg1
+          // VOPC instructions can only write to the VCC register. We can't
+          // force them to use VCC here, because this is only one register and
+          // cannot deal with sequences which would require multiple copies of
+          // VCC, e.g. S_AND_B64 (vcc = V_CMP_...), (vcc = V_CMP_...)
           //
           // So, instead of forcing the instruction to write to VCC, we provide
           // a hint to the register allocator to use VCC and then we we will run
@@ -288,9 +299,19 @@ bool SIShrinkInstructions::runOnMachineFunction(MachineFunction &MF) {
         Inst32.addOperand(*Src1);
 
       const MachineOperand *Src2 =
-          TII->getNamedOperand(MI, AMDGPU::OpName::src2);
-      if (Src2)
-        Inst32.addOperand(*Src2);
+        TII->getNamedOperand(MI, AMDGPU::OpName::src2);
+      if (Src2) {
+        int Op32Src2Idx = AMDGPU::getNamedOperandIdx(Op32, AMDGPU::OpName::src2);
+        if (Op32Src2Idx != -1) {
+          Inst32.addOperand(*Src2);
+        } else {
+          // In the case of V_CNDMASK_B32_e32, the explicit operand src2 is
+          // replaced with an implicit read of vcc.
+          assert(Src2->getReg() == AMDGPU::VCC &&
+                 "Unexpected missing register operand");
+          Inst32.addOperand(copyRegOperandAsImplicit(*Src2));
+        }
+      }
 
       ++NumInstructionsShrunk;
       MI.eraseFromParent();