[NVPTX] Move NVPTXPeephole after NVPTXPrologEpilogPass

Summary:
Offset of frame index is calculated by NVPTXPrologEpilogPass. Before
that the correct offset of stack objects cannot be obtained, which
leads to wrong offset if there are more than 2 frame objects. This patch
move NVPTXPeephole after NVPTXPrologEpilogPass. Because the frame index
is already replaced by %VRFrame in NVPTXPrologEpilogPass, we check
VRFrame register instead, and try to remove the VRFrame if there
is no usage after NVPTXPeephole pass.

Patched by Xuetian Weng. 

Test Plan:
Strengthened test/CodeGen/NVPTX/local-stack-frame.ll to check the
offset calculation based on SP and SPL.

Reviewers: jholewinski, jingyue

Reviewed By: jingyue

Subscribers: jholewinski, llvm-commits

Differential Revision: http://reviews.llvm.org/D10853

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@241185 91177308-0d34-0410-b5e6-96231b3b80d8

lib/Target/NVPTX/NVPTXPeephole.cpp

@@ -22,7 +22,7 @@
 // This peephole pass optimizes these cases, for example
 //
 // It will transform the following pattern
-//    %vreg0<def> = LEA_ADDRi64 <fi#0>, 4
+//    %vreg0<def> = LEA_ADDRi64 %VRFrame, 4
 //    %vreg1<def> = cvta_to_local_yes_64 %vreg0
 //
 // into
@@ -36,7 +36,6 @@
 #include "llvm/CodeGen/MachineFunctionPass.h"
 #include "llvm/CodeGen/MachineInstrBuilder.h"
 #include "llvm/CodeGen/MachineRegisterInfo.h"
-#include "llvm/CodeGen/MachineFrameInfo.h"
 #include "llvm/Target/TargetRegisterInfo.h"
 #include "llvm/Target/TargetInstrInfo.h"
 
@@ -96,7 +95,7 @@ static bool isCVTAToLocalCombinationCandidate(MachineInstr &Root) {
 
   // Check the LEA_ADDRi operand is Frame index
   auto &BaseAddrOp = GenericAddrDef->getOperand(1);
-  if (BaseAddrOp.getType() == MachineOperand::MO_FrameIndex) {
+  if (BaseAddrOp.isReg() && BaseAddrOp.getReg() == NVPTX::VRFrame) {
     return true;
   }
 
@@ -110,16 +109,11 @@ static void CombineCVTAToLocal(MachineInstr &Root) {
   const TargetInstrInfo *TII = MF.getSubtarget().getInstrInfo();
   auto &Prev = *MRI.getUniqueVRegDef(Root.getOperand(1).getReg());
 
-  // Get the correct offset
-  int FrameIndex = Prev.getOperand(1).getIndex();
-  int Offset = MF.getFrameInfo()->getObjectOffset(FrameIndex) +
-               Prev.getOperand(2).getImm();
-
   MachineInstrBuilder MIB =
       BuildMI(MF, Root.getDebugLoc(), TII->get(Prev.getOpcode()),
               Root.getOperand(0).getReg())
           .addReg(NVPTX::VRFrameLocal)
-          .addOperand(MachineOperand::CreateImm(Offset));
+          .addOperand(Prev.getOperand(2));
 
   MBB.insert((MachineBasicBlock::iterator)&Root, MIB);
 
@@ -145,6 +139,15 @@ bool NVPTXPeephole::runOnMachineFunction(MachineFunction &MF) {
       }
     }  // Instruction
   }    // Basic Block
+
+  // Remove unnecessary %VRFrame = cvta.local %VRFrameLocal
+  const auto &MRI = MF.getRegInfo();
+  if (MRI.use_empty(NVPTX::VRFrame)) {
+    if (auto MI = MRI.getUniqueVRegDef(NVPTX::VRFrame)) {
+      MI->eraseFromParentAndMarkDBGValuesForRemoval();
+    }
+  }
+
   return Changed;
 }
 

lib/Target/NVPTX/NVPTXTargetMachine.cpp

@@ -205,13 +205,15 @@ bool NVPTXPassConfig::addInstSelector() {
   if (!ST.hasImageHandles())
     addPass(createNVPTXReplaceImageHandlesPass());
 
-  addPass(createNVPTXPeephole());
-
   return false;
 }
 
 void NVPTXPassConfig::addPostRegAlloc() {
   addPass(createNVPTXPrologEpilogPass(), false);
+  // NVPTXPrologEpilogPass calculates frame object offset and replace frame
+  // index with VRFrame register. NVPTXPeephole need to be run after that and
+  // will replace VRFrame with VRFrameLocal when possible.
+  addPass(createNVPTXPeephole());
 }
 
 FunctionPass *NVPTXPassConfig::createTargetRegisterAllocator(bool) {

test/CodeGen/NVPTX/local-stack-frame.ll

@@ -59,10 +59,16 @@ define void @foo3(i32 %a) {
 
 ; PTX32:        cvta.local.u32   %SP, %SPL;
 ; PTX32:        add.u32          {{%r[0-9]+}}, %SP, 0;
+; PTX32:        add.u32          {{%r[0-9]+}}, %SPL, 0;
+; PTX32:        add.u32          {{%r[0-9]+}}, %SP, 4;
+; PTX32:        add.u32          {{%r[0-9]+}}, %SPL, 4;
 ; PTX32:        st.local.u32     [{{%r[0-9]+}}], {{%r[0-9]+}}
 ; PTX32:        st.local.u32     [{{%r[0-9]+}}], {{%r[0-9]+}}
 ; PTX64:        cvta.local.u64   %SP, %SPL;
 ; PTX64:        add.u64          {{%rd[0-9]+}}, %SP, 0;
+; PTX64:        add.u64          {{%rd[0-9]+}}, %SPL, 0;
+; PTX64:        add.u64          {{%rd[0-9]+}}, %SP, 4;
+; PTX64:        add.u64          {{%rd[0-9]+}}, %SPL, 4;
 ; PTX64:        st.local.u32     [{{%rd[0-9]+}}], {{%r[0-9]+}}
 ; PTX64:        st.local.u32     [{{%rd[0-9]+}}], {{%r[0-9]+}}
 define void @foo4() {