AMDGPU/SI: Reserve appropriate number of sgprs for flat scratch init.

[oota-llvm.git] / lib / Target / PowerPC / PPCFrameLowering.cpp

diff --git a/lib/Target/PowerPC/PPCFrameLowering.cpp b/lib/Target/PowerPC/PPCFrameLowering.cpp
index 60da158b058c518452ff0e682805e4af05005273..beab844c6025f424d16595228f7a28572db00928 100644 (file)
--- a/lib/Target/PowerPC/PPCFrameLowering.cpp
+++ b/lib/Target/PowerPC/PPCFrameLowering.cpp
@@ -30,7 +30,7 @@ using namespace llvm;
 
 /// VRRegNo - Map from a numbered VR register to its enum value.
 ///
-static const uint16_t VRRegNo[] = {
+static const MCPhysReg VRRegNo[] = {
  PPC::V0 , PPC::V1 , PPC::V2 , PPC::V3 , PPC::V4 , PPC::V5 , PPC::V6 , PPC::V7 ,
  PPC::V8 , PPC::V9 , PPC::V10, PPC::V11, PPC::V12, PPC::V13, PPC::V14, PPC::V15,
  PPC::V16, PPC::V17, PPC::V18, PPC::V19, PPC::V20, PPC::V21, PPC::V22, PPC::V23,
@@ -270,7 +270,7 @@ static void RemoveVRSaveCode(MachineInstr *MI) {
   // epilog blocks.
   for (MachineFunction::iterator I = MF->begin(), E = MF->end(); I != E; ++I) {
     // If last instruction is a return instruction, add an epilogue
-    if (!I->empty() && I->back().isReturn()) {
+    if (I->isReturnBlock()) {
       bool FoundIt = false;
       for (MBBI = I->end(); MBBI != I->begin(); ) {
         --MBBI;
@@ -326,7 +326,7 @@ static void HandleVRSaveUpdate(MachineInstr *MI, const TargetInstrInfo &TII) {
   for (MachineFunction::const_iterator BI = MF->begin(), BE = MF->end();
        UsedRegMask != 0 && BI != BE; ++BI) {
     const MachineBasicBlock &MBB = *BI;
-    if (MBB.empty() || !MBB.back().isReturn())
+    if (!MBB.isReturnBlock())
       continue;
     const MachineInstr &Ret = MBB.back();
     for (unsigned I = 0, E = Ret.getNumOperands(); I != E; ++I) {
@@ -556,6 +556,65 @@ void PPCFrameLowering::replaceFPWithRealFP(MachineFunction &MF) const {
     }
 }
 
+bool PPCFrameLowering::findScratchRegister(MachineBasicBlock *MBB,
+                                           bool UseAtEnd,
+                                           unsigned *ScratchRegister) const {
+  RegScavenger RS;
+  unsigned     R0 = Subtarget.isPPC64() ? PPC::X0 : PPC::R0;
+
+  if (ScratchRegister)
+    *ScratchRegister = R0;
+
+  // If MBB is an entry or exit block, use R0 as the scratch register
+  if ((UseAtEnd && MBB->isReturnBlock()) ||
+      (!UseAtEnd && (&MBB->getParent()->front() == MBB)))
+    return true;
+
+  RS.enterBasicBlock(MBB);
+
+  if (UseAtEnd && !MBB->empty()) {
+    // The scratch register will be used at the end of the block, so must consider
+    // all registers used within the block
+
+    MachineBasicBlock::iterator MBBI = MBB->getFirstTerminator();
+    // If no terminator, back iterator up to previous instruction.
+    if (MBBI == MBB->end())
+      MBBI = std::prev(MBBI);
+
+    if (MBBI != MBB->begin())
+      RS.forward(MBBI);
+  }
+  
+  if (!RS.isRegUsed(R0)) 
+    return true;
+
+  unsigned Reg = RS.FindUnusedReg(Subtarget.isPPC64() ? &PPC::G8RCRegClass
+                                  : &PPC::GPRCRegClass);
+  
+  // Make sure the register scavenger was able to find an available register
+  // If not, use R0 but return false to indicate no register was available and
+  // R0 must be used (as recommended by the ABI)
+  if (Reg == 0)
+    return false;
+
+  if (ScratchRegister)
+    *ScratchRegister = Reg;
+
+  return true;
+}
+
+bool PPCFrameLowering::canUseAsPrologue(const MachineBasicBlock &MBB) const {
+  MachineBasicBlock *TmpMBB = const_cast<MachineBasicBlock *>(&MBB);
+
+  return findScratchRegister(TmpMBB, false, nullptr);
+}
+
+bool PPCFrameLowering::canUseAsEpilogue(const MachineBasicBlock &MBB) const {
+  MachineBasicBlock *TmpMBB = const_cast<MachineBasicBlock *>(&MBB);
+
+  return findScratchRegister(TmpMBB, true, nullptr);
+}
+
 void PPCFrameLowering::emitPrologue(MachineFunction &MF,
                                     MachineBasicBlock &MBB) const {
   MachineBasicBlock::iterator MBBI = MBB.begin();
@@ -613,7 +672,7 @@ void PPCFrameLowering::emitPrologue(MachineFunction &MF,
   unsigned BPReg       = RegInfo->getBaseRegister(MF);
   unsigned FPReg       = isPPC64 ? PPC::X31 : PPC::R31;
   unsigned LRReg       = isPPC64 ? PPC::LR8 : PPC::LR;
-  unsigned ScratchReg  = isPPC64 ? PPC::X0  : PPC::R0;
+  unsigned ScratchReg  = 0;
   unsigned TempReg     = isPPC64 ? PPC::X12 : PPC::R12; // another scratch reg
   //  ...(R12/X12 is volatile in both Darwin & SVR4, & can't be a function arg.)
   const MCInstrDesc& MFLRInst = TII.get(isPPC64 ? PPC::MFLR8
@@ -642,6 +701,9 @@ void PPCFrameLowering::emitPrologue(MachineFunction &MF,
   assert((isPPC64 || !isSVR4ABI || !(!FrameSize && (MustSaveLR || HasFP))) &&
          "FrameSize must be >0 to save/restore the FP or LR for 32-bit SVR4.");
 
+  findScratchRegister(&MBB, false, &ScratchReg);
+  assert(ScratchReg && "No scratch register!");
+         
   int LROffset = getReturnSaveOffset();
 
   int FPOffset = 0;
@@ -950,7 +1012,7 @@ void PPCFrameLowering::emitEpilogue(MachineFunction &MF,
   unsigned SPReg      = isPPC64 ? PPC::X1  : PPC::R1;
   unsigned BPReg      = RegInfo->getBaseRegister(MF);
   unsigned FPReg      = isPPC64 ? PPC::X31 : PPC::R31;
-  unsigned ScratchReg  = isPPC64 ? PPC::X0  : PPC::R0;
+  unsigned ScratchReg = 0;
   unsigned TempReg     = isPPC64 ? PPC::X12 : PPC::R12; // another scratch reg
   const MCInstrDesc& MTLRInst = TII.get( isPPC64 ? PPC::MTLR8
                                                  : PPC::MTLR );
@@ -964,10 +1026,14 @@ void PPCFrameLowering::emitEpilogue(MachineFunction &MF,
                                                    : PPC::ADDI );
   const MCInstrDesc& AddInst = TII.get( isPPC64 ? PPC::ADD8
                                                 : PPC::ADD4 );
-
+  
   int LROffset = getReturnSaveOffset();
 
   int FPOffset = 0;
+
+  findScratchRegister(&MBB, true, &ScratchReg);
+  assert(ScratchReg && "No scratch register!");
+  
   if (HasFP) {
     if (isSVR4ABI) {
       MachineFrameInfo *FFI = MF.getFrameInfo();
@@ -999,7 +1065,7 @@ void PPCFrameLowering::emitEpilogue(MachineFunction &MF,
     PBPOffset = FFI->getObjectOffset(PBPIndex);
   }
 
-  bool IsReturnBlock = MBBI->isReturn();
+  bool IsReturnBlock = (MBBI != MBB.end() && MBBI->isReturn());
   
   if (IsReturnBlock) {
     unsigned RetOpcode = MBBI->getOpcode();
@@ -1062,7 +1128,6 @@ void PPCFrameLowering::emitEpilogue(MachineFunction &MF,
         .addImm(0)
         .addReg(SPReg);
     }
-
   }
 
   if (MustSaveLR)