Improve PPC VR (Altivec) register spilling

This change cleans up two issues with Altivec register spilling:

  1. The spilling code was inefficient (using two instructions, and add and a
     load, when just one would do)

  2. The code assumed that r0 would always be available (true for now, but this
     will change)

The new code handles VR spilling just like GPR spills but forced into r+r mode.
As a result, when any VR spills are present, we must now always allocate the
register-scavenger spill slot.

llvm-svn: 177231

1 files changed, 29 insertions, 34 deletions

diff --git a/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp b/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
index 7ae34423309..7fe78809343 100644
--- a/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
+++ b/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
@@ -439,7 +439,8 @@ PPCInstrInfo::StoreRegToStackSlot(MachineFunction &MF,
                                   unsigned SrcReg, bool isKill,
                                   int FrameIdx,
                                   const TargetRegisterClass *RC,
-                                  SmallVectorImpl<MachineInstr*> &NewMIs) const{
+                                  SmallVectorImpl<MachineInstr*> &NewMIs,
+                                  bool &NonRI) const{
   DebugLoc DL;
   if (PPC::GPRCRegClass.hasSubClassEq(RC)) {
     if (SrcReg != PPC::LR) {
@@ -521,23 +522,14 @@ PPCInstrInfo::StoreRegToStackSlot(MachineFunction &MF,
       Reg = PPC::CR7;
 
     return StoreRegToStackSlot(MF, Reg, isKill, FrameIdx,
-                               &PPC::CRRCRegClass, NewMIs);
+                               &PPC::CRRCRegClass, NewMIs, NonRI);
 
   } else if (PPC::VRRCRegClass.hasSubClassEq(RC)) {
-    // We don't have indexed addressing for vector loads.  Emit:
-    // R0 = ADDI FI#
-    // STVX VAL, 0, R0
-    //
-    // FIXME: We use R0 here, because it isn't available for RA.
-    bool Is64Bit = TM.getSubtargetImpl()->isPPC64();
-    unsigned Instr = Is64Bit ? PPC::ADDI8 : PPC::ADDI;
-    unsigned GPR0  = Is64Bit ? PPC::X0    : PPC::R0;
-    NewMIs.push_back(addFrameReference(BuildMI(MF, DL, get(Instr), GPR0),
-                                       FrameIdx, 0, 0));
-    NewMIs.push_back(BuildMI(MF, DL, get(PPC::STVX))
-                     .addReg(SrcReg, getKillRegState(isKill))
-                     .addReg(GPR0)
-                     .addReg(GPR0));
+    NewMIs.push_back(addFrameReference(BuildMI(MF, DL, get(PPC::STVX))
+                                       .addReg(SrcReg,
+                                               getKillRegState(isKill)),
+                                       FrameIdx));
+    NonRI = true;
   } else {
     llvm_unreachable("Unknown regclass!");
   }
@@ -557,9 +549,13 @@ PPCInstrInfo::storeRegToStackSlot(MachineBasicBlock &MBB,
   PPCFunctionInfo *FuncInfo = MF.getInfo<PPCFunctionInfo>();
   FuncInfo->setHasSpills();
 
-  if (StoreRegToStackSlot(MF, SrcReg, isKill, FrameIdx, RC, NewMIs))
+  bool NonRI = false;
+  if (StoreRegToStackSlot(MF, SrcReg, isKill, FrameIdx, RC, NewMIs, NonRI))
     FuncInfo->setSpillsCR();
 
+  if (NonRI)
+    FuncInfo->setHasNonRISpills();
+
   for (unsigned i = 0, e = NewMIs.size(); i != e; ++i)
     MBB.insert(MI, NewMIs[i]);
 
@@ -576,7 +572,8 @@ bool
 PPCInstrInfo::LoadRegFromStackSlot(MachineFunction &MF, DebugLoc DL,
                                    unsigned DestReg, int FrameIdx,
                                    const TargetRegisterClass *RC,
-                                   SmallVectorImpl<MachineInstr*> &NewMIs)const{
+                                   SmallVectorImpl<MachineInstr*> &NewMIs,
+                                   bool &NonRI) const{
   if (PPC::GPRCRegClass.hasSubClassEq(RC)) {
     if (DestReg != PPC::LR) {
       NewMIs.push_back(addFrameReference(BuildMI(MF, DL, get(PPC::LWZ),
@@ -635,21 +632,12 @@ PPCInstrInfo::LoadRegFromStackSlot(MachineFunction &MF, DebugLoc DL,
       Reg = PPC::CR7;
 
     return LoadRegFromStackSlot(MF, DL, Reg, FrameIdx,
-                                &PPC::CRRCRegClass, NewMIs);
+                                &PPC::CRRCRegClass, NewMIs, NonRI);
 
   } else if (PPC::VRRCRegClass.hasSubClassEq(RC)) {
-    // We don't have indexed addressing for vector loads.  Emit:
-    // R0 = ADDI FI#
-    // Dest = LVX 0, R0
-    //
-    // FIXME: We use R0 here, because it isn't available for RA.
-    bool Is64Bit = TM.getSubtargetImpl()->isPPC64();
-    unsigned Instr = Is64Bit ? PPC::ADDI8 : PPC::ADDI;
-    unsigned GPR0  = Is64Bit ? PPC::X0    : PPC::R0;
-    NewMIs.push_back(addFrameReference(BuildMI(MF, DL, get(Instr), GPR0),
-                                       FrameIdx, 0, 0));
-    NewMIs.push_back(BuildMI(MF, DL, get(PPC::LVX),DestReg).addReg(GPR0)
-                     .addReg(GPR0));
+    NewMIs.push_back(addFrameReference(BuildMI(MF, DL, get(PPC::LVX), DestReg),
+                                       FrameIdx));
+    NonRI = true;
   } else {
     llvm_unreachable("Unknown regclass!");
   }
@@ -667,10 +655,17 @@ PPCInstrInfo::loadRegFromStackSlot(MachineBasicBlock &MBB,
   SmallVector<MachineInstr*, 4> NewMIs;
   DebugLoc DL;
   if (MI != MBB.end()) DL = MI->getDebugLoc();
-  if (LoadRegFromStackSlot(MF, DL, DestReg, FrameIdx, RC, NewMIs)) {
-    PPCFunctionInfo *FuncInfo = MF.getInfo<PPCFunctionInfo>();
+
+  PPCFunctionInfo *FuncInfo = MF.getInfo<PPCFunctionInfo>();
+  FuncInfo->setHasSpills();
+
+  bool NonRI = false;
+  if (LoadRegFromStackSlot(MF, DL, DestReg, FrameIdx, RC, NewMIs, NonRI))
     FuncInfo->setSpillsCR();
-  }
+
+  if (NonRI)
+    FuncInfo->setHasNonRISpills();
+
   for (unsigned i = 0, e = NewMIs.size(); i != e; ++i)
     MBB.insert(MI, NewMIs[i]);