R600/SI: Choose the correct MOV instruction for copying immediates

[oota-llvm.git] / lib / Target / R600 / R600OptimizeVectorRegisters.cpp

diff --git a/lib/Target/R600/R600OptimizeVectorRegisters.cpp b/lib/Target/R600/R600OptimizeVectorRegisters.cpp
index 0c123a101e7b20108aac1942ed638a31bba6818e..acacffa85d5f19eea98e63a05271e3e3ba18b668 100644 (file)
--- a/lib/Target/R600/R600OptimizeVectorRegisters.cpp
+++ b/lib/Target/R600/R600OptimizeVectorRegisters.cpp
@@ -104,7 +104,7 @@ private:
 public:
   static char ID;
   R600VectorRegMerger(TargetMachine &tm) : MachineFunctionPass(ID),
-  TII (static_cast<const R600InstrInfo *>(tm.getInstrInfo())) { }
+  TII(0) { }
 
   void getAnalysisUsage(AnalysisUsage &AU) const {
     AU.setPreservesCFG();
@@ -183,10 +183,6 @@ MachineInstr *R600VectorRegMerger::RebuildVector(
   std::vector<unsigned> UpdatedUndef = BaseRSI->UndefReg;
   for (DenseMap<unsigned, unsigned>::iterator It = RSI->RegToChan.begin(),
       E = RSI->RegToChan.end(); It != E; ++It) {
-    if (BaseRSI->RegToChan.find((*It).first) != BaseRSI->RegToChan.end()) {
-      UpdatedRegToChan[(*It).first] = (*It).second;
-      continue;
-    }
     unsigned DstReg = MRI->createVirtualRegister(&AMDGPU::R600_Reg128RegClass);
     unsigned SubReg = (*It).first;
     unsigned Swizzle = (*It).second;
@@ -198,9 +194,13 @@ MachineInstr *R600VectorRegMerger::RebuildVector(
         .addReg(SubReg)
         .addImm(Chan);
     UpdatedRegToChan[SubReg] = Chan;
-    UpdatedUndef.erase(
-        std::remove(UpdatedUndef.begin(), UpdatedUndef.end(), Chan),
-        UpdatedUndef.end());
+    std::vector<unsigned>::iterator ChanPos =
+        std::find(UpdatedUndef.begin(), UpdatedUndef.end(), Chan);
+    if (ChanPos != UpdatedUndef.end())
+      UpdatedUndef.erase(ChanPos);
+    assert(std::find(UpdatedUndef.begin(), UpdatedUndef.end(), Chan) ==
+               UpdatedUndef.end() &&
+           "UpdatedUndef shouldn't contain Chan more than once!");
     DEBUG(dbgs() << "    ->"; Tmp->dump(););
     (void)Tmp;
     SrcVec = DstReg;
@@ -310,6 +310,7 @@ void R600VectorRegMerger::trackRSI(const RegSeqInfo &RSI) {
 }
 
 bool R600VectorRegMerger::runOnMachineFunction(MachineFunction &Fn) {
+  TII = static_cast<const R600InstrInfo *>(Fn.getTarget().getInstrInfo());
   MRI = &(Fn.getRegInfo());
   for (MachineFunction::iterator MBB = Fn.begin(), MBBe = Fn.end();
        MBB != MBBe; ++MBB) {
@@ -321,8 +322,17 @@ bool R600VectorRegMerger::runOnMachineFunction(MachineFunction &Fn) {
     for (MachineBasicBlock::iterator MII = MB->begin(), MIIE = MB->end();
          MII != MIIE; ++MII) {
       MachineInstr *MI = MII;
-      if (MI->getOpcode() != AMDGPU::REG_SEQUENCE)
+      if (MI->getOpcode() != AMDGPU::REG_SEQUENCE) {
+        if (TII->get(MI->getOpcode()).TSFlags & R600_InstFlag::TEX_INST) {
+          unsigned Reg = MI->getOperand(1).getReg();
+          for (MachineRegisterInfo::def_iterator It = MRI->def_begin(Reg),
+              E = MRI->def_end(); It != E; ++It) {
+            RemoveMI(&(*It));
+          }
+        }
         continue;
+      }
+
 
       RegSeqInfo RSI(*MRI, MI);