R600: Do not predicated basic block with multiple alu clause

Test is not included as it is several 1000 lines long.
To test this functionnality, a test case must generate at least 2 ALU clauses,
where an ALU clause is ~110 instructions long.

NOTE: This is a candidate for the stable branch.
llvm-svn: 185943

llvm/lib/Target/R600/AMDGPUTargetMachine.cpp

@@ -148,7 +148,11 @@ bool AMDGPUPassConfig::addPostRegAlloc() {
 }
 
 bool AMDGPUPassConfig::addPreSched2() {
+  const AMDGPUSubtarget &ST = TM->getSubtarget<AMDGPUSubtarget>();
 
+  if (ST.getGeneration() <= AMDGPUSubtarget::NORTHERN_ISLANDS) {
+    addPass(createR600EmitClauseMarkers(*TM));
+  }
   addPass(&IfConverterID);
   return false;
 }
@@ -158,7 +162,6 @@ bool AMDGPUPassConfig::addPreEmitPass() {
   if (ST.getGeneration() <= AMDGPUSubtarget::NORTHERN_ISLANDS) {
     addPass(createAMDGPUCFGPreparationPass(*TM));
     addPass(createAMDGPUCFGStructurizerPass(*TM));
-    addPass(createR600EmitClauseMarkers(*TM));
     addPass(createR600ExpandSpecialInstrsPass(*TM));
     addPass(&FinalizeMachineBundlesID);
     addPass(createR600Packetizer(*TM));

llvm/lib/Target/R600/R600ControlFlowFinalizer.cpp

@@ -256,6 +256,7 @@ private:
         ClauseContent.push_back(MILit);
       }
     }
+    assert(ClauseContent.size() < 128 && "ALU clause is too big");
     ClauseHead->getOperand(7).setImm(ClauseContent.size() - 1);
     return ClauseFile(ClauseHead, ClauseContent);
   }
@@ -276,6 +277,7 @@ private:
   void
   EmitALUClause(MachineBasicBlock::iterator InsertPos, ClauseFile &Clause,
       unsigned &CfCount) {
+    Clause.first->getOperand(0).setImm(0);
     CounterPropagateAddr(Clause.first, CfCount);
     MachineBasicBlock *BB = Clause.first->getParent();
     BuildMI(BB, InsertPos->getDebugLoc(), TII->get(AMDGPU::ALU_CLAUSE))

llvm/lib/Target/R600/R600EmitClauseMarkers.cpp

@@ -32,6 +32,7 @@ class R600EmitClauseMarkersPass : public MachineFunctionPass {
 private:
   static char ID;
   const R600InstrInfo *TII;
+  int Address;
 
   unsigned OccupiedDwords(MachineInstr *MI) const {
     switch (MI->getOpcode()) {
@@ -159,7 +160,7 @@ private:
   }
 
   MachineBasicBlock::iterator
-  MakeALUClause(MachineBasicBlock &MBB, MachineBasicBlock::iterator I) const {
+  MakeALUClause(MachineBasicBlock &MBB, MachineBasicBlock::iterator I) {
     MachineBasicBlock::iterator ClauseHead = I;
     std::vector<std::pair<unsigned, unsigned> > KCacheBanks;
     bool PushBeforeModifier = false;
@@ -199,20 +200,25 @@ private:
     unsigned Opcode = PushBeforeModifier ?
         AMDGPU::CF_ALU_PUSH_BEFORE : AMDGPU::CF_ALU;
     BuildMI(MBB, ClauseHead, MBB.findDebugLoc(ClauseHead), TII->get(Opcode))
-        .addImm(0) // ADDR
+    // We don't use the ADDR field until R600ControlFlowFinalizer pass, where
+    // it is safe to assume it is 0. However if we always put 0 here, the ifcvt
+    // pass may assume that identical ALU clause starter at the beginning of a 
+    // true and false branch can be factorized which is not the case.
+        .addImm(Address++) // ADDR
         .addImm(KCacheBanks.empty()?0:KCacheBanks[0].first) // KB0
         .addImm((KCacheBanks.size() < 2)?0:KCacheBanks[1].first) // KB1
         .addImm(KCacheBanks.empty()?0:2) // KM0
         .addImm((KCacheBanks.size() < 2)?0:2) // KM1
         .addImm(KCacheBanks.empty()?0:KCacheBanks[0].second) // KLINE0
         .addImm((KCacheBanks.size() < 2)?0:KCacheBanks[1].second) // KLINE1
-        .addImm(AluInstCount); // COUNT
+        .addImm(AluInstCount) // COUNT
+        .addImm(1); // Enabled
     return I;
   }
 
 public:
   R600EmitClauseMarkersPass(TargetMachine &tm) : MachineFunctionPass(ID),
-    TII(0) { }
+    TII(0), Address(0) { }
 
   virtual bool runOnMachineFunction(MachineFunction &MF) {
     TII = static_cast<const R600InstrInfo *>(MF.getTarget().getInstrInfo());

llvm/lib/Target/R600/R600InstrInfo.cpp

@@ -651,6 +651,17 @@ int R600InstrInfo::getBranchInstr(const MachineOperand &op) const {
   };
 }
 
+static
+MachineBasicBlock::iterator FindLastAluClause(MachineBasicBlock &MBB) {
+  for (MachineBasicBlock::reverse_iterator It = MBB.rbegin(), E = MBB.rend();
+      It != E; ++It) {
+    if (It->getOpcode() == AMDGPU::CF_ALU ||
+        It->getOpcode() == AMDGPU::CF_ALU_PUSH_BEFORE)
+      return llvm::prior(It.base());
+  }
+  return MBB.end();
+}
+
 unsigned
 R600InstrInfo::InsertBranch(MachineBasicBlock &MBB,
                             MachineBasicBlock *TBB,
@@ -672,6 +683,11 @@ R600InstrInfo::InsertBranch(MachineBasicBlock &MBB,
       BuildMI(&MBB, DL, get(AMDGPU::JUMP_COND))
              .addMBB(TBB)
              .addReg(AMDGPU::PREDICATE_BIT, RegState::Kill);
+      MachineBasicBlock::iterator CfAlu = FindLastAluClause(MBB);
+      if (CfAlu == MBB.end())
+        return 1;
+      assert (CfAlu->getOpcode() == AMDGPU::CF_ALU);
+      CfAlu->setDesc(get(AMDGPU::CF_ALU_PUSH_BEFORE));
       return 1;
     }
   } else {
@@ -683,6 +699,11 @@ R600InstrInfo::InsertBranch(MachineBasicBlock &MBB,
             .addMBB(TBB)
             .addReg(AMDGPU::PREDICATE_BIT, RegState::Kill);
     BuildMI(&MBB, DL, get(AMDGPU::JUMP)).addMBB(FBB);
+    MachineBasicBlock::iterator CfAlu = FindLastAluClause(MBB);
+    if (CfAlu == MBB.end())
+      return 2;
+    assert (CfAlu->getOpcode() == AMDGPU::CF_ALU);
+    CfAlu->setDesc(get(AMDGPU::CF_ALU_PUSH_BEFORE));
     return 2;
   }
 }
@@ -706,6 +727,11 @@ R600InstrInfo::RemoveBranch(MachineBasicBlock &MBB) const {
     MachineInstr *predSet = findFirstPredicateSetterFrom(MBB, I);
     clearFlag(predSet, 0, MO_FLAG_PUSH);
     I->eraseFromParent();
+    MachineBasicBlock::iterator CfAlu = FindLastAluClause(MBB);
+    if (CfAlu == MBB.end())
+      break;
+    assert (CfAlu->getOpcode() == AMDGPU::CF_ALU_PUSH_BEFORE);
+    CfAlu->setDesc(get(AMDGPU::CF_ALU));
     break;
   }
   case AMDGPU::JUMP:
@@ -726,6 +752,11 @@ R600InstrInfo::RemoveBranch(MachineBasicBlock &MBB) const {
     MachineInstr *predSet = findFirstPredicateSetterFrom(MBB, I);
     clearFlag(predSet, 0, MO_FLAG_PUSH);
     I->eraseFromParent();
+    MachineBasicBlock::iterator CfAlu = FindLastAluClause(MBB);
+    if (CfAlu == MBB.end())
+      break;
+    assert (CfAlu->getOpcode() == AMDGPU::CF_ALU_PUSH_BEFORE);
+    CfAlu->setDesc(get(AMDGPU::CF_ALU));
     break;
   }
   case AMDGPU::JUMP:
@@ -760,6 +791,15 @@ R600InstrInfo::isPredicable(MachineInstr *MI) const {
 
   if (MI->getOpcode() == AMDGPU::KILLGT) {
     return false;
+  } else if (MI->getOpcode() == AMDGPU::CF_ALU) {
+    // If the clause start in the middle of MBB then the MBB has more
+    // than a single clause, unable to predicate several clauses.
+    if (MI->getParent()->begin() != MachineBasicBlock::iterator(MI))
+      return false;
+    // TODO: We don't support KC merging atm
+    if (MI->getOperand(3).getImm() != 0 || MI->getOperand(4).getImm() != 0)
+      return false;
+    return true;
   } else if (isVector(*MI)) {
     return false;
   } else {
@@ -855,6 +895,11 @@ R600InstrInfo::PredicateInstruction(MachineInstr *MI,
                       const SmallVectorImpl<MachineOperand> &Pred) const {
   int PIdx = MI->findFirstPredOperandIdx();
 
+  if (MI->getOpcode() == AMDGPU::CF_ALU) {
+    MI->getOperand(8).setImm(0);
+    return true;
+  }
+
   if (PIdx != -1) {
     MachineOperand &PMO = MI->getOperand(PIdx);
     PMO.setReg(Pred[2].getReg());

llvm/lib/Target/R600/R600Instructions.td

@@ -563,7 +563,7 @@ class ALU_CLAUSE<bits<4> inst, string OpName> : AMDGPUInst <(outs),
 (ins i32imm:$ADDR, i32imm:$KCACHE_BANK0, i32imm:$KCACHE_BANK1,
 KCACHE:$KCACHE_MODE0, KCACHE:$KCACHE_MODE1,
 i32imm:$KCACHE_ADDR0, i32imm:$KCACHE_ADDR1,
-i32imm:$COUNT),
+i32imm:$COUNT, i32imm:$Enabled),
 !strconcat(OpName, " $COUNT, @$ADDR, "
 "KC0[$KCACHE_MODE0], KC1[$KCACHE_MODE1]"),
 [] >, CF_ALU_WORD0, CF_ALU_WORD1 {

llvm/lib/Target/R600/R600Packetizer.cpp

@@ -304,7 +304,8 @@ bool R600Packetizer::runOnMachineFunction(MachineFunction &Fn) {
     MachineBasicBlock::iterator End = MBB->end();
     MachineBasicBlock::iterator MI = MBB->begin();
     while (MI != End) {
-      if (MI->isKill()) {
+      if (MI->isKill() ||
+          (MI->getOpcode() == AMDGPU::CF_ALU && !MI->getOperand(8).getImm())) {
         MachineBasicBlock::iterator DeleteMI = MI;
         ++MI;
         MBB->erase(DeleteMI);

llvm/test/CodeGen/R600/jump-address.ll

@@ -1,6 +1,6 @@
 ;RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
 
-; CHECK: JUMP @3
+; CHECK: JUMP @7
 ; CHECK: EXPORT
 ; CHECK-NOT: EXPORT