//===----------------------------------------------------------------------===//
#define DEBUG_TYPE "branchfolding"
+#include "BranchFolding.h"
#include "llvm/CodeGen/Passes.h"
#include "llvm/CodeGen/MachineModuleInfo.h"
#include "llvm/CodeGen/MachineFunctionPass.h"
#include "llvm/Target/TargetRegisterInfo.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
+#include "llvm/Support/ErrorHandling.h"
+#include "llvm/Support/raw_ostream.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/ADT/STLExtras.h"
STATISTIC(NumTailMerge , "Number of block tails merged");
static cl::opt<cl::boolOrDefault> FlagEnableTailMerge("enable-tail-merge",
cl::init(cl::BOU_UNSET), cl::Hidden);
-namespace {
- // Throttle for huge numbers of predecessors (compile speed problems)
- static cl::opt<unsigned>
- TailMergeThreshold("tail-merge-threshold",
- cl::desc("Max number of predecessors to consider tail merging"),
- cl::init(100), cl::Hidden);
-
- struct VISIBILITY_HIDDEN BranchFolder : public MachineFunctionPass {
- static char ID;
- explicit BranchFolder(bool defaultEnableTailMerge) :
- MachineFunctionPass((intptr_t)&ID) {
- switch (FlagEnableTailMerge) {
- case cl::BOU_UNSET: EnableTailMerge = defaultEnableTailMerge; break;
- case cl::BOU_TRUE: EnableTailMerge = true; break;
- case cl::BOU_FALSE: EnableTailMerge = false; break;
- }
- }
+// Throttle for huge numbers of predecessors (compile speed problems)
+static cl::opt<unsigned>
+TailMergeThreshold("tail-merge-threshold",
+ cl::desc("Max number of predecessors to consider tail merging"),
+ cl::init(150), cl::Hidden);
- virtual bool runOnMachineFunction(MachineFunction &MF);
- virtual const char *getPassName() const { return "Control Flow Optimizer"; }
- const TargetInstrInfo *TII;
- MachineModuleInfo *MMI;
- bool MadeChange;
- private:
- // Tail Merging.
- bool EnableTailMerge;
- bool TailMergeBlocks(MachineFunction &MF);
- bool TryMergeBlocks(MachineBasicBlock* SuccBB,
- MachineBasicBlock* PredBB);
- void ReplaceTailWithBranchTo(MachineBasicBlock::iterator OldInst,
- MachineBasicBlock *NewDest);
- MachineBasicBlock *SplitMBBAt(MachineBasicBlock &CurMBB,
- MachineBasicBlock::iterator BBI1);
- unsigned ComputeSameTails(unsigned CurHash, unsigned minCommonTailLength);
- void RemoveBlocksWithHash(unsigned CurHash, MachineBasicBlock* SuccBB,
- MachineBasicBlock* PredBB);
-
- typedef std::pair<unsigned,MachineBasicBlock*> MergePotentialsElt;
- typedef std::vector<MergePotentialsElt>::iterator MPIterator;
- std::vector<MergePotentialsElt> MergePotentials;
- typedef std::pair<MPIterator, MachineBasicBlock::iterator> SameTailElt;
- std::vector<SameTailElt> SameTails;
-
- const TargetRegisterInfo *RegInfo;
- RegScavenger *RS;
- // Branch optzn.
- bool OptimizeBranches(MachineFunction &MF);
- void OptimizeBlock(MachineBasicBlock *MBB);
- void RemoveDeadBlock(MachineBasicBlock *MBB);
- bool OptimizeImpDefsBlock(MachineBasicBlock *MBB);
-
- bool CanFallThrough(MachineBasicBlock *CurBB);
- bool CanFallThrough(MachineBasicBlock *CurBB, bool BranchUnAnalyzable,
- MachineBasicBlock *TBB, MachineBasicBlock *FBB,
- const std::vector<MachineOperand> &Cond);
- };
- char BranchFolder::ID = 0;
-}
+
+char BranchFolderPass::ID = 0;
FunctionPass *llvm::createBranchFoldingPass(bool DefaultEnableTailMerge) {
- return new BranchFolder(DefaultEnableTailMerge); }
+ return new BranchFolderPass(DefaultEnableTailMerge);
+}
+
+bool BranchFolderPass::runOnMachineFunction(MachineFunction &MF) {
+ return OptimizeFunction(MF,
+ MF.getTarget().getInstrInfo(),
+ MF.getTarget().getRegisterInfo(),
+ getAnalysisIfAvailable<MachineModuleInfo>());
+}
+
+
+
+BranchFolder::BranchFolder(bool defaultEnableTailMerge) {
+ switch (FlagEnableTailMerge) {
+ case cl::BOU_UNSET: EnableTailMerge = defaultEnableTailMerge; break;
+ case cl::BOU_TRUE: EnableTailMerge = true; break;
+ case cl::BOU_FALSE: EnableTailMerge = false; break;
+ }
+}
/// RemoveDeadBlock - Remove the specified dead machine basic block from the
/// function, updating the CFG.
void BranchFolder::RemoveDeadBlock(MachineBasicBlock *MBB) {
assert(MBB->pred_empty() && "MBB must be dead!");
- DOUT << "\nRemoving MBB: " << *MBB;
+ DEBUG(errs() << "\nRemoving MBB: " << *MBB);
MachineFunction *MF = MBB->getParent();
// drop all successors.
while (!MBB->succ_empty())
MBB->removeSuccessor(MBB->succ_end()-1);
- // If there is DWARF info to active, check to see if there are any LABEL
- // records in the basic block. If so, unregister them from MachineModuleInfo.
+ // If there are any labels in the basic block, unregister them from
+ // MachineModuleInfo.
if (MMI && !MBB->empty()) {
for (MachineBasicBlock::iterator I = MBB->begin(), E = MBB->end();
I != E; ++I) {
- if ((unsigned)I->getOpcode() == TargetInstrInfo::LABEL) {
+ if (I->isLabel())
// The label ID # is always operand #0, an immediate.
MMI->InvalidateLabel(I->getOperand(0).getImm());
- }
}
}
// Remove the block.
- MF->getBasicBlockList().erase(MBB);
+ MF->erase(MBB);
}
/// OptimizeImpDefsBlock - If a basic block is just a bunch of implicit_def
break;
unsigned Reg = I->getOperand(0).getReg();
ImpDefRegs.insert(Reg);
- for (const unsigned *SubRegs = RegInfo->getSubRegisters(Reg);
+ for (const unsigned *SubRegs = TRI->getSubRegisters(Reg);
unsigned SubReg = *SubRegs; ++SubRegs)
ImpDefRegs.insert(SubReg);
++I;
return true;
}
-bool BranchFolder::runOnMachineFunction(MachineFunction &MF) {
- TII = MF.getTarget().getInstrInfo();
- if (!TII) return false;
+/// OptimizeFunction - Perhaps branch folding, tail merging and other
+/// CFG optimizations on the given function.
+bool BranchFolder::OptimizeFunction(MachineFunction &MF,
+ const TargetInstrInfo *tii,
+ const TargetRegisterInfo *tri,
+ MachineModuleInfo *mmi) {
+ if (!tii) return false;
- RegInfo = MF.getTarget().getRegisterInfo();
+ TII = tii;
+ TRI = tri;
+ MMI = mmi;
+
+ RS = TRI->requiresRegisterScavenging(MF) ? new RegScavenger() : NULL;
// Fix CFG. The later algorithms expect it to be right.
- bool EverMadeChange = false;
+ bool MadeChange = false;
for (MachineFunction::iterator I = MF.begin(), E = MF.end(); I != E; I++) {
MachineBasicBlock *MBB = I, *TBB = 0, *FBB = 0;
- std::vector<MachineOperand> Cond;
- if (!TII->AnalyzeBranch(*MBB, TBB, FBB, Cond))
- EverMadeChange |= MBB->CorrectExtraCFGEdges(TBB, FBB, !Cond.empty());
- EverMadeChange |= OptimizeImpDefsBlock(MBB);
+ SmallVector<MachineOperand, 4> Cond;
+ if (!TII->AnalyzeBranch(*MBB, TBB, FBB, Cond, true))
+ MadeChange |= MBB->CorrectExtraCFGEdges(TBB, FBB, !Cond.empty());
+ MadeChange |= OptimizeImpDefsBlock(MBB);
}
- RS = RegInfo->requiresRegisterScavenging(MF) ? new RegScavenger() : NULL;
-
- MMI = getAnalysisToUpdate<MachineModuleInfo>();
bool MadeChangeThisIteration = true;
while (MadeChangeThisIteration) {
MadeChangeThisIteration = false;
MadeChangeThisIteration |= TailMergeBlocks(MF);
MadeChangeThisIteration |= OptimizeBranches(MF);
- EverMadeChange |= MadeChangeThisIteration;
+ MadeChange |= MadeChangeThisIteration;
}
// See if any jump tables have become mergable or dead as the code generator
// Scan the jump tables, seeing if there are any duplicates. Note that this
// is N^2, which should be fixed someday.
- for (unsigned i = 1, e = JTs.size(); i != e; ++i)
- JTMapping.push_back(JTI->getJumpTableIndex(JTs[i].MBBs));
+ for (unsigned i = 1, e = JTs.size(); i != e; ++i) {
+ if (JTs[i].MBBs.empty())
+ JTMapping.push_back(i);
+ else
+ JTMapping.push_back(JTI->getJumpTableIndex(JTs[i].MBBs));
+ }
// If a jump table was merge with another one, walk the function rewriting
// references to jump tables to reference the new JT ID's. Keep track of
I != E; ++I)
for (unsigned op = 0, e = I->getNumOperands(); op != e; ++op) {
MachineOperand &Op = I->getOperand(op);
- if (!Op.isJumpTableIndex()) continue;
+ if (!Op.isJTI()) continue;
unsigned NewIdx = JTMapping[Op.getIndex()];
Op.setIndex(NewIdx);
for (unsigned i = 0, e = JTIsLive.size(); i != e; ++i)
if (!JTIsLive.test(i)) {
JTI->RemoveJumpTable(i);
- EverMadeChange = true;
+ MadeChange = true;
}
}
-
+
delete RS;
- return EverMadeChange;
+ return MadeChange;
}
//===----------------------------------------------------------------------===//
// If OldBB isn't immediately before OldBB, insert a branch to it.
if (++MachineFunction::iterator(OldBB) != MachineFunction::iterator(NewDest))
- TII->InsertBranch(*OldBB, NewDest, 0, std::vector<MachineOperand>());
+ TII->InsertBranch(*OldBB, NewDest, 0, SmallVector<MachineOperand, 0>());
OldBB->addSuccessor(NewDest);
++NumTailMerge;
}
/// iterator. This returns the new MBB.
MachineBasicBlock *BranchFolder::SplitMBBAt(MachineBasicBlock &CurMBB,
MachineBasicBlock::iterator BBI1) {
+ MachineFunction &MF = *CurMBB.getParent();
+
// Create the fall-through block.
MachineFunction::iterator MBBI = &CurMBB;
- MachineBasicBlock *NewMBB = new MachineBasicBlock(CurMBB.getBasicBlock());
- CurMBB.getParent()->getBasicBlockList().insert(++MBBI, NewMBB);
+ MachineBasicBlock *NewMBB =MF.CreateMachineBasicBlock(CurMBB.getBasicBlock());
+ CurMBB.getParent()->insert(++MBBI, NewMBB);
// Move all the successors of this block to the specified block.
- while (!CurMBB.succ_empty()) {
- MachineBasicBlock *S = *(CurMBB.succ_end()-1);
- NewMBB->addSuccessor(S);
- CurMBB.removeSuccessor(S);
- }
+ NewMBB->transferSuccessors(&CurMBB);
// Add an edge from CurMBB to NewMBB for the fall-through.
CurMBB.addSuccessor(NewMBB);
RS->enterBasicBlock(&CurMBB);
if (!CurMBB.empty())
RS->forward(prior(CurMBB.end()));
- BitVector RegsLiveAtExit(RegInfo->getNumRegs());
+ BitVector RegsLiveAtExit(TRI->getNumRegs());
RS->getRegsUsed(RegsLiveAtExit, false);
- for (unsigned int i=0, e=RegInfo->getNumRegs(); i!=e; i++)
+ for (unsigned int i=0, e=TRI->getNumRegs(); i!=e; i++)
if (RegsLiveAtExit[i])
NewMBB->addLiveIn(i);
}
const TargetInstrDesc &TID = I->getDesc();
if (TID.isCall())
Time += 10;
- else if (TID.isSimpleLoad() || TID.mayStore())
+ else if (TID.mayLoad() || TID.mayStore())
Time += 2;
else
++Time;
return Time;
}
-/// ShouldSplitFirstBlock - We need to either split MBB1 at MBB1I or MBB2 at
-/// MBB2I and then insert an unconditional branch in the other block. Determine
-/// which is the best to split
-static bool ShouldSplitFirstBlock(MachineBasicBlock *MBB1,
- MachineBasicBlock::iterator MBB1I,
- MachineBasicBlock *MBB2,
- MachineBasicBlock::iterator MBB2I,
- MachineBasicBlock *PredBB) {
- // If one block is the entry block, split the other one; we can't generate
- // a branch to the entry block, as its label is not emitted.
- MachineBasicBlock *Entry = MBB1->getParent()->begin();
- if (MBB1 == Entry)
- return false;
- if (MBB2 == Entry)
- return true;
-
- // If one block falls through into the common successor, choose that
- // one to split; it is one instruction less to do that.
- if (PredBB) {
- if (MBB1 == PredBB)
- return true;
- else if (MBB2 == PredBB)
- return false;
- }
- // TODO: if we had some notion of which block was hotter, we could split
- // the hot block, so it is the fall-through. Since we don't have profile info
- // make a decision based on which will hurt most to split.
- unsigned MBB1Time = EstimateRuntime(MBB1->begin(), MBB1I);
- unsigned MBB2Time = EstimateRuntime(MBB2->begin(), MBB2I);
-
- // If the MBB1 prefix takes "less time" to run than the MBB2 prefix, split the
- // MBB1 block so it falls through. This will penalize the MBB2 path, but will
- // have a lower overall impact on the program execution.
- return MBB1Time < MBB2Time;
-}
-
// CurMBB needs to add an unconditional branch to SuccMBB (we removed these
// branches temporarily for tail merging). In the case where CurMBB ends
// with a conditional branch to the next block, optimize by reversing the
MachineFunction *MF = CurMBB->getParent();
MachineFunction::iterator I = next(MachineFunction::iterator(CurMBB));
MachineBasicBlock *TBB = 0, *FBB = 0;
- std::vector<MachineOperand> Cond;
+ SmallVector<MachineOperand, 4> Cond;
if (I != MF->end() &&
- !TII->AnalyzeBranch(*CurMBB, TBB, FBB, Cond)) {
+ !TII->AnalyzeBranch(*CurMBB, TBB, FBB, Cond, true)) {
MachineBasicBlock *NextBB = I;
if (TBB == NextBB && !Cond.empty() && !FBB) {
if (!TII->ReverseBranchCondition(Cond)) {
}
}
}
- TII->InsertBranch(*CurMBB, SuccBB, NULL, std::vector<MachineOperand>());
+ TII->InsertBranch(*CurMBB, SuccBB, NULL, SmallVector<MachineOperand, 0>());
}
static bool MergeCompare(const std::pair<unsigned,MachineBasicBlock*> &p,
// _GLIBCXX_DEBUG checks strict weak ordering, which involves comparing
// an object with itself.
#ifndef _GLIBCXX_DEBUG
- assert(0 && "Predecessor appears twice");
+ llvm_unreachable("Predecessor appears twice");
#endif
- return(false);
+ return false;
}
}
CurMPIter->second,
I->second,
TrialBBI1, TrialBBI2);
- if (CommonTailLen >= minCommonTailLength) {
+ // If we will have to split a block, there should be at least
+ // minCommonTailLength instructions in common; if not, at worst
+ // we will be replacing a fallthrough into the common tail with a
+ // branch, which at worst breaks even with falling through into
+ // the duplicated common tail, so 1 instruction in common is enough.
+ // We will always pick a block we do not have to split as the common
+ // tail if there is one.
+ // (Empty blocks will get forwarded and need not be considered.)
+ if (CommonTailLen >= minCommonTailLength ||
+ (CommonTailLen > 0 &&
+ (TrialBBI1==CurMPIter->second->begin() ||
+ TrialBBI2==I->second->begin()))) {
if (CommonTailLen > maxCommonTailLength) {
SameTails.clear();
maxCommonTailLength = CommonTailLen;
void BranchFolder::RemoveBlocksWithHash(unsigned CurHash,
MachineBasicBlock* SuccBB,
MachineBasicBlock* PredBB) {
- for (MPIterator CurMPIter = prior(MergePotentials.end()),
- B = MergePotentials.begin();
+ MPIterator CurMPIter, B;
+ for (CurMPIter = prior(MergePotentials.end()), B = MergePotentials.begin();
CurMPIter->first==CurHash;
--CurMPIter) {
// Put the unconditional branch back, if we need one.
MachineBasicBlock *CurMBB = CurMPIter->second;
if (SuccBB && CurMBB != PredBB)
FixTail(CurMBB, SuccBB, TII);
- MergePotentials.erase(CurMPIter);
- if (CurMPIter==B)
+ if (CurMPIter==B)
break;
}
+ if (CurMPIter->first!=CurHash)
+ CurMPIter++;
+ MergePotentials.erase(CurMPIter, MergePotentials.end());
+}
+
+/// CreateCommonTailOnlyBlock - None of the blocks to be tail-merged consist
+/// only of the common tail. Create a block that does by splitting one.
+unsigned BranchFolder::CreateCommonTailOnlyBlock(MachineBasicBlock *&PredBB,
+ unsigned maxCommonTailLength) {
+ unsigned i, commonTailIndex;
+ unsigned TimeEstimate = ~0U;
+ for (i=0, commonTailIndex=0; i<SameTails.size(); i++) {
+ // Use PredBB if possible; that doesn't require a new branch.
+ if (SameTails[i].first->second==PredBB) {
+ commonTailIndex = i;
+ break;
+ }
+ // Otherwise, make a (fairly bogus) choice based on estimate of
+ // how long it will take the various blocks to execute.
+ unsigned t = EstimateRuntime(SameTails[i].first->second->begin(),
+ SameTails[i].second);
+ if (t<=TimeEstimate) {
+ TimeEstimate = t;
+ commonTailIndex = i;
+ }
+ }
+
+ MachineBasicBlock::iterator BBI = SameTails[commonTailIndex].second;
+ MachineBasicBlock *MBB = SameTails[commonTailIndex].first->second;
+
+ DEBUG(errs() << "\nSplitting " << MBB->getNumber() << ", size "
+ << maxCommonTailLength);
+
+ MachineBasicBlock *newMBB = SplitMBBAt(*MBB, BBI);
+ SameTails[commonTailIndex].first->second = newMBB;
+ SameTails[commonTailIndex].second = newMBB->begin();
+ // If we split PredBB, newMBB is the new predecessor.
+ if (PredBB==MBB)
+ PredBB = newMBB;
+
+ return commonTailIndex;
}
// See if any of the blocks in MergePotentials (which all have a common single
bool BranchFolder::TryMergeBlocks(MachineBasicBlock *SuccBB,
MachineBasicBlock* PredBB) {
- // We cannot jump to the entry block, which affects various choices below.
- MachineBasicBlock *Entry = MergePotentials.begin()->second->
- getParent()->begin();
+ bool MadeChange = false;
// It doesn't make sense to save a single instruction since tail merging
// will add a jump.
// FIXME: Ask the target to provide the threshold?
unsigned minCommonTailLength = (SuccBB ? 1 : 2) + 1;
- MadeChange = false;
- DOUT << "\nTryMergeBlocks " << MergePotentials.size();
+ DEBUG(errs() << "\nTryMergeBlocks " << MergePotentials.size() << '\n');
// Sort by hash value so that blocks with identical end sequences sort
// together.
}
// If one of the blocks is the entire common tail (and not the entry
- // block, which we can't jump to), treat all blocks with this same
- // tail at once.
- unsigned int i;
- for (i=0; i<SameTails.size(); i++) {
- MachineBasicBlock *MBB = SameTails[i].first->second;
- if (MBB->begin() == SameTails[i].second && MBB != Entry)
- break;
- }
- if (i!=SameTails.size()) {
+ // block, which we can't jump to), we can treat all blocks with this same
+ // tail at once. Use PredBB if that is one of the possibilities, as that
+ // will not introduce any extra branches.
+ MachineBasicBlock *EntryBB = MergePotentials.begin()->second->
+ getParent()->begin();
+ unsigned int commonTailIndex, i;
+ for (commonTailIndex=SameTails.size(), i=0; i<SameTails.size(); i++) {
MachineBasicBlock *MBB = SameTails[i].first->second;
- // MBB is common tail. Adjust all other BB's to jump to this one.
- // Traversal must be forwards so erases work.
- DOUT << "\nUsing common tail " << MBB->getNumber() << " for ";
- for (unsigned int j=0; j<SameTails.size(); ++j) {
- if (i==j)
- continue;
- DOUT << SameTails[j].first->second->getNumber() << ",";
- // Hack the end off BB j, making it jump to BB i instead.
- ReplaceTailWithBranchTo(SameTails[j].second, MBB);
- // This modifies BB j, so remove it from the worklist.
- MergePotentials.erase(SameTails[j].first);
+ if (MBB->begin() == SameTails[i].second && MBB != EntryBB) {
+ commonTailIndex = i;
+ if (MBB==PredBB)
+ break;
}
- DOUT << "\n";
- // We leave i in the worklist in case there are other blocks that
- // match it with a smaller number of instructions.
- MadeChange = true;
- continue;
}
- // Otherwise, merge the 2 blocks in SameTails that are latest in
- // MergePotentials; these are at indices 0 and 1 in SameTails.
- MachineBasicBlock::iterator BBI1 = (SameTails[0]).second;
- MachineBasicBlock::iterator BBI2 = (SameTails[1]).second;
- MachineBasicBlock *MBB1 = (SameTails[0]).first->second;
- MachineBasicBlock *MBB2 = (SameTails[1]).first->second;
-
- DOUT << "\nMerging " << MBB1->getNumber() << "," <<
- MBB2->getNumber() << ", size " << maxCommonTailLength;
-
- // Neither block is the entire common tail; split the tail of one block
- // to make it redundant with the other tail. We cannot jump to the
- // entry block, so if one block is the entry block, split the other one.
-
- // The second half of the split block will remain in SameTails, and will
- // consist entirely of common code. Thus in the case where there are
- // multiple blocks that would all need to be split, the next iteration of
- // the outer loop will handle all the rest of them.
-
- // Decide whether we want to split MBB1 or MBB2.
- if (ShouldSplitFirstBlock(MBB1, BBI1, MBB2, BBI2, PredBB)) {
- MBB1 = SplitMBBAt(*MBB1, BBI1);
- BBI1 = MBB1->begin();
- SameTails[0].first->second = MBB1;
- } else {
- MBB2 = SplitMBBAt(*MBB2, BBI2);
- BBI2 = MBB2->begin();
- SameTails[1].first->second = MBB2;
+ if (commonTailIndex==SameTails.size()) {
+ // None of the blocks consist entirely of the common tail.
+ // Split a block so that one does.
+ commonTailIndex = CreateCommonTailOnlyBlock(PredBB, maxCommonTailLength);
}
-
- if (MBB2->begin() == BBI2 && MBB2 != Entry) {
- // Hack the end off MBB1, making it jump to MBB2 instead.
- ReplaceTailWithBranchTo(BBI1, MBB2);
- // This modifies MBB1, so remove it from the worklist.
- MergePotentials.erase(SameTails[0].first);
- } else {
- assert(MBB1->begin() == BBI1 && MBB1 != Entry &&
- "Didn't split block correctly?");
- // Hack the end off MBB2, making it jump to MBB1 instead.
- ReplaceTailWithBranchTo(BBI2, MBB1);
- // This modifies MBB2, so remove it from the worklist.
- MergePotentials.erase(SameTails[1].first);
+
+ MachineBasicBlock *MBB = SameTails[commonTailIndex].first->second;
+ // MBB is common tail. Adjust all other BB's to jump to this one.
+ // Traversal must be forwards so erases work.
+ DEBUG(errs() << "\nUsing common tail " << MBB->getNumber() << " for ");
+ for (unsigned int i=0; i<SameTails.size(); ++i) {
+ if (commonTailIndex==i)
+ continue;
+ DEBUG(errs() << SameTails[i].first->second->getNumber() << ",");
+ // Hack the end off BB i, making it jump to BB commonTailIndex instead.
+ ReplaceTailWithBranchTo(SameTails[i].second, MBB);
+ // BB i is no longer a predecessor of SuccBB; remove it from the worklist.
+ MergePotentials.erase(SameTails[i].first);
}
+ DEBUG(errs() << "\n");
+ // We leave commonTailIndex in the worklist in case there are other blocks
+ // that match it with a smaller number of instructions.
MadeChange = true;
}
return MadeChange;
if (!EnableTailMerge) return false;
- MadeChange = false;
+ bool MadeChange = false;
// First find blocks with no successors.
MergePotentials.clear();
// transformations.)
for (MachineFunction::iterator I = MF.begin(), E = MF.end(); I != E; ++I) {
- if (!I->succ_empty() && I->pred_size() >= 2 &&
- I->pred_size() < TailMergeThreshold) {
+ if (I->pred_size() >= 2 && I->pred_size() < TailMergeThreshold) {
+ SmallPtrSet<MachineBasicBlock *, 8> UniquePreds;
MachineBasicBlock *IBB = I;
MachineBasicBlock *PredBB = prior(I);
MergePotentials.clear();
// Skip blocks that loop to themselves, can't tail merge these.
if (PBB==IBB)
continue;
+ // Visit each predecessor only once.
+ if (!UniquePreds.insert(PBB))
+ continue;
MachineBasicBlock *TBB = 0, *FBB = 0;
- std::vector<MachineOperand> Cond;
- if (!TII->AnalyzeBranch(*PBB, TBB, FBB, Cond)) {
+ SmallVector<MachineOperand, 4> Cond;
+ if (!TII->AnalyzeBranch(*PBB, TBB, FBB, Cond, true)) {
// Failing case: IBB is the target of a cbr, and
// we cannot reverse the branch.
- std::vector<MachineOperand> NewCond(Cond);
+ SmallVector<MachineOperand, 4> NewCond(Cond);
if (!Cond.empty() && TBB==IBB) {
if (TII->ReverseBranchCondition(NewCond))
continue;
if (MergePotentials.size() >= 2)
MadeChange |= TryMergeBlocks(I, PredBB);
// Reinsert an unconditional branch if needed.
- // The 1 below can be either an original single predecessor, or a result
- // of removing blocks in TryMergeBlocks.
+ // The 1 below can occur as a result of removing blocks in TryMergeBlocks.
PredBB = prior(I); // this may have been changed in TryMergeBlocks
if (MergePotentials.size()==1 &&
- (MergePotentials.begin())->second != PredBB)
- FixTail((MergePotentials.begin())->second, I, TII);
+ MergePotentials.begin()->second != PredBB)
+ FixTail(MergePotentials.begin()->second, I, TII);
}
}
return MadeChange;
//===----------------------------------------------------------------------===//
bool BranchFolder::OptimizeBranches(MachineFunction &MF) {
- MadeChange = false;
+ bool MadeChange = false;
// Make sure blocks are numbered in order
MF.RenumberBlocks();
for (MachineFunction::iterator I = ++MF.begin(), E = MF.end(); I != E; ) {
MachineBasicBlock *MBB = I++;
- OptimizeBlock(MBB);
+ MadeChange |= OptimizeBlock(MBB);
// If it is dead, remove it.
if (MBB->pred_empty()) {
///
bool BranchFolder::CanFallThrough(MachineBasicBlock *CurBB,
bool BranchUnAnalyzable,
- MachineBasicBlock *TBB, MachineBasicBlock *FBB,
- const std::vector<MachineOperand> &Cond) {
+ MachineBasicBlock *TBB,
+ MachineBasicBlock *FBB,
+ const SmallVectorImpl<MachineOperand> &Cond) {
MachineFunction::iterator Fallthrough = CurBB;
++Fallthrough;
// If FallthroughBlock is off the end of the function, it can't fall through.
///
bool BranchFolder::CanFallThrough(MachineBasicBlock *CurBB) {
MachineBasicBlock *TBB = 0, *FBB = 0;
- std::vector<MachineOperand> Cond;
- bool CurUnAnalyzable = TII->AnalyzeBranch(*CurBB, TBB, FBB, Cond);
+ SmallVector<MachineOperand, 4> Cond;
+ bool CurUnAnalyzable = TII->AnalyzeBranch(*CurBB, TBB, FBB, Cond, true);
return CanFallThrough(CurBB, CurUnAnalyzable, TBB, FBB, Cond);
}
/// OptimizeBlock - Analyze and optimize control flow related to the specified
/// block. This is never called on the entry block.
-void BranchFolder::OptimizeBlock(MachineBasicBlock *MBB) {
+bool BranchFolder::OptimizeBlock(MachineBasicBlock *MBB) {
+ bool MadeChange = false;
+
MachineFunction::iterator FallThrough = MBB;
++FallThrough;
// points to this block.
if (MBB->empty() && !MBB->isLandingPad()) {
// Dead block? Leave for cleanup later.
- if (MBB->pred_empty()) return;
+ if (MBB->pred_empty()) return MadeChange;
if (FallThrough == MBB->getParent()->end()) {
// TODO: Simplify preds to not branch here if possible!
MachineBasicBlock *Pred = *(MBB->pred_end()-1);
Pred->ReplaceUsesOfBlockWith(MBB, FallThrough);
}
-
// If MBB was the target of a jump table, update jump tables to go to the
// fallthrough instead.
MBB->getParent()->getJumpTableInfo()->
ReplaceMBBInJumpTables(MBB, FallThrough);
MadeChange = true;
}
- return;
+ return MadeChange;
}
// Check to see if we can simplify the terminator of the block before this
MachineBasicBlock &PrevBB = *prior(MachineFunction::iterator(MBB));
MachineBasicBlock *PriorTBB = 0, *PriorFBB = 0;
- std::vector<MachineOperand> PriorCond;
+ SmallVector<MachineOperand, 4> PriorCond;
bool PriorUnAnalyzable =
- TII->AnalyzeBranch(PrevBB, PriorTBB, PriorFBB, PriorCond);
+ TII->AnalyzeBranch(PrevBB, PriorTBB, PriorFBB, PriorCond, true);
if (!PriorUnAnalyzable) {
// If the CFG for the prior block has extra edges, remove them.
MadeChange |= PrevBB.CorrectExtraCFGEdges(PriorTBB, PriorFBB,
// if the branch condition is reversible, reverse the branch to create a
// fall-through.
if (PriorTBB == MBB) {
- std::vector<MachineOperand> NewPriorCond(PriorCond);
+ SmallVector<MachineOperand, 4> NewPriorCond(PriorCond);
if (!TII->ReverseBranchCondition(NewPriorCond)) {
TII->RemoveBranch(PrevBB);
TII->InsertBranch(PrevBB, PriorFBB, 0, NewPriorCond);
}
}
- // If this block doesn't fall through (e.g. it ends with an uncond branch or
- // has no successors) and if the pred falls through into this block, and if
- // it would otherwise fall through into the block after this, move this
- // block to the end of the function.
+ // If this block has no successors (e.g. it is a return block or ends with
+ // a call to a no-return function like abort or __cxa_throw) and if the pred
+ // falls through into this block, and if it would otherwise fall through
+ // into the block after this, move this block to the end of the function.
//
// We consider it more likely that execution will stay in the function (e.g.
// due to loops) than it is to exit it. This asserts in loops etc, moving
// the assert condition out of the loop body.
- if (!PriorCond.empty() && PriorFBB == 0 &&
+ if (MBB->succ_empty() && !PriorCond.empty() && PriorFBB == 0 &&
MachineFunction::iterator(PriorTBB) == FallThrough &&
!CanFallThrough(MBB)) {
bool DoTransform = true;
if (DoTransform) {
// Reverse the branch so we will fall through on the previous true cond.
- std::vector<MachineOperand> NewPriorCond(PriorCond);
+ SmallVector<MachineOperand, 4> NewPriorCond(PriorCond);
if (!TII->ReverseBranchCondition(NewPriorCond)) {
- DOUT << "\nMoving MBB: " << *MBB;
- DOUT << "To make fallthrough to: " << *PriorTBB << "\n";
+ DEBUG(errs() << "\nMoving MBB: " << *MBB
+ << "To make fallthrough to: " << *PriorTBB << "\n");
TII->RemoveBranch(PrevBB);
TII->InsertBranch(PrevBB, MBB, 0, NewPriorCond);
MBB->moveAfter(--MBB->getParent()->end());
MadeChange = true;
++NumBranchOpts;
- return;
+ return MadeChange;
}
}
}
// Analyze the branch in the current block.
MachineBasicBlock *CurTBB = 0, *CurFBB = 0;
- std::vector<MachineOperand> CurCond;
- bool CurUnAnalyzable = TII->AnalyzeBranch(*MBB, CurTBB, CurFBB, CurCond);
+ SmallVector<MachineOperand, 4> CurCond;
+ bool CurUnAnalyzable= TII->AnalyzeBranch(*MBB, CurTBB, CurFBB, CurCond, true);
if (!CurUnAnalyzable) {
// If the CFG for the prior block has extra edges, remove them.
MadeChange |= MBB->CorrectExtraCFGEdges(CurTBB, CurFBB, !CurCond.empty());
// we want:
// Loop: xxx; jncc Loop; jmp Out
if (CurTBB && CurFBB && CurFBB == MBB && CurTBB != MBB) {
- std::vector<MachineOperand> NewCond(CurCond);
+ SmallVector<MachineOperand, 4> NewCond(CurCond);
if (!TII->ReverseBranchCondition(NewCond)) {
TII->RemoveBranch(*MBB);
TII->InsertBranch(*MBB, CurFBB, CurTBB, NewCond);
} else {
DidChange = true;
PMBB->ReplaceUsesOfBlockWith(MBB, CurTBB);
+ // If this change resulted in PMBB ending in a conditional
+ // branch where both conditions go to the same destination,
+ // change this to an unconditional branch (and fix the CFG).
+ MachineBasicBlock *NewCurTBB = 0, *NewCurFBB = 0;
+ SmallVector<MachineOperand, 4> NewCurCond;
+ bool NewCurUnAnalyzable = TII->AnalyzeBranch(*PMBB, NewCurTBB,
+ NewCurFBB, NewCurCond, true);
+ if (!NewCurUnAnalyzable && NewCurTBB && NewCurTBB == NewCurFBB) {
+ TII->RemoveBranch(*PMBB);
+ NewCurCond.clear();
+ TII->InsertBranch(*PMBB, NewCurTBB, 0, NewCurCond);
+ MadeChange = true;
+ ++NumBranchOpts;
+ PMBB->CorrectExtraCFGEdges(NewCurTBB, NewCurFBB, false);
+ }
}
}
if (DidChange) {
++NumBranchOpts;
MadeChange = true;
- if (!HasBranchToSelf) return;
+ if (!HasBranchToSelf) return MadeChange;
}
}
}
PrevBB.isSuccessor(FallThrough)) {
MBB->moveAfter(--MBB->getParent()->end());
MadeChange = true;
- return;
+ return MadeChange;
}
}
}
+
+ return MadeChange;
}
projects / oota-llvm.git / blobdiff