//
//===----------------------------------------------------------------------===//
//
-// This pass performs 'jump threading', which looks at blocks that have multiple
-// predecessors and multiple successors. If one or more of the predecessors of
-// the block can be proven to always jump to one of the successors, we forward
-// the edge from the predecessor to the successor by duplicating the contents of
-// this block.
+// This file implements the Jump Threading pass.
//
//===----------------------------------------------------------------------===//
#define DEBUG_TYPE "jump-threading"
#include "llvm/Transforms/Scalar.h"
+#include "llvm/IntrinsicInst.h"
#include "llvm/Pass.h"
+#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/Statistic.h"
+#include "llvm/Transforms/Utils/BasicBlockUtils.h"
+#include "llvm/Transforms/Utils/Local.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Compiler.h"
+#include "llvm/Support/Debug.h"
using namespace llvm;
-//STATISTIC(NumThreads, "Number of jumps threaded");
+STATISTIC(NumThreads, "Number of jumps threaded");
+STATISTIC(NumFolds, "Number of terminators folded");
+
+static cl::opt<unsigned>
+Threshold("jump-threading-threshold",
+ cl::desc("Max block size to duplicate for jump threading"),
+ cl::init(6), cl::Hidden);
namespace {
- cl::opt<unsigned>
- Threshold("jump-threading-threshold",
- cl::desc("Max block size to duplicate for jump threading"),
- cl::init(6), cl::Hidden);
+ /// This pass performs 'jump threading', which looks at blocks that have
+ /// multiple predecessors and multiple successors. If one or more of the
+ /// predecessors of the block can be proven to always jump to one of the
+ /// successors, we forward the edge from the predecessor to the successor by
+ /// duplicating the contents of this block.
+ ///
+ /// An example of when this can occur is code like this:
+ ///
+ /// if () { ...
+ /// X = 4;
+ /// }
+ /// if (X < 3) {
+ ///
+ /// In this case, the unconditional branch at the end of the first if can be
+ /// revectored to the false side of the second if.
+ ///
class VISIBILITY_HIDDEN JumpThreading : public FunctionPass {
public:
static char ID; // Pass identification
JumpThreading() : FunctionPass((intptr_t)&ID) {}
bool runOnFunction(Function &F);
+ bool ThreadBlock(BasicBlock *BB);
+ void ThreadEdge(BasicBlock *BB, BasicBlock *PredBB, BasicBlock *SuccBB);
+ BasicBlock *FactorCommonPHIPreds(PHINode *PN, Constant *CstVal);
+
+ bool ProcessJumpOnPHI(PHINode *PN);
+ bool ProcessBranchOnLogical(Value *V, BasicBlock *BB, bool isAnd);
+ bool ProcessBranchOnCompare(CmpInst *Cmp, BasicBlock *BB);
};
- char JumpThreading::ID = 0;
- RegisterPass<JumpThreading> X("jump-threading", "Jump Threading");
}
+char JumpThreading::ID = 0;
+static RegisterPass<JumpThreading>
+X("jump-threading", "Jump Threading");
+
// Public interface to the Jump Threading pass
FunctionPass *llvm::createJumpThreadingPass() { return new JumpThreading(); }
/// runOnFunction - Top level algorithm.
///
bool JumpThreading::runOnFunction(Function &F) {
- bool Changed = false;
- return Changed;
+ DOUT << "Jump threading on function '" << F.getNameStart() << "'\n";
+
+ bool AnotherIteration = true, EverChanged = false;
+ while (AnotherIteration) {
+ AnotherIteration = false;
+ bool Changed = false;
+ for (Function::iterator I = F.begin(), E = F.end(); I != E; ++I)
+ while (ThreadBlock(I))
+ Changed = true;
+ AnotherIteration = Changed;
+ EverChanged |= Changed;
+ }
+ return EverChanged;
+}
+
+/// FactorCommonPHIPreds - If there are multiple preds with the same incoming
+/// value for the PHI, factor them together so we get one block to thread for
+/// the whole group.
+/// This is important for things like "phi i1 [true, true, false, true, x]"
+/// where we only need to clone the block for the true blocks once.
+///
+BasicBlock *JumpThreading::FactorCommonPHIPreds(PHINode *PN, Constant *CstVal) {
+ SmallVector<BasicBlock*, 16> CommonPreds;
+ for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i)
+ if (PN->getIncomingValue(i) == CstVal)
+ CommonPreds.push_back(PN->getIncomingBlock(i));
+
+ if (CommonPreds.size() == 1)
+ return CommonPreds[0];
+
+ DOUT << " Factoring out " << CommonPreds.size()
+ << " common predecessors.\n";
+ return SplitBlockPredecessors(PN->getParent(),
+ &CommonPreds[0], CommonPreds.size(),
+ ".thr_comm", this);
+}
+
+
+/// getJumpThreadDuplicationCost - Return the cost of duplicating this block to
+/// thread across it.
+static unsigned getJumpThreadDuplicationCost(const BasicBlock *BB) {
+ BasicBlock::const_iterator I = BB->begin();
+ /// Ignore PHI nodes, these will be flattened when duplication happens.
+ while (isa<PHINode>(*I)) ++I;
+
+ // Sum up the cost of each instruction until we get to the terminator. Don't
+ // include the terminator because the copy won't include it.
+ unsigned Size = 0;
+ for (; !isa<TerminatorInst>(I); ++I) {
+ // Debugger intrinsics don't incur code size.
+ if (isa<DbgInfoIntrinsic>(I)) continue;
+
+ // If this is a pointer->pointer bitcast, it is free.
+ if (isa<BitCastInst>(I) && isa<PointerType>(I->getType()))
+ continue;
+
+ // All other instructions count for at least one unit.
+ ++Size;
+
+ // Calls are more expensive. If they are non-intrinsic calls, we model them
+ // as having cost of 4. If they are a non-vector intrinsic, we model them
+ // as having cost of 2 total, and if they are a vector intrinsic, we model
+ // them as having cost 1.
+ if (const CallInst *CI = dyn_cast<CallInst>(I)) {
+ if (!isa<IntrinsicInst>(CI))
+ Size += 3;
+ else if (isa<VectorType>(CI->getType()))
+ Size += 1;
+ }
+ }
+
+ // Threading through a switch statement is particularly profitable. If this
+ // block ends in a switch, decrease its cost to make it more likely to happen.
+ if (isa<SwitchInst>(I))
+ Size = Size > 6 ? Size-6 : 0;
+
+ return Size;
+}
+
+
+/// ThreadBlock - If there are any predecessors whose control can be threaded
+/// through to a successor, transform them now.
+bool JumpThreading::ThreadBlock(BasicBlock *BB) {
+ // See if this block ends with a branch of switch. If so, see if the
+ // condition is a phi node. If so, and if an entry of the phi node is a
+ // constant, we can thread the block.
+ Value *Condition;
+ if (BranchInst *BI = dyn_cast<BranchInst>(BB->getTerminator())) {
+ // Can't thread an unconditional jump.
+ if (BI->isUnconditional()) return false;
+ Condition = BI->getCondition();
+ } else if (SwitchInst *SI = dyn_cast<SwitchInst>(BB->getTerminator()))
+ Condition = SI->getCondition();
+ else
+ return false; // Must be an invoke.
+
+ // If the terminator of this block is branching on a constant, simplify the
+ // terminator to an unconditional branch. This can occur due to threading in
+ // other blocks.
+ if (isa<ConstantInt>(Condition)) {
+ DOUT << " In block '" << BB->getNameStart()
+ << "' folding terminator: " << *BB->getTerminator();
+ ++NumFolds;
+ ConstantFoldTerminator(BB);
+ return true;
+ }
+
+ // If there is only a single predecessor of this block, nothing to fold.
+ if (BB->getSinglePredecessor())
+ return false;
+
+ // See if this is a phi node in the current block.
+ PHINode *PN = dyn_cast<PHINode>(Condition);
+ if (PN && PN->getParent() == BB)
+ return ProcessJumpOnPHI(PN);
+
+ // If this is a conditional branch whose condition is and/or of a phi, try to
+ // simplify it.
+ if (BinaryOperator *CondI = dyn_cast<BinaryOperator>(Condition)) {
+ if ((CondI->getOpcode() == Instruction::And ||
+ CondI->getOpcode() == Instruction::Or) &&
+ isa<BranchInst>(BB->getTerminator()) &&
+ ProcessBranchOnLogical(CondI, BB,
+ CondI->getOpcode() == Instruction::And))
+ return true;
+ }
+
+ // If we have "br (phi != 42)" and the phi node has any constant values as
+ // operands, we can thread through this block.
+ if (CmpInst *CondCmp = dyn_cast<CmpInst>(Condition))
+ if (isa<PHINode>(CondCmp->getOperand(0)) &&
+ isa<Constant>(CondCmp->getOperand(1)) &&
+ ProcessBranchOnCompare(CondCmp, BB))
+ return true;
+
+ return false;
+}
+
+/// ProcessJumpOnPHI - We have a conditional branch of switch on a PHI node in
+/// the current block. See if there are any simplifications we can do based on
+/// inputs to the phi node.
+///
+bool JumpThreading::ProcessJumpOnPHI(PHINode *PN) {
+ // See if the phi node has any constant values. If so, we can determine where
+ // the corresponding predecessor will branch.
+ ConstantInt *PredCst = 0;
+ for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i)
+ if ((PredCst = dyn_cast<ConstantInt>(PN->getIncomingValue(i))))
+ break;
+
+ // If no incoming value has a constant, we don't know the destination of any
+ // predecessors.
+ if (PredCst == 0)
+ return false;
+
+ // See if the cost of duplicating this block is low enough.
+ BasicBlock *BB = PN->getParent();
+ unsigned JumpThreadCost = getJumpThreadDuplicationCost(BB);
+ if (JumpThreadCost > Threshold) {
+ DOUT << " Not threading BB '" << BB->getNameStart()
+ << "' - Cost is too high: " << JumpThreadCost << "\n";
+ return false;
+ }
+
+ // If so, we can actually do this threading. Merge any common predecessors
+ // that will act the same.
+ BasicBlock *PredBB = FactorCommonPHIPreds(PN, PredCst);
+
+ // Next, figure out which successor we are threading to.
+ BasicBlock *SuccBB;
+ if (BranchInst *BI = dyn_cast<BranchInst>(BB->getTerminator()))
+ SuccBB = BI->getSuccessor(PredCst == ConstantInt::getFalse());
+ else {
+ SwitchInst *SI = cast<SwitchInst>(BB->getTerminator());
+ SuccBB = SI->getSuccessor(SI->findCaseValue(PredCst));
+ }
+
+ // If threading to the same block as we come from, we would infinite loop.
+ if (SuccBB == BB) {
+ DOUT << " Not threading BB '" << BB->getNameStart()
+ << "' - would thread to self!\n";
+ return false;
+ }
+
+ // And finally, do it!
+ DOUT << " Threading edge from '" << PredBB->getNameStart() << "' to '"
+ << SuccBB->getNameStart() << "' with cost: " << JumpThreadCost
+ << ", across block:\n "
+ << *BB << "\n";
+
+ ThreadEdge(BB, PredBB, SuccBB);
+ ++NumThreads;
+ return true;
+}
+
+/// ProcessJumpOnLogicalPHI - PN's basic block contains a conditional branch
+/// whose condition is an AND/OR where one side is PN. If PN has constant
+/// operands that permit us to evaluate the condition for some operand, thread
+/// through the block. For example with:
+/// br (and X, phi(Y, Z, false))
+/// the predecessor corresponding to the 'false' will always jump to the false
+/// destination of the branch.
+///
+bool JumpThreading::ProcessBranchOnLogical(Value *V, BasicBlock *BB,
+ bool isAnd) {
+ // If this is a binary operator tree of the same AND/OR opcode, check the
+ // LHS/RHS.
+ if (BinaryOperator *BO = dyn_cast<BinaryOperator>(V))
+ if (isAnd && BO->getOpcode() == Instruction::And ||
+ !isAnd && BO->getOpcode() == Instruction::Or) {
+ if (ProcessBranchOnLogical(BO->getOperand(0), BB, isAnd))
+ return true;
+ if (ProcessBranchOnLogical(BO->getOperand(1), BB, isAnd))
+ return true;
+ }
+
+ // If this isn't a PHI node, we can't handle it.
+ PHINode *PN = dyn_cast<PHINode>(V);
+ if (!PN || PN->getParent() != BB) return false;
+
+ // We can only do the simplification for phi nodes of 'false' with AND or
+ // 'true' with OR. See if we have any entries in the phi for this.
+ unsigned PredNo = ~0U;
+ ConstantInt *PredCst = ConstantInt::get(Type::Int1Ty, !isAnd);
+ for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i) {
+ if (PN->getIncomingValue(i) == PredCst) {
+ PredNo = i;
+ break;
+ }
+ }
+
+ // If no match, bail out.
+ if (PredNo == ~0U)
+ return false;
+
+ // See if the cost of duplicating this block is low enough.
+ unsigned JumpThreadCost = getJumpThreadDuplicationCost(BB);
+ if (JumpThreadCost > Threshold) {
+ DOUT << " Not threading BB '" << BB->getNameStart()
+ << "' - Cost is too high: " << JumpThreadCost << "\n";
+ return false;
+ }
+
+ // If so, we can actually do this threading. Merge any common predecessors
+ // that will act the same.
+ BasicBlock *PredBB = FactorCommonPHIPreds(PN, PredCst);
+
+ // Next, figure out which successor we are threading to. If this was an AND,
+ // the constant must be FALSE, and we must be targeting the 'false' block.
+ // If this is an OR, the constant must be TRUE, and we must be targeting the
+ // 'true' block.
+ BasicBlock *SuccBB = BB->getTerminator()->getSuccessor(isAnd);
+
+ // If threading to the same block as we come from, we would infinite loop.
+ if (SuccBB == BB) {
+ DOUT << " Not threading BB '" << BB->getNameStart()
+ << "' - would thread to self!\n";
+ return false;
+ }
+
+ // And finally, do it!
+ DOUT << " Threading edge through bool from '" << PredBB->getNameStart()
+ << "' to '" << SuccBB->getNameStart() << "' with cost: "
+ << JumpThreadCost << ", across block:\n "
+ << *BB << "\n";
+
+ ThreadEdge(BB, PredBB, SuccBB);
+ ++NumThreads;
+ return true;
+}
+
+/// ProcessBranchOnCompare - We found a branch on a comparison between a phi
+/// node and a constant. If the PHI node contains any constants as inputs, we
+/// can fold the compare for that edge and thread through it.
+bool JumpThreading::ProcessBranchOnCompare(CmpInst *Cmp, BasicBlock *BB) {
+ PHINode *PN = cast<PHINode>(Cmp->getOperand(0));
+ Constant *RHS = cast<Constant>(Cmp->getOperand(1));
+
+ // If the phi isn't in the current block, an incoming edge to this block
+ // doesn't control the destination.
+ if (PN->getParent() != BB)
+ return false;
+
+ // We can do this simplification if any comparisons fold to true or false.
+ // See if any do.
+ Constant *PredCst = 0;
+ bool TrueDirection = false;
+ for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i) {
+ PredCst = dyn_cast<Constant>(PN->getIncomingValue(i));
+ if (PredCst == 0) continue;
+
+ Constant *Res;
+ if (ICmpInst *ICI = dyn_cast<ICmpInst>(Cmp))
+ Res = ConstantExpr::getICmp(ICI->getPredicate(), PredCst, RHS);
+ else
+ Res = ConstantExpr::getFCmp(cast<FCmpInst>(Cmp)->getPredicate(),
+ PredCst, RHS);
+ // If this folded to a constant expr, we can't do anything.
+ if (ConstantInt *ResC = dyn_cast<ConstantInt>(Res)) {
+ TrueDirection = ResC->getZExtValue();
+ break;
+ }
+ // If this folded to undef, just go the false way.
+ if (isa<UndefValue>(Res)) {
+ TrueDirection = false;
+ break;
+ }
+
+ // Otherwise, we can't fold this input.
+ PredCst = 0;
+ }
+
+ // If no match, bail out.
+ if (PredCst == 0)
+ return false;
+
+ // See if the cost of duplicating this block is low enough.
+ unsigned JumpThreadCost = getJumpThreadDuplicationCost(BB);
+ if (JumpThreadCost > Threshold) {
+ DOUT << " Not threading BB '" << BB->getNameStart()
+ << "' - Cost is too high: " << JumpThreadCost << "\n";
+ return false;
+ }
+
+ // If so, we can actually do this threading. Merge any common predecessors
+ // that will act the same.
+ BasicBlock *PredBB = FactorCommonPHIPreds(PN, PredCst);
+
+ // Next, get our successor.
+ BasicBlock *SuccBB = BB->getTerminator()->getSuccessor(!TrueDirection);
+
+ // If threading to the same block as we come from, we would infinite loop.
+ if (SuccBB == BB) {
+ DOUT << " Not threading BB '" << BB->getNameStart()
+ << "' - would thread to self!\n";
+ return false;
+ }
+
+
+ // And finally, do it!
+ DOUT << " Threading edge through bool from '" << PredBB->getNameStart()
+ << "' to '" << SuccBB->getNameStart() << "' with cost: "
+ << JumpThreadCost << ", across block:\n "
+ << *BB << "\n";
+
+ ThreadEdge(BB, PredBB, SuccBB);
+ ++NumThreads;
+ return true;
+}
+
+
+/// ThreadEdge - We have decided that it is safe and profitable to thread an
+/// edge from PredBB to SuccBB across BB. Transform the IR to reflect this
+/// change.
+void JumpThreading::ThreadEdge(BasicBlock *BB, BasicBlock *PredBB,
+ BasicBlock *SuccBB) {
+
+ // Jump Threading can not update SSA properties correctly if the values
+ // defined in the duplicated block are used outside of the block itself. For
+ // this reason, we spill all values that are used outside of BB to the stack.
+ for (BasicBlock::iterator I = BB->begin(); I != BB->end(); ++I) {
+ if (!I->isUsedOutsideOfBlock(BB))
+ continue;
+
+ // We found a use of I outside of BB. Create a new stack slot to
+ // break this inter-block usage pattern.
+ if (!isa<StructType>(I->getType())) {
+ DemoteRegToStack(*I);
+ continue;
+ }
+
+ // Alternatively, I must be a call or invoke that returns multiple retvals.
+ // We can't use 'DemoteRegToStack' because that will create loads and
+ // stores of aggregates which is not valid yet. If I is a call, we can just
+ // pull all the getresult instructions up to this block. If I is an invoke,
+ // we are out of luck.
+ BasicBlock::iterator IP = I; ++IP;
+ for (Value::use_iterator UI = I->use_begin(), E = I->use_end();
+ UI != E; ++UI)
+ cast<GetResultInst>(UI)->moveBefore(IP);
+ }
+
+ // We are going to have to map operands from the original BB block to the new
+ // copy of the block 'NewBB'. If there are PHI nodes in BB, evaluate them to
+ // account for entry from PredBB.
+ DenseMap<Instruction*, Value*> ValueMapping;
+
+ BasicBlock *NewBB =
+ BasicBlock::Create(BB->getName()+".thread", BB->getParent(), BB);
+ NewBB->moveAfter(PredBB);
+
+ BasicBlock::iterator BI = BB->begin();
+ for (; PHINode *PN = dyn_cast<PHINode>(BI); ++BI)
+ ValueMapping[PN] = PN->getIncomingValueForBlock(PredBB);
+
+ // Clone the non-phi instructions of BB into NewBB, keeping track of the
+ // mapping and using it to remap operands in the cloned instructions.
+ for (; !isa<TerminatorInst>(BI); ++BI) {
+ Instruction *New = BI->clone();
+ New->setName(BI->getNameStart());
+ NewBB->getInstList().push_back(New);
+ ValueMapping[BI] = New;
+
+ // Remap operands to patch up intra-block references.
+ for (unsigned i = 0, e = New->getNumOperands(); i != e; ++i)
+ if (Instruction *Inst = dyn_cast<Instruction>(New->getOperand(i)))
+ if (Value *Remapped = ValueMapping[Inst])
+ New->setOperand(i, Remapped);
+ }
+
+ // We didn't copy the terminator from BB over to NewBB, because there is now
+ // an unconditional jump to SuccBB. Insert the unconditional jump.
+ BranchInst::Create(SuccBB, NewBB);
+
+ // Check to see if SuccBB has PHI nodes. If so, we need to add entries to the
+ // PHI nodes for NewBB now.
+ for (BasicBlock::iterator PNI = SuccBB->begin(); isa<PHINode>(PNI); ++PNI) {
+ PHINode *PN = cast<PHINode>(PNI);
+ // Ok, we have a PHI node. Figure out what the incoming value was for the
+ // DestBlock.
+ Value *IV = PN->getIncomingValueForBlock(BB);
+
+ // Remap the value if necessary.
+ if (Instruction *Inst = dyn_cast<Instruction>(IV))
+ if (Value *MappedIV = ValueMapping[Inst])
+ IV = MappedIV;
+ PN->addIncoming(IV, NewBB);
+ }
+
+ // Finally, NewBB is good to go. Update the terminator of PredBB to jump to
+ // NewBB instead of BB. This eliminates predecessors from BB, which requires
+ // us to simplify any PHI nodes in BB.
+ TerminatorInst *PredTerm = PredBB->getTerminator();
+ for (unsigned i = 0, e = PredTerm->getNumSuccessors(); i != e; ++i)
+ if (PredTerm->getSuccessor(i) == BB) {
+ BB->removePredecessor(PredBB);
+ PredTerm->setSuccessor(i, NewBB);
+ }
}