Gericom/teak-llvm
1
Fork 0
mirror of https://github.com/Gericom/teak-llvm.git synced 2025-06-19 19:45:40 -04:00
Code Issues Actions Packages Projects Releases Wiki Activity

[LoopRotate] add ability to repeat loop rotation until non-deoptimizing exit is found

Browse Source 
In case of loops with multiple exit where all-but-one exit are deoptimizing
it might happen that the first rotation will end up with latch having a deoptimizing
exit. This makes the loop unsuitable for trip-count analysis (say, getLoopEstimatedTripCount)
as well as for loop transformations that know how to handle multple deoptimizing exits.

It pretty much means that canonical form in multple-deoptimizing-exits case should be
with non-deoptimizing exit at latch.
Teach loop-rotation to reach this canonical form by repeating rotation.

-loop-rotate-multi option introduced to control this behavior, currently disabled by default.

Reviewers: skatkov, asbirlea, reames, fhahn
Reviewed By: skatkov

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D73058
This commit is contained in:
Fedor Sergeev 2020-01-23 15:55:32 +03:00
parent 279fa8e006
commit 2f6987ba61
4 changed files with 670 additions and 271 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

609
llvm/lib/Transforms/Utils/LoopRotationUtils.cpp
View File

@ -46,6 +46,11 @@ using namespace llvm;
STATISTIC(NumRotated, "Number of loops rotated"); STATISTIC(NumRotated, "Number of loops rotated");
static cl::opt<bool>
MultiRotate("loop-rotate-multi", cl::init(false), cl::Hidden,
cl::desc("Allow loop rotation multiple times in order to reach "
"a better latch exit"));
namespace { namespace {
/// A simple loop rotation transformation. /// A simple loop rotation transformation.
class LoopRotate { class LoopRotate {
@ -177,14 +182,16 @@ static void RewriteUsesOfClonedInstructions(BasicBlock *OrigHeader,
} }
} }
// Look for a phi which is only used outside the loop (via a LCSSA phi) // Assuming both header and latch are exiting, look for a phi which is only
// in the exit from the header. This means that rotating the loop can // used outside the loop (via a LCSSA phi) in the exit from the header.
// remove the phi. // This means that rotating the loop can remove the phi.
static bool shouldRotateLoopExitingLatch(Loop *L) { static bool profitableToRotateLoopExitingLatch(Loop *L) {
BasicBlock *Header = L->getHeader(); BasicBlock *Header = L->getHeader();
BasicBlock *HeaderExit = Header->getTerminator()->getSuccessor(0); BranchInst *BI = dyn_cast<BranchInst>(Header->getTerminator());
assert(BI && BI->isConditional() && "need header with conditional exit");
BasicBlock *HeaderExit = BI->getSuccessor(0);
if (L->contains(HeaderExit)) if (L->contains(HeaderExit))
HeaderExit = Header->getTerminator()->getSuccessor(1); HeaderExit = BI->getSuccessor(1);
for (auto &Phi : Header->phis()) { for (auto &Phi : Header->phis()) {
// Look for uses of this phi in the loop/via exits other than the header. // Look for uses of this phi in the loop/via exits other than the header.
@ -194,7 +201,50 @@ static bool shouldRotateLoopExitingLatch(Loop *L) {
continue; continue;
return true; return true;
} }
return false;
}
// Check that latch exit is deoptimizing (which means - very unlikely to happen)
// and there is another exit from the loop which is non-deoptimizing.
// If we rotate latch to that exit our loop has a better chance of being fully
// canonical.
//
// It can give false positives in some rare cases.
static bool canRotateDeoptimizingLatchExit(Loop *L) {
BasicBlock *Latch = L->getLoopLatch();
assert(Latch && "need latch");
BranchInst *BI = dyn_cast<BranchInst>(Latch->getTerminator());
// Need normal exiting latch.
if (!BI || !BI->isConditional())
return false;
BasicBlock *Exit = BI->getSuccessor(1);
if (L->contains(Exit))
Exit = BI->getSuccessor(0);
// Latch exit is non-deoptimizing, no need to rotate.
if (!Exit->getPostdominatingDeoptimizeCall())
return false;
SmallVector<BasicBlock *, 4> Exits;
L->getUniqueExitBlocks(Exits);
if (!Exits.empty()) {
// There is at least one non-deoptimizing exit.
//
// Note, that BasicBlock::getPostdominatingDeoptimizeCall is not exact,
// as it can conservatively return false for deoptimizing exits with
// complex enough control flow down to deoptimize call.
//
// That means here we can report success for a case where
// all exits are deoptimizing but one of them has complex enough
// control flow (e.g. with loops).
//
// That should be a very rare case and false positives for this function
// have compile-time effect only.
return any_of(Exits, [](const BasicBlock *BB) {
return !BB->getPostdominatingDeoptimizeCall();
});
}
return false; return false;
} }
@ -208,319 +258,336 @@ static bool shouldRotateLoopExitingLatch(Loop *L) {
/// rotation. LoopRotate should be repeatable and converge to a canonical /// rotation. LoopRotate should be repeatable and converge to a canonical
/// form. This property is satisfied because simplifying the loop latch can only /// form. This property is satisfied because simplifying the loop latch can only
/// happen once across multiple invocations of the LoopRotate pass. /// happen once across multiple invocations of the LoopRotate pass.
///
/// If -loop-rotate-multi is enabled we can do multiple rotations in one go
/// so to reach a suitable (non-deoptimizing) exit.
bool LoopRotate::rotateLoop(Loop *L, bool SimplifiedLatch) { bool LoopRotate::rotateLoop(Loop *L, bool SimplifiedLatch) {
// If the loop has only one block then there is not much to rotate. // If the loop has only one block then there is not much to rotate.
if (L->getBlocks().size() == 1) if (L->getBlocks().size() == 1)
return false; return false;
BasicBlock *OrigHeader = L->getHeader(); bool Rotated = false;
BasicBlock *OrigLatch = L->getLoopLatch(); do {
BasicBlock *OrigHeader = L->getHeader();
BasicBlock *OrigLatch = L->getLoopLatch();
BranchInst *BI = dyn_cast<BranchInst>(OrigHeader->getTerminator()); BranchInst *BI = dyn_cast<BranchInst>(OrigHeader->getTerminator());
if (!BI || BI->isUnconditional()) if (!BI || BI->isUnconditional())
return false; return Rotated;
// If the loop header is not one of the loop exiting blocks then // If the loop header is not one of the loop exiting blocks then
// either this loop is already rotated or it is not // either this loop is already rotated or it is not
// suitable for loop rotation transformations. // suitable for loop rotation transformations.
if (!L->isLoopExiting(OrigHeader)) if (!L->isLoopExiting(OrigHeader))
return false; return Rotated;
// If the loop latch already contains a branch that leaves the loop then the // If the loop latch already contains a branch that leaves the loop then the
// loop is already rotated. // loop is already rotated.
if (!OrigLatch) if (!OrigLatch)
return false; return Rotated;
// Rotate if either the loop latch does *not* exit the loop, or if the loop // Rotate if either the loop latch does *not* exit the loop, or if the loop
// latch was just simplified. Or if we think it will be profitable. // latch was just simplified. Or if we think it will be profitable.
if (L->isLoopExiting(OrigLatch) && !SimplifiedLatch && IsUtilMode == false && if (L->isLoopExiting(OrigLatch) && !SimplifiedLatch && IsUtilMode == false &&
!shouldRotateLoopExitingLatch(L)) !profitableToRotateLoopExitingLatch(L) &&
return false; !canRotateDeoptimizingLatchExit(L))
return Rotated;
// Check size of original header and reject loop if it is very big or we can't // Check size of original header and reject loop if it is very big or we can't
// duplicate blocks inside it. // duplicate blocks inside it.
{ {
SmallPtrSet<const Value *, 32> EphValues; SmallPtrSet<const Value *, 32> EphValues;
CodeMetrics::collectEphemeralValues(L, AC, EphValues); CodeMetrics::collectEphemeralValues(L, AC, EphValues);
CodeMetrics Metrics; CodeMetrics Metrics;
Metrics.analyzeBasicBlock(OrigHeader, *TTI, EphValues); Metrics.analyzeBasicBlock(OrigHeader, *TTI, EphValues);
if (Metrics.notDuplicatable) { if (Metrics.notDuplicatable) {
LLVM_DEBUG( LLVM_DEBUG(
dbgs() << "LoopRotation: NOT rotating - contains non-duplicatable" dbgs() << "LoopRotation: NOT rotating - contains non-duplicatable"
<< " instructions: "; << " instructions: ";
L->dump()); L->dump());
return false; return Rotated;
}
if (Metrics.convergent) {
LLVM_DEBUG(dbgs() << "LoopRotation: NOT rotating - contains convergent "
"instructions: ";
L->dump());
return Rotated;
}
if (Metrics.NumInsts > MaxHeaderSize)
return Rotated;
} }
if (Metrics.convergent) {
LLVM_DEBUG(dbgs() << "LoopRotation: NOT rotating - contains convergent "
"instructions: ";
L->dump());
return false;
}
if (Metrics.NumInsts > MaxHeaderSize)
return false;
}
// Now, this loop is suitable for rotation. // Now, this loop is suitable for rotation.
BasicBlock *OrigPreheader = L->getLoopPreheader(); BasicBlock *OrigPreheader = L->getLoopPreheader();
// If the loop could not be converted to canonical form, it must have an // If the loop could not be converted to canonical form, it must have an
// indirectbr in it, just give up. // indirectbr in it, just give up.
if (!OrigPreheader || !L->hasDedicatedExits()) if (!OrigPreheader || !L->hasDedicatedExits())
return false; return Rotated;
// Anything ScalarEvolution may know about this loop or the PHI nodes // Anything ScalarEvolution may know about this loop or the PHI nodes
// in its header will soon be invalidated. We should also invalidate // in its header will soon be invalidated. We should also invalidate
// all outer loops because insertion and deletion of blocks that happens // all outer loops because insertion and deletion of blocks that happens
// during the rotation may violate invariants related to backedge taken // during the rotation may violate invariants related to backedge taken
// infos in them. // infos in them.
if (SE) if (SE)
SE->forgetTopmostLoop(L); SE->forgetTopmostLoop(L);
LLVM_DEBUG(dbgs() << "LoopRotation: rotating "; L->dump()); LLVM_DEBUG(dbgs() << "LoopRotation: rotating "; L->dump());
if (MSSAU && VerifyMemorySSA) if (MSSAU && VerifyMemorySSA)
MSSAU->getMemorySSA()->verifyMemorySSA(); MSSAU->getMemorySSA()->verifyMemorySSA();
// Find new Loop header. NewHeader is a Header's one and only successor // Find new Loop header. NewHeader is a Header's one and only successor
// that is inside loop. Header's other successor is outside the // that is inside loop. Header's other successor is outside the
// loop. Otherwise loop is not suitable for rotation. // loop. Otherwise loop is not suitable for rotation.
BasicBlock *Exit = BI->getSuccessor(0); BasicBlock *Exit = BI->getSuccessor(0);
BasicBlock *NewHeader = BI->getSuccessor(1); BasicBlock *NewHeader = BI->getSuccessor(1);
if (L->contains(Exit)) if (L->contains(Exit))
std::swap(Exit, NewHeader); std::swap(Exit, NewHeader);
assert(NewHeader && "Unable to determine new loop header"); assert(NewHeader && "Unable to determine new loop header");
assert(L->contains(NewHeader) && !L->contains(Exit) && assert(L->contains(NewHeader) && !L->contains(Exit) &&
"Unable to determine loop header and exit blocks"); "Unable to determine loop header and exit blocks");
// This code assumes that the new header has exactly one predecessor. // This code assumes that the new header has exactly one predecessor.
// Remove any single-entry PHI nodes in it. // Remove any single-entry PHI nodes in it.
assert(NewHeader->getSinglePredecessor() && assert(NewHeader->getSinglePredecessor() &&
"New header doesn't have one pred!"); "New header doesn't have one pred!");
FoldSingleEntryPHINodes(NewHeader); FoldSingleEntryPHINodes(NewHeader);
// Begin by walking OrigHeader and populating ValueMap with an entry for // Begin by walking OrigHeader and populating ValueMap with an entry for
// each Instruction. // each Instruction.
BasicBlock::iterator I = OrigHeader->begin(), E = OrigHeader->end(); BasicBlock::iterator I = OrigHeader->begin(), E = OrigHeader->end();
ValueToValueMapTy ValueMap, ValueMapMSSA; ValueToValueMapTy ValueMap, ValueMapMSSA;
// For PHI nodes, the value available in OldPreHeader is just the // For PHI nodes, the value available in OldPreHeader is just the
// incoming value from OldPreHeader. // incoming value from OldPreHeader.
for (; PHINode *PN = dyn_cast<PHINode>(I); ++I) for (; PHINode *PN = dyn_cast<PHINode>(I); ++I)
InsertNewValueIntoMap(ValueMap, PN, InsertNewValueIntoMap(ValueMap, PN,
PN->getIncomingValueForBlock(OrigPreheader)); PN->getIncomingValueForBlock(OrigPreheader));
// For the rest of the instructions, either hoist to the OrigPreheader if // For the rest of the instructions, either hoist to the OrigPreheader if
// possible or create a clone in the OldPreHeader if not. // possible or create a clone in the OldPreHeader if not.
Instruction *LoopEntryBranch = OrigPreheader->getTerminator(); Instruction *LoopEntryBranch = OrigPreheader->getTerminator();
// Record all debug intrinsics preceding LoopEntryBranch to avoid duplication. // Record all debug intrinsics preceding LoopEntryBranch to avoid duplication.
using DbgIntrinsicHash = using DbgIntrinsicHash =
std::pair<std::pair<Value *, DILocalVariable *>, DIExpression *>; std::pair<std::pair<Value *, DILocalVariable *>, DIExpression *>;
auto makeHash = [](DbgVariableIntrinsic *D) -> DbgIntrinsicHash { auto makeHash = [](DbgVariableIntrinsic *D) -> DbgIntrinsicHash {
return {{D->getVariableLocation(), D->getVariable()}, D->getExpression()}; return {{D->getVariableLocation(), D->getVariable()}, D->getExpression()};
}; };
SmallDenseSet<DbgIntrinsicHash, 8> DbgIntrinsics; SmallDenseSet<DbgIntrinsicHash, 8> DbgIntrinsics;
for (auto I = std::next(OrigPreheader->rbegin()), E = OrigPreheader->rend(); for (auto I = std::next(OrigPreheader->rbegin()), E = OrigPreheader->rend();
I != E; ++I) { I != E; ++I) {
if (auto *DII = dyn_cast<DbgVariableIntrinsic>(&*I)) if (auto *DII = dyn_cast<DbgVariableIntrinsic>(&*I))
DbgIntrinsics.insert(makeHash(DII)); DbgIntrinsics.insert(makeHash(DII));
else else
break; break;
}
while (I != E) {
Instruction *Inst = &*I++;
// If the instruction's operands are invariant and it doesn't read or write
// memory, then it is safe to hoist. Doing this doesn't change the order of
// execution in the preheader, but does prevent the instruction from
// executing in each iteration of the loop. This means it is safe to hoist
// something that might trap, but isn't safe to hoist something that reads
// memory (without proving that the loop doesn't write).
if (L->hasLoopInvariantOperands(Inst) && !Inst->mayReadFromMemory() &&
!Inst->mayWriteToMemory() && !Inst->isTerminator() &&
!isa<DbgInfoIntrinsic>(Inst) && !isa<AllocaInst>(Inst)) {
Inst->moveBefore(LoopEntryBranch);
continue;
} }
// Otherwise, create a duplicate of the instruction. while (I != E) {
Instruction *C = Inst->clone(); Instruction *Inst = &*I++;
// Eagerly remap the operands of the instruction. // If the instruction's operands are invariant and it doesn't read or write
RemapInstruction(C, ValueMap, // memory, then it is safe to hoist. Doing this doesn't change the order of
RF_NoModuleLevelChanges | RF_IgnoreMissingLocals); // execution in the preheader, but does prevent the instruction from
// executing in each iteration of the loop. This means it is safe to hoist
// Avoid inserting the same intrinsic twice. // something that might trap, but isn't safe to hoist something that reads
if (auto *DII = dyn_cast<DbgVariableIntrinsic>(C)) // memory (without proving that the loop doesn't write).
if (DbgIntrinsics.count(makeHash(DII))) { if (L->hasLoopInvariantOperands(Inst) && !Inst->mayReadFromMemory() &&
C->deleteValue(); !Inst->mayWriteToMemory() && !Inst->isTerminator() &&
!isa<DbgInfoIntrinsic>(Inst) && !isa<AllocaInst>(Inst)) {
Inst->moveBefore(LoopEntryBranch);
continue; continue;
} }
// With the operands remapped, see if the instruction constant folds or is // Otherwise, create a duplicate of the instruction.
// otherwise simplifyable. This commonly occurs because the entry from PHI Instruction *C = Inst->clone();
// nodes allows icmps and other instructions to fold.
Value *V = SimplifyInstruction(C, SQ); // Eagerly remap the operands of the instruction.
if (V && LI->replacementPreservesLCSSAForm(C, V)) { RemapInstruction(C, ValueMap,
// If so, then delete the temporary instruction and stick the folded value RF_NoModuleLevelChanges | RF_IgnoreMissingLocals);
// in the map.
InsertNewValueIntoMap(ValueMap, Inst, V); // Avoid inserting the same intrinsic twice.
if (!C->mayHaveSideEffects()) { if (auto *DII = dyn_cast<DbgVariableIntrinsic>(C))
C->deleteValue(); if (DbgIntrinsics.count(makeHash(DII))) {
C = nullptr; C->deleteValue();
continue;
}
// With the operands remapped, see if the instruction constant folds or is
// otherwise simplifyable. This commonly occurs because the entry from PHI
// nodes allows icmps and other instructions to fold.
Value *V = SimplifyInstruction(C, SQ);
if (V && LI->replacementPreservesLCSSAForm(C, V)) {
// If so, then delete the temporary instruction and stick the folded value
// in the map.
InsertNewValueIntoMap(ValueMap, Inst, V);
if (!C->mayHaveSideEffects()) {
C->deleteValue();
C = nullptr;
}
} else {
InsertNewValueIntoMap(ValueMap, Inst, C);
}
if (C) {
// Otherwise, stick the new instruction into the new block!
C->setName(Inst->getName());
C->insertBefore(LoopEntryBranch);
if (auto *II = dyn_cast<IntrinsicInst>(C))
if (II->getIntrinsicID() == Intrinsic::assume)
AC->registerAssumption(II);
// MemorySSA cares whether the cloned instruction was inserted or not, and
// not whether it can be remapped to a simplified value.
if (MSSAU)
InsertNewValueIntoMap(ValueMapMSSA, Inst, C);
} }
} else {
InsertNewValueIntoMap(ValueMap, Inst, C);
} }
if (C) {
// Otherwise, stick the new instruction into the new block!
C->setName(Inst->getName());
C->insertBefore(LoopEntryBranch);
if (auto *II = dyn_cast<IntrinsicInst>(C)) // Along with all the other instructions, we just cloned OrigHeader's
if (II->getIntrinsicID() == Intrinsic::assume) // terminator into OrigPreHeader. Fix up the PHI nodes in each of OrigHeader's
AC->registerAssumption(II); // successors by duplicating their incoming values for OrigHeader.
// MemorySSA cares whether the cloned instruction was inserted or not, and for (BasicBlock *SuccBB : successors(OrigHeader))
// not whether it can be remapped to a simplified value. for (BasicBlock::iterator BI = SuccBB->begin();
if (MSSAU) PHINode *PN = dyn_cast<PHINode>(BI); ++BI)
InsertNewValueIntoMap(ValueMapMSSA, Inst, C); PN->addIncoming(PN->getIncomingValueForBlock(OrigHeader), OrigPreheader);
}
}
// Along with all the other instructions, we just cloned OrigHeader's // Now that OrigPreHeader has a clone of OrigHeader's terminator, remove
// terminator into OrigPreHeader. Fix up the PHI nodes in each of OrigHeader's // OrigPreHeader's old terminator (the original branch into the loop), and
// successors by duplicating their incoming values for OrigHeader. // remove the corresponding incoming values from the PHI nodes in OrigHeader.
for (BasicBlock *SuccBB : successors(OrigHeader)) LoopEntryBranch->eraseFromParent();
for (BasicBlock::iterator BI = SuccBB->begin();
PHINode *PN = dyn_cast<PHINode>(BI); ++BI)
PN->addIncoming(PN->getIncomingValueForBlock(OrigHeader), OrigPreheader);
// Now that OrigPreHeader has a clone of OrigHeader's terminator, remove
// OrigPreHeader's old terminator (the original branch into the loop), and
// remove the corresponding incoming values from the PHI nodes in OrigHeader.
LoopEntryBranch->eraseFromParent();
// Update MemorySSA before the rewrite call below changes the 1:1
// instruction:cloned_instruction_or_value mapping.
if (MSSAU) {
InsertNewValueIntoMap(ValueMapMSSA, OrigHeader, OrigPreheader);
MSSAU->updateForClonedBlockIntoPred(OrigHeader, OrigPreheader,
ValueMapMSSA);
}
SmallVector<PHINode*, 2> InsertedPHIs;
// If there were any uses of instructions in the duplicated block outside the
// loop, update them, inserting PHI nodes as required
RewriteUsesOfClonedInstructions(OrigHeader, OrigPreheader, ValueMap,
&InsertedPHIs);
// Attach dbg.value intrinsics to the new phis if that phi uses a value that
// previously had debug metadata attached. This keeps the debug info
// up-to-date in the loop body.
if (!InsertedPHIs.empty())
insertDebugValuesForPHIs(OrigHeader, InsertedPHIs);
// NewHeader is now the header of the loop.
L->moveToHeader(NewHeader);
assert(L->getHeader() == NewHeader && "Latch block is our new header");
// Inform DT about changes to the CFG.
if (DT) {
// The OrigPreheader branches to the NewHeader and Exit now. Then, inform
// the DT about the removed edge to the OrigHeader (that got removed).
SmallVector<DominatorTree::UpdateType, 3> Updates;
Updates.push_back({DominatorTree::Insert, OrigPreheader, Exit});
Updates.push_back({DominatorTree::Insert, OrigPreheader, NewHeader});
Updates.push_back({DominatorTree::Delete, OrigPreheader, OrigHeader});
DT->applyUpdates(Updates);
// Update MemorySSA before the rewrite call below changes the 1:1
// instruction:cloned_instruction_or_value mapping.
if (MSSAU) { if (MSSAU) {
MSSAU->applyUpdates(Updates, *DT); InsertNewValueIntoMap(ValueMapMSSA, OrigHeader, OrigPreheader);
if (VerifyMemorySSA) MSSAU->updateForClonedBlockIntoPred(OrigHeader, OrigPreheader,
MSSAU->getMemorySSA()->verifyMemorySSA(); ValueMapMSSA);
} }
}
// At this point, we've finished our major CFG changes. As part of cloning SmallVector<PHINode*, 2> InsertedPHIs;
// the loop into the preheader we've simplified instructions and the // If there were any uses of instructions in the duplicated block outside the
// duplicated conditional branch may now be branching on a constant. If it is // loop, update them, inserting PHI nodes as required
// branching on a constant and if that constant means that we enter the loop, RewriteUsesOfClonedInstructions(OrigHeader, OrigPreheader, ValueMap,
// then we fold away the cond branch to an uncond branch. This simplifies the &InsertedPHIs);
// loop in cases important for nested loops, and it also means we don't have
// to split as many edges.
BranchInst *PHBI = cast<BranchInst>(OrigPreheader->getTerminator());
assert(PHBI->isConditional() && "Should be clone of BI condbr!");
if (!isa<ConstantInt>(PHBI->getCondition()) ||
PHBI->getSuccessor(cast<ConstantInt>(PHBI->getCondition())->isZero()) !=
NewHeader) {
// The conditional branch can't be folded, handle the general case.
// Split edges as necessary to preserve LoopSimplify form.
// Right now OrigPreHeader has two successors, NewHeader and ExitBlock, and // Attach dbg.value intrinsics to the new phis if that phi uses a value that
// thus is not a preheader anymore. // previously had debug metadata attached. This keeps the debug info
// Split the edge to form a real preheader. // up-to-date in the loop body.
BasicBlock *NewPH = SplitCriticalEdge( if (!InsertedPHIs.empty())
OrigPreheader, NewHeader, insertDebugValuesForPHIs(OrigHeader, InsertedPHIs);
CriticalEdgeSplittingOptions(DT, LI, MSSAU).setPreserveLCSSA());
NewPH->setName(NewHeader->getName() + ".lr.ph");
// Preserve canonical loop form, which means that 'Exit' should have only // NewHeader is now the header of the loop.
// one predecessor. Note that Exit could be an exit block for multiple L->moveToHeader(NewHeader);
// nested loops, causing both of the edges to now be critical and need to assert(L->getHeader() == NewHeader && "Latch block is our new header");
// be split.
SmallVector<BasicBlock *, 4> ExitPreds(pred_begin(Exit), pred_end(Exit)); // Inform DT about changes to the CFG.
bool SplitLatchEdge = false; if (DT) {
for (BasicBlock *ExitPred : ExitPreds) { // The OrigPreheader branches to the NewHeader and Exit now. Then, inform
// We only need to split loop exit edges. // the DT about the removed edge to the OrigHeader (that got removed).
Loop *PredLoop = LI->getLoopFor(ExitPred); SmallVector<DominatorTree::UpdateType, 3> Updates;
if (!PredLoop || PredLoop->contains(Exit) || Updates.push_back({DominatorTree::Insert, OrigPreheader, Exit});
ExitPred->getTerminator()->isIndirectTerminator()) Updates.push_back({DominatorTree::Insert, OrigPreheader, NewHeader});
continue; Updates.push_back({DominatorTree::Delete, OrigPreheader, OrigHeader});
SplitLatchEdge |= L->getLoopLatch() == ExitPred; DT->applyUpdates(Updates);
BasicBlock *ExitSplit = SplitCriticalEdge(
ExitPred, Exit, if (MSSAU) {
CriticalEdgeSplittingOptions(DT, LI, MSSAU).setPreserveLCSSA()); MSSAU->applyUpdates(Updates, *DT);
ExitSplit->moveBefore(Exit); if (VerifyMemorySSA)
MSSAU->getMemorySSA()->verifyMemorySSA();
}
} }
assert(SplitLatchEdge &&
"Despite splitting all preds, failed to split latch exit?");
} else {
// We can fold the conditional branch in the preheader, this makes things
// simpler. The first step is to remove the extra edge to the Exit block.
Exit->removePredecessor(OrigPreheader, true /*preserve LCSSA*/);
BranchInst *NewBI = BranchInst::Create(NewHeader, PHBI);
NewBI->setDebugLoc(PHBI->getDebugLoc());
PHBI->eraseFromParent();
// With our CFG finalized, update DomTree if it is available. // At this point, we've finished our major CFG changes. As part of cloning
if (DT) DT->deleteEdge(OrigPreheader, Exit); // the loop into the preheader we've simplified instructions and the
// duplicated conditional branch may now be branching on a constant. If it is
// branching on a constant and if that constant means that we enter the loop,
// then we fold away the cond branch to an uncond branch. This simplifies the
// loop in cases important for nested loops, and it also means we don't have
// to split as many edges.
BranchInst *PHBI = cast<BranchInst>(OrigPreheader->getTerminator());
assert(PHBI->isConditional() && "Should be clone of BI condbr!");
if (!isa<ConstantInt>(PHBI->getCondition()) ||
PHBI->getSuccessor(cast<ConstantInt>(PHBI->getCondition())->isZero()) !=
NewHeader) {
// The conditional branch can't be folded, handle the general case.
// Split edges as necessary to preserve LoopSimplify form.
// Update MSSA too, if available. // Right now OrigPreHeader has two successors, NewHeader and ExitBlock, and
if (MSSAU) // thus is not a preheader anymore.
MSSAU->removeEdge(OrigPreheader, Exit); // Split the edge to form a real preheader.
} BasicBlock *NewPH = SplitCriticalEdge(
OrigPreheader, NewHeader,
CriticalEdgeSplittingOptions(DT, LI, MSSAU).setPreserveLCSSA());
NewPH->setName(NewHeader->getName() + ".lr.ph");
assert(L->getLoopPreheader() && "Invalid loop preheader after loop rotation"); // Preserve canonical loop form, which means that 'Exit' should have only
assert(L->getLoopLatch() && "Invalid loop latch after loop rotation"); // one predecessor. Note that Exit could be an exit block for multiple
// nested loops, causing both of the edges to now be critical and need to
// be split.
SmallVector<BasicBlock *, 4> ExitPreds(pred_begin(Exit), pred_end(Exit));
bool SplitLatchEdge = false;
for (BasicBlock *ExitPred : ExitPreds) {
// We only need to split loop exit edges.
Loop *PredLoop = LI->getLoopFor(ExitPred);
if (!PredLoop || PredLoop->contains(Exit) ||
ExitPred->getTerminator()->isIndirectTerminator())
continue;
SplitLatchEdge |= L->getLoopLatch() == ExitPred;
BasicBlock *ExitSplit = SplitCriticalEdge(
ExitPred, Exit,
CriticalEdgeSplittingOptions(DT, LI, MSSAU).setPreserveLCSSA());
ExitSplit->moveBefore(Exit);
}
assert(SplitLatchEdge &&
"Despite splitting all preds, failed to split latch exit?");
} else {
// We can fold the conditional branch in the preheader, this makes things
// simpler. The first step is to remove the extra edge to the Exit block.
Exit->removePredecessor(OrigPreheader, true /*preserve LCSSA*/);
BranchInst *NewBI = BranchInst::Create(NewHeader, PHBI);
NewBI->setDebugLoc(PHBI->getDebugLoc());
PHBI->eraseFromParent();
if (MSSAU && VerifyMemorySSA) // With our CFG finalized, update DomTree if it is available.
MSSAU->getMemorySSA()->verifyMemorySSA(); if (DT) DT->deleteEdge(OrigPreheader, Exit);
// Now that the CFG and DomTree are in a consistent state again, try to merge // Update MSSA too, if available.
// the OrigHeader block into OrigLatch. This will succeed if they are if (MSSAU)
// connected by an unconditional branch. This is just a cleanup so the MSSAU->removeEdge(OrigPreheader, Exit);
// emitted code isn't too gross in this common case. }
DomTreeUpdater DTU(DT, DomTreeUpdater::UpdateStrategy::Eager);
MergeBlockIntoPredecessor(OrigHeader, &DTU, LI, MSSAU);
if (MSSAU && VerifyMemorySSA) assert(L->getLoopPreheader() && "Invalid loop preheader after loop rotation");
MSSAU->getMemorySSA()->verifyMemorySSA(); assert(L->getLoopLatch() && "Invalid loop latch after loop rotation");
if (MSSAU && VerifyMemorySSA)
MSSAU->getMemorySSA()->verifyMemorySSA();
// Now that the CFG and DomTree are in a consistent state again, try to merge
// the OrigHeader block into OrigLatch. This will succeed if they are
// connected by an unconditional branch. This is just a cleanup so the
// emitted code isn't too gross in this common case.
DomTreeUpdater DTU(DT, DomTreeUpdater::UpdateStrategy::Eager);
MergeBlockIntoPredecessor(OrigHeader, &DTU, LI, MSSAU);
if (MSSAU && VerifyMemorySSA)
MSSAU->getMemorySSA()->verifyMemorySSA();
LLVM_DEBUG(dbgs() << "LoopRotation: into "; L->dump());
++NumRotated;
Rotated = true;
SimplifiedLatch = false;
// Check that new latch is a deoptimizing exit and then repeat rotation if possible.
// Deoptimizing latch exit is not a generally typical case, so we just loop over.
// TODO: if it becomes a performance bottleneck extend rotation algorithm
// to handle multiple rotations in one go.
} while (MultiRotate && canRotateDeoptimizingLatchExit(L));
LLVM_DEBUG(dbgs() << "LoopRotation: into "; L->dump());
++NumRotated;
return true; return true;
} }

165
llvm/test/Transforms/LoopRotate/multiple-deopt-exits.ll Normal file
View File

@ -0,0 +1,165 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -S < %s -loop-rotate -loop-rotate-multi=true | FileCheck %s
; RUN: opt -S < %s -passes='loop(rotate)' -loop-rotate-multi=true | FileCheck %s
; Test loop rotation with multiple exits, some of them - deoptimizing.
; We should end up with a latch which exit is non-deoptimizing, so we should rotate
; more than once.
declare i32 @llvm.experimental.deoptimize.i32(...)
define i32 @test_cond_with_one_deopt_exit(i32 * nonnull %a, i64 %x) {
; Rotation done twice.
; Latch should be at the 2nd condition (for.cond2), exiting to %return.
;
; CHECK-LABEL: @test_cond_with_one_deopt_exit(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[VAL_A_IDX3:%.*]] = load i32, i32* %a, align 4
; CHECK-NEXT: [[ZERO_CHECK4:%.*]] = icmp eq i32 [[VAL_A_IDX3]], 0
; CHECK-NEXT: br i1 [[ZERO_CHECK4]], label %deopt.exit, label %for.cond2.lr.ph
; CHECK: for.cond2.lr.ph:
; CHECK-NEXT: [[FOR_CHECK8:%.*]] = icmp ult i64 0, %x
; CHECK-NEXT: br i1 [[FOR_CHECK8]], label %for.body.lr.ph, label %return
; CHECK: for.body.lr.ph:
; CHECK-NEXT: br label %for.body
; CHECK: for.cond2:
; CHECK: [[FOR_CHECK:%.*]] = icmp ult i64 {{%.*}}, %x
; CHECK-NEXT: br i1 [[FOR_CHECK]], label %for.body, label %for.cond2.return_crit_edge
; CHECK: for.body:
; CHECK: br label %for.tail
; CHECK: for.tail:
; CHECK: [[VAL_A_IDX:%.*]] = load i32, i32*
; CHECK-NEXT: [[ZERO_CHECK:%.*]] = icmp eq i32 [[VAL_A_IDX]], 0
; CHECK-NEXT: br i1 [[ZERO_CHECK]], label %for.cond1.deopt.exit_crit_edge, label %for.cond2
; CHECK: for.cond2.return_crit_edge:
; CHECK-NEXT: {{%.*}} = phi i32
; CHECK-NEXT: br label %return
; CHECK: return:
; CHECK-NEXT: [[SUM_LCSSA2:%.*]] = phi i32
; CHECK-NEXT: ret i32 [[SUM_LCSSA2]]
; CHECK: for.cond1.deopt.exit_crit_edge:
; CHECK-NEXT: {{%.*}} = phi i32
; CHECK-NEXT: br label %deopt.exit
; CHECK: deopt.exit:
; CHECK: [[DEOPT_VAL:%.*]] = call i32 (...) @llvm.experimental.deoptimize.i32() [ "deopt"(i32 {{%.*}}) ]
; CHECK-NEXT: ret i32 [[DEOPT_VAL]]
;
entry:
br label %for.cond1
for.cond1:
%idx = phi i64 [ 0, %entry ], [ %idx.next, %for.tail ]
%sum = phi i32 [ 0, %entry ], [ %sum.next, %for.tail ]
%a.idx = getelementptr inbounds i32, i32 *%a, i64 %idx
%val.a.idx = load i32, i32* %a.idx, align 4
%zero.check = icmp eq i32 %val.a.idx, 0
br i1 %zero.check, label %deopt.exit, label %for.cond2
for.cond2:
%for.check = icmp ult i64 %idx, %x
br i1 %for.check, label %for.body, label %return
for.body:
br label %for.tail
for.tail:
%sum.next = add i32 %sum, %val.a.idx
%idx.next = add nuw nsw i64 %idx, 1
br label %for.cond1
return:
ret i32 %sum
deopt.exit:
%deopt.val = call i32(...) @llvm.experimental.deoptimize.i32() [ "deopt"(i32 %val.a.idx) ]
ret i32 %deopt.val
}
define i32 @test_cond_with_two_deopt_exits(i32 ** nonnull %a, i64 %x) {
; Rotation done three times.
; Latch should be at the 3rd condition (for.cond3), exiting to %return.
;
; CHECK-LABEL: @test_cond_with_two_deopt_exits(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[A_IDX_DEREF4:%.*]] = load i32*, i32** %a
; CHECK-NEXT: [[NULL_CHECK5:%.*]] = icmp eq i32* [[A_IDX_DEREF4]], null
; CHECK-NEXT: br i1 [[NULL_CHECK5]], label %deopt.exit1, label %for.cond2.lr.ph
; CHECK: for.cond2.lr.ph:
; CHECK-NEXT: [[VAL_A_IDX9:%.*]] = load i32, i32* [[A_IDX_DEREF4]], align 4
; CHECK-NEXT: [[ZERO_CHECK10:%.*]] = icmp eq i32 [[VAL_A_IDX9]], 0
; CHECK-NEXT: br i1 [[ZERO_CHECK10]], label %deopt.exit2, label %for.cond3.lr.ph
; CHECK: for.cond3.lr.ph:
; CHECK-NEXT: [[FOR_CHECK14:%.*]] = icmp ult i64 0, %x
; CHECK-NEXT: br i1 [[FOR_CHECK14]], label %for.body.lr.ph, label %return
; CHECK: for.body.lr.ph:
; CHECK-NEXT: br label %for.body
; CHECK: for.cond2:
; CHECK: [[VAL_A_IDX:%.*]] = load i32, i32*
; CHECK-NEXT: [[ZERO_CHECK:%.*]] = icmp eq i32 [[VAL_A_IDX]], 0
; CHECK-NEXT: br i1 [[ZERO_CHECK]], label %for.cond2.deopt.exit2_crit_edge, label %for.cond3
; CHECK: for.cond3:
; CHECK: [[FOR_CHECK:%.*]] = icmp ult i64 {{%.*}}, %x
; CHECK-NEXT: br i1 [[FOR_CHECK]], label %for.body, label %for.cond3.return_crit_edge
; CHECK: for.body:
; CHECK: br label %for.tail
; CHECK: for.tail:
; CHECK: [[IDX_NEXT:%.*]] = add nuw nsw i64 {{%.*}}, 1
; CHECK: [[NULL_CHECK:%.*]] = icmp eq i32* {{%.*}}, null
; CHECK-NEXT: br i1 [[NULL_CHECK]], label %for.cond1.deopt.exit1_crit_edge, label %for.cond2
; CHECK: for.cond3.return_crit_edge:
; CHECK-NEXT: [[SPLIT18:%.*]] = phi i32
; CHECK-NEXT: br label %return
; CHECK: return:
; CHECK-NEXT: [[SUM_LCSSA2:%.*]] = phi i32
; CHECK-NEXT: ret i32 [[SUM_LCSSA2]]
; CHECK: for.cond1.deopt.exit1_crit_edge:
; CHECK-NEXT: br label %deopt.exit1
; CHECK: deopt.exit1:
; CHECK-NEXT: [[DEOPT_VAL1:%.*]] = call i32 (...) @llvm.experimental.deoptimize.i32() [ "deopt"(i32 0) ]
; CHECK-NEXT: ret i32 [[DEOPT_VAL1]]
; CHECK: for.cond2.deopt.exit2_crit_edge:
; CHECK-NEXT: [[SPLIT:%.*]] = phi i32
; CHECK-NEXT: br label %deopt.exit2
; CHECK: deopt.exit2:
; CHECK-NEXT: [[VAL_A_IDX_LCSSA:%.*]] = phi i32
; CHECK-NEXT: [[DEOPT_VAL2:%.*]] = call i32 (...) @llvm.experimental.deoptimize.i32() [ "deopt"(i32 [[VAL_A_IDX_LCSSA]]) ]
; CHECK-NEXT: ret i32 [[DEOPT_VAL2]]
;
entry:
br label %for.cond1
for.cond1:
%idx = phi i64 [ 0, %entry ], [ %idx.next, %for.tail ]
%sum = phi i32 [ 0, %entry ], [ %sum.next, %for.tail ]
%a.idx = getelementptr inbounds i32*, i32 **%a, i64 %idx
%a.idx.deref = load i32*, i32** %a.idx
%null.check = icmp eq i32* %a.idx.deref, null
br i1 %null.check, label %deopt.exit1, label %for.cond2
for.cond2:
%val.a.idx = load i32, i32* %a.idx.deref, align 4
%zero.check = icmp eq i32 %val.a.idx, 0
br i1 %zero.check, label %deopt.exit2, label %for.cond3
for.cond3:
%for.check = icmp ult i64 %idx, %x
br i1 %for.check, label %for.body, label %return
for.body:
br label %for.tail
for.tail:
%sum.next = add i32 %sum, %val.a.idx
%idx.next = add nuw nsw i64 %idx, 1
br label %for.cond1
return:
ret i32 %sum
deopt.exit1:
%deopt.val1 = call i32(...) @llvm.experimental.deoptimize.i32() [ "deopt"(i32 0) ]
ret i32 %deopt.val1
deopt.exit2:
%deopt.val2 = call i32(...) @llvm.experimental.deoptimize.i32() [ "deopt"(i32 %val.a.idx) ]
ret i32 %deopt.val2
}

1
llvm/unittests/Transforms/Utils/CMakeLists.txt
View File

@ -15,6 +15,7 @@ add_llvm_unittest(UtilsTests
FunctionComparatorTest.cpp FunctionComparatorTest.cpp
IntegerDivisionTest.cpp IntegerDivisionTest.cpp
LocalTest.cpp LocalTest.cpp
LoopRotationUtilsTest.cpp
LoopUtilsTest.cpp LoopUtilsTest.cpp
SizeOptsTest.cpp SizeOptsTest.cpp
SSAUpdaterBulkTest.cpp SSAUpdaterBulkTest.cpp

166
llvm/unittests/Transforms/Utils/LoopRotationUtilsTest.cpp Normal file
View File

@ -0,0 +1,166 @@
//===- LoopRotationUtilsTest.cpp - Unit tests for LoopRotation utility ----===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
#include "llvm/Transforms/Utils/LoopRotationUtils.h"
#include "llvm/Analysis/AssumptionCache.h"
#include "llvm/Analysis/InstructionSimplify.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/ScalarEvolution.h"
#include "llvm/Analysis/TargetLibraryInfo.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/AsmParser/Parser.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/LLVMContext.h"
#include "llvm/Support/SourceMgr.h"
#include "gtest/gtest.h"
using namespace llvm;
static std::unique_ptr<Module> parseIR(LLVMContext &C, const char *IR) {
SMDiagnostic Err;
std::unique_ptr<Module> Mod = parseAssemblyString(IR, Err, C);
if (!Mod)
Err.print("LoopRotationUtilsTest", errs());
return Mod;
}
/// This test contains multi-deopt-exits pattern that might allow loop rotation
/// to trigger multiple times if multiple rotations are enabled.
/// At least one rotation should be performed, no matter what loop rotation settings are.
TEST(LoopRotate, MultiDeoptExit) {
LLVMContext C;
std::unique_ptr<Module> M = parseIR(
C,
R"(
declare i32 @llvm.experimental.deoptimize.i32(...)
define i32 @test(i32 * nonnull %a, i64 %x) {
entry:
br label %for.cond1
for.cond1:
%idx = phi i64 [ 0, %entry ], [ %idx.next, %for.tail ]
%sum = phi i32 [ 0, %entry ], [ %sum.next, %for.tail ]
%a.idx = getelementptr inbounds i32, i32 *%a, i64 %idx
%val.a.idx = load i32, i32* %a.idx, align 4
%zero.check = icmp eq i32 %val.a.idx, 0
br i1 %zero.check, label %deopt.exit, label %for.cond2
for.cond2:
%for.check = icmp ult i64 %idx, %x
br i1 %for.check, label %for.body, label %return
for.body:
br label %for.tail
for.tail:
%sum.next = add i32 %sum, %val.a.idx
%idx.next = add nuw nsw i64 %idx, 1
br label %for.cond1
return:
ret i32 %sum
deopt.exit:
%deopt.val = call i32(...) @llvm.experimental.deoptimize.i32() [ "deopt"(i32 %val.a.idx) ]
ret i32 %deopt.val
})"
);
auto *F = M->getFunction("test");
DominatorTree DT(*F);
LoopInfo LI(DT);
AssumptionCache AC(*F);
TargetTransformInfo TTI(M->getDataLayout());
TargetLibraryInfoImpl TLII;
TargetLibraryInfo TLI(TLII);
ScalarEvolution SE(*F, TLI, AC, DT, LI);
SimplifyQuery SQ(M->getDataLayout());
Loop *L = *LI.begin();
bool ret = LoopRotation(L, &LI, &TTI,
&AC, &DT,
&SE, nullptr,
SQ, true, -1, false);
EXPECT_TRUE(ret);
}
/// Checking a special case of multi-deopt exit loop that can not perform
/// required amount of rotations due to the desired header containing
/// non-duplicatable code.
/// Similar to MultiDeoptExit test this one should do at least one rotation and
/// pass no matter what loop rotation settings are.
TEST(LoopRotate, MultiDeoptExit_Nondup) {
LLVMContext C;
std::unique_ptr<Module> M = parseIR(
C,
R"(
; Rotation should be done once, attempted twice.
; Second time fails due to non-duplicatable header.
declare i32 @llvm.experimental.deoptimize.i32(...)
declare void @nondup()
define i32 @test_nondup(i32 * nonnull %a, i64 %x) {
entry:
br label %for.cond1
for.cond1:
%idx = phi i64 [ 0, %entry ], [ %idx.next, %for.tail ]
%sum = phi i32 [ 0, %entry ], [ %sum.next, %for.tail ]
%a.idx = getelementptr inbounds i32, i32 *%a, i64 %idx
%val.a.idx = load i32, i32* %a.idx, align 4
%zero.check = icmp eq i32 %val.a.idx, 0
br i1 %zero.check, label %deopt.exit, label %for.cond2
for.cond2:
call void @nondup() noduplicate
%for.check = icmp ult i64 %idx, %x
br i1 %for.check, label %for.body, label %return
for.body:
br label %for.tail
for.tail:
%sum.next = add i32 %sum, %val.a.idx
%idx.next = add nuw nsw i64 %idx, 1
br label %for.cond1
return:
ret i32 %sum
deopt.exit:
%deopt.val = call i32(...) @llvm.experimental.deoptimize.i32() [ "deopt"(i32 %val.a.idx) ]
ret i32 %deopt.val
})"
);
auto *F = M->getFunction("test_nondup");
DominatorTree DT(*F);
LoopInfo LI(DT);
AssumptionCache AC(*F);
TargetTransformInfo TTI(M->getDataLayout());
TargetLibraryInfoImpl TLII;
TargetLibraryInfo TLI(TLII);
ScalarEvolution SE(*F, TLI, AC, DT, LI);
SimplifyQuery SQ(M->getDataLayout());
Loop *L = *LI.begin();
bool ret = LoopRotation(L, &LI, &TTI,
&AC, &DT,
&SE, nullptr,
SQ, true, -1, false);
/// LoopRotation should properly report "true" as we still perform the first rotation
/// so we do change the IR.
EXPECT_TRUE(ret);
}