mirror of
https://github.com/apple/swift.git
synced 2025-12-14 20:36:38 +01:00
309 lines
10 KiB
C++
309 lines
10 KiB
C++
//===--- COWOpts.cpp - Optimize COW operations ----------------------------===//
|
|
//
|
|
// This source file is part of the Swift.org open source project
|
|
//
|
|
// Copyright (c) 2014 - 2020 Apple Inc. and the Swift project authors
|
|
// Licensed under Apache License v2.0 with Runtime Library Exception
|
|
//
|
|
// See https://swift.org/LICENSE.txt for license information
|
|
// See https://swift.org/CONTRIBUTORS.txt for the list of Swift project authors
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
//
|
|
// This pass optimizes begin_cow_mutation and end_cow_mutation patterns.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#define DEBUG_TYPE "cow-opts"
|
|
#include "swift/SILOptimizer/PassManager/Transforms.h"
|
|
#include "swift/SILOptimizer/Analysis/AliasAnalysis.h"
|
|
#include "swift/SIL/NodeBits.h"
|
|
#include "swift/SIL/SILFunction.h"
|
|
#include "swift/SIL/SILBasicBlock.h"
|
|
#include "swift/SIL/SILArgument.h"
|
|
#include "swift/SIL/SILBuilder.h"
|
|
#include "swift/SIL/StackList.h"
|
|
#include "llvm/Support/Debug.h"
|
|
|
|
using namespace swift;
|
|
|
|
namespace {
|
|
|
|
/// Constant folds the uniqueness result of begin_cow_mutation instructions.
|
|
///
|
|
/// If it can be proved that the buffer argument is uniquely referenced, the
|
|
/// uniqueness result is replaced with a constant boolean "true".
|
|
/// For example:
|
|
///
|
|
/// \code
|
|
/// %buffer = end_cow_mutation %mutable_buffer
|
|
/// // ...
|
|
/// // %buffer does not escape here
|
|
/// // ...
|
|
/// (%is_unique, %mutable_buffer2) = begin_cow_mutation %buffer
|
|
/// cond_br %is_unique, ...
|
|
/// \endcode
|
|
///
|
|
/// is replaced with
|
|
///
|
|
/// \code
|
|
/// %buffer = end_cow_mutation [keep_unique] %mutable_buffer
|
|
/// // ...
|
|
/// (%not_used, %mutable_buffer2) = begin_cow_mutation %buffer
|
|
/// %true = integer_literal 1
|
|
/// cond_br %true, ...
|
|
/// \endcode
|
|
///
|
|
/// Note that the keep_unique flag is set on the end_cow_mutation because the
|
|
/// code now relies on that the buffer is really uniquely referenced.
|
|
///
|
|
/// The optimization can also handle def-use chains between end_cow_mutation and
|
|
/// begin_cow_mutation which involve phi-arguments.
|
|
///
|
|
class COWOptsPass : public SILFunctionTransform {
|
|
public:
|
|
COWOptsPass() {}
|
|
|
|
void run() override;
|
|
|
|
private:
|
|
AliasAnalysis *AA = nullptr;
|
|
|
|
bool optimizeBeginCOW(BeginCOWMutationInst *BCM);
|
|
|
|
static void collectEscapePoints(SILValue v,
|
|
InstructionSetWithSize &escapePoints,
|
|
ValueSet &handled);
|
|
};
|
|
|
|
void COWOptsPass::run() {
|
|
SILFunction *F = getFunction();
|
|
if (!F->shouldOptimize())
|
|
return;
|
|
|
|
LLVM_DEBUG(llvm::dbgs() << "*** COW optimization on function: "
|
|
<< F->getName() << " ***\n");
|
|
|
|
AA = PM->getAnalysis<AliasAnalysis>(F);
|
|
|
|
bool changed = false;
|
|
for (SILBasicBlock &block : *F) {
|
|
|
|
for (SILInstruction &inst : block) {
|
|
if (auto *beginCOW = dyn_cast<BeginCOWMutationInst>(&inst)) {
|
|
if (optimizeBeginCOW(beginCOW))
|
|
changed = true;
|
|
}
|
|
}
|
|
}
|
|
|
|
if (changed) {
|
|
invalidateAnalysis(SILAnalysis::InvalidationKind::Instructions);
|
|
}
|
|
}
|
|
|
|
static SILValue skipStructAndExtract(SILValue value) {
|
|
while (true) {
|
|
if (auto *si = dyn_cast<StructInst>(value)) {
|
|
if (si->getNumOperands() != 1)
|
|
return value;
|
|
value = si->getOperand(0);
|
|
continue;
|
|
}
|
|
if (auto *sei = dyn_cast<StructExtractInst>(value)) {
|
|
value = sei->getOperand();
|
|
continue;
|
|
}
|
|
return value;
|
|
}
|
|
}
|
|
|
|
bool COWOptsPass::optimizeBeginCOW(BeginCOWMutationInst *BCM) {
|
|
LLVM_DEBUG(llvm::dbgs() << "Looking at: ");
|
|
LLVM_DEBUG(BCM->dump());
|
|
|
|
SILFunction *function = BCM->getFunction();
|
|
StackList<EndCOWMutationInst *> endCOWMutationInsts(function);
|
|
InstructionSet endCOWMutationsFound(function);
|
|
|
|
{
|
|
// Collect all end_cow_mutation instructions, used by the begin_cow_mutation,
|
|
// looking through block phi-arguments.
|
|
StackList<SILValue> workList(function);
|
|
ValueSet handled(function);
|
|
workList.push_back(BCM->getOperand());
|
|
while (!workList.empty()) {
|
|
SILValue v = skipStructAndExtract(workList.pop_back_val());
|
|
if (SILPhiArgument *arg = dyn_cast<SILPhiArgument>(v)) {
|
|
if (handled.insert(arg)) {
|
|
SmallVector<SILValue, 4> incomingVals;
|
|
if (!arg->getIncomingPhiValues(incomingVals))
|
|
return false;
|
|
for (SILValue incomingVal : incomingVals) {
|
|
workList.push_back(incomingVal);
|
|
}
|
|
}
|
|
} else if (auto *ECM = dyn_cast<EndCOWMutationInst>(v)) {
|
|
if (endCOWMutationsFound.insert(ECM))
|
|
endCOWMutationInsts.push_back(ECM);
|
|
} else {
|
|
return false;
|
|
}
|
|
}
|
|
}
|
|
|
|
// Collect all uses of the end_cow_instructions, where the buffer can
|
|
// potentially escape.
|
|
InstructionSetWithSize potentialEscapePoints(function);
|
|
{
|
|
ValueSet handled(function);
|
|
for (EndCOWMutationInst *ECM : endCOWMutationInsts) {
|
|
collectEscapePoints(ECM, potentialEscapePoints, handled);
|
|
}
|
|
}
|
|
|
|
if (!potentialEscapePoints.empty()) {
|
|
// Now, this is the complicated part: check if there is an escape point
|
|
// within the liverange between the end_cow_mutation(s) and
|
|
// begin_cow_mutation.
|
|
//
|
|
// For store instructions we do a little bit more: only count a store as an
|
|
// escape if there is a (potential) load from the same address within the
|
|
// liverange.
|
|
StackList<SILInstruction *> instWorkList(function);
|
|
StackList<SILInstruction *> potentialLoadInsts(function);
|
|
StackList<SILValue> storeAddrs(function);
|
|
ValueSet storeAddrsFound(function);
|
|
BasicBlockSet handled(function);
|
|
int numStoresFound = 0;
|
|
int numLoadsFound = 0;
|
|
|
|
// This is a simple worklist-based backward dataflow analysis.
|
|
// Start at the initial begin_cow_mutation and go backward.
|
|
instWorkList.push_back(BCM);
|
|
|
|
while (!instWorkList.empty()) {
|
|
SILInstruction *inst = instWorkList.pop_back_val();
|
|
for (;;) {
|
|
if (potentialEscapePoints.contains(inst)) {
|
|
if (auto *store = dyn_cast<StoreInst>(inst)) {
|
|
// Don't immediately bail on a store instruction. Instead, remember
|
|
// it and check if it interferes with any (potential) load.
|
|
if (storeAddrsFound.insert(store->getDest())) {
|
|
LLVM_DEBUG(llvm::dbgs() << "Found store escape, record: ");
|
|
LLVM_DEBUG(inst->dump());
|
|
storeAddrs.push_back(store->getDest());
|
|
numStoresFound += 1;
|
|
}
|
|
} else {
|
|
LLVM_DEBUG(llvm::dbgs() << "Found non-store escape, bailing out: ");
|
|
LLVM_DEBUG(inst->dump());
|
|
return false;
|
|
}
|
|
}
|
|
if (inst->mayReadFromMemory()) {
|
|
LLVM_DEBUG(llvm::dbgs() << "Found a may read inst, record: ");
|
|
LLVM_DEBUG(inst->dump());
|
|
potentialLoadInsts.push_back(inst);
|
|
numLoadsFound += 1;
|
|
}
|
|
|
|
// An end_cow_mutation marks the begin of the liverange. It's the end
|
|
// point of the dataflow analysis.
|
|
auto *ECM = dyn_cast<EndCOWMutationInst>(inst);
|
|
if (ECM && endCOWMutationsFound.contains(ECM))
|
|
break;
|
|
|
|
if (inst == &inst->getParent()->front()) {
|
|
for (SILBasicBlock *pred : inst->getParent()->getPredecessorBlocks()) {
|
|
if (handled.insert(pred))
|
|
instWorkList.push_back(pred->getTerminator());
|
|
}
|
|
break;
|
|
}
|
|
|
|
inst = &*std::prev(inst->getIterator());
|
|
}
|
|
}
|
|
|
|
// Check if there is any (potential) load from a memory location where the
|
|
// buffer is stored to.
|
|
if (numStoresFound != 0) {
|
|
// Avoid quadratic behavior. Usually this limit is not exceeded.
|
|
if (numStoresFound * numLoadsFound > 128)
|
|
return false;
|
|
for (SILInstruction *load : potentialLoadInsts) {
|
|
for (SILValue storeAddr : storeAddrs) {
|
|
if (!AA || AA->mayReadFromMemory(load, storeAddr)) {
|
|
LLVM_DEBUG(llvm::dbgs() << "Found a store address aliasing with a load:");
|
|
LLVM_DEBUG(load->dump());
|
|
LLVM_DEBUG(storeAddr->dump());
|
|
return false;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
// Replace the uniqueness result of the begin_cow_mutation with an integer
|
|
// literal of "true".
|
|
SILBuilderWithScope B(BCM);
|
|
auto *IL = B.createIntegerLiteral(BCM->getLoc(),
|
|
BCM->getUniquenessResult()->getType(), 1);
|
|
BCM->getUniquenessResult()->replaceAllUsesWith(IL);
|
|
|
|
for (EndCOWMutationInst *ECM : endCOWMutationInsts) {
|
|
// This is important for other optimizations: The code is now relying on
|
|
// the buffer to be unique.
|
|
ECM->setKeepUnique();
|
|
}
|
|
|
|
return true;
|
|
}
|
|
|
|
void COWOptsPass::collectEscapePoints(SILValue v,
|
|
InstructionSetWithSize &escapePoints,
|
|
ValueSet &handled) {
|
|
if (!handled.insert(v))
|
|
return;
|
|
|
|
for (Operand *use : v->getUses()) {
|
|
SILInstruction *user = use->getUser();
|
|
switch (user->getKind()) {
|
|
case SILInstructionKind::BeginCOWMutationInst:
|
|
case SILInstructionKind::RefElementAddrInst:
|
|
case SILInstructionKind::RefTailAddrInst:
|
|
case SILInstructionKind::DebugValueInst:
|
|
break;
|
|
case SILInstructionKind::BranchInst:
|
|
collectEscapePoints(cast<BranchInst>(user)->getArgForOperand(use),
|
|
escapePoints, handled);
|
|
break;
|
|
case SILInstructionKind::CondBranchInst:
|
|
if (use->getOperandNumber() != CondBranchInst::ConditionIdx) {
|
|
collectEscapePoints(cast<CondBranchInst>(user)->getArgForOperand(use),
|
|
escapePoints, handled);
|
|
}
|
|
break;
|
|
case SILInstructionKind::StructInst:
|
|
case SILInstructionKind::StructExtractInst:
|
|
case SILInstructionKind::TupleInst:
|
|
case SILInstructionKind::TupleExtractInst:
|
|
case SILInstructionKind::UncheckedRefCastInst:
|
|
collectEscapePoints(cast<SingleValueInstruction>(user),
|
|
escapePoints, handled);
|
|
break;
|
|
default:
|
|
// Everything else is considered to be a potential escape of the buffer.
|
|
escapePoints.insert(user);
|
|
}
|
|
}
|
|
}
|
|
|
|
} // end anonymous namespace
|
|
|
|
SILTransform *swift::createCOWOpts() {
|
|
return new COWOptsPass();
|
|
}
|
|
|