Skip to content

[AMDGPU] Improve StructurizeCFG pass performance: avoid redundant DebugLoc map initialization. NFC. #130568

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Mar 11, 2025
Merged

[AMDGPU] Improve StructurizeCFG pass performance: avoid redundant DebugLoc map initialization. NFC. #130568

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
56294b8
[AMDGPU] Improve compile time of StructurizeCFG pass: avoid unnecessa…
vpykhtin Mar 10, 2025
b7ae8c8
Avoid using temporary DebugLoc object when possible.
vpykhtin Mar 10, 2025
e30c1cf
Improved comment
vpykhtin Mar 10, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
40 changes: 25 additions & 15 deletions llvm/lib/Transforms/Scalar/StructurizeCFG.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -361,6 +361,8 @@ class StructurizeCFG {

void rebuildSSA();

void setDebugLoc(BranchInst *Br, BasicBlock *BB);

public:
void init(Region *R);
bool run(Region *R, DominatorTree *DT);
Expand Down Expand Up @@ -595,14 +597,6 @@ void StructurizeCFG::collectInfos() {
// Find the last back edges
analyzeLoops(RN);
}

// Reset the collected term debug locations
TermDL.clear();

for (BasicBlock &BB : *Func) {
if (const DebugLoc &DL = BB.getTerminator()->getDebugLoc())
TermDL[&BB] = DL;
}
}

/// Insert the missing branch conditions
Expand Down Expand Up @@ -924,12 +918,24 @@ void StructurizeCFG::simplifyAffectedPhis() {
} while (Changed);
}

void StructurizeCFG::setDebugLoc(BranchInst *Br, BasicBlock *BB) {
auto I = TermDL.find(BB);
if (I == TermDL.end())
return;

Br->setDebugLoc(I->second);
TermDL.erase(I);
}

/// Remove phi values from all successors and then remove the terminator.
void StructurizeCFG::killTerminator(BasicBlock *BB) {
Instruction *Term = BB->getTerminator();
if (!Term)
return;

if (const DebugLoc &DL = Term->getDebugLoc())
TermDL[BB] = DL;

for (BasicBlock *Succ : successors(BB))
delPhiValues(BB, Succ);

Expand Down Expand Up @@ -974,7 +980,7 @@ void StructurizeCFG::changeExit(RegionNode *Node, BasicBlock *NewExit,
BasicBlock *BB = Node->getNodeAs<BasicBlock>();
killTerminator(BB);
BranchInst *Br = BranchInst::Create(NewExit, BB);
Br->setDebugLoc(TermDL[BB]);
setDebugLoc(Br, BB);
addPhiValues(BB, NewExit);
if (IncludeDominator)
DT->changeImmediateDominator(NewExit, BB);
Expand All @@ -990,10 +996,14 @@ BasicBlock *StructurizeCFG::getNextFlow(BasicBlock *Dominator) {
Func, Insert);
FlowSet.insert(Flow);

// use a temporary variable to avoid a use-after-free if the map's storage is
// reallocated
DebugLoc DL = TermDL[Dominator];
TermDL[Flow] = std::move(DL);
if (auto *Term = Dominator->getTerminator()) {
if (const DebugLoc &DL = Term->getDebugLoc())
TermDL[Flow] = DL;
} else if (DebugLoc DLTemp = TermDL.lookup(Dominator)) {
// Use a temporary copy to avoid a use-after-free if the map's storage
// is reallocated.
TermDL[Flow] = DLTemp;
}

DT->addNewBlock(Flow, Dominator);
ParentRegion->getRegionInfo()->setRegionFor(Flow, ParentRegion);
Expand Down Expand Up @@ -1088,7 +1098,7 @@ void StructurizeCFG::wireFlow(bool ExitUseAllowed,

// let it point to entry and next block
BranchInst *Br = BranchInst::Create(Entry, Next, BoolPoison, Flow);
Br->setDebugLoc(TermDL[Flow]);
setDebugLoc(Br, Flow);
Conditions.push_back(Br);
addPhiValues(Flow, Entry);
DT->changeImmediateDominator(Entry, Flow);
Expand Down Expand Up @@ -1129,7 +1139,7 @@ void StructurizeCFG::handleLoops(bool ExitUseAllowed,
LoopEnd = needPrefix(false);
BasicBlock *Next = needPostfix(LoopEnd, ExitUseAllowed);
BranchInst *Br = BranchInst::Create(Next, LoopStart, BoolPoison, LoopEnd);
Br->setDebugLoc(TermDL[LoopEnd]);
setDebugLoc(Br, LoopEnd);
LoopConds.push_back(Br);
addPhiValues(LoopEnd, LoopStart);
setPrevNode(Next);
Expand Down