[ASTMatcher] Fix redundant macro expansion checks in getExpansionLocOfMacro

[mandymimi]

A performance issue was descibed in #114521

Root Cause: The function getExpansionLocOfMacro is responsible for finding the expansion location of macros. When dealing with macro parameters, it recursively calls itself to check the expansion of macro arguments. This recursive logic redundantly checks previous macro expansions, leading to significant performance degradation when macros are heavily nested.
Solution
Modification: Track already processed macros during recursion.
Implementation Details:
Introduced a data structure to record processed macros.
Before each recursive call, check if the macro has already been processed to avoid redundant calculations.
Testing:

1. refer to [clang-tidy] ASTMatchersInternal.cpp getExpansionLocOfMacro() occurs an infinite loop on freecad/src/Mod/Path/App/AreaPyImp.cpp #114521
   Finder->addMatcher(expr(isExpandedFromMacro("NULL")).bind("E"), this);
   run clang-tidy on freecad/src/Mod/Path/App/AreaPyImp.cpp
   clang-tidy src/Mod/Path/App/AreaPyImp.cpp -checks=-*,testchecker -p=build/compile_commands.json
   checker runs normally
2. check-clang-unit pass

[github-actions]

Thank you for submitting a Pull Request (PR) to the LLVM Project!

This PR will be automatically labeled and the relevant teams will be notified.

If you wish to, you can add reviewers by using the "Reviewers" section on this page.

If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using @ followed by their GitHub username.

If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers.

If you have further questions, they may be answered by the LLVM GitHub User Guide.

You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums.

[llvmbot]

@llvm/pr-subscribers-clang

Author: mandymi (mandymimi)

Changes

A performance issue was descibed in #114521

Root Cause: The function getExpansionLocOfMacro is responsible for finding the expansion location of macros. When dealing with macro parameters, it recursively calls itself to check the expansion of macro arguments. This recursive logic redundantly checks previous macro expansions, leading to significant performance degradation when macros are heavily nested.
Solution
Modification: Track already processed macros during recursion.
Implementation Details:
Introduced a data structure to record processed macros.
Before each recursive call, check if the macro has already been processed to avoid redundant calculations.
Testing:

1. refer to #114521
   Finder->addMatcher(expr(isExpandedFromMacro("NULL")).bind("E"), this);
   run clang-tidy on freecad/src/Mod/Path/App/AreaPyImp.cpp
   clang-tidy src/Mod/Path/App/AreaPyImp.cpp -checks=-*,testchecker -p=build/compile_commands.json
   checker runs normally
2. check-clang-unit pass

---

Full diff: https://github.com/llvm/llvm-project/pull/117143.diff

1 Files Affected:

• (modified) clang/lib/ASTMatchers/ASTMatchersInternal.cpp (+61-23)

diff --git a/clang/lib/ASTMatchers/ASTMatchersInternal.cpp b/clang/lib/ASTMatchers/ASTMatchersInternal.cpp
index 46dd44e6f2b24f..80eb82949831bc 100644
--- a/clang/lib/ASTMatchers/ASTMatchersInternal.cpp
+++ b/clang/lib/ASTMatchers/ASTMatchersInternal.cpp
@@ -36,6 +36,7 @@
 #include <cstddef>
 #include <optional>
 #include <string>
+#include <unordered_set>
 #include <utility>
 #include <vector>
 
@@ -136,7 +137,8 @@ class IdDynMatcher : public DynMatcherInterface {
   bool dynMatches(const DynTypedNode &DynNode, ASTMatchFinder *Finder,
                   BoundNodesTreeBuilder *Builder) const override {
     bool Result = InnerMatcher->dynMatches(DynNode, Finder, Builder);
-    if (Result) Builder->setBinding(ID, DynNode);
+    if (Result)
+      Builder->setBinding(ID, DynNode);
     return Result;
   }
 
@@ -354,7 +356,8 @@ bool DynTypedMatcher::canConvertTo(ASTNodeKind To) const {
   auto TypeKind = ASTNodeKind::getFromNodeKind<Type>();
   /// Mimic the implicit conversions of Matcher<>.
   /// - From Matcher<Type> to Matcher<QualType>
-  if (From.isSame(TypeKind) && To.isSame(QualKind)) return true;
+  if (From.isSame(TypeKind) && To.isSame(QualKind))
+    return true;
   /// - From Matcher<Base> to Matcher<Derived>
   return From.isBaseOf(To);
 }
@@ -440,8 +443,8 @@ optionallyVariadicOperator(const DynTypedNode &DynNode, ASTMatchFinder *Finder,
   return true;
 }
 
-inline static
-std::vector<std::string> vectorFromRefs(ArrayRef<const StringRef *> NameRefs) {
+inline static std::vector<std::string>
+vectorFromRefs(ArrayRef<const StringRef *> NameRefs) {
   std::vector<std::string> Names;
   Names.reserve(NameRefs.size());
   for (auto *Name : NameRefs)
@@ -454,8 +457,8 @@ Matcher<NamedDecl> hasAnyNameFunc(ArrayRef<const StringRef *> NameRefs) {
       new internal::HasNameMatcher(vectorFromRefs(NameRefs)));
 }
 
-Matcher<ObjCMessageExpr> hasAnySelectorFunc(
-    ArrayRef<const StringRef *> NameRefs) {
+Matcher<ObjCMessageExpr>
+hasAnySelectorFunc(ArrayRef<const StringRef *> NameRefs) {
   return hasAnySelectorMatcher(vectorFromRefs(NameRefs));
 }
 
@@ -697,27 +700,61 @@ static bool isTokenAtLoc(const SourceManager &SM, const LangOptions &LangOpts,
   return !Invalid && Text == TokenText;
 }
 
-std::optional<SourceLocation>
-getExpansionLocOfMacro(StringRef MacroName, SourceLocation Loc,
-                       const ASTContext &Context) {
+namespace {
+struct SourceLocationHash {
+  std::size_t operator()(const SourceLocation &Loc) const {
+    return Loc.getHashValue();
+  }
+};
+
+struct SourceLocationEqual {
+  bool operator()(const SourceLocation &LHS, const SourceLocation &RHS) const {
+    return LHS == RHS;
+  }
+};
+
+} // namespace
+
+static std::optional<SourceLocation> getExpansionLocOfMacroRecursive(
+    StringRef MacroName, SourceLocation Loc, const ASTContext &Context,
+    std::unordered_set<SourceLocation, SourceLocationHash, SourceLocationEqual>
+        &CheckedLocations) {
   auto &SM = Context.getSourceManager();
   const LangOptions &LangOpts = Context.getLangOpts();
   while (Loc.isMacroID()) {
+    if (CheckedLocations.count(Loc)) {
+      return std::nullopt;
+    }
+    CheckedLocations.insert(Loc);
     SrcMgr::ExpansionInfo Expansion =
         SM.getSLocEntry(SM.getFileID(Loc)).getExpansion();
-    if (Expansion.isMacroArgExpansion())
+    if (Expansion.isMacroArgExpansion()) {
       // Check macro argument for an expansion of the given macro. For example,
       // `F(G(3))`, where `MacroName` is `G`.
-      if (std::optional<SourceLocation> ArgLoc = getExpansionLocOfMacro(
-              MacroName, Expansion.getSpellingLoc(), Context))
+      if (std::optional<SourceLocation> ArgLoc =
+              getExpansionLocOfMacroRecursive(MacroName,
+                                              Expansion.getSpellingLoc(),
+                                              Context, CheckedLocations)) {
         return ArgLoc;
+      }
+    }
     Loc = Expansion.getExpansionLocStart();
-    if (isTokenAtLoc(SM, LangOpts, MacroName, Loc))
+    if (isTokenAtLoc(SM, LangOpts, MacroName, Loc)) {
       return Loc;
+    }
   }
   return std::nullopt;
 }
 
+std::optional<SourceLocation>
+getExpansionLocOfMacro(StringRef MacroName, SourceLocation Loc,
+                       const ASTContext &Context) {
+  std::unordered_set<SourceLocation, SourceLocationHash, SourceLocationEqual>
+      CheckedLocations;
+  return getExpansionLocOfMacroRecursive(MacroName, Loc, Context,
+                                         CheckedLocations);
+}
+
 std::shared_ptr<llvm::Regex> createAndVerifyRegex(StringRef Regex,
                                                   llvm::Regex::RegexFlags Flags,
                                                   StringRef MatcherID) {
@@ -744,7 +781,8 @@ const internal::VariadicDynCastAllOfMatcher<Decl, TypeAliasDecl> typeAliasDecl;
 const internal::VariadicDynCastAllOfMatcher<Decl, TypeAliasTemplateDecl>
     typeAliasTemplateDecl;
 const internal::VariadicAllOfMatcher<Decl> decl;
-const internal::VariadicDynCastAllOfMatcher<Decl, DecompositionDecl> decompositionDecl;
+const internal::VariadicDynCastAllOfMatcher<Decl, DecompositionDecl>
+    decompositionDecl;
 const internal::VariadicDynCastAllOfMatcher<Decl, BindingDecl> bindingDecl;
 const internal::VariadicDynCastAllOfMatcher<Decl, LinkageSpecDecl>
     linkageSpecDecl;
@@ -845,8 +883,7 @@ const internal::VariadicDynCastAllOfMatcher<Decl, ObjCCategoryImplDecl>
     objcCategoryImplDecl;
 const internal::VariadicDynCastAllOfMatcher<Decl, ObjCMethodDecl>
     objcMethodDecl;
-const internal::VariadicDynCastAllOfMatcher<Decl, BlockDecl>
-    blockDecl;
+const internal::VariadicDynCastAllOfMatcher<Decl, BlockDecl> blockDecl;
 const internal::VariadicDynCastAllOfMatcher<Decl, ObjCIvarDecl> objcIvarDecl;
 const internal::VariadicDynCastAllOfMatcher<Decl, ObjCPropertyDecl>
     objcPropertyDecl;
@@ -907,7 +944,8 @@ const internal::VariadicDynCastAllOfMatcher<Stmt, CXXRewrittenBinaryOperator>
 const internal::VariadicDynCastAllOfMatcher<Stmt, CXXFoldExpr> cxxFoldExpr;
 const internal::VariadicDynCastAllOfMatcher<Stmt, Expr> expr;
 const internal::VariadicDynCastAllOfMatcher<Stmt, DeclRefExpr> declRefExpr;
-const internal::VariadicDynCastAllOfMatcher<Stmt, ObjCIvarRefExpr> objcIvarRefExpr;
+const internal::VariadicDynCastAllOfMatcher<Stmt, ObjCIvarRefExpr>
+    objcIvarRefExpr;
 const internal::VariadicDynCastAllOfMatcher<Stmt, BlockExpr> blockExpr;
 const internal::VariadicDynCastAllOfMatcher<Stmt, IfStmt> ifStmt;
 const internal::VariadicDynCastAllOfMatcher<Stmt, ForStmt> forStmt;
@@ -937,13 +975,15 @@ const internal::VariadicDynCastAllOfMatcher<Stmt, AsmStmt> asmStmt;
 const internal::VariadicDynCastAllOfMatcher<Stmt, CXXBoolLiteralExpr>
     cxxBoolLiteral;
 const internal::VariadicDynCastAllOfMatcher<Stmt, StringLiteral> stringLiteral;
-const internal::VariadicDynCastAllOfMatcher<Stmt, ObjCStringLiteral> objcStringLiteral;
+const internal::VariadicDynCastAllOfMatcher<Stmt, ObjCStringLiteral>
+    objcStringLiteral;
 const internal::VariadicDynCastAllOfMatcher<Stmt, CharacterLiteral>
     characterLiteral;
 const internal::VariadicDynCastAllOfMatcher<Stmt, IntegerLiteral>
     integerLiteral;
 const internal::VariadicDynCastAllOfMatcher<Stmt, FloatingLiteral> floatLiteral;
-const internal::VariadicDynCastAllOfMatcher<Stmt, ImaginaryLiteral> imaginaryLiteral;
+const internal::VariadicDynCastAllOfMatcher<Stmt, ImaginaryLiteral>
+    imaginaryLiteral;
 const internal::VariadicDynCastAllOfMatcher<Stmt, FixedPointLiteral>
     fixedPointLiteral;
 const internal::VariadicDynCastAllOfMatcher<Stmt, UserDefinedLiteral>
@@ -955,12 +995,10 @@ const internal::VariadicDynCastAllOfMatcher<Stmt, CXXNullPtrLiteralExpr>
 const internal::VariadicDynCastAllOfMatcher<Stmt, ChooseExpr> chooseExpr;
 const internal::VariadicDynCastAllOfMatcher<Stmt, ConvertVectorExpr>
     convertVectorExpr;
-const internal::VariadicDynCastAllOfMatcher<Stmt, CoawaitExpr>
-    coawaitExpr;
+const internal::VariadicDynCastAllOfMatcher<Stmt, CoawaitExpr> coawaitExpr;
 const internal::VariadicDynCastAllOfMatcher<Stmt, DependentCoawaitExpr>
     dependentCoawaitExpr;
-const internal::VariadicDynCastAllOfMatcher<Stmt, CoyieldExpr>
-    coyieldExpr;
+const internal::VariadicDynCastAllOfMatcher<Stmt, CoyieldExpr> coyieldExpr;
 const internal::VariadicDynCastAllOfMatcher<Stmt, GNUNullExpr> gnuNullExpr;
 const internal::VariadicDynCastAllOfMatcher<Stmt, GenericSelectionExpr>
     genericSelectionExpr;

[AaronBallman]

Thank you for working on a fix for this! Please be sure to update clang/docs/ReleaseNotes.rst so users know about the fix.

Test coverage for the changes is interesting in that I'm not certain there's a good test to add for this (the test would be one which is pretty expensive to run, I'd imagine). But if you can think of some good coverage to add to clang/unittests/ASTMatchers/, that'd be good to add as well.

[AaronBallman]

LGTM!

[github-actions]

@mandymimi Congratulations on having your first Pull Request (PR) merged into the LLVM Project!

Your changes will be combined with recent changes from other authors, then tested by our build bots. If there is a problem with a build, you may receive a report in an email or a comment on this PR.

Please check whether problems have been caused by your change specifically, as the builds can include changes from many authors. It is not uncommon for your change to be included in a build that fails due to someone else's changes, or infrastructure issues.

How to do this, and the rest of the post-merge process, is covered in detail here.

If your change does cause a problem, it may be reverted, or you can revert it yourself. This is a normal part of LLVM development. You can fix your changes and open a new PR to merge them again.

If you don't get any reports, no action is required from you. Your changes are working as expected, well done!