[clang] nolock/noalloc attributes (PR #84983)
via cfe-commits
cfe-commits at lists.llvm.org
Tue Mar 12 16:22:56 PDT 2024
================
@@ -3922,6 +3922,42 @@ bool Sema::MergeFunctionDecl(FunctionDecl *New, NamedDecl *&OldD, Scope *S,
return true;
}
+ const auto OldFX = Old->getFunctionEffects();
+ const auto NewFX = New->getFunctionEffects();
+ if (OldFX != NewFX) {
+ const auto Diffs = FunctionEffectSet::differences(OldFX, NewFX);
+ for (const auto &Item : Diffs) {
+ const FunctionEffect *Effect = Item.first;
+ const bool Adding = Item.second;
+ if (Effect->diagnoseRedeclaration(Adding, *Old, OldFX, *New, NewFX)) {
+ Diag(New->getLocation(),
+ diag::warn_mismatched_func_effect_redeclaration)
+ << Effect->name();
+ Diag(Old->getLocation(), diag::note_previous_declaration);
+ }
+ }
+
+ const auto MergedFX = OldFX | NewFX;
+
+ // Having diagnosed any problems, prevent further errors by applying the
+ // merged set of effects to both declarations.
+ auto applyMergedFX = [&](FunctionDecl *FD) {
+ const auto *FPT = FD->getType()->getAs<FunctionProtoType>();
+ FunctionProtoType::ExtProtoInfo EPI = FPT->getExtProtoInfo();
+ EPI.FunctionEffects = MergedFX;
+ QualType ModQT = Context.getFunctionType(FD->getReturnType(),
+ FPT->getParamTypes(), EPI);
+
+ FD->setType(ModQT);
+ };
+
+ applyMergedFX(Old);
+ applyMergedFX(New);
+
+ OldQType = Old->getType();
----------------
Sirraide wrote:
Assigning to `OldQType` doesn’t seem right here since it gets overwritten five lines down in C++ mode. Generally, I don’t think we really want to modify the `Old` decl here, only the new one.
Also, this function mostly checks if declarations are compatible, from what I recall from the last time I had to change something here. There’s also `Sema::MergeCompatibleFunctionDecls`, which does most of the actual merging (as well as `Sema::MergeCXXFunctionDecl` for C++-specific things).
It might make sense to allow effects to be specified on the first declaration of a function only, and then either copy them from the old to the new decl when merging declarations or, perhaps better, only ever check the first declaration (and diagnosing/ignoring effects specified on later declarations). Unless there is a specific use case for adding/removing effects in redeclarations (or the definition), I would probably recommend going with that—if anything, it should make dealing w/ all of this here easier.
https://github.com/llvm/llvm-project/pull/84983
More information about the cfe-commits
mailing list