[clang] [clang][ThreadSafety] Instantiate a parameter's attribute naming a later parameter (PR #232784)
Aaron Puchert via cfe-commits
cfe-commits at lists.llvm.org
Sun Nov 1 14:52:15 PST 2026
================
@@ -5493,23 +5520,25 @@ TypeSourceInfo *TemplateDeclInstantiator::SubstFunctionType(
// the function parameters themselves.
const FunctionProtoType *OldProto =
cast<FunctionProtoType>(OldProtoLoc.getType());
- for (unsigned i = 0, i_end = OldProtoLoc.getNumParams(); i != i_end;
- ++i) {
- ParmVarDecl *OldParam = OldProtoLoc.getParam(i);
- if (!OldParam) {
- Params.push_back(SemaRef.BuildParmVarDeclForTypedef(
- D, D->getLocation(), OldProto->getParamType(i)));
- continue;
- }
-
- ParmVarDecl *Parm = SemaRef.SubstParmVarDecl(
- OldParam, TemplateArgs, /*indexAdjustment=*/0,
- /*NumExpansions=*/std::nullopt,
- /*ExpectParameterPack=*/false, EvaluateConstraints);
- if (!Parm)
- return nullptr;
- Params.push_back(Parm);
- }
+ SmallVector<QualType, 4> ParamTypes;
+ Sema::ExtParameterInfoBuilder ExtParamInfos;
+ if (SemaRef.SubstParmTypes(D->getLocation(), OldProtoLoc.getParams(),
+ nullptr, TemplateArgs, ParamTypes, &Params,
+ ExtParamInfos, OldProto->param_type_begin(),
+ EvaluateConstraints))
+ return nullptr;
+ // Declare a parameter that was only a type in the prototype. A function
+ // declared through a typedef with an attribute that rewrites the
+ // function type in place, such as noreturn, gets a prototype whose
+ // parameters have no ParmVarDecl, so no attribute of the typedef's
+ // parameters reaches it; see TypeSpecLocFiller in SemaType.
+ // FIXME: Keep the typedef's parameters there instead. The dependent
+ // branch above returns null for such a prototype, dropping the member.
----------------
aaronpuchert wrote:
> That's certainly a possible cause, though I think noreturn might affect overloads?
I guess what's tricky about `__attribute__((noreturn))` is that it's a type attribute, whereas standard `_Noreturn` (C) and `[[noreturn]]` (C++) are declaration attributes:
```c++
[[noreturn]] void f();
_Noreturn void g();
__attribute__((noreturn)) void h();
```
produces
```
FunctionDecl f 'void ()' external-linkage
`-CXX11NoReturnAttr noreturn
FunctionDecl g 'void ()' external-linkage
`-C11NoReturnAttr
FunctionDecl h 'void () __attribute__((noreturn))' external-linkage
```
My guess is that this causes subtle issues in the handling of these attributes.
> This is also likely one of the hardest problems with getting TSA to be a type attribute.
Yes, that's a problem. But static analysis attributes can be stripped on canonicalization, and even if we do that we still have better coverage than declaration attributes, and (I hope) in some cases a more straightforward implementation.
We don't ever want the attributes to affect mangling, because it shouldn't affect the ABI. I have been thinking in the past about instantiating "shadow" templates which are only used for static analysis and not for code generation, but it's a pipe dream for now and we have more pressing issues.
https://github.com/llvm/llvm-project/pull/232784
More information about the cfe-commits
mailing list