[clang] [clang] Diagnose invalid identifiers in gnu::abi_tag attribute arguments (PR #169060)

[via cfe-commits]

https://github.com/camc created https://github.com/llvm/llvm-project/pull/169060

Fixes #168901 

Adds a `err_attribute_string_literal_invalid_ident` diagnostic and emits it when a `gnu::abi_tag` attribute has a string argument with an invalid identifier.

>From 02ec7f0e3948454b02c2fffd87d26010c9bdf9b7 Mon Sep 17 00:00:00 2001
From: camc <pushy-crop-cartel at duck.com>
Date: Fri, 21 Nov 2025 16:13:54 +0000
Subject: [PATCH] [clang] Diagnose invalid identifiers in gnu::abi_tag
 attribute arguments

---
 clang/docs/ReleaseNotes.rst                     |  2 ++
 .../include/clang/Basic/DiagnosticSemaKinds.td  |  2 ++
 clang/lib/Sema/SemaDeclAttr.cpp                 | 17 +++++++++++++++++
 clang/test/SemaCXX/attr-abi-tag-syntax.cpp      | 10 ++++++++++
 4 files changed, 31 insertions(+)

diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
index 62c8c0130c3d0..9a5efe97d0595 100644
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -449,6 +449,8 @@ Improvements to Clang's diagnostics
 - A new warning ``-Wenum-compare-typo`` has been added to detect potential erroneous
   comparison operators when mixed with bitwise operators in enum value initializers.
   This can be locally disabled by explicitly casting the initializer value.
+- Clang now emits a diagnostic when an invalid identifier string is passed to
+  the `gnu::abi_tag` attribute (#GH168901).
 
 Improvements to Clang's time-trace
 ----------------------------------
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index 53aa86a7dabde..ce52066d38957 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -3326,6 +3326,8 @@ def err_attribute_no_member_function : Error<
 def err_attribute_parameter_types : Error<
   "%0 attribute parameter types do not match: parameter %1 of function %2 has type %3, "
   "but parameter %4 of function %5 has type %6">;
+def err_attribute_string_literal_invalid_ident : Error<
+  "%0 attribute parameters must be string literals containing valid identifiers">;
 
 def err_attribute_too_many_arguments : Error<
   "%0 attribute takes no more than %1 argument%s1">;
diff --git a/clang/lib/Sema/SemaDeclAttr.cpp b/clang/lib/Sema/SemaDeclAttr.cpp
index e3af5023c74d0..bd941f0aa05f3 100644
--- a/clang/lib/Sema/SemaDeclAttr.cpp
+++ b/clang/lib/Sema/SemaDeclAttr.cpp
@@ -6172,12 +6172,29 @@ static void handleMSConstexprAttr(Sema &S, Decl *D, const ParsedAttr &AL) {
   D->addAttr(::new (S.Context) MSConstexprAttr(S.Context, AL));
 }
 
+static bool checkIsRawIdentifier(const Sema &S, StringRef StrRef) {
+  std::string Str = StrRef.str();
+  auto FileLoc = S.getSourceManager().getLocForStartOfFile(
+      S.getSourceManager().getMainFileID());
+  Lexer Lex(FileLoc, S.getLangOpts(), Str.c_str(), Str.c_str(),
+      Str.c_str() + Str.size());
+  Token Tok;
+  bool LexedToEnd = Lex.LexFromRawLexer(Tok);
+  return LexedToEnd && Tok.is(tok::raw_identifier);
+}
+
 static void handleAbiTagAttr(Sema &S, Decl *D, const ParsedAttr &AL) {
   SmallVector<StringRef, 4> Tags;
   for (unsigned I = 0, E = AL.getNumArgs(); I != E; ++I) {
     StringRef Tag;
     if (!S.checkStringLiteralArgumentAttr(AL, I, Tag))
       return;
+    if (!checkIsRawIdentifier(S, Tag)) {
+      S.Diag(AL.getArgAsExpr(I)->getBeginLoc(),
+             diag::err_attribute_string_literal_invalid_ident)
+          << AL;
+      return;
+    }
     Tags.push_back(Tag);
   }
 
diff --git a/clang/test/SemaCXX/attr-abi-tag-syntax.cpp b/clang/test/SemaCXX/attr-abi-tag-syntax.cpp
index 4f14a3c043b5c..ea3f5b184aba2 100644
--- a/clang/test/SemaCXX/attr-abi-tag-syntax.cpp
+++ b/clang/test/SemaCXX/attr-abi-tag-syntax.cpp
@@ -31,3 +31,13 @@ extern int a2;
 // expected-note at -1 {{previous declaration is here}}
 __attribute__((abi_tag("A")))extern int a2;
 // expected-error at -1 {{cannot add 'abi_tag' attribute in a redeclaration}}
+
+[[gnu::abi_tag("[[something]]")]] // expected-error {{'gnu::abi_tag' attribute parameters must be string literals containing valid identifiers}}
+int f1() { 
+	return 0;
+}
+
+[[gnu::abi_tag("something")]]
+int f2() { 
+	return 0;
+}