[Clang][C] Fixed a bug where we reject an _Atomic qualified integer in a switch statment

[to268]

Summary

We are currently rejecting an _Atomic qualified integer in a switch statment.
This fixes the issue by doing an Lvalue conversion before trying to match on the type.
Fixes #65557

[aaron.ballman]

Thank you for working on this! You should also add a release note for the fix.

clang/lib/Sema/SemaOverload.cpp

6306–6310

I think this is a slightly cleaner version of what you were doing.

[to268]

I have added the missing ReleaseNote.

[to268]

Added recommendation by @aaron.ballman

[aaron.ballman]

LGTM (with a tiny nit with the formatting in the release notes), thank you for the fix!

clang/docs/ReleaseNotes.rst

229

[to268]

I have fixed the ReleaseNote formatting error.
Please land this patch on my behalf "Guillot Tony <[e-mail]>"

[cor3ntin]

@to268 as discussed, i landed the change on your behalf. Thanks for the patch!

clang/docs/ReleaseNotes.rst

 - Support MSVC predefined macro expressions in constant expressions and in
   local structs.
 - Correctly parse non-ascii identifiers that appear immediately after a line splicing
   (`#65156 <https://github.com/llvm/llvm-project/issues/65156>`_)
 - Clang no longer considers the loss of ``__unaligned`` qualifier from objects as
   an invalid conversion during method function overload resolution.
 - Fix parser crash when dealing with ill-formed objective C++ header code. Fixes
   (`#64836 <https://github.com/llvm/llvm-project/issues/64836>`_)
+- Clang now allows an ``_Atomic`` qualified integer in a switch statement. Fixes
+  (`#65557 <https://github.com/llvm/llvm-project/issues/65557>`_)
 
 Bug Fixes to Compiler Builtins
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Bug Fixes to Attribute Support
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
 Bug Fixes to C++ Support

clang/lib/Sema/SemaOverload.cpp

ExprResult Sema::PerformContextualImplicitConversion(
   // Process placeholders immediately.
   if (From->hasPlaceholderType()) {
     ExprResult result = CheckPlaceholderExpr(From);
     if (result.isInvalid())
       return result;
     From = result.get();
   }
 
+  // Try converting the expression to an Lvalue first, to get rid of qualifiers.
+  ExprResult Converted = DefaultLvalueConversion(From);
+  QualType T = Converted.isUsable() ? Converted.get()->getType() : QualType();
   // If the expression already has a matching type, we're golden.
-  QualType T = From->getType();
   if (Converter.match(T))
-    return DefaultLvalueConversion(From);
+    return Converted;
 
   // FIXME: Check for missing '()' if T is a function type?
 
   // We can only perform contextual implicit conversions on objects of class
   // type.
   const RecordType *RecordTy = T->getAs<RecordType>();
   if (!RecordTy || !getLangOpts().CPlusPlus) {
     if (!Converter.Suppress)

clang/test/Sema/atomic-expr.c

void func_18(void) {
   data2 = (_Atomic int)0;
   (void)(_Atomic(int *))0;
 
   // But that we correctly reject casts to atomic aggregate types.
   struct S { int a; } s;
   struct T { int a; };
   (void)(_Atomic struct T)s; // expected-error {{used type 'struct T' where arithmetic or pointer type is required}}
 }
+
+// Test if we can handle an _Atomic qualified integer in a switch statement.
+void func_19(void) {
+  _Atomic int a = 0;
+  switch (a) { }
+}