[cfe-commits] r154659 - in /cfe/trunk: include/clang/Basic/Builtins.def lib/AST/Expr.cpp lib/CodeGen/CGExpr.cpp lib/Sema/SemaChecking.cpp test/CodeGen/atomic-ops.c

[Chris Lattner]

On Apr 13, 2012, at 1:20 PM, Richard Smith wrote:
> * clang implements __atomic_fetch_nand but does not impement __sync_fetch_and_nand.
> 
> That all seems fine to me. However...
> 
> Clang lowers __atomic_fetch_nand(&p, 123, 5) to:
> 
>   %0 = atomicrmw nand i32* @p, i32 123 seq_cst
> 
> ... which is documented as performing *x = ~(*x & y) operation, but actually performs *x = ~*x & y (at least on x86):

Sounds like a (bad!) codegen bug.  It probably dates back to when we did support __sync_fetch_and_nand.  I think that GCC 4.4 is old enough that we could add support for __sync_fetch_and_nand in, with the proper semantics (and fix the codegen bug!)

-Chris

> 
>         movl    $123, %ecx
> .LBB0_1:                                # %entry
>         movl    p, %eax
>         movl    %eax, %edx
>         notl    %edx
>         andl    %ecx, %edx
>         lock
>         cmpxchgl        %edx, p
>         jne     .LBB0_1
> 
> So, is this an LLVM documentation bug (in which case I'll remove these builtins again, since LLVM doesn't provide support for them) or is it an LLVM codegen bug?
> 
> Thanks!
> Richard