[test-suite] r176903 - Perform the atomic operations on a global, rather then a local stack variable.

[Jeffrey Yasskin]

Doesn't this indicate a bug in the atomic support rather than the
test? User code shouldn't need to worry about infinite loops caused by
compare_and_swap. Either the codegen should notice that 'x' isn't
shared and so compile away the atomics entirely, or it should ensure
that even the fast codegen can't interfere with the LL/SC enough to
form an infinite loop.

On Tue, Mar 12, 2013 at 3:07 PM, Chad Rosier <mcrosier at apple.com> wrote:
> Author: mcrosier
> Date: Tue Mar 12 17:07:00 2013
> New Revision: 176903
>
> URL: http://llvm.org/viewvc/llvm-project?rev=176903&view=rev
> Log:
> Perform the atomic operations on a global, rather then a local stack variable.
> This was causing failures on one of our internal -O0 testers.
>
> The issue appears to be that spills/refills to/from the stack are clearing the
> monitor and causing the atomic operations to fail, which results in the benchmark
> going into an infinite loop.  This seems to only happen at -O0 because the
> fast-regalloc is generating a reload in-between the excusive load and exclusive
> store to a stack slot adjacent to the local variable x.
> rdar://13363219
>
> Modified:
>     test-suite/trunk/SingleSource/UnitTests/AtomicOps.c
>
> Modified: test-suite/trunk/SingleSource/UnitTests/AtomicOps.c
> URL: http://llvm.org/viewvc/llvm-project/test-suite/trunk/SingleSource/UnitTests/AtomicOps.c?rev=176903&r1=176902&r2=176903&view=diff
> ==============================================================================
> --- test-suite/trunk/SingleSource/UnitTests/AtomicOps.c (original)
> +++ test-suite/trunk/SingleSource/UnitTests/AtomicOps.c Tue Mar 12 17:07:00 2013
> @@ -5,8 +5,8 @@ int foo(volatile *mem, int val, int c) {
>    return oldval + c;
>  }
>
> +volatile int x = 0;
>  int main() {
> -  volatile int x = 0;
>    int y = foo(&x, 1, 2);
>    printf("%d, %d\n", y, x);
>    y = __sync_val_compare_and_swap(&x, 1, 2);
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits