On Thu, Jul 30, 2015 at 3:02 PM, Richard Sandiford
<richard.sandif...@arm.com> wrote:
> expand_builtin_atomic_clear has code to handle an atomic_clear instruction.
> However, no target defines such a pattern, and the code wouldn't compile
> on any target that tried to define it, since the generator is passed the
> memory model enum directly rather than via an rtx wrapper. It seems better
> to remove the code rather than try to update it.
>
> No documentation changes are needed because the pattern isn't mentioned
> in md.texi.
>
> Bootstrapped & regression-tested on x86_64-linux-gnu. Grepped for references
> to atomic_clear to make sure that no such pattern exists. OK to install?
Ok.
Thanks,
Richard.
> Thanks,
> Richard
>
> gcc/
> * builtins.c (HAVE_atomic_clear, gen_atomic_clear): Delete.
> (expand_builtin_atomic_clear): Remove support for atomic_clear
> pattern.
>
> diff --git a/gcc/builtins.c b/gcc/builtins.c
> index 1750e25..d7eb65f 100644
> --- a/gcc/builtins.c
> +++ b/gcc/builtins.c
> @@ -5587,12 +5587,6 @@ expand_builtin_atomic_fetch_op (machine_mode mode,
> tree exp, rtx target,
> return ret;
> }
>
> -
> -#ifndef HAVE_atomic_clear
> -# define HAVE_atomic_clear 0
> -# define gen_atomic_clear(x,y) (gcc_unreachable (), NULL_RTX)
> -#endif
> -
> /* Expand an atomic clear operation.
> void _atomic_clear (BOOL *obj, enum memmodel)
> EXP is the call expression. */
> @@ -5615,12 +5609,6 @@ expand_builtin_atomic_clear (tree exp)
> model = MEMMODEL_SEQ_CST;
> }
>
> - if (HAVE_atomic_clear)
> - {
> - emit_insn (gen_atomic_clear (mem, model));
> - return const0_rtx;
> - }
> -
> /* Try issuing an __atomic_store, and allow fallback to
> __sync_lock_release.
> Failing that, a store is issued by __atomic_store. The only way this
> can
> fail is if the bool type is larger than a word size. Unlikely, but
>
