On Fri, 8 Nov 2013, Uros Bizjak wrote:
> On Fri, Nov 8, 2013 at 2:13 PM, Joseph S. Myers <jos...@codesourcery.com>
> wrote:
>
> >> Can we introduce a target-dependant source here, in the same way as
> >
> > Sure, that seems a reasonable thing to do. I think putting a file fenv.c
> > in an appropriate subdirectory of libatomic/config will result in it being
> > found automatically by the existing search path logic, but you'll need to
> > test that.
>
> Attached is the x86 optimized implementation of fenv.c. The source
> depends as little as possible on fenv.h definitions - these are
> defined by hardware, and for sure won't change soon in
> hardware-dependant file.
This is fine with me.
> Bootstrapped and regression tested on x86_64-pc-linux-gnu {,-m32}. The
> testing of atomics currently fails on 32bit target due to missing
> __atomic_{load,store}_16 functions, so the patch is not adequately
> tested yet.
(See <http://gcc.gnu.org/ml/gcc-patches/2013-11/msg00894.html> for
commentary on that issue and what might be the right approach for a fix.)
--
Joseph S. Myers
jos...@codesourcery.com
