PATCH: optimized sincosf with SSE for x86_64 and x86_32

[Dmitrieva Liubov]

We didn't check in what other places the same bug exists, we were just
aiming to fix "make check" for the proposed patch with optimized
sincos since new implementations raise that latent bug.
That's why I've joined that fix to the patch.

--
Liubov Dmitrieva
Intel Corporation

2012/9/8 Joseph S. Myers <joseph@codesourcery.com>:
> On Sat, 8 Sep 2012, Dmitrieva Liubov wrote:
>
>> 4. Fix latent bugs we have already discussed  regarding with wrong
>> using of sincosf routine for subnormal arguments.
>
> Those should be fixed separately, in a patch that fixes the affected
> functions for all of float, double and long double rather than letting the
> float versions diverge from the double and long double versions.  To keep
> down divergence, the versions fixed should include
> sysdeps/ieee754/ldbl-128ibm/{s_ctanhl.c,s_ctanl.c} as well.
>
> --
> Joseph S. Myers
> joseph@codesourcery.com