This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] Disable x87 inline functions for x86_64 and SSE [BZ #17262]
- From: "H.J. Lu" <hjl dot tools at gmail dot com>
- To: Siddhesh Poyarekar <siddhesh at redhat dot com>
- Cc: GNU C Library <libc-alpha at sourceware dot org>, Jakub Jelinek <jakub at redhat dot com>
- Date: Tue, 12 Aug 2014 11:48:44 -0700
- Subject: Re: [PATCH] Disable x87 inline functions for x86_64 and SSE [BZ #17262]
- Authentication-results: sourceware.org; auth=none
- References: <20140812183008 dot GO9284 at spoyarek dot pnq dot redhat dot com>
On Tue, Aug 12, 2014 at 11:30 AM, Siddhesh Poyarekar
<siddhesh@redhat.com> wrote:
> Hi,
>
> Since:
>
> commit 409e00bd69b8d8dd74d7327085351d26769ea6fc
> Author: H.J. Lu <hjl.tools@gmail.com>
> Date: Wed Jan 29 07:51:41 2014 -0800
>
> Disable x87 inline functions for SSE2 math
>
> When i386 and x86-64 mathinline.h was merged into a single mathinline.h,
> "gcc -m32" enables x87 inline functions on x86-64 even when -mfpmath=sse
> and SSE2 is enabled. It is a regression on x86-64. We should check
> __SSE2_MATH__ instead of __x86_64__ when disabling x87 inline functions.
>
>
> gcc-3.2 is unable to correctly compile x86_64 routines for llrint
> since it gets redefined. This is because gcc 3.2 does not set
> __SSE2_MATH__ for x86_64, thus exposing the duplicate definition.
>
> The correct fix ought to be to check for both __SSE2_MATH__ and
> __x86_64__ and enable those bits only when neither are defined.
>
> Tested fix with the reproducer for
> 409e00bd69b8d8dd74d7327085351d26769ea6fc as well as with gcc-3.2.
>
> OK for 2.20?
>
> Thanks,
> Siddhesh
>
> [BZ #17262]
> * sysdeps/x86/fpu/bits/mathinline.h: Check both __SSE2_MATH__
> and __x86_64__ when disabling x87 inline functions.
>
> diff --git a/sysdeps/x86/fpu/bits/mathinline.h b/sysdeps/x86/fpu/bits/mathinline.h
> index 9c32e95..ee88b66 100644
> --- a/sysdeps/x86/fpu/bits/mathinline.h
> +++ b/sysdeps/x86/fpu/bits/mathinline.h
> @@ -384,7 +384,10 @@ __END_NAMESPACE_C99
> # endif
> #endif
>
> -#ifndef __SSE2_MATH__
> +/* Disable x87 inlines when -fpmath=sse is passed and also when we're building
> + on x86_64. Older gcc (gcc-3.2 for example) does not set __SSE2_MATH__
> + for x86_64. */
> +#if !defined __SSE2_MATH__ && !defined __x86_64__
> # if ((!defined __NO_MATH_INLINES || defined __LIBC_INTERNAL_MATH_INLINES) \
> && defined __OPTIMIZE__)
>
> @@ -970,4 +973,4 @@ __inline_mathcode2 (__ieee754_atan2, __y, __x,
> return __value;)
> # endif
>
> -#endif /* !__SSE2_MATH__ */
> +#endif /* !__SSE2_MATH__ && !__x86_64__ */
Will this patch get x86 inline functions with -fpmath=sse -m32 for
GCC 3.2?
--
H.J.