This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 6/7] Use SSE optimized strcmp in x86-64 ld.so
- From: OndÅej BÃlka <neleai at seznam dot cz>
- To: "H.J. Lu" <hjl dot tools at gmail dot com>
- Cc: GNU C Library <libc-alpha at sourceware dot org>
- Date: Tue, 25 Aug 2015 18:54:25 +0200
- Subject: Re: [PATCH 6/7] Use SSE optimized strcmp in x86-64 ld.so
- Authentication-results: sourceware.org; auth=none
- References: <20150825122408 dot GE1588 at gmail dot com> <20150825140022 dot GA19842 at domone> <CAMe9rOqXzF9Dmu67EUOBHWxppq7d5jUD3XB8aWG=pUWQbgDONQ at mail dot gmail dot com>
On Tue, Aug 25, 2015 at 07:35:50AM -0700, H.J. Lu wrote:
> On Tue, Aug 25, 2015 at 7:00 AM, OndÅej BÃlka <neleai@seznam.cz> wrote:
> > On Tue, Aug 25, 2015 at 05:24:08AM -0700, H.J. Lu wrote:
> >> Since ld.so preserves vector registers now, we can SSE optimized strcmp
> >> in x86-64 ld.so.
> >>
> >> OK for master?
> >>
> > Not completely, while I agree that we should use optimized strcmp I
> > think that better default would be strcmp_sse2_unaligned.
> >
> > Its best for processors from core2 upwards and works everywhere as it
> > uses only sse2.
> >
> > I forgot to ask about changing that when I submitted that patch.
>
> I think it is a good idea to make strcmp_sse2_unaligned as the
> default and uses __strcmp_sse2 if Fast_Unaligned_Load is
> off.
>
Which does some thing as ifunc, but its possible.