This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH v4] faster strlen on x64
- From: OndÅej BÃlka <neleai at seznam dot cz>
- To: Richard Henderson <rth at twiddle dot net>
- Cc: Andreas Jaeger <aj at suse dot com>, libc-alpha at sourceware dot org
- Date: Thu, 7 Mar 2013 01:01:05 +0100
- Subject: Re: [PATCH v4] faster strlen on x64
- References: <20130213113840.GA7781@domone.kolej.mff.cuni.cz><5137BB03.7020206@suse.com><5137C254.4080100@twiddle.net>
On Wed, Mar 06, 2013 at 02:25:24PM -0800, Richard Henderson wrote:
> On 2013-03-06 13:54, Andreas Jaeger wrote:
> >>+/* Used in linker - use only %xmm8-%xmm15. */
> >
> >Better expand
> >Say explictely - if it's true: This is used inside the dynamic linker during
> >relocation, thus we cannot touch the usual registers. We can use %xmm8 to %xmm15.
>
> I did already object to using *any* xmm registers in the dynamic linker.
>
> I know for a fact that Intel has a __regcall calling convention in
> place that uses all 16 xmm registers for parameter passing.
>
A press relase introducing it dates 13.2.2013. I moved rtld to separate
patch.
>
> r~