Re: [PING] [PATCH] x86_32: memcpy, mempcpy, memmove, memcpy_chk, mempcpy_chk, memmove_chk optimized with SSE2 unaligned loads/stores

[Andrew Senkevich <andrew dot n dot senkevich at gmail dot com>]

2014-12-30 18:25 GMT+03:00 H.J. Lu <hjl.tools@gmail.com>:
> On Tue, Dec 23, 2014 at 08:15:00PM +0100, OndÅej BÃlka wrote:
>> On Mon, Dec 22, 2014 at 02:56:10PM +0300, Andrew Senkevich wrote:
>> > Hi,
>> >
>> > it is ping for patch which was discussed at the beginning of august,
>> > end of discussion is
>> > https://sourceware.org/ml/libc-alpha/2014-08/msg00078.html
>> >
>> > Is it Ok for trunk?
>> >
>> It looked OK for me performancewise. I planned to check that for bugs
>> which is still in my TODO list. If somebody else checked that it would
>> be ok to commit.
>
> I checked it in with a couple fixes:
>
> 1. Add __bcopy_sse2_unaligned.
> 2. Check HAS_SSE2 in ifunc-impl-list.c.
> 3. Replace !defined NOT_IN_libc with IS_IN (libc).
> 4. Check bit_SSE2 before using __xxx_sse2_unaligned.
> 5. Replace CPUID_OFFSET with FEATURE_OFFSET in bit_Fast_Unaligned_Load
> check.
>
> Andrew, please double check my checkin.
>
> Thanks.

Make check ok.
Thank you.


--
WBR,
Andrew