Remove unconditional -msse3 -mssse3 CFLAGS from media.gyp.

Remove unconditional -msse3 -mssse3 CFLAGS from media.gyp.

This is confirmed to work by a user who experienced
the crash on illegal instruction because of this,
https://bugs.gentoo.org/show_bug.cgi?id=401537#c27

BUG=107532
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=120485

[Paweł Hajdan Jr., 2012-02-03 09:22:30]

Any objections?

[Ami GONE FROM CHROMIUM, 2012-02-03 18:52:29]

Is this really enough, considering media/base/simd/convert_rgb_to_yuv_ssse3.asm
still uses SSE3/SSSE3 instructions?

[Paweł Hajdan Jr., 2012-02-03 20:16:13]

On 2012/02/03 18:52:29, Ami Fischman wrote:
> Is this really enough, considering
media/base/simd/convert_rgb_to_yuv_ssse3.asm
> still uses SSE3/SSSE3 instructions?

The user reported that crashes stopped after applying this, and I've tested it
on my system (which didn't have the crashes in the first place), and things
continue to build and work fine.

My guess is that there is proper fallback for convert_rgb_to_yuv_ssse3.asm
(run-time detection), but -msse3 -mssse3 tells gcc to _assume_ those are always
available.

[Ami GONE FROM CHROMIUM, 2012-02-03 20:56:20]

LGTM

> The user reported that crashes stopped after applying this

The user reported: "Still many pbs to read videos during one www session,
but that bug is gone."
Do you know what other problems(?) he was seeing?

> My guess is that there is proper fallback for convert_rgb_to_yuv_ssse3.asm
> (run-time detection), but -msse3 -mssse3 tells gcc to _assume_ those are
> always available.

Comparing the diffs in the original submission (r100915) and the resubmit
(r101067) where a bunch of SSSE3-by-default was removed, I'm inclined to
agree with you.

[commit-bot: I haz the power, 2012-02-04 07:16:18]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/phajdan.jr@chromium.org/9309075/1

[commit-bot: I haz the power, 2012-02-04 16:53:40]

Change committed as 120485