Disable assembly code in MemorySanitizer builds.

Disable assembly code in MemorySanitizer builds.

MemorySanitizer is an unitialized memory use detector which is used in
Chromium, and does not presently support assembly code.

BUG=chromium:344505, chromium:373739
R=mtklein@google.com

Committed: https://skia.googlesource.com/skia/+/6a7fdef

[earthdok, 2014-07-02 18:01:49]

Please take a look. If this looks good, could you please land it for me?

[mtklein, 2014-07-02 18:32:52]

LGTM

So, intrinsics are ok, assembly's not.  What about inline assembly?  (This is
just curiosity: the only other assembly we use on x86 is for cpuid, but on ARM
and MIPS we're using assembly all over the place.)

[mtklein, 2014-07-02 18:33:07]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-07-02 18:35:39]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-07-02 18:35:40]

Commit queue rejected this change because it did not recognize the base URL.
Please commit your change manually.

[mtklein_C, 2014-07-02 18:37:32]

Committed patchset #1 manually as r6a7fdef (presubmit successful).

[earthdok, 2014-07-02 18:43:20]

On 2014/07/02 18:32:52, mtklein wrote:
> LGTM
Thanks!

> So, intrinsics are ok, assembly's not.  What about inline assembly?
Not supported yet.

> (This is
> just curiosity: the only other assembly we use on x86 is for cpuid, but on ARM
> and MIPS we're using assembly all over the place.)
MSan is x86-64 only so it's fine.

I think commit-bot fails because I uploaded this patch from a Chromium checkout.
Should I reupload it from a standalone Skia checkout instead?

[earthdok, 2014-07-02 18:43:39]

On 2014/07/02 18:43:20, earthdok wrote:
> On 2014/07/02 18:32:52, mtklein wrote:
> > LGTM
> Thanks!
> 
> > So, intrinsics are ok, assembly's not.  What about inline assembly?
> Not supported yet.
> 
> > (This is
> > just curiosity: the only other assembly we use on x86 is for cpuid, but on
ARM
> > and MIPS we're using assembly all over the place.)
> MSan is x86-64 only so it's fine.
> 
> I think commit-bot fails because I uploaded this patch from a Chromium
checkout.
> Should I reupload it from a standalone Skia checkout instead?

Nevermind.

[mtklein, 2014-07-02 18:45:10]

On 2014/07/02 18:43:39, earthdok wrote:
> On 2014/07/02 18:43:20, earthdok wrote:
> > On 2014/07/02 18:32:52, mtklein wrote:
> > > LGTM
> > Thanks!
> > 
> > > So, intrinsics are ok, assembly's not.  What about inline assembly?
> > Not supported yet.
> > 
> > > (This is
> > > just curiosity: the only other assembly we use on x86 is for cpuid, but on
> ARM
> > > and MIPS we're using assembly all over the place.)
> > MSan is x86-64 only so it's fine.

Gotcha, thanks!
 
> > I think commit-bot fails because I uploaded this patch from a Chromium
> checkout.
> > Should I reupload it from a standalone Skia checkout instead?
> 
> Nevermind.

Sorry, got all proactive on you there.  Yes, uploading from a standalone Skia
checkout is the way to get that to work.  This is something that drives us nuts
too.