[arm/Linux] Don't rely on KUSER_HELPERS feature

[arm/Linux] Don't rely on KUSER_HELPERS feature

ARM specific CONFIG_KUSER_HELPERS kernel feature for Linux can be disabled,
and in this case, we shouldn't crash. Use a __sync_synchronize() call
instead for Linux platforms.

BUG=chromium:599051
LOG=Y

Committed: https://crrev.com/0b557da25fd26e56fb97f634542844de8cd63327
Cr-Commit-Position: refs/heads/master@{#35170}

[jbriance, 2016-03-30 12:00:51]

Description was changed from

==========
[arm/Linux] Don't rely on KUSER_HELPERS feature

ARM specific CONFIG_KUSER_HELPERS kernel feature for Linux can be disabled,
and in this case, we shouldn't crash. Use a __sync_synchronize() call
instead for Linux platforms.

BUG=chromium:599051
==========

to

==========
[arm/Linux] Don't rely on KUSER_HELPERS feature

ARM specific CONFIG_KUSER_HELPERS kernel feature for Linux can be disabled,
and in this case, we shouldn't crash. Use a __sync_synchronize() call
instead for Linux platforms.

BUG=chromium:599051
==========

[jbriance, 2016-03-30 12:00:51]

jbriance@cisco.com changed reviewers:
+ bmeurer@chromium.org, danno@chromium.org, rodolph.perfetta@arm.com

[jbriance, 2016-03-30 13:44:55]

Note: change done in chromium/src:
https://chromium.googlesource.com/chromium/src/+/d65c3c2511404c3c5a5a5f304eb3...
(https://codereview.chromium.org/1842003002/)

[jbriance, 2016-03-30 16:07:14]

jbriance@cisco.com changed reviewers:
+ jochen@chromium.org

[jochen (gone - plz use gerrit), 2016-03-30 16:53:48]

Rodolph, can you comment on this? Is it safe performance-wise to always fallback
to the GCC builtin on linux?

https://cs.chromium.org/0xffff0fa0 implies that this function is also used in
several other libraries included in chromium btw..

https://codereview.chromium.org/1840203004/diff/1/AUTHORS
File AUTHORS (right):

https://codereview.chromium.org/1840203004/diff/1/AUTHORS#newcode116
AUTHORS:116: 柳荣一 <admin@web-tinker.com>
can you undo this change please?

[jbriance, 2016-03-30 20:49:16]

On 2016/03/30 16:53:48, jochen wrote:
> https://cs.chromium.org/0xffff0fa0 implies that this function is also used in
> several other libraries included in chromium btw..

Yes, I saw that. At least the re2.git one is obsolete now (thanks to
https://chromium.googlesource.com/external/github.com/google/re2/+/cd505f4597...),
but I plan to update the remaining references to these hardcoded calls for
ARM/Linux.

[jbriance, 2016-03-30 20:50:34]

https://codereview.chromium.org/1840203004/diff/1/AUTHORS
File AUTHORS (right):

https://codereview.chromium.org/1840203004/diff/1/AUTHORS#newcode116
AUTHORS:116: 柳荣一 <admin@web-tinker.com>
On 2016/03/30 16:53:47, jochen wrote:
> can you undo this change please?

Sure, will do tomorrow morning

[jbriance, 2016-03-31 07:09:01]

https://codereview.chromium.org/1840203004/diff/1/AUTHORS
File AUTHORS (right):

https://codereview.chromium.org/1840203004/diff/1/AUTHORS#newcode116
AUTHORS:116: 柳荣一 <admin@web-tinker.com>
On 2016/03/30 20:50:33, jbriance wrote:
> On 2016/03/30 16:53:47, jochen wrote:
> > can you undo this change please?
> 
> Sure, will do tomorrow morning

Done.

[Rodolph Perfetta (ARM), 2016-03-31 09:50:51]

On 2016/03/30 16:53:48, jochen wrote:
> Rodolph, can you comment on this? Is it safe performance-wise to always
fallback
> to the GCC builtin on linux?
> 
> https://cs.chromium.org/0xffff0fa0 implies that this function is also used in
> several other libraries included in chromium btw..
> 
> https://codereview.chromium.org/1840203004/diff/1/AUTHORS
> File AUTHORS (right):
> 
> https://codereview.chromium.org/1840203004/diff/1/AUTHORS#newcode116
> AUTHORS:116: 柳荣一 <mailto:admin@web-tinker.com>
> can you undo this change please?

Using the GCC builtin should be fine.

[jochen (gone - plz use gerrit), 2016-03-31 15:27:32]

lgtm

[jbriance, 2016-03-31 15:28:28]

The CQ bit was checked by jbriance@cisco.com

[commit-bot: I haz the power, 2016-03-31 15:28:38]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1840203004/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1840203004/20001

[commit-bot: I haz the power, 2016-03-31 15:31:29]

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

[commit-bot: I haz the power, 2016-03-31 15:31:30]

Try jobs failed on following builders:
  v8_presubmit on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/13061)

[jbriance, 2016-03-31 15:52:14]

Description was changed from

==========
[arm/Linux] Don't rely on KUSER_HELPERS feature

ARM specific CONFIG_KUSER_HELPERS kernel feature for Linux can be disabled,
and in this case, we shouldn't crash. Use a __sync_synchronize() call
instead for Linux platforms.

BUG=chromium:599051
==========

to

==========
[arm/Linux] Don't rely on KUSER_HELPERS feature

ARM specific CONFIG_KUSER_HELPERS kernel feature for Linux can be disabled,
and in this case, we shouldn't crash. Use a __sync_synchronize() call
instead for Linux platforms.

BUG=chromium:599051
LOG=Y
==========

[jbriance, 2016-03-31 15:52:32]

The CQ bit was checked by jbriance@cisco.com

[commit-bot: I haz the power, 2016-03-31 15:52:41]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1840203004/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1840203004/20001

[commit-bot: I haz the power, 2016-03-31 15:54:16]

Description was changed from

==========
[arm/Linux] Don't rely on KUSER_HELPERS feature

ARM specific CONFIG_KUSER_HELPERS kernel feature for Linux can be disabled,
and in this case, we shouldn't crash. Use a __sync_synchronize() call
instead for Linux platforms.

BUG=chromium:599051
LOG=Y
==========

to

==========
[arm/Linux] Don't rely on KUSER_HELPERS feature

ARM specific CONFIG_KUSER_HELPERS kernel feature for Linux can be disabled,
and in this case, we shouldn't crash. Use a __sync_synchronize() call
instead for Linux platforms.

BUG=chromium:599051
LOG=Y
==========

[commit-bot: I haz the power, 2016-03-31 15:54:17]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-03-31 15:56:12]

Description was changed from

==========
[arm/Linux] Don't rely on KUSER_HELPERS feature

ARM specific CONFIG_KUSER_HELPERS kernel feature for Linux can be disabled,
and in this case, we shouldn't crash. Use a __sync_synchronize() call
instead for Linux platforms.

BUG=chromium:599051
LOG=Y
==========

to

==========
[arm/Linux] Don't rely on KUSER_HELPERS feature

ARM specific CONFIG_KUSER_HELPERS kernel feature for Linux can be disabled,
and in this case, we shouldn't crash. Use a __sync_synchronize() call
instead for Linux platforms.

BUG=chromium:599051
LOG=Y

Committed: https://crrev.com/0b557da25fd26e56fb97f634542844de8cd63327
Cr-Commit-Position: refs/heads/master@{#35170}
==========

[commit-bot: I haz the power, 2016-03-31 15:56:13]

Patchset 2 (id:??) landed as
https://crrev.com/0b557da25fd26e56fb97f634542844de8cd63327
Cr-Commit-Position: refs/heads/master@{#35170}