Update sandbox/linux from upstream

Update sandbox/linux/ from https://crrev.com/344610

Pull sandbox changes from upstream, up to
https://crrev.com/344610

This mostly include fixes and cleanups, including for a complex TLS
corruption issue fixed in crbug.com/512623

This required a minimal workaround to work with the older
version of base/ of Mojo.

R=jamesr@chromium.org, mdempsky@chromium.org

Committed: https://chromium.googlesource.com/external/mojo/+/06f35a69f16372fb4f7953f0b83e16725db517de

[jln (very slow on Chromium), 2015-08-22 00:11:31]

jln@chromium.org changed reviewers:
+ mdempsky@chromium.org

[jln (very slow on Chromium), 2015-08-22 00:13:43]

mdempsky: could you take a quick look? I'm rolling upstream Bauxite to Mojo. The
only change from upstream is in thread_helpers.cc to adapt to an older version
of base/

This is probably a rubberstamp.

[mdempsky, 2015-08-22 01:33:36]

lgtm

Looks like difference is GetThreadId vs thread_id?  For future rolls, it might
be easier if one of the Rietveld patch sets is the unmodified Bauxite copy.

[jln (very slow on Chromium), 2015-08-22 01:36:42]

On 2015/08/22 01:33:36, mdempsky wrote:
> lgtm

Thanks!

> Looks like difference is GetThreadId vs thread_id?  For future rolls, it might
> be easier if one of the Rietveld patch sets is the unmodified Bauxite copy.

Yes. And yes, you're right I will do that next time. One issue is that "git cl
upload" starts the trybots automatically with Mojo, so they'll be red for a run
(obviously not a big deal).

[jln (very slow on Chromium), 2015-08-22 01:36:56]

jln@chromium.org changed reviewers:
+ jamesr@chromium.org

[jln (very slow on Chromium), 2015-08-22 01:37:23]

James, do you mind rubberstamping this?

Next time, should I just commit if it only touches sandbox/?

[jln (very slow on Chromium), 2015-08-22 16:55:19]

(This message is sent from the chromium.org origin to see if it makes a
difference)

[jamesr, 2015-08-24 17:27:39]

lgtm

in the future if you're pulling in changes from chromium (or wherever) to
sandbox/ that have already been reviewed feel free to just TBR.  If you
need/want to make some changes in addition to //sandbox (i.e. change up //base
or whatnot) I'd recommend getting a second pair of eyes on that

[jamesr, 2015-08-24 17:29:10]

Also as a style matter whenever pulling from chromium I'd recommend putting an
https://crrev.com/{commit position} link in the commit message (or even title). 
It's terse and unambiguous and having a uniform style for this makes it easier
to find code being copied around.  For this you want https://crrev.com/344610

[jln (very slow on Chromium), 2015-08-24 17:38:36]

On 2015/08/24 17:29:10, jamesr wrote:
> Also as a style matter whenever pulling from chromium I'd recommend putting an
> https://crrev.com/%7Bcommit position} link in the commit message (or even
title). 
> It's terse and unambiguous and having a uniform style for this makes it easier
> to find code being copied around.  For this you want https://crrev.com/344610

Done, thanks!

[jln (very slow on Chromium), 2015-08-24 17:42:13]

Committed patchset #2 (id:20001) manually as
06f35a69f16372fb4f7953f0b83e16725db517de (presubmit successful).