Extract device_sensors to /device via Mojofication

Extract device_sensors to /device via Mojofication

Mojofy /content/browser/device_sensors and extract it
into /device to slim /content module.
Also, Extract shared memory seqLock related codes
from /content into /components/shared_memory_seqlock.

Done: Linux gn/gyp build/run OK.
Todo: unit_tests and browser_tests, Android/Win/IOS/ChromeOS build/run.
IntentToImpl is discussed in
https://groups.google.com/a/chromium.org/forum/#!topic/mojo-dev/wa_Wn3HWdFs

TBR=
BUG=

[leonhsl(Using Gerrit), 2015-05-29 11:05:00]

leon.han@intel.com changed reviewers:
+ darin@chromium.org, hans@chromium.org, jamesr@chromium.org,
leandrogracia@chromium.org, timvolodine@chromium.org

[leonhsl(Using Gerrit), 2015-05-29 11:05:02]

This is a draft implementation and could you PTAL?
If you think that it makes sense and worth a try, I want to continue to
accomplish this work. Thanks.

[timvolodine, 2015-05-29 12:22:16]

On 2015/05/29 11:05:02, leon.han wrote:
> This is a draft implementation and could you PTAL?
> If you think that it makes sense and worth a try, I want to continue to
> accomplish this work. Thanks.

wow that's a huge patch thanks! ;)

IMHO this patch makes sense from the perspective of componentization, i.e.
having sensors outside of content and using mojo. It would be nice of course to
use whatever shared memory support there is in mojo. At the time (over half a
year ago) shared memory was not yet available in mojo, but maybe things have
changed now.

think would be good to have an opinion of higher level owners on this before
proceeding.

[leonhsl(Using Gerrit), 2015-06-02 02:57:17]

leon.han@intel.com changed reviewers:
+ armansito@chromium.org, avi@chromium.org, rockot@chromium.org

[leonhsl(Using Gerrit), 2015-06-02 02:57:18]

Thanks timvolodine very much for review comments!
Although this draft CL just employs Mojo MessagePipe, I think you are right and
I'm going to dip into Mojo SharedBuffer to explore how to employ it here to make
the overall work more reasonable.

And, waiting for more comments :)
Loop /device OWNERs armansito@, rockot@ and /content OWNERs avi@.

[leonhsl(Using Gerrit), 2015-06-02 11:44:13]

Please hold review for this CL until I got a solution to employ Mojo
SharedBuffer here. Thanks a lot.

[timvolodine, 2015-06-03 12:09:19]

On 2015/06/02 11:44:13, leon.han wrote:
> Please hold review for this CL until I got a solution to employ Mojo
> SharedBuffer here. Thanks a lot.

leon.han@:
thanks for investigating this. depending on the amount of changes needed for
using mojo shared buffer it may be easier to review in more than one patch
though (i.e. add the shared buffer approach in a separate patch). I'll be happy
to review this patch if you think it's ready and assuming content owners are ok
with it ;)

[leonhsl(Using Gerrit), 2015-06-04 02:35:20]

On 2015/06/03 12:09:19, timvolodine wrote:
> On 2015/06/02 11:44:13, leon.han wrote:
> > Please hold review for this CL until I got a solution to employ Mojo
> > SharedBuffer here. Thanks a lot.
> 
> leon.han@:
> thanks for investigating this. depending on the amount of changes needed for
> using mojo shared buffer it may be easier to review in more than one patch
> though (i.e. add the shared buffer approach in a separate patch). I'll be
happy
> to review this patch if you think it's ready and assuming content owners are
ok
> with it ;)

timvolodine@
Thank you very much for kindly advice. I'll continue to go deep into :)

[leonhsl(Using Gerrit), 2015-06-29 06:55:45]

progress update:
As for Mojo DataPipe, it's employing platform pipe underlying now, maybe employ
platform shared memory in future.
As for Mojo SharedBuffer, currently it can't support sandboxed process receiving
SharedBuffer handle which can be seen from RawChannelWin implementation.
RawChannelWin let receiver process duplicate the platform handle from peer
process but sandboxed process have no such duplicate permission. Obviously our
Renderer Process is sandboxed so we can't employ Mojo SharedBuffer for now.

[Ken Rockot(use gerrit already), 2015-09-29 19:26:31]

rockot@chromium.org changed reviewers:
- rockot@chromium.org