Fix SGI_video_sync cpu usage and rendering issues with Nvidia driver.

Fix SGI_video_sync cpu usage and rendering issues with Nvidia driver.

Using the GLX_SGI_video_sync extension has caused high cpu usage and
rendering problems. The cpu usage was worked around by throttling the
use of the extension to once every 5 seconds. The rendering problems
started happening with Nvidia driver 361.42 and was being worked around
by not using the extension.

This CL fixes both problems by doing the following:
1. Using a separate unmapped child window (and GLX drawable) for use
   with SGI_video_sync.
2. Create a dummy unmapped window on both the main X Display and the
   video sync Display before the sandbox is set up.
3. Whitelist the Nvidia device files that are opened after the sandbox
   is set up.

This CL also makes all vsync providers use a single GLXContext because
there's only one thread that's used for all of them.

R=piman@chromium.org,rickyz@chromium.org
BUG=636644, 483721
TEST=manual testing with i3 wm (no compositing)

Committed: https://crrev.com/ffc0e467329470df9c2da085d133d59cc8ae2414
Cr-Commit-Position: refs/heads/master@{#417747}

[sunnyps, 2016-08-30 22:05:08]

The CQ bit was checked by sunnyps@chromium.org to run a CQ dry run

[sunnyps, 2016-08-30 22:05:26]

PTAL

[commit-bot: I haz the power, 2016-08-30 22:05:31]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[sunnyps, 2016-08-30 22:07:07]

Let me know if you want to split this CL. Also note that /dev/nvidiactl is not
strictly needed to be sandboxed given that we create the dummy window for video
sync before the sandbox is set up but /dev/nvidia0 and
/proc/driver/nvidia/params do need to be whitelisted.

[commit-bot: I haz the power, 2016-08-30 22:08:00]

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

[commit-bot: I haz the power, 2016-08-30 22:08:01]

Dry run: Try jobs failed on following builders:
  ios-device on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/ios-device/builds...)
  ios-simulator on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/ios-simulator/bui...)
  mac_chromium_compile_dbg_ng on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_comp...)
  mac_chromium_rel_ng on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[sunnyps, 2016-08-30 22:53:54]

The CQ bit was checked by sunnyps@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-30 22:54:32]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[sunnyps, 2016-08-30 22:54:55]

Description was changed from

==========
Fix SGI_video_sync cpu usage and rendering issues with Nvidia driver.

Using the GLX_SGI_video_sync extension has caused high cpu usage and
rendering problems. The cpu usage was worked around by throttling the
use of the extension to once every 5 seconds. The rendering problems
started happening with Nvidia driver 361.42 and was being worked around
by not using the extension.

This CL fixes both problems by doing the following:
1. Using a separate unmapped child window (and GLX drawable) for use
   with SGI_video_sync.
2. Create a dummy unmapped window on both the main X Display and the
   video sync Display before the sandbox is set up.
3. Whitelist the Nvidia device files that are opened after the sandbox
   is set up.

This CL also makes all vsync providers use a single GLXContext because
there is only one thread that's used for all of them.

R=piman@chromium.org,rickyz@chromium.org
BUG=636644,483721
TEST=manual testing with i3 wm (no compositing)
==========

to

==========
Fix SGI_video_sync cpu usage and rendering issues with Nvidia driver.

Using the GLX_SGI_video_sync extension has caused high cpu usage and
rendering problems. The cpu usage was worked around by throttling the
use of the extension to once every 5 seconds. The rendering problems
started happening with Nvidia driver 361.42 and was being worked around
by not using the extension.

This CL fixes both problems by doing the following:
1. Using a separate unmapped child window (and GLX drawable) for use
   with SGI_video_sync.
2. Create a dummy unmapped window on both the main X Display and the
   video sync Display before the sandbox is set up.
3. Whitelist the Nvidia device files that are opened after the sandbox
   is set up.

This CL also makes all vsync providers use a single GLXContext because
there's only one thread that's used for all of them.

R=piman@chromium.org,rickyz@chromium.org
BUG=636644,483721
TEST=manual testing with i3 wm (no compositing)
==========

[commit-bot: I haz the power, 2016-08-30 23:56:36]

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

[commit-bot: I haz the power, 2016-08-30 23:56:37]

Dry run: This issue passed the CQ dry run.

[piman, 2016-08-31 00:28:03]

lgtm

https://codereview.chromium.org/2291373002/diff/20001/ui/gl/gl_surface_glx.cc
File ui/gl/gl_surface_glx.cc (right):

https://codereview.chromium.org/2291373002/diff/20001/ui/gl/gl_surface_glx.cc...
ui/gl/gl_surface_glx.cc:208: // This ensures that creation of |parnet_window_|
has occured when this shim
nit: typo parnet->parent

https://codereview.chromium.org/2291373002/diff/20001/ui/gl/gl_surface_glx.cc...
ui/gl/gl_surface_glx.cc:425: if (!CreateDummyWindow(g_display)) {
Are we doing this to force the driver to connect/pre-allocate things? If so,
could you add a comment about it?

[sunnyps, 2016-09-01 18:42:28]

The CQ bit was checked by sunnyps@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-09-01 18:42:53]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[sunnyps, 2016-09-01 18:45:15]

Addressed comments. I made the code more robust for multi-GPU setups where
/dev/nvidiaN could be the GPU that's used for output. I also did the same for
/dev/dri/cardN and added these permissions to the Linux non-ChromeOS i.e.
desktop case only.

rickyz@ Can you PTAL?

https://codereview.chromium.org/2291373002/diff/20001/ui/gl/gl_surface_glx.cc
File ui/gl/gl_surface_glx.cc (right):

https://codereview.chromium.org/2291373002/diff/20001/ui/gl/gl_surface_glx.cc...
ui/gl/gl_surface_glx.cc:208: // This ensures that creation of |parnet_window_|
has occured when this shim
On 2016/08/31 at 00:28:02, piman wrote:
> nit: typo parnet->parent

Done.

https://codereview.chromium.org/2291373002/diff/20001/ui/gl/gl_surface_glx.cc...
ui/gl/gl_surface_glx.cc:425: if (!CreateDummyWindow(g_display)) {
On 2016/08/31 at 00:28:02, piman wrote:
> Are we doing this to force the driver to connect/pre-allocate things? If so,
could you add a comment about it?

Done.

[commit-bot: I haz the power, 2016-09-01 19:32:19]

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

[commit-bot: I haz the power, 2016-09-01 19:32:20]

Dry run: This issue passed the CQ dry run.

[rickyz (no longer on Chrome), 2016-09-08 05:01:38]

lgtm

[sunnyps, 2016-09-09 21:49:06]

The CQ bit was checked by sunnyps@chromium.org

[sunnyps, 2016-09-09 21:49:06]

The patchset sent to the CQ was uploaded after l-g-t-m from piman@chromium.org,
rickyz@chromium.org
Link to the patchset: https://codereview.chromium.org/2291373002/#ps100001
(title: "rebase")

[commit-bot: I haz the power, 2016-09-09 21:49:43]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-09 22:53:18]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2016-09-09 22:55:48]

Description was changed from

==========
Fix SGI_video_sync cpu usage and rendering issues with Nvidia driver.

Using the GLX_SGI_video_sync extension has caused high cpu usage and
rendering problems. The cpu usage was worked around by throttling the
use of the extension to once every 5 seconds. The rendering problems
started happening with Nvidia driver 361.42 and was being worked around
by not using the extension.

This CL fixes both problems by doing the following:
1. Using a separate unmapped child window (and GLX drawable) for use
   with SGI_video_sync.
2. Create a dummy unmapped window on both the main X Display and the
   video sync Display before the sandbox is set up.
3. Whitelist the Nvidia device files that are opened after the sandbox
   is set up.

This CL also makes all vsync providers use a single GLXContext because
there's only one thread that's used for all of them.

R=piman@chromium.org,rickyz@chromium.org
BUG=636644,483721
TEST=manual testing with i3 wm (no compositing)
==========

to

==========
Fix SGI_video_sync cpu usage and rendering issues with Nvidia driver.

Using the GLX_SGI_video_sync extension has caused high cpu usage and
rendering problems. The cpu usage was worked around by throttling the
use of the extension to once every 5 seconds. The rendering problems
started happening with Nvidia driver 361.42 and was being worked around
by not using the extension.

This CL fixes both problems by doing the following:
1. Using a separate unmapped child window (and GLX drawable) for use
   with SGI_video_sync.
2. Create a dummy unmapped window on both the main X Display and the
   video sync Display before the sandbox is set up.
3. Whitelist the Nvidia device files that are opened after the sandbox
   is set up.

This CL also makes all vsync providers use a single GLXContext because
there's only one thread that's used for all of them.

R=piman@chromium.org,rickyz@chromium.org
BUG=636644,483721
TEST=manual testing with i3 wm (no compositing)

Committed: https://crrev.com/ffc0e467329470df9c2da085d133d59cc8ae2414
Cr-Commit-Position: refs/heads/master@{#417747}
==========

[commit-bot: I haz the power, 2016-09-09 22:55:50]

Patchset 6 (id:??) landed as
https://crrev.com/ffc0e467329470df9c2da085d133d59cc8ae2414
Cr-Commit-Position: refs/heads/master@{#417747}