The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/56b183a214c15e788a6eeeafa869578b3ad90365

commit 56b183a214c15e788a6eeeafa869578b3ad90365
Author: Jonathan <jonross@chromium.org>
Date: Mon Jul 10 21:31:20 2017

Don't try to init mash_service in mus_browser_tests

The mash_service is not available/used when running chrome --mus
The mus_browser_tests were still trying to launch the service, leading to spammy
logs of the failure to launch the service.

This updates MojoTestState to only launch the service in --run-in-mash mode.

TBR=sky@chromium.org
TEST=mash_browser_tests, mus_browser_tests

Bug:  740655 
Change-Id: I1f7618f5db5106b6f92e34647b6b79f5395dc857
Reviewed-on: https://chromium-review.googlesource.com/565171
Reviewed-by: Jonathan Ross <jonross@chromium.org>
Commit-Queue: Jonathan Ross <jonross@chromium.org>
Cr-Commit-Position: refs/heads/master@{#485400}
[modify] https://crrev.com/56b183a214c15e788a6eeeafa869578b3ad90365/chrome/test/base/mojo_test_connector.cc

This is currently blocked by an error being reached in PlatformChannelPair's ctor:

https://cs.chromium.org/chromium/src/mojo/edk/embedder/platform_channel_pair_posix.cc?rcl=eb332a66dcdd2191f6410c7c69137caa4701e265&l=68

We are hitting the socked limit on trybots.

This is because the test framework creates N MojoTestStates all at once, before running any tests. So we end up trying to create N ServiceManagers and the pipes for each child test process, along with all the related services.

We need to update how content::LaunchTests does it's batching of TestStates, or find a later point to do initialization.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/25bed6f775cd0b22e31f462c9930c2c17dcd1452

commit 25bed6f775cd0b22e31f462c9930c2c17dcd1452
Author: Scott Violet <sky@chromium.org>
Date: Wed Aug 09 21:28:59 2017

chromeos: fix potential use-after-free

BrowserProcessPlatformPartChromeos has an
ImageCursorsSet. ImageCursorsSet may indirectly hold a reference to
CursorDataFactoryOzone, which is indirectly owned by Shell. To avoid
potential use after free the ImageCursorsSet needs to be destroyed
before the Shell.

This was caught in trying to bringup more browser_tests for mushrome.

BUG= 740655 
TEST=none

Change-Id: I903bb5016f378eb225c8f009630a3298ec1fb57d
Reviewed-on: https://chromium-review.googlesource.com/608902
Reviewed-by: Elliot Glaysher <erg@chromium.org>
Commit-Queue: Scott Violet <sky@chromium.org>
Cr-Commit-Position: refs/heads/master@{#493145}
[modify] https://crrev.com/25bed6f775cd0b22e31f462c9930c2c17dcd1452/chrome/browser/browser_process_platform_part_chromeos.cc
[modify] https://crrev.com/25bed6f775cd0b22e31f462c9930c2c17dcd1452/chrome/browser/browser_process_platform_part_chromeos.h
[modify] https://crrev.com/25bed6f775cd0b22e31f462c9930c2c17dcd1452/chrome/browser/ui/ash/ash_init.cc
[modify] https://crrev.com/25bed6f775cd0b22e31f462c9930c2c17dcd1452/services/ui/ws/threaded_image_cursors.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/986ac474e59cc1e47cd29ea09fb9eadf29010dd4

commit 986ac474e59cc1e47cd29ea09fb9eadf29010dd4
Author: Scott Violet <sky@chromium.org>
Date: Thu Aug 10 16:38:34 2017

chromeos: fix renderer crash in ChromeOS with UI service (aka mus)

The browser_test
AppViewTests/AppViewTest.TestAppViewMultipleConnects/0 pretty
regularly triggers RequestNewLayerTreeFrameSink() being called with a
routing_id that has no RendererWindowTreeClient. This makes the code
run the callback with null if that happens rather than crashing.

Non-chromeos-mus code has a similar check in
FrameSinkProviderImpl::CreateForWidget(), which verifies there is
a RenderWidgetHostImpl before continuing. Running this test on a
non-chromeos-mus build triggers RenderWidgetHostImpl::FromID() returning null.

BUG= 740655 
TEST=covered by tests

Change-Id: I647e0d3009353f12373470dec1018a331ca9dd41
Reviewed-on: https://chromium-review.googlesource.com/608834
Reviewed-by: Sadrul Chowdhury <sadrul@chromium.org>
Commit-Queue: Scott Violet <sky@chromium.org>
Cr-Commit-Position: refs/heads/master@{#493425}
[modify] https://crrev.com/986ac474e59cc1e47cd29ea09fb9eadf29010dd4/content/renderer/render_thread_impl.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/71a7993bdd05e2e84dbf0d5f8ed5f5232250ce80

commit 71a7993bdd05e2e84dbf0d5f8ed5f5232250ce80
Author: Scott Violet <sky@chromium.org>
Date: Mon Aug 21 20:17:25 2017

chromeos: enable more tests for mus_browser_tests

Now that mushrome (chromeos with ui service in process) runs
everything in process more tests are passing. I'm going to an
exclusion list as at this point the majority pass.

BUG= 740655 
TEST=test only changes

Cq-Include-Trybots: master.tryserver.chromium.android:android_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
Change-Id: I39c726c7d627a3214f5756580501d680a84eee63
Reviewed-on: https://chromium-review.googlesource.com/604498
Commit-Queue: Scott Violet <sky@chromium.org>
Reviewed-by: Jonathan Ross <jonross@chromium.org>
Cr-Commit-Position: refs/heads/master@{#496041}
[modify] https://crrev.com/71a7993bdd05e2e84dbf0d5f8ed5f5232250ce80/testing/buildbot/chromium.chromiumos.json
[modify] https://crrev.com/71a7993bdd05e2e84dbf0d5f8ed5f5232250ce80/testing/buildbot/chromium.fyi.json
[modify] https://crrev.com/71a7993bdd05e2e84dbf0d5f8ed5f5232250ce80/testing/buildbot/filters/mojo.fyi.mus.browser_tests.filter

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/3d625ee07a4ab4358b0312c731ad3882310f9c49

commit 3d625ee07a4ab4358b0312c731ad3882310f9c49
Author: Scott Violet <sky@chromium.org>
Date: Tue Aug 22 14:58:42 2017

chromeos: disable couple more tests for mus_browser_tests

These look to be continually failing. Will file bugs separately.

BUG= 740655 
TEST=test only changes
TBR=jonross@chromium.org

Change-Id: Ief4442d5ee2dbe7ae6a92c39d8d64c8fb00fc4c7
Reviewed-on: https://chromium-review.googlesource.com/625071
Reviewed-by: Scott Violet <sky@chromium.org>
Reviewed-by: Jonathan Ross <jonross@chromium.org>
Commit-Queue: Jonathan Ross <jonross@chromium.org>
Cr-Commit-Position: refs/heads/master@{#496305}
[modify] https://crrev.com/3d625ee07a4ab4358b0312c731ad3882310f9c49/testing/buildbot/filters/mojo.fyi.mus.browser_tests.filter

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/9dfc0188632bd280ef2aacbe653e0a2e351b7bc7

commit 9dfc0188632bd280ef2aacbe653e0a2e351b7bc7
Author: Scott Violet <sky@chromium.org>
Date: Fri Aug 25 23:05:34 2017

chromeos: make mus_browser_tests run more tests

I haven't seen any consistent new failures in a while, so moving the
fyi config to the main waterfall.

BUG= 740655 
TEST=test only changes

Change-Id: Ib4a6c96f01db8c613a71333d4d2127584e56e1d8
Reviewed-on: https://chromium-review.googlesource.com/636108
Commit-Queue: Jonathan Ross <jonross@chromium.org>
Reviewed-by: Jonathan Ross <jonross@chromium.org>
Cr-Commit-Position: refs/heads/master@{#497583}
[modify] https://crrev.com/9dfc0188632bd280ef2aacbe653e0a2e351b7bc7/testing/buildbot/filters/mojo.fyi.mus.browser_tests.filter
[modify] https://crrev.com/9dfc0188632bd280ef2aacbe653e0a2e351b7bc7/testing/buildbot/filters/mus.browser_tests.filter

The majority of mus_browser_tests are now being ran. Separate bugs have been filed for each of the respective failures.