Decouple some graphics-related IsLowEndDevice() policies for 1GB devices.

Decouple some graphics-related IsLowEndDevice() policies for 1GB devices.

In http://crrev.com/2814323002, IsLowEndDevice() was changed to apply to
some 1GB devices, which was not the original intent of the flag (used to
be strictly 512MB).  I triaged the current uses of it, and found the two
following inappropriate ones:

1) Use of 16-bit color in the graphics subsystem is a particularly
desperate measure unique to Chrome on 512MB devices (not even resorted
to by any other Android app on 512MB devices).  Because 16-bit GL driver
features are little-used or tested, it poses a high risk of severe
performance and stability issues (which are browser crashes in
IsLowEndDevice mode).  So replace those callsites with an explicit
memory check.

2) The video code has some settings to disable autoplay and idle
playback on Jellybean or low-RAM devices.  Disabling autoplay is a
web-standards-violating change so we shouldn't apply it to any new
devices.  As far as I can tell from the history on
http://crbug.com/612909, this was added to workaround Jellybean bugs
which have never been observed on any device meeting the Svelte criteria
(KitKat or above, 512MB).  isLowEndDevice() should only be used for RAM
optimizations, so I simply removed the calls.

BUG=711079
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

Review-Url: https://codereview.chromium.org/2820643002
Cr-Commit-Position: refs/heads/master@{#464813}
Committed: https://chromium.googlesource.com/chromium/src/+/e678aa49f3fbb21132296a1356239c79d8db276a

[aelias_OOO_until_Jul13, 2017-04-14 02:49:12]

Description was changed from

==========
Decouple some graphics-related IsLowEndDevice() policies for 1GB devices.

In http://crrev.com/2814323002, IsLowEndDevice() was changed to apply to
some 1GB devices, which was not the original intent of the flag (used to
be strictly 512MB).  I triaged the current uses of it, and found the two
following inappropriate ones:

1) Use of 16-bit color in the graphics subsystem is a particularly
desperate measure unique to Chrome on 512MB devices (not even resorted
to by any other Android app on 512MB devices).  Because 16-bit GL driver
features are little-used or tested, it poses a high risk of severe
performance and stability issues (which are browser crashes in
IsLowEndDevice mode).  So replace those callsites with an explicit
memory check.

2) The video code has some settings to disable autoplay and idle
playback on Jellybean or low-RAM devices.  Disabling autoplay is a
web-standards-violating change so we shouldn't apply it to any new
devices.  As far as I can tell from the history on
http://crbug.com/612909, this was added to workaround Jellybean bugs
which have never been observed on any device meeting the Svelte criteria
(KitKat or above, 512MB).  isLowEndDevice() should only be used for RAM
optimizations, so I simply removed the calls.

BUG=711079
==========

to

==========
Decouple some graphics-related IsLowEndDevice() policies for 1GB devices.

In http://crrev.com/2814323002, IsLowEndDevice() was changed to apply to
some 1GB devices, which was not the original intent of the flag (used to
be strictly 512MB).  I triaged the current uses of it, and found the two
following inappropriate ones:

1) Use of 16-bit color in the graphics subsystem is a particularly
desperate measure unique to Chrome on 512MB devices (not even resorted
to by any other Android app on 512MB devices).  Because 16-bit GL driver
features are little-used or tested, it poses a high risk of severe
performance and stability issues (which are browser crashes in
IsLowEndDevice mode).  So replace those callsites with an explicit
memory check.

2) The video code has some settings to disable autoplay and idle
playback on Jellybean or low-RAM devices.  Disabling autoplay is a
web-standards-violating change so we shouldn't apply it to any new
devices.  As far as I can tell from the history on
http://crbug.com/612909, this was added to workaround Jellybean bugs
which have never been observed on any device meeting the Svelte criteria
(KitKat or above, 512MB).  isLowEndDevice() should only be used for RAM
optimizations, so I simply removed the calls.

BUG=711079
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
==========

[aelias_OOO_until_Jul13, 2017-04-14 02:52:03]

aelias@chromium.org changed reviewers:
+ dalecurtis@chromium.org, jbauman@chromium.org

[aelias_OOO_until_Jul13, 2017-04-14 02:52:04]

PTAL,

dalecurtis@ for video
jbauman@ for ui/gl

[aelias_OOO_until_Jul13, 2017-04-14 02:52:51]

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

[commit-bot: I haz the power, 2017-04-14 02:53:14]

Dry run: 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, 2017-04-14 02:56:02]

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

[commit-bot: I haz the power, 2017-04-14 02:56:03]

Dry run: Try jobs failed on following builders:
  mac_optional_gpu_tests_rel on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_optional_gpu_...)

[aelias_OOO_until_Jul13, 2017-04-14 03:14:04]

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

[commit-bot: I haz the power, 2017-04-14 03:14:20]

Dry run: 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, 2017-04-14 05:01:01]

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

[commit-bot: I haz the power, 2017-04-14 05:01:02]

Dry run: This issue passed the CQ dry run.

[DaleCurtis, 2017-04-14 18:24:25]

media/ lgtm, mlamouri@ never liked that change anyways and I only have a sample
size of 1 to back it up.

[jbauman, 2017-04-14 22:06:57]

lgtm

[aelias_OOO_until_Jul13, 2017-04-14 22:35:34]

The CQ bit was checked by aelias@chromium.org

[commit-bot: I haz the power, 2017-04-14 22:36:04]

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, 2017-04-14 22:41:11]

CQ is committing da patch.

Bot data: {"patchset_id": 20001, "attempt_start_ts": 1492209334546680,
"parent_rev": "45deedd3bd5ff89fa359ff892ad596a9aec3be8c", "commit_rev":
"e678aa49f3fbb21132296a1356239c79d8db276a"}

[commit-bot: I haz the power, 2017-04-14 22:41:57]

Description was changed from

==========
Decouple some graphics-related IsLowEndDevice() policies for 1GB devices.

In http://crrev.com/2814323002, IsLowEndDevice() was changed to apply to
some 1GB devices, which was not the original intent of the flag (used to
be strictly 512MB).  I triaged the current uses of it, and found the two
following inappropriate ones:

1) Use of 16-bit color in the graphics subsystem is a particularly
desperate measure unique to Chrome on 512MB devices (not even resorted
to by any other Android app on 512MB devices).  Because 16-bit GL driver
features are little-used or tested, it poses a high risk of severe
performance and stability issues (which are browser crashes in
IsLowEndDevice mode).  So replace those callsites with an explicit
memory check.

2) The video code has some settings to disable autoplay and idle
playback on Jellybean or low-RAM devices.  Disabling autoplay is a
web-standards-violating change so we shouldn't apply it to any new
devices.  As far as I can tell from the history on
http://crbug.com/612909, this was added to workaround Jellybean bugs
which have never been observed on any device meeting the Svelte criteria
(KitKat or above, 512MB).  isLowEndDevice() should only be used for RAM
optimizations, so I simply removed the calls.

BUG=711079
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
==========

to

==========
Decouple some graphics-related IsLowEndDevice() policies for 1GB devices.

In http://crrev.com/2814323002, IsLowEndDevice() was changed to apply to
some 1GB devices, which was not the original intent of the flag (used to
be strictly 512MB).  I triaged the current uses of it, and found the two
following inappropriate ones:

1) Use of 16-bit color in the graphics subsystem is a particularly
desperate measure unique to Chrome on 512MB devices (not even resorted
to by any other Android app on 512MB devices).  Because 16-bit GL driver
features are little-used or tested, it poses a high risk of severe
performance and stability issues (which are browser crashes in
IsLowEndDevice mode).  So replace those callsites with an explicit
memory check.

2) The video code has some settings to disable autoplay and idle
playback on Jellybean or low-RAM devices.  Disabling autoplay is a
web-standards-violating change so we shouldn't apply it to any new
devices.  As far as I can tell from the history on
http://crbug.com/612909, this was added to workaround Jellybean bugs
which have never been observed on any device meeting the Svelte criteria
(KitKat or above, 512MB).  isLowEndDevice() should only be used for RAM
optimizations, so I simply removed the calls.

BUG=711079
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

Review-Url: https://codereview.chromium.org/2820643002
Cr-Commit-Position: refs/heads/master@{#464813}
Committed:
https://chromium.googlesource.com/chromium/src/+/e678aa49f3fbb21132296a135623...
==========

[commit-bot: I haz the power, 2017-04-14 22:41:58]

Committed patchset #2 (id:20001) as
https://chromium.googlesource.com/chromium/src/+/e678aa49f3fbb21132296a135623...