Revert of Version of MessagePumpForUI optimized for GPU process

Revert of Version of MessagePumpForUI optimized for GPU process (patchset #8 id:220001 of https://codereview.chromium.org/1714263002/ )

Reason for revert:
It looks like this is causing issues with UI responsiveness due to shared message queue.

Original issue's description:
> Version of MessagePumpForUI optimized for GPU process.
>
> The purpose of this change is to reduce overhead of
> MessagePump in GPU process. According to my testing on
> Windows 10 Dell XPS laptop this change improved the
> power usage by 0.15 Wt in one test and 0.12 Wt in another.
> The test was a windowed playback of 1080p @ 60 Hz H.264
> video.
>
> Based on ETW sample profile the overhead of the message
> pump is reduced from 13-15% to 7-10%. The overhead is
> calculated as a number of inclusive samples in DoRunLoop
> that are not part of DoWork and DoDelayedWork.
>
> The change eliminates some system calls that shouldn't be needed
> in the context of GPU process such as GetQueueStatus,
> SetTimer, KillTimer, and minimizes other system calls that
> came up high in the sampling profile - PeekMessage,
> MsgWaitForMultipleObjectsEx, PostMessage.
> While this version of MessagePump is still capable to peek
> and dispatch Windows messages which apparently is needed
> in the GPU process on Windows version above Win7,
> based on the usage pattern analysis, the emphasis gets
> shifted to handling the work (vs. handling Windows messages in MessagePumpForUI).
>
> See the bug for more details.
>
> If this work well in a limited context of the GPU process
> some of these ideas might later be applied to other
> types of MessagePump.
>
> BUG=588798
>
> Committed: https://crrev.com/f71e6d2cbb90cacc88ce41a48475c47c6b7a3543
> Cr-Commit-Position: refs/heads/master@{#386898}

TBR=jbauman@chromium.org,thestig@chromium.org,ananta@chromium.org,iyengar@google.com
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=588798
Committed: https://crrev.com/06711f5f9fad01f4a9dc89457bfc583b45646a8f
Cr-Commit-Position: refs/heads/master@{#387790}

[stanisc, 2016-04-16 00:23:28]

Created Revert of Version of MessagePumpForUI optimized for GPU process

[stanisc, 2016-04-16 00:25:13]

The CQ bit was checked by stanisc@chromium.org

[commit-bot: I haz the power, 2016-04-16 00:26:03]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1889293003/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1889293003/1

[commit-bot: I haz the power, 2016-04-16 02:20:46]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-04-16 02:23:05]

Description was changed from

==========
Revert of Version of MessagePumpForUI optimized for GPU process (patchset #8
id:220001 of https://codereview.chromium.org/1714263002/ )

Reason for revert:
It looks like this is causing issues with UI responsiveness due to shared
message queue.

Original issue's description:
> Version of MessagePumpForUI optimized for GPU process.
> 
> The purpose of this change is to reduce overhead of
> MessagePump in GPU process. According to my testing on
> Windows 10 Dell XPS laptop this change improved the
> power usage by 0.15 Wt in one test and 0.12 Wt in another.
> The test was a windowed playback of 1080p @ 60 Hz H.264
> video.
> 
> Based on ETW sample profile the overhead of the message
> pump is reduced from 13-15% to 7-10%. The overhead is
> calculated as a number of inclusive samples in DoRunLoop
> that are not part of DoWork and DoDelayedWork.
> 
> The change eliminates some system calls that shouldn't be needed
> in the context of GPU process such as GetQueueStatus,
> SetTimer, KillTimer, and minimizes other system calls that
> came up high in the sampling profile - PeekMessage,
> MsgWaitForMultipleObjectsEx, PostMessage.
> While this version of MessagePump is still capable to peek
> and dispatch Windows messages which apparently is needed
> in the GPU process on Windows version above Win7,
> based on the usage pattern analysis, the emphasis gets
> shifted to handling the work (vs. handling Windows messages in
MessagePumpForUI).
> 
> See the bug for more details.
> 
> If this work well in a limited context of the GPU process
> some of these ideas might later be applied to other
> types of MessagePump.
> 
> BUG=588798
> 
> Committed: https://crrev.com/f71e6d2cbb90cacc88ce41a48475c47c6b7a3543
> Cr-Commit-Position: refs/heads/master@{#386898}

TBR=jbauman@chromium.org,thestig@chromium.org,ananta@chromium.org,iyengar@goo...
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=588798
==========

to

==========
Revert of Version of MessagePumpForUI optimized for GPU process (patchset #8
id:220001 of https://codereview.chromium.org/1714263002/ )

Reason for revert:
It looks like this is causing issues with UI responsiveness due to shared
message queue.

Original issue's description:
> Version of MessagePumpForUI optimized for GPU process.
>
> The purpose of this change is to reduce overhead of
> MessagePump in GPU process. According to my testing on
> Windows 10 Dell XPS laptop this change improved the
> power usage by 0.15 Wt in one test and 0.12 Wt in another.
> The test was a windowed playback of 1080p @ 60 Hz H.264
> video.
>
> Based on ETW sample profile the overhead of the message
> pump is reduced from 13-15% to 7-10%. The overhead is
> calculated as a number of inclusive samples in DoRunLoop
> that are not part of DoWork and DoDelayedWork.
>
> The change eliminates some system calls that shouldn't be needed
> in the context of GPU process such as GetQueueStatus,
> SetTimer, KillTimer, and minimizes other system calls that
> came up high in the sampling profile - PeekMessage,
> MsgWaitForMultipleObjectsEx, PostMessage.
> While this version of MessagePump is still capable to peek
> and dispatch Windows messages which apparently is needed
> in the GPU process on Windows version above Win7,
> based on the usage pattern analysis, the emphasis gets
> shifted to handling the work (vs. handling Windows messages in
MessagePumpForUI).
>
> See the bug for more details.
>
> If this work well in a limited context of the GPU process
> some of these ideas might later be applied to other
> types of MessagePump.
>
> BUG=588798
>
> Committed: https://crrev.com/f71e6d2cbb90cacc88ce41a48475c47c6b7a3543
> Cr-Commit-Position: refs/heads/master@{#386898}

TBR=jbauman@chromium.org,thestig@chromium.org,ananta@chromium.org,iyengar@goo...
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=588798

Committed: https://crrev.com/06711f5f9fad01f4a9dc89457bfc583b45646a8f
Cr-Commit-Position: refs/heads/master@{#387790}
==========

[commit-bot: I haz the power, 2016-04-16 02:23:06]

Patchset 1 (id:??) landed as
https://crrev.com/06711f5f9fad01f4a9dc89457bfc583b45646a8f
Cr-Commit-Position: refs/heads/master@{#387790}