Add a new pair of IPC categories for messages that need handling as input events

Add a new pair of IPC categories for messages that need handling as input events

The compositor thread input filtering mechanism adds the possibility of
handling some IPCs in the render process main thread in a different order than
they were sent from the browser process since it filters some messages from the
IO thread.  For the most part, this is OK since the exact order of different
types of messages is not important but there are some input-related messages
that do need to processed in order - for instance MouseCaptureLost
(see crbug.com/231532).

This adds a new IPC class, InputMsg, for messages that need input-like
filtering and InputHostMsg for their ACKs. All messages of this class are
routed through the compositor thread to preserve order.

BUG=233365
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=196907

[jamesr, 2013-04-24 23:38:13]

This adds a new IPC class to do the compositor thread filtering.  Hopefully this
method makes it a bit easier to maintain this set.

It's hard to know whether to put a new IPC in this class or in ViewMsg_.  The
tradeoff is that putting it in ViewInputMsg_ will guarantee that the message is
received on the render process main thread in the order it was sent relative to
other ViewInputMsgs, but it will be copied extra times.  IME events, for
instance, use a separate ACKing mechanism to avoid going through this trip.

[jam, 2013-04-25 15:31:16]

question: why ViewInputMsg instead of InputMsg?

https://codereview.chromium.org/14487003/diff/8001/content/common/view_input_...
File content/common/view_input_messages.h (right):

https://codereview.chromium.org/14487003/diff/8001/content/common/view_input_...
content/common/view_input_messages.h:130: // independent ViewHostMsg, then
ifdefs for platform specific ViewHostMsg.
nit: update this comment

[jamesr, 2013-04-26 00:23:28]

On 2013/04/25 15:31:16, jam wrote:
> question: why ViewInputMsg instead of InputMsg?

I didn't think of that.  InputMsg is much better, done.

>
https://codereview.chromium.org/14487003/diff/8001/content/common/view_input_...
> content/common/view_input_messages.h:130: // independent ViewHostMsg, then
> ifdefs for platform specific ViewHostMsg.
> nit: update this comment

Done

[jam, 2013-04-26 00:36:09]

lgtm

https://codereview.chromium.org/14487003/diff/17001/content/renderer/render_v...
File content/renderer/render_view_impl.cc (right):

https://codereview.chromium.org/14487003/diff/17001/content/renderer/render_v...
content/renderer/render_view_impl.cc:1007: IPC_MESSAGE_HANDLER(InputMsg_Undo,
OnUndo)
nit: please sort all the InputMsg at the top. also make sure their message
handlers are sorted together in the cc file and header

https://codereview.chromium.org/14487003/diff/17001/content/renderer/render_w...
File content/renderer/render_widget.cc (right):

https://codereview.chromium.org/14487003/diff/17001/content/renderer/render_w...
content/renderer/render_widget.cc:308:
IPC_MESSAGE_HANDLER(InputMsg_HandleInputEvent, OnHandleInputEvent)
ditto

https://codereview.chromium.org/14487003/diff/17001/ipc/ipc_message_start.h
File ipc/ipc_message_start.h (right):

https://codereview.chromium.org/14487003/diff/17001/ipc/ipc_message_start.h#n...
ipc/ipc_message_start.h:14: ViewInputMsgStart,
nit: InputMsgStart

[commit-bot: I haz the power, 2013-04-26 01:21:09]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/jamesr@chromium.org/14487003/28001

[commit-bot: I haz the power, 2013-04-26 01:21:20]

Presubmit check for 14487003-28001 failed and returned exit status 1.

INFO:root:Found 28 file(s).

Running presubmit commit checks ...
Running /b/commit-queue/workdir/chromium/PRESUBMIT.py

** Presubmit ERRORS **
Missing LGTM from an OWNER for these files:
    content/common/input_messages.h
    content/common/view_messages.h
    ipc/ipc_message_start.h

Presubmit checks took 2.6s to calculate.

[jamesr, 2013-04-26 02:23:36]

Cris - could you security review this change?  I'm moving a bunch of IPCs to a
different header / class, but not changing them.

[darin (slow to review), 2013-04-26 04:59:13]

LGTM

[palmer, 2013-04-26 21:50:41]

IPC security LGTM

[commit-bot: I haz the power, 2013-04-26 21:51:04]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/jamesr@chromium.org/14487003/28001

[commit-bot: I haz the power, 2013-04-27 01:08:17]

Change committed as 196907