ozone: Cancel all touches once a single touch is cancelled

ozone: Cancel all touches once a single touch is cancelled

Not all areas of Chrome handle single cancelled touches well, to improve
palm suppression support, we will cancel all touches if a single touch
was reported as cancelled by the touch device. No further events are
coming through until the touches are released.

BUG=chrome-os-partner:60780
TEST=None

Review-Url: https://codereview.chromium.org/2640163003
Cr-Commit-Position: refs/heads/master@{#445126}
Committed: https://chromium.googlesource.com/chromium/src/+/ab976a7548aa9fb6f1e5221804ac6e44a134329d

[denniskempin, 2017-01-18 22:00:00]

denniskempin@google.com changed reviewers:
+ spang@chromium.org

[spang, 2017-01-19 01:08:46]

This seems fine to me but I think we should fix MotionEventAura regardless of
what we do here.

Does this still work if you have two touchscreens? TouchEventConverterEvdev can
only cancel its own touches. Not sure if it matters.

[spang, 2017-01-19 01:25:51]

https://codereview.chromium.org/2640163003/diff/20001/ui/events/ozone/evdev/t...
File ui/events/ozone/evdev/touch_event_converter_evdev.cc (right):

https://codereview.chromium.org/2640163003/diff/20001/ui/events/ozone/evdev/t...
ui/events/ozone/evdev/touch_event_converter_evdev.cc:192: CancelAllTouches();
Can this move after the loop so it runs at most once?

After this change Initialize() will never set |cancelled| to false. Is that
intended?

https://codereview.chromium.org/2640163003/diff/20001/ui/events/ozone/evdev/t...
ui/events/ozone/evdev/touch_event_converter_evdev.cc:377: CancelAllTouches();
What if we get a new contact with a lower numbered slot before the syn_report ?
Such a press will be reordered before all of the cancels and will still be alive
despite the attempt to cancel everything.

We really shouldn't to do anything here other than update events_[current_slot_]
to match our view the kernel internal state.

[denniskempin, 2017-01-20 02:04:39]

PTAL again Michael. Thanks!!

https://codereview.chromium.org/2640163003/diff/20001/ui/events/ozone/evdev/t...
File ui/events/ozone/evdev/touch_event_converter_evdev.cc (right):

https://codereview.chromium.org/2640163003/diff/20001/ui/events/ozone/evdev/t...
ui/events/ozone/evdev/touch_event_converter_evdev.cc:192: CancelAllTouches();
On 2017/01/19 01:25:51, spang wrote:
> Can this move after the loop so it runs at most once?
> 
> After this change Initialize() will never set |cancelled| to false. Is that
> intended?

Done.

https://codereview.chromium.org/2640163003/diff/20001/ui/events/ozone/evdev/t...
ui/events/ozone/evdev/touch_event_converter_evdev.cc:377: CancelAllTouches();
On 2017/01/19 01:25:51, spang wrote:
> What if we get a new contact with a lower numbered slot before the syn_report
?
> Such a press will be reordered before all of the cancels and will still be
alive
> despite the attempt to cancel everything.
> 
> We really shouldn't to do anything here other than update
events_[current_slot_]
> to match our view the kernel internal state.

Done.

[spang, 2017-01-20 18:32:59]

lgtm

Can you put a TODO in to undo this once the problems in the layers above are
worked out? We should leave the bug open or file a new one.

[denniskempin, 2017-01-20 18:59:13]

Done. Thanks Michael!

[denniskempin, 2017-01-20 18:59:18]

The CQ bit was checked by denniskempin@google.com

[denniskempin, 2017-01-20 18:59:18]

The patchset sent to the CQ was uploaded after l-g-t-m from spang@chromium.org
Link to the patchset: https://codereview.chromium.org/2640163003/#ps80001
(title: "added TODO note")

[commit-bot: I haz the power, 2017-01-20 18:59:40]

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, 2017-01-20 19:25:46]

CQ is committing da patch.

Bot data: {"patchset_id": 80001, "attempt_start_ts": 1484938758575160,
"parent_rev": "8d8e31c003abeaaad83a74f9212e67fcd6897cbd", "commit_rev":
"ab976a7548aa9fb6f1e5221804ac6e44a134329d"}

[commit-bot: I haz the power, 2017-01-20 19:26:43]

Description was changed from

==========
ozone: Cancel all touches once a single touch is cancelled

Not all areas of Chrome handle single cancelled touches well, to improve
palm suppression support, we will cancel all touches if a single touch
was reported as cancelled by the touch device. No further events are
coming through until the touches are released.

BUG=chrome-os-partner:60780
TEST=None
==========

to

==========
ozone: Cancel all touches once a single touch is cancelled

Not all areas of Chrome handle single cancelled touches well, to improve
palm suppression support, we will cancel all touches if a single touch
was reported as cancelled by the touch device. No further events are
coming through until the touches are released.

BUG=chrome-os-partner:60780
TEST=None

Review-Url: https://codereview.chromium.org/2640163003
Cr-Commit-Position: refs/heads/master@{#445126}
Committed:
https://chromium.googlesource.com/chromium/src/+/ab976a7548aa9fb6f1e5221804ac...
==========

[commit-bot: I haz the power, 2017-01-20 19:26:44]

Committed patchset #5 (id:80001) as
https://chromium.googlesource.com/chromium/src/+/ab976a7548aa9fb6f1e5221804ac...