Use the task scheduler for blocking tasks in the USB service on Linux

Use the task scheduler for blocking tasks in the USB service on Linux

This patch switches UsbServiceLinux to being initialized with a
SequencedTaskRunner that puts it on its own non-shutdown-blocking
sequence. Individual device connections are also put on their own
sequences because there is no dependency between them and the global
service.

As DeviceMonitorLinux is global and owned by the FILE thread a new
UdevWatcher class has been added which can be owned by each service
that needs to monitor udev and lives on its own sequence.

This should resolve issue 694496 by allowing Chrome to exit even if a
faulty driver or device has locked up the USB service.

BUG=694496
Review-Url: https://codereview.chromium.org/2849953003
Cr-Commit-Position: refs/heads/master@{#468456}
Committed: https://chromium.googlesource.com/chromium/src/+/221e92f4f397ace9c60dd7361c306e7f7a1c4851

[Reilly Grant (use Gerrit), 2017-04-29 00:32:04]

Description was changed from

==========
Use the task scheduler for blocking tasks in the USB service on Linux

This patch switches UsbServiceLinux to being initialized with a
SequencedTaskRunner that puts it on its own non-shutdown-blocking
sequence. Individual device connections are also put on their own
sequences because there is no dependency between them and the global
service.

This should resolve issue 694496 by allowing Chrome to exit even if a
faulty driver or device has locked up the USB service.

BUG=694496
==========

to

==========
Use the task scheduler for blocking tasks in the USB service on Linux

This patch switches UsbServiceLinux to being initialized with a
SequencedTaskRunner that puts it on its own non-shutdown-blocking
sequence. Individual device connections are also put on their own
sequences because there is no dependency between them and the global
service.

As DeviceMonitorLinux is global and owned by the FILE thread a new
UdevWatcher class has been added which can be owned by each service
that needs to monitor udev and lives on its own sequence.

This should resolve issue 694496 by allowing Chrome to exit even if a
faulty driver or device has locked up the USB service.

BUG=694496
==========

[Reilly Grant (use Gerrit), 2017-04-29 00:32:35]

reillyg@chromium.org changed reviewers:
+ robliao@chromium.org

[Reilly Grant (use Gerrit), 2017-04-29 00:32:35]

Please take a look.

[Reilly Grant (use Gerrit), 2017-04-29 00:33:36]

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

[commit-bot: I haz the power, 2017-04-29 00:34:06]

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, 2017-04-29 01:41:04]

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

[commit-bot: I haz the power, 2017-04-29 01:41:05]

Dry run: This issue passed the CQ dry run.

[robliao, 2017-05-01 16:36:19]

Task API Usage lgtm.

https://codereview.chromium.org/2849953003/diff/1/device/usb/usb_device_linux.cc
File device/usb/usb_device_linux.cc (right):

https://codereview.chromium.org/2849953003/diff/1/device/usb/usb_device_linux...
device/usb/usb_device_linux.cc:33: scoped_refptr<base::SequencedTaskRunner>
CreateBlockingTaskRunner() {
It might be useful to have this in a shared header file to ensure the same
blocking task runner policy on all of the platforms.

[Reilly Grant (use Gerrit), 2017-05-01 20:39:37]

Add a shared CreateBlockingTaskRunner() method

[Reilly Grant (use Gerrit), 2017-05-01 20:40:16]

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

[commit-bot: I haz the power, 2017-05-01 20:41:11]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Reilly Grant (use Gerrit), 2017-05-01 20:46:59]

Remove more passing of the SequencedTaskRunner

[Reilly Grant (use Gerrit), 2017-05-01 20:57:43]

https://codereview.chromium.org/2849953003/diff/1/device/usb/usb_device_linux.cc
File device/usb/usb_device_linux.cc (right):

https://codereview.chromium.org/2849953003/diff/1/device/usb/usb_device_linux...
device/usb/usb_device_linux.cc:33: scoped_refptr<base::SequencedTaskRunner>
CreateBlockingTaskRunner() {
On 2017/05/01 16:36:19, robliao wrote:
> It might be useful to have this in a shared header file to ensure the same
> blocking task runner policy on all of the platforms.

Done.

[Reilly Grant (use Gerrit), 2017-05-01 20:57:52]

The CQ bit was checked by reillyg@chromium.org

[Reilly Grant (use Gerrit), 2017-05-01 20:57:52]

The patchset sent to the CQ was uploaded after l-g-t-m from robliao@chromium.org
Link to the patchset: https://codereview.chromium.org/2849953003/#ps40001
(title: "Remove more passing of the SequencedTaskRunner")

[commit-bot: I haz the power, 2017-05-01 20:58:39]

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, 2017-05-01 22:26:12]

CQ is committing da patch.

Bot data: {"patchset_id": 40001, "attempt_start_ts": 1493672272343130,
"parent_rev": "e8e3f43d16a81d9b13795f885d011e29bd8d4164", "commit_rev":
"221e92f4f397ace9c60dd7361c306e7f7a1c4851"}

[commit-bot: I haz the power, 2017-05-01 22:26:21]

Description was changed from

==========
Use the task scheduler for blocking tasks in the USB service on Linux

This patch switches UsbServiceLinux to being initialized with a
SequencedTaskRunner that puts it on its own non-shutdown-blocking
sequence. Individual device connections are also put on their own
sequences because there is no dependency between them and the global
service.

As DeviceMonitorLinux is global and owned by the FILE thread a new
UdevWatcher class has been added which can be owned by each service
that needs to monitor udev and lives on its own sequence.

This should resolve issue 694496 by allowing Chrome to exit even if a
faulty driver or device has locked up the USB service.

BUG=694496
==========

to

==========
Use the task scheduler for blocking tasks in the USB service on Linux

This patch switches UsbServiceLinux to being initialized with a
SequencedTaskRunner that puts it on its own non-shutdown-blocking
sequence. Individual device connections are also put on their own
sequences because there is no dependency between them and the global
service.

As DeviceMonitorLinux is global and owned by the FILE thread a new
UdevWatcher class has been added which can be owned by each service
that needs to monitor udev and lives on its own sequence.

This should resolve issue 694496 by allowing Chrome to exit even if a
faulty driver or device has locked up the USB service.

BUG=694496

Review-Url: https://codereview.chromium.org/2849953003
Cr-Commit-Position: refs/heads/master@{#468456}
Committed:
https://chromium.googlesource.com/chromium/src/+/221e92f4f397ace9c60dd7361c30...
==========

[commit-bot: I haz the power, 2017-05-01 22:26:22]

Committed patchset #3 (id:40001) as
https://chromium.googlesource.com/chromium/src/+/221e92f4f397ace9c60dd7361c30...