Use USB hotplug events from libusb when possible.

Use USB hotplug events from libusb when possible.

When possible subscribe to the USB device add and remove events exposed
by libusb instead of refreshing the device list every time a client
requests a list of devices. In practice this means that hotplug events
are used on Linux (when udev is available) and OS X.

This is the first step to providing USB hotplug events on all platforms.

BUG=411715
Committed: https://crrev.com/0ae3a999fb424785077a9a34817d96bc79417ddf
Cr-Commit-Position: refs/heads/master@{#309946}

[Reilly Grant (use Gerrit), 2014-12-18 19:00:18]

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

[Reilly Grant (use Gerrit), 2014-12-18 19:00:19]

Renaud, please take a look at this. It is not urgent as this is only the first
step in a series of patches that should probably go in together.

[rpaquay, 2014-12-18 21:08:54]

https://codereview.chromium.org/803653003/diff/1/device/usb/usb_service_impl.cc
File device/usb/usb_service_impl.cc (right):

https://codereview.chromium.org/803653003/diff/1/device/usb/usb_service_impl....
device/usb/usb_service_impl.cc:225: base::Unretained(self), device));
nit: Comment why using Unretained is ok?
Also, how do we know "device" is ok until the callback is invoked?

https://codereview.chromium.org/803653003/diff/1/device/usb/usb_service_impl....
device/usb/usb_service_impl.cc:234: base::Unretained(self), device));
nit: Comment why using Unretained is ok?
Also, how do we know "device" is ok until the callback is invoked?

[Reilly Grant (use Gerrit), 2014-12-18 22:28:13]

https://codereview.chromium.org/803653003/diff/1/device/usb/usb_service_impl.cc
File device/usb/usb_service_impl.cc (right):

https://codereview.chromium.org/803653003/diff/1/device/usb/usb_service_impl....
device/usb/usb_service_impl.cc:225: base::Unretained(self), device));
On 2014/12/18 21:08:54, rpaquay wrote:
> nit: Comment why using Unretained is ok?
> Also, how do we know "device" is ok until the callback is invoked?

libusb is nice and guarantees this is safe for us. I've added a comment to this
function explaining.

[rpaquay, 2014-12-18 22:52:47]

lgtm

https://codereview.chromium.org/803653003/diff/20001/device/usb/usb_service_i...
File device/usb/usb_service_impl.cc (right):

https://codereview.chromium.org/803653003/diff/20001/device/usb/usb_service_i...
device/usb/usb_service_impl.cc:219: // and so guarantees that this funciton will
not be called by the event
funciton -> function

[Reilly Grant (use Gerrit), 2015-01-05 19:06:33]

The CQ bit was checked by reillyg@chromium.org

[commit-bot: I haz the power, 2015-01-05 19:07:43]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/803653003/80001

[commit-bot: I haz the power, 2015-01-05 20:13:11]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2015-01-05 20:14:02]

Patchset 5 (id:??) landed as
https://crrev.com/0ae3a999fb424785077a9a34817d96bc79417ddf
Cr-Commit-Position: refs/heads/master@{#309946}