Parse USB interface association descriptors.

Parse USB interface association descriptors.

USB interface association descriptors are used to combine multiple
interfaces into a single functional group. This patch adds support for
parsing them out of the |extra_data| field left by libusb's parsing of
configuration, interface and endpoint descriptors. The resulting
association is then represented by setting the |first_interface| field
of each interface in a function to the |interface_number| of the first
interface in the function.

WebUSB will use these associations to set permissions for an entire
function with a single descriptor.

BUG=492204
Committed: https://crrev.com/b61e467059e55ad71f1b7ae5e1412beb3d4803e0
Cr-Commit-Position: refs/heads/master@{#375324}

[Reilly Grant (use Gerrit), 2016-01-06 22:15:35]

Patchset #1 (id:1) has been deleted

[Reilly Grant (use Gerrit), 2016-01-06 23:17:55]

Patchset #1 (id:20001) has been deleted

[Reilly Grant (use Gerrit), 2016-01-06 23:18:46]

Description was changed from

==========
Parse USB interface association descriptors.

USB interface association descriptors are used to combine multiple
interfaces into a single functional group. This patch adds support for
parsing them out of the |extra_data| field left by libusb's parsing of
configuration, interface and endpoint descriptors. The resulting
association is then represented by setting the |first_interface| field
of each interface in a function to the |interface_number| of the first
interface in the function.

WebUSB uses these associations to set permissions for an entire function
with one descriptor.

BUG=492204
==========

to

==========
Parse USB interface association descriptors.

USB interface association descriptors are used to combine multiple
interfaces into a single functional group. This patch adds support for
parsing them out of the |extra_data| field left by libusb's parsing of
configuration, interface and endpoint descriptors. The resulting
association is then represented by setting the |first_interface| field
of each interface in a function to the |interface_number| of the first
interface in the function.

WebUSB will use these associations to set permissions for an entire
function with a single descriptor.

BUG=492204
==========

[Reilly Grant (use Gerrit), 2016-01-06 23:18:46]

reillyg@chromium.org changed reviewers:
+ pfeldman@chromium.org, rockot@chromium.org

[Reilly Grant (use Gerrit), 2016-01-06 23:20:04]

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

[Reilly Grant (use Gerrit), 2016-01-06 23:20:04]

rockot@, please look at //device
pfeldman@, please look at //chrome/browser/devtools
stevenjb@, please look at //chrome/browser/chromeos

[pfeldman, 2016-01-06 23:28:00]

devtools lgtm

[stevenjb, 2016-01-07 18:04:29]

c/b/chromeos lgtm

[Reilly Grant (use Gerrit), 2016-02-12 18:11:26]

reillyg@chromium.org changed reviewers:
- pfeldman@chromium.org, stevenjb@chromium.org

[Reilly Grant (use Gerrit), 2016-02-12 18:27:25]

Patchset #2 (id:60001) has been deleted

[Reilly Grant (use Gerrit), 2016-02-12 18:35:16]

rockot@, ping. Now that I've landed https://codereview.chromium.org/1468423003/
I didn't actually need Pavel and Steven's reviews.

[Ken Rockot(use gerrit already), 2016-02-12 19:25:01]

rockot@chromium.org changed reviewers:
+ pfeldman@chromium.org, stevenjb@chromium.org

[Ken Rockot(use gerrit already), 2016-02-12 19:25:01]

lgtm

https://codereview.chromium.org/1568673002/diff/80001/device/usb/usb_descript...
File device/usb/usb_descriptors.cc (right):

https://codereview.chromium.org/1568673002/diff/80001/device/usb/usb_descript...
device/usb/usb_descriptors.cc:34: UsbInterfaceAssociationDescriptor() = delete;
I'm fairly sure that this is implied by defining a non-zero-argument constructor

https://codereview.chromium.org/1568673002/diff/80001/device/usb/usb_descript...
device/usb/usb_descriptors.cc:35: ~UsbInterfaceAssociationDescriptor() =
default;
Isn't this also implied given that the struct is only POD?

[Reilly Grant (use Gerrit), 2016-02-12 21:06:21]

https://codereview.chromium.org/1568673002/diff/80001/device/usb/usb_descript...
File device/usb/usb_descriptors.cc (right):

https://codereview.chromium.org/1568673002/diff/80001/device/usb/usb_descript...
device/usb/usb_descriptors.cc:34: UsbInterfaceAssociationDescriptor() = delete;
On 2016/02/12 19:25:01, Ken Rockot wrote:
> I'm fairly sure that this is implied by defining a non-zero-argument
constructor

You're right.

https://codereview.chromium.org/1568673002/diff/80001/device/usb/usb_descript...
device/usb/usb_descriptors.cc:35: ~UsbInterfaceAssociationDescriptor() =
default;
On 2016/02/12 19:25:01, Ken Rockot wrote:
> Isn't this also implied given that the struct is only POD?

Indeed. I just got overly excited about default and delete.

[Reilly Grant (use Gerrit), 2016-02-12 21:06:26]

The CQ bit was checked by reillyg@chromium.org

[Reilly Grant (use Gerrit), 2016-02-12 21:06:26]

The patchset sent to the CQ was uploaded after l-g-t-m from
stevenjb@chromium.org, pfeldman@chromium.org, rockot@chromium.org
Link to the patchset: https://codereview.chromium.org/1568673002/#ps100001
(title: "Remove unnecessary default/delete.")

[commit-bot: I haz the power, 2016-02-12 21:06:44]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1568673002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1568673002/100001

[commit-bot: I haz the power, 2016-02-12 21:17:51]

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

[commit-bot: I haz the power, 2016-02-12 21:17:51]

Try jobs failed on following builders:
  android_chromium_gn_compile_dbg on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_chro...)

[Reilly Grant (use Gerrit), 2016-02-13 00:58:15]

The CQ bit was checked by reillyg@chromium.org

[commit-bot: I haz the power, 2016-02-13 00:59:17]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1568673002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1568673002/100001

[commit-bot: I haz the power, 2016-02-13 01:54:28]

Description was changed from

==========
Parse USB interface association descriptors.

USB interface association descriptors are used to combine multiple
interfaces into a single functional group. This patch adds support for
parsing them out of the |extra_data| field left by libusb's parsing of
configuration, interface and endpoint descriptors. The resulting
association is then represented by setting the |first_interface| field
of each interface in a function to the |interface_number| of the first
interface in the function.

WebUSB will use these associations to set permissions for an entire
function with a single descriptor.

BUG=492204
==========

to

==========
Parse USB interface association descriptors.

USB interface association descriptors are used to combine multiple
interfaces into a single functional group. This patch adds support for
parsing them out of the |extra_data| field left by libusb's parsing of
configuration, interface and endpoint descriptors. The resulting
association is then represented by setting the |first_interface| field
of each interface in a function to the |interface_number| of the first
interface in the function.

WebUSB will use these associations to set permissions for an entire
function with a single descriptor.

BUG=492204
==========

[commit-bot: I haz the power, 2016-02-13 01:54:29]

Committed patchset #3 (id:100001)

[gab, 2016-02-14 21:02:16]

gab@chromium.org changed reviewers:
+ gab@chromium.org

[gab, 2016-02-14 21:02:17]

https://codereview.chromium.org/1568673002/diff/100001/device/usb/usb_descrip...
File device/usb/usb_descriptors.cc (right):

https://codereview.chromium.org/1568673002/diff/100001/device/usb/usb_descrip...
device/usb/usb_descriptors.cc:191: interface_it != interfaces.end() &&
This is wrong as if the inner loop exits because this hit the end, the outer
loop will proceed with another increment prior to verifying its condition and
thus will go past the end.

http://crbug.com/586824

I will revert this CL for now as it otherwise makes Chrome hang when using USB
devices. Hopefully tomorrow's Canary picks up the revert.

Also: feels like this should be unit-tested.

[gab, 2016-02-14 21:03:04]

A revert of this CL (patchset #3 id:100001) has been created in
https://codereview.chromium.org/1697863003/ by gab@chromium.org.

The reason for reverting is: http://crbug.com/586824

Makes Chrome hang when using USB devices..

[gab, 2016-02-14 21:10:07]

https://codereview.chromium.org/1568673002/diff/100001/device/usb/usb_descrip...
File device/usb/usb_descriptors.cc (right):

https://codereview.chromium.org/1568673002/diff/100001/device/usb/usb_descrip...
device/usb/usb_descriptors.cc:191: interface_it != interfaces.end() &&
On 2016/02/14 21:02:17, gab (OOO until Tuesday 16th) wrote:
> This is wrong as if the inner loop exits because this hit the end, the outer
> loop will proceed with another increment prior to verifying its condition and
> thus will go past the end.
> 
> http://crbug.com/586824
> 
> I will revert this CL for now as it otherwise makes Chrome hang when using USB
> devices. Hopefully tomorrow's Canary picks up the revert.
> 
> Also: feels like this should be unit-tested.

Apologies, looks like there is a unit test, but it doesn't trigger this issue,
should probably add a case that hits the end in the inner-loop.

Cheers!

[Reilly Grant (use Gerrit), 2016-02-15 00:35:32]

Patchset #4 (id:120001) has been deleted

[commit-bot: I haz the power, 2016-02-16 22:46:56]

Description was changed from

==========
Parse USB interface association descriptors.

USB interface association descriptors are used to combine multiple
interfaces into a single functional group. This patch adds support for
parsing them out of the |extra_data| field left by libusb's parsing of
configuration, interface and endpoint descriptors. The resulting
association is then represented by setting the |first_interface| field
of each interface in a function to the |interface_number| of the first
interface in the function.

WebUSB will use these associations to set permissions for an entire
function with a single descriptor.

BUG=492204
==========

to

==========
Parse USB interface association descriptors.

USB interface association descriptors are used to combine multiple
interfaces into a single functional group. This patch adds support for
parsing them out of the |extra_data| field left by libusb's parsing of
configuration, interface and endpoint descriptors. The resulting
association is then represented by setting the |first_interface| field
of each interface in a function to the |interface_number| of the first
interface in the function.

WebUSB will use these associations to set permissions for an entire
function with a single descriptor.

BUG=492204

Committed: https://crrev.com/b61e467059e55ad71f1b7ae5e1412beb3d4803e0
Cr-Commit-Position: refs/heads/master@{#375324}
==========

[commit-bot: I haz the power, 2016-02-16 22:46:57]

Patchset 3 (id:??) landed as
https://crrev.com/b61e467059e55ad71f1b7ae5e1412beb3d4803e0
Cr-Commit-Position: refs/heads/master@{#375324}

[gab, 2016-02-16 22:48:45]

On 2016/02/16 22:46:57, commit-bot: I haz the power wrote:
> Patchset 3 (id:??) landed as
> https://crrev.com/b61e467059e55ad71f1b7ae5e1412beb3d4803e0
> Cr-Commit-Position: refs/heads/master@{#375324}

Hmmm? This was relanded without the fix?

[gab, 2016-02-16 22:49:55]

On 2016/02/16 22:48:45, gab wrote:
> On 2016/02/16 22:46:57, commit-bot: I haz the power wrote:
> > Patchset 3 (id:??) landed as
> > https://crrev.com/b61e467059e55ad71f1b7ae5e1412beb3d4803e0
> > Cr-Commit-Position: refs/heads/master@{#375324}
> 
> Hmmm? This was relanded without the fix?

Nvm, looks like commit-bot is going nuts commenting on all the issues it landed
long ago..