Add EME Layout test for unclosed the key session

Add EME Layout test for unclosed key session

EME spec notes that "the User Agent must use the CDM to close the
session before User Agent state associated with the session is
deleted"
So add test close the key session associated with the object
when the page is inaccessible.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
TEST=EME layout tests pass

Committed: https://crrev.com/f6f57160426b548a43d8ebc1b1c70c3796dc7923
Cr-Commit-Position: refs/heads/master@{#434087}

[Dongheun Kang, 2016-10-28 06:43:03]

Dear jrunmell

If call contextDestroyed before close session, clear session without closeTask.
I found this issues during integrating vudu tv app on tv platform.

App use iframe include video player for playing commercial contents(ex.
dash.js). This issues occurs remove iframe after close MediaKeySession.

Please review this patch.

Thanks.

[ddorwin, 2016-10-28 14:53:07]

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

[ddorwin, 2016-10-28 14:53:08]

Thanks. Please file a bug at crbug.com, including what happens. Is there a
crash? Also, we should try to write a layout test for this.

[Dongheun Kang, 2016-10-31 01:19:16]

On 2016/10/28 14:53:08, ddorwin wrote:
> Thanks. Please file a bug at http://crbug.com, including what happens. Is
there a
> crash? Also, we should try to write a layout test for this.

I updated commit massages but didn't update this pages's Description.
I submit new bug (https://bugs.chromium.org/p/chromium/issues/detail?id=660750).

please check this issues. thanks.

[Dongheun Kang, 2016-11-01 01:13:29]

On 2016/10/28 14:53:08, ddorwin wrote:
> Thanks. Please file a bug at http://crbug.com, including what happens. Is
there a
> crash? Also, we should try to write a layout test for this.

Dear ddorwin, jrummell

I added a layout test. I don't know the method of checking result(call closeTask
or not).
But test easily this issues.

content_shell --run-layout-test --no-sandbox
./third_party/WebKit/LayoutTests/media/encrypted-media/encrypted-media-session-close-and-context-destroyed.html
--enable-logging --vmodule=MediaKey*=3,*=0

Please check this patch with layout test.

Thanks.

[haraken, 2016-11-01 02:18:33]

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

[haraken, 2016-11-01 02:18:34]

https://codereview.chromium.org/2454333002/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp
(right):

https://codereview.chromium.org/2454333002/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1041: if
(m_actionTimer.isActive()) {

Do we need to check this?

https://codereview.chromium.org/2454333002/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1046:
break;

Isn't it possible that there are more than one pending actions?

[Dongheun Kang, 2016-11-01 03:32:56]

Dear haraken please check my comment. thanks.

https://codereview.chromium.org/2454333002/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp
(right):

https://codereview.chromium.org/2454333002/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1041: if
(m_actionTimer.isActive()) {
On 2016/11/01 02:18:33, haraken wrote:
> 
> Do we need to check this?

I think there seems to be no need.

https://codereview.chromium.org/2454333002/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1046:
break;
On 2016/11/01 02:18:33, haraken wrote:
> 
> Isn't it possible that there are more than one pending actions?

So I can change as like following code.

  if (!m_pendingActions.isEmpty())
    actionTimerFired(0);

Do you this about?

[haraken, 2016-11-01 03:53:12]

On 2016/11/01 03:32:56, Dongheun Kang wrote:
> Dear haraken please check my comment. thanks.
> 
>
https://codereview.chromium.org/2454333002/diff/40001/third_party/WebKit/Sour...
> File third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp
> (right):
> 
>
https://codereview.chromium.org/2454333002/diff/40001/third_party/WebKit/Sour...
> third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1041: if
> (m_actionTimer.isActive()) {
> On 2016/11/01 02:18:33, haraken wrote:
> > 
> > Do we need to check this?
> 
> I think there seems to be no need.
> 
>
https://codereview.chromium.org/2454333002/diff/40001/third_party/WebKit/Sour...
> third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1046:
> break;
> On 2016/11/01 02:18:33, haraken wrote:
> > 
> > Isn't it possible that there are more than one pending actions?
> 
> So I can change as like following code.
> 
>   if (!m_pendingActions.isEmpty())
>     actionTimerFired(0);
> 
> Do you this about?

I meant, we shouldn't add 'break' here.

[Dongheun Kang, 2016-11-01 04:05:37]

https://codereview.chromium.org/2454333002/diff/40001/third_party/WebKit/Sour...
File third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp
(right):

https://codereview.chromium.org/2454333002/diff/40001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1046:
break;
On 2016/11/01 02:18:33, haraken wrote:
> 
> Isn't it possible that there are more than one pending actions?
Sorry. I miss-understand your comments. This patch is just for differed
closeTask. So add break.
But I will simply change the patch.
Thanks.

[Dongheun Kang, 2016-11-01 04:05:45]

Description was changed from

==========
Call closeTask before MediaKeySession::contextDestroyed

If call contextDestroyed before close session, clear session without closeTask.
So call session closeTask if there is a delayed pendingAction::Close.

R=jrummell@chromium.org
BUG=
==========

to

==========
Call closeTask before MediaKeySession::contextDestroyed

If call contextDestroyed before close session, clear session without closeTask.
So call session closeTask if there is a delayed pendingAction::Close.

R=jrummell@chromium.org, ddorwin@chromium.org
BUG=660750
==========

[Dongheun Kang, 2016-11-01 04:14:18]

Description was changed from

==========
Call closeTask before MediaKeySession::contextDestroyed

If call contextDestroyed before close session, clear session without closeTask.
So call session closeTask if there is a delayed pendingAction::Close.

R=jrummell@chromium.org, ddorwin@chromium.org
BUG=660750
==========

to

==========
Call closeTask before MediaKeySession::contextDestroyed

If call contextDestroyed before close session, clear session without closeTask.
So call actionTimerFired(0) if there is a delayed pendingAction.

R=jrummell@chromium.org, ddorwin@chromium.org
BUG=660750
==========

[Dongheun Kang, 2016-11-01 05:15:07]

Dear all
I upload new patch set (#4).
please review this patch.
thanks.

[Dongheun Kang, 2016-11-01 05:15:41]

Description was changed from

==========
Call closeTask before MediaKeySession::contextDestroyed

If call contextDestroyed before close session, clear session without closeTask.
So call actionTimerFired(0) if there is a delayed pendingAction.

R=jrummell@chromium.org, ddorwin@chromium.org
BUG=660750
==========

to

==========
Call closeTask before MediaKeySession::contextDestroyed

If call contextDestroyed before close session, clear session without closeTask.
So call actionTimerFired(0) if there is a delayed pendingAction.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
==========

[haraken, 2016-11-01 06:28:25]

https://codereview.chromium.org/2454333002/diff/60001/third_party/WebKit/Sour...
File third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp
(right):

https://codereview.chromium.org/2454333002/diff/60001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1042:
actionTimerFired(0);

If there is a pending task with PendingAction::GenerateRequest, you'll end up
with initiating a new request in contextDestroyed(). Is it really what you want
to do here?

[Dongheun Kang, 2016-11-01 06:48:32]

Description was changed from

==========
Call closeTask before MediaKeySession::contextDestroyed

If call contextDestroyed before close session, clear session without closeTask.
So call actionTimerFired(0) if there is a delayed pendingAction.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
==========

to

==========
Call closeTask before MediaKeySession::contextDestroyed

If call contextDestroyed before close session, clear session without closeTask.
So call closeTask if there is a delayed pendingAction::Close.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
==========

[Dongheun Kang, 2016-11-01 06:54:44]

I rollback to a previous patch.

https://codereview.chromium.org/2454333002/diff/60001/third_party/WebKit/Sour...
File third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp
(right):

https://codereview.chromium.org/2454333002/diff/60001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1042:
actionTimerFired(0);
On 2016/11/01 06:28:25, haraken wrote:
> 
> If there is a pending task with PendingAction::GenerateRequest, you'll end up
> with initiating a new request in contextDestroyed(). Is it really what you
want
> to do here?
> 

No. I'd better rollback to a previous patch.

[haraken, 2016-11-01 06:56:15]

https://codereview.chromium.org/2454333002/diff/80001/third_party/WebKit/Sour...
File third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp
(right):

https://codereview.chromium.org/2454333002/diff/80001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1045:
break;

Sorry, would you explain why we need 'break' here? What happens if we have more
than one tasks with PendingAction::Close?

[Dongheun Kang, 2016-11-01 07:09:30]

https://codereview.chromium.org/2454333002/diff/80001/third_party/WebKit/Sour...
File third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp
(right):

https://codereview.chromium.org/2454333002/diff/80001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1045:
break;
On 2016/11/01 06:56:15, haraken wrote:
> 
> Sorry, would you explain why we need 'break' here? What happens if we have
more
> than one tasks with PendingAction::Close?

I think can't add duplicated close action in pendingAction.
And I don't need to call close in same key session. Once a call closeTask is
enough.

[haraken, 2016-11-01 07:21:38]

LGTM

Please wait for jrummell or ddorwin's approval.

https://codereview.chromium.org/2454333002/diff/80001/third_party/WebKit/Sour...
File third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp
(right):

https://codereview.chromium.org/2454333002/diff/80001/third_party/WebKit/Sour...
third_party/WebKit/Source/modules/encryptedmedia/MediaKeySession.cpp:1040: // So
call closeTask if there is a delayed pendingAction::Close.

// When the context is destroyed, we need to close a pending action with
PendingAction::Close.

[Dongheun Kang, 2016-11-02 23:10:57]

Dear jrummell, ddorwin

please review this patch.

Thanks.

[jrummell, 2016-11-03 18:05:15]

Currently the EME spec states "If a MediaKeySession object becomes inaccessible
to the page and is not closed, the User Agent must run the MediaKeySession
destroyed algorithm before User Agent state associated with the session is
deleted." To me this implies that we should call close() regardless of whether
the user has called it or not, if the session is not marked as closed.

My preference is that this is done in WebContentDecryptionModuleSessionImpl,
which is the Chromium side object that matches blink::MediaKeySession (what
|m_session| points to). It will take a little bit of work since
WebContentDecryptionModuleSessionImpl doesn't keep track whether close() has
been called or not.

I don't like it being done in MediaKeySession since calling
WebContentDecryptionModuleSessionImpl::close() requires a promise to be passed
as a parameter. MKS is going away, so it seems unnecessary to to have to
(possibly) create another blink object that will not really be used. Doing it on
the Chromium side you'll still have to create a dummy CdmPromise, but that
should be trivial (since it will just ignore resolve/reject calls).

[Dongheun Kang, 2016-11-03 18:59:20]

On 2016/11/03 18:05:15, jrummell wrote:
> Currently the EME spec states "If a MediaKeySession object becomes
inaccessible
> to the page and is not closed, the User Agent must run the MediaKeySession
> destroyed algorithm before User Agent state associated with the session is
> deleted." To me this implies that we should call close() regardless of whether
> the user has called it or not, if the session is not marked as closed.
> 
> My preference is that this is done in WebContentDecryptionModuleSessionImpl,
> which is the Chromium side object that matches blink::MediaKeySession (what
> |m_session| points to). It will take a little bit of work since
> WebContentDecryptionModuleSessionImpl doesn't keep track whether close() has
> been called or not.
> 
> I don't like it being done in MediaKeySession since calling
> WebContentDecryptionModuleSessionImpl::close() requires a promise to be passed
> as a parameter. MKS is going away, so it seems unnecessary to to have to
> (possibly) create another blink object that will not really be used. Doing it
on
> the Chromium side you'll still have to create a dummy CdmPromise, but that
> should be trivial (since it will just ignore resolve/reject calls).

This patch was created because of a very serious problem that paid content could
no longer be played on our platform.

There is also a problem with simply saying that the app did not follow the spec.

So I think we need this patch until modify eme spec for specific user agent
actions.

Thanks you for reviewing this patch.

[jrummell, 2016-11-03 20:16:00]

On 2016/11/03 18:59:20, Dongheun Kang wrote:
> On 2016/11/03 18:05:15, jrummell wrote:
> > Currently the EME spec states "If a MediaKeySession object becomes
> inaccessible
> > to the page and is not closed, the User Agent must run the MediaKeySession
> > destroyed algorithm before User Agent state associated with the session is
> > deleted." To me this implies that we should call close() regardless of
whether
> > the user has called it or not, if the session is not marked as closed.
> > 
> > My preference is that this is done in WebContentDecryptionModuleSessionImpl,
> > which is the Chromium side object that matches blink::MediaKeySession (what
> > |m_session| points to). It will take a little bit of work since
> > WebContentDecryptionModuleSessionImpl doesn't keep track whether close() has
> > been called or not.
> > 
> > I don't like it being done in MediaKeySession since calling
> > WebContentDecryptionModuleSessionImpl::close() requires a promise to be
passed
> > as a parameter. MKS is going away, so it seems unnecessary to to have to
> > (possibly) create another blink object that will not really be used. Doing
it
> on
> > the Chromium side you'll still have to create a dummy CdmPromise, but that
> > should be trivial (since it will just ignore resolve/reject calls).
> 
> This patch was created because of a very serious problem that paid content
could
> no longer be played on our platform.
> 
> There is also a problem with simply saying that the app did not follow the
spec.
> 
> So I think we need this patch until modify eme spec for specific user agent
> actions.
> 
> Thanks you for reviewing this patch.

I am not disagreeing with the problem, just the solution as presented so far.
The EME spec is pretty clear, so I'd prefer that we fix it to match the spec
rather than just addressing part of the problem.

[Dongheun Kang, 2016-11-03 21:59:16]

On 2016/11/03 20:16:00, jrummell wrote:
> On 2016/11/03 18:59:20, Dongheun Kang wrote:
> > On 2016/11/03 18:05:15, jrummell wrote:
> > > Currently the EME spec states "If a MediaKeySession object becomes
> > inaccessible
> > > to the page and is not closed, the User Agent must run the MediaKeySession
> > > destroyed algorithm before User Agent state associated with the session is
> > > deleted." To me this implies that we should call close() regardless of
> whether
> > > the user has called it or not, if the session is not marked as closed.
> > > 
> > > My preference is that this is done in
WebContentDecryptionModuleSessionImpl,
> > > which is the Chromium side object that matches blink::MediaKeySession
(what
> > > |m_session| points to). It will take a little bit of work since
> > > WebContentDecryptionModuleSessionImpl doesn't keep track whether close()
has
> > > been called or not.
> > > 
> > > I don't like it being done in MediaKeySession since calling
> > > WebContentDecryptionModuleSessionImpl::close() requires a promise to be
> passed
> > > as a parameter. MKS is going away, so it seems unnecessary to to have to
> > > (possibly) create another blink object that will not really be used. Doing
> it
> > on
> > > the Chromium side you'll still have to create a dummy CdmPromise, but that
> > > should be trivial (since it will just ignore resolve/reject calls).
> > 
> > This patch was created because of a very serious problem that paid content
> could
> > no longer be played on our platform.
> > 
> > There is also a problem with simply saying that the app did not follow the
> spec.
> > 
> > So I think we need this patch until modify eme spec for specific user agent
> > actions.
> > 
> > Thanks you for reviewing this patch.
> 
> I am not disagreeing with the problem, just the solution as presented so far.
> The EME spec is pretty clear, so I'd prefer that we fix it to match the spec
> rather than just addressing part of the problem.

This is not to say that the specification is wrong, but we need additional
clarification.

Dear ddorwin

Do you have the same opinion?

[ddorwin, 2016-11-07 19:54:34]

On 2016/11/03 21:59:16, Dongheun Kang wrote:
> 
> This is not to say that the specification is wrong, but we need additional
> clarification.
> 
> Dear ddorwin
> 
> Do you have the same opinion?

My understanding is that you need the session to be explicitly closed in order
to free its resources. Destroying the CDM instance should also close those
resources. How this is implemented is a bit of an implementation detail since
non-closed sessions remain accessible until the MediaKeys is inaccessible.
Please see https://bugs.chromium.org/p/chromium/issues/detail?id=660750#c7.

While https://codereview.chromium.org/2484873002/ attempts to explicitly close
the session, the CDM implementation should make sure it is destroying keys /
releasing resources related to the session when the CDM instance is destroyed.
Note that the CDM instance is a concept related to the MediaKeys object. The
implementation is responsible for ensuring this concept, including that keys are
not shared across MediaKeys objects.

[Dongheun Kang, 2016-11-08 23:33:07]

On 2016/11/07 19:54:34, ddorwin wrote:
> On 2016/11/03 21:59:16, Dongheun Kang wrote:
> > 
> > This is not to say that the specification is wrong, but we need additional
> > clarification.
> > 
> > Dear ddorwin
> > 
> > Do you have the same opinion?
> 
> My understanding is that you need the session to be explicitly closed in order
> to free its resources. Destroying the CDM instance should also close those
> resources. How this is implemented is a bit of an implementation detail since
> non-closed sessions remain accessible until the MediaKeys is inaccessible.
> Please see https://bugs.chromium.org/p/chromium/issues/detail?id=660750#c7.
> 
> While https://codereview.chromium.org/2484873002/ attempts to explicitly close
> the session, the CDM implementation should make sure it is destroying keys /
> releasing resources related to the session when the CDM instance is destroyed.
> Note that the CDM instance is a concept related to the MediaKeys object. The
> implementation is responsible for ensuring this concept, including that keys
are
> not shared across MediaKeys objects.

Thank you for your comment.

I understand that your comments should clear up the resources in cdm when the
key session object is cleaned up. is right?
and Is this spec change(https://github.com/w3c/encrypted-media/pull/359) related
to this issues?

[ddorwin, 2016-11-09 21:42:25]

On 2016/11/08 23:33:07, Dongheun Kang wrote:
> On 2016/11/07 19:54:34, ddorwin wrote:
> > On 2016/11/03 21:59:16, Dongheun Kang wrote:
> > > 
> > > This is not to say that the specification is wrong, but we need additional
> > > clarification.
> > > 
> > > Dear ddorwin
> > > 
> > > Do you have the same opinion?
> > 
> > My understanding is that you need the session to be explicitly closed in
order
> > to free its resources. Destroying the CDM instance should also close those
> > resources. How this is implemented is a bit of an implementation detail
since
> > non-closed sessions remain accessible until the MediaKeys is inaccessible.
> > Please see https://bugs.chromium.org/p/chromium/issues/detail?id=660750#c7.
> > 
> > While https://codereview.chromium.org/2484873002/ attempts to explicitly
close
> > the session, the CDM implementation should make sure it is destroying keys /
> > releasing resources related to the session when the CDM instance is
destroyed.
> > Note that the CDM instance is a concept related to the MediaKeys object. The
> > implementation is responsible for ensuring this concept, including that keys
> are
> > not shared across MediaKeys objects.
> 
> Thank you for your comment.
> 
> I understand that your comments should clear up the resources in cdm when the
> key session object is cleaned up. is right?
> and Is this spec change(https://github.com/w3c/encrypted-media/pull/359)
related
> to this issues?

Yes. The associated issue (#358) was filed after reviewing
https://codereview.chromium.org/2484873002/. The pull request removes the
problematic text and clarifies what implementations should do and what authors
should expect. As noted in the final version of the pull request and now the
spec, applications that are concerned about timing should call close() and wait
for the returned promise to be resolved.

[jrummell, 2016-11-12 00:34:32]

Now that https://codereview.chromium.org/2484873002/ has landed (which should
fix the problem), you can remove the changes to MediaKeySession and get the test
added to Chromium.

[Dongheun Kang, 2016-11-14 01:38:58]

Description was changed from

==========
Call closeTask before MediaKeySession::contextDestroyed

If call contextDestroyed before close session, clear session without closeTask.
So call closeTask if there is a delayed pendingAction::Close.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
==========

to

==========
Add EME Layout test for unclosed the key session
 
EME spec notes that "the User Agent must use the CDM to close the 
session before User Agent state associated with the session is
deleted"
So add test close the key session associated with the object
when the page is inaccessible.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
TEST=EME layout tests pass
    TEST=EME layout tests pass
==========

[Dongheun Kang, 2016-11-14 01:39:11]

Description was changed from

==========
Add EME Layout test for unclosed the key session
 
EME spec notes that "the User Agent must use the CDM to close the 
session before User Agent state associated with the session is
deleted"
So add test close the key session associated with the object
when the page is inaccessible.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
TEST=EME layout tests pass
    TEST=EME layout tests pass
==========

to

==========
Add EME Layout test for unclosed the key session
 
EME spec notes that "the User Agent must use the CDM to close the 
session before User Agent state associated with the session is
deleted"
So add test close the key session associated with the object
when the page is inaccessible.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
TEST=EME layout tests pass
==========

[Dongheun Kang, 2016-11-14 01:52:14]

On 2016/11/12 00:34:32, jrummell wrote:
> Now that https://codereview.chromium.org/2484873002/ has landed (which should
> fix the problem), you can remove the changes to MediaKeySession and get the
test
> added to Chromium.

Thank you.

I tested https://codereview.chromium.org/2484873002/. works well.
And also update layout test that removed the changes to MediaKeySession.

I sincerely thank all those who have reviewed this issue.

[Dongheun Kang, 2016-11-15 00:49:22]

Dear all

please review this layout test patch.

Thanks.

[Dongheun Kang, 2016-11-16 22:35:43]

Dear jrummell, ddorwin
please review this patch. thanks.

[Dongheun Kang, 2016-11-22 10:07:02]

The CQ bit was checked by dongheun.kang@lge.com

[Dongheun Kang, 2016-11-22 10:07:02]

The patchset sent to the CQ was uploaded after l-g-t-m from haraken@chromium.org
Link to the patchset: https://codereview.chromium.org/2454333002/#ps100001
(title: "Add EME Layout test for unclosed the key session")

[commit-bot: I haz the power, 2016-11-22 10:07:17]

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, 2016-11-22 11:00:53]

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

[commit-bot: I haz the power, 2016-11-22 11:00:54]

Failed to apply patch for AUTHORS:
While running git apply --index -p1;
  error: patch failed: AUTHORS:176
  error: AUTHORS: patch does not apply

Patch:       AUTHORS
Index: AUTHORS
diff --git a/AUTHORS b/AUTHORS
index
e22a2bd840ac926c38340b47c145f8b601c84156..cd1d61e9f5b1ac7ef8b7d1242c4a3b973acee67b
100644
--- a/AUTHORS
+++ b/AUTHORS
@@ -176,6 +176,7 @@ Dominic Jodoin <dominic.jodoin@gmail.com>
 Dominik Röttsches <dominik.rottsches@intel.com>
 Don Woodward <woodward@adobe.com>
 Donghee Na <corona10@gmail.com>
+Dongheun Kang <dongheun.kang@lge.com>
 Dongie Agnir <dongie.agnir@gmail.com>
 Dongjun Kim <djmix.kim@samsung.com>
 Dongseong Hwang <dongseong.hwang@intel.com>

[ddorwin, 2016-11-22 16:45:56]

Description was changed from

==========
Add EME Layout test for unclosed the key session
 
EME spec notes that "the User Agent must use the CDM to close the 
session before User Agent state associated with the session is
deleted"
So add test close the key session associated with the object
when the page is inaccessible.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
TEST=EME layout tests pass
==========

to

==========
Add EME Layout test for unclosed key session
 
EME spec notes that "the User Agent must use the CDM to close the 
session before User Agent state associated with the session is
deleted"
So add test close the key session associated with the object
when the page is inaccessible.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
TEST=EME layout tests pass
==========

[Dongheun Kang, 2016-11-22 23:42:14]

The CQ bit was checked by dongheun.kang@lge.com

[Dongheun Kang, 2016-11-22 23:42:15]

The patchset sent to the CQ was uploaded after l-g-t-m from haraken@chromium.org
Link to the patchset: https://codereview.chromium.org/2454333002/#ps120001
(title: "Add EME Layout test for unclosed the key session")

[commit-bot: I haz the power, 2016-11-22 23:43:12]

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, 2016-11-23 02:30:24]

CQ is committing da patch.

Bot data: {"patchset_id": 120001, "attempt_start_ts": 1479858134724490,
"parent_rev": "3f6b0b2b8bf7ec7e9399856b3ab2f8f185999fa2", "commit_rev":
"3cfb19dfb662f8261d437faec929cec26ba76cb9"}

[commit-bot: I haz the power, 2016-11-23 02:30:50]

Description was changed from

==========
Add EME Layout test for unclosed key session
 
EME spec notes that "the User Agent must use the CDM to close the 
session before User Agent state associated with the session is
deleted"
So add test close the key session associated with the object
when the page is inaccessible.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
TEST=EME layout tests pass
==========

to

==========
Add EME Layout test for unclosed key session
 
EME spec notes that "the User Agent must use the CDM to close the 
session before User Agent state associated with the session is
deleted"
So add test close the key session associated with the object
when the page is inaccessible.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
TEST=EME layout tests pass
==========

[commit-bot: I haz the power, 2016-11-23 02:30:51]

Committed patchset #7 (id:120001)

[commit-bot: I haz the power, 2016-11-23 02:33:13]

Description was changed from

==========
Add EME Layout test for unclosed key session
 
EME spec notes that "the User Agent must use the CDM to close the 
session before User Agent state associated with the session is
deleted"
So add test close the key session associated with the object
when the page is inaccessible.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
TEST=EME layout tests pass
==========

to

==========
Add EME Layout test for unclosed key session

EME spec notes that "the User Agent must use the CDM to close the
session before User Agent state associated with the session is
deleted"
So add test close the key session associated with the object
when the page is inaccessible.

R=jrummell@chromium.org, ddorwin@chromium.org, haraken@chromium.org
BUG=660750
TEST=EME layout tests pass

Committed: https://crrev.com/f6f57160426b548a43d8ebc1b1c70c3796dc7923
Cr-Commit-Position: refs/heads/master@{#434087}
==========

[commit-bot: I haz the power, 2016-11-23 02:33:14]

Patchset 7 (id:??) landed as
https://crrev.com/f6f57160426b548a43d8ebc1b1c70c3796dc7923
Cr-Commit-Position: refs/heads/master@{#434087}