Add terminateServiceWorker method for layout tests - chromium side

Add terminateServiceWorker method for layout tests.

The blink side of this change: https://codereview.chromium.org/695643002/

BUG=None

Committed: https://crrev.com/408cf16324c95887c4893f7c4e923338b9573080
Cr-Commit-Position: refs/heads/master@{#305275}

[Marijn Kruisselbrink, 2014-11-04 22:49:53]

Patchset #1 (id:1) has been deleted

[Marijn Kruisselbrink, 2014-11-17 23:18:39]

mek@chromium.org changed reviewers:
+ dominicc@chromium.org

[dominicc (has gone to gerrit), 2014-11-20 05:49:22]

dominicc@chromium.org changed reviewers:
+ jam@chromium.org, nhiroki@chromium.org

[dominicc (has gone to gerrit), 2014-11-20 05:49:34]

On 2014/11/17 23:18:39, Marijn Kruisselbrink wrote:

Informally LGTM.

[jam, 2014-11-20 21:17:24]

lgtm

[nhiroki, 2014-11-21 02:18:51]

LGTM with nits

https://codereview.chromium.org/688413002/diff/40001/content/common/service_w...
File content/common/service_worker/service_worker_messages.h (right):

https://codereview.chromium.org/688413002/diff/40001/content/common/service_w...
content/common/service_worker/service_worker_messages.h:246:
IPC_MESSAGE_CONTROL1(ServiceWorkerHostMsg_TerminateWorker, int /* handle_id */)
nit: How about moving this to line 162 to group messages by handler?

Can you add a line break after the comma?

[Marijn Kruisselbrink, 2014-11-21 18:41:22]

mek@chromium.org changed reviewers:
+ tsepez@chromium.org

[Marijn Kruisselbrink, 2014-11-21 18:41:22]

+tsepez for IPC OWNERS

https://codereview.chromium.org/688413002/diff/40001/content/common/service_w...
File content/common/service_worker/service_worker_messages.h (right):

https://codereview.chromium.org/688413002/diff/40001/content/common/service_w...
content/common/service_worker/service_worker_messages.h:246:
IPC_MESSAGE_CONTROL1(ServiceWorkerHostMsg_TerminateWorker, int /* handle_id */)
On 2014/11/21 02:18:51, nhiroki wrote:
> nit: How about moving this to line 162 to group messages by handler?
> 
> Can you add a line break after the comma?

Done.

[Tom Sepez, 2014-11-21 18:46:52]

What ensure that the renderer has the right to terminate the service worker?  If
its a compromised renderer, could it terminate service workers unrelated to the
origin that it is currently rendering?

[Marijn Kruisselbrink, 2014-11-21 18:53:43]

On 2014/11/21 18:46:52, Tom Sepez wrote:
> What ensure that the renderer has the right to terminate the service worker? 
If
> its a compromised renderer, could it terminate service workers unrelated to
the
> origin that it is currently rendering?
The handle_id that is used to identify the service worker is a
per-renderer-process identifier. ServiceWorkerDispatcherHost has its own mapping
to map this to an actual Service Worker instance. So a renderer can't terminate
service workers that aren't already associated with that renderer process in
some manner (so assuming one origin per renderer a renderer can't do anything to
service workers for other origins).

[Marijn Kruisselbrink, 2014-11-21 19:12:49]

On 2014/11/21 18:53:43, Marijn Kruisselbrink wrote:
> On 2014/11/21 18:46:52, Tom Sepez wrote:
> > What ensure that the renderer has the right to terminate the service worker?

> If
> > its a compromised renderer, could it terminate service workers unrelated to
> the
> > origin that it is currently rendering?
> The handle_id that is used to identify the service worker is a
> per-renderer-process identifier. ServiceWorkerDispatcherHost has its own
mapping
> to map this to an actual Service Worker instance. So a renderer can't
terminate
> service workers that aren't already associated with that renderer process in
> some manner (so assuming one origin per renderer a renderer can't do anything
to
> service workers for other origins).
For even more security it might actually be nice to just completely
reject/ignore
this IPC if chrome isn't running layout tests (I'm not entirely sure how to
detect
that this is a content_shell --dump-render-tree invocation, but there must be
other
code already doing the same).

[Tom Sepez, 2014-11-21 19:15:42]

On 2014/11/21 19:12:49, Marijn Kruisselbrink wrote:
> On 2014/11/21 18:53:43, Marijn Kruisselbrink wrote:
> > On 2014/11/21 18:46:52, Tom Sepez wrote:
> > > What ensure that the renderer has the right to terminate the service
worker?
> 
> > If
> > > its a compromised renderer, could it terminate service workers unrelated
to
> > the
> > > origin that it is currently rendering?
> > The handle_id that is used to identify the service worker is a
> > per-renderer-process identifier. ServiceWorkerDispatcherHost has its own
> mapping
> > to map this to an actual Service Worker instance. So a renderer can't
> terminate
> > service workers that aren't already associated with that renderer process in
> > some manner (so assuming one origin per renderer a renderer can't do
anything
> to
> > service workers for other origins).
OK LGTM.

> For even more security it might actually be nice to just completely
> reject/ignore
> this IPC if chrome isn't running layout tests (I'm not entirely sure how to
> detect
> that this is a content_shell --dump-render-tree invocation, but there must be
> other
> code already doing the same).
Sounds good.  Please take a stab at this as a follow-up.

[Marijn Kruisselbrink, 2014-11-21 19:21:40]

The CQ bit was checked by mek@chromium.org

[commit-bot: I haz the power, 2014-11-21 19:22:37]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/688413002/60001

[commit-bot: I haz the power, 2014-11-21 20:31:59]

Committed patchset #3 (id:60001)

[commit-bot: I haz the power, 2014-11-21 20:32:29]

Patchset 3 (id:??) landed as
https://crrev.com/408cf16324c95887c4893f7c4e923338b9573080
Cr-Commit-Position: refs/heads/master@{#305275}