Prepare to replicate sandbox flags for OOPIF

Prepare to replicate sandbox flags for OOPIF.

Corresponding Chromium side CL: https://codereview.chromium.org/837283003/

* Add a new enum to use sandbox flags from the Web layer.
* Add plumbing to use flags passed from the browser process when creating LocalFrames with a remote parent (via RemoteBridgeFrameOwner).
* Allow a RemoteSecurityContext to set replicated SandboxFlags.  This will be used at SecurityContext init time to let a LocalFrame check whether its parent is sandboxed when the parent is a RemoteFrame (to ensure children of sandboxed frames are also sandboxed).

This CL doesn't actually send sandbox flags to the browser process when new frames are created.  This will happen in a follow-up CL (https://codereview.chromium.org/838903002), once the Chromium-side CL lands.

BUG=426512
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=188409

[alexmos, 2015-01-08 01:21:58]

alexmos@chromium.org changed reviewers:
+ dcheng@chromium.org

[alexmos, 2015-01-08 01:21:59]

Daniel, could you please take a look?  This is the first part of replicating
sandbox flags on the Blink side.  Feel free also to take a look at the linked
CLs to see if the overall approach makes sense.

I included moving sandbox flags from ExecutionContext to SecurityContext here,
which I know Nate also has in https://codereview.chromium.org/799923006/.  Just
wanted to get this review started, but happy to wait/rebase on that.

[dcheng, 2015-01-08 22:50:45]

https://codereview.chromium.org/793493003/diff/20001/Source/web/WebRemoteFram...
File Source/web/WebRemoteFrameImpl.cpp (right):

https://codereview.chromium.org/793493003/diff/20001/Source/web/WebRemoteFram...
Source/web/WebRemoteFrameImpl.cpp:809: OwnPtr<RemoteBridgeFrameOwner> owner =
RemoteBridgeFrameOwner::create(child);
OwnPtrWillBeRawPtr

https://codereview.chromium.org/793493003/diff/20001/public/web/WebSandboxFla...
File public/web/WebSandboxFlags.h (right):

https://codereview.chromium.org/793493003/diff/20001/public/web/WebSandboxFla...
public/web/WebSandboxFlags.h:26: #ifndef WebSandboxFlags_h
I'd probably reupload this with --no-find-copies.

Also, headers for new files should be:
// Copyright 2015 The Chromium Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

per http://www.chromium.org/blink/coding-style#TOC-License

https://codereview.chromium.org/793493003/diff/20001/public/web/WebSandboxFla...
public/web/WebSandboxFlags.h:36: enum WebSandboxFlags : int {
How inconvenient would it be to make this a strongly typed enum, e.g. enum class
WebSandboxFlags?

[alexmos, 2015-01-09 19:49:57]

Responses below, and I also rebased to properly include Nate's changes.

https://codereview.chromium.org/793493003/diff/20001/Source/web/WebRemoteFram...
File Source/web/WebRemoteFrameImpl.cpp (right):

https://codereview.chromium.org/793493003/diff/20001/Source/web/WebRemoteFram...
Source/web/WebRemoteFrameImpl.cpp:809: OwnPtr<RemoteBridgeFrameOwner> owner =
RemoteBridgeFrameOwner::create(child);
On 2015/01/08 22:50:45, dcheng wrote:
> OwnPtrWillBeRawPtr

Done.

https://codereview.chromium.org/793493003/diff/20001/public/web/WebSandboxFla...
File public/web/WebSandboxFlags.h (right):

https://codereview.chromium.org/793493003/diff/20001/public/web/WebSandboxFla...
public/web/WebSandboxFlags.h:26: #ifndef WebSandboxFlags_h
On 2015/01/08 22:50:45, dcheng wrote:
> I'd probably reupload this with --no-find-copies.
> 
> Also, headers for new files should be:
> // Copyright 2015 The Chromium Authors. All rights reserved.
> // Use of this source code is governed by a BSD-style license that can be
> // found in the LICENSE file.
> 
> per http://www.chromium.org/blink/coding-style#TOC-License

Done.  Thanks for the pointers!

https://codereview.chromium.org/793493003/diff/20001/public/web/WebSandboxFla...
public/web/WebSandboxFlags.h:36: enum WebSandboxFlags : int {
On 2015/01/08 22:50:45, dcheng wrote:
> How inconvenient would it be to make this a strongly typed enum, e.g. enum
class
> WebSandboxFlags?

Done.  I'm all for it, especially since I already used "enum class" in my
chromium-side CL.

[dcheng, 2015-01-09 22:46:12]

https://codereview.chromium.org/793493003/diff/60001/Source/web/WebRemoteFram...
File Source/web/WebRemoteFrameImpl.cpp (right):

https://codereview.chromium.org/793493003/diff/60001/Source/web/WebRemoteFram...
Source/web/WebRemoteFrameImpl.cpp:51: SandboxFlags m_sandboxFlags;
Since this can't change after creation, let's mark this const and initialize it
in the constructor.

[alexmos, 2015-01-09 23:12:30]

https://codereview.chromium.org/793493003/diff/60001/Source/web/WebRemoteFram...
File Source/web/WebRemoteFrameImpl.cpp (right):

https://codereview.chromium.org/793493003/diff/60001/Source/web/WebRemoteFram...
Source/web/WebRemoteFrameImpl.cpp:51: SandboxFlags m_sandboxFlags;
On 2015/01/09 22:46:12, dcheng wrote:
> Since this can't change after creation, let's mark this const and initialize
it
> in the constructor.

As we just discussed, the flags are mutable, with updates taking effect on next
navigation, so we may need to update them after creation.  I'm planning to add
the plumbing to handle dynamic updates in a later CL.

[dcheng, 2015-01-10 00:01:00]

lgtm with nits

https://codereview.chromium.org/793493003/diff/60001/Source/web/WebRemoteFram...
File Source/web/WebRemoteFrameImpl.cpp (right):

https://codereview.chromium.org/793493003/diff/60001/Source/web/WebRemoteFram...
Source/web/WebRemoteFrameImpl.cpp:78: m_sandboxFlags = flags;
Nit: just inline this into the declaration.

https://codereview.chromium.org/793493003/diff/60001/Source/web/WebRemoteFram...
Source/web/WebRemoteFrameImpl.cpp:814:
OwnPtrWillBeRawPtr<RemoteBridgeFrameOwner> owner =
RemoteBridgeFrameOwner::create(child);
Nit: it's probably still nicer to pass this in the constructor, since we have to
initialize it anyway.

[alexmos, 2015-01-10 00:51:54]

https://codereview.chromium.org/793493003/diff/60001/Source/web/WebRemoteFram...
File Source/web/WebRemoteFrameImpl.cpp (right):

https://codereview.chromium.org/793493003/diff/60001/Source/web/WebRemoteFram...
Source/web/WebRemoteFrameImpl.cpp:78: m_sandboxFlags = flags;
On 2015/01/10 00:01:00, dcheng wrote:
> Nit: just inline this into the declaration.

Done.  Inlined setSandboxFlags(), sandboxFlags(), and isLocal().

https://codereview.chromium.org/793493003/diff/60001/Source/web/WebRemoteFram...
Source/web/WebRemoteFrameImpl.cpp:814:
OwnPtrWillBeRawPtr<RemoteBridgeFrameOwner> owner =
RemoteBridgeFrameOwner::create(child);
On 2015/01/10 00:01:00, dcheng wrote:
> Nit: it's probably still nicer to pass this in the constructor, since we have
to
> initialize it anyway.

Done.

[alexmos, 2015-01-10 00:56:49]

alexmos@chromium.org changed reviewers:
+ japhet@chromium.org

[alexmos, 2015-01-10 00:56:49]

Nate, could you please review this for Source/web owner approval?

[alexmos, 2015-01-14 17:48:54]

On 2015/01/10 00:56:49, alexmos wrote:
> Nate, could you please review this for Source/web owner approval?

Hi Nate - friendly ping.

[Nate Chapin, 2015-01-14 18:14:54]

lgtm

[alexmos, 2015-01-14 18:17:13]

The CQ bit was checked by alexmos@chromium.org

[commit-bot: I haz the power, 2015-01-14 18:17:31]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/793493003/80001

[commit-bot: I haz the power, 2015-01-14 18:53:33]

Committed patchset #5 (id:80001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=188409