Eliminate kMojoChannelToken usage for render processes

Eliminate kMojoChannelToken usage for render processes

Changes RenderProcessHostImpl to connect its IPC Channel
by requesting a nominal IPC::mojom::ChannelBootstrap
interface from its shell::Connection to the render process.

Changes ChildThreadImpl to support binding such requests
by fusing them to its local IPC Channel if the (now deprecated)
kMojoChannelToken was not provided.

BUG=623396
[email protected]

Committed: https://crrev.com/7a7f8f7e6977500d09270ff9d9b3817dda5e5b8a
Committed: https://crrev.com/ff1b853cb6bfc3d5402fc970b2d856cbe73418cc
Cr-Original-Commit-Position: refs/heads/master@{#410656}
Cr-Commit-Position: refs/heads/master@{#410699}

[Ken Rockot(use gerrit already), 2016-08-03 20:41:27]

The CQ bit was checked by [email protected] to run a CQ dry run

[commit-bot: I haz the power, 2016-08-03 20:42:07]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Ken Rockot(use gerrit already), 2016-08-03 20:57:42]

The CQ bit was checked by [email protected] to run a CQ dry run

[commit-bot: I haz the power, 2016-08-03 20:58:20]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Ken Rockot(use gerrit already), 2016-08-03 21:04:29]

[email protected] changed reviewers:
+ [email protected]

[Ken Rockot(use gerrit already), 2016-08-03 21:04:30]

Also +tsepez@ for trivial mojom change

[Tom Sepez, 2016-08-03 21:09:50]

Description was changed from

==========
Eliminate kMojoChannelToken usage for render processes

Changes RenderProcessHostImpl to connect its IPC Channel
by requesting a nominal IPC::mojom::ChannelBootstrap
interface from its shell::Connection to the render process.

Changes ChildThreadImpl to support binding such requests
by fusing them to its local IPC Channel if the (now deprecated)
kMojoChannelToken was not provided.

BUG=623396
[email protected]
==========

to

==========
Eliminate kMojoChannelToken usage for render processes

Changes RenderProcessHostImpl to connect its IPC Channel
by requesting a nominal IPC::mojom::ChannelBootstrap
interface from its shell::Connection to the render process.

Changes ChildThreadImpl to support binding such requests
by fusing them to its local IPC Channel if the (now deprecated)
kMojoChannelToken was not provided.

BUG=623396
[email protected]
==========

[Tom Sepez, 2016-08-03 21:09:50]

[email protected] changed reviewers:
+ [email protected]

[Ben Goodger (Google), 2016-08-03 21:10:39]

lgtm !!

[Tom Sepez, 2016-08-03 21:11:21]

On 2016/08/03 21:04:30, Ken Rockot wrote:
> Also +tsepez@ for trivial mojom change

My hazy memory is that this token was important on windows to prevent evil
processes from stealing each other's pipes during startup.  Adding jschuh who
may have more windows context.

[Ben Goodger (Google), 2016-08-03 21:12:21]

On 2016/08/03 21:11:21, Tom Sepez wrote:
> On 2016/08/03 21:04:30, Ken Rockot wrote:
> > Also +tsepez@ for trivial mojom change
> 
> My hazy memory is that this token was important on windows to prevent evil
> processes from stealing each other's pipes during startup.  Adding jschuh who
> may have more windows context.

There's still a token for the service request channel though...

[Tom Sepez, 2016-08-03 21:16:13]

On 2016/08/03 21:12:21, Ben Goodger (Google) wrote:
> On 2016/08/03 21:11:21, Tom Sepez wrote:
> > On 2016/08/03 21:04:30, Ken Rockot wrote:
> > > Also +tsepez@ for trivial mojom change
> > 
> > My hazy memory is that this token was important on windows to prevent evil
> > processes from stealing each other's pipes during startup.  Adding jschuh
who
> > may have more windows context.
> 
> There's still a token for the service request channel though...

Ok, that's good. So why did we need the random token here?

[Ken Rockot(use gerrit already), 2016-08-03 21:16:21]

On 2016/08/03 at 21:11:21, tsepez wrote:
> On 2016/08/03 21:04:30, Ken Rockot wrote:
> > Also +tsepez@ for trivial mojom change
> 
> My hazy memory is that this token was important on windows to prevent evil
processes from stealing each other's pipes during startup.  Adding jschuh who
may have more windows context.

Different kind of token I think - you're probably thinking of the token used by
legacy Channel implementations to achieve that purpose.

These tokens are just a way of connecting message pipes across processes using
the mojo::edk::Create[Parent/Child]MessagePipe API, and there's really no
security motivation to use them. They're large random numbers to make them
easier to use, but the parent process already knows which children have which
tokens.

[Tom Sepez, 2016-08-03 21:18:19]

On 2016/08/03 21:16:21, Ken Rockot wrote:
> On 2016/08/03 at 21:11:21, tsepez wrote:
> > On 2016/08/03 21:04:30, Ken Rockot wrote:
> > > Also +tsepez@ for trivial mojom change
> > 
> > My hazy memory is that this token was important on windows to prevent evil
> processes from stealing each other's pipes during startup.  Adding jschuh who
> may have more windows context.
> 
> Different kind of token I think - you're probably thinking of the token used
by
> legacy Channel implementations to achieve that purpose.
> 
> These tokens are just a way of connecting message pipes across processes using
> the mojo::edk::Create[Parent/Child]MessagePipe API, and there's really no
> security motivation to use them. They're large random numbers to make them
> easier to use, but the parent process already knows which children have which
> tokens.

Great, LGTM.

[Ken Rockot(use gerrit already), 2016-08-03 21:22:23]

The CQ bit was unchecked by [email protected]

[Ken Rockot(use gerrit already), 2016-08-03 21:22:24]

The CQ bit was checked by [email protected]

[commit-bot: I haz the power, 2016-08-03 21:22:47]

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, 2016-08-03 23:23:03]

The CQ bit was unchecked by [email protected]

[commit-bot: I haz the power, 2016-08-03 23:23:04]

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

[Ken Rockot(use gerrit already), 2016-08-04 17:38:24]

The CQ bit was checked by [email protected] to run a CQ dry run

[commit-bot: I haz the power, 2016-08-04 17:38:44]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Ken Rockot(use gerrit already), 2016-08-04 17:49:12]

The CQ bit was checked by [email protected] to run a CQ dry run

[commit-bot: I haz the power, 2016-08-04 17:49:39]

Dry run: 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, 2016-08-04 19:10:58]

The CQ bit was unchecked by [email protected]

[commit-bot: I haz the power, 2016-08-04 19:11:00]

Dry run: Try jobs failed on following builders:
  linux_chromium_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[Ken Rockot(use gerrit already), 2016-08-08 20:40:44]

The CQ bit was checked by [email protected] to run a CQ dry run

[commit-bot: I haz the power, 2016-08-08 20:41:15]

Dry run: 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, 2016-08-08 21:49:10]

The CQ bit was unchecked by [email protected]

[commit-bot: I haz the power, 2016-08-08 21:49:11]

Dry run: This issue passed the CQ dry run.

[Ken Rockot(use gerrit already), 2016-08-09 14:05:41]

The CQ bit was checked by [email protected]

[Ken Rockot(use gerrit already), 2016-08-09 14:05:42]

The patchset sent to the CQ was uploaded after l-g-t-m from [email protected],
[email protected]
Link to the patchset: https://codereview.chromium.org/2201183005/#ps80001
(title: "rebase")

[commit-bot: I haz the power, 2016-08-09 14:06:04]

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, 2016-08-09 14:10:00]

Description was changed from

==========
Eliminate kMojoChannelToken usage for render processes

Changes RenderProcessHostImpl to connect its IPC Channel
by requesting a nominal IPC::mojom::ChannelBootstrap
interface from its shell::Connection to the render process.

Changes ChildThreadImpl to support binding such requests
by fusing them to its local IPC Channel if the (now deprecated)
kMojoChannelToken was not provided.

BUG=623396
[email protected]
==========

to

==========
Eliminate kMojoChannelToken usage for render processes

Changes RenderProcessHostImpl to connect its IPC Channel
by requesting a nominal IPC::mojom::ChannelBootstrap
interface from its shell::Connection to the render process.

Changes ChildThreadImpl to support binding such requests
by fusing them to its local IPC Channel if the (now deprecated)
kMojoChannelToken was not provided.

BUG=623396
[email protected]
==========

[commit-bot: I haz the power, 2016-08-09 14:10:02]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2016-08-09 14:11:37]

Description was changed from

==========
Eliminate kMojoChannelToken usage for render processes

Changes RenderProcessHostImpl to connect its IPC Channel
by requesting a nominal IPC::mojom::ChannelBootstrap
interface from its shell::Connection to the render process.

Changes ChildThreadImpl to support binding such requests
by fusing them to its local IPC Channel if the (now deprecated)
kMojoChannelToken was not provided.

BUG=623396
[email protected]
==========

to

==========
Eliminate kMojoChannelToken usage for render processes

Changes RenderProcessHostImpl to connect its IPC Channel
by requesting a nominal IPC::mojom::ChannelBootstrap
interface from its shell::Connection to the render process.

Changes ChildThreadImpl to support binding such requests
by fusing them to its local IPC Channel if the (now deprecated)
kMojoChannelToken was not provided.

BUG=623396
[email protected]

Committed: https://crrev.com/7a7f8f7e6977500d09270ff9d9b3817dda5e5b8a
Cr-Commit-Position: refs/heads/master@{#410656}
==========

[commit-bot: I haz the power, 2016-08-09 14:11:39]

Patchset 5 (id:??) landed as
https://crrev.com/7a7f8f7e6977500d09270ff9d9b3817dda5e5b8a
Cr-Commit-Position: refs/heads/master@{#410656}

[huangs, 2016-08-09 14:19:03]

[email protected] changed reviewers:
+ [email protected]

[huangs, 2016-08-09 14:19:05]

This is causing failure in Webkit Mac Builder.  Rolling back.

[huangs, 2016-08-09 14:20:12]

A revert of this CL (patchset #5 id:80001) has been created in
https://codereview.chromium.org/2222373002/ by [email protected].

The reason for reverting is: Compile error in Webkit Mac Builder.

child_thread_impl.cc:272:47: error: non-virtual member function marked
'override' hides virtual member function
                 shell::Connector* connector) override {
.

[Ken Rockot(use gerrit already), 2016-08-09 14:21:17]

A revert of this CL (patchset #5 id:80001) has been created in
https://codereview.chromium.org/2227053002/ by [email protected].

The reason for reverting is: Needs rebase.

[findit-for-me, 2016-08-09 14:29:36]

FYI: Findit try jobs (rerunning failed compile or tests) identified this CL
at revision 410656 as the culprit for failures in the build cycles as shown on:
https://findit-for-me.appspot.com/waterfall/culprit?key=ag9zfmZpbmRpdC1mb3Itb...

[Ken Rockot(use gerrit already), 2016-08-09 15:17:27]

The CQ bit was checked by [email protected]

[Ken Rockot(use gerrit already), 2016-08-09 15:17:28]

The patchset sent to the CQ was uploaded after l-g-t-m from [email protected],
[email protected]
Link to the patchset: https://codereview.chromium.org/2201183005/#ps100001
(title: "rebase")

[commit-bot: I haz the power, 2016-08-09 15:17:46]

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, 2016-08-09 16:30:45]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2016-08-09 16:38:58]

Description was changed from

==========
Eliminate kMojoChannelToken usage for render processes

Changes RenderProcessHostImpl to connect its IPC Channel
by requesting a nominal IPC::mojom::ChannelBootstrap
interface from its shell::Connection to the render process.

Changes ChildThreadImpl to support binding such requests
by fusing them to its local IPC Channel if the (now deprecated)
kMojoChannelToken was not provided.

BUG=623396
[email protected]

Committed: https://crrev.com/7a7f8f7e6977500d09270ff9d9b3817dda5e5b8a
Cr-Commit-Position: refs/heads/master@{#410656}
==========

to

==========
Eliminate kMojoChannelToken usage for render processes

Changes RenderProcessHostImpl to connect its IPC Channel
by requesting a nominal IPC::mojom::ChannelBootstrap
interface from its shell::Connection to the render process.

Changes ChildThreadImpl to support binding such requests
by fusing them to its local IPC Channel if the (now deprecated)
kMojoChannelToken was not provided.

BUG=623396
[email protected]

Committed: https://crrev.com/7a7f8f7e6977500d09270ff9d9b3817dda5e5b8a
Committed: https://crrev.com/ff1b853cb6bfc3d5402fc970b2d856cbe73418cc
Cr-Original-Commit-Position: refs/heads/master@{#410656}
Cr-Commit-Position: refs/heads/master@{#410699}
==========

[commit-bot: I haz the power, 2016-08-09 16:39:00]

Patchset 6 (id:??) landed as
https://crrev.com/ff1b853cb6bfc3d5402fc970b2d856cbe73418cc
Cr-Commit-Position: refs/heads/master@{#410699}

[huangs, 2016-08-09 19:08:44]

Looks like this fails mac now?
https://build.chromium.org/p/chromium.win/builders/Win%20x64%20Builder/builds...

FAILED: obj/content/child/child/child_thread_impl.obj 
ninja -t msvc -e environment.x64 -- C:\b\c\cipd\goma/gomacc.exe
"C:\b\depot_tools\win_toolchain\vs_files\95ddda401ec5678f15eeed01d2bee08fcbc5ee97\VC\bin\amd64/cl.exe"
/nologo /showIncludes /FC @obj/content/child/child/child_thread_impl.obj.rsp /c
../../content/child/child_thread_impl.cc
/Foobj/content/child/child/child_thread_impl.obj
/Fd"obj/content/child/child_cc.pdb"
c:\b\c\b\win\src\content\child\child_thread_impl.cc(271): error C3668:
'content::`anonymous-namespace'::ChannelBootstrapFilter::OnConnect': method with
override specifier 'override' did not override any base class methods
c:\b\c\b\win\src\content\child\child_thread_impl.cc(276): error C2039:
'AddInterface': is not a member of 'shell::Connection'

[Ken Rockot(use gerrit already), 2016-08-10 17:01:00]

A revert of this CL (patchset #6 id:100001) has been created in
https://codereview.chromium.org/2231133002/ by [email protected].

The reason for reverting is: Speculative revert for failures on
https://build.chromium.org/p/chromium.win/builders/Win7%20Tests%20%281%29 -
can't repro locally but wanna see if this greens the tests..