Use a raw pointer instead of Singleton for PnaclHost and leak it.

Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead
of the IO thread it's bound to) with live WeakPtrs.

Issue 631547 results in random flakes for all browser_tests exercising this
code one way or another, it should be fixed by this.

The WeakPtrs served no purpose as the task running system is long gone by the time
static destruction kicks in and no task would thus ever have been cancelled through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - Replace callback_forward.h include by callback.h (typedefs need it
   and were getting it transitively through singleton.h->at_exit.h).
 - The attempted un-initialization through DeInitIfSafe() remains the same.

BUG=631547
Committed: https://crrev.com/459e9aa8dd4682aaaf5b8bcbb0fd030632b2d787
Cr-Commit-Position: refs/heads/master@{#409365}

[gab, 2016-08-02 19:46:38]

The CQ bit was checked by gab@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-02 19:47:18]

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

[gab, 2016-08-02 19:53:27]

Description was changed from

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

BUG=631547
==========

to

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead

of the IO thread it's bound to) with live WeakPtrs.

The WeakPtrs served no purpose as the task running system is long gone by the
time
static destruction kicks in and no task would thus ever have been cancelled
through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - The attempted uninitialization logic through DeInitIfSafe() remains the same.

BUG=631547
==========

[gab, 2016-08-02 19:55:29]

The CQ bit was checked by gab@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-02 19:55:35]

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, 2016-08-02 20:04:20]

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

[commit-bot: I haz the power, 2016-08-02 20:04:21]

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

[gab, 2016-08-02 20:16:50]

The CQ bit was checked by gab@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-02 20:17:32]

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

[gab, 2016-08-02 20:18:27]

Description was changed from

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead

of the IO thread it's bound to) with live WeakPtrs.

The WeakPtrs served no purpose as the task running system is long gone by the
time
static destruction kicks in and no task would thus ever have been cancelled
through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - The attempted uninitialization logic through DeInitIfSafe() remains the same.

BUG=631547
==========

to

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead

of the IO thread it's bound to) with live WeakPtrs.

The WeakPtrs served no purpose as the task running system is long gone by the
time
static destruction kicks in and no task would thus ever have been cancelled
through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - Replace callback_forward.h include by callback.h (typedefs need it
   and were getting it transitively through singleton.h->at_exit.h).
 - The attempted un-initialization through DeInitIfSafe() remains the same.

BUG=631547
==========

[gab, 2016-08-02 21:02:07]

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

[gab, 2016-08-02 21:02:08]

@mseaborn: PTAL, this addresses waterfall flakes caused by PNaclHost.

[commit-bot: I haz the power, 2016-08-02 21:13:17]

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

[commit-bot: I haz the power, 2016-08-02 21:13:18]

Dry run: This issue passed the CQ dry run.

[Mark Seaborn, 2016-08-02 21:47:32]

> @mseaborn: PTAL, this addresses waterfall flakes caused by PNaclHost.

Can you mention in the commit message that this fixes a flaky failure?  Thanks
-- LGTM.

https://codereview.chromium.org/2207683002/diff/60001/components/nacl/browser...
File components/nacl/browser/pnacl_host.h (right):

https://codereview.chromium.org/2207683002/diff/60001/components/nacl/browser...
components/nacl/browser/pnacl_host.h:47: // time, it will be too late when
AtExitManager kicks in anways so subscribing
Typo: "anyway"

[gab, 2016-08-02 21:56:10]

Description was changed from

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead

of the IO thread it's bound to) with live WeakPtrs.

The WeakPtrs served no purpose as the task running system is long gone by the
time
static destruction kicks in and no task would thus ever have been cancelled
through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - Replace callback_forward.h include by callback.h (typedefs need it
   and were getting it transitively through singleton.h->at_exit.h).
 - The attempted un-initialization through DeInitIfSafe() remains the same.

BUG=631547
==========

to

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead

of the IO thread it's bound to) with live WeakPtrs.

Issue 631547 results in random flakes for all browser_tests exercising this
code somehow, it should be fixed by this.

The WeakPtrs served no purpose as the task running system is long gone by the
time
static destruction kicks in and no task would thus ever have been cancelled
through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - Replace callback_forward.h include by callback.h (typedefs need it
   and were getting it transitively through singleton.h->at_exit.h).
 - The attempted un-initialization through DeInitIfSafe() remains the same.

BUG=631547
==========

[gab, 2016-08-02 21:56:32]

Description was changed from

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead

of the IO thread it's bound to) with live WeakPtrs.

Issue 631547 results in random flakes for all browser_tests exercising this
code somehow, it should be fixed by this.

The WeakPtrs served no purpose as the task running system is long gone by the
time
static destruction kicks in and no task would thus ever have been cancelled
through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - Replace callback_forward.h include by callback.h (typedefs need it
   and were getting it transitively through singleton.h->at_exit.h).
 - The attempted un-initialization through DeInitIfSafe() remains the same.

BUG=631547
==========

to

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead

of the IO thread it's bound to) with live WeakPtrs.

Issue 631547 results in random flakes for all browser_tests exercising this
code one way or another, it should be fixed by this.

The WeakPtrs served no purpose as the task running system is long gone by the
time
static destruction kicks in and no task would thus ever have been cancelled
through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - Replace callback_forward.h include by callback.h (typedefs need it
   and were getting it transitively through singleton.h->at_exit.h).
 - The attempted un-initialization through DeInitIfSafe() remains the same.

BUG=631547
==========

[gab, 2016-08-02 21:57:42]

On 2016/08/02 21:47:32, Mark Seaborn wrote:
> > @mseaborn: PTAL, this addresses waterfall flakes caused by PNaclHost.
> 
> Can you mention in the commit message that this fixes a flaky failure?  Thanks
> -- LGTM.

Done, thanks.

https://codereview.chromium.org/2207683002/diff/60001/components/nacl/browser...
File components/nacl/browser/pnacl_host.h (right):

https://codereview.chromium.org/2207683002/diff/60001/components/nacl/browser...
components/nacl/browser/pnacl_host.h:47: // time, it will be too late when
AtExitManager kicks in anways so subscribing
On 2016/08/02 21:47:31, Mark Seaborn wrote:
> Typo: "anyway"

Done.

[gab, 2016-08-02 21:58:04]

The CQ bit was checked by gab@chromium.org

[gab, 2016-08-02 21:58:05]

The patchset sent to the CQ was uploaded after l-g-t-m from
mseaborn@chromium.org
Link to the patchset: https://codereview.chromium.org/2207683002/#ps80001
(title: "typo")

[commit-bot: I haz the power, 2016-08-02 21:58:47]

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, 2016-08-02 22:53:01]

Description was changed from

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead

of the IO thread it's bound to) with live WeakPtrs.

Issue 631547 results in random flakes for all browser_tests exercising this
code one way or another, it should be fixed by this.

The WeakPtrs served no purpose as the task running system is long gone by the
time
static destruction kicks in and no task would thus ever have been cancelled
through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - Replace callback_forward.h include by callback.h (typedefs need it
   and were getting it transitively through singleton.h->at_exit.h).
 - The attempted un-initialization through DeInitIfSafe() remains the same.

BUG=631547
==========

to

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead

of the IO thread it's bound to) with live WeakPtrs.

Issue 631547 results in random flakes for all browser_tests exercising this
code one way or another, it should be fixed by this.

The WeakPtrs served no purpose as the task running system is long gone by the
time
static destruction kicks in and no task would thus ever have been cancelled
through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - Replace callback_forward.h include by callback.h (typedefs need it
   and were getting it transitively through singleton.h->at_exit.h).
 - The attempted un-initialization through DeInitIfSafe() remains the same.

BUG=631547
==========

[commit-bot: I haz the power, 2016-08-02 22:53:03]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2016-08-02 22:55:19]

Description was changed from

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead

of the IO thread it's bound to) with live WeakPtrs.

Issue 631547 results in random flakes for all browser_tests exercising this
code one way or another, it should be fixed by this.

The WeakPtrs served no purpose as the task running system is long gone by the
time
static destruction kicks in and no task would thus ever have been cancelled
through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - Replace callback_forward.h include by callback.h (typedefs need it
   and were getting it transitively through singleton.h->at_exit.h).
 - The attempted un-initialization through DeInitIfSafe() remains the same.

BUG=631547
==========

to

==========
Use a raw pointer instead of Singleton for PnaclHost and leak it.

This CL was brought up by issue 631547 which highlights that PnaclHost is
sometimes deleted off-thread (on the main thread at static unitilization instead
of the IO thread it's bound to) with live WeakPtrs.

Issue 631547 results in random flakes for all browser_tests exercising this
code one way or another, it should be fixed by this.

The WeakPtrs served no purpose as the task running system is long gone by the
time
static destruction kicks in and no task would thus ever have been cancelled
through
them.

It already leaked its state on static destruction from AtExitManager.
It could instead have used base::LeakySingletonTraits but it really doesn't
need to even be a Singleton per being intended for single-threaded usage.

So this CL:
 - Replaces WeakPtrs usage with Unretained(this).
 - Adds missing thread_checker_ calls now that they will no longer be implicitly
   provided by GetWeakPtr() calls.
 - Replaces Singleton with a leaked static raw pointer.
 - Replace callback_forward.h include by callback.h (typedefs need it
   and were getting it transitively through singleton.h->at_exit.h).
 - The attempted un-initialization through DeInitIfSafe() remains the same.

BUG=631547

Committed: https://crrev.com/459e9aa8dd4682aaaf5b8bcbb0fd030632b2d787
Cr-Commit-Position: refs/heads/master@{#409365}
==========

[commit-bot: I haz the power, 2016-08-02 22:55:20]

Patchset 5 (id:??) landed as
https://crrev.com/459e9aa8dd4682aaaf5b8bcbb0fd030632b2d787
Cr-Commit-Position: refs/heads/master@{#409365}