service worker: Attempt to send WorkerStopped before ProviderDestroyed.

service worker: Attempt to send WorkerStopped before ProviderDestroyed.

Previously in IPC when a worker stops, the renderer sends the browser a
WorkerStopped message followed by a ProviderDestroyed message. This
ordering got reversed in https://crrev.com/6ea839368fd4dbf, causing an
unexpected case where a ServiceWorkerProviderHost is destroyed,
releasing a ServiceWorkerVersion before OnStopped can be invoked on it.
This patch returns to the previous ordering.

However even with this patch, for Mojo the ordering is still not
guaranteed, since the pipes are different. Since the IPC land is no
longer the default, the real fix is to figure out how to do this in
Mojo. This patch is just a mitigation, so no tests are added yet.

BUG=629701, 676526
Committed: https://crrev.com/af96c9a9e8d491929f32f2e693fd61fd1b93557d
Cr-Commit-Position: refs/heads/master@{#440356}

[falken, 2016-12-22 06:12:55]

falken@chromium.org changed reviewers:
+ shimazu@chromium.org

[falken, 2016-12-22 06:12:55]

ptal

[falken, 2016-12-22 06:13:07]

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

[commit-bot: I haz the power, 2016-12-22 06:13:27]

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

[shimazu, 2016-12-22 07:09:40]

lgtm, thanks!

[falken, 2016-12-22 07:13:56]

The CQ bit was unchecked by falken@chromium.org

[falken, 2016-12-22 07:14:01]

The CQ bit was checked by falken@chromium.org

[commit-bot: I haz the power, 2016-12-22 07:14:15]

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, 2016-12-22 07:48:09]

CQ is committing da patch.

Bot data: {"patchset_id": 1, "attempt_start_ts": 1482390841939440, "parent_rev":
"466210b333690d00d46be2289e74eae85d2880c8", "commit_rev":
"90c4acd93acd20c401ce7390d1aacea11ea1ff47"}

[commit-bot: I haz the power, 2016-12-22 07:48:36]

Description was changed from

==========
service worker: Attempt to send WorkerStopped before ProviderDestroyed.

Previously in IPC when a worker stops, the renderer sends the browser a
WorkerStopped message followed by a ProviderDestroyed message. This
ordering got reversed in https://crrev.com/6ea839368fd4dbf, causing an
unexpected case where a ServiceWorkerProviderHost is destroyed,
releasing a ServiceWorkerVersion before OnStopped can be invoked on it.
This patch returns to the previous ordering.

However even with this patch, for Mojo the ordering is still not
guaranteed, since the pipes are different. Since the IPC land is no
longer the default, the real fix is to figure out how to do this in
Mojo. This patch is just a mitigation, so no tests are added yet.

BUG=629701,676526
==========

to

==========
service worker: Attempt to send WorkerStopped before ProviderDestroyed.

Previously in IPC when a worker stops, the renderer sends the browser a
WorkerStopped message followed by a ProviderDestroyed message. This
ordering got reversed in https://crrev.com/6ea839368fd4dbf, causing an
unexpected case where a ServiceWorkerProviderHost is destroyed,
releasing a ServiceWorkerVersion before OnStopped can be invoked on it.
This patch returns to the previous ordering.

However even with this patch, for Mojo the ordering is still not
guaranteed, since the pipes are different. Since the IPC land is no
longer the default, the real fix is to figure out how to do this in
Mojo. This patch is just a mitigation, so no tests are added yet.

BUG=629701,676526

Review-Url: https://codereview.chromium.org/2596153002
==========

[commit-bot: I haz the power, 2016-12-22 07:48:38]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-12-22 07:50:36]

Description was changed from

==========
service worker: Attempt to send WorkerStopped before ProviderDestroyed.

Previously in IPC when a worker stops, the renderer sends the browser a
WorkerStopped message followed by a ProviderDestroyed message. This
ordering got reversed in https://crrev.com/6ea839368fd4dbf, causing an
unexpected case where a ServiceWorkerProviderHost is destroyed,
releasing a ServiceWorkerVersion before OnStopped can be invoked on it.
This patch returns to the previous ordering.

However even with this patch, for Mojo the ordering is still not
guaranteed, since the pipes are different. Since the IPC land is no
longer the default, the real fix is to figure out how to do this in
Mojo. This patch is just a mitigation, so no tests are added yet.

BUG=629701,676526

Review-Url: https://codereview.chromium.org/2596153002
==========

to

==========
service worker: Attempt to send WorkerStopped before ProviderDestroyed.

Previously in IPC when a worker stops, the renderer sends the browser a
WorkerStopped message followed by a ProviderDestroyed message. This
ordering got reversed in https://crrev.com/6ea839368fd4dbf, causing an
unexpected case where a ServiceWorkerProviderHost is destroyed,
releasing a ServiceWorkerVersion before OnStopped can be invoked on it.
This patch returns to the previous ordering.

However even with this patch, for Mojo the ordering is still not
guaranteed, since the pipes are different. Since the IPC land is no
longer the default, the real fix is to figure out how to do this in
Mojo. This patch is just a mitigation, so no tests are added yet.

BUG=629701,676526

Committed: https://crrev.com/af96c9a9e8d491929f32f2e693fd61fd1b93557d
Cr-Commit-Position: refs/heads/master@{#440356}
==========

[commit-bot: I haz the power, 2016-12-22 07:50:37]

Patchset 1 (id:??) landed as
https://crrev.com/af96c9a9e8d491929f32f2e693fd61fd1b93557d
Cr-Commit-Position: refs/heads/master@{#440356}

[falken, 2016-12-26 08:52:08]

Patchset #2 (id:20001) has been deleted