[Device Sensors] Fix crash on shutdown

[Device Sensors] Fix crash on shutdown

DataFetcherSharedMemoryBase has a thread to which it posts tasks. It
currently stops this thread in its destructor. However, this behavior
is problematic: the tasks call pure virtual methods of
DataFetcherSharedMemoryBase, and stopping the thread ensures that all
pending tasks are run. The combination of these facts mean that code
gets run in the DataFetcherSharedMemoryBase that calls pure virtual
methods of that class, which naturally causes crashes (as the
subclass' vtable has been torn down at that point).

This CL fixes the problem by stopping the thread in
DataFetcherSharedMemoryBase::Shutdown(), at which point the subclass
instance is guaranteed to still be alive.

BUG=699496
Review-Url: https://codereview.chromium.org/2743493003
Cr-Commit-Position: refs/heads/master@{#455797}
Committed: https://chromium.googlesource.com/chromium/src/+/55462c0bffc90caf9c5e11bdd6a5d611f187330d

[blundell, 2017-03-09 04:25:52]

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

[commit-bot: I haz the power, 2017-03-09 04:26:09]

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, 2017-03-09 04:48:00]

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

[commit-bot: I haz the power, 2017-03-09 04:48:01]

Dry run: Try jobs failed on following builders:
  mac_chromium_rel_ng on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[blundell, 2017-03-09 15:39:16]

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

[commit-bot: I haz the power, 2017-03-09 15:39:31]

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, 2017-03-09 16:28:21]

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

[commit-bot: I haz the power, 2017-03-09 16:28:22]

Dry run: This issue passed the CQ dry run.

[blundell, 2017-03-09 16:31:40]

blundell@chromium.org changed reviewers:
+ timvolodine@chromium.org

[timvolodine, 2017-03-09 18:19:23]

Yes most definitely, thanks for fixing this. How embarrassing, calling pure
virtuals from destructors is not a good idea ). I guess we didn't see this crash
before because of a different destruct/shutdown sequence..

lgtm!

[blundell, 2017-03-09 18:21:20]

The CQ bit was checked by blundell@chromium.org

[commit-bot: I haz the power, 2017-03-09 18:22:01]

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, 2017-03-09 18:26:20]

CQ is committing da patch.

Bot data: {"patchset_id": 40001, "attempt_start_ts": 1489083680218430,
"parent_rev": "7fae73cb6d6bef3c2c8246c867961ed33e4eed57", "commit_rev":
"55462c0bffc90caf9c5e11bdd6a5d611f187330d"}

[commit-bot: I haz the power, 2017-03-09 18:27:02]

Description was changed from

==========
[Device Sensors] Fix crash on shutdown

DataFetcherSharedMemoryBase has a thread to which it posts tasks. It
currently stops this thread in its destructor. However, this behavior
is problematic: the tasks call pure virtual methods of
DataFetcherSharedMemoryBase, and stopping the thread ensures that all
pending tasks are run. The combination of these facts mean that code
gets run in the DataFetcherSharedMemoryBase that calls pure virtual
methods of that class, which naturally causes crashes (as the
subclass' vtable has been torn down at that point).

This CL fixes the problem by stopping the thread in
DataFetcherSharedMemoryBase::Shutdown(), at which point the subclass
instance is guaranteed to still be alive.

BUG=699496
==========

to

==========
[Device Sensors] Fix crash on shutdown

DataFetcherSharedMemoryBase has a thread to which it posts tasks. It
currently stops this thread in its destructor. However, this behavior
is problematic: the tasks call pure virtual methods of
DataFetcherSharedMemoryBase, and stopping the thread ensures that all
pending tasks are run. The combination of these facts mean that code
gets run in the DataFetcherSharedMemoryBase that calls pure virtual
methods of that class, which naturally causes crashes (as the
subclass' vtable has been torn down at that point).

This CL fixes the problem by stopping the thread in
DataFetcherSharedMemoryBase::Shutdown(), at which point the subclass
instance is guaranteed to still be alive.

BUG=699496

Review-Url: https://codereview.chromium.org/2743493003
Cr-Commit-Position: refs/heads/master@{#455797}
Committed:
https://chromium.googlesource.com/chromium/src/+/55462c0bffc90caf9c5e11bdd6a5...
==========

[commit-bot: I haz the power, 2017-03-09 18:27:02]

Committed patchset #3 (id:40001) as
https://chromium.googlesource.com/chromium/src/+/55462c0bffc90caf9c5e11bdd6a5...