Fix use after free in SearchProviderInstallData GoogleURLObserver.

Fix use after free in SearchProviderInstallData GoogleURLObserver.

Follow-up to r246358.

BUG=338341
R=avi@chromium.org
TBR=pkasting@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=247898

[Peter Lundblad, 2014-01-29 00:12:37]

Here is a dump (from a trybot run of some tests I am trying to run when I
stumbled on this):

http://build.chromium.org/p/tryserver.chromium/builders/linux_chromeos_asan/b...

[Avi (use Gerrit), 2014-01-29 04:54:45]

On 2014/01/29 00:12:37, Peter Lundblad wrote:
> Here is a dump

You can repro this!?!

The bug is http://crbug.com/338341 . The problem with your fix is that it
shouldn't be necessary; the RenderProcessHostImpl should be dead after the
RenderProcessHostDestroyed callback and there should never be another call.

Someone is delaying a call to RenderProcessHostImpl::ProcessDied :

     #0 0x7f254937acc8 in content::RenderProcessHostImpl::ProcessDied(bool)
src/content/browser/renderer_host/render_process_host_impl.cc:1831
    #1 0x7f2543988547 in base::internal::InvokeHelper<false, void,
base::internal::RunnableAdapter<void (IPC::ChannelProxy::Context::*)()>, void
(IPC::ChannelProxy::Context*
const&)>::MakeItSo(base::internal::RunnableAdapter<void
(IPC::ChannelProxy::Context::*)()>, IPC::ChannelProxy::Context* const&)
src/base/bind_internal.h:871
    #2 0x7f2542b82df9 in base::MessageLoop::RunTask(base::PendingTask const&)
src/base/message_loop/message_loop.cc:448

If you can reproduce this, can you help me debug this? Can you find out who is
putting that ProcessDied task on the MessageLoop?

[Avi (use Gerrit), 2014-01-29 16:14:26]

Let's commit this bandaid, merge it, and I'll keep looking.

[Avi (use Gerrit), 2014-01-29 16:15:11]

On 2014/01/29 16:14:26, Avi wrote:
> Let's commit this bandaid, merge it, and I'll keep looking.

LGTM for now.

[commit-bot: I haz the power, 2014-01-29 16:17:46]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/plundblad@chromium.org/137833009/1

[commit-bot: I haz the power, 2014-01-29 17:01:43]

Retried try job too often on chromium_presubmit for step(s) presubmit
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=chromium_p...

[commit-bot: I haz the power, 2014-01-29 17:04:28]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/plundblad@chromium.org/137833009/1

[commit-bot: I haz the power, 2014-01-30 08:38:26]

Change committed as 247898

[Peter Kasting, 2014-02-04 02:07:30]

On 2014/01/29 16:14:26, Avi wrote:
> Let's commit this bandaid, merge it, and I'll keep looking.

(Random note, maybe there should have been a TODO added on the bandaid noting
its bandaid status...)