[sql] Serialize calls to sqlite3_initialize().

[sql] Serialize calls to sqlite3_initialize().

sqlite3_initialize() uses double-check locking, and is thus prone to
data races.

BUG=248101
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=212984

[Alexander Potapenko, 2013-07-12 12:35:44]

This won't fix the race in
https://code.google.com/p/chromium/issues/detail?id=248101#c5, since one of the
memory accesses happens outside sqlite3_initialize()

https://codereview.chromium.org/19000007/diff/1/sql/sql.gyp
File sql/sql.gyp (right):

https://codereview.chromium.org/19000007/diff/1/sql/sql.gyp#newcode16
sql/sql.gyp:16:
'../base/third_party/dynamic_annotations/dynamic_annotations.gyp:dynamic_annotations',
Why do you need a dependency on the dynamic annotations here?

[Scott Hess - ex-Googler, 2013-07-12 17:29:26]

On 2013/07/12 12:35:44, Alexander Potapenko wrote:
> This won't fix the race in
> https://code.google.com/p/chromium/issues/detail?id=248101#c5, since one of
the
> memory accesses happens outside sqlite3_initialize()

It will prevent sqlite3_open() from being called before sqlite3_initialize() is
completed.

https://codereview.chromium.org/19000007/diff/1/sql/sql.gyp
File sql/sql.gyp (right):

https://codereview.chromium.org/19000007/diff/1/sql/sql.gyp#newcode16
sql/sql.gyp:16:
'../base/third_party/dynamic_annotations/dynamic_annotations.gyp:dynamic_annotations',
On 2013/07/12 12:35:44, Alexander Potapenko wrote:
> Why do you need a dependency on the dynamic annotations here?

It's necessary for the lazy initializer.  I think, I don't fully understand why.
 sql/ had never used a lazy initializer before.

[Scott Hess - ex-Googler, 2013-07-19 20:24:42]

On 2013/07/12 17:29:26, shess wrote:
> On 2013/07/12 12:35:44, Alexander Potapenko wrote:
> > This won't fix the race in
> > https://code.google.com/p/chromium/issues/detail?id=248101#c5, since one of
> the
> > memory accesses happens outside sqlite3_initialize()
> 
> It will prevent sqlite3_open() from being called before sqlite3_initialize()
is
> completed.

Did this make sense?

[Alexander Potapenko, 2013-07-22 15:37:12]

I've double-checked that your CL removes the data races in sqlite3_initialize()
LGTM

[commit-bot: I haz the power, 2013-07-22 18:32:05]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/shess@chromium.org/19000007/13001

[commit-bot: I haz the power, 2013-07-22 19:08:19]

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

[Scott Hess - ex-Googler, 2013-07-22 19:24:58]

On 2013/07/22 19:08:19, I haz the power (commit-bot) wrote:
> Retried try job too often on linux_chromeos for step(s) components_unittests
>
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_chro...

AFAICT, that was someone else and has been reverted, though I haven't looked
super close.  So clicking again.

[commit-bot: I haz the power, 2013-07-22 19:25:35]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/shess@chromium.org/19000007/13001

[commit-bot: I haz the power, 2013-07-22 22:02:19]

Change committed as 212984