Temporary instrumentation to help understand lock errors on mac debug bots.

Temporary instrumentation to help understand lock errors on mac debug bots.

BUG=102161
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=107835

[eroman, 2011-10-28 19:55:12]

I figured I would throw in some asserts to catch use-after-free of locks, in
case that is the reason why the lock handle is invalid on the bots.

I don't really have any evidence that shows this is the case though. If this
ends up being true, we will still need symbols to figure out where that lock
came from...

Dunno, what do you think?

[Mark Mentovai, 2011-10-28 19:58:26]

http://codereview.chromium.org/8414031/diff/3001/base/synchronization/lock_im...
File base/synchronization/lock_impl.h (right):

http://codereview.chromium.org/8414031/diff/3001/base/synchronization/lock_im...
base/synchronization/lock_impl.h:56: #if defined(OS_MAC) && !defined(NDEBUG)
This won’t ever work. It’s OS_MACOSX.

You can add another #define inside this #ifdef so that you don’t need to keep
checking the same condition over and over again in the .cc file.

[eroman, 2011-10-28 20:00:52]

http://codereview.chromium.org/8414031/diff/3001/base/synchronization/lock_im...
File base/synchronization/lock_impl.h (right):

http://codereview.chromium.org/8414031/diff/3001/base/synchronization/lock_im...
base/synchronization/lock_impl.h:56: #if defined(OS_MAC) && !defined(NDEBUG)
On 2011/10/28 19:58:26, Mark Mentovai wrote:
> This won’t ever work. It’s OS_MACOSX.
> 
> You can add another #define inside this #ifdef so that you don’t need to keep
> checking the same condition over and over again in the .cc file.

Thanks Mark!

I'll do that, and ping back once I have actually tested this :)  I just wanted
to get feedback early.

[eroman, 2011-10-28 22:03:08]

Please take another look.

[Mark Mentovai, 2011-10-28 22:06:42]

LGTM

[commit-bot: I haz the power, 2011-10-28 22:37:24]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/eroman@chromium.org/8414031/2005

[commit-bot: I haz the power, 2011-10-29 00:46:40]

Change committed as 107835

[eroman, 2011-11-08 00:18:58]

FYI, I will be reverting this now as the underlying issue has been resolved
(confirmed it hasn't reproduced in past few days on bots).

I will be TBRing the revert; no need for comment.