Cleanup histogram syncronizer a bit...

Cleanup histogram syncronizer a bit

I adjusted the code so tha there was no chance
for the pending response count to get to zero
until all the increments had been applied, and
all the rendereres had been notified.  Without
this change, there was a chance that the tally
would hit zero prematurely, triggering a
a premature completion.

I went through the code clarifying where
a lock was needed, and where it wasn't.
In all cases, we now either take a lock, or
assert we're consistently on a single
thread (meaning we don't need a lock).

r=davemoore

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=64781

[jar (doing other things), 2010-11-01 19:19:17]

Dave,

I looked closely and found one more issue.  I decided we should keep the count
greater than zero until we had sent out all the renderer notifications.  This
ensures that we don't bounce-off-zero prematurely.  You wouldn't notice this in
your test (where you previously hung), but you might not get all the up-to-date
data you were expecting.

I also added a bunch of clarifying comments and assertions, so that the places
where a lock is needed are much more apprarent.

[DaveMoore, 2010-11-02 17:53:00]

LGTM