ThreadWatcher - Added jankiness monitor for FILE thread.

ThreadWatcher - Added jankiness monitor for FILE thread.

Extended crash-on-hang-threads command line switch to
accept the unresponsive seconds. 

For example, if  --crash-on-hang-threads=UI:3:18,IO:3:18 is passed as
command line argument, ThreadWatcher crashes the browser if UI or IO
is not responsive for 18 seconds and the number of browser threads that
are responding is less than or equal to 3.

If there is no command line switch, it defaults to
"UI:3:18,IO:3:18,FILE:3,90" as the value in Canary and Dev.

It defaults to empty string in STABLE channel (i.e., we don't crash
the browser due to jankiness).

For beta channel it defaults to "UI:3:36,IO:3:36,FILE:3:180".

Deleted crash-on-hang-threads command line switch because
crash-on-hang-threads supports this.

R=jar@chromium.org
TEST=unit tests for ThreadWatcher

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=181874

[ramant (doing other things), 2013-02-06 03:44:11]

Hi Jim,
  Made the changes we had discussed to detect the jankiness on FILE thread to
crash the browser if FILE thread doesn't respond for 90 secs. Would appreciate
your comments.

thanks
raman

[ramant (doing other things), 2013-02-11 23:17:10]

Hi Jim,
  Updated the ThreadWatcher's comments. PTAL.
thanks
raman

[jar (doing other things), 2013-02-12 01:27:27]

Just some nits below, then LGTM.

https://codereview.chromium.org/12183008/diff/41005/chrome/browser/metrics/th...
File chrome/browser/metrics/thread_watcher.h (right):

https://codereview.chromium.org/12183008/diff/41005/chrome/browser/metrics/th...
chrome/browser/metrics/thread_watcher.h:341: // UI and IO threads, we would
crash if it doesn't respond for 18 seconds (or
nit: it doesn't --> those threads don't

https://codereview.chromium.org/12183008/diff/41005/chrome/browser/metrics/th...
chrome/browser/metrics/thread_watcher.h:345: // ThreadWatcher watches six (UI,
IO, DB, FILE, FILE_USER_BLOCKING and CACHE)
This paragraph should move up earlier, before you list the example on line 335
that starts "It parses command line..."

Change sentence to:
ThreadWatcher MONITORS six BROWSER THREADS (i.e., UI, IO, DB, FILE,
FILE_USER_BLOCKING and CACHE).  Out of the 6 threads, any subset may be watched,
to potentially cause a crash.  The following example's command line causes
exactly 3 threads to be watched.  The example command line argument consists of 
"UI...."  In that string, ....

https://codereview.chromium.org/12183008/diff/41005/chrome/browser/metrics/th...
chrome/browser/metrics/thread_watcher.h:355: // responding, as we'd assume such
large hang counts mean that the system is
nit: should have read:

We would not crash if none of the threads are responding.

There is an extra "not" before "responding" in your text.

https://codereview.chromium.org/12183008/diff/41005/chrome/browser/metrics/th...
chrome/browser/metrics/thread_watcher.h:357: // Example 2: If the
|live_threads_threshold| value for "UI" was INT_MAX and
Instead of INT_MAX, "any number higher than 6"

[ramant (doing other things), 2013-02-12 01:54:30]

https://codereview.chromium.org/12183008/diff/41005/chrome/browser/metrics/th...
File chrome/browser/metrics/thread_watcher.h (right):

https://codereview.chromium.org/12183008/diff/41005/chrome/browser/metrics/th...
chrome/browser/metrics/thread_watcher.h:341: // UI and IO threads, we would
crash if it doesn't respond for 18 seconds (or
On 2013/02/12 01:27:27, jar wrote:
> nit: it doesn't --> those threads don't

Done.

https://codereview.chromium.org/12183008/diff/41005/chrome/browser/metrics/th...
chrome/browser/metrics/thread_watcher.h:345: // ThreadWatcher watches six (UI,
IO, DB, FILE, FILE_USER_BLOCKING and CACHE)
On 2013/02/12 01:27:27, jar wrote:
> This paragraph should move up earlier, before you list the example on line 335
> that starts "It parses command line..."
> 
> Change sentence to:
> ThreadWatcher MONITORS six BROWSER THREADS (i.e., UI, IO, DB, FILE,
> FILE_USER_BLOCKING and CACHE).  Out of the 6 threads, any subset may be
watched,
> to potentially cause a crash.  The following example's command line causes
> exactly 3 threads to be watched.  The example command line argument consists
of 
> "UI...."  In that string, ....

Done.

https://codereview.chromium.org/12183008/diff/41005/chrome/browser/metrics/th...
chrome/browser/metrics/thread_watcher.h:355: // responding, as we'd assume such
large hang counts mean that the system is
On 2013/02/12 01:27:27, jar wrote:
> nit: should have read:
> 
> We would not crash if none of the threads are responding.
> 
> There is an extra "not" before "responding" in your text.

Done.

https://codereview.chromium.org/12183008/diff/41005/chrome/browser/metrics/th...
chrome/browser/metrics/thread_watcher.h:357: // Example 2: If the
|live_threads_threshold| value for "UI" was INT_MAX and
On 2013/02/12 01:27:27, jar wrote:
> Instead of INT_MAX, "any number higher than 6"

Done.

[jar (doing other things), 2013-02-12 01:56:41]

lgtm

[commit-bot: I haz the power, 2013-02-12 02:00:56]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rtenneti@chromium.org/12183008/29006

[commit-bot: I haz the power, 2013-02-12 05:29:33]

Change committed as 181874