On Monday 2010-02-15 15:49, Jan Kara wrote:
quoted text
>On Sat 13-02-10 13:58:19, Jan Engelhardt wrote:
>> >> 
>> >> This fixes it by using the passed in page writeback count, instead of
>> >> doing MAX_WRITEBACK_PAGES batches, which gets us much better performance
>> >> (Jan reports it's up from ~400KB/sec to 10MB/sec) and makes sync(1)
>> >> finish properly even when new pages are being dirted.
>> >
>> >This seems broken.
>> 
>> It seems so. Jens, Jan Kara, your patch does not entirely fix this.
>> While there is no sync/fsync to be seen in these traces, I can
>> tell there's a livelock, without Dirty decreasing at all.
>
>  I don't think this is directly connected with my / Jens' patch.

I start to think so too.

quoted text
>Similar traces happen even without the patch (see e.g.
>http://bugzilla.kernel.org/show_bug.cgi?id=14830). But maybe the patch
>makes it worse... So are you able to reproduce these warnings and
>without the patch they did not happen?

Your patch speeds up the slow sync; without the patch, there was
no real chance to observ the hard lockup, as the slow sync would
take up all time.

So far, no reproduction. It seems to be just as you say.

quoted text
>  Where in the code is jbd2_journal_commit_transaction+0x218/0x15e0?

0000000000569554 <jbd2_journal_commit_transaction>:
  56976c:       40 04 ee 62     call  6a50f4 <schedule>

Since there is an obvious schedule() call in jbd2_journal_commit_transaction's
C code, I think that's where it is.
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/