Re: regression in page writeback

[Wu Fengguang]

On Wed, Sep 23, 2009 at 10:09:15AM +0800, Andrew Morton wrote:
> On Wed, 23 Sep 2009 10:01:04 +0800 Wu Fengguang <fengguang.wu@xxxxxxxxx> wrote:
> 
> > > If there's still outstanding dirty data for any of those queues, both
> > > wb_kupdate() and background_writeout() will take a teeny sleep and then
> > > will re-poll the queues.
> > > 
> > > Did that logic get broken?
> > 
> > No, but the "teeny sleep" is normally much smaller. When io queue is
> > not congested, every io completion event will wakeup the congestion
> > waiters. Also A's event could wake up B's waiters.
> > 
> > __freed_request() always calls blk_clear_queue_congested() if under
> > congestion threshold which in turn wakes up congestion waiters:
> > 
> >         if (rl->count[sync] < queue_congestion_off_threshold(q))
> >                 blk_clear_queue_congested(q, sync);
> > 
> 
> Yes.  Have any problems been demonstrated due to that?

Hmm, I was merely clarifying a fact.

Chris Mason listed some reasons to convert the congestion_wait() based
polls to the some queue waiting:
 
  http://lkml.org/lkml/2009/9/8/210

My impression is, it changed obviously too fast without enough discussions.

> And what's _sufficiently_ wrong with that to justify adding potentially
> thousands of kernel threads?  It was always a design objective to avoid
> doing that.

Yeah, the number of threads could be a problem. 

I guess the per-bdi threads and congestion_wait are mostly two
independent changes.  congestion_wait is not the reason to do per-bdi
threads.  Just that Jens piggy backed some changes to congestion_wait.

Thanks,
Fengguang
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/