Re: [PATCH 03/10] writeback: Do not congestion sleep if there are no congested BDIs or significant writeback

[Mel Gorman <mel@csn.ul.ie>]

On Wed, Sep 08, 2010 at 11:52:45PM +0900, Minchan Kim wrote:
> On Wed, Sep 08, 2010 at 12:04:03PM +0100, Mel Gorman wrote:
> > On Wed, Sep 08, 2010 at 12:25:33AM +0900, Minchan Kim wrote:
> > > > + * @zone: A zone to consider the number of being being written back from
> > > > + * @sync: SYNC or ASYNC IO
> > > > + * @timeout: timeout in jiffies
> > > > + *
> > > > + * Waits for up to @timeout jiffies for a backing_dev (any backing_dev) to exit
> > > > + * write congestion.  If no backing_devs are congested then the number of
> > > > + * writeback pages in the zone are checked and compared to the inactive
> > > > + * list. If there is no sigificant writeback or congestion, there is no point
> > >                                                 and 
> > > 
> > 
> > Why and? "or" makes sense because we avoid sleeping on either condition.
> 
> if (nr_bdi_congested[sync]) == 0) {
>         if (writeback < inactive / 2) {
>                 cond_resched();
>                 ..
>                 goto out
>         }
> }
> 
> for avoiding sleeping, above two condition should meet. 

This is a terrible comment that is badly written. Is this any clearer?

/**
 * wait_iff_congested - Conditionally wait for a backing_dev to become uncongested or a zone to complete writes
 * @zone: A zone to consider the number of being being written back from
 * @sync: SYNC or ASYNC IO
 * @timeout: timeout in jiffies
 *
 * In the event of a congested backing_dev (any backing_dev) or a given @zone
 * having a large number of pages in writeback, this waits for up to @timeout
 * jiffies for either a BDI to exit congestion or a write to complete.
 *
 * If there is no congestion and few pending writes, then cond_resched()
 * is called to yield the processor if necessary but otherwise does not
 * sleep.
 */

> > 
> > > > + * in sleeping but cond_resched() is called in case the current process has
> > > > + * consumed its CPU quota.
> > > > + */
> > > > +long wait_iff_congested(struct zone *zone, int sync, long timeout)
> > > > +{
> > > > +	long ret;
> > > > +	unsigned long start = jiffies;
> > > > +	DEFINE_WAIT(wait);
> > > > +	wait_queue_head_t *wqh = &congestion_wqh[sync];
> > > > +
> > > > +	/*
> > > > +	 * If there is no congestion, check the amount of writeback. If there
> > > > +	 * is no significant writeback and no congestion, just cond_resched
> > > > +	 */
> > > > +	if (atomic_read(&nr_bdi_congested[sync]) == 0) {
> > > > +		unsigned long inactive, writeback;
> > > > +
> > > > +		inactive = zone_page_state(zone, NR_INACTIVE_FILE) +
> > > > +				zone_page_state(zone, NR_INACTIVE_ANON);
> > > > +		writeback = zone_page_state(zone, NR_WRITEBACK);
> > > > +
> > > > +		/*
> > > > +		 * If less than half the inactive list is being written back,
> > > > +		 * reclaim might as well continue
> > > > +		 */
> > > > +		if (writeback < inactive / 2) {
> > > 
> > > I am not sure this is best.
> > > 
> > 
> > I'm not saying it is. The objective is to identify a situation where
> > sleeping until the next write or congestion clears is pointless. We have
> > already identified that we are not congested so the question is "are we
> > writing a lot at the moment?". The assumption is that if there is a lot
> > of writing going on, we might as well sleep until one completes rather
> > than reclaiming more.
> > 
> > This is the first effort at identifying pointless sleeps. Better ones
> > might be identified in the future but that shouldn't stop us making a
> > semi-sensible decision now.
> 
> nr_bdi_congested is no problem since we have used it for a long time.
> But you added new rule about writeback. 
> 

Yes, I'm trying to add a new rule about throttling in the page allocator
and from vmscan. As you can see from the results in the leader, we are
currently sleeping more than we need to.

> Why I pointed out is that you added new rule and I hope let others know
> this change since they have a good idea or any opinions. 
> I think it's a one of roles as reviewer.
> 

Of course.

> > 
> > > 1. Without considering various speed class storage, could we fix it as half of inactive?
> > 
> > We don't really have a good means of identifying speed classes of
> > storage. Worse, we are considering on a zone-basis here, not a BDI
> > basis. The pages being written back in the zone could be backed by
> > anything so we cannot make decisions based on BDI speed.
> 
> True. So it's why I have below question.
> As you said, we don't have enough information in vmscan.
> So I am not sure how effective such semi-sensible decision is. 
> 

What additional metrics would you apply than the ones I used in the
leader mail?

> I think best is to throttle in page-writeback well. 

I do not think there is a problem as such in page writeback throttling.
The problem is that we are going to sleep without any congestion or without
writes in progress. We sleep for a full timeout in this case for no reason
and this is what I'm trying to avoid.

> But I am not a expert about that and don't have any idea. Sorry.

Don't be, this is something that needs thinking about!

> So I can't insist on my nitpick. If others don't have any objection,
> I don't mind this, either. 
> 
> Wu, Do you have any opinion?
> 
> > 
> > > 2. Isn't there any writeback throttling on above layer? Do we care of it in here?
> > > 
> > 
> > There are but congestion_wait() and now wait_iff_congested() are part of
> > that. We can see from the figures in the leader that congestion_wait()
> > is sleeping more than is necessary or smart.
> > 
> > > Just out of curiosity. 
> > > 
> > 
> > -- 
> > Mel Gorman
> > Part-time Phd Student                          Linux Technology Center
> > University of Limerick                         IBM Dublin Software Lab
> 
> -- 
> Kind regards,
> Minchan Kim
> 

-- 
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab