Re: [patch] mm: clean up zone flags

[Johannes Weiner]

On Tue, Sep 02, 2014 at 02:42:14PM -0700, David Rientjes wrote:
> On Tue, 2 Sep 2014, Johannes Weiner wrote:
> > @@ -631,7 +631,7 @@ long wait_iff_congested(struct zone *zone, int sync, long timeout)
> >  	 * of sleeping on the congestion queue
> >  	 */
> >  	if (atomic_read(&nr_bdi_congested[sync]) == 0 ||
> > -			!zone_is_reclaim_congested(zone)) {
> > +	    test_bit(ZONE_CONGESTED, &zone->flags)) {
> >  		cond_resched();
> >  
> >  		/* In case we scheduled, work out time remaining */
> 
> That's not equivalent.
> 
> [snip]
> 
> > diff --git a/mm/vmscan.c b/mm/vmscan.c
> > index 2836b5373b2e..590a92bec6a4 100644
> > --- a/mm/vmscan.c
> > +++ b/mm/vmscan.c
> > @@ -920,7 +920,7 @@ static unsigned long shrink_page_list(struct list_head *page_list,
> >  			/* Case 1 above */
> >  			if (current_is_kswapd() &&
> >  			    PageReclaim(page) &&
> > -			    zone_is_reclaim_writeback(zone)) {
> > +			    test_bit(ZONE_WRITEBACK, &zone->flags)) {
> >  				nr_immediate++;
> >  				goto keep_locked;
> >  
> > @@ -1002,7 +1002,7 @@ static unsigned long shrink_page_list(struct list_head *page_list,
> >  			 */
> >  			if (page_is_file_cache(page) &&
> >  					(!current_is_kswapd() ||
> > -					 !zone_is_reclaim_dirty(zone))) {
> > +					 test_bit(ZONE_DIRTY, &zone->flags))) {
> >  				/*
> >  				 * Immediately reclaim when written back.
> >  				 * Similar in principal to deactivate_page()
> 
> Nor is this.
>
> After fixed, for the oom killer bits:
> 
> 	Acked-by: David Rientjes <rientjes@xxxxxxxxxx>
> 
> since this un-obscurification is most welcome.

Yikes, thanks for catching those and acking.  Updated patch:

---