Re: [RFC 1/7] release_lru_pages(): Generic release of pages to the LRU

[Christoph Lameter <clameter@sgi.com>]

On Tue, 21 Aug 2007, Mel Gorman wrote:

> > + */
> > +void release_lru_pages(struct list_head *page_list)
> > +{
> 
> Can the migrate.c#putback_lru_pages() be replaced with this?

Correct. We can get rid of the putback_lru_pages in migrate.c 
with this.

> > +	struct page *page;
> > +	struct pagevec pvec;
> > +	struct zone *zone = NULL;
> > +
> > +	pagevec_init(&pvec, 1);
> > +	while (!list_empty(page_list)) {
> > +		page = lru_to_page(page_list);
> > +		VM_BUG_ON(PageLRU(page));
> > +		if (zone != page_zone(page)) {
> > +			if (zone)
> > +				spin_unlock_irq(&zone->lru_lock);
> > +			zone = page_zone(page);
> > +			spin_lock_irq(&zone->lru_lock);
> 
> Is this really necessary? Why situation would occur that would have a
> list of pages in multiple zones?

Because we reclaim from multiple zones and gather laundry from different 
zones.

> Also, it may be worth commenting here that __pagevec_release() is able to
> handle lists of pages in multiple zones.

Ok.

> >  			__count_zone_vm_events(PGSCAN_DIRECT, zone, nr_scan);
> >  		__count_zone_vm_events(PGSTEAL, zone, nr_freed);
> > +		local_irq_enable();
> > +		release_lru_pages(&page_list);
> >  
> 
> Separate these apart by a line. I thought the local_irq_enable() was related
> to the call to release_lru_pages(&page_list) while reading the patch
> which isn't the case at all.

Will do.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>