On Wed, Apr 21, 2010 at 09:30:20AM -0500, Christoph Lameter wrote:
quoted text
> On Tue, 20 Apr 2010, Mel Gorman wrote:
> 
> > @@ -520,10 +521,12 @@ static int move_to_new_page(struct page *newpage, struct page *page)
> >  	else
> >  		rc = fallback_migrate_page(mapping, newpage, page);
> >
> > -	if (!rc)
> > -		remove_migration_ptes(page, newpage);
> > -	else
> > +	if (rc) {
> >  		newpage->mapping = NULL;
> > +	} else {
> > +		if (remap_swapcache)
> > +			remove_migration_ptes(page, newpage);
> > +	}
> 
> You are going to keep the migration ptes after the page has been unlocked?

Yes, because it's not known if the anon_vma for the unmapped swapcache page
still exists or not.  Now, a bug has been reported where a migration PTE is
found where the page is not locked. I'm trying to determine if it's the same
page or not but the problem takes ages to reproduce.

quoted text
> Or is remap_swapcache true if its not a swapcache page?
> 
> Maybe you meant
> 
> if (!remap_swapcache)
> 
> ?
> 

No, remap_swapcache could just be called "remap". If it's 0, it's
considered unsafe to remap the page.

quoted text
> >  	unlock_page(newpage);
> >
> 
> >
> >  skip_unmap:
> >  	if (!page_mapped(page))
> > -		rc = move_to_new_page(newpage, page);
> > +		rc = move_to_new_page(newpage, page, remap_swapcache);
> >
> > -	if (rc)
> > +	if (rc && remap_swapcache)
> >  		remove_migration_ptes(page, page);
> >  rcu_unlock:
> >
> 
> Looks like you meant !remap_swapcache
> 

If remap_swapcache is 1, the anon_vma is valid (or irrelevant because
it's a file) and it's safe to remap the page by removing the migration
PTEs.

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