Re: [PATCH v2] page cache: Store only head pages in i_pages

[Kirill A. Shutemov]

On Thu, Feb 14, 2019 at 02:11:58PM -0800, Matthew Wilcox wrote:
> On Fri, Feb 15, 2019 at 01:08:10AM +0300, Kirill A. Shutemov wrote:
> > On Thu, Feb 14, 2019 at 01:17:57PM -0800, Matthew Wilcox wrote:
> > > On Thu, Feb 14, 2019 at 04:30:04PM +0300, Kirill A. Shutemov wrote:
> > > >  - migrate_page_move_mapping() has to be converted too.
> > > 
> > > I think that's as simple as:
> > > 
> > > +++ b/mm/migrate.c
> > > @@ -465,7 +465,7 @@ int migrate_page_move_mapping(struct address_space *mapping,
> > >  
> > >                 for (i = 1; i < HPAGE_PMD_NR; i++) {
> > >                         xas_next(&xas);
> > > -                       xas_store(&xas, newpage + i);
> > > +                       xas_store(&xas, newpage);
> > >                 }
> > >         }
> > >  
> > > 
> > > or do you see something else I missed?
> > 
> > Looks right to me.
> > 
> > BTW, maybe some add syntax sugar from XArray side?
> > 
> > Replace the loop and xas_store() before it with:
> > 
> > 		xas_fill(&xas, newpage, 1UL << compound_order(newpage));
> > 
> > or something similar?
> 
> If we were keeping this code longterm, then yes, something like that
> would be great.  I'm hoping this code is a mere stepping stone towards
> using multi-slot entries for the page cache.

Fair enough.

-- 
 Kirill A. Shutemov