Re: [PATCH 5/5] Mark page cache pages as __GFP_PAGECACHE instead of __GFP_MOVABLE

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

On Thu, 17 May 2007, Andrew Morton wrote:

> On Thu, 17 May 2007 11:12:03 +0100 (IST)
> Mel Gorman <mel@csn.ul.ie> wrote:
>
> > --- linux-2.6.22-rc1-mm1-025_gfphighuser/fs/buffer.c	2007-05-16 22:55:50.000000000 +0100
> > +++ linux-2.6.22-rc1-mm1-030_pagecache_mark/fs/buffer.c	2007-05-16 23:07:30.000000000 +0100
> > @@ -1009,7 +1009,7 @@ grow_dev_page(struct block_device *bdev,
> >  	struct buffer_head *bh;
> >
> >  	page = find_or_create_page(inode->i_mapping, index,
> > -					GFP_NOFS|__GFP_RECLAIMABLE);
> > +					GFP_NOFS_PAGECACHE);
> >  	if (!page)
> >  		return NULL;
>
> I ended up with
>
>        page = find_or_create_page(inode->i_mapping, index,
>                (mapping_gfp_mask(inode->i_mapping) & ~__GFP_FS)|__GFP_MOVABLE);
>
> here.

That looks like it'll work fine with respects to grouping by mobility but 
there is a slight functional difference worth noting. Specifically, the 
old code did not obey cpuset limits because __GFP_HARDWALL was not set. 
This change gets it's GFP mask from bdget() calling mapping_set_gfp_mask() 
which is GFP_USER and so will obey CPUSET limits. This new version looks 
more correct.

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