From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner+w=401wt.eu-S1760183AbYEBCdV@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1760183AbYEBCdV (ORCPT <rfc822;w@1wt.eu>);
	Thu, 1 May 2008 22:33:21 -0400
Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1756019AbYEBCdM
	(ORCPT <rfc822;linux-kernel-outgoing>);
	Thu, 1 May 2008 22:33:12 -0400
Received: from netops-testserver-3-out.sgi.com ([192.48.171.28]:39949 "EHLO
	relay.sgi.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org
	with ESMTP id S1751564AbYEBCdL (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Thu, 1 May 2008 22:33:11 -0400
Date: Thu, 1 May 2008 21:33:08 -0500
From: Russ Anderson <rja@sgi.com>
To: Christoph Lameter <clameter@sgi.com>
Cc: linux-kernel@vger.kernel.org, linux-ia64@vger.kernel.org,
       Linus Torvalds <torvalds@linux-foundation.org>,
       Andrew Morton <akpm@linux-foundation.org>,
       Tony Luck <tony.luck@intel.com>
Subject: Re: [PATCH 2/3] mm: Avoid putting a bad page back on the LRU v2
Message-ID: <20080502023307.GA1591@sgi.com>
Reply-To: Russ Anderson <rja@sgi.com>
References: <20080502004248.GC12006@sgi.com> <Pine.LNX.4.64.0805011810410.13527@schroedinger.engr.sgi.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Pine.LNX.4.64.0805011810410.13527@schroedinger.engr.sgi.com>
User-Agent: Mutt/1.5.9i
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Thu, May 01, 2008 at 06:12:49PM -0700, Christoph Lameter wrote:
> On Thu, 1 May 2008, Russ Anderson wrote:
> 
> > Index: linus/mm/migrate.c
> > ===================================================================
> > --- linus.orig/mm/migrate.c	2008-05-01 19:05:33.000000000 -0500
> > +++ linus/mm/migrate.c	2008-05-01 19:06:15.000000000 -0500
> > @@ -380,6 +380,7 @@ static void migrate_page_copy(struct pag
> >  		SetPageChecked(newpage);
> >  	if (PageMappedToDisk(page))
> >  		SetPageMappedToDisk(newpage);
> > +	/* Do not migrate PG_memerror to the new page */
> 
> Why is the comment here?

To let anyone that may think PG_memerror should get copied to 
the new page that it should not.  I can remove the comment.
  
> >  	if (PageDirty(page)) {
> >  		clear_page_dirty_for_io(page);
> > @@ -721,6 +722,8 @@ unlock:
> >   		 */
> >   		list_del(&page->lru);
> >   		move_to_lru(page);
> > +		if (PageMemError(page))
> > +			totalbad_pages++;
> 
> Wouldnt this be taken care of by the lru handling?

It is only for accounting purposes, keeping a count
of the bad pages.

> > +#ifdef CONFIG_PAGEFLAGS_EXTENDED
> > +PAGEFLAG(MemError, memerror)
> > +#else
> > +#define PageMemError(page)	0
> 
> Use
> 
> PAGEFLAG_FLAGS(MemError)

OK.

> >  #include <linux/page-flags.h>
> >  
> > -#define PAGE_FLAGS	(1 << PG_lru   | 1 << PG_private   | 1 << PG_locked | \
> > +#define PAGE_FLAGS_BASE	(1 << PG_lru   | 1 << PG_private   | 1 << PG_locked | \
> >  			 1 << PG_buddy | 1 << PG_writeback | \
> >  			 1 << PG_slab  | 1 << PG_swapcache | 1 << PG_active)
> > +
> > +#ifdef CONFIG_PAGEFLAGS_EXTENDED
> > +#define	PAGE_FLAGS	(PAGE_FLAGS_BASE | 1UL << PG_memerror)
> > +#else
> > +#define	PAGE_FLAGS	(PAGE_FLAGS_BASE)
> > +#endif
> >  
> >  #define PAGE_FLAGS_RECLAIM      (PAGE_FLAGS | 1 << PG_reclaim | 1 << PG_dirty)
> >  #define PAGE_FLAGS_RESERVE	(PAGE_FLAGS | 1 << PG_reserved)
> 
> The groups of page flags could also be put into page-flags.h.

page-flags.h is where I originally put them.  Since they are only used in 
mm/page_alloc.c, I thought mm.h would be a more appropriate place.  
I can put them in page-flags.h.

-- 
Russ Anderson, OS RAS/Partitioning Project Lead  
SGI - Silicon Graphics Inc          rja@sgi.com