LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Nick Piggin <npiggin@suse.de>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	Hugh Dickins <hugh@veritas.com>,
	Linux Kernel <linux-kernel@vger.kernel.org>,
	Linux Memory Management <linux-mm@kvack.org>,
	Linux Filesystems <linux-fsdevel@vger.kernel.org>
Subject: Re: [patch 2/3] fs: buffer don't PageUptodate without page locked
Date: Tue, 6 Feb 2007 09:31:03 +0100	[thread overview]
Message-ID: <20070206083103.GB16965@wotan.suse.de> (raw)
In-Reply-To: <20070206002140.4030a11f.akpm@linux-foundation.org>

On Tue, Feb 06, 2007 at 12:21:40AM -0800, Andrew Morton wrote:
> On Tue,  6 Feb 2007 09:02:23 +0100 (CET) Nick Piggin <npiggin@suse.de> wrote:
> 
> > __block_write_full_page is calling SetPageUptodate without the page locked.
> > This is unusual, but not incorrect, as PG_writeback is still set.
> > 
> > However with the previous patch, this is now a problem: so don't bother
> > setting the page uptodate in this case (it is weird that the write path
> > does such a thing anyway). Instead just leave it to the read side to bring
> > the page uptodate when it notices that all buffers are uptodate.
> > 
> > Signed-off-by: Nick Piggin <npiggin@suse.de>
> > 
> > Index: linux-2.6/fs/buffer.c
> > ===================================================================
> > --- linux-2.6.orig/fs/buffer.c
> > +++ linux-2.6/fs/buffer.c
> > @@ -1679,6 +1679,7 @@ static int __block_write_full_page(struc
> >  	 */
> >  	BUG_ON(PageWriteback(page));
> >  	set_page_writeback(page);
> > +	unlock_page(page);
> >  
> >  	do {
> >  		struct buffer_head *next = bh->b_this_page;
> > @@ -1688,7 +1689,6 @@ static int __block_write_full_page(struc
> >  		}
> >  		bh = next;
> >  	} while (bh != head);
> > -	unlock_page(page);
> >  
> >  	err = 0;
> >  done:
> 
> Why this change?  Without looking at it too hard, it seems that if
> submit_bh() completes synchronously, this thread can end up playing with
> the buffers on a non-locked, non-PageWriteback page.  Someone else could
> whip the buffers away and oops?

Hmm, it definitely shouldn't be there, it leaked in from another patch
to bring partiy with the error handling...

Here is an updated patch.

--

__block_write_full_page is calling SetPageUptodate without the page locked.
This is unusual, but not incorrect, as PG_writeback is still set.

However with the previous patch, this is now a problem: so don't bother
setting the page uptodate in this case (it is weird that the write path
does such a thing anyway). Instead just leave it to the read side to bring
the page uptodate when it notices that all buffers are uptodate.

Signed-off-by: Nick Piggin <npiggin@suse.de>

Index: linux-2.6/fs/buffer.c
===================================================================
--- linux-2.6.orig/fs/buffer.c
+++ linux-2.6/fs/buffer.c
@@ -1698,17 +1698,8 @@ done:
 		 * clean.  Someone wrote them back by hand with
 		 * ll_rw_block/submit_bh.  A rare case.
 		 */
-		int uptodate = 1;
-		do {
-			if (!buffer_uptodate(bh)) {
-				uptodate = 0;
-				break;
-			}
-			bh = bh->b_this_page;
-		} while (bh != head);
-		if (uptodate)
-			SetPageUptodate(page);
 		end_page_writeback(page);
+
 		/*
 		 * The page and buffer_heads can be released at any time from
 		 * here on.

  reply	other threads:[~2007-02-06  8:31 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-02-06  8:02 [patch 0/3] 2.6.20 fix for PageUptodate memorder problem Nick Piggin
2007-02-06  8:02 ` [patch 1/3] mm: fix PageUptodate memorder Nick Piggin
2007-02-06  8:25   ` Andrew Morton
2007-02-06  8:51     ` Nick Piggin
2007-02-06  8:02 ` [patch 2/3] fs: buffer don't PageUptodate without page locked Nick Piggin
2007-02-06  8:21   ` Andrew Morton
2007-02-06  8:31     ` Nick Piggin [this message]
2007-02-06  8:02 ` [patch 3/3] mm: make read_cache_page synchronous Nick Piggin
2007-02-06  8:28   ` Andrew Morton
2007-02-06  8:56     ` Nick Piggin
2007-02-06  8:58     ` Nick Piggin
2007-02-06 22:58 ` [patch 0/3] 2.6.20 fix for PageUptodate memorder problem David Chinner
2007-02-07  3:13   ` Nick Piggin
2007-02-08 13:26 [patch 0/3] 2.6.20 fix for PageUptodate memorder problem (try 2) Nick Piggin
2007-02-08 13:27 ` [patch 2/3] fs: buffer don't PageUptodate without page locked Nick Piggin
2007-02-10  2:31 [patch 0/3] 2.6.20 fix for PageUptodate memorder problem (try 3) Nick Piggin
2007-02-10  2:31 ` [patch 2/3] fs: buffer don't PageUptodate without page locked Nick Piggin
2007-02-15  7:31 [patch 0/3] 2.6.20 fix for PageUptodate memorder problem (try 4) Nick Piggin
2007-02-15  7:31 ` [patch 2/3] fs: buffer don't PageUptodate without page locked Nick Piggin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20070206083103.GB16965@wotan.suse.de \
    --to=npiggin@suse.de \
    --cc=akpm@linux-foundation.org \
    --cc=hugh@veritas.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=torvalds@linux-foundation.org \
    --subject='Re: [patch 2/3] fs: buffer don'\''t PageUptodate without page locked' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).