linux-kernel - Re: [PATCH] write-back: fix break condition

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue, 10 Feb 2009 07:22:49 +0100
From:	Damien Wyart <damien.wyart@...e.fr>
To:	Nick Piggin <npiggin@...e.de>
Cc:	Andrew Morton <akpm@...ux-foundation.org>,
	Federico Cuello <fedux@...men.org.ar>,
	linux-kernel@...r.kernel.org,
	Artem Bityutskiy <Artem.Bityutskiy@...ia.com>
Subject: Re: [PATCH] write-back: fix break condition

* Nick Piggin <npiggin@...e.de> [2009-02-10 01:34]:
> On Mon, Feb 09, 2009 at 03:21:40PM -0800, Andrew Morton wrote:

> > Thanks, but please do cc the people who were involved with a patch when
> > you find a problem with it!

> > On Sat,  7 Feb 2009 01:33:30 -0200
> > Federico Cuello <fedux@...men.org.ar> wrote:

> > > Commit 673353723e7a6550625fb719059c5f31cfaecd18 fixed nr_to_write
> > > counter, but didn't set the break condition properly.

> > It's actually commit dcf6a79dda5cc2a2bec183e50d829030c0972aaa
> > ("write-back: fix nr_to_write counter").

> > > If nr_to_write == 0 after being decremented it will loop one more time
> > > before setting done = 1 and breaking the loop.

> > We prefer that patches include the author's Signed-off-by:, as per
> > Documentation/SubmittingPatches, please.


> > > diff --git a/mm/page-writeback.c b/mm/page-writeback.c
> > > index bb5fa2b..9e2ae50 100644
> > > --- a/mm/page-writeback.c
> > > +++ b/mm/page-writeback.c
> > > @@ -981,20 +981,21 @@ continue_unlock:
> > >  				}
> > >   			}

> > > -			if (nr_to_write > 0)
> > > +			if (nr_to_write > 0) {
> > >  				nr_to_write--;
> > > -			else if (wbc->sync_mode == WB_SYNC_NONE) {
> > > -				/*
> > > -				 * We stop writing back only if we are not
> > > -				 * doing integrity sync. In case of integrity
> > > -				 * sync we have to keep going because someone
> > > -				 * may be concurrently dirtying pages, and we
> > > -				 * might have synced a lot of newly appeared
> > > -				 * dirty pages, but have not synced all of the
> > > -				 * old dirty pages.
> > > -				 */
> > > -				done = 1;
> > > -				break;
> > > +				if (nr_to_write == 0 && wbc->sync_mode == WB_SYNC_NONE) {
> > > +					/*
> > > +					 * We stop writing back only if we are not
> > > +					 * doing integrity sync. In case of integrity
> > > +					 * sync we have to keep going because someone
> > > +					 * may be concurrently dirtying pages, and we
> > > +					 * might have synced a lot of newly appeared
> > > +					 * dirty pages, but have not synced all of the
> > > +					 * old dirty pages.
> > > +					 */
> > > +					done = 1;
> > > +					break;
> > > +				}
> > >  			}

> > >  			if (wbc->nonblocking && bdi_write_congested(bdi)) {

> > Artem, Nick, please check?

> Yes, this looks OK by me.

> Acked-by: Nick Piggin <npiggin@...e.de>

Shouldn't this go to stable 2.6.28, too ? They have not been in Cc for
now. I think the problem was at first detected in 2.6.28.X, but it is
not yet sure if it is solved or if this interacts with some ext4
problems.

-- 
Damien Wyart
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/