LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Andrew Morton <akpm@linux-foundation.org>
To: Arthur Jones <ajones@riverbed.com>
Cc: sandeen@redhat.com, linux-ext4@vger.kernel.org, sct@redhat.com,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH] ext3: wait on all pending commits in ext3_sync_fs
Date: Mon, 3 Nov 2008 12:37:50 -0800	[thread overview]
Message-ID: <20081103123750.67c96224.akpm@linux-foundation.org> (raw)
In-Reply-To: <20081103201428.GB30565@ajones-laptop.nbttech.com>

On Mon, 3 Nov 2008 12:14:28 -0800
Arthur Jones <ajones@riverbed.com> wrote:

> Hi Andrew, ...
> 
> On Mon, Nov 03, 2008 at 11:33:18AM -0800, Andrew Morton wrote:
> > [...]
> > > --- a/fs/ext3/super.c
> > > +++ b/fs/ext3/super.c
> > > @@ -2392,7 +2392,13 @@ static int ext3_sync_fs(struct super_block *sb, int wait)
> > >       if (journal_start_commit(EXT3_SB(sb)->s_journal, &target)) {
> > >               if (wait)
> > >                       log_wait_commit(EXT3_SB(sb)->s_journal, target);
> > > -     }
> > > +     } else if (wait)
> > > +             /*
> > > +              * We may have a commit in progress, clear it out
> > > +              * before we go on...
> > > +              */
> > > +             ext3_force_commit(sb);
> > > +
> > >       return 0;
> > >  }
> > 
> > Can we do
> > 
> >         sb->s_dirt = 0;
> >         if (wait)
> >                 ext3_force_commit(...);
> >         else
> >                 journal_start_commit(...);
> 
> I think this is what you had in mind:

yup.

> 
> diff --git a/fs/ext3/super.c b/fs/ext3/super.c
> index 18eaa78..2b48b85 100644
> --- a/fs/ext3/super.c
> +++ b/fs/ext3/super.c
> @@ -2386,13 +2386,12 @@ static void ext3_write_super (struct super_block * sb)
>  
>  static int ext3_sync_fs(struct super_block *sb, int wait)
>  {
> -	tid_t target;
> -
>  	sb->s_dirt = 0;
> -	if (journal_start_commit(EXT3_SB(sb)->s_journal, &target)) {
> -		if (wait)
> -			log_wait_commit(EXT3_SB(sb)->s_journal, target);
> -	}
> +	if (wait)
> +		ext3_force_commit(sb);
> +	else
> +		journal_start_commit(EXT3_SB(sb)->s_journal, NULL);
> +
>  	return 0;
>  }
>  
> I tried this and it too fixes the problem.  FWIW I agree it
> looks better...
> 

OK.  But still, questions remain.

- should we clear s_dirt if log_wait_commit() didn't work?

- ext3_force_commit() clears s_dirt also

- should ext3_sync_fs() be dropping the ext3_force_commit() return
  value on the floor?

This is all rather worrisome.

  reply	other threads:[~2008-11-03 20:40 UTC|newest]

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20081024183733.GA25797@ajones-laptop.nbttech.com>
2008-10-27 16:54 ` ext3: slow symlink corruption on umount Arthur Jones
2008-10-29 19:54   ` Arthur Jones
2008-10-29 20:36     ` Eric Sandeen
2008-10-29 21:09       ` Theodore Tso
2008-10-30 13:38         ` Eric Sandeen
2008-10-30 13:55           ` Arthur Jones
2008-10-31  9:47           ` Nick Piggin
2008-10-30 17:40       ` Arthur Jones
2008-10-30 18:03         ` Eric Sandeen
2008-10-30 21:34           ` Arthur Jones
2008-10-31 17:24             ` Arthur Jones
2008-10-31 18:37               ` Eric Sandeen
2008-10-30 18:32         ` Arthur Jones
2008-11-03 18:44       ` [PATCH] ext3: wait on all pending commits in ext3_sync_fs Arthur Jones
2008-11-03 19:33         ` Andrew Morton
2008-11-03 20:14           ` Arthur Jones
2008-11-03 20:37             ` Andrew Morton [this message]
2008-11-03 20:58               ` Arthur Jones
2008-11-03 21:13                 ` Andrew Morton
2008-11-03 21:19                   ` Theodore Tso
2008-11-03 21:27                     ` Andrew Morton
2008-11-03 21:48                       ` Theodore Tso
2008-11-03 22:01                       ` Theodore Tso
2008-11-03 22:18                         ` Arthur Jones
2008-11-03 22:27                         ` Andrew Morton
2008-11-03 22:55                           ` Theodore Tso
2008-11-03 23:01                             ` Arthur Jones
2008-11-03 23:12                               ` Theodore Tso
2008-11-04 16:26                                 ` Arthur Jones
2008-11-03 21:48               ` Arthur Jones
2008-11-03 22:47                 ` Theodore Tso
2008-12-18 23:17             ` Jan Kara
2008-12-18 23:37               ` Eric Sandeen
2008-12-19  0:27                 ` Jan Kara
2008-12-19  1:34                   ` Eric Sandeen
2008-12-22 19:15                     ` Ric Wheeler
2008-12-22 22:57                       ` Andreas Dilger
2008-12-23  0:09                         ` Ric Wheeler
2008-12-23 15:56                         ` Eric Sandeen
2009-01-12 22:28                 ` Jan Kara
2009-01-13 17:21                   ` Eric Sandeen
2009-01-13 22:14               ` Eric Sandeen
2009-01-14  4:24                 ` Theodore Tso
2009-01-14 17:26                   ` Eric Sandeen
2009-01-14 17:27                   ` Jan Kara
2009-01-29 18:27                     ` Mike Snitzer
2009-01-29 20:05                       ` Eric Sandeen
2008-11-03 19:59         ` Eric Sandeen

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=20081103123750.67c96224.akpm@linux-foundation.org \
    --to=akpm@linux-foundation.org \
    --cc=ajones@riverbed.com \
    --cc=linux-ext4@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=sandeen@redhat.com \
    --cc=sct@redhat.com \
    --subject='Re: [PATCH] ext3: wait on all pending commits in ext3_sync_fs' \
    /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).