public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Josef Bacik <jbacik@redhat.com>
To: Andreas Dilger <adilger@sun.com>
Cc: Josef Bacik <jbacik@redhat.com>,
	linux-ext4@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] jbd: fix assertion failure in journal_next_log_block
Date: Thu, 31 Jan 2008 16:52:26 -0500	[thread overview]
Message-ID: <20080131215226.GC29679@unused.rdu.redhat.com> (raw)
In-Reply-To: <20080131193543.GS23836@webber.adilger.int>

On Thu, Jan 31, 2008 at 12:35:43PM -0700, Andreas Dilger wrote:
> On Jan 31, 2008  11:14 -0500, Josef Bacik wrote:
> [snip excellent analysis]
> > So you get into this situation where
> > t_nr_buffers (the actual number of buffers that are on the transaction) is
> > greater than the number of buffers accounted for via t_outstanding_credits.
> > This presents a problem since as we loop through writting buffers to the
> > journal, we decrement t_outstanding_credits, and if t_nr_buffers is more than
> > t_outstanding_credits then we end up with a negative number for
> > t_outstanding_credits
> > 
> > Signed-off-by: Josef Bacik <jbacik@redhat.com>
> 
> Do you know what kernel this problem was introduced in, or is this a
> long standing problem?  Presumably the same is needed for jbd2?
> 
> Once we have some decent amount of testing going on with ext4, I think
> it makes sense to merge the jbd2 changes back into jbd and return to
> a single code base, since there is nothing in the jbd2 code that ext3
> can't also work with (i.e. all of the changes are properly isolated
> with compatibility flags and such).
> 
> > @@ -1056,7 +1056,7 @@ static inline int jbd_space_needed(journal_t *journal)
> >  	int nblocks = journal->j_max_transaction_buffers;
> >  	if (journal->j_committing_transaction)
> >  		nblocks += journal->j_committing_transaction->
> > -					t_outstanding_credits;
> > +					t_nr_buffers;
> 
> (trivial) this can be moved back onto the previous line.
> 
> > @@ -1168,7 +1168,7 @@ static inline int jbd_space_needed(journal_t *journal)
> >  	int nblocks = journal->j_max_transaction_buffers;
> >  	if (journal->j_committing_transaction)
> >  		nblocks += journal->j_committing_transaction->
> > -					t_outstanding_credits;
> > +					t_nr_buffers;
> 
> Same...
>

The original issue was reported on RHEL4, so thats 2.6.9, and looking through
the old-bkcvs git tree I can't see where this was introduced, so it's probably
existed before that.  The same problem looks to exist in jbd2 though I haven't
tested it myself, I just went ahead and included the fixes.  Here is the updated
patch, thanks much for the comments.

Signed-off-by: Josef Bacik <jbacik@redhat.com>

 
diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c
index 31853eb..e385a5c 100644
--- a/fs/jbd/commit.c
+++ b/fs/jbd/commit.c
@@ -561,13 +561,6 @@ void journal_commit_transaction(journal_t *journal)
 			continue;
 		}
 
-		/*
-		 * start_this_handle() uses t_outstanding_credits to determine
-		 * the free space in the log, but this counter is changed
-		 * by journal_next_log_block() also.
-		 */
-		commit_transaction->t_outstanding_credits--;
-
 		/* Bump b_count to prevent truncate from stumbling over
                    the shadowed buffer!  @@@ This can go if we ever get
                    rid of the BJ_IO/BJ_Shadow pairing of buffers. */
diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c
index 4f302d2..c0f93f5 100644
--- a/fs/jbd2/commit.c
+++ b/fs/jbd2/commit.c
@@ -580,7 +580,7 @@ void jbd2_journal_commit_transaction(journal_t *journal)
 	stats.u.run.rs_logging = jiffies;
 	stats.u.run.rs_flushing = jbd2_time_diff(stats.u.run.rs_flushing,
 						 stats.u.run.rs_logging);
-	stats.u.run.rs_blocks = commit_transaction->t_outstanding_credits;
+	stats.u.run.rs_blocks = commit_transaction->t_nr_buffers;
 	stats.u.run.rs_blocks_logged = 0;
 
 	descriptor = NULL;
@@ -655,13 +655,6 @@ void jbd2_journal_commit_transaction(journal_t *journal)
 			continue;
 		}
 
-		/*
-		 * start_this_handle() uses t_outstanding_credits to determine
-		 * the free space in the log, but this counter is changed
-		 * by jbd2_journal_next_log_block() also.
-		 */
-		commit_transaction->t_outstanding_credits--;
-
 		/* Bump b_count to prevent truncate from stumbling over
                    the shadowed buffer!  @@@ This can go if we ever get
                    rid of the BJ_IO/BJ_Shadow pairing of buffers. */
diff --git a/include/linux/jbd.h b/include/linux/jbd.h
index d9ecd13..eaeb3db 100644
--- a/include/linux/jbd.h
+++ b/include/linux/jbd.h
@@ -1055,8 +1055,7 @@ static inline int jbd_space_needed(journal_t *journal)
 {
 	int nblocks = journal->j_max_transaction_buffers;
 	if (journal->j_committing_transaction)
-		nblocks += journal->j_committing_transaction->
-					t_outstanding_credits;
+		nblocks += journal->j_committing_transaction->t_nr_buffers;
 	return nblocks;
 }
 
diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h
index 2cbf6fd..acf9d34 100644
--- a/include/linux/jbd2.h
+++ b/include/linux/jbd2.h
@@ -1167,8 +1167,7 @@ static inline int jbd_space_needed(journal_t *journal)
 {
 	int nblocks = journal->j_max_transaction_buffers;
 	if (journal->j_committing_transaction)
-		nblocks += journal->j_committing_transaction->
-					t_outstanding_credits;
+		nblocks += journal->j_committing_transaction->t_nr_buffers;
 	return nblocks;
 }
 

  reply	other threads:[~2008-01-31 21:58 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-01-31 16:14 [PATCH] jbd: fix assertion failure in journal_next_log_block Josef Bacik
2008-01-31 19:35 ` Andreas Dilger
2008-01-31 21:52   ` Josef Bacik [this message]
2008-02-01  0:50     ` Mingming Cao
2008-02-05 17:27 ` Jan Kara
2008-02-05 18:59   ` Josef Bacik
2008-02-06 11:06     ` Jan Kara
2008-02-06 18:47     ` Mingming Cao

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=20080131215226.GC29679@unused.rdu.redhat.com \
    --to=jbacik@redhat.com \
    --cc=adilger@sun.com \
    --cc=linux-ext4@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox