On Jul 21, 2009 09:19 -0700, Andrew Morton wrote:
> On Tue, 21 Jul 2009 12:04:15 +0200 Jan Kara <jack@xxxxxxx> wrote:
> > Due to on disk corruption, it can happen that journal is too short. Fail
> > to load it in such case so that we don't oops somewhere later.
> >
> > Reported-by: Nageswara R Sastry <rnsastry@xxxxxxxxxxxxxxxxxx>
> > Signed-off-by: Jan Kara <jack@xxxxxxx>
> > ---
> > fs/jbd/journal.c | 6 ++++++
> > 1 files changed, 6 insertions(+), 0 deletions(-)
> >
> > diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c
> > index 737f724..94a64a1 100644
> > --- a/fs/jbd/journal.c
> > +++ b/fs/jbd/journal.c
> > @@ -848,6 +848,12 @@ static int journal_reset(journal_t *journal)
> >
> > first = be32_to_cpu(sb->s_first);
> > last = be32_to_cpu(sb->s_maxlen);
> > + if (first + JFS_MIN_JOURNAL_BLOCKS > last + 1) {
> > + printk(KERN_ERR "JBD: Journal too short (blocks %lu-%lu).\n",
> > + first, last);
> > + journal_fail_superblock(journal);
> > + return -EINVAL;
> > + }
> >
> > journal->j_first = first;
> > journal->j_last = last;
>
> It's odd that sb->s_first/s_maxlen are 32-bit and
> journal->j_first/j_last are unsigned long.
>
> These things will only ever be 32-bit unless we change the journal
> superblock.
The jbd on disk structure and APIs cannot handle 64-bit block numbers.
That is one of the first changes we made for jbd2 so that it is possible
to store either 32-bit or 64-bit block numbers in a transaction. I don't
think that needs to be fixed for the jbd code.
Cheers, Andreas
--
Andreas Dilger
Sr. Staff Engineer, Lustre Group
Sun Microsystems of Canada, Inc.
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
