2023-04-10 17:28:32

by Guoqing Cai

Subject: [PATCH] fs: jbd2: fix an incorrect warn log

In jbd2_journal_load(), when journal_reset fails, it prints an incorrect
warn log.

Fix this by changing the goto statement to return statement.

Signed-off-by: Guoqing Cai <[email protected]>
Reviewed-by: Dongliang Mu <[email protected]>
---
fs/jbd2/journal.c | 12 +++++-------
1 file changed, 5 insertions(+), 7 deletions(-)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index e80c781731f8..555f09ca1d99 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -2082,8 +2082,10 @@ int jbd2_journal_load(journal_t *journal)

/* Let the recovery code check whether it needs to recover any
* data from the journal. */
- if (jbd2_journal_recover(journal))
- goto recovery_error;
+ if (jbd2_journal_recover(journal)) {
+ printk(KERN_WARNING "JBD2: recovery failed\n");
+ return -EIO;
+ }

if (journal->j_failed_commit) {
printk(KERN_ERR "JBD2: journal transaction %u on %s "
@@ -2101,14 +2103,10 @@ int jbd2_journal_load(journal_t *journal)
* reinitialise the dynamic contents of the superblock in memory
* and reset them on disk. */
if (journal_reset(journal))
- goto recovery_error;
+ return -EIO;

journal->j_flags |= JBD2_LOADED;
return 0;
-
-recovery_error:
- printk(KERN_WARNING "JBD2: recovery failed\n");
- return -EIO;
}

/**
--
2.40.0

2023-04-10 17:51:28

by Darrick J. Wong

Subject: Re: [PATCH] fs: jbd2: fix an incorrect warn log

On Tue, Apr 11, 2023 at 01:20:38AM +0800, Guoqing Cai wrote:
> In jbd2_journal_load(), when journal_reset fails, it prints an incorrect
> warn log.
>
> Fix this by changing the goto statement to return statement.
>
> Signed-off-by: Guoqing Cai <[email protected]>
> Reviewed-by: Dongliang Mu <[email protected]>
> ---
> fs/jbd2/journal.c | 12 +++++-------
> 1 file changed, 5 insertions(+), 7 deletions(-)
>
> diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
> index e80c781731f8..555f09ca1d99 100644
> --- a/fs/jbd2/journal.c
> +++ b/fs/jbd2/journal.c
> @@ -2082,8 +2082,10 @@ int jbd2_journal_load(journal_t *journal)
>
> /* Let the recovery code check whether it needs to recover any
> * data from the journal. */
> - if (jbd2_journal_recover(journal))
> - goto recovery_error;
> + if (jbd2_journal_recover(journal)) {
> + printk(KERN_WARNING "JBD2: recovery failed\n");
> + return -EIO;
> + }
>
> if (journal->j_failed_commit) {
> printk(KERN_ERR "JBD2: journal transaction %u on %s "
> @@ -2101,14 +2103,10 @@ int jbd2_journal_load(journal_t *journal)
> * reinitialise the dynamic contents of the superblock in memory
> * and reset them on disk. */
> if (journal_reset(journal))
> - goto recovery_error;
> + return -EIO;

Why wouldn't you warn about the journal reset failing here?

--D

>
> journal->j_flags |= JBD2_LOADED;
> return 0;
> -
> -recovery_error:
> - printk(KERN_WARNING "JBD2: recovery failed\n");
> - return -EIO;
> }
>
> /**
> --
> 2.40.0
>

2023-04-11 10:43:35

by Jan Kara

Subject: Re: [PATCH] fs: jbd2: fix an incorrect warn log

On Mon 10-04-23 10:45:45, Darrick J. Wong wrote:
> On Tue, Apr 11, 2023 at 01:20:38AM +0800, Guoqing Cai wrote:
> > In jbd2_journal_load(), when journal_reset fails, it prints an incorrect
> > warn log.
> >
> > Fix this by changing the goto statement to return statement.
> >
> > Signed-off-by: Guoqing Cai <[email protected]>
> > Reviewed-by: Dongliang Mu <[email protected]>
> > ---
> > fs/jbd2/journal.c | 12 +++++-------
> > 1 file changed, 5 insertions(+), 7 deletions(-)
> >
> > diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
> > index e80c781731f8..555f09ca1d99 100644
> > --- a/fs/jbd2/journal.c
> > +++ b/fs/jbd2/journal.c
> > @@ -2082,8 +2082,10 @@ int jbd2_journal_load(journal_t *journal)
> >
> > /* Let the recovery code check whether it needs to recover any
> > * data from the journal. */
> > - if (jbd2_journal_recover(journal))
> > - goto recovery_error;
> > + if (jbd2_journal_recover(journal)) {
> > + printk(KERN_WARNING "JBD2: recovery failed\n");
> > + return -EIO;
> > + }
> >
> > if (journal->j_failed_commit) {
> > printk(KERN_ERR "JBD2: journal transaction %u on %s "
> > @@ -2101,14 +2103,10 @@ int jbd2_journal_load(journal_t *journal)
> > * reinitialise the dynamic contents of the superblock in memory
> > * and reset them on disk. */
> > if (journal_reset(journal))
> > - goto recovery_error;
> > + return -EIO;
>
> Why wouldn't you warn about the journal reset failing here?

Well, ext4 will still print:

ext4_msg(sb, KERN_ERR, "error loading journal");

which I guess is good enough for such strange failure.

Honza
--
Jan Kara <[email protected]>
SUSE Labs, CR