[PATCH] fix up thinko in jbd2 revoke code

[PATCH] fix up thinko in jbd2 revoke code

[James Bottomley]

This commit:

commit e898f3154d78d1520265aade0992981866d827fb
Author: Duane Griffin <duaneg@dghda.com>
Date:   Tue Mar 25 19:47:39 2008 -0400

    jbd2: tidy up revoke cache initialisation and destruction

Introduced a thinko (it doesn't check whether jbd2_revoke_record cache
is null, it checks jbd2_revoke_table_cache twice).

Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>

---

diff --git a/fs/jbd2/revoke.c b/fs/jbd2/revoke.c
index e13584a..4dc5b85 100644
--- a/fs/jbd2/revoke.c
+++ b/fs/jbd2/revoke.c
@@ -192,7 +192,7 @@ int __init jbd2_journal_init_revoke_caches(void)
 	jbd2_revoke_table_cache = kmem_cache_create("jbd2_revoke_table",
 					   sizeof(struct jbd2_revoke_table_s),
 					   0, SLAB_TEMPORARY, NULL);
-	if (jbd2_revoke_table_cache && jbd2_revoke_table_cache) {
+	if (jbd2_revoke_record_cache && jbd2_revoke_table_cache) {
 		return 0;
 	} else {
 		jbd2_journal_destroy_revoke_caches();

Re: [PATCH] fix up thinko in jbd2 revoke code

[Mingming Cao]

On Mon, 2008-03-31 at 10:57 -0500, James Bottomley wrote:
> This commit:
> 
> commit e898f3154d78d1520265aade0992981866d827fb
> Author: Duane Griffin <duaneg@dghda.com>
> Date:   Tue Mar 25 19:47:39 2008 -0400
> 
>     jbd2: tidy up revoke cache initialisation and destruction
> 
> Introduced a thinko (it doesn't check whether jbd2_revoke_record cache
> is null, it checks jbd2_revoke_table_cache twice).
> 
> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
> 
Signed-off-by: Mingming Cao <cmm@us.ibm.com>

Thanks, I picked this up in ext4 patch queue.

Mingming
> ---
> 
> diff --git a/fs/jbd2/revoke.c b/fs/jbd2/revoke.c
> index e13584a..4dc5b85 100644
> --- a/fs/jbd2/revoke.c
> +++ b/fs/jbd2/revoke.c
> @@ -192,7 +192,7 @@ int __init jbd2_journal_init_revoke_caches(void)
>  	jbd2_revoke_table_cache = kmem_cache_create("jbd2_revoke_table",
>  					   sizeof(struct jbd2_revoke_table_s),
>  					   0, SLAB_TEMPORARY, NULL);
> -	if (jbd2_revoke_table_cache && jbd2_revoke_table_cache) {
> +	if (jbd2_revoke_record_cache && jbd2_revoke_table_cache) {
>  		return 0;
>  	} else {
>  		jbd2_journal_destroy_revoke_caches();
> 
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] fix up thinko in jbd2 revoke code

[Christoph Hellwig]

On Mon, Mar 31, 2008 at 11:30:58AM -0700, Mingming Cao wrote:
> On Mon, 2008-03-31 at 10:57 -0500, James Bottomley wrote:
> > This commit:
> > 
> > commit e898f3154d78d1520265aade0992981866d827fb
> > Author: Duane Griffin <duaneg@dghda.com>
> > Date:   Tue Mar 25 19:47:39 2008 -0400
> > 
> >     jbd2: tidy up revoke cache initialisation and destruction
> > 
> > Introduced a thinko (it doesn't check whether jbd2_revoke_record cache
> > is null, it checks jbd2_revoke_table_cache twice).
> > 
> > Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
> > 
> Signed-off-by: Mingming Cao <cmm@us.ibm.com>
> 
> Thanks, I picked this up in ext4 patch queue.

anyone care to convert this function to proper goto-based error
unwinding?

Re: [PATCH] fix up thinko in jbd2 revoke code

[Aneesh Kumar K.V]

On Thu, Apr 03, 2008 at 12:17:27PM -0400, Christoph Hellwig wrote:
> On Mon, Mar 31, 2008 at 11:30:58AM -0700, Mingming Cao wrote:
> > On Mon, 2008-03-31 at 10:57 -0500, James Bottomley wrote:
> > > This commit:
> > > 
> > > commit e898f3154d78d1520265aade0992981866d827fb
> > > Author: Duane Griffin <duaneg@dghda.com>
> > > Date:   Tue Mar 25 19:47:39 2008 -0400
> > > 
> > >     jbd2: tidy up revoke cache initialisation and destruction
> > > 
> > > Introduced a thinko (it doesn't check whether jbd2_revoke_record cache
> > > is null, it checks jbd2_revoke_table_cache twice).
> > > 
> > > Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
> > > 
> > Signed-off-by: Mingming Cao <cmm@us.ibm.com>
> > 
> > Thanks, I picked this up in ext4 patch queue.
> 
> anyone care to convert this function to proper goto-based error
> unwinding?

Like below. ?

Mingming,

I folded jbd2-tidy-up-revoe-cache-intialisation-and-destruction.patch
jbd2-fix-up-thinko-in-jbd2-revoke-code.patch in the patch queue to 
this.

-aneesh


jbd2: tidy up revoke cache initialisation and destruction

From: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>

Make revocation cache destruction safe to call if initialisation fails
partially or entirely.  This allows it to be used to cleanup in the case of
initialisation failure, simplifying the code slightly.

Signed-off-by: Duane Griffin <duaneg@dghda.com>
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
---

 fs/jbd2/revoke.c |   36 ++++++++++++++++++++++--------------
 1 files changed, 22 insertions(+), 14 deletions(-)


diff --git a/fs/jbd2/revoke.c b/fs/jbd2/revoke.c
index 2e1453a..72b2608 100644
--- a/fs/jbd2/revoke.c
+++ b/fs/jbd2/revoke.c
@@ -167,33 +167,41 @@ static struct jbd2_revoke_record_s *find_revoke_record(journal_t *journal,
 	return NULL;
 }
 
+void jbd2_journal_destroy_revoke_caches(void)
+{
+	if (jbd2_revoke_record_cache) {
+		kmem_cache_destroy(jbd2_revoke_record_cache);
+		jbd2_revoke_record_cache = NULL;
+	}
+	if (jbd2_revoke_table_cache) {
+		kmem_cache_destroy(jbd2_revoke_table_cache);
+		jbd2_revoke_table_cache = NULL;
+	}
+}
+
 int __init jbd2_journal_init_revoke_caches(void)
 {
+	J_ASSERT(!jbd2_revoke_record_cache);
+	J_ASSERT(!jbd2_revoke_table_cache);
+
 	jbd2_revoke_record_cache = kmem_cache_create("jbd2_revoke_record",
 					   sizeof(struct jbd2_revoke_record_s),
 					   0,
 					   SLAB_HWCACHE_ALIGN|SLAB_TEMPORARY,
 					   NULL);
 	if (!jbd2_revoke_record_cache)
-		return -ENOMEM;
+		goto record_cache_failure;
 
 	jbd2_revoke_table_cache = kmem_cache_create("jbd2_revoke_table",
 					   sizeof(struct jbd2_revoke_table_s),
 					   0, SLAB_TEMPORARY, NULL);
-	if (!jbd2_revoke_table_cache) {
-		kmem_cache_destroy(jbd2_revoke_record_cache);
-		jbd2_revoke_record_cache = NULL;
-		return -ENOMEM;
-	}
+	if (!jbd2_revoke_table_cache)
+		goto table_cache_failure;
 	return 0;
-}