[PATCH] ext4: fix wrong return err in ext4_load_and_init_journal()

[PATCH] ext4: fix wrong return err in ext4_load_and_init_journal()

[Jason Yan]

The return value is wrong in ext4_load_and_init_journal(). The local
variable 'err' need to be initialized before goto out. The original code
in __ext4_fill_super() is fine because it has two return values 'ret'
and 'err' and 'ret' is initialized as -EINVAL. After we factor out
ext4_load_and_init_journal(), this code is broken. So fix it by directly
returning -EINVAL in the error handler path.

Fixes: 9c1dd22d7422 (ext4: factor out ext4_load_and_init_journal())
Signed-off-by: Jason Yan <yanaijie@huawei.com>
---
 fs/ext4/super.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 989365b878a6..89c6bad28a8a 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -4885,7 +4885,7 @@ static int ext4_load_and_init_journal(struct super_block *sb,
 	flush_work(&sbi->s_error_work);
 	jbd2_journal_destroy(sbi->s_journal);
 	sbi->s_journal = NULL;
-	return err;
+	return -EINVAL;
 }
 
 static int ext4_journal_data_mode_check(struct super_block *sb)
-- 
2.31.1

Re: [PATCH] ext4: fix wrong return err in ext4_load_and_init_journal()

[Jan Kara]

On Sat 22-10-22 21:07:39, Jason Yan wrote:
> The return value is wrong in ext4_load_and_init_journal(). The local
> variable 'err' need to be initialized before goto out. The original code
> in __ext4_fill_super() is fine because it has two return values 'ret'
> and 'err' and 'ret' is initialized as -EINVAL. After we factor out
> ext4_load_and_init_journal(), this code is broken. So fix it by directly
> returning -EINVAL in the error handler path.
> 
> Fixes: 9c1dd22d7422 (ext4: factor out ext4_load_and_init_journal())

We format the tag usually as:

Fixes: 9c1dd22d7422 ("ext4: factor out ext4_load_and_init_journal()")

> Signed-off-by: Jason Yan <yanaijie@huawei.com>

Otherwise the patch looks good. Feel free to add:

Reviewed-by: Jan Kara <jack@suse.cz>

								Honza

> ---
>  fs/ext4/super.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/ext4/super.c b/fs/ext4/super.c
> index 989365b878a6..89c6bad28a8a 100644
> --- a/fs/ext4/super.c
> +++ b/fs/ext4/super.c
> @@ -4885,7 +4885,7 @@ static int ext4_load_and_init_journal(struct super_block *sb,
>  	flush_work(&sbi->s_error_work);
>  	jbd2_journal_destroy(sbi->s_journal);
>  	sbi->s_journal = NULL;
> -	return err;
> +	return -EINVAL;
>  }
>  
>  static int ext4_journal_data_mode_check(struct super_block *sb)
> -- 
> 2.31.1
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

Re: [PATCH] ext4: fix wrong return err in ext4_load_and_init_journal()

[Jason Yan]

On 2022/10/24 23:29, Jan Kara wrote:
> On Sat 22-10-22 21:07:39, Jason Yan wrote:
>> The return value is wrong in ext4_load_and_init_journal(). The local
>> variable 'err' need to be initialized before goto out. The original code
>> in __ext4_fill_super() is fine because it has two return values 'ret'
>> and 'err' and 'ret' is initialized as -EINVAL. After we factor out
>> ext4_load_and_init_journal(), this code is broken. So fix it by directly
>> returning -EINVAL in the error handler path.
>>
>> Fixes: 9c1dd22d7422 (ext4: factor out ext4_load_and_init_journal())
> 
> We format the tag usually as:
> 
> Fixes: 9c1dd22d7422 ("ext4: factor out ext4_load_and_init_journal()")
> 

Oh, sorry I didn't notice it. Thank you so much.

I generate this tag by the following script:

#cat .gitconfig
  [alias]
          fixes = log --abbrev=12 -1 --format='Fixes: %h ("%s")'


#git fixes 9c1dd22d742249cfae7bbf3680a7c188d194d3ce
Fixes: 9c1dd22d7422 (ext4: factor out ext4_load_and_init_journal())

This works fine before but it fails recently. I don't know what makes 
the behavior changed.

However if I directly call the whole command:
#git log --abbrev=12 -1 --format='Fixes: %h ("%s")' 9c1dd22d742249cfae
Fixes: 9c1dd22d7422 ("ext4: factor out ext4_load_and_init_journal()")

It works fine.

>> Signed-off-by: Jason Yan <yanaijie@huawei.com>
> 
> Otherwise the patch looks good. Feel free to add:
> 
> Reviewed-by: Jan Kara <jack@suse.cz>
> 

Thanks.

> 								Honza
> 
>> ---
>>   fs/ext4/super.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/fs/ext4/super.c b/fs/ext4/super.c
>> index 989365b878a6..89c6bad28a8a 100644
>> --- a/fs/ext4/super.c
>> +++ b/fs/ext4/super.c
>> @@ -4885,7 +4885,7 @@ static int ext4_load_and_init_journal(struct super_block *sb,
>>   	flush_work(&sbi->s_error_work);
>>   	jbd2_journal_destroy(sbi->s_journal);
>>   	sbi->s_journal = NULL;
>> -	return err;
>> +	return -EINVAL;
>>   }
>>   
>>   static int ext4_journal_data_mode_check(struct super_block *sb)
>> -- 
>> 2.31.1
>>

Re: [PATCH] ext4: fix wrong return err in ext4_load_and_init_journal()

[Jan Kara]

On Tue 25-10-22 11:27:01, Jason Yan wrote:
> 
> On 2022/10/24 23:29, Jan Kara wrote:
> > On Sat 22-10-22 21:07:39, Jason Yan wrote:
> > > The return value is wrong in ext4_load_and_init_journal(). The local
> > > variable 'err' need to be initialized before goto out. The original code
> > > in __ext4_fill_super() is fine because it has two return values 'ret'
> > > and 'err' and 'ret' is initialized as -EINVAL. After we factor out
> > > ext4_load_and_init_journal(), this code is broken. So fix it by directly
> > > returning -EINVAL in the error handler path.
> > > 
> > > Fixes: 9c1dd22d7422 (ext4: factor out ext4_load_and_init_journal())
> > 
> > We format the tag usually as:
> > 
> > Fixes: 9c1dd22d7422 ("ext4: factor out ext4_load_and_init_journal()")
> > 
> 
> Oh, sorry I didn't notice it. Thank you so much.
> 
> I generate this tag by the following script:
> 
> #cat .gitconfig
>  [alias]
>          fixes = log --abbrev=12 -1 --format='Fixes: %h ("%s")'
> 
> 
> #git fixes 9c1dd22d742249cfae7bbf3680a7c188d194d3ce
> Fixes: 9c1dd22d7422 (ext4: factor out ext4_load_and_init_journal())
> 
> This works fine before but it fails recently. I don't know what makes the
> behavior changed.

I guess something does one more round of expansion on the string. I guess
you could trace it with strace and see where the quotes get lost...

								Honza
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR