Re: [PATCH 19/24] udf: check if udf_load_logicalvol failed

[Jan Kara <jack@suse.cz>]

On Sun 23-12-07 02:51:09, marcin.slusarz@gmail.com wrote:
> udf_load_logicalvol may fail eg in out of memory conditions - check it
> and propagate error further
> 
> Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
> CC: Ben Fennema <bfennema@falcon.csc.calpoly.edu>
> CC: Jan Kara <jack@suse.cz>
  I certainly agree with the idea of the change. Just note that
udf_process_sequence() returns 1 in case of failure so propagating an error
is not ideal - we should keep things consistent. And also note that the
actual value is disregarded in udf_load_partition() anyway.

									Honza
> ---
>  fs/udf/super.c |    7 ++++++-
>  1 files changed, 6 insertions(+), 1 deletions(-)
> 
> diff --git a/fs/udf/super.c b/fs/udf/super.c
> index 4aef612..f8c5794 100644
> --- a/fs/udf/super.c
> +++ b/fs/udf/super.c
> @@ -1141,6 +1141,7 @@ static int udf_process_sequence(struct super_block *sb, long block,
>  	uint32_t vdsn;
>  	uint16_t ident;
>  	long next_s = 0, next_e = 0;
> +	int ret;
>  
>  	memset(vds, 0, sizeof(struct udf_vds_record) * VDS_POS_LENGTH);
>  
> @@ -1216,7 +1217,11 @@ static int udf_process_sequence(struct super_block *sb, long block,
>  			if (i == VDS_POS_PRIMARY_VOL_DESC) {
>  				udf_load_pvoldesc(sb, bh);
>  			} else if (i == VDS_POS_LOGICAL_VOL_DESC) {
> -				udf_load_logicalvol(sb, bh, fileset); /* TODO: check return value */
> +				ret = udf_load_logicalvol(sb, bh, fileset);
> +				if (ret != 0) {
> +					brelse(bh);
> +					return ret;
> +				}
>  			} else if (i == VDS_POS_PARTITION_DESC) {
>  				struct buffer_head *bh2 = NULL;
>  				if (udf_load_partdesc(sb, bh)) {
-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR