Re: [PATCH][next] dm zoned: fix memory leak of newly allocated zone on xa_insert failure

From: Mike Snitzer
Date: Wed Jun 03 2020 - 14:18:57 EST


On Wed, Jun 03 2020 at 12:02pm -0400,
Colin King <colin.king@xxxxxxxxxxxxx> wrote:

> From: Colin Ian King <colin.king@xxxxxxxxxxxxx>
>
> Currently if an xa_insert fails then there is a memory lead of the
> recently allocated zone object. Fix this by kfree'ing zone before
> returning on the error return path.
>
> Addresses-Coverity: ("Resource leak")
> Fixes: 1a311efa3916 ("dm zoned: convert to xarray")
> Signed-off-by: Colin Ian King <colin.king@xxxxxxxxxxxxx>
> ---
> drivers/md/dm-zoned-metadata.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/md/dm-zoned-metadata.c b/drivers/md/dm-zoned-metadata.c
> index b23ff090c056..130b5a6d9f12 100644
> --- a/drivers/md/dm-zoned-metadata.c
> +++ b/drivers/md/dm-zoned-metadata.c
> @@ -313,8 +313,10 @@ static struct dm_zone *dmz_insert(struct dmz_metadata *zmd,
> if (!zone)
> return ERR_PTR(-ENOMEM);
>
> - if (xa_insert(&zmd->zones, zone_id, zone, GFP_KERNEL))
> + if (xa_insert(&zmd->zones, zone_id, zone, GFP_KERNEL)) {
> + kfree(zone);
> return ERR_PTR(-EBUSY);
> + }
>
> INIT_LIST_HEAD(&zone->link);
> atomic_set(&zone->refcount, 0);
> --
> 2.25.1
>

Thanks, I folded this in.