On Wed, Jul 12, 2017 at 09:32:15AM +0300, Nikolay Borisov wrote:
> insert_into_bitmap has only one caller which always allocates the info struct
> passed. As such remove the any NULL checks for info and also remove code
> to allocate info in case it was NULL.
>
> Signed-off-by: Nikolay Borisov <nborisov@xxxxxxxx>
> ---
> fs/btrfs/free-space-cache.c | 13 +------------
> 1 file changed, 1 insertion(+), 12 deletions(-)
>
> diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
> index c5e6180cdb8c..fd24fb99d6dc 100644
> --- a/fs/btrfs/free-space-cache.c
> +++ b/fs/btrfs/free-space-cache.c
> @@ -2101,7 +2101,7 @@ static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl,
> goto again;
>
> new_bitmap:
> - if (info && info->bitmap) {
> + if (info->bitmap) {
> add_new_bitmap(ctl, info, offset);
> added = 1;
> info = NULL;
What if we reach this point, go back to label "again:", come back to
this check again, then info would be NULL and dereferencing info->bitmap
would crash. Then the below code is still required.
> @@ -2109,17 +2109,6 @@ static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl,
> } else {
> spin_unlock(&ctl->tree_lock);
>
> - /* no pre-allocated info, allocate a new one */
> - if (!info) {
> - info = kmem_cache_zalloc(btrfs_free_space_cachep,
> - GFP_NOFS);
> - if (!info) {
> - spin_lock(&ctl->tree_lock);
> - ret = -ENOMEM;
> - goto out;
> - }
> - }
> -
> /* allocate the bitmap */
> info->bitmap = kzalloc(PAGE_SIZE, GFP_NOFS);
> spin_lock(&ctl->tree_lock);
> --
> 2.7.4
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
