On Wed, Jun 23, 2021 at 04:21:14PM -0700, Stephen Brennan wrote:
> Commit 5d1b1b3f492f ("ext4: fix BUG when calling ext4_error with locked
> block group") introduces ext4_grp_locked_error to handle unlocking a
> group in error cases. Otherwise, there is a possibility of a sleep while
> atomic. However, since 43c73221b3b1 ("ext4: replace BUG_ON with WARN_ON
> in mb_find_extent()"), mb_find_extent() has contained a ext4_error()
> call while a group spinlock is held. Replace this with
> ext4_grp_locked_error.
>
> Fixes: 43c73221b3b1 ("ext4: replace BUG_ON with WARN_ON in mb_find_extent()")
> Cc: <stable@xxxxxxxxxxxxxxx> # 4.14+
> Signed-off-by: Stephen Brennan <stephen.s.brennan@xxxxxxxxxx>
> Reviewed-by: Junxiao Bi <junxiao.bi@xxxxxxxxxx>
Good catch, looks good to me.
Reviewed-by: Lukas Czerner <lczerner@xxxxxxxxxx>
Thanks!
-Lukas
> ---
> fs/ext4/mballoc.c | 9 +++++----
> 1 file changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
> index c2c22c2baac0..089c958aa2c3 100644
> --- a/fs/ext4/mballoc.c
> +++ b/fs/ext4/mballoc.c
> @@ -1909,10 +1909,11 @@ static int mb_find_extent(struct ext4_buddy *e4b, int block,
> if (ex->fe_start + ex->fe_len > EXT4_CLUSTERS_PER_GROUP(e4b->bd_sb)) {
> /* Should never happen! (but apparently sometimes does?!?) */
> WARN_ON(1);
> - ext4_error(e4b->bd_sb, "corruption or bug in mb_find_extent "
> - "block=%d, order=%d needed=%d ex=%u/%d/%d@%u",
> - block, order, needed, ex->fe_group, ex->fe_start,
> - ex->fe_len, ex->fe_logical);
> + ext4_grp_locked_error(e4b->bd_sb, e4b->bd_group, 0, 0,
> + "corruption or bug in mb_find_extent "
> + "block=%d, order=%d needed=%d ex=%u/%d/%d@%u",
> + block, order, needed, ex->fe_group, ex->fe_start,
> + ex->fe_len, ex->fe_logical);
> ex->fe_len = 0;
> ex->fe_start = 0;
> ex->fe_group = 0;
> --
> 2.30.2
>
