The updating of 'bfqg->ref' should be protected by 'bfqd->lock', however,
during code review, we found that bfq_pd_free() update 'bfqg->ref'
without holding the lock, which is problematic:
1) bfq_pd_free() triggered by removing cgroup is called asynchronously;
2) bfqq will grab bfqg reference, and exit bfqq will drop the reference,
which can concurrent with 1).
Unfortunately, 'bfqd->lock' can't be held here because 'bfqd' might already
be freed in bfq_pd_free(). Fix the problem by using atomic refcount apis.
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20230103084755.1256479-1-yukuai1@huaweicloud.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
static void bfqg_get(struct bfq_group *bfqg)
{
- bfqg->ref++;
+ refcount_inc(&bfqg->ref);
}
static void bfqg_put(struct bfq_group *bfqg)
{
- bfqg->ref--;
-
- if (bfqg->ref == 0)
+ if (refcount_dec_and_test(&bfqg->ref))
kfree(bfqg);
}
}
/* see comments in bfq_bic_update_cgroup for why refcounting */
- bfqg_get(bfqg);
+ refcount_set(&bfqg->ref, 1);
return &bfqg->pd;
}
char blkg_path[128];
/* reference counter (see comments in bfq_bic_update_cgroup) */
- int ref;
+ refcount_t ref;
/* Is bfq_group still online? */
bool online;