On Thu, Apr 22, 2010 at 05:32:28PM +0800, Li Zefan wrote:
> with CONFIG_PROVE_RCU=y, a warning can be triggered:
>
> # mount -t cgroup -o blkio xxx /mnt
> # mkdir /mnt/subgroup
>
> ...
> kernel/cgroup.c:4442 invoked rcu_dereference_check() without protection!
> ...
>
IIUC, so blkiocg_create() is being called with cgroup_mutex held and not
with rcu read lock held. Hence rcu_dereference() in css_depth() gives
warning.
So one easy solution is to don't use css_depth() at all. In this case
simple check like cgroup->top_cgroup should suffice. Makese sense to
me.
Acked-by: Vivek Goyal <vgoyal@xxxxxxxxxx>
Thanks
Vivek
> To fix this, we avoid caling css_depth() here, which is a bit simpler
> than the original code.
>
> Signed-off-by: Li Zefan <lizf@xxxxxxxxxxxxxx>
> ---
> block/blk-cgroup.c | 8 ++++----
> 1 files changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
> index 5fe03de..2cc682b 100644
> --- a/block/blk-cgroup.c
> +++ b/block/blk-cgroup.c
> @@ -286,16 +286,16 @@ done:
> static struct cgroup_subsys_state *
> blkiocg_create(struct cgroup_subsys *subsys, struct cgroup *cgroup)
> {
> - struct blkio_cgroup *blkcg, *parent_blkcg;
> + struct blkio_cgroup *blkcg;
> + struct cgroup *parent = cgroup->parent;
>
> - if (!cgroup->parent) {
> + if (!parent) {
> blkcg = &blkio_root_cgroup;
> goto done;
> }
>
> /* Currently we do not support hierarchy deeper than two level (0,1) */
> - parent_blkcg = cgroup_to_blkio_cgroup(cgroup->parent);
> - if (css_depth(&parent_blkcg->css) > 0)
> + if (parent != cgroup->top_cgroup)
> return ERR_PTR(-EINVAL);
>
> blkcg = kzalloc(sizeof(*blkcg), GFP_KERNEL);
> --
> 1.6.3
_______________________________________________
Containers mailing list
Containers@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linux-foundation.org/mailman/listinfo/containers
