Hello, Jens.
How should this one be routed? It can go through either block or
cgroup tree.
Thanks!
------------------------- 8< -------------------------
From: Tejun Heo <tj@xxxxxxxxxx>
Subject: blkcg: make blkcg_policy_register() correctly handle cgroup_add_cftypes() failures
blkcg_policy_register() currently triggers WARN when
cgroup_add_cftypes() fails and proceeds, which is a rather crappy way
to handle errors. It was written that way as cgroup_add_cftypes()
itself didn't handle errors correctly. Now that cgroup_add_cftypes()
correctly handles errors in itself, let's handle its failure properly.
Remove the WARN_ON() and cancel and fail policy registration on
cgroup_add_cftypes() failure.
Signed-off-by: Tejun Heo <tj@xxxxxxxxxx>
Cc: Vivek Goyal <vgoyal@xxxxxxxxxx>
Cc: Jens Axboe <axboe@xxxxxxxxx>
---
block/blk-cgroup.c | 9 ++++++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index e8918ff..b27a9d2 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -1127,10 +1127,13 @@ int blkcg_policy_register(struct blkcg_policy *pol)
pol->plid = i;
blkcg_policy[i] = pol;
- /* everything is in place, add intf files for the new policy */
- if (pol->cftypes)
- WARN_ON(cgroup_add_cftypes(&blkio_subsys, pol->cftypes));
+ /* try to add intf files for the new policy */
ret = 0;
+ if (pol->cftypes) {
+ ret = cgroup_add_cftypes(&blkio_subsys, pol->cftypes);
+ if (ret)
+ blkcg_policy[i] = NULL;
+ }
out_unlock:
mutex_unlock(&blkcg_pol_mutex);
return ret;
--
1.8.3.1
_______________________________________________
Containers mailing list
Containers@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linuxfoundation.org/mailman/listinfo/containers
