Patrick McHardy wrote:
> Chris Madden wrote:
>
>>Ok, I replaced ingress_lock with queue_lock in ing_filter and it died in
>>the same place. Trace below (looks to be substantively the same)...
>>
>>If I am reading tc_ctl_tfilter correctly, we are adding our new
>>tcf_proto to the end of the list, and it is getting used before the
>>change function is getting executed on it. Locking the ingress_lock in
>>qdisc_lock_tree ( in addition to the extant queue_lock )seems to have
>>the same effect; I can get a traceback from that if its useful.
>
>
>
> The problem is that some classifiers (like basic, fw and route)
> implement empty ->init functions, although fw and route properly
> deal with it in their classification functions. The ->init function
> should allocate and initalize tp->root, which basic fails to do.
> If you give me a few hours I'll cook up a patch for this.
Actually its only cls_basic thats broken, in case of route and fw
its intentional for backwards-compatibility.
Can you try this patch please?
diff --git a/net/sched/cls_basic.c b/net/sched/cls_basic.c
index fad08e5..ed788d0 100644
--- a/net/sched/cls_basic.c
+++ b/net/sched/cls_basic.c
@@ -81,6 +81,13 @@ static void basic_put(struct tcf_proto *
static int basic_init(struct tcf_proto *tp)
{
+ struct basic_head *head;
+
+ head = kzalloc(sizeof(*head), GFP_KERNEL);
+ if (head == NULL)
+ return -ENOMEM;
+ INIT_LIST_HEAD(&head->flist);
+ tp->root = head;
return 0;
}
@@ -175,16 +182,6 @@ static int basic_change(struct tcf_proto
return basic_set_parms(tp, f, base, tb, tca[TCA_RATE-1]);
}
- err = -ENOBUFS;
- if (head == NULL) {
- head = kzalloc(sizeof(*head), GFP_KERNEL);
- if (head == NULL)
- goto errout;
-
- INIT_LIST_HEAD(&head->flist);
- tp->root = head;
- }
-
f = kzalloc(sizeof(*f), GFP_KERNEL);
if (f == NULL)
goto errout;
