diff mbox series

[net] net_sched: sch_mqprio: handle return value of mqprio_queue_get

Message ID 20190610063821.27007-1-jian.w.wen@oracle.com
State Changes Requested
Delegated to: David Miller
Headers show
Series [net] net_sched: sch_mqprio: handle return value of mqprio_queue_get | expand

Commit Message

Jacob Wen June 10, 2019, 6:38 a.m. UTC 
It may return NULL thus we can't ignore it.
---
 net/sched/sch_mqprio.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

Comments

David Miller June 10, 2019, 4:05 p.m. UTC | #1 
From: Jacob Wen <jian.w.wen@oracle.com>
Date: Mon, 10 Jun 2019 14:38:21 +0800

> It may return NULL thus we can't ignore it.

Please repost with a proper signoff.
David Miller June 10, 2019, 4:06 p.m. UTC | #2 
From: David Miller <davem@davemloft.net>
Date: Mon, 10 Jun 2019 09:05:26 -0700 (PDT)

> From: Jacob Wen <jian.w.wen@oracle.com>
> Date: Mon, 10 Jun 2019 14:38:21 +0800
> 
>> It may return NULL thus we can't ignore it.
> 
> Please repost with a proper signoff.

Also, you are breaking the reverse christmas tree ordering of the local
variables with your change.  Please do not do that.
Cong Wang June 10, 2019, 6:19 p.m. UTC | #3 
On Sun, Jun 9, 2019 at 11:41 PM Jacob Wen <jian.w.wen@oracle.com> wrote:
>
> It may return NULL thus we can't ignore it.

How is this possible? All of the callers should have validated
the 'cl' before calling this, for example by calling ->find().

I don't see it is possible to be NULL at this point.

Did you see a real crash? If so, please put the full stack trace
in your changelog.

Thanks!
Jacob Wen June 11, 2019, 12:27 a.m. UTC | #4 
Hi Cong,

This was detected by a tool.

Thanks for the review :)

I will abandon the patch.

On 6/11/19 2:19 AM, Cong Wang wrote:
> On Sun, Jun 9, 2019 at 11:41 PM Jacob Wen <jian.w.wen@oracle.com> wrote:
>> It may return NULL thus we can't ignore it.
> How is this possible? All of the callers should have validated
> the 'cl' before calling this, for example by calling ->find().
>
> I don't see it is possible to be NULL at this point.
>
> Did you see a real crash? If so, please put the full stack trace
> in your changelog.
>
> Thanks!
diff mbox series

Patch

diff --git a/net/sched/sch_mqprio.c b/net/sched/sch_mqprio.c
index d05086dc3866..d926056f72ac 100644
--- a/net/sched/sch_mqprio.c
+++ b/net/sched/sch_mqprio.c
@@ -491,9 +491,12 @@  static int mqprio_dump_class(struct Qdisc *sch, unsigned long cl,
 			 struct sk_buff *skb, struct tcmsg *tcm)
 {
 	if (cl < TC_H_MIN_PRIORITY) {
-		struct netdev_queue *dev_queue = mqprio_queue_get(sch, cl);
 		struct net_device *dev = qdisc_dev(sch);
 		int tc = netdev_txq_to_tc(dev, cl - 1);
+		struct netdev_queue *dev_queue = mqprio_queue_get(sch, cl);
+
+		if (!dev_queue)
+			return -EINVAL;
 
 		tcm->tcm_parent = (tc < 0) ? 0 :
 			TC_H_MAKE(TC_H_MAJ(sch->handle),
@@ -558,6 +561,8 @@  static int mqprio_dump_class_stats(struct Qdisc *sch, unsigned long cl,
 			return -1;
 	} else {
 		struct netdev_queue *dev_queue = mqprio_queue_get(sch, cl);
+		if (!dev_queue)
+			return -1;
 
 		sch = dev_queue->qdisc_sleeping;
 		if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch),