On Wed, Apr 6, 2016 at 6:10 AM, Lars Persson <lars.pers...@axis.com> wrote:
> A failure in validate_xmit_skb_list() triggered an unconditional call
> to dev_requeue_skb with skb=NULL. This slowly grows the queue
> discipline's qlen count until all traffic through the queue stops.
>
Sounds reasonable.
> Fixes: 55a93b3ea780 ("qdisc: validate skb without holding lock")
> Signed-off-by: Lars Persson <lar...@axis.com>
> ---
> net/sched/sch_generic.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
> index f18c350..1031536 100644
> --- a/net/sched/sch_generic.c
> +++ b/net/sched/sch_generic.c
> @@ -165,6 +165,9 @@ int sch_direct_xmit(struct sk_buff *skb, struct Qdisc *q,
> skb = dev_hard_start_xmit(skb, dev, txq, &ret);
>
> HARD_TX_UNLOCK(dev, txq);
> + } else {
> + spin_lock(root_lock);
> + return qdisc_qlen(q);
I think we should return 0 for this failure case so that qdisc_restart()
will stop. How about teaching dev_requeue_skb() to skip skb==NULL
case?
