>> -static inline int dev_requeue_skb(struct sk_buff *skb, struct Qdisc *q) >> +static inline int __dev_requeue_skb(struct sk_buff *skb, struct Qdisc *q) >> { > > Perhaps dev_requeue_skb_qdisc_locked is more descriptive. Or > adding a lockdep_is_held(..) also documents that the __locked variant > below is not just a lock/unlock wrapper around this inner function. > >> - q->gso_skb = skb; >> + __skb_queue_head(&q->gso_skb, skb); >> q->qstats.requeues++; >> qdisc_qstats_backlog_inc(q, skb); >> q->q.qlen++; /* it's still part of the queue */ >> @@ -57,6 +56,30 @@ static inline int dev_requeue_skb(struct sk_buff *skb, >> struct Qdisc *q) >> return 0; >> } >> >> +static inline int dev_requeue_skb_locked(struct sk_buff *skb, struct Qdisc >> *q) >> +{ >> + spinlock_t *lock = qdisc_lock(q); >> + >> + spin_lock(lock); >> + __skb_queue_tail(&q->gso_skb, skb); > > why does this requeue at the tail, unlike __dev_requeue_skb?
I guess that requeue has to queue at the tail in the lockless case, and it does not matter in the qdisc_locked case, as then there can only ever be at most one outstanding gso_skb.