net: sched: red: delay destroying child qdisc on replace

Move destroying of the old child qdisc outside of the sch_tree_lock()
section.  This should improve the software qdisc replace but is even
more important for offloads.  Firstly calling offloads under a spin
lock is best avoided.  Secondly the destroy event of existing child
would have been sent to the offload device before the replace, causing
confusion.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: John Hurley <john.hurley@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/sched/sch_red.c

@@ -193,10 +193,10 @@ static const struct nla_policy red_policy[TCA_RED_MAX + 1] = {
 static int red_change(struct Qdisc *sch, struct nlattr *opt,
 		      struct netlink_ext_ack *extack)
 {
+	struct Qdisc *old_child = NULL, *child = NULL;
 	struct red_sched_data *q = qdisc_priv(sch);
 	struct nlattr *tb[TCA_RED_MAX + 1];
 	struct tc_red_qopt *ctl;
-	struct Qdisc *child = NULL;
 	int err;
 	u32 max_P;
 
@@ -233,7 +233,7 @@ static int red_change(struct Qdisc *sch, struct nlattr *opt,
 	if (child) {
 		qdisc_tree_reduce_backlog(q->qdisc, q->qdisc->q.qlen,
 					  q->qdisc->qstats.backlog);
-		qdisc_put(q->qdisc);
+		old_child = q->qdisc;
 		q->qdisc = child;
 	}
 
@@ -252,7 +252,11 @@ static int red_change(struct Qdisc *sch, struct nlattr *opt,
 		red_start_of_idle_period(&q->vars);
 
 	sch_tree_unlock(sch);
+
 	red_offload(sch, true);
+
+	if (old_child)
+		qdisc_put(old_child);
 	return 0;
 }