sch_hfsc: make hfsc_qlen_notify() idempotent

hfsc_qlen_notify() is not idempotent either and not friendly
to its callers, like fq_codel_dequeue(). Let's make it idempotent
to ease qdisc_tree_reduce_backlog() callers' life:

1. update_vf() decreases cl->cl_nactive, so we can check whether it is
non-zero before calling it.

2. eltree_remove() always removes RB node cl->el_node, but we can use
   RB_EMPTY_NODE() + RB_CLEAR_NODE() to make it safe.

Reported-by: Gerrard Tai <[email protected]>
Signed-off-by: Cong Wang <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Link: https://patch.msgid.link/[email protected]
Acked-by: Jamal Hadi Salim <[email protected]>
Signed-off-by: Paolo Abeni <[email protected]>

Author: congwang

net/sched/sch_hfsc.c

@@ -203,7 +203,10 @@ eltree_insert(struct hfsc_class *cl)
 static inline void
 eltree_remove(struct hfsc_class *cl)
 {
-	rb_erase(&cl->el_node, &cl->sched->eligible);
+	if (!RB_EMPTY_NODE(&cl->el_node)) {
+		rb_erase(&cl->el_node, &cl->sched->eligible);
+		RB_CLEAR_NODE(&cl->el_node);
+	}
 }
 
 static inline void
@@ -1220,7 +1223,8 @@ hfsc_qlen_notify(struct Qdisc *sch, unsigned long arg)
 	/* vttree is now handled in update_vf() so that update_vf(cl, 0, 0)
 	 * needs to be called explicitly to remove a class from vttree.
 	 */
-	update_vf(cl, 0, 0);
+	if (cl->cl_nactive)
+		update_vf(cl, 0, 0);
 	if (cl->cl_flags & HFSC_RSC)
 		eltree_remove(cl);
 }