On Thu, Jul 24, 2008 at 11:57:15AM +0200, Patrick McHardy wrote:
...

IIRC the ESFQ for SFQ patches even use
"flows" in ->init, so it might make things a bit more ugly if we
ever merge those patches.

This argument looks right to me.

David, please, apply this reverting patch.

Thanks,
Jarek P.

-------------------->

pkt_sched: sch_sfq: revert "dump a real number of flows" patch

Patrick McHardy [off-list ref] noticed that since "flows" could be
used by sfq in the future as a parameter, we shouldn't currently
change it's meaning. This patch reverts my previous patch:
commit f867e6af94239a04ec23aeec2fcda5aa58e41db7


Signed-off-by: Jarek Poplawski <redacted>

---

diff -Nurp net-2.6-/net/sched/sch_sfq.c net-2.6+/net/sched/sch_sfq.c

--- net-2.6-/net/sched/sch_sfq.c	2008-07-24 10:26:08.000000000 +0000
+++ net-2.6+/net/sched/sch_sfq.c	2008-07-21 09:32:53.000000000 +0000

@@ -536,14 +536,7 @@ static int sfq_dump(struct Qdisc *sch, s
 
 	opt.limit = q->limit;
 	opt.divisor = SFQ_HASH_DIVISOR;
-	opt.flows = 0;
-	if (q->tail != SFQ_DEPTH) {
-		unsigned int i;
-
-		for (i = 0; i < SFQ_HASH_DIVISOR; i++)
-			if (q->ht[i] != SFQ_DEPTH)
-				opt.flows++;
-	}
+	opt.flows = q->limit;
 
 	NLA_PUT(skb, TCA_OPTIONS, sizeof(opt), &opt);