Fix the queue delay estimation in PIE/FQ-PIE when the timestamp

(TS) method is used.  When packet timestamp is used, the "current_qdelay"
keeps storing the last queue delay value calculated in the dequeue
function.  Therefore, when a burst of packets arrives followed by
a pause, the "current_qdelay" will store a high value caused by the
burst and stick to that value during the pause because the queue
delay measurement is done inside the dequeue function.  This causes
the drop probability calculation function to calculate high drop
probability value instead of zero and prevents the burst allowance
mechanism from working properly.  Fix this problem by resetting
"current_qdelay" inside the drop probability calculation function
when the queue length is zero and TS option is used.

Submitted by:	Rasool Al-Saadi <ralsaadi@swin.edu.au>
MFC after:	1 week

1 files changed, 8 insertions, 3 deletions

diff --git a/sys/netpfil/ipfw/dn_aqm_pie.c b/sys/netpfil/ipfw/dn_aqm_pie.c
index 8259d0853df6..c306a4caac9d 100644
--- a/sys/netpfil/ipfw/dn_aqm_pie.c
+++ b/sys/netpfil/ipfw/dn_aqm_pie.c
@@ -211,11 +211,16 @@ calculate_drop_prob(void *x)
 	pprms = pst->parms;
 	prob = pst->drop_prob;
 
-	/* calculate current qdelay */
-	if (pprms->flags & PIE_DEPRATEEST_ENABLED) {
+	/* calculate current qdelay using DRE method.
+	 * If TS is used and no data in the queue, reset current_qdelay
+	 * as it stays at last value during dequeue process. 
+	*/
+	if (pprms->flags & PIE_DEPRATEEST_ENABLED)
 		pst->current_qdelay = ((uint64_t)pst->pq->ni.len_bytes *
 			pst->avg_dq_time) >> PIE_DQ_THRESHOLD_BITS;
-	}
+	else 
+		if (!pst->pq->ni.len_bytes)
+			 pst->current_qdelay = 0;
 
 	/* calculate drop probability */
 	p = (int64_t)pprms->alpha *