kernel_optimize_test/net/sched
Eric Dumazet ddecf0f4db net_sched: sfq: add optional RED on top of SFQ
Adds an optional Random Early Detection on each SFQ flow queue.

Traditional SFQ limits count of packets, while RED permits to also
control number of bytes per flow, and adds ECN capability as well.

1) We dont handle the idle time management in this RED implementation,
since each 'new flow' begins with a null qavg. We really want to address
backlogged flows.

2) if headdrop is selected, we try to ecn mark first packet instead of
currently enqueued packet. This gives faster feedback for tcp flows
compared to traditional RED [ marking the last packet in queue ]

Example of use :

tc qdisc add dev $DEV parent 1:1 handle 10: est 1sec 4sec sfq \
	limit 3000 headdrop flows 512 divisor 16384 \
	redflowlimit 100000 min 8000 max 60000 probability 0.20 ecn

qdisc sfq 10: parent 1:1 limit 3000p quantum 1514b depth 127 headdrop
flows 512/16384 divisor 16384
 ewma 6 min 8000b max 60000b probability 0.2 ecn
 prob_mark 0 prob_mark_head 4876 prob_drop 6131
 forced_mark 0 forced_mark_head 0 forced_drop 0
 Sent 1175211782 bytes 777537 pkt (dropped 6131, overlimits 11007
requeues 0)
 rate 99483Kbit 8219pps backlog 689392b 456p requeues 0

In this test, with 64 netperf TCP_STREAM sessions, 50% using ECN enabled
flows, we can see number of packets CE marked is smaller than number of
drops (for non ECN flows)

If same test is run, without RED, we can check backlog is much bigger.

qdisc sfq 10: parent 1:1 limit 3000p quantum 1514b depth 127 headdrop
flows 512/16384 divisor 16384
 Sent 1148683617 bytes 795006 pkt (dropped 0, overlimits 0 requeues 0)
 rate 98429Kbit 8521pps backlog 1221290b 841p requeues 0

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Stephen Hemminger <shemminger@vyatta.com>
CC: Dave Taht <dave.taht@gmail.com>
Tested-by: Dave Taht <dave.taht@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2012-01-12 20:05:28 -08:00
..
act_api.c net: Fix files explicitly needing to include module.h 2011-10-31 19:30:28 -04:00
act_csum.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
act_gact.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
act_ipt.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
act_mirred.c net_sched: fix port mirror/redirect stats reporting 2011-08-17 20:10:20 -07:00
act_nat.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
act_pedit.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
act_police.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
act_simple.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
act_skbedit.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
cls_api.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
cls_basic.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
cls_cgroup.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
cls_flow.c cls_flow: remove one dynamic array 2011-12-14 13:34:55 -05:00
cls_fw.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
cls_route.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
cls_rsvp6.c
cls_rsvp.c
cls_rsvp.h pkt_sched: cls_rsvp.h was outdated 2011-09-15 14:49:43 -04:00
cls_tcindex.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
cls_u32.c net: sched: constify tcf_proto and tc_action 2011-07-06 02:52:16 -07:00
em_cmp.c
em_meta.c net: remove SK_ROUTE_CAPS from meta ematch 2011-07-14 14:45:59 -07:00
em_nbyte.c
em_text.c
em_u32.c
ematch.c
Kconfig networking: NET_CLS_ROUTE4 depends on INET 2011-05-19 19:23:28 -04:00
Makefile pkt_sched: QFQ - quick fair queue scheduler 2011-04-04 11:10:24 -07:00
sch_api.c net_sched: qdisc_alloc_handle() can be too slow 2012-01-03 13:03:20 -05:00
sch_atm.c net: remove interrupt.h inclusion from netdevice.h 2011-06-06 22:55:11 -07:00
sch_blackhole.c
sch_cbq.c
sch_choke.c net_sched: red: split red_parms into parms and vars 2012-01-05 14:01:21 -05:00
sch_drr.c
sch_dsmark.c
sch_fifo.c
sch_generic.c net: Add queue state xoff flag for stack 2011-11-29 12:46:19 -05:00
sch_gred.c net_sched: red: split red_parms into parms and vars 2012-01-05 14:01:21 -05:00
sch_hfsc.c sch_hfsc: report backlog information 2011-12-23 16:51:18 -05:00
sch_htb.c Fix common misspellings 2011-03-31 11:26:23 -03:00
sch_ingress.c
sch_mq.c net: Add export.h for EXPORT_SYMBOL/THIS_MODULE to non-modules 2011-10-31 19:30:30 -04:00
sch_mqprio.c mqprio: Avoid panic if no options are provided 2011-12-22 22:34:56 -05:00
sch_multiq.c net: Add queue state xoff flag for stack 2011-11-29 12:46:19 -05:00
sch_netem.c net_sched: Bug in netem reordering 2012-01-05 13:27:39 -05:00
sch_prio.c net_sched: prio: use qdisc_dequeue_peeked 2011-08-09 21:52:11 -07:00
sch_qfq.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2012-01-04 21:35:43 -05:00
sch_red.c net_sched: red: split red_parms into parms and vars 2012-01-05 14:01:21 -05:00
sch_sfb.c sch_sfb: use skb_flow_dissect() 2011-11-28 19:09:28 -05:00
sch_sfq.c net_sched: sfq: add optional RED on top of SFQ 2012-01-12 20:05:28 -08:00
sch_tbf.c sch_tbf: report backlog information 2011-12-29 15:07:21 -05:00
sch_teql.c net: Rename dst_get_neighbour{, _raw} to dst_get_neighbour_noref{, _raw}. 2011-12-05 15:20:19 -05:00