linux/net/sched
Eric Dumazet f52ed89971 pkt_sched: fq: fix pacing for small frames
For performance reasons, sch_fq tried hard to not setup timers for every
sent packet, using a quantum based heuristic : A delay is setup only if
the flow exhausted its credit.

Problem is that application limited flows can refill their credit
for every queued packet, and they can evade pacing.

This problem can also be triggered when TCP flows use small MSS values,
as TSO auto sizing builds packets that are smaller than the default fq
quantum (3028 bytes)

This patch adds a 40 ms delay to guard flow credit refill.

Fixes: afe4fd0624 ("pkt_sched: fq: Fair Queue packet scheduler")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Maciej Żenczykowski <maze@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-11-15 21:01:52 -05:00
..
act_api.c
act_csum.c
act_gact.c
act_ipt.c
act_mirred.c
act_nat.c
act_pedit.c
act_police.c net_sched: add u64 rate to psched_ratecfg_precompute() 2013-09-20 14:41:02 -04:00
act_simple.c
act_skbedit.c
cls_api.c
cls_basic.c qdisc: basic classifier - remove unnecessary initialization 2013-09-30 15:47:43 -04:00
cls_bpf.c net: sched: cls_bpf: add BPF-based classifier 2013-10-29 17:33:17 -04:00
cls_cgroup.c cgroup: cls: remove unnecessary task_cls_classid 2013-10-08 16:27:34 -04:00
cls_flow.c
cls_fw.c
cls_route.c
cls_rsvp6.c
cls_rsvp.c
cls_rsvp.h
cls_tcindex.c
cls_u32.c
em_canid.c
em_cmp.c
em_ipset.c em_ipset: use dev_net() accessor 2013-10-18 16:23:06 -04:00
em_meta.c qdisc: meta return ENOMEM on alloc failure 2013-09-30 15:47:43 -04:00
em_nbyte.c
em_text.c
em_u32.c
ematch.c
Kconfig net: sched: cls_bpf: add BPF-based classifier 2013-10-29 17:33:17 -04:00
Makefile net: sched: cls_bpf: add BPF-based classifier 2013-10-29 17:33:17 -04:00
sch_api.c net_sched: increment drop counters in qdisc_tree_decrease_qlen() 2013-10-08 16:27:33 -04:00
sch_atm.c
sch_blackhole.c
sch_cbq.c
sch_choke.c treewide: Add __GFP_NOWARN to k.alloc calls with v.alloc fallbacks 2013-08-20 13:06:40 +02:00
sch_codel.c
sch_drr.c
sch_dsmark.c
sch_fifo.c
sch_fq_codel.c
sch_fq.c pkt_sched: fq: fix pacing for small frames 2013-11-15 21:01:52 -05:00
sch_generic.c net: Add layer 2 hardware acceleration operations for macvlan devices 2013-11-07 19:11:41 -05:00
sch_gred.c
sch_hfsc.c
sch_htb.c net_sched: htb: support of 64bit rates 2013-09-20 14:41:03 -04:00
sch_ingress.c
sch_mq.c qdisc: allow setting default queuing discipline 2013-08-31 00:32:32 -04:00
sch_mqprio.c qdisc: allow setting default queuing discipline 2013-08-31 00:32:32 -04:00
sch_multiq.c
sch_netem.c netem: markov loss model transition fix 2013-10-25 19:03:39 -04:00
sch_plug.c
sch_prio.c
sch_qfq.c
sch_red.c
sch_sfb.c
sch_sfq.c
sch_tbf.c net_sched: tbf: support of 64bit rates 2013-11-09 14:53:37 -05:00
sch_teql.c