openwrtv3/target/linux/generic/patches-4.4/660-fq_codel_defaults.patch

--- a/net/sched/sch_fq_codel.c
+++ b/net/sched/sch_fq_codel.c
@@ -440,8 +440,8 @@ static int fq_codel_init(struct Qdisc *s
 	struct fq_codel_sched_data *q = qdisc_priv(sch);
 	int i;
 
-	sch->limit = 10*1024;
-	q->flows_cnt = 1024;
+	sch->limit = 1024;
+	q->flows_cnt = 128;
 	q->drop_batch_size = 64;
 	q->quantum = psched_mtu(qdisc_dev(sch));
 	q->perturbation = prandom_u32();
kernel: add linux 4.4 support Based on 4.4-rc3. Runtime tested on MIPS. Signed-off-by: Jonas Gorski <jogo@openwrt.org> SVN-Revision: 47701 2015-12-02 22:23:22 +00:00			`--- a/net/sched/sch_fq_codel.c`
			`+++ b/net/sched/sch_fq_codel.c`
kernel: backport patches improving fq_codel drop behavior Signed-off-by: Felix Fietkau <nbd@nbd.name> 2016-05-12 19:04:50 +00:00			`@@ -440,8 +440,8 @@ static int fq_codel_init(struct Qdisc *s`
kernel: add linux 4.4 support Based on 4.4-rc3. Runtime tested on MIPS. Signed-off-by: Jonas Gorski <jogo@openwrt.org> SVN-Revision: 47701 2015-12-02 22:23:22 +00:00			`struct fq_codel_sched_data *q = qdisc_priv(sch);`
			`int i;`

			`- sch->limit = 10*1024;`
kernel: fq_codel match flows_cnt to limit sizing OpenWRT changed the default fq_codel sch->limit from 10240 to 1024, without also adjusting q->flows_cnt. Eric Dumazet explains below that you must also adjust the buckets (q->flows_cnt) for this not to break. Eric explains: Limit of 1024 packets and 1024 flows is not wise I think. (If all buckets are in use, each bucket has a virtual queue of 1 packet, which is almost the same than having no queue at all) I suggest to have at least 8 packets per bucket, to let Codel have a chance to trigger. So you could either reduce number of buckets to 128 (if memory is tight), or increase limit to 8192. flows_cnt is now set to 1024/8=128 Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk> 2016-05-07 08:57:18 +00:00			`- q->flows_cnt = 1024;`
kernel: add linux 4.4 support Based on 4.4-rc3. Runtime tested on MIPS. Signed-off-by: Jonas Gorski <jogo@openwrt.org> SVN-Revision: 47701 2015-12-02 22:23:22 +00:00			`+ sch->limit = 1024;`
kernel: fq_codel match flows_cnt to limit sizing OpenWRT changed the default fq_codel sch->limit from 10240 to 1024, without also adjusting q->flows_cnt. Eric Dumazet explains below that you must also adjust the buckets (q->flows_cnt) for this not to break. Eric explains: Limit of 1024 packets and 1024 flows is not wise I think. (If all buckets are in use, each bucket has a virtual queue of 1 packet, which is almost the same than having no queue at all) I suggest to have at least 8 packets per bucket, to let Codel have a chance to trigger. So you could either reduce number of buckets to 128 (if memory is tight), or increase limit to 8192. flows_cnt is now set to 1024/8=128 Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk> 2016-05-07 08:57:18 +00:00			`+ q->flows_cnt = 128;`
kernel: backport patches improving fq_codel drop behavior Signed-off-by: Felix Fietkau <nbd@nbd.name> 2016-05-12 19:04:50 +00:00			`q->drop_batch_size = 64;`
kernel: revert fq_codel quantum override to prevent it from causing too much cpu load with higher speed (#21326) Signed-off-by: Felix Fietkau <nbd@openwrt.org> SVN-Revision: 47811 2015-12-08 00:14:41 +00:00			`q->quantum = psched_mtu(qdisc_dev(sch));`
kernel: add linux 4.4 support Based on 4.4-rc3. Runtime tested on MIPS. Signed-off-by: Jonas Gorski <jogo@openwrt.org> SVN-Revision: 47701 2015-12-02 22:23:22 +00:00			`q->perturbation = prandom_u32();`