Mmh the last tc-s qdisc output did not look like working cake, but maybe the interface/device was wrong. Could I ask you again to post the output of:
tc -s qdisc
note this will be larger than the earlier output as it contains all interfaces, but there should be 2 cake instances in there.
1 Like
dtaht
82
I am impressed. A couple notes:
-
I was aware of the twitch streaming problem. I'm glad this seems to address it. Optimizing some for the increasingly common userspace vpn might also help - tossing wireguard traffic into the video queue for example. During covid I switched to diffserv4
for cake.
-
We showed in general (from the AP) that it was better to toss as much stuff as possible into the BE queue and to maximize aggregation, for greedy traffic. Going hog wild on using up the the wifi queues generally made things worse. What I wanted to propose at the time was that in a case where there were multiple packets scheduled for a station potentially using different hardware queues, that they be merged into a single txop. If there was a high priority packet in that stations queue(s), then everything would end up in the higher priority queue. I know this is hard, but if you look at the outputs of a rrul test vs a rrul_be test, or (worse) any of the multistation rrul_var tests, you'll see the damage currently caused by streaming traffic in each of these queues.
-
In general my focus is on deprioritizing packets (like torrent), not prioritizing them. I like very much the idea of a "blessed" default config containing all the known protocols and port numbers - or multiple such files. Having one enormous shared file that could be jointly maintained would help.
-
I have always preferred the idea of packets being marked and preserved end to end, not in the middleboxes, and applications marking their own traffic appropriately. Remarking from 0 to something else would be allowed, but not from one class to another, except in the case of abuse, or, in the case of LE, deprecation of CS1.
2 Likes
dtaht
83
It looks to me like qosify could actually also be used by applicatons running on clients. Has android gained ebf support yet?
1 Like
di_Niko
84
Note that my device its wan.20 so its strange that the qosify works only if I set device as
config interface wan
option name wan
option disabled 0
option bandwidth_up 1000mbit
option bandwidth_down 1000mbit
Summary
BusyBox v1.34.1 (2021-11-13 20:29:41 UTC) built-in shell (ash)
_______ ________ __
| |.-----.-----.-----.| | | |.----.| |_
| - || _ | -__| || | | || _|| _|
|_______|| __|_____|__|__||________||__| |____|
|__| W I R E L E S S F R E E D O M
-----------------------------------------------------
OpenWrt SNAPSHOT, r18084-b5ab6d0bc5
-----------------------------------------------------
root@OpenWrt:~# tc -s qdisc
qdisc noqueue 0: dev lo root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc mq 0: dev eth0 root
Sent 1121221 bytes 3340 pkt (dropped 0, overlimits 0 requeues 10)
backlog 0b 0p requeues 10
qdisc fq_codel 0: dev eth0 parent :8 limit 10240p flows 1024 quantum 1522 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
maxpacket 0 drop_overlimit 0 new_flow_count 0 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc fq_codel 0: dev eth0 parent :7 limit 10240p flows 1024 quantum 1522 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
maxpacket 0 drop_overlimit 0 new_flow_count 0 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc fq_codel 0: dev eth0 parent :6 limit 10240p flows 1024 quantum 1522 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
maxpacket 0 drop_overlimit 0 new_flow_count 0 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc fq_codel 0: dev eth0 parent :5 limit 10240p flows 1024 quantum 1522 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
maxpacket 0 drop_overlimit 0 new_flow_count 0 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc fq_codel 0: dev eth0 parent :4 limit 10240p flows 1024 quantum 1522 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
maxpacket 0 drop_overlimit 0 new_flow_count 0 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc fq_codel 0: dev eth0 parent :3 limit 10240p flows 1024 quantum 1522 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
maxpacket 0 drop_overlimit 0 new_flow_count 0 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc fq_codel 0: dev eth0 parent :2 limit 10240p flows 1024 quantum 1522 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
maxpacket 0 drop_overlimit 0 new_flow_count 0 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc fq_codel 0: dev eth0 parent :1 limit 10240p flows 1024 quantum 1522 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 1121221 bytes 3340 pkt (dropped 0, overlimits 0 requeues 10)
backlog 0b 0p requeues 10
maxpacket 1482 drop_overlimit 0 new_flow_count 4 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc noqueue 0: dev lan4 root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc noqueue 0: dev lan3 root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc noqueue 0: dev lan2 root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc noqueue 0: dev lan1 root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc noqueue 0: dev wan root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc noqueue 0: dev br-lan root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc noqueue 0: dev wan.20 root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc cake 1: dev pppoe-wan root refcnt 2 bandwidth 1Gbit diffserv4 dual-srchost nat nowash no-ack-filter split-gso rtt 100ms raw overhead 0
Sent 65512 bytes 200 pkt (dropped 0, overlimits 1 requeues 0)
backlog 0b 0p requeues 0
memory used: 8960b of 15180Kb
capacity estimate: 1Gbit
min/max network layer size: 40 / 1492
min/max overhead-adjusted size: 40 / 1492
average network hdr offset: 0
Bulk Best Effort Video Voice
thresh 62500Kbit 1Gbit 500Mbit 250Mbit
target 5ms 5ms 5ms 5ms
interval 100ms 100ms 100ms 100ms
pk_delay 0us 20us 11us 14us
av_delay 0us 0us 1us 1us
sp_delay 0us 0us 1us 1us
backlog 0b 0b 0b 0b
pkts 0 11 102 87
bytes 0 10010 15169 40333
way_inds 0 0 0 0
way_miss 0 1 10 45
way_cols 0 0 0 0
drops 0 0 0 0
marks 0 0 0 0
ack_drop 0 0 0 0
sp_flows 0 0 0 1
bk_flows 0 1 1 1
un_flows 0 0 0 0
max_len 0 5808 1859 1368
quantum 1514 1514 1514 1514
qdisc ingress ffff: dev pppoe-wan parent ffff:fff1 ----------------
Sent 80684 bytes 181 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc cake 1: dev ifb-pppoe-wan root refcnt 2 bandwidth 1Gbit diffserv4 dual-dsthost nat nowash ingress no-ack-filter split-gso rtt 100ms raw overhead 0
Sent 80684 bytes 181 pkt (dropped 0, overlimits 35 requeues 0)
backlog 0b 0p requeues 0
memory used: 6Kb of 15140Kb
capacity estimate: 1Gbit
min/max network layer size: 40 / 1492
min/max overhead-adjusted size: 40 / 1492
average network hdr offset: 0
Bulk Best Effort Video Voice
thresh 62500Kbit 1Gbit 500Mbit 250Mbit
target 5ms 5ms 5ms 5ms
interval 100ms 100ms 100ms 100ms
pk_delay 0us 6us 37us 18us
av_delay 0us 0us 2us 2us
sp_delay 0us 0us 2us 2us
backlog 0b 0b 0b 0b
pkts 0 5 95 81
bytes 0 2431 41418 36835
way_inds 0 0 0 0
way_miss 0 1 11 45
way_cols 0 0 0 0
drops 0 0 0 0
marks 0 0 0 0
ack_drop 0 0 0 0
sp_flows 0 0 0 1
bk_flows 0 0 1 1
un_flows 0 0 0 0
max_len 0 1452 1492 1452
quantum 1514 1514 1514 1514
root@OpenWrt:~#
dtaht
85
1 Like
That's my problem as well.
I set mine to pppoe-wan like I did for sqm-script and didn't work.
Set to wan and it works now tc showing it's working on the correct interface
2 Likes
hi @di_Niko replace by
option dscp_bulk bulk
by
option dscp_bulk CS1
like my post page 53 for bad argument and then
as i mentioned above do this
then so that it starts automatically add a file in hotplug iface and put this in this file ..
#!/bin/sh
[ -n "$DEVICE" ] || exit 0
[ "$ACTION" = ifup ] && {
echo y | /etc/init.d/qosify reload | start
logger -t qosify "Reloading qosify due to $ACTION of $INTERFACE ($DEVICE)"
}
like my post page 71
1 Like
IIRC is wan an OpenWrt internal name that needs translating into someting tc understands, which in your case seems to be pppoe-wan.... and ifb-pppoe-wan (not using sqm convention and calling this ifb4pppoe-wan, which given that qosify is not sqm seems reasonable).
3 Likes
That seems unfortunate, because it will probably no work with more interesting set-ups that avoid using the costly IFB indirection.
sorry i have another problem qosify generates this error
Sun Nov 14 14:27:38 2021 daemon.info procd: Instance qosify::instance1 s in a crash loop 6 crashes, 1 seconds since last crash
do you know what the problem could be?
@moeller0
my config actual is like this
config defaults
list defaults /etc/qosify/*.conf
option dscp_prio CS5
option dscp_icmp CS6
option dscp_bulk CS1
option dscp_default_udp CS4
option bulk_trigger_timeout 5
option bulk_trigger_pps 100
option prio_max_avg_pkt_len 500
config alias bulk
option ingress LE
option egress LE
config alias video
option ingress AF41
option egress AF41
config alias voice
option ingress CS6
option egress CS6
config interface wan
option name wan
option disabled 0
option bandwidth_up 4mbit
option bandwidth_down 56mbit
# defaults:
option ingress 1
option egress 1
option mode diffserv4
option nat 1
option host_isolate 1
option autorate_ingress 1
option ingress_options ""
option egress_options ""
option options "overhead 44"
config device wandev
option disabled 1
option name wan
option bandwidth 100mbit
(with diffserv4)
and i has put for my priority gaming +CS6
how can I be sure that my console is released as a priority for games
# DNS
tcp:53 CS5
tcp:5353 CS5
udp:53 CS5
udp:5353 CS5
# NTP
udp:123 CS6
# SSH
tcp:22 +CS4
# HTTP/QUIC
tcp:80 +CS3
tcp:443 +CS3
udp:80 +CS3
udp:443 +CS3
-->only priorize that
#cod
udp:3074 +CS6
#BF2042
udp:3659 +CS6
But sometimes i heard than cake the cs6 is good for wireless and EF for wired is right ?
@moeller just after restart qosify
my tc qdisc is like this, why
all my traffic for the video ?
qdisc noqueue 0: dev lo root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc fq_codel 0: dev eth0 root refcnt 2 limit 10240p flows 1024 quantum 1518 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 18801293244 bytes 26367621 pkt (dropped 0, overlimits 0 requeues 19)
backlog 0b 0p requeues 19
maxpacket 6072 drop_overlimit 0 new_flow_count 218 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc noqueue 0: dev lan1 root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc noqueue 0: dev lan2 root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc noqueue 0: dev lan3 root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc noqueue 0: dev lan4 root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc cake 1: dev wan root refcnt 2 bandwidth 4Mbit diffserv4 dual-srchost nat nowash no-ack-filter split-gso rtt 100ms noatm overhead 44
Sent 10495190 bytes 142687 pkt (dropped 8879, overlimits 293304 requeues 0)
backlog 5868b 78p requeues 0
memory used: 1242496b of 4Mb
capacity estimate: 4Mbit
min/max network layer size: 40 / 1500
min/max overhead-adjusted size: 84 / 1544
average network hdr offset: 14
Bulk Best Effort Video Voice
thresh 250Kbit 4Mbit 2Mbit 1Mbit
target 72.7ms 5ms 9.08ms 18.2ms
interval 168ms 100ms 104ms 113ms
pk_delay 0us 4.85ms 87.4ms 214us
av_delay 0us 272us 16ms 7us
sp_delay 0us 271us 907us 7us
backlog 0b 0b 5868b 0b
pkts 0 40 151674 28
bytes 0 2808 11100642 2314
way_inds 0 0 0 0
way_miss 0 5 15 23
way_cols 0 0 0 0
drops 0 0 8879 0
marks 0 0 0 0
ack_drop 0 0 0 0
sp_flows 0 5 3 2
bk_flows 0 0 3 0
un_flows 0 0 0 0
max_len 0 78 2039 400
quantum 300 300 300 300
qdisc ingress ffff: dev wan parent ffff:fff1 ----------------
Sent 165869120 bytes 111794 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc noqueue 0: dev br-lan root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc cake 1: dev ifb-wan root refcnt 2 bandwidth 56Mbit diffserv4 dual-dsthost nat nowash ingress no-ack-filter split-gso rtt 100ms noatm overhead 44
Sent 168184444 bytes 111786 pkt (dropped 8, overlimits 155374 requeues 0)
backlog 0b 0p requeues 0
memory used: 178104b of 4Mb
capacity estimate: 56Mbit
min/max network layer size: 46 / 1500
min/max overhead-adjusted size: 90 / 1544
average network hdr offset: 14
Bulk Best Effort Video Voice
thresh 3500Kbit 56Mbit 28Mbit 14Mbit
target 5.19ms 5ms 5ms 5ms
interval 100ms 100ms 100ms 100ms
pk_delay 0us 65us 1.02ms 166us
av_delay 0us 1us 299us 8us
sp_delay 0us 1us 29us 8us
backlog 0b 0b 0b 0b
pkts 0 5 111767 22
bytes 0 300 168194435 1821
way_inds 0 0 0 0
way_miss 0 3 15 18
way_cols 0 0 0 0
drops 0 0 8 0
marks 0 0 0 0
ack_drop 0 0 0 0
sp_flows 0 1 4 1
bk_flows 0 0 2 0
un_flows 0 0 0 0
max_len 0 60 3028 119
quantum 300 1514 854 427
root@OpenWrt:~#
All your normal unmarked web traffic is getting CS3 added, which maps to video. You can change it or comment it out.
1 Like
thanks @dave14305 i has uncomment
and now
Bulk Best Effort Video Voice
thresh 250Kbit 4Mbit 2Mbit 1Mbit
target 72.7ms 5ms 9.08ms 18.2ms
interval 168ms 100ms 104ms 113ms
pk_delay 2.94ms 12.8ms 0us 1.71ms
av_delay 276us 2.28ms 0us 292us
sp_delay 8us 156us 0us 7us
backlog 0b 0b 0b 0b
pkts 361960 430 0 5807
bytes 27463801 327562 0 840516
way_inds 0 0 0 0
way_miss 11 36 0 238
way_cols 0 0 0 0
drops 50 1 0 4
marks 0 0 0 0
ack_drop 0 0 0 0
sp_flows 1 1 0 1
bk_flows 1 0 0 0
un_flows 0 0 0 0
max_len 2047 4542 0 1606
quantum 300 300 300 300
qdisc ingress ffff: dev wan parent ffff:fff1 ----------------
Sent 413429057 bytes 282312 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc noqueue 0: dev br-lan root refcnt 2
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
backlog 0b 0p requeues 0
qdisc cake 1: dev ifb-wan root refcnt 2 bandwidth 56Mbit diffserv4 dual-dsthost nat nowash ingress no-ack-filter split-gso rtt 100ms noatm overhead 44
Sent 419028075 bytes 282299 pkt (dropped 11, overlimits 350939 requeues 0)
backlog 0b 0p requeues 0
memory used: 353568b of 4Mb
capacity estimate: 56Mbit
min/max network layer size: 46 / 1500
min/max overhead-adjusted size: 90 / 1544
average network hdr offset: 14
Bulk Best Effort Video Voice
thresh 3500Kbit 56Mbit 28Mbit 14Mbit
target 5.19ms 5ms 5ms 5ms
interval 100ms 100ms 100ms 100ms
pk_delay 230us 860us 0us 187us
av_delay 45us 252us 0us 64us
sp_delay 6us 7us 0us 5us
backlog 0b 0b 0b 0b
pkts 276240 1791 0 4279
bytes 415702038 2404131 0 938560
way_inds 0 0 0 0
way_miss 11 98 0 199
way_cols 0 0 0 0
drops 10 1 0 0
marks 0 0 0 0
ack_drop 0 0 0 0
sp_flows 1 1 0 2
bk_flows 1 0 0 0
un_flows 0 0 0 0
max_len 10598 4542 0 1514
quantum 300 1514 854 427
root@OpenWrt:~#
i suppose than traffic voice is my 2 rules in CS6 ?
#cod
udp:3074 +CS6
#BF2042
udp:3659 +CS6
Or DNS, NTP, or prio packets.
1 Like
As I noticed before cake as fq_codel are designed such that sparse protocols like DNS and NTP tend to not need additional prioritization, by virtue of being naturally sparse. I would probably first keep as much in CS0 as possible and only down-up prioritize if there is empirical evidence that doing so helps (or like with gaming stuff) there is strong a priori intent to prioritize no matter what (DNS will happily work with an occasional latency spike, on-line FPS games will tolerant, or rather the players 
)
1 Like
what do you recommend to improve the game
diffserv4 is a mode that I do not master, and I mainly play fps games, the rest does not matter to me
# DNS
#tcp:53 CS5
#tcp:5353 CS5
#udp:53 CS5
#udp:5353 CS5
# NTP
#udp:123 CS6
# SSH
#tcp:22 +CS4
# HTTP/QUIC
#tcp:80 +CS3
#tcp:443 +CS3
#udp:80 +CS3
#udp:443 +CS3
#cod
udp:3074 +CS6
#BF2042
udp:3659 +CS6
nbd
98
Sorry about the missing alias support. Turns out I forgot to push the init script change 
3 Likes
no problem
for the moment my game is bad compared to iptables and sqm but it's in development ...
my wireshark show all
but while i have changed to changed to CS6 why he not make apply the new dscp ?
#cod
udp:3074 +CS6
#BF2042
udp:3659 +CS6
anyone has an explication ?
hnyman
100
@nbd
Some questions & ideas for you:
-
Is the "interfaces" option intentionally missing from the default config file's "defaults" section?
I got qosify to work once I added there "option interfaces eth0.2".
-
Does the "name" option defined in the "interface wan" section have any actual meaning or relation to the netifd interface name? I left that as "wan", and qosify seems to still apply the settings for "eth0.2" and seems to work ok.
-
There could be clear example of a working config in README.
Possibly separately for a DSA based router (with "wan" etc.) and for a non-DSA router (with "eth0.2" etc.).
-
README (or possibly the default config file, so that the user actually sees it ??) might clearly explain the intended use and relations to other packages. E.g. apparently SQM should be disabled, so that qosify starts the cake qdisc. Or did I get that wrong?
3 Likes
