Hi!
I had my SQM settings dialed for an A+ on Bufferbloat, but after adding a dumb AP downstream, I am experiencing latency again on the Bufferbloat tests. Have I maybe configured something incorrectly? I followed the instructions on: https://openwrt.org/docs/guide-user/network/wifi/wifiextenders/bridgedap
My main router is a Nanopi r4s that has an old router (NETGER r6900 v2) downstream of it. Any help or direction would be much appreciated. Thank you!
Install bridger on the downstream router, and post links to waveform test results. Saying A B C does not help anything.
A+ is zero latency under load for download and upload. But thank you for the advice, I will look into bridger.
Show comparative results. You can add simple tbfs both ways on the bridge's exit to wire to make CPU relaxed.
So I take it you are comparing the measured bufferbloat between:
A) router (Nanopi r4s) <-ethernet-> test computer
B) router (Nanopi r4s) <-ethernet-> AP (NETGEAR r6900 v2) <-WiFi-> test computer
And with A) you get A+ but with B) you get lower scores?
This would imply that maybe the WiFi is the bottleneck and even good WiFi generally has larger queueing delays under load.
But please elaborate on your tests and if possible post screenshots and links to the test results for the good and bad situation.
Also, could you post the output of the following commands please?
ifstatus wan
cat /etc/config/sqm
tc -s qdisc
Gotchya, sorry I am a new openwrt user. So this is odd, I'm getting good results again. I haven't changed anything, the old router is still downstream of the new one and here are the results: https://www.waveform.com/tools/bufferbloat?test-id=a001c8c4-9ac4-4a6d-8b32-50609628dd54
Trivially mt7621 can forward 300-400Mbps without offload, and double with offload (bridger)
Sorry, I was not detailed enough in my original post. Test "A" is as you described, however I'm now connected to the AP via ethernet. So it would look like:
B) router (Nanopi r4s) <-ethernet-> AP (NETGEAR r6900 v2) <-Ethernet-> test computer
My current bufferbloat grade with that setup is fine all of the sudden...? Here's the results
ifstatus wan
{
"up": true,
"pending": false,
"available": true,
"autostart": true,
"dynamic": false,
"uptime": 7423,
"l3_device": "eth0",
"proto": "dhcp",
"device": "eth0",
"updated": [
"addresses",
"routes",
"data"
],
"metric": 0,
"dns_metric": 0,
"delegation": true,
"ipv4-address": [
{
"address": "xxxx",
"mask": 20
}
],
"ipv6-address": [
],
"ipv6-prefix": [
],
"ipv6-prefix-assignment": [
],
"route": [
{
"target": "0.0.0.0",
"mask": 0,
"nexthop": "xxx",
"source": "xxx"
}
],
"dns-server": [
"8.8.8.8",
"8.8.4.4"
],
"dns-search": [
],
"neighbors": [
],
"inactive": {
"ipv4-address": [
],
"ipv6-address": [
],
"route": [
],
"dns-server": [
],
"dns-search": [
],
"neighbors": [
]
},
"data": {
"dhcpserver": "xxxx",
"hostname": "xxx",
"leasetime": 85200,
"ntpserver": "xxxx"
}
}
root@xxxx:~# cat /etc/config/sqm
config queue 'eth1'
option enabled '1'
option interface 'eth1'
option download '975000'
option upload '950000'
option qdisc 'cake'
option script 'piece_of_cake.qos'
option linklayer 'ethernet'
option debug_logging '0'
option verbosity '5'
option overhead '44'
option linklayer_advanced '1'
option tcMTU '2047'
option tcTSIZE '128'
option tcMPU '84'
option linklayer_adaptation_mechanism 'default'
root@xxx:~# 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 8582239900 bytes 8768203 pkt (dropped 0, overlimits 0 requeues 3)
backlog 0b 0p requeues 3
qdisc fq_codel 0: dev eth0 parent :1 limit 10240p flows 1024 quantum 1514 target 5ms interval 100ms memory_limit 4Mb ecn drop_batch 64
Sent 8582239900 bytes 8768203 pkt (dropped 0, overlimits 0 requeues 3)
backlog 0b 0p requeues 3
maxpacket 1514 drop_overlimit 0 new_flow_count 285 ecn_mark 0
new_flows_len 0 old_flows_len 0
qdisc cake 800d: dev eth1 root refcnt 2 bandwidth 950Mbit besteffort triple-isolate nonat nowash no-ack-filter split-gso rtt 100ms noatm overhead 44 mpu 84
Sent 3259199660 bytes 3023019 pkt (dropped 2362, overlimits 3824515 requeues 0)
backlog 0b 0p requeues 0
memory used: 6123264b of 15140Kb
capacity estimate: 950Mbit
min/max network layer size: 28 / 1500
min/max overhead-adjusted size: 84 / 1544
average network hdr offset: 14
Tin 0
thresh 950Mbit
target 5ms
interval 100ms
pk_delay 21us
av_delay 13us
sp_delay 5us
backlog 0b
pkts 3025381
bytes 3262179427
way_inds 81149
way_miss 1805
way_cols 0
drops 2362
marks 0
ack_drop 0
sp_flows 1
bk_flows 1
un_flows 0
max_len 5633
quantum 1514
qdisc ingress ffff: dev eth1 parent ffff:fff1 ----------------
Sent 2122187167 bytes 2131440 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 800e: dev ifb4eth1 root refcnt 2 bandwidth 975Mbit besteffort triple-isolate nonat wash no-ack-filter split-gso rtt 100ms noatm overhead 44 mpu 84
Sent 2182487628 bytes 2130973 pkt (dropped 467, overlimits 15104 requeues 0)
backlog 0b 0p requeues 0
memory used: 1225472b of 15140Kb
capacity estimate: 975Mbit
min/max network layer size: 46 / 1500
min/max overhead-adjusted size: 90 / 1544
average network hdr offset: 14
Tin 0
thresh 975Mbit
target 5ms
interval 100ms
pk_delay 38us
av_delay 18us
sp_delay 6us
backlog 0b
pkts 2131440
bytes 2183151303
way_inds 244
way_miss 1925
way_cols 0
drops 467
marks 0
ack_drop 0
sp_flows 1
bk_flows 1
un_flows 0
max_len 67184
quantum 1514
Please explain about "bridger". How it works, when to benefit from it, and how to install
Please use the "Preformatted text </>" button for logs, scripts, configs and general console output.
Please edit your post accordingly. Thank you! 
No idea yet, conf files are from ONE system when two are involved.
thank you! much nicer to look at now
There should be documentation on it on the openwrt site, but I think it's down atm.
For internet debloating SQM likely should be on eth0...
I assume that you hooked up the r6900's wan port to the r4s and connected the test computer on one of the LAN ports?
No, the r6000's WAN port is unused. The test computer is connected to the r6000's lan port though, yes. I'd like to refer to openwrt's instructions on how I set it up, but the site is down atm.
So Test B) was entirely using the r6900's switch then? That is pretty odd...
Yeah, no idea whats going on lol. The test is getting good results now though.... 
bufferresults
Are my ethernet cables in the slots? The R4S's lan cable goes to one of the R6000's lan ports, right?
edit: Openwrt's site is back up, and according to the bridgedap topic, I should have everything setup correctly. Under "Fine tuning" it mentions:
Use an Ethernet cable to connect one of the LAN ports on your main router to one of the LAN ports (not the WAN /Internet port) of the wireless AP.
I actually could not find much on bridger.
For internet debloating SQM likely should be on eth0...
You are 100% correct. I went back and looked over my settings and I had the wrong eth selected 
. I have no idea how things were working so smoothly 
.