The primary difference I noticed is mt76 (wifi driver) is more up to date right now on main snapshot, they haven't moved the new driver to 23.05.
1 Like
Not sure if this works correctly as designed, as we all know sfo never worked correctly with sqm, but, following Asus tuf ax6000 thread, attaching results for sqm with ctinfo + sfo + packet steering and irqbalance for 900 mb speed:
tc -s qdisc:
Summary
qdisc cake 8005: dev eth1 root refcnt 17 bandwidth 900Mbit diffserv4 dual-srchost nat nowash no-ack-filter split-gso rtt 100ms raw overhead 0 mpu 84
Sent 2979062138 bytes 2795487 pkt (dropped 50, overlimits 2534959 requeues 2395)
backlog 0b 0p requeues 2395
memory used: 3548655b of 15140Kb
capacity estimate: 900Mbit
min/max network layer size: 42 / 1514
min/max overhead-adjusted size: 84 / 1514
average network hdr offset: 14
Bulk Best Effort Video Voice
thresh 56250Kbit 900Mbit 450Mbit 225Mbit
target 5ms 5ms 5ms 5ms
interval 100ms 100ms 100ms 100ms
pk_delay 0us 8us 5us 3us
av_delay 0us 2us 0us 0us
sp_delay 0us 1us 0us 0us
backlog 0b 0b 0b 0b
pkts 0 2795490 33 14
bytes 0 2979129723 3026 1695
way_inds 0 106797 0 0
way_miss 0 1238 33 8
way_cols 0 0 0 0
drops 0 50 0 0
marks 0 0 0 0
ack_drop 0 0 0 0
sp_flows 0 3 0 0
bk_flows 0 1 0 0
un_flows 0 0 0 0
max_len 0 18902 118 151
quantum 1514 1514 1514 1514
qdisc cake 8006: dev ifb4eth1 root refcnt 2 bandwidth 900Mbit diffserv4 dual-dsthost nat nowash ingress no-ack-filter split-gso rtt 100ms raw overhead 0
Sent 2985309171 bytes 2822908 pkt (dropped 31968, overlimits 902537 requeues 0)
backlog 0b 0p requeues 0
memory used: 10695502b of 15140Kb
capacity estimate: 900Mbit
min/max network layer size: 56 / 1514
min/max overhead-adjusted size: 56 / 1514
average network hdr offset: 14
Bulk Best Effort Video Voice
thresh 56250Kbit 900Mbit 450Mbit 225Mbit
target 5ms 5ms 5ms 5ms
interval 100ms 100ms 100ms 100ms
pk_delay 0us 41us 5us 5us
av_delay 0us 5us 0us 0us
sp_delay 0us 1us 0us 0us
backlog 0b 0b 0b 0b
pkts 0 2854842 31 3
bytes 0 3033698497 2790 848
way_inds 0 215 0 0
way_miss 0 1085 31 2
way_cols 0 0 0 0
drops 0 31968 0 0
marks 0 0 0 0
ack_drop 0 0 0 0
sp_flows 0 3 0 1
bk_flows 0 1 0 0
un_flows 0 0 0 0
max_len 0 68130 90 394
quantum 1514 1514 1514 1514
Got the router today! Anyone had issues installing OpenWRT? I followed the instructions in this thread, specifically:
1) Remove Power from the router
2) Press and Hold the reset button
3) Apply power whilst still holding the reset button
4) After 5 seconds head to http://192.168.1.1
5) Flash OpenWRT https://downloads.openwrt.org/snapshots/targets/mediatek/filogic/openwrt-mediatek-filogic-glinet_gl-mt6000-squashfs-sysupgrade.bin
I flashed the image using the upgrade firmware function in the WebUI. It installed, restarted and since then I cannot log in the router or do anything really. LED light also stopped to blink and is now still on all the time. It blinks only once after restart. What are my options? 
Snapshot builds don't have the Luci web UI installed by default.
Connect LAN port to your computer, open a terminal
ssh root@192.168.1.1
opkg update
opkg install luci
Reboot
Navigate to http://192.168.1.1
OR read above about customising a sysupgrade image to include Luci on the firmware selector page and then install that from the recovery web UI.
1 Like
Everything's working fine on my end, but everyone is seeing all of these issues and warnings in their system log, right?
I'm currently using OpenWrt SNAPSHOT r24668-5a85aada3f.
kern.warn kernel: [ 1.765932] GPT:Primary header thinks Alt. header is not at the end of the disk.
kern.warn kernel: [ 1.782670] GPT:1 != 15269887
kern.warn kernel: [ 1.785628] GPT:Alternate GPT header not at the end of the disk.
kern.warn kernel: [ 1.785630] GPT:1 != 15269887
kern.warn kernel: [ 1.804891] GPT: Use GNU Parted to correct GPT errors.
kern.info kernel: [ 1.810031] mmcblk0: p1 p2 p3 p4 p5 p6 p7
kern.info kernel: [ 1.825524] mmcblk0boot0: mmc0:0001 8GTF4R 4.00 MiB
kern.info kernel: [ 1.831288] mmcblk0boot1: mmc0:0001 8GTF4R 4.00 MiB
kern.info kernel: [ 1.836455] mmcblk0rpmb: mmc0:0001 8GTF4R 512 KiB, chardev (250:0)
...
kern.notice kernel: [ 5.056455] F2FS-fs (loop0): Disable nat_bits due to incorrect cp_ver (5561039705580362896, 0)
kern.notice kernel: [ 5.067419] F2FS-fs (loop0): Mounted with checkpoint version = 8f5dc90
user.info kernel: [ 5.074435] mount_root: switching to f2fs overlay
kern.warn kernel: [ 5.079750] overlayfs: null uuid detected in lower fs '/', falling back to xino=off,index=off,nfs_export=off.
...
kern.info kernel: [ 6.003253] platform 15010000.wed: MTK WED WO Firmware Version: DEV_000000, Build Time: 20221012175005
kern.info kernel: [ 6.012600] platform 15010000.wed: MTK WED WO Chip ID 00 Region 3
kern.info kernel: [ 6.591821] mt798x-wmac 18000000.wifi: HW/SW Version: 0x8a108a10, Build Time: 20221012174648a
kern.info kernel: [ 6.697391] mt798x-wmac 18000000.wifi: WM Firmware Version: ____000000, Build Time: 20221012174725
kern.info kernel: [ 6.765435] mt798x-wmac 18000000.wifi: WA Firmware Version: DEV_000000, Build Time: 20221012174937
kern.warn kernel: [ 6.911952] mt798x-wmac 18000000.wifi: eeprom load fail, use default bin
...
daemon.notice procd: /etc/rc.d/S25packet_steering: uci: Entry not found
...
kern.info kernel: [ 2.512184] xhci-mtk 11200000.usb: xHCI Host Controller
kern.info kernel: [ 2.517415] xhci-mtk 11200000.usb: new USB bus registered, assigned bus number 1
kern.info kernel: [ 2.527881] xhci-mtk 11200000.usb: hcc params 0x01403f99 hci version 0x110 quirks 0x0000000000210010
kern.info kernel: [ 2.537031] xhci-mtk 11200000.usb: irq 123, io mem 0x11200000
kern.info kernel: [ 2.542865] xhci-mtk 11200000.usb: xHCI Host Controller
kern.info kernel: [ 2.548081] xhci-mtk 11200000.usb: new USB bus registered, assigned bus number 2
kern.info kernel: [ 2.555458] xhci-mtk 11200000.usb: Host supports USB 3.2 Enhanced SuperSpeed
kern.info kernel: [ 2.562814] hub 1-0:1.0: USB hub found
kern.info kernel: [ 2.566568] hub 1-0:1.0: 2 ports detected
kern.info kernel: [ 2.570843] usb usb2: We don't know the algorithms for LPM for this host, disabling LPM.
kern.info kernel: [ 2.579213] hub 2-0:1.0: USB hub found
kern.info kernel: [ 2.582973] hub 2-0:1.0: 1 port detected
Also, if you configure Adguard Home to listen on br-lan then this patch seems to be necessary, otherwise the service starts too early and then fails.
some 5G wifi questions:
- as someone reported already, it seems AX 160Mhz does not work?
- the tx power options seems fluctuating? I mean if I choose US, sometimes the max available is 199mW, sometimes it is 251mW, am I the only person seeing this? Does it correlate with channel? I mean I set the channel to auto, and maybe max available tx power depends on the actual channel?
You are correct.
I installed latest snapshot with luci from the web ui and everything went fine but I'll wait to the official release to change it as my main router.
Spectrum DOCSIS 500/20
Seems I straightened things out. I finally got around to testing and ended up switching back to the SQM package from qosify and I think the issue was me referencing the device by the wan alias instead of eth1. I also noticed some packet loss a few hops downstream under load that may be throwing the algorithm off but I've had several A+ results with minimum, single digit, latency increases. There were times where I consistently had 0 increased latency with the RT3200 and qosify but I haven't bother to test for months.
# cat /etc/config/sqm
config queue 'eth1'
option enabled '1'
option interface 'eth1'
option download '520000'
option upload '22000'
option qdisc 'cake'
option script 'piece_of_cake.qos'
option linklayer 'ethernet'
option debug_logging '0'
option verbosity '5'
option overhead '22'
Previously working qosify config from the RT3200
# cat /etc/config/qosify
...
config interface wan
option name wan
option disabled 0
option bandwidth_up 22mbit
option bandwidth_down 520mbit
option overhead_type docsis
# defaults:
option ingress 1
option egress 1
option mode diffserv4
option nat 1
option host_isolate 1
option autorate_ingress 1
#option ingress_options "overhead 46 mpu 92 no-split-gso"
#option egress_options "overhead 38 mpu 76 no-split-gso"
option options ""
This is my /etc/config/sqm which works well for me.
ISP: Comcast Xfinity
Plan: 200/10 Mbps ("Connect More" Plan) (usually with 20% over-provisioning, i.e. 240/12 Mbps)
Modem: Arris S33 DOCSIS 3.1
Router: GL.iNet GL-MT6000 (Flint 2) flashed with OpenWrt SNAPSHOT
Same /etc/config/sqm used with Dynalink DL-WRX36 before that, and Belkin RT3200 before that (with only the corresponding "wan" interface changed).
I had tried qosify during it's early days (long back) but I went back to regular sqm as qosify had issues (failed to start or random crash) at that time.
config queue
option enabled '1'
option interface 'eth1'
option verbosity '5'
option qdisc_advanced '1'
option qdisc_really_really_advanced '1'
option egress_ecn 'NOECN'
option squash_ingress '1'
option squash_dscp '1'
option debug_logging '1'
option ingress_ecn 'ECN'
option qdisc 'cake'
option script 'layer_cake.qos'
option download '200000'
option upload '10000'
option linklayer 'none'
option iqdisc_opts 'ingress docsis diffserv4 nat dual-dsthost'
option eqdisc_opts 'egress docsis diffserv4 nat dual-srchost ack-filter'
For your setup (500/20 Mbps), you can re-use my /etc/config/sqm with just the following change:
option download '500000'
option upload '20000'
I suggest just setting 500000/20000 instead of
option download '520000'
option upload '22000'
as the value you specify in /etc/config/sqm has to be about 80% to 90% of max download/upload bandwidth.
Setting the values to 100% of plan speed rather than 110% to 120% (like in your case) ends up being 83.33% (1/1.2) or 90.91% (1/1.1) of the max speed which works well for sqm.
1 Like
actually, 5G AX 160MHz works.
1 Like
Can confirm, 160 MHZ works. I get iperf3 results of 1.7 Gbps over wifi and 2.3 Gbps over wired at the same time.
4 Likes
I wonder if anyone did measure the power consumption of this router?
Was scrolling through the thread, but did not find anything.
Yes I did but I didn't took the time to write about it yet.
It comes with a 48 W power supply. Does it need all of that? Short answer: no. Long answer: probably still no.
Idle it draws around 6 or 7 Watts. Connecting 1G devices adds around 0.3 W and 2.5G devices around 0.7W. Under full load I got around 12W with a single device connected via both 2.5G ethernet and 160 MHz wifi) and two iperf3 tests running like before. (Having the router be the server for both iperf3 tests is close to full cpu load. Adding more cpu load didn't do much to power usage).
However, it also has a USB 3.0 port. Which can deliver up to 2 A at 5 V. Those 10 W translate to around 12 W at the wall. Which puts the total at around 24 W. So if they'd ship it with a 24 W supply, it wouldn't leave any margin. So I guess that's why they gave us this supply.
I guess they could have given us a 30 W or 36 W supply but they didn't.
I did some more testing on the power supply itself. It can actually handle 48 W (and a bit more, I tested up to 55 W) without overheating. It seems to be around 85% - 88% efficient at normal power levels (tested at 6 W and 20 W).
9 Likes
I'm debating if I should get this or the TUF-AX6000 for OpenWRT. This has double the RAM of the TUF but Asus has a completely different stock firmware to fall back on.
I went with this since it's a much better looking device (imo), and 2x the RAM made it a no brainer. Got it for $110 preorder price, but I think it went up a bit since.
I really like the rear USB 3.0 port on the Asus though. I have a 2TB drive plugged into the side port of this (using Samba) and would be nice to have all cables in the back.
The GLiNet is easy to install and has a great rescue web GUI, so not much need for a fallback firmware.
For 5GHz wifi 6 oddly I couldn't get my device to connect to Ch36 160MHz width (iPhone 15 is all I tried, didn't try my laptop yet).
At 80MHz width I consistently get 650-690 Mbps at 15-ft distance. Load-%: 31/20/18/22 on wifi at that speed shown with htop. (edit: 760 Mbps with WED see posts below).
Wired to my PC I get 950-960 Mbps. Load-%: 0/0/0/0 with hardware flow offload enabled (most mediatek devices support HFO ).
I'm sure I read somewhere that Apple products don't support 160MHz on the 5GHz band. But in general I think you should only be using 160MHz on channels 100 to 144? You can check for yourself by using iw reg get.
It might be faster if you enable WED.
2 Likes
Wow never heard of WED, they should get that exposed to the Wireless advanced settings on LuCI asap. Enabled it, reboot, and ran a few more speedtests:
5GHz 80MHz width, 700-760 Mbps, Load% dropped to 5/0/0/0.