Is there any estimate on when version 4.5.9 will be available?
new firmware, will it improve anything?
1 Like
I have firmware version 4.5.8 installed.
And the router is quite stable for my use.
It is cascaded to my provider’s ONT and provides me with a 1G-2G connection on the ETH interface and 1.5G on WiFi 5Ghz simultaneously.
The 2.4Ghz WiFi is more unstable, but for home automation it is sufficient.
It also handles IPv6 perfectly.
Then I use AdGuard Home (which, thanks to a script, is constantly updated to the latest version available) and WireGuard (as clients a GL-iNet Beryl AX and Notebooks) and these also work perfectly.
So, for my use, it is an excellent router (of course, always waiting for a firmware that will make the device even more stable) and then the GL-iNet GUI is incomparable while maintaining the possibility via LuCI to make more specific configurations on OpenWRT.
1 Like
I’m eagerly waiting to also report that this device is stable for me.
Unfortunately I seem to have 1 client in the wifi network that seems to be able to cause an issue.
Fri Apr 12 08:17:47 2024 daemon.info hostapd: wlan1: STA mac_address IEEE 802.11: authenticated
Fri Apr 12 08:17:47 2024 daemon.info hostapd: wlan1: STA mac_address IEEE 802.11: associated (aid 4)
Fri Apr 12 08:17:47 2024 daemon.notice hostapd: wlan1: STA-OPMODE-SMPS-MODE-CHANGED mac_address off
Fri Apr 12 08:17:48 2024 daemon.notice hostapd: wlan1: AP-STA-CONNECTED mac_address auth_alg=open
Fri Apr 12 08:17:48 2024 daemon.info hostapd: wlan1: STA mac_address WPA: pairwise key handshake completed (RSN)
Fri Apr 12 08:17:48 2024 daemon.notice hostapd: wlan1: EAPOL-4WAY-HS-COMPLETED mac_address
Fri Apr 12 08:18:00 2024 daemon.info dnsmasq-dhcp[1]: DHCPDISCOVER(br-lan) mac_address
Fri Apr 12 08:18:00 2024 daemon.info dnsmasq-dhcp[1]: DHCPOFFER(br-lan) 192.168.8.151 mac_address
Fri Apr 12 08:18:00 2024 daemon.info dnsmasq-dhcp[1]: DHCPREQUEST(br-lan) 192.168.8.151 mac_address
Fri Apr 12 08:18:00 2024 daemon.info dnsmasq-dhcp[1]: DHCPACK(br-lan) 192.168.8.151 mac_address
....
Fri Apr 12 11:14:03 2024 daemon.notice hostapd: wlan1: STA-OPMODE-SMPS-MODE-CHANGED mac_address dynamic
Fri Apr 12 11:16:02 2024 daemon.notice hostapd: wlan1: STA-OPMODE-SMPS-MODE-CHANGED mac_address off
Fri Apr 12 11:18:11 2024 daemon.notice hostapd: wlan1: STA-OPMODE-SMPS-MODE-CHANGED mac_address dynamic
Fri Apr 12 11:19:55 2024 daemon.notice hostapd: wlan1: STA-OPMODE-SMPS-MODE-CHANGED mac_address off
Fri Apr 12 11:19:57 2024 daemon.notice hostapd: wlan1: STA-OPMODE-SMPS-MODE-CHANGED mac_address dynamic
Fri Apr 12 11:19:58 2024 daemon.notice hostapd: wlan1: STA-OPMODE-SMPS-MODE-CHANGED mac_address off
Fri Apr 12 11:20:00 2024 daemon.notice hostapd: wlan1: STA-OPMODE-SMPS-MODE-CHANGED mac_address dynamic
Fri Apr 12 11:20:01 2024 daemon.notice hostapd: wlan1: STA-OPMODE-SMPS-MODE-CHANGED mac_address off
This seems to be an openwrt bug.
From the moment I observe the last log lines the client in question gets limited to 50Mbps.
I also had the impression the whole wifi network got slowed down, but that is a bit harder too prove.
After some searching I found that you can go to the advanced setting of the wifi card and change the “MIMO Power Save Mode” to No SMPS. This seems to resolve the issue. (Credits go to Fredo on the OpenWRT issues tracker - "STA-OPMODE-SMPS-MODE-CHANGED" what that mean? - #26 by Fredo - Network and Wireless Configuration - OpenWrt Forum)
Update:
The issue doesn’t seem to be entirely gone.
After a period of inactivity, I got back at the laptop in question and noticed the wlan speed got stuck again at 50 Mbps.
I observe this in the router log:
Fri Apr 12 12:02:04 2024 daemon.info hostapd: wlan0: STA mac_address IEEE 802.11: authenticated
Fri Apr 12 12:02:04 2024 daemon.info hostapd: wlan0: STA mac_address IEEE 802.11: associated (aid 3)
Fri Apr 12 12:02:05 2024 daemon.notice hostapd: wlan0: AP-STA-CONNECTED mac_address auth_alg=open
Fri Apr 12 12:02:05 2024 daemon.info hostapd: wlan0: STA mac_address WPA: pairwise key handshake completed (RSN)
Fri Apr 12 12:02:05 2024 daemon.notice hostapd: wlan0: EAPOL-4WAY-HS-COMPLETED mac_address
Fri Apr 12 12:02:07 2024 daemon.info dnsmasq-dhcp[1]: DHCPREQUEST(br-lan) 192.168.8.151 mac_address
Fri Apr 12 12:02:07 2024 daemon.info dnsmasq-dhcp[1]: DHCPACK(br-lan) 192.168.8.151 mac_address
Fri Apr 12 12:02:16 2024 daemon.notice hostapd: wlan1: DFS-CAC-COMPLETED success=1 freq=5260 ht_enabled=0 chan_offset=0 chan_width=3 cf1=5290 cf2=0
...
Fri Apr 12 12:11:39 2024 daemon.info hostapd: wlan1: STA mac_address IEEE 802.11: authenticated
Fri Apr 12 12:11:39 2024 daemon.info hostapd: wlan1: STA mac_address IEEE 802.11: associated (aid 2)
Fri Apr 12 12:11:39 2024 daemon.notice hostapd: wlan0: Prune association for mac_address
Fri Apr 12 12:11:39 2024 daemon.notice hostapd: wlan0: AP-STA-DISCONNECTED mac_address
Fri Apr 12 12:11:41 2024 daemon.notice hostapd: wlan1: AP-STA-CONNECTED mac_address auth_alg=open
Fri Apr 12 12:11:41 2024 daemon.info hostapd: wlan1: STA mac_address WPA: pairwise key handshake completed (RSN)
Fri Apr 12 12:11:41 2024 daemon.notice hostapd: wlan1: EAPOL-4WAY-HS-COMPLETED mac_address
Fri Apr 12 12:11:44 2024 daemon.info dnsmasq-dhcp[1]: DHCPREQUEST(br-lan) 192.168.8.151 mac_address
Fri Apr 12 12:11:44 2024 daemon.info dnsmasq-dhcp[1]: DHCPACK(br-lan) 192.168.8.151 mac_address
Fri Apr 12 12:12:09 2024 daemon.info hostapd: wlan0: STA mac_address IEEE 802.11: deauthenticated due to inactivity (timer DEAUTH/REMOVE)
Fri Apr 12 13:08:56 2024 daemon.notice hostapd: wlan1: STA-OPMODE-SMPS-MODE-CHANGED mac_address dynamic
Disconnecting and connecting again to the wifi resolves the speed issue…
2 Likes
Can you give us some details about how to do this? Thanks.
This might be a long shot, but is there a way I can take a backup of the entire flash inside the router?
Like how you can create an image of a disk… I’d like to experiment with some of these beta versions and also mess about with using OpenWRT natively etc, however I’d like a way to quickly restore everything back to how I have it set up right now.
I’ve got quite a few customisations that currently take more than an hour to get everything back to how I like it, so being able to image the flash and restore it back would be useful.
I’ve seen a method for SD cards but apparently it doesn’t work with NAND.
1 Like
You mean creating a image with custom configuration?
If you managed to get the build environment working you need to create a folder called files inside the root directory of the build tools and there you can mirror a copy of /etc/config like files/etc/config but you have to be sure these are stable because it will be written to /rom/ to in the image with other words resets will put these files back.
For mt76 patches you can download them to package/kernel/mt76/patches and if you want to test a pull request as a patch just write .patch in the web address in the end and then download it.
Though setting up build tools alone can be difficult the build tools dependencies described on the readme are not all the dependencies, this is really try and error with logging, also you need gcc12 and g++12 but the current wsl ubuntu does not have it, you have to add a newer ubuntu apt repo inside the sources list and then fix both gcc and g++ by using a symlink, or use a vm with a more up to date ubuntu.
ln -s /usr/bin/gcc12 /usr/bin/gcc -f
ln -s /usr/bin/g++12 /usr/bin/g++ -f
Though maybe gcc12 and g++12 are no longer a requirement, but i’m not sure about that since i noticed commits for the build bot to allow it, but compiling with these higher versions of gcc and g++ seem to increase performance on the cpu of the router somehow👍
with make menuconfig you can enter a menu where you can set pre installed packages, you probably also want to install all feeds before that or the packages are not visible do so on each git pull you do: ./scripts/feeds update -a ./scripts/feeds install -a and then use make menuconfig so it knows the feeds are updated.
Here is also a wiki from openwrt for setting up build tools:
Just to see if I’m the only one on this “problem”:
If you insert an USB Stick on Flint 2, after 30 minutes can you notice it’s quite warm, around 38°C? The room temperature is 20°C
No readings/writings. Just leave it in standby, wait 30 minutes, remove it and check the temperature of the Flash drive Tip.
Hello, i’m trying to set speed limit, but on three dot button i can edit only device type and name…! any idea?
reference:
Clients - GL.iNet Router Docs 4.
You need to disable hardware acceleration: Hardware Acceleration - GL.iNet Router Docs 4
Tbh I would not recommend it, but that would be the way to go.
1 Like
2.4 GHz network performance, N mode, in stable version (v4.5.8 - do dia 12/04) is very poor.
Smartphone: Motorola Edge 30 Neo (not compatible with WiFi 6)
2.4 & 5 Ghz Channel Analysis doesn’t work with 4.5.7 firmware, and doesn’t work with 4.5.8.
1 Like
As I reported here Flint 2 FW 4.5.8 release 4 - #22 by Merciful
Power on 5ghz tops out at 20dbm in regulatory area where it should be 23dbm (UK). Changing it to another similar regulatory area results in the same limit. 2.4ghz band already set as UK as well.
Edited to add - with 4.5.8 stable firmware.
Also as above channel analysis is non-functional.
SQM seems to work better with the current drivers in 4.5.8 rel 4 compared to 4.5.6 (open source) at least on my experience. I almost cannot use the SQM in 4.5.6 since I can only see latency changes if the speed was set to around 80-50mbps in download. However, 4.5.8 already shows decent latency improvement by just reducing 15% of my download speed. However, I am not 100% sure that the improvement was indeed caused by the proprietary drivers but my setup is pretty much the same when I was still driving 4.5.6. (Modem RM520N-GL with M.2 to 2.5Gb ETH board connected to Flint 2 via ETH WAN)
I can get the same results with OpenWRT on kernel 6.1 and kernel 6.6.
Additionally, it appears there has been an update that changes the way packages are targeted in the latest OpenWRT snapshot, which should bring improvements to SQM: git.openwrt.org Git - openwrt/openwrt.git/commit
Another important piece of information is that support for kernel 5.15, the same one used in GL.inet firmwares, is being abandoned and continuing to use it could cause security problems in the long term: pistachio: remove 5.15 support · openwrt/openwrt@12137cb · GitHub
1 Like
Im still waiting for their Openwrt 23 release so I dont really have a choice right now. I don’t want to also abandon the GL inet GUI since Im using some of their features thats not present / too complex to setup, on vanilla Openwrt. Im just hoping that they will choose the best vanilla Openwrt release for their 4.5.8 counterpart.
7 Likes
I wish they had optimized the SMB Server since that’s something really I use, I can’t even transfer a 10GB file from Windows most of the time, it is unstable and a bit unreliable. Seems like we need to wait a long time for it
4.5.8 here, again problem with old configuration? Webgui down… I need to reset again… 
