OK OK Just making sure 
Now how about this..........
They use straight up OpenWRT and OpenWRT repositories AND compile their forked OS interface into a package that can only be used in an imagebuilder.
Also put an on/off switch on the damn thing just because ..........
You got yourself a giant killer and a cult following and a money maker.
Honestly the Gl.inet interface is so close to Argon I would rather the Argon author just add his theme to the OpenWrt repos already and call it a day...
If you really want that look, install luci-theme-material it's sorta close enough, or in a few more steps install the Argon repo and add that.
Is the 2.4 GHz bug or whatever discussed on the gl inet forums specific to their own openwrt or would it also happen on the official openwrt builds if you flashed that?
I've tried with the GL.iNet firmware and the OpenWrt firmware and the speed and range isn't great with either of them, which isn't surprising since they're both using the same WiFi drivers.
I guess someone could try using ImmortalWrt or LEDE, since both of them can apply more tweaks for performance. But as they're both using the mt76 WiFi driver I wouldn't expect any drastic changes.
It seems like GL.iNet might be including patches from MediaTek into their current test firmware (v4.5.5) and these patches haven't been merged into the mt76 driver yet. And GL.iNet don't appear to be sharing their fork either.
if you have active ipv6 on kernel 6.1, we can forget about Software flow offloading-On
Hardware flow offloading-On
Just got one of these to replace an R7800. Is it just me or is Samba performance on this router worse? I used to get 70mb/s on wifi and sometimes 112mb/s by ethernet, but it seems like I'm stuck at 65mb/s on both with this router.
@Reznale I posted my Samba results above: 85 MB/s read, 110 MB/s write. There seems to be an 85 MB/s cap on read right now not sure why but I posted about it on the Samba thread too. My old WRT32X is a little faster, but I don't consider this slow results by any means.
@u251 Mesh is not something I've ever looked into, although I do have an unused U6-Lite that I think supports it. Someone on your link said to use batman-adv. Is there a wifi 6 + mesh support list somewhere?
I see that the Flint 2 does not have tri-band radios, but 802.11s should still work. On the plus side the 5 GHz 4x4 160 MHz AX band should offer plenty of bandwidth even if the 5 GHz radio is time shared between 802.11s and AP.
There are plenty of systems and users who use mesh (802.11s) on systems with two radios.
I think the problem is that gl-inet in their own fork do not allow wpads that have proper 802.11s support. You would have to get a snapshot image of OpenWRT to install the necessary wpad and set up mesh that way.
I got my MT6000 yesterday and flashed it today. I noticed pretty quickly that the blue LED is continuously on during the boot process, so there is no indication about the progress of the boot process. The DTS for MT6000 seems to be missing all OpenWrt LED functionality definitions and just the blue LEDs is forced to be always on.
So, I authored the fixes for LED definitions and issued a PR for that:
After the PR gets accepted, the LED will show during the boot process this way:
During sysupgrade flashing: white LED blinks
Until the PR gets accepted, you can patch your own builds (if compiling with the full toolchain).
EDIT:
The LED colors have been modified as follows:
Thank you very much!
And because it's a dts-only change, it can be (easily 
) imported to Image Builder! 
@hynman Thanks for your efforts, I must admit it boots so quickly didn't even notice the LED.
After a great first week with this as my main router decided to start testing kernel 6.1 along with Docker containers since the 8GB eMMC is so spacious. So far so good, the new luci-app-irqbalance package is interesting too 
I have the feeling by the time 24.0x hits with kernel 6.1 and some further mt76 refinements we have a top tier target on our hands here.
Do you see the same issue as @oli with the newer kernel?
Kernel 5.15:
https://forum.openwrt.org/t/gl-mt6000-discussions/173524/262
Kernel 6.1:
https://forum.openwrt.org/t/gl-mt6000-discussions/173524/303
EDIT: Unless oli was demonstrating how good the newer kernel is without software and hardware flow offloading? But I assume they're enabled and he's seeing increased CPU usage.
For the Image Builder users I created a repository to easily try such PRs:
The laziest can try the prebuilt images, containing:
I am always curious as to how well the fq_codel implementation on the wifi works...
Hello! Thank you for your hard work in assisting in the development of Gl.Inet.
What would be the difference between the openwrt-23.05-snapshot and openwrt-patch-mediatek firmwares?
Also, what is the function of preloader.bin?
Thanks for your PR. I've only one thing to add: With the original firmware the led is white during runtime. Using that as default would be helpful, because than it's clear that the device is not (hanging) at the bootloader.
I installed luci-app-wifischedule and scheduled like this:
The PR was merged in its original form, so blue LED for running normally.
This command does nothing on this target. If you have router you can test it. It does worked for example for ax3600 but not for this one.
