Flint 2 (GL-MT6000 ) - bug reports - collective thread

Does this happen on OpenWrt master? :thinking:

I cannot replicate this with mt76 alone on kernel 6.6, but there are indeed some quirks i noticed.

For the increase of slowness:
One example is when i use something like banip, if the ban list is kinda above 20000 elements i get these strange time outs aswell, though my expectation is that the cpu should ‘handle’ this expensive load, but instead doesn’t handle it very well, i noticed the same type behaviours on sqm when banip was turned off and irqbalance, so somehow i suspect mt79 being vulnerable to cpu load in where the speed can malfunction.

When i ended disabling them it works stable to me, and when i removed all my interfering devices such as bluetooth and usb stick my range also normalized, i’m aware there is also a bug report for end devices using bluetooth and wifi causing issues can you check if this improves your connection?

It happens for me with the snapshot version which I got from: OpenWrt Firmware Selector

I also created an issue report: MT7986 - temporarily ping spikes and terrible WiFi performance - 5GHz, 160 MHz · Issue #877 · openwrt/mt76 · GitHub

SQM was never activated by me. I have a VLAN network switch connected to the Flints USB port, but its just using the power, no data is transmitted.

The mentioned problems disappear with the stock firmware 4.5.8 or by using the Flint 1 instead.

1 Like

This could be part of the issue even though no data is going through there might be electricity interference, it is how well shielded the connection is from the cables connector.

But why it is not an issue for the firmware 4.5.8 and even the Flint 1?

My suspicioun is that it runs on usb2 or usb1 drivers, but the OpenWrt version by default uses usb3, i could be wrong but that seem the most reasonable explaination.

You could try opkg update && opkg install usbutils and then type lsusb inside the cli if you see xhci its using usb 3, ehci is usb 2.

2 Likes

Hello all,

@GL.iNet Staff Please focus on the latest stable version of OpenWrt
23.05.3, I returned to the official GL-MT6000 software and I am satisfied, I have been using it for 5 days on a pppoe ipv6 dual stack connection with IPV6 active, wireguard server, everything works without problems!

GL-MT6000 Ver.4.5.6 Xiaomi 14 wifi speed test 2Ghz & 5Ghz!





2 Likes

grafik

Firmware 4.5.8

1 Like

Jup in that case it must be a issue then since the usb3 is mutual present in both firmwares :+1:

I’m having regular also WiFi crashes with 4.5.8 every 1-2 hrs for around 5min. My VLAN switch is now powered externally.

@alzhao I can provide you the full log file if you want.

Fri May  3 17:15:33 2024 kern.debug kernel: [56525.492372] mtwf_dbg_prt: vsnprintf error
Fri May  3 17:15:33 2024 kern.err kernel: [56525.496399] 7986@C08L1,wifi_sys_op_lock() 812: wifi_sys_disconn_act+0xec/0x2b8 [mt_wifi](0): get wdev_op_lock fail!! The latest caller with the lock: wifi_sys_disconn_act+/../mt_wifi/embedded/common/wifi_sys_info.c:813 assert FALSEfailed
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.517420] CPU: 0 PID: 6525 Comm: RtmpMlmeTask Tainted: P                  5.4.238 #0
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.525319] Hardware name: GL.iNet GL-MT6000 (DT)
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.530008] Call trace:
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.532447]  dump_backtrace+0x0/0x198
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.536096]  show_stack+0x14/0x20
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.539399]  dump_stack+0xb4/0xf4
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.542865]  wifi_sys_op_lock+0x17c/0x1a8 [mt_wifi]
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.547832]  wifi_sys_disconn_act+0xec/0x2b8 [mt_wifi]
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.553037]  wdev_do_disconn_act+0x20/0x38 [mt_wifi]
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.558060]  MacTableDeleteEntry+0x370/0x1160 [mt_wifi]
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.563340]  ap_peer_auth_req_at_idle_action+0x894/0x1820 [mt_wifi]
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.569663]  auth_fsm_mlme_auth_req_action+0x5dc/0x778 [mt_wifi]
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.575723]  StateMachinePerformAction+0x6c/0x80 [mt_wifi]
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.581264]  MlmeHandler+0x6d4/0x958 [mt_wifi]
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.585763]  MlmeHandler+0x898/0x958 [mt_wifi]
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.590194]  kthread+0x120/0x128
Fri May  3 17:15:33 2024 kern.debug kernel: [56525.593410]  ret_from_fork+0x10/0x1c
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.691950] mtwf_dbg_prt: vsnprintf error
Fri May  3 17:16:24 2024 kern.err kernel: [56576.695970] 7986@C08L1,wifi_sys_op_lock() 812: wifi_sys_disconn_act+0xec/0x2b8 [mt_wifi](0): get wdev_op_lock fail!! The latest caller with the lock: wifi_sys_disconn_act+/../mt_wifi/embedded/common/wifi_sys_info.c:813 assert FALSEfailed
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.717298] CPU: 0 PID: 4897 Comm: RtmpCmdQTask Tainted: P                  5.4.238 #0
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.725197] Hardware name: GL.iNet GL-MT6000 (DT)
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.729887] Call trace:
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.732326]  dump_backtrace+0x0/0x198
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.735976]  show_stack+0x14/0x20
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.739280]  dump_stack+0xb4/0xf4
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.742772]  wifi_sys_op_lock+0x17c/0x1a8 [mt_wifi]
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.747757]  wifi_sys_disconn_act+0xec/0x2b8 [mt_wifi]
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.752960]  wdev_do_disconn_act+0x20/0x38 [mt_wifi]
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.757981]  MacTableDeleteEntry+0x370/0x1160 [mt_wifi]
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.763260]  SetRtsThenCtsRetryCnt+0x37c/0xcc8 [mt_wifi]
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.768628]  CMDHandler+0x13c/0x2b0 [mt_wifi]
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.773042]  RTPCICmdThread+0x1c8/0x278 [mt_wifi]
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.777803]  hif_cmd_thread+0x2c/0x88 [mt_wifi]
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.782321]  kthread+0x120/0x128
Fri May  3 17:16:24 2024 kern.debug kernel: [56576.785538]  ret_from_fork+0x10/0x1c

Edit: It looks like this is caused by a unspecific client. It didn’t happen so far in the evening.

I’m also wondering. Why is Gl-inet using an old MTK version?

Currently it is:

Filogic 630/830 MP2.1 Release (20230508)

while there is already:

Filogic 830 MP2.2 Release (20231027)

Source:

https://git01.mediatek.com/plugins/gitiles/openwrt/feeds/mtk-openwrt-feeds/+/refs/heads/master/autobuild_mac80211_release/Release.md

I tried to compile MP2.2, which worked fine but unfortunately it seems to need some further configuration so that it can boot from the MT6000.

ethernet port 2 like wan 2 does not work as load balance

What’s your cypher? WPAx?

None, Open (public) network.

@xize11 I also tried the OpenWRT snapshot version without anything connected to USB. The WiFi is still timing out/freezing and the performance often suddenly terrible. There is not much traffic load during this time.

Here a comparison between openwrt snapshot and the stock firmware 4.5.8

1 Like

This is indeed strange, could it be something device specific?

For the random speed i noticed this too, often it fixes itself by reconnecting to wifi or by repositioning the device angle towards the router.

Though my router hangs vertically and also some of the antennas point to the wrong direction, it isn’t super bad since i pulled up a few antenna vertical but it could explain why this happens to me.

I think that your device lost connection for some reason… Just to understand if you set a wpa2 do you get same issue?

Do you get this issue from a specific device? Do you measure the ping from the device or from router or?

Thank you!

Yes brand new Flint2 and can’t even get in to the routers browser. Password doesn’t work!

What a shame!

There is no password.

Are you sure it’s a brand new one?
Reset it so the initial setup will begin.

1 Like

Hi,

Since I updated Flint2 to v4.5.8 my Fairphone3 has difficulties to reconnect to wifi. The wifi networks from Flint2 does not appear anymore in stored wifi networks of the phone.

So it works fine at the first connection and it won’t or hardly reconnect after. Even if I click on one of my wifi network (2.4G or 5G, LAN or Guest) the FP3 ask again for the password (weird?) and then nothing more happens.

I tried to reset wifi data from FP3 but it changed nothing.

On other wifi networks I don’t have this problem.

From which version did you update to 4.5.8?

Hi, I updated from v4.5.6

(I used to install v4.5.7. As I needed docker that did not work anymore I rolled back to v4.5.6 and reset the router)