[TRACKING] Unstable and unreliable WLAN RZ616/MT7922 [FW13 AMD DIY]

[Ceremony]

I tested ubuntu 22 LTS (kernel 6.2) and 23 (kernel 6.5) live images: after 5 tries each, the initial issue didn’t pop up.

So back to testing my current manjaro setup: issue popped up reliably with both kernel 6.6 and 6.5, but not with 6.1 (LTS)… ill stick with 6.1 for now and see if the bandwidth issue also pops up.

Can someone else test manjaro live distro with their Framework+MT7922? I used KDE variant with Linux 6.5: https://download.manjaro.org/kde/23.0.4/manjaro-kde-23.0.4-231015-linux65.iso

P.S. I previously stated in the thread that 6.1 was affected… guess I was wrong about that?

[D.H]

What’s your Linux-firmware package version date?

There’s also instructions to manually update to the latest 7922 firmware files at GitHub - morrownr/USB-WiFi: USB WiFi Adapter Information for Linux

[Ceremony]

I checked: manjaro already comes with the latest firmware for the mt7922.
the version of the linux-firmware package is 20231030.2b304bfe-1.

[lbkNhubert]

To add a data point, I’m using that card on the 11th gen Manjaro with no issues. Best of luck getting thinga sorted out.

[Ceremony]

@lbkNhubert what kernel are you running and do you have issues with 6.5 or 6.6?

[dimitris]

@Ceremony your log entries also reminded me of this thread, issue there was resolved in the kernel as of 6.5.7. Your use case is different (no dock involved) but I wonder if that or some similar issue is at play in the kernel:

[RESPONDED] Sleep while on TB dock, disconnect, resume: ~ 1min delay to resume

Thanks for the heads-up dimitris. On 6.5.9-arch2-1 I see the same - no more delay. Looks like it was probably this patch PCI/PM: Mark devices disconnected if upstream PCIe link is down on resume, which landed in 6.5.7.

[lbkNhubert]

@Ceremony - currently running 6.6.1-1-Manjaro, before that 6.5.11-1-Manjaro, no issues on either. Router is an Asus running the Merlin firmware.

[Ceremony]

One more thing @lbkNhubert: If you are on 11th gen Intel, don’t you have an AX200/AX210 chip? Did you swap it with an mt7922?

[lbkNhubert]

@Ceremony - you are correct regarding the card. I swapped out the ax210 for the mt7922 in order to make use of the 6e band(s).

[Ceremony]

but… the AX210 is a 6e card? Did you mean AX200?

[lbkNhubert]

No, on the 11th Gen there is an issue with newer kernels that causes the 6e band(s) to be disabled.

[Ceremony]

Well, while it does work on my new Endeavour OS on a fresh boot, the slow transfer speed issue persists.
It pops up real often, while dmesg doesn’t show any mt7922/wlan0 issues.

I am swapping it with my old AX210 now, as that one was reliable in my previous laptop.

[ouaibe]

Running into the same issue here with Manjaro Gnome on Kernel 6.5.11 on my FW13 AMD DIY: there’s a ~20-30s delay post graphical boot/GDM login to get the mt7921e to show up and I have the same messages in dmesg:

[   10.873751] mt7921e 0000:01:00.0: not ready 1023ms after FLR; waiting
[   11.914400] mt7921e 0000:01:00.0: not ready 2047ms after FLR; waiting
[   14.020826] mt7921e 0000:01:00.0: not ready 4095ms after FLR; waiting
[   18.287488] mt7921e 0000:01:00.0: not ready 8191ms after FLR; waiting
[   26.607311] mt7921e 0000:01:00.0: not ready 16383ms after FLR; waiting
[   44.100756] mt7921e 0000:01:00.0: not ready 32767ms after FLR; waiting
[   78.234430] mt7921e 0000:01:00.0: not ready 65535ms after FLR; giving up
[   78.387363] mt7921e 0000:01:00.0: enabling device (0000 -> 0002)
[   78.410109] mt7921e 0000:01:00.0: ASIC revision: 79220010
[   78.488252] mt7921e 0000:01:00.0: HW/SW Version: 0x8a108a10, Build Time: 20230627143702a
[   78.864614] mt7921e 0000:01:00.0: WM Firmware Version: ____000000, Build Time: 20230627143946
[   79.952194] mt7921e 0000:01:00.0 wlp1s0: renamed from wlan0

Note that this behavior is NOT systematic, it happens from time to time after certain reboots, hard to predict when or what triggers this.

[fw13amd]

Hi, let me join the party:

[RESOLVED] FW13AMD authentication with WiFi BSSID timed out

Thanks, I will give that a try if the issue comes back. Of course now it’s been working fine for the past couple of days, go figure.

I have a different error in dmesg, but apart from that everything you guys say here resonates a lot with me. Especially the hard-to-reproduce part.
Sometimes it works just fine for entire sessions. Then I reboot/shutdown, and on the next one it’s completely useless (in my case the authentication to the base station times out).
There seems to be some correlation with the router (I have a TP-Link Archer C80 - no WiFi 6 on that one). Because on other APs (my ISP’s router and my Pixel 8 Pro’s hotspot) it works flawlessly and consistenly so.

With my previous laptop I had to stop taking firmware updates for my AX201 because it would destroy bandwidth whenever my BT earbuds were also connected (and that was an Intel board), now this…

Hardware is hard nowadays

[dump]

Also experiencing sudden slowdowns/flakiness on FW13 AMD DIY with an AX1800. Dunno what’s going on.

[lbkNhubert]

This thread has some things that you might try to alleviate this: Poor Wi-Fi performance with AMD RZ616 - #4 by Mario_Limonciello

[Matt_Hartley]

Hi @Ceremony,

As @lbkNhubert indicated via Mario:

[RESPONDED] Poor Wi-Fi performance with AMD RZ616

This patch is not going to be merged. But it does confirm there is some issue with power management and the APs you use. Would you be able to share more details about them? Perhaps in joining that email thread discussion?

I did locate what I believe was the patch in question.

https://patchwork.kernel.org/project/linux-wireless/patch/20231212090852.162787-1-mario.limonciello@amd.com/

Now, here is what I would suggest trying to see if it provides any relief.

Grab your device name (on Ubuntu, you will need to apt install iw):

iw dev | grep Interface | awk '{print $2}'

With that interface name, use nano/gedit or whichever editor you feel comfortable with:

sudo nano /etc/NetworkManager/conf.d/default-wifi-powersave-on.conf

edit 3 into 2, save the file.

Reboot

Open the terminal back up again, check your power save status:

iw dev <interface-name> get power_save

You should see Power save: off

Now, this may not necessarily help. On my own WAP, I am not seeing this issue and I have kept power save on.

Internally, our engineers have had success disabling power saving on their 6E networks.

[stenosis]

Same situation and configuration here. So far I haven’t found a solution, and unfortunately the tip with the adjusted configuration doesn’t work either. The WLAN card is not recognized for approx 20 to 30 seconds after login.

$ iw dev <interface-name> get power_save
command failed: No such device (-19)

And the dmesg output looks like this:

[   76.490501] mt7921e 0000:01:00.0: not ready 65535ms after FLR; giving up
[   76.644294] mt7921e 0000:01:00.0: enabling device (0000 -> 0002)
[   76.660565] mt7921e 0000:01:00.0: ASIC revision: 79220010
[   76.745805] mt7921e 0000:01:00.0: HW/SW Version: 0x8a108a10, Build Time: 20231120183400a
[   77.118386] mt7921e 0000:01:00.0: WM Firmware Version: ____000000, Build Time: 20231120183441
[   78.253446] mt7921e 0000:01:00.0 wlp1s0: renamed from wlan0

After a while, the WiFi card appears in the system, connects to the network and the output looks like this:

$ iw dev <interface-name> get power_save
Power save: off

So I guess it doesn’t necessarily have to do with the power settings of the WLAN card.

[Matt_Hartley]

Is this Ubuntu by chance? If so, I will be filing this bug Tues.

[stenosis]

It’s Manjaro, not Ubuntu.