AP stops with mt76x2u: error: mt76x02u_mcu_wait_resp failed with -110 #405
Comments
|
[ 12.917747] mt7615e 0000:02:00.0: Invalid firmware |
|
I wasn't even using it and it died again. |
Sorry I don't understand. |
|
I upgraded to 2020-06-02. Firmware Version: OpenWrt SNAPSHOT r13427-a47acae319 / LuCI Master git-20.154.25027-09014e0 |
|
It happened again but it seems to have more information in the system log. Kernel log: |
|
It's happening without any app updates on my |
|
The "interface" name keeps increasing when I reinsert the USB, it is "wlan4" now. |
|
|
It lasted longer this time.
I noticed luci still says my cellphone is associated. |
|
Raspberry Pi 4 Model B Rev 1.1 Thu Jun 18 11:06:34 2020 kern.err kernel: [220856.547498] mt76x2u 2-1:1.0: error: mt76x02u_mcu_wait_resp failed with -110 |
I've ordered a powered hub to see if that helps. |
|
Raspberry Pi 4 Model B Rev 1.1 |
|
I am downloading a big patch for my Android phone and it keeps crashing. :-( |
|
|
AP died again when I was playing around with iperf.
|
|
That's all that's in the log but my AP is dead.
|
|
My cell phone was downloading updates in the background. |
|
Cell phone downloading updates. |
|
Raspberry Pi 4 Model B Rev 1.1
I tried a script to simulate manually reinserting the USB but that didn't work: |
|
Powered hub, short/no cables. |
|
Kernel Version 5.4.48 Adding disable_usb_sg=1 to mt76_usb module. |
|
This is new: Raspberry Pi 4 Model B Rev 1.1 |
Same problem here but happens only when more that one device is connected on USB (I have mediatek on USB 3.0 and an atheros ath9k_htc on USB 2.0) This "disable_usb_sg=1 to mt76_usb module" is in current snapshot? |
|
I compiled latest MT76 code and now I get:
|
|
without: echo 1 > /sys/module/mt76_usb/parameters/disable_usb_sg in case the trace helps. |
|
So, I have built b5df0fb and the problem is still there on a RPi4 with Openwrt snapshot Apparently, but I need to do some more testing, the problem comes when there is some sustained (even if nothing crazy) traffic going over the air and for the moment I have seen it only with an android phone. I usually have a Windows10 PC with Intel WiFi card, and so far I did not see the problem happening with it, but I try to make some more tests |
|
Are using it with scatter-gather on demand disabled? |
|
No, but I can give it a shot. It is enough to run it runtime? Or shall I do something else, like disconnect/connect? |
Unplug the USB, run that command, plug in the USB. I created so it happens on reboot: This will show the status: |
|
Hi @bjlockie Yes apparently it solves the issue Thanks, bye |
|
Same boat as everyone above (RPi4 + My understanding is that, at this stage, the only hope for a fix, is in a firmware update for RPi? As a workaround in the meantime, has anyone come up with a script to reset the connection regularly? (physically unplugging the USB stick works, but there has to be a more efficient way to detect when the error occurred and reload the device…? |
|
Maybe removing the mt76 driver and readdiing it?
|
|
Any updates on this issue? |
|
@TrylobitBroccoli unfortunately, even after ensuring drivers/firmware were the very latest, and playing around every possible port config, the bottom line is that RPi + USB3 + wifi is a no-go… Which is a pretty huge limitation, and ought to be a little more well-known, imho. |
|
@zedrdave That's so disappointing. I've got the RPi 4B just for this purpose... Is there an alternative that can be used that is OpenWRT compatible and similar to RPi and doesn't produce this kind of problems? |
|
@TrylobitBroccoli @zedrdave have you tried using the the USB2 ports? This improves stability for me, but occasional |
|
@TrylobitBroccoli I actually use a NanoPi R4S as a router with a mt76x2u type wifi adapter for the access point, and after switching off SG as part of the module initialization, the system appears to be working stable. I have this setup running now for over a week without issues (so yes, still pretty new). when leaving SG enabled, the -110 error would appear after 2-3 days. |
|
I had a solid AP after I used a powered USB 3 hub and disabled scatter gather.
VIA is apparently working on a firmware fix.
|
|
Unfortunately USB 2 is even less stable. Running on a powered USB 3 (with SG disabled) leads to random disconnects, every few GB of traffic… |
|
Just got this error after putting adapter behind a USB3 hub. I put it back directly. @LorenzoBianconi expected? |
|
A powered USB 3 hub?
|
|
No, passive |
|
Some people have had 0 luck with any kind of hub or extension cable.
I use a powered USB 3 hub but probably got lucky with what I bought.
|
|
I tried both passive and powered USB hubs between the WNDR3800 and the mt76, does not help. |
|
Is there something about power saving you can turn of in mt76? |
|
I had the exact same issue when I added a USB 3.0 ALFA AWUS036ACM to my Raspberry Pi 4 router and Raspberry Pi 4 Access Point. I could reproduce the problems simply by running the speed test from my Android phone and either the router or the access point WiFi would stop working. To get the WiFi working again I often would perform a reboot. Now that I have the fix from @bjlockie (to disable the scatter gather in USB) I seem to be able to run all the speed tests that I want. Fingers crossed I will report how long my OpenWRT Raspberry Pi devices continue to function. |
|
Unfortunately disabling SG (or any of the other USB port fiddling suggestions) never solved the issue for me. Kept having random disconnects every few days (presumably when reaching some arbitrary traffic threshold). Eventually gave up and plugged it to a Wifi repeater through ethernet (works like a charm). |
|
I have been running with no issues for over 11 days since I disabled usb scatter gather feature. I am only using 5 GHZ and I have two different Raspberry Pi's one with OpenWRT 22.03.5 and the main router with OpenWRT 22.03.3. I have fast transition turned on so I can roam between the two OpenWRT systems. |
|
I also have (had?) this issue, and disabling SG didn't help.
These actions usually make me reboot Raspberry Pi after an hour of usage because Wi-Fi gets lost. |
|
@Qualtagh , just wondering, how did you disable SG? |
|
I think this is it: |
|
@bjlockie that sound about right. in the past I has problems when I changed the parameter after module initialization, but doing it in a way as you describe worked well for me. |
|
quick update..... I used the fix from @bjlockie to disable scatter gather on OpenWRT version 22.03.5 and version 22.03.3 and it has been stable for months. Today I setup a new RasPi 4 with a USB 3.0 ALFA AWUS036ACM with OpenWRT version 23.05.2 and disabling scatter gather actually makes the WiFi AP unstable under high load. I removed the file /etc/modules.d/mymoduleparam.conf and rebooted and now the WiFi AP is very stable with good performance. I am guessing there is a patch that fixed this issue in 23.05.2 and disabling scatter gather is no longer required. Maybe this issue can be closed now? |
|
I encounter this kind of AP stops with my Raspberry Pi 4B, with Alfa AWUS036ACM, connected to USB 2, though. /etc/init.d/network restart does not help. OpenWrt version 23.05.2, r23630-842932a63d |
|
This is not just with the RPI4. I am seeing this regularly with a haswell laptop with an usb mt76 wifi adapter (comfast CF-WU782ACV2). Wifi hangs, then after some time it spontaneously recovers. Now trying with scatter/gather disabled. |
|
Personally, I have gave up with the Comfast CF-WU782ACV2 which was not just the source of a significant frustration, but also originally not cheap. Scatter/Gather was not the origin of the problem and possibly my board had some inherent design/hardware/firmware flaw or was a counterfeit. I have bought a Fenvi AX1800 on AliExpress that is also Mediatek based (mt7921u) that came at 1/5 of the cost (7€) and differently from the Comfast so far works fine out of the box. It has smaller antennas and a somehow lower antenna gain than the Comfast, though. The immediately evident thing that does not work with the linux driver is the led. Have not tried as an AP, yet. |
|
FWIW confirming that problem is still present for v23.05.3 on the WNDR3800/AR7161 platform, even with /sys/module/mt76_usb/parameters/disable_usb_sg=y Possibly less severe than earlier versions, but I had one lockup after around 8 days uptime with SG disabled, and it only took one day to crash with SG enabled. Other platforms stable. Really wish I hadn't bought a bunch of the AWUS036ACMs, now looking at alternative hardware.. |
|
I am really scratching my head...I have two openWRTs running on Raspberry Pi. OpenWrt 22.03.5 r20134-5f15225c1e. One has up-time of 112 days the other has up-time of 169 days. I use the WiFi everyday with no issues. I can upgrade to OpenWrt 23 and see what happens. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
My AP usually stops when I update all the packages on my Android phone that were installed via google play.
I tried to reproduce it with iperf (client on my phone and server on my desktop).
It happened once but works alot.
I haven't found a way to reproduce it on demand.
I use a Raspberry Pi4 with an imagebuilder 2020-05-14 build of openwrt
and an Alfa AWUS036ACM (Mediatek MT7612U).
I have it set as a 5GHz access point.
I found this bug report that might be related:
mt76x2u mac specific condition occurred #403
After remove and insert USB device:
I reset USB device from Luci and it works.
Any ideas on how to reproduce it?
The text was updated successfully, but these errors were encountered: