Re: RTW88 USB Wi-Fi Adapter (8822BU) Disconnects and Fails with USB Write Errors - Status -71 & -110

[Michal Pecio <michal.pecio@gmail.com>]

On Fri, 17 Oct 2025 17:25:03 +0000, Samuil wrote:
> > Does it happen every time on every boot?  
> The disconnect happens on every boot although rarely the device is
> able to reconnect again. Usually, however, the adapter just stops
> working until I reboot again.

Well, I had no luck reproducing this on v6.18-rc1.
I left it running overnight with ping traffic ~700KB/s each way.

> > May we know what 'lspci -nn' says about this 0000:02:00.0 controller?  
> 02:00.0 USB controller [0c03]: Advanced Micro Devices, Inc. [AMD] 400 Series Chipset USB 3.1 xHCI Compliant Host Controller [1022:43d5] (rev 01)

Similar to my B350 chipset.

In some stress tests it showed a HW bug which currently isn't being
worked around by xhci_hcd and I think this bug may be responsible for
at least some of the noise you get, namely this sequence with constant
"event dma" but increasing "TD at":

Event dma 0x0000000105a9e110 for ep 0 status 4 not part of TD at 000000011fd2bd30 - 000000011fd2bd50
Event dma 0x0000000105a9e110 for ep 0 status 4 not part of TD at 000000011fd2bd90 - 000000011fd2bdb0  

It's a broken state which rtw88 can't do anything about. The dynamic
debug I asked for may shed some light on how you got into this state.

Whether fixing this would enable rtw88 to continue is unclear, because
everything likely starts with a "Transaction Error" on endpoint 0. That
could either be a one-off error (harmless if xHCI recoveres from it and
rtw88 retries whatever it wanted to do), or the USB device crashes due
to some traffic pattern unique to your network or whatever.

> Are you using a USB 2 port for the adapter to be working properly
> without configuration?

Nope, I made sure to run it SuperSpeed on my buggy AMD chipset.
This is probably unique to your network traffic or hardware (poor
USB link quality? maybe try other ports.)

Still, I would be curious to see this dynamic debug if possible.

Are you able to compile and run a patched kernel?

Regards,
Michal