Re: Regression xhci_hcd cmd failed due to incorrect slot or ep state

[Mathias Nyman <mathias.nyman@linux.intel.com>]

On 7.4.2019 13.32, Bernhard Gebetsberger wrote:
> When using  the kernel version 4.20 or above my usb wifi stick doesn't
> work until I reboot my system, dmesg shows this message when the issue
> occurs:
> 
> xhci_hcd 0000:15:00.0: WARN Set TR Deq Ptr cmd failed due to incorrect slot or ep state
> 
> I've found that the problem is caused by this commit:
> 
> commit f8f80be501aa2f10669585c3e328fad079d8cb3a
> Author: Mathias Nyman <mathias.nyman@linux.intel.com  <mailto:mathias.nyman@linux.intel.com>>
> Date:   Thu Sep 20 19:13:37 2018 +0300
> 
>      xhci: Use soft retry to recover faster from transaction errors
> 
> It looks like it only affects wifi sticks(various models) on specific
> hardware configurations(I have this issue on my Ryzen pc, my Intel pc
> isn´t affected, but someone in the kernel bug tracker mentioned that he
> has this issue on an Intel notebook as well (only with USB 3.0 ports
> though)).

Thanks for the report.

I can see how these would be related, setting a new Transfer Ring dequeue pointer
requires the ep ring to be in Error or stopped state, otherwise you see the above message.

The soft retry patch will automatically restart a ring from stopped to running state.

But to know the details where the issue exactly is, or how we ended up there, I would need to
look at logs and traces.

Logs and traces can be taken with:

mount -t debugfs none /sys/kernel/debug
echo 'module xhci_hcd =p' >/sys/kernel/debug/dynamic_debug/control
echo 'module usbcore =p' >/sys/kernel/debug/dynamic_debug/control
echo 81920 > /sys/kernel/debug/tracing/buffer_size_kb
echo 1 > /sys/kernel/debug/tracing/events/xhci-hcd/enable

< trigger failure >

Send output of dmesg to me, and
copy /sys/kernel/debug/tracing/trace to me.

-Mathias