Warning from USC core on kernel 3.13

Warning from USC core on kernel 3.13

[Larry Finger]

Alan,

The Lenovo Yogi 13 tablet comes with a Realtek RTL8723AU wireless device built 
in. Realtek sent me a driver that I modified so that it would build on new 
kernels, and created a GitHub repo so that it would be available to the 
community. One of the users of this driver is reporting intermittent warnings 
that say

WARNING: CPU: 1 PID: 0 at drivers/usb/core/urb.c:452 usb_submit_urb+0x205/0x5d0()
usb 1-1.4: BOGUS urb xfer, pipe 3 != type 1

The full splat is available at http://pastebin.com/VcFPd4Yt.

As no other user has reported this problem, I think this is some kind of problem 
with the wireless device or the adapter on this box. Even so, I would like to 
know more about what is happening. In addition, Jes Sorensen at RedHat is 
working on making the driver suitable for submission to the staging part of the 
tree. Just in case other uses run into the problem, it would be good to know if 
a fix is possible.

Can you suggest any addition logging for when this situation occurs? The user is 
generating his own kernels, thus we can add any patches we want.

Thanks,

Larry

Re: Warning from USC core on kernel 3.13

[Alan Stern]

On Thu, 20 Feb 2014, Larry Finger wrote:

> Alan,
> 
> The Lenovo Yogi 13 tablet comes with a Realtek RTL8723AU wireless device built 
> in. Realtek sent me a driver that I modified so that it would build on new 
> kernels, and created a GitHub repo so that it would be available to the 
> community. One of the users of this driver is reporting intermittent warnings 
> that say
> 
> WARNING: CPU: 1 PID: 0 at drivers/usb/core/urb.c:452 usb_submit_urb+0x205/0x5d0()
> usb 1-1.4: BOGUS urb xfer, pipe 3 != type 1
> 
> The full splat is available at http://pastebin.com/VcFPd4Yt.
> 
> As no other user has reported this problem, I think this is some kind of problem 
> with the wireless device or the adapter on this box. Even so, I would like to 
> know more about what is happening. In addition, Jes Sorensen at RedHat is 
> working on making the driver suitable for submission to the staging part of the 
> tree. Just in case other uses run into the problem, it would be good to know if 
> a fix is possible.
> 
> Can you suggest any addition logging for when this situation occurs? The user is 
> generating his own kernels, thus we can add any patches we want.

You probably don't need any additional logging.

The WARNING is telling you that the usb_read_interrupt_complete() 
routine in the 8723au driver called usb_submit_urb() with an incorrect 
pipe value.  The URB specified a BULK pipe, but the endpoint in 
question is actually an INTERRUPT endpoint.

Alan Stern