panic in uhci_abort_xfer: not in process context

[Dan Bilik]

Hi.

>From time to time I have to connect to the Net through USB
Qualcomm CDMA modem which really cannot be called "reliable piece of
hardware (tm)". It often experiences spontaneous shutoffs and sometimes
it just stops communicating on the port for a while and decides there
are better things to do (like blinking or catching stronger signal or
something). I'm OK with all this but unfortunately it triggers a panic
in RELENG_6...

(kgdb) bt 
#0  doadump () at pcpu.h:165

#1  0xc04f401c in boot (howto=260)
 at /usr/src/sys/kern/kern_shutdown.c:399

#2  0xc04f45bf in panic (fmt=0xc0667796 "uhci_abort_xfer: not in
 process context") at /usr/src/sys/kern/kern_shutdown.c:555

#3  0xc0499cb2 in uhci_abort_xfer (xfer=0xc37e2d00,
 status=USBD_NORMAL_COMPLETION) at /usr/src/sys/dev/usb/uhci.c:1957

#4  0xc0499e10 in uhci_timeout_task (addr=0x0)
 at /usr/src/sys/dev/usb/uhci.c:1534

#5  0xc049d9f0 in usb_task_thread (arg=0x0)
 at /usr/src/sys/dev/usb/usb.c:476

#6  0xc04d7f6d in fork_exit (callout=0xc049d9b0 <usb_task_thread>,
 arg=0x0, frame=0x0) at /usr/src/sys/kern/kern_fork.c:789

#7  0xc062b29c in fork_trampoline ()
 at /usr/src/sys/i386/i386/exception.s:208

There is similar PR (usb/76204), opened a year ago for 5.3-STABLE.
Anyone else seeing this behaviour on RELENG_6?

I wasn't able to trigger it manually by turning off the modem in the
middle of transfer. But the modem strikes quite often so I can call it
reproducable on that box, and it's possible to get more debugging
info on demand. Dmesg available at
http://neosystem.cz/freebsd/uhci_abort_xfer.dmesg

Dan