6.2, USB wedged

[Charles Sprickman]

Hi all,

I'm not too familiar with USB in 5.x or 6.x.  After some great one-on-one 
work with one of the NUT UPS tools devs, he got many linuxisms out of the 
tripplite_usb code so I can monitor a UPS here.

The code is likely not perfect, and sometimes the NUT daemon needs to be 
restarted to get fresh data.

Today it did this and I accidentally started two instances of the daemon. 
This seems to have completely wedged up the USB stack.  The tripplite_usb 
process is unkillable, and is in disk wait state, which seems odd:

nut       59693  0.0  0.1  1480  1028  p1  D     1:34PM   0:00.00 
/usr/local/libexec/nut/tripplite_usb -a colo1

All the "kill -9"'s in the world won't stop it.

I ran "usbdevs" to see if it was able to probe any devices and that too 
hung and is now unkillable:

[root at miko /home/spork]# usbdevs
addr 1: UHCI root hub, VIA
^C^C^C

root      31627  0.0  0.1  1264   580  p1  D+    5:33PM   0:00.00 usbdevs

Also in disk wait...

Any ideas, short of a reboot to kick things back into shape?

Here's the other processes, they all seem to be in disk wait and "waiting 
to acquire a lock" (DL):

root         25  0.0  0.0     0     8  ??  DL   14Mar07   0:00.97 [usb0]
root         26  0.0  0.0     0     8  ??  DL   14Mar07   0:00.00 
[usbtask]
root         28  0.0  0.0     0     8  ??  DL   14Mar07   0:00.98 [usb1]
root         29  0.0  0.0     0     8  ??  DL   14Mar07   0:00.73 [usb2]
root         31  0.0  0.0     0     8  ??  DL   14Mar07   0:00.72 [usb3]
root      34496  0.0  0.1  1288   712  ??  Is    5:42PM   0:00.00 
/usr/sbin/usbd
root      31627  0.0  0.1  1264   580  p1  D+    5:33PM   0:00.00 usbdevs
nut       59693  0.0  0.1  1480  1028  p1  D     1:34PM   0:00.00 
/usr/local/libexec/nut/tripplite_usb -a colo1

Also any pointers to more info on USB for users would be appreciated.

Thanks,

Charles

___
Charles Sprickman
NetEng/SysAdmin
Bway.net - New York's Best Internet - www.bway.net
spork at bway.net - 212.655.9344