how can I troubleshoot ulpt0 causing errors with my ural0?

(If you have a suggestion for a tech-* list to also post at,
please let me 
know.)

I have been using ural for over a year.

ral0 at uhub2 port 1
ral0: Cisco-Linksys Wireless-G USB Network Adapter, rev
2.00/0.04, addr 2
ral0: MAC/BBP RT2570 (rev 0x03), RF RT2526
ral0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps
ral0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps
12Mbps 18Mbps 
24Mbps 36Mbps 48Mbps 54Mbps

Today, I plugged in my printer via USB.

ulpt0 at uhub2 port 2 configuration 1 interface 0
ulpt0: Brother HL-5150D, rev 2.00/1.00, addr 3, iclass 7/1
ulpt0: using bi-directional mode

My networking started hanging a lot. I looked at netstat -i
and saw a 
bunch of "errs".

netstat man page says:

	BUGS
     		The notion of errors is ill-defined.

Why?

Watching "netstat -I ral0 -w 60" for about an hour
showed the input errors 
are staying the same, but the output errors are growing from
186 (where I 
started watching) to 224 errors. My network is hanging at
same point when 
that goes up. I don't recall having any errors before (even
though my 
networking would be up for over a week at a time). It
sometimes is just 
slow and sometimes hangs for more thanm 5 seconds.

I powered down my printer:

ulpt0: at uhub2 port 2 (addr 3) disconnected
ulpt0 detached

And now about five minutes later no more "errs"
(with ulpt0 disabled).

Any suggestions on how I can troubleshoot this? (Anything to
look at with 
"systat vmstat" for example??)

This is NetBSD/i386 3.99.24. I will upgrade soon to see if
this fixes it. 
(Nevertheless, I'd like to learn some troubleshooting
techniques for my 
education too.)

  Jeremy C. Reed

"Jeremy C. Reed" <reedreedmedia.net> writes:

> ral0: Cisco-Linksys Wireless-G USB Network Adapter, rev
2.00/0.04, addr 2
> ral0: MAC/BBP RT2570 (rev 0x03), RF RT2526

> ulpt0 at uhub2 port 2 configuration 1 interface 0
> ulpt0: Brother HL-5150D, rev 2.00/1.00, addr 3, iclass
7/1
> ulpt0: using bi-directional mode
>
> My networking started hanging a lot. I looked at
netstat -i and saw a 
> bunch of "errs".
>
> netstat man page says:
>
>      		The notion of errors is ill-defined.

At the netstat level it is ill-defined because there's no
specification
for how drivers count errors.  At the driver level, there's
a counter and
when incremented that's an error.  So see the ral source
for
ifp->if_oerrors++.   This can happen due to missing
interrupts, or the
card reporting that it never got an ack (excessive
retries).

  [output errors increase while printer is attached]

> Any suggestions on how I can troubleshoot this?
(Anything to look at with 
> "systat vmstat" for example??)

Look in the ral source (sys/dev/ic/rt2560.c?), and turn on
debugging so
you'll get a printf at each oerrors++.   When you know which
case is
being hit, you can try to figure out why.

I wonder if something in the usb stack is locking out
interrupts for too
long.