kern/71155: [usb] misbehaving usb-printer hangs processes, causes panics

[Mikhail Teterin]

The following reply was made to PR kern/71155; it has been noted by GNATS.

From: Mikhail Teterin <mi at aldan.algebra.com>
To: freebsd-gnats-submit at freebsd.org
Cc:  
Subject: Re: kern/71155: [usb] misbehaving usb-printer hangs processes, causes panics
Date: Thu, 27 Jan 2005 06:41:41 -0500

 I'm seeing the same thing with a different system running 5.3-stable and
 a different HP printer.
 
 A freshly booted system prints fine, but after some uptime an attempt
 to print results in some commotion on the printer and a flood of
 messages like:
 
 Jan 26 12:19:32 aldan hpijs: unable to write to output, fd=6, count=4096: m
 Jan 26 12:19:32 aldan hpijs: unable to send DeviceStatus: m
 Jan 26 12:19:32 aldan hpijs: unable to write to output, fd=6, count=4096: m
 Jan 26 12:19:32 aldan hpijs: unable to write to output, fd=6, count=4096: m
 Jan 26 12:19:32 aldan hpijs: unable to send DeviceStatus: m
 
 (Indeed, hpijs is programmed to keep trying after a write(2) fails.)
 
 After some time, the whole machine panics -- even if I manage to kill
 the "offending" hpijs process.
 
 Could it be, that after some time it is harder for the driver to get
 memory and the shortage-handling code is broken?
 
 This is quite embarassing, actually. A simple print job to panic a
 server?
 
 	-mi