Hello folks.  I have a machine with a Realtek 8139 dcard in it that I'm 
trying to get running.  If I cause the chip to go out to lunch, and the
rl_watchdog() routine to get called, the system panics just after the
watchdog timer goes off, when the next packet is due to be transmitted.
The problem seems to be that m_freem() is getting called with an invalid
address from rl_encap(). I think what's happening is that the rl_softc
structure is holding invalid memory references to the chip after the reset
occurs, but I don't completely understand what needs to happen when
rl_watchdog goes off in order to get things rolling again.  Has anyone else
encountered this problem, or is there a fix in a later driver I can pull
back to this version?  
I'm using:
/* $NetBSD: if_rl.c,v 1.2.6.1 1999/10/20 23:39:22 he Exp $ */

Or, if someone can tell me what I need to do, I'll be happy to do it and
submit a pr.
-thanks
-Brian