Subject: telnet localhost w/o IPv6 network
To: None <port-alpha@netbsd.org, current-users@netbsd.org>
From: Michael G. Schabert <mikeride@mac.com>
List: current-users
Date: 03/21/2002 04:04:54
Hi guys,
I'm running 1.5ZC current on an AlphaStation 200, with a custom 
kernel. The kernel is basically a pared-down GENERIC, getting rid of 
devices I don't have. The only things added to the kernel config are 
uncommenting options/netatalk and changing the two instances of "pms" 
to "pmsi". I left in IPv6 support, but am not running any IPv6 
network, & have nothing setup wrt that. I was testing a daemon, & 
tried "telnet localhost ###" (### being requested port) and the 
everything came to a grinding halt. All disk i/o stopped, all remote 
sessions stopped responding. I had X running on the TGA console, 
which aparently just had not cleared and moved back into console 
mode. But I found that no matter what port I requested, this would 
happen. When doing it when not running X, I could see that what 
really happened was that I dropped into the debugger. Here's what was 
on screen:

CPU 0	trap entry = 0x2 (memory management fault)
CPU 0	a0	= 0x10
CPU 0	a1	=	0x1
CPU 0	a2	=	0x0
CPU 0	pc	=	0xfffffc0000465610
CPU 0	ra	=	0xfffffc00004655fc
CPU 0	pv	=	0xfffffc00005f0fec
CPU 0	curproc	=	fffffc0001aace70
CPU 0	pid	=	1152, comm	=	telnet

panic: trap
stopped in pid 1152 (telnet) at cpu_debugger+0x4:	ret	zero,(ra)

Note that I can successfully connect to myself if I use 192.168.1.1, 
and even 127.0.0.1. Just not localhost. Trying "c" from debugger 
starts for a second and then drops back into debugger (although 
admittedly I forgot to kill 1152 first), so I rebooted from the 
debugger. As this is consistent, let me know if I also need a 
backtrace to help.

Oh, and the reason I consider it to be an IPv6 issue is because of 
what happens as I do the connection...it tries connecting to ::1: or 
something like that.

Mike
-- 
Bikers don't *DO* taglines.