Subject: bin/12418: dhcrelay -i core dumps
To: None <gnats-bugs@gnats.netbsd.org>
From: None <hag@linnaean.org>
List: netbsd-bugs
Date: 03/15/2001 20:11:40
>Number:         12418
>Category:       bin
>Synopsis:       dhcrelay -i core dumps
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    bin-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Mar 15 17:12:00 PST 2001
>Closed-Date:
>Last-Modified:
>Originator:     Daniel Hagerty
>Release:        NetBSD-1.5.1_ALPHA-current
>Organization:
	
>Environment:
	
System: NetBSD neuralgia.linnaean.org 1.5_BETA NetBSD 1.5_BETA (NEURALGIA) #92: Sun Oct 29 17:15:25 EST 2000 hag@neuralgia.linnaean.org:/fs/neuralgia/home/hag/work/hacking/os/netbsd/src/sys/arch/i386/compile/NEURALGIA i386
Architecture: i386

>Description:
	dhcrelay will core dump if started with -i option to limit it
to specific interfaces.

>How-To-Repeat:
	dhcrelay -i <yourethernet> 10.1.1.1

>Fix:
	Not known.  A cursory glance at dhcrelay.c shows that the -i
option calls interface_allocate (defined by a macro in
common/discover.c) without having initialized the global
dhcp_type_interface defined by discover.c .  The fact that omapi_init
hasn't been called yet probably doesn't help things.

	dhcpd does things in a very different order; omapi_init is
called almost immediately, and dhcp_common_objects_setup is called to
initialize dhcp_type_interface .

	I'll keep pounding on it until someone with more clue tells me
to stop; I need this.
>Release-Note:
>Audit-Trail:
>Unformatted: