The following reply was made to PR kern/35273; it has been noted by GNATS.

From: Denis Lagno <dlagno@rambler.ru>
To: <gnats-bugs@NetBSD.org>
Cc: <joerg@NetBSD.org>, <netbsd-bugs@netbsd.org>,
	<gnats-admin@netbsd.org>
Subject: Re: kern/35273 (ex(4) is broken and may cause kernel crash)
Date: Mon, 22 Jan 2007 14:11:55 +0300

 * joerg@netbsd.org [Mon, 22 Jan 2007 10:30:26 +0000 (UTC)]:
 > About the route behaviour: the documented behaviour for 4.4BSD
 > was preference of host over net routes, so it is kind of intended.
 > It shouldn't have changed though, I'll take a look at it.
 
 I'll describe how I've noticed it:
 I have a script that connects to my provider via pppd and pptp.
 For example, 195.14.38.12 is address of pptp-vpn server.
 Then after pppd established connection, pppd automatically establishes 
 -iface -host route to vpn server via newly created ppp interface:
 195.14.38.12           89.178.41.185    UH
 
 (89.178.41.185 is address assigned to ppp interface):
 ppp0: flags=8051<UP,POINTOPOINT,RUNNING,MULTICAST> mtu 1500
         inet 89.178.41.185 -> 195.14.38.12 netmask 0xff000000
         inet6 fe80::210:4bff:fe67:8727%ppp0 ->  prefixlen 64 scopeid 0x5
 
 However in case of my provider this route does not work.  Packets for 
 195.14.38.12 should be routed instead via interface ex0 and local 
 gateway 10.239.0.17:
 
 ex0: flags=8863<UP,BROADCAST,NOTRAILERS,RUNNING,SIMPLEX,MULTICAST> mtu 
 1500
         capabilities=3f00<IP4CSUM_Rx,IP4CSUM_Tx,TCP4CSUM_Rx,TCP4CSUM_Tx,UDP4CSUM_Rx,UDP4CSUM_Tx>
         enabled=3f00<IP4CSUM_Rx,IP4CSUM_Tx,TCP4CSUM_Rx,TCP4CSUM_Tx,UDP4CSUM_Rx,UDP4CSUM_Tx>
         address: 00:10:4b:67:87:27
         media: Ethernet autoselect (100baseTX full-duplex)
         status: active
         inet 10.239.17.213 netmask 0xffff0000 broadcast 10.239.255.255
         inet6 fe80::210:4bff:fe67:8727%ex0 prefixlen 64 scopeid 0x1
 
 Previously I solved this by setting following route before executing 
 pppd:
 195.14.38.12/32    10.239.0.17        UG
 
 This route somehow took preference over -host -iface route created by 
 pppd.
 However with netbsd-4.99.8 it does not work, and correct solution is to 
 delete route created by pppd in /etc/ppp/ip-up
 
 -- denis