From owner-freebsd-hackers  Thu Mar  2 15:29:08 1995
Return-Path: hackers-owner
Received: (from majordom@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id PAA14157 for hackers-outgoing; Thu, 2 Mar 1995 15:29:08 -0800
Received: from mwunix.mitre.org (mwunix.mitre.org [128.29.154.1]) by freefall.cdrom.com (8.6.9/8.6.6) with ESMTP id PAA14136 for <freebsd-hackers@freebsd.org>; Thu, 2 Mar 1995 15:29:00 -0800
Received: from fluky.mitre.org (fluky.mitre.org [128.29.113.24]) by mwunix.mitre.org (8.6.10/8.6.4) with SMTP id SAA27251; Thu, 2 Mar 1995 18:28:41 -0500
Received: from [128.29.113.163] (m21378-mac2.mitre.org) by fluky.mitre.org (4.1/SMI-4.0)
	id AA15577; Thu, 2 Mar 95 18:24:31 EST
X-Sender: slr@fluky (Unverified)
Message-Id: <v01510108ab7bfd8c0636@[128.29.113.163]>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Thu, 2 Mar 1995 18:33:48 -0500
To: tinguely@plains.nodak.edu, wollman@halloran-eldar.lcs.mit.edu
From: slr@mitre.org (Soochon Radee)
Subject: multicast pruning
Cc: atn-group@mitre.org, freebsd-hackers@FreeBSD.org, ajit@udel.edu
Sender: hackers-owner@FreeBSD.org
Precedence: bulk

Here's another piece of the multicast problem solved. FreeBSD was unable to
prune with big-endian machines (Sun) because some differences in byte order
were not dealt with in the mrouted distribution code. These two diffs fix
that problem:

lusty# diff igmp.c igmp.c.orig
184c184
<                                (char *)(igmp+1), igmpdatalen, ntohl(group));
---
>                                (char *)(igmp+1), igmpdatalen, group);
189c189
<                                 (char *)(igmp+1), igmpdatalen, ntohl(group));
---
>                                 (char *)(igmp+1), igmpdatalen, group);

lusty# diff prune.c prune.c.orig
533d532
< #if BYTE_ORDER == BIG_ENDIAN
535,538d533
< #endif
< #if BYTE_ORDER == LITTLE_ENDIAN
<     for (i = 3; i >= 0; i--)
< #endif

The first set of diffs fixes the version number other multicast routers
send (that's why 'kill -USR1' dumps show 0.0 peer versions), and the other
one writes the prune lifetime in network order.

With the upcall and prune problems solved, all that's left is to check that
the LKM version works properly and some testing, and mrouted should be
ready for the 2.1 Release.

Soochon