Date: Mon, 28 Jul 1997 02:10:22 -0700 (PDT) From: denny1@home.com To: freebsd-gnats-submit@FreeBSD.ORG Subject: kern/4184: minor nits in sys/netatalk Message-ID: <199707280910.CAA19754@hub.freebsd.org> Resent-Message-ID: <199707280920.CAA20222@hub.freebsd.org>
next in thread | raw e-mail | index | archive | help
>Number: 4184
>Category: kern
>Synopsis: minor nits in sys/netatalk
>Confidential: no
>Severity: non-critical
>Priority: low
>Responsible: freebsd-bugs
>State: open
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Mon Jul 28 02:20:02 PDT 1997
>Last-Modified:
>Originator: Denton Gentry
>Organization:
>Release: N/A
>Environment:
Found while porting netatalk to OpenBSD.
>Description:
Several minor problems were uncovered while bringing up netatalk
under OpenBSD. They are attached below.
Also a number of changes to tcpdump to better support AppleTalk
were made. Those changes have been submitted to LBL, but I do not
know when they might be included in a new tcpdump release. If FreeBSD
wants to incorporate tcpdump changes for AppleTalk before that I will
happily mail the diffs.
>How-To-Repeat:
>Fix:
The first change was made in umich 1.4a2 relative to 1.3.3.
In aarp.c incoming AARP packets are checked against the broadcast
address. In some environments they saw broadcast wars...
coventry[145] diff -c aarp.c aarp.c.orig
*** aarp.c Sun Jul 27 23:04:01 1997
--- aarp.c.orig Sun Jul 27 23:03:00 1997
***************
*** 319,335 ****
return;
}
- /*
- * Check if from broadcast address. This could be a more robust
- * check, since we could look for multicasts. XXX
- */
- if ( !bcmp(( caddr_t )ea->aarp_sha, ( caddr_t )etherbroadcastaddr,
- sizeof( etherbroadcastaddr ))) {
- log( LOG_ERR, "aarp: source is broadcast!\n" );
- m_freem( m );
- return;
- }
-
op = ntohs( ea->aarp_op );
bcopy( ea->aarp_tpnet, &net, sizeof( net ));
--- 319,324 ----
The second change is in ddp_output.c. FreeBSD's port of 1.3.3
removed an m_pullup, probably because it failed for length
MLEN (can only pullup for MHLEN w/ 4.4BSD). That means phase2->phase1
routing results in mbufs at odd byte alignment. For x86 this is not
an issue, but if FreeBSD does an alpha port it will be.
coventry[146] diff -c ddp_output.c ddp_output.c.orig
*** ddp_output.c Wed Mar 5 03:13:49 1997
--- ddp_output.c.orig Sun Jul 27 23:05:56 1997
***************
*** 141,147 ****
* This is bad for transition routing, since phase 1 and phase 2
* packets end up poorly aligned due to the three byte elap header.
*/
! if ( !(aa->aa_flags & AFA_PHASE2) ) {
MGET( m0, M_WAIT, MT_HEADER );
if ( m0 == 0 ) {
m_freem( m );
--- 141,151 ----
* This is bad for transition routing, since phase 1 and phase 2
* packets end up poorly aligned due to the three byte elap header.
*/
! if ( aa->aa_flags & AFA_PHASE2 ) {
! if (( m = m_pullup( m, MIN( MHLEN, m->m_pkthdr.len ))) == 0 ) {
! return( ENOBUFS );
! }
! } else {
MGET( m0, M_WAIT, MT_HEADER );
if ( m0 == 0 ) {
m_freem( m );
>Audit-Trail:
>Unformatted:
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199707280910.CAA19754>