NetBSD Problem Report #8343
Received: (qmail 11490 invoked from network); 7 Sep 1999 23:33:43 -0000
Message-Id: <199909072338.AAA00546@buzzard.cambridge.arm.com>
Date: Wed, 8 Sep 1999 00:38:50 +0100 (BST)
From: rearnsha@arm.com
Reply-To: rearnsha@cambridge.arm.com
To: gnats-bugs@gnats.netbsd.org
Subject: bpf bug in if_ie.c
X-Send-Pr-Version: 3.95
>Number: 8343
>Category: port-arm32
>Synopsis: bpf bug in if_ie.c
>Confidential: no
>Severity: serious
>Priority: high
>Responsible: port-arm32-maintainer
>State: closed
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Tue Sep 07 16:35:00 +0000 1999
>Closed-Date: Tue Sep 14 06:14:13 +0000 1999
>Last-Modified: Tue Sep 14 06:16:51 +0000 1999
>Originator: Richard Earnshaw
>Release: 1.4 or later
>Organization:
ARM
>Environment:
System: NetBSD buzzard.cambridge.arm.com 1.4 NetBSD 1.4 (BUZZARD) #161: Tue Sep 7 21:21:53 BST 1999 rearnsha@buzzard.cambridge.arm.com:/home/rearnsha/netbsd/src/sys/arch/arm32/compile/BUZZARD arm32
>Description:
The #ifdef BPFILTER in podulebus/if_ie.c is miss-spelt. Further, the
header of ethernet package is not passed to the filter, which causes
it to interpret the packet incorrectly.
Without this patch it is impossible to serve DHCP on this
interface.
>How-To-Repeat:
Try to start dhcp on a subnet that is on a if_ie interface.
Observe that nothing happens :-(.
Try to run tcpdump on the same interface; observe that no input
packets are dumped :-(
Fix the name of the #ifdef; observe that dumped input packets are
garbage :-(
>Fix:
Patch attached. The patch uses the same trick as that in
net/if_loop.c (and probably other places).
*** /usr/src/sys/arch/arm32/podulebus/if_ie.c Sun Mar 28 18:50:01 1999
--- if_ie.c Tue Sep 7 21:20:53 1999
*************** ie_read_frame(sc, num)
*** 1314,1322 ****
ether_sprintf(eh.ether_shost), (u_int)eh.ether_type );
*/
! #if NBFILTER > 0
if ( ifp->if_bpf ) {
! bpf_mtap(ifp->if_bpf, m );
};
#endif
--- 1314,1331 ----
ether_sprintf(eh.ether_shost), (u_int)eh.ether_type );
*/
! #if NBPFILTER > 0
if ( ifp->if_bpf ) {
! /* We need to pass the ethernet header; cons up an mbuf to hold it.
! This is safe since the bpf will only read the packet, it will
! never try to free it. */
! struct mbuf m0;
!
! m0.m_next = m;
! m0.m_len = sizeof (struct ether_header);
! m0.m_data = (char *)&eh;
!
! bpf_mtap(ifp->if_bpf, &m0 );
};
#endif
>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->analyzed
State-Changed-By: tron
State-Changed-When: Mon Sep 13 13:20:39 PDT 1999
State-Changed-Why:
In NetBSD-current the mac header is passed to the BPF correctly. The "#ifdef"
was still broken but is fixed now.
I'll check if we can apply your patch to the release branch.
State-Changed-From-To: analyzed->closed
State-Changed-By: tron
State-Changed-When: Mon Sep 13 23:14:13 PDT 1999
State-Changed-Why:
Your patch has been applied to the release branch, thanks.
>Unformatted:
(Contact us)
$NetBSD: query-full-pr,v 1.51 2026/08/10 02:28:17 riastradh Exp $
$NetBSD: gnats_config.sh,v 1.10 2026/05/13 22:00:09 riastradh Exp $
Copyright © 1994-2026
The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.