Warn about some broken drivers *cough* Broadcom *cough* ignoring

DLIOCRAW and overwriting the Ethertype with the SAP we happened to bind
to (the fact that we requested SAP promiscuity nonwithstanding).

diff --git a/pcap.3 b/pcap.3
index aa5939e32efe7a4d9c96226f49cb8d8ba1719084..2be5697614bdbc7de3238894411d662c8efd0d17 100644 (file)
--- a/pcap.3
+++ b/pcap.3
@@ -1,4 +1,4 @@
-.\" @(#) $Header: /tcpdump/master/libpcap/Attic/pcap.3,v 1.64.2.7 2005-07-07 02:04:35 guy Exp $
+.\" @(#) $Header: /tcpdump/master/libpcap/Attic/pcap.3,v 1.64.2.8 2005-09-07 08:29:17 guy Exp $
 .\"
 .\" Copyright (c) 1994, 1996, 1997
 .\"    The Regents of the University of California.  All rights reserved.
@@ -689,7 +689,11 @@ supplied to
 as the source link-layer address, if the header contains such an
 address, might be changed to be the address assigned to the interface on
 which the packet it sent, if the platform doesn't support sending
-completely raw and unchanged packets.
+completely raw and unchanged packets.  Even worse, some drivers on some
+platforms might change the link-layer type field to whatever value
+libpcap used when attaching to the device, even on platforms that
+.I do
+nominally support sending completely raw and unchanged packets.
 .PP
 .B pcap_sendpacket()
 is like