kern/63982: em0 hardware checksum offloading causes bogus errors in
tcpdump
ian j hart
ict at cardinalnewman.coventry.sch.uk
Tue Mar 9 06:20:16 PST 2004
The following reply was made to PR kern/63982; it has been noted by GNATS.
From: ian j hart <ict at cardinalnewman.coventry.sch.uk>
To: freebsd-gnats-submit at FreeBSD.org
Cc:
Subject: kern/63982: em0 hardware checksum offloading causes bogus errors in
tcpdump
Date: Tue, 9 Mar 2004 03:59:05 -0800 (PST)
>Number: 63982
>Category: kern
>Synopsis: em0 hardware checksum offloading causes bogus errors in tcpdump
>Confidential: no
>Severity: non-critical
>Priority: low
>Responsible: freebsd-bugs
>State: open
>Quarter:
>Keywords:
>Date-Required:
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Tue Mar 09 04:00:31 PST 2004
>Closed-Date:
>Last-Modified:
>Originator: ian j hart
>Release: 4.9-RELEASEp3
>Organization:
Cardinal Newman School
>Environment:
backup.cardinalnewman.lan 4.9-RELEASE-p3 FreeBSD 4.9-RELEASE-p3 #12: Tue Mar 9 10:13:10 GMT 2004 root at backup.cardinalnewman.lan:/usr/obj/usr/src/sys/GENERIC i386
patched with src/sys/dev/em from RELENG_4
>Description:
Running tcpdump -v shows many packets marked as having bad
checksums. The packets are delivered okay with correct checksums. This
mainly affects UDP.
I booted from disc2.iso for 4.7, 4.8, and 5.2.1RC and confirmed the
problem there.
One motherboard (GA-7VAXP-A Ultra) DOES work but all others tried so far, fail. Card in different slots, natch`. irq/timing?
>How-To-Repeat:
tcpdump -v | grep bad
Then generate some UDP traffic
>Fix:
A workaround and proof of the source of the problem is to
disable the hardware checksum.
/usr/src/sys/dev/em/if_em.h
#define EM_CHECKSUM_FEATURES (CSUM_TCP | CSUM_UDP)
Set to zero
>Release-Note:
>Audit-Trail:
>Unformatted:
_______________________________________________
freebsd-bugs at freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-bugs
To unsubscribe, send any mail to "freebsd-bugs-unsubscribe at freebsd.org"
More information about the freebsd-bugs
mailing list