tcpdump versus TSO

Rick Jones <rick.jones2@hp.com> · Fri, 21 May 2004 16:52:48 -0700

David Mosberger suggested it might be good for me to ask this question 
here - if anyone has already encountered issues with TSO and tcpdump.

rick jones
--- Begin Message ---

To: Rick Jones <rick.jones2@hp.com>
Subject: Stalled post to tcpdump-workers
From: tcpdump-workers-owner@lists.tcpdump.org
Date: Fri, 21 May 2004 18:46:34 -0400

Your message to tcpdump-workers has been delayed, and requires the approval 
of the moderators, for the following reason(s):

The postblock flag is set for Rick Jones <rick.jones2@hp.com>.

If you do not wish the message to be posted, or have other concerns, 
please send a message to the list owners at the following address: 
  tcpdump-workers-owner@lists.tcpdump.org
--- Begin Message ---

To: tcpdump-workers@tcpdump.org
Subject: tracing with TSO enabled on linux 2.6
From: Rick Jones <rick.jones2@hp.com>
Date: Fri, 21 May 2004 15:45:51 -0700
User-agent: Mozilla/5.0 (X11; U; HP-UX 9000/785; en-US; rv:1.6) Gecko/20040304

Does anyone have pointers as to what might need to be done to "fix" 
tracing under Linux 2.6 with TSO enabled on an interface?  It seems that 
for locally transmitted packets when TSO is being used the total length 
field in the IP header is being set to 0, and this makes the IP print 
routine somewhat grumpy, emitting "bad-len" messages.

rick jones

--- End Message ---

--- End Message ---