Ethereal-dev: Re: [Ethereal-dev] Searching for packets with incorrect checksums

Note: This archive is from the project's previous web site, ethereal.com. This list is no longer active.

From: "Jim Fleming" <jfleming@xxxxxxxx>
Date: Tue, 18 Dec 2001 16:27:03 -0600
----- Original Message -----
From: "Guy Harris" <guy@xxxxxxxxxx>
To: "Jim Fleming" <jfleming@xxxxxxxx>
Cc: "Guy Harris" <guy@xxxxxxxxxx>; "Gilbert Ramirez" <gram@xxxxxxxxxxxxxxx>; "Michael Tuexen" <Michael.Tuexen@xxxxxxxxxxxxxxxxx>;
<ethereal-dev@xxxxxxxxxxxx>
Sent: Tuesday, December 18, 2001 3:55 PM
Subject: Re: [Ethereal-dev] Searching for packets with incorrect checksums


> > Sorry, IPv7 is an internal code-name for using the 0x8000 bit
> > of the offset field. My point was that when that bit is set (=1)
> > it shows up as 0 in the display and as an 0x80 in the hex dump.
>
> In what RFC or Internet draft is that specified?  Currently, we show
> only the DF and MF bits in that field.  There's a #define for 8000 for
> IP_CE, listed as a "Congestion" flag, but RFC 3168 seems to suggest that
> the ECN stuff shows up in the TOS field instead.
>

The left-most bit of the off-set field is normally 0.

http://www.dot-biz.com/IPv4/Tutorial/RIFRAFBSD4.4.txt

Apparently, someone thought they could use that bit for congestion
stuff that belongs in Layer 3 (TCP|UDP) not Layer 2 (IP), and they
left it hanging around in the code. The same can be said for TOS values.
Layer 2 needs all the bits it can get for addressing, to expand beyond
the 32-bits. A 20-byte header only has 160 bits, and as they say with
land, they are not making any more.

Jim Fleming
http://www.ddj.com/articles/search/search.cgi?q=fleming
Oct93: The C+@ Programming Language