tcpdump mailing list archives

Re: Proposed new pcap format


From: "Fulvio Risso" <fulvio.risso () polito it>
Date: Wed, 14 Apr 2004 08:14:03 +0200



-----Original Message-----
From: tcpdump-workers-owner () lists sandelman ca
[mailto:tcpdump-workers-owner () lists sandelman ca]On Behalf Of Loris
Degioanni
Sent: martedì 13 aprile 2004 20.18
To: tcpdump-workers () tcpdump org
Subject: Re: [tcpdump-workers] Proposed new pcap format


Hi,



----- Original Message -----
From: "Loris Degioanni"
Sent: Monday, April 12, 2004 2:55 PM
Subject: Re: [tcpdump-workers] Proposed new pcap format



Essentially, what you propose is that the SHB CONTAINS a
section rather
than
MARKING its beginning. The SHB, in fact, as any other block,
includes a
Block Total Length field, which could be used to specify the length of
data
that follows the header.
However, this field is 32 bit only. Do you think
it's too short, considering that we could put another SHB after 4 GB?

32 bits is too short.  (some) People are already today using
2GB capture
files with all
the pains that brings with pcap and pcap-supposed-to-be-compliant
implementations
that treat file-length/offset as signed integer.

So, the solutions are:

a. to use a field of the SHB instead than its length, to specify
the length
of a section. This field can be 64 bit.

b. to use the length of the SHB, modifying the block header to
contain a 64
bit "Block Total Length" rather than a 32 bit one. This solution has the
disadvantage of making the capture files longer and more complex to parse,
but allows easy backward file navigation (the Block Total Length is
replicated at the end of the block exactly for this purpose),
which can be a
useful feature.

c. to use the length of the SHB, maintaning the 32 bit "Block
Total Length".
This solution allows backward navigation, but requires to create a new SHB
every 4 gigabites.

What do you vote?

Personally I don't like to transform the Section Header Block from a MARKER
to a CONTAINER.
I don't like to rewind the file in case of large capture in order to update
such a value.
And what about if the application crashes before updating that value? The
format of the file is wrong, because the section length is set to a wrong
value.

Personally, I would like to keep the SHB a marker, and add and option that
says "the size of this section is XXX", where XXX is a 64 bit number.

        fulvio

-
This is the tcpdump-workers list.
Visit https://lists.sandelman.ca/ to unsubscribe.


Current thread: