Re: Definition of "packet"

[Jacob Nevins <jacobn+secsh%chiark.greenend.org.uk@localhost>]

Henrick Hellstrom writes:
> This ought to be a simple question, but it seems it is not spelled out 
> anywhere in specification:
> 
> Exactly what is to be included in the calculation of the length of a 
> "packet", when checking if it exceeds the Maximum Packet Size parameter 
> sent during SSH Connection establishment?

I assume that you're referring to the parameter in the
SSH_MSG_CHANNEL_OPEN packet.

> a) Just the raw packet load
> b) Packet load + SSH specific formatting (e.g. 4 for the length 
> specifier of a "string")
> c) Packet load + formatting + channel packet header (e.g. 5 for byte 
> SSH_MSG_CHANNEL_DATA; uint32 recipient channel)
> d) Packet load + formatting + channel packet header + transport packet 
> formatting (as per SSH Transport Layer Protocol draft paragraph 6)
> e) Packet load + formatting + channel packet header + transport packet 
> formatting + TCP/IP headers
> ...etc

I agree this is not absolutely clear.

I think, given the context, that I'd expect it to refer to the maximum
amount of actual channel data in a packet (i.e., the maximum length of
the `data' field, not including the string length, in the
SSH_MSG_CHANNEL_DATA packet), and similar for related types of packet. I
think this is your option `a'.

PuTTY at least appears to agree with this.

If there's consensus on this, can we get a note into the RFC before it's
published?