Re: [lwip-users] Pbuf pool size and network interface

[Enrico Murador - Research & Development - CET]

Dear Kieran,

> Is such a small MSS valid?  All IP networks are supposed to guarantee
> that they can receive a segment of 576 bytes without fragmenting, so
> assuming the default header sizes that would make a TCP segment of 536
> bytes.  While setting it smaller might be OK in most situations, you
> might find that this causes interoperability problems with other stacks,
> as they could ignore your MSS request and just use 536 instead.
thank you for the tip, I was not aware of this kind of situation.

By "googling out" a while, I found this discussion:

http://www.ietf.org/mail-archive/web/tsvwg/current/msg04483.html

In particular, this explanation seems useful:

> For IPv4, RFC 791 defines the minimum IP packet size as 68 bytes. On 
> page 25:
>
> Every internet module must be able to forward a datagram of 68
> octets without further fragmentation. This is because an internet
> header may be up to 60 octets, and the minimum fragment is 8 octets.
>
> Every internet destination must be able to receive a datagram of 576
> octets either in one piece or in fragments to be reassembled.

From this description, it seems to me that a MSS equal or more than 68 
bytes "should" be OK, as long as there are sufficient resources to 
reassemble a 576 bytes long (minimum) fragmented packet...

Enrico