[lwip-users] Re: [lwip] incoming MSS > lwip MSS

[Kieran Mansley]

On 19 Sep 2002, Dirk Koopman wrote:

> Has anyone any experience of handling incoming packets greater than the
> defined MSS size, thus leading to the lower level creating a chain of
> (rather than the default of exactly one) pbuf(s) for each incoming
> packet?
>
> At first sight it doen't appear to work, but it may (possibly :-) be my
> low-level driver code.

Sounds like you're confusing two things here - the MSS and the pbuf size.
Having the low-level netif layer split an incoming segment over multiple
pbufs (when PBUF_POOL_BUFSIZE > TCP_MSS) should work.  I do it anyway (for
different reasons - header in one pbuf, payload in the second), but can't
remember if I had to fiddle around to get it going.  There's nothing wrong
with the concept anyway.

Getting the netif layer to split an incoming segment that is greater than
the MSS though is harder.  You'd have to effectively make it two packets I
think, each with it's own header, which even if possible would be
distinctly dodgy.  Someone sending you a packet that is larger than the
MSS you've advertised is a Bad Thing, so I don't think this should work.

Kieran

[This message was sent through the lwip discussion list.]