Re: [lwip-users] buffering outgoing packets

[address@hidden]

Kieran Mansley wrote:
> > transmit is only an issue if there is IP fragmentation.  Is that
> > correct?  If so, then is it true that I would not have problems if I
> > have a known MTU and never send a "PDU" larger than the MTU?  Or in
> > other words, if I never send a UDP datagram that doesn't fit in the
> > MTU, and ensure that the TCP maximum segment size fits within the MTU?
> >      
> MTU is a property of the local network.  IP fragmentation could happen
> at any point in the path between the communicating devices, so I don't
> think that would help.
>
>    
I think it would: as we don't do path MTU discovery in lwIP we cannot 
find out whith which MTU to send and thus are always sending with our 
configured MTU. Therefore, fragmentation won't take place as long as you 
don't try to send a UDP or ICMP packet bigger than fits the MTU - TCP 
doesn't create such packets anyway.
> > In fact, it sounds like maybe I should just disable support for IP
> > fragmentation and reassembly altogether.
> >      
> I don't fully understand the problem, but Simon mentions a workaround in
> the bug tracker.  Disabling IP fragmentation sounds like it would work
> but perhaps not necessary to go that far.
>    
As I stated above, I think there's no need for IP_FRAG when you know 
your application that well - it's what I did, too, although rather to 
save code size...

However, I wouldn't disable IP_REASSEMBLY as you never know which host 
sends fragmented packets for you, so while you know you never will 
create such packets, you might still have to receive them and be able to 
reassemble them.

Simon