Re: [lwip-users] Improving LWIP performance

[Kieran Mansley]

On Mon, 2010-08-09 at 19:17 -0400, Karthik Vadambacheri Manian wrote:
> Can you please give your insights on where it is going wrong?

So to summarise: you can send one packet that results in failure because
the checksum of received data differs from the checksum of sent data,
but you can send a second packet that has a couple of bytes swapped (and
so has the same checksum), but this is received OK.

1) Can you reproduce this behaviour reliably?  Is the first packet
always rejected.  Is the second always accepted OK?

2) You need to work out if the problem is at the sender or receiver.
Can you get a packet capture of the problem packet?  If it has the wrong
checksum in the packet capture that suggests the sender is at fault (or
something between the sender and the capture has corrupted the packet).
If it has the right checksum in the capture that suggests the receiver
is at fault (or something between the capture and receiver has corrupted
the packet).

3) Is the data that is checksummed at the receiver (including headers
and payload) exactly the same as sent at the receiver, or has some error
been introduced.  i.e. is it a checksum problem or is the failing
checksum correct because the data have been corrupted.

Kieran