RE: [Xen-devel] Question about TCP checksum offload in Xen

[James Harper <james.harper@xxxxxxxxxxxxxxxx>]

> With xen checksum offloading, domain to domain traffic appears to be
> received with a bad checksum.  This is OK, there is no point in
> calculating a checksum if the packets are only going through memory.  If
> your memory is going to randomly corrupt packets you have more bigger
> problems to worry about.   However, this does upset at least Solaris: if
> you're using a Solaris guest for NAT then the NAT module on Solaris gets
> all upset if the checksum is wrong and drops the packets.  (This is
> Solaris's NAT module being overly picky, it may need to recalculate or
> at least invalidate the existing checksum, but it doesn't need to check
> it as well.)

Windows 2003 can't handle receiving packets that have a bad checksum either. 
Despite the fact that the 'checksum good' flag is set, Windows seems to go 
ahead and check it anyway. I wonder if it's the firewall doing that in Windows 
too... I always just assumed it was NDIS itself.

Windows also seems to differ from Linux in its opinion on whether the total 
packet length in a gso packet is included in the psuedoheader checksum. That 
took a bit of mucking around to find as some hardware seems to ignore the 
psuedoheader checksum and calculate the whole checksum again anyway, so you 
only see the problem on some cards.

James