Biswas, Anumita wrote:
>>> If keeping the standard TCP CRC intact (and covering TCP header &
>>> data) and having an RFC1146 alternate checksum that works
>>> than defined therein, perhaps it would be better to make it
>> a new TCP
>>> option then.
>> Defined therein where?
> Section 3 - "Negotiating the use of CRC 32C"
There are two differences:
a) different processing rules
computed over different portions of the segment
different rules for what happens to the original TCP checksum area
different rules for what to do when the option fails on SYNs
These are hard to fix without rewriting RFC-1146, which is feasible only if
there is a real need.
b) different algorithm
this is easily fixed without modifying RFC-1146, by
defining a new algorithm
this assumes only the algorithm changes, though
tcpm mailing list