> This re-assembly simplifies some code, but it not only has a bad impact
> on memory management, it also involves a "useless" copy operation.
That "useless" operation is somewhat temporary - its only useless because
we dont reassemble and checksum in one pass. Also ATM already and HIPPI
very soon will be giving us 65280 byte frames that are not fragmented.
And you will note that all these large MTU crazy bandwidth cards have
scatter gather chain capabilities (and in some cases checksumming as
well) so we will need to re-architect the skb mechanism to be able to
send chains or iovector type stuff to the driver.
This is in the works and will coincide with the networking page
flipping stuff Eric and myself will be coding...
---------------------------------------------////
Yow! 11.26 MB/s remote host TCP bandwidth & ////
199 usec remote TCP latency over 100Mb/s ////
ethernet. Beat that! ////
-----------------------------------------////__________ o
David S. Miller, davem@caip.rutgers.edu /_____________/ / // /_/ ><