Paul Jakma wrote: > On Wed, 17 Oct 2007, Paul Jakma wrote: > >> Firmware 'near' the NIC (I've never seen an explanation of the exact >> mechanism) interposes itself between hardware and OS and 'hijacks' >> traffic to that port. It never makes it to the OS. > > Ah, googling suggests that when this "Not quite a NIC" has this > management-port-hijack feature enabled, that it would limit the NIC to > 100Mb/s mode (due to firmware implementation not supporting GigE > somehow).
I think this is only true when the host operating system is not running. If the host OS is running, then gigabit should be fine. > > Is this performance limitation still present? See above. > > Another question I would have: If the management-hijack is active and > the OS requests promiscious receive, does the OS get the > management-port traffic? I.e. I suspect this feature *must* either > break AMT or else network-observability through promiscious-mode > operation, is this the case? If not, I'd be curious as to how it's done. I'm not sure (not on the project team), but as this was asked from Intel at their AMT presentation at the Intel Developer Forum, I believe the answer is that the host doesn't receive these packets, even in promiscuous mode. - -Garrett
