> Let's be honest. We are SO far away from having a working stack that > while > these discussions are interesting I'm more inclined to say, "let's > start > getting something that works and worry about some of this later". Seems to be the right attitude to get this thing done :-) > Each cards gets an rx and tx queue, each managed by a thread. The > thread > basically just gets data from the card/stack and passes it onto the > stack/card, transferring from/to mbufs as required. From the exit/ > entry > point of that function up to the final passing back to the user is > mbuf's. The card manages its own queues and sends filled buffers in the rx queue and asks for new ones in the tx queue, the stack just waits for new packets - that would be the zero-copying driver model. In our current model, between the stack and the driver we could have those threads to fill the driver's rx queue, and send out buffers in the tx queue. Should work, yes. Adios... Axel.