Re: PCI DMA into USER space

From: Jes Sorensen (jes@linuxcare.com)
Date: Thu Jun 15 2000 - 15:09:33 EST


>>>>> "David" == David Hinds <dhinds@valinux.com> writes:

David> No. Read what I said and think about it. My argument is that
David> if you compare a network card that has PCI shared memory, with
David> a network card that does bus mastering to buffers in host RAM,
David> you'll see that exactly the same number of system bus and PCI
David> bus transactions are required to send or receive a packet. In
David> one case the CPU is initiator; in the other the card is
David> initiator. So there is a difference in CPU utilization. There
David> is not a difference in bus utilization. Bus mastering is
David> "faster" because it offloads work from the CPU. PCI bus
David> mastering transactions are not more efficient at the bus level
David> than shared memory transactions: the PCI bus does not
David> distinguish between them, they are all just PCI reads or
David> writes.

They are often very different because you get the streaming effect
when doing bus mastering DMA which you do not get when doing
individual read/writes. Yes there is a big difference.

Jes

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Thu Jun 15 2000 - 21:00:36 EST