Re: API and driver interface

From: Jens David (terhi.victor@logonet.com)
Date: Thu Feb 10 2000 - 21:36:17 EET

  • Next message: Terry Dawson: "Re: Status of kernel AX.25 implementation."

    > > And for Linux: Due to the lack of information about how FlexNet in the low
    > > level stage really works

    How the "low level stage" works comes pretty clear if you take a look at the
    device driver development kit. It fact it is not much different from what we
    are doing now. It is *very* important to implement it in that way to achieve
    a reasonable channel access behaviour. Another advantage flexnet has over us,
    is that it vitally has the CPU for its own. We are pretty close to that in
    kernel mode. However, we have at least two *really* big advantages over
    flexnet:
    a reasonable memory buffer concept and very powerful subsystems for card
    configuration (PCI) and low-level port access. We also have very nice debugging
    tools (kdb et al.).
    Writing a flexnet/16 device driver is a pain in the *ss.

      -- Jens



    This archive was generated by hypermail 2b29 : Fri Feb 11 2000 - 02:27:45 EET