In article <2930.825406405@NIAGARA.NECTAR.CS.CMU.EDU> Chris_G_Demetriou@NIAGARA.NECTAR.CS.CMU.EDU writes:
>> So, what is going to be necessary to make the BT742/747/757 use the
>> EISA stuff?  Make a BT7XX-specific driver, separate from the ISA one?
>
>I see it as being mostly a matter of splitting the driver up into
>front-ends which touch the busses, and a back end that touches the
>hardware.  This shouldn't be too hard.
>
>Some people (I honestly forget who, right now) are working on
>code/infrastructure to make this a bit easier...  they can/should
>speak up, and discuss their plans.

Charles, Jason and I were discussing the idea of changing config so that
it needs:
    struct cfdriver <device><bus>cd;
instead of:
    struct cfdriver <device>cd;

This will allow different probe/attach methods for different busses on
the same device names. I think Jason was working on it and was almost done...

>
>Worth noting that to do this right, you really need to create some way
>to deal with the aha/bt conflicts/interactions, perhaps with one
>'universal' probe routine, to check for any of the boards.

Same deal with the 3com isa ethernet cards; you probe once to find all
the cards.

>chris