Martin Mares wrote:
>
> Hi Jeff!
>
> > It would be nice to have devices compiled into the kernel probe in slot
> > order, ie.
> >
> > slot 1 - video card - fbdev/nothing
> > slot 2 - ethernet card - tulip
> > slot 3 - 2nd ethernet card - eepro100
> >
> > However this makes initialization hairy, because you cannot init the
> > hardware at init_module time. For new PCI drivers I think this is an
> > easy problem -- just delay when probe() is called -- but it seems like
> > the new PCI API will need a bit of reworking before slot order probing
> > is possible.
>
> By the way, what's the advantage in probing in slot order? I guess it
> should be enough to preserve slot order for devices belonging to single
> driver which is exactly what we currently do.
>
> (Assuming we're speaking of `natural' slot ordering (i.e., by increasing
> value of devfn, not according to slot numbers reported by the PCI BIOS).)
Probing in slot order has the advantage of being an ordering which is
known and obvious to the user. It is counter intuitive to have cards in
slots 1, 2, 3 get attached to drivers in the order 3, 1, 2, for example.
Further, several kernel subsystems have their own command line options
to allow the user to specify an arbitrary order. It might be useful to
replace or supplement that with slot-based ordering.
-- Jeff Garzik | Tact is the ability to tell a man Building 1024 | he has an open mind when he has a MandrakeSoft, Inc. | hole in his head. (-random fortune)- 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 Mar 23 2000 - 21:00:34 EST