Subject: Re: wt driver conflict with ????
To: Chris G. Demetriou <cgd@pa.dec.com>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: port-i386
Date: 06/22/1998 13:47:02
Manuel Bouyer <bouyer@antioche.lip6.fr> writes:
> Since the ed driver split, this comment is not accurate. [snip]

Chris G. Demetriou <cgd@pa.dec.com> replies:
>i seem to recall that the 'wt' and 'el' drivers combined to hose any
>'we' card at their address.

Uh, Is this to update the comment, or to re-enable the driver?

All those drivers use port accounting, so we could turn it on in
GENERIC as long as we make sure that 'wt' proves after 'we'-- same
trick used for ISA bha vs aha probes.

Since all the drivers honour bus_space_map() of their port ranges, the
only way this could lose is booting a kernel configured *with* 'wt'
and 'el' at the address of a WD/SMC card, but *without* 'we' at that
address. Which can't happen for a GENERIC kernel.  Right?  

OTOH I dont see a compelling case for 'wt' for generic kernels.  We
dont build tape distributions, do we?

mcd might be useful for installs; my lab still has a couple of PCs
with mcd CDrom drives.  How about re-enabling mcd -- with the
promiscuous default: case in mcd_find() disabled by an option.  Does
anyone recall just what the mcd probe curdled? (Perry?)