Subject: Re: mcd and wt, was DUMMY_NOPS
To: None <port-i386@NetBSD.ORG>
From: Bill Studenmund <skippy@macro.stanford.edu>
List: port-i386
Date: 10/21/1997 21:50:22
On Tue, 21 Oct 1997, Bill Studenmund wrote:

> On Tue, 21 Oct 1997, David Maxwell wrote:
> 
> > [split up the kernels]    And make a
> > separate BASIC/INSTALL kernel to distribute with a set of options which
> > don't cause problems for older systems. (mcd and wd come to mind?)
> 
> I thought the problem w/ mcd was that it would trash a network card if it
> tried to probe. I've just had the experience that the lc0 driver probe
> will wedge my ne2000 card. I've had the same experience with the mcd
> driver. So the mcd problem's more one of equipment than one of MB age (I
> believe if I put my ethernet card in a 386 and booted a kernel w/ mcd
> turned on, it'd wedge). 
> 
> One option, though, would be to re-arrange the tests (change order in
> /sys/conf/files) so that the mcd doesn't get probed until after any ed2 or
> ne1's get probed. That's what's fixed things for me and the lc0 driver. It
> might also fix things for wt0 and mcd0. I'll check.

Is there any way to get config to order device tests other than
re-aranging the config files? Because if the ne2000, and probably the ed
driver, could get included (and thus probed) before the mcd and wt drivers
do, now that we have the bus-space allocation, we could probably turn them
back on in the GENERIC configuration.

The problem is that the ne2000 and ed drivers are defined in
/sys/conf/files which is read after /sys/arch/i386/conf/files.i386 (which
includes /sys/dev/isa/files.isa which has these two drivers in it).

Take care,

Bill