> > IWM
> 
> Good luck!

Well, I wish the NetBSD device driver framework was only _half_ as easy
to see through as the IWM code. And that's disassembler dump instead of
source code. Many unix/k&r c hackers seem not to have realized that
comments are *not* turned into object code no matter how many you add.
So they don't use any. :-> 

 
> > For the IWM, I always receive an empty string. The IWM attaches to obio
> > (on-board I/O) - does obio hold back the name string or not generate it?
> 
> It doesn't generate it.  It should, I think, call each match function
> once for each configured device.

My iwm match function _is_ called - that's not the problem.
What puzzles me is that other ports' floppy drivers are actually given a
string (referenced by aux) to compare their name with. And this string
is "" for a device attached to mac68k obio. And obio_attach() itself
gets aux from somewhere else - out of the blue? That's why I asked for a
description of the flow of control. You cant' just set a breakpoint with
gdb and step through, can you? ;-) 
Sure, I can just ignore this and be happy with looking at IWMBase, but
I'd rather know what goes on here...

> It's a pain, I know...  I think that there is a paper in the tree
> somewhere under /usr/share/doc, but I've never tried to look it up.

Does any NetBSD person actually update the information there? Never seen
any changes... 
I've got the O'Reilly prints of the 4.4BSD manuals here and they are
usually pretty helpful, but the SMM chapters on kernel configuration are

a) rather high level,
b) a strange mix between the old MD config scheme used for the HP port
and the new MI one used for the sparc port, and
c) when it comes to examples, they fall back to the VAX mba/uba crap
which is totally outdated. AFAIK there isn't even a VAX port of 4.4BSD,
so it would seem people copied over stuff from the 4.3BSD manuals.

Same with the Egan/Teixira book (19989/92) -- mostly SysV and the BSD
parts are 4.3/VAX. :-(


> Cheers,
> -allen

        hauke

-- 
"It's never straight up and down"  (DEVO)