I'd like to see us "fix" this problem.  But I don't have Sun3 hardware
to test with, and while I can find an emulator, I suspect its not quite
the same thing.

Is anyone still "working" on Sun3 stuff anymore?

    -- Garrett

Chuck Silvers wrote:
> On Tue, Jun 06, 2006 at 08:21:56AM +0900, Izumi Tsutsui wrote:
>   
>> garrett_damore@tadpole.com wrote:
>>
>>     
>>> I notice that we still have the __BROKEN_CONFIG_UNIT_USAGE logic in
>>> sys/kern/subr_autconf.c.
>>>
>>> The sun3 port declares this macro, but it *appears* that it may not
>>> actually require it.  It is not declared anywhere else.
>>>
>>> I would like to propose removing the macro (and the code it enables). 
>>> However, I lack sufficient experience to be 100% sure, so I am asking
>>> for knowledgable parties to step forward and either approve the change,
>>> or just make the change themselves -- or explain why this change is a
>>> bad idea and will cause untold grief amongst millions of sun3 users... :-)
>>>       
>> IIUC, the abuse is in sun3/sun3/vme.c. It uses cf_unit to specify
>> VME bus types. Of cource sun3 should switch to the MI vme driver,
>> but we have to implement bus_space(9) for sun3 first.
>> It should be trivial (see sun68k/sun68k/bus.c), but...
>>     
>
> that's one part, the other is in the xy/xd drivers where it requires
> that the disk unit number be a function of the controller unit number
> (and there's some stuff in sun3/autoconf.c that knows that hard-wired
> mapping as well).
>
> -Chuck
>   


-- 
Garrett D'Amore, Principal Software Engineer
Tadpole Computer / Computing Technologies Division,
General Dynamics C4 Systems
http://www.tadpolecomputer.com/
Phone: 951 325-2134  Fax: 951 325-2191