Subject: Re: -current kernel build fails (unofficial patches)[SOLVED]
To: Michael <macallan18@earthlink.net>
From: dixie <dixieml@imap.cc>
List: port-macppc
Date: 08/04/2005 15:24:49
Michael wrote:

>Hello,
>
>  
>
>>Uhm, something has changed :)
>>    
>>
>
>Yes ;)
>
>  
>
>>#define adt7467c {}
>>#device adt7467c: sysmon_envsys
>>#attach adt7467c at ki2c
>>#file arch/macppc/dev/adt7467.c                 adt7467c
>>
>>config still failed:
>>
>>HORUS:529: `adt7467c' cannot attach to `ki2c'
>>*** Stop.
>>    
>>
>
>Of course - if you remove the information how to attach adt7467c to ki2c
>config can't attach adt7467c to ki2c ;)
>
>  
>
>>referring to that line:
>>
>>adt7467c*       at ki2c?                # Analog Devices ADT7467
>>thermal monitor, fan controller and so on
>>
>>How can I make config rightly aware of the new files and devices?
>>Are the old .c files and the patch still of any use?
>>Thank you for your support :)
>>    
>>
>
>My bad, I forgot to include an updated files.macppc. Change the
>statements about adt7467c to this:
>
>attach adt7467c at ki2c with adt7467_ki2c
>file arch/macppc/dev/adt7467_ki2c.c             adt7467c
>
>or use the attached files.macppc.
>
>The reason behind this is that I made the drivers generic i2c devices (
>so they live in sys/dev/i2c now ) and what's left in macppc/dev is just
>some glue to attach directly to ki2c so you won't have to state an
>i2c-address in your kernel config.
>
>I guess I'll make a real patch with all my local changes now...
>
>have fun
>Michael
>  
>
Finally it builds and it boots too :)

...
uni_n0 at mainbus0 address 0xf8000000
hw-clock at uni_n0 not configured
ki2c0 at uni_n0 address 0xf8001000
iic0 at ki2c0: I2C bus
adt7467c0 at ki2c0 ADT7467 thermal monitor and fan controller
...
deq0 at ki2c1 Apple Digital Equalizer, addr 6a
...

In the last archive you uploaded dev/i2c/files.i2c is missing and 
macppc.patch adds this line to arch/macppc/conf/files.macppc:

file arch/macppc/dev/ofb_cons.c

which probably is useful to you but in a "clean" environment causes 
'make depend' to fail :)

--
Reading dmesg I noticed that line which I cannot remember:

uninorth2 at mainbus0

is that due to the presence of uni_n0?
--

Anyway many thanks again for that great set of drivers :)

regards

d.