on 2/12/03 9:18 AM, Matt Thomas at matt@3am-software.com wrote:

> At 02:13 AM 2/12/2003, Monroe Williams wrote:
> 
>> I think I know why my kernel can't find the root device, but I'm not sure
>> how to fix it.
>> 
>> The OF tree for Sonnet ATA-100 and ATA-133 cards (which use a Promise
>> chipset and UltraTek firmware) looks approximately like this:
>> 
>> /UltraTek100+
>>   /FrmTk-0@10,0
>>     /sd@0
>>   /FrmTk-0@11,0
>>     /sd@0
>> 
>> The top level ("UltraTek100+") has a "device_type" property of "scsi", while
>> the second level ("FrmTk-0") has a "device_type" property of "ata".
>> 
>> I suspect that the code in device_register() that matches up OF device tree
>> nodes with the kernel's idea of devices can't cope with an extra layer in
>> the OF tree between the pciide controller and the drive(s).
>> 
>> Does this sound plausible?
>> 
>> If so, how might one work around this problem without committing an
>> egregious hack?
> 
> I thought of fixing device_register but found it was easier to force my
> config file to have root on wd0 :)

While that works for you and it works for me, wouldn't it be much better if
it worked for everyone?  ;)

-- monroe
------------------------------------------------------------------------
Monroe Williams                                         monroe@pobox.com