[PATCH] [POWERPC] 85xx: Add basic Uniprocessor MPC8572 DS port

Kumar Gala galak at kernel.crashing.org
Wed Sep 12 13:35:01 EST 2007


On Sep 11, 2007, at 10:00 PM, David Gibson wrote:

> On Tue, Sep 11, 2007 at 12:59:22PM -0500, Kumar Gala wrote:
>>
>> On Sep 11, 2007, at 12:48 PM, Scott Wood wrote:
>>
>>> Olof Johansson wrote:
>>>> On Tue, Sep 11, 2007 at 12:21:30PM -0500, Scott Wood wrote:
>>>>> Olof Johansson wrote:
>>>>>> On Tue, Sep 11, 2007 at 11:00:28AM -0500, Kumar Gala wrote:
>>>>>>> well the ifdefs are orthogonal.  We don't have a way of knowing
>>>>>>> primary from the device tree today.
>>>>>> How about something like "fsl,primary-phb" in the bus device
>>>>>> node? I don't
>>>>>> know, maybe it's already been discussed and turned down for some
>>>>>> reason.
>>>>> It's more of a Linux issue than anything to do with the hardware.
>>>>
>>>> That doesn't stop firmware from telling linux which bus is the
>>>> primary
>>>> one on the system to help out.
>>>
>>> The entire notion of a "primary" PCI bus is due to a Linux flaw.
>>>
>>> If we did put it in the device tree, it should be something like
>>> "linux,primary-phb".  But since Linux can tell from the node's
>>> children,
>>> there doesn't seem to be much point.
>>
>> Once someone rights code to do this I'm happy to change over.  I took
>> this model of explicitly knowing the primary PHB from the pmac code.
>
> In the meantime, couldn't the code still be merged, using an explicit
> test of the root node's 'compatible' or 'model' properties to decide
> on the right primary bus.

I will be, I'm not going to wait on having some device tree spec for  
this.  The board code can handle it until we come to some agreement  
on how to do this.  I'm in agreement with Scott in that code should  
be added to scan or allow explicit determination.  Adding a 'prop' to  
the device tree just for linux seems a bit silly.

- k



More information about the Linuxppc-dev mailing list