Re: change to of MTD partition parsing

[Kumar Gala <galak@kernel.crashing.org>]

Ben,

We need a better solution to the problem.  What does the device tree  
on SLOF look like?

- k

On Aug 12, 2009, at 10:07 AM, Benjamin Krill wrote:

> Hi Kumar,
>
> slof has a further node inside the flash node which is not a partition
> entry. The old code just used all children and since the further node
> isn't a partition not all needed properties are there.
>
> cheers
> ben
>
>
> * Kumar Gala | 2009-08-12 09:46:10 [-0500]:
>
>> Ben,
>>
>> The following commit breaks the previous definition of flash  
>> partitions according to
>> Documentation/powerpc/dts-bindings/mtd-physmap.txt.  Using the  
>> 'name' field is bad practice.  What was going
>> on w/the SLOF case?
>>
>> - k
>>
>>
>> commit 4b08e149c0e02e97ec49c2a31d14a0d3a02f8074
>> Author: Benjamin Krill <ben@codiert.org>
>> Date:   Fri Jan 23 17:18:05 2009 +0100
>>
>>   [MTD] ofpart: Check name property to determine partition nodes.
>>
>>   SLOF has a further node which could not be evaluated
>>   by the current routine. The current routine returns
>>   because the node hasn't the required reg property. As
>>   fix this patch adds a check to determine the partition
>>   child nodes. If the node is not a partition the number
>>   of total partitions will be decreased and loop continues
>>   with the next nodes.
>>
>>   Signed-off-by: Benjamin Krill <ben@codiert.org>
>>   Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
>>