Re: [Qemu-devel] [Qemu-ppc] [PATCH 2/2] [v3] target-ppc: Enhance CPU nodes of device tree to be PAPR compliant.

[Andreas Färber]

Am 14.08.2013 12:18, schrieb Alexander Graf:
> 
> On 12.08.2013, at 23:22, Benjamin Herrenschmidt wrote:
> 
>> On Mon, 2013-08-12 at 21:17 +0200, Thomas Huth wrote:
>>> Am Mon, 12 Aug 2013 16:03:24 +1000
>>> schrieb Benjamin Herrenschmidt <address@hidden>:
>>>
>>>> On Mon, 2013-08-12 at 10:07 +0530, Prerna Saxena wrote:
>>>>
>>>> .../...
>>>>
>>>>> I dont know what context lead to this observation.
>>>>> However, PAPR mentions the following nomenclature guideline:
>>>>>
>>>>> "The value of this property shall be of the form: “PowerPC,<name>”,
>>>>> where <name> is the name of the processor chip which may be displayed to
>>>>> the user. <name> shall not contain underscores."
>>>>
>>>> This actually comes from the original Open Firmware binding for PowerPC
>>>> processors, which PAPR inherits largely from. Thus this naming scheme
>>>> should apply to all PowerPC processors when a device-tree is involved.
>>>
>>> Well, I think it should be used when an Open Firmware environment is
>>> used. When you boot via ePAPR device tree, the name should be "cpu"
>>> instead, according to the ePAPR specification.
>>
>> Yeah well ... this is a gratuituous change in ePAPR, I don't think it
>> matters really what the name is anyway. I'd suggest sticking to the
>> original OF binding.
> 
> Can't we just include the PowerPC, bit as part of the fw_name field in the 
> class? I don't think we have any CPUs that can be used both in ePAPR and 
> sPAPR environments. So the POWER7 fw_name field would just contain 
> "PowerPC,POWER7" and the device tree creation code merely appends the @%d 
> piece.

That's exactly what my patch series does, just be patient. :)

Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg