Re: [PATCH v2 5/6] xen/arm: Dissociate logical and hardware CPU ID

[Julien Grall <julien.grall@linaro.org>]

On 09/17/2013 04:08 PM, Ian Campbell wrote:
> On Tue, 2013-09-17 at 16:02 +0100, Julien Grall wrote:
>> On 09/17/2013 03:39 PM, Ian Campbell wrote:
>>> On Wed, 2013-09-11 at 12:59 +0100, Julien Grall wrote:
>>>
>>>> +    dt_for_each_child_node( cpus, cpu )
>>>> +    {
>>>> +        u32 hwid;
>>>> +
>>>> +        if ( !dt_device_type_is_equal(cpu, "cpu") )
>>>> +            continue;
>>>
>>> This could eventually use dt_find_node_by_type which I added in my start
>>> of day rework. I would assume your patch will go in first so I'll try
>>> and remember to do that when I rebase...
>>
>> cpu node must be under /cpus.
> 
> Must it? Documentation/devicetree/bindings/arm/cpus.txt doesn't mention
> that.

In Documentation/devicetree/booting-without-of.txt:
 b) The /cpus node

  This node is the parent of all individual CPU nodes. It doesn't
  have any specific requirements, though it's generally good practice
  to have at least:

               #address-cells = <00000001>
               #size-cells    = <00000000>

  This defines that the "address" for a CPU is a single cell, and has
  no meaningful size. This is not necessary but the kernel will assume
  that format when reading the "reg" properties of a CPU node, see
  below

> But if it were required then wouldn't it be invalid to have a node with
> type cpu outside that subtree? IOW looking up by type would still be OK.
> FYI this is what arm64 Linux does.

On arm32 Linux it's only looks in /cpus :).
I'm fine to replace this loop with dt_find_node_by_type.
Will you take care of this change, or do I need to add your patch on my
series and modify the code?

>>  dt_find_node_by_type will look at all the
>> nodes (not only the child) so we can't replace by this call.


-- 
Julien Grall