[Skiboot] [RFC PATCH 1/5] dt: add dt_new_check()

[Oliver O'Halloran]

On Fri, Sep 16, 2016 at 6:41 PM, Vasant Hegde
<hegdevasant at linux.vnet.ibm.com> wrote:
> On 09/16/2016 12:14 PM, Oliver O'Halloran wrote:
>>
>> This is similar to dt_new(), but if the node already exists it will
>> return the existing node. This is useful because some init code depends
>> on the presence of certain nodes, but where the node is actually created
>> is unimportant.
>>
>> Signed-off-by: Oliver O'Halloran <oohall at gmail.com>
>> ---
>>   core/device.c    | 14 ++++++++++++++
>>   include/device.h |  1 +
>>   2 files changed, 15 insertions(+)
>>
>> diff --git a/core/device.c b/core/device.c
>> index 9e7ef0d448d9..bd0674e31fda 100644
>> --- a/core/device.c
>> +++ b/core/device.c
>> @@ -352,6 +352,20 @@ struct dt_node *dt_find_by_name(struct dt_node *root,
>> const char *name)
>>         return NULL;
>>   }
>>
>> +
>> +struct dt_node *dt_new_check(struct dt_node *parent, const char *name)
>
>
> Patch itself looks good.
>
> Reviewed-by: Vasant Hegde <hegdevasant at linux.vnet.ibm.com>
>
> May be its better to rename function as  dt_new_get() ?

I'd like to avoid using *_get and *_put considering how they are used
in Linux. I know dt_new_check() isn't a great name, but it seemed like
the least bad option.

>
> Also it doesn't takes care of handling node creation with address etc..
> which is probably fine for now ..as we have issues with node name without
> address (like ibm,opal).

Add it when we need it IMO.

>
>
> -Vasant
>