Re: [PATCH] of: Export __of_find_all_nodes()

[Frank Rowand <frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>]

On 01/19/17 01:43, Alexander Sverdlin wrote:
> Hi!
> 
> On 18/01/17 22:33, Frank Rowand wrote:
>>> This is necessary for for_each_of_allnodes() to work in the modules.
>>>
>>> Signed-off-by: Alexander Sverdlin <alexander.sverdlin-xNZwKgViW5gAvxtiuMwx3w@public.gmane.org>
>>> Cc: Rob Herring <robh+dt-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
>>> Cc: Frank Rowand <frowand.list-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>>> Cc: devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
>>> ---
>>>  drivers/of/base.c | 1 +
>>>  1 file changed, 1 insertion(+)
>>>
>>> diff --git a/drivers/of/base.c b/drivers/of/base.c
>>> index 2eb4dea62b84..f576c33e0b84 100644
>>> --- a/drivers/of/base.c
>>> +++ b/drivers/of/base.c
>>> @@ -270,6 +270,7 @@ struct device_node *__of_find_all_nodes(struct device_node *prev)
>>>  	}
>>>  	return np;
>>>  }
>>> +EXPORT_SYMBOL_GPL(__of_find_all_nodes);
>>>  
>>>  /**
>>>   * of_find_all_nodes - Get next node in global list
>>>
>> Quoting Rob from the last request for this:
>>
>> "What driver needs this? This isn't really a function I'd expect drivers to use."
> 
> This one and the one from previous patch/mail are used in a module which reads DT overlays

If I understand correctly, that was a request to be able to use of_node_init() directly.
Is that correct?

> from pluggable extension boards, eventually performs some sanity checks/corrections and
> applies these overlays. This eventual corrections require low-level access to DT.

Are you saying that you want to modify the device tree data structure after the overlay
is applied?

The idea of sanity checks is one that I am very interested in.  Could the sanity checks
be done statically, either with the overlay source, or with the compiled overlay source?
The idea behind that question is whether the checks could be done while the overlay is
being developed, instead of at run time.

I am also curious about what sort of corrections are made and how it is determined what
corrections to make.


> I cannot judge, how relevant this usecase is for the community, but I was driven by the fact,
> that the functions which require the functions I export are defined statically in .h files
> and therefore are exposed to the modules.
> 

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html