[PATCH v4 02/10] property: Add device_get_child_node_count_named()

Matti Vaittinen mazziesaccount at gmail.com
Tue Feb 25 02:56:56 PST 2025


On 25/02/2025 12:21, Andy Shevchenko wrote:
> On Tue, Feb 25, 2025 at 11:40:16AM +0200, Heikki Krogerus wrote:
>>> +/**
>>> + * device_get_child_node_count_named - number of child nodes with given name
>>> + *
>>> + * Scan device's child nodes and find all the nodes with a specific name and
>>> + * return the number of found nodes. Potential '@number' -ending for scanned
>>> + * names is ignored. Eg,
>>> + * device_get_child_node_count(dev, "channel");
>>> + * would match all the nodes:
>>> + * channel { }, channel at 0 {}, channel at 0xabba {}...
>>> + *
>>> + * @dev: Device to count the child nodes for
> 
> This has an inconsistent kernel doc structure in comparison to the rest in this
> file.

Hmm. I'll take a look at the differences for v5.

>>> + * Return: the number of child nodes with a matching name for a given device.
>>> + */
>>> +unsigned int device_get_child_node_count_named(const struct device *dev,
>>> +					       const char *name)
>>> +{
>>> +	struct fwnode_handle *child;
>>> +	unsigned int count = 0;
>>> +
>>> +	device_for_each_child_node(dev, child)
>>> +		if (fwnode_name_eq(child, "channel"))
>>
>> s/"channel"/name/ ?
>>
>>> +			count++;
>>> +
>>> +	return count;
>>> +}
>>> +EXPORT_SYMBOL_GPL(device_get_child_node_count_named);
>>
>> I did not check how many users are you proposing for this, but if
>> there's only one, then IMO this should not be a global function yet.
>> It just feels to special case to me. But let's see what the others
>> think.
> 
> The problem is that if somebody hides it, we might potentially see
> a duplication in the future. So I _slightly_ prefer to publish and
> then drop that after a few cycles if no users appear.
> 
> Also this misses the test cases.

I'll also take a look at the tests, but I have a bit of an attitude 
problem what comes to unit testing. Adding tests for the sake of having 
tests just hinders the development. It makes improving functions less 
appealing (as tests need to be changed as well) and adds bunch of 
inertia & maintenance cost. Sure, on complex functions having tests 
increases the confidence that changes work - but I don't see much value 
here.

Do we have tests for all the property.h functions?

Yours,
	-- Matti



More information about the linux-arm-kernel mailing list