Re: [PATCH v2] net: dm9000: Allow instantiation using device tree

[Sylwester Nawrocki <sylvester.nawrocki@gmail.com>]

On 05/19/2013 03:41 PM, Tomasz Figa wrote:
>>> +static struct dm9000_plat_data *dm9000_parse_dt(struct device *dev)
>>> >  >  +{
>>> >  >  +	struct dm9000_plat_data *pdata;
>>> >  >  +	struct device_node *np = dev->of_node;
>>> >  >  +	const void *mac_addr;
>>> >  >  +
>>> >  >  +	if (!IS_ENABLED(CONFIG_OF) || !np)
>>> >  >  +		return NULL;
>> >
>> >  Shouldn't ERR_PTR() value be returned here ?
>
> Nope. No platform data is a valid case, so no error here.

OK, sorry, I should have checked that.

>>> >  >  +	pdata = devm_kzalloc(dev, sizeof(*pdata), GFP_KERNEL);
>>> >  >  +	if (!pdata) {
>>> >  >  +		dev_err(dev, "failed to allocate platform data struct\n");
>> >
>> >  There is no need for this error log, k*alloc already logs any failures.
>
> Hmm. Does it print what allocation exactly failed? (e.g. a backtrace)

It seems the caller tracking is supported in subset of the kernel 
configurations.

> Not that it would give anything that could help you in an out of memory
> condition like this, but in general it's good to know in what point the
> failure happened.

As long as it is ENOMEM and the driver core reports failed probe I wouldn't
really care. Nevertheless I saw patch series removing such already existing
"Not enough memory" kind of error logs from the kernel, pointing out that mm
already provides relevant error logging.

Thanks,
Sylwester