Re: [PATCH v3 06/10] pwm: Add NVIDIA Tegra SoC support

[Ryan Mallon <rmallon-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>]

On 23/02/12 19:14, Thierry Reding wrote:

> * Ryan Mallon wrote:
>> On 23/02/12 02:17, Thierry Reding wrote:
>>
>>> This commit adds a generic PWM framework driver for the PWFM controller
>>> found on NVIDIA Tegra SoCs. The driver is based on code from the
>>> Chromium kernel tree and was originally written by Gary King (NVIDIA)
>>> and later modified by Simon Que (Chromium).
>>>
>>> Signed-off-by: Thierry Reding <thierry.reding-RM9K5IK7kjKj5M59NBduVrNAH6kLmebB@public.gmane.org>
>>> ---
>>
>> <snip>
>>
>>> +
>>> +	pwm->chip.dev = &pdev->dev;
>>> +	pwm->chip.ops = &tegra_pwm_ops;
>>> +	pwm->chip.base = -1;
>>> +	pwm->chip.npwm = NUM_PWM;
>>
>>> +
>>> +	ret = pwmchip_add(&pwm->chip);
>>
>>
>> If a driver fails to initialise the pwm_chip structure correctly it can
>> cause problems in the pwm core. For example, if the dev field doesn't
>> get set, then you will get an oops if you try to cat the pwm debugfs file.
>>
>> pwmchip_add should probably verify that the initialisation of the
>> pwm_chip structure is sane to avoid problems like this.
> 
> Absolutely. What would be the best response to an invalid struct pwm_chip? I
> suppose at least returning -EINVAL, perhaps complemented with WARN_ON?


Just returning -EINVAL should be okay. I don't think you need a WARN_ON,
since failing to register the hardware should be enough of a reason for
a user to report a problem.

~Ryan