On Thu, Oct 9, 2014 at 1:46 AM, RR <rajaram.officemails@xxxxxxxxx> wrote:
> On Wed, Oct 8, 2014 at 12:18 AM, Alexandre Courbot <gnurou@xxxxxxxxx> wrote:
>> On Wed, Oct 8, 2014 at 4:09 PM, Muthu Mani <muth@xxxxxxxxxxx> wrote:
>>>> > +static int cy_gpio_direction_output(struct gpio_chip *chip,
>>>> > + unsigned offset, int value) {
>>>> > + return 0;
>>>> > +}
>>>>
>>>> If that chip is capable of both output and input, shouldn't these functions be
>>>> implemented? I think this has already been pointed out in a previous version
>>>> but you did not reply.
>>>
>>> Thanks for your inputs.
>>>
>>> Only the GPIOs which are configured to be output GPIO can be set.
>>
>> In that case cy_gpio_set() should return an error for GPIOs which are
>> not configured as outputs. Is that guaranteed by the current
>> implementation?
>>
>>> The set operation would fail trying to set the input or unconfigured GPIOs.
>>> In this version of driver, this support is not added, it can be introduced in future versions.
>>> I will add a TODO note in the code.
>>
>> Argh, no TODO please. Actual code that will turn this code into a
>> solid driver that can be merged.
>
> Does a driver targeted for a custom device has to implement every
> functionality in the 1st version ?
When you post a driver to the GPIO maintainers it is *NOT* tageted
at a consumer device, it is targeted at the kernel community and
upstream maintainers.
Of course you can deliver add-on patches out-of-tree to your
customers, it's generally a bad idea for the long term and maintenance
of your driver, but it's your pick.
Yours,
Linus Walleij
--
To unsubscribe from this list: send the line "unsubscribe linux-i2c" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
