On Mon, Nov 7, 2011 at 11:02 AM, Stephen Warren <swar...@nvidia.com> wrote:
> Mike Frysinger wrote at Monday, November 07, 2011 10:40 AM:
>> * PGP Signed by an unknown key
>>
>> On Monday 07 November 2011 11:35:33 Stephen Warren wrote:
>> > Joe Hershberger wrote at Friday, November 04, 2011 8:25 PM:
>> > > -void gpio_free(int gp)
>> > > +int gpio_free(unsigned gpio)
>> > >  {
>> > > + return 0;
>> > >  }
>> >
>> > If you're doing a cleanup pass on this driver, you may as well make
>> > gpio_free() do something; it should probably clear gpio_names[gpio].name
>> > and perhaps set the pin back to SFIO - in other words, undo gpio_reqeust().
>>
>> i think the decision made in Linux was that freeing a GPIO should not cause 
>> it
>> to change tristate or anything.  all it should do is mark it as "available" 
>> so
>> something else can request it.
>> -mike
>
> OK, I'll buy that, but presumably gpio_names[gpio].name should still be 
> cleared
> to indicate the pin is free?

Besides, that would directly undo this patch:
http://patchwork.ozlabs.org/patch/119277/

I'll clear the string.

Thanks,
-Joe
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot

Reply via email to