On Mon, Mar 02, 2009 at 07:05:28PM -0600, Lopez Cruz, Misael wrote:
This is much better than last time, thanks! Some issues do remain,
though (as well as those that Takashi mentioned):
> +int snd_soc_jack_add_gpios(struct snd_soc_jack *jack, int count,
> + struct snd_soc_jack_gpio *gpios);
> +void snd_soc_jack_free_gpios(struct snd_soc_jack *jack);
snd_soc_jack_free_gpios() shouild take the same arguments as the add...
> struct snd_soc_jack {
> struct snd_jack *jack;
> struct snd_soc_card *card;
> struct list_head pins;
> + struct snd_soc_jack_gpio *gpios;
> + int gpio_count;
> int status;
> };
...so (as previously discussed) you don't have this. The jack shouldn't
need this information and it's only used to free things at the minute.
Consider how this will scale: if the jack detection method for every
codec were to add back pointers to the snd_soc_jack structure then you'd
rapidly find that the bulk of the structure is taken up with these back
pointers.
> + enable = gpio_get_value(gpio->gpio);
> + if (gpio->invert)
> + enable != enable;
I think you mean enable = !enable here (I'm surprised the compiler
doesn't warn about this).
_______________________________________________
Alsa-devel mailing list
Alsa-devel@xxxxxxxxxxxxxxxx
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
