Re: [2.5.58][PnP] Some small points.

[Adam Belay <ambx1@neo.rr.com>]

Hi Ruslan,

Thanks for pointing these out. I appreciate your help.


On Wed, Jan 15, 2003 at 06:16:11PM +0300, Ruslan U. Zakirov wrote:
> Hello All.
> 1) __pnp_remove_device(dev) call twice. First time in
> pnpc_remove_device(dev). I think it's wrong.
>

Agreed, I'll apply your patch here.

> _______________________________
> 2) We've forgot to unreg driver on module unload in opl3sa2 driver.
> --- sound/isa/opl3sa2.c~        2003-01-15 17:45:25.000000000 +0300
> +++ sound/isa/opl3sa2.c 2003-01-15 17:47:00.000000000 +0300
> @@ -881,6 +881,9 @@
>
>         for (idx = 0; idx < SNDRV_CARDS; idx++)
>                 snd_card_free(snd_opl3sa2_cards[idx]);
> +#ifdef CONFIG_PNP
> +       pnpc_unregister_driver(&opl3sa2_pnpc_driver);
> +#endif
>  }
>

Yep, it should unregister it there.

>  module_init(alsa_card_opl3sa2_init)
> _________________________________
> 3) Queston: Why we do free in this way?
> static int snd_opl3sa2_free(opl3sa2_t *chip)
> {
> #ifdef CONFIG_PNP
>         chip->dev = NULL;   -> Here. Why NULL? Who realy free resources and how?

The structure that chip->dev points to is freed by the pnp layer on the release call.
Is this what you were asking or are you talking about something else?

> #endif
> #ifdef CONFIG_PM
>         if (chip->pm_dev)
>                 pm_unregister(chip->pm_dev);
> #endif
>         if (chip->irq >= 0)
>                 free_irq(chip->irq, (void *)chip);
>         if (chip->res_port) {
>                 release_resource(chip->res_port);
>                 kfree_nocheck(chip->res_port);
>         }
>         snd_magic_kfree(chip);
>         return 0;
> }



> _______________________________________
> 4) Have we got ALSA driver that work absolutly and use PnP layer in
> right ways?

I believe Jaroslav is working on that now.  Jaroslav, if you would like me to 
convert ALSA drivers please let me know.

Regards,
Adam