[PATCH 3/3] ASoC: rsnd: add null CLOCKIN support

[Geert Uytterhoeven]

Hi Morimoto-san,

On Thu, May 27, 2021 at 12:06 AM Kuninori Morimoto
<kuninori.morimoto.gx at renesas.com> wrote:
> > Oh right, I missed the static clk_hw pointer.
> > What if you unload the snd-soc-rcar.ko module?
>
> Hmm.. indeed.
> It needs something..
> Thank you for poining it.
>
> >     #define for_each_rsnd_clk(pos, adg, i)          \
> >             for (i = 0; (pos) = adg->clk[i], i < CLKMAX; i++)            \
> >                     if (pos) {                      \
> >                             continue;               \
> >                     } else
>
> Wow!! I didn't know this technique.
> Indeed it can use NULL pointer.
>
> But, I want to avoid "if (pos) else" code as much as possible,
> and keep simple code.
> It can handle all clk case without thinking it if it has null_clk.
>
> Why you don't want null_clk ??

It adds a dummy object, which needs to be cleaned up.  Basically you
are trading a simple NULL pointer check for a zero clock rate check
deeper inside the driver, with the additional burden of needing to
take care of the dummy clock's life cycle.

Note that most clk_*() calls happily operate on a NULL pointer, and
just return success.  This includes clk_get_rate(), which returns
a zero rate.

Mark might have a different view, though, due to his experience with
dummy regulators?

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds