Hi Morimoto-san,
On Fri, Sep 14, 2018 at 2:53 AM Kuninori Morimoto
<kuninori.morimoto.gx@xxxxxxxxxxx> wrote:
> > > Before 6c92d5a2744e2761 patch, driver will forcibly ignore
> > > non-SSI modules, and try to use PIO mode.
> > > I'm not sure it is "kindly support" or "overkill support".
> > >
> > > After this patch, it needs DMA, otherwise, probe will be failed.
> > > DT shouldn't have non-SSI modules if you want to use PIO mode.
> > >
> > > + /* use PIO mode */
> > > - playback = <&ssi0 &src0 &dvc0>;
> > > + playback = <&ssi0>;
> >
> > OK, so falling back to PIO was really a "best effort" fallback, and the user
> > should really enable DMA?
>
> Yeah, I'm thinking this is better.
> PIO fallback is of course "nice to have" if possible.
> But, because of this new iommu patch, keeping this feature
> needs "big complicated patch", and we can get "small effect" I think.
> Thus, I think this is the time to remove this feature.
> Can you agree ?
You're the rcar-sound expert ;-)
If you think there's not much value in keeping PIO fallback, it can be
removed.
We can change behavior once these dependency issues have been resolved
in a generic way.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx
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
