Re: [PATCH v2 3/5] cpufreq: sun50i: Add D1 support

[Andre Przywara]

On Mon, 18 Dec 2023 14:55:30 +0000
Conor Dooley <conor@xxxxxxxxxx> wrote:

Hi,

> On Mon, Dec 18, 2023 at 12:05:41PM +0100, Brandon Cheo Fusi wrote:
> > Add support for D1 based devices to the Allwinner H6 cpufreq
> > driver
> > 
> > Signed-off-by: Brandon Cheo Fusi <fusibrandon13@xxxxxxxxx>
> > ---
> >  drivers/cpufreq/sun50i-cpufreq-nvmem.c | 1 +
> >  1 file changed, 1 insertion(+)
> > 
> > diff --git a/drivers/cpufreq/sun50i-cpufreq-nvmem.c b/drivers/cpufreq/sun50i-cpufreq-nvmem.c
> > index 32a9c88f8..ccf83780f 100644
> > --- a/drivers/cpufreq/sun50i-cpufreq-nvmem.c
> > +++ b/drivers/cpufreq/sun50i-cpufreq-nvmem.c
> > @@ -160,6 +160,7 @@ static struct platform_driver sun50i_cpufreq_driver = {
> >  
> >  static const struct of_device_id sun50i_cpufreq_match_list[] = {
> >  	{ .compatible = "allwinner,sun50i-h6" },
> > +	{ .compatible = "allwinner,sun20i-d1" },  
> 
> I thought the feedback in v2 was to drop this change, since the
> devicetree has the sun50i-h6 as a fallback compatible?

Well, this is the *board* (fallback) compatible string, so we cannot assign
it as we like. The whole (existing) scheme is admittedly somewhat weird,
because we not only match on a particular device compatible
(like allwinner,sun20i-d1-operating-points), but also need to blocklist and
re-match some parts against the *board compatible*, owing to the
cpufreq-dt driver. The board name is basically used as a placeholder to
find out the SoC, because there is (or was?) no other good way - the
CPU DT nodes don't work for this. Back when this was introduced, this was
the "least worst" solution.

I don't remember all the details, and didn't find time yet to look into
this in more detail, but fixing this is non-trivial. If this isn't 6.8
material, I might have a look at this later this week/month.

Cheers,
Andre