Re: [PATCH 11/13] clk: renesas: rzg2l-cpg: drop determine_rate op and use CLK_ROUNDING_FW_MANAGED flag

[Geert Uytterhoeven]

Hi Biju,

On Fri, 27 Feb 2026 at 16:57, Biju Das <biju.das.jz@xxxxxxxxxxxxxx> wrote:
> > From: Brian Masney <bmasney@xxxxxxxxxx>
> > On Fri, Feb 27, 2026 at 03:23:19PM +0000, Biju Das wrote:
> > > > From: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>
> > > > On Fri, 27 Feb 2026 at 16:01, Brian Masney <bmasney@xxxxxxxxxx> wrote:
> > > > > On Fri, Feb 27, 2026 at 09:20:09AM +0100, Geert Uytterhoeven wrote:
> > > > > > On Thu, 26 Feb 2026 at 19:18, Brian Masney <bmasney@xxxxxxxxxx> wrote:
> > > > > > > This clk driver has a noop determine_rate clk op. Drop this
> > > > > > > empty function, and enable the CLK_ROUNDING_FW_MANAGED flag.
> > > > > > >
> > > > > > > Signed-off-by: Brian Masney <bmasney@xxxxxxxxxx>
> > > > > >
> > > > > > Thanks for your patch!
> > > > > >
> > > > > > > --- a/drivers/clk/renesas/rzg2l-cpg.c
> > > > > > > +++ b/drivers/clk/renesas/rzg2l-cpg.c
> > > >
> > > > > > > @@ -1041,7 +1034,7 @@ rzg2l_cpg_sipll5_register(const struct cpg_core_clk *core,
> > > > > > >         init.name = core->name;
> > > > > > >         parent_name = __clk_get_name(parent);
> > > > > > >         init.ops = &rzg2l_cpg_sipll5_ops;
> > > > > > > -       init.flags = 0;
> > > > > > > +       init.flags = CLK_ROUNDING_FW_MANAGED;
> > > > > >
> > > > > > Iff this is the Right Thing To Do (TM), it needs a comment, as
> > > > > > this clock is not managed by firmware.
> > > > >
> > > > > Before I start a larger discussion on patch 1 with more people
> > > > > about a name for this flag, help me understand why this provider
> > > > > has a noop determine rate. Is the hardware eventually programmed
> > > > > with a rate that's close enough to what was passed in? Or it
> > > > > doesn't really matter what the clock rate is, just as long as it
> > > > > is running? Or should the determine_rate function be filled out in this particular case?
> > > >
> > > > I'd like to defer to Biju, who added the empty round^Wdetermine rate function.
> > >
> > > PLL is capable of generating any frequency. that is the reason.
> > >
> > > But we could, call rzg2l_cpg_get_foutpostdiv_rate() from
> > > determine_rate() as modified rzg2l_cpg_get_foutpostdiv_rate()[1] can
> > > return errors
> > >
> > > [1]
> > > https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/tr
> > > ee/drivers/clk/renesas/rzg2l-cpg.c?h=next-20260227#n590
> >
> > OK. So how do you all want to proceed here?
> >
> > Do you want to fill out the determine rate in this driver? If so, we can just drop this particular
> > patch from this series.
>
> We plan to fill out the determine rate later, as it can return error.
>
> >
> > Since the PLL is capable of any frequency, do you want to keep the empty determine rate functionality?
> > If so, we could use a different name for this flag. Maybe CLK_ROUNDING_NOOP?
>
> I guess, maybe we could use CLK_ROUNDING_NOOP, till we have proper .determine_rate() for this driver???

Just keeping the empty function should work, too, right?

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