Re: [PATCH v2 3/8] riscv: dts: thead: Add TH1520 pin control nodes

[Conor Dooley]

On Tue, Jan 09, 2024 at 06:28:19AM -0800, Emil Renner Berthing wrote:
> Conor Dooley wrote:
> > On Tue, Jan 09, 2024 at 04:02:01AM -0800, Emil Renner Berthing wrote:
> > > Conor Dooley wrote:
> > > > On Wed, Jan 03, 2024 at 02:28:40PM +0100, Emil Renner Berthing wrote:
> > > > > Add nodes for pin controllers on the T-Head TH1520 RISC-V SoC.
> > > > >
> > > > > Signed-off-by: Emil Renner Berthing <emil.renner.berthing@xxxxxxxxxxxxx>
> > > > > ---
> > > > >  .../boot/dts/thead/th1520-beaglev-ahead.dts   |  4 ++++
> > > > >  .../dts/thead/th1520-lichee-module-4a.dtsi    |  4 ++++
> > > > >  arch/riscv/boot/dts/thead/th1520.dtsi         | 24 +++++++++++++++++++
> > > > >  3 files changed, 32 insertions(+)
> > > > >
> > > > > diff --git a/arch/riscv/boot/dts/thead/th1520-beaglev-ahead.dts b/arch/riscv/boot/dts/thead/th1520-beaglev-ahead.dts
> > > > > index 70e8042c8304..6c56318a8705 100644
> > > > > --- a/arch/riscv/boot/dts/thead/th1520-beaglev-ahead.dts
> > > > > +++ b/arch/riscv/boot/dts/thead/th1520-beaglev-ahead.dts
> > > > > @@ -44,6 +44,10 @@ &osc_32k {
> > > > >  	clock-frequency = <32768>;
> > > > >  };
> > > > >
> > > > > +&aonsys_clk {
> > > > > +	clock-frequency = <73728000>;
> > > > > +};
> > > > > +
> > > > >  &apb_clk {
> > > > >  	clock-frequency = <62500000>;
> > > > >  };
> > > > > diff --git a/arch/riscv/boot/dts/thead/th1520-lichee-module-4a.dtsi b/arch/riscv/boot/dts/thead/th1520-lichee-module-4a.dtsi
> > > > > index a802ab110429..9865925be372 100644
> > > > > --- a/arch/riscv/boot/dts/thead/th1520-lichee-module-4a.dtsi
> > > > > +++ b/arch/riscv/boot/dts/thead/th1520-lichee-module-4a.dtsi
> > > > > @@ -25,6 +25,10 @@ &osc_32k {
> > > > >  	clock-frequency = <32768>;
> > > > >  };
> > > > >
> > > > > +&aonsys_clk {
> > > > > +	clock-frequency = <73728000>;
> > > > > +};
> > > > > +
> > > > >  &apb_clk {
> > > > >  	clock-frequency = <62500000>;
> > > > >  };
> > > > > diff --git a/arch/riscv/boot/dts/thead/th1520.dtsi b/arch/riscv/boot/dts/thead/th1520.dtsi
> > > > > index ba4d2c673ac8..e65a306ff575 100644
> > > > > --- a/arch/riscv/boot/dts/thead/th1520.dtsi
> > > > > +++ b/arch/riscv/boot/dts/thead/th1520.dtsi
> > > > > @@ -134,6 +134,12 @@ osc_32k: 32k-oscillator {
> > > > >  		#clock-cells = <0>;
> > > > >  	};
> > > > >
> > > > > +	aonsys_clk: aonsys-clk {
> > > > > +		compatible = "fixed-clock";
> > > > > +		clock-output-names = "aonsys_clk";
> > > > > +		#clock-cells = <0>;
> > > > > +	};
> > > >
> > > > Did this stuff sneak into this commit accidentally?
> > >
> > > Not really by accident no. It turns out the clock tree has gates for the bus
> > > clock of each pinctrl block and I think it's better to add this clock
> > > dependency to the bindings and driver up front.
> >
> > Maybe if I had looked a wee bit more deeply I would've noticed that it
> > was used there, but it's always good to mention the rationale in the
> > commit message so that it's more obvious why you're doin it.
> 
> You absolutely right. I forgot to update the commit message.
> 
> > > Since there is not yet any clock driver the initial device tree for the TH1520
> > > included the dummy apb_clk that two of the pinctrl blocks derive their clock
> > > from, but not the "aonsys" clock needed by the "always-on" pinctrl. I thought
> > > it was better to add this dummy clock with the only (so far) user of it, but if
> > > you have a better idea, let me know.
> >
> > No, that's fine. I was just wondering why there was an unmentioned set
> > of clocks being added. If they're stubbed fixed clocks I dunno if it
> > makes sense to add them to the board.dts/module.dtsi files though. Where
> > do the initial values come from for the rates? Out of reset values or
> > set by firmware that may vary from board to board?
> 
> The vendor u-boot sets the PLLs different from the reset values. For now I
> think it's the same code for every board using the Lichee Pi 4A module (and
> probably also for the BeagleV Ahead), but it might still make sense to move the
> freqency to the board instead of the module device tree.

Yeah, think so. Only temporarily though, do you have a clue if anyone is
working on the actual clock driver stuff? Seems pretty Deadge?
https://lore.kernel.org/linux-clk/?q=th1520

Cheers,
Conor.

Attachment: signature.asc
Description: PGP signature