[PATCH 1/2] riscv: dts: cv1800b: add pinctrl node for cv1800b

[Conor Dooley]

On Mon, Nov 13, 2023 at 09:03:11PM +0800, Jisheng Zhang wrote:
> On Sun, Nov 12, 2023 at 08:51:20PM -0500, Samuel Holland wrote:
> > On 2023-11-12 6:57 PM, Jisheng Zhang wrote:
> > > Add the reset device tree node to cv1800b SoC reusing the
> >           ^^^^^
> >           I assume you mean pinctrl here?
> 
> oops copy and paste the commit msg ;) thanks
> > 
> > > pinctrl-single driver.
> > > 
> > > Signed-off-by: Jisheng Zhang <jszhang at kernel.org>
> > > ---
> > >  arch/riscv/boot/dts/sophgo/cv-pinctrl.h | 19 +++++++++++++++++++
> > >  arch/riscv/boot/dts/sophgo/cv1800b.dtsi | 10 ++++++++++
> > >  2 files changed, 29 insertions(+)
> > >  create mode 100644 arch/riscv/boot/dts/sophgo/cv-pinctrl.h
> > > 
> > > diff --git a/arch/riscv/boot/dts/sophgo/cv-pinctrl.h b/arch/riscv/boot/dts/sophgo/cv-pinctrl.h
> > > new file mode 100644
> > > index 000000000000..ed78b6fb3142
> > > --- /dev/null
> > > +++ b/arch/riscv/boot/dts/sophgo/cv-pinctrl.h
> > 
> > A couple of questions: Should this go in include/dt-bindings? And is it worth
> 
> When I cooked this series two weeks ago, I did put it in dt-binding, but
> then I found commit fe49f2d776f799 ("arm64: dts: ti: Use local header for
> pinctrl register values"), "These definitions were previously put in the
> bindings header to avoid code duplication and to provide some context
> meaning (name), but they do not fit the purpose of bindings." which is
> suggested and acked by Krzysztof, so I just want to follow the style
> here.
> 
> 
> > including macros for the actual function mappings, like in the vendor source[1]?
> 
> Do you want something as the following?
> 
> #define UART0_TX	0
> #define CAM_MCLK1	1
> ...
> 
> #define REG_UART0_TX	0x24
> ...
> 
> pinctrl-single,pins = <REG_UART0_TX UART0_TX>;
> 
> Other pinctl-single users just uses the register value directly, I have
> no preference. But I'd like to get suggestions from DT and pinctl-single
> maintainers. Hi Rob, Krzysztof, Conor, Tony, what's your opinion?

Basically, if the definitions map directly to registers and are just
used to make writing your devicetree easier then they do not belong
in a binding. This differs from clock or reset indices, where we
essentially make up a set of indices that may or may not correlate to
offsets in the hardware as using the register values without any sort of
abstraction is not defining an ABI.

Cheers,
Conor.
> 
> > 
> > [1]:
> > https://github.com/milkv-duo/duo-buildroot-sdk/blob/develop/linux_5.10/drivers/pinctrl/cvitek/cv180x_pinlist_swconfig.h
> > 
> > > @@ -0,0 +1,19 @@
> > > +/* SPDX-License-Identifier: GPL-2.0 */
> > > +/*
> > > + * This header provides constants for pinctrl bindings for Sophgo CV* SoC.
> > > + *
> > > + * Copyright (C) 2023 Jisheng Zhang <jszhang at kernel.org>
> > > + */
> > > +#ifndef _DTS_RISCV_SOPHGO_CV_PINCTRL_H
> > > +#define _DTS_RISCV_SOPHGO_CV_PINCTRL_H
> > > +
> > > +#define MUX_M0		0
> > > +#define MUX_M1		1
> > > +#define MUX_M2		2
> > > +#define MUX_M3		3
> > > +#define MUX_M4		4
> > > +#define MUX_M5		5
> > > +#define MUX_M6		6
> > > +#define MUX_M7		7
> > > +
> > > +#endif
> > > diff --git a/arch/riscv/boot/dts/sophgo/cv1800b.dtsi b/arch/riscv/boot/dts/sophgo/cv1800b.dtsi
> > > index e04df04a91c0..7a44d8e8672b 100644
> > > --- a/arch/riscv/boot/dts/sophgo/cv1800b.dtsi
> > > +++ b/arch/riscv/boot/dts/sophgo/cv1800b.dtsi
> > > @@ -6,6 +6,8 @@
> > >  #include <dt-bindings/interrupt-controller/irq.h>
> > >  #include <dt-bindings/reset/sophgo,cv1800b-reset.h>
> > >  
> > > +#include "cv-pinctrl.h"
> > > +
> > >  / {
> > >  	compatible = "sophgo,cv1800b";
> > >  	#address-cells = <1>;
> > > @@ -55,6 +57,14 @@ soc {
> > >  		dma-noncoherent;
> > >  		ranges;
> > >  
> > > +		pinctrl0: pinctrl at 3001000 {
> > > +			compatible = "pinctrl-single";
> > > +			reg = <0x3001000 0x130>;
> > > +			#pinctrl-cells = <1>;
> > > +			pinctrl-single,register-width = <32>;
> > > +			pinctrl-single,function-mask = <0x00000007>;
> > > +		};
> > 
> > From the vendor driver[2], it looks like this peripheral block only handles
> > pinmuxing, so indeed this looks like a good use of pinctrl-single.
> 
> This is deffinitely pinctrl-single style pinmux controller ;)
> 
> > 
> > [2]:
> > https://github.com/milkv-duo/duo-buildroot-sdk/blob/develop/linux_5.10/drivers/pinctrl/cvitek/pinctrl-cv180x.h
> > 
> > > +
> > >  		rst: reset-controller at 3003000 {
> > >  			compatible = "sophgo,cv1800b-reset";
> > >  			reg = <0x03003000 0x1000>;
> > 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-riscv/attachments/20231113/7545be8a/attachment.sig>