Re: [PATCH v2 3/3] arm64: dts: exynos: Add SPI nodes for Exynos850

[Sam Protsenko]

On Mon, Jan 29, 2024 at 11:51 AM Tudor Ambarus <tudor.ambarus@xxxxxxxxxx> wrote:
>
>
>
> On 1/25/24 01:38, Sam Protsenko wrote:
> > diff --git a/arch/arm64/boot/dts/exynos/exynos850.dtsi b/arch/arm64/boot/dts/exynos/exynos850.dtsi
> > index 618bc674896e..ca257da74b50 100644
> > --- a/arch/arm64/boot/dts/exynos/exynos850.dtsi
> > +++ b/arch/arm64/boot/dts/exynos/exynos850.dtsi
> > @@ -738,6 +738,24 @@ usi_spi_0: usi@139400c0 {
> >                                <&cmu_peri CLK_GOUT_SPI0_IPCLK>;
> >                       clock-names = "pclk", "ipclk";
> >                       status = "disabled";
> > +
> > +                     spi_0: spi@13940000 {
> > +                             compatible = "samsung,exynos850-spi";
> > +                             reg = <0x13940000 0x30>;
> > +                             interrupts = <GIC_SPI 221 IRQ_TYPE_LEVEL_HIGH>;
> > +                             pinctrl-0 = <&spi0_pins>;
> > +                             pinctrl-names = "default";
> > +                             clocks = <&cmu_peri CLK_GOUT_SPI0_IPCLK>,
> > +                                      <&cmu_peri CLK_GOUT_SPI0_PCLK>;
> > +                             clock-names = "spi_busclk0", "spi";
> > +                             samsung,spi-src-clk = <0>;
>
> this optional property
>

The reason this property is provided here despite being optional, is
to avoid corresponding dev_warn() message from spi-s3c64xx.c driver:

        if (of_property_read_u32(dev->of_node, "samsung,spi-src-clk", &temp)) {
            dev_warn(dev, "spi bus clock parent not specified, using
clock at index 0 as parent\n");

The same usage (samsung,spi-src-clk = <0>) can be encountered in
multiple other Exynos dts in arch/arm/ and arch/arm64/, and it's also
used in bindings example. Probably for the same reason explained
above. Even if dev_warn() is removed in the driver, I guess the older
kernels will still print it if spi-src-clk is omitted. So I'd like to
keep it here.

> > +                             dmas = <&pdma0 5>, <&pdma0 4>;
> > +                             dma-names = "tx", "rx";
> > +                             num-cs = <1>;
>
> and this one, are already defaults in the driver. Shall you remove them?
>

For exactly the same reasoning as stated above, I'd like to keep this
here to keep dmesg clean and tidy. Otherwise it prints this warning:

        if (of_property_read_u32(dev->of_node, "num-cs", &temp)) {
            dev_warn(dev, "number of chip select lines not specified,
assuming 1 chip select line\n");

And even if the warning is removed in the driver, older kernels will
still print it.

> > +                             #address-cells = <1>;
> > +                             #size-cells = <0>;
> > +                             status = "disabled";
> > +                     };
> >               };