[PATCH v1 3/3] riscv: dts: starfive: Add mmc node

[William Qiu]

On 2022/12/8 0:31, Conor Dooley wrote:
> On Wed, Dec 07, 2022 at 04:14:53PM +0100, Krzysztof Kozlowski wrote:
>> On 07/12/2022 14:17, William Qiu wrote:
>> > This adds the mmc node for the StarFive JH7110 SoC.
>> > Set sdioo node to emmc and set sdio1 node to sd.
>> > 
>> > Signed-off-by: William Qiu <william.qiu at starfivetech.com>
>> > ---
>> >  .../jh7110-starfive-visionfive-v2.dts         | 25 ++++++++++++
>> >  arch/riscv/boot/dts/starfive/jh7110.dtsi      | 38 +++++++++++++++++++
>> >  2 files changed, 63 insertions(+)
>> > 
>> > diff --git a/arch/riscv/boot/dts/starfive/jh7110-starfive-visionfive-v2.dts b/arch/riscv/boot/dts/starfive/jh7110-starfive-visionfive-v2.dts
>> > index c8946cf3a268..6ef8e303c2e6 100644
>> > --- a/arch/riscv/boot/dts/starfive/jh7110-starfive-visionfive-v2.dts
>> > +++ b/arch/riscv/boot/dts/starfive/jh7110-starfive-visionfive-v2.dts
>> > @@ -47,6 +47,31 @@ &clk_rtc {
>> >  	clock-frequency = <32768>;
>> >  };
>> >  
>> > +&sdio0 {
>> > +	max-frequency = <100000000>;
>> > +	card-detect-delay = <300>;
>> > +	bus-width = <8>;
>> > +	cap-mmc-highspeed;
>> > +	mmc-ddr-1_8v;
>> > +	mmc-hs200-1_8v;
>> > +	non-removable;
>> > +	cap-mmc-hw-reset;
>> > +	post-power-on-delay-ms = <200>;
>> > +	status = "okay";
>> > +};
>> > +
>> > +&sdio1 {
>> > +	max-frequency = <100000000>;
>> > +	card-detect-delay = <300>;
>> > +	bus-width = <4>;
>> > +	no-sdio;
>> > +	no-mmc;
>> > +	broken-cd;
>> > +	cap-sd-highspeed;
>> > +	post-power-on-delay-ms = <200>;
>> > +	status = "okay";
>> > +};
>> > +
>> >  &gmac0_rmii_refin {
>> >  	clock-frequency = <50000000>;
>> >  };
>> > diff --git a/arch/riscv/boot/dts/starfive/jh7110.dtsi b/arch/riscv/boot/dts/starfive/jh7110.dtsi
>> > index c22e8f1d2640..e90b085d7e41 100644
>> > --- a/arch/riscv/boot/dts/starfive/jh7110.dtsi
>> > +++ b/arch/riscv/boot/dts/starfive/jh7110.dtsi
>> > @@ -331,6 +331,11 @@ aoncrg: clock-controller at 17000000 {
>> >  			#reset-cells = <1>;
>> >  		};
>> >  
>> > +		sys_syscon: sys_syscon at 13030000 {
>> 
>> No underscores in node names, generic node names (syscon or
>> system-controller)
>> 
>> > +			compatible = "syscon";
>> 
>> This is not allowed alone.
>> 
>> > +			reg = <0x0 0x13030000 0x0 0x1000>;
>> > +		};
>> > +
>> >  		gpio: gpio at 13040000 {
>> >  			compatible = "starfive,jh7110-sys-pinctrl";
>> >  			reg = <0x0 0x13040000 0x0 0x10000>;
>> > @@ -433,5 +438,38 @@ uart5: serial at 12020000 {
>> >  			reg-shift = <2>;
>> >  			status = "disabled";
>> >  		};
>> > +
>> > +		/* unremovable emmc as mmcblk0 */
>> > +		sdio0: mmc at 16010000 {
>> > +			compatible = "starfive,jh7110-sdio";
>> > +			reg = <0x0 0x16010000 0x0 0x10000>;
>> > +			clocks = <&syscrg JH7110_SYSCLK_SDIO0_AHB>,
>> > +				 <&syscrg JH7110_SYSCLK_SDIO0_SDCARD>;
>> > +			clock-names = "biu","ciu";
>> > +			resets = <&syscrg JH7110_SYSRST_SDIO0_AHB>;
>> > +			reset-names = "reset";
>> > +			interrupts = <74>;
>> > +			fifo-depth = <32>;
>> > +			fifo-watermark-aligned;
>> > +			data-addr = <0>;
>> > +			starfive,sys-syscon = <&sys_syscon 0x14 0x1a 0x7c000000>;
>> 
>> This does not match your bindings at all. "&sys_syscon" is a phandle,
>> not a number of tuning retries, as you expect in your bindings.
> 
> Additionally, a Link: to the documentation for where-ever these "random"
> numbers that are being used would be nice.
> 

I will update in next version.

> +static int dw_mci_starfive_parse_dt(struct dw_mci *host)
> +{
> +	struct of_phandle_args args;
> +	struct starfive_priv *priv;
> +	int ret;
> +
> +	priv = devm_kzalloc(host->dev, sizeof(*priv), GFP_KERNEL);
> +	if (!priv)
> +		return -ENOMEM;
> +
> +	ret = of_parse_phandle_with_fixed_args(host->dev->of_node,
> +						"starfive,sys-syscon", 3, 0, &args);
> +	if (ret) {
> +		dev_err(host->dev, "Failed to parse starfive,sys-syscon\n");
> +		return -EINVAL;
> +	}
> +
> +	priv->reg_syscon = syscon_node_to_regmap(args.np);
> +	of_node_put(args.np);
> +	if (IS_ERR(priv->reg_syscon))
> +		return PTR_ERR(priv->reg_syscon);
> +
> +	priv->syscon_offset = args.args[0];
> +	priv->syscon_shift  = args.args[1];
> +	priv->syscon_mask   = args.args[2];
> 
> Given the driver, the property description just seems incorrect and this
> is actually the bit of the syscon that is relevant to the tuning process
> (perhaps where the find the tuning values?). Without public docs or a
> better description it is hard for (me at least) to know :)
> 

Hi Conor,

Thank you for taking time to review and provide helpful comments for this patch.
I'll make the description better and more detailed in next version.

Best regards,
William Qiu
> +
> +	host->priv = priv;
> +
> +	return 0;
> +}