Received-SPF: pass (google.com: domain of linux-kernel+bounces-4047-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
Feedback-ID: i0ad843c9:Fastmail
Message-ID: <aad8302d-a015-44ee-ad11-1a4c6e00074c@sholland.org>
Date: Mon, 18 Dec 2023 10:14:42 -0600
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH v2 1/5] riscv: dts: allwinner: Update opp table to allow
 CPU frequency scaling
Content-Language: en-US
To: Brandon Cheo Fusi <fusibrandon13@gmail.com>,
 Yangtao Li <tiny.windzz@gmail.com>, Viresh Kumar <vireshk@kernel.org>,
 Nishanth Menon <nm@ti.com>, Stephen Boyd <sboyd@kernel.org>,
 Rob Herring <robh+dt@kernel.org>,
 Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
 Conor Dooley <conor+dt@kernel.org>, Chen-Yu Tsai <wens@csie.org>,
 Jernej Skrabec <jernej.skrabec@gmail.com>,
 Paul Walmsley <paul.walmsley@sifive.com>, Palmer Dabbelt
 <palmer@dabbelt.com>, Albert Ou <aou@eecs.berkeley.edu>,
 "Rafael J . Wysocki" <rafael@kernel.org>, Maxime Ripard <mripard@kernel.org>
Cc: linux-pm@vger.kernel.org, devicetree@vger.kernel.org,
 linux-arm-kernel@lists.infradead.org, linux-sunxi@lists.linux.dev,
 linux-kernel@vger.kernel.org, linux-riscv@lists.infradead.org
References: <20231218110543.64044-1-fusibrandon13@gmail.com>
 <20231218110543.64044-2-fusibrandon13@gmail.com>
From: Samuel Holland <samuel@sholland.org>
In-Reply-To: <20231218110543.64044-2-fusibrandon13@gmail.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

On 12/18/23 05:05, Brandon Cheo Fusi wrote:
> Two OPPs are currently defined for the D1/D1s; one at 408MHz and
> another at 1.08GHz. Switching between these can be done with the
> "sun50i-cpufreq-nvmem" driver. This patch populates the opp table
> appropriately, inspired by
> https://github.com/Tina-Linux/linux-5.4/blob/master/arch/riscv/boot/dts/sunxi/sun20iw1p1.dtsi
> 
> The supply voltages are PWM-controlled, but support for that IP
> is still in the works. So stick to a target vdd-cpu supply of 0.9V,
> which seems to be the default on most D1 boards.
> 
> Signed-off-by: Brandon Cheo Fusi <fusibrandon13@gmail.com>
> ---
>  arch/riscv/boot/dts/allwinner/sun20i-d1s.dtsi | 18 +++++++++++++++---
>  1 file changed, 15 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/riscv/boot/dts/allwinner/sun20i-d1s.dtsi b/arch/riscv/boot/dts/allwinner/sun20i-d1s.dtsi
> index 64c3c2e6c..2f1771c19 100644
> --- a/arch/riscv/boot/dts/allwinner/sun20i-d1s.dtsi
> +++ b/arch/riscv/boot/dts/allwinner/sun20i-d1s.dtsi
> @@ -39,16 +39,22 @@ cpu0_intc: interrupt-controller {
>  	};
>  
>  	opp_table_cpu: opp-table-cpu {
> -		compatible = "operating-points-v2";
> +		compatible = "allwinner,sun20i-d1-operating-points",
> +				 "allwinner,sun50i-h6-operating-points";

This is not correct. The compatible string specifies the algorithm for
looking up the speed grade from the NVMEM cell, and D1 uses a totally
different algorithm compared to H6:

https://github.com/Tina-Linux/linux-5.4/blob/master/drivers/cpufreq/sun50i-cpufreq-nvmem.c#L293-L338

So you cannot use a fallback compatible string here, and you need to add
driver support for the D1's speed bin encoding.

Regards,
Samuel

> +		nvmem-cells = <&cpu_speed_grade>;
> +		nvmem-cell-names = "speed";
> +		opp-shared;
>  
>  		opp-408000000 {
> +			clock-latency-ns = <244144>; /* 8 32k periods */
>  			opp-hz = /bits/ 64 <408000000>;
> -			opp-microvolt = <900000 900000 1100000>;
> +			opp-microvolt-speed0 = <900000 900000 1100000>;
>  		};
>  
>  		opp-1080000000 {
> +			clock-latency-ns = <244144>; /* 8 32k periods */
>  			opp-hz = /bits/ 64 <1008000000>;
> -			opp-microvolt = <900000 900000 1100000>;
> +			opp-microvolt-speed0 = <900000 900000 1100000>;
>  		};
>  	};
>  
> @@ -115,3 +121,9 @@ pmu {
>  			<0x00000000 0x0000000f 0xffffffff 0xffffffff 0x00020000>;
>  	};
>  };
> +
> +&sid {
> +	cpu_speed_grade: cpu-speed-grade@0 {
> +		reg = <0x00 0x2>;
> +	};
> +};