Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <898fe00e-43f4-a8aa-b2b4-5b293fc77640@denx.de>
Date:   Thu, 11 May 2023 21:25:46 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.10.0
Subject: Re: [PATCH] cpufreq: imx6q: Disable only available frequencies
Content-Language: en-US
To:     Christoph Niedermaier <cniedermaier@dh-electronics.com>,
        linux-arm-kernel@lists.infradead.org
Cc:     Viresh Kumar <viresh.kumar@linaro.org>,
        "Rafael J. Wysocki" <rafael@kernel.org>,
        Shawn Guo <shawnguo@kernel.org>,
        Fabio Estevam <festevam@denx.de>,
        NXP Linux Team <linux-imx@nxp.com>,
        "open list:CPU FREQUENCY SCALING FRAMEWORK" 
        <linux-pm@vger.kernel.org>,
        open list <linux-kernel@vger.kernel.org>
References: <20230511092334.3017-1-cniedermaier@dh-electronics.com>
From:   Marek Vasut <marex@denx.de>
In-Reply-To: <20230511092334.3017-1-cniedermaier@dh-electronics.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 5/11/23 11:23, Christoph Niedermaier wrote:
> In the example in Documentation/power/opp.rst, an availability check
> is present before disabling a specific frequency. If a frequency isn't
> available, the warning of a failed disabling of a non-existent
> frequency is misleading. Therefore, check the availability of the
> frequency in a separate inline function before disabling it.

[...]

> diff --git a/drivers/cpufreq/imx6q-cpufreq.c b/drivers/cpufreq/imx6q-cpufreq.c
> index 48e1772e98fd..4e2d2bc47aba 100644
> --- a/drivers/cpufreq/imx6q-cpufreq.c
> +++ b/drivers/cpufreq/imx6q-cpufreq.c
> @@ -209,6 +209,21 @@ static struct cpufreq_driver imx6q_cpufreq_driver = {
>   	.suspend = cpufreq_generic_suspend,
>   };
>   
> +static inline int disable_freq_if_available(struct device *dev,

The inline isn't needed, esp. on static function, let the compiler 
figure it out.

Also, "disable if available" should rather be "disable if unavailable" I 
think ?

> +					    unsigned long freq)
> +{
> +	struct dev_pm_opp *opp;
> +	int ret = 0;
> +
> +	opp = dev_pm_opp_find_freq_exact(dev, freq, true);
> +	if (!IS_ERR(opp)) {
> +		dev_pm_opp_put(opp);
> +		ret = dev_pm_opp_disable(dev, freq);
> +	}
> +
> +	return ret;
> +}
> +
>   #define OCOTP_CFG3			0x440
>   #define OCOTP_CFG3_SPEED_SHIFT		16
>   #define OCOTP_CFG3_SPEED_1P2GHZ		0x3
> @@ -254,16 +269,16 @@ static int imx6q_opp_check_speed_grading(struct device *dev)
>   	val &= 0x3;
>   
>   	if (val < OCOTP_CFG3_SPEED_996MHZ)
> -		if (dev_pm_opp_disable(dev, 996000000))
> +		if (disable_freq_if_available(dev, 996000000))
>   			dev_warn(dev, "failed to disable 996MHz OPP\n");
>   
>   	if (of_machine_is_compatible("fsl,imx6q") ||
>   	    of_machine_is_compatible("fsl,imx6qp")) {

Can we introduce a function like:

void imx_disable_freq_if_unavailable(struct device *dev, u32 freq_mhz, 
u32 val, u32 mask)
{
    if (val == mask)
      return;
    if (!disable_freq_if_available(dev, freq_mhz * 1000000))
      return;
    dev_warn(dev, "failed to disable %dMHz OPP\n", mhz);
}

And then just call it multiple times in here, to reduce duplication ?

>   		if (val != OCOTP_CFG3_SPEED_852MHZ)
> -			if (dev_pm_opp_disable(dev, 852000000))
> +			if (disable_freq_if_available(dev, 852000000))
>   				dev_warn(dev, "failed to disable 852MHz OPP\n");

[...]