[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <58c4b791-fe7c-6b34-0c13-9e41a6960c7f@gmail.com>
Date: Sat, 27 Feb 2021 00:23:07 +0900
From: Chanwoo Choi <cwchoi00@...il.com>
To: Abel Vesa <abel.vesa@....com>, Rob Herring <robh@...nel.org>,
Shawn Guo <shawnguo@...nel.org>,
Sascha Hauer <kernel@...gutronix.de>,
Lucas Stach <l.stach@...gutronix.de>,
Fabio Estevam <festevam@...il.com>,
Chanwoo Choi <cw00.choi@...sung.com>,
Georgi Djakov <djakov@...nel.org>,
Dong Aisheng <aisheng.dong@....com>,
Peng Fan <peng.fan@....com>,
Martin Kepplinger <martink@...teo.de>,
devicetree@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
linux-clk@...r.kernel.org,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Cc: NXP Linux Team <linux-imx@....com>
Subject: Re: [RFC 07/19] devfreq: imx8m-ddrc: Add late system sleep PM ops
On 21. 2. 20. 오전 1:00, Abel Vesa wrote:
> Seems that, in order to be able to resume from suspend, the dram rate
> needs to be the highest one available. Therefore, add the late system
> suspend/resume PM ops which set the highest rate on suspend and the
> latest one used before suspending on resume.
>
> Signed-off-by: Abel Vesa <abel.vesa@....com>
> ---
> drivers/devfreq/imx8m-ddrc.c | 26 ++++++++++++++++++++++++++
> 1 file changed, 26 insertions(+)
>
> diff --git a/drivers/devfreq/imx8m-ddrc.c b/drivers/devfreq/imx8m-ddrc.c
> index 33de83acfd8b..04347dee781b 100644
> --- a/drivers/devfreq/imx8m-ddrc.c
> +++ b/drivers/devfreq/imx8m-ddrc.c
> @@ -72,6 +72,8 @@ struct imx8m_ddrc {
> struct clk *dram_alt;
> struct clk *dram_apb;
>
> + unsigned long suspend_rate;
> + unsigned long resume_rate;
> int freq_count;
> struct imx8m_ddrc_freq freq_table[IMX8M_DDRC_MAX_FREQ_COUNT];
> };
> @@ -271,6 +273,22 @@ static int imx8m_ddrc_target(struct device *dev, unsigned long *freq, u32 flags)
> return ret;
> }
>
> +static int imx8m_ddrc_suspend(struct device *dev)
> +{
> + struct imx8m_ddrc *priv = dev_get_drvdata(dev);
> +
> + priv->resume_rate = clk_get_rate(priv->dram_core);
> +
> + return imx8m_ddrc_target(dev, &priv->suspend_rate, 0);
> +}
> +
> +static int imx8m_ddrc_resume(struct device *dev)
> +{
> + struct imx8m_ddrc *priv = dev_get_drvdata(dev);
> +
> + return imx8m_ddrc_target(dev, &priv->resume_rate, 0);
> +}
> +
> static int imx8m_ddrc_get_cur_freq(struct device *dev, unsigned long *freq)
> {
> struct imx8m_ddrc *priv = dev_get_drvdata(dev);
> @@ -336,6 +354,9 @@ static int imx8m_ddrc_init_freq_info(struct device *dev)
>
> if (dev_pm_opp_add(dev, freq->rate * 250000, 0))
> return -ENODEV;
> +
> + if (index == 0)
> + priv->suspend_rate = freq->rate * 250000;
> }
>
> return 0;
> @@ -412,10 +433,15 @@ static const struct of_device_id imx8m_ddrc_of_match[] = {
> };
> MODULE_DEVICE_TABLE(of, imx8m_ddrc_of_match);
>
> +static const struct dev_pm_ops imx8m_ddrc_pm_ops = {
> + SET_LATE_SYSTEM_SLEEP_PM_OPS(imx8m_ddrc_suspend, imx8m_ddrc_resume)
> +};
Are there any reason to use suspend_late instead of suspend?
Usually, it is enough to change the frequency on normal suspend() step.
The devfreq supports the 'opp-suspend' property[1]. If you keep the OPP
entries on DT, you simply support your goal with 'opp-suspend'.
[1]
https://elixir.bootlin.com/linux/latest/source/Documentation/devicetree/bindings/opp/opp.txt#L156
> +
> static struct platform_driver imx8m_ddrc_platdrv = {
> .probe = imx8m_ddrc_probe,
> .driver = {
> .name = "imx8m-ddrc-devfreq",
> + .pm = &imx8m_ddrc_pm_ops,
> .of_match_table = of_match_ptr(imx8m_ddrc_of_match),
> },
> };
>
Powered by blists - more mailing lists