[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <b6668968-897f-4864-913c-d4d557f1d7cc@roeck-us.net>
Date: Mon, 24 Mar 2025 16:30:10 -0700
From: Guenter Roeck <linux@...ck-us.net>
To: Your Name <naresh.solanki@...ements.com>, Jean Delvare
<jdelvare@...e.com>, linux-hwmon@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] hwmon: (max6639) : Allow setting target RPM
On 3/24/25 11:57, Your Name wrote:
> From: Naresh Solanki <naresh.solanki@...ements.com>
>
> Currently, during startup, the fan is set to its maximum RPM by default,
> which may not be suitable for all use cases.
> This patch introduces support for specifying a target RPM via the Device
> Tree property "target-rpm".
>
> Changes:
> - Added `target_rpm` field to `max6639_data` structure to store the
> target RPM for each fan channel.
> - Modified `max6639_probe_child_from_dt()` to read the `"target-rpm"`
> property from the Device Tree and set `target_rpm` accordingly.
> - Updated `max6639_init_client()` to use `target_rpm` to compute the
> initial PWM duty cycle instead of defaulting to full speed (120/120).
>
> Behavior:
> - If `"target-rpm"` is specified, the fan speed is set accordingly.
> - If `"target-rpm"` is not specified, the previous behavior (full speed
> at startup) is retained.
>
Unless I am missing something, that is not really correct. See below.
> This allows better control over fan speed during system initialization.
>
> Signed-off-by: Naresh Solanki <naresh.solanki@...ements.com>
> ---
> drivers/hwmon/max6639.c | 15 ++++++++++++---
> 1 file changed, 12 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/hwmon/max6639.c b/drivers/hwmon/max6639.c
> index 32b4d54b2076..ca8a8f58d133 100644
> --- a/drivers/hwmon/max6639.c
> +++ b/drivers/hwmon/max6639.c
> @@ -80,6 +80,7 @@ struct max6639_data {
> /* Register values initialized only once */
> u8 ppr[MAX6639_NUM_CHANNELS]; /* Pulses per rotation 0..3 for 1..4 ppr */
> u8 rpm_range[MAX6639_NUM_CHANNELS]; /* Index in above rpm_ranges table */
> + u32 target_rpm[MAX6639_NUM_CHANNELS];
>
> /* Optional regulator for FAN supply */
> struct regulator *reg;
> @@ -560,8 +561,14 @@ static int max6639_probe_child_from_dt(struct i2c_client *client,
> }
>
target_rpm[] is 0 here.
> err = of_property_read_u32(child, "max-rpm", &val);
> - if (!err)
> + if (!err) {
> data->rpm_range[i] = rpm_range_to_reg(val);
> + data->target_rpm[i] = val;
> + }
If there is no max-rpm property, or if there is no devicetree support,
target_rpm[i] is still 0.
> +
> + err = of_property_read_u32(child, "target-rpm", &val);
> + if (!err)
> + data->target_rpm[i] = val;
If there is neither max-rpm nor target-rpm, target_rpm[i] is still 0.
>
> return 0;
> }
> @@ -573,6 +580,7 @@ static int max6639_init_client(struct i2c_client *client,
> const struct device_node *np = dev->of_node;
> struct device_node *child;
> int i, err;
> + u8 target_duty;
>
> /* Reset chip to default values, see below for GCONFIG setup */
> err = regmap_write(data->regmap, MAX6639_REG_GCONFIG, MAX6639_GCONFIG_POR);
> @@ -639,8 +647,9 @@ static int max6639_init_client(struct i2c_client *client,
> if (err)
> return err;
>
> - /* PWM 120/120 (i.e. 100%) */
> - err = regmap_write(data->regmap, MAX6639_REG_TARGTDUTY(i), 120);
> + /* Set PWM based on target RPM if specified */
> + target_duty = 120 * data->target_rpm[i] / rpm_ranges[data->rpm_range[i]];
If there is no devicetree support, or if neither max-rpm nor target-rpm are
provided, target_duty will be 0, and the fans will stop.
Maybe my interpretation is wrong, but I think both target_rpm[] and rpm_range[]
will need to be initialized. Also, it seems to me that there will need to be an
upper bound for target_rpm[]; without it, it is possible that target_duty > 120,
which would probably not be a good idea.
Guenter
> + err = regmap_write(data->regmap, MAX6639_REG_TARGTDUTY(i), target_duty);
> if (err)
> return err;
> }
>
> base-commit: 2115cbeec8a3ccc69e3b7ecdf97b4472b0829cfc
Powered by blists - more mailing lists