* [PATCH v3 0/3] Export fan control and register fans as cooling devices
@ 2025-05-12 7:11 Sung-Chi Li via B4 Relay
2025-05-12 7:11 ` [PATCH v3 1/3] platform/chrome: update pwm fan control host commands Sung-Chi Li via B4 Relay
` (2 more replies)
0 siblings, 3 replies; 10+ messages in thread
From: Sung-Chi Li via B4 Relay @ 2025-05-12 7:11 UTC (permalink / raw)
To: Benson Leung, Guenter Roeck, Thomas Weißschuh, Jean Delvare,
Guenter Roeck, Jonathan Corbet
Cc: chrome-platform, linux-kernel, linux-hwmon, linux-doc,
Sung-Chi Li, Sung-Chi Li
This is a continuation of the previous series "Export the target RPM fan
control by ChromeOS EC under hwmon"
(https://lore.kernel.org/lkml/20250313-extend_ec_hwmon_fan-v1-0-5c566776f2c4@chromium.org/T/#t).
There is a change from controlling the target fan RPM value to control
the PWM value.
We anticipate to involve fans connected to EC as thermal cooling
devices, so we can utilize the thermal framework to have further thermal
control strategies.
This series updates the required EC controls definitions, implements the
mechanism for controlling fan PWM values, and registers these fans under
thermal framework as cooling devices.
Adapting comments from the previous series, the driver probes the host
command capability at beginning to see whether a fan is controllable:
- if command `EC_CMD_PWM_GET_FAN_DUTY` is supported (v0, this is a
new command).
- if command `EC_CMD_THERMAL_AUTO_FAN_CTRL` v2 is supported.
- if command `EC_CMD_PWM_SET_FAN_DUTY` v1 is supported.
This combination is selected as this is the minimum requirement for a
fan to be fully controllable under hwmon framework.
The driver supports changing the fan control mode, and allows to change
the fan PWM value only if the fan is in manual control mode. The power
management hook is implemented as well for keeping the fan control
settings, as EC will automatically restore the control method to auto
when device is suspended.
Change-Id: I4e2fdc8c4bc50778c0d04cfbefeaab7088d3181e
Signed-off-by: Sung-Chi Li <lschyi@google.com>
---
Changes in v3:
- Make required EC command versions macros.
- Add `CONFIG_THERMAL` guarding for registering as thermal fan cooling
devices.
- Add error handling during registering thermal cooling devices, and
immediately abort the registration if any error occurred to align with
the thermal sensor registration in hwmon core.
- Add error handling for EC fan communication during suspend and resume.
- Add `CONFIG_PM` guarding for checking whether the EC supports a
complete fan control in hwmon driver.
- Sort variables order in declaration.
- Separate declaration and logic to different sections.
- Move `cros_ec_thermal_cooling_ops` next right after the operation
functions declaration.
- Improve describing the resume behavior in documentation.
Changes in v2:
- Change column from 80 to 100 and fix styles.
- Directly store driver data into platform dev with
platform_set_drvdata.
- Unify the PWM unit (from 0 ~ 255) between hwmon and thermal cooling
devices.
- Only fetch the fan control mode and PWM value when suspending rather
than caching values when writing. The suspend hook is thus added.
- Link to v1: https://lore.kernel.org/r/20250429-cros_ec_fan-v1-0-a8d9e3efbb1a@chromium.org
---
Sung-Chi Li (3):
platform/chrome: update pwm fan control host commands
hwmon: (cros_ec) add PWM control over fans
hwmon: (cros_ec) register fans into thermal framework cooling devices
Documentation/hwmon/cros_ec_hwmon.rst | 7 +-
drivers/hwmon/cros_ec_hwmon.c | 306 +++++++++++++++++++++++++
include/linux/platform_data/cros_ec_commands.h | 29 ++-
3 files changed, 340 insertions(+), 2 deletions(-)
---
base-commit: 33035b665157558254b3c21c3f049fd728e72368
change-id: 20250429-cros_ec_fan-da3b64ac9c10
Best regards,
--
Sung-Chi Li <lschyi@chromium.org>
^ permalink raw reply [flat|nested] 10+ messages in thread
* [PATCH v3 1/3] platform/chrome: update pwm fan control host commands
2025-05-12 7:11 [PATCH v3 0/3] Export fan control and register fans as cooling devices Sung-Chi Li via B4 Relay
@ 2025-05-12 7:11 ` Sung-Chi Li via B4 Relay
2025-05-12 7:11 ` [PATCH v3 2/3] hwmon: (cros_ec) add PWM control over fans Sung-Chi Li via B4 Relay
2025-05-12 7:11 ` [PATCH v3 3/3] hwmon: (cros_ec) register fans into thermal framework cooling devices Sung-Chi Li via B4 Relay
2 siblings, 0 replies; 10+ messages in thread
From: Sung-Chi Li via B4 Relay @ 2025-05-12 7:11 UTC (permalink / raw)
To: Benson Leung, Guenter Roeck, Thomas Weißschuh, Jean Delvare,
Guenter Roeck, Jonathan Corbet
Cc: chrome-platform, linux-kernel, linux-hwmon, linux-doc,
Sung-Chi Li, Sung-Chi Li
From: Sung-Chi Li <lschyi@chromium.org>
Update cros_ec_commands.h to include definitions for getting PWM fan
duty, getting and setting the fan control mode.
Signed-off-by: Sung-Chi Li <lschyi@chromium.org>
---
include/linux/platform_data/cros_ec_commands.h | 29 +++++++++++++++++++++++++-
1 file changed, 28 insertions(+), 1 deletion(-)
diff --git a/include/linux/platform_data/cros_ec_commands.h b/include/linux/platform_data/cros_ec_commands.h
index 1f4e4f2b89bb936b4b1c3f4162fec203b196cbc8..2ac1a30f9a3195bfc9dffc72fe5c5b92d83f1ef2 100644
--- a/include/linux/platform_data/cros_ec_commands.h
+++ b/include/linux/platform_data/cros_ec_commands.h
@@ -1825,6 +1825,16 @@ struct ec_response_pwm_get_duty {
uint16_t duty; /* Duty cycle, EC_PWM_MAX_DUTY = 100% */
} __ec_align2;
+#define EC_CMD_PWM_GET_FAN_DUTY 0x0027
+
+struct ec_params_pwm_get_fan_duty {
+ uint8_t fan_idx;
+} __ec_align1;
+
+struct ec_response_pwm_get_fan_duty {
+ uint32_t percent; /* Percentage of duty cycle, ranging from 0 ~ 100 */
+} __ec_align4;
+
/*****************************************************************************/
/*
* Lightbar commands. This looks worse than it is. Since we only use one HOST
@@ -3105,14 +3115,31 @@ struct ec_params_thermal_set_threshold_v1 {
/****************************************************************************/
-/* Toggle automatic fan control */
+/* Set or get fan control mode */
#define EC_CMD_THERMAL_AUTO_FAN_CTRL 0x0052
+enum ec_auto_fan_ctrl_cmd {
+ EC_AUTO_FAN_CONTROL_CMD_SET = 0,
+ EC_AUTO_FAN_CONTROL_CMD_GET,
+};
+
/* Version 1 of input params */
struct ec_params_auto_fan_ctrl_v1 {
uint8_t fan_idx;
} __ec_align1;
+/* Version 2 of input params */
+struct ec_params_auto_fan_ctrl_v2 {
+ uint8_t fan_idx;
+ uint8_t cmd; /* enum ec_auto_fan_ctrl_cmd */
+ uint8_t set_auto; /* only used with EC_AUTO_FAN_CONTROL_CMD_SET - bool
+ */
+} __ec_align4;
+
+struct ec_response_auto_fan_control {
+ uint8_t is_auto; /* bool */
+} __ec_align1;
+
/* Get/Set TMP006 calibration data */
#define EC_CMD_TMP006_GET_CALIBRATION 0x0053
#define EC_CMD_TMP006_SET_CALIBRATION 0x0054
--
2.49.0.1015.ga840276032-goog
^ permalink raw reply related [flat|nested] 10+ messages in thread
* [PATCH v3 2/3] hwmon: (cros_ec) add PWM control over fans
2025-05-12 7:11 [PATCH v3 0/3] Export fan control and register fans as cooling devices Sung-Chi Li via B4 Relay
2025-05-12 7:11 ` [PATCH v3 1/3] platform/chrome: update pwm fan control host commands Sung-Chi Li via B4 Relay
@ 2025-05-12 7:11 ` Sung-Chi Li via B4 Relay
2025-05-12 7:30 ` Thomas Weißschuh
2025-05-12 7:11 ` [PATCH v3 3/3] hwmon: (cros_ec) register fans into thermal framework cooling devices Sung-Chi Li via B4 Relay
2 siblings, 1 reply; 10+ messages in thread
From: Sung-Chi Li via B4 Relay @ 2025-05-12 7:11 UTC (permalink / raw)
To: Benson Leung, Guenter Roeck, Thomas Weißschuh, Jean Delvare,
Guenter Roeck, Jonathan Corbet
Cc: chrome-platform, linux-kernel, linux-hwmon, linux-doc,
Sung-Chi Li, Sung-Chi Li
From: Sung-Chi Li <lschyi@chromium.org>
Newer EC firmware supports controlling fans through host commands, so
adding corresponding implementations for controlling these fans in the
driver for other kernel services and userspace to control them.
The driver will first probe the supported host command versions (get and
set of fan PWM values, get and set of fan control mode) to see if the
connected EC fulfills the requirements of controlling the fan, then
exposes corresponding sysfs nodes for userspace to control the fan with
corresponding read and write implementations.
As EC will automatically change the fan mode to auto when the device is
suspended, the power management hooks are added as well to keep the fan
control mode and fan PWM value consistent during suspend and resume. As
we need to access the hwmon device in the power management hook, update
the driver by storing the hwmon device in the driver data as well.
Signed-off-by: Sung-Chi Li <lschyi@chromium.org>
---
Documentation/hwmon/cros_ec_hwmon.rst | 5 +-
drivers/hwmon/cros_ec_hwmon.c | 228 ++++++++++++++++++++++++++++++++++
2 files changed, 232 insertions(+), 1 deletion(-)
diff --git a/Documentation/hwmon/cros_ec_hwmon.rst b/Documentation/hwmon/cros_ec_hwmon.rst
index 47ecae983bdbef4bfcafc5dd2fff3de039f77f8e..355557a08c9a54b4c177bafde3743e7dc02218be 100644
--- a/Documentation/hwmon/cros_ec_hwmon.rst
+++ b/Documentation/hwmon/cros_ec_hwmon.rst
@@ -23,4 +23,7 @@ ChromeOS embedded controller used in Chromebooks and other devices.
The channel labels exposed via hwmon are retrieved from the EC itself.
-Fan and temperature readings are supported.
+Fan and temperature readings are supported. PWM fan control is also supported if
+the EC also supports setting fan PWM values and fan mode. Note that EC will
+switch fan control mode back to auto when suspended. This driver will restore
+the fan state to what they were before suspended when resumed.
diff --git a/drivers/hwmon/cros_ec_hwmon.c b/drivers/hwmon/cros_ec_hwmon.c
index 9991c3fa020ac859cbbff29dfb669e53248df885..884be1d34000c1a7485a381373c78329108dd427 100644
--- a/drivers/hwmon/cros_ec_hwmon.c
+++ b/drivers/hwmon/cros_ec_hwmon.c
@@ -7,6 +7,7 @@
#include <linux/device.h>
#include <linux/hwmon.h>
+#include <linux/math.h>
#include <linux/mod_devicetable.h>
#include <linux/module.h>
#include <linux/platform_device.h>
@@ -17,10 +18,17 @@
#define DRV_NAME "cros-ec-hwmon"
+#define CROS_EC_HWMON_PWM_GET_FAN_DUTY_CMD_VERSION 0
+#define CROS_EC_HWMON_PWM_SET_FAN_DUTY_CMD_VERSION 1
+#define CROS_EC_HWMON_THERMAL_AUTO_FAN_CTRL_CMD_VERSION 2
+
struct cros_ec_hwmon_priv {
struct cros_ec_device *cros_ec;
const char *temp_sensor_names[EC_TEMP_SENSOR_ENTRIES + EC_TEMP_SENSOR_B_ENTRIES];
u8 usable_fans;
+ bool fan_control_supported;
+ u8 manual_fans; /* bits to indicate whether the fan is set to manual */
+ u8 manual_fan_pwm_values[EC_FAN_SPEED_ENTRIES];
};
static int cros_ec_hwmon_read_fan_speed(struct cros_ec_device *cros_ec, u8 index, u16 *speed)
@@ -36,6 +44,42 @@ static int cros_ec_hwmon_read_fan_speed(struct cros_ec_device *cros_ec, u8 index
return 0;
}
+static int cros_ec_hwmon_read_pwm_value(struct cros_ec_device *cros_ec, u8 index, u8 *pwm_value)
+{
+ struct ec_params_pwm_get_fan_duty req = {
+ .fan_idx = index,
+ };
+ struct ec_response_pwm_get_fan_duty resp;
+ int ret;
+
+ ret = cros_ec_cmd(cros_ec, CROS_EC_HWMON_PWM_GET_FAN_DUTY_CMD_VERSION,
+ EC_CMD_PWM_GET_FAN_DUTY, &req, sizeof(req), &resp, sizeof(resp));
+ if (ret < 0)
+ return ret;
+
+ *pwm_value = (u8)DIV_ROUND_CLOSEST(le32_to_cpu(resp.percent) * 255, 100);
+ return 0;
+}
+
+static int cros_ec_hwmon_read_pwm_enable(struct cros_ec_device *cros_ec, u8 index,
+ u8 *control_method)
+{
+ struct ec_params_auto_fan_ctrl_v2 req = {
+ .cmd = EC_AUTO_FAN_CONTROL_CMD_GET,
+ .fan_idx = index,
+ };
+ struct ec_response_auto_fan_control resp;
+ int ret;
+
+ ret = cros_ec_cmd(cros_ec, CROS_EC_HWMON_THERMAL_AUTO_FAN_CTRL_CMD_VERSION,
+ EC_CMD_THERMAL_AUTO_FAN_CTRL, &req, sizeof(req), &resp, sizeof(resp));
+ if (ret < 0)
+ return ret;
+
+ *control_method = resp.is_auto ? 2 : 1;
+ return 0;
+}
+
static int cros_ec_hwmon_read_temp(struct cros_ec_device *cros_ec, u8 index, u8 *temp)
{
unsigned int offset;
@@ -73,7 +117,9 @@ static long cros_ec_hwmon_temp_to_millicelsius(u8 temp)
static int cros_ec_hwmon_read(struct device *dev, enum hwmon_sensor_types type,
u32 attr, int channel, long *val)
{
+ u8 control_method;
struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
+ u8 pwm_value;
int ret = -EOPNOTSUPP;
u16 speed;
u8 temp;
@@ -92,6 +138,16 @@ static int cros_ec_hwmon_read(struct device *dev, enum hwmon_sensor_types type,
if (ret == 0)
*val = cros_ec_hwmon_is_error_fan(speed);
}
+ } else if (type == hwmon_pwm) {
+ if (attr == hwmon_pwm_enable) {
+ ret = cros_ec_hwmon_read_pwm_enable(priv->cros_ec, channel, &control_method);
+ if (ret == 0)
+ *val = control_method;
+ } else if (attr == hwmon_pwm_input) {
+ ret = cros_ec_hwmon_read_pwm_value(priv->cros_ec, channel, &pwm_value);
+ if (ret == 0)
+ *val = pwm_value;
+ }
} else if (type == hwmon_temp) {
if (attr == hwmon_temp_input) {
ret = cros_ec_hwmon_read_temp(priv->cros_ec, channel, &temp);
@@ -124,6 +180,74 @@ static int cros_ec_hwmon_read_string(struct device *dev, enum hwmon_sensor_types
return -EOPNOTSUPP;
}
+static int cros_ec_hwmon_set_fan_pwm_val(struct cros_ec_device *cros_ec, u8 index, u8 val)
+{
+ struct ec_params_pwm_set_fan_duty_v1 req = {
+ .fan_idx = index,
+ .percent = DIV_ROUND_CLOSEST((uint32_t)val * 100, 255),
+ };
+ int ret;
+
+ ret = cros_ec_cmd(cros_ec, CROS_EC_HWMON_PWM_SET_FAN_DUTY_CMD_VERSION,
+ EC_CMD_PWM_SET_FAN_DUTY, &req, sizeof(req), NULL, 0);
+ if (ret < 0)
+ return ret;
+ return 0;
+}
+
+static int cros_ec_hwmon_write_pwm_input(struct cros_ec_device *cros_ec, u8 index, u8 val)
+{
+ u8 control_method;
+ int ret;
+
+ ret = cros_ec_hwmon_read_pwm_enable(cros_ec, index, &control_method);
+ if (ret)
+ return ret;
+ if (control_method != 1)
+ return -EOPNOTSUPP;
+
+ return cros_ec_hwmon_set_fan_pwm_val(cros_ec, index, val);
+}
+
+static int cros_ec_hwmon_write_pwm_enable(struct cros_ec_device *cros_ec, u8 index, u8 val)
+{
+ struct ec_params_auto_fan_ctrl_v2 req = {
+ .fan_idx = index,
+ .cmd = EC_AUTO_FAN_CONTROL_CMD_SET,
+ };
+ int ret;
+
+ /* No CROS EC supports no fan speed control */
+ if (val == 0)
+ return -EOPNOTSUPP;
+
+ req.set_auto = (val != 1) ? true : false;
+ ret = cros_ec_cmd(cros_ec, CROS_EC_HWMON_THERMAL_AUTO_FAN_CTRL_CMD_VERSION,
+ EC_CMD_THERMAL_AUTO_FAN_CTRL, &req, sizeof(req), NULL, 0);
+ if (ret < 0)
+ return ret;
+ return 0;
+}
+
+static int cros_ec_hwmon_write(struct device *dev, enum hwmon_sensor_types type, u32 attr,
+ int channel, long val)
+{
+ struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
+
+ if (type == hwmon_pwm) {
+ switch (attr) {
+ case hwmon_pwm_input:
+ return cros_ec_hwmon_write_pwm_input(priv->cros_ec, channel, val);
+ case hwmon_pwm_enable:
+ return cros_ec_hwmon_write_pwm_enable(priv->cros_ec, channel, val);
+ default:
+ return -EOPNOTSUPP;
+ }
+ }
+
+ return -EOPNOTSUPP;
+}
+
static umode_t cros_ec_hwmon_is_visible(const void *data, enum hwmon_sensor_types type,
u32 attr, int channel)
{
@@ -132,6 +256,9 @@ static umode_t cros_ec_hwmon_is_visible(const void *data, enum hwmon_sensor_type
if (type == hwmon_fan) {
if (priv->usable_fans & BIT(channel))
return 0444;
+ } else if (type == hwmon_pwm) {
+ if (priv->fan_control_supported && priv->usable_fans & BIT(channel))
+ return 0644;
} else if (type == hwmon_temp) {
if (priv->temp_sensor_names[channel])
return 0444;
@@ -147,6 +274,11 @@ static const struct hwmon_channel_info * const cros_ec_hwmon_info[] = {
HWMON_F_INPUT | HWMON_F_FAULT,
HWMON_F_INPUT | HWMON_F_FAULT,
HWMON_F_INPUT | HWMON_F_FAULT),
+ HWMON_CHANNEL_INFO(pwm,
+ HWMON_PWM_INPUT | HWMON_PWM_ENABLE,
+ HWMON_PWM_INPUT | HWMON_PWM_ENABLE,
+ HWMON_PWM_INPUT | HWMON_PWM_ENABLE,
+ HWMON_PWM_INPUT | HWMON_PWM_ENABLE),
HWMON_CHANNEL_INFO(temp,
HWMON_T_INPUT | HWMON_T_FAULT | HWMON_T_LABEL,
HWMON_T_INPUT | HWMON_T_FAULT | HWMON_T_LABEL,
@@ -178,6 +310,7 @@ static const struct hwmon_channel_info * const cros_ec_hwmon_info[] = {
static const struct hwmon_ops cros_ec_hwmon_ops = {
.read = cros_ec_hwmon_read,
.read_string = cros_ec_hwmon_read_string,
+ .write = cros_ec_hwmon_write,
.is_visible = cros_ec_hwmon_is_visible,
};
@@ -233,6 +366,25 @@ static void cros_ec_hwmon_probe_fans(struct cros_ec_hwmon_priv *priv)
}
}
+static inline bool is_cros_ec_cmd_fulfilled(struct cros_ec_device *cros_ec,
+ u16 cmd, u8 version)
+{
+ int ret;
+
+ ret = cros_ec_get_cmd_versions(cros_ec, cmd);
+ return ret >= 0 && (ret & EC_VER_MASK(version));
+}
+
+static bool cros_ec_hwmon_probe_fan_control_supported(struct cros_ec_device *cros_ec)
+{
+ if (!IS_ENABLED(CONFIG_PM))
+ return false;
+
+ return is_cros_ec_cmd_fulfilled(cros_ec, EC_CMD_PWM_GET_FAN_DUTY, CROS_EC_HWMON_PWM_GET_FAN_DUTY_CMD_VERSION) &&
+ is_cros_ec_cmd_fulfilled(cros_ec, EC_CMD_PWM_SET_FAN_DUTY, CROS_EC_HWMON_PWM_SET_FAN_DUTY_CMD_VERSION) &&
+ is_cros_ec_cmd_fulfilled(cros_ec, EC_CMD_THERMAL_AUTO_FAN_CTRL, CROS_EC_HWMON_THERMAL_AUTO_FAN_CTRL_CMD_VERSION);
+}
+
static int cros_ec_hwmon_probe(struct platform_device *pdev)
{
struct device *dev = &pdev->dev;
@@ -259,13 +411,87 @@ static int cros_ec_hwmon_probe(struct platform_device *pdev)
cros_ec_hwmon_probe_temp_sensors(dev, priv, thermal_version);
cros_ec_hwmon_probe_fans(priv);
+ priv->fan_control_supported = cros_ec_hwmon_probe_fan_control_supported(priv->cros_ec);
hwmon_dev = devm_hwmon_device_register_with_info(dev, "cros_ec", priv,
&cros_ec_hwmon_chip_info, NULL);
+ platform_set_drvdata(pdev, priv);
return PTR_ERR_OR_ZERO(hwmon_dev);
}
+static int cros_ec_hwmon_suspend(struct platform_device *pdev, pm_message_t state)
+{
+ u8 control_method;
+ size_t i;
+ struct cros_ec_hwmon_priv *priv = platform_get_drvdata(pdev);
+ int ret;
+
+ if (!priv->fan_control_supported)
+ return 0;
+
+ /* EC sets fan control to auto after suspended, store settings before suspending. */
+ for (i = 0; i < EC_FAN_SPEED_ENTRIES; i++) {
+ if (!(priv->usable_fans & BIT(i)))
+ continue;
+
+ ret = cros_ec_hwmon_read_pwm_enable(priv->cros_ec, i, &control_method);
+ if (ret) {
+ dev_warn(&pdev->dev, "failed to get mode setting for fan %zu: %d", i, ret);
+ continue;
+ }
+
+ if (control_method != 1) {
+ priv->manual_fans &= ~BIT(i);
+ continue;
+ } else {
+ priv->manual_fans |= BIT(i);
+ }
+
+ ret = cros_ec_hwmon_read_pwm_value(
+ priv->cros_ec, i, &priv->manual_fan_pwm_values[i]);
+ /*
+ * If failed for storing the value, invalidate the stored mode value by setting it
+ * to auto control. EC will automatically switch to auto mode for that fan after
+ * suspended.
+ */
+ if (ret) {
+ priv->manual_fans &= ~BIT(i);
+ dev_warn(&pdev->dev, "failed to get PWM setting for fan %zu: %d", i, ret);
+ continue;
+ }
+ }
+
+ return 0;
+}
+
+static int cros_ec_hwmon_resume(struct platform_device *pdev)
+{
+ size_t i;
+ const struct cros_ec_hwmon_priv *priv = platform_get_drvdata(pdev);
+ int ret;
+
+ if (!priv->fan_control_supported)
+ return 0;
+
+ /* EC sets fan control to auto after suspended, restore to settings before suspended. */
+ for (i = 0; i < EC_FAN_SPEED_ENTRIES; i++) {
+ if (!(priv->manual_fans & BIT(i)))
+ continue;
+
+ /*
+ * Setting fan PWM value to EC will change the mode to manual for that fan in EC as
+ * well, so we do not need to issue a separate fan mode to manual call.
+ */
+ ret = cros_ec_hwmon_set_fan_pwm_val(
+ priv->cros_ec, i, priv->manual_fan_pwm_values[i]);
+ if (ret)
+ dev_warn(&pdev->dev, "failed to restore settings for fan %zu: %d", i, ret);
+ }
+
+ return 0;
+}
+
static const struct platform_device_id cros_ec_hwmon_id[] = {
{ DRV_NAME, 0 },
{}
@@ -274,6 +500,8 @@ static const struct platform_device_id cros_ec_hwmon_id[] = {
static struct platform_driver cros_ec_hwmon_driver = {
.driver.name = DRV_NAME,
.probe = cros_ec_hwmon_probe,
+ .suspend = pm_ptr(cros_ec_hwmon_suspend),
+ .resume = pm_ptr(cros_ec_hwmon_resume),
.id_table = cros_ec_hwmon_id,
};
module_platform_driver(cros_ec_hwmon_driver);
--
2.49.0.1015.ga840276032-goog
^ permalink raw reply related [flat|nested] 10+ messages in thread
* [PATCH v3 3/3] hwmon: (cros_ec) register fans into thermal framework cooling devices
2025-05-12 7:11 [PATCH v3 0/3] Export fan control and register fans as cooling devices Sung-Chi Li via B4 Relay
2025-05-12 7:11 ` [PATCH v3 1/3] platform/chrome: update pwm fan control host commands Sung-Chi Li via B4 Relay
2025-05-12 7:11 ` [PATCH v3 2/3] hwmon: (cros_ec) add PWM control over fans Sung-Chi Li via B4 Relay
@ 2025-05-12 7:11 ` Sung-Chi Li via B4 Relay
2025-05-12 7:35 ` Thomas Weißschuh
2 siblings, 1 reply; 10+ messages in thread
From: Sung-Chi Li via B4 Relay @ 2025-05-12 7:11 UTC (permalink / raw)
To: Benson Leung, Guenter Roeck, Thomas Weißschuh, Jean Delvare,
Guenter Roeck, Jonathan Corbet
Cc: chrome-platform, linux-kernel, linux-hwmon, linux-doc,
Sung-Chi Li, Sung-Chi Li
From: Sung-Chi Li <lschyi@chromium.org>
Register fans connected under EC as thermal cooling devices as well, so
these fans can then work with the thermal framework.
During the driver probing phase, we will also try to register each fan
as a thermal cooling device based on previous probe result (whether the
there are fans connected on that channel, and whether EC supports fan
control). The basic get max state, get current state, and set current
state methods are then implemented as well.
Signed-off-by: Sung-Chi Li <lschyi@chromium.org>
---
Documentation/hwmon/cros_ec_hwmon.rst | 2 +
drivers/hwmon/cros_ec_hwmon.c | 78 +++++++++++++++++++++++++++++++++++
2 files changed, 80 insertions(+)
diff --git a/Documentation/hwmon/cros_ec_hwmon.rst b/Documentation/hwmon/cros_ec_hwmon.rst
index 355557a08c9a54b4c177bafde3743e7dc02218be..6db812708325f7abb6d319af3312b4079e6923c6 100644
--- a/Documentation/hwmon/cros_ec_hwmon.rst
+++ b/Documentation/hwmon/cros_ec_hwmon.rst
@@ -27,3 +27,5 @@ Fan and temperature readings are supported. PWM fan control is also supported if
the EC also supports setting fan PWM values and fan mode. Note that EC will
switch fan control mode back to auto when suspended. This driver will restore
the fan state to what they were before suspended when resumed.
+If a fan is controllable, this driver will register that fan as a cooling device
+in the thermal framework as well.
diff --git a/drivers/hwmon/cros_ec_hwmon.c b/drivers/hwmon/cros_ec_hwmon.c
index 884be1d34000c1a7485a381373c78329108dd427..557d268d5ef52e1c6f61f2eba71b449231849f2a 100644
--- a/drivers/hwmon/cros_ec_hwmon.c
+++ b/drivers/hwmon/cros_ec_hwmon.c
@@ -13,6 +13,7 @@
#include <linux/platform_device.h>
#include <linux/platform_data/cros_ec_commands.h>
#include <linux/platform_data/cros_ec_proto.h>
+#include <linux/thermal.h>
#include <linux/types.h>
#include <linux/units.h>
@@ -31,6 +32,11 @@ struct cros_ec_hwmon_priv {
u8 manual_fan_pwm_values[EC_FAN_SPEED_ENTRIES];
};
+struct cros_ec_hwmon_cooling_priv {
+ struct cros_ec_hwmon_priv *hwmon_priv;
+ u8 index;
+};
+
static int cros_ec_hwmon_read_fan_speed(struct cros_ec_device *cros_ec, u8 index, u16 *speed)
{
int ret;
@@ -307,6 +313,42 @@ static const struct hwmon_channel_info * const cros_ec_hwmon_info[] = {
NULL
};
+static int cros_ec_hwmon_cooling_get_max_state(struct thermal_cooling_device *cdev,
+ unsigned long *val)
+{
+ *val = 255;
+ return 0;
+}
+
+static int cros_ec_hwmon_cooling_get_cur_state(struct thermal_cooling_device *cdev,
+ unsigned long *val)
+{
+ const struct cros_ec_hwmon_cooling_priv *priv = cdev->devdata;
+ u8 read_val;
+ int ret;
+
+ ret = cros_ec_hwmon_read_pwm_value(priv->hwmon_priv->cros_ec, priv->index, &read_val);
+ if (ret)
+ return ret;
+
+ *val = read_val;
+ return ret;
+}
+
+static int cros_ec_hwmon_cooling_set_cur_state(struct thermal_cooling_device *cdev,
+ unsigned long val)
+{
+ const struct cros_ec_hwmon_cooling_priv *priv = cdev->devdata;
+
+ return cros_ec_hwmon_write_pwm_input(priv->hwmon_priv->cros_ec, priv->index, val);
+}
+
+static const struct thermal_cooling_device_ops cros_ec_thermal_cooling_ops = {
+ .get_max_state = cros_ec_hwmon_cooling_get_max_state,
+ .get_cur_state = cros_ec_hwmon_cooling_get_cur_state,
+ .set_cur_state = cros_ec_hwmon_cooling_set_cur_state,
+};
+
static const struct hwmon_ops cros_ec_hwmon_ops = {
.read = cros_ec_hwmon_read,
.read_string = cros_ec_hwmon_read_string,
@@ -385,6 +427,41 @@ static bool cros_ec_hwmon_probe_fan_control_supported(struct cros_ec_device *cro
is_cros_ec_cmd_fulfilled(cros_ec, EC_CMD_THERMAL_AUTO_FAN_CTRL, CROS_EC_HWMON_THERMAL_AUTO_FAN_CTRL_CMD_VERSION);
}
+static void cros_ec_hwmon_register_fan_cooling_devices(struct device *dev,
+ struct cros_ec_hwmon_priv *priv)
+{
+ struct thermal_cooling_device *cdev;
+ struct cros_ec_hwmon_cooling_priv *cpriv;
+ size_t i;
+ char *type;
+
+ if (!IS_ENABLED(CONFIG_THERMAL))
+ return;
+
+ if (!priv->fan_control_supported)
+ return;
+
+ for (i = 0; i < EC_FAN_SPEED_ENTRIES; i++) {
+ if (!(priv->usable_fans & BIT(i)))
+ continue;
+
+ cpriv = devm_kzalloc(dev, sizeof(*cpriv), GFP_KERNEL);
+ if (!cpriv)
+ return;
+
+ type = devm_kasprintf(dev, GFP_KERNEL, "%s-fan%zu", dev_name(dev), i);
+ if (!type)
+ return;
+
+ cpriv->hwmon_priv = priv;
+ cpriv->index = i;
+ cdev = devm_thermal_of_cooling_device_register(dev, NULL, type, cpriv,
+ &cros_ec_thermal_cooling_ops);
+ if (!cdev)
+ return;
+ }
+}
+
static int cros_ec_hwmon_probe(struct platform_device *pdev)
{
struct device *dev = &pdev->dev;
@@ -412,6 +489,7 @@ static int cros_ec_hwmon_probe(struct platform_device *pdev)
cros_ec_hwmon_probe_temp_sensors(dev, priv, thermal_version);
cros_ec_hwmon_probe_fans(priv);
priv->fan_control_supported = cros_ec_hwmon_probe_fan_control_supported(priv->cros_ec);
+ cros_ec_hwmon_register_fan_cooling_devices(dev, priv);
hwmon_dev = devm_hwmon_device_register_with_info(dev, "cros_ec", priv,
&cros_ec_hwmon_chip_info, NULL);
--
2.49.0.1015.ga840276032-goog
^ permalink raw reply related [flat|nested] 10+ messages in thread
* Re: [PATCH v3 2/3] hwmon: (cros_ec) add PWM control over fans
2025-05-12 7:11 ` [PATCH v3 2/3] hwmon: (cros_ec) add PWM control over fans Sung-Chi Li via B4 Relay
@ 2025-05-12 7:30 ` Thomas Weißschuh
2025-06-18 7:26 ` Sung-Chi Li
0 siblings, 1 reply; 10+ messages in thread
From: Thomas Weißschuh @ 2025-05-12 7:30 UTC (permalink / raw)
To: lschyi
Cc: Benson Leung, Guenter Roeck, Jean Delvare, Guenter Roeck,
Jonathan Corbet, chrome-platform, linux-kernel, linux-hwmon,
linux-doc, Sung-Chi Li
On 2025-05-12 15:11:56+0800, Sung-Chi Li via B4 Relay wrote:
> From: Sung-Chi Li <lschyi@chromium.org>
>
> Newer EC firmware supports controlling fans through host commands, so
> adding corresponding implementations for controlling these fans in the
> driver for other kernel services and userspace to control them.
>
> The driver will first probe the supported host command versions (get and
> set of fan PWM values, get and set of fan control mode) to see if the
> connected EC fulfills the requirements of controlling the fan, then
> exposes corresponding sysfs nodes for userspace to control the fan with
> corresponding read and write implementations.
> As EC will automatically change the fan mode to auto when the device is
> suspended, the power management hooks are added as well to keep the fan
> control mode and fan PWM value consistent during suspend and resume. As
> we need to access the hwmon device in the power management hook, update
> the driver by storing the hwmon device in the driver data as well.
>
> Signed-off-by: Sung-Chi Li <lschyi@chromium.org>
> ---
> Documentation/hwmon/cros_ec_hwmon.rst | 5 +-
> drivers/hwmon/cros_ec_hwmon.c | 228 ++++++++++++++++++++++++++++++++++
> 2 files changed, 232 insertions(+), 1 deletion(-)
<snip>
> static int cros_ec_hwmon_read_temp(struct cros_ec_device *cros_ec, u8 index, u8 *temp)
> {
> unsigned int offset;
> @@ -73,7 +117,9 @@ static long cros_ec_hwmon_temp_to_millicelsius(u8 temp)
> static int cros_ec_hwmon_read(struct device *dev, enum hwmon_sensor_types type,
> u32 attr, int channel, long *val)
> {
> + u8 control_method;
> struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
> + u8 pwm_value;
> int ret = -EOPNOTSUPP;
> u16 speed;
> u8 temp;
Ordering again.
This should be:
struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
int ret = -EOPNOTSUPP;
u8 control_method;
u8 pwm_value;
u16 speed;
u8 temp;
or:
struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
u8 control_method, pwm_value, temp;
int ret = -EOPNOTSUPP;
u16 speed;
<snip>
> +static inline bool is_cros_ec_cmd_fulfilled(struct cros_ec_device *cros_ec,
> + u16 cmd, u8 version)
"fulfilled" -> "available" or "present"
> +{
> + int ret;
> +
> + ret = cros_ec_get_cmd_versions(cros_ec, cmd);
> + return ret >= 0 && (ret & EC_VER_MASK(version));
> +}
> +
> +static bool cros_ec_hwmon_probe_fan_control_supported(struct cros_ec_device *cros_ec)
> +{
> + if (!IS_ENABLED(CONFIG_PM))
> + return false;
Why? This should generally work fine without CONFIG_PM.
Only the suspend/resume callbacks are unnecessary in that case.
> +
> + return is_cros_ec_cmd_fulfilled(cros_ec, EC_CMD_PWM_GET_FAN_DUTY, CROS_EC_HWMON_PWM_GET_FAN_DUTY_CMD_VERSION) &&
> + is_cros_ec_cmd_fulfilled(cros_ec, EC_CMD_PWM_SET_FAN_DUTY, CROS_EC_HWMON_PWM_SET_FAN_DUTY_CMD_VERSION) &&
> + is_cros_ec_cmd_fulfilled(cros_ec, EC_CMD_THERMAL_AUTO_FAN_CTRL, CROS_EC_HWMON_THERMAL_AUTO_FAN_CTRL_CMD_VERSION);
> +}
<snip>
> +static int cros_ec_hwmon_suspend(struct platform_device *pdev, pm_message_t state)
> +{
> + u8 control_method;
> + size_t i;
> + struct cros_ec_hwmon_priv *priv = platform_get_drvdata(pdev);
> + int ret;
Ordering.
> +
> + if (!priv->fan_control_supported)
> + return 0;
> +
> + /* EC sets fan control to auto after suspended, store settings before suspending. */
> + for (i = 0; i < EC_FAN_SPEED_ENTRIES; i++) {
> + if (!(priv->usable_fans & BIT(i)))
> + continue;
> +
> + ret = cros_ec_hwmon_read_pwm_enable(priv->cros_ec, i, &control_method);
> + if (ret) {
> + dev_warn(&pdev->dev, "failed to get mode setting for fan %zu: %d", i, ret);
> + continue;
> + }
> +
> + if (control_method != 1) {
> + priv->manual_fans &= ~BIT(i);
> + continue;
> + } else {
> + priv->manual_fans |= BIT(i);
> + }
> +
> + ret = cros_ec_hwmon_read_pwm_value(
> + priv->cros_ec, i, &priv->manual_fan_pwm_values[i]);
> + /*
> + * If failed for storing the value, invalidate the stored mode value by setting it
"If storing the value failed"
> + * to auto control. EC will automatically switch to auto mode for that fan after
> + * suspended.
> + */
> + if (ret) {
> + priv->manual_fans &= ~BIT(i);
> + dev_warn(&pdev->dev, "failed to get PWM setting for fan %zu: %d", i, ret);
Print the warning first, so it's easier to follow the logic.
> + continue;
> + }
> + }
> +
> + return 0;
> +}
> +
> +static int cros_ec_hwmon_resume(struct platform_device *pdev)
> +{
> + size_t i;
> + const struct cros_ec_hwmon_priv *priv = platform_get_drvdata(pdev);
> + int ret;
Ordering.
> +
> + if (!priv->fan_control_supported)
> + return 0;
> +
> + /* EC sets fan control to auto after suspended, restore to settings before suspended. */
"after suspend"
> + for (i = 0; i < EC_FAN_SPEED_ENTRIES; i++) {
> + if (!(priv->manual_fans & BIT(i)))
> + continue;
> +
> + /*
> + * Setting fan PWM value to EC will change the mode to manual for that fan in EC as
> + * well, so we do not need to issue a separate fan mode to manual call.
> + */
> + ret = cros_ec_hwmon_set_fan_pwm_val(
> + priv->cros_ec, i, priv->manual_fan_pwm_values[i]);
You could rename "manual_fan_pwm_values" to "manual_fan_pwm" to save
some linebreaks here and in _suspend().
> + if (ret)
> + dev_warn(&pdev->dev, "failed to restore settings for fan %zu: %d", i, ret);
You can use %pe to print errors instead of %d:
"%zu: %pe\n", i, ERR_PTR(ret);
Also missing newline at the end of the message.
> + }
> +
> + return 0;
> +}
<snip>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v3 3/3] hwmon: (cros_ec) register fans into thermal framework cooling devices
2025-05-12 7:11 ` [PATCH v3 3/3] hwmon: (cros_ec) register fans into thermal framework cooling devices Sung-Chi Li via B4 Relay
@ 2025-05-12 7:35 ` Thomas Weißschuh
2025-06-19 7:31 ` Sung-Chi Li
0 siblings, 1 reply; 10+ messages in thread
From: Thomas Weißschuh @ 2025-05-12 7:35 UTC (permalink / raw)
To: lschyi
Cc: Benson Leung, Guenter Roeck, Jean Delvare, Guenter Roeck,
Jonathan Corbet, chrome-platform, linux-kernel, linux-hwmon,
linux-doc, Sung-Chi Li
On 2025-05-12 15:11:57+0800, Sung-Chi Li via B4 Relay wrote:
> From: Sung-Chi Li <lschyi@chromium.org>
>
> Register fans connected under EC as thermal cooling devices as well, so
> these fans can then work with the thermal framework.
>
> During the driver probing phase, we will also try to register each fan
> as a thermal cooling device based on previous probe result (whether the
> there are fans connected on that channel, and whether EC supports fan
> control). The basic get max state, get current state, and set current
> state methods are then implemented as well.
>
> Signed-off-by: Sung-Chi Li <lschyi@chromium.org>
> ---
> Documentation/hwmon/cros_ec_hwmon.rst | 2 +
> drivers/hwmon/cros_ec_hwmon.c | 78 +++++++++++++++++++++++++++++++++++
> 2 files changed, 80 insertions(+)
<snip>
> +static void cros_ec_hwmon_register_fan_cooling_devices(struct device *dev,
> + struct cros_ec_hwmon_priv *priv)
> +{
> + struct thermal_cooling_device *cdev;
> + struct cros_ec_hwmon_cooling_priv *cpriv;
> + size_t i;
> + char *type;
Ordering.
> +
> + if (!IS_ENABLED(CONFIG_THERMAL))
> + return;
> +
> + if (!priv->fan_control_supported)
> + return;
> +
> + for (i = 0; i < EC_FAN_SPEED_ENTRIES; i++) {
> + if (!(priv->usable_fans & BIT(i)))
> + continue;
> +
> + cpriv = devm_kzalloc(dev, sizeof(*cpriv), GFP_KERNEL);
> + if (!cpriv)
> + return;
The failures are swallowed silently. They should be propagated.
> +
> + type = devm_kasprintf(dev, GFP_KERNEL, "%s-fan%zu", dev_name(dev), i);
> + if (!type)
> + return;
> +
> + cpriv->hwmon_priv = priv;
> + cpriv->index = i;
> + cdev = devm_thermal_of_cooling_device_register(dev, NULL, type, cpriv,
> + &cros_ec_thermal_cooling_ops);
> + if (!cdev)
..._cooling_device_register() returns an error pointer on failure, not NULL.
> + return;
> + }
> +}
> +
> static int cros_ec_hwmon_probe(struct platform_device *pdev)
> {
> struct device *dev = &pdev->dev;
> @@ -412,6 +489,7 @@ static int cros_ec_hwmon_probe(struct platform_device *pdev)
> cros_ec_hwmon_probe_temp_sensors(dev, priv, thermal_version);
> cros_ec_hwmon_probe_fans(priv);
> priv->fan_control_supported = cros_ec_hwmon_probe_fan_control_supported(priv->cros_ec);
> + cros_ec_hwmon_register_fan_cooling_devices(dev, priv);
>
> hwmon_dev = devm_hwmon_device_register_with_info(dev, "cros_ec", priv,
> &cros_ec_hwmon_chip_info, NULL);
>
> --
> 2.49.0.1015.ga840276032-goog
>
>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v3 2/3] hwmon: (cros_ec) add PWM control over fans
2025-05-12 7:30 ` Thomas Weißschuh
@ 2025-06-18 7:26 ` Sung-Chi Li
2025-06-18 10:40 ` Thomas Weißschuh
2025-06-18 12:53 ` Guenter Roeck
0 siblings, 2 replies; 10+ messages in thread
From: Sung-Chi Li @ 2025-06-18 7:26 UTC (permalink / raw)
To: Thomas Weißschuh
Cc: Benson Leung, Guenter Roeck, Jean Delvare, Guenter Roeck,
Jonathan Corbet, chrome-platform, linux-kernel, linux-hwmon,
linux-doc
On Mon, May 12, 2025 at 09:30:39AM +0200, Thomas Weißschuh wrote:
Sorry for the late reply, I missed mails for this series.
> On 2025-05-12 15:11:56+0800, Sung-Chi Li via B4 Relay wrote:
> > From: Sung-Chi Li <lschyi@chromium.org>
> > static int cros_ec_hwmon_read_temp(struct cros_ec_device *cros_ec, u8 index, u8 *temp)
> > {
> > unsigned int offset;
> > @@ -73,7 +117,9 @@ static long cros_ec_hwmon_temp_to_millicelsius(u8 temp)
> > static int cros_ec_hwmon_read(struct device *dev, enum hwmon_sensor_types type,
> > u32 attr, int channel, long *val)
> > {
> > + u8 control_method;
> > struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
> > + u8 pwm_value;
> > int ret = -EOPNOTSUPP;
> > u16 speed;
> > u8 temp;
>
> Ordering again.
>
> This should be:
>
> struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
> int ret = -EOPNOTSUPP;
> u8 control_method;
> u8 pwm_value;
> u16 speed;
> u8 temp;
>
> or:
>
> struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
> u8 control_method, pwm_value, temp;
> int ret = -EOPNOTSUPP;
> u16 speed;
>
> <snip>
>
Would you mind to share the sorting logic, so I do not bother you with checking
these styling issue? Initially, I thought the sorting is based on the variable
name, but after seeing your example (which I am appreciated), I am not sure how
the sorting works. Is it sorted along with the variable types (
"u8 control_method", "u16 speed", etc)? If that is the case, why "u16 speed" is
in the middle of the u8 group variables?
> > +static inline bool is_cros_ec_cmd_fulfilled(struct cros_ec_device *cros_ec,
> > + u16 cmd, u8 version)
>
> "fulfilled" -> "available" or "present"
>
> > +{
> > + int ret;
> > +
> > + ret = cros_ec_get_cmd_versions(cros_ec, cmd);
> > + return ret >= 0 && (ret & EC_VER_MASK(version));
> > +}
> > +
> > +static bool cros_ec_hwmon_probe_fan_control_supported(struct cros_ec_device *cros_ec)
> > +{
> > + if (!IS_ENABLED(CONFIG_PM))
> > + return false;
>
> Why? This should generally work fine without CONFIG_PM.
> Only the suspend/resume callbacks are unnecessary in that case.
>
I treat fan control should include restoring the fan setting after resume, so
I think if no CONFIG_PM, the fan control is not complete. I am good with
removing this check, and if you have any thoughts after this explanation, please
share with me, otherwise I will remove it in the next series.
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v3 2/3] hwmon: (cros_ec) add PWM control over fans
2025-06-18 7:26 ` Sung-Chi Li
@ 2025-06-18 10:40 ` Thomas Weißschuh
2025-06-18 12:53 ` Guenter Roeck
1 sibling, 0 replies; 10+ messages in thread
From: Thomas Weißschuh @ 2025-06-18 10:40 UTC (permalink / raw)
To: Sung-Chi Li
Cc: Benson Leung, Guenter Roeck, Jean Delvare, Guenter Roeck,
Jonathan Corbet, chrome-platform, linux-kernel, linux-hwmon,
linux-doc
On 2025-06-18 15:26:38+0800, Sung-Chi Li wrote:
> On Mon, May 12, 2025 at 09:30:39AM +0200, Thomas Weißschuh wrote:
>
> Sorry for the late reply, I missed mails for this series.
>
> > On 2025-05-12 15:11:56+0800, Sung-Chi Li via B4 Relay wrote:
> > > From: Sung-Chi Li <lschyi@chromium.org>
> > > static int cros_ec_hwmon_read_temp(struct cros_ec_device *cros_ec, u8 index, u8 *temp)
> > > {
> > > unsigned int offset;
> > > @@ -73,7 +117,9 @@ static long cros_ec_hwmon_temp_to_millicelsius(u8 temp)
> > > static int cros_ec_hwmon_read(struct device *dev, enum hwmon_sensor_types type,
> > > u32 attr, int channel, long *val)
> > > {
> > > + u8 control_method;
> > > struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
> > > + u8 pwm_value;
> > > int ret = -EOPNOTSUPP;
> > > u16 speed;
> > > u8 temp;
> >
> > Ordering again.
> >
> > This should be:
> >
> > struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
> > int ret = -EOPNOTSUPP;
> > u8 control_method;
> > u8 pwm_value;
> > u16 speed;
> > u8 temp;
> >
> > or:
> >
> > struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
> > u8 control_method, pwm_value, temp;
> > int ret = -EOPNOTSUPP;
> > u16 speed;
> >
> > <snip>
> >
>
> Would you mind to share the sorting logic, so I do not bother you with checking
> these styling issue? Initially, I thought the sorting is based on the variable
> name, but after seeing your example (which I am appreciated), I am not sure how
> the sorting works. Is it sorted along with the variable types (
> "u8 control_method", "u16 speed", etc)? If that is the case, why "u16 speed" is
> in the middle of the u8 group variables?
Welp, I really should have explained this, instead of assuming it to be
obvious, sorry. The sorting is based on the length of the line, AKA
the "reverse christmas tree".
While this is not a strict rule, the driver is already using it, so I'd
like to stick to it. "Bonus points" for preserving preserving
semantically useful ordering within the line-length based one.
> > > +static inline bool is_cros_ec_cmd_fulfilled(struct cros_ec_device *cros_ec,
> > > + u16 cmd, u8 version)
> >
> > "fulfilled" -> "available" or "present"
> >
> > > +{
> > > + int ret;
> > > +
> > > + ret = cros_ec_get_cmd_versions(cros_ec, cmd);
> > > + return ret >= 0 && (ret & EC_VER_MASK(version));
> > > +}
> > > +
> > > +static bool cros_ec_hwmon_probe_fan_control_supported(struct cros_ec_device *cros_ec)
> > > +{
> > > + if (!IS_ENABLED(CONFIG_PM))
> > > + return false;
> >
> > Why? This should generally work fine without CONFIG_PM.
> > Only the suspend/resume callbacks are unnecessary in that case.
> >
>
> I treat fan control should include restoring the fan setting after resume, so
> I think if no CONFIG_PM, the fan control is not complete. I am good with
> removing this check, and if you have any thoughts after this explanation, please
> share with me, otherwise I will remove it in the next series.
As far as I understand, with CONFIG_PM=n the kernel won't be able to do
any suspend/resume. So cros_ec_hwmon won't need to handle the
suspend/resume case. However even then the regular UAPI for fan duty
management is useful for users.
Thomas
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v3 2/3] hwmon: (cros_ec) add PWM control over fans
2025-06-18 7:26 ` Sung-Chi Li
2025-06-18 10:40 ` Thomas Weißschuh
@ 2025-06-18 12:53 ` Guenter Roeck
1 sibling, 0 replies; 10+ messages in thread
From: Guenter Roeck @ 2025-06-18 12:53 UTC (permalink / raw)
To: Sung-Chi Li, Thomas Weißschuh
Cc: Benson Leung, Guenter Roeck, Jean Delvare, Jonathan Corbet,
chrome-platform, linux-kernel, linux-hwmon, linux-doc
On 6/18/25 00:26, Sung-Chi Li wrote:
> On Mon, May 12, 2025 at 09:30:39AM +0200, Thomas Weißschuh wrote:
>
> Sorry for the late reply, I missed mails for this series.
>
>> On 2025-05-12 15:11:56+0800, Sung-Chi Li via B4 Relay wrote:
>>> From: Sung-Chi Li <lschyi@chromium.org>
>>> static int cros_ec_hwmon_read_temp(struct cros_ec_device *cros_ec, u8 index, u8 *temp)
>>> {
>>> unsigned int offset;
>>> @@ -73,7 +117,9 @@ static long cros_ec_hwmon_temp_to_millicelsius(u8 temp)
>>> static int cros_ec_hwmon_read(struct device *dev, enum hwmon_sensor_types type,
>>> u32 attr, int channel, long *val)
>>> {
>>> + u8 control_method;
>>> struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
>>> + u8 pwm_value;
>>> int ret = -EOPNOTSUPP;
>>> u16 speed;
>>> u8 temp;
>>
>> Ordering again.
>>
>> This should be:
>>
>> struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
>> int ret = -EOPNOTSUPP;
>> u8 control_method;
>> u8 pwm_value;
>> u16 speed;
>> u8 temp;
>>
>> or:
>>
>> struct cros_ec_hwmon_priv *priv = dev_get_drvdata(dev);
>> u8 control_method, pwm_value, temp;
>> int ret = -EOPNOTSUPP;
>> u16 speed;
>>
>> <snip>
>>
>
> Would you mind to share the sorting logic, so I do not bother you with checking
> these styling issue? Initially, I thought the sorting is based on the variable
> name, but after seeing your example (which I am appreciated), I am not sure how
> the sorting works. Is it sorted along with the variable types (
> "u8 control_method", "u16 speed", etc)? If that is the case, why "u16 speed" is
> in the middle of the u8 group variables?
>
It is called "reverse christmas tree". Longer lines first.
Guenter
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH v3 3/3] hwmon: (cros_ec) register fans into thermal framework cooling devices
2025-05-12 7:35 ` Thomas Weißschuh
@ 2025-06-19 7:31 ` Sung-Chi Li
0 siblings, 0 replies; 10+ messages in thread
From: Sung-Chi Li @ 2025-06-19 7:31 UTC (permalink / raw)
To: Thomas Weißschuh
Cc: Benson Leung, Guenter Roeck, Jean Delvare, Guenter Roeck,
Jonathan Corbet, chrome-platform, linux-kernel, linux-hwmon,
linux-doc
On Mon, May 12, 2025 at 09:35:47AM +0200, Thomas Weißschuh wrote:
> On 2025-05-12 15:11:57+0800, Sung-Chi Li via B4 Relay wrote:
> > + if (!IS_ENABLED(CONFIG_THERMAL))
> > + return;
> > +
> > + if (!priv->fan_control_supported)
> > + return;
> > +
> > + for (i = 0; i < EC_FAN_SPEED_ENTRIES; i++) {
> > + if (!(priv->usable_fans & BIT(i)))
> > + continue;
> > +
> > + cpriv = devm_kzalloc(dev, sizeof(*cpriv), GFP_KERNEL);
> > + if (!cpriv)
> > + return;
>
> The failures are swallowed silently. They should be propagated.
>
After rethinking about the logic, I think I should try and register all fans
with best effort, so I should not return immediately; instead, I will log an
warning log, and continue with the next fan.
> > +
> > + type = devm_kasprintf(dev, GFP_KERNEL, "%s-fan%zu", dev_name(dev), i);
> > + if (!type)
> > + return;
> > +
> > + cpriv->hwmon_priv = priv;
> > + cpriv->index = i;
> > + cdev = devm_thermal_of_cooling_device_register(dev, NULL, type, cpriv,
> > + &cros_ec_thermal_cooling_ops);
> > + if (!cdev)
>
> ..._cooling_device_register() returns an error pointer on failure, not NULL.
>
Thank you, will use IS_ERR for checking and %pe to print error.
> > + return;
> > + }
> > +}
> > +
> > static int cros_ec_hwmon_probe(struct platform_device *pdev)
> > {
> > struct device *dev = &pdev->dev;
> > @@ -412,6 +489,7 @@ static int cros_ec_hwmon_probe(struct platform_device *pdev)
> > cros_ec_hwmon_probe_temp_sensors(dev, priv, thermal_version);
> > cros_ec_hwmon_probe_fans(priv);
> > priv->fan_control_supported = cros_ec_hwmon_probe_fan_control_supported(priv->cros_ec);
> > + cros_ec_hwmon_register_fan_cooling_devices(dev, priv);
> >
> > hwmon_dev = devm_hwmon_device_register_with_info(dev, "cros_ec", priv,
> > &cros_ec_hwmon_chip_info, NULL);
> >
> > --
> > 2.49.0.1015.ga840276032-goog
> >
> >
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2025-06-19 7:31 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-05-12 7:11 [PATCH v3 0/3] Export fan control and register fans as cooling devices Sung-Chi Li via B4 Relay
2025-05-12 7:11 ` [PATCH v3 1/3] platform/chrome: update pwm fan control host commands Sung-Chi Li via B4 Relay
2025-05-12 7:11 ` [PATCH v3 2/3] hwmon: (cros_ec) add PWM control over fans Sung-Chi Li via B4 Relay
2025-05-12 7:30 ` Thomas Weißschuh
2025-06-18 7:26 ` Sung-Chi Li
2025-06-18 10:40 ` Thomas Weißschuh
2025-06-18 12:53 ` Guenter Roeck
2025-05-12 7:11 ` [PATCH v3 3/3] hwmon: (cros_ec) register fans into thermal framework cooling devices Sung-Chi Li via B4 Relay
2025-05-12 7:35 ` Thomas Weißschuh
2025-06-19 7:31 ` Sung-Chi Li
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).