[lm-sensors] [PATCH] Add driver for SMSC EMC2103 temperature monitor and fan controller
Jean Delvare
khali at linux-fr.org
Fri Jun 25 13:27:39 CEST 2010
Hi Steve,
On Sun, 6 Jun 2010 17:32:00 +0100, Steve Glendinning wrote:
> SMSC's EMC2103 family of temperature/fan controllers have 1
> onboard and up to 3 external temperature sensors, and allow
> closed-loop control of one fan. This patch adds support for
> them.
>
> Changes since v1:
> - rename driver to emc2103
> - expand return-hiding macros
>
> Signed-off-by: Steve Glendinning <steve.glendinning at smsc.com>
> ---
> MAINTAINERS | 6 +
> drivers/hwmon/Kconfig | 10 +
> drivers/hwmon/Makefile | 1 +
> drivers/hwmon/emc2103.c | 723 +++++++++++++++++++++++++++++++++++++++++++++++
It would be appreciated if you could also contribute some documentation
as Documentation/hwmon/emc2103. Take example on other files in this
directory. It doesn't need to be extra verbose.
> 4 files changed, 740 insertions(+), 0 deletions(-)
> create mode 100644 drivers/hwmon/emc2103.c
>
Finally found some time to review your driver:
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 67accd7..2f662a2 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -5208,6 +5208,12 @@ M: Nicolas Pitre <nico at fluxnic.net>
> S: Odd Fixes
> F: drivers/net/smc91x.*
>
> +SMSC EMC2103 HARDWARE MONITOR DRIVER
> +M: Steve Glendinning <steve.glendinning at smsc.com>
> +L: lm-sensors at lm-sensors.org
> +S: Supported
> +F: drivers/hwmon/emc2103.c
> +
> SMSC47B397 HARDWARE MONITOR DRIVER
> M: "Mark M. Hoffman" <mhoffman at lightlink.com>
> L: lm-sensors at lm-sensors.org
> diff --git a/drivers/hwmon/Kconfig b/drivers/hwmon/Kconfig
> index e19cf8e..569082c 100644
> --- a/drivers/hwmon/Kconfig
> +++ b/drivers/hwmon/Kconfig
> @@ -794,6 +794,16 @@ config SENSORS_SMSC47M192
> This driver can also be built as a module. If so, the module
> will be called smsc47m192.
>
> +config SENSORS_EMC2103
> + tristate "SMSC EMC2103"
> + depends on I2C
> + help
> + If you say yes here you get support for the temperature
> + and fan sensors of the SMSC EMC2103 chips.
> +
> + This driver can also be built as a module. If so, the module
> + will be called emc2103.
> +
Please add this entry right after SMSC EMC1403, so that the
alphanumeric order is preserved.
> config SENSORS_SMSC47B397
> tristate "SMSC LPC47B397-NC"
> depends on EXPERIMENTAL
> diff --git a/drivers/hwmon/Makefile b/drivers/hwmon/Makefile
> index 2138ceb..bca0d45 100644
> --- a/drivers/hwmon/Makefile
> +++ b/drivers/hwmon/Makefile
> @@ -89,6 +89,7 @@ obj-$(CONFIG_SENSORS_SIS5595) += sis5595.o
> obj-$(CONFIG_SENSORS_SMSC47B397)+= smsc47b397.o
> obj-$(CONFIG_SENSORS_SMSC47M1) += smsc47m1.o
> obj-$(CONFIG_SENSORS_SMSC47M192)+= smsc47m192.o
> +obj-$(CONFIG_SENSORS_EMC2103) += emc2103.o
Here again, please respect the alphanumeric order.
> obj-$(CONFIG_SENSORS_AMC6821) += amc6821.o
> obj-$(CONFIG_SENSORS_THMC50) += thmc50.o
> obj-$(CONFIG_SENSORS_TMP102) += tmp102.o
> diff --git a/drivers/hwmon/emc2103.c b/drivers/hwmon/emc2103.c
> new file mode 100644
> index 0000000..14f1c29
> --- /dev/null
> +++ b/drivers/hwmon/emc2103.c
> @@ -0,0 +1,723 @@
> +/*
> + emc2103.c - Support for SMSC EMC2103
> + Copyright (c) 2010 SMSC
> +
> + This program is free software; you can redistribute it and/or modify
> + it under the terms of the GNU General Public License as published by
> + the Free Software Foundation; either version 2 of the License, or
> + (at your option) any later version.
> +
> + This program is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + GNU General Public License for more details.
> +
> + You should have received a copy of the GNU General Public License
> + along with this program; if not, write to the Free Software
> + Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
> +*/
> +
> +#include <linux/module.h>
> +#include <linux/init.h>
> +#include <linux/slab.h>
> +#include <linux/jiffies.h>
> +#include <linux/i2c.h>
> +#include <linux/hwmon.h>
> +#include <linux/hwmon-sysfs.h>
> +#include <linux/err.h>
> +#include <linux/mutex.h>
> +
> +/* Addresses scanned */
> +static const unsigned short normal_i2c[] = { 0x2E, I2C_CLIENT_END };
> +
> +static u8 REG_TEMP[4] = { 0x00, 0x02, 0x04, 0x06 };
> +static u8 REG_TEMP_MIN[4] = { 0x3c, 0x38, 0x39, 0x3a };
> +static u8 REG_TEMP_MAX[4] = { 0x34, 0x30, 0x31, 0x32 };
All 3 should be const.
> +
> +#define REG_CONF1 (0x20)
> +#define REG_TEMP_MAX_ALARM (0x24)
> +#define REG_TEMP_MIN_ALARM (0x25)
> +#define REG_FAN_CONF1 (0x42)
> +#define REG_FAN_TARGET_LO (0x4c)
> +#define REG_FAN_TARGET_HI (0x4d)
> +#define REG_FAN_TACH_HI (0x4e)
> +#define REG_FAN_TACH_LO (0x4f)
> +#define REG_PRODUCT_ID (0xfd)
> +#define REG_MFG_ID (0xfe)
Parentheses around all these constants aren't needed.
A blank line here would be welcome, as the following value isn't a
register address.
> +#define FAN_RPM_FACTOR (3932160)
This constant is quite obscure. Can it be written as an expression
the reader can make sense of?
> +
> +static int apd = 1;
> +module_param(apd, bool, 0);
> +MODULE_PARM_DESC(init, "Set to zero to disable anti-parallel diode mode");
> +
> +struct temperature {
> + s8 degrees;
> + u8 fraction; /* 0-7 multiples of 0.125 */
> +};
This doesn't make much sense. Just use a left-aligned s16. Your code
will be much nicer.
> +
> +struct emc2103_data {
> + struct device *hwmon_dev;
> + struct mutex update_lock;
> + bool valid; /* registers are valid */
> + bool fan_rpm_control;
> + bool have_temp3;
> + bool have_temp4;
> + unsigned long last_updated; /* in jiffies */
> + struct temperature temp[4]; /* internal + 3 external */
> + s8 temp_min[4]; /* no fractional part */
> + s8 temp_max[4]; /* no fractional part */
> + u8 temp_min_alarm;
> + u8 temp_max_alarm;
> + u8 fan_range;
> + u16 fan_tach;
> + u16 fan_target;
> +};
> +
> +static void read_u8_from_i2c(struct i2c_client *client, u8 i2c_reg, u8 *output)
> +{
> + int status = i2c_smbus_read_byte_data(client, i2c_reg);
> + if (status < 0) {
> + dev_dbg(&client->dev, "reg 0x%02x, err %d\n",
> + i2c_reg, status);
This should be reported more loudly. Such errors aren't supposed to
happen! I suggest dev_warn().
> + } else {
> + *output = status;
> + }
> +}
> +
> +static void read_s8_from_i2c(struct i2c_client *client, u8 i2c_reg, s8 *output)
> +{
> + int status = i2c_smbus_read_byte_data(client, i2c_reg);
> + if (status < 0) {
> + dev_dbg(&client->dev, "reg 0x%02x, err %d\n",
> + i2c_reg, status);
> + } else {
> + *output = status;
> + }
> +}
This is really the same function twice. Just cast the parameter as
needed, and you need only one function.
> +
> +static void read_temp_from_i2c(struct i2c_client *client, u8 i2c_reg,
> + struct temperature *temp)
> +{
> + /* read integer part */
> + int status = i2c_smbus_read_byte_data(client, i2c_reg);
> + if (status < 0) {
> + dev_dbg(&client->dev, "reg 0x%02x, err %d\n",
> + i2c_reg, status);
> + } else {
> + temp->degrees = status;
> + }
You're duplicating read_*8_from_i2c().
> +
> + /* Read fractional part from the next register offset */
> + status = i2c_smbus_read_byte_data(client, i2c_reg + 1);
> + if (status < 0) {
> + dev_dbg(&client->dev, "reg 0x%02x, err %d\n",
> + i2c_reg, status);
> + } else {
> + temp->fraction = (status & 0xe0) >> 5;
> + }
And here again.
Did you check if the chip supports i2c_smbus_read_word_data()? This
would be faster. Might even improve reliability if the device doesn't
do register latching (I confess I didn't read the datasheet.)
> +}
> +
> +static void read_fan_from_i2c(struct i2c_client *client, u16 *output,
> + u8 hi_addr, u8 lo_addr)
Always called with hi_addr == lo_addr + 1, so it could be simplified
(would be consistent with what you do for temperatures above.)
> +{
> + u16 high_byte;
> +
> + int status = i2c_smbus_read_byte_data(client, hi_addr);
> + if (status < 0) {
> + dev_dbg(&client->dev, "reg 0x%02x, err %d\n",
> + hi_addr, status);
> + } else {
> + high_byte = status & 0xff;
Pointless masking.
> +
> + status = i2c_smbus_read_byte_data(client, lo_addr);
> + if (status < 0) {
> + dev_dbg(&client->dev, "reg 0x%02x, err %d\n",
> + lo_addr, status);
> + } else {
> + *output = (high_byte << 5) | ((status & 0xf8) >> 3);
> + }
> + }
Here again you're duplicating read_*8_from_i2c().
> +}
> +
> +static void write_fan_target_to_i2c(struct i2c_client *client, u16 new_target)
> +{
> + u8 high_byte = (new_target & 0x1fe0) >> 5;
> + u8 low_byte = (new_target & 0x001f) << 3;
> + i2c_smbus_write_byte_data(client, REG_FAN_TARGET_LO, low_byte);
> + i2c_smbus_write_byte_data(client, REG_FAN_TARGET_HI, high_byte);
> +}
> +
> +static void read_fan_range_from_i2c(struct i2c_client *client, u8 *output)
> +
> +{
> + int status = i2c_smbus_read_byte_data(client, REG_FAN_CONF1);
> + if (status < 0) {
> + dev_dbg(&client->dev, "reg 0x%02x, err %d\n",
> + REG_FAN_CONF1, status);
> + } else {
> + *output = (status & 0x60) >> 5;
> + }
Code duplication again.
> +}
> +
> +static struct emc2103_data *emc2103_update_device(struct device *dev)
> +{
> + struct i2c_client *client = to_i2c_client(dev);
> + struct emc2103_data *data = i2c_get_clientdata(client);
> +
> + mutex_lock(&data->update_lock);
> +
> + if (time_after(jiffies, data->last_updated + HZ + HZ / 2)
> + || !data->valid) {
> + int i;
> +
> + dev_dbg(&client->dev, "Starting emc2103 update\n");
I know that many hwmon drivers have a debug message at this location,
but I con't think it is desirable. Thus function is called so frequently
that it floods the log quickly, and it doesn't add much value.
> +
> + for (i = 0; i < 4; i++) {
> + read_temp_from_i2c(client, REG_TEMP[i], &data->temp[i]);
> + read_s8_from_i2c(client, REG_TEMP_MIN[i],
> + &data->temp_min[i]);
> + read_s8_from_i2c(client, REG_TEMP_MAX[i],
> + &data->temp_max[i]);
You are reading registers which you may not need (if temp3 and/or temp4
aren't present.) This should be avoided, as I2C access can be quite
slow.
> + }
> +
> + read_u8_from_i2c(client, REG_TEMP_MIN_ALARM,
> + &data->temp_min_alarm);
> + read_u8_from_i2c(client, REG_TEMP_MAX_ALARM,
> + &data->temp_max_alarm);
> +
> + read_fan_from_i2c(client, &data->fan_tach,
> + REG_FAN_TACH_HI, REG_FAN_TACH_LO);
> + read_fan_from_i2c(client, &data->fan_target,
> + REG_FAN_TARGET_HI, REG_FAN_TARGET_LO);
> + read_fan_range_from_i2c(client, &data->fan_range);
> +
> + data->last_updated = jiffies;
> + data->valid = true;
> + }
> +
> + mutex_unlock(&data->update_lock);
> +
> + return data;
> +}
> +
> +static ssize_t
> +show_temp(struct device *dev, struct device_attribute *da, char *buf)
> +{
> + struct sensor_device_attribute *attr = to_sensor_dev_attr(da);
I strongly recommend the following syntax for code easier to read:
int nr = to_sensor_dev_attr(da)->index;
> + struct emc2103_data *data = emc2103_update_device(dev);
> + int millidegrees = data->temp[attr->index].degrees * 1000
> + + data->temp[attr->index].fraction * 125;
> + return sprintf(buf, "%d\n", millidegrees);
> +}
> +
> +static ssize_t
> +show_temp_min(struct device *dev, struct device_attribute *da, char *buf)
> +{
> + struct sensor_device_attribute *attr = to_sensor_dev_attr(da);
> + struct emc2103_data *data = emc2103_update_device(dev);
> + int millidegrees = data->temp_min[attr->index] * 1000;
> + return sprintf(buf, "%d\n", millidegrees);
> +}
> +
> +static ssize_t
> +show_temp_max(struct device *dev, struct device_attribute *da, char *buf)
> +{
> + struct sensor_device_attribute *attr = to_sensor_dev_attr(da);
> + struct emc2103_data *data = emc2103_update_device(dev);
> + int millidegrees = data->temp_max[attr->index] * 1000;
> + return sprintf(buf, "%d\n", millidegrees);
> +}
> +
> +static ssize_t
> +show_temp_fault(struct device *dev, struct device_attribute *da, char *buf)
> +{
> + struct sensor_device_attribute *attr = to_sensor_dev_attr(da);
> + struct emc2103_data *data = emc2103_update_device(dev);
> + bool fault = (data->temp[attr->index].degrees == -128);
> + return sprintf(buf, "%d\n", fault ? 1 : 0);
> +}
> +
> +static ssize_t
> +show_temp_min_alarm(struct device *dev, struct device_attribute *da, char *buf)
> +{
> + struct sensor_device_attribute *attr = to_sensor_dev_attr(da);
> + struct emc2103_data *data = emc2103_update_device(dev);
> + bool alarm = data->temp_min_alarm & (1 << attr->index);
> + return sprintf(buf, "%d\n", alarm ? 1 : 0);
> +}
> +
> +static ssize_t
> +show_temp_max_alarm(struct device *dev, struct device_attribute *da, char *buf)
> +{
> + struct sensor_device_attribute *attr = to_sensor_dev_attr(da);
> + struct emc2103_data *data = emc2103_update_device(dev);
> + bool alarm = data->temp_max_alarm & (1 << attr->index);
> + return sprintf(buf, "%d\n", alarm ? 1 : 0);
> +}
> +
> +static ssize_t set_temp_min(struct device *dev, struct device_attribute *da,
> + const char *buf, size_t count)
> +{
> + struct sensor_device_attribute *attr = to_sensor_dev_attr(da);
> + struct emc2103_data *data = emc2103_update_device(dev);
> + struct i2c_client *client = to_i2c_client(dev);
> + long val;
> +
> + int result = strict_strtol(buf, 10, &val);
> + if (result < 0)
> + return -EINVAL;
> +
> + val = val / 1000;
I suggest using DIV_ROUND_CLOSEST().
> + if ((val < -63) || (val > 127))
> + return -EINVAL;
> +
> + mutex_lock(&data->update_lock);
> + data->temp_min[attr->index] = val;
> + i2c_smbus_write_byte_data(client, REG_TEMP_MIN[attr->index], val);
> + mutex_unlock(&data->update_lock);
> +
> + return count;
> +}
> +
> +static ssize_t set_temp_max(struct device *dev, struct device_attribute *da,
> + const char *buf, size_t count)
> +{
> + struct sensor_device_attribute *attr = to_sensor_dev_attr(da);
> + struct emc2103_data *data = emc2103_update_device(dev);
> + struct i2c_client *client = to_i2c_client(dev);
> + long val;
> +
> + int result = strict_strtol(buf, 10, &val);
> + if (result < 0)
> + return -EINVAL;
> +
Missing division by 1000. You didn't test this function :p
> + if ((val < -63) || (val > 127))
> + return -EINVAL;
> +
> + mutex_lock(&data->update_lock);
> + data->temp_max[attr->index] = val;
> + i2c_smbus_write_byte_data(client, REG_TEMP_MAX[attr->index], val);
> + mutex_unlock(&data->update_lock);
> +
> + return count;
> +}
> +
> +static ssize_t
> +show_fan(struct device *dev, struct device_attribute *da, char *buf)
> +{
> + struct emc2103_data *data = emc2103_update_device(dev);
> + int fan_div = 1 << data->fan_range;
> + int rpm = (FAN_RPM_FACTOR * fan_div) / data->fan_tach;
> + return sprintf(buf, "%d\n", rpm);
> +}
> +
> +static ssize_t
> +show_fan_div(struct device *dev, struct device_attribute *da, char *buf)
> +{
> + struct emc2103_data *data = emc2103_update_device(dev);
> + int fan_div = 1 << data->fan_range;
> + return sprintf(buf, "%d\n", fan_div);
> +}
> +
> +/* Note: we also update the fan target here, because its value is
> + determined in part by the fan clock divider. This follows the principle
> + of least surprise; the user doesn't expect the fan target to change just
> + because the divider changed. */
> +static ssize_t set_fan_div(struct device *dev, struct device_attribute *da,
> + const char *buf, size_t count)
> +{
> + struct emc2103_data *data = emc2103_update_device(dev);
> + struct i2c_client *client = to_i2c_client(dev);
> + int status, old_div = 1 << data->fan_range;
> + long new_div;
> +
> + int result = strict_strtol(buf, 10, &new_div);
> + if (result < 0)
> + return -EINVAL;
> +
> + if (new_div == old_div) /* No change */
> + return count;
> +
> + mutex_lock(&data->update_lock);
> + switch (new_div) {
> + case 1:
> + data->fan_range = 0;
> + break;
> + case 2:
> + data->fan_range = 1;
> + break;
> + case 4:
> + data->fan_range = 2;
> + break;
> + case 8:
> + data->fan_range = 3;
> + break;
> + default:
> + mutex_unlock(&data->update_lock);
> + return -EINVAL;
> + }
> +
> + status = i2c_smbus_read_byte_data(client, REG_FAN_CONF1);
> + if (status < 0) {
> + dev_dbg(&client->dev, "reg 0x%02x, err %d\n",
> + REG_FAN_CONF1, status);
> + mutex_unlock(&data->update_lock);
> + return -EIO;
> + }
> + status &= 0x9F;
> + status |= (data->fan_range << 5) & 0x60;
Mask is superfluous, will always be a no-op by construction.
> + i2c_smbus_write_byte_data(client, REG_FAN_CONF1, status);
> +
> + /* update fan target if high word is not disabled */
You mean high byte, not word.
> + if ((data->fan_target & 0x1fe0) != 0x1fe0) {
> + data->fan_target = (data->fan_target * new_div) / old_div;
This might overflow. You must clamp to 0x1fff.
> + write_fan_target_to_i2c(client, data->fan_target);
> + }
> +
> + /* invalidate data to force re-read from hardware */
> + data->valid = false;
> +
> + mutex_unlock(&data->update_lock);
> + return count;
> +}
> +
> +static ssize_t
> +show_fan_target(struct device *dev, struct device_attribute *da, char *buf)
> +{
> + struct emc2103_data *data = emc2103_update_device(dev);
> + int fan_div = 1 << data->fan_range;
> + int rpm = (FAN_RPM_FACTOR * fan_div) / data->fan_target;
data->fan_target is read from the device, it could be 0, and then you
just crashed your kernel.
> +
> + /* high byte of 0xff = disabled so return 0 */
> + if ((data->fan_target & 0x1fe0) == 0x1fe0)
> + rpm = 0;
> +
> + return sprintf(buf, "%d\n", rpm);
> +}
> +
> +static ssize_t set_fan_target(struct device *dev, struct device_attribute *da,
> + const char *buf, size_t count)
> +{
> + struct emc2103_data *data = emc2103_update_device(dev);
> + struct i2c_client *client = to_i2c_client(dev);
> + long rpm_target;
> + int fan_div = 1 << data->fan_range;
> +
> + int result = strict_strtol(buf, 10, &rpm_target);
> + if (result < 0)
> + return -EINVAL;
> +
> + if ((rpm_target < 0) || (rpm_target > 16384))
Odd arbitrary limit. Rationale please? The chip can handle speeds
faster than this.
> + return -EINVAL;
> +
> + mutex_lock(&data->update_lock);
> +
> + if (rpm_target == 0)
> + data->fan_target = 0x1fff;
> + else
> + data->fan_target =
> + ((FAN_RPM_FACTOR * fan_div) / rpm_target) & 0x1fff;
Masking isn't OK. Either the result can be off-limit and you have to
clamp it (SENSORS_LIMIT() may help) or it can't and the masking isn't
needed.
> +
> + write_fan_target_to_i2c(client, data->fan_target);
> +
> + if (!data->fan_rpm_control) {
> + /* RPM control mode is not currently enabled */
> + int status = i2c_smbus_read_byte_data(client, REG_FAN_CONF1);
> + if (status < 0) {
> + dev_dbg(&client->dev, "reg 0x%02x, err %d\n",
> + REG_FAN_CONF1, status);
> + mutex_unlock(&data->update_lock);
> + return -EIO;
> + }
> + status |= 0x80;
> + i2c_smbus_write_byte_data(client, REG_FAN_CONF1, status);
This isn't OK. Fan speed control should be enabled (and disabled! -
your driver currently doesn't even allow it) explicitly by the user.
This should be controlled by file pwm1_enable: 0 for no control and 3
for target speed mode.
> +
> + data->fan_rpm_control = true;
> + }
> +
> + mutex_unlock(&data->update_lock);
> + return count;
> +}
> +
> +static ssize_t
> +show_fan_fault(struct device *dev, struct device_attribute *da, char *buf)
> +{
> + struct emc2103_data *data = emc2103_update_device(dev);
> + bool fault = ((data->fan_tach & 0x1fe0) == 0x1fe0);
> + return sprintf(buf, "%d\n", fault ? 1 : 0);
> +}
> +
> +static SENSOR_DEVICE_ATTR(temp1_input, S_IRUGO, show_temp, NULL, 0);
> +static SENSOR_DEVICE_ATTR(temp1_min, S_IRUGO | S_IWUSR, show_temp_min,
> + set_temp_min, 0);
> +static SENSOR_DEVICE_ATTR(temp1_max, S_IRUGO | S_IWUSR, show_temp_max,
Tab should be space.
> + set_temp_max, 0);
> +static SENSOR_DEVICE_ATTR(temp1_fault, S_IRUGO, show_temp_fault, NULL, 0);
> +static SENSOR_DEVICE_ATTR(temp1_min_alarm, S_IRUGO, show_temp_min_alarm,
> + NULL, 0);
> +static SENSOR_DEVICE_ATTR(temp1_max_alarm, S_IRUGO, show_temp_max_alarm,
> + NULL, 0);
> +
> +static SENSOR_DEVICE_ATTR(temp2_input, S_IRUGO, show_temp, NULL, 1);
> +static SENSOR_DEVICE_ATTR(temp2_min, S_IRUGO | S_IWUSR, show_temp_min,
> + set_temp_min, 1);
> +static SENSOR_DEVICE_ATTR(temp2_max, S_IRUGO | S_IWUSR, show_temp_max,
Tab should be space, and again 3 times below.
> + set_temp_max, 1);
> +static SENSOR_DEVICE_ATTR(temp2_fault, S_IRUGO, show_temp_fault, NULL, 1);
> +static SENSOR_DEVICE_ATTR(temp2_min_alarm, S_IRUGO, show_temp_min_alarm,
> + NULL, 1);
> +static SENSOR_DEVICE_ATTR(temp2_max_alarm, S_IRUGO, show_temp_max_alarm,
> + NULL, 1);
> +
> +static SENSOR_DEVICE_ATTR(temp3_input, S_IRUGO, show_temp, NULL, 2);
> +static SENSOR_DEVICE_ATTR(temp3_min, S_IRUGO | S_IWUSR, show_temp_min,
> + set_temp_min, 2);
> +static SENSOR_DEVICE_ATTR(temp3_max, S_IRUGO | S_IWUSR, show_temp_max,
> + set_temp_max, 2);
> +static SENSOR_DEVICE_ATTR(temp3_fault, S_IRUGO, show_temp_fault, NULL, 2);
> +static SENSOR_DEVICE_ATTR(temp3_min_alarm, S_IRUGO, show_temp_min_alarm,
> + NULL, 2);
> +static SENSOR_DEVICE_ATTR(temp3_max_alarm, S_IRUGO, show_temp_max_alarm,
> + NULL, 2);
> +
> +static SENSOR_DEVICE_ATTR(temp4_input, S_IRUGO, show_temp, NULL, 3);
> +static SENSOR_DEVICE_ATTR(temp4_min, S_IRUGO | S_IWUSR, show_temp_min,
> + set_temp_min, 3);
> +static SENSOR_DEVICE_ATTR(temp4_max, S_IRUGO | S_IWUSR, show_temp_max,
> + set_temp_max, 3);
> +static SENSOR_DEVICE_ATTR(temp4_fault, S_IRUGO, show_temp_fault, NULL, 3);
> +static SENSOR_DEVICE_ATTR(temp4_min_alarm, S_IRUGO, show_temp_min_alarm,
> + NULL, 3);
> +static SENSOR_DEVICE_ATTR(temp4_max_alarm, S_IRUGO, show_temp_max_alarm,
> + NULL, 3);
> +
> +static SENSOR_DEVICE_ATTR(fan1_input, S_IRUGO, show_fan, NULL, 0);
> +static SENSOR_DEVICE_ATTR(fan1_div, S_IRUGO | S_IWUSR, show_fan_div,
> + set_fan_div, 0);
> +static SENSOR_DEVICE_ATTR(fan1_target, S_IRUGO | S_IWUSR, show_fan_target,
> + set_fan_target, 0);
> +static SENSOR_DEVICE_ATTR(fan1_fault, S_IRUGO, show_fan_fault, NULL, 0);
DEVICE_ATTR() would do for all fan attributes, and is cheaper.
> +
> +/* sensors present on all models */
> +static struct attribute *emc2103_attributes[] = {
> + &sensor_dev_attr_temp1_input.dev_attr.attr,
> + &sensor_dev_attr_temp1_min.dev_attr.attr,
> + &sensor_dev_attr_temp1_max.dev_attr.attr,
> + &sensor_dev_attr_temp1_fault.dev_attr.attr,
> + &sensor_dev_attr_temp1_min_alarm.dev_attr.attr,
> + &sensor_dev_attr_temp1_max_alarm.dev_attr.attr,
> + &sensor_dev_attr_temp2_input.dev_attr.attr,
> + &sensor_dev_attr_temp2_min.dev_attr.attr,
> + &sensor_dev_attr_temp2_max.dev_attr.attr,
> + &sensor_dev_attr_temp2_fault.dev_attr.attr,
> + &sensor_dev_attr_temp2_min_alarm.dev_attr.attr,
> + &sensor_dev_attr_temp2_max_alarm.dev_attr.attr,
> + &sensor_dev_attr_fan1_input.dev_attr.attr,
> + &sensor_dev_attr_fan1_div.dev_attr.attr,
> + &sensor_dev_attr_fan1_target.dev_attr.attr,
> + &sensor_dev_attr_fan1_fault.dev_attr.attr,
> + NULL
> +};
> +
> +/* extra temperature sensors only present on 2103-2 and 2103-4 */
> +static struct attribute *emc2103_attributes_temp3[] = {
> + &sensor_dev_attr_temp3_input.dev_attr.attr,
> + &sensor_dev_attr_temp3_min.dev_attr.attr,
> + &sensor_dev_attr_temp3_max.dev_attr.attr,
> + &sensor_dev_attr_temp3_fault.dev_attr.attr,
> + &sensor_dev_attr_temp3_min_alarm.dev_attr.attr,
> + &sensor_dev_attr_temp3_max_alarm.dev_attr.attr,
> + NULL
> +};
> +
> +/* extra temperature sensors only present on 2103-2 and 2103-4 in APD mode */
> +static struct attribute *emc2103_attributes_temp4[] = {
> + &sensor_dev_attr_temp4_input.dev_attr.attr,
> + &sensor_dev_attr_temp4_min.dev_attr.attr,
> + &sensor_dev_attr_temp4_max.dev_attr.attr,
> + &sensor_dev_attr_temp4_fault.dev_attr.attr,
> + &sensor_dev_attr_temp4_min_alarm.dev_attr.attr,
> + &sensor_dev_attr_temp4_max_alarm.dev_attr.attr,
> + NULL
> +};
> +
> +static const struct attribute_group emc2103_group = {
> + .attrs = emc2103_attributes,
> +};
> +
> +static const struct attribute_group emc2103_temp3_group = {
> + .attrs = emc2103_attributes_temp3,
> +};
> +
> +static const struct attribute_group emc2103_temp4_group = {
> + .attrs = emc2103_attributes_temp4,
> +};
> +
> +static int
> +emc2103_probe(struct i2c_client *client, const struct i2c_device_id *id)
> +{
> + struct emc2103_data *data;
> + int status;
> +
> + if (!i2c_check_functionality(client->adapter,
> + I2C_FUNC_SMBUS_BYTE_DATA | I2C_FUNC_SMBUS_WORD_DATA))
At the moment, your driver doesn't use word reads nor writes.
> + return -EIO;
> +
> + data = kzalloc(sizeof(struct emc2103_data), GFP_KERNEL);
> + if (!data)
> + return -ENOMEM;
> +
> + i2c_set_clientdata(client, data);
> + mutex_init(&data->update_lock);
> +
> + /* 2103-2 and 2103-4 have 3 external diodes, 2103-1 has 1 */
> + status = i2c_smbus_read_byte_data(client, REG_PRODUCT_ID);
> + data->have_temp3 = (status == 0x26);
> +
> + /* check if the device is already in RPM control mode */
> + status = i2c_smbus_read_byte_data(client, REG_FAN_CONF1);
> + if (status < 0) {
> + dev_dbg(&client->dev, "reg 0x%02x, err %d\n",
> + REG_FAN_CONF1, status);
> + return -EIO;
Memory leak!
> + }
> + data->fan_rpm_control = (status & 0x80);
As said before, this piece of code doesn't belong there.
> +
> + if (data->have_temp3) {
> + /* Configure anti-parallel diode mode */
> + data->have_temp4 = apd;
> + status = i2c_smbus_read_byte_data(client, REG_CONF1);
> + if (status < 0) {
> + dev_dbg(&client->dev, "reg 0x%02x, err %d\n", REG_CONF1,
> + status);
> + return -EIO;
Memory leak!
> + }
> +
> + if (data->have_temp4)
> + status |= 0x01;
> + else
> + status &= ~(0x01);
> +
> + i2c_smbus_write_byte_data(client, REG_CONF1, status & 0xff);
The default should be dictated by the chip itself, not the driver. By
default, the driver should simply follow the current chip configuration.
> + }
> +
> + /* Register sysfs hooks */
> + status = sysfs_create_group(&client->dev.kobj, &emc2103_group);
> + if (status)
> + goto exit_free;
> +
> + if (data->have_temp3) {
> + status = sysfs_create_group(&client->dev.kobj,
> + &emc2103_temp3_group);
> + if (status)
> + goto exit_remove;
> + }
> +
> + if (data->have_temp4) {
> + status = sysfs_create_group(&client->dev.kobj,
> + &emc2103_temp4_group);
> + if (status)
> + goto exit_remove_temp3;
> + }
> +
> + data->hwmon_dev = hwmon_device_register(&client->dev);
> + if (IS_ERR(data->hwmon_dev)) {
> + status = PTR_ERR(data->hwmon_dev);
> + goto exit_remove_temp4;
> + }
> +
> + dev_info(&client->dev, "%s: sensor '%s'\n",
> + dev_name(data->hwmon_dev), client->name);
> +
> + return 0;
> +
> +exit_remove_temp4:
> + if (data->have_temp4)
> + sysfs_remove_group(&client->dev.kobj, &emc2103_temp4_group);
> +exit_remove_temp3:
> + if (data->have_temp3)
> + sysfs_remove_group(&client->dev.kobj, &emc2103_temp3_group);
> +exit_remove:
> + sysfs_remove_group(&client->dev.kobj, &emc2103_group);
> +exit_free:
> + i2c_set_clientdata(client, NULL);
No longer needed, i2c-core takes care now.
> + kfree(data);
> + return status;
> +}
> +
> +static int emc2103_remove(struct i2c_client *client)
> +{
> + struct emc2103_data *data = i2c_get_clientdata(client);
> +
> + hwmon_device_unregister(data->hwmon_dev);
> +
> + if (data->have_temp4)
> + sysfs_remove_group(&client->dev.kobj, &emc2103_temp4_group);
> +
> + if (data->have_temp3)
> + sysfs_remove_group(&client->dev.kobj, &emc2103_temp3_group);
> +
> + sysfs_remove_group(&client->dev.kobj, &emc2103_group);
> +
> + i2c_set_clientdata(client, NULL);
No longer needed.
> + kfree(data);
> + return 0;
> +}
> +
> +static const struct i2c_device_id emc2103_ids[] = {
> + { "emc2103", 0, },
> + { /* LIST END */ }
> +};
> +MODULE_DEVICE_TABLE(i2c, emc2103_ids);
> +
> +/* Return 0 if detection is successful, -ENODEV otherwise */
> +static int
> +emc2103_detect(struct i2c_client *new_client, struct i2c_board_info *info)
> +{
> + struct i2c_adapter *adapter = new_client->adapter;
> + int manufacturer, product;
> +
> + if (!i2c_check_functionality(adapter, I2C_FUNC_SMBUS_BYTE_DATA))
> + return -ENODEV;
> +
> + manufacturer = i2c_smbus_read_byte_data(new_client, REG_MFG_ID);
> + if (manufacturer != 0x5D)
> + return -ENODEV;
> +
> + product = i2c_smbus_read_byte_data(new_client, REG_PRODUCT_ID);
> + if ((product != 0x24) && (product != 0x26))
> + return -ENODEV;
> +
> + strlcpy(info->type, "emc2103", I2C_NAME_SIZE);
> +
> + return 0;
> +}
> +
> +static struct i2c_driver emc2103_driver = {
> + .class = I2C_CLASS_HWMON,
> + .driver = {
> + .name = "emc2103",
> + },
> + .probe = emc2103_probe,
> + .remove = emc2103_remove,
> + .id_table = emc2103_ids,
> + .detect = emc2103_detect,
> + .address_list = normal_i2c,
> +};
> +
> +static int __init sensors_emc2103_init(void)
> +{
> + return i2c_add_driver(&emc2103_driver);
> +}
> +
> +static void __exit sensors_emc2103_exit(void)
> +{
> + i2c_del_driver(&emc2103_driver);
> +}
> +
> +MODULE_AUTHOR("Steve Glendinning <steve.glendinning at smsc.com>");
> +MODULE_DESCRIPTION("SMSC EMCxxxx hwmon driver");
EMC2103
> +MODULE_LICENSE("GPL");
> +
> +module_init(sensors_emc2103_init);
> +module_exit(sensors_emc2103_exit);
Overall it's rather good, especially for a first submission :) Please
address my comments and resubmit, and your driver should be almost
ready for upstream.
Thanks,
--
Jean Delvare
More information about the lm-sensors
mailing list