------------------------------------------------
The libsensors library offers an interface to the raw sensors data
-through the sysfs interface. See libsensors documentation and source for
-more further information. As of writing this document, libsensors
-(from lm_sensors 2.8.3) is heavily chip-dependant. Adding or updating
-support for any given chip requires modifying the library's code.
-This is because libsensors was written for the procfs interface
-older kernel modules were using, which wasn't standardized enough.
-Recent versions of libsensors (from lm_sensors 2.8.2 and later) have
-support for the sysfs interface, though.
-
-The new sysfs interface was designed to be as chip-independant as
-possible.
+through the sysfs interface. Since lm-sensors 3.0.0, libsensors is
+completely chip-independent. It assumes that all the kernel drivers
+implement the standard sysfs interface described in this document.
+This makes adding or updating support for any given chip very easy, as
+libsensors, and applications using it, do not need to be modified.
+This is a major improvement compared to lm-sensors 2.
Note that motherboards vary widely in the connections to sensor chips.
There is no standard that ensures, for example, that the second
can change from motherboard to motherboard, the conversions cannot be
hard coded into the driver and have to be done in user space.
-For this reason, even if we aim at a chip-independant libsensors, it will
+For this reason, even if we aim at a chip-independent libsensors, it will
still require a configuration file (e.g. /etc/sensors.conf) for proper
values conversion, labeling of inputs and hiding of unused inputs.
will have to implement conversion, labeling and hiding of inputs. For
this reason, it is still not recommended to bypass the library.
-If you are developing a userspace application please send us feedback on
-this standard.
-
-Note that this standard isn't completely established yet, so it is subject
-to changes, even important ones. One more reason to use the library instead
-of accessing sysfs files directly.
-
Each chip gets its own directory in the sysfs /sys/devices tree. To
-find all sensor chips, it is easier to follow the symlinks from
-/sys/i2c/devices/
+find all sensor chips, it is easier to follow the device symlinks from
+/sys/class/hwmon/hwmon*.
-All sysfs values are fixed point numbers. To get the true value of some
-of the values, you should divide by the specified value.
+Up to lm-sensors 3.0.0, libsensors looks for hardware monitoring attributes
+in the "physical" device directory. Since lm-sensors 3.0.1, attributes found
+in the hwmon "class" device directory are also supported. Complex drivers
+(e.g. drivers for multifunction chips) may want to use this possibility to
+avoid namespace pollution. The only drawback will be that older versions of
+libsensors won't support the driver in question.
+
+All sysfs values are fixed point numbers.
There is only one value per file, unlike the older /proc specification.
The common scheme for files naming is: <type><number>_<item>. Usual
alarm (for example, whether a threshold must be met or must be exceeded
to cause an alarm) is chip-dependent.
+When setting values of hwmon sysfs attributes, the string representation of
+the desired value must be written, note that strings which are not a number
+are interpreted as 0! For more on how written strings are interpreted see the
+"sysfs attribute writes interpretation" section at the end of this file.
-------------------------------------------------------------------------
+[0-*] denotes any positive number starting from 0
+[1-*] denotes any positive number starting from 1
+RO read only value
+WO write only value
+RW read/write value
+
+Read/write values may be read-only for some chips, depending on the
+hardware implementation.
+
+All entries (except name) are optional, and should only be created in a
+given driver if the chip has the feature.
+
+
+********
+* Name *
+********
+
+name The chip name.
+ This should be a short, lowercase string, not containing
+ spaces nor dashes, representing the chip name. This is
+ the only mandatory attribute.
+ I2C devices get this attribute created automatically.
+ RO
+
+
************
* Voltages *
************
-in[0-8]_min Voltage min value.
+in[0-*]_min Voltage min value.
Unit: millivolt
- Read/Write
+ RW
-in[0-8]_max Voltage max value.
+in[0-*]_max Voltage max value.
Unit: millivolt
- Read/Write
+ RW
-in[0-8]_input Voltage input value.
+in[0-*]_input Voltage input value.
Unit: millivolt
- Read only
+ RO
+ Voltage measured on the chip pin.
Actual voltage depends on the scaling resistors on the
motherboard, as recommended in the chip datasheet.
This varies by chip and by motherboard.
Because of this variation, values are generally NOT scaled
by the chip driver, and must be done by the application.
However, some drivers (notably lm87 and via686a)
- do scale, with various degrees of success.
- These drivers will output the actual voltage.
-
- Typical usage:
- in0_* CPU #1 voltage (not scaled)
- in1_* CPU #2 voltage (not scaled)
- in2_* 3.3V nominal (not scaled)
- in3_* 5.0V nominal (scaled)
- in4_* 12.0V nominal (scaled)
- in5_* -12.0V nominal (scaled)
- in6_* -5.0V nominal (scaled)
- in7_* varies
- in8_* varies
-
-cpu[0-1]_vid CPU core reference voltage.
+ do scale, because of internal resistors built into a chip.
+ These drivers will output the actual voltage. Rule of
+ thumb: drivers should report the voltage values at the
+ "pins" of the chip.
+
+in[0-*]_label Suggested voltage channel label.
+ Text string
+ Should only be created if the driver has hints about what
+ this voltage channel is being used for, and user-space
+ doesn't. In all other cases, the label is provided by
+ user-space.
+ RO
+
+cpu[0-*]_vid CPU core reference voltage.
Unit: millivolt
- Read only.
+ RO
Not always correct.
vrm Voltage Regulator Module version number.
- Read only.
- Two digit number, first is major version, second is
- minor version.
+ RW (but changing it should no more be necessary)
+ Originally the VRM standard version multiplied by 10, but now
+ an arbitrary number, as not all standards have a version
+ number.
Affects the way the driver calculates the CPU core reference
voltage from the vid pins.
+Also see the Alarms section for status flags associated with voltages.
+
********
* Fans *
********
-fan[1-3]_min Fan minimum value
+fan[1-*]_min Fan minimum value
+ Unit: revolution/min (RPM)
+ RW
+
+fan[1-*]_max Fan maximum value
Unit: revolution/min (RPM)
- Read/Write.
+ Only rarely supported by the hardware.
+ RW
-fan[1-3]_input Fan input value.
+fan[1-*]_input Fan input value.
Unit: revolution/min (RPM)
- Read only.
+ RO
-fan[1-3]_div Fan divisor.
+fan[1-*]_div Fan divisor.
Integer value in powers of two (1, 2, 4, 8, 16, 32, 64, 128).
+ RW
Some chips only support values 1, 2, 4 and 8.
Note that this is actually an internal clock divisor, which
affects the measurable speed range, not the read value.
+fan[1-*]_target
+ Desired fan speed
+ Unit: revolution/min (RPM)
+ RW
+ Only makes sense if the chip supports closed-loop fan speed
+ control based on the measured fan speed.
+
+fan[1-*]_label Suggested fan channel label.
+ Text string
+ Should only be created if the driver has hints about what
+ this fan channel is being used for, and user-space doesn't.
+ In all other cases, the label is provided by user-space.
+ RO
+
+Also see the Alarms section for status flags associated with fans.
+
+
*******
* PWM *
*******
-pwm[1-3] Pulse width modulation fan control.
+pwm[1-*] Pulse width modulation fan control.
Integer value in the range 0 to 255
- Read/Write
+ RW
255 is max or 100%.
-pwm[1-3]_enable
- Switch PWM on and off.
- Not always present even if fan*_pwm is.
- 0 to turn off
- 1 to turn on in manual mode
- 2 to turn on in automatic mode
- Read/Write
+pwm[1-*]_enable
+ Fan speed control method:
+ 0: no fan speed control (i.e. fan at full speed)
+ 1: manual fan speed control enabled (using pwm[1-*])
+ 2+: automatic fan speed control enabled
+ Check individual chip documentation files for automatic mode
+ details.
+ RW
+
+pwm[1-*]_mode 0: DC mode (direct current)
+ 1: PWM mode (pulse-width modulation)
+ RW
+
+pwm[1-*]_freq Base PWM frequency in Hz.
+ Only possibly available when pwmN_mode is PWM, but not always
+ present even then.
+ RW
pwm[1-*]_auto_channels_temp
Select which temperature channels affect this PWM output in
auto mode. Bitfield, 1 is temp1, 2 is temp2, 4 is temp3 etc...
Which values are possible depend on the chip used.
+ RW
pwm[1-*]_auto_point[1-*]_pwm
pwm[1-*]_auto_point[1-*]_temp
Define the PWM vs temperature curve. Number of trip points is
chip-dependent. Use this for chips which associate trip points
to PWM output channels.
+ RW
OR
Define the PWM vs temperature curve. Number of trip points is
chip-dependent. Use this for chips which associate trip points
to temperature channels.
+ RW
****************
* Temperatures *
****************
-temp[1-3]_type Sensor type selection.
- Integers 1 to 4 or thermistor Beta value (typically 3435)
- Read/Write.
+temp[1-*]_type Sensor type selection.
+ Integers 1 to 6
+ RW
1: PII/Celeron Diode
2: 3904 transistor
3: thermal diode
- 4: thermistor (default/unknown Beta)
+ 4: thermistor
+ 5: AMD AMDSI
+ 6: Intel PECI
Not all types are supported by all chips
-temp[1-4]_max Temperature max value.
- Unit: millidegree Celcius
- Read/Write value.
+temp[1-*]_max Temperature max value.
+ Unit: millidegree Celsius (or millivolt, see below)
+ RW
-temp[1-3]_min Temperature min value.
- Unit: millidegree Celcius
- Read/Write value.
+temp[1-*]_min Temperature min value.
+ Unit: millidegree Celsius
+ RW
-temp[1-3]_max_hyst
+temp[1-*]_max_hyst
Temperature hysteresis value for max limit.
- Unit: millidegree Celcius
+ Unit: millidegree Celsius
Must be reported as an absolute temperature, NOT a delta
from the max value.
- Read/Write value.
+ RW
-temp[1-4]_input Temperature input value.
- Unit: millidegree Celcius
- Read only value.
+temp[1-*]_input Temperature input value.
+ Unit: millidegree Celsius
+ RO
-temp[1-4]_crit Temperature critical value, typically greater than
+temp[1-*]_crit Temperature critical value, typically greater than
corresponding temp_max values.
- Unit: millidegree Celcius
- Read/Write value.
+ Unit: millidegree Celsius
+ RW
-temp[1-2]_crit_hyst
+temp[1-*]_crit_hyst
Temperature hysteresis value for critical limit.
- Unit: millidegree Celcius
+ Unit: millidegree Celsius
Must be reported as an absolute temperature, NOT a delta
from the critical value.
- Read/Write value.
+ RW
-temp[1-4]_offset
+temp[1-*]_offset
Temperature offset which is added to the temperature reading
by the chip.
Unit: millidegree Celsius
Read/Write value.
- If there are multiple temperature sensors, temp1_* is
- generally the sensor inside the chip itself,
- reported as "motherboard temperature". temp2_* to
- temp4_* are generally sensors external to the chip
- itself, for example the thermal diode inside the CPU or
- a thermistor nearby.
+temp[1-*]_label Suggested temperature channel label.
+ Text string
+ Should only be created if the driver has hints about what
+ this temperature channel is being used for, and user-space
+ doesn't. In all other cases, the label is provided by
+ user-space.
+ RO
+
+temp[1-*]_lowest
+ Historical minimum temperature
+ Unit: millidegree Celsius
+ RO
+
+temp[1-*]_highest
+ Historical maximum temperature
+ Unit: millidegree Celsius
+ RO
+
+temp[1-*]_reset_history
+ Reset temp_lowest and temp_highest
+ WO
+
+temp_reset_history
+ Reset temp_lowest and temp_highest for all sensors
+ WO
+
+Some chips measure temperature using external thermistors and an ADC, and
+report the temperature measurement as a voltage. Converting this voltage
+back to a temperature (or the other way around for limits) requires
+mathematical functions not available in the kernel, so the conversion
+must occur in user space. For these chips, all temp* files described
+above should contain values expressed in millivolt instead of millidegree
+Celsius. In other words, such temperature channels are handled as voltage
+channels by the driver.
+
+Also see the Alarms section for status flags associated with temperatures.
************
Note that no known chip provides current measurements as of writing,
so this part is theoretical, so to say.
-curr[1-n]_max Current max value
+curr[1-*]_max Current max value
Unit: milliampere
- Read/Write.
+ RW
-curr[1-n]_min Current min value.
+curr[1-*]_min Current min value.
Unit: milliampere
- Read/Write.
+ RW
-curr[1-n]_input Current input value
+curr[1-*]_input Current input value
Unit: milliampere
- Read only.
+ RO
+
+*********
+* Power *
+*********
+
+power[1-*]_average Average power use
+ Unit: microWatt
+ RO
+
+power[1-*]_average_interval Power use averaging interval
+ Unit: milliseconds
+ RW
+
+power[1-*]_average_highest Historical average maximum power use
+ Unit: microWatt
+ RO
+
+power[1-*]_average_lowest Historical average minimum power use
+ Unit: microWatt
+ RO
+
+power[1-*]_input Instantaneous power use
+ Unit: microWatt
+ RO
+
+power[1-*]_input_highest Historical maximum power use
+ Unit: microWatt
+ RO
+
+power[1-*]_input_lowest Historical minimum power use
+ Unit: microWatt
+ RO
+
+power[1-*]_reset_history Reset input_highest, input_lowest,
+ average_highest and average_lowest.
+ WO
+
+**********
+* Energy *
+**********
+
+energy[1-*]_input Cumulative energy use
+ Unit: microJoule
+ RO
**********
limit-related alarms, not both. The driver should just reflect the hardware
implementation.
-in[0-n]_alarm
-fan[1-n]_alarm
-temp[1-n]_alarm
+in[0-*]_alarm
+fan[1-*]_alarm
+temp[1-*]_alarm
Channel alarm
- Boolean
- Read-only
+ 0: no alarm
+ 1: alarm
+ RO
OR
-in[0-n]_min_alarm
-in[0-n]_max_alarm
-fan[1-n]_min_alarm
-temp[1-n]_min_alarm
-temp[1-n]_max_alarm
-temp[1-n]_crit_alarm
+in[0-*]_min_alarm
+in[0-*]_max_alarm
+fan[1-*]_min_alarm
+fan[1-*]_max_alarm
+temp[1-*]_min_alarm
+temp[1-*]_max_alarm
+temp[1-*]_crit_alarm
Limit alarm
- Boolean
- Read-only
+ 0: no alarm
+ 1: alarm
+ RO
Each input channel may have an associated fault file. This can be used
to notify open diodes, unconnected fans etc. where the hardware
supports it. When this boolean has value 1, the measurement for that
channel should not be trusted.
-in[0-n]_input_fault
-fan[1-n]_input_fault
-temp[1-n]_input_fault
+in[0-*]_fault
+fan[1-*]_fault
+temp[1-*]_fault
Input fault condition
- Boolean
- Read-only
+ 0: no fault occured
+ 1: fault condition
+ RO
Some chips also offer the possibility to get beeped when an alarm occurs:
beep_enable Master beep enable
- Boolean
- Read/Write
+ 0: no beeps
+ 1: beeps
+ RW
-in[0-n]_beep
-fan[1-n]_beep
-temp[1-n]_beep
+in[0-*]_beep
+fan[1-*]_beep
+temp[1-*]_beep
Channel beep
- 0 to disable.
- 1 to enable.
- Read/write
+ 0: disable
+ 1: enable
+ RW
In theory, a chip could provide per-limit beep masking, but no such chip
was seen so far.
for compatibility reasons:
alarms Alarm bitmask.
- Read only.
+ RO
Integer representation of one to four bytes.
A '1' bit means an alarm.
Chips should be programmed for 'comparator' mode so that
Same format as 'alarms' with the same bit locations,
use discouraged for the same reason. Use individual
*_beep files instead.
- Read/Write
-
-
-*********
-* Other *
-*********
-
-eeprom Raw EEPROM data in binary form.
- Read only.
-
-pec Enable or disable PEC (SMBus only)
- Read/Write
+ RW
+
+
+***********************
+* Intrusion detection *
+***********************
+
+intrusion[0-*]_alarm
+ Chassis intrusion detection
+ 0: OK
+ 1: intrusion detected
+ RW
+ Contrary to regular alarm flags which clear themselves
+ automatically when read, this one sticks until cleared by
+ the user. This is done by writing 0 to the file. Writing
+ other values is unsupported.
+
+intrusion[0-*]_beep
+ Chassis intrusion beep
+ 0: disable
+ 1: enable
+ RW
+
+
+sysfs attribute writes interpretation
+-------------------------------------
+
+hwmon sysfs attributes always contain numbers, so the first thing to do is to
+convert the input to a number, there are 2 ways todo this depending whether
+the number can be negative or not:
+unsigned long u = simple_strtoul(buf, NULL, 10);
+long s = simple_strtol(buf, NULL, 10);
+
+With buf being the buffer with the user input being passed by the kernel.
+Notice that we do not use the second argument of strto[u]l, and thus cannot
+tell when 0 is returned, if this was really 0 or is caused by invalid input.
+This is done deliberately as checking this everywhere would add a lot of
+code to the kernel.
+
+Notice that it is important to always store the converted value in an
+unsigned long or long, so that no wrap around can happen before any further
+checking.
+
+After the input string is converted to an (unsigned) long, the value should be
+checked if its acceptable. Be careful with further conversions on the value
+before checking it for validity, as these conversions could still cause a wrap
+around before the check. For example do not multiply the result, and only
+add/subtract if it has been divided before the add/subtract.
+
+What to do if a value is found to be invalid, depends on the type of the
+sysfs attribute that is being set. If it is a continuous setting like a
+tempX_max or inX_max attribute, then the value should be clamped to its
+limits using SENSORS_LIMIT(value, min_limit, max_limit). If it is not
+continuous like for example a tempX_type, then when an invalid value is
+written, -EINVAL should be returned.
+
+Example1, temp1_max, register is a signed 8 bit value (-128 - 127 degrees):
+
+ long v = simple_strtol(buf, NULL, 10) / 1000;
+ v = SENSORS_LIMIT(v, -128, 127);
+ /* write v to register */
+
+Example2, fan divider setting, valid values 2, 4 and 8:
+
+ unsigned long v = simple_strtoul(buf, NULL, 10);
+
+ switch (v) {
+ case 2: v = 1; break;
+ case 4: v = 2; break;
+ case 8: v = 3; break;
+ default:
+ return -EINVAL;
+ }
+ /* write v to register */