1 Naming and data format standards for sysfs files
2 ------------------------------------------------
4 The libsensors library offers an interface to the raw sensors data
5 through the sysfs interface. See libsensors documentation and source for
6 further information. As of writing this document, libsensors
7 (from lm_sensors 2.8.3) is heavily chip-dependent. Adding or updating
8 support for any given chip requires modifying the library's code.
9 This is because libsensors was written for the procfs interface
10 older kernel modules were using, which wasn't standardized enough.
11 Recent versions of libsensors (from lm_sensors 2.8.2 and later) have
12 support for the sysfs interface, though.
14 The new sysfs interface was designed to be as chip-independent as
17 Note that motherboards vary widely in the connections to sensor chips.
18 There is no standard that ensures, for example, that the second
19 temperature sensor is connected to the CPU, or that the second fan is on
20 the CPU. Also, some values reported by the chips need some computation
21 before they make full sense. For example, most chips can only measure
22 voltages between 0 and +4V. Other voltages are scaled back into that
23 range using external resistors. Since the values of these resistors
24 can change from motherboard to motherboard, the conversions cannot be
25 hard coded into the driver and have to be done in user space.
27 For this reason, even if we aim at a chip-independent libsensors, it will
28 still require a configuration file (e.g. /etc/sensors.conf) for proper
29 values conversion, labeling of inputs and hiding of unused inputs.
31 An alternative method that some programs use is to access the sysfs
32 files directly. This document briefly describes the standards that the
33 drivers follow, so that an application program can scan for entries and
34 access this data in a simple and consistent way. That said, such programs
35 will have to implement conversion, labeling and hiding of inputs. For
36 this reason, it is still not recommended to bypass the library.
38 If you are developing a userspace application please send us feedback on
41 Note that this standard isn't completely established yet, so it is subject
42 to changes. If you are writing a new hardware monitoring driver those
43 features can't seem to fit in this interface, please contact us with your
44 extension proposal. Keep in mind that backward compatibility must be
47 Each chip gets its own directory in the sysfs /sys/devices tree. To
48 find all sensor chips, it is easier to follow the device symlinks from
49 /sys/class/hwmon/hwmon*.
51 All sysfs values are fixed point numbers.
53 There is only one value per file, unlike the older /proc specification.
54 The common scheme for files naming is: <type><number>_<item>. Usual
55 types for sensor chips are "in" (voltage), "temp" (temperature) and
56 "fan" (fan). Usual items are "input" (measured value), "max" (high
57 threshold, "min" (low threshold). Numbering usually starts from 1,
58 except for voltages which start from 0 (because most data sheets use
59 this). A number is always used for elements that can be present more
60 than once, even if there is a single element of the given type on the
61 specific chip. Other files do not refer to a specific element, so
62 they have a simple name, and no number.
64 Alarms are direct indications read from the chips. The drivers do NOT
65 make comparisons of readings to thresholds. This allows violations
66 between readings to be caught and alarmed. The exact definition of an
67 alarm (for example, whether a threshold must be met or must be exceeded
68 to cause an alarm) is chip-dependent.
70 When setting values of hwmon sysfs attributes, the string representation of
71 the desired value must be written, note that strings which are not a number
72 are interpreted as 0! For more on how written strings are interpreted see the
73 "sysfs attribute writes interpretation" section at the end of this file.
75 -------------------------------------------------------------------------
77 [0-*] denotes any positive number starting from 0
78 [1-*] denotes any positive number starting from 1
82 Read/write values may be read-only for some chips, depending on the
83 hardware implementation.
85 All entries (except name) are optional, and should only be created in a
86 given driver if the chip has the feature.
94 This should be a short, lowercase string, not containing
95 spaces nor dashes, representing the chip name. This is
96 the only mandatory attribute.
97 I2C devices get this attribute created automatically.
105 in[0-*]_min Voltage min value.
109 in[0-*]_max Voltage max value.
113 in[0-*]_input Voltage input value.
116 Voltage measured on the chip pin.
117 Actual voltage depends on the scaling resistors on the
118 motherboard, as recommended in the chip datasheet.
119 This varies by chip and by motherboard.
120 Because of this variation, values are generally NOT scaled
121 by the chip driver, and must be done by the application.
122 However, some drivers (notably lm87 and via686a)
123 do scale, because of internal resistors built into a chip.
124 These drivers will output the actual voltage. Rule of
125 thumb: drivers should report the voltage values at the
128 in[0-*]_label Suggested voltage channel label.
130 Should only be created if the driver has hints about what
131 this voltage channel is being used for, and user-space
132 doesn't. In all other cases, the label is provided by
136 cpu[0-*]_vid CPU core reference voltage.
141 vrm Voltage Regulator Module version number.
142 RW (but changing it should no more be necessary)
143 Originally the VRM standard version multiplied by 10, but now
144 an arbitrary number, as not all standards have a version
146 Affects the way the driver calculates the CPU core reference
147 voltage from the vid pins.
149 Also see the Alarms section for status flags associated with voltages.
156 fan[1-*]_min Fan minimum value
157 Unit: revolution/min (RPM)
160 fan[1-*]_input Fan input value.
161 Unit: revolution/min (RPM)
164 fan[1-*]_div Fan divisor.
165 Integer value in powers of two (1, 2, 4, 8, 16, 32, 64, 128).
167 Some chips only support values 1, 2, 4 and 8.
168 Note that this is actually an internal clock divisor, which
169 affects the measurable speed range, not the read value.
173 Unit: revolution/min (RPM)
175 Only makes sense if the chip supports closed-loop fan speed
176 control based on the measured fan speed.
178 fan[1-*]_label Suggested fan channel label.
180 Should only be created if the driver has hints about what
181 this fan channel is being used for, and user-space doesn't.
182 In all other cases, the label is provided by user-space.
185 Also see the Alarms section for status flags associated with fans.
192 pwm[1-*] Pulse width modulation fan control.
193 Integer value in the range 0 to 255
198 Fan speed control method:
199 0: no fan speed control (i.e. fan at full speed)
200 1: manual fan speed control enabled (using pwm[1-*])
201 2+: automatic fan speed control enabled
202 Check individual chip documentation files for automatic mode
206 pwm[1-*]_mode 0: DC mode (direct current)
207 1: PWM mode (pulse-width modulation)
210 pwm[1-*]_freq Base PWM frequency in Hz.
211 Only possibly available when pwmN_mode is PWM, but not always
215 pwm[1-*]_auto_channels_temp
216 Select which temperature channels affect this PWM output in
217 auto mode. Bitfield, 1 is temp1, 2 is temp2, 4 is temp3 etc...
218 Which values are possible depend on the chip used.
221 pwm[1-*]_auto_point[1-*]_pwm
222 pwm[1-*]_auto_point[1-*]_temp
223 pwm[1-*]_auto_point[1-*]_temp_hyst
224 Define the PWM vs temperature curve. Number of trip points is
225 chip-dependent. Use this for chips which associate trip points
226 to PWM output channels.
231 temp[1-*]_auto_point[1-*]_pwm
232 temp[1-*]_auto_point[1-*]_temp
233 temp[1-*]_auto_point[1-*]_temp_hyst
234 Define the PWM vs temperature curve. Number of trip points is
235 chip-dependent. Use this for chips which associate trip points
236 to temperature channels.
244 temp[1-*]_type Sensor type selection.
253 Not all types are supported by all chips
255 temp[1-*]_max Temperature max value.
256 Unit: millidegree Celsius (or millivolt, see below)
259 temp[1-*]_min Temperature min value.
260 Unit: millidegree Celsius
264 Temperature hysteresis value for max limit.
265 Unit: millidegree Celsius
266 Must be reported as an absolute temperature, NOT a delta
270 temp[1-*]_input Temperature input value.
271 Unit: millidegree Celsius
274 temp[1-*]_crit Temperature critical value, typically greater than
275 corresponding temp_max values.
276 Unit: millidegree Celsius
280 Temperature hysteresis value for critical limit.
281 Unit: millidegree Celsius
282 Must be reported as an absolute temperature, NOT a delta
283 from the critical value.
287 Temperature offset which is added to the temperature reading
289 Unit: millidegree Celsius
292 temp[1-*]_label Suggested temperature channel label.
294 Should only be created if the driver has hints about what
295 this temperature channel is being used for, and user-space
296 doesn't. In all other cases, the label is provided by
300 Some chips measure temperature using external thermistors and an ADC, and
301 report the temperature measurement as a voltage. Converting this voltage
302 back to a temperature (or the other way around for limits) requires
303 mathematical functions not available in the kernel, so the conversion
304 must occur in user space. For these chips, all temp* files described
305 above should contain values expressed in millivolt instead of millidegree
306 Celsius. In other words, such temperature channels are handled as voltage
307 channels by the driver.
309 Also see the Alarms section for status flags associated with temperatures.
316 Note that no known chip provides current measurements as of writing,
317 so this part is theoretical, so to say.
319 curr[1-*]_max Current max value
323 curr[1-*]_min Current min value.
327 curr[1-*]_input Current input value
336 Each channel or limit may have an associated alarm file, containing a
337 boolean value. 1 means than an alarm condition exists, 0 means no alarm.
339 Usually a given chip will either use channel-related alarms, or
340 limit-related alarms, not both. The driver should just reflect the hardware
364 Each input channel may have an associated fault file. This can be used
365 to notify open diodes, unconnected fans etc. where the hardware
366 supports it. When this boolean has value 1, the measurement for that
367 channel should not be trusted.
372 Input fault condition
377 Some chips also offer the possibility to get beeped when an alarm occurs:
379 beep_enable Master beep enable
392 In theory, a chip could provide per-limit beep masking, but no such chip
395 Old drivers provided a different, non-standard interface to alarms and
396 beeps. These interface files are deprecated, but will be kept around
397 for compatibility reasons:
399 alarms Alarm bitmask.
401 Integer representation of one to four bytes.
402 A '1' bit means an alarm.
403 Chips should be programmed for 'comparator' mode so that
404 the alarm will 'come back' after you read the register
405 if it is still valid.
406 Generally a direct representation of a chip's internal
407 alarm registers; there is no standard for the position
408 of individual bits. For this reason, the use of this
409 interface file for new drivers is discouraged. Use
410 individual *_alarm and *_fault files instead.
411 Bits are defined in kernel/include/sensors.h.
413 beep_mask Bitmask for beep.
414 Same format as 'alarms' with the same bit locations,
415 use discouraged for the same reason. Use individual
416 *_beep files instead.
420 sysfs attribute writes interpretation
421 -------------------------------------
423 hwmon sysfs attributes always contain numbers, so the first thing to do is to
424 convert the input to a number, there are 2 ways todo this depending whether
425 the number can be negative or not:
426 unsigned long u = simple_strtoul(buf, NULL, 10);
427 long s = simple_strtol(buf, NULL, 10);
429 With buf being the buffer with the user input being passed by the kernel.
430 Notice that we do not use the second argument of strto[u]l, and thus cannot
431 tell when 0 is returned, if this was really 0 or is caused by invalid input.
432 This is done deliberately as checking this everywhere would add a lot of
435 Notice that it is important to always store the converted value in an
436 unsigned long or long, so that no wrap around can happen before any further
439 After the input string is converted to an (unsigned) long, the value should be
440 checked if its acceptable. Be careful with further conversions on the value
441 before checking it for validity, as these conversions could still cause a wrap
442 around before the check. For example do not multiply the result, and only
443 add/subtract if it has been divided before the add/subtract.
445 What to do if a value is found to be invalid, depends on the type of the
446 sysfs attribute that is being set. If it is a continuous setting like a
447 tempX_max or inX_max attribute, then the value should be clamped to its
448 limits using SENSORS_LIMIT(value, min_limit, max_limit). If it is not
449 continuous like for example a tempX_type, then when an invalid value is
450 written, -EINVAL should be returned.
452 Example1, temp1_max, register is a signed 8 bit value (-128 - 127 degrees):
454 long v = simple_strtol(buf, NULL, 10) / 1000;
455 v = SENSORS_LIMIT(v, -128, 127);
456 /* write v to register */
458 Example2, fan divider setting, valid values 2, 4 and 8:
460 unsigned long v = simple_strtoul(buf, NULL, 10);
463 case 2: v = 1; break;
464 case 4: v = 2; break;
465 case 8: v = 3; break;
469 /* write v to register */