R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 1 | Kernel driver it87 |
| 2 | ================== |
| 3 | |
| 4 | Supported chips: |
| 5 | * IT8705F |
| 6 | Prefix: 'it87' |
Jean Delvare | 9174999 | 2005-10-08 00:10:00 +0200 | [diff] [blame] | 7 | Addresses scanned: from Super I/O config space (8 I/O ports) |
Jean Delvare | 404a552 | 2010-03-05 22:17:21 +0100 | [diff] [blame] | 8 | Datasheet: Once publicly available at the ITE website, but no longer |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 9 | * IT8712F |
| 10 | Prefix: 'it8712' |
Jean Delvare | 8e9afcb | 2006-12-12 18:18:28 +0100 | [diff] [blame] | 11 | Addresses scanned: from Super I/O config space (8 I/O ports) |
Jean Delvare | 404a552 | 2010-03-05 22:17:21 +0100 | [diff] [blame] | 12 | Datasheet: Once publicly available at the ITE website, but no longer |
Rudolf Marek | 08a8f6e | 2007-06-09 10:11:16 -0400 | [diff] [blame] | 13 | * IT8716F/IT8726F |
Jean Delvare | 17d648b | 2006-08-28 14:23:46 +0200 | [diff] [blame] | 14 | Prefix: 'it8716' |
| 15 | Addresses scanned: from Super I/O config space (8 I/O ports) |
Jean Delvare | 404a552 | 2010-03-05 22:17:21 +0100 | [diff] [blame] | 16 | Datasheet: Once publicly available at the ITE website, but no longer |
Jean Delvare | 87673dd | 2006-08-28 14:37:19 +0200 | [diff] [blame] | 17 | * IT8718F |
| 18 | Prefix: 'it8718' |
| 19 | Addresses scanned: from Super I/O config space (8 I/O ports) |
Jean Delvare | 404a552 | 2010-03-05 22:17:21 +0100 | [diff] [blame] | 20 | Datasheet: Once publicly available at the ITE website, but no longer |
Jean-Marc Spaggiari | b4da93e | 2009-01-07 16:37:32 +0100 | [diff] [blame] | 21 | * IT8720F |
| 22 | Prefix: 'it8720' |
| 23 | Addresses scanned: from Super I/O config space (8 I/O ports) |
Jean Delvare | 404a552 | 2010-03-05 22:17:21 +0100 | [diff] [blame] | 24 | Datasheet: Not publicly available |
Jean Delvare | 44c1bcd | 2010-10-28 20:31:51 +0200 | [diff] [blame] | 25 | * IT8721F/IT8758E |
| 26 | Prefix: 'it8721' |
| 27 | Addresses scanned: from Super I/O config space (8 I/O ports) |
| 28 | Datasheet: Not publicly available |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 29 | * SiS950 [clone of IT8705F] |
Jean Delvare | 9174999 | 2005-10-08 00:10:00 +0200 | [diff] [blame] | 30 | Prefix: 'it87' |
| 31 | Addresses scanned: from Super I/O config space (8 I/O ports) |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 32 | Datasheet: No longer be available |
| 33 | |
Jean Delvare | b19367c | 2006-08-28 14:39:26 +0200 | [diff] [blame] | 34 | Authors: |
Jean Delvare | f1d8e33 | 2007-11-25 16:14:44 +0100 | [diff] [blame] | 35 | Christophe Gauthron |
Jean Delvare | b19367c | 2006-08-28 14:39:26 +0200 | [diff] [blame] | 36 | Jean Delvare <khali@linux-fr.org> |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 37 | |
| 38 | |
| 39 | Module Parameters |
| 40 | ----------------- |
| 41 | |
| 42 | * update_vbat: int |
| 43 | |
| 44 | 0 if vbat should report power on value, 1 if vbat should be updated after |
| 45 | each read. Default is 0. On some boards the battery voltage is provided |
| 46 | by either the battery or the onboard power supply. Only the first reading |
| 47 | at power on will be the actual battery voltage (which the chip does |
| 48 | automatically). On other boards the battery voltage is always fed to |
| 49 | the chip so can be read at any time. Excessive reading may decrease |
| 50 | battery life but no information is given in the datasheet. |
| 51 | |
| 52 | * fix_pwm_polarity int |
| 53 | |
| 54 | Force PWM polarity to active high (DANGEROUS). Some chips are |
| 55 | misconfigured by BIOS - PWM values would be inverted. This option tries |
| 56 | to fix this. Please contact your BIOS manufacturer and ask him for fix. |
| 57 | |
Jean Delvare | 8e9afcb | 2006-12-12 18:18:28 +0100 | [diff] [blame] | 58 | |
| 59 | Hardware Interfaces |
| 60 | ------------------- |
| 61 | |
| 62 | All the chips suported by this driver are LPC Super-I/O chips, accessed |
| 63 | through the LPC bus (ISA-like I/O ports). The IT8712F additionally has an |
| 64 | SMBus interface to the hardware monitoring functions. This driver no |
| 65 | longer supports this interface though, as it is slower and less reliable |
| 66 | than the ISA access, and was only available on a small number of |
| 67 | motherboard models. |
| 68 | |
| 69 | |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 70 | Description |
| 71 | ----------- |
| 72 | |
Jean Delvare | 87673dd | 2006-08-28 14:37:19 +0200 | [diff] [blame] | 73 | This driver implements support for the IT8705F, IT8712F, IT8716F, |
Jean Delvare | 44c1bcd | 2010-10-28 20:31:51 +0200 | [diff] [blame] | 74 | IT8718F, IT8720F, IT8721F, IT8726F, IT8758E and SiS950 chips. |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 75 | |
| 76 | These chips are 'Super I/O chips', supporting floppy disks, infrared ports, |
| 77 | joysticks and other miscellaneous stuff. For hardware monitoring, they |
| 78 | include an 'environment controller' with 3 temperature sensors, 3 fan |
| 79 | rotation speed sensors, 8 voltage sensors, and associated alarms. |
| 80 | |
Jean Delvare | 17d648b | 2006-08-28 14:23:46 +0200 | [diff] [blame] | 81 | The IT8712F and IT8716F additionally feature VID inputs, used to report |
| 82 | the Vcore voltage of the processor. The early IT8712F have 5 VID pins, |
| 83 | the IT8716F and late IT8712F have 6. They are shared with other functions |
| 84 | though, so the functionality may not be available on a given system. |
Jean Delvare | 17d648b | 2006-08-28 14:23:46 +0200 | [diff] [blame] | 85 | |
Jean-Marc Spaggiari | b4da93e | 2009-01-07 16:37:32 +0100 | [diff] [blame] | 86 | The IT8718F and IT8720F also features VID inputs (up to 8 pins) but the value |
| 87 | is stored in the Super-I/O configuration space. Due to technical limitations, |
Jean Delvare | 87673dd | 2006-08-28 14:37:19 +0200 | [diff] [blame] | 88 | this value can currently only be read once at initialization time, so |
| 89 | the driver won't notice and report changes in the VID value. The two |
| 90 | upper VID bits share their pins with voltage inputs (in5 and in6) so you |
| 91 | can't have both on a given board. |
Jean Delvare | 17d648b | 2006-08-28 14:23:46 +0200 | [diff] [blame] | 92 | |
Jean Delvare | 44c1bcd | 2010-10-28 20:31:51 +0200 | [diff] [blame] | 93 | The IT8716F, IT8718F, IT8720F, IT8721F/IT8758E and later IT8712F revisions |
| 94 | have support for 2 additional fans. The additional fans are supported by the |
| 95 | driver. |
Jean Delvare | 87673dd | 2006-08-28 14:37:19 +0200 | [diff] [blame] | 96 | |
Jean Delvare | 44c1bcd | 2010-10-28 20:31:51 +0200 | [diff] [blame] | 97 | The IT8716F, IT8718F, IT8720F and IT8721F/IT8758E, and late IT8712F and |
| 98 | IT8705F also have optional 16-bit tachometer counters for fans 1 to 3. This |
| 99 | is better (no more fan clock divider mess) but not compatible with the older |
| 100 | chips and revisions. The 16-bit tachometer mode is enabled by the driver when |
| 101 | one of the above chips is detected. |
Jean Delvare | 17d648b | 2006-08-28 14:23:46 +0200 | [diff] [blame] | 102 | |
Rudolf Marek | 08a8f6e | 2007-06-09 10:11:16 -0400 | [diff] [blame] | 103 | The IT8726F is just bit enhanced IT8716F with additional hardware |
| 104 | for AMD power sequencing. Therefore the chip will appear as IT8716F |
| 105 | to userspace applications. |
| 106 | |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 107 | Temperatures are measured in degrees Celsius. An alarm is triggered once |
| 108 | when the Overtemperature Shutdown limit is crossed. |
| 109 | |
| 110 | Fan rotation speeds are reported in RPM (rotations per minute). An alarm is |
Jean Delvare | 17d648b | 2006-08-28 14:23:46 +0200 | [diff] [blame] | 111 | triggered if the rotation speed has dropped below a programmable limit. When |
| 112 | 16-bit tachometer counters aren't used, fan readings can be divided by |
| 113 | a programmable divider (1, 2, 4 or 8) to give the readings more range or |
| 114 | accuracy. With a divider of 2, the lowest representable value is around |
| 115 | 2600 RPM. Not all RPM values can accurately be represented, so some rounding |
| 116 | is done. |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 117 | |
| 118 | Voltage sensors (also known as IN sensors) report their values in volts. An |
| 119 | alarm is triggered if the voltage has crossed a programmable minimum or |
| 120 | maximum limit. Note that minimum in this case always means 'closest to |
| 121 | zero'; this is important for negative voltage measurements. All voltage |
| 122 | inputs can measure voltages between 0 and 4.08 volts, with a resolution of |
Jean Delvare | 44c1bcd | 2010-10-28 20:31:51 +0200 | [diff] [blame] | 123 | 0.016 volt (except IT8721F/IT8758E: 0.012 volt.) The battery voltage in8 does |
| 124 | not have limit registers. |
| 125 | |
| 126 | On the IT8721F/IT8758E, some voltage inputs are internal and scaled inside |
| 127 | the chip (in7, in8 and optionally in3). The driver handles this transparently |
| 128 | so user-space doesn't have to care. |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 129 | |
Jean-Marc Spaggiari | b4da93e | 2009-01-07 16:37:32 +0100 | [diff] [blame] | 130 | The VID lines (IT8712F/IT8716F/IT8718F/IT8720F) encode the core voltage value: |
Jean Delvare | 87673dd | 2006-08-28 14:37:19 +0200 | [diff] [blame] | 131 | the voltage level your processor should work with. This is hardcoded by |
| 132 | the mainboard and/or processor itself. It is a value in volts. |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 133 | |
| 134 | If an alarm triggers, it will remain triggered until the hardware register |
| 135 | is read at least once. This means that the cause for the alarm may already |
| 136 | have disappeared! Note that in the current implementation, all hardware |
| 137 | registers are read whenever any data is read (unless it is less than 1.5 |
| 138 | seconds since the last update). This means that you can easily miss |
| 139 | once-only alarms. |
| 140 | |
Jean Delvare | d9b327c | 2010-03-05 22:17:17 +0100 | [diff] [blame] | 141 | Out-of-limit readings can also result in beeping, if the chip is properly |
| 142 | wired and configured. Beeping can be enabled or disabled per sensor type |
| 143 | (temperatures, voltages and fans.) |
| 144 | |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 145 | The IT87xx only updates its values each 1.5 seconds; reading it more often |
| 146 | will do no harm, but will return 'old' values. |
| 147 | |
Jean Delvare | 4ed1077 | 2008-10-17 17:51:16 +0200 | [diff] [blame] | 148 | To change sensor N to a thermistor, 'echo 4 > tempN_type' where N is 1, 2, |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 149 | or 3. To change sensor N to a thermal diode, 'echo 3 > tempN_type'. |
| 150 | Give 0 for unused sensor. Any other value is invalid. To configure this at |
Jean Delvare | 4ed1077 | 2008-10-17 17:51:16 +0200 | [diff] [blame] | 151 | startup, consult lm_sensors's /etc/sensors.conf. (4 = thermistor; |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 152 | 3 = thermal diode) |
| 153 | |
Jean Delvare | f8d0c19 | 2007-02-14 21:15:02 +0100 | [diff] [blame] | 154 | |
| 155 | Fan speed control |
| 156 | ----------------- |
| 157 | |
R.Marek@sh.cvut.cz | 7f15b66 | 2005-05-26 12:42:19 +0000 | [diff] [blame] | 158 | The fan speed control features are limited to manual PWM mode. Automatic |
Jean Delvare | 4f3f51b | 2010-03-05 22:17:21 +0100 | [diff] [blame] | 159 | "Smart Guardian" mode control handling is only implemented for older chips |
| 160 | (see below.) However if you want to go for "manual mode" just write 1 to |
| 161 | pwmN_enable. |
Jean Delvare | f8d0c19 | 2007-02-14 21:15:02 +0100 | [diff] [blame] | 162 | |
| 163 | If you are only able to control the fan speed with very small PWM values, |
| 164 | try lowering the PWM base frequency (pwm1_freq). Depending on the fan, |
| 165 | it may give you a somewhat greater control range. The same frequency is |
| 166 | used to drive all fan outputs, which is why pwm2_freq and pwm3_freq are |
| 167 | read-only. |
Jean Delvare | 4f3f51b | 2010-03-05 22:17:21 +0100 | [diff] [blame] | 168 | |
| 169 | |
| 170 | Automatic fan speed control (old interface) |
| 171 | ------------------------------------------- |
| 172 | |
| 173 | The driver supports the old interface to automatic fan speed control |
| 174 | which is implemented by IT8705F chips up to revision F and IT8712F |
| 175 | chips up to revision G. |
| 176 | |
| 177 | This interface implements 4 temperature vs. PWM output trip points. |
| 178 | The PWM output of trip point 4 is always the maximum value (fan running |
| 179 | at full speed) while the PWM output of the other 3 trip points can be |
| 180 | freely chosen. The temperature of all 4 trip points can be freely chosen. |
| 181 | Additionally, trip point 1 has an hysteresis temperature attached, to |
| 182 | prevent fast switching between fan on and off. |
| 183 | |
| 184 | The chip automatically computes the PWM output value based on the input |
| 185 | temperature, based on this simple rule: if the temperature value is |
| 186 | between trip point N and trip point N+1 then the PWM output value is |
| 187 | the one of trip point N. The automatic control mode is less flexible |
| 188 | than the manual control mode, but it reacts faster, is more robust and |
| 189 | doesn't use CPU cycles. |
Jean Delvare | cccfc9c | 2010-03-05 22:17:21 +0100 | [diff] [blame] | 190 | |
| 191 | Trip points must be set properly before switching to automatic fan speed |
| 192 | control mode. The driver will perform basic integrity checks before |
| 193 | actually switching to automatic control mode. |