cachepc-linux

Fork of AMDESE/linux with modifications for CachePC side-channel attack
git clone https://git.sinitax.com/sinitax/cachepc-linux
Log | Files | Refs | README | LICENSE | sfeed.txt

acpi_power_meter.rst (2194B)


      1Kernel driver power_meter
      2=========================
      3
      4This driver talks to ACPI 4.0 power meters.
      5
      6Supported systems:
      7
      8  * Any recent system with ACPI 4.0.
      9
     10    Prefix: 'power_meter'
     11
     12    Datasheet: https://uefi.org/specifications, section 10.4.
     13
     14Author: Darrick J. Wong
     15
     16Description
     17-----------
     18
     19This driver implements sensor reading support for the power meters exposed in
     20the ACPI 4.0 spec (Chapter 10.4).  These devices have a simple set of
     21features--a power meter that returns average power use over a configurable
     22interval, an optional capping mechanism, and a couple of trip points.  The
     23sysfs interface conforms with the specification outlined in the "Power" section
     24of Documentation/hwmon/sysfs-interface.rst.
     25
     26Special Features
     27----------------
     28
     29The `power[1-*]_is_battery` knob indicates if the power supply is a battery.
     30Both `power[1-*]_average_{min,max}` must be set before the trip points will work.
     31When both of them are set, an ACPI event will be broadcast on the ACPI netlink
     32socket and a poll notification will be sent to the appropriate
     33`power[1-*]_average` sysfs file.
     34
     35The `power[1-*]_{model_number, serial_number, oem_info}` fields display
     36arbitrary strings that ACPI provides with the meter.  The measures/ directory
     37contains symlinks to the devices that this meter measures.
     38
     39Some computers have the ability to enforce a power cap in hardware.  If this is
     40the case, the `power[1-*]_cap` and related sysfs files will appear.  When the
     41average power consumption exceeds the cap, an ACPI event will be broadcast on
     42the netlink event socket and a poll notification will be sent to the
     43appropriate `power[1-*]_alarm` file to indicate that capping has begun, and the
     44hardware has taken action to reduce power consumption.  Most likely this will
     45result in reduced performance.
     46
     47There are a few other ACPI notifications that can be sent by the firmware.  In
     48all cases the ACPI event will be broadcast on the ACPI netlink event socket as
     49well as sent as a poll notification to a sysfs file.  The events are as
     50follows:
     51
     52`power[1-*]_cap` will be notified if the firmware changes the power cap.
     53`power[1-*]_interval` will be notified if the firmware changes the averaging
     54interval.