ThinkPad ACPI Extras Driver
Version 0.14
April 21st, 2007
Borislav Deianov <borislav@users.sf.net>
Henrique de Moraes Holschuh <hmh@hmh.eng.br>
http://ibm-acpi.sf.net/
This is a Linux driver for the IBM and Lenovo ThinkPad laptops. It
supports various features of these laptops which are accessible
through the ACPI and ACPI EC framework, but not otherwise fully
supported by the generic Linux ACPI drivers.
This driver used to be named ibm-acpi until kernel 2.6.21 and release
0.13-20070314. It used to be in the drivers/acpi tree, but it was
moved to the drivers/misc tree and renamed to thinkpad-acpi for kernel
2.6.22, and release 0.14.
Status
------
The features currently supported are the following (see below for
detailed description):
- Fn key combinations
- Bluetooth enable and disable
- video output switching, expansion control
- ThinkLight on and off
- limited docking and undocking
- UltraBay eject
- CMOS control
- LED control
- ACPI sounds
- temperature sensors
- Experimental: embedded controller register dump
- LCD brightness control
- Volume control
- Fan control and monitoring: fan speed, fan enable/disable
- Experimental: WAN enable and disable
A compatibility table by model and feature is maintained on the web
site, http://ibm-acpi.sf.net/. I appreciate any success or failure
reports, especially if they add to or correct the compatibility table.
Please include the following information in your report:
- ThinkPad model name
- a copy of your DSDT, from /proc/acpi/dsdt
- a copy of the output of dmidecode, with serial numbers
and UUIDs masked off
- which driver features work and which don't
- the observed behavior of non-working features
Any other comments or patches are also more than welcome.
Installation
------------
If you are compiling this driver as included in the Linux kernel
sources, simply enable the CONFIG_THINKPAD_ACPI option, and optionally
enable the CONFIG_THINKPAD_ACPI_BAY option if you want the
thinkpad-specific bay functionality.
Features
--------
The driver exports two different interfaces to userspace, which can be
used to access the features it provides. One is a legacy procfs-based
interface, which will be removed at some time in the distant future.
The other is a new sysfs-based interface which is not complete yet.
The procfs interface creates the /proc/acpi/ibm directory. There is a
file under that directory for each feature it supports. The procfs
interface is mostly frozen, and will change very little if at all: it
will not be extended to add any new functionality in the driver, instead
all new functionality will be implemented on the sysfs interface.
The sysfs interface tries to blend in the generic Linux sysfs subsystems
and classes as much as possible. Since some of these subsystems are not
yet ready or stabilized, it is expected that this interface will change,
and any and all userspace programs must deal with it.
Notes about the sysfs interface:
Unlike what was done with the procfs interface, correctness when talking
to the sysfs interfaces will be enforced, as will correctness in the
thinkpad-acpi's implementation of sysfs interfaces.
Also, any bugs in the thinkpad-acpi sysfs driver code or in the
thinkpad-acpi's implementation of the sysfs interfaces will be fixed for
maximum correctness, even if that means changing an interface in
non-compatible ways. As these interfaces mature both in the kernel and
in thinkpad-acpi, such changes should become quite rare.
Applications interfacing to the thinkpad-acpi sysfs interfaces must
follow all sysfs guidelines and correctly process all errors (the sysfs
interface makes extensive use of errors). File descriptors and open /
close operations to the sysfs inodes must also be properly implemented.
The version of thinkpad-acpi's sysfs interface is exported by the driver
as a driver attribute (see below).
Sysfs driver attributes are on the driver's sysfs attribute space,
for 2.6.20 this is /sys/bus/platform/drivers/thinkpad-acpi/.
Sysfs device attributes are on the driver's sysfs attribute space,
for 2.6.20 this is /sys/devices/platform/thinkpad-acpi/.
Driver version
--------------
procfs: /proc/acpi/ibm/driver
sysfs driver attribute: version
The driver name and version. No commands can be written to this file.
Sysfs interface version
-----------------------
sysfs driver attribute: interface_version
Version of the thinkpad-acpi sysfs interface, as an unsigned long
(output in hex format: 0xAAAABBCC), where:
AAAA - major revision
BB - minor revision
CC - bugfix revision
The sysfs interface version changelog for the driver can be found at the
end of this document. Changes to the sysfs interface done by the kernel
subsystems are not documented here, nor are they tracked by this
attribute.
Changes to the thinkpad-acpi sysfs interface are only considered
non-experimental when they are submitted to Linux mainline, at which
point the changes in this interface are documented and interface_version
may be updated. If you are using any thinkpad-acpi features not yet
sent to mainline for merging, you do so on your own risk: these features
may disappear, or be implemented in a different and incompatible way by
the time they are merged in Linux mainline.
Changes that are backwards-compatible by nature (e.g. the addition of
attributes that do not change the way the other attributes work) do not
always warrant an update of interface_version. Therefore, one must
expect that an attribute might not be there, and deal with it properly
(an attribute not being there *is* a valid way to make it clear that a
feature is not available in sysfs).
Hot keys
--------
procfs: /proc/acpi/ibm/hotkey
sysfs device attribute: hotkey_*
In a ThinkPad, the ACPI HKEY handler is responsible for comunicating
some important events and also keyboard hot key presses to the operating
system. Enabling the hotkey functionality of thinkpad-acpi signals the
firmware that such a driver is present, and modifies how the ThinkPad
firmware will behave in many situations.
When the hotkey feature is enabled and the hot key mask is set (see
below), the various hot keys either generate ACPI events in the
following format:
ibm/hotkey HKEY 00000080 0000xxxx
or events over the input layer. The input layer support accepts the
standard IOCTLs to remap the keycodes assigned to each hotkey.
When the input device is open, the driver will suppress any ACPI hot key
events that get translated into a meaningful input layer event, in order
to avoid sending duplicate events to userspace. Hot keys that are
mapped to KEY_RESERVED in the keymap are not translated, and will always
generate an ACPI ibm/hotkey HKEY event, and no input layer events.
The hot key bit mask allows some control over which hot keys generate
events. If a key is "masked" (bit set to 0 in the mask), the firmware
will handle it. If it is "unmasked", it signals the firmware that
thinkpad-acpi would prefer to handle it, if the firmware would be so
kind to allow it (and it often doesn't!).
Not all bits in the mask can be modified. Not all bits that can be
modified do anything. Not all hot keys can be individually controlled
by the mask. Some models do not support the mask at all, and in those
models, hot keys cannot be controlled individually. The behaviour of
the mask is, therefore, higly dependent on the ThinkPad model.
Note that unmasking some keys prevents their default behavior. For
example, if Fn+F5 is unmasked, that key will no longer enable/disable
Bluetooth by itself.
Note also that not all Fn key combinations are supported through ACPI.
For example, on the X40, the brightness, volume and "Access IBM" buttons
do not generate ACPI events even with this driver. They *can* be used
through the "ThinkPad Buttons" utility, see http://www.nongnu.org/tpb/
procfs notes:
The following commands can be written to the /proc/acpi/ibm/hotkey file:
echo enable > /proc/acpi/ibm/hotkey -- enable the hot keys feature
echo disable > /proc/acpi/ibm/hotkey -- disable the hot keys feature
echo 0xffffffff > /proc/acpi/ibm/hotkey -- enable all hot keys
echo 0 > /proc/acpi/ibm/hotkey -- disable all possible hot keys
... any other 8-hex-digit mask ...
echo reset > /proc/acpi/ibm/hotkey -- restore the original mask
sysfs notes:
hotkey_bios_enabled:
Returns the status of the hot keys feature when
thinkpad-acpi was loaded. Upon module unload, the hot
key feature status will be restored to this value.
0: hot keys were disabled
1: hot keys were enabled (unusual)
hotkey_bios_mask:
Returns the hot keys mask when thinkpad-acpi was loaded.
Upon module unload, the hot keys mask will be restored
to this value.
hotkey_enable:
Enables/disables the hot keys feature, and reports
current status of the hot keys feature.
0: disables the hot keys feature / feature disabled
1: enables the hot keys feature / feature enabled
hotkey_mask:
bit mask to enable driver-handling and ACPI event
generation for each hot key (see above). Returns the
current status of the hot keys mask, and allows one to
modify it.
hotkey_all_mask:
bit mask that should enable event reporting for all
supported hot keys, when echoed to hotkey_mask above.
Unless you know which events need to be handled
passively (because the firmware *will* handle them
anyway), do *not* use hotkey_all_mask. Use
hotkey_recommended_mask, instead. You have been warned.
hotkey_recommended_mask:
bit mask that should enable event reporting for all
supported hot keys, except those which are always
handled by the firmware anyway. Echo it to
hotkey_mask above, to use.
hotkey_radio_sw:
if the ThinkPad has a hardware radio switch, this
attribute will read 0 if the switch is in the "radios
disabled" postition, and 1 if the switch is in the
"radios enabled" position.
input layer notes:
A Hot key is mapped to a single input layer EV_KEY event, possibly
followed by an EV_MSC MSC_SCAN event that shall contain that key's scan
code. An EV_SYN event will always be generated to mark the end of the
event block.
Do not use the EV_MSC MSC_SCAN events to process keys. They are to be
used as a helper to remap keys, only. They are particularly useful when
remapping KEY_UNKNOWN keys.
The events are available in an input device, with the following id:
Bus: BUS_HOST
vendor: 0x1014 (PCI_VENDOR_ID_IBM)
product: 0x5054 ("TP")
version: 0x4101
The version will have its LSB incremented if the keymap changes in a
backwards-compatible way. The MSB shall always be 0x41 for this input
device. If the MSB is not 0x41, do not use the device as described in
this section, as it is either something else (e.g. another input device
exported by a thinkpad driver, such as HDAPS) or its functionality has
been changed in a non-backwards compatible way.
Adding other event types for other functionalities shall be considered a
backwards-compatible change for this input device.
Thinkpad-acpi Hot Key event map (version 0x4101):
ACPI Scan
event code Key Notes
0x1001 0x00 FN+F1 -
0x1002 0x01 FN+F2 -
0x1003 0x02 FN+F3 Many models always report this
hot key, even with hot keys
disabled or with Fn+F3 masked
off
0x1004 0x03 FN+F4 Sleep button (ACPI sleep button
semanthics, i.e. sleep-to-RAM).
It is always generate some kind
of event, either the hot key
event or a ACPI sleep button
event. The firmware may
refuse to generate further FN+F4
key presses until a S3 or S4 ACPI
sleep cycle is performed or some
time passes.
0x1005 0x04 FN+F5 Radio. Enables/disables
the internal BlueTooth hardware
and W-WAN card if left in control
of the firmware. Does not affect
the WLAN card.
0x1006 0x05 FN+F6 -
0x1007 0x06 FN+F7 Video output cycle.
Do you feel lucky today?
0x1008 0x07 FN+F8 -
.. .. ..
0x100B 0x0A FN+F11 -
0x100C 0x0B FN+F12 Sleep to disk. You are always
supposed to handle it yourself,
either through the ACPI event,
or through a hotkey event.
The firmware may refuse to
generate further FN+F4 key
press events until a S3 or S4
ACPI sleep cycle is performed,
or some time passes.
0x100D 0x0C FN+BACKSPACE -
0x100E 0x0D FN+INSERT -
0x100F 0x0E FN+DELETE -
0x1010 0x0F FN+HOME Brightness up. This key is
always handled by the firmware,
even when unmasked. Just leave
it alone.
0x1011 0x10 FN+END Brightness down. This key is
always handled by the firmware,
even when unmasked. Just leave
it alone.
0x1012 0x11 FN+PGUP Thinklight toggle. This key is
always handled by the firmware,
even when unmasked.
0x1013 0x12 FN+PGDOWN -
0x1014 0x13 FN+SPACE Zoom key
0x1015 0x14 VOLUME UP Internal mixer volume up. This
key is always handled by the
firmware, even when unmasked.
0x1016 0x15 VOLUME DOWN Internal mixer volume up. This
key is always handled by the
firmware, even when unmasked.
0x1017 0x16 MUTE Mute internal mixer. This
key is always handled by the
firmware, even when unmasked.
0x1018 0x17 THINKPAD Thinkpad/Access IBM/Lenovo key
0x1019 0x18 unknown
.. .. ..
0x1020 0x1F unknown
The ThinkPad firmware does not allow one to differentiate when most hot
keys are pressed or released (either that, or we don't know how to, yet).
For these keys, the driver generates a set of events for a key press and
immediately issues the same set of events for a key release. It is
unknown by the driver if the ThinkPad firmware triggered these events on
hot key press or release, but the firmware will do it for either one, not
both.
If a key is mapped to KEY_RESERVED, it generates no input events at all,
and it may generate a legacy thinkpad-acpi ACPI hotkey event.
If a key is mapped to KEY_UNKNOWN, it generates an input event that
includes an scan code, and it may also generate a legacy thinkpad-acpi
ACPI hotkey event.
If a key is mapped to anything else, it will only generate legacy
thinkpad-acpi ACPI hotkey events if nobody has opened the input device.
Non hot-key ACPI HKEY event map:
0x5001 Lid closed
0x5002 Lid opened
0x7000 Radio Switch may have changed state
Bluetooth
---------
procfs: /proc/acpi/ibm/bluetooth
sysfs device attribute: bluetooth_enable
This feature shows the presence and current state of a ThinkPad
Bluetooth device in the internal ThinkPad CDC slot.
Procfs notes:
If Bluetooth is installed, the following commands can be used:
echo enable > /proc/acpi/ibm/bluetooth
echo disable > /proc/acpi/ibm/bluetooth
Sysfs notes:
If the Bluetooth CDC card is installed, it can be enabled /
disabled through the "bluetooth_enable" thinkpad-acpi device
attribute, and its current status can also be queried.
enable:
0: disables Bluetooth / Bluetooth is disabled
1: enables Bluetooth / Bluetooth is enabled.
Note: this interface will be probably be superseeded by the
generic rfkill class, so it is NOT to be considered stable yet.
Video output control -- /proc/acpi/ibm/video
--------------------------------------------
This feature allows control over the devices used for video output -
LCD, CRT or DVI (if available). The following commands are available:
echo lcd_enable > /proc/acpi/ibm/video
echo lcd_disable > /proc/acpi/ibm/video
echo crt_enable > /proc/acpi/ibm/video
echo crt_disable > /proc/acpi/ibm/video
echo dvi_enable > /proc/acpi/ibm/video
echo dvi_disable > /proc/acpi/ibm/video
echo auto_enable > /proc/acpi/ibm/video
echo auto_disable > /proc/acpi/ibm/video
echo expand_toggle > /proc/acpi/ibm/video
echo video_switch > /proc/acpi/ibm/video
Each video output device can be enabled or disabled individually.
Reading /proc/acpi/ibm/video shows the status of each device.
Automatic video switching can be enabled or disabled. When automatic
video switching is enabled, certain events (e.g. opening the lid,
docking or undocking) cause the video output device to change
automatically. While this can be useful, it also causes flickering
and, on the X40, video corruption. By disabling automatic switching,
the flickering or video corruption can be avoided.
The video_switch command cycles through the available video outputs
(it simulates the behavior of Fn-F7).
Video expansion can be toggled through this feature. This controls
whether the display is expanded to fill the entire LCD screen when a
mode with less than full resolution is used. Note that the current
video expansion status cannot be determined through this feature.
Note that on many models (particularly those using Radeon graphics
chips) the X driver configures the video card in a way which prevents
Fn-F7 from working. This also disables the video output switching
features of this driver, as it uses the same ACPI methods as
Fn-F7. Video switching on the console should still work.
UPDATE: There's now a patch for the X.org Radeon driver which
addresses this issue. Some people are reporting success with the patch
while others are still having problems. For more information:
https://bugs.freedesktop.org/show_bug.cgi?id=2000
ThinkLight control -- /proc/acpi/ibm/light
------------------------------------------
The current status of the ThinkLight can be found in this file. A few
models which do not make the status available will show it as
"unknown". The available commands are:
echo on > /proc/acpi/ibm/light
echo off > /proc/acpi/ibm/light
Docking / undocking -- /proc/acpi/ibm/dock
------------------------------------------
Docking and undocking (e.g. with the X4 UltraBase) requires some
actions to be taken by the operating system to safely make or break
the electrical connections with the dock.
The docking feature of this driver generates the following ACPI events:
ibm/dock GDCK 00000003 00000001 -- eject request
ibm/dock GDCK 00000003 00000002 -- undocked
ibm/dock GDCK 00000000 00000003 -- docked
NOTE: These events will only be generated if the laptop was docked
when originally booted. This is due to the current lack of support for
hot plugging of devices in the Linux ACPI framework. If the laptop was
booted while not in the dock, the following message is shown in the
logs:
Mar 17 01:42:34 aero kernel: thinkpad_acpi: dock device not present
In this case, no dock-related events are generated but the dock and
undock commands described below still work. They can be executed
manually or triggered by Fn key combinations (see the example acpid
configuration files included in the driver tarball package available
on the web site).
When the eject request button on the dock is pressed, the first event
above is generated. The handler for this event should issue the
following command:
echo undock > /proc/acpi/ibm/dock
After the LED on the dock goes off, it is safe to eject the laptop.
Note: if you pressed this key by mistake, go ahead and eject the
laptop, then dock it back in. Otherwise, the dock may not function as
expected.
When the laptop is docked, the third event above is generated. The
handler for this event should issue the following command to fully
enable the dock:
echo dock > /proc/acpi/ibm/dock
The contents of the /proc/acpi/ibm/dock file shows the current status
of the dock, as provided by the ACPI framework.
The docking support in this driver does not take care of enabling or
disabling any other devices you may have attached to the dock. For
example, a CD drive plugged into the UltraBase needs to be disabled or
enabled separately. See the provided example acpid configuration files
for how this can be accomplished.
There is no support yet for PCI devices that may be attached to a
docking station, e.g. in the ThinkPad Dock II. The driver currently
does not recognize, enable or disable such devices. This means that
the only docking stations currently supported are the X-series
UltraBase docks and "dumb" port replicators like the Mini Dock (the
latter don't need any ACPI support, actually).
UltraBay eject -- /proc/acpi/ibm/bay
------------------------------------
Inserting or ejecting an UltraBay device requires some actions to be
taken by the operating system to safely make or break the electrical
connections with the device.
This feature generates the following ACPI events:
ibm/bay MSTR 00000003 00000000 -- eject request
ibm/bay MSTR 00000001 00000000 -- eject lever inserted
NOTE: These events will only be generated if the UltraBay was present
when the laptop was originally booted (on the X series, the UltraBay
is in the dock, so it may not be present if the laptop was undocked).
This is due to the current lack of support for hot plugging of devices
in the Linux ACPI framework. If the laptop was booted without the
UltraBay, the following message is shown in the logs:
Mar 17 01:42:34 aero kernel: thinkpad_acpi: bay device not present
In this case, no bay-related events are generated but the eject
command described below still works. It can be executed manually or
triggered by a hot key combination.
Sliding the eject lever generates the first event shown above. The
handler for this event should take whatever actions are necessary to
shut down the device in the UltraBay (e.g. call idectl), then issue
the following command:
echo eject > /proc/acpi/ibm/bay
After the LED on the UltraBay goes off, it is safe to pull out the
device.
When the eject lever is inserted, the second event above is
generated. The handler for this event should take whatever actions are
necessary to enable the UltraBay device (e.g. call idectl).
The contents of the /proc/acpi/ibm/bay file shows the current status
of the UltraBay, as provided by the ACPI framework.
EXPERIMENTAL warm eject support on the 600e/x, A22p and A3x (To use
this feature, you need to supply the experimental=1 parameter when
loading the module):
These models do not have a button near the UltraBay device to request
a hot eject but rather require the laptop to be put to sleep
(suspend-to-ram) before the bay device is ejected or inserted).
The sequence of steps to eject the device is as follows:
echo eject > /proc/acpi/ibm/bay
put the ThinkPad to sleep
remove the drive
resume from sleep
cat /proc/acpi/ibm/bay should show that the drive was removed
On the A3x, both the UltraBay 2000 and UltraBay Plus devices are
supported. Use "eject2" instead of "eject" for the second bay.
Note: the UltraBay eject support on the 600e/x, A22p and A3x is
EXPERIMENTAL and may not work as expected. USE WITH CAUTION!
CMOS control
------------
procfs: /proc/acpi/ibm/cmos
sysfs device attribute: cmos_command
This feature is mostly used internally by the ACPI firmware to keep the legacy
CMOS NVRAM bits in sync with the current machine state, and to record this
state so that the ThinkPad will retain such settings across reboots.
Some of these commands actually perform actions in some ThinkPad models, but
this is expected to disappear more and more in newer models. As an example, in
a T43 and in a X40, commands 12 and 13 still control the ThinkLight state for
real, but commands 0 to 2 don't control the mixer anymore (they have been
phased out) and just update the NVRAM.
The range of valid cmos command numbers is 0 to 21, but not all have an
effect and the behavior varies from model to model. Here is the behavior
on the X40 (tpb is the ThinkPad Buttons utility):
0 - Related to "Volume down" key press
1 - Related to "Volume up" key press
2 - Related to "Mute on" key press
3 - Related to "Access IBM" key press
4 - Related to "LCD brightness up" key pess
5 - Related to "LCD brightness down" key press
11 - Related to "toggle screen expansion" key press/function
12 - Related to "ThinkLight on"
13 - Related to "ThinkLight off"
14 - Related to "ThinkLight" key press (toggle thinklight)
The cmos command interface is prone to firmware split-brain problems, as
in newer ThinkPads it is just a compatibility layer. Do not use it, it is
exported just as a debug tool.
LED control -- /proc/acpi/ibm/led
---------------------------------
Some of the LED indicators can be controlled through this feature. The
available commands are:
echo '<led number> on' >/proc/acpi/ibm/led
echo '<led number> off' >/proc/acpi/ibm/led
echo '<led number> blink' >/proc/acpi/ibm/led
The <led number> range is 0 to 7. The set of LEDs that can be
controlled varies from model to model. Here is the mapping on the X40:
0 - power
1 - battery (orange)
2 - battery (green)
3 - UltraBase
4 - UltraBay
7 - standby
All of the above can be turned on and off and can be made to blink.
ACPI sounds -- /proc/acpi/ibm/beep
----------------------------------
The BEEP method is used internally by the ACPI firmware to provide
audible alerts in various situations. This feature allows the same
sounds to be triggered manually.
The commands are non-negative integer numbers:
echo <number> >/proc/acpi/ibm/beep
The valid <number> range is 0 to 17. Not all numbers trigger sounds
and the sounds vary from model to model. Here is the behavior on the
X40:
0 - stop a sound in progress (but use 17 to stop 16)
2 - two beeps, pause, third beep ("low battery")
3 - single beep
4 - high, followed by low-pitched beep ("unable")
5 - single beep
6 - very high, followed by high-pitched beep ("AC/DC")
7 - high-pitched beep
9 - three short beeps
10 - very long beep
12 - low-pitched beep
15 - three high-pitched beeps repeating constantly, stop with 0
16 - one medium-pitched beep repeating constantly, stop with 17
17 - stop 16
Temperature sensors
-------------------
procfs: /proc/acpi/ibm/thermal
sysfs device attributes: (hwmon) temp*_input
Most ThinkPads include six or more separate temperature sensors but
only expose the CPU temperature through the standard ACPI methods.
This feature shows readings from up to eight different sensors on older
ThinkPads, and it has experimental support for up to sixteen different
sensors on newer ThinkPads.
EXPERIMENTAL: The 16-sensors feature is marked EXPERIMENTAL because the
implementation directly accesses hardware registers and may not work as
expected. USE WITH CAUTION! To use this feature, you need to supply the
experimental=1 parameter when loading the module. When EXPERIMENTAL
mode is enabled, reading the first 8 sensors on newer ThinkPads will
also use an new experimental thermal sensor access mode.
For example, on the X40, a typical output may be:
temperatures: 42 42 45 41 36 -128 33 -128
EXPERIMENTAL: On the T43/p, a typical output may be:
temperatures: 48 48 36 52 38 -128 31 -128 48 52 48 -128 -128 -128 -128 -128
The mapping of thermal sensors to physical locations varies depending on
system-board model (and thus, on ThinkPad model).
http://thinkwiki.org/wiki/Thermal_Sensors is a public wiki page that
tries to track down these locations for various models.
Most (newer?) models seem to follow this pattern:
1: CPU
2: (depends on model)
3: (depends on model)
4: GPU
5: Main battery: main sensor
6: Bay battery: main sensor
7: Main battery: secondary sensor
8: Bay battery: secondary sensor
9-15: (depends on model)
For the R51 (source: Thomas Gruber):
2: Mini-PCI
3: Internal HDD
For the T43, T43/p (source: Shmidoax/Thinkwiki.org)
http://thinkwiki.org/wiki/Thermal_Sensors#ThinkPad_T43.2C_T43p
2: System board, left side (near PCMCIA slot), reported as HDAPS temp
3: PCMCIA slot
9: MCH (northbridge) to DRAM Bus
10: Clock-generator, mini-pci card and ICH (southbridge), under Mini-PCI
card, under touchpad
11: Power regulator, underside of system board, below F2 key
The A31 has a very atypical layout for the thermal sensors
(source: Milos Popovic, http://thinkwiki.org/wiki/Thermal_Sensors#ThinkPad_A31)
1: CPU
2: Main Battery: main sensor
3: Power Converter
4: Bay Battery: main sensor
5: MCH (northbridge)
6: PCMCIA/ambient
7: Main Battery: secondary sensor
8: Bay Battery: secondary sensor
Procfs notes:
Readings from sensors that are not available return -128.
No commands can be written to this file.
Sysfs notes:
Sensors that are not available return the ENXIO error. This
status may change at runtime, as there are hotplug thermal
sensors, like those inside the batteries and docks.
thinkpad-acpi thermal sensors are reported through the hwmon
subsystem, and follow all of the hwmon guidelines at
Documentation/hwmon.
EXPERIMENTAL: Embedded controller register dump -- /proc/acpi/ibm/ecdump
------------------------------------------------------------------------
This feature is marked EXPERIMENTAL because the implementation
directly accesses hardware registers and may not work as expected. USE
WITH CAUTION! To use this feature, you need to supply the
experimental=1 parameter when loading the module.
This feature dumps the values of 256 embedded controller
registers. Values which have changed since the last time the registers
were dumped are marked with a star:
[root@x40 ibm-acpi]# cat /proc/acpi/ibm/ecdump
EC +00 +01 +02 +03 +04 +05 +06 +07 +08 +09 +0a +0b +0c +0d +0e +0f
EC 0x00: a7 47 87 01 fe 96 00 08 01 00 cb 00 00 00 40 00
EC 0x10: 00 00 ff ff f4 3c 87 09 01 ff 42 01 ff ff 0d 00
EC 0x20: 00 00 00 00 00 00 00 00 00 00 00 03 43 00 00 80
EC 0x30: 01 07 1a 00 30 04 00 00 *85 00 00 10 00 50 00 00
EC 0x40: 00 00 00 00 00 00 14 01 00 04 00 00 00 00 00 00
EC 0x50: 00 c0 02 0d 00 01 01 02 02 03 03 03 03 *bc *02 *bc
EC 0x60: *02 *bc *02 00 00 00 00 00 00 00 00 00 00 00 00 00
EC 0x70: 00 00 00 00 00 12 30 40 *24 *26 *2c *27 *20 80 *1f 80
EC 0x80: 00 00 00 06 *37 *0e 03 00 00 00 0e 07 00 00 00 00
EC 0x90: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
EC 0xa0: *ff 09 ff 09 ff ff *64 00 *00 *00 *a2 41 *ff *ff *e0 00
EC 0xb0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
EC 0xc0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
EC 0xd0: 03 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
EC 0xe0: 00 00 00 00 00 00 00 00 11 20 49 04 24 06 55 03
EC 0xf0: 31 55 48 54 35 38 57 57 08 2f 45 73 07 65 6c 1a
This feature can be used to determine the register holding the fan
speed on some models. To do that, do the following:
- make sure the battery is fully charged
- make sure the fan is running
- run 'cat /proc/acpi/ibm/ecdump' several times, once per second or so
The first step makes sure various charging-related values don't
vary. The second ensures that the fan-related values do vary, since
the fan speed fluctuates a bit. The third will (hopefully) mark the
fan register with a star:
[root@x40 ibm-acpi]# cat /proc/acpi/ibm/ecdump
EC +00 +01 +02 +03 +04 +05 +06 +07 +08 +09 +0a +0b +0c +0d +0e +0f
EC 0x00: a7 47 87 01 fe 96 00 08 01 00 cb 00 00 00 40 00
EC 0x10: 00 00 ff ff f4 3c 87 09 01 ff 42 01 ff ff 0d 00
EC 0x20: 00 00 00 00 00 00 00 00 00 00 00 03 43 00 00 80
EC 0x30: 01 07 1a 00 30 04 00 00 85 00 00 10 00 50 00 00
EC 0x40: 00 00 00 00 00 00 14 01 00 04 00 00 00 00 00 00
EC 0x50: 00 c0 02 0d 00 01 01 02 02 03 03 03 03 bc 02 bc
EC 0x60: 02 bc 02 00 00 00 00 00 00 00 00 00 00 00 00 00
EC 0x70: 00 00 00 00 00 12 30 40 24 27 2c 27 21 80 1f 80
EC 0x80: 00 00 00 06 *be 0d 03 00 00 00 0e 07 00 00 00 00
EC 0x90: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
EC 0xa0: ff 09 ff 09 ff ff 64 00 00 00 a2 41 ff ff e0 00
EC 0xb0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
EC 0xc0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
EC 0xd0: 03 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
EC 0xe0: 00 00 00 00 00 00 00 00 11 20 49 04 24 06 55 03
EC 0xf0: 31 55 48 54 35 38 57 57 08 2f 45 73 07 65 6c 1a
Another set of values that varies often is the temperature
readings. Since temperatures don't change vary fast, you can take
several quick dumps to eliminate them.
You can use a similar method to figure out the meaning of other
embedded controller registers - e.g. make sure nothing else changes
except the charging or discharging battery to determine which
registers contain the current battery capacity, etc. If you experiment
with this, do send me your results (including some complete dumps with
a description of the conditions when they were taken.)
LCD brightness control
----------------------
procfs: /proc/acpi/ibm/brightness
sysfs backlight device "thinkpad_screen"
This feature allows software control of the LCD brightness on ThinkPad
models which don't have a hardware brightness slider.
It has some limitations: the LCD backlight cannot be actually turned on or off
by this interface, and in many ThinkPad models, the "dim while on battery"
functionality will be enabled by the BIOS when this interface is used, and
cannot be controlled.
The backlight control has eight levels, ranging from 0 to 7. Some of the
levels may not be distinct.
Procfs notes:
The available commands are:
echo up >/proc/acpi/ibm/brightness
echo down >/proc/acpi/ibm/brightness
echo 'level <level>' >/proc/acpi/ibm/brightness
Sysfs notes:
The interface is implemented through the backlight sysfs class, which is poorly
documented at this time.
Locate the thinkpad_screen device under /sys/class/backlight, and inside it
there will be the following attributes:
max_brightness:
Reads the maximum brightness the hardware can be set to.
The minimum is always zero.
actual_brightness:
Reads what brightness the screen is set to at this instant.
brightness:
Writes request the driver to change brightness to the given
value. Reads will tell you what brightness the driver is trying
to set the display to when "power" is set to zero and the display
has not been dimmed by a kernel power management event.
power:
power management mode, where 0 is "display on", and 1 to 3 will
dim the display backlight to brightness level 0 because
thinkpad-acpi cannot really turn the backlight off. Kernel
power management events can temporarily increase the current
power management level, i.e. they can dim the display.
Volume control -- /proc/acpi/ibm/volume
---------------------------------------
This feature allows volume control on ThinkPad models which don't have
a hardware volume knob. The available commands are:
echo up >/proc/acpi/ibm/volume
echo down >/proc/acpi/ibm/volume
echo mute >/proc/acpi/ibm/volume
echo 'level <level>' >/proc/acpi/ibm/volume
The <level> number range is 0 to 15 although not all of them may be
distinct. The unmute the volume after the mute command, use either the
up or down command (the level command will not unmute the volume).
The current volume level and mute state is shown in the file.
Fan control and monitoring: fan speed, fan enable/disable
---------------------------------------------------------
procfs: /proc/acpi/ibm/fan
sysfs device attributes: (hwmon) fan_input, pwm1, pwm1_enable
NOTE NOTE NOTE: fan control operations are disabled by default for
safety reasons. To enable them, the module parameter "fan_control=1"
must be given to thinkpad-acpi.
This feature attempts to show the current fan speed, control mode and
other fan data that might be available. The speed is read directly
from the hardware registers of the embedded controller. This is known
to work on later R, T, X and Z series ThinkPads but may show a bogus
value on other models.
Fan levels:
Most ThinkPad fans work in "levels" at the firmware interface. Level 0
stops the fan. The higher the level, the higher the fan speed, although
adjacent levels often map to the same fan speed. 7 is the highest
level, where the fan reaches the maximum recommended speed.
Level "auto" means the EC changes the fan level according to some
internal algorithm, usually based on readings from the thermal sensors.
There is also a "full-speed" level, also known as "disengaged" level.
In this level, the EC disables the speed-locked closed-loop fan control,
and drives the fan as fast as it can go, which might exceed hardware
limits, so use this level with caution.
The fan usually ramps up or down slowly from one speed to another, and
it is normal for the EC to take several seconds to react to fan
commands. The full-speed level may take up to two minutes to ramp up to
maximum speed, and in some ThinkPads, the tachometer readings go stale
while the EC is transitioning to the full-speed level.
WARNING WARNING WARNING: do not leave the fan disabled unless you are
monitoring all of the temperature sensor readings and you are ready to
enable it if necessary to avoid overheating.
An enabled fan in level "auto" may stop spinning if the EC decides the
ThinkPad is cool enough and doesn't need the extra airflow. This is
normal, and the EC will spin the fan up if the varios thermal readings
rise too much.
On the X40, this seems to depend on the CPU and HDD temperatures.
Specifically, the fan is turned on when either the CPU temperature
climbs to 56 degrees or the HDD temperature climbs to 46 degrees. The
fan is turned off when the CPU temperature drops to 49 degrees and the
HDD temperature drops to 41 degrees. These thresholds cannot
currently be controlled.
The ThinkPad's ACPI DSDT code will reprogram the fan on its own when
certain conditions are met. It will override any fan programming done
through thinkpad-acpi.
The thinkpad-acpi kernel driver can be programmed to revert the fan
level to a safe setting if userspace does not issue one of the procfs
fan commands: "enable", "disable", "level" or "watchdog", or if there
are no writes to pwm1_enable (or to pwm1 *if and only if* pwm1_enable is
set to 1, manual mode) within a configurable amount of time of up to
120 seconds. This functionality is called fan safety watchdog.
Note that the watchdog timer stops after it enables the fan. It will be
rearmed again automatically (using the same interval) when one of the
above mentioned fan commands is received. The fan watchdog is,
therefore, not suitable to protect against fan mode changes made through
means other than the "enable", "disable", and "level" procfs fan
commands, or the hwmon fan control sysfs interface.
Procfs notes:
The fan may be enabled or disabled with the following commands:
echo enable >/proc/acpi/ibm/fan
echo disable >/proc/acpi/ibm/fan
Placing a fan on level 0 is the same as disabling it. Enabling a fan
will try to place it in a safe level if it is too slow or disabled.
The fan level can be controlled with the command:
echo 'level <level>' > /proc/acpi/ibm/fan
Where <level> is an integer from 0 to 7, or one of the words "auto" or
"full-speed" (without the quotes). Not all ThinkPads support the "auto"
and "full-speed" levels. The driver accepts "disengaged" as an alias for
"full-speed", and reports it as "disengaged" for backwards
compatibility.
On the X31 and X40 (and ONLY on those models), the fan speed can be
controlled to a certain degree. Once the fan is running, it can be
forced to run faster or slower with the following command:
echo 'speed <speed>' > /proc/acpi/ibm/fan
The sustainable range of fan speeds on the X40 appears to be from about
3700 to about 7350. Values outside this range either do not have any
effect or the fan speed eventually settles somewhere in that range. The
fan cannot be stopped or started with this command. This functionality
is incomplete, and not available through the sysfs interface.
To program the safety watchdog, use the "watchdog" command.
echo 'watchdog <interval in seconds>' > /proc/acpi/ibm/fan
If you want to disable the watchdog, use 0 as the interval.
Sysfs notes:
The sysfs interface follows the hwmon subsystem guidelines for the most
part, and the exception is the fan safety watchdog.
Writes to any of the sysfs attributes may return the EINVAL error if
that operation is not supported in a given ThinkPad or if the parameter
is out-of-bounds, and EPERM if it is forbidden. They may also return
EINTR (interrupted system call), and EIO (I/O error while trying to talk
to the firmware).
Features not yet implemented by the driver return ENOSYS.
hwmon device attribute pwm1_enable:
0: PWM offline (fan is set to full-speed mode)
1: Manual PWM control (use pwm1 to set fan level)
2: Hardware PWM control (EC "auto" mode)
3: reserved (Software PWM control, not implemented yet)
Modes 0 and 2 are not supported by all ThinkPads, and the
driver is not always able to detect this. If it does know a
mode is unsupported, it will return -EINVAL.
hwmon device attribute pwm1:
Fan level, scaled from the firmware values of 0-7 to the hwmon
scale of 0-255. 0 means fan stopped, 255 means highest normal
speed (level 7).
This attribute only commands the fan if pmw1_enable is set to 1
(manual PWM control).
hwmon device attribute fan1_input:
Fan tachometer reading, in RPM. May go stale on certain
ThinkPads while the EC transitions the PWM to offline mode,
which can take up to two minutes. May return rubbish on older
ThinkPads.
driver attribute fan_watchdog:
Fan safety watchdog timer interval, in seconds. Minimum is
1 second, maximum is 120 seconds. 0 disables the watchdog.
To stop the fan: set pwm1 to zero, and pwm1_enable to 1.
To start the fan in a safe mode: set pwm1_enable to 2. If that fails
with EINVAL, try to set pwm1_enable to 1 and pwm1 to at least 128 (255
would be the safest choice, though).
EXPERIMENTAL: WAN
-----------------
procfs: /proc/acpi/ibm/wan
sysfs device attribute: wwan_enable
This feature is marked EXPERIMENTAL because the implementation
directly accesses hardware registers and may not work as expected. USE
WITH CAUTION! To use this feature, you need to supply the
experimental=1 parameter when loading the module.
This feature shows the presence and current state of a W-WAN (Sierra
Wireless EV-DO) device.
It was tested on a Lenovo Thinkpad X60. It should probably work on other
Thinkpad models which come with this module installed.
Procfs notes:
If the W-WAN card is installed, the following commands can be used:
echo enable > /proc/acpi/ibm/wan
echo disable > /proc/acpi/ibm/wan
Sysfs notes:
If the W-WAN card is installed, it can be enabled /
disabled through the "wwan_enable" thinkpad-acpi device
attribute, and its current status can also be queried.
enable:
0: disables WWAN card / WWAN card is disabled
1: enables WWAN card / WWAN card is enabled.
Note: this interface will be probably be superseeded by the
generic rfkill class, so it is NOT to be considered stable yet.
Multiple Commands, Module Parameters
------------------------------------
Multiple commands can be written to the proc files in one shot by
separating them with commas, for example:
echo enable,0xffff > /proc/acpi/ibm/hotkey
echo lcd_disable,crt_enable > /proc/acpi/ibm/video
Commands can also be specified when loading the thinkpad-acpi module,
for example:
modprobe thinkpad_acpi hotkey=enable,0xffff video=auto_disable
Enabling debugging output
-------------------------
The module takes a debug paramater which can be used to selectively
enable various classes of debugging output, for example:
modprobe ibm_acpi debug=0xffff
will enable all debugging output classes. It takes a bitmask, so
to enable more than one output class, just add their values.
Debug bitmask Description
0x0001 Initialization and probing
0x0002 Removal
There is also a kernel build option to enable more debugging
information, which may be necessary to debug driver problems.
The level of debugging information output by the driver can be changed
at runtime through sysfs, using the driver attribute debug_level. The
attribute takes the same bitmask as the debug module parameter above.
Force loading of module
-----------------------
If thinkpad-acpi refuses to detect your ThinkPad, you can try to specify
the module parameter force_load=1. Regardless of whether this works or
not, please contact ibm-acpi-devel@lists.sourceforge.net with a report.
Sysfs interface changelog:
0x000100: Initial sysfs support, as a single platform driver and
device.
0x000200: Hot key support for 32 hot keys, and radio slider switch
support.
0x010000: Hot keys are now handled by default over the input
layer, the radio switch generates input event EV_RADIO,
and the driver enables hot key handling by default in
the firmware.
