CHANGES

Change history for the tp_smapi package:

0.44 2023-07-31
---------------------
- Support for 6.4 kernels

0.43 2018-02-24
---------------------
- Support for 4.15 kernels

0.42 2016-04-23
---------------------
- fixes for newer kernels
- kbuild support
- dkms support
- quirks update
- RPM support

0.41 2011-07-29
---------------------
- new maintainer
- fixes for newer kernels
- fixes for newer ThinkPads (esp. SandyBridge ones)

0.40 2008-12-16
---------------------
- thinkpad_ec: Added a "force_io=1" module parameter, to let it load on
  recent ThinkPad models (e.g., T400 and T500). This is a kludge to work
  around an ACPI DSDT which claims the ports we need.
- tp_smapi: Don't restore battery charging thresholds when resuming from
  suspend-to-RAM.
- When building, use /lib/modules/*/build instead of /lib/modules/*/src
  and give a clearer error message if that's missing.

0.39 2008-09-27
---------------------
- Fixed compilation on kernels <2.6.26
  (thanks to Evgeni Golov and Whoopie!)

0.38 2008-09-26
---------------------
- Fixed compilation on kernel 2.6.27-rc7.
- tp_smapi: Added a new battery attribute
  /sys/devices/platform/smapi/BAT?/remaining_percent_error
  This is the error margin for the remaing_percent attribute.
  Unfortunately, it doesn't seem to be reset by battery calibration.
- hdaps: Removed the fake_data_mode attribute.

0.37 2008-03-28
---------------------
- Fix compilation on kernel 2.6.25
- hdaps whitelist: changed default orientation of ThinkPad R60, R61, X40, X41.
- tp_smapi: Added 3 new read-only battery attributes
  /sys/devices/platform/smapi/BAT?/remaining_running_time_now
    (time remaining according to instantenous power instead of average)
  /sys/devices/platform/smapi/BAT?/charging_max_current
    (maximum current allowed during charging)
  /sys/devices/platform/smapi/BAT?/charging_max_voltage
    (maximum voltage allowed during charging)
- tp_smapi and README: improved documentation of battery info attributes.

0.36 2008-01-27
---------------------
- tp_smapi: Added a 4th cell group voltage attribute:
  /sys/devices/platform/smapi/BAT0/group3_voltage
- tp_smapi: Reversed order of the group{0,1,2}_voltage attributes.
- hdaps whitelist: changed default orientation of ThinkPad X61.
- README: Improve attributes documentation.
- Sourcecode cleanups.

0.35-test1 2008-01-21
---------------------
- Allow loading on new T61 firmware that caused an "hdaps_check_ec failed
  error. Correct operation with this firmware is not yet fully tested.
- tp_smapi: Added 3 new sysfs attributes:
  /sys/devices/platform/smapi/BAT0/group0_voltage
  /sys/devices/platform/smapi/BAT0/group1_voltage
  /sys/devices/platform/smapi/BAT0/group2_voltage
  ThinkPad batteries (at least on the 600, X3x, T4x and R5x models) contains
  3 cell groups in series, where each group consisting of 2 or 3 cells
  connected in parallel. The voltage of each group is given by these attributes.
  (The effective performance of the battery is determined by the weakest
  group, i.e., the one those voltage changes most rapidly during dis/charging.)
- hdaps whitelist: changed default orientation of ThinkPad X60 Tablet,
  X61 Tablet and ThinkPad X60s and ThinkPd T60.
- hdaps: when hdaps_ec_check fails, report the values received from the EC.

0.34  2008-01-06
---------------------
- hdaps: Fixed handling of "invert" module parameter (make it a uint,
  and don't let whitelist override module parameter).
  Contributed by Ostap Cherkashin.
- hdaps whitelist: change ThinkPad X61s and T61 default orientation.
  X60 Tablet and X61 Tablet remain wrong until someone reports their DMI ID.

0.33  2007-12-22
---------------------
- Ported to kernel 2.6.24-rc6.
  Building on older kernels will cause harmless warnings about pointer types.
  (Makefile patch contributed by Tim Niemeyer and Whoopie.)
- README: Document the additional hdaps attributes and the reduced timer
  interrupts introduced on 0.32.
- hdaps: Extends the "invert" parameter to cover all possible axis
  orientations.
  The possible values are as follows.
  Let X,Y denote the hardware readouts.
  Let R denote the laptop's roll (tilt left/right).
  Let P denote the laptop's pitch (tilt forward/backward).
    invert=0:   R= X  P= Y   (same as mainline)
    invert=1:   R=-X  P=-Y   (same as mainline)
    invert=2:   R=-X  P= Y   (new)
    invert=3:   R= X  P=-Y   (new)
    invert=4:   R= Y  P= X   (new)
    invert=5:   R=-Y  P=-X   (new)
    invert=6:   R=-Y  P= X   (new)
    invert=7:   R= Y  P=-X   (new)
  It's probably easiest to just try all 8 possibilities and see which yields
  correct results (e.g., in the hdaps-gl visualisation).
- hdaps: Adds a whitelist which automatically sets the correct axis orientation
  for some models. If the value for your model is wrong or missing, you can
  override it using the "invert" parameter. Please also update the tables at
  http://www.thinkwiki.org/wiki/tp_smapi and
  http://www.thinkwiki.org/wiki/List_of_DMI_IDs
  and submit a patch for the whitelist in hdaps.c.
- tp_smapi: Report EOPNOTSUPP instead of ENOSYS for unimplemented SMAPI
  functions.

0.32 2007-07-28
---------------------
- hdaps: Added a second input device which publishes the raw sensor position,
  without joystick fuzz. In conjunction with a new hdapsd, this lets us
  avoid the frequent redundant interrupts caused by hdapsd.
- hdaps: Stop polling the hardware when no one is listening. This prevents
  unnecessary interrupts on tickless kernel. (Contributed by Michael Riepe.)
- In the patch generated by "make patch", thinkpad_ec and tp_smapi now reside
  under drivers/misc instead of drivers/platform.
- "make patch" fixed for kernel 2.6.22.
- Removed kludge for kernels lacking the DMI OEM String decode patch.
- Now requires kernel >= 2.6.19 (due to above).
- Removed the useless "enable_pci_power_saving_on_boot" sysfs attribute.

0.31  2007-03-07
---------------------
Visible changes:
- Fix #includes to allow compilation on 64-bit platforms
  (patch by Emil Renner Berthing).
- thinkpad_ec: Avoid spurious "bad end STR3: (...)->0x80" warnings
  by relaxing status check (still OK according to H8 EC docs).
- Invert axes on all T60, not just T60p.

Internal changes:
- Add __init to find_smapi_port() and thinkpad_ec_test()
- tp_smapi.c: Fix "initialization from incompatible pointer type" warnings
  on 64-bit platforms

0.30  2006-09-03
---------------------
Visible changes:
- Makefile: now supports kernels built with separate source and build
  directories (thanks to Lenz Grimmer).
  To work against against some kernel installed under /lib/modules/
  add an appropriate KVER= parameter:
  # make patch KVER=2.6.16-rc2
  If it's not installed you'll need to set KSRC and KBUILD too:
  # make patch KVER=2.6.16-rc2 KSRC=$HOME/2.6.16-rc2 KBUILD=$HOME/2.6.16-rc2
- Changed the format of /sys/devices/platform/smapi/smapi_request
  (no more register names).

Internal changes:
- tp_smapi: the 'dump' sysfs attribute now outputs only rows 0x00 through 0x0a,
  since 0x0b causes an EC hang on some firmware (e.g., all T42 and old T43).
  Thanks to Henrique and Sukant for tracking this down!
- thinkpad_ec: Log a warning if 0x161F returns 0x80.
- thinkpad_ec: Increase TPC_REQUEST_RETRIES
- thinkpad_ec: Report row args upon data read error.
- hdaps: Apply hwmon-hdaps-handle-errors-from-input-register-device.patch
  from -mm.

0.29  2006-08-17
---------------------
Visible changes:
- Added new sysfs attributes to tp_smapi:
    /sys/devices/platform/smapi/BAT0/remaining_percent
    /sys/devices/platform/smapi/BAT0/remaining_running_time
    /sys/devices/platform/smapi/BAT0/remaining_charging_time
    /sys/devices/platform/smapi/BAT0/temperature    # milli-Celsius
  All of these are read directly from the embedded controller, so they are
  more accurate than doing your own computation.

Internal changes:
- tp_smapi: Extended and commented battery attribute show functions.

0.28  2006-08-16
---------------------
Visible changes:
- thinkpad_ec: Removed 'debug' module parameter, it no longer affected much.
- tp_smapi: Remove optical drive speed control (formerly #ifdefed out).

Internal changes:
- hdaps: Removed __init from hdaps_check_ec(). This fixes crashes on resume
  (Thanks, Igor!).
- hdaps: Moved axis transformation to dedicated function transform_axes().
- tp_smapi: Cleaned up printk() output, changed macros.
- tp_smapi: Added "E" to SMAPI parameter names (BX->EBX) - they're 32-bit.
- tp_Smapi: Removed verbose printks()s on bad sysfs args.
- A lot of small cleanups.

0.27 2006-08-05
---------------------
Visible changes:
- Renamed tp_base.* to thinkpad_ec.*.
  NOTE: If you have manual modprobe commands or configuration mentioning 
  tp_base, you'll need to change them.
- hdaps: All HDAPS-equippedThinkPads should be now be automatically
  recognized. Whitelisting is needed only for changing the axis
  configuration.
- hdaps: Removed 'force' module parameter, it's no longer needed.

Internal changes:
- Renamed all tp_controller_* functions to thinkpad_ec_*.
- Small change to invalidation in hdaps_device_init().
- Separate Kconfig changes into separate patches.
- Clean up generated patch format a bit.

0.26 2006-08-04
---------------------
- The modified hdaps.c is now included as a full file (based on linux git)
  instead of a patch, to solve compilation and versioning problems.
- Makefile: Fixed a problem with "make patch" on new kernels.
- tp_smapi.c: renamed internal attribute functions.

0.25 2006-08-03
---------------------
- Added new attribute: 
  /sys/devices/platform/smapi/enable_pci_power_saving_on_boot
  This controls the "PCI bus power saving" option in the BIOS (takes
  effect at the next boot).
- Added new attribute: 
  /sys/devices/platform/smapi/smapi_request
  This performs raw SMAPI calls. It uses a bad interface that cannot handle
  multiple simultaneous accesses. Don't touch it, it's for development only.
- "make patch" works again. If needed, the resulting patch will include
  the dmi-decode-and-save-oem-string-information patch.
- "make load" now gives an informative error if ran as non-root.
- Makefile: Give nicer error message on kernel <= 2.6.16.
- Makefile: Verify that /usr/sbin/dmidecode works.

0.24-test3 2006-08-01
---------------------
- tp_base: fix compilation error on unpatched kernels.

0.24-test1 2006-07-31
---------------------
This is an experimental version with a few quirks necessary for testing new
code. It's OK to run it yourself to try the new features, but it's not 
suitable for packaging in distributions. The stable version remains 0.22.

Changes to loading, whitelisting and building:
- Detection of the embedded controller (needed for all our modules) is now 
  done according to OEM Strings in the DMI information. If the EC is not
  mentioned there (you can check it in 'dmidecode' too) you'll get a 
  "tp_base: no ThinkPad embedded controller!" message in dmesg and the 
  modules won't load. If tp_smapi used to work and was broken by this,
  please report. A few known exceptions are already 
- The above wants the following kernel patch (soon to be included in -mm):
  diff/dmi-decode-and-save-oem-string-information.patch
  If this patch is not already applied to the current kernel, the Makefile 
  and tp_base.c default to an ugly kludge where the DMI information is
  hardcoded into your driver instead of being read in runtime. You'll get 
  a warning about it, but it will work nicely. Just don't run the compiled 
  modules on a different machine. 
  Sorry for the mess, but we need to exercise the DMI check code.
- "make patch" is temporarily broken, due to the above.
- hdaps: Ignore the "initial mode latch" during init, we're not using it and 
  Lenovo keeps adding new ones.

Changes to hdaps functionality:
- hdaps: New attribute /sys/devices/platform/hdaps/sampling_rate. 
  This determines how frequently of accelerometer state refresh, independently
  of userspace polling (i.e., reading the position attribute more frequently
  than this will yield duplicate readouts). Default=50.
- Multiple hdaps applications can now run without stealing each other's data.
  Set /sys/devices/platform/hdaps/sampling_rate to the highest rate required
  by any app.
- hdaps: New attribute /sys/devices/platform/hdaps/oversampling_ratio.
  When set to X, the embedded controller is told to do physical accelerometer
  measurements at a rate that is X times higher than the rate at which
  the driver reads those measurements (i.e., X*sampling_rate). This
  reduces sample phase difference is, and useful for the running average
  filter (see next). Default=5.
- hdaps: New attribute  /sys/devices/platform/hdaps/running_avg_filter_order.
  When set to X, reported readouts will be the average of the last X physical
  accelerometer measurements. Current firmware allows 1<=X<=8. Setting to a
  high value decreases readout fluctuations. The averaging is handled
  by the embedded controller, so no CPU resources are used. 
  Default=2 (this implicit in previous versions).
- hdaps: New attribute  /sys/devices/platform/hdaps/fake_data_mode (write only).
  If set to 1, enables a test mode where the physical accelerometer readouts
  are replaced with an incrementing counter. This is useful for checking the
  regularity of the sampling interval and driver<->userspace communication.

Changes to tp_smapi functionality:
- tp_smapi: Removed the units ("mV", "mWh" etc.) from all /sys readouts, to
  follow kernel convention.
- tp_smapi: /sys/devices/platform/smapi/BAT?/dump now contains additional 
  data lines (but we don't know what they mean). Default=0.
- tp_smapi: /sys/devices/platform/smapi/BAT?/dump contains some unused values;
  they're now omitted (replaced by "--") instead of showing semi-random junk.

Other changes:
- tp_smapi: fixed signness bug in current_now and current_avg when 
  discharging (thanks to Pavel Machek!).
- hdaps: Turn off power and EC polling upon suspend or module unload.
- hdaps: Check more EC function return values.
- hdaps: Refactored some functions.
- tp_base: Test EC on module load and refuse loading if EC not found.
- tp_base, hdaps, tp_smapi: changed args passing and made it more concise.
- tp_smapi: cleaned up attribute macros.
- tp_smapi: cleaned up smapi retry loop.
- tp_smapi: removed smapi_write and dropped outAX param from smapi_request.
- tp_smapi: improved comments.
- tp_base: moved comments from tp_base.h to tp_base.c.

Many of the above changes were suggested or contributed by
Henrique de Moraes Holschuh.

0.22  2006-07-20
----------------
- hdaps: Added support for X60s (but X axis is still inverted).
- tp_base: Wait until EC starts replying before considering a request 
  successful (this improves stability and eliminates abnormal events.)
- hdaps: Removed /sys/devices/platform/hdaps/{variance,temp2}. They actually
  show an older pending readout (if there is one), so it never makes sense
  to read them via the existing interface.
- tp_base: Added many additional status and consistency checks on EC, 
  to give early warning on any strange behavior.
- tp_base: New constants and comments greatly clarify access to the EC.
- hdaps: Convert all init code to use tp_base transactions. Mo more direct 
  IO port access.
- hdaps: Move some init code to new functions hdaps_set_power() and 
  hdaps_set_ec_config(). These will be exposed as attributes in the future.
- hdaps: If init failed, tell at which step it happened.
- tp_base: Allow additional inputs and optional outputs to EC requests.
- tp_base: Decrease timeout for reply to request.
- Kconfig: Make TP_BASE an invisible tristate option
- tp_base: Rename tp_controller_trylock() to tp_controller_try_lock().
- tp_base: make tp_controller_lock() return an int; have it checked in 
  tp_smapi and hdaps.
- tp_base: Rewrote all printk()s using macros, and made debug=1 quieter.
- Makefile: work even if $SHELL is not bash.
- Makefile: Autopatch hdaps patch to work with 2.6.18-rc1

Many of the above changes were suggested or contributed by
Henrique de Moraes Holschuh.

0.21  2006-06-21
----------------
- Compatibility with kernel 2.6.17.
- hdaps: Fixed a locking bug. This further reduces (completely solves?) the
  EC hangs problem.
  (Thanks to Rasto, Andrew and Whoopie for helping track this down!)
- hdaps: Support for ThinkPad T60: added 0x04 as valid initial latch value
- hdaps: Mousedev poll timer was not restarted after resume from suspend.
- tp_base: Added verbose debug output, enabled by the "debug=1" module
  parameter (and by "make load DEBUG=1")
- tp_base: Cosmetic fix to tp_base DMI vendor reporting (patch by Whoopie).
- Cleaned up some printk()s.
- Makefile: "make load" now enables dmesg debug output only if given DEBUG=1.
- Makefile: will not delete hdaps.c if edited manually after generation.

0.20  2006-04-30
----------------
- Makefile: fix for "linux/tp_base.h: No such file or directory" problem.
- tp_base: whitelist "LENOVO" vendor string too.
- hdaps: add "force=1" parameter to force loading on models that are
  not whitelisted (patch by Whoopie).
- hdaps: fix init on several models by using tp_controller_read_row()
  instead of direct port IO. This apparently also makes the initial latch
  value is meaningful even when the device is already initialized.
- Now requires kernel 2.6.15 or newer (compatibility cruft removed).

0.19  2006-04-08
----------------
- Added new battery status attributes:
  /sys/devices/platform/smapi/BAT?/first_use_date
  /sys/devices/platform/smapi/BAT?/manufacture_date
- Bugfix in tp_base embedded controller readout 
  (missing prefetch invalidation). This apparently solves remnants of the
  the embedded-controller-lockup problem.
- "make load" now demands "HDAPS=1" if the hdaps module is loaded.
- hdaps driver patch:
  - Remember keyboard/mouse activity for 0.1sec, so that all applications
    get a chance to read it (the hardware resets its activity flag
    when it's read).

0.18  2006-04-01
----------------
- Solved embedded controller lockups with HDAPS=1 (see hdaps comments below).
- Added new function tp_controller_try_read_row() to tp_base.
- Added extra status checks in tp_base, to catch abnormal conditions earlier.
- Restructured several functions in tp_base for clarity and reusability.
- Restructured tp_smapi attribute registration code using fancy macros, to 
  remove redundancy.
- Reduced prefetch delay and maximum retries in tp_base.
- Now locks tp_controller when making SMAPI call (just in case).
- Makefile change to fix compilation in Debian.
- Minor cleanups and added comments in tp_smapi.c.
- Major changes to hdaps patch:
  - Whenever we read data from the controller, parse all of it and
    remember the values in global vars.
  - Simplified the device model *_show functions, via the above change.
  - If the mousedev poll timer handler experiences a transient fault, use
    use last saved data from the global vars.
  - Handle delayed calibration at first opportunity, not just mousedev polls.
  - Disable mousedev poll timer when suspending, to prevent readouts before
    the resume code re-initializes the sensor.
  - Don't accept initial status 0x00, it was probably an artifact of 
    the premature mousepoll invocation.
  - When doing a sensor readout in mousedev poll, avoid time-consuming
    retries and fetches, to minimize time in softirq. This effectively 
    exorcises an embedded controller lockup Heisenbug. There may still 
    a nonzero probability of lockup, but now it's not worse than the 
    vanilla hdaps (since they do essentially the same).

0.17  2006-02-10
----------------
- Fixed off-by-one bugs in charge threshold handling + cleanup
- Enforce bounds on threshold the same way Battery Maximizer does
- Recognize more battry status values
- Fixes in error reporting
- Fixed a warning in tp_smapi.c
- hdaps driver patches:
  - Informatively report init errors
  - Recognize 0x00 as a valid latch value
  - Increase init timeouts

0.16  2006-01-11
----------------
- Renamed smapi/BAT?/force_discharge1 to smapi/BAT?/force_discharge
- Removed smapi/BAT?/force_discharge2 (doesn't work on any model)
- HDAPS patch changes:
  - Check ready status (instead of occasionally returning junk) also for 
    variance, not just for position.
  - Calibrate unsynchronously (reduces load and resume time by 1-2 seconds)
  - Fix position readouts - read word, not byte.
  - Fixed driver lockup when writing to hdaps/variance.
- Changes in diff handling.

0.15  2006-01-10
----------------
- Bugfix: 0.14 had broken SM BIOS call code.

0.14  2006-01-09
----------------
This version has no user-visible functionality changes, but improves
the reliability of coordinating hardware access with the hdaps driver.
- Moved controller access code to tp_base.
- Changes in HDAPS driver patch:
  - Use tp_base for controller access instead of direct port IO. Using
    the "row read" and "prefetch" abstraction of tp_base makes the hdaps
    code shorter, clearer and safer.
  - Added checking of the STATUS port and automatic retries if device is not
    ready (previously it just returned junk values).
  - Added missing lock in hdaps_invert_store.
  - A few local code simplifications
  - Redue the retry delay (200ms was excessive).
- Changed SMAPI calls to use 32-bit args.
- Made "make patch" produce an LKML-compliant patches
  (include a diffstat, remove CD_SPEED, remove <2.6.15 #ifdefs).
- Changes in Makefile and directory structure.

0.13  2005-12-21
----------------
- First step toward resolving conflict with the hdaps module:
  - Added a tp_base module, which handles coordination of access to the 
    ThinkPad controller (thanks to Alan Cox and Rovert Love).
  - Changed tp_smapi to require and use tp_base.
  - "make load HDAPS=1" and "make install HDAPS=1" will copy and patch hdaps.c
    for compatibility with tp_smapi and tp_base, and then load or install it.
- Added a "make patch" target, which creates a stand-alone patch against the
  current kernel tree (thanks, Spiney!).
- Future kernel compatibility: avoids platform_device_register_simple.

0.12  2005-12-16
----------------
- Added smapi/BAT?/force_discharge1 and
        smapi/BAT?/force_discharge2.
  When set to 1, they stop forces discharging of battery even if on AC power.
  The two files have the same functionaliy but use different SMAPI calls to
  achieve it. On ThinkPad T43 only force_discharge1 works; the other one is
  completely untested.
- Removed smapi/cd_speed (unless you set "#define PROVIDE_CD_SPEED"), since
  its function is provided in a safer way by the combination of "hdparm -E"
  (for CD) and speedcontrol (for DVD, http://safari.iki.fi/speedcontrol.c).

0.10  2005-12-13
----------------
- Added smapi/BAT?/state (idle/charging/discharing).
- Added smapi/BAT?/{power_now,power_avg}.
- Renamed smapi/BAT?/{current1,current2} to {current_now,current_avg}.
- If stop_charge_thresh is unsupported, when trying to set it don't affect
  start_charge_thresh.
- smapi/BAT?/{design_capacity,last_full_capacity} were reversed.
- smapi/cdrom_speed renamed to cd_speed, and safety mechanism added: you must 
  use "echo 1 yes_crash_my_computer > /sys/devices/platform/smapi/cd_speed".
- Added smapi/ac_connected added (actually in 0.09).

0.09  2005-12-12
----------------
- Dual-battery support: moved all battery-related sysfs files to
  /sys/devices/platform/smapi/BAT0/*
  and made the 2nd battery accessible via 
  /sys/devices/platform/smapi/BAT1/*
- Added numerous read-only battery status files:
  /sys/devices/platform/smapi/BAT?/{installed,cycle_count,current1,current2,
  last_full_capacity,remaining_capacity,design_capacity,voltage,
  design_voltage,manufacturer,model,serial,barcoding,chemistry}
  These are incompatible with HDAPS - see README.
- /sys/devices/platform/smapi/BAT?/dump gives the raw status dump.
- Added "debug" module parameter, default (debug=0) reduces printk verbosity.
  For bug reports, please use "modprobe debug=1" (or just use "make load").
- Now requires kernel >= 2.6.13 (stick with v0.08 for earlier kernels).
- Suspend+resume now correctly handles default thresholds (start==stop==0).
- Extended whitelist to cover all ThinkPads.
- Cleanup of init/probing code.
- Reduced kernel logging level
- Fix in set_inhibit_charge

0.08  2005-12-09
----------------
- Fixes in README and dmesg outputs.

0.07 2005-12-07
----------------
- Added /sys/devices/platform/smapi/cdrom_speed to set/get the CD drive speed
  level (0=slow, 1=medium, 2=fast).
  WARNING: 
  Writing to this file when the CD is being accessed will hang your computer.
- Fixed some dmesg outputs

0.06  2005-12-06
----------------
- /sys/devices/platform/smapi/inhibit_charge renamed to
  /sys/devices/platform/smapi/inhibit_charge_minutes and now accepts
  the number of minutes to inhibit charging.
- Compatibility with kernel 2.6.12 (thanks to Guenther Starnberger)

0.05  2005-12-05
----------------
- Kernel 2.6.15 compatibility (thanks to Volker Gropp)
- Improved error reporting
- Cleared confusing dmesg output

0.04  2005-12-05
----------------
- Made start_charge_thresh work even with stop_start_thresh is
  not available.

0.03  2005-12-05
----------------
- Added /sys/devices/platform/smapi/inhibit_charge
- Fixed #includes
- Kernel 2.6.15 compatibility
- Added versionless IBM machines to DMI whitelist
- Added retries to SMAPI request code
- Added list of supported models to README

0.02  2005-12-05
----------------
- improved SMAPI request code and error reporting

0.01  2005-12-04
----------------
- initial release