root/lm-sensors/tags/i2c-2-8-km2/TODO

LM-Sensors TODO list
Contact us if you have comments or wish to help.
------------------------------------------------

I2C CODE CHANGES
================

See TODO in the i2c package.


KERNEL MODULES
==============

* Assistance on how to fix module refcounting is welcome.
  To my understanding, it does not handle client->adapter->owner
  correctly now, and it is possible to rmmod adapter while one of its
  clients is in use.

* mkpatch. Trace back Configure changes from 2.5 tree and
  have it make clean diffs.
  Applies to both i2c and sensors.

* ACPI and SMBus host
 - ACPI subsystem may access SMBus host too. Locks?
 - On my ancient board using i2c-via, suspend and power switch status
   seems to be in the same register with SCL and SDA.

* Suspend-to-disk, losing Vcc
 - Reset i2c client to avoid alarms and SCI interrupts. Currently
   sensors do not create interrupt with alarms.

* Return from suspend
 - Reinitialize i2c client.

* 2.4 / 2.5 cleanup
  See adm1021, lm75, i2c-amd756, and i2c-amd8111 in kernel 2.5.54 for examples
        - Remove inc_use and dec_use in adapters and chips, convert to owner.
          Modules shouldn't adjust their own refcounts.
        - Convert PCI drivers to use pci_driver structs,
          registration, etc.
        - Use module_init and module_exit
        - Check return value of request_region(), don't use check_region()
        - C99 initializers
        - Remove excess #includes
        - interrupts/locking
        - Remove #ifdef MODULE, THIS_MODULE, MODULE_LICENSE, etc.
        - Remove EXPORT_NO_SYMBOLS
        - Remove other 2.1/2.2 #ifdefs

* Handle -1 returns from i2c layer rather than just delivering 0xFF to
  /proc and libsensors; report errors through /proc 'alarms' or
  new 'fail' entry?
* Add string function in sensors.o
* ALL: cleanup_module is void; check also that cleaning up is done
  properly, now we know that the module will be unloaded whatever we
  do.
* ALL chip drivers: add a readonly insmod option so we don't cause APM/ACPI
  to go insane.
* gl518sm: using iterate==2 causes a tread to be started. this thread remains
  even after setting iterate==1 and/or removing the gl518sm module
  <koenig@uranus.tat.physik.uni-tuebingen.de>
* LM78 detection: Tom Webster has proven that the reset bit in the 
  id register can be one (strange!)
* maxilife.c: Introduce new insmod variables
* maxilife.c: Round REG_TO_VID correctly
* icspll.c: Rewrite. The current implementation simply can't work at all.
  It needs i2c-level access (too bad for SMBus-only adapters).
* w83781d: Some features are still unsupported for the W8378[23].
  Also, alarms seem to give strange results
  sometimes, and there are some other minor problems, as indicated by
  Jonathan Bradshaw <jonathan@NrgUp.Com>
* Support 10-bit addresses. At this moment, they are supported nowhere, except
  in Simon Vogl's i2c modules.
* Better general locking, or at least a once-over to check no race-conditions
  remain. This is part of the SMP-safeness, and can better be done at once.
* adm9240: check whether the current voltage computations are correct.
  Probably not, as they are different from the datasheet specifications.
  Also check for the supported dallas chip.
* lm80: Check how OS alarms work. At page 17 of the sheet, it tells
  something completely different from the description at page 25.
* gl518sm: Assume that new values are close to old values, so start with
  trying to use a small range near the old values (already partially done)
* lm77: Add detection in lm75.c and sensors-detect to distinguish from
  an lm75. Add support for the two new registers.
* Add MTP006F chip: http://www.myson.com/Pcd/MTP006/Mtp006.pdf
  gfiala@s.netic.de (Guido Fiala) has one on his mainboard


LIBRARY
=======

* Easier use of /proc and libsensors without knowing the chip type
  At the moment a programmer wishing to use the lm_sensors package needs to know
  what chips are supported and needs to add all those chips into his package.
  If you were to add a field or function to your structures to the effect of
  sensor-function a programmer could use the package a lot more simply.
  For values you could do something like -
    enum sensor-type { volt, fan, temp, volt_min, volt_max, fan_div, fan_min,
    temp_over, temp_hyst }

* reload does not work (Bison/Flex problem?).
* Some library routines are pretty inefficient right now.
* Library should be split in more separate files, for better linking
  behaviour.
* Some adm9240 labels start with a digit; change this, it leads to confusing
  syntax in the conf file (they have to be quoted)


PROGRAMS
========

* Sensors program needs to print a + before lm75 temperatures
* Sensors program needs to be extended for many new chips


OTHERS
======

* Examine watchdog-4.3 or later of meskes@debian.org (Michael Meskes)
  tsx-11.mit.edu /pub/linux/sources/sbin or 
  sunsite.unc.edu /pub/linux/system/daemons/watchdog