Sophie

Sophie

distrib > Mandriva > 2008.1 > x86_64 > media > main-release > by-pkgid > 19963f2f7ae3136d8ae7abee1eb1a862 > files > 50

nut-2.2.1-2mdv2008.1.x86_64.rpm

Desc: Issues and direction for the megatec driver
File: megatec.txt
Date: 29 March 2006
Auth: Carlos Rodrigues <carlos.efr@mail.telepac.pt>

Overall concept
===============

The "powermust" driver was originally developed to support the Mustek Powermust
series of UPSes, but since there are some other drivers implementing (mostly)
the same protocol, an idea came up to unify all these other drivers under a
renamed version of "powermust": "megatec".

The idea is to support all the hardware previously supported by the other
drivers with a minimum number of special cases. Most of the time this isn't
difficult, as the protocol states that all unknown/unimplemented commands
should be ignored (the commands are simply echoed back).

Target drivers
==============

The following drivers are targeted for "assimilation":

  * blazer (1)
  * fentonups (1)
  * mustek (1)
  * esupssmart (1)
  * ippon (1)
  * sms (1)
  * masterguard (3)

(1) megatec already supercedes this driver in functionality.
(2) megatec partially supercedes this driver in functionality, but some
    small features are missing (non-critical for normal operation).
(3) megatec partially supercedes this driver in functionality, but some
    important features are missing, which means some hardware may not be
    supported.

The above status is infered from some testing and from looking at the code
of the drivers. Although I have a high degree of certainty about it, more
testing is needed/welcome.

Notes
=====

The "masterguard" driver seems to be the most problematic: it implements a
"Q3" command (similar to "Q1" but with extra information), and the command
to get information from the UPS is "WH" instead of "I" (with a different
response also).

All drivers cancel the battery tests with "CT". The protocol seems to imply
that the 10 second battery test can't be cancelled (which makes sense).

The beeper can be toggled with "Q", but there doesn't seem to exist a way to
check if it is on or off, which may confuse users when a "beeper.on" command
actually turns it off and vice-versa (maybe some models default to "on", and
others to "off", this may make this command fit better as a blind
"beeper.toggle").

The "sms" driver implements a simple watchdog (send and "SxxRxxx" command to
power cycle the UPS every time the "reset.watchdog" command is called).
There are two battery tests: failure ("T" - 10sec?) and until LOWBATT ("TL").
Moreover, there are some "S.3" or "S.3R0003" commands which fly in the face
of "S05R0003" which the driver also issues (this last one matches all the
other drivers, the first two seem to imply that the UPS also accepts those
format variations to the standard "SxxRxxxx" shutdown+return commands).

-- 
Carlos Rodrigues