[Nut-upsdev] ups.firmware

Peter Selinger selinger at mathstat.dal.ca
Thu Aug 24 16:58:19 UTC 2006


Kjell Claesson wrote:
> 
> Hi all.
> 
> As the subject has come up. Powerware ups may contain
> several cpu's and different firmware versions.
> 
> Five of them is always ordered in a specific way.
> 
> 1 control panel
> 2 inverter
> 3 rectifier
> 4 net interface card
> 5 display control
> 
> So if a ups have a inverter cpu but not a control panel
> it sets the firmware revision to 00.00 to the control
> panel cpu, and get the version for the inverter.
> 
> At the moment the driver is fetching the first version it
> find. And if it have a control panel it set ups.firmware
> to the version of that.
> 
> One buggy thing is that if the ups is missing control panel
> it may give the firmware version of the communication board,
> as it skips the firmware with revision 00.00. 
> 
> So it would be nice with
> 
> ups.control.firmware
> ups.inverter.firmware
> ups.rectifier.firmware
> ups.interface.firmware
> ups.display.firmware

Surely this would be more logical:

ups.firmware.control
ups.firmware.inverter
ups.firmware.rectifier
ups.firmware.interface
ups.firmware.display

Another possibility is to let the driver construct a compound firmware
string.  How about ups.firmware = 'c:00.01 i:01.10 r:02.00 n:00.10
d:02.00'?  The firmware variable is an opaque string anyway, and as
far as I know, there is no restriction on its formatting. Is there a
reason why the driver shouldn't synthesize this string?

> Is this something that may apply to other ups'es then powerware ?

There is already a variable ups.firmware.aux, which is used by some
drivers.

-- Peter




More information about the Nut-upsdev mailing list