[Nut-upsdev] Important regression in usbhid-ups (r1113)

Peter Selinger selinger at mathstat.dal.ca
Wed Feb 20 17:08:30 UTC 2008


Hi Arnaud,

I noticed a similar problem: on APC, the new shutdown method actually
includes a delay of many seconds. Presumably the intende behavior was:
shutdown immediately, and return to power after n seconds. The
actual effect was the opposite: sit there for n seconds, then shut
down and return immediately.

The previous, model-specific shutdown routines had the advantage that
they actually did the correct thing for each model.

-- Peter

Arnaud Quette wrote:
> 
> while finishing some work on the HAL integration (I'm now mapping DBus
> method with NUT commands, to allow at least the final UPS poweroff), I
> was horrified to realize (and so late) the changes in
> usbhid-ups->upsdrv_shutdown().
> 
> We've lost a lot there, but making a generic code, instead of keeping
> the subdrivers delegation.
> 
> For example, the shutdown.return command (which is or should be the
> standard UPS poweroff method) has to set both delay.shutdown *and*
> delay.start does only set the latter.
> It results in an immediate UPS poweroff... The same applies to
> shutdown.return. And shutdown.stop should clear both shutdown and
> start delay to be sure.
> 
> The reason I previously kept the subdrv delegation (apart from the
> lack of visibility on other mfrs implementation) was the same as Arjen
> comment: the var. and cmd dealing with shutdown clearly have to be
> reworked since we have not a 1 to 1 mapping here.
> 
> This has made its way into 2.2.1 only (r1127) and the trunk (since r1113).
> 
> The fix will obviously have to go into 2.2.2, and the best would be to
> keep the generic method (not subdrv delegation).
> 
> So we have 2 possibilities (non exclusive. We may apply (1) quickly,
> while waiting for (2))
> 1) change upsdrv_shutdown() to proceed with setvar instead of (or
> with) the current instcmd
> 2) modify the hid2nut mapping to better deal with the shutdown related
> vars and commands.
> 
> Can you (all subdrivers maintainers) please ack this, make some tests
> on your side and report back your feeling about that.
> 
> Arnaud
> -- 
> Linux / Unix Expert R&D - MGE Office Protection Systems - http://www.mgeops.com
> Network UPS Tools (NUT) Project Leader - http://www.networkupstools.org/
> Debian Developer - http://people.debian.org/~aquette/
> Free Software Developer - http://arnaud.quette.free.fr/
> 




More information about the Nut-upsdev mailing list