Nut upsuser - Sep 2014 - upd and OpenBSD (was Nut 2.7.2 on OpenBSD 5.6 with APC USB UPS)

* Charles Lepple <clepple at gmail.com> [2014-09-20
01:12]:
> On Sep 19, 2014, at 5:18 PM, Stan Gammons <sg063015 at gmail.com>
wrote:
> > Looks like this patch fixed the problem.  I see the status change now.
> Good to hear.
> I'm not familiar with the OpenBSD release process. Is that patch on
track to be a part of the final 5.6 release?
no, 5.6 is already cut. 5.7 will have it.

-- 
Henning Brauer, hb at bsws.de, henning at openbsd.org
BS Web Services GmbH, AG Hamburg HRB 128289, http://bsws.de
Full-Service ISP - Secure Hosting, Mail and DNS Services
Dedicated Servers, Rootservers, VMs/PVS, Application Hosting

* Henning Brauer <hb-nut at bsws.de> [2014-09-20
13:39]:
> * Charles Lepple <clepple at gmail.com> [2014-09-20 01:12]:
> > On Sep 19, 2014, at 5:18 PM, Stan Gammons <sg063015 at
gmail.com> wrote:
> > > Looks like this patch fixed the problem.  I see the status change
now.
> > Good to hear.
> > I'm not familiar with the OpenBSD release process. Is that patch
on track to be a part of the final 5.6 release?
> no, 5.6 is already cut. 5.7 will have it.
scratch that, thinko. while 5.6 is already cut this won't make 5.7
either.

We have upd(4) attaching to these devices now. Either nut needs a
driver getting status from upd (trivial, will show up in the
hw.sensors sysctl tree) or we need some way for nut still being able
to talk to the ups directly, without upd and nut interfering with each
other.

-- 
Henning Brauer, hb at bsws.de, henning at openbsd.org
BS Web Services GmbH, AG Hamburg HRB 128289, http://bsws.de
Full-Service ISP - Secure Hosting, Mail and DNS Services
Dedicated Servers, Rootservers, VMs/PVS, Application Hosting

On Sep 23, 2014, at 3:55 AM, Henning Brauer <hb-nut at bsws.de> wrote:
> We have upd(4) attaching to these devices now. Either nut needs a
> driver getting status from upd (trivial, will show up in the
> hw.sensors sysctl tree) or we need some way for nut still being able
> to talk to the ups directly, without upd and nut interfering with each
> other.
While a upd(4) driver is certainly technically feasible, and probably would not
be too complicated to write, I will leave that to someone else to work on. From
a quick glance at the upd(4) man page[*], it looks like a step backwards in
terms of available information, and I don't see any information about
sending commands to the UPS (to initiate an UPS self-test or shutdown, for
instance).

[*]
http://www.openbsd.org/cgi-bin/man.cgi/OpenBSD-current/man4/upd.4?query=upd&sec=4

If OpenBSD implements the logic behind the libusb_detach_kernel_driver() call
(does not seem to be upstream, as of libusb-1.0.18), this would allow users to
decide at runtime which method they would like to use for monitoring the UPS.

I have never really liked side-stepping the OS HID drivers, so if there is a way
to do this from userspace with uhid(4) instead of ugen+libusb, that would be
handy. But that's a lot of driver rewriting.

In the mean time, it looks like we will have to document around this. Back to
one of my earlier questions, what is the status of that patch that Stan posted
from Stuart Henderson? From 5.6 onwards, is that sufficient for users to apply,
or is there another patch needed to deconflict ugen and upd?

-- 
Charles Lepple
clepple at gmail