[Nut-upsuser] usbhid-ups and "data stale"

[Arjen de Korte]

>> First of all, it helps if you post the *full* line from the logs and
>> also show some context (how often these messages occur).
> That was the *full* line from the logs, except the syslog timestamp.
> But I hardly think the timestamp and the hostname could be at any use
> here.

We need the timestamp in order to find out what goes wrong and how often
it goes wrong. This is *very* relevant information.

[...]

>> Before all, upgrade to nut-2.2.1.
> Now, I've done the upgrade, after a few adjustments to the ports
> Makefile I will submit the patches to the FreeBSD ports collection, so
> everyone can make use of 2.2.1.

I don't know if you mentioned this before, but FreeBSD might be part of
the problem here. We have seen numerous problems with that in the past
because of the 'libusb' implementation is different from what is common in
Linux. Especially handling interrupt reports is handled different. I'm not
an expert in FreeBSD, so I don't know of the fine details.

> The problem is still there, after a few hours all the UPSes become
> stale, and just nothing happens until I restart the driver processes.

This sounds familiar.

> I've checked the logs, but there's nothing before/after the messages
> that is become stale, no USB errors, no driver notices, nothing at all.

This may be due to the way FreeBSD handles USB connection. I believe there
is something odd in the 'libusb' port for FreeBSD that causes this
behaviour. We've had some postings on this on either the users or
development mailinglist before.

Best regards, Arjen