[Nut-upsdev] Newhidups gets unbound after a while

[Arnaud Quette]

2006/5/8, Peter Selinger <selinger at mathstat.dal.ca>:
> I believe this is the bug that was fixed in r392 (April 5):
>
> ----------------------------------------------------------------------
> r392 | aquette | 2006-04-05 05:04:37 -0300 (Wed, 05 Apr 2006) | 3 lines
>
> newhidups: fix the "double free or corruption" crash upon device
>    reconnexion (Debian bug #354305)
> ----------------------------------------------------------------------
>
> However, the fix is a bit of cheating, because the device data
> structure now simply isn't freed at all.  I would still like to find
> out what caused this "double free" problem. I spent a few hours
> reading the libusb sources, and I could not find any memory management
> code that was suspicious.  Very unsatisfying. It works for now, but I
> think I will try to scrutinize this bug a little more when I have
> time, particularly since I think it did not occur before December or
> so. -- Peter

I fully agree that there is something strange there!
and that I wasn't much more satisfied with my "partial" solution.
I'll also try to have another look at it on my side, but if you find
something, that would be great.
I think that the problem appeared in a libusb release since december
as we have not done much on this part in newhidups...

Finally, 2.0.4 will soon be released as there is a good bunch of bug
fixes, so that will give us some time to bug hunt that libusb issue.

Arnaud
--
Linux / Unix Expert - MGE UPS SYSTEMS - R&D Dpt
Network UPS Tools (NUT) Project Leader - http://www.networkupstools.org/
Debian Developer - http://people.debian.org/~aquette/
OpenSource Developer - http://arnaud.quette.free.fr/