[Nut-upsdev] upsd segfault.

Kjell Claesson kjell.claesson at epost.tidanet.se
Mon Dec 22 15:53:09 UTC 2008


Den Monday 22 December 2008 14.46.50 skrev Charles Lepple:

> The code in http://boxster.ghz.cc/projects/nut/changeset/1429 handles
> things differently, but I am confused as to why the new code doesn't
> catch that case properly if the old code worked.

As I wrote in an earlier mail, I can't realy know if the fault existed
before. As normal I have the ups connected to the computer running
NUT, and in that case it has to be online when I start the upsd.

Then the segfault would not show. But now I have 2 ups'es connected
to the computer, one supplying the computer and the other is for testing.

So now I can start the upsd when this extra ups is offline, and get the
fault when I put it online.

>
> It may be something in the code that calls this function, too.

Think it has to do with the ALARM. But I can not understand the difference 
between having it online first and then shifting on and off and going from 
offline to online.

Maybe it is that the node Alarm is in the middle of the struckt and that it is 
removed. But when I have it online it add it at the end.

Have to check how the status look (ALARM OFF) when I start it in standby,
v.s. if I start it online and put it in standby.

And I don't understand why it is calling st_tree_node_add as it is going to 
remove the alarm node and the status alarm. 

/Kjell




More information about the Nut-upsdev mailing list