[Nut-upsdev] Fun with TrippLite

Peter Selinger selinger at mathstat.dal.ca
Fri Dec 16 17:16:50 UTC 2005


Jonathan Freedman wrote:
> 
> Well this is cute.
> 
> After speaking with Charles a bit last night I ended up with some success.
> I made a few modifications to activate the tripplite subdriver and like
> magic it worked. Now the system in question hung last night for an
> entierly different reason, and now it doesn't work. But sometimes it does
> work....
> 
> Running upsdrvctl, when it works, is beautiful. Occasionally it will hang
> due to newhidups giving a segfault. When I run newhidups with debug level
> 1 or less, it segfaults. When I run it with 2 or more it's fine...

Can you find out where the segfault happens? I think the newest
development version might contain a memory allocation bug somewhere,
but I have not yet been able to track it down. If you compile with
debugging enabled (make CFLAGS="-I../include -O -Wall -Wsign-compare
-g" usb) or similar, you should be able to run the debugger on the
resulting core file (gdb drivers/newhidups core.XXXXX) to find out
where the segmentation fault happened.

-- Peter



More information about the Nut-upsdev mailing list