Nut upsdev - Apr 2006 - NUT and debuggers (was: newhidups crashes intermittently (double free or corruption))

2006/4/4, Niels Baggesen
<nba@users.sourceforge.net>:
> On Tue, Apr 04, 2006 at 10:57:51PM +0200, Nick Rosier wrote:
> > Again, with an unstripped version of newhidups. But it doesn't
seem to
> > give more information; no core dump either. Anything else I can try?
> > Strace it and hope the file doesn't fill up my disk :-)
>
> It is probably better to try to run it from gdb, and then get a
> traceback when it fails.
>
> But if it really if memory corruption, it may have happened at any time
> before it is detected.
>
> Did anybody ever try compiling wnd running ith a memory debugger (dmalloc,
> valgrind, ...)
Russell did (and traced) some things: grep "valgrind" and you'll
have
a bunch of traces...

Arnaud

On 4/5/06, Arnaud Quette <aquette.dev@gmail.com>
wrote:
> 2006/4/4, Niels Baggesen <nba@users.sourceforge.net>:
> > On Tue, Apr 04, 2006 at 10:57:51PM +0200, Nick Rosier wrote:
> > > Again, with an unstripped version of newhidups. But it
doesn't seem to
> > > give more information; no core dump either. Anything else I can
try?
> > > Strace it and hope the file doesn't fill up my disk :-)
> >
> > It is probably better to try to run it from gdb, and then get a
> > traceback when it fails.
> >
> > But if it really if memory corruption, it may have happened at any
time
> > before it is detected.
> >
> > Did anybody ever try compiling wnd running ith a memory debugger
(dmalloc,
> > valgrind, ...)
>
> Russell did (and traced) some things: grep "valgrind" and
you'll have
> a bunch of traces...
Where are you referring to?

There is a paragraph in docs/developers.txt on valgrind.

I ran the newhidups driver under valgrind a while back, and I was
unable to suppress all of the spurious libc messages (things from
optimized string operations, mostly). Do you have any suggested
command lines?

--
- Charles Lepple