[omniORB] omni_thread memory leak?

Paul Nader naderp@d22mail.alcatel.com.au
Fri, 20 Mar 1998 09:09:44 +1100 (EST)


Tristan,

The platform is HPUX 10.20, using libcma.a (v 1.5). The problem still
remains even after _dispatch calls locks lock prior to calling wait.
I havent't been able to test it on other platforms and I suspected it
might be something specific to HP's implementation of pthreads. I'll take
it up with them then and let you know how I go.

BTW : the version of omnithread I'm using is the one relased in the
      2.5.0 beta (i.e prior to final release).

Regards, Paul.

> 
> > Is there a known bug iun relation to omni_thread not releasing memory
> > when a run_undetached method terminates?
> 
> Which platform are you running it on?  We've tried your example quickly on
> Solaris and NT and it doesn't seem to leak, although we did find a bug in your
> Object::_dispatch() function in that the lock mutex isn't held when
> shut.wait() is called.
> 
> Also there was a resource leak on NT in earlier versions of omnithread but
> this has been fixed.
> 
> Cheers
> 
> Tristan
> 
> +--------------------------------------------------------------------+
> |  Tristan Richardson                 Email:  tjr@orl.co.uk          |
> |  ORL                                  Tel:  +44 1223 343000        |
> |  24a Trumpington Street               Fax:  +44 1223 313542        |
> |  Cambridge, CB2 1QA, UK               WWW:  http://www.orl.co.uk/  |
> +--------------------------------------------------------------------+
> |          ORL - The Olivetti & Oracle Research Laboratory           |
> +--------------------------------------------------------------------+
>