[PATCH] TALLOC_FREE(domain->check_online_event) refers to null pointer in winbindd_reinit_after_fork() [Was Re: [PATCH] clean the event context after fork in fork_domain_child()]

Jeremy Allison jra at samba.org
Tue Jan 13 01:43:21 GMT 2009


On Mon, Jan 12, 2009 at 07:17:30PM +0800, boyang wrote:
> boyang wrote:
> > Jeremy Allison wrote:
> >   
> >> On Mon, Jan 12, 2009 at 04:08:34PM +0800, boyang wrote:
> >>   
> >>     
> >>> Jeremy Allison wrote:
> >>>     
> >>>       
> >>>> On Sun, Jan 11, 2009 at 12:58:37PM +0800, boyang wrote:
> >>>>
> >>>>   
> >>>>       
> >>>>         
> >>>>> It is a pity that I have no testing or building environment around at
> >>>>> present, they are in company and the vpn is quite lagging. :-(  I'll
> >>>>> test the new code tomorrow(Monday). I'll get back to you as soon as I
> >>>>> finished the test.
> >>>>>     
> >>>>>         
> >>>>>           
> >>>> Oh, no problem. I will do the same, but your Monday comes
> >>>> before mine :-).
> >>>>   
> >>>>       
> >>>>         
> >>> Hi, Jeremy:
> >>>      There is still bug in winbindd_reinit_after_fork(), when
> >>> TALLOC_FREE(domain->check_online_event), the destructor is called, which
> >>> refers to null pointer, because te->event_ctx is set to NULL in
> >>> event_context_destructor(), which is called by reinit_after_fork(). So,
> >>> te->event_ctx->timer_events in timed_event_destructor() crashes the process.
> >>>
> >>> Patches are in attachment for v3-0-test and v3-2-test, please review them.
> >>>
> >>> After the fix applied, everything works for me very well. :-)
> >>>     
> >>>       
> >> Got it, pushed to 3.2 and 3.0, thanks ! I'm glad you
> >> got to your test environment first :-).
> >>   
> >>     
> > Hi, Jeremy:
> >      I am sorry to inform you that "make test" failed in
> > test_net_registry.sh and test_net_s3.sh... I am still try to catch it,
> > but have no idea at present. :-(
> >   
> failed in v3-2-test. net_registry rpc failed. :-(

I can't reproduce this either, it's passing for me (Ubuntu).

Jeremy.


More information about the samba-technical mailing list