> > In my case fd 0 was unexpectedly closed and tap interface opened on this fd. Can you elaborate?
> So, this assert was triggered. It is not a common case, I agree, but it is not > right anyway to expect, that we will never get fd 0 as a result of open(). > If you want to revert this patch you should forbid opening fd 0 at least > inside netdev-linux. > > Best regards, Ilya Maximets. > >> >> >> On Tue, Sep 29, 2015 at 5:07 PM, Gurucharan Shetty <shet...@nicira.com> >> wrote: >>> hash_2words takes unsigned int. I guess that is the problem? >>> >>> On Tue, Sep 29, 2015 at 5:00 PM, Alin Serdean >>> <aserd...@cloudbasesolutions.com> wrote: >>>> I think so. I think for once we should do >>>> https://github.com/openvswitch/ovs/blob/master/lib/poll-loop.c#L119-L123 >>>> before >>>> https://github.com/openvswitch/ovs/blob/master/lib/poll-loop.c#L115-L116 >>>> >>>> I am compiling under debug and attaching a debugger to it. >>>> >>>> Alin. >>>> >>>>> -----Mesaj original----- >>>>> De la: Gurucharan Shetty [mailto:shet...@nicira.com] >>>>> Trimis: Wednesday, September 30, 2015 2:34 AM >>>>> Către: Alin Serdean <aserd...@cloudbasesolutions.com> >>>>> Cc: Ben Pfaff <b...@nicira.com>; Ilya Maximets <i.maxim...@samsung.com>; >>>>> dev <dev@openvswitch.org>; Dyasly Sergey <s.dya...@samsung.com> >>>>> Subiect: Re: [ovs-dev] [PATCH] poll-loop: fix assertion in >>>>> poll_create_node >>>>> >>>>> Alin, >>>>> Reverting this commit on tip of master does not give me any unit test >>>>> failures. When you say there is more to it, do you mean that this commit >>>>> actually exposed some other bug? >> _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
