Skip to first unread message
Zoraze Ali
Jun 17, 2016, 6:07:52 PM6/17/16
to ns-3-users
Hi,
I am trying to view the GUI for the ConfigStore while executing lena-simple-epc.cc but it is crashing due to the Null pointer assignment for the attribute "RootQueueDiscList" of TrafficControlLayer class. I am attaching an example script to reproduce the crash and the NS_LOGs of attribute-iterator.cc ( you can find it under ns-3.25/src/config-store/model) where the crash happens. I am also attaching the back logs from gdb for the reference.
With the help of NS_LOGs and debugging i know that the SIGSEGV error is due to the null pointer returned by " Ptr<Object> tmp = (*it).second; " line 237, in DoIterate (Ptr<Object> object) function of attribute-iterator.cc. But unfortunately i was unable to resolve the issue. I will be very thankful if someone could look into it and help me to resolve it. I am using ns-3.25.
Kind regards,
Zoraze
I am trying to view the GUI for the ConfigStore while executing lena-simple-epc.cc but it is crashing due to the Null pointer assignment for the attribute "RootQueueDiscList" of TrafficControlLayer class. I am attaching an example script to reproduce the crash and the NS_LOGs of attribute-iterator.cc ( you can find it under ns-3.25/src/config-store/model) where the crash happens. I am also attaching the back logs from gdb for the reference.
With the help of NS_LOGs and debugging i know that the SIGSEGV error is due to the null pointer returned by " Ptr<Object> tmp = (*it).second; " line 237, in DoIterate (Ptr<Object> object) function of attribute-iterator.cc. But unfortunately i was unable to resolve the issue. I will be very thankful if someone could look into it and help me to resolve it. I am using ns-3.25.
Kind regards,
Zoraze
Tommaso Pecorella
Jun 18, 2016, 7:09:31 PM6/18/16
to ns-3-users
Why are you trying to configure the attributes *after* the simulation is ended ?
Still, you have found a bug. Please follow the bug resolution here:
T.
Zoraze Ali
Jun 18, 2016, 8:04:06 PM6/18/16
to ns-3-users
Hi Tommaso,
Thanks for looking into it and filing the big report. I agree the Config call should be before the Simulator::Run, but with LTE simulations UEs get connected to the eNB at run time, therefore, invoking the call before Simulation:Run will not show the context of the connected UEs which every eNB maintain under its attribute "UeMap",
Kind regards,
Zoraze
Thanks for looking into it and filing the big report. I agree the Config call should be before the Simulator::Run, but with LTE simulations UEs get connected to the eNB at run time, therefore, invoking the call before Simulation:Run will not show the context of the connected UEs which every eNB maintain under its attribute "UeMap",
Kind regards,
Zoraze
Zoraze Ali
Jun 18, 2016, 8:08:02 PM6/18/16
to ns-3-users
Sorry for the Typo, i mean bug not big.
Tommaso Pecorella
Jun 19, 2016, 5:30:18 AM6/19/16
to ns-3-users
I see... well, in that case you'll see the attributes after the simulation, and changing them have no point. Still, you'll see what they have been set to.
Cheers,
T.
Zoraze Ali
Jun 19, 2016, 6:47:08 AM6/19/16
to ns-3-users
Yes, I just wanted to see the configured attributes.
Kind regards,
Zoraze
Stefano Avallone
Jun 23, 2016, 1:34:41 PM6/23/16
to ns-3-users
It looks like that the script only crashes when GtkConfigStore is enabled. And the reason appears to be that, despite RootQueueDiscList is empty, something tries to access a field of a map element (Ptr<Object> tmp = (*it).second;). Thus, I guess the bug does not lie within the traffic control but within the code that attempts to do that (GtkConfigStore?)
Stefano
Zoraze Ali
Jun 24, 2016, 6:10:58 PM6/24/16
to ns-3-users
Hi Stefano and Tom,
Thanks alot for solving this bug!. I have been following the bug resolution at bugzilla. I am sorry i didn't contributed in the solution as, i didn't have much deep understanding of the cause. But i learned from the explanation on bugzilla. I have tested the solution with my script and the simulation ends successfully.
Kind regards,
Zoraze
Thanks alot for solving this bug!. I have been following the bug resolution at bugzilla. I am sorry i didn't contributed in the solution as, i didn't have much deep understanding of the cause. But i learned from the explanation on bugzilla. I have tested the solution with my script and the simulation ends successfully.
Kind regards,
Zoraze
Reply all
Reply to author
Forward
0 new messages