[srslte-users] Segmentation fault during attach with mobile phone

Andre Puschmann andre.puschmann at softwareradiosystems.com
Wed Jul 26 11:45:09 UTC 2017


Umberto,



On 25.07.2017 18:06, Umberto Fattore wrote:
> Dear Andre, 
> 
> I tried what you suggested. Attached, you can find the new log file. 
> 
> cnfg->dl_am_rlc.t_status_prohibit always has value 0 (in the log, look at the number next to "step 6"). 
> For each rrc->add_bearer we have added a printf before and after the call and apparently it does not execute the function for the DRB1 setup. 

Ok, so the crash happens when setting up DRB1. That's good to know.

I think I accidentally asked you to print "dl_am_rlc.t_status_prohibit"
but really meant "poll_pdu", because this is where the segfault happens,
right? Could you try to print that value too?

Cheers
Andre

> 
> Best Regards, 
> Umberto
> 
> 
> -----Original Message-----
> From: Andre Puschmann [mailto:andre.puschmann at softwareradiosystems.com] 
> Sent: Donnerstag, 20. Juli 2017 22:34
> To: Umberto Fattore
> Cc: srslte-users at lists.softwareradiosystems.com
> Subject: Re: [srslte-users] Segmentation fault during attach with mobile phone
> 
> Hi Umberto,
> 
> On 20.07.2017 14:28, Umberto Fattore wrote:
>> Yes, I coded a printf after each instruction in the body of the 
>> function rlc_am::configure( … ) and I realized that (for an single 
>> UE), all printf are executed in sequence more than once. But, when the 
>> segmentation fault occurs, the last printf executed is the one before 
>> the  instruction: “poll_pdu = liblte_rrc_poll_pdu_num[“
> 
> Could you try to print the value of cnfg->dl_am_rlc.t_status_prohibit in that printf as well? I.e. "printf("%d\n",
> cnfg->dl_am_rlc.t_status_prohibit);"
> 
> It shouldn't go above 64 and is valid only below 56.
> 
> Also in srsenb/upper/rrc.cc in send_connection_reconf() there are a few calls to "parent->rlc->add_bearer()" could you add printfs before all of them too to see for which bearer it fails?
> 
> 
>>
>> {Side note: I checked again my subscription, but I result already 
>> subscribed}
> 
> Not sure what's going on there either but could you please try to turn off HTML when posting on the list?
> 
> Thanks
> Andre
> 
>>
>>
>> Greetings,
>> Umberto
>>
>>
>>  
>>
>> *From:*Andre Puschmann 
>> [mailto:andre.puschmann at softwareradiosystems.com]
>> *Sent:* Donnerstag, 20. Juli 2017 14:08
>> *To:* Umberto Fattore
>> *Cc:* srslte-users at lists.softwareradiosystems.com
>> *Subject:* RE: [srslte-users] Segmentation fault during attach with 
>> mobile phone
>>
>>  
>>
>> Hi,
>> The RLC AM configure function is called whenever an bearer in 
>> acknowledged mode (AM) is created. By "runs for more times" you mean 
>> that it's called multiple times when a single UE attaches before it crashes?
>>
>> On a side note, could you please register for the list before you post 
>> the next time?
>>
>> Thanks
>> Andre
>>
>> On July 19, 2017 2:16:36 PM GMT+02:00, Umberto Fattore 
>> <Umberto.Fattore at neclab.eu <mailto:Umberto.Fattore at neclab.eu>> wrote:
>>
>> Hi Andre,
>>
>>  
>>
>> Thank you for answer. Yes, I have logs obtained running with gdb. You 
>> can find them attached in this message.
>>
>> In my understanding, it seems that the segmentation fault happens in 
>> function rlc_am::configure( … ). In particular, I notice that this 
>> runs for more times correctly, but at a certain point it arrives only 
>> at 2^nd code line (poll_pdu = liblte_rrc_poll_pdu_num[ … etc  …] ) and 
>> then segmentation signal is detected.
>>
>> Thanks Again,
>> Umberto
>>
>>  
>>
>> *From:*Andre Puschmann 
>> [mailto:andre.puschmann at softwareradiosystems.com]
>> *Sent:* Dienstag, 18. Juli 2017 22:41
>> *To:* Umberto Fattore; srslte-users at lists.softwareradiosystems.com
>> <mailto:srslte-users at lists.softwareradiosystems.com>
>> *Subject:* Re: [srslte-users] Segmentation fault during attach with 
>> mobile phone
>>
>>  
>>
>> Hi Umberto,
>> In case of a segfault it's difficult to rely on logs or pcap because 
>> they may not contain the "last messages" before death. Do you have a 
>> chance to run with gdb and tell where the segfault happens?
>>
>> Thanks
>> Andre
>>
>> On July 18, 2017 2:10:07 PM GMT+02:00, Umberto Fattore 
>> <Umberto.Fattore at neclab.eu <mailto:Umberto.Fattore at neclab.eu>> wrote:
>>
>> Dear all,
>>
>> I am encountering a problem with srslte software. I am trying to make 
>> an attach with my mobile phone. Obviously, the board I am using as 
>> eNodeB is connected to an MME (OpenEPC MME on VMWare virtual machine). 
>> If I try an attach with my phone, the S1AP exchange between eNodeB and 
>> MME is not able to be completed, because the srslte crashes for a segmentation fault.
>>
>> Could you help me understing the reason of this ?
>>
>> I attach in the message both the Wireshark capture and the log srslte, 
>> that I suppose could help figure out the thing.
>>
>>  
>>
>>  
>>
>> Thanks,
>>
>> Umberto
>>
>>  
>>
>>  
>>
> 
> 
> --
> Andre Puschmann
> 
> Software Radio Systems (SRS)
> http://www.softwareradiosystems.com
> 
> PGP/GnuPG key: 6C42AB31
> fingerprint: 137A AE49 785B A445 257C 8AD7 D877 A498 6C42 AB31
> 


-- 
Andre Puschmann

Software Radio Systems (SRS)
http://www.softwareradiosystems.com

PGP/GnuPG key: 6C42AB31
fingerprint: 137A AE49 785B A445 257C 8AD7 D877 A498 6C42 AB31


More information about the srslte-users mailing list