[srslte-users] srsUE USIM Support => Late Warnings

[Paul Sutton]

Hi David,

I've just pushed a branch which decouples the RLC from the PDCP in the
DL path called "dl_decouple".
Can you try it out and let me know how it goes?

Best regards,
Paul

On 05/12/16 11:21, David Rupprecht wrote:
> Hi Paul,
>
> thanks for your answer.
>
> Just for understanding: The NAS layer does *not* run in a separate yet?
> The DSP threads in the PHY layer call directly the upper layer functions
> without any decoupling and as the NAS layer takes so long the calling
> DSP thread blocks the further process?
>
> Just a short question to the message path. Considering the schematic
> picture [1], the DL_INFO message comes from the MAX DEMUX over the RLC
> layer to the NAS layer. However, looking at the code the RRC unit calls
> nas->write_pdu(). So for me the RRC_DL_INFO message uses a path with the
> RRC unit?
>
> For handing off the message to a separate thread. I think I have two
> options:
> - Running the whole NAS unit in a separate thread using and define an
> inter thread communication.
> - Starting a new thread when receiving the an authentication request for
> handling just this message.
>
> Which option would you recommend?
>
> Best regards,
> David
>
> [1]
> http://www.softwareradiosystems.com/wp-content/uploads/2015/08/srsue_layer.png
>  05.12.2016 10:33, Paul Sutton wrote:
>> Hi David,
>>
>> Have you tried handing off the message to a separate thread at the NAS
>> layer?
>> In the srsUE receive path, our high-priority DSP threads push messages
>> all the way up through the stack. It looks like the hard USIM delay is
>> blocking one of these threads for 43ms, resulting in RF timing misses.
>> We have multiple DSP threads in a pool which can take up the slack at
>> lower bandwidths but, as this is a 20MHz carrier, we can't afford to
>> lose one for them for that long.
>>
>> Best regards,
>> Paul
>>
>>
>> On 05/12/16 09:07, David Rupprecht wrote:
>>> Dear all,
>>>
>>> I have implemented USIM Support for the srsUE using the PCSClib.
>>> Unfortunately, the communication with the SIM Card takes to long (43 us
>>> in the log) and the Authentication Response can not be sent correctly
>>> indicated by RF late warnings with a crash.
>>>
>>> My question is, can I some how delay the Authentication Response? To my
>>> understanding that should be specification conform, as the communication
>>> on the NAS layer independent of the RF layer. Do you have any advises
>>> concerning this problem?
>>>
>>> Best regards,
>>> David
>>>
>>>
>>> _______________________________________________
>>> srslte-users mailing list
>>> srslte-users at lists.softwareradiosystems.com
>>> http://www.softwareradiosystems.com/mailman/listinfo/srslte-users
>> -- 
>> ________________________________________________________________
>> Paul Sutton Ph.D.
>>
>> Software Radio Systems (SRS)
>> http://www.softwareradiosystems.com
>>
>> +353-87-9813473 | paul at softwareradiosystems.com
>>
>> PGP Key ID: 3B4A5292
>> Fingerprint: B0AC 19C9 B228 A6EB 86E1 82B2 90C7 EC95 3B4A 5292
>> ________________________________________________________________ 
>>

-- 
________________________________________________________________
Paul Sutton Ph.D.

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

+353-87-9813473 | paul at softwareradiosystems.com

PGP Key ID: 3B4A5292
Fingerprint: B0AC 19C9 B228 A6EB 86E1 82B2 90C7 EC95 3B4A 5292
________________________________________________________________