Bug 2531 - False retransmission on second layer when using MIMO
False retransmission on second layer when using MIMO
Status: RESOLVED FIXED
Product: ns-3
Classification: Unclassified
Component: lte
ns-3.26
PC Linux
: P5 normal
Assigned To: Biljana Bojović
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2016-10-26 10:52 UTC by Biljana Bojović
Modified: 2016-11-03 17:34 UTC (History)
2 users (show)

See Also:


Attachments
stats (760 bytes, application/zip)
2016-10-26 10:52 UTC, Biljana Bojović
Details
patch (778 bytes, patch)
2016-10-26 10:55 UTC, Biljana Bojović
Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Biljana Bojović 2016-10-26 10:52:28 UTC
Created attachment 2637 [details]
stats

The issue was originally detected by Zoraze Ali. 

Here is what he noticed:

"I am troubleshooting a behavior where i am receiving a NACK in spite of no errors in the reception when using MIMO (2 layers). I have logged ltespectrumphy and have noticed following behavior,

2nd column contains the UE node id, 

2.225s 2 LteSpectrumPhy:EndRxData(): [DEBUG] 0x2500dd0 No. of burts 1
2.225s 2 LteSpectrumPhy:EndRxData(): [DEBUG] 0x2500dd0 Expected TBs 2
2.225s 2 LteSpectrumPhy:EndRxData(): [DEBUG] 0x2500dd0 txMode 2 gain 0.524807
2.225s 2 LteSpectrumPhy:EndRxData(): [DEBUG] 0x2500dd0RNTI 1 size 1095 mcs 28 bitmap 12 layer 0 TBLER 0 corrupted 0
2.225s 2 LteSpectrumPhy:EndRxData(): [DEBUG] 0x2500dd0RNTI 1 size 1095 mcs 28 bitmap 12 layer 1 TBLER 0 corrupted 0
2.225s 2 LteSpectrumPhy:EndRxData(): [INFO ] 0x2500dd0 Packet of 1 layer 0
2.225s 2 LteSpectrumPhy:EndRxData(): [DEBUG] 0x2500dd0 RNTI 1 harqId 1 layer 0 size 1095 send DL-HARQ-ACK
2.225s 2 LteSpectrumPhy:ChangeState(): [LOGIC] 0x2500dd0 state: RX_DATA -> IDLE


In above logs, i am receiving the ACK for layer 0 but not for layer 1. Due to this, in MAC stats you can notice that there is a retransmission of the TB for layer 1 only.  Could you please comment if this is the right behavior? "
Comment 1 Biljana Bojović 2016-10-26 10:55:28 UTC
Created attachment 2638 [details]
patch
Comment 2 Biljana Bojović 2016-10-26 11:12:22 UTC
Just added a patch provided by Zoraze Ali. 

Current behavior:
In the cases when only one layer (first) is being used for transmission (out of two layers - MIMO), HARQ feedback will provide feedback for both layers, and since at the second layer nothing is  being received the default value will be used, which is NACK. This NACK will cause retransmission on the second layer.

The patch is changing the default value of the HARQ feedback value from NACK to ACK.

The patch is approved by Marco Miozzo. We will proceed toward committing it to ns-3-dev.
Comment 3 Biljana Bojović 2016-11-03 17:34:43 UTC
Pushed in changeset: 8a412f99b744