[FS#44] Ath9k - TX performance regression on greater coverage settings - Task opened

[LEDE Bugs]

A new Flyspray task has been opened.  Details are below. 

User who did this - Koen Vandeputte (xback) 

Attached to Project - LEDE Project
Summary - Ath9k - TX performance regression on greater coverage settings
Task Type - Bug Report
Category - Base system
Status - Unconfirmed
Assigned To - 
Operating System - All
Severity - Medium
Priority - Very Low
Reported Version - Trunk
Due in Version - Undecided
Due Date - Undecided
Details - Synopsis:

When increasing coverage distance on the transmitting device, frame aggregation is heavily reduced.
Even when conditions are perfect.

This is regression compared to the BarrierBreaker release.


Hardware used:
2x cns3xxx GW-2388, each containing a single SR71-15


Setup:
- 2 devices are 1 meter apart.
- RSSI = -40
- Completely free channel
- Using wpa_supplicant, AES2, HT20 rates, IBSS


Repro rate:
100%

Repro steps:
- Set coverage distance on device 1 to >10000m
- Set Device 1 as iperf client
- Set Device 2 as iperf server

Obervations:
- iwinfo and iw wlanx station dump reveal 130Mbit/s raw link rate and 45Mbps net throughput
- The actual throughput is stuck @ 7Mbit/s
- Checking debugfs-xmit reveals that frame aggregation is at a rate of 2 frames/s
- When setting the coverage range of device 1 to 50m, the aggregation rate is easily >5000 frames/s
- Switching to HT40 yields the same issues.


Other observations:
- Setting coverage to 12000m and starting the iperf yields 7Mbit/s
- setting coverage to 50m and starting iperf yields >55Mbit/s
- Setting coverage to 50m, starting iperf and increase coverage to >10000m while running reduces the rate to 25Mbit.
  When stopping and restarting iperf it drops again to 7Mbit/s


In barrierbreaker release, increasing coverage distance to 10000m dropped to the throughput to ~50Mbit iso 7

More information can be found at the following URL:
https://bugs.lede-project.org/index.php?do=details&task_id=44