[FS#608] wds AP/STA disconnecting every 10 minutes

[LEDE Bugs]

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

User who did this - Enrico (enryIT) 

Attached to Project - LEDE Project
Summary - wds AP/STA disconnecting every 10 minutes
Task Type - Bug Report
Category - Base system
Status - Unconfirmed
Assigned To - 
Operating System - All
Severity - High
Priority - Very Low
Reported Version - Trunk
Due in Version - Undecided
Due Date - Undecided
Details - Devices:
AP WDS: Netgear r7800
STA WDS: TP-LINK Archer c7 v2

Problem exists on:
  - with both latest trunk and 17.01
  - with both ath10k and ath10k-ct
  - with STP enabled or not on both sides
  - with STP enabled or not on only one side
  - with both WPA2 tkip and aes
  - with 2.4Ghz and 5Ghz bands

Steps to reproduce
  - Configure AP WDS
  - Configure STA WDS
  - Wait **exactly 10 minutes** and you'll see this in the log:

AP:
daemon.info hostapd: wlan1-1: STA 30:b5:c2:08:b9:e1 IEEE 802.11: authenticated
daemon.info hostapd: wlan1-1: STA 30:b5:c2:08:b9:e1 IEEE 802.11: associated (aid 1)
daemon.info hostapd: wlan1-1: STA 30:b5:c2:08:b9:e1 RADIUS: starting accounting session F72C76B14648E205
daemon.info hostapd: wlan1-1: STA 30:b5:c2:08:b9:e1 WPA: pairwise key handshake completed (RSN)
STA:
daemon.notice netifd: Network device 'wlan1' link is down
kern.info kernel: [  387.179978] wlan1: deauthenticated from de:ef:09:f2:99:75 (Reason: 2=PREV_AUTH_NOT_VALID)
kern.info kernel: [  387.188344] br-lan: port 3(wlan1) entered disabled state
kern.info kernel: [  388.139233] wlan1: authenticate with de:ef:09:f2:99:75
kern.info kernel: [  388.170697] wlan1: send auth to de:ef:09:f2:99:75 (try 1/3)
kern.info kernel: [  388.195249] wlan1: authenticated
Wkern.info kernel: [  388.225092] wlan1: associate with de:ef:09:f2:99:75 (try 1/3)
kern.info kernel: [  388.235224] wlan1: RX AssocResp from de:ef:09:f2:99:75 (capab=0x431 status=0 aid=1)
kern.info kernel: [  388.243356] wlan1: associated
daemon.notice netifd: Network device 'wlan1' link is up

I think this has to do with wpa_supplicant in some way, maybe it doesn't receive the correct packets and force disconnect the client?

Now, this only happens in WDS bridge mode. I'm currently using relayd in order to overcome this issue and it works like a charm.
I want to solve this problem since relayd uses much more cpu power (indeed this is a true routing procotol after all) and total troughput is halved right now.

 

More information can be found at the following URL:
https://bugs.lede-project.org/index.php?do=details&task_id=608