[FS#368] issues with airtime fairness on the ath9k

LEDE Bugs lede-bugs at lists.infradead.org
Thu Jan 5 06:11:20 PST 2017


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

User who did this - Dave Täht (dtaht) 

Attached to Project - LEDE Project
Summary - issues with airtime fairness on the ath9k
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 - The airtime fairness patch for the ath9k builds on top of the fq_codel code to provide fairness between stations, allocating similar amounts of airtime to everything on the network.

 - Device problem occurs on: unknown
 - Software versions of LEDE release: folded into lede dec 23 2016, and reverted jan 3 "due to stability issues"
 - Steps to reproduce: unknown
   
As features go, ATF is pretty killer - wifi network performance degrades much less as you add stations, particularly as you add more distant or unreliable stations. 
( https://blog.tohojo.dk/2016/06/fixing-the-wifi-performance-anomaly-on-ath9k.html)

Prior versions of the patch (for lede and kernel mainline) were extensively tested on platforms such as the archer c7v2, uap-lite, openmesh, wndr3800, and x86, in AP, client, and adhoc mode, with and without crypto enabled. 802.11s was not tested. Bridge mode was not particularly well tested either (at least, not by me). 

The test matrix for this stuff is hyperdimensional, and no doubt we missed something, somewhere. 

Some discussion at: http://lists.infradead.org/pipermail/lede-dev/2017-January/005132.html

* is it an RCU issue? (see discussion above)

* what is going wrong for people? (throughput? latency? falling off the net?)

* Is it a specific device or set of devices?

* Is it traffic patterns? (slow stations WILL get slower with this patch, fast stations get faster. UDP flooding traffic will behave differently. These are not bugs IMHO, but a feature)

* Is it something else? (I am having issues with ipv6 multicast, I think, there are several bugs filed on this topic)

More information can be found at the following URL:
https://bugs.lede-project.org/index.php?do=details&task_id=368



More information about the lede-bugs mailing list