On Tue, 2016-02-16 at 16:28 -0500, Avery Pennarun wrote:

Since around the beginning of time, ath9k aggregates have timed out
after
5000 TU (around 5000ms) of inactivity, but nobody seems to be quite
sure
why, and this magic number seems to have migrated around from one
place to
another.  An openbsd mailing list recently had a patch to disable the
timeout completely, which they say matches some commercial routers:
https://www.mail-archive.com/tech@openbsd.org/msg29456.html

Even in Linux, several non-ath9k drivers default to no timeout
already.  I
think changing it directly to zero would be safe, but to allow a more
structured investigation, let's make it configurable for now.

I believe the original timeout came from some ancient Intel code.
Nobody gave it any thought at the time, and it was merged and preserved
to this day.

The reason for Intel specifically is that every aggregation session
(both RX and TX), even idle ones, takes up hardware resources, so
there's a limited need to drop completely idle sessions.

As a result, I believe that changing the default for every driver other
than iwlegacy, iwlwifi/dvm and iwlwifi/mvm would be perfectly fine.

johannes