[Codel] fq_codel: revenge of the standing queue

[Eric Dumazet]

On Tue, 2012-09-04 at 12:51 -0700, Dave Taht wrote:

> - and we exit dropping state when that queue empties, when globally,
> across all the fq_codel queues, we still need to be dropping in order
> to get to the target.
> 

Not at all. You are trying to view fq_codel as a global controller,
which it is not.

> And that first_above time for that cvars for that fq_codel queue is
> reset to 0 for that fq_codel queue when it empties, forcing a recalc
> of the right interval (sojourn) for re-entering dropping state, with a
> "hands off" interval...
> 
> It's seems reasonably ok for a fq_codel queue to go empty for a while
> but not have to go through a sojourn again to start dropping. It makes
> sense to always deliver one packet after going empty...
> 
> these thoughts are half formed, and I did my damnedest to describe the
> behavior I was seeing.

Then you describe a codel bug.

But its not a bug. If a queue is empty, we set dropping to false, so
that next time we dequeue a packet, we can check how long queue was
empty.

Once again, this has nothing to do with fq_codel.