Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Toke =?utf-8?Q?H=C3=B8iland-J=C3=B8rgensen?= <toke@redhat.com>
To:     Sven Eckelmann <sven@narfation.org>, linux-wireless@vger.kernel.org
Cc:     ath10k@lists.infradead.org, Yibo Zhao <yiboz@codeaurora.org>
Subject: Re: ath10k: Missing airtime scheduler parts
In-Reply-To: <2631481.BEx9A2HvPv@ripper>
References: <2631481.BEx9A2HvPv@ripper>
Date:   Tue, 01 Dec 2020 11:56:57 +0100
Message-ID: <87a6uxx092.fsf@toke.dk>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk

Sven Eckelmann <sven@narfation.org> writes:

> Hi,
>
> I was asked what parts are currently missing for (a better) airtime fairness 
> with ath10k. I know that the AQL were merged and enabled for ath10k (and 
> mt76). But there was also the virtual time-based airtime scheduler [1] which 
> was proposed. I think the development on that one didn't continue since last 
> year.

I did recently rebase that patch to the current mac80211-next, actually.
Sent it off to Felix for some testing, but if you wan to give it a go I
can post it to the list as well :)

> Maybe someone else knows if there were some other parts worked on
> which I've am missed and which were not yet merged.

I don't think there have been any other patches being actively worked
on, but I seem to recall something about AQL (partly?) breaking airtime
fairness because the scheduler only keeps fairness between stations that
are actually scheduled, and AQL interrupting that...

-Toke