Thank you very much for the clarification. That makes a lot of sense. Do you happen to know where the mt76 driver stands with respect to these different phases? I am in the market for a new router, and due to the openness of the driver I'm currently looking at a mediatek based platform, namely the dir-860l rev b1.
On Sun, Jan 29, 2017 at 6:33 AM, Weedy <weedy2...@gmail.com> wrote: > On 28 January 2017 at 15:56, Rich Brown <richb.hano...@gmail.com> wrote: >> Hi Jaap, >> >>> On Jan 28, 2017, at 6:36 AM, Jaap Buurman <jaapbuur...@gmail.com> wrote: >>> Also about the following: "Improvements to the WiFi stack eliminating >>> bufferbloat on ath9k (and some ath10k) chipsets" >>> >>> Is this about the Airtime Fairness Patch? AFAIK this is only for ath9k >>> and not for ath10k? Or am I mixing things up here? >> >> I am not sure. I was hoping someone who has been following this more closely >> would weigh in. > > As far as I know the make-wifi-fast work was done in stages. Something like: > Phase 1: Reducing buffering all over the stack/drivers. Pretty sure > atk10k got some of this > Phase 2: Replace most/all txqueue stuff with fq_codel inside the > driver. I know Toke made patches for ath10k, I don't know if upstream > accepted them > Phase 3: Air time fairness stuff. Only for ath9k, and I don't know if > ath10k can ever get this since the chip firmware is basically a full > operating system and it does whatever it wants too. > > NB: I'm just a user, not a dev. But I like WiFi things and try to keep > informed. _______________________________________________ Lede-dev mailing list Lede-dev@lists.infradead.org http://lists.infradead.org/mailman/listinfo/lede-dev
