[Cerowrt-devel] WNDR3800 improvements?

[Matt Taggart]

Aaron Wood writes:
> I thought it did better than that with Cake?

Yes, I was referring to newer cake, possible BQL improvements, possible 
ath9k improvements, newer kernel, the stuff listed at 
https://www.bufferbloat.net/projects/make-wifi-fast/wiki/Wifi_Stack_Rework/ 
 etc

I've also seen discussion of "policing" rather than full SQM, etc.

Basically, what is the state of the art we should be running on the 3800?


> On Tue, Dec 13, 2016 at 8:41 PM, Dave T=C3=A4ht <dave at taht.net> wrote:
> 
> >
> >
> > On 12/13/16 4:58 PM, Matt Taggart wrote:
> > > Hi,
> > >
> > > I love the WNDR3800 platform, it's been great over the years first with
> > > cerowrt and then openwrt. Of the many I've deployed I have only had
> > > hardware problems with 2 of them, and usually uptimes go over 100 days.
> > You
> > > can also still buy them used for $20 w/free shipping on amazon!
> > >
> > > With the recent improvement for cake,  make-wifi-fast, driver
> > improvements,
> > > etc is there any chance in seeing some of these things land for the
> > > WNDR3800 specifically? It would be really nice if this hardware could
> > > continue to do SQM, etc for some of the faster broadband speeds the cab=
> le
> > > providers are offering (comcast xfinity has 100, 200, 250 plans now) an=
> d
> > > see some of the wifi improvements too.
> >
> > We already support the 3800, although it peaks at 60 mbits of inbound
> > rate shaping. If your primary use case is wifi, with the latest fq_codel
> > code, you can live without inbound shaping and probably get 150mbits
> > well managed.


Thanks,

-- 
Matt Taggart
matt at lackof.org