Re: [Make-wifi-fast] Make-wifi-fast Digest, Vol 29, Issue 9

["Toke Høiland-Jørgensen" <toke@toke.dk>]

Jon Pike <jonpike54@gmail.com> writes:

> Hey Toke...
>
> Did you notice a guy responded to your commit on LEDE-DEV?  He's reporting
> no speed problems on his C2600...   This is turning into a head
> scratcher.

Yeah, that is odd. Not really sure what to do about it at this point;
maybe just leave it in and see if more people complain?

> I was going to finally get back to this thread and suggest that at
> least the C5/7 special build that rOOt manages on the LEDE forum could
> benefit from this, (uh, except for that pesky lack of latency
> reduction?) after it gets reverted from the nightly.
>
> Now, I'm just wondering what benifits there would be with this. I
> realize I've pretty much forgotten now what kinds of bufferbloat
> reduction and or airtime fairness bits ath10k had before, or would
> have with this. So if you wouldn't mind going over that again, at
> least I'd have the right info to pass along on what the benifits would
> be... assuming it's reccomendable at this point.

I don't have any ath10k cards in my own testbed, so I don't have a good
dataset for the potential gains. Not sure if someone else on the list
has done comparisons and can share results?

-Toke