[Make-wifi-fast] possible TSQ patch for better aggregation

[Make-wifi-fast] possible TSQ patch for better aggregation

[Dave Taht]

Eric strikes again, with a TSQ improvement that might account for some
of the issues we've encountered with testing single tcps against wifi
clients:

https://www.mail-archive.com/netdev@vger.kernel.org/msg129509.html

I've got a pic showing the issue somewhere - basically flows to the
routers did poorer aggregation than flows through the router, even
with only minimal cpu in use.

thx eric!

-- 
Dave Täht
Let's go make home routers and wifi faster! With better software!
http://blog.cerowrt.org

Re: [Make-wifi-fast] possible TSQ patch for better aggregation

[Eric Dumazet]

On Thu, 2016-09-22 at 12:29 -0700, Dave Taht wrote:
> Eric strikes again, with a TSQ improvement that might account for some
> of the issues we've encountered with testing single tcps against wifi
> clients:
> 
> https://www.mail-archive.com/netdev@vger.kernel.org/msg129509.html
> 
> I've got a pic showing the issue somewhere - basically flows to the
> routers did poorer aggregation than flows through the router, even
> with only minimal cpu in use.
> 
> thx eric!
> 

You are very welcome, although I am not sure the patch actually changes
something in the case you described ?

Unless you run your TCP flows from the router itself ?