> On 9 Feb, 2017, at 22:46, Dave Taht <[email protected]> wrote:
> 
> My original thought was we could leverage skb->hash just as fq_codel
> is now in linux head, but that seems to be exclusive of doing
> nat/conntrack...

We’d also still need to calculate additional hashes to get the endpoint 
addresses for the dual and triple modes.  Hard to see whether it would often be 
a win.

 - Jonathan Morton

_______________________________________________
Cake mailing list
[email protected]
https://lists.bufferbloat.net/listinfo/cake

Reply via email to