> On Oct 12, 2016, at 12:21 , Dave Taht <dave.t...@gmail.com> wrote:
> I still defend the idea of the diffserv "squash" option cake once had.
> It was essentially RFC compliant, simple to use, and because iptables
> was too late on inbound, needed, no matter the layer violation.
As stated I am not sure whether that is a relevant layering violation to begin
with… I believe that Jonathan found a way to have a dscp qdisc stacked on top
of cake that does the squashing afterwards. I am uncertain whether that only
worked in theory (so whether research stopped after finding the dscp qdisc) or
whether that was actually tested to DTRT in allowing cake to use the internally
set DSC values but re-map them to 0 before putting them on the wire?
Cake mailing list