[Cake] [PATCH net-next 1/5] sch_cake: fix IP protocol handling in the presence of VLAN tags

[Sebastian Moeller]

Hi Jonathan,

thinking this over, I wonder whether a hypothetical carrier grade cake, might not actually grow a classify-by-vlan-priority keyword to allow switching over to using VLAN priority tags instead of dscps? That would avoid tempting carriers to re-map deeep-encapsulated dscps if they can just ignore them for good. And it scratches my pet itch, that 3 bits of classification should be enough for >80 % of the cases ;)

What do you think?

Best Regards
	Sebastian

P.S.: I reduced the CC list since I doubt that netdev is the right venue for mere hypotheticals ;)


> On Jun 26, 2020, at 15:11, Jonathan Morton <chromatix99 at gmail.com> wrote:
> 
> Toke has already replied, but:
> 
>> Sure, my proposal does not cover the problem of mangling the CE bit inside
>> VLAN-tagged packets, i.e. if we should understand if qdiscs should allow
>> it or not.
> 
> This is clearly wrong-headed by itself.
> 
> Everything I've heard about VLAN tags thus far indicates that they should be *transparent* to nodes which don't care about them; they determine where the packet goes within the LAN, but not how it behaves.  In particular this means that AQM should be able to apply congestion control signals to them in the normal way, by modifying the ECN field of the IP header encapsulated within.
> 
> The most I would entertain is to incorporate a VLAN tag into the hashes that Cake uses to distinguish hosts and/or flows.  This would account for the case where two hosts on different VLANs of the same physical network have the same IP address.
> 
> - Jonathan Morton
> 
> _______________________________________________
> Cake mailing list
> Cake at lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/cake