[j-nsp] Setting CoS bits on ingress frames

[Victor Sudakov]

Dave Bell wrote:
> When frames enter the switch, the switch will need to classify them
> and place them into a queue. By default everything goes into queue 0,
> regardless of what other switches have done.

I am afraid you are mistaken. If a frame enters a switch via a trunk
port and the frame already has a non-zero COS value, this COS value is
forwarded *intact* out the egress interface. 

I have experimental proof of this. In my lab, Sw1 (Cisco) classifies
frames and their COS value set by Sw1 is visible at HostC.

> 
> You have a rewrite rule on your port facing host C that sets the .p
> bits on queue 0 to 0.

Yes I do, and this rewrite rule mangles frames which are already
classified by other switches.  I would prefer to do away with the
rewrite rule altogether or set it up so that it does not modify frames
arriving into the port facing Sw2.

> 
> To fix this, you need to have a classifier on ingress that moves the
> frames into the correct queue. This means when it egresses the switch,
> it goes through the correct part of the rewrite rule and keeps its
> markings.

This is a chart of what I want to achieve eventually (what what Cisco
does): ftp://ftp.sibptus.ru/pub/vas/chart2.pdf



-- 
Victor Sudakov,  VAS4-RIPE, VAS47-RIPN
sip:sudakov at sibptus.tomsk.ru