[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: new WRR sched, updated TBF and "real" ingres implementation



> cascade policers. Look at the "color aware" example.

thanks I will, please where can I found these samples ? I looked
at http://icawww1.epfl.ch/linux-diffserv/ with no success :-(

> Why is packet dropping when a flow oversubscribes such a bad thing?

ugh .. probably gaping hole in my english .. I can't understand
the sentence ..

> To add queueing to ingress is trivial as i said in the earlier email.

It is not so trivial. As I looked thru code, there is one point
I'm not able to understand. It is why there is dev->qdisc and
dev->qdisc_sleeping. It seems to me that second one could be
ommited but probably I'm missing something. But I can't realize
what :-|. Have you any clue ?

> I just havent heard anything that is convincing to say we need this.

I would like to look at the "color-aware" example as you mentioned
above. If it can solve the problem with drop fairnes between tcp
flow then I also don't need it.
I really don't want to reinvent wheel. I only want to solve problem
here with my best effort and share experience/code with another people.

regards, devik