Re: [PATCH] Inbound Connection Control mechanism: Prioritized Accept

[Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>]

Hello!

> I am not sure how much overhead is involved in maintaining the the no. of
> slots left for each priority class. Also what should be the ratio of slots 
> that need to reserved for each class? 

It is an experimental value like total size of accept queue,
which is also unknown apriori. No differences.


> Do you think that the existing PAQ patch with SYN policing is a reasonable
> way for prioritizing incoming connection requests?

I still did not look at this patch, I have just got some url from netdev.
(that blamed by Jamal. :-) Guys, tell your managers they should reserve
a bit of money for admins to replace bogus firewalls. ibm site is really
not accessible, it is not a joke. :-)). I will look at it tonight.


> Preempting existing low priority connections in acceptq with high priority 
> ones may not be good idea as we need to abort them by sending a RST.

Of course. It is _very_ bad idea. :-)

Actually, true preemption can be relaized here with moving socket
back to SYN-RECV state, converting it to open_request. We just pretend
that we did not receive ACK, it is fully legal. 

But in this case we also have room for effective preemption,
stopping process SYN_RECV->ESTABLISHED for low priorities.
I.e. exactly, which SYN policing makes.

Alexey