Subject: Re: is this bad (UDP stats)?
To: Erik E. Fair <fair@clock.org>
From: Jason Thorpe <thorpej@nas.nasa.gov>
List: tech-kern
Date: 10/25/1998 23:02:34
On Sun, 25 Oct 1998 09:15:46 -0800 
 "Erik E. Fair" <fair@clock.org> wrote:

 >         443201 PCB hash misses

 > What's up with all those PCB hash misses?

That stat is incremented if in_pcblookup_connect() fails on the incoming
packet.  I.e. whatever application the packet is for isn't using connected
UDP.

For a while now we've used a hash for unconnected sockets, too, but
I suppose the stat counter was never updated.  In any case, the
lookup_connect, by its very nature, is more efficient than the lookup_bind;
just take a look at the code in in_pcb.c to see why.

Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                            Home: +1 408 866 1912
NAS: M/S 258-5                                       Work: +1 650 604 0935
Moffett Field, CA 94035                             Pager: +1 650 940 5942