Richard A Steenbergen wrote:
> 
> On Thu, May 30, 2002 at 05:21:17PM +0200, Andre Oppermann wrote:
> > > This is a non sequitur. All routes will be available through the kernel
> > > RIB, but for exact matches only. When is a longest prefix match needed
> > > there?
> >
> > When the routing daemon instructs us to remove the prefix 10.0.0.0/8
> > when we also have 10.0.0.0/9 and 10.128.0.0/9.
> 
> That is not a longest prefix match, this is an exact match.

Ah, well. You're right.

> > Where? Do you mean rt_metrics?
> 
> Yes.

I'm axing that right now and will provide a tcp_hostcache that will
assume that role (tcp is the only consumer of rt_metrics, except for
rmx_mtu and rmx_pksent). By moving this every node/leaf in the routing
table shrinks by 48 bytes. On a default free view of the Internet this
gives us a whopping 5.5 Mbytes in kernel memory savings (110k routes).

-- 
Andre

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-net" in the body of the message

Reply via email to