David S. Miller wrote: > From: Ben Greear <greearb@candelatech.com> > Date: Wed, 28 Aug 2002 00:13:31 -0700 > > For machines with large numbers of interfaces (VLANs for instance), > this could be a real performance drag. > > Have a look at our VLAN implementation before you jump to > conclusions :-) At the time I wrote the vlan code, I also wrote a patch to hash the devices on both index and on name. It made ifconfig of 4k vlans happen in 20 seconds instead of 25 minutes...but it was not accepted, and I didn't push the matter... I am not sure what part of vlan would make the lookup by index not 0(n), unless things have changed significantly since the patch was accepted. > And the route.c case you mention is only when the route > cache lookup misses (ie. it happens rarely) I wonder if the rest of the cases are as rare? I know I assumed that it was not an expensive method to call, though that doesn't mean everyone designed like that. So, I still think it would be good to hash. Ben > -- Ben Greear <greearb@candelatech.com> <Ben_Greear AT excite.com> President of Candela Technologies Inc http://www.candelatech.com ScryMUD: http://scry.wanfear.com http://scry.wanfear.com/~greear - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
