From owner-freebsd-questions Sun Sep 10 14:12:15 1995 Return-Path: questions-owner Received: (from majordom@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id OAA22777 for questions-outgoing; Sun, 10 Sep 1995 14:12:15 -0700 Received: from Root.COM (implode.Root.COM [198.145.90.17]) by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id OAA22771 for ; Sun, 10 Sep 1995 14:12:13 -0700 Received: from corbin.Root.COM (corbin [198.145.90.34]) by Root.COM (8.6.12/8.6.5) with ESMTP id OAA20016; Sun, 10 Sep 1995 14:09:36 -0700 Received: from localhost (localhost [127.0.0.1]) by corbin.Root.COM (8.6.12/8.6.5) with SMTP id OAA02963; Sun, 10 Sep 1995 14:11:49 -0700 Message-Id: <199509102111.OAA02963@corbin.Root.COM> To: Terry Lambert cc: jc@irbs.com, shorty@iii.net, freebsd-questions@freebsd.org Subject: Re: Network problems, please assist In-reply-to: Your message of "Sun, 10 Sep 95 13:53:50 PDT." <199509102053.NAA15245@phaeton.artisoft.com> From: David Greenman Reply-To: davidg@Root.COM Date: Sun, 10 Sep 1995 14:10:53 -0700 Sender: questions-owner@freebsd.org Precedence: bulk >> date: 1995/06/29 08:21:32; author: davidg; state: Exp; lines: +1 -5 >> Removed "GATEWAY" consideration when calculating number of mbuf clusters. >> It now always uses the value that was used for the GATEWAY case. > >Yeah. This is probably the wrong thing to do for the non-GATEWAY case. No, it is correct for both cases. >> It isn't practical to make it soft configurable because the size of the >> mb_map must be calculated at startup time. We could (should!) make some >> extensions to 'userconfig' to allow you to change things like this. > >Or make mb_map resizeable. I hate static configuration. maps aren't resizable. >> I highly recommend that you subscribe to the cvs-all mailing list so that >> you don't give out so much disinformation. > >I prefer to run "cvs log" on specific files; too bad it doesn't work >with SUP instead of CTM. > >Having too many buffers for the non-GATEWAY case is little better than >having few in the GATWAY case. Both are equally broken. You don't understand what this does. "nmbclusters" sets the *limit*. No buffers are actually allocated until they are needed. There are no additional buffers or space allocated (except in the map, but this is a virtual memory thing and has nothing to do with physical memory). -DG