From owner-freebsd-hackers Thu Dec 6 13:12:43 2001 Delivered-To: freebsd-hackers@freebsd.org Received: from bogslab.ucdavis.edu (bogslab.ucdavis.edu [169.237.68.34]) by hub.freebsd.org (Postfix) with ESMTP id 7429837B416 for ; Thu, 6 Dec 2001 13:12:34 -0800 (PST) Received: from thistle.bogs.org (thistle.bogs.org [198.137.203.61]) by bogslab.ucdavis.edu (8.9.3/8.9.3) with ESMTP id NAA17962 for ; Thu, 6 Dec 2001 13:12:27 -0800 (PST) (envelope-from greg@bogslab.ucdavis.edu) Received: from thistle.bogs.org (localhost [127.0.0.1]) by thistle.bogs.org (8.11.3/8.11.3) with ESMTP id fB6LCmI41514 for ; Thu, 6 Dec 2001 13:12:49 -0800 (PST) (envelope-from greg@thistle.bogs.org) Message-Id: <200112062112.fB6LCmI41514@thistle.bogs.org> To: hackers@FreeBSD.ORG X-To: Leo Bicknell X-Sender: owner-freebsd-hackers@FreeBSD.ORG Subject: Re: Can TCP changes be put in RELENG_4? In-reply-to: Your message of "Thu, 06 Dec 2001 09:42:09 EST." <20011206094209.A60489@ussenterprise.ufp.org> Reply-To: gkshenaut@ucdavis.edu Date: Thu, 06 Dec 2001 13:12:48 -0800 From: Greg Shenaut Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG In message <20011206094209.A60489@ussenterprise.ufp.org>, Leo Bicknell cleopede: >On Thu, Dec 06, 2001 at 06:23:31AM -0800, Terry Lambert wrote: >> > I'd suggest our target should be a P-III 600 with around 256M of >> > RAM as what Generic should be tuned for.... >> >> Can't. The static allocations for that much assumed RAM would >> result in the machine not booting, with the amount of RAM for >> the page tables alone ~1/4M. By default, the 120M KVA space >> mappings are arguably overlarge for small memory machines. > >Would result in what machines not booting? As long as a 64M PC >can boot (even if it has only 10 Meg free for user apps) that's ok >in my book. If we're still trying to boot on 4, 8, or 16 meg >machines that's just dumb. > >As I've said before, there are two types of FreeBSD users. There >are "users", who want something to replace windows and who really >like the Linux distro's with KDE and all that. These people are >unlikely to build a kernel, and as time goes by are even less likely >to know what a kernel is. They are also likely to have a < 3 year >old PC, probably that they are dual booting. Linux recognized this, >and targest this sort of hardware out-of-the-box. > >The second type of user is someone like you, or me, or most of the >people on this list. They will build a custom kernel no matter >how appropriate the default settings. They will tune things for >odd application boxes, like IRC and News servers and the like. >The defaults are virtually irrevelant for these people, provided >sysinstall can finish. > >As far as I'm concerned any machine with < 64M these days falls >into the second catagory, where someone should have to futz with >it to make it work. When 256M DIMMs are $18 we need to get with >the program. Speaking as someone with a lab full of older machines, including some 8MB 386SXs happily humming along under FreeBSD, and no machine with > 32 MB, I obviously am going to disagree vehemently with this entire line of argumentation. (I always build a custom kernel for my machines, but if you required 64MB or more just to boot the installation floppy, I would have to go buy RAM just to have a machine to install the next version on. Ptui!) However, I do see the value in making it easier to have a faster, more memory-intensive kernel, so why not just provide a "turbo kernel" in the standard root distribution along with the current "generic kernel"? Shoot, I think even casual, non kernel-configuring users might be interested in comparing performance under the two kernels; plus, having the config file for the turbo kernel available for perusal in the kernel source would be a big plus in my book. Greg Shenaut To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message