From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Fri, 29 Jun 2001 20:09:40 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Fri, 29 Jun 2001 20:09:30 -0400 Received: from ash.lnxi.com ([207.88.130.242]:5617 "EHLO DLT.linuxnetworx.com") by vger.kernel.org with ESMTP id ; Fri, 29 Jun 2001 20:09:19 -0400 To: Subject: Results of playing with linux-lite.... From: ebiederman@lnxi.com (Eric W. Biederman) Date: 29 Jun 2001 18:09:03 -0600 Message-ID: User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.5 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I keep playing with different ideas on how to do the the minimal bootloader seen, and wishing I had a full OS behind me todo it. So today I played with linux-lite. A slightly modified linux-1.0.9. It looks like you can get a pretty useable kernel out of it in about 128K, after compression. I have built up 2 patches .eb1 This allows the code to compile with gcc-2.95.2 (hopefully the code works). .eb2 This patch enables compiling out large sections of the kernel like the VFS layer, so I can experiment and see where the kernel bloat is. I don't understand why the linux kernel when you disable everything is much larger than linuxBIOS. The code is available at: ftp://download.linuxnetworx.com/pub/src/linux-kernel-patches/ Even when I compile out the VFS layer, and the network stack, swap support, and mmap, all character and block devices. Uncompressed the linux kernel is still about 100KB. It feels like there is a significant source of bloat there. My hope is to build a very modular kernel without nearly zero bloat and then port the infrastructure up to the latest kernels. Linux 1.0 is very much linux in structure, so if nothing else it should be able to serve as a good comparision point. Anyway. A kernel with nothing in it that compresses to 47KB is kind of fun :) Eric