From owner-freebsd-hackers Wed Feb 15 05:08:26 1995 Return-Path: hackers-owner Received: (from root@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id FAA07455 for hackers-outgoing; Wed, 15 Feb 1995 05:08:26 -0800 Received: from sbstark.cs.sunysb.edu (sbstark.cs.sunysb.edu [130.245.1.47]) by freefall.cdrom.com (8.6.9/8.6.6) with ESMTP id FAA07443 for ; Wed, 15 Feb 1995 05:08:25 -0800 Received: from starkhome.UUCP (root@localhost) by sbstark.cs.sunysb.edu (8.6.9/8.6.9) with UUCP id IAA20274 for hackers@freebsd.org; Wed, 15 Feb 1995 08:05:32 -0500 Received: by starkhome.cs.sunysb.edu (8.6.9/1.34) id IAA03485; Wed, 15 Feb 1995 08:07:51 -0500 Date: Wed, 15 Feb 1995 08:07:51 -0500 From: starkhome!gene@sbstark.cs.sunysb.edu (Gene Stark) Message-Id: <199502151307.IAA03485@starkhome.cs.sunysb.edu> To: hackers@FreeBSD.org Subject: Diskless boot questions Sender: hackers-owner@FreeBSD.org Precedence: bulk Does anybody know the answer to this one? We are using the FreeBSD diskless boot code in an EPROM on an SMC Ultra network interface. This will be used in a student lab where students will either be booting FreeBSD over the net, or else will be booting MSDOS off of the local hard drive. The mechanism for loading MSDOS is to exit the EPROM boot code and continue with the normal BIOS boot chain. We have found that once the EPROM boot code runs (in particular, it is only necessary for the code in start2.S to run to cause the problem) once DOS comes up, something is left hosed so that it crashes SoundBlaster 16 driver software. DOS, Windows, etc. seem to work fine, but when one attempts to load the SoundBlaster drivers from config.sys the system crashes, requiring a hard reset. The only things that occurs in the code in start2.S are: 1. An initial segment that does some fiddling to apparently place a return point into some low memory locations 0x302, 0x304. 2. Relocation of the entire boot code to 0x90000. 3. Switch to protected mode. 4. Switch back to real mode. 5. Return via the saved return point. My question is: does anybody know what kind of cruft this might leave around in the PC memory that might screw up the SoundBlaster drivers? If some PC whiz knows this, it might save me some debugging time. It seems somewhat obscure to me, since DOS boots fine, but there is apparently some lurking modification to the address space that screws up the SoundBlaster. Thanks! - Gene Stark