Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!agate!howland.reston.ans.net!spool.mu.edu!bloom-beacon.mit.edu!ai-lab!life.ai.mit.edu!mycroft From: mycroft@duality.gnu.ai.mit.edu (Charles Hannum) Newsgroups: comp.os.386bsd.questions Subject: Re: Building a kernel larger than 640K Date: 11 Nov 1993 17:54:10 GMT Organization: MIT Artificial Intelligence Lab Lines: 23 Distribution: inet Message-ID: <MYCROFT.93Nov11125410@duality.gnu.ai.mit.edu> References: <2ahdk8$rjk@alva.ge.com> <JKH.93Oct26164254@whisker.lotus.ie> <2bohik$1g1@keltia.frmug.fr.net> NNTP-Posting-Host: duality.ai.mit.edu In-reply-to: Ollivier.Robert@keltia.frmug.fr.net's message of 9 Nov 1993 16:47:02 GMT In article <2bohik$1g1@keltia.frmug.fr.net> Ollivier.Robert@keltia.frmug.fr.net (Ollivier Robert) writes: Jordan K. Hubbard (jkh@whisker.lotus.ie) wrote: > config "386bsd" at 0xFE100000 root on sd0 swap on sd0 When one builds such a kernel, are the 640 KB lost or do they get back into the memory pool ? From my experience, it seems that they're lost... In NetBSD, you still get to use the 640K. I'm not sure offhand whether those changed made it in before 0.9 or not. Do you plan to implement such bounce buffers or does one must get an EISA motherboard in order to use more than 16 MB ? I've heard that they are/were used by the floppy driver. `Bounce buffers' have been used for quite a while in the floppy driver. Their use in the SCSI code is really a matter to discuss with Julian Elischer, as he pretty much maintains the SCSI code in both systems.