Return to BSD News archive
Xref: sserve comp.os.linux.development:8703 comp.os.minix:23648 comp.unix.bsd:13875 comp.unix.pc-clone.32bit:6143 comp.os.386bsd.development:2057 comp.sys.ibm.pc.hardware.misc:5020
Newsgroups: comp.os.linux.development,comp.os.minix,comp.unix.bsd,comp.unix.pc-clone.32bit,comp.os.386bsd.development,comp.sys.ibm.pc.hardware.misc
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msuinfo!agate!darkstar.UCSC.EDU!news.hal.COM!decwrl!netcomsv!calcite!vjs
From: vjs@calcite.rhyolite.com (Vernon Schryver)
Subject: Re: [Announcement] 386BSD Release 1.0
Message-ID: <CpHLJE.J5w@calcite.rhyolite.com>
Organization: Rhyolite Software
Date: Sun, 8 May 1994 13:59:38 GMT
References: <newcombe.142.00141E4A@aa.csc.peachnet.edu> <2qdvvp$r@bmerha64.bnr.ca> <2qht16$msc@news.u.washington.edu>
Lines: 46
In article <2qht16$msc@news.u.washington.edu> tzs@u.washington.edu (Tim Smith) writes:
> ...
>I wrote a SCSI BIOS that doesn't support booting from a CD-ROM, because it
>never occurred to me or the people we wrote it for that this would be a
>cool thing, but if it had, it would have taken about 5 minutes to modify
>the BIOS to support it. It doesn't surprise me to find that someone did
>think of it and do it. It really is trivial.
While that is no doubt true, the start of this thread is that someone
says that Jolitzes say that their new CDROM is intended to be booted
directly from CDROM using Adaptec 1540's.
That is between "impossible" and "impractical" on purely practical
grounds. Everyone agrees it would require you to make your CDROM drive
be target 0, which would mean that when you booted your system, it might
come up from the CDROM. Since most CDROM drives are read-only, that
would make it rather inconvenient to configure the drives (and so forth)
you actually have in your system. That would make it a little inconvenient
to actually use the magnetic drives in your system. Would you want the
read-only UNIX kernel on the CDROM to poke at every SCSI, IDE, SMD, and
floppy to see if it finds volume labels it thinks it recognizes and
mount any apparent UNIX volumes read-write?--I certainly wouldn't!
Would you want to stick to only CDROM with no disk? Or have to manually
mount the real drives on each reboot? Or be stuck with having what ever
combination of SCSI targets the authors of the CDROM think reasonable?
(The Ciprico RF5500 lets you configure the mapping between "drive c"
and SCSI target, but the quote said "Adaptec 1540", not "Ciprico".)
It would be mildly useful to be able
1. temporarily remove or reconfigure your hard drive on target 0,
2. reconfigure your CDROM to target 0 and boot from it.
3. write a conventially miniroot boot floppy.
4. restore the correct configuration of your SCSI devices,
5. boot from the miniroot floppy and procede as has been usual
since at least the early 1980's for 4.1-BSD systems.
On the other hand, if I paid $100 for a CDROM that costs about $2 to
manufacture, I'd expect it to come with a suitable boot floppy that
costs about the same as the CDROM's jewel case to manufacture.
(I'd also expect a $100 CDROM to come with a jewel case.) (Yes, we all
know software NRE costs are generally much more than software manufacturing
cost.)
Vernon Schryver vjs@rhyolite.com