*BSD News Article 13449


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!agate!spool.mu.edu!uunet!pipex!marble.uknet.ac.uk!uknet!edcastle!spider!raft.spider.co.uk!mark
From: mark@spider.co.uk (Mark Valentine)
Newsgroups: comp.os.386bsd.bugs
Subject: BIOS problems: RTC diagnostic error & memory refresh
Message-ID: <1993Mar27.162025.9225@spider.co.uk>
Date: 27 Mar 93 16:20:25 GMT
Sender: news@spider.co.uk (USENET News System)
Organization: Spider Systems Limited, Edinburgh, UK
Lines: 47
Nntp-Posting-Host: lynx.spider.co.uk

I've identified a piece of code in the clock initialisation routine (in
/usr/src/sys.i386/i386/isa/clock.c) which gives me problems.

  Configuration:
    Elonex PC425X (25MHz 486SX - Symphony chipset)
    MR (Microid Research) BIOS
    local bus video (Headland/Video 7 - HT216-32)
    both IDE & SCSI disk (AHA1450B adapter)

There are two symptoms.  The first is the following error message at
bootup:

    RTC BIOS diagnostic error 10

This appears to be harmless in itself.  However, the second symptom is
that I can't boot from floppy after booting 386BSD (sometimes I'll get
an error (or part of one) about a non-system disk, at other times the
floppy activity light just goes off after a while and the system hangs).

On closer inspection, one of my BIOS settings is changed from the default:
the Memory Refresh is normally "decoupled", booting 386BSD makes it become
"standard".  I've also seen the AT bus speed changed back to the default of
~8 MHz if I've set it higher.

If I change the BIOS setting back in Setup, I can then boot from floppy(?!).

If I comment out the following lines in clock.c, I still get the diagnostic
error (obviously), but my BIOS remains intact and my floopy is bootable.

    outb (IO_RTC, RTC_DIAG);
    outb (IO_RTC+1, 0);

None of the other code in the routine makes any difference with respect to
the BIOS being changed.  I'm using the original 0.1 code for this routine,
but the findcpuspeed() patch makes no difference.  (The rest of my system
is patched with patchkit-0.2.1 [up to patch83, I think].)

Anyone know enough about this stuff to know what's going on?  Might my BIOS
be suspect?

		Cheers,

		Mark.
-- 
Mark Valentine, Spider Systems <mark@spider.co.uk>            /\oo/\
"Tigers will do ANYTHING for a tuna fish sandwich."
"We're kind of stupid that way."