Very Computer

And the long and short of it is, I'm pretty much out of ideas and
feeling pretty low, and could *REALLY* use a little bit o' help.
If somebody could bear with me while I relate*details, I'd
be very much obliged.

Here's the deal.  I had a Generic Brand compuer, an "Easy Data
486-GVT-2" motherboard 486 w/256k cache and 16Mb in an early-'94 vintage
mid-tower package, a CL-GD542X VESA video card w/Cyrix chipset, a
no-name IDE/floppy/2-serial/1-parallel controller board with 1 504Mb
Maxtor drive, running a pre-1.0 linux kernel, no problems, very happy.

One day, I decided that I should install a more recent linux (since
none of the WWW browsers will run with this moldy version of X I'm typing at).
The thought of 60 diskettes was intimidating, and no network card, so I
decided to get a CD rom and some spare disk space while I was at it.

A shopping trip later, with WD 2850RTL IDE drive, Reveal CD-100 ATAPI
drive, BOCA IDE 100 generic secondary IDE controller in hand, (I didn't
want to depend on the CD-only controller; I wanted the two disks on
separate controllers) I sat down to what turned out to be the Night of
the Dead Drives.  (Unfortunately, the only information I have on the
primary IDE controller is that it was imported from China: no specs).

So.  Anyway.  After trying MANY things, the symptoms are these:

 1) Neither DOS, Windows 3.1, or the idecd linux boot kernel
    (slackware 2.3, kernel 1.2) will take any notice of the CD when it
    is configured for the secondary IDE card.  And yes, I'm *very* VERY
    sure I configured it appropriately, with IRQ15, address 170, and the
    power management features on the motherboard disabled.  I've also
    tried configuring it for interupt 12 to avoid the power management
    features.  No joy.

 2) DOS and idecd both seem to work acceptably when the CD is configured
    as the slave drive on the primary card.  Windows hangs upon any attempt
    to access the CD drive when configured this way.  With the CD configured
    in this way, I could make boot disks, and start an installation, partway
    through which the CD ROM started getting "buffer botch" errors, and
    eventually degenerated into complete IO-error-itude.  Sigh.

 3) DOS completely ignores any disk configured on the secondary IDE card,
    (and, as per (1 above) *everybody* ignores the CD when so configured.
    But I thought I had some hope when the idecd boot kernel recognized
    both cards and all three devices when configured as
             ide0   maxtor = master
                    reveal CD = slave
             ide1   WD = solitary
    Alas, during initialization, the ide1 controller got bizarre intermittant
    errors, was reset several times, and was eventually abandoned.  And
    the WD disk manager program wouldn't even recognize the disk when
    configured on that controller.

 4) The BOCA secondary controller seems sound: when configured as ide0
    at IRQ 14, the system boots just fine.  Thus, I thought the system had
    problems with ide1 only.   But when retrying a linux install from CD rom
    configured as slave on the primary (no name) IDE card, the "buffer botch"
    errors reappeared, even with no other IDE card in the system.  Sigh.

My current expectations are that it's either the no-name IDE card (which
I don't trust), or the semi-no-name motherboard (which I also don't
trust).  Perhaps an IRQ or IO map address conflict.  But rather than
rush out and swap the (presumably offending) components, or paying a
repair shop lots of money to come to the conclusion that they need to
swap out offending components, I thought I'd beg for help, in case I've
done something incredibly stupid (and hence perhaps easy to fix without
throwing too much more money at it).  

Imagine me with large puppy-dog brown eyes, teardrops starting to
collect in the corners, and my chin starting to quiver, and you'll have
some idea of my current burnt-out state.

Any ideas?

(Cheap no-name hardware.  Mutter, grumble.)
--

I've narrowed things down, and it looks like I have a bad (or at best
marginal) motherboard: symptoms change mysteriously when boards are moved
to differing slots, and so on.  I'm pursuing that.

But I had a configuration that seemed to work at least minimally, namely
with 1 HD and 1 CD as master and slave (respectively) on the primary
IDE controller.  But, during the attempt to install linux from the newly
installed CD ROM drive, it halts in setup with the claim that it can't
find the slackware directory on the CD.  Using "ls" shows it there,
plain as day, but poking at the CD a bit more shows an intermittant
problem, that gets worse with time: an error such as

    hdb: cdrom_read_from_bufffer: buffer botch (12131)
    end_request: IO error, dev 340, sector 12131

Subsequent ls usage sometimes gets past that problem, and shows the whole
directory, but apparently the initial probe that setup does to find
the slackware directory fails with one of these messages (which is quickly
wiped off the screen by the "friendly" version of setup).

Can anybody interpet this arcana for me?  What is likely going wrong here,
and is there anything I can do about it?  Will a more recent idecd kernel
help?  (The one I'm using is 1.2.8)

As to the other problem (which I think is a bogus motherboard),
when I install ide0(master=HD,slave=CD) + ide1(solo=HD), then boot
the idecd kernel from diskette, it correctly identifies all the hardware,
and the fact that there is an ide0 and ide1.  Then it correctly reads the
partition table from hda, skips hdb (as a CD), then it gets

     hdc: IRQ probe failed (0)
     hdd: IRQ probe failed (0)
     hdd: IRQ probe failed (0)

Note specifically, the probe at the non-existant drive is repeated,
the probe at the drive it had identified correctly just before occurs
only once.

Can somebody interpret this behavior for me?  What do these tea leaves mean?
Specifically, is this consistent with my current best guess that it's
the motherboard that's doing me wrong when I try to install a
secondary IDE controller?

Ah well.   At least I can listen to CDs on headphones while I work now.
Nothing else I bought is workable yet, but the CDs internal electronics is.
--

Quote:>A shopping trip later, with WD 2850RTL IDE drive, Reveal CD-100 ATAPI
>drive, BOCA IDE 100 generic secondary IDE controller in hand, (I didn't
>want to depend on the CD-only controller; I wanted the two disks on
>separate controllers) I sat down to what turned out to be the Night of
>the Dead Drives.  (Unfortunately, the only information I have on the
>primary IDE controller is that it was imported from China: no specs).
>So.  Anyway.  After trying MANY things, the symptoms are these:
> 1) Neither DOS, Windows 3.1, or the idecd linux boot kernel
>    (slackware 2.3, kernel 1.2) will take any notice of the CD when it
>    is configured for the secondary IDE card.  And yes, I'm *very* VERY
>    sure I configured it appropriately, with IRQ15, address 170, and the
>    power management features on the motherboard disabled.  I've also
>    tried configuring it for interupt 12 to avoid the power management
>    features.  No joy.

[ snip snip ]

Kevin.  (http://www.io.org/~kcozens/)

In decreasing likelihood of getting mail through to me:

#include <disclaimer/favourite>

Thanks in advance (+ please post because my mail isn't 100% at
the moment).