Well, this becomes a long story (3 months of hassle so far :) ), so
here's a keyword list:
ATAPI CDROM
Linux distributions
Debian Linux
NEC-260 (250?) CDROM
Creative Labs CD1220E CDROM
Award BIOS
IDT-C6 "Winchip", 200MHz
Shuttle Spacewalker Motherboard (HOT-569)
I'm trying to get my first Linux system working (I have big plans for
it, if I can ever get anywhere with it!). The first go at this I used
an old Soundblaster Pro with a connected CDROM, which seemed to work
okay, though I'm not sure I thoroughly tested it.
Unfortunately, I gave that to my Mom for unrelated reasons, thinking,
"No problem, it's old anyway, I'll get an IDE CDROM and it'll be
simpler" Heh, heh, heh, little did I suspect. :)
So here we go,
I am attempting to install Debian version 2.0.2 from a Linux Systems
Labs CDROM (the dmesg utility reports "Kernel version 2.0.34").
I liked the Debian webpage and attitude, and since I'm
used to Solaris 2.6 at work, I thought that the "packages" approach
to software loading was nice (don't know if other distributions
have this [?] ).
I have two CDROM ATAPI IDE drives: one is an NEC 2x, which says
"NEC-250"
on the drive itself, but the BIOS reads it as "NEC-260", and so does
the debian "dmesg" utility. The other is a Creative Labs 12x, which
has a goofy infra-red remote with MS Windows support. It says
"Infra-1800"
on it, but the BIOS calls it a "Creative Labs CD1220E".
Either of these drives works okay on my Windows system. Both fail
to work on the Linux box, but in completely different ways, as follows:
The NEC-260:
Is not correctly recognized. First of all, the distribution disk
will not boot on it (I assume no CD boot disk would work, but
the Debian one, is the only bootable CD I have).
The "dmesg" utility reports the following:
ide: i82371 PIIX (Triton) on PCI bus 0 function 57
ide0: BM-DMA at 0xf000-0xf007
ide1: BM-DMA at 0xf008-0xf00f
hda: SAMSUNG VG38404A (8.40GB), 8056MB w/490kB Cache, CHS=1027/255/63
hdb: NEC CD-ROM DRIVE:260, ATAPI cdrom or floppy?, assuming FLOPPY drive
ide-floppy: Packet size is not 12 bytes long
ide-floppy: hdb: not supported by this version of ide-floppy
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
Several fixes were recommended to me from the Debian-user mailing
list, which basically boil down to adding:
/dev/hdb=cdrom
to the /etc/lilo.conf file which Debian (evidently) uses during
boot (perhaps all Linux dists do?).
This seems to have no effect, however. Anyway, the system does
not acknowledge /dev/hdb to exist at all, so I can't mount it.
Yuck!
Since this was getting difficult, I tried swapping the drives:
Creative CD1220E:
Better: it boots the Debian installation disk, up to the initial
message. However, attempting to proceed (pressing at
the "boot:" prompt), results very quickly in either a hang-up,
or a reboot (it appears to be random which).
If I install from the floppy disks, I can then mount the CDROM
and attempt to install packages from it, but whenever it tries
to load a .tgz file, zcat reports CRC errors -- so the files
must be getting corrupted (?). Copying the data to available
space on the hard disk, I tried again, with the same problem
(I used "cp -dpRv /cdrom /home/cd1" to do this). So, it would
appear that the corruption problem exists with the copy command
as well, and is not just a problem with the installation program
(I thought perhaps the cp command might catch and correct errors
that got past the installation program).
So I'm up a creek without a CDROM drive (or something like that :) ).
Anyway, I'd like to get to the bottom of this, or at least find
a way around it. Both drives claim to be IDE ATAPI type. Both
use a driver in DOS/Windows (loaded by config.sys), which the
MSCDEX program then uses. I tried installing the CD1220E on
the DOS/Windows machine while leaving the NEC driver in place. A
DOS game ("Return to Zork", for anyone who's interested) reported
numerous read errors in this configuration, although it mostly
worked. Replacing with the driver supplied with it fixed the read
errors.
Based on that, I'm guessing maybe the DOS driver traps and corrects
bad reads, and that the Linux ATAPI CDROM driver (which must be
a generic driver, right?) does not have that facility. Then on large
files, it just corrupts the data, and that's my problem.
But if so, how do I know another CDROM won't just do the same thing?
Could this instead be a problem with the BIOS or motherboard?
This is:
Award Modular BIOS v4.5PG, An Energy Star Ally
Copyright (C) 1984-97, Award Software, Inc.
(569WIQ05) Intel i430TX PCIset(TM)
Award Plug and Play BIOS Extension v1.0A
Copywrite (C) 1997, Award Software, Inc.
The motherboard is a Shuttle, "Spacewalker" HOT-569 V2.x
The CPU is an IDT-C6 "Winchip" running at 200MHz
Let's see, I also have 8 Mb RAM (the RAM is a little oddball,
I'm using a 4Mb 72-pin SIMM with four 1 Mb 30-pin SIMMs in
a 30-to-72 pin SIMM adapter (from http://www.bgmicro.com, BTW),
so there may be a reliability
