I'm trying to use ZipSlack. I've unzipped ZipSlack to the Zip disk and
created the boot floppy as explained in the docs. When I boot, I get the
following error:

Checking if this processor honours the WP bit even in supervisor mode... No.
Kernel panic: This kernel doesn't support CPU's with a broken WP. Recompile
it for a 386!
In idle task - not syncing

I guess the first question is "What's a WP bit?" The machine I'm trying
to run ZipSlack on is a 386sx and the ZipSlack System Reqirements page says
that ZipSlack runs on a 386, so I'm a little confused as to why I'm getting
this error. When I look at /kernels/zipslack.s/config for the ZipSlack
kernel, it says:

#
# Processor type and features
#
# CONFIG_M386 is not set
CONFIG_M486=y

Does this mean that ZipSlack doesn't support a 386, and to get it to support
a 386, I have to recompile the kernel with CONFIG_M386 set to y? What about
the CONFIG_X86_WP_WORKS_OK=y setting? Perhaps recompile the kernel without
this option?

Is there some other (hopefully easier) voodoo I need to do?

Thanks for any help!

Kevin

On Wed, 21 Apr 2004 20:11:59 +0000, Fodor wrote:
> I'm trying to use ZipSlack. I've unzipped ZipSlack to the Zip disk and

Which version?

> I guess the first question is "What's a WP bit?" The machine I'm trying
> to run ZipSlack on is a 386sx and the ZipSlack System Reqirements page says
> that ZipSlack runs on a 386, so I'm a little confused as to why I'm getting

Unfortunately those Web pages have not been maintained for quite some
time. When that was written it was true. But Slackware 8.1 was the last
version which would run on a 386.

> Does this mean that ZipSlack doesn't support a 386, and to get it to support

Yes.

> a 386, I have to recompile the kernel with CONFIG_M386 set to y? What about

No. I think the problem with 386's was a result of a change in glibc.
Even with an i386 kernel, your glibc won't work, so nothing else will,
either.

> Is there some other (hopefully easier) voodoo I need to do?

8.1.
--
/dev/rob0 - preferred_email=i$((28*28+28))@softhome.net
or put "not-spam" or "/dev/rob0" in Subject header to reply

"/dev/rob0" <rob0@gmx.co.uk> wrote in message
newsan.2004.04.21.20.50.46.476666@gmx.co.uk...
> On Wed, 21 Apr 2004 20:11:59 +0000, Fodor wrote:
> > I'm trying to use ZipSlack. I've unzipped ZipSlack to the Zip disk and
>
> Which version?

I was using 9.0, but after reading your reply, I tried 8.1. It gets past the
WP bit problem, or at least seems to. The text scrolls by so quickly, it's
too fast to read. It also accesses the zip drive and detects the hard drive.
Shortly after that, the screen goes black and the system reboots.

Any ideas? Is there something I can do to slow down the boot process to make
it easier to diagnose?

Thanks again.

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

In alt.os.linux.slackware, Fodor dared to utter,
> the screen goes black and the system reboots.

Are you possibly using framebuffer with this kernel? It's been a very
long time since I've used Zipslack so I couldn't tell you off-hand if
it uses framebuffer or not. Try passing "vga=normal" to LILO when it
asks you what to boot.

- --
It is better to hear the rebuke of the wise,
Than for a man to hear the song of fools.
Ecclesiastes 7:5
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQFAhxFwL3KiNGOqr6ERAg+lAJ9Z6rUgSyCpe8AZtVwkrb 76pZAdlgCfQsJ9
azKaGaL8anXkKt4UFRoTBl4=
=0bex
-----END PGP SIGNATURE-----

On Wed, 21 Apr 2004 22:00:55 +0000, Fodor wrote:
> I was using 9.0, but after reading your reply, I tried 8.1. It gets past the
> WP bit problem, or at least seems to. The text scrolls by so quickly, it's
> too fast to read. It also accesses the zip drive and detects the hard drive.
> Shortly after that, the screen goes black and the system reboots.
>
> Any ideas? Is there something I can do to slow down the boot process to make
> it easier to diagnose?

First, try your best to see the LAST text which is on screen. Focus
your eyes at the spot on the screen where you see it. Another set of
eyes might help, too. Go through this several times until you can get
an idea what the last boot message is.

Second, in fact, I do have an idea. Could it be that the IDE controller
is pre-ATAPI? There's a kernel option for that, CONFIG_BLK_DEV_HD_IDE:
"Use old disk-only driver on primary interface". Try using the xt.i
kernel, but I looked at its config, and this does not seem to be set.
This might be a mistake in the config, or the config itself does not
match the xt.i kernel.

Okay, if slackware-8.1/bootdisks/xt.i still does not work (or try
../kernels/xt.i/bzImage with loadlin) try going back to older versions.
You can *probably* (never tried it) use the older 2.2.x kernels with
your 8.1 ZipSlack. Do not bother with 8.0's 2.4.5 kernels, as these
had a known issue with UMSDOS root. 8.0's 2.2.19 xt.i *might* work. If
it does, get slackware-8.0/slackware/a/modules.tgz to have modules for
that kernel.

I take it that you have access to a "real" Linux machine, and you know
something about configuring and compiling kernels? I hope so, because
a 386 is one class of machine which can really benefit from a custom
kernel, but it's not up to the chore of compiling it. (I once did a
2.0.36 on mine in around 8 hours, but that was a much smaller code
base.) You'll want to remove (modularise) big drivers you won't be
using, such as reiserfs. I would suggest starting with the 2.4.26
config for xt.i and make changes as needed. Cut out PCI and other
built-in stuff you can't use ... vesafb comes to mind as another.

Old common wisdom was that Slackware was best for old junk. It
probably still is true for 486-class and above. But you might want to
look at the routerlinux.org project for this one. I read about it here
not long ago (thanks to Menno); it's a Slackware-inspired Linux which
is much lighter-weight, with uClibc and busybox as shell.

Good luck! BTW what do you have planned for this dinosaur? It should be
adequate as a router and a low-use server of various tasks.
--
/dev/rob0 - preferred_email=i$((28*28+28))@softhome.net
or put "not-spam" or "/dev/rob0" in Subject header to reply

"Alan Hicks" <alan@lizella.netWORK> wrote in message
news:4087117a$1_2@127.0.0.1...
> Are you possibly using framebuffer with this kernel? It's been a very
> long time since I've used Zipslack so I couldn't tell you off-hand if
> it uses framebuffer or not. Try passing "vga=normal" to LILO when it
> asks you what to boot.

Honestly, I don't know. I'm using the Slackware 8.1 ZipSlack kernel right
from the distro. In any case, I tried the parameter you suggested, but it's
still doing the same thing.

Thanks all the same. I'll try anything at this point.

"/dev/rob0" <rob0@gmx.co.uk> wrote in message
newsan.2004.04.22.00.58.23.115933@gmx.co.uk...
> First, try your best to see the LAST text which is on screen. Focus
> your eyes at the spot on the screen where you see it. Another set of
> eyes might help, too. Go through this several times until you can get
> an idea what the last boot message is.

I can do better than that! I attempted to boot with the console on a serial
port (ramdisk console=ttyS0). As a result, I have the exact text of the
entire boot process right up to the reboot. Here's the last four lines:

Uniform Multi-Platform E-IDE driver Revision: 6.31
ide: Assuming 50MHz system bus speed for PIO modes; override with idebus=xx
hda: Conner Peripherals 84MB - CP2084A, ATA DISK drive
hde: probing with STATUS(0x74) instead of ALTSTATUS(0xff)

.... then a restart. Does that help? It definitely looks hard drive related.
What is a "hde"?

> Second, in fact, I do have an idea. Could it be that the IDE controller
> is pre-ATAPI? There's a kernel option for that, CONFIG_BLK_DEV_HD_IDE:
> "Use old disk-only driver on primary interface". Try using the xt.i
> kernel, but I looked at its config, and this does not seem to be set.
> This might be a mistake in the config, or the config itself does not
> match the xt.i kernel.

I'm not sure. The above says it's an ATA DISK, or at least it was detected
as one. In any case, I tried the xt.i kernel. It reboots in the same place.

> I take it that you have access to a "real" Linux machine, and you know
> something about configuring and compiling kernels? I hope so, because
> a 386 is one class of machine which can really benefit from a custom
> kernel, but it's not up to the chore of compiling it.

Actually, I don't have a "real" Linux box, at least not yet. On my modern
system, I have one 20GB hard drive full of archived data. Once decide to
either delete it or burn it to DVDs, that drive will be for Linux.

> Good luck! BTW what do you have planned for this dinosaur? It should be
> adequate as a router and a low-use server of various tasks.

I'm looking for it to serve static web pages on a LAN/Intranet, not much
more. It's an old system that I'm trying to get some more life out of. I
dunno - maybe it's a waste with how cheap hardware is these days.

On Thu, 22 Apr 2004 21:20:43 +0000, Fodor wrote:
>> First, try your best to see the LAST text which is on screen. Focus your
>
> I can do better than that! I attempted to boot with the console on a

Woohoo! Good for you.

> Uniform Multi-Platform E-IDE driver Revision: 6.31 ide: Assuming 50MHz
> system bus speed for PIO modes; override with idebus=xx hda: Conner

Hmmm, I don't know about that speed. ISA is what, about 8MHz? Oh, that
says "for PIO modes", which would probably have no effect on an ISA IDE
controller.

> Peripherals 84MB - CP2084A, ATA DISK drive hde: probing with STATUS(0x74)
> instead of ALTSTATUS(0xff)
>
> ... then a restart. Does that help? It definitely looks hard drive
> related. What is a "hde"?

That helps. That about pinpoints it at the IDE controller, just as I
thought. I guessed right.

Note: fixed-width font assumed!
#v+
IDE device table - Linux standard names
=======================================
Bus Master Slave
::: :::::: :::::
first channel (ide0) hda hdb
second channel (ide1) hdc hdd
third channel (ide2) hde hdf
fourth channel (ide3) hdg hdh
#v-

Your 84MB primary master drive seems to have been detected, but the
third channel seems to be giving us the trouble. Can you try moving
that drive to the second channel, or as slave on the first channel?

Is that the Zip drive? Did you try to get it working in any other OS,
such as perhaps DOS? (See http://www.freedos.org/ for a GPL'ed DOS
clone.) I wonder if this controller works with ATAPI devices (which
most IDE Zip drives are)?

What is the Zip drive's hardware interface? They made two types of
parallel, SCSI, ATAPI and some IDE. (I'm not really sure what is the
difference between IDE and ATAPI - I think it is the command set.
Obviously both use the same hardware connection.)

> as one. In any case, I tried the xt.i kernel. It reboots in the same

I'm not surprised. Somewhere along the way I suspect that Pat has
broken the configuration of xt.i. Did you do what I suggested with
OLDER versions of xt.i? There are at least 4 likely suspects you could
try, and I'd do them in this order: 8.0 (2.2.19), 7.1 (2.2.16), 7.0
(2.2.13) and 4.0 (2.2.5). I *think* your ZipSlack 8.1 would run with a
2.2.x kernel.

>> I take it that you have access to a "real" Linux machine, and you know
>
> Actually, I don't have a "real" Linux box, at least not yet. On my

If you have a FAT partition with at least 100MB free, you could
"borrow" that machine from yourself for ZipSlack.

>> Good luck! BTW what do you have planned for this dinosaur? It should be
>> adequate as a router and a low-use server of various tasks.
>
> I'm looking for it to serve static web pages on a LAN/Intranet, not much
> more. It's an old system that I'm trying to get some more life out of. I
> dunno - maybe it's a waste with how cheap hardware is these days.

Yes and no. You'll learn a lot from it - things that you might not get
from a "real" computer. What you're doing is exactly what I did as a
beginner in 1999. My ZipSlack 386 (v. 3.6) was a modem router gateway
and firewall. It rose again running 8.1 and kernel 2.4.20 in late 2002
when I got cable. It was decommissioned again due to a shortage of
electrical outlets! (I needed a more modern server for other purposes,
so that machine took over the router / firewall functions.)

Whilst it should be possible to run X on such a thing, it will be
painful. That will force you into use of the command line for system
administration, rather than coddling you with friendlier-looking GUI
tools. (IMO that's all they are ... they LOOK more friendly, but in
reality they are not.)

I do think you should look at Router Linux, but I'd encourage you to
stick with the 386 for now. We can get it running.

Another coincidence: I had the exact same booting problem at first. At
the time I did not figure it out. I *did* figure that it was failing at
the IDE controller, but I didn't find out WHY. I gave up on RedHat 5.0
to try ZS 3.6, which of course had the same problem.

I started pulling out hardware, going at it in that way. I got it to
boot by using a newer (486-era) IDE controller. That's what's still in
there. If you have a newer ISA IDE controller, you'll do fine with any
of the 8.1 kernels. (Knock on wood. But you should get a custom
kernel built.

Oh, that reminds me to ask: how much memory is there? If it's 8MB or
more, you'll do okay, although I would put in at least 2-3x that amount
of swap space. I really would not recommend trying Linux on 4MB. It is
technically possible, but unbearably slow, as every process must swap.
I remember booting a 7.0 bare.i kernel on a 486 with 4MB, and waiting
20 minutes for "ls" to complete, right after boot and first login.

You could try using a "legacy" version of Slackware, too. Back in the
old days Pat provided more kernels. He used to have a "no-pci.i" which
is likely to work for you.

About the kernel, I would offer to compile one for you, but I'm not
sure if I'll have time in the next few days. Maybe someone else will?
--
/dev/rob0 - preferred_email=i$((28*28+28))@softhome.net
or put "not-spam" or "/dev/rob0" in Subject header to reply

On Thu, 22 Apr 2004 18:26:04 -0500, I tried to write:
#v+
IDE device table - Linux standard names
=======================================
Bus Master Slave
::: :::::: :::::
first channel (ide0) hda hdb
second channel (ide1) hdc hdd
third channel (ide2) hde hdf
fourth channel (ide3) hdg hdh
#v-

Maybe it works without the leading whitespace? I didn't think Pan would
mangle my text like that. When I posted originally it looked like a
nice, even table, with 2 spaces at the beginning of each line, and the
automatic word-wrap in the editor was turned off.
--
/dev/rob0 - preferred_email=i$((28*28+28))@softhome.net
or put "not-spam" or "/dev/rob0" in Subject header to reply