PCMCIA A: drive? That is probably just not going to work. PCMCIA
services under linux are user-level processes that are never, ever, as far
as I know, available at boot up. Linux does _NOT_ use the BIOS to read
the floppy drive, therefore the floppy disappears as soon as the kernel
takes over and tries to load the root filesystem. The BIOS is making the
PCMCIA A: drive look like a floppy at boot, but it really isn't a normal
floppy controller that can run without pcmcia services being started. I
doubt there is any way to boot linux successfully from it unless _maybe_
if you use initrd to load a root filesystem that then enables card
services. LILO might use the BIOS to load an initrd, so that might work.
-Tom
On Thu, 23 Sep 1999, Joe Wronkowski wrote:
> Date: Thu, 23 Sep 1999 14:25:41 -0700
> From: Joe Wronkowski <[EMAIL PROTECTED]>
> To: Tom Oehser <[EMAIL PROTECTED]>
> Subject: RE: Linux Question
>
> The reason for the lock up is that the libretto looses the touch with the
> pcmcia A: drive at a certain point hda check and then dosn't have the module
> built in the kernal to start the service after the check.
>
> --Joe
>
> -----Original Message-----
> From: Tom Oehser [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, September 23, 1999 7:45 AM
> To: Joe Wronkowski
> Subject: RE: Linux Question
>
>
>
> > ok the dd technique will produce a copy of my hd on the linux box, I can
> see
> > this, but how can I create an IMG file to store for use? The librettos
>
> dd if=/dev/hda | gzip -9 > filename.img
> then
> dd if=filename.img | gzip -d | dd of=/dev/hda
>
> actually, even
> "gzip </dev/hda >filename.img"
> and
> "gzip <filename.img >/dev/hda"
> will work. Of course, combine with nfs or rsh or nc to go to/from
> different machines.
>
> > working becasue the img dosn't boot linux side. I have tried to boot from
> > your 1.7.140 disk on a computer that had an img downloaded and the
> computer
> > stops after the partition check.
>
> This should not happen. It stops exactly right after the partition check
> and locks up? That implies a bug in the kernel. You could make the
> device a module and maybe get around it, but, that is strange. Anything
> more you can tell me about that will help.
>
> -Tom
>
>
