Re: [cobalt-users] Raq4 - Kernel 2.4 Upgrade Woes

["Josh Trutwin" <josh@xxxxxxxxxxxxxxxxxxx>]

Thanks for the quick response.

> JT> request_module[block-major-3]: Root fs not mounted).  Seems
> JT> odd, the ROM kernel wasn't changed, it should still have IDE
> JT> drivers...
>
> Different CHS translations?  Can you supply us with the disk
> geometry of the drive you installed, as it was detected by the
> host machine?

-- from K6 running slack (booted from floppies due to missing kernel)
# cat /proc/ide/ide0/hda/geometry
physical 65531/16/63
logical 4111/255/63

Other info:

# fdisk /dev/hda

p (print partition table)

Disk /dev/hda: 33.8 GB, 33820286976 bytes
255 heads, 63 sectors/track, 4111 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

Device     Boot Start        End    Blocks   Id  System
/dev/hda1    *     1          13    104391   83  Linux
/dev/hda2         14          39    208845   82  Linux swap
/dev/hda3         40          65    208845   83  Linux
/dev/hda4         66        4111  32499495   5   Extended
/dev/hda5         66          91    208813+  83  Linux
/dev/hda6         92         117    208813+  83  Linux
/dev/hda7        118         181    514048+  83  Linux
/dev/hda8        182        4111  31567693+  83  Linux

Anything else I can provide?

> JT> 1.) Does it matter that I'm using an Athlon K6 to build this?
>
> Uh, Athlon != K6.  Athlon is the successor to K6.

Doh.  (slap forehead)  My bad, I meant AMD...

> JT> I don't have a good grasp on the architecture of the Raq4,
> JT> but I read in one of the instructions to use a spare PIII.
> JT> Will an Athlon work?  (it's i386 right?)  If not, will a PIV
> JT> work, no PIII sitting around....
>
> Compile for Pentium or K6 architecture.  Do not use instructions
> introduced in PII/PIII/P4/Athlon, such as CMOVxx.  If you stick
> with "-march=pentium" or "-march=k6" or earlier, you'll be fine.

With the slackware instructions, I do not actually compile anything (yet).
 Just install a tiny slackware w/o a kernel and attempting to have the
Cobalt ROM kernel recognize the drive.  Further on in the process I can
build a kernel...  (http://www.traumatized.org/linux/cobalt/kernel.php)

> JT> 3.) Does it matter where in the hard drive /dev/hda1 is?
> JT> Somewhere I think I remember reading to use the last
> JT> cylinders of the hard drive.  I used the beginning of the
> JT> hard drive during my partitioning.
>
> Until IDE drivers are loaded, one is confined to the first 1024
> cylinders.  Having a kernel in ROM helps. ;-)

I read that as: have /dev/hda1 somewhere in cylinders 1-1024

> JT> FYI, I have a copy of the output from attempting to boot at
> JT> http://trutwins.homeip.net/raq4.txt that might be helpful.
>
> Hmmmm.  "Unable to mount root fs on 03:01", which is /dev/hda1.

I know, maybe the info I provided will shed some light???  I'm certainly
not seeing it!

Thx again,

Josh