Re: initial console

On Thu, Oct 19, 2000 at 02:16:35PM +0200, Xavier Grave wrote:

On Wed, 18 Oct 2000, Alois Fertl wrote:

quoted

The problem you are seeing here is probably prep_init in prep_setup.c
not copying the arguments passed by the boot loader into the kernels command
line. Your root=/dev/sda5 is probably ignored.
Look at your early log messages during boot. There should be a line
"Boot arguments: ..." telling you those things seen by the kernel.

Indeed, The kernel argument I give is :
root=/dev/sda5
and during the boot when I look more closely :
PReP architecture
Boot arguments: /dev/sda1

quoted

In case the arguments aren't as you typed them, try putting the following
lines at the beginning of prep_init:

+       if (r6) {
+               *(char *)(r7 + KERNELBASE) = 0;
+               strcpy(cmd_line, (char *)(r6 + KERNELBASE));
+       }

Ok, this is fine my kernel boots ! Thank you very much.

quoted

What hardware is this running on? If it is Motorola VME this failure is only
the starting point and other trouble will follow. There are additional fixes
but I can't verify them against a working source at the moment. test9 and
test10
from fmlabs currently both panic during init as soon as the process starts
using the hard drive (this is on a PowerStack).

I have a MVME2303 from motorola diskless + a PMC with a 53c895 pci board
for my disk. With the 2.2.17 kernel all works well.
And now with the latest fsmlabs + your hack all function !

Sorry, I had this change locally but hadn't pushed it yet. :)  The other
option on PReP is to build with /dev/nvram support and then (on a
PPCBUG system) use 'gevedit bootargs' to set a kernel argument command
line like "root=/dev/sda5".  That's often more convenient then hard
coding in your parameters like with the PReP cmd_line config option.

--
Matt Porter
MontaVista Software, Inc.
mporter@mvista.com

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/