Re: 43p-140 install issues
On Thursday 13 January 2005 11:36, Sven Luther wrote:
> On Thu, Jan 13, 2005 at 10:23:11AM +0100, Philippe Guyot wrote:
> > On Wednesday 12 January 2005 18:03, Sven Luther wrote:
> > > Alternatively, you can boot into d-i, and then chroot to your system,
> > > install ssh or another network client, and move
> > > /boot/vmlinuz-2.6.8-powerpc to your /tftpboot partition.
I did that, ftp-ing vmlinux, not vmlinuz. may be a mistake ?
And netbooting gave this:
0 > boot net
LOAD: Waiting 60 seconds for Spanning Tree
BOOTP R = 1 BOOTP S = 1
FILE: /tftpboot/vmlinux
Load Addr=0x1000000 Max Size=0x7000000
FINAL Packet Count = 7744 Final File Size = 3964695 bytes.
INIT-PROGRAM: ELF-LOADER
Endian mode : Load Address = 1000000
signature 0x7f454c46 : 7f454c46
class 0x01 : 1
ei_data 0x01 : 2
ei_version 0x01 : 1
type 0x0002 : 2
machine 0x0014 : 14
version 0x00000001 : 1
entry 0x02001b40 : c0000000
phoff 0x00000034 : 34
shoff 0x000180f4 : 2c259c
flags 0x00000000 : 0
ehsize 0x00000034 : 34
phentsize 0x00000020 : 20
phnum 0x00000002 : 2
shentsize 0x00000028 : 28
shnum 0x0000000a : 26
shstrndx 0x00000008 : 23
No ELF note section found in the image !!! Number of program headers : 2
pheader # 1 -------------------------------------
type = PT_LOAD
offset = 10000
vaddr = c0000000
paddr = 1010000
filesz = 2aca40
memsz = 2e2f2c
flags = 7
align = 10000
about to $sync-cache ... ... done with $sync-cache
pheader # 2 -------------------------------------
type = unknown type of Program Header in ELF file
offset = 0
vaddr = 0
paddr = 0
filesz = 0
memsz = 0
flags = 7
align = 4
exiting ELF section
exit INIT-PROGRAM: ELF-LOADER
Is it because of vmlinux instead of vmlinuz ??
And why an ELF on a PReP machine ?
BTW, I believed that vmlinuz is compressed and vmlinux not, but in /boot,
vmlinux is about 3 MB and vmlinuz about 6 MB.....
> >
> > I should have think of that myself!
> > So this kernel should assume that root is /dev/sda3 (or /dev/scsi/.....)
> > right ?
>
> There are two things here. Effectively, the kernel assumes root=/dev/sda3
> (it is hardcoded to it), but naturally, if no root= argument was given at
> all, the initrd does know about what the root= partition is, and know how
> to go to it.
>
> i wonder if we should disable root=/dev/sda3 in the -prep kernel, since i
> believe this is what breaks the without arg d-i loading and means you need
> to add root=/dev/ram.
>
> Let me experiment a bit with this, ok ?
ok. I'll make tests when you want.
>
> > > > > Seems like the scsi driver didn't get loaded or something,
> > > >
> > > > Any relationship with the PCI errors ?
> > > >
> > > > > can you check if
> > > > > it is present on the initrd ? mount -t cramfs /boot/initrd.img /mnt
> > > > > or something such should do.
> > > >
> > > > Just now I can't boot, I have to netboot and go reinstall until
> > > > /target will be available, no ?
> > >
> > > Yep.
> > >
> > > > And.. /boot/initrd.img is not a device, I don't see how to mount it.
> > >
> > > Ah, forgot -o loop, but this is not needed anymore, as we detected your
> > > problem, you where not using the installed kernel, but the
> > > debian-installer one, and thus altough it is the same kernel, using the
> > > wrong initrd.
> >
> > I'll do it however, for my knowledge.
>
> Ok
Mount needs -o loop and was OK. I saw a lot of files in
lib/modules/2.6.8-powerpc/kernel/drivers/scsi/
Shoul be cool.
>
> Friendly,
>
> Sven Luther
Reply to: