Discussion:
after new kernel-build booting stops
(too old to reply)
Dieter Geyer
2008-01-13 04:12:03 UTC
Permalink
Hi all,

i have an old Motorola PowerStack II (with 300 MHz PPC 604, 256 MB RAM)
and installed (with some tricks) a debian 4.0 R1 (Etch) on this maschine.
This is a PPC PReP achitecture.
It boots from the original boot-kernel into text-based multi-user console.

Now i want to create a new kernel, after compiling and linking (with tool
mkvmlinuz version 32) i get a file "mykernel" without any errors.

But after installing the vmlinuz-file into partition /dev/sda1 (PReP boot)
i get only this messages at boot time (recorded via RS232 console)

#############

loaded at: 00400400 00A75F4C
relocated to: 00800000 00E75B4C
zimage at: 00809DB4 009570CC
initrd at: 00958000 00E6C910
avail ram: 00400000 00800000

Linux/PPC load: console=ttyS0,9600 console=tty0
Uncompressing Linux...done.
Now booting the kernel

##############

after this, no more output occurs, no hdd activity, no keys are accepted.

now my question : what's wrong ?

Thanks
Dieter

if nessesary, i can give the output of compiling/linking process.
tortoise
2008-01-15 06:47:06 UTC
Permalink
Post by Dieter Geyer
Hi all,
i have an old Motorola PowerStack II (with 300 MHz PPC 604, 256 MB RAM)
and installed (with some tricks) a debian 4.0 R1 (Etch) on this maschine.
This is a PPC PReP achitecture.
It boots from the original boot-kernel into text-based multi-user console.
Now i want to create a new kernel, after compiling and linking (with tool
mkvmlinuz version 32) i get a file "mykernel" without any errors.
But after installing the vmlinuz-file into partition /dev/sda1 (PReP boot)
i get only this messages at boot time (recorded via RS232 console)
you didn't just copy the kernel did you ? you also need to install
/boot and /lib/modules, and new initrd.

did you use debian kernel package to make a .deb or did you build
linus
style ?
Post by Dieter Geyer
#############
loaded at: 00400400 00A75F4C
relocated to: 00800000 00E75B4C
zimage at: 00809DB4 009570CC
initrd at: 00958000 00E6C910
avail ram: 00400000 00800000
Linux/PPC load: console=ttyS0,9600 console=tty0
Uncompressing Linux...done.
Now booting the kernel
##############
after this, no more output occurs, no hdd activity, no keys are accepted.
now my question : what's wrong ?
maybe you made an error in configuring your kernel.

what did you change ? what config tool did you use ? (makeconfig or
what)
Post by Dieter Geyer
Thanks
Dieter
if nessesary, i can give the output of compiling/linking process.
Dieter Geyer
2008-01-15 21:49:31 UTC
Permalink
Post by tortoise
Post by Dieter Geyer
Hi all,
i have an old Motorola PowerStack II (with 300 MHz PPC 604, 256 MB RAM)
and installed (with some tricks) a debian 4.0 R1 (Etch) on this maschine.
This is a PPC PReP achitecture.
It boots from the original boot-kernel into text-based multi-user console.
Now i want to create a new kernel, after compiling and linking (with tool
mkvmlinuz version 32) i get a file "mykernel" without any errors.
But after installing the vmlinuz-file into partition /dev/sda1 (PReP boot)
i get only this messages at boot time (recorded via RS232 console)
you didn't just copy the kernel did you ? you also need to install
/boot and /lib/modules, and new initrd.
did you use debian kernel package to make a .deb or did you build
linus
style ?
Post by Dieter Geyer
#############
loaded at: 00400400 00A75F4C
relocated to: 00800000 00E75B4C
zimage at: 00809DB4 009570CC
initrd at: 00958000 00E6C910
avail ram: 00400000 00800000
Linux/PPC load: console=ttyS0,9600 console=tty0
Uncompressing Linux...done.
Now booting the kernel
##############
after this, no more output occurs, no hdd activity, no keys are accepted.
now my question : what's wrong ?
maybe you made an error in configuring your kernel.
what did you change ? what config tool did you use ? (makeconfig or
what)
I did the old way, make oldconfig, NOTHING changed, i want to test the
compiling/linking/installing a kernel. initrd is the old one, it was read
at the building process into the new kernelfile.
Yes, i have copied the new kernelfile to /dev/sda1,
attention : the PReP actitecture doesnt support bootloader (that's why all
things have to be in one big file) , so /boot is not so important,
but i have copied the new kernel also into it.

I'm wondering, that the boot process stops after "Now booting the kernel".
As you can see, the RAM is containing the zimage and a initrd,
i'm missing the "normal" boot messages .... (this should be working without
modules and initrd, these files will be activated later)

like
...
Linux version 2.6.18 (***@yyyyy) (gcc version x.y.z) #3 Fri Feb 08
02:47:49 EST 2006
Built 1 zonelists
Kernel command line: console=ttyS0,9600 console=tty0
PID hash table entries: 256 (order: 8, 4096 bytes)
...

Thanks
Dieter
tortoise
2008-01-16 09:42:29 UTC
Permalink
Post by Dieter Geyer
Post by tortoise
Post by Dieter Geyer
Hi all,
i have an old Motorola PowerStack II (with 300 MHz PPC 604, 256 MB RAM)
and installed (with some tricks) a debian 4.0 R1 (Etch) on this maschine.
This is a PPC PReP achitecture.
It boots from the original boot-kernel into text-based multi-user console.
Now i want to create a new kernel, after compiling and linking (with tool
mkvmlinuz version 32) i get a file "mykernel" without any errors.
But after installing the vmlinuz-file into partition /dev/sda1 (PReP boot)
i get only this messages at boot time (recorded via RS232 console)
you didn't just copy the kernel did you ? you also need to install
/boot and /lib/modules, and new initrd.
did you use debian kernel package to make a .deb or did you build
linus
style ?
Post by Dieter Geyer
#############
loaded at: 00400400 00A75F4C
relocated to: 00800000 00E75B4C
zimage at: 00809DB4 009570CC
initrd at: 00958000 00E6C910
avail ram: 00400000 00800000
Linux/PPC load: console=ttyS0,9600 console=tty0
Uncompressing Linux...done.
Now booting the kernel
##############
after this, no more output occurs, no hdd activity, no keys are accepted.
now my question : what's wrong ?
maybe you made an error in configuring your kernel.
what did you change ? what config tool did you use ? (makeconfig or
what)
I did the old way, make oldconfig, NOTHING changed, i want to test the
compiling/linking/installing a kernel. initrd is the old one, it was read
at the building process into the new kernelfile.
Yes, i have copied the new kernelfile to /dev/sda1,
attention : the PReP actitecture doesnt support bootloader (that's why all
things have to be in one big file) , so /boot is not so important,
but i have copied the new kernel also into it.
I'm wondering, that the boot process stops after "Now booting the kernel".
As you can see, the RAM is containing the zimage and a initrd,
i'm missing the "normal" boot messages .... (this should be working without
modules and initrd, these files will be activated later)
like
...
02:47:49 EST 2006
Built 1 zonelists
Kernel command line: console=ttyS0,9600 console=tty0
PID hash table entries: 256 (order: 8, 4096 bytes)
...
Thanks
Dieter
i would strongly advise using debian kernel package rather than the
linus
upstream way.

are you sure (this is unlikely) that the kernel version of your
sources
be exactly the same as the first one. if mismatch then you have
a problem.

also check anyway you can if anything is not the same. for instance
are the kernels exactly the same size ? is the .deb package the same
size as the downloadable package ?

i am curious what "tricks" as you say you needed to install. i have
install on a couple "oldworld" mac, one a 604 and another a 750.
they had some problems with the initrd. i built the kernel with the
options to build in the necessary modules rather than load through
initrd and then read remaining modules from the root partition
directory.
(i had to do this on another machine or on an older system before
i could get etch to boot -- also for a while i used an ubuntu kernel).

i did have in these circumstances similar symptom -- that is a black
screen or hanging early. sometimes the earlier screen it flashes very
quickly if you can see it at all. i never really saw until i had it
hang at
that point, but sometimes it hung before or after (some of this was
earlier
with different machines during testing of kernel migration).

Continue reading on narkive:
Loading...