Hi,
I'm trying to upgrade my kernel from 2.6.6 to 2.6.7 and all seems to work fine, but my system won't boot anymore with the new kernel.

At first I thought I got a kernel panic when it got to the ACPI stuff, so I disabled the ACPI from the kernel and the system works fine without it.

I would like to have ACPI support though since it will auto-shutdown the system

A friend gave me his .config for 2.6.7 from another distro (slackware I think) and I used that one to compair it with my own and made some changes.

The system still won't boot with ACPI enabled, but it doesn't seem to give a Kernel Panic anymore.

anybody who knows if this is a know problem with kernel 2.6.7 or who knows a way to fix it ?

my system is :
Pentium 4 @ 2.88Ghz HTT (Hyper Threading enabled and SMP enabled in the kernel)
Intel Desktop Board 865 Chipset
1024 MB DDR Memory @ 400 Mhz

I think that's the most important stuff

You don't mention whether you're using Debian's precompiled kernel-image or if you've compiled your own kernel. I've heard that APM and ACPI don't work well together, so check that you don't have both enabled. The poweroff shutdown works OK in my computer with 2.6.7 (Debian's kernel-image). With kernel 2.4 I used to add "apm power_off=1" to /etc/modules to get APM to do poweroff. Perhaps you could try that if you don't get ACPI working.

Sorry, I forgot to mention that,
It's a custom made kernel from source with make-kpkg (found the instructions on OSNews and it always worked fine for me with with 2.4.x and 2.6.x kernels).

I'll have a look at the .config to see if APM is enabled in the kernel when I get home from work later today, but I think it's not enabled since APM was reported not to work with SMP kernels.

also thanks for the "apm power_off=1" I'll try that with APM if ACPI won't work.

Tnx.

OK, just compiled the kernel with only ACPI support, and it still won't boot
here are the last few lines :

At this point the system freezes.

Well the apm power_off=1 also doesn't seem to work,
I think I see an error about APM flashing by during boot but it's to quick to read it
and I can't get it with dmesg.

There's a bug report on similar problem:

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=251207

You may have no alternative at this time than booting Debian with ACPI disabled. Hopefully this gets fixed soon.

However, I think that you can add the APM power-off option to your boot loader. If you're using lilo, try this: http://www.linuxquestions.org/questi...hreadid=187824 or if you're using grub, add the "apm=on" in menu.lst to the end of the "kernel" line, after all the other stuff.

Well I added the line : append="apm=on" to my lilo.conf and ran /sbin/lilo
the system boots fine but it still won't shutdown, just show the power-down message and then just sits there.

Guess I'll have to wait for this bug to be fixed.
Strange thing is that the link you have states that the problem is knows since kernel 2.6.3 but I never had a problem with kernel 2.6.0 through 2.6.6

Yup, it's strange. Computers are strange things. Try append="apm=power-off" in lilo.conf instead of append="apm=on".

Thanks !
That did the job !

Im running gentoo and I tried to upgrade to 2.6.7 from 2.6.5 with similar results.

It seems to happen to Intel motherboards. I have 2 computers with intel mobo's and the same thing happens on both of them...quite frustrating

I also had to disable ACPI when I upgraded my kernel to 2.6.7. I used the debian packaged kernel for my PIV 2.8GHz, and everything booted fine, but after 30 seconds or so my machine would automatically restart itself. It was quite annoying

RE:

"Unable to handle kernel paging request at virtual address..."

and

"It seems to happen to Intel motherboards"

I wish....MSI boxes riding Soyo motherboards with VIA chipsets, and I see this chronically in Fedora Core 2, 1, Redhat 9...pick your kernel version out of those distros (currently 2.6.7-1.494.2.2). Along with your hardcore intterupt-based kernel errors that get written to the console but never make it to log for spice, and your occasional interesting kernel grafitti in the message log liike "Aug 12 19:38:33 iseeu kernel: spurious 8259A interrupt: IRQ7." for dessert.

Now I'm stressing the box a bit (driving 16 video cameras and writing to disk), but all capture access is controlled through a semaphore shared across 16 processes...in short, its sequential (although I let the kernels do the disk accesses however they want to schedule it...perhaps I'm too trusting).

Thinking, quite seriously, of moving to FreeBSD. (And no, I don't have these problems under Windows.)

kernel-source-2.6.7-1 worked on my inspiron 8200 w/M50 A12 BIOS
kernel-source-2.6.7-2 worked on my inspiron 8200 w/M50 A12 BIOS

-2.6.7-3 & -2.6.7-4 both broke my system at the same point...
(Got everything up-to-date w/ apt-get update && apt-get upgrade && apt-get dist-upgrade *that's how i got the new kernel-sources)

Removed linux symlink and renamed (mv 'ed) the kernel-sources to different directories
thought that 2.6.7-4 would work

Nope: Here is the error at each end...

VFS: Cannot open root device /dev/hda5
or unknown-block

Please append a correct "root" boot option

Kernel panic: VFS: Unable to mount root fs on unknown-block

I have not tried with ACPI disabled as this is a laptop; however, this is a custom made kernel that has gone through very little change since 2.6.2 ... think I had problems with 2.6.5

all boot stuff (IDE support ... ext3 ... support is monolithic)
Why am I getting what I think is an "initrd" goof up?

After all that tried an make-kpkg initrd or make-initrd, can't remember, but it went through all of the kernel options and replied with
nothing to do for "all"

So I'm stumped! Please help.
Again, I have checked the different versions each time and see no difference that would make me change configurations. Please help!

Is 2.6.7 a development kernel?

Check this out-->http://www.linuxquestions.org/questi...hreadid=216424


Thanks in advance!

That sounds like a grub/lilo config issue...the kernel is telling you it doesn't know where to boot from, hence its waiting for you to "append" the correct /dev/hd?? to the kernel boot line.

Or, if you know that is indeed the correct /dev/hd??, then your disk/init RD got toasted somehow. Ya gotta rescue disk (cd or floppy)?

Then why does linux.old still boot fine...its 2.6.7 version 2?

When i replaced everything again with good ole backup, same thing
Tried updating from 2.6.6-final to 2.6.7-3 and 2.6.7-4, no dice
but again, 2.6.6 and 2.6.7-1 & -2 are fine

Have a kernel development partition without X for checking out kernels specifically
nothing fancy shmancy here, however...they work without kiccups

mmmm, im stumped but think you may be right about init scripts...but never messed with them before

What's a good place to start...kernel script (when the penguin would appear if i boot-icons) ?

thinking that an apt-get upgrade or dist-upgrade modified them but againi, they work for 2.6.7-1 & -2
dunno

Please continue the feed and check out my url -->http://www.linuxquestions.org/questi...hreadid=217098 ---> same thing

Thanks in advance cause no one is attempting to answer my original thread posts.