Bug 79470

Summary: Kernel panic on reboot (but not on powerup), IBM Thinkpad 570 + RH8
Product: [Retired] Red Hat Linux Reporter: Need Real Name <jahf>
Component: kernelAssignee: Arjan van de Ven <arjanv>
Status: CLOSED CURRENTRELEASE QA Contact: Brian Brock <bbrock>
Severity: high Docs Contact:
Priority: medium    
Version: 8.0CC: me
Target Milestone: ---   
Target Release: ---   
Hardware: i686   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2004-09-30 15:40:17 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Need Real Name 2002-12-11 23:35:28 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.2.1) Gecko/20021130

Description of problem:
I'm setting up an older IBM Thinkpad 570 to be my gateway/router (I
use a PCMCIA 802.11b card for broadband connectivity and need a good
NAT box).

I installed RH8.0 with no problem. It's up and running. If I power
down the system before each boot, I have no issues.

However, if I reboot the machine (reboot, not power down/power up) it
hangs with a kernel panic "No init found". Powering down the system
and powering it up always works. Rebooting always panics and hangs.
Obviously not a workable condition for a system that is going to be
used as a gateway.

NOTES: 

* I have RH8.0 installed from the same CDs on a newer IBM Thinkpad X20
and everything works appropriately from a reboot. I've applied the
same patches to both systems such that up2date is happy on both.

* I've read many posts about how not having a /initrd directory, even
if it's empty, will cause this kernel panic. However, /initrd is
there. If it weren't then I would expect this kernel panic to also
happen on powering up.

* I also read a few posts about drive geometry being different in
Linux vs. the system BIOS causing this. If this is the case I have a
secondary problem: I don't seem to have a BIOS option to specific disk
geometry on this laptop!

* I couldn't find anything closer to the problem on groups.google or
bugzilla.redhat.


Any help is appreciated. The exact text of the last few lines of the
boot screen before the panic are:

Loading ext3 module
Mounting /proc filesystem
Creating block devices
Creating root device
mkrootdev: label / not found
Mounting root filesystems
mount: error 2 mounting ext3
pivotroot: pivot_root(/sysroot,sysroot/initrd) failed: 2
umount /initrd/proc failed: 2
Freeing unused kernel memory: 172k freed
Kernel panic: No init found.  Try passing init= option to kernel.



PS. I posted this on comp.os.linux.redhat ... if I get a fix from either source
I'll cross-post the answer both here and the mailing list.

Version-Release number of selected component (if applicable):


How reproducible:
Always

Steps to Reproduce:
1. Install RH8
2. run up2date to get latest versions (as of 12/11/02 I have all updates)
3. reboot, see kernel panic
4. power down/power up, see boot succeed
	

Actual Results:  kernel panic when rebooted

Expected Results:  boot up normally

Additional info: wasn't sure which component to file under. There is not
category for "boot" or "init".

Comment 1 Need Real Name 2002-12-12 02:14:44 UTC
FYI, I decided to confirm that the problem existed before performing any
updates. I reinstalled RH8 from CD onto the system, formatting the old RH8
partitions. 

After a fresh install rebooting still causes the kernel panic with the older
2.4.18-14 kernel.

The only other anomaly I could think of was that I had the CD tray open this
whole time. I powered up and rebooted with the CD tray closed (and empty) the
entire time. No change, still panics if I reboot but will load properly from a
cold boot.

Also, as an FYI, I've had Mandrake 8.1, Gentoo 1.4 (that was fun :), Win98 and
Win2K Pro (Win2K until this morning) all running on this machine in the past few
months (it was my lab laptop until I got a 2nd one). I haven't experienced this
problem on other OSes, so I'm assuming that it is related to RH8 and not my
hardware.






Comment 2 Arjan van de Ven 2002-12-12 07:34:40 UTC
can you try adding
"reboot=c"
to the vmlinuz line of /boot/grub/grub.conf ?
that changes the method the kernel uses for rebooting.
(of course you need to reboot after this change, and then reboot AGAIN to see if
it has effect)

Comment 3 Need Real Name 2002-12-12 18:09:55 UTC
Steps: 
* edited /boot/grub/grub.conf
* appended " reboot=c" to the vmlinux line after root=LABEL=/
* powered down/powered up
* rebooted machine from gdm

Result: no change

...

I read up on the reboot= parameter in the kernel and decided to try a
combination of options to see if I have any luck. 

Questions:

1) Can you tell me what the default is with nothing specified? I'm assuming
warm,hard.

2) Documentation on a page about lilo with redhat shows the order of options to
be {hard|bios},{warm|cold}. However other pages reverse that to
{warm|cold},{hard|bios}. Does the order matter with RH8+2.4.18+Grub?

...
NEW TESTS: Combinations of reboot=

SUMMARY: no combination seems to work, though changing combinations produced a 
new result.

DETAILS:

* no "reboot=" parameter (default) - case described by initial bug report

* reboot=c - fail as per initial bug

* reboot=cold - fail as per initial bug

* reboot=cold,hard - fails as initial bug report

* reboot=cold,bios - fails as initial bug report and as per =cold,hard

* reboot=warm - ugly freeze after reboot, new result:

(note: lots of lines similar to the first 4 quoted, only typed in the last 4)

...
[<c0114ed5>] apm_cpu_idle [kernel] 0 (0xc0341fc4))
[<c0107040>] default_idle [kernel] 0x0 (0xc0341fd0))
[<c0105000>] stext [kernel] 0x0 (0xc0341fd4))
[<c01070d4>] cpu_idle [kernel] 0x24 (0xc0341fdc))

Code:  Bad EIP value.
 <0>Kernel panic: Aiee, killing interrupt handler!
In interrupt handler - not syncing
...

* reboot=warm,hard - fail during reboot as per initial report

* reboot=warm,bios - fail with new ugly error as per =warm

ERRATA: I tried to reproduce the weird error on =warm and =warm,bios and on
repeats they both failed as per the initial bug. Unsure what the cause would
have been.

RESULT: no change, still no successes on rebooting. Possibly the problem lies
outside the reboot= kernel code?


Comment 4 Arjan van de Ven 2002-12-12 18:21:58 UTC
> RESULT: no change, still no successes on rebooting. Possibly the problem lies
> outside the reboot= kernel code?

quite possible; the reboot= code is obviously first suspect in a problem as you
reported. the APM oops you gave makes me wonder if you can try "apm=off" and see
if that helps ? maybe using apm confuses things ;(


Comment 5 Need Real Name 2002-12-12 18:42:42 UTC
* removed "reboot=c" from the vmlinuz line

* added apm=off to the vmlinuz line

No change, assuming that's where apm=off should have gone.

Problem is still consistently the same as from initial report ... ie:

Loading ext3 module
Mounting /proc filesystem
Creating block devices
Creating root device
mkrootdev: label / not found
Mounting root filesystems
mount: error 2 mounting ext3
pivotroot: pivot_root(/sysroot,sysroot/initrd) failed: 2
umount /initrd/proc failed: 2
Freeing unused kernel memory: 172k freed
Kernel panic: No init found.  Try passing init= option to kernel.

The first error in the listing above is "mrootdev: label / not found" ... is
there anything that would cause the system to loose the drive geometry during a
reboot but would not be a problem during boot from a power up?



Comment 6 Need Real Name 2002-12-12 20:03:26 UTC
Addendum ... I decided to keep reading on other peoples' problems involving
pivotroot messages and ended up trying the following with no success. Just
adding this to streamline possible tests.

* edited grub.conf from "root=LABEL=/" to "root=/dev/hda5", reverted on failure

* edited /etc/fstab to change "LABEL=/" to "/dev/hda5" and "LABEL=/boot" to
"/dev/hda2" (/dev/hda1 is an empty 100MB vfat partition in case I have to use
DOS to update firmware at some point), reverted on failure

* created a new initrd image with mkinitrd ... didn't expect this to work,
reverted on failure

* I logged the IDE section of dmesg during a successful boot, follows this. I
tried to compare it to a failed reboot but it goes by very fast. Is there a way
for me to keep the dmesg info from a failed boot attempt? When I boot again from
power-up it's obviously replaced with the new, successful boot.

... (note: the following is from a successful boot)
Uniform Multi-Platform E-IDE driver Revision: 6.31
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
PIIX4: IDE controller on PCI bus 00 dev 31
PIIX4: chipset revision 1
PIIX4: not 100% native mode: will probe irqs later
    ide0: BM-DMA at 0x1800-0x1807, BIOS settings: hda:DMA, hdb:pio
    ide1: BM-DMA at 0x1808-0x180f, BIOS settings: hdc:pio, hdd:pio
hda: IBM-DBCA-204860, ATA DISK drive
hdc: CRN-8241B, ATAPI CD/DVD-ROM drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide1 at 0x170-0x177,0x376 on irq 15
blk: queue c03afd84, I/O limit 4095Mb (mask 0xffffffff)
blk: queue c03afd84, I/O limit 4095Mb (mask 0xffffffff)
hda: 8007552 sectors (4100 MB) w/420KiB Cache, CHS=993/128/63, UDMA(33)
ide-floppy driver 0.99.newide
Partition check:
 hda: hda1 hda2 hda3 hda4 < hda5 >
Floppy drive(s): fd0 is 1.44M
...



Comment 7 Need Real Name 2002-12-12 20:08:20 UTC
NOTE: 

I have noticed each time that during a fresh power-on boot, the "Partition
check:" segment of booting takes virtually no time to complete and spit out
"hda: hda1 hda2 hda3 hda4 < hda5 >".

However, when it does the "Partition check:" during reboot, it takes 4-6 seconds
during which the hard drive is very active, before it spits out the result
(which passes to quickly for me to confirm 100% that it succeeds) and proceeds
to halt the boot as mentioned above.

...

I'll quit pegging you with fresh entries until you've had a chance to read
through what I've posted today. Thanks for looking at this.


Comment 8 Need Real Name 2002-12-31 23:45:34 UTC
Haven't heard anything recently here. I didn't find any answers on Google, just
someone else who claims to have had this same problem since Red Hat 7.3 ... see
<a
href="http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&threadm=1cf0b4c8.0212111527.1ec7e5b3%40posting.google.com&rnum=1&prev=/groups%3Fq%3Dkernel%2Bpanic%2Bon%2Breboot%252C%2BIBM%2BThinkpad%2B570%26ie%3DUTF-8%26oe%3DUTF-8%26hl%3Den%26btnG%3DGoogle%2BSearch">
this article</a>.

If there is anything else I can do to help this along, please let me know.

Thanks,
/Geoff


Comment 9 Need Real Name 2003-02-24 23:47:29 UTC
Requesting an update ... will this be fixed? It was working in previous Red Hat
and Mandrake versions, so it would seem to be something that could be tracked down. 

Comment 10 Alan Cox 2003-06-05 13:10:48 UTC
It appears that your bios doesnt restart hard disks that the OS placed
(correctly) into power saving mode when soft booting and decides no disk is
present. Its hard to be sure.


Comment 11 MCSquared 2004-04-27 15:07:51 UTC
For what it's worth, I also have a Thinkpad 570 which exhibits the
exact same behavior after I install Fedora Core 1.

Comment 12 Bugzilla owner 2004-09-30 15:40:17 UTC
Thanks for the bug report. However, Red Hat no longer maintains this version of
the product. Please upgrade to the latest version and open a new bug if the problem
persists.

The Fedora Legacy project (http://fedoralegacy.org/) maintains some older releases, 
and if you believe this bug is interesting to them, please report the problem in
the bug tracker at: http://bugzilla.fedora.us/