Bug 119244 - Kernel won't boot; Kernel panic Bad EIP Value
Summary: Kernel won't boot; Kernel panic Bad EIP Value
Alias: None
Product: Red Hat Enterprise Linux 3
Classification: Red Hat
Component: kernel
Version: 3.0
Hardware: All
OS: Linux
Target Milestone: ---
Assignee: Dave Anderson
QA Contact: Brian Brock
Depends On:
TreeView+ depends on / blocked
Reported: 2004-03-27 01:22 UTC by Mike Andrewjeski
Modified: 2007-11-30 22:07 UTC (History)
2 users (show)

Clone Of:
Last Closed: 2004-08-26 16:12:28 UTC

Attachments (Terms of Use)

Description Mike Andrewjeski 2004-03-27 01:22:46 UTC
Description of problem:

RHEL AS 2.1 on an IBM eserver 345 Bios level = 1.13

Kernel Locked up and on reboot won't come up. Screen output at boot:

EIP: 0060:[<8491a899>] Not tainted 
EFLAGS: 00010246

EIP is at Using_versions[]
eax: 0000003b ebx: f6c4a000 ecx: 00000000 edx: 00000000 
esi: 00008136 edi: 00000296 ebp: 67890000 esp: c8ae7dd0 
ds: 0058 es: 000 ss: 0018 

Process init (pid 25, stackpage=f6a5f000) 

call Trace : 

Code: Bad EIP value 

Kernel Panic: Fatal Exception

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

How reproducible: completely

Steps to Reproduce:
Actual results:
kernel panic

Expected results:

Additional info:

Comment 1 Mike Andrewjeski 2004-03-29 23:24:18 UTC
This box is actually RHEL 3.0 and not 2.1

Comment 3 Ernie Petrides 2004-08-25 23:58:58 UTC
Mike, do you have a complete RHEL3 installation on that box?

I couldn't even find a reference to a "Using_versions" symbol
in the latest sources (corresponding to the soon-to-be-released
Update 3).  Does the same problem occur with a more recent RHEL3
installation (such as 2.4.21-15.EL, i.e. Update 2)?

Comment 4 Dave Anderson 2004-08-26 13:06:00 UTC
This must have been a manual copy?  Because it should be
"Using_Versions" (capital "V"), which I believe ends up as
the last kernel symbol stored.  So when a bogus EIP like this
one is encountered, the kernel's ksymoops output typically
consists of "EIP is at Using_Versions+<some bizarre offset>".

So the bogus EIP is 0x8491a899, and when it was cycled through
lookup_symbol(), it looks to have come out as:

  EIP is at Using_Versions+0x8491a898 (2.4.21-4.EL)

or (EIP-1), if it was transposed correctly.  And that doesn't make
any mathematical sense AFAICT.

In any case, like Jason said, there's nothing to work with here.

Comment 5 Mike Andrewjeski 2004-08-26 16:11:47 UTC
Thanks Dave & Ernie,

I appreciate that you've taken this bug even after such a long time.

I do need to let you both know that I have resolved this issue and that 
problem was that init was corrupted. Once we restored a good copy the
machine runs well again. My apologies in not closing this bug out
myself. Can either of close this bug? or tell me how I can?

Once Again thanks for efforts they are greatly appreciated


Note You need to log in before you can comment on or make changes to this bug.