Bug 245662 - [LSI 5.2 bug] creating a partition on a RHEL 5 ppc64 system causes a kernel page fault.
Summary: [LSI 5.2 bug] creating a partition on a RHEL 5 ppc64 system causes a kernel p...
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: Red Hat Enterprise Linux 5
Classification: Red Hat
Component: kernel
Version: 5.0
Hardware: ppc64
OS: Linux
low
urgent
Target Milestone: rc
: ---
Assignee: Scott Moser
QA Contact: Martin Jenner
URL:
Whiteboard:
Depends On:
Blocks: 217105 425461
TreeView+ depends on / blocked
 
Reported: 2007-06-25 22:36 UTC by Abdel Sadek
Modified: 2009-06-19 23:03 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2007-12-19 16:31:04 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)
Server RHEL 5.0 /proc/slbinfo /proc/meminfo /proc/cpuinfo (3.39 KB, application/x-zip-compressed)
2007-06-25 22:36 UTC, Abdel Sadek
no flags Details

Description Abdel Sadek 2007-06-25 22:36:45 UTC
Description of problem:
creating a partition on a RHEL 5 ppc64 system causes a kernel page fault.

Version-Release number of selected component (if applicable):
2.6.18-8.el5
 
How reproducible:
 

Steps to Reproduce:
1. Connect a RHEL 5.0 ppc64 system to a LSI storage array.
2. map 9 volumes of sizes between 1 Gb and 5 Gb.
3.attempt to unregister the device partitions with DM:
  kpartx -d -pp $device
  kpartx -d -p-part $device
  ($device is lun presented from the LSI array)
4. Clear up any previous partition tables: 
   dd if=/dev/zero of=$device bs=16384 count=1 1>/dev/null 2>&1
  ($device is lun presented from the LSI array)
5. make a partition label:
   parted -s $device mklabel msdos 1>/dev/null 2>&1
6. make 4 equal size partitions on each volume:
   my $start = 0;
   my $end = $partSize;
   for(my $i = 0; $i < $partnum; $i++) {
     `parted -s $device mkpart primary $start $end`;
     $start = $end;
     $end = $start + $partSize;
   }
7. all the above steps were included in a script to run on all 9 volumes.
  
Actual results:
when running the script, sometimes the host hangs at begining at the command "dd
if=/dev/zero of=$device bs=16384 count=1 1>/dev/null 2>&1".(step 3).
Sometimes, it passes on the first volume. but it fails at step 3 on the second
volume.
In all cases, the console on the host report the error below:
halfa Unable to handle kernel paging request for data at address 0x00000000
Faulting instruction address: 0xd0000000004680bc


Expected results:


Additional info:

Comment 1 Abdel Sadek 2007-06-25 22:36:45 UTC
Created attachment 157829 [details]
Server RHEL 5.0  /proc/slbinfo /proc/meminfo /proc/cpuinfo

Comment 2 RHEL Program Management 2007-06-27 00:03:54 UTC
This request was evaluated by Red Hat Product Management for inclusion in a Red
Hat Enterprise Linux maintenance release.  Product Management has requested
further review of this request by Red Hat Engineering, for potential
inclusion in a Red Hat Enterprise Linux Update release for currently deployed
products.  This request is not yet committed for inclusion in an Update
release.

Comment 3 Andrius Benokraitis 2007-07-06 05:52:28 UTC
Scott - does this ppc issue look familiar?

Comment 4 Tom Coughlan 2007-07-09 15:08:46 UTC
Abdel,

I would expect to see an oops message, including a stack trace, after the
"Unable to handle kernel paging request". Please check for this. If there is
none, then force a crash dump after the system hangs and post the resulting info. 

Scott, re-assign as appropriate, based on whatever the stack trace shows. 

Tom

Comment 6 Andrius Benokraitis 2007-08-01 04:56:23 UTC
Haven't heard from Abdel from LSI on this - deferring to 5.2 since we are
already past Beta.

Scott - please re-assign.

Comment 8 Scott Moser 2007-11-07 17:47:05 UTC
I'm not really sure who to reassign this to.  Perhaps Tom could do a better job
of re-assigning ?

Comment 10 Andrius Benokraitis 2007-11-08 19:44:49 UTC
Abdel, this bugzilla is going to be CLOSED shortly due to lack of response from LSI.

Comment 11 Peter Martuccelli 2007-12-19 16:20:12 UTC
No resonse for two update cycles, dev NAK.  Feel free to reopen a new request if
this problem persists in RHEL 5.2.


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