I seem to have a similar problem.
Linux  2.6.24-18-xen #1 SMP Wed May 28 22:50:01 UTC 2008 x86_64 GNU/Linux
Hardware: Dell PowerEdge 2900 with PERC5/i

Stumbled across this when doing some disk performance tests. 
/dev/sda and /dev/sdb are disks on the PERC (setup as two separate single disk 
raid0)
/dev/md3 is a software raid1 from /dev/sda5 and /dev/sdb5.

I changed the scheduler from the default cfq to deadline
echo deadline > /sys/block/sda/queue/scheduler
echo deadline > /sys/block/sdb/queue/scheduler

When I run this commands (or similar):
dd if=/dev/zero of=/dev/md3 bs=4K count=16K
I get a kernel panic, but _only_ when I'm not using the cfq scheduler. noop, 
anticipatory and deadline all give the same behaviour (kernelpanic) but not cfq 
(at least I haven't been
able to make it crash)
It doesn't happen with a non Xen kernel either.

I get these messages in the console

PCI-DMA: Out of SW-IOMMU space for 65536 bytes at device 0000:02:0e.0
[  269.070347] ------------[ cut here ]------------
[  269.097818] kernel BUG at 
/build/buildd/linux-2.6.24/debian/build/custom-source-xen/drivers/scsi/megaraid/megaraid_sas.c:466!
[  269.165440] invalid opcode: 0000 [1] SMP

See attachment for full console output...



** Attachment added: "kernelpanic.txt"
   http://launchpadlibrarian.net/15355881/kernelpanic.txt

-- 
Xen dom0 crash when writing to LVM volume with DELL PERC 5/i RAID
https://bugs.launchpad.net/bugs/227120
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to