On Wed, 2010-12-22 at 11:12 +0200, Ilan Tayari wrote: > Package: linux-2.6 > Version: 2.6.26-26lenny1 > Severity: normal > > I've recently installed Lenny on an old HP NetServer E60 machine. > The installer worked perfectly, but after the reboot I got a kernel > panic with flashing keyboard lights. > > A search lead me to bug 486168, which seems very similar (same > hardware, same panic message). > The comments there suggest using rescue mode to install a 486 kernel > and boot it. > I did that, and the system now works, but the 486 kernel recognizes > just one CPU and not both. > (I'm not sure if this is a limitation of the 486 kernel or not) That's correct, it is a limitation of the 486 kernel. > A late comment in bug 486168 suggests disabling the memory hole in the > BIOS settings. > I checked the BIOS settings on my machine and the memory hole is > disabled, yet the kernel panic still happens. > > I'd like to load the 686 kernel for increased performance, and > possibly to take advantage of SMP. > > -- Package-specific info: > ** Kernel log: boot messages should be attached > > Can't make it save the boot messages to a file, so uploaded as YouTube video at: > http://www.youtube.com/watch?v=0ahFTUIY4ZY > I hope the quality is enough. crash it around 00:22 [...] The key messages are clear: [ 5.812886] ahc_pci:0:5:0: Using left over BIOS settings ... [ 21.396440] Kernel panic - not syncing: HOST_MSG_LOOP with invalid SCB ff The panic is generated by the driver code used for various Adaptec SCSI adapters. I think that this error would normally indicate a fault in the SCSI adapter and the driver should not panic, but that may not be easy to avoid. However, since the 486 kernel works, the fault is clearly in software. Could you test the 686 kernel with the added kernel parameter 'maxcpus=1'? Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse.
Attachment:
signature.asc
Description: This is a digitally signed message part