Subject: Re: siop problems
To: Lloyd Parkes <lloyd@must-have-coffee.gen.nz>
From: Manuel Bouyer <bouyer@antioche.eu.org>
List: port-sparc64
Date: 03/03/2002 16:43:00
On Sun, Mar 03, 2002 at 07:50:23PM +1300, Lloyd Parkes wrote:
> > Does a 'tr' give usefull infos ?
> 
> The panic was in siop_attach, so I didn't bother getting a stack trace. I have done one now, and here it is
> 
> db> tr
> siop_pci_attach(1f87900, 1f97800, 1c08ce0, 12b5b08, 0, 1f97828) at siop_pci_atta
> ch+0x28                                                                        

This is really strange; siop_pci_attach is a 3 line function.
Can you confirm with gdb where it is ?
On my kernel gdb says that siop_pci_attach+0x28 is at the closing brace of
the funtion. I have a 53c875 in my machine:
siop0 at pci1 dev 1 function 0: Symbios Logic 53c875 (ultra-wide scsi)
siop0: using on-board RAM
siop0: interrupting at ivec 10
scsibus0 at siop0: 16 targets, 8 luns per target

Can you try to boot a kernel without siop to see what PCI vendor/device
ID this adapter has ?
Also, maybe add a few printfs in siop_pci_attach and siop_pci_attach_common
to narrow down the place where the kernel dies.

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
--