FreeBSD IA64 5.3/RELEASE on Compaq DL590. (boot error)

[Marcel Moolenaar]

On Wed, Feb 09, 2005 at 07:30:30PM -0500, bigcat at runningleopard.com wrote:
> 
> pci25: <ACPI PCI bus> on pcib6
> ida0: <Compaq Integrated Array controller> port 0xc000-0xc0ff mem \
	0x43000000-0x43ffffff,0x41000000-0x41ffffff irq 110 at device \
	8.0 on pci25
> ida0: [GIANT-LOCKED]
> 
> fatal kernel trap (cpu 0):
> 
>     trap vector = 0x14 (Page Not Present)
>     cr.iip      = 0xe000000004288350
>     cr.ipsr     = 0x1210080a2010 (mfl,ic,dt,dfh,rt,cpl=0,it,ri=1,bn)
>     cr.isr      = 0x20400000000 (code=0,vector=0,r,ei=1)
>     cr.ifa      = 0xa0000000dbd0610c
>     curthread   = 0xe0000000049815e8
>         pid = 0, comm = swapper
> 
> [thread 0]
> Stopped at      ida_command+0x2e1:      [M1]    ld8 r14=[r14]
> db>

Ouch...

> Any ideas?

First of all: A backtrace would definitely help. At the "db>" prompt
type:
	trace

Secondly: It looks like the ida(4) driver is not 64-bit clean. The
faulting instruction is an 8-byte wide load from an address that's
not 8-byte aligned. This is always fatal in the kernel, although
it's not the reason for the panic in this case. Nonetheless, the
code is suspicious...

> I'm suspecting that there is no kernel loadable for the Compaq Disk Array,
> perhaps?

The ida(4) driver comes as a loadable kernel module, but happens
to be compiled into the kernel by default. Unfortunately this means
that I need to create a new snapshot with the ida(4) driver removed
from the kernel, or you remove the Compaq Disk Array from the machine
until you can build your own kernel without the ida(4) driver.

As for fixing the ida(4) driver: I don't have the hardware, so I
can't do that and I need to see the backtrace to see if one can
fix the ida(4) driver on any 64-bit machine, or if it has to be
an ia64 box...

FYI,

-- 
 Marcel Moolenaar	  USPA: A-39004		 marcel at xcllnt.net