6.0-beta4 on 2100A: unexpected machine check

[Tillman Hodgson]

Howdy folks,

I recently acquired a nice 4-CPU 512MB 2100A "Lynx" machine. This is my
first FreeBSD install on Alpha (though I've run it on i386 and sparc64
for years). I ran into rouble while trying to boot the install kernel.
THis might be related to the thread I found at
http://lists.freebsd.org/pipermail/freebsd-alpha/2005-August/002659.html
... at least, the symptoms seem similar to my uneducated-in-Alpha eye.

Unfortunately, I don't have a serial console hooked up at the moment
(it's in the wrong room to connect to my digiport and it's much too
large to move up the stairs), so all of the following is tped in from my
handwritten notes.

After issuing a 'boot DKA600':

----------
FreeBSD/SMP: Multiprocessor system detected: 4 CPUs
CPU0 (BSP): PAL ID: 0
CPU1  (AP): PAL ID: 1
CPU2  (AP): PAL ID: 2
CPU3  (AP): PAL ID: 3
t20: using interrupt typ1 on pci bus 0
t20: <T2 Core Logic chipset>
pcib0: <T2 PCI host bus adapter> on t20
eisab0: <PCI-EISA bridge> at device 2.0 on pci0
----------

(Note that there are no EISA cards, just some PCI cards)

The error then occurs, and looks like this:

----------
unexpected machine check:

    mces    = 0x1
    vector  = 0x670
    param   = 0xfffffc0000006000
    pc      = 0xfffffc00003d49f0
    ra      = 0xfffffc00003d49d4
    curproc = 0xfffffc0000912be0
        pid = 0, comm = swapper

[thread pid 0 tid 0]
stopped at eisa_probe_slot+0x80: add s0,#01,s0 <s0=0x0>
----------

I notice that at this point I've been dumped to the db> debugger prompt,
but the key mapping is messed up ... pressing 'g' gives me ',' for
example.

The output of 'show device' is:
----------
DKA0.0.0.2001.0   DKA0   Seagate ST32550W 9104
DKA600.6.0.2001.0 DKA600 RRD45 1645
DRA0.0.0.2007.0   DRA0   5 member RAID 5
DVA0.0.0.1000.0   DVA0   RX25/RX23
EWA0.0.0.2008.0   EWA0   (the MAC address)
PKA0.7.0.2001.0   PKA0   SCSI Bus ID 7
PKB0.7.0.7.0      PKB0   SCSI Bus ID 7 R01 A12
----------

The output of 'show FRU' doesn't include any errors. Aside from the
network card and the video card, all the remaining PCI cards are SCSI
adapters (including the one full-length DAC960 RAID controller). I did
pull 2 additional SCSI adapter that weren't in use to reduce the
device-related possibilities.

Does this problem look familiar to anyone? What's the best way to
approach an "unexpected machine check" error?

Thanks in advance for your time,

-T



-- 
Infinity attracts us like a floodlight in the night, blinding us to the 
excesses it can inflict upon the finite.
	- Meditations from Bifrost Eyrie