In article <mtuzo9jyo0s.fsf@contents-vnder-pressvre.mit.edu>
nathanw@MIT.EDU writes:

> I should also note that if you have a kernel built from sources after
> July 3, the machine check handler will print out the machine check
> code, which is located at param + 0x10; you should also be able to
> examine it from ddb or SRM.

I tried with newer kernel.

unexpected machine check:

    mces    = 0x1
    vector  = 0x670
    param   = 0xfffffc0000006060
    pc      = 0xfffffc00003002d8
    ra      = 0xfffffc00003003e0
    code    = 0x98
    curproc = 0xfffffc000703e2f0
        pid = 485, comm = skkdic-sort


So the check is reporting...

> 0x98  Processor Detected Hard error 

Hmm, what is detected?

In article <mtu3d7b0zcy.fsf@contents-vnder-pressvre.mit.edu>
nathanw@MIT.EDU writes:

> The 500/600 line has a pdf document from DEC on cgd's documentation
> mirror,
> ftp://ftp.netbsd.org/pub/NetBSD/misc/dec-docs/index.html#alpha-sys-as600
> 
> It does describe the logout frame in enough detail to write the
> as500/600 version of the code I wrote for the as200.

The logout frame seems similar to the ev5 one.
Okay, I'll try and write some code in the weekend.

Thanks,
-- 
ITOH Yasufumi