itohy@netbsd.org (ITOH Yasufumi) writes:
>   - On AlphaStation 500, is it possible to detect which part is damaged
>     from machine check?
>     The AS200 code seems to have such diag message.

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.

Here are the values for the machine check code, from the table on page
8-30 of the document I mentioned previously:

Processor-detected errors:
0x80  Tag parity error 
0x82  Tag control parity error 
0x84  Generic Hard error 
0x88  Processor Detected Uncorrectable ECC error 
0x8A  Bugcheck generated by OS specific PALcode 
0x90  Bugcheck generated by PALcode 
0x96  I-Cache read retryable error 
0x98  Processor Detected Hard error 

CIA-detected errors:
0x203 System Detected Uncorrectable ECC error
0x205 Parity error detected by CIA
0x207 Non-existent Memory Error
0x209 PCI SERR detected
0x20b PCI Data Parity Error detected
0x20d PCD Address Parity Error detected
0x20f PCI Master Abort Error
0x211 PCI Target Abort Error
0x213 Scatter/Gather PTE invalid error
0x215 Flash ROM Write Error
0x217 IO A Timeout detected
0x219 IO CHK #, EISA add-in board parity or other catastrophic
0x21b EISA Fail-safe timer time-out
0x21d EISA Bus time-out
0x21f EISA Software generated NMI
0x221 Unexpected EV5 IRQ[3] interrupt

        - Nathan