boot - "Memory Address Not Aligned"

[jacko24]

On a Sparc5, I am getting "Stack Underflow" (3 times) followed by "Memory Address Not Aligned" and then stopping at Ok prompt. When doing a "test-all" I see IAR errors such as:
IAR Error: Actual=80 Expected=0
IAR Error: Actual=80 Expected=1
.... etc to about 5 or 6

I have not been able to find a resources explaining what those errors mean.

Most reading on the web suggests memory modules placement, however the machine ran fine for several weeks and several reboots, until one day this started. I have swapped in/out 4 diff memory modules, and I am beginning to doubt that the modules are to blame.

If anyone has any suggestions please reply.

Thanks

 

[ianf]

Yuk ! Sound motherboardy to me if you are getting the same message off different MMs.

Have you tried putting just one MM in at a time rather than replacing all of them ? Are the "new" MMs from another Sparc5 or a different machine altogether ?

Silly questions, I know, but a new Sparc 5 motherboard may be hard to find nowadays (depending on your location & service agreements). Ian

"IF" is not a word it's a way of life

 

[teaboy]

Actually, if it comes to that (and unfortunately it sounds like it might), check ebay. They used to have sparc 5 parts quite often.

T

 

[Guest_imported]

I have seen this problem as well across 10+ SPARC ULTRA 10's. It appears to be a problem with the L2 Cache memory on the processor. After replacing the processors in these SPARC's our problems went away. We were also getting PARITY ERRORS and DATA ACCESS ERRORS. Check your /var/admin/messages files for other possible errors occuring.

 

[rj5132]

Sunsolve.sun.com is a good online resource for all types of issues.

 

[Guest_imported]

It's very important not to leave any unpowered SCSI CD-ROM drives connected (am guessing any unpowered SCSI devices) to any Sparc box at power-up as you'll get that same [red]"Stack Underflow" (3 times) followed by "Memory Address Not Aligned"[/red] message. I had this problem (on a Sparc 5 Box) but was solved by obviously removing any SCSI devices before power-up. It may be better to check this before forking out any cash for any extra memory as the problem will remain regardless of the memory you put in. [red][/red]