Product Fault Management,
Continued
Crash process or system, dependent upon PSL (current mode)
with a fatal bugcheck for the following situations:
—
Retry is not possible.
—
Memory page could not be replaced for uncorrectable
ECC memory error.
—
Uncorrectable tag store ECC errors present in Bcache.
—
Uncorrectable data store ECC errors present in Bcache
for marked as OWNED.
—
Most INT60 errors.
—
Threshold is exceeded (except for cache errors).
—
A few other errors of the sort considered nonrecoverable
are present.
Disable cache(s) permanently if error threshold is exceeded.
Flush and re-enable those caches which have been marked as
good.
Clear the error flags.
Perform return from exception or interrupt (REI) to recover
and restart or continue the instruction stream for the
following situations:
—
Most INT54 errors.
—
Those INT60 and INT54 errors which result in bad ECC
written to a memory location. (These errors can provide
clues that the problem is not memory related.)
—
Machine check conditions where instruction retry is
possible.
—
Memory uncorrectable ECC error where page
replacement is possible and instruction retry is possible.
—
Threshold exceeded (for cache errors only).
—
Return from subroutine (RSB) and return from all polled
errors.
Continued on next page
5–59
