Nortel DMS 100 Series Maintenance Manual page 68

2-52 Maintenance overview
Parity interrupt handler
The interrupt handler generates data on a memory fault. The interrupt handler
stores the data in the appropriate area for reporting and action. The
requirements for this function follow:
•
•
•
•
•
•
•
•
•
XPM code protection
The parity audit process examines protected and non-protected memory areas.
This feature allows writing to a write-protected memory area. This condition
allows for pattern testing a parity fault byte in protected space. This condition
also allows querying if a physical address lies in loaded code.
The parity exception handler determines if the parity fault occurred in
execution space. The parity exception handler can initiate a quick recovery
process because of this action.
The parity audit process examines protected and non-protected memory areas.
The XPM code protection allows writing to a write-protected memory area.
This condition allows for pattern testing a parity fault location byte. This
condition also allows querying if a given physical address lies in loaded code.
The parity exception handler determines if the parity fault occurred in
execution space. The parity exception handler can initiate a quick recovery
process because of this action.
Memory allocation
The memory allocation feature provides additional flexibility in memory use
to address memory management issues. This flexibility protects codes and
critical data from damage that errant software causes. The flexibility also
reduces the risk of outages.
297-8223-550 Standard 04.02 November 2000
Define a single entry point that can be bound as an interrupt handler.
Be continuously available and correctly handle interrupts that start from
the time that the interrupt handle is bound.
Be able to distinguish between the two conditions when an interrupt can be
raised. The interrupt can be raised if entered from parity audit or not
entered from parity audit.
Localize the fault to a specified byte address.
Perform low-level categorization of the fault (hard/soft/intermittent)
regardless of the interrupt source.
Determine if the fault address is in code space or is not in code space.
Leave the fault indication in place if the fault is not corrected.
Save the fault data according to where the interrupt originated.
Exit the handler according to where the interrupt originated.