4-6 Computer Group Literature Center Web Site
Hawk Programming Details
4
Error Notification and Handling

The Hawk ASIC can detect certain hardware errors and can be

programmed to report these errors via the MPIC interr upts or the Machine

Check Interrupt. The following table summarizes how the hardware errors

are handled by the MVME5100 series:

Table 4-3. Error Notification and Handling

Cause Action
Single-bit ECC Store: Write corrected data to memory
Load: Present corrected data to the MPC master
Generate interrupt via MPIC if so enabled
Double-bit ECC Store: Terminate the bus cycle normally without writing to SDRAM
Load: Present uncorrected data to the MPC master
Generate interrupt via MPIC if so enabled
Generate Machine Check Interrupt to the Processor(s) if so enabled
MPC Bus Time Out Store: Discard write data and terminate bus cycle normally
Load: Present undefined data to the MPC master
Generate interrupt via MPIC if so enabled
Generate Machine Check Interrupt to the Processor(s) if so enabled
PCI Target Abort Store: Discard write data and terminate bus cycle normally
Load: Return all 1s and terminate bus cycle normally
Generate interrupt via MPIC if so enabled
Generate Machine Check Interrupt to the Processor(s) if so enabled
PCI Master Abort Store: Discard write data and terminate bus cycle normally
Load: Return all 1s and terminate bus cycle normally
Generate interrupt via MPIC if so enabled
Generate Machine Check Interrupt to the Processor(s) if so enabled
PERR# Detected Generate interrupt via MPIC if so enabled
Generate Machine Check Interrupt to the Processor(s) if so enabled
SERR# Detected Generate interrupt via MPIC if so enabled
Generate Machine Check Interrupt to the Processor(s) if so enabled