Motorola MVME5100 Manual Do Utilizador

Página de 330
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 interrupts 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