MPC8533EVTAQGA Freescale Semiconductor, MPC8533EVTAQGA Datasheet - Page 243

• Manufacturer Part Number: MPC8533EVTAQGA
• Description: MPU POWERQUICC 783-PBGA
• Manufacturer: Freescale Semiconductor

Specifications of MPC8533EVTAQGA

• Processor Type: MPC85xx PowerQUICC III 32-Bit
• Speed: 1.0GHz
• Voltage: 0.95 V ~ 1.05 V
• Mounting Type: Surface Mount
• Package / Case: 783-BBGA, FCBGA
• Lead Free Status / RoHS Status: Lead free / RoHS Compliant
• Features: -

Freescale Semiconductor
40–45
47–49
Bits
46
50
ASTME
Name
RFXE
—
—
MPC8533E PowerQUICC III Integrated Host Processor Family Reference Manual, Rev. 1
Reserved, should be cleared.
Read fault exception enable. Enables the core to internally generate a machine check interrupt when
core_fault_in is asserted. Depending on the value of MSR[ME], this results in either a machine check
interrupt or a checkstop.
0 Assertion of core_fault_in cannot cause a machine check. The core does not execute any instructions
Note: The e500v1 does not stall when faulty instructions or data are received. Instead, it continues
processing with faulty instructions or data. The only reliable way to prevent such behavior is to set RFXE,
which causes a machine check before the faulty instructions or data are used. To avoid the use of faulty
instructions or data and to have good error determination, software must set RFXE and program the PIC
to interrupt the processor when errors occur. As a result, software must deal with multiple interrupts for
the same fundamental problem.
1 Assertion of core_fault_in causes a machine check if MSR[ME] = 1 or a checkstop if MSR[ME] = 0.
The RFXE bit provides flexibility in error recovery. Typically, devices outside the core have some way
other than the assertion of core_fault_in to signal the core that an error occurred. Usually, this is done
by channeling interrupt requests through a programmable interrupt controller (PIC) to the core. In these
cases, the assertion of core_fault_in is used only to prevent the core from using bad data before
receiving an interrupt from the PIC (for example, an external or critical input interrupt). Possible
combinations of RFXE and PIC configuration are as follows:
Reserved, should be cleared.
Address bus streaming mode enable. This bit, along with the ECM stream control bits in the EEBACR,
enables address bus streaming on the CCB. See
Register (EEBACR).”
0 Address bus streaming mode disabled
1 Address bus streaming mode enabled
• RFXE = 0 and the PIC is configured to interrupt the processor. In this configuration, the assertion of
• RFXE = 1 and the PIC is not configured to interrupt the processor. This configuration provides quick
• RFXE = 1 and the PIC is configured to interrupt the processor. In this configuration, the core may
• RFXE = 0 and the PIC is not configured to interrupt the processor. This is not a recommended
from a faulty instruction fetch and does not execute any load instructions that get their data from a
faulty data fetch.
On the e500v2, if these instructions are eventually required by the sequential programming model
(that is, they are not in a speculative execution path), the e500v2 stalls until an asynchronous interrupt
is taken. The e500v1 does not stall when faulty instructions or data are received, as described in the
following note.
The core_fault_in signal is asserted to the core when logic outside of the core has a problem
delivering good data to the core. For example, the front-side L2 cache asserts core_fault_in when an
ECC error occurs and ECC is enabled. As a second example, it is asserted when there is a master
abort on a PCI transaction. See “Proper Reporting of Bus Faults” in the core complex bus chapter of
the PowerPC™ e500 Core Family Reference Manual .
core_fault_in does not trigger a machine check interrupt. The core does not use the faulty instructions
or data and may stall. The PIC interrupts the core so that error recovery can begin. This configuration
allows the core to query the PIC and the rest of the system for more information about the cause of
the interrupt, and generally provides the best error recovery capabilities.
error detection without the overhead of configuring the PIC. When the PIC is not configured, setting
RFXE avoids stalling the core when core_fault_in is asserted. Determination of the root cause of the
problem may be somewhat more difficult than it would be if the PIC were enabled.
receive two interrupts for the same fundamental error. The two interrupts may occur in any order,
which may complicate error handling. Therefore, this is usually not an interesting configuration for a
single-core device. This may, however, be an interesting configuration for multi-core devices in which
the PIC may steer interrupts to a processor other than the one that attempted to fetch the faulty data.
configuration. The processor may stall indefinitely due to an unreported error.
Table 6-19. HID1 Field Descriptions (continued)
Description
Section 8.2.1.1, “ECM CCB Address Configuration
Core Register Summary
6-27