MPC8544COMEDEV Freescale Semiconductor, MPC8544COMEDEV Datasheet - Page 314

• Manufacturer Part Number: MPC8544COMEDEV
• Description: KIT DEV EXPRESS MPC8544COM
• Manufacturer: Freescale Semiconductor
• Type: MPUr

Specifications of MPC8544COMEDEV

• Contents: Board
• For Use With/related Products: MPC8544
• Lead Free Status / RoHS Status: Lead free / RoHS Compliant

e500 Coherency Module
8.1.1
The ECM routes transactions initiated by the e500 core to the appropriate target interface on the device.
In a manner analogous to a bridging router in a local area network, the ECM forwards I/O-initiated
transactions that are tagged with the global attribute onto the core complex bus (CCB). This allows on-chip
caches to snoop these transactions as if they were locally initiated and to take actions to maintain
coherency across cacheable memory.
8.1.2
The ECM includes these distinctive features:
8-2
•
•
•
•
•
•
•
•
•
Support for the e500 core and an L2/SRAM on the CCB, including a CCB arbiter.
It sources a 64-bit data bus for returning read data from the ECM to the e500 core and routing write
data from the ECM to the L2/SRAM. It sinks a 128-bit data bus for receiving data from the
L2/SRAM and a 128-bit write data bus from the e500 core.
Four connection points for I/O initiating (mastering into the device) interfaces. The ECM supports
five connection points for I/O targets. The DDR memory controller, local bus, OCeaN targets, and
configuration register access block all have a target port connection to the ECM.
Split transaction support—separate address and data tenures allow for pipelining of transactions
and out-of-order data tenures between initiators and targets.
Proper ordering of I/O-initiated transactions.
Speculative read bus for low-latency dispatch of reads to the DDR controller.
Low-latency path for returning read data from DDR to the e500 core.
Error registers trap transactions with invalid addresses. Errors can be programmed to generate
interrupts to the e500 core, as described in the following sections:
—
—
—
—
—
Errors from reading I/O devices (for example a master-aborted read transaction on the PCI
interface) terminate with data sent to the master with a corrupt attribute. If the master is the e500
core, the ECM asserts core_fault_in to the core, which causes the core to generate a machine check
interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and one of these errors
occurs, appropriate interrupts must be enabled to ensure that an interrupt is generated. See
Section 6.10.2, “Hardware Implementation-Dependent Register 1
Section 8.2.1.5, “ECM Error Detect Register (EEDR)”
Section 8.2.1.6, “ECM Error Enable Register (EEER)”
Section 8.2.1.7, “ECM Error Attributes Capture Register (EEATR)”
Section 8.2.1.8, “ECM Error Low Address Capture Register (EELADR)”
Section 8.2.1.9, “ECM Error High Address Capture Register (EEHADR)”
Overview
Features
MPC8544E PowerQUICC III Integrated Host Processor Family Reference Manual, Rev. 1
(HID1).”
Freescale Semiconductor