Intel IXP2800 2.2 Intel XScale Core Microarchitecture

30 Hardware Reference Manual

Intel® IXP2800 Network Processor

Technical Description

2.2 Intel XScale® Core Microarchitecture

The Intel XScale® microarchitecture consists of a 32-bit general purpose RISC processor that

incorporates an extensive list of architecture features that allows it to achieve high performance.

2.2.1 ARM* Compatibility

The Intel XScale® microarchitecture is ARM* Version 5 (V5) Architecture compliant. It

implements the integer instruction set of ARM* V5, but does not provide hardware support of the

floating point instructions.

The Intel XScale® microarchitecture provides the Thumb instruction set (ARM V5T) and the

ARM V5E DSP extensions.

Backward compatibility with the first generation of StrongARM* products is maintained for user-

mode applications. Operating systems may require modifications to match the specific hardware

features of the Intel XScale® microarchitecture and to take advantage of the performance

enhancements added to the Intel XScale® core.

2.2.2 Features

2.2.2.1 Multiply/Accumulate (MAC)

The MAC unit supports early termination of multiplies/accumulates in two cycles and can sustain a

throughput of a MAC operation every cycle. Several architectural enhancements were made to the

MAC to support audio coding algorithms, which include a 40-bit accumulator and support for

16-bit packed values.

2.2.2.2 Memory Management

The Intel XScale® microarchitecture implements the Memory Management Unit (MMU)

Architecture specified in the ARM Architecture Reference Manual. The MMU provides access

protection and virtual to physical address translation.

The MMU Architecture also specifies the caching policies for the instruction cache and data

memory. These policies are specified as page attributes and include:

•identifying code as cacheable or non-cacheable

•selecting between the mini-data cache or data cache

•write-back or write-through data caching

•enabling data write allocation policy

•and enabling the write buffer to coalesce stores to external memory

2.2.2.3 Instruction Cache

The Intel XScale® microarchitecture implements a 32-Kbyte, 32-way set associative instruction

cache with a line size of 32 bytes. All requests that “miss” the instruction cache generate a 32-byte

read request to external memory. A mechanism to lock critical code within the cache is also

provided.

Contents

Main 2Hardware Reference Manual Revision History Contents Page Page Page Page Page Page Page Page Page Page Page Page Figures Page Page Tables Page Page Page Page Page Introduction 1 1.1 About This Document 1.2 Related Documentation 1.3 Terminology Technical Description 2 2.1 Overview Intel IXP2800 Network Processor Figure 1. IXP2800 Network Processor Functional Block Diagram Figure 2. IXP2800 Network Processor Detailed Diagram 2.2 Intel XScale Core Microarchitecture 2.2.1 ARM* Compatibility 2.2.2 Features 2.2.2.1 Multiply/Accumulate (MAC) 2.2.2.2 Memory Management 2.2.2.4 Branch Target Buffer 2.2.2.5 Data Cache 2.2.2.6 Interrupt Controller 2.2.2.7 Address Map 2.3 Microengines Figure 4. Microengine Block Diagram NNData_In (from previous ME) D_Push (from DRAM) S_Push (from SRAM Scratchpad, MSF, Hash, PCI, CAP) Control Store 2.3.1 Microengine Bus Arrangement 2.3.2 Control Store 2.3.3 Contexts Page 2.3.4 Datapath Registers 2.3.4.1 General-Purpose Registers (GPRs) 2.3.4.2 Transfer Registers 2.3.4.3 Next Neighbor Registers 2.3.4.4 Local Memory Page 2.3.5 Addressing Modes 2.3.5.1 Context-Relative Addressing Mode 2.3.5.2 Absolute Addressing Mode 2.3.5.3 Indexed Addressing Mode 2.3.6 Local CSRs 2.3.7 Execution Datapath 2.3.7.1 Byte Align Page 2.3.7.2 CAM Page Page 2.3.8 CRC Unit 2.3.9 Event Signals 2.4 DRAM 2.4.1 Size Configuration 2.4.2 Read and Write Access 2.5 SRAM 2.5.1 QDR Clocking Scheme 2.5.2 SRAM Controller Configurations 2.5.3 SRAM Atomic Operations 2.5.4 Queue Data Structure Commands 2.5.5 Reference Ordering Table 12. Address Reference Order Table 13. Q_array Entry Reference Order 2.5.5.2 Microengine Software Restrictions to Maintain Ordering 2.6 Scratchpad Memory 2.6.1 Scratchpad Atomic Operations 2.6.2 Ring Commands Page 2.7 Media and Switch Fabric Interface 2.7.1 SPI-4 2.7.2 CSIX 2.7.3 Receive 2.7.3.1 RBUF 2.7.3.2 Full Element List 2.7.3.3 RX_THREAD_FREELIST 2.7.3.4 Receive Operation Summary 2.7.4 Transmit 2.7.4.1 TBUF The definitions of the fields are shown in Table15. Table 15. TBUF SPI-4 Control Definition Hardware Reference Manual 67 The definitions of the fields are shown in Table16. 2.7.4.2 Transmit Operation Summary Table 16. TBUF CSIX Control Definition 2.7.5 The Flow Control Interface 2.7.5.1 SPI-4 2.7.5.2 CSIX 2.8 Hash Unit Figure 14. Hash Unit Block Diagram 2.9 PCI Controller 2.9.1 Target Access 2.9.2 Master Access 2.9.3 DMA Channels 2.9.3.1 DMA Descriptor 2.9.3.2 DMA Channel Operation 2.9.3.3 DMA Channel End Operation 2.9.3.4 Adding Descriptors to an Unterminated Chain 2.9.4 Mailbox and Message Registers 2.9.5 PCI Arbiter 2.10 Control and Status Register Access Proxy 2.11 Intel XScale Core Peripherals 2.11.1 Interrupt Controller 2.11.2 Timers 2.11.3 General Purpose I/O 2.11.4 Universal Asynchronous Receiver/Transmitter 2.11.5 Slowport 2.12 I/O Latency 2.13 Performance Monitor Intel XScale Core 3 3.1 Introduction 3.2 Features 3.2.1 Multiply/ACcumulate (MAC) 3.2.2 Memory Management 3.2.3 Instruction Cache 3.2.4 Branch Target Buffer (BTB) 3.2.5 Data Cache 3.2.6 Performance Monitoring 3.2.7 Power Management 3.3 Memory Management 3.3.1 Architecture Model Page 3.3.2 Exceptions 3.3.3 Interaction of the MMU, Instruction Cache, and Data Cache 3.3.4 Control 3.3.4.1 Invalidate (Flush) Operation 3.3.4.2 Enabling/Disabling 3.3.4.3 Locking Entries 3.3.4.4 Round-Robin Replacement Algorithm 3.4 Instruction Cache 3.4.1 Instruction Cache Operation 3.4.1.1 Operation when Instruction Cache is Enabled 3.4.1.2 Operation when Instruction Cache is Disabled 3.4.1.3 Fetch Policy 3.4.1.4 Round-Robin Replacement Algorithm 3.4.1.5 Parity Protection 3.4.1.6 Instruction Cache Coherency 3.4.2 Instruction Cache Control 3.4.2.1 Instruction Cache State at Reset 3.4.2.2 Enabling/Disabling 3.4.2.3 Invalidating the Instruction Cache 3.4.2.4 Locking Instructions in the Instruction Cache Page 3.4.2.5 Unlocking Instructions in the Instruction Cache 3.5 Branch Target Buffer (BTB) 3.5.1 Branch Target Buffer Operation 3.5.1.1 Reset SN WN WT ST 3.5.2 Update Policy 3.5.3 BTB Control 3.6 Data Cache 3.6.1 Overviews 3.6.1.1 Data Cache Overview 3.6.1.2 Mini-Data Cache Overview 3.6.1.3 Write Buffer and Fill Buffer Overview 3.6.2 Data Cache and Mini-Data Cache Operation 3.6.2.1 Operation when Caching is Enabled 3.6.2.2 Operation when Data Caching is Disabled 3.6.2.3 Cache Policies Page 3.6.2.4 Round-Robin Replacement Algorithm 3.6.2.5 Parity Protection 3.6.2.6 Atomic Accesses 3.6.3 Data Cache and Mini-Data Cache Control 3.6.3.1 Data Memory State After Reset 3.6.3.2 Enabling/Disabling 3.6.3.3 Invalidate and Clean Operations Page 3.6.4 Reconfiguring the Data Cache as Data RAM 3.6.5 Write Buffer/Fill Buffer Operation and Control 3.7 Configuration 3.8 Performance Monitoring 3.8.1 Performance Monitoring Events 3.8.1.1 Instruction Cache Efficiency Mode 3.8.1.2 Data Cache Efficiency Mode 3.8.1.3 Instruction Fetch Latency Mode 3.8.1.4 Data/Bus Request Buffer Full Mode 3.8.1.5 Stall/Writeback Statistics 3.8.1.6 Instruction TLB Efficiency Mode 3.9 Performance Considerations 3.9.1 Interrupt Latency 3.9.2 Branch Prediction 3.9.3 Addressing Modes 3.9.4 Instruction Latencies 3.9.4.1 Performance Terms Page 3.9.4.2 Branch Instruction Timings 3.9.4.3 Data Processing Instruction Timings Table 28. Branch Instruction Timings (Predicted by the BTB) Table 29. Branch Instruction Timings (Not Predicted by the BTB) Table 30. Data Processing Instruction Timings 3.9.4.4 Multiply Instruction Timings Table 31. Multiply Instruction Timings (Sheet 1 of 2) 3.9.4.5 Saturated Arithmetic Instructions Table 32. Multiply Implicit Accumulate Instruction Timings Table 33. Implicit Accumulator Access Instruction Timings Table 34. Saturated Data Processing Instruction Timings Table 31. Multiply Instruction Timings (Sheet 2 of 2) 3.9.4.6 Status Register Access Instructions 3.9.4.7 Load/Store Instructions 3.9.4.8 Semaphore Instructions Table 35. Status Register Access Instruction Timings Table 36. Load and Store Instruction Timings 3.10 Test Features 3.10.1 IXP2800 Network Processor Endianness 3.10.1.1 Read and Write Transactions Initiated by the Intel XScale Core Page Page 3.11 Intel XScale Core Gasket Unit 3.11.1 Overview Figure 26. Intel XScale Core-Initiated Write to the IXP2800 Network Processor (Continued) Word Write by Intel XScale Core Network Processor 3.11.2 Intel XScale Core Gasket Functional Description 3.11.2.1 Command Memory Bus to Command Push/Pull Conversion 3.11.3 CAM Operation 3.11.4 Atomic Operations 3.11.4.1 Summary of Rules for the Atomic Command Regarding I/O 3.11.4.2 Intel XScale Core Access to SRAM Q-Array 3.11.5 I/O Transaction 3.11.6 Hash Access 3.11.7 Gasket Local CSR 3.11.8 Interrupt Figure 29. Interrupt Mask Block Diagram 3.12 Intel XScale Core Peripheral Interface 3.12.1 XPI Overview 3.12.1.1 Data Transfers 3.12.1.2 Data Alignment 3.12.1.3 Address Spaces for XPI Internal Devices Table 5 3 shows the address space assignment for XPI devices. Table 52. Data Transaction Alignment Table 53. Address Spaces for XPI Internal Devices 3.12.2 UART Overview 3.12.3 UART Operation 3.12.3.1 UART FIFO OPERATION 3.12.4 Baud Rate Generator 3.12.5 General Purpose I/O (GPIO) 3.12.6 Timers 3.12.6.1 Timer Operation 3.12.7 Slowport Unit 3.12.7.1 PROM Device Support 3.12.7.2 Microprocessor Interface Support for the Framer 3.12.7.3 Slowport Unit Interfaces Figure 35 shows the Slowport unit interface diagram. Table 55. SONET/SDH Devices (Sheet 2 of 2) Figure 35. Slowport Unit Interface Diagram 3.12.7.4 Address Space 3.12.7.5 Slowport Interfacing Topology 3.12.7.6 Slowport 8-Bit Device Bus Protocols Figure 37. Slowport Example Application Topology Page Page Page 3.12.7.7 SONET/SDH Microprocessor Access Support Page Figure 42. An Interface Topology with Lucent* TDAT042G5 SONET/SDH Lucent TDAT042G5* Intel IXP2000 Page Page Intel IXP2800 PMC-Sierra* PM5351 Page Page Page Figure 49. Mode 3 Second Interface Topology with Intel / AMCC* SONET/SDH Device Page Page Page Figure 53. Second Interface Topology with Intel / AMCC* SONET/SDH Device Page Page Page Microengines 4 4.1 Overview Figure 56. Microengine Block Diagram NNData_In (from previous ME) D_Push (from DRAM) S_Push (from SRAM Scratchpad, MSF, Hash, PCI, CAP) Control Store 4.1.1 Control Store 4.1.2 Contexts Page 4.1.3 Datapath Registers 4.1.3.3 Next Neighbor Registers 4.1.3.4 Local Memory 4.1.4 Addressing Modes 4.1.4.1 Context-Relative Addressing Mode 4.1.4.2 Absolute Addressing Mode 4.1.4.3 Indexed Addressing Mode 4.2 Local CSRs 4.3 Execution Datapath 4.3.1 Byte Align Page 4.3.2 CAM Page 4.4 CRC Unit 4.5 Event Signals 4.5.1 Microengine Endianness 4.5.1.2 Write to TBUF 4.5.1.3 Read/Write from/to SRAM 4.5.1.4 Read/Write from/to DRAM 4.5.1.5 Read/Write from/to SHaC and Other CSRs 4.5.1.6 Write to Hash Unit 4.5.2 Media Access 4.5.2.1 Read from RBUF 4.5.2.2 Write to TBUF 4.5.2.3 TBUF to SPI-4 Transfer DRAM 5 5.1 Overview 5.2 Size Configuration 5.3 DRAM Clocking DRAM 5.4 Bank Policy Figure 67. IXP2800 Clocking for RDRAM at 400 MHz RMC RAC System Ref_Clk DRCG 5.5 Interleaving 5.5.1 Three Channels Active (3-Way Interleave) DRAM Table 63. Address Rearrangement for 3-Way Interleave (Sheet 1 of 2) 5.5.2 Two Channels Active (2-Way Interleave) 5.5.3 One Channel Active (No Interleave) 5.5.4 Interleaving Across RDRAMs and Banks 5.6 Parity and ECC 5.6.1 Parity and ECC Disabled 5.6.2 Parity Enabled 5.6.3 ECC Enabled 5.6.4 ECC Calculation and Syndrome 5.7 Timing Configuration 5.8 Microengine Signals 5.9 Serial Port 5.10 RDRAM Controller Block Diagram 5.10.1 Commands 5.10.2 DRAM Write 5.10.2.1 Masked Write 5.10.3 DRAM Read 5.10.4 CSR Write 5.10.5 CSR Read 5.10.6 Arbitration 5.11 DRAM Push/Pull Arbiter 5.11.1 Arbiter Push/Pull Operation 5.11.2 DRAM Push Arbiter Description 5.12 DRAM Pull Arbiter Description Page Page SRAM Interface 6 6.1 Overview 6.2 SRAM Interface Configurations 6.2.1 Internal Interface 6.2.2 Number of Channels 6.2.3 Coprocessor and/or SRAMs Attached to a Channel 6.3 SRAM Controller Configurations Page 6.4 Command Overview 6.4.1 Basic Read/Write Commands 6.4.2 Atomic Operations Page 6.4.3 Queue Data Structure Commands Page Page 6.4.3.1 Read_Q_Descriptor Commands 6.4.3.2 Write_Q_Descriptor Commands 6.4.3.3 ENQ and DEQ Commands 6.4.4 Ring Data Structure Commands 6.4.5 Journaling Commands 6.5 Parity 6.6 Address Map 6.7 Reference Ordering 6.7.1 Reference Order Tables SRAM Interface 6.7.2 Microcode Restrictions to Maintain Ordering Table 78. Q_array Entry Reference Order Example 31. Table Update Microcode 6.8 Coprocessor Mode Page Page Page SHaC Unit Expansion 7 7.1 Overview 7.1.1 SHaC Unit Block Diagram Figure 84. SHaC Top Level Diagram 7.1.2 Scratchpad 7.1.2.1 Scratchpad Description Figure 85. Scratchpad Block Diagram Scratchpad State Machine Scratchpad RAM 7.1.2.2 Scratchpad Interface 7.1.2.3 Scratchpad Block Level Diagram Page Writes Reads Signal Done Page 7.1.3 Hash Unit Hash Algorithm HASH_RESULT 7.1.3.1 Hashing Operation Page 7.1.3.2 Hash Algorithm Page Media and Switch Fabric Interface 8 8.1 Overview Page 8.1.1 SPI-4 Page Table84 shows the order of bytes on SPI-4; this example shows a 43-byte packet. Table 84. Order of Bytes1 within the SPI-4 Data Burst Figure 90. Receive and Transmit Clock Generation 8.1.2 CSIX 8.1.3 CSIX/SPI-4 Interleave Mode 8.2 Receive The receive section consists of: 8.2.1 Receive Pins 8.2.2 RBUF Page 8.2.2.1 SPI-4 Page The status contains the following information: The definitions of the fields are shown in Table90. Table 90. RBUF SPIF-4 Status Definition 8.2.2.2 CSIX The definitions of the fields are shown in Table91. Table 91. RBUF CSIX Status Definition 8.2.3 Full Element List 8.2.4 Rx_Thread_Freelist_# 8.2.5 Rx_Thread_Freelist_Timeout_# 8.2.6 Receive Operation Summary Page 8.2.7 Receive Flow Control Status 8.2.7.1 SPI-4 8.2.7.2 CSIX 8.2.8 Parity 8.2.8.1 SPI-4 8.2.8.2 CSIX 8.2.9 Error Cases 8.3 Transmit The transmit section consists of: Each of these is described below. Figure 94 is a simplified block diagram of the MSF transmit block. 8.3.1 Transmit Pins 8.3.2 TBUF Page Page 8.3.2.1 SPI-4 The definitions of the fields are shown in Table98. Table 98. TBUF SPI-4 Control Definition 8.3.2.2 CSIX The definitions of the fields are shown in Table99. Table 99. TBUF CSIX Control Definition 8.3.3 Transmit Operation Summary 8.3.3.1 SPI-4 8.3.3.2 CSIX 8.3.3.3 Transmit Summary 8.3.4 Transmit Flow Control Status 8.3.4.1 SPI-4 Page 8.3.4.2 CSIX 8.3.5 Parity 8.3.5.1 SPI-4 8.3.5.2 CSIX 8.4 RBUF and TBUF Summary 8.5 CSIX Flow Control Interface 8.5.1 TXCSRB and RXCSRB Signals 8.5.2 FCIFIFO and FCEFIFO Buffers 8.5.2.1 Full Duplex CSIX 8.5.2.2 Simplex CSIX Page 8.5.3 TXCDAT/RXCDAT, TXCSOF/RXCSOF, TXCPAR/RXCPAR, and TXCFC/RXCFC Signals 8.6 Deskew and Training Table 105. Calendar Deskew Functions Table 106. Flow Control Deskew Functions 8.6.1 Data Training Pattern 8.6.2 Flow Control Training Pattern 8.6.3 Use of Dynamic Training Table 110. IXP2800 Network Processor Requires Data Training Table 111. Switch Fabric or SPI-4 Framer Requires Data Training Page 8.7 CSIX Startup Sequence 8.7.1 CSIX Full Duplex 8.7.1.1 Ingress IXP2800 Network Processor 8.7.1.2 Egress IXP2800 Network Processor 8.7.1.3 Single IXP2800 Network Processor 8.7.2 CSIX Simplex 8.7.2.1 Ingress IXP2800 Network Processor 8.7.2.2 Egress IXP2800 Network Processor 8.7.2.3 Single IXP2800 Network Processor 8.8 Interface to Command and Push and Pull Buses Figure 100. MSF to Command and Push and Pull Buses Interface Block Diagram TBUF RBUF 8.8.1 RBUF or MSF CSR to Microengine S_TRANSFER_IN Register for Instruction: 8.8.2 Microengine S_TRANSFER_OUT Register to TBUF or MSF CSR for Instruction: 8.8.3 Microengine to MSF CSR for Instruction: 8.8.4 From RBUF to DRAM for Instruction: 8.8.5 From DRAM to TBUF for Instruction: 8.9 Receiver and Transmitter Interoperation with Framers and Switch Fabrics 8.9.1 Receiver and Transmitter Configurations Receiver Transmitter 8.9.1.2 Hybrid Simplex Configuration Processor Network Transmitter Fabric Receiver 8.9.1.3 Dual Network Processor Full Duplex Configuration 8.9.1.4 Single Network Processor Full Duplex Configuration (SPI-4.2) 8.9.1.5 Single Network Processor, Full Duplex Configuration (SPI-4.2 and CSIX-L1) 8.9.2 System Configurations 8.9.2.1 Framer, Single Network Processor Ingress and Egress, and Fabric Interface Chip Figure 107. Framer, Single Network Processor Ingress and Egress, and Fabric Interface Chip Figure 108. Framer, Dual Processor Ingress, Single Processor Egress, and Fabric Interface Chip 8.9.2.4 CPU Complex, Network Processor, and Fabric Interface Chip 8.9.2.5 Framer, Single Network Processor, Co-Processor, and Fabric Interface Chip 8.9.3 SPI-4.2 Support 8.9.3.1 SPI-4.2 Receiver 8.9.3.2 SPI-4.2 Transmitter 8.9.4 CSIX-L1 Protocol Support 8.9.4.1 CSIX-L1 Interface Reference Model: Traffic Manager and Fabric Interface Chip 8.9.4.2 Intel IXP2800 Support of the CSIX-L1 Protocol Page Page Page 8.9.4.3 CSIX-L1 Protocol Receiver Support 8.9.4.4 CSIX-L1 Protocol Transmitter Support 8.9.4.5 Implementation of a Bridge Chip to CSIX-L1 8.9.5 Dual Protocol (SPI and CSIX-L1) Support 8.9.5.1 Dual Protocol Receiver Support 8.9.5.2 Dual Protocol Transmitter Support 8.9.5.3 Implementation of a Bridge Chip to CSIX-L1 and SPI-4.2 8.9.6 Transmit State Machine 8.9.6.1 SPI-4.2 Transmitter State Machine 8.9.6.2 Training Transmitter State Machine 8.9.6.3 CSIX-L1 Transmitter State Machine 8.9.7 Dynamic De-Skew 8.9.8 Summary of Receiver and Transmitter Signals Intel IXP2800 Network Processor Page PCI Unit 9 9.1 Overview Figure 118. PCI Functional Blocks 64-bit PCI Bus Command Bus Master Command Bus Slave Master Interface 9.2 PCI Pin Protocol Interface Block 9.2.1 PCI Commands 9.2.2 IXP2800 Network Processor Initialization 9.2.2.1 Initialization by the Intel XScale Core 9.2.2.2 Initialization by a PCI Host 9.2.3 PCI Type 0 Configuration Cycles 9.2.3.1 Configuration Write 9.2.3.2 Configuration Read 9.2.4 PCI 64-Bit Bus Extension 9.2.5 PCI Target Cycles 9.2.5.5 Target Read Accesses from the PCI Bus 9.2.6 PCI Initiator Transactions 9.2.6.1 PCI Request Operation Page 9.2.6.6 Special Cycle 9.2.7 PCI Fast Back-to-Back Cycles 9.2.8 PCI Retry 9.2.9 PCI Disconnect 9.2.11 PCI Central Functions 9.2.11.1 PCI Interrupt Inputs 9.2.11.2 PCI Reset Output 9.2.11.3 PCI Internal Arbiter 9.3 Slave Interface Block 9.3.1 CSR Interface 9.3.2 SRAM Interface 9.3.2.1 SRAM Slave Writes 9.3.2.2 SRAM Slave Reads 9.3.3 DRAM Interface 9.3.3.1 DRAM Slave Writes 9.3.3.2 DRAM Slave Reads 9.3.4 Mailbox and Doorbell Registers Page Page 9.3.5 PCI Interrupt Pin An external PCI interrupt can be generated in the following way: Figure 126 shows how PCI interrupts are managed via the PCI and the Intel XScale core. Table125 describes how IRQ are generated for each silicon stepping. 9.4 Master Interface Block 9.4.1 DMA Interface 9.4.1.1 Allocation of the DMA Channels 9.4.1.2 Special Registers for Microengine Channels 9.4.1.3 DMA Descriptor 9.4.1.4 DMA Channel Operation 9.4.1.5 DMA Channel End Operation 9.4.1.6 Adding Descriptor to an Unterminated Chain 9.4.1.7 DRAM to PCI Transfer 9.4.1.8 PCI to DRAM Transfer 9.4.2 Push/Pull Command Bus Target Interface 9.4.2.1 Command Bus Master Access to Local Configuration Registers 9.4.2.2 Command Bus Master Access to Local Control and Status Registers 9.4.2.3 Command Bus Master Direct Access to PCI Bus Page 9.5 PCI Unit Error Behavior 9.5.1 PCI Target Error Behavior 9.5.2 As a PCI Initiator During a DMA Transfer 9.5.2.2 DMA Read from SRAM (Descriptor Read) Gets a Memory Error 9.5.2.3 DMA from DRAM Transfer (Write to PCI) Receives PCI_PERR_L on PCI Bus 9.5.2.4 DMA To DRAM (Read from PCI) Has Bad Data Parity 9.5.2.5 DMA Transfer Experiences a Master Abort (Time-Out) on PCI 9.5.2.6 DMA Transfer Receives a Target Abort Response During a Data Phase 9.5.3 As a PCI Initiator During a Direct Access from the Intel XScale Core or Microengine 9.5.3.1 Master Transfer Experiences a Master Abort (Time-Out) on PCI 9.5.3.2 Master Transfer Receives a Target Abort Response During a Data Phase 9.5.3.4 Master Read from PCI (Read from PCI) Has Bad Data Parity 9.5.3.5 Master Transfer Receives PCI_SERR_L from the PCI Bus 9.6 PCI Data Byte Lane Alignment PCI Data SRAM Data Table 131. Byte Lane Alignment for 64-Bit PCI Data In (64 Bits PCI Big-Endian to Big-Endian PCI Data SRAM Data PCI Add[2]=1 PCI Add[2]=0 PCI Data SRAM Data Table 133. Byte Lane Alignment for 32-Bit PCI Data In (32 Bits PCI Big-Endian to Big-Endian Table 134. Byte Lane Alignment for 64-Bit PCI Data Out (Big-Endian to 64 Bits PCI Little Table 135. Byte Lane Alignment for 64-Bit PCI Data Out (Big-Endian to 64 Bits PCI Big-Endian Table 136. Byte Lane Alignment for 32-Bit PCI Data Out (Big-Endian to 32 Bits PCI Little Table 137. Byte Lane Alignment for 32-Bit PCI Data Out (Big-Endian to 32 Bits PCI Big-Endian 9.6.1 Endian for Byte Enable Table 141. Byte Enable Alignment for 32-Bit PCI Data In (32 Bits PCI Big-Endian to Big-Endian Table 142. Byte Enable Alignment for 64-Bit PCI Data Out (Big-Endian to 64 Bits PCI Little PCI Add[2]=1 PCI Add[2]=0 SRAM Data PCI Side Page Table 146. PCI I/O Cycles with Data Swap Enable Stepping Description Clocks and Reset 10 10.1 Clocks Figure 130. Overall Clock Generation and Distribution Clock Unit with PLL Intel IXP2800 Network Processor Table 147. Clock Usage Summary (Sheet 1 of 2) Page Table 148. Clock Rates Examples 10.2 Synchronization Between Frequency Domains 10.3 Reset 10.3.1 Hardware Reset Using nRESET or PCI_RST_L Page 10.3.2 PCI-Initiated Reset 10.3.3 Watchdog Timer-Initiated Reset 10.3.4 Software-Initiated Reset 10.3.5 Reset Removal Operation Based on CFG_PROM_BOOT 10.3.5.1 When CFG_PROM_BOOT is 1 (BOOT_PROM is Present) 10.3.5.2 When CFG_PROM_BOOT is 0 (BOOT_PROM is Not Present) 10.3.6 Strap Pins Table 149. IXP2800 Network Processor Strap Pins 10.3.7 Powerup Reset Sequence 10.4 Boot Mode Figure 135. Boot Process 10.4.1 Flash ROM 10.4.2 PCI Host Download 10.5 Initialization Page Performance Monitor Unit 11 11.1 Introduction 11.1.1 Motivation for Performance Monitors 11.1.2 Motivation for Choosing CHAP Counters Performance Monitoring Unit 11.1.3 Functional Overview of CHAP Counters 11.1.4 Basic Operation of the Performance Monitor Unit 11.1.5 Definition of CHAP Terminology Figure 138. Basic Block Diagram of IXP2800 Network Processor with PMU 11.1.6 Definition of Clock Domains 11.2 Interface and CSR Description 11.2. 1 APB P eriph eral 11.2.2 CAP Description 11.2.2.1 Selecting the Access Mode 11.2.2.2 PMU CSR 11.2.2.3 CAP Writes 11.3 Performance Measurements Table 152. Hardware Blocks and Their Performance Measurement Events (Sheet 1 of 2) Table 152. Hardware Blocks and Their Performance Measurement Events (Sheet 2 of 2) 11.4 Events Monitored in Hardware 11.4.1 Queue Statistics Events 11.4.1.1 Queue Latency 11.4.1.2 Queue Utilization 11.4.2 Count Events 11.4.3 Design Block Select Definitions 11.4.4 Null Event Table 153. PMU Design Unit Selection (Sheet 2 of 2) 11.4.5 Threshold Events 11.4.6 External Input Events 11.4.6.1 XPI Events Target ID(000001) / Design Block #(0100) Table 155. XPI PMU Event List (Sheet 1 of 4) Table 155. XPI PMU Event List (Sheet 2 of 4) Table 155. XPI PMU Event List (Sheet 3 of 4) Table 155. XPI PMU Event List (Sheet 4 of 4) 11.4.6.2 SHaC Events Target ID(000010) / Design Block #(0101) Table 156. SHaC PMU Event List (Sheet 1 of 4) Table 156. SHaC PMU Event List (Sheet 2 of 4) Table 156. SHaC PMU Event List (Sheet 3 of 4) 11.4.6.3 IXP2800 Network Processor MSF Events Target ID(000011) / Design Block #(0110) Table 156. SHaC PMU Event List (Sheet 4 of 4) Table 157. IXP2800 Network Processor MSF PMU Event List (Sheet 1 of 6) Table 157. IXP2800 Network Processor MSF PMU Event List (Sheet 2 of 6) Table 157. IXP2800 Network Processor MSF PMU Event List (Sheet 3 of 6) Table 157. IXP2800 Network Processor MSF PMU Event List (Sheet 4 of 6) Table 157. IXP2800 Network Processor MSF PMU Event List (Sheet 5 of 6) Table 157. IXP2800 Network Processor MSF PMU Event List (Sheet 6 of 6) 11.4.6.4 Intel XScale Core Events Target ID(000100) / Design Block #(0111) Table 158. Intel XScale Core Gasket PMU Event List (Sheet 1 of 4) Table 158. Intel XScale Core Gasket PMU Event List (Sheet 2 of 4) Table 158. Intel XScale Core Gasket PMU Event List (Sheet 3 of 4) 11.4.6.5 PCI Events Target ID(000101) / Design Block #(1000) Table 158. Intel XScale Core Gasket PMU Event List (Sheet 4 of 4) Table 159. PCI PMU Event List (Sheet 1 of 5) Table 159. PCI PMU Event List (Sheet 2 of 5) Table 159. PCI PMU Event List (Sheet 3 of 5) Table 159. PCI PMU Event List (Sheet 4 of 5) 11.4.6.6 ME00 Events Target ID(100000) / Design Block #(1001) Table 159. PCI PMU Event List (Sheet 5 of 5) Table 160. ME00 PMU Event List (Sheet 1 of 2) 11.4.6.7 ME01 Events Target ID(100001) / Design Block #(1001) Table 160. ME00 PMU Event List (Sheet 2 of 2) Table 161. ME01 PMU Event List 11.4.6.8 ME02 Events Target ID(100010) / Design Block #(1001) 11.4.6.9 ME03 Events Target ID(100011) / Design Block #(1001) Table 162. ME02 PMU Event List Table 163. ME03 PMU Event List 11.4.6.10 ME04 Events Target ID(100100) / Design Block #(1001) 11.4.6.11 ME05 Events Target ID(100101) / Design Block #(1001) Table 164. ME04 PMU Event List Table 165. ME05 PMU Event List 11.4.6.12 ME06 Events Target ID(100110) / Design Block #(1001) 11.4.6.13 ME07 Events Target ID(100111) / Design Block #(1001) Table 166. ME06 PMU Event List Table 167. ME07 PMU Event List 11.4.6.14 ME10 Events Target ID(110000) / Design Block #(1010) 11.4.6.15 ME11 Events Target ID(110001) / Design Block #(1010) Table 168. ME10 PMU Event List Table 169. ME11 PMU Event List 11.4.6.16 ME12 Events Target ID(110010) / Design Block #(1010) 11.4.6.17 ME13 Events Target ID(110011) / Design Block #(1010) Table 170. ME12 PMU Event List Table 171. ME13 PMU Event List 11.4.6.18 ME14 Events Target ID(110100) / Design Block #(1010) 11.4.6.19 ME15 Events Target ID(110101) / Design Block #(1010) Table 172. ME14 PMU Event List Table 173. ME15 PMU Event List 11.4.6.20 ME16 Events Target ID(100110) / Design Block #(1010) 11.4.6.21 ME17 Events Target ID(110111) / Design Block #(1010) Table 174. ME16 PMU Event List Table 175. ME17 PMU Event List 11.4.6.22 SRAM DP1 Events Target ID(001001) / Design Block #(0010) 11.4.6.23 SRAM DP0 Events Target ID(001010) / Design Block #(0010) Table 176. SRAM DP1 PMU Event List Table 177. SRAM DP0 PMU Event List (Sheet 1 of 3) Table 177. SRAM DP0 PMU Event List (Sheet 2 of 3) 11.4.6.24 SRAM CH3 Events Target ID(001011) / Design Block #(0010) Table 177. SRAM DP0 PMU Event List (Sheet 3 of 3) Table 178. SRAM CH3 PMU Event List 11.4.6.25 SRAM CH2 Events Target ID(001100) / Design Block #(0010) 11.4.6.26 SRAM CH1 Events Target ID(001101) / Design Block #(0010) Table 179. SRAM CH3 PMU Event List Table 180. SRAM CH3 PMU Event List 11.4.6.27 SRAM CH0 Events Target ID(001110) / Design Block #(0010) Table 181. SRAM CH0 PMU Event List (Sheet 1 of 2) 11.4.6.28 DRAM DPLA Events Target ID(010010) / Design Block #(0011) Table 181. SRAM CH0 PMU Event List (Sheet 2 of 2) Table 182. IXP2800 Network Processor Dram DPLA PMU Event List (Sheet 1 of 2) 11.4.6.29 DRAM DPSA Events Target ID(010011) / Design Block #(0011) Table 182. IXP2800 Network Processor Dram DPLA PMU Event List (Sheet 2 of 2) Table 183. IXP2800 Network Processor Dram DPSA PMU Event List (Sheet 1 of 2) 11.4.6.30 IXP2800 Network Processor DRAM CH2 Events Target ID(010100) / Table 183. IXP2800 Network Processor Dram DPSA PMU Event List (Sheet 2 of 2) Table 184. IXP2800 Network Processor Dram CH2 PMU Event List (Sheet 1 of 5) Table 184. IXP2800 Network Processor Dram CH2 PMU Event List (Sheet 2 of 5) Table 184. IXP2800 Network Processor Dram CH2 PMU Event List (Sheet 3 of 5) Table 184. IXP2800 Network Processor Dram CH2 PMU Event List (Sheet 4 of 5) 11.4.6.31 IXP2800 Network Processor DRAM CH1 Events Target ID(010101) / 11.4.6.32 IXP2800 Network Processor DRAM CH0 Events Target ID(010110) / Table 184. IXP2800 Network Processor Dram CH2 PMU Event List (Sheet 5 of 5) Table 185. IXP2800 Network Processor Dram CH1 PMU Event List Table 186. IXP2800 Network Processor Dram CH0 PMU Event List