Programmer’s Guide

• 1. Introduction
  • 1.1. Documentation Roadmap
  • 1.2. Related Publications
• 2. Overview
  • 2.1. Development Environment
  • 2.2. Environment Abstraction Layer
  • 2.3. Core Components
    • 2.3.1. Ring Manager (librte_ring)
    • 2.3.2. Memory Pool Manager (librte_mempool)
    • 2.3.3. Network Packet Buffer Management (librte_mbuf)
    • 2.3.4. Timer Manager (librte_timer)
  • 2.4. Ethernet* Poll Mode Driver Architecture
  • 2.5. Packet Forwarding Algorithm Support
  • 2.6. librte_net
• 3. Environment Abstraction Layer
  • 3.1. EAL in a Linux-userland Execution Environment
    • 3.1.1. Initialization and Core Launching
    • 3.1.2. Shutdown and Cleanup
    • 3.1.3. Multi-process Support
    • 3.1.4. Memory Mapping Discovery and Memory Reservation
    • 3.1.5. Support for Externally Allocated Memory
    • 3.1.6. Per-lcore and Shared Variables
    • 3.1.7. Logs
    • 3.1.8. CPU Feature Identification
    • 3.1.9. User Space Interrupt Event
    • 3.1.10. Blacklisting
    • 3.1.11. Misc Functions
    • 3.1.12. IOVA Mode Configuration
  • 3.2. Memory Segments and Memory Zones (memzone)
  • 3.3. Multiple pthread
    • 3.3.1. EAL pthread and lcore Affinity
    • 3.3.2. non-EAL pthread support
    • 3.3.3. Public Thread API
    • 3.3.4. Control Thread API
    • 3.3.5. Known Issues
    • 3.3.6. cgroup control
  • 3.4. Malloc
    • 3.4.1. Cookies
    • 3.4.2. Alignment and NUMA Constraints
    • 3.4.3. Use Cases
    • 3.4.4. Internal Implementation
• 4. Service Cores
  • 4.1. Service Core Initialization
  • 4.2. Enabling Services on Cores
  • 4.3. Service Core Statistics
• 5. Ring Library
  • 5.1. References for Ring Implementation in FreeBSD*
  • 5.2. Lockless Ring Buffer in Linux*
  • 5.3. Additional Features
    • 5.3.1. Name
  • 5.4. Use Cases
  • 5.5. Anatomy of a Ring Buffer
    • 5.5.1. Single Producer Enqueue
    • 5.5.2. Single Consumer Dequeue
    • 5.5.3. Multiple Producers Enqueue
    • 5.5.4. Modulo 32-bit Indexes
  • 5.6. References
• 6. Mempool Library
  • 6.1. Cookies
  • 6.2. Stats
  • 6.3. Memory Alignment Constraints
  • 6.4. Local Cache
  • 6.5. Mempool Handlers
  • 6.6. Use Cases
• 7. Mbuf Library
  • 7.1. Design of Packet Buffers
  • 7.2. Buffers Stored in Memory Pools
  • 7.3. Constructors
  • 7.4. Allocating and Freeing mbufs
  • 7.5. Manipulating mbufs
  • 7.6. Meta Information
  • 7.7. Direct and Indirect Buffers
  • 7.8. Debug
  • 7.9. Use Cases
• 8. Poll Mode Driver
  • 8.1. Requirements and Assumptions
  • 8.2. Design Principles
  • 8.3. Logical Cores, Memory and NIC Queues Relationships
  • 8.4. Device Identification, Ownership and Configuration
    • 8.4.1. Device Identification
    • 8.4.2. Port Ownership
    • 8.4.3. Device Configuration
    • 8.4.4. On-the-Fly Configuration
    • 8.4.5. Configuration of Transmit Queues
    • 8.4.6. Free Tx mbuf on Demand
    • 8.4.7. Hardware Offload
  • 8.5. Poll Mode Driver API
    • 8.5.1. Generalities
    • 8.5.2. Generic Packet Representation
    • 8.5.3. Ethernet Device API
    • 8.5.4. Ethernet Device Standard Device Arguments
    • 8.5.5. Extended Statistics API
    • 8.5.6. NIC Reset API
• 9. Generic flow API (rte_flow)
  • 9.1. Overview
  • 9.2. Flow rule
    • 9.2.1. Description
    • 9.2.2. Attributes
    • 9.2.3. Pattern item
    • 9.2.4. Matching pattern
    • 9.2.5. Meta item types
    • 9.2.6. Data matching item types
    • 9.2.7. Actions
    • 9.2.8. Action types
    • 9.2.9. Negative types
    • 9.2.10. Planned types
  • 9.3. Rules management
    • 9.3.1. Validation
    • 9.3.2. Creation
    • 9.3.3. Destruction
    • 9.3.4. Flush
    • 9.3.5. Query
  • 9.4. Isolated mode
  • 9.5. Verbose error reporting
  • 9.6. Helpers
    • 9.6.1. Error initializer
    • 9.6.2. Object conversion
  • 9.7. Caveats
  • 9.8. PMD interface
  • 9.9. Device compatibility
    • 9.9.1. Global bit-masks
    • 9.9.2. Unsupported layer types
    • 9.9.3. ANY pattern item
    • 9.9.4. Unsupported actions
    • 9.9.5. Flow rules priority
  • 9.10. Future evolutions
• 10. Switch Representation within DPDK Applications
  • 10.1. Introduction
  • 10.2. Port Representors
  • 10.3. Basic SR-IOV
  • 10.4. Controlled SR-IOV
    • 10.4.1. Initialization
    • 10.4.2. VF Representors
    • 10.4.3. Traffic Steering
  • 10.5. Flow API (rte_flow)
    • 10.5.1. Extensions
    • 10.5.2. Traffic Direction
    • 10.5.3. Transferring Traffic
    • 10.5.4. Pattern Items And Actions
    • 10.5.5. Actions Order and Repetition
  • 10.6. Switching Examples
    • 10.6.1. Associating VF 1 with Physical Port 0
    • 10.6.2. Sharing Broadcasts
    • 10.6.3. Encapsulating VF 2 Traffic in VXLAN
• 11. Traffic Metering and Policing API
  • 11.1. Overview
  • 11.2. Configuration steps
  • 11.3. Run-time processing
• 12. Traffic Management API
  • 12.1. Overview
  • 12.2. Capability API
  • 12.3. Scheduling Algorithms
  • 12.4. Traffic Shaping
  • 12.5. Congestion Management
  • 12.6. Packet Marking
  • 12.7. Steps to Setup the Hierarchy
    • 12.7.1. Initial Hierarchy Specification
    • 12.7.2. Hierarchy Commit
    • 12.7.3. Run-Time Hierarchy Updates
• 13. Wireless Baseband Device Library
  • 13.1. Design Principles
  • 13.2. Device Management
    • 13.2.1. Device Creation
    • 13.2.2. Device Identification
    • 13.2.3. Device Configuration
    • 13.2.4. Queues Configuration
    • 13.2.5. Device & Queues Management
    • 13.2.6. Logical Cores, Memory and Queues Relationships
  • 13.3. Device Operation Capabilities
    • 13.3.1. Capabilities Discovery
  • 13.4. Operation Processing
    • 13.4.1. Enqueue / Dequeue Burst APIs
    • 13.4.2. Operation Representation
    • 13.4.3. Operation Management and Allocation
    • 13.4.4. BBDEV Inbound/Outbound Memory
    • 13.4.5. BBDEV Turbo Encode Operation
    • 13.4.6. BBDEV Turbo Decode Operation
  • 13.5. Sample code
    • 13.5.1. BBDEV Device API
• 14. Cryptography Device Library
  • 14.1. Design Principles
  • 14.2. Device Management
    • 14.2.1. Device Creation
    • 14.2.2. Device Identification
    • 14.2.3. Device Configuration
    • 14.2.4. Configuration of Queue Pairs
    • 14.2.5. Logical Cores, Memory and Queues Pair Relationships
  • 14.3. Device Features and Capabilities
    • 14.3.1. Device Features
    • 14.3.2. Device Operation Capabilities
    • 14.3.3. Capabilities Discovery
  • 14.4. Operation Processing
    • 14.4.1. Private data
    • 14.4.2. Enqueue / Dequeue Burst APIs
    • 14.4.3. Operation Representation
    • 14.4.4. Operation Management and Allocation
  • 14.5. Symmetric Cryptography Support
    • 14.5.1. Session and Session Management
    • 14.5.2. Transforms and Transform Chaining
    • 14.5.3. Symmetric Operations
  • 14.6. Sample code
  • 14.7. Asymmetric Cryptography
    • 14.7.1. Session and Session Management
    • 14.7.2. Asymmetric Sessionless Support
    • 14.7.3. Transforms and Transform Chaining
    • 14.7.4. Asymmetric Operations
  • 14.8. Asymmetric crypto Sample code
    • 14.8.1. Asymmetric Crypto Device API
• 15. Compression Device Library
  • 15.1. Device Management
    • 15.1.1. Device Creation
    • 15.1.2. Device Identification
    • 15.1.3. Device Configuration
    • 15.1.4. Configuration of Queue Pairs
    • 15.1.5. Logical Cores, Memory and Queues Pair Relationships
  • 15.2. Device Features and Capabilities
    • 15.2.1. Capabilities
    • 15.2.2. Capabilities Discovery
  • 15.3. Compression Operation
    • 15.3.1. Operation Representation
    • 15.3.2. Operation Management and Allocation
    • 15.3.3. Passing source data as mbuf-chain
    • 15.3.4. Operation Status
    • 15.3.5. Operation status after enqueue / dequeue
    • 15.3.6. Produced, Consumed And Operation Status
  • 15.4. Transforms
  • 15.5. Compression API Hash support
  • 15.6. Compression API Stateless operation
    • 15.6.1. priv_xform in Stateless operation
    • 15.6.2. Stateless and OUT_OF_SPACE
    • 15.6.3. Hash in Stateless
    • 15.6.4. Checksum in Stateless
  • 15.7. Compression API Stateful operation
    • 15.7.1. Stream in Stateful operation
    • 15.7.2. Stateful and OUT_OF_SPACE
    • 15.7.3. Hash in Stateful
    • 15.7.4. Checksum in Stateful
  • 15.8. Burst in compression API
    • 15.8.1. Enqueue / Dequeue Burst APIs
  • 15.9. Sample code
    • 15.9.1. Compression Device API
• 16. Security Library
  • 16.1. Design Principles
    • 16.1.1. Inline Crypto
    • 16.1.2. Inline protocol offload
    • 16.1.3. Lookaside protocol offload
    • 16.1.4. PDCP Flow Diagram
  • 16.2. Device Features and Capabilities
    • 16.2.1. Device Capabilities For Security Operations
    • 16.2.2. Capabilities Discovery
    • 16.2.3. Security Session Create/Free
    • 16.2.4. Security session configuration
    • 16.2.5. Security API
    • 16.2.6. Flow based Security Session
• 17. Rawdevice Library
  • 17.1. Introduction
  • 17.2. Design
    • 17.2.1. Device Identification
• 18. Link Bonding Poll Mode Driver Library
  • 18.1. Link Bonding Modes Overview
  • 18.2. Implementation Details
    • 18.2.1. Link Status Change Interrupts / Polling
    • 18.2.2. Requirements / Limitations
    • 18.2.3. Configuration
  • 18.3. Using Link Bonding Devices
    • 18.3.1. Using the Poll Mode Driver from an Application
    • 18.3.2. Using Link Bonding Devices from the EAL Command Line
• 19. Timer Library
  • 19.1. Implementation Details
  • 19.2. Use Cases
  • 19.3. References
• 20. Hash Library
  • 20.1. Hash API Overview
  • 20.2. Multi-process support
  • 20.3. Multi-thread support
  • 20.4. Extendable Bucket Functionality support
  • 20.5. Implementation Details (non Extendable Bucket Case)
  • 20.6. Implementation Details (with Extendable Bucket)
  • 20.7. Entry distribution in hash table
  • 20.8. Use Case: Flow Classification
  • 20.9. References
• 21. Elastic Flow Distributor Library
  • 21.1. Introduction
  • 21.2. Flow Based Distribution
    • 21.2.1. Computation Based Schemes
    • 21.2.2. Flow-Table Based Schemes
    • 21.2.3. EFD Based Scheme
  • 21.3. Example of EFD Library Usage
  • 21.4. Library API Overview
    • 21.4.1. EFD Table Create
    • 21.4.2. EFD Insert and Update
    • 21.4.3. EFD Lookup
    • 21.4.4. EFD Delete
  • 21.5. Library Internals
    • 21.5.1. Insert Function Internals
    • 21.5.2. Lookup Function Internals
    • 21.5.3. Group Rebalancing Function Internals
  • 21.6. References
• 22. Membership Library
  • 22.1. Introduction
  • 22.2. Vector of Bloom Filters
  • 22.3. Hash-Table based Set-Summaries
    • 22.3.1. Set-Summaries with False Negative Probability
  • 22.4. Library API Overview
    • 22.4.1. Set-summary Create
    • 22.4.2. Set-summary Element Insertion
    • 22.4.3. Set-summary Element Lookup
    • 22.4.4. Set-summary Element Delete
  • 22.5. References
• 23. LPM Library
  • 23.1. LPM API Overview
  • 23.2. Implementation Details
    • 23.2.1. Addition
    • 23.2.2. Lookup
    • 23.2.3. Limitations in the Number of Rules
    • 23.2.4. Use Case: IPv4 Forwarding
    • 23.2.5. References
• 24. LPM6 Library
  • 24.1. LPM6 API Overview
    • 24.1.1. Implementation Details
    • 24.1.2. Addition
    • 24.1.3. Lookup
    • 24.1.4. Limitations in the Number of Rules
  • 24.2. Use Case: IPv6 Forwarding
• 25. Flow Classification Library
  • 25.1. Overview
    • 25.1.1. Classifier creation
    • 25.1.2. Adding a table to the Classifier
    • 25.1.3. Flow Parsing
    • 25.1.4. Adding Flow Rules
    • 25.1.5. Deleting Flow Rules
    • 25.1.6. Packet Matching
• 26. Packet Distributor Library
  • 26.1. Distributor Core Operation
  • 26.2. Worker Operation
• 27. Reorder Library
  • 27.1. Operation
  • 27.2. Implementation Details
  • 27.3. Use Case: Packet Distributor
• 28. IP Fragmentation and Reassembly Library
  • 28.1. Packet fragmentation
  • 28.2. Packet reassembly
    • 28.2.1. IP Fragment Table
    • 28.2.2. Packet Reassembly
    • 28.2.3. Debug logging and Statistics Collection
• 29. Generic Receive Offload Library
  • 29.1. Overview
  • 29.2. Two Sets of API
    • 29.2.1. Lightweight Mode API
    • 29.2.2. Heavyweight Mode API
  • 29.3. Reassembly Algorithm
    • 29.3.1. Challenges
    • 29.3.2. Key-based Reassembly Algorithm
  • 29.4. TCP/IPv4 GRO
  • 29.5. VxLAN GRO
  • 29.6. GRO Library Limitations
• 30. Generic Segmentation Offload Library
  • 30.1. Overview
  • 30.2. Limitations
  • 30.3. Packet Segmentation
    • 30.3.1. GSO Output Segment Format
  • 30.4. Supported GSO Packet Types
    • 30.4.1. TCP/IPv4 GSO
    • 30.4.2. UDP/IPv4 GSO
    • 30.4.3. VxLAN GSO
    • 30.4.4. GRE GSO
  • 30.5. How to Segment a Packet
• 31. The librte_pdump Library
  • 31.1. Operation
  • 31.2. Implementation Details
  • 31.3. Use Case: Packet Capturing
• 32. Multi-process Support
  • 32.1. Memory Sharing
  • 32.2. Deployment Models
    • 32.2.1. Symmetric/Peer Processes
    • 32.2.2. Asymmetric/Non-Peer Processes
    • 32.2.3. Running Multiple Independent DPDK Applications
    • 32.2.4. Running Multiple Independent Groups of DPDK Applications
  • 32.3. Multi-process Limitations
  • 32.4. Communication between multiple processes
    • 32.4.1. Registering for incoming messages
    • 32.4.2. Sending messages
    • 32.4.3. Sending requests
    • 32.4.4. Receiving and responding to messages
    • 32.4.5. Misc considerations
• 33. Kernel NIC Interface
  • 33.1. The DPDK KNI Kernel Module
    • 33.1.1. Loopback Mode
    • 33.1.2. Kernel Thread Mode
    • 33.1.3. Default Carrier State
  • 33.2. KNI Creation and Deletion
  • 33.3. DPDK mbuf Flow
  • 33.4. Use Case: Ingress
  • 33.5. Use Case: Egress
  • 33.6. Ethtool
• 34. Thread Safety of DPDK Functions
  • 34.1. Fast-Path APIs
  • 34.2. Performance Insensitive API
  • 34.3. Library Initialization
  • 34.4. Interrupt Thread
• 35. Event Device Library
  • 35.1. Event struct
    • 35.1.1. Event Metadata
    • 35.1.2. Event Payload
    • 35.1.3. Queues
    • 35.1.4. Ports
  • 35.2. API Walk-through
    • 35.2.1. Init and Config
    • 35.2.2. Setting up Queues
    • 35.2.3. Setting up Ports
    • 35.2.4. Linking Queues and Ports
    • 35.2.5. Starting the EventDev
    • 35.2.6. Ingress of New Events
    • 35.2.7. Forwarding of Events
    • 35.2.8. Egress of Events
  • 35.3. Summary
• 36. Event Ethernet Rx Adapter Library
  • 36.1. API Walk-through
    • 36.1.1. Creating an Adapter Instance
    • 36.1.2. Adding Rx Queues to the Adapter Instance
    • 36.1.3. Querying Adapter Capabilities
    • 36.1.4. Configuring the Service Function
    • 36.1.5. Starting the Adapter Instance
    • 36.1.6. Getting Adapter Statistics
    • 36.1.7. Interrupt Based Rx Queues
    • 36.1.8. Rx Callback for SW Rx Adapter
• 37. Event Ethernet Tx Adapter Library
  • 37.1. API Walk-through
    • 37.1.1. Creating an Adapter Instance
    • 37.1.2. Adding Tx Queues to the Adapter Instance
    • 37.1.3. Querying Adapter Capabilities
    • 37.1.4. Linking a Queue to the Adapter’s Event Port
    • 37.1.5. Configuring the Service Function
    • 37.1.6. Starting the Adapter Instance
    • 37.1.7. Enqueuing Packets to the Adapter
    • 37.1.8. Getting Adapter Statistics
• 38. Event Timer Adapter Library
  • 38.1. Event Timer struct
    • 38.1.1. Timer Expiry Event
    • 38.1.2. Timeout Ticks
    • 38.1.3. State
    • 38.1.4. User Metadata
  • 38.2. API Overview
    • 38.2.1. Create and Configure an Adapter Instance
    • 38.2.2. Retrieve Event Timer Adapter Contextual Information
    • 38.2.3. Configuring the Service Component
    • 38.2.4. Starting the Adapter Instance
    • 38.2.5. Arming Event Timers
    • 38.2.6. Canceling Event Timers
  • 38.3. Processing Timer Expiry Events
  • 38.4. Summary
• 39. Event Crypto Adapter Library
  • 39.1. Adapter Mode
    • 39.1.1. RTE_EVENT_CRYPTO_ADAPTER_OP_NEW mode
    • 39.1.2. RTE_EVENT_CRYPTO_ADAPTER_OP_FORWARD mode
  • 39.2. API Overview
    • 39.2.1. Create an adapter instance
    • 39.2.2. Querying adapter capabilities
    • 39.2.3. Adding queue pair to the adapter instance
    • 39.2.4. Configure the service function
    • 39.2.5. Set event request/response information
    • 39.2.6. Start the adapter instance
    • 39.2.7. Get adapter statistics
• 40. Quality of Service (QoS) Framework
  • 40.1. Packet Pipeline with QoS Support
  • 40.2. Hierarchical Scheduler
    • 40.2.1. Overview
    • 40.2.2. Scheduling Hierarchy
    • 40.2.3. Application Programming Interface (API)
    • 40.2.4. Implementation
    • 40.2.5. Worst Case Scenarios for Performance
  • 40.3. Dropper
    • 40.3.1. Configuration
    • 40.3.2. Enqueue Operation
    • 40.3.3. Queue Empty Operation
    • 40.3.4. Source Files Location
    • 40.3.5. Integration with the DPDK QoS Scheduler
    • 40.3.6. Integration with the DPDK QoS Scheduler Sample Application
    • 40.3.7. Application Programming Interface (API)
  • 40.4. Traffic Metering
    • 40.4.1. Functional Overview
    • 40.4.2. Implementation Overview
• 41. Power Management
  • 41.1. CPU Frequency Scaling
  • 41.2. Core-load Throttling through C-States
  • 41.3. Per-core Turbo Boost
  • 41.4. Use of Power Library in a Hyper-Threaded Environment
  • 41.5. API Overview of the Power Library
  • 41.6. User Cases
  • 41.7. Empty Poll API
    • 41.7.1. Abstract
    • 41.7.2. Proposed Solution
    • 41.7.3. API Overview for Empty Poll Power Management
  • 41.8. User Cases
  • 41.9. References
• 42. Packet Classification and Access Control
  • 42.1. Overview
    • 42.1.1. Rule definition
    • 42.1.2. RT memory size limit
    • 42.1.3. Classification methods
  • 42.2. Application Programming Interface (API) Usage
    • 42.2.1. Classify with Multiple Categories
• 43. Packet Framework
  • 43.1. Design Objectives
  • 43.2. Overview
  • 43.3. Port Library Design
    • 43.3.1. Port Types
    • 43.3.2. Port Interface
  • 43.4. Table Library Design
    • 43.4.1. Table Types
    • 43.4.2. Table Interface
    • 43.4.3. Hash Table Design
  • 43.5. Pipeline Library Design
    • 43.5.1. Connectivity of Ports and Tables
    • 43.5.2. Port Actions
    • 43.5.3. Table Actions
  • 43.6. Multicore Scaling
    • 43.6.1. Shared Data Structures
  • 43.7. Interfacing with Accelerators
• 44. Vhost Library
  • 44.1. Vhost API Overview
  • 44.2. Vhost-user Implementations
  • 44.3. Guest memory requirement
  • 44.4. Vhost supported vSwitch reference
  • 44.5. Vhost data path acceleration (vDPA)
• 45. Metrics Library
  • 45.1. Initializing the library
  • 45.2. Registering metrics
  • 45.3. Updating metric values
  • 45.4. Querying metrics
  • 45.5. Bit-rate statistics library
    • 45.5.1. Initialization
    • 45.5.2. Controlling the sampling rate
  • 45.6. Latency statistics library
    • 45.6.1. Initialization
    • 45.6.2. Triggering statistic updates
    • 45.6.3. Library shutdown
    • 45.6.4. Timestamp and latency calculation
• 46. Berkeley Packet Filter Library
  • 46.1. Not currently supported eBPF features
• 47. Source Organization
  • 47.1. Makefiles and Config
  • 47.2. Libraries
  • 47.3. Drivers
  • 47.4. Applications
• 48. Development Kit Build System
  • 48.1. Building the Development Kit Binary
    • 48.1.1. Build Directory Concept
  • 48.2. Building External Applications
  • 48.3. Makefile Description
    • 48.3.1. General Rules For DPDK Makefiles
    • 48.3.2. Makefile Types
    • 48.3.3. Internally Generated Build Tools
    • 48.3.4. Useful Variables Provided by the Build System
    • 48.3.5. Variables that Can be Set/Overridden in a Makefile Only
    • 48.3.6. Variables that can be Set/Overridden by the User on the Command Line Only
    • 48.3.7. Variables that Can be Set/Overridden by the User in a Makefile or Command Line
• 49. Development Kit Root Makefile Help
  • 49.1. Configuration Targets
  • 49.2. Build Targets
  • 49.3. Install Targets
  • 49.4. Test Targets
  • 49.5. Documentation Targets
  • 49.6. Misc Targets
  • 49.7. Other Useful Command-line Variables
  • 49.8. Make in a Build Directory
  • 49.9. Compiling for Debug
• 50. Extending the DPDK
  • 50.1. Example: Adding a New Library libfoo
    • 50.1.1. Example: Using libfoo in the Test Application
• 51. Building Your Own Application
  • 51.1. Compiling a Sample Application in the Development Kit Directory
  • 51.2. Build Your Own Application Outside the Development Kit
  • 51.3. Customizing Makefiles
    • 51.3.1. Application Makefile
    • 51.3.2. Library Makefile
    • 51.3.3. Customize Makefile Actions
• 52. External Application/Library Makefile help
  • 52.1. Prerequisites
  • 52.2. Build Targets
  • 52.3. Help Targets
  • 52.4. Other Useful Command-line Variables
  • 52.5. Make from Another Directory
• 53. Performance Optimization Guidelines
  • 53.1. Introduction
• 54. Writing Efficient Code
  • 54.1. Memory
    • 54.1.1. Memory Copy: Do not Use libc in the Data Plane
    • 54.1.2. Memory Allocation
    • 54.1.3. Concurrent Access to the Same Memory Area
    • 54.1.4. NUMA
    • 54.1.5. Distribution Across Memory Channels
    • 54.1.6. Locking memory pages
  • 54.2. Communication Between lcores
  • 54.3. PMD Driver
    • 54.3.1. Lower Packet Latency
  • 54.4. Locks and Atomic Operations
  • 54.5. Coding Considerations
    • 54.5.1. Inline Functions
    • 54.5.2. Branch Prediction
  • 54.6. Setting the Target CPU Type
• 55. Profile Your Application
  • 55.1. Profiling on x86
    • 55.1.1. Profiling with VTune
  • 55.2. Profiling on ARM64
    • 55.2.1. Using Linux perf
    • 55.2.2. High-resolution cycle counter
• 56. Glossary