master_doc = 'index'
+# Figures, tables and code-blocks automatically numbered if they have caption
+numfig = True
+
latex_documents = [
('index',
'doc.tex',
**Figures**
-:ref:`Figure 1. Virtualization for a Single Port NIC in SR-IOV Mode <nic_figure_1>`
+:numref:`figure_single_port_nic` :ref:`figure_single_port_nic`
-:ref:`Figure 2. SR-IOV Performance Benchmark Setup <nic_figure_2>`
+:numref:`figure_perf_benchmark` :ref:`figure_perf_benchmark`
-:ref:`Figure 3. Fast Host-based Packet Processing <nic_figure_3>`
+:numref:`figure_fast_pkt_proc` :ref:`figure_fast_pkt_proc`
-:ref:`Figure 4. SR-IOV Inter-VM Communication <nic_figure_4>`
+:numref:`figure_inter_vm_comms` :ref:`figure_inter_vm_comms`
-:ref:`Figure 5. Virtio Host2VM Communication Example Using KNI vhost Back End <nic_figure_5>`
+:numref:`figure_host_vm_comms` :ref:`figure_host_vm_comms`
-:ref:`Figure 6. Virtio Host2VM Communication Example Using Qemu vhost Back End <nic_figure_6>`
+:numref:`figure_host_vm_comms_qemu` :ref:`figure_host_vm_comms_qemu`
+
+:numref:`figure_vmxnet3_int` :ref:`figure_vmxnet3_int`
+
+:numref:`figure_vswitch_vm` :ref:`figure_vswitch_vm`
+
+:numref:`figure_vm_vm_comms` :ref:`figure_vm_vm_comms`
The DPDK uses the SR-IOV feature for hardware-based I/O sharing in IOV mode.
Therefore, it is possible to partition SR-IOV capability on Ethernet controller NIC resources logically and
expose them to a virtual machine as a separate PCI function called a "Virtual Function".
-Refer to Figure 10.
+Refer to :numref:`figure_single_port_nic`.
-Therefore, a NIC is logically distributed among multiple virtual machines (as shown in Figure 10),
+Therefore, a NIC is logically distributed among multiple virtual machines (as shown in :numref:`figure_single_port_nic`),
while still having global data in common to share with the Physical Function and other Virtual Functions.
The DPDK fm10kvf, i40evf, igbvf or ixgbevf as a Poll Mode Driver (PMD) serves for the Intel® 82576 Gigabit Ethernet Controller,
Intel® Ethernet Controller I350 family, Intel® 82599 10 Gigabit Ethernet Controller NIC,
* `Scalable I/O Virtualized Servers <http://www.intel.com/content/www/us/en/virtualization/server-virtualization/scalable-i-o-virtualized-servers-paper.html>`_
-.. _nic_figure_1:
+.. _figure_single_port_nic:
-**Figure 1. Virtualization for a Single Port NIC in SR-IOV Mode**
+.. figure:: img/single_port_nic.*
+
+ Virtualization for a Single Port NIC in SR-IOV Mode
-.. image:: img/single_port_nic.*
Physical and Virtual Function Infrastructure
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
can also be used to bind and unbind devices to a virtual machine in Ubuntu.
If this option is used, step 6 in the instructions provided will be different.
- * The Virtual Machine Monitor (see Figure 11) is equivalent to a Host OS with KVM installed as described in the instructions.
+ * The Virtual Machine Monitor (see :numref:`figure_perf_benchmark`) is equivalent to a Host OS with KVM installed as described in the instructions.
+
+.. _figure_perf_benchmark:
-.. _nic_figure_2:
+.. figure:: img/perf_benchmark.*
-**Figure 2. Performance Benchmark Setup**
+ Performance Benchmark Setup
-.. image:: img/perf_benchmark.*
DPDK SR-IOV PMD PF/VF Driver Usage Model
----------------------------------------
With such host instance fast packet processing, lots of services such as filtering, QoS,
DPI can be offloaded on the host fast path.
-Figure 12 shows the scenario where some VMs directly communicate externally via a VFs,
+:numref:`figure_fast_pkt_proc` shows the scenario where some VMs directly communicate externally via a VFs,
while others connect to a virtual switch and share the same uplink bandwidth.
-.. _nic_figure_3:
+.. _figure_fast_pkt_proc:
+
+.. figure:: img/fast_pkt_proc.*
-**Figure 3. Fast Host-based Packet Processing**
+ Fast Host-based Packet Processing
-.. image:: img/fast_pkt_proc.*
SR-IOV (PF/VF) Approach for Inter-VM Communication
--------------------------------------------------
However, when VF crosses physical ports (VM0<->VM2), there is no such hardware bridge.
In this case, the DPDK PMD PF driver provides host forwarding between such VMs.
-Figure 13 shows an example.
+:numref:`figure_inter_vm_comms` shows an example.
In this case an update of the MAC address lookup tables in both the NIC and host DPDK application is required.
In the NIC, writing the destination of a MAC address belongs to another cross device VM to the PF specific pool.
The SR-IOV NIC switch forwards the packet to a specific VM according to the MAC destination address
which belongs to the destination VF on the VM.
-.. _nic_figure_4:
+.. _figure_inter_vm_comms:
-**Figure 4. Inter-VM Communication**
+.. figure:: img/inter_vm_comms.*
-.. image:: img/inter_vm_comms.*
+ Inter-VM Communication
This section demonstrates kni vhost back end example setup for Phy-VM Communication.
-.. _nic_figure_5:
+.. _figure_host_vm_comms:
-**Figure 5. Host2VM Communication Example Using kni vhost Back End**
+.. figure:: img/host_vm_comms.*
+
+ Host2VM Communication Example Using kni vhost Back End
-.. image:: img/host_vm_comms.*
Host2VM communication example
We use testpmd as the forwarding application in this example.
- .. image:: img/console.*
+ .. figure:: img/console.*
+
+ Running testpmd
#. Use IXIA packet generator to inject a packet stream into the KNI physical port.
Virtio with qemu virtio Back End
--------------------------------
-.. _nic_figure_6:
+.. _figure_host_vm_comms_qemu:
+
+.. figure:: img/host_vm_comms_qemu.*
-**Figure 6. Host2VM Communication Example Using qemu vhost Back End**
+ Host2VM Communication Example Using qemu vhost Back End
-.. image:: img/host_vm_comms_qemu.*
.. code-block:: console
* Before starting a VM, a VMXNET3 interface to a VM through VMware vSphere Client must be assigned.
This is shown in the figure below.
-.. image:: img/vmxnet3_int.*
+.. _figure_vmxnet3_int:
+
+.. figure:: img/vmxnet3_int.*
+
+ Assigning a VMXNET3 interface to a VM using VMware vSphere Client
.. note::
This section describes an example setup for Phy-vSwitch-VM-Phy communication.
-.. image:: img/vswitch_vm.*
+.. _figure_vswitch_vm:
+
+.. figure:: img/vswitch_vm.*
+
+ VMXNET3 with a Native NIC Connected to a vSwitch
.. note::
The following figure shows an example VM-to-VM communication over a Phy-VM-vSwitch-VM-Phy communication channel.
-.. image:: img/vm_vm_comms.*
+.. _figure_vm_vm_comms:
+
+.. figure:: img/vm_vm_comms.*
+
+ VMXNET3 Chaining VMs Connected to a vSwitch
.. note::
Then, the main() function is called. The core initialization and launch is done in rte_eal_init() (see the API documentation).
It consist of calls to the pthread library (more specifically, pthread_self(), pthread_create(), and pthread_setaffinity_np()).
-.. _pg_figure_2:
+.. _figure_linuxapp_launch:
-**Figure 2. EAL Initialization in a Linux Application Environment**
+.. figure:: img/linuxapp_launch.*
-.. image3_png has been replaced
+ EAL Initialization in a Linux Application Environment
-|linuxapp_launch|
.. note::
echo 50000 > pkt_io/cpu.cfs_quota_us
-.. |linuxapp_launch| image:: img/linuxapp_launch.*
**Figures**
-:ref:`Figure 1. Core Components Architecture <pg_figure_1>`
+:numref:`figure_architecture-overview` :ref:`figure_architecture-overview`
-:ref:`Figure 2. EAL Initialization in a Linux Application Environment <pg_figure_2>`
+:numref:`figure_linuxapp_launch` :ref:`figure_linuxapp_launch`
-:ref:`Figure 3. Example of a malloc heap and malloc elements within the malloc library <pg_figure_3>`
+:numref:`figure_malloc_heap` :ref:`figure_malloc_heap`
-:ref:`Figure 4. Ring Structure <pg_figure_4>`
+:numref:`figure_ring1` :ref:`figure_ring1`
-:ref:`Figure 5. Two Channels and Quad-ranked DIMM Example <pg_figure_5>`
+:numref:`figure_ring-enqueue1` :ref:`figure_ring-enqueue1`
-:ref:`Figure 6. Three Channels and Two Dual-ranked DIMM Example <pg_figure_6>`
+:numref:`figure_ring-enqueue2` :ref:`figure_ring-enqueue2`
-:ref:`Figure 7. A mempool in Memory with its Associated Ring <pg_figure_7>`
+:numref:`figure_ring-enqueue3` :ref:`figure_ring-enqueue3`
-:ref:`Figure 8. An mbuf with One Segment <pg_figure_8>`
+:numref:`figure_ring-dequeue1` :ref:`figure_ring-dequeue1`
-:ref:`Figure 9. An mbuf with Three Segments <pg_figure_9>`
+:numref:`figure_ring-dequeue2` :ref:`figure_ring-dequeue2`
-:ref:`Figure 16. Memory Sharing in the Intel® DPDK Multi-process Sample Application <pg_figure_16>`
+:numref:`figure_ring-dequeue3` :ref:`figure_ring-dequeue3`
-:ref:`Figure 17. Components of an Intel® DPDK KNI Application <pg_figure_17>`
+:numref:`figure_ring-mp-enqueue1` :ref:`figure_ring-mp-enqueue1`
-:ref:`Figure 18. Packet Flow via mbufs in the Intel DPDK® KNI <pg_figure_18>`
+:numref:`figure_ring-mp-enqueue2` :ref:`figure_ring-mp-enqueue2`
-:ref:`Figure 19. vHost-net Architecture Overview <pg_figure_19>`
+:numref:`figure_ring-mp-enqueue3` :ref:`figure_ring-mp-enqueue3`
-:ref:`Figure 20. KNI Traffic Flow <pg_figure_20>`
+:numref:`figure_ring-mp-enqueue4` :ref:`figure_ring-mp-enqueue4`
-:ref:`Figure 21. Complex Packet Processing Pipeline with QoS Support <pg_figure_21>`
+:numref:`figure_ring-mp-enqueue5` :ref:`figure_ring-mp-enqueue5`
-:ref:`Figure 22. Hierarchical Scheduler Block Internal Diagram <pg_figure_22>`
+:numref:`figure_ring-modulo1` :ref:`figure_ring-modulo1`
-:ref:`Figure 23. Scheduling Hierarchy per Port <pg_figure_23>`
+:numref:`figure_ring-modulo2` :ref:`figure_ring-modulo2`
-:ref:`Figure 24. Internal Data Structures per Port <pg_figure_24>`
+:numref:`figure_memory-management` :ref:`figure_memory-management`
-:ref:`Figure 25. Prefetch Pipeline for the Hierarchical Scheduler Enqueue Operation <pg_figure_25>`
+:numref:`figure_memory-management2` :ref:`figure_memory-management2`
-:ref:`Figure 26. Pipe Prefetch State Machine for the Hierarchical Scheduler Dequeue Operation <pg_figure_26>`
+:numref:`figure_mempool` :ref:`figure_mempool`
-:ref:`Figure 27. High-level Block Diagram of the Intel® DPDK Dropper <pg_figure_27>`
+:numref:`figure_mbuf1` :ref:`figure_mbuf1`
-:ref:`Figure 28. Flow Through the Dropper <pg_figure_28>`
+:numref:`figure_mbuf2` :ref:`figure_mbuf2`
-:ref:`Figure 29. Example Data Flow Through Dropper <pg_figure_29>`
+:numref:`figure_multi_process_memory` :ref:`figure_multi_process_memory`
-:ref:`Figure 30. Packet Drop Probability for a Given RED Configuration <pg_figure_30>`
+:numref:`figure_kernel_nic_intf` :ref:`figure_kernel_nic_intf`
-:ref:`Figure 31. Initial Drop Probability (pb), Actual Drop probability (pa) Computed Using a Factor 1 (Blue Curve) and a Factor 2 (Red Curve) <pg_figure_31>`
+:numref:`figure_pkt_flow_kni` :ref:`figure_pkt_flow_kni`
-:ref:`Figure 32. Example of packet processing pipeline. The input ports 0 and 1 are connected with the output ports 0, 1 and 2 through tables 0 and 1. <pg_figure_32>`
+:numref:`figure_vhost_net_arch2` :ref:`figure_vhost_net_arch2`
-:ref:`Figure 33. Sequence of steps for hash table operations in packet processing context <pg_figure_33>`
+:numref:`figure_kni_traffic_flow` :ref:`figure_kni_traffic_flow`
-:ref:`Figure 34. Data structures for configurable key size hash tables <pg_figure_34>`
-:ref:`Figure 35. Bucket search pipeline for key lookup operation (configurable key size hash tables) <pg_figure_35>`
+:numref:`figure_pkt_proc_pipeline_qos` :ref:`figure_pkt_proc_pipeline_qos`
-:ref:`Figure 36. Pseudo-code for match, match_many and match_pos <pg_figure_36>`
+:numref:`figure_hier_sched_blk` :ref:`figure_hier_sched_blk`
-:ref:`Figure 37. Data structures for 8-byte key hash tables <pg_figure_37>`
+:numref:`figure_sched_hier_per_port` :ref:`figure_sched_hier_per_port`
-:ref:`Figure 38. Data structures for 16-byte key hash tables <pg_figure_38>`
+:numref:`figure_data_struct_per_port` :ref:`figure_data_struct_per_port`
+
+:numref:`figure_prefetch_pipeline` :ref:`figure_prefetch_pipeline`
+
+:numref:`figure_pipe_prefetch_sm` :ref:`figure_pipe_prefetch_sm`
+
+:numref:`figure_blk_diag_dropper` :ref:`figure_blk_diag_dropper`
+
+:numref:`figure_flow_tru_droppper` :ref:`figure_flow_tru_droppper`
+
+:numref:`figure_ex_data_flow_tru_dropper` :ref:`figure_ex_data_flow_tru_dropper`
+
+:numref:`figure_pkt_drop_probability` :ref:`figure_pkt_drop_probability`
+
+:numref:`figure_drop_probability_graph` :ref:`figure_drop_probability_graph`
+
+:numref:`figure_figure32` :ref:`figure_figure32`
+
+:numref:`figure_figure33` :ref:`figure_figure33`
+
+:numref:`figure_figure34` :ref:`figure_figure34`
+
+:numref:`figure_figure35` :ref:`figure_figure35`
+
+:numref:`figure_figure37` :ref:`figure_figure37`
+
+:numref:`figure_figure38` :ref:`figure_figure38`
+
+:numref:`figure_figure39` :ref:`figure_figure39`
-:ref:`Figure 39. Bucket search pipeline for key lookup operation (single key size hash tables) <pg_figure_39>`
**Tables**
A typical DPDK IVSHMEM use case looks like the following.
-.. image28_png has been renamed
-|ivshmem|
+.. figure:: img/ivshmem.*
+
+ Typical Ivshmem use case
+
The same could work with several virtual machines, providing host-to-VM or VM-to-VM communication.
The maximum number of metadata files is 32 (by default) and each metadata file can contain different (or even the same) hugepages.
As a result, if the user wishes to shut down or restart the IVSHMEM host application,
it is not enough to simply shut the application down.
The virtual machine must also be shut down (if not, it will hold onto outdated host data).
-
-.. |ivshmem| image:: img/ivshmem.*
* Allows an interface with the kernel network stack.
-The components of an application using the DPDK Kernel NIC Interface are shown in Figure 17.
+The components of an application using the DPDK Kernel NIC Interface are shown in :numref:`figure_kernel_nic_intf`.
-.. _pg_figure_17:
+.. _figure_kernel_nic_intf:
-**Figure 17. Components of a DPDK KNI Application**
+.. figure:: img/kernel_nic_intf.*
-.. image43_png has been renamed
+ Components of a DPDK KNI Application
-|kernel_nic_intf|
The DPDK KNI Kernel Module
--------------------------
The kernel module will be aware of mbufs,
but all mbuf allocation and free operations will be handled by the DPDK application only.
-Figure 18 shows a typical scenario with packets sent in both directions.
+:numref:`figure_pkt_flow_kni` shows a typical scenario with packets sent in both directions.
-.. _pg_figure_18:
+.. _figure_pkt_flow_kni:
-**Figure 18. Packet Flow via mbufs in the DPDK KNI**
+.. figure:: img/pkt_flow_kni.*
-.. image44_png has been renamed
+ Packet Flow via mbufs in the DPDK KNI
-|pkt_flow_kni|
Use Case: Ingress
-----------------
There is a little penalty that comes from the non-polling mode of vhost.
However, it scales throughput well when using KNI in multi-thread mode.
-.. _pg_figure_19:
+.. _figure_vhost_net_arch2:
-**Figure 19. vHost-net Architecture Overview**
+.. figure:: img/vhost_net_arch.*
-.. image45_png has been renamed
+ vHost-net Architecture Overview
-|vhost_net_arch|
Packet Flow
~~~~~~~~~~~
happens in the context of vhost kthread.
Every vhost-net device is exposed to a front end virtio device in the guest.
-.. _pg_figure_20:
+.. _figure_kni_traffic_flow:
-**Figure 20. KNI Traffic Flow**
+.. figure:: img/kni_traffic_flow.*
-.. image46_png has been renamed
+ KNI Traffic Flow
-|kni_traffic_flow|
Sample Usage
~~~~~~~~~~~~
Even if the option is turned on, kni-vhost will ignore the information that the header contains.
When working with legacy virtio on the guest, it is better to turn off unsupported offload features using ethtool -K.
Otherwise, there may be problems such as an incorrect L4 checksum error.
-
-.. |kni_traffic_flow| image:: img/kni_traffic_flow.*
-
-.. |vhost_net_arch| image:: img/vhost_net_arch.*
-
-.. |pkt_flow_kni| image:: img/pkt_flow_kni.*
-
-.. |kernel_nic_intf| image:: img/kernel_nic_intf.*
DPDK also includes a pure-software library that
allows physical PMD's to be bonded together to create a single logical PMD.
-|bond-overview|
+.. figure:: img/bond-overview.*
+
+ Bonded PMDs
+
The Link Bonding PMD library(librte_pmd_bond) supports bonding of groups of
``rte_eth_dev`` ports of the same speed and duplex to provide
* **Round-Robin (Mode 0):**
-|bond-mode-0|
+.. figure:: img/bond-mode-0.*
+
+ Round-Robin (Mode 0)
+
This mode provides load balancing and fault tolerance by transmission of
packets in sequential order from the first available slave device through
* **Active Backup (Mode 1):**
-|bond-mode-1|
+.. figure:: img/bond-mode-1.*
+
+ Active Backup (Mode 1)
+
In this mode only one slave in the bond is active at any time, a different
slave becomes active if, and only if, the primary active slave fails,
* **Balance XOR (Mode 2):**
-|bond-mode-2|
+.. figure:: img/bond-mode-2.*
+
+ Balance XOR (Mode 2)
+
This mode provides transmit load balancing (based on the selected
transmission policy) and fault tolerance. The default policy (layer2) uses
* **Broadcast (Mode 3):**
-|bond-mode-3|
+.. figure:: img/bond-mode-3.*
+
+ Broadcast (Mode 3)
+
This mode provides fault tolerance by transmission of packets on all slave
ports.
* **Link Aggregation 802.3AD (Mode 4):**
-|bond-mode-4|
+.. figure:: img/bond-mode-4.*
+
+ Link Aggregation 802.3AD (Mode 4)
+
This mode provides dynamic link aggregation according to the 802.3ad
specification. It negotiates and monitors aggregation groups that share the
* **Transmit Load Balancing (Mode 5):**
-|bond-mode-5|
+.. figure:: img/bond-mode-5.*
+
+ Transmit Load Balancing (Mode 5)
+
This mode provides an adaptive transmit load balancing. It dynamically
changes the transmitting slave, according to the computed load. Statistics
.. code-block:: console
$RTE_TARGET/app/testpmd -c '0xf' -n 4 --vdev 'eth_bond0,mode=2, slave=0000:00a:00.01,slave=0000:004:00.00,xmit_policy=l34' -- --port-topology=chained
-
-.. |bond-overview| image:: img/bond-overview.*
-.. |bond-mode-0| image:: img/bond-mode-0.*
-.. |bond-mode-1| image:: img/bond-mode-1.*
-.. |bond-mode-2| image:: img/bond-mode-2.*
-.. |bond-mode-3| image:: img/bond-mode-3.*
-.. |bond-mode-4| image:: img/bond-mode-4.*
-.. |bond-mode-5| image:: img/bond-mode-5.*
By splitting the process in different tables/levels and limiting the number of tbl8s,
we can greatly reduce memory consumption while maintaining a very good lookup speed (one memory access per level).
-.. image40_png has been renamed
-|tbl24_tbl8_tbl8|
+.. figure:: img/tbl24_tbl8_tbl8.*
+
+ Table split into different levels
+
An entry in a table contains the following fields:
-------------------------
The LPM algorithm is used to implement the Classless Inter-Domain Routing (CIDR) strategy used by routers implementing IP forwarding.
-
-.. |tbl24_tbl8_tbl8| image:: img/tbl24_tbl8_tbl8.*
By splitting the process in two different tables/levels and limiting the number of tbl8s,
we can greatly reduce memory consumption while maintaining a very good lookup speed (one memory access, most of the times).
-.. image39 has been renamed
-|tbl24_tbl8|
+.. figure:: img/tbl24_tbl8.*
+
+ Table split into different levels
+
An entry in tbl24 contains the following fields:
* Pankaj Gupta, Algorithms for Routing Lookups and Packet Classification, PhD Thesis, Stanford University,
2000 (`http://klamath.stanford.edu/~pankaj/thesis/ thesis_1sided.pdf <http://klamath.stanford.edu/~pankaj/thesis/%20thesis_1sided.pdf>`_ )
-
-.. |tbl24_tbl8| image:: img/tbl24_tbl8.*
since these are never touched except when they are to be freed again -
at which point the pointer to the block is an input to the free() function.
-.. _pg_figure_3:
+.. _figure_malloc_heap:
-**Figure 3. Example of a malloc heap and malloc elements within the malloc library**
+.. figure:: img/malloc_heap.*
-.. image4_png has been renamed
+ Example of a malloc heap and malloc elements within the malloc library
-|malloc_heap|
Structure: malloc_elem
^^^^^^^^^^^^^^^^^^^^^^
and if so, they are merged with the current elements.
This means that we can never have two free memory blocks adjacent to one another,
they are always merged into a single block.
-
-.. |malloc_heap| image:: img/malloc_heap.*
and so on between different entities in the system.
Message buffers may also use their buffer pointers to point to other message buffer data sections or other structures.
-Figure 8 and Figure 9 show some of these scenarios.
+:numref:`figure_mbuf1` and :numref:`figure_mbuf2` show some of these scenarios.
-.. _pg_figure_8:
+.. _figure_mbuf1:
-**Figure 8. An mbuf with One Segment**
+.. figure:: img/mbuf1.*
-.. image22_png has been replaced
+ An mbuf with One Segment
-|mbuf1|
-.. _pg_figure_9:
+.. _figure_mbuf2:
-**Figure 9. An mbuf with Three Segments**
+.. figure:: img/mbuf2.*
-.. image23_png has been replaced
+ An mbuf with Three Segments
-|mbuf2|
The Buffer Manager implements a fairly standard set of buffer access functions to manipulate network packets.
---------
All networking application should use mbufs to transport network packets.
-
-.. |mbuf1| image:: img/mbuf1.*
-
-.. |mbuf2| image:: img/mbuf2.*
The command line must always have the number of memory channels specified for the processor.
-Examples of alignment for different DIMM architectures are shown in Figure 5 and Figure 6.
+Examples of alignment for different DIMM architectures are shown in
+:numref:`figure_memory-management` and :numref:`figure_memory-management2`.
-.. _pg_figure_5:
+.. _figure_memory-management:
-**Figure 5. Two Channels and Quad-ranked DIMM Example**
+.. figure:: img/memory-management.*
-.. image19_png has been replaced
+ Two Channels and Quad-ranked DIMM Example
-|memory-management|
In this case, the assumption is that a packet is 16 blocks of 64 bytes, which is not true.
The Intel® 5520 chipset has three channels, so in most cases,
no padding is required between objects (except for objects whose size are n x 3 x 64 bytes blocks).
-.. _pg_figure_6:
+.. _figure_memory-management2:
-**Figure 6. Three Channels and Two Dual-ranked DIMM Example**
+.. figure:: img/memory-management2.*
-.. image20_png has been replaced
+ Three Channels and Two Dual-ranked DIMM Example
-|memory-management2|
When creating a new pool, the user can specify to use this feature or not.
The maximum size of the cache is static and is defined at compilation time (CONFIG_RTE_MEMPOOL_CACHE_MAX_SIZE).
-Figure 7 shows a cache in operation.
+:numref:`figure_mempool` shows a cache in operation.
-.. _pg_figure_7:
+.. _figure_mempool:
-**Figure 7. A mempool in Memory with its Associated Ring**
+.. figure:: img/mempool.*
-.. image21_png has been replaced
+ A mempool in Memory with its Associated Ring
-|mempool|
Use Cases
---------
* :ref:`Environment Abstraction Layer <Environment_Abstraction_Layer>` , for logging service
* Any application that needs to allocate fixed-sized objects in the data plane and that will be continuously utilized by the system.
-
-.. |memory-management| image:: img/memory-management.*
-
-.. |memory-management2| image:: img/memory-management2.*
-
-.. |mempool| image:: img/mempool.*
Refer to Section 23.3 "Multi-process Limitations" for details of
how Linux kernel Address-Space Layout Randomization (ASLR) can affect memory sharing.
-.. _pg_figure_16:
+.. _figure_multi_process_memory:
-**Figure 16. Memory Sharing in the DPDK Multi-process Sample Application**
+.. figure:: img/multi_process_memory.*
-.. image42_png has been replaced
+ Memory Sharing in the DPDK Multi-process Sample Application
-|multi_process_memory|
The EAL also supports an auto-detection mode (set by EAL --proc-type=auto flag ),
whereby an DPDK process is started as a secondary instance if a primary instance is already running.
which means that only the first, primary DPDK process instance can open and mmap /dev/hpet.
If the number of required DPDK processes exceeds that of the number of available HPET comparators,
the TSC (which is the default timer in this release) must be used as a time source across all processes instead of the HPET.
-
-.. |multi_process_memory| image:: img/multi_process_memory.*
The *core components* are a set of libraries that provide all the elements needed
for high-performance packet processing applications.
-.. _pg_figure_1:
+.. _figure_architecture-overview:
-**Figure 1. Core Components Architecture**
+.. figure:: img/architecture-overview.*
-.. image2_png has been replaced
+ Core Components Architecture
-|architecture-overview|
Memory Manager (librte_malloc)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The librte_net library is a collection of IP protocol definitions and convenience macros.
It is based on code from the FreeBSD* IP stack and contains protocol numbers (for use in IP headers),
IP-related macros, IPv4/IPv6 header structures and TCP, UDP and SCTP header structures.
-
-.. |architecture-overview| image:: img/architecture-overview.*
and a set of worker lcores which are responsible for receiving the packets from the distributor and operating on them.
The model of operation is shown in the diagram below.
-|packet_distributor1|
+.. figure:: img/packet_distributor1.*
+
+ Packet Distributor mode of operation
+
Distributor Core Operation
--------------------------
Using the process and returned_pkts API, the following application workflow can be used,
while allowing packet order within a packet flow -- identified by a tag -- to be maintained.
-.. image41_png has been renamed
-|packet_distributor2|
+.. figure:: img/packet_distributor2.*
+
+ Application workflow
+
The flush and clear_returns API calls, mentioned previously,
are likely of less use that the process and returned_pkts APIS, and are principally provided to aid in unit testing of the library.
i.e. to save power at times of lighter load,
it is possible to have a worker stop processing packets by calling "rte_distributor_return_pkt()" to indicate that
it has finished the current packet and does not want a new one.
-
-.. |packet_distributor1| image:: img/packet_distributor1.*
-
-.. |packet_distributor2| image:: img/packet_distributor2.*
provides the set of actions to be applied on the current packet,
as well as the next hop for the packet, which can be either another table, an output port or packet drop.
-An example of packet processing pipeline is presented in Figure 32:
+An example of packet processing pipeline is presented in :numref:`figure_figure32`:
-.. _pg_figure_32:
+.. _figure_figure32:
-**Figure 32 Example of Packet Processing Pipeline where Input Ports 0 and 1 are Connected with Output Ports 0, 1 and 2 through Tables 0 and 1**
+.. figure:: img/figure32.*
-.. Object_1_png has been renamed
+ Example of Packet Processing Pipeline where Input Ports 0 and 1
+ are Connected with Output Ports 0, 1 and 2 through Tables 0 and 1
-|figure32|
Port Library Design
-------------------
this means that all the keys that end up in the same hash table bucket have the lower *n_bits* of their signature identical.
In order to reduce the number of keys in the same bucket (collisions), the number of hash table buckets needs to be increased.
-In packet processing context, the sequence of operations involved in hash table operations is described in Figure 33:
+In packet processing context, the sequence of operations involved in hash table operations is described in :numref:`figure_figure33`:
-.. _pg_figure_33:
+.. _figure_figure33:
-**Figure 33 Sequence of Steps for Hash Table Operations in a Packet Processing Context**
+.. figure:: img/figure33.*
+
+ Sequence of Steps for Hash Table Operations in a Packet Processing Context
-|figure33|
Hash Table Use Cases
Configurable Key Size Hash Table
""""""""""""""""""""""""""""""""
-Figure 34, Table 25 and Table 26 detail the main data structures used to implement configurable key size hash tables (either LRU or extendable bucket,
+:numref:`figure_figure34`, Table 25 and Table 26 detail the main data structures used to implement configurable key size hash tables (either LRU or extendable bucket,
either with pre-computed signature or "do-sig").
-.. _pg_figure_34:
+.. _figure_figure34:
-**Figure 34 Data Structures for Configurable Key Size Hash Tables**
+.. figure:: img/figure34.*
-.. image65_png has been renamed
+ Data Structures for Configurable Key Size Hash Tables
-|figure34|
.. _pg_table_25:
+---+------------------+--------------------+------------------------------------------------------------------+
-Figure 35 and Table 27 detail the bucket search pipeline stages (either LRU or extendable bucket,
+:numref:`figure_figure35` and Table 27 detail the bucket search pipeline stages (either LRU or extendable bucket,
either with pre-computed signature or "do-sig").
For each pipeline stage, the described operations are applied to each of the two packets handled by that stage.
-.. _pg_figure_35:
+.. _figure_figure35:
+
+.. figure:: img/figure35.*
-**Figure 35 Bucket Search Pipeline for Key Lookup Operation (Configurable Key Size Hash Tables)**
+ Bucket Search Pipeline for Key Lookup Operation (Configurable Key Size Hash
+ Tables)
-|figure35|
.. _pg_table_27:
| | | |
+------------+------------------------------------------+-------------------+
-The pseudo-code is displayed in Figure 36.
-
-.. _pg_figure_36:
-**Figure 36 Pseudo-code for match, match_many and match_pos**
+The pseudo-code for match, match_many and match_pos is::
match = (0xFFFELLU >> mask) & 1;
Single Key Size Hash Tables
"""""""""""""""""""""""""""
-Figure 37, Figure 38, Table 30 and 31 detail the main data structures used to implement 8-byte and 16-byte key hash tables
+:numref:`figure_figure37`, :numref:`figure_figure38`, Table 30 and 31 detail the main data structures used to implement 8-byte and 16-byte key hash tables
(either LRU or extendable bucket, either with pre-computed signature or "do-sig").
-.. _pg_figure_37:
+.. _figure_figure37:
-**Figure 37 Data Structures for 8-byte Key Hash Tables**
+.. figure:: img/figure37.*
-.. image66_png has been renamed
+ Data Structures for 8-byte Key Hash Tables
-|figure37|
-.. _pg_figure_38:
+.. _figure_figure38:
-**Figure 38 Data Structures for 16-byte Key Hash Tables**
+.. figure:: img/figure38.*
-.. image67_png has been renamed
+ Data Structures for 16-byte Key Hash Tables
-|figure38|
.. _pg_table_30:
either with pre-computed signature or "do-sig").
For each pipeline stage, the described operations are applied to each of the two packets handled by that stage.
-.. _pg_figure_39:
+.. _figure_figure39:
-**Figure 39 Bucket Search Pipeline for Key Lookup Operation (Single Key Size Hash Tables)**
+.. figure:: img/figure39.*
+
+ Bucket Search Pipeline for Key Lookup Operation (Single Key Size Hash
+ Tables)
-|figure39|
.. _pg_table_32:
with all the implementations sharing the same API: pure SW implementation (no acceleration), implementation using accelerator A, implementation using accelerator B, etc.
The selection between these implementations could be done at build time or at run-time (recommended), based on which accelerators are present in the system,
with no application changes required.
-
-.. |figure33| image:: img/figure33.*
-
-.. |figure35| image:: img/figure35.*
-
-.. |figure39| image:: img/figure39.*
-
-.. |figure34| image:: img/figure34.*
-
-.. |figure32| image:: img/figure32.*
-
-.. |figure37| image:: img/figure37.*
-
-.. |figure38| image:: img/figure38.*
An example of a complex packet processing pipeline with QoS support is shown in the following figure.
-.. _pg_figure_21:
+.. _figure_pkt_proc_pipeline_qos:
-**Figure 21. Complex Packet Processing Pipeline with QoS Support**
+.. figure:: img/pkt_proc_pipeline_qos.*
-.. image47_png has been renamed
+ Complex Packet Processing Pipeline with QoS Support
-|pkt_proc_pipeline_qos|
This pipeline can be built using reusable DPDK software libraries.
The main blocks implementing QoS in this pipeline are: the policer, the dropper and the scheduler.
as the NIC TX is requesting more packets for transmission,
these packets are later on removed and handed over to the NIC TX with the packet selection logic observing the predefined SLAs (dequeue operation).
-.. _pg_figure_22:
+.. _figure_hier_sched_blk:
-**Figure 22. Hierarchical Scheduler Block Internal Diagram**
+.. figure:: img/hier_sched_blk.*
-.. image48_png has been renamed
+ Hierarchical Scheduler Block Internal Diagram
-|hier_sched_blk|
The hierarchical scheduler is optimized for a large number of packet queues.
When only a small number of queues are needed, message passing queues should be used instead of this block.
Scheduling Hierarchy
~~~~~~~~~~~~~~~~~~~~
-The scheduling hierarchy is shown in Figure 23.
+The scheduling hierarchy is shown in :numref:`figure_sched_hier_per_port`.
The first level of the hierarchy is the Ethernet TX port 1/10/40 GbE,
with subsequent hierarchy levels defined as subport, pipe, traffic class and queue.
delay and jitter requirements, such as voice, video or data transfers.
Each queue hosts packets from one or multiple connections of the same type belonging to the same user.
-.. _pg_figure_23:
+.. _figure_sched_hier_per_port:
-**Figure 23. Scheduling Hierarchy per Port**
+.. figure:: img/sched_hier_per_port.*
-.. image49_png has been renamed
+ Scheduling Hierarchy per Port
-|sched_hier_per_port|
The functionality of each hierarchical level is detailed in the following table.
A schematic of the internal data structures in shown in with details in.
-.. _pg_figure_24:
+.. _figure_data_struct_per_port:
-**Figure 24. Internal Data Structures per Port**
+.. figure:: img/data_struct_per_port.*
-.. image50_png has been renamed
+ Internal Data Structures per Port
-|data_struct_per_port|
.. _pg_table_4:
The only other work available is to execute different stages of the enqueue sequence of operations on other input packets,
thus resulting in a pipelined implementation for the enqueue operation.
-Figure 25 illustrates a pipelined implementation for the enqueue operation with 4 pipeline stages and each stage executing 2 different input packets.
+:numref:`figure_prefetch_pipeline` illustrates a pipelined implementation for the enqueue operation with 4 pipeline stages and each stage executing 2 different input packets.
No input packet can be part of more than one pipeline stage at a given time.
-.. _pg_figure_25:
+.. _figure_prefetch_pipeline:
-**Figure 25. Prefetch Pipeline for the Hierarchical Scheduler Enqueue Operation**
+.. figure:: img/prefetch_pipeline.*
-.. image51 has been renamed
+ Prefetch Pipeline for the Hierarchical Scheduler Enqueue Operation
-|prefetch_pipeline|
The congestion management scheme implemented by the enqueue pipeline described above is very basic:
packets are enqueued until a specific queue becomes full,
therefore it tries to send as many packets from the same pipe TC and pipe as possible (up to the available packets and credits) before
moving to the next active TC from the same pipe (if any) or to another active pipe.
-.. _pg_figure_26:
+.. _figure_pipe_prefetch_sm:
-**Figure 26. Pipe Prefetch State Machine for the Hierarchical Scheduler Dequeue Operation**
+.. figure:: img/pipe_prefetch_sm.*
-.. image52 has been renamed
+ Pipe Prefetch State Machine for the Hierarchical Scheduler Dequeue
+ Operation
-|pipe_prefetch_sm|
Timing and Synchronization
^^^^^^^^^^^^^^^^^^^^^^^^^^
The purpose of the DPDK dropper is to drop packets arriving at a packet scheduler to avoid congestion.
The dropper supports the Random Early Detection (RED),
Weighted Random Early Detection (WRED) and tail drop algorithms.
-Figure 1 illustrates how the dropper integrates with the scheduler.
+:numref:`figure_blk_diag_dropper` illustrates how the dropper integrates with the scheduler.
The DPDK currently does not support congestion management
so the dropper provides the only method for congestion avoidance.
-.. _pg_figure_27:
+.. _figure_blk_diag_dropper:
-**Figure 27. High-level Block Diagram of the DPDK Dropper**
+.. figure:: img/blk_diag_dropper.*
-.. image53_png has been renamed
+ High-level Block Diagram of the DPDK Dropper
-|blk_diag_dropper|
The dropper uses the Random Early Detection (RED) congestion avoidance algorithm as documented in the reference publication.
The purpose of the RED algorithm is to monitor a packet queue,
This occurs when a packet queue has reached maximum capacity and cannot store any more packets.
In this situation, all arriving packets are dropped.
-The flow through the dropper is illustrated in Figure 28.
+The flow through the dropper is illustrated in :numref:`figure_flow_tru_droppper`.
The RED/WRED algorithm is exercised first and tail drop second.
-.. _pg_figure_28:
+.. _figure_flow_tru_droppper:
-**Figure 28. Flow Through the Dropper**
+.. figure:: img/flow_tru_droppper.*
-.. image54_png has been renamed
+ Flow Through the Dropper
-|flow_tru_droppper|
The use cases supported by the dropper are:
Enqueue Operation
~~~~~~~~~~~~~~~~~
-In the example shown in Figure 29, q (actual queue size) is the input value,
+In the example shown in :numref:`figure_ex_data_flow_tru_dropper`, q (actual queue size) is the input value,
avg (average queue size) and count (number of packets since the last drop) are run-time values,
decision is the output value and the remaining values are configuration parameters.
-.. _pg_figure_29:
+.. _figure_ex_data_flow_tru_dropper:
-**Figure 29. Example Data Flow Through Dropper**
+.. figure:: img/ex_data_flow_tru_dropper.*
-.. image55_png has been renamed
+ Example Data Flow Through Dropper
-|ex_data_flow_tru_dropper|
EWMA Filter Microblock
^^^^^^^^^^^^^^^^^^^^^^
The definition of the EWMA filter is given in the following equation.
-**Equation 1.**
-
-.. image56_png has been renamed
-
-|ewma_filter_eq_1|
+.. image:: img/ewma_filter_eq_1.*
Where:
or remaining stagnant at the last computed value.
When a packet is enqueued on an empty queue, the average queue size is computed using the following formula:
-**Equation 2.**
-
-.. image57_png has been renamed
-
-|ewma_filter_eq_2|
+.. image:: img/ewma_filter_eq_2.*
Where:
In the dropper module, *m* is defined as:
-.. image58_png has been renamed
-
-|m_definition|
+.. image:: img/m_definition.*
Where:
This method is based on the following identity:
-.. image59_png has been renamed
+.. image:: img/eq2_factor.*
-|eq2_factor|
This allows us to express the following:
-.. image60_png has been renamed
+.. image:: img/eq2_expression.*
-|eq2_expression|
In the dropper module, a look-up table is used to compute log2(1-wq) for each value of wq supported by the dropper module.
The factor (1-wq)^m can then be obtained by multiplying the table value by *m* and applying shift operations.
The initial drop probability is calculated using the following equation.
-**Equation 3.**
-
-.. image61_png has been renamed
-
-|drop_probability_eq3|
+.. image:: img/drop_probability_eq3.*
Where:
* *maxth* = maximum threshold
-The calculation of the packet drop probability using Equation 3 is illustrated in Figure 30.
+The calculation of the packet drop probability using Equation 3 is illustrated in :numref:`figure_pkt_drop_probability`.
If the average queue size is below the minimum threshold, an arriving packet is enqueued.
If the average queue size is at or above the maximum threshold, an arriving packet is dropped.
If the average queue size is between the minimum and maximum thresholds,
a drop probability is calculated to determine if the packet should be enqueued or dropped.
-.. _pg_figure_30:
+.. _figure_pkt_drop_probability:
-**Figure 30. Packet Drop Probability for a Given RED Configuration**
+.. figure:: img/pkt_drop_probability.*
-.. image62_png has been renamed
+ Packet Drop Probability for a Given RED Configuration
-|pkt_drop_probability|
Actual Drop Probability
"""""""""""""""""""""""
If the average queue size is between the minimum and maximum thresholds,
then the actual drop probability is calculated from the following equation.
-**Equation 4.**
-
-.. image63_png has been renamed
-
-|drop_probability_eq4|
+.. image:: img/drop_probability_eq4.*
Where:
It should be noted that the value pa computed from can be negative or greater than 1.
If this is the case, then a value of 1 should be used instead.
-The initial and actual drop probabilities are shown in Figure 31.
+The initial and actual drop probabilities are shown in :numref:`figure_drop_probability_graph`.
The actual drop probability is shown for the case where
the formula given in the reference document1 is used (blue curve)
and also for the case where the formula implemented in the dropper module,
The choice to deviate from the reference document is simply a design decision and
one that has been taken by other RED implementations, for example, FreeBSD* ALTQ RED.
-.. _pg_figure_31:
+.. _figure_drop_probability_graph:
-**Figure 31. Initial Drop Probability (pb), Actual Drop probability (pa) Computed Using a Factor 1 (Blue Curve) and a Factor 2 (Red Curve)**
+.. figure:: img/drop_probability_graph.*
-.. image64_png has been renamed
+ Initial Drop Probability (pb), Actual Drop probability (pa) Computed Using
+ a Factor 1 (Blue Curve) and a Factor 2 (Red Curve)
-|drop_probability_graph|
.. _Queue_Empty_Operation:
the input color of the packet is also considered.
When the output color is not red, a number of tokens equal to the length of the IP packet are
subtracted from the C or E /P or both buckets, depending on the algorithm and the output color of the packet.
-
-.. |flow_tru_droppper| image:: img/flow_tru_droppper.*
-
-.. |drop_probability_graph| image:: img/drop_probability_graph.*
-
-.. |drop_probability_eq3| image:: img/drop_probability_eq3.*
-
-.. |eq2_expression| image:: img/eq2_expression.*
-
-.. |drop_probability_eq4| image:: img/drop_probability_eq4.*
-
-.. |pkt_drop_probability| image:: img/pkt_drop_probability.*
-
-.. |pkt_proc_pipeline_qos| image:: img/pkt_proc_pipeline_qos.*
-
-.. |ex_data_flow_tru_dropper| image:: img/ex_data_flow_tru_dropper.*
-
-.. |ewma_filter_eq_1| image:: img/ewma_filter_eq_1.*
-
-.. |ewma_filter_eq_2| image:: img/ewma_filter_eq_2.*
-
-.. |data_struct_per_port| image:: img/data_struct_per_port.*
-
-.. |prefetch_pipeline| image:: img/prefetch_pipeline.*
-
-.. |pipe_prefetch_sm| image:: img/pipe_prefetch_sm.*
-
-.. |blk_diag_dropper| image:: img/blk_diag_dropper.*
-
-.. |m_definition| image:: img/m_definition.*
-
-.. |eq2_factor| image:: img/eq2_factor.*
-
-.. |sched_hier_per_port| image:: img/sched_hier_per_port.*
-
-.. |hier_sched_blk| image:: img/hier_sched_blk.*
A simplified representation of a Ring is shown in with consumer and producer head and tail pointers to objects stored in the data structure.
-.. _pg_figure_4:
+.. _figure_ring1:
-**Figure 4. Ring Structure**
+.. figure:: img/ring1.*
-.. image5_png has been replaced
+ Ring Structure
-|ring1|
References for Ring Implementation in FreeBSD*
----------------------------------------------
If there is not enough room in the ring (this is detected by checking cons_tail), it returns an error.
-.. image6_png has been replaced
-|ring-enqueue1|
+.. _figure_ring-enqueue1:
+
+.. figure:: img/ring-enqueue1.*
+
+ Enqueue first step
+
Enqueue Second Step
^^^^^^^^^^^^^^^^^^^
A pointer to the added object is copied in the ring (obj4).
-.. image7_png has been replaced
-|ring-enqueue2|
+.. _figure_ring-enqueue2:
+
+.. figure:: img/ring-enqueue2.*
+
+ Enqueue second step
+
Enqueue Last Step
^^^^^^^^^^^^^^^^^
Once the object is added in the ring, ring->prod_tail in the ring structure is modified to point to the same location as *ring->prod_head*.
The enqueue operation is finished.
-.. image8_png has been replaced
-|ring-enqueue3|
+.. _figure_ring-enqueue3:
+
+.. figure:: img/ring-enqueue3.*
+
+ Enqueue last step
+
Single Consumer Dequeue
~~~~~~~~~~~~~~~~~~~~~~~
If there are not enough objects in the ring (this is detected by checking prod_tail), it returns an error.
-.. image9_png has been replaced
-|ring-dequeue1|
+.. _figure_ring-dequeue1:
+
+.. figure:: img/ring-dequeue1.*
+
+ Dequeue last step
+
Dequeue Second Step
^^^^^^^^^^^^^^^^^^^
The pointer to the dequeued object (obj1) is copied in the pointer given by the user.
-.. image10_png has been replaced
-|ring-dequeue2|
+.. _figure_ring-dequeue2:
+
+.. figure:: img/ring-dequeue2.*
+
+ Dequeue second step
+
Dequeue Last Step
^^^^^^^^^^^^^^^^^
Finally, ring->cons_tail in the ring structure is modified to point to the same location as ring->cons_head.
The dequeue operation is finished.
-.. image11_png has been replaced
-|ring-dequeue3|
+.. _figure_ring-dequeue3:
+
+.. figure:: img/ring-dequeue3.*
+
+ Dequeue last step
+
Multiple Producers Enqueue
~~~~~~~~~~~~~~~~~~~~~~~~~~
The initial state is to have a prod_head and prod_tail pointing at the same location.
-MC Enqueue First Step
-^^^^^^^^^^^^^^^^^^^^^
+Multiple Consumer Enqueue First Step
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
On both cores, *ring->prod_head* and ring->cons_tail are copied in local variables.
The prod_next local variable points to the next element of the table,
If there is not enough room in the ring (this is detected by checking cons_tail), it returns an error.
-.. image12_png has been replaced
-|ring-mp-enqueue1|
+.. _figure_ring-mp-enqueue1:
+
+.. figure:: img/ring-mp-enqueue1.*
+
+ Multiple consumer enqueue first step
+
-MC Enqueue Second Step
-^^^^^^^^^^^^^^^^^^^^^^
+Multiple Consumer Enqueue Second Step
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
The second step is to modify ring->prod_head in the ring structure to point to the same location as prod_next.
This operation is done using a Compare And Swap (CAS) instruction, which does the following operations atomically:
In the figure, the operation succeeded on core 1, and step one restarted on core 2.
-.. image13_png has been replaced
-|ring-mp-enqueue2|
+.. _figure_ring-mp-enqueue2:
-MC Enqueue Third Step
-^^^^^^^^^^^^^^^^^^^^^
+.. figure:: img/ring-mp-enqueue2.*
+
+ Multiple consumer enqueue second step
+
+
+Multiple Consumer Enqueue Third Step
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
The CAS operation is retried on core 2 with success.
The core 1 updates one element of the ring(obj4), and the core 2 updates another one (obj5).
-.. image14_png has been replaced
-|ring-mp-enqueue3|
+.. _figure_ring-mp-enqueue3:
+
+.. figure:: img/ring-mp-enqueue3.*
+
+ Multiple consumer enqueue third step
+
-MC Enqueue Fourth Step
-^^^^^^^^^^^^^^^^^^^^^^
+Multiple Consumer Enqueue Fourth Step
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Each core now wants to update ring->prod_tail.
A core can only update it if ring->prod_tail is equal to the prod_head local variable.
This is only true on core 1. The operation is finished on core 1.
-.. image15_png has been replaced
-|ring-mp-enqueue4|
+.. _figure_ring-mp-enqueue4:
-MC Enqueue Last Step
-^^^^^^^^^^^^^^^^^^^^
+.. figure:: img/ring-mp-enqueue4.*
+
+ Multiple consumer enqueue fourth step
+
+
+Multiple Consumer Enqueue Last Step
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Once ring->prod_tail is updated by core 1, core 2 is allowed to update it too.
The operation is also finished on core 2.
-.. image16_png has been replaced
-|ring-mp-enqueue5|
+.. _figure_ring-mp-enqueue5:
+
+.. figure:: img/ring-mp-enqueue5.*
+
+ Multiple consumer enqueue last step
+
Modulo 32-bit Indexes
~~~~~~~~~~~~~~~~~~~~~
In addition, the four indexes are defined as unsigned 16-bit integers,
as opposed to unsigned 32-bit integers in the more realistic case.
-.. image17_png has been replaced
-|ring-modulo1|
+.. _figure_ring-modulo1:
+
+.. figure:: img/ring-modulo1.*
+
+ Modulo 32-bit indexes - Example 1
+
This ring contains 11000 entries.
-.. image18_png has been replaced
-|ring-modulo2|
+.. _figure_ring-modulo2:
+
+.. figure:: img/ring-modulo2.*
+
+ Modulo 32-bit indexes - Example 2
+
This ring contains 12536 entries.
* `bufring.c in FreeBSD <http://svn.freebsd.org/viewvc/base/release/8.0.0/sys/kern/subr_bufring.c?revision=199625&view=markup>`_ (version 8)
* `Linux Lockless Ring Buffer Design <http://lwn.net/Articles/340400/>`_
-
-.. |ring1| image:: img/ring1.*
-
-.. |ring-enqueue1| image:: img/ring-enqueue1.*
-
-.. |ring-enqueue2| image:: img/ring-enqueue2.*
-
-.. |ring-enqueue3| image:: img/ring-enqueue3.*
-
-.. |ring-dequeue1| image:: img/ring-dequeue1.*
-
-.. |ring-dequeue2| image:: img/ring-dequeue2.*
-
-.. |ring-dequeue3| image:: img/ring-dequeue3.*
-
-.. |ring-mp-enqueue1| image:: img/ring-mp-enqueue1.*
-
-.. |ring-mp-enqueue2| image:: img/ring-mp-enqueue2.*
-
-.. |ring-mp-enqueue3| image:: img/ring-mp-enqueue3.*
-
-.. |ring-mp-enqueue4| image:: img/ring-mp-enqueue4.*
-
-.. |ring-mp-enqueue5| image:: img/ring-mp-enqueue5.*
-
-.. |ring-modulo1| image:: img/ring-modulo1.*
-
-.. |ring-modulo2| image:: img/ring-modulo2.*
This application can be used to benchmark performance using the traffic
generator as shown in the figure below.
-.. _figure_22:
+.. _figure_dist_perf:
-**Figure 22. Performance Benchmarking Setup (Basic Environment)**
+.. figure:: img/dist_perf.*
+
+ Performance Benchmarking Setup (Basic Environment)
-|dist_perf|
Compiling the Application
-------------------------
The distributor application consists of three types of threads: a receive
thread (lcore_rx()), a set of worker threads(lcore_worker())
and a transmit thread(lcore_tx()). How these threads work together is shown
-in Fig2 below. The main() function launches threads of these three types.
+in :numref:`figure_dist_app` below. The main() function launches threads of these three types.
Each thread has a while loop which will be doing processing and which is
terminated only upon SIGINT or ctrl+C. The receive and transmit threads
communicate using a software ring (rte_ring structure).
in the application will terminate all running threads gracefully and print
final statistics to the user.
-.. _figure_23:
+.. _figure_dist_app:
+
+.. figure:: img/dist_app.*
-**Figure 23. Distributor Sample Application Layout**
+ Distributor Sample Application Layout
-|dist_app|
Debug Logging Support
---------------------
TX queue initialization is done in the same way as it is done in the L2 Forwarding
Sample Application. See Section 9.4.5, "TX Queue Initialization".
-
-.. |dist_perf| image:: img/dist_perf.*
-
-.. |dist_app| image:: img/dist_app.*
The packet flow through the exception path application is as shown in the following figure.
-.. _figure_1:
+.. _figure_exception_path_example:
-**Figure 1. Packet Flow**
+.. figure:: img/exception_path_example.*
-.. image2_png has been replaced
+ Packet Flow
-|exception_path_example|
To make throughput measurements, kernel bridges must be setup to forward data between the bridges appropriately.
brctl delbr br0
openvpn --rmtun --dev tap_dpdk_00
-.. |exception_path_example| image:: img/exception_path_example.*
**Figures**
-:ref:`Figure 1.Packet Flow <figure_1>`
+:numref:`figure_exception_path_example` :ref:`figure_exception_path_example`
-:ref:`Figure 2.Kernel NIC Application Packet Flow <figure_2>`
+:numref:`figure_kernel_nic` :ref:`figure_kernel_nic`
-:ref:`Figure 3.Performance Benchmark Setup (Basic Environment) <figure_3>`
+:numref:`figure_l2_fwd_benchmark_setup_jobstats` :ref:`figure_l2_fwd_benchmark_setup_jobstats`
-:ref:`Figure 4.Performance Benchmark Setup (Virtualized Environment) <figure_4>`
+:numref:`figure_l2_fwd_virtenv_benchmark_setup_jobstats` :ref:`figure_l2_fwd_virtenv_benchmark_setup_jobstats`
-:ref:`Figure 5.Load Balancer Application Architecture <figure_5>`
+:numref:`figure_l2_fwd_benchmark_setup` :ref:`figure_l2_fwd_benchmark_setup`
-:ref:`Figure 5.Example Rules File <figure_5_1>`
+:numref:`figure_l2_fwd_virtenv_benchmark_setup` :ref:`figure_l2_fwd_virtenv_benchmark_setup`
-:ref:`Figure 6.Example Data Flow in a Symmetric Multi-process Application <figure_6>`
+:numref:`figure_ipv4_acl_rule` :ref:`figure_ipv4_acl_rule`
-:ref:`Figure 7.Example Data Flow in a Client-Server Symmetric Multi-process Application <figure_7>`
+:numref:`figure_example_rules` :ref:`figure_example_rules`
-:ref:`Figure 8.Master-slave Process Workflow <figure_8>`
+:numref:`figure_load_bal_app_arch` :ref:`figure_load_bal_app_arch`
-:ref:`Figure 9.Slave Process Recovery Process Flow <figure_9>`
+:numref:`figure_sym_multi_proc_app` :ref:`figure_sym_multi_proc_app`
-:ref:`Figure 10.QoS Scheduler Application Architecture <figure_10>`
+:numref:`figure_client_svr_sym_multi_proc_app` :ref:`figure_client_svr_sym_multi_proc_app`
-:ref:`Figure 11.Intel®QuickAssist Technology Application Block Diagram <figure_11>`
+:numref:`figure_master_slave_proc` :ref:`figure_master_slave_proc`
-:ref:`Figure 12.Pipeline Overview <figure_12>`
+:numref:`figure_slave_proc_recov` :ref:`figure_slave_proc_recov`
-:ref:`Figure 13.Ring-based Processing Pipeline Performance Setup <figure_13>`
+:numref:`figure_qos_sched_app_arch` :ref:`figure_qos_sched_app_arch`
-:ref:`Figure 14.Threads and Pipelines <figure_14>`
+:numref:`figure_quickassist_block_diagram` :ref:`figure_quickassist_block_diagram`
-:ref:`Figure 15.Packet Flow Through the VMDQ and DCB Sample Application <figure_15>`
+:numref:`figure_pipeline_overview` :ref:`figure_pipeline_overview`
-:ref:`Figure 16.QEMU Virtio-net (prior to vhost-net) <figure_16>`
+:numref:`figure_ring_pipeline_perf_setup` :ref:`figure_ring_pipeline_perf_setup`
-:ref:`Figure 17.Virtio with Linux* Kernel Vhost <figure_17>`
+:numref:`figure_threads_pipelines` :ref:`figure_threads_pipelines`
-:ref:`Figure 18.Vhost-net Architectural Overview <figure_18>`
+:numref:`figure_vmdq_dcb_example` :ref:`figure_vmdq_dcb_example`
-:ref:`Figure 19.Packet Flow Through the vhost-net Sample Application <figure_19>`
+:numref:`figure_qemu_virtio_net` :ref:`figure_qemu_virtio_net`
-:ref:`Figure 20.Packet Flow on TX in DPDK-testpmd <figure_20>`
+:numref:`figure_virtio_linux_vhost` :ref:`figure_virtio_linux_vhost`
-:ref:`Figure 21.Test Pipeline Application <figure_21>`
+:numref:`figure_vhost_net_arch` :ref:`figure_vhost_net_arch`
-:ref:`Figure 22.Performance Benchmarking Setup (Basic Environment) <figure_22>`
+:numref:`figure_vhost_net_sample_app` :ref:`figure_vhost_net_sample_app`
-:ref:`Figure 23.Distributor Sample Application Layout <figure_23>`
+:numref:`figure_tx_dpdk_testpmd` :ref:`figure_tx_dpdk_testpmd`
-:ref:`Figure 24.High level Solution <figure_24>`
+:numref:`figure_test_pipeline_app` :ref:`figure_test_pipeline_app`
-:ref:`Figure 25.VM request to scale frequency <figure_25>`
+:numref:`figure_dist_perf` :ref:`figure_dist_perf`
+
+:numref:`figure_dist_app` :ref:`figure_dist_app`
+
+:numref:`figure_vm_power_mgr_highlevel` :ref:`figure_vm_power_mgr_highlevel`
+
+:numref:`figure_vm_power_mgr_vm_request_seq` :ref:`figure_vm_power_mgr_vm_request_seq`
**Tables**
Overview
--------
-An overview of the application is provided in Figure 11.
+An overview of the application is provided in :numref:`figure_quickassist_block_diagram`.
For simplicity, only two NIC ports and one Intel® QuickAssist Technology device are shown in this diagram,
although the number of NIC ports and Intel® QuickAssist Technology devices can be different.
-.. _figure_11:
+.. _figure_quickassist_block_diagram:
-**Figure 11. Intel® QuickAssist Technology Application Block Diagram**
+.. figure:: img/quickassist_block_diagram.*
-.. image14_png has been renamed
+ Intel® QuickAssist Technology Application Block Diagram
-|quickassist_block_diagram|
The application allows the configuration of the following items:
Refer to the *DPDK Test Report* for more examples of traffic generator setup and the application startup command lines.
If no errors are generated in response to the startup commands, the application is running correctly.
-
-.. |quickassist_block_diagram| image:: img/quickassist_block_diagram.*
The packet flow through the Kernel NIC Interface application is as shown in the following figure.
-.. _figure_2:
+.. _figure_kernel_nic:
-**Figure 2. Kernel NIC Application Packet Flow**
+.. figure:: img/kernel_nic.*
-.. image3_png has been renamed to kernel_nic.*
+ Kernel NIC Application Packet Flow
-|kernel_nic|
Compiling the Application
-------------------------
RTE_LOG(ERR, APP, "Failed to start port %d\n", port_id);
return ret;
}
-
-.. |kernel_nic| image:: img/kernel_nic.*
* The destination MAC address is replaced by 02:00:00:00:00:TX_PORT_ID
-This application can be used to benchmark performance using a traffic-generator, as shown in the Figure 3.
+This application can be used to benchmark performance using a traffic-generator, as shown in the :numref:`figure_l2_fwd_benchmark_setup_jobstats`.
-The application can also be used in a virtualized environment as shown in Figure 4.
+The application can also be used in a virtualized environment as shown in :numref:`figure_l2_fwd_virtenv_benchmark_setup_jobstats`.
The L2 Forwarding application can also be used as a starting point for developing a new application based on the DPDK.
-.. _figure_3:
+.. _figure_l2_fwd_benchmark_setup_jobstats:
-**Figure 3. Performance Benchmark Setup (Basic Environment)**
+.. figure:: img/l2_fwd_benchmark_setup.*
-.. image4_png has been replaced
+ Performance Benchmark Setup (Basic Environment)
-|l2_fwd_benchmark_setup|
+.. _figure_l2_fwd_virtenv_benchmark_setup_jobstats:
-.. _figure_4:
+.. figure:: img/l2_fwd_virtenv_benchmark_setup.*
-**Figure 4. Performance Benchmark Setup (Virtualized Environment)**
+ Performance Benchmark Setup (Virtualized Environment)
-.. image5_png has been renamed
-
-|l2_fwd_virtenv_benchmark_setup|
Virtual Function Setup Instructions
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
* in which it was called. */
rte_jobstats_finish(&qconf->flush_job, qconf->flush_job.target);
}
-
-.. |l2_fwd_benchmark_setup| image:: img/l2_fwd_benchmark_setup.*
-
-.. |l2_fwd_virtenv_benchmark_setup| image:: img/l2_fwd_virtenv_benchmark_setup.*
* The destination MAC address is replaced by 02:00:00:00:00:TX_PORT_ID
-This application can be used to benchmark performance using a traffic-generator, as shown in the Figure 3.
+This application can be used to benchmark performance using a traffic-generator, as shown in the :numref:`figure_l2_fwd_benchmark_setup`.
-The application can also be used in a virtualized environment as shown in Figure 4.
+The application can also be used in a virtualized environment as shown in :numref:`figure_l2_fwd_virtenv_benchmark_setup`.
The L2 Forwarding application can also be used as a starting point for developing a new application based on the DPDK.
-.. _figure_3:
+.. _figure_l2_fwd_benchmark_setup:
-**Figure 3. Performance Benchmark Setup (Basic Environment)**
+.. figure:: img/l2_fwd_benchmark_setup.*
-.. image4_png has been replaced
+ Performance Benchmark Setup (Basic Environment)
-|l2_fwd_benchmark_setup|
-.. _figure_4:
+.. _figure_l2_fwd_virtenv_benchmark_setup:
-**Figure 4. Performance Benchmark Setup (Virtualized Environment)**
+.. figure:: img/l2_fwd_virtenv_benchmark_setup.*
-.. image5_png has been renamed
+ Performance Benchmark Setup (Virtualized Environment)
-|l2_fwd_virtenv_benchmark_setup|
Virtual Function Setup Instructions
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
prev_tsc = cur_tsc;
}
-
-.. |l2_fwd_benchmark_setup| image:: img/l2_fwd_benchmark_setup.*
-
-.. |l2_fwd_virtenv_benchmark_setup| image:: img/l2_fwd_virtenv_benchmark_setup.*
* A typical IPv4 ACL rule line should have a format as shown below:
-.. image6_png has been renamed
-|ipv4_acl_rule|
+.. _figure_ipv4_acl_rule:
+
+.. figure:: img/ipv4_acl_rule.*
+
+ A typical IPv4 ACL rule
+
IPv4 addresses are specified in CIDR format as specified in RFC 4632.
They consist of the dot notation for the address and a prefix length separated by '/'.
Rules File Example
~~~~~~~~~~~~~~~~~~
-.. _figure_5_1:
+.. _figure_example_rules:
-Figure 5 is an example of a rules file. This file has three rules, one for ACL and two for route information.
+.. figure:: img/example_rules.*
-**Figure 5.Example Rules File**
+ Rules example
-.. image7_png has been renamed
-
-|example_rules|
Each rule is explained as follows:
adds rules parsed from the file into the database and build an ACL trie.
It is important to note that the application creates an independent copy of each database for each socket CPU
involved in the task to reduce the time for remote memory access.
-
-.. |ipv4_acl_rule| image:: img/ipv4_acl_rule.*
-
-.. |example_rules| image:: img/example_rules.*
The architecture of the Load Balance application is presented in the following figure.
-.. _figure_5:
+.. _figure_load_bal_app_arch:
-**Figure 5. Load Balancer Application Architecture**
+.. figure:: img/load_bal_app_arch.*
-.. image8_png has been renamed
+ Load Balancer Application Architecture
-|load_bal_app_arch|
For the sake of simplicity, the diagram illustrates a specific case of two I/O RX and two I/O TX lcores off loading the packet I/O
overhead incurred by four NIC ports from four worker cores, with each I/O lcore handling RX/TX for two NIC ports.
#. ABC: The packet is received on socket A, it is processed by an lcore on socket B,
then it has to be transmitted out by a NIC connected to socket C.
The performance price for crossing the CPU socket boundary is paid twice for this packet.
-
-.. |load_bal_app_arch| image:: img/load_bal_app_arch.*
where different processes perform different tasks, yet co-operate to form a packet-processing system.)
The following diagram shows the data-flow through the application, using two processes.
-.. _figure_6:
+.. _figure_sym_multi_proc_app:
-**Figure 6. Example Data Flow in a Symmetric Multi-process Application**
+.. figure:: img/sym_multi_proc_app.*
-.. image9_png has been renamed
+ Example Data Flow in a Symmetric Multi-process Application
-|sym_multi_proc_app|
As the diagram shows, each process reads packets from each of the network ports in use.
RSS is used to distribute incoming packets on each port to different hardware RX queues.
The following diagram shows the data-flow through the application, using two client processes.
-.. _figure_7:
+.. _figure_client_svr_sym_multi_proc_app:
-**Figure 7. Example Data Flow in a Client-Server Symmetric Multi-process Application**
+.. figure:: img/client_svr_sym_multi_proc_app.*
-.. image10_png has been renamed
+ Example Data Flow in a Client-Server Symmetric Multi-process Application
-|client_svr_sym_multi_proc_app|
Running the Application
^^^^^^^^^^^^^^^^^^^^^^^
memory, CPU cores, driver, ports, and so on, as the other examples do.
Thereafter, it creates slave processes, as shown in the following figure.
-.. _figure_8:
+.. _figure_master_slave_proc:
-**Figure 8. Master-slave Process Workflow**
+.. figure:: img/master_slave_proc.*
-.. image11_png has been renamed
+ Master-slave Process Workflow
-|master_slave_proc|
The master process calls the rte_eal_mp_remote_launch() EAL function to launch an application function for each pinned thread through the pipe.
Then, it waits to check if any slave processes have exited.
The following diagram describes slave process recovery.
-.. _figure_9:
+.. _figure_slave_proc_recov:
-**Figure 9. Slave Process Recovery Process Flow**
+.. figure:: img/slave_proc_recov.*
-.. image12_png has been renamed
+ Slave Process Recovery Process Flow
-|slave_proc_recov|
Floating Process Support
^^^^^^^^^^^^^^^^^^^^^^^^
}
return 0;
}
-
-.. |sym_multi_proc_app| image:: img/sym_multi_proc_app.*
-
-.. |client_svr_sym_multi_proc_app| image:: img/client_svr_sym_multi_proc_app.*
-
-.. |master_slave_proc| image:: img/master_slave_proc.*
-
-.. |slave_proc_recov| image:: img/slave_proc_recov.*
The architecture of the QoS scheduler application is shown in the following figure.
-.. _figure_10:
+.. _figure_qos_sched_app_arch:
-**Figure 10. QoS Scheduler Application Architecture**
+.. figure:: img/qos_sched_app_arch.*
-.. image13_png has been renamed
+ QoS Scheduler Application Architecture
-|qos_sched_app_arch|
There are two flavors of the runtime execution for this application,
with two or three threads per each packet flow configuration being used.
+----------------+-------------------------+--------------------------------------------------+----------------------------------+
Please refer to the "QoS Scheduler" chapter in the *DPDK Programmer's Guide* for more information about these parameters.
-
-.. |qos_sched_app_arch| image:: img/qos_sched_app_arch.*
The MAC addresses of the forwarded Ethernet frames are not affected.
Internally, packets are pulled from the ports by the master logical core and put on a variable length processing pipeline,
-each stage of which being connected by rings, as shown in Figure 12.
+each stage of which being connected by rings, as shown in :numref:`figure_pipeline_overview`.
-.. _figure_12:
+.. _figure_pipeline_overview:
-**Figure 12. Pipeline Overview**
+.. figure:: img/pipeline_overview.*
-.. image15_png has been renamed
+ Pipeline Overview
-|pipeline_overview|
An adjustable quota value controls how many packets are being moved through the pipeline per enqueue and dequeue.
Adjustable watermark values associated with the rings control a back-off mechanism that
On top of serving as an example of quota and watermark usage,
this application can be used to benchmark ring based processing pipelines performance using a traffic- generator,
-as shown in Figure 13.
+as shown in :numref:`figure_ring_pipeline_perf_setup`.
-.. _figure_13:
+.. _figure_ring_pipeline_perf_setup:
-**Figure 13. Ring-based Processing Pipeline Performance Setup**
+.. figure:: img/ring_pipeline_perf_setup.*
-.. image16_png has been renamed
+ Ring-based Processing Pipeline Performance Setup
-|ring_pipeline_perf_setup|
Compiling the Application
-------------------------
The application uses the master logical core to poll all the ports for new packets and enqueue them on a ring associated with the port.
Each logical core except the last runs pipeline_stage() after a ring for each used port is initialized on that core.
-pipeline_stage() on core X dequeues packets from core X-1's rings and enqueue them on its own rings. See Figure 14.
+pipeline_stage() on core X dequeues packets from core X-1's rings and enqueue them on its own rings. See :numref:`figure_threads_pipelines`.
.. code-block:: c
Receive, Process and Transmit Packets
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-.. _figure_14:
+.. _figure_threads_pipelines:
-Figure 14 shows where each thread in the pipeline is.
-It should be used as a reference while reading the rest of this section.
+.. figure:: img/threads_pipelines.*
-**Figure 14. Threads and Pipelines**
+ Threads and Pipelines
-.. image17_png has been renamed
-
-|threads_pipelines|
In the receive_stage() function running on the master logical core,
the main task is to read ingress packets from the RX ports and enqueue them
low_watermark = (unsigned int *) qw_memzone->addr + sizeof(int);
}
-
-.. |pipeline_overview| image:: img/pipeline_overview.*
-
-.. |ring_pipeline_perf_setup| image:: img/ring_pipeline_perf_setup.*
-
-.. |threads_pipelines| image:: img/threads_pipelines.*
* Core C ("TX core") receives traffic from core B through software queues and sends it to the NIC ports for transmission.
-.. _figure_21:
+.. _figure_test_pipeline_app:
-**Figure 21.Test Pipeline Application**
+.. figure:: img/test_pipeline_app.*
-.. image24_png has been renamed
+ Test Pipeline Application
-|test_pipeline_app|
Compiling the Application
-------------------------
* destination TCP port fixed to 0
* source TCP port fixed to 0
-
-.. |test_pipeline_app| image:: img/test_pipeline_app.*
It was found that virtio-net performance was poor due to context switching and packet copying between host, guest, and QEMU.
The following figure shows the system architecture for a virtio-based networking (virtio-net).
-.. _figure_16:
+.. _figure_qemu_virtio_net:
-**Figure16. QEMU Virtio-net (prior to vhost-net)**
+.. figure:: img/qemu_virtio_net.*
-.. image19_png has been renamed
+ System Architecture for Virtio-based Networking (virtio-net).
-|qemu_virtio_net|
The Linux* Kernel vhost-net module was developed as an offload mechanism for virtio-net.
The vhost-net module enables KVM (QEMU) to offload the servicing of virtio-net devices to the vhost-net kernel module,
The following figure shows the system architecture for virtio-net networking with vhost-net offload.
-.. _figure_17:
+.. _figure_virtio_linux_vhost:
-**Figure 17. Virtio with Linux* Kernel Vhost**
+.. figure:: img/virtio_linux_vhost.*
-.. image20_png has been renamed
+ Virtio with Linux
-|virtio_linux_vhost|
Sample Code Overview
--------------------
The following figure shows the architecture of the Vhost sample application based on vhost-cuse.
-.. _figure_18:
+.. _figure_vhost_net_arch:
-**Figure 18. Vhost-net Architectural Overview**
+.. figure:: img/vhost_net_arch.*
-.. image21_png has been renamed
+ Vhost-net Architectural Overview
-|vhost_net_arch|
The following figure shows the flow of packets through the vhost-net sample application.
-.. _figure_19:
+.. _figure_vhost_net_sample_app:
-**Figure 19. Packet Flow Through the vhost-net Sample Application**
+.. figure:: img/vhost_net_sample_app.*
-.. image22_png has been renamed
+ Packet Flow Through the vhost-net Sample Application
-|vhost_net_sample_app|
Supported Distributions
-----------------------
The "wait and retry" algorithm is implemented in DPDK testpmd as a forwarding method call "mac_retry".
The following sequence diagram describes the algorithm in detail.
-.. _figure_20:
+.. _figure_tx_dpdk_testpmd:
-**Figure 20. Packet Flow on TX in DPDK-testpmd**
+.. figure:: img/tx_dpdk_testpmd.*
-.. image23_png has been renamed
+ Packet Flow on TX in DPDK-testpmd
-|tx_dpdk_testpmd|
Running Testpmd
~~~~~~~~~~~~~~~
The above message indicates that device 0 has been registered with MAC address cc:bb:bb:bb:bb:bb and VLAN tag 1000.
Any packets received on the NIC with these values is placed on the devices receive queue.
When a virtio-net device transmits packets, the VLAN tag is added to the packet by the DPDK vhost sample code.
-
-.. |vhost_net_arch| image:: img/vhost_net_arch.*
-
-.. |qemu_virtio_net| image:: img/qemu_virtio_net.*
-
-.. |tx_dpdk_testpmd| image:: img/tx_dpdk_testpmd.*
-
-.. |vhost_net_sample_app| image:: img/vhost_net_sample_app.*
-
-.. |virtio_linux_vhost| image:: img/virtio_linux_vhost.*
The l3fwd-power application will use this implementation when deployed on a VM
(see Chapter 11 "L3 Forwarding with Power Management Application").
-.. _figure_24:
+.. _figure_vm_power_mgr_highlevel:
-**Figure 24. Highlevel Solution**
+.. figure:: img/vm_power_mgr_highlevel.*
+
+ Highlevel Solution
-|vm_power_mgr_highlevel|
Overview
--------
Upon receiving a request, the host translates the vCPU to a pCPU via
the libvirt API before forwarding to the host librte_power.
-.. _figure_25:
+.. _figure_vm_power_mgr_vm_request_seq:
+
+.. figure:: img/vm_power_mgr_vm_request_seq.*
-**Figure 25. VM request to scale frequency**
+ VM request to scale frequency
-|vm_power_mgr_vm_request_seq|
Performance Considerations
~~~~~~~~~~~~~~~~~~~~~~~~~~
.. code-block:: console
set_cpu_freq {core_num} up|down|min|max
-
-.. |vm_power_mgr_highlevel| image:: img/vm_power_mgr_highlevel.*
-
-.. |vm_power_mgr_vm_request_seq| image:: img/vm_power_mgr_vm_request_seq.*
The traffic is split into 128 queues on input, where each thread of the application reads from multiple queues.
For example, when run with 8 threads, that is, with the -c FF option, each thread receives and forwards packets from 16 queues.
-As supplied, the sample application configures the VMDQ feature to have 16 pools with 8 queues each as indicated in Figure 15.
+As supplied, the sample application configures the VMDQ feature to have 16 pools with 8 queues each as indicated in :numref:`figure_vmdq_dcb_example`.
The Intel® 82599 10 Gigabit Ethernet Controller NIC also supports the splitting of traffic into 32 pools of 4 queues each and
this can be used by changing the NUM_POOLS parameter in the supplied code.
The NUM_POOLS parameter can be passed on the command line, after the EAL parameters:
where, NP can be 16 or 32.
-.. _figure_15:
+.. _figure_vmdq_dcb_example:
-**Figure 15. Packet Flow Through the VMDQ and DCB Sample Application**
+.. figure:: img/vmdq_dcb_example.*
-.. image18_png has been replaced
+ Packet Flow Through the VMDQ and DCB Sample Application
-|vmdq_dcb_example|
In Linux* user space, the application can display statistics with the number of packets received on each queue.
To have the application display the statistics, send a SIGHUP signal to the running application process, as follows:
Please note that the statistics output will appear on the terminal where the vmdq_dcb_app is running,
rather than the terminal from which the HUP signal was sent.
-
-.. |vmdq_dcb_example| image:: img/vmdq_dcb_example.*
The following diagram describes the functionality of the DPDK Xen Packet- Switching Solution.
-.. image35_png has been renamed
-|dpdk_xen_pkt_switch|
+.. _figure_dpdk_xen_pkt_switch:
+
+.. figure:: img/dpdk_xen_pkt_switch.*
+
+ Functionality of the DPDK Xen Packet Switching Solution.
+
Note 1 The Xen hypervisor uses a mechanism called a Grant Table to share memory between domains
(`http://wiki.xen.org/wiki/Grant Table <http://wiki.xen.org/wiki/Grant%20Table>`_).
A diagram of the design is shown below, where "gva" is the Guest Virtual Address,
which is the data pointer of the mbuf, and "hva" is the Host Virtual Address:
-.. image36_png has been renamed
-|grant_table|
+.. _figure_grant_table:
+
+.. figure:: img/grant_table.*
+
+ DPDK Xen Layout
+
In this design, a Virtio ring is used as a para-virtualized interface for better performance over a Xen private ring
when packet switching to and from a VM.
The real Grant reference information is stored in this virtual address space,
where (gref, pfn) pairs follow each other with -1 as the terminator.
-.. image37_pnng has been renamed
-|grant_refs|
+.. _figure_grant_refs:
+
+.. figure:: img/grant_refs.*
+
+ Mapping Grant references to a continuous virtual address space
+
After all gref# IDs are retrieved, the host maps them to a continuous virtual address space.
With the guest mempool virtual address, the host establishes 1:1 address mapping.
The packet flow is:
packet generator->Virtio in guest VM1->switching backend->Virtio in guest VM2->switching backend->wire
-
-.. |grant_table| image:: img/grant_table.*
-
-.. |grant_refs| image:: img/grant_refs.*
-
-.. |dpdk_xen_pkt_switch| image:: img/dpdk_xen_pkt_switch.*
git.droids-corp.org / dpdk.git / commitdiff