1.. SPDX-License-Identifier: BSD-3-Clause 2 Copyright(c) 2010-2014 Intel Corporation. 3 4.. _Environment_Abstraction_Layer: 5 6Environment Abstraction Layer 7============================= 8 9The Environment Abstraction Layer (EAL) is responsible for gaining access to low-level resources such as hardware and memory space. 10It provides a generic interface that hides the environment specifics from the applications and libraries. 11It is the responsibility of the initialization routine to decide how to allocate these resources 12(that is, memory space, devices, timers, consoles, and so on). 13 14Typical services expected from the EAL are: 15 16* DPDK Loading and Launching: 17 The DPDK and its application are linked as a single application and must be loaded by some means. 18 19* Core Affinity/Assignment Procedures: 20 The EAL provides mechanisms for assigning execution units to specific cores as well as creating execution instances. 21 22* System Memory Reservation: 23 The EAL facilitates the reservation of different memory zones, for example, physical memory areas for device interactions. 24 25* Trace and Debug Functions: Logs, dump_stack, panic and so on. 26 27* Utility Functions: Spinlocks and atomic counters that are not provided in libc. 28 29* CPU Feature Identification: Determine at runtime if a particular feature, for example, Intel® AVX is supported. 30 Determine if the current CPU supports the feature set that the binary was compiled for. 31 32* Interrupt Handling: Interfaces to register/unregister callbacks to specific interrupt sources. 33 34* Alarm Functions: Interfaces to set/remove callbacks to be run at a specific time. 35 36EAL in a Linux-userland Execution Environment 37--------------------------------------------- 38 39In a Linux user space environment, the DPDK application runs as a user-space application using the pthread library. 40 41The EAL performs physical memory allocation using mmap() in hugetlbfs (using huge page sizes to increase performance). 42This memory is exposed to DPDK service layers such as the :ref:`Mempool Library <Mempool_Library>`. 43 44At this point, the DPDK services layer will be initialized, then through pthread setaffinity calls, 45each execution unit will be assigned to a specific logical core to run as a user-level thread. 46 47The time reference is provided by the CPU Time-Stamp Counter (TSC) or by the HPET kernel API through a mmap() call. 48 49Initialization and Core Launching 50~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 51 52Part of the initialization is done by the start function of glibc. 53A check is also performed at initialization time to ensure that the micro architecture type chosen in the config file is supported by the CPU. 54Then, the main() function is called. The core initialization and launch is done in rte_eal_init() (see the API documentation). 55It consist of calls to the pthread library (more specifically, pthread_self(), pthread_create(), and pthread_setaffinity_np()). 56 57.. _figure_linux_launch: 58 59.. figure:: img/linuxapp_launch.* 60 61 EAL Initialization in a Linux Application Environment 62 63 64.. note:: 65 66 Initialization of objects, such as memory zones, rings, memory pools, lpm tables and hash tables, 67 should be done as part of the overall application initialization on the main lcore. 68 The creation and initialization functions for these objects are not multi-thread safe. 69 However, once initialized, the objects themselves can safely be used in multiple threads simultaneously. 70 71Shutdown and Cleanup 72~~~~~~~~~~~~~~~~~~~~ 73 74During the initialization of EAL resources such as hugepage backed memory can be 75allocated by core components. The memory allocated during ``rte_eal_init()`` 76can be released by calling the ``rte_eal_cleanup()`` function. Refer to the 77API documentation for details. 78 79Multi-process Support 80~~~~~~~~~~~~~~~~~~~~~ 81 82The Linux EAL allows a multi-process as well as a multi-threaded (pthread) deployment model. 83See chapter 84:ref:`Multi-process Support <Multi-process_Support>` for more details. 85 86Memory Mapping Discovery and Memory Reservation 87~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 88 89The allocation of large contiguous physical memory is done using hugepages. 90The EAL provides an API to reserve named memory zones in this contiguous memory. 91The physical address of the reserved memory for that memory zone is also returned to the user by the memory zone reservation API. 92 93There are two modes in which DPDK memory subsystem can operate: dynamic mode, 94and legacy mode. Both modes are explained below. 95 96.. note:: 97 98 Memory reservations done using the APIs provided by rte_malloc 99 are also backed by hugepages unless ``--no-huge`` option is given. 100 101Dynamic Memory Mode 102^^^^^^^^^^^^^^^^^^^ 103 104Currently, this mode is only supported on Linux and Windows. 105 106In this mode, usage of hugepages by DPDK application will grow and shrink based 107on application's requests. Any memory allocation through ``rte_malloc()``, 108``rte_memzone_reserve()`` or other methods, can potentially result in more 109hugepages being reserved from the system. Similarly, any memory deallocation can 110potentially result in hugepages being released back to the system. 111 112Memory allocated in this mode is not guaranteed to be IOVA-contiguous. If large 113chunks of IOVA-contiguous are required (with "large" defined as "more than one 114page"), it is recommended to either use VFIO driver for all physical devices (so 115that IOVA and VA addresses can be the same, thereby bypassing physical addresses 116entirely), or use legacy memory mode. 117 118For chunks of memory which must be IOVA-contiguous, it is recommended to use 119``rte_memzone_reserve()`` function with ``RTE_MEMZONE_IOVA_CONTIG`` flag 120specified. This way, memory allocator will ensure that, whatever memory mode is 121in use, either reserved memory will satisfy the requirements, or the allocation 122will fail. 123 124There is no need to preallocate any memory at startup using ``-m`` or 125``--socket-mem`` command-line parameters, however it is still possible to do so, 126in which case preallocate memory will be "pinned" (i.e. will never be released 127by the application back to the system). It will be possible to allocate more 128hugepages, and deallocate those, but any preallocated pages will not be freed. 129If neither ``-m`` nor ``--socket-mem`` were specified, no memory will be 130preallocated, and all memory will be allocated at runtime, as needed. 131 132Another available option to use in dynamic memory mode is 133``--single-file-segments`` command-line option. This option will put pages in 134single files (per memseg list), as opposed to creating a file per page. This is 135normally not needed, but can be useful for use cases like userspace vhost, where 136there is limited number of page file descriptors that can be passed to VirtIO. 137 138If the application (or DPDK-internal code, such as device drivers) wishes to 139receive notifications about newly allocated memory, it is possible to register 140for memory event callbacks via ``rte_mem_event_callback_register()`` function. 141This will call a callback function any time DPDK's memory map has changed. 142 143If the application (or DPDK-internal code, such as device drivers) wishes to be 144notified about memory allocations above specified threshold (and have a chance 145to deny them), allocation validator callbacks are also available via 146``rte_mem_alloc_validator_callback_register()`` function. 147 148A default validator callback is provided by EAL, which can be enabled with a 149``--socket-limit`` command-line option, for a simple way to limit maximum amount 150of memory that can be used by DPDK application. 151 152.. warning:: 153 Memory subsystem uses DPDK IPC internally, so memory allocations/callbacks 154 and IPC must not be mixed: it is not safe to allocate/free memory inside 155 memory-related or IPC callbacks, and it is not safe to use IPC inside 156 memory-related callbacks. See chapter 157 :ref:`Multi-process Support <Multi-process_Support>` for more details about 158 DPDK IPC. 159 160Legacy Memory Mode 161^^^^^^^^^^^^^^^^^^ 162 163This mode is enabled by specifying ``--legacy-mem`` command-line switch to the 164EAL. This switch will have no effect on FreeBSD as FreeBSD only supports 165legacy mode anyway. 166 167This mode mimics historical behavior of EAL. That is, EAL will reserve all 168memory at startup, sort all memory into large IOVA-contiguous chunks, and will 169not allow acquiring or releasing hugepages from the system at runtime. 170 171If neither ``-m`` nor ``--socket-mem`` were specified, the entire available 172hugepage memory will be preallocated. 173 174Hugepage Allocation Matching 175^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 176 177This behavior is enabled by specifying the ``--match-allocations`` command-line 178switch to the EAL. This switch is Linux-only and not supported with 179``--legacy-mem`` nor ``--no-huge``. 180 181Some applications using memory event callbacks may require that hugepages be 182freed exactly as they were allocated. These applications may also require 183that any allocation from the malloc heap not span across allocations 184associated with two different memory event callbacks. Hugepage allocation 185matching can be used by these types of applications to satisfy both of these 186requirements. This can result in some increased memory usage which is 187very dependent on the memory allocation patterns of the application. 188 18932-bit Support 190^^^^^^^^^^^^^^ 191 192Additional restrictions are present when running in 32-bit mode. In dynamic 193memory mode, by default maximum of 2 gigabytes of VA space will be preallocated, 194and all of it will be on main lcore NUMA node unless ``--socket-mem`` flag is 195used. 196 197In legacy mode, VA space will only be preallocated for segments that were 198requested (plus padding, to keep IOVA-contiguousness). 199 200Maximum Amount of Memory 201^^^^^^^^^^^^^^^^^^^^^^^^ 202 203All possible virtual memory space that can ever be used for hugepage mapping in 204a DPDK process is preallocated at startup, thereby placing an upper limit on how 205much memory a DPDK application can have. DPDK memory is stored in segment lists, 206each segment is strictly one physical page. It is possible to change the amount 207of virtual memory being preallocated at startup by editing the following config 208variables: 209 210* ``RTE_MAX_MEMSEG_LISTS`` controls how many segment lists can DPDK have 211* ``RTE_MAX_MEM_MB_PER_LIST`` controls how much megabytes of memory each 212 segment list can address 213* ``RTE_MAX_MEMSEG_PER_LIST`` controls how many segments each segment list 214 can have 215* ``RTE_MAX_MEMSEG_PER_TYPE`` controls how many segments each memory type 216 can have (where "type" is defined as "page size + NUMA node" combination) 217* ``RTE_MAX_MEM_MB_PER_TYPE`` controls how much megabytes of memory each 218 memory type can address 219* ``RTE_MAX_MEM_MB`` places a global maximum on the amount of memory 220 DPDK can reserve 221 222Normally, these options do not need to be changed. 223 224.. note:: 225 226 Preallocated virtual memory is not to be confused with preallocated hugepage 227 memory! All DPDK processes preallocate virtual memory at startup. Hugepages 228 can later be mapped into that preallocated VA space (if dynamic memory mode 229 is enabled), and can optionally be mapped into it at startup. 230 231Hugepage Mapping 232^^^^^^^^^^^^^^^^ 233 234Below is an overview of methods used for each OS to obtain hugepages, 235explaining why certain limitations and options exist in EAL. 236See the user guide for a specific OS for configuration details. 237 238FreeBSD uses ``contigmem`` kernel module 239to reserve a fixed number of hugepages at system start, 240which are mapped by EAL at initialization using a specific ``sysctl()``. 241 242Windows EAL allocates hugepages from the OS as needed using Win32 API, 243so available amount depends on the system load. 244It uses ``virt2phys`` kernel module to obtain physical addresses, 245unless running in IOVA-as-VA mode (e.g. forced with ``--iova-mode=va``). 246 247Linux allows to select any combination of the following: 248 249* use files in hugetlbfs (the default) 250 or anonymous mappings (``--in-memory``); 251* map each hugepage from its own file (the default) 252 or map multiple hugepages from one big file (``--single-file-segments``). 253 254Mapping hugepages from files in hugetlbfs is essential for multi-process, 255because secondary processes need to map the same hugepages. 256EAL creates files like ``rtemap_0`` 257in directories specified with ``--huge-dir`` option 258(or in the mount point for a specific hugepage size). 259The ``rte`` prefix can be changed using ``--file-prefix``. 260This may be needed for running multiple primary processes 261that share a hugetlbfs mount point. 262Each backing file by default corresponds to one hugepage, 263it is opened and locked for the entire time the hugepage is used. 264This may exhaust the number of open files limit (``NOFILE``). 265See :ref:`segment-file-descriptors` section 266on how the number of open backing file descriptors can be reduced. 267 268In dynamic memory mode, EAL removes a backing hugepage file 269when all pages mapped from it are freed back to the system. 270However, backing files may persist after the application terminates 271in case of a crash or a leak of DPDK memory (e.g. ``rte_free()`` is missing). 272This reduces the number of hugepages available to other processes 273as reported by ``/sys/kernel/mm/hugepages/hugepages-*/free_hugepages``. 274EAL can remove the backing files after opening them for mapping 275if ``--huge-unlink`` is given to avoid polluting hugetlbfs. 276However, since it disables multi-process anyway, 277using anonymous mapping (``--in-memory``) is recommended instead. 278 279:ref:`EAL memory allocator <malloc>` relies on hugepages being zero-filled. 280Hugepages are cleared by the kernel when a file in hugetlbfs or its part 281is mapped for the first time system-wide 282to prevent data leaks from previous users of the same hugepage. 283EAL ensures this behavior by removing existing backing files at startup 284and by recreating them before opening for mapping (as a precaution). 285 286One exception is ``--huge-unlink=never`` mode. 287It is used to speed up EAL initialization, usually on application restart. 288Clearing memory constitutes more than 95% of hugepage mapping time. 289EAL can save it by remapping existing backing files 290with all the data left in the mapped hugepages ("dirty" memory). 291Such segments are marked with ``RTE_MEMSEG_FLAG_DIRTY``. 292Memory allocator detects dirty segments and handles them accordingly, 293in particular, it clears memory requested with ``rte_zmalloc*()``. 294In this mode EAL also does not remove a backing file 295when all pages mapped from it are freed, 296because they are intended to be reusable at restart. 297 298Anonymous mapping does not allow multi-process architecture. 299This mode does not use hugetlbfs 300and thus does not require root permissions for memory management 301(the limit of locked memory amount, ``MEMLOCK``, still applies). 302It is free of filename conflict and leftover file issues. 303If ``memfd_create(2)`` is supported both at build and run time, 304DPDK memory manager can provide file descriptors for memory segments, 305which are required for VirtIO with vhost-user backend. 306This can exhaust the number of open files limit (``NOFILE``) 307despite not creating any visible files. 308See :ref:`segment-file-descriptors` section 309on how the number of open file descriptors used by EAL can be reduced. 310 311.. _segment-file-descriptors: 312 313Segment File Descriptors 314^^^^^^^^^^^^^^^^^^^^^^^^ 315 316On Linux, in most cases, EAL will store segment file descriptors in EAL. This 317can become a problem when using smaller page sizes due to underlying limitations 318of ``glibc`` library. For example, Linux API calls such as ``select()`` may not 319work correctly because ``glibc`` does not support more than certain number of 320file descriptors. 321 322There are two possible solutions for this problem. The recommended solution is 323to use ``--single-file-segments`` mode, as that mode will not use a file 324descriptor per each page, and it will keep compatibility with Virtio with 325vhost-user backend. This option is not available when using ``--legacy-mem`` 326mode. 327 328Another option is to use bigger page sizes. Since fewer pages are required to 329cover the same memory area, fewer file descriptors will be stored internally 330by EAL. 331 332Support for Externally Allocated Memory 333~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 334 335It is possible to use externally allocated memory in DPDK. There are two ways in 336which using externally allocated memory can work: the malloc heap API's, and 337manual memory management. 338 339+ Using heap API's for externally allocated memory 340 341Using a set of malloc heap API's is the recommended way to use externally 342allocated memory in DPDK. In this way, support for externally allocated memory 343is implemented through overloading the socket ID - externally allocated heaps 344will have socket ID's that would be considered invalid under normal 345circumstances. Requesting an allocation to take place from a specified 346externally allocated memory is a matter of supplying the correct socket ID to 347DPDK allocator, either directly (e.g. through a call to ``rte_malloc``) or 348indirectly (through data structure-specific allocation API's such as 349``rte_ring_create``). Using these API's also ensures that mapping of externally 350allocated memory for DMA is also performed on any memory segment that is added 351to a DPDK malloc heap. 352 353Since there is no way DPDK can verify whether memory is available or valid, this 354responsibility falls on the shoulders of the user. All multiprocess 355synchronization is also user's responsibility, as well as ensuring that all 356calls to add/attach/detach/remove memory are done in the correct order. It is 357not required to attach to a memory area in all processes - only attach to memory 358areas as needed. 359 360The expected workflow is as follows: 361 362* Get a pointer to memory area 363* Create a named heap 364* Add memory area(s) to the heap 365 - If IOVA table is not specified, IOVA addresses will be assumed to be 366 unavailable, and DMA mappings will not be performed 367 - Other processes must attach to the memory area before they can use it 368* Get socket ID used for the heap 369* Use normal DPDK allocation procedures, using supplied socket ID 370* If memory area is no longer needed, it can be removed from the heap 371 - Other processes must detach from this memory area before it can be removed 372* If heap is no longer needed, remove it 373 - Socket ID will become invalid and will not be reused 374 375For more information, please refer to ``rte_malloc`` API documentation, 376specifically the ``rte_malloc_heap_*`` family of function calls. 377 378+ Using externally allocated memory without DPDK API's 379 380While using heap API's is the recommended method of using externally allocated 381memory in DPDK, there are certain use cases where the overhead of DPDK heap API 382is undesirable - for example, when manual memory management is performed on an 383externally allocated area. To support use cases where externally allocated 384memory will not be used as part of normal DPDK workflow, there is also another 385set of API's under the ``rte_extmem_*`` namespace. 386 387These API's are (as their name implies) intended to allow registering or 388unregistering externally allocated memory to/from DPDK's internal page table, to 389allow API's like ``rte_mem_virt2memseg`` etc. to work with externally allocated 390memory. Memory added this way will not be available for any regular DPDK 391allocators; DPDK will leave this memory for the user application to manage. 392 393The expected workflow is as follows: 394 395* Get a pointer to memory area 396* Register memory within DPDK 397 - If IOVA table is not specified, IOVA addresses will be assumed to be 398 unavailable 399 - Other processes must attach to the memory area before they can use it 400* Perform DMA mapping with ``rte_dev_dma_map`` if needed 401* Use the memory area in your application 402* If memory area is no longer needed, it can be unregistered 403 - If the area was mapped for DMA, unmapping must be performed before 404 unregistering memory 405 - Other processes must detach from the memory area before it can be 406 unregistered 407 408Since these externally allocated memory areas will not be managed by DPDK, it is 409therefore up to the user application to decide how to use them and what to do 410with them once they're registered. 411 412Per-lcore and Shared Variables 413~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 414 415.. note:: 416 417 lcore refers to a logical execution unit of the processor, sometimes called a hardware *thread*. 418 419Shared variables are the default behavior. 420Per-lcore variables are implemented using *Thread Local Storage* (TLS) to provide per-thread local storage. 421 422Logs 423~~~~ 424 425A logging API is provided by EAL. 426By default, in a Linux application, logs are sent to syslog and also to the console. 427However, the log function can be overridden by the user to use a different logging mechanism. 428 429Trace and Debug Functions 430^^^^^^^^^^^^^^^^^^^^^^^^^ 431 432There are some debug functions to dump the stack in glibc. 433The rte_panic() function can voluntarily provoke a SIG_ABORT, 434which can trigger the generation of a core file, readable by gdb. 435 436CPU Feature Identification 437~~~~~~~~~~~~~~~~~~~~~~~~~~ 438 439The EAL can query the CPU at runtime (using the rte_cpu_get_features() function) to determine which CPU features are available. 440 441User Space Interrupt Event 442~~~~~~~~~~~~~~~~~~~~~~~~~~ 443 444+ User Space Interrupt and Alarm Handling in Host Thread 445 446The EAL creates a host thread to poll the UIO device file descriptors to detect the interrupts. 447Callbacks can be registered or unregistered by the EAL functions for a specific interrupt event 448and are called in the host thread asynchronously. 449The EAL also allows timed callbacks to be used in the same way as for NIC interrupts. 450 451.. note:: 452 453 In DPDK PMD, the only interrupts handled by the dedicated host thread are those for link status change 454 (link up and link down notification) and for sudden device removal. 455 456 457+ RX Interrupt Event 458 459The receive and transmit routines provided by each PMD don't limit themselves to execute in polling thread mode. 460To ease the idle polling with tiny throughput, it's useful to pause the polling and wait until the wake-up event happens. 461The RX interrupt is the first choice to be such kind of wake-up event, but probably won't be the only one. 462 463EAL provides the event APIs for this event-driven thread mode. 464Taking Linux as an example, the implementation relies on epoll. Each thread can monitor an epoll instance 465in which all the wake-up events' file descriptors are added. The event file descriptors are created and mapped to 466the interrupt vectors according to the UIO/VFIO spec. 467From FreeBSD's perspective, kqueue is the alternative way, but not implemented yet. 468 469EAL initializes the mapping between event file descriptors and interrupt vectors, while each device initializes the mapping 470between interrupt vectors and queues. In this way, EAL actually is unaware of the interrupt cause on the specific vector. 471The eth_dev driver takes responsibility to program the latter mapping. 472 473.. note:: 474 475 Per queue RX interrupt event is only allowed in VFIO which supports multiple MSI-X vector. In UIO, the RX interrupt 476 together with other interrupt causes shares the same vector. In this case, when RX interrupt and LSC(link status change) 477 interrupt are both enabled(intr_conf.lsc == 1 && intr_conf.rxq == 1), only the former is capable. 478 479The RX interrupt are controlled/enabled/disabled by ethdev APIs - 'rte_eth_dev_rx_intr_*'. They return failure if the PMD 480hasn't support them yet. The intr_conf.rxq flag is used to turn on the capability of RX interrupt per device. 481 482+ Device Removal Event 483 484This event is triggered by a device being removed at a bus level. Its 485underlying resources may have been made unavailable (i.e. PCI mappings 486unmapped). The PMD must make sure that on such occurrence, the application can 487still safely use its callbacks. 488 489This event can be subscribed to in the same way one would subscribe to a link 490status change event. The execution context is thus the same, i.e. it is the 491dedicated interrupt host thread. 492 493Considering this, it is likely that an application would want to close a 494device having emitted a Device Removal Event. In such case, calling 495``rte_eth_dev_close()`` can trigger it to unregister its own Device Removal Event 496callback. Care must be taken not to close the device from the interrupt handler 497context. It is necessary to reschedule such closing operation. 498 499Block list 500~~~~~~~~~~ 501 502The EAL PCI device block list functionality can be used to mark certain NIC ports as unavailable, 503so they are ignored by the DPDK. 504The ports to be blocked are identified using the PCIe* description (Domain:Bus:Device.Function). 505 506Misc Functions 507~~~~~~~~~~~~~~ 508 509Locks and atomic operations are per-architecture (i686 and x86_64). 510 511IOVA Mode Detection 512~~~~~~~~~~~~~~~~~~~ 513 514IOVA Mode is selected by considering what the current usable Devices on the 515system require and/or support. 516 517On FreeBSD, RTE_IOVA_PA is always the default. On Linux, the IOVA mode is 518detected based on a 2-step heuristic detailed below. 519 520For the first step, EAL asks each bus its requirement in terms of IOVA mode 521and decides on a preferred IOVA mode. 522 523- if all buses report RTE_IOVA_PA, then the preferred IOVA mode is RTE_IOVA_PA, 524- if all buses report RTE_IOVA_VA, then the preferred IOVA mode is RTE_IOVA_VA, 525- if all buses report RTE_IOVA_DC, no bus expressed a preference, then the 526 preferred mode is RTE_IOVA_DC, 527- if the buses disagree (at least one wants RTE_IOVA_PA and at least one wants 528 RTE_IOVA_VA), then the preferred IOVA mode is RTE_IOVA_DC (see below with the 529 check on Physical Addresses availability), 530 531If the buses have expressed no preference on which IOVA mode to pick, then a 532default is selected using the following logic: 533 534- if physical addresses are not available, RTE_IOVA_VA mode is used 535- if /sys/kernel/iommu_groups is not empty, RTE_IOVA_VA mode is used 536- otherwise, RTE_IOVA_PA mode is used 537 538In the case when the buses had disagreed on their preferred IOVA mode, part of 539the buses won't work because of this decision. 540 541The second step checks if the preferred mode complies with the Physical 542Addresses availability since those are only available to root user in recent 543kernels. Namely, if the preferred mode is RTE_IOVA_PA but there is no access to 544Physical Addresses, then EAL init fails early, since later probing of the 545devices would fail anyway. 546 547.. note:: 548 549 The RTE_IOVA_VA mode is preferred as the default in most cases for the 550 following reasons: 551 552 - All drivers are expected to work in RTE_IOVA_VA mode, irrespective of 553 physical address availability. 554 - By default, the mempool, first asks for IOVA-contiguous memory using 555 ``RTE_MEMZONE_IOVA_CONTIG``. This is slow in RTE_IOVA_PA mode and it may 556 affect the application boot time. 557 - It is easy to enable large amount of IOVA-contiguous memory use cases 558 with IOVA in VA mode. 559 560 It is expected that all PCI drivers work in both RTE_IOVA_PA and 561 RTE_IOVA_VA modes. 562 563 If a PCI driver does not support RTE_IOVA_PA mode, the 564 ``RTE_PCI_DRV_NEED_IOVA_AS_VA`` flag is used to dictate that this PCI 565 driver can only work in RTE_IOVA_VA mode. 566 567 When the KNI kernel module is detected, RTE_IOVA_PA mode is preferred as a 568 performance penalty is expected in RTE_IOVA_VA mode. 569 570IOVA Mode Configuration 571~~~~~~~~~~~~~~~~~~~~~~~ 572 573Auto detection of the IOVA mode, based on probing the bus and IOMMU configuration, may not report 574the desired addressing mode when virtual devices that are not directly attached to the bus are present. 575To facilitate forcing the IOVA mode to a specific value the EAL command line option ``--iova-mode`` can 576be used to select either physical addressing('pa') or virtual addressing('va'). 577 578.. _max_simd_bitwidth: 579 580 581Max SIMD bitwidth 582~~~~~~~~~~~~~~~~~ 583 584The EAL provides a single setting to limit the max SIMD bitwidth used by DPDK, 585which is used in determining the vector path, if any, chosen by a component. 586The value can be set at runtime by an application using the 587'rte_vect_set_max_simd_bitwidth(uint16_t bitwidth)' function, 588which should only be called once at initialization, before EAL init. 589The value can be overridden by the user using the EAL command-line option '--force-max-simd-bitwidth'. 590 591When choosing a vector path, along with checking the CPU feature support, 592the value of the max SIMD bitwidth must also be checked, and can be retrieved using the 593'rte_vect_get_max_simd_bitwidth()' function. 594The value should be compared against the enum values for accepted max SIMD bitwidths: 595 596.. code-block:: c 597 598 enum rte_vect_max_simd { 599 RTE_VECT_SIMD_DISABLED = 64, 600 RTE_VECT_SIMD_128 = 128, 601 RTE_VECT_SIMD_256 = 256, 602 RTE_VECT_SIMD_512 = 512, 603 RTE_VECT_SIMD_MAX = INT16_MAX + 1, 604 }; 605 606 if (rte_vect_get_max_simd_bitwidth() >= RTE_VECT_SIMD_512) 607 /* Take AVX-512 vector path */ 608 else if (rte_vect_get_max_simd_bitwidth() >= RTE_VECT_SIMD_256) 609 /* Take AVX2 vector path */ 610 611 612Memory Segments and Memory Zones (memzone) 613------------------------------------------ 614 615The mapping of physical memory is provided by this feature in the EAL. 616As physical memory can have gaps, the memory is described in a table of descriptors, 617and each descriptor (called rte_memseg ) describes a physical page. 618 619On top of this, the memzone allocator's role is to reserve contiguous portions of physical memory. 620These zones are identified by a unique name when the memory is reserved. 621 622The rte_memzone descriptors are also located in the configuration structure. 623This structure is accessed using rte_eal_get_configuration(). 624The lookup (by name) of a memory zone returns a descriptor containing the physical address of the memory zone. 625 626Memory zones can be reserved with specific start address alignment by supplying the align parameter 627(by default, they are aligned to cache line size). 628The alignment value should be a power of two and not less than the cache line size (64 bytes). 629Memory zones can also be reserved from either 2 MB or 1 GB hugepages, provided that both are available on the system. 630 631Both memsegs and memzones are stored using ``rte_fbarray`` structures. Please 632refer to *DPDK API Reference* for more information. 633 634 635Multiple pthread 636---------------- 637 638DPDK usually pins one pthread per core to avoid the overhead of task switching. 639This allows for significant performance gains, but lacks flexibility and is not always efficient. 640 641Power management helps to improve the CPU efficiency by limiting the CPU runtime frequency. 642However, alternately it is possible to utilize the idle cycles available to take advantage of 643the full capability of the CPU. 644 645By taking advantage of cgroup, the CPU utilization quota can be simply assigned. 646This gives another way to improve the CPU efficiency, however, there is a prerequisite; 647DPDK must handle the context switching between multiple pthreads per core. 648 649For further flexibility, it is useful to set pthread affinity not only to a CPU but to a CPU set. 650 651EAL pthread and lcore Affinity 652~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 653 654The term "lcore" refers to an EAL thread, which is really a Linux/FreeBSD pthread. 655"EAL pthreads" are created and managed by EAL and execute the tasks issued by *remote_launch*. 656In each EAL pthread, there is a TLS (Thread Local Storage) called *_lcore_id* for unique identification. 657As EAL pthreads usually bind 1:1 to the physical CPU, the *_lcore_id* is typically equal to the CPU ID. 658 659When using multiple pthreads, however, the binding is no longer always 1:1 between an EAL pthread and a specified physical CPU. 660The EAL pthread may have affinity to a CPU set, and as such the *_lcore_id* will not be the same as the CPU ID. 661For this reason, there is an EAL long option '--lcores' defined to assign the CPU affinity of lcores. 662For a specified lcore ID or ID group, the option allows setting the CPU set for that EAL pthread. 663 664The format pattern: 665 --lcores='<lcore_set>[@cpu_set][,<lcore_set>[@cpu_set],...]' 666 667'lcore_set' and 'cpu_set' can be a single number, range or a group. 668 669A number is a "digit([0-9]+)"; a range is "<number>-<number>"; a group is "(<number|range>[,<number|range>,...])". 670 671If a '\@cpu_set' value is not supplied, the value of 'cpu_set' will default to the value of 'lcore_set'. 672 673 :: 674 675 For example, "--lcores='1,2@(5-7),(3-5)@(0,2),(0,6),7-8'" which means start 9 EAL thread; 676 lcore 0 runs on cpuset 0x41 (cpu 0,6); 677 lcore 1 runs on cpuset 0x2 (cpu 1); 678 lcore 2 runs on cpuset 0xe0 (cpu 5,6,7); 679 lcore 3,4,5 runs on cpuset 0x5 (cpu 0,2); 680 lcore 6 runs on cpuset 0x41 (cpu 0,6); 681 lcore 7 runs on cpuset 0x80 (cpu 7); 682 lcore 8 runs on cpuset 0x100 (cpu 8). 683 684Using this option, for each given lcore ID, the associated CPUs can be assigned. 685It's also compatible with the pattern of corelist('-l') option. 686 687non-EAL pthread support 688~~~~~~~~~~~~~~~~~~~~~~~ 689 690It is possible to use the DPDK execution context with any user pthread (aka. non-EAL pthreads). 691There are two kinds of non-EAL pthreads: 692 693- a registered non-EAL pthread with a valid *_lcore_id* that was successfully assigned by calling ``rte_thread_register()``, 694- a non registered non-EAL pthread with a LCORE_ID_ANY, 695 696For non registered non-EAL pthread (with a LCORE_ID_ANY *_lcore_id*), some libraries will use an alternative unique ID (e.g. TID), some will not be impacted at all, and some will work but with limitations (e.g. timer and mempool libraries). 697 698All these impacts are mentioned in :ref:`known_issue_label` section. 699 700Public Thread API 701~~~~~~~~~~~~~~~~~ 702 703There are two public APIs ``rte_thread_set_affinity()`` and ``rte_thread_get_affinity()`` introduced for threads. 704When they're used in any pthread context, the Thread Local Storage(TLS) will be set/get. 705 706Those TLS include *_cpuset* and *_socket_id*: 707 708* *_cpuset* stores the CPUs bitmap to which the pthread is affinitized. 709 710* *_socket_id* stores the NUMA node of the CPU set. If the CPUs in CPU set belong to different NUMA node, the *_socket_id* will be set to SOCKET_ID_ANY. 711 712 713Control Thread API 714~~~~~~~~~~~~~~~~~~ 715 716It is possible to create Control Threads using the public API 717``rte_ctrl_thread_create()``. 718Those threads can be used for management/infrastructure tasks and are used 719internally by DPDK for multi process support and interrupt handling. 720 721Those threads will be scheduled on CPUs part of the original process CPU 722affinity from which the dataplane and service lcores are excluded. 723 724For example, on a 8 CPUs system, starting a dpdk application with -l 2,3 725(dataplane cores), then depending on the affinity configuration which can be 726controlled with tools like taskset (Linux) or cpuset (FreeBSD), 727 728- with no affinity configuration, the Control Threads will end up on 729 0-1,4-7 CPUs. 730- with affinity restricted to 2-4, the Control Threads will end up on 731 CPU 4. 732- with affinity restricted to 2-3, the Control Threads will end up on 733 CPU 2 (main lcore, which is the default when no CPU is available). 734 735.. _known_issue_label: 736 737Known Issues 738~~~~~~~~~~~~ 739 740+ rte_mempool 741 742 The rte_mempool uses a per-lcore cache inside the mempool. 743 For unregistered non-EAL pthreads, ``rte_lcore_id()`` will not return a valid number. 744 So for now, when rte_mempool is used with unregistered non-EAL pthreads, the put/get operations will bypass the default mempool cache and there is a performance penalty because of this bypass. 745 Only user-owned external caches can be used in an unregistered non-EAL context in conjunction with ``rte_mempool_generic_put()`` and ``rte_mempool_generic_get()`` that accept an explicit cache parameter. 746 747+ rte_ring 748 749 rte_ring supports multi-producer enqueue and multi-consumer dequeue. 750 However, it is non-preemptive, this has a knock on effect of making rte_mempool non-preemptible. 751 752 .. note:: 753 754 The "non-preemptive" constraint means: 755 756 - a pthread doing multi-producers enqueues on a given ring must not 757 be preempted by another pthread doing a multi-producer enqueue on 758 the same ring. 759 - a pthread doing multi-consumers dequeues on a given ring must not 760 be preempted by another pthread doing a multi-consumer dequeue on 761 the same ring. 762 763 Bypassing this constraint may cause the 2nd pthread to spin until the 1st one is scheduled again. 764 Moreover, if the 1st pthread is preempted by a context that has an higher priority, it may even cause a dead lock. 765 766 This means, use cases involving preemptible pthreads should consider using rte_ring carefully. 767 768 1. It CAN be used for preemptible single-producer and single-consumer use case. 769 770 2. It CAN be used for non-preemptible multi-producer and preemptible single-consumer use case. 771 772 3. It CAN be used for preemptible single-producer and non-preemptible multi-consumer use case. 773 774 4. It MAY be used by preemptible multi-producer and/or preemptible multi-consumer pthreads whose scheduling policy are all SCHED_OTHER(cfs), SCHED_IDLE or SCHED_BATCH. User SHOULD be aware of the performance penalty before using it. 775 776 5. It MUST not be used by multi-producer/consumer pthreads, whose scheduling policies are SCHED_FIFO or SCHED_RR. 777 778 Alternatively, applications can use the lock-free stack mempool handler. When 779 considering this handler, note that: 780 781 - It is currently limited to the aarch64 and x86_64 platforms, because it uses 782 an instruction (16-byte compare-and-swap) that is not yet available on other 783 platforms. 784 - It has worse average-case performance than the non-preemptive rte_ring, but 785 software caching (e.g. the mempool cache) can mitigate this by reducing the 786 number of stack accesses. 787 788+ rte_timer 789 790 Running ``rte_timer_manage()`` on an unregistered non-EAL pthread is not allowed. However, resetting/stopping the timer from a non-EAL pthread is allowed. 791 792+ rte_log 793 794 In unregistered non-EAL pthreads, there is no per thread loglevel and logtype, global loglevels are used. 795 796+ misc 797 798 The debug statistics of rte_ring, rte_mempool and rte_timer are not supported in an unregistered non-EAL pthread. 799 800cgroup control 801~~~~~~~~~~~~~~ 802 803The following is a simple example of cgroup control usage, there are two pthreads(t0 and t1) doing packet I/O on the same core ($CPU). 804We expect only 50% of CPU spend on packet IO. 805 806 .. code-block:: console 807 808 mkdir /sys/fs/cgroup/cpu/pkt_io 809 mkdir /sys/fs/cgroup/cpuset/pkt_io 810 811 echo $cpu > /sys/fs/cgroup/cpuset/cpuset.cpus 812 813 echo $t0 > /sys/fs/cgroup/cpu/pkt_io/tasks 814 echo $t0 > /sys/fs/cgroup/cpuset/pkt_io/tasks 815 816 echo $t1 > /sys/fs/cgroup/cpu/pkt_io/tasks 817 echo $t1 > /sys/fs/cgroup/cpuset/pkt_io/tasks 818 819 cd /sys/fs/cgroup/cpu/pkt_io 820 echo 100000 > pkt_io/cpu.cfs_period_us 821 echo 50000 > pkt_io/cpu.cfs_quota_us 822 823.. _malloc: 824 825Malloc 826------ 827 828The EAL provides a malloc API to allocate any-sized memory. 829 830The objective of this API is to provide malloc-like functions to allow 831allocation from hugepage memory and to facilitate application porting. 832The *DPDK API Reference* manual describes the available functions. 833 834Typically, these kinds of allocations should not be done in data plane 835processing because they are slower than pool-based allocation and make 836use of locks within the allocation and free paths. 837However, they can be used in configuration code. 838 839Refer to the rte_malloc() function description in the *DPDK API Reference* 840manual for more information. 841 842 843Alignment and NUMA Constraints 844~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 845 846The rte_malloc() takes an align argument that can be used to request a memory 847area that is aligned on a multiple of this value (which must be a power of two). 848 849On systems with NUMA support, a call to the rte_malloc() function will return 850memory that has been allocated on the NUMA socket of the core which made the call. 851A set of APIs is also provided, to allow memory to be explicitly allocated on a 852NUMA socket directly, or by allocated on the NUMA socket where another core is 853located, in the case where the memory is to be used by a logical core other than 854on the one doing the memory allocation. 855 856Use Cases 857~~~~~~~~~ 858 859This API is meant to be used by an application that requires malloc-like 860functions at initialization time. 861 862For allocating/freeing data at runtime, in the fast-path of an application, 863the memory pool library should be used instead. 864 865Internal Implementation 866~~~~~~~~~~~~~~~~~~~~~~~ 867 868Data Structures 869^^^^^^^^^^^^^^^ 870 871There are two data structure types used internally in the malloc library: 872 873* struct malloc_heap - used to track free space on a per-socket basis 874 875* struct malloc_elem - the basic element of allocation and free-space 876 tracking inside the library. 877 878Structure: malloc_heap 879"""""""""""""""""""""" 880 881The malloc_heap structure is used to manage free space on a per-socket basis. 882Internally, there is one heap structure per NUMA node, which allows us to 883allocate memory to a thread based on the NUMA node on which this thread runs. 884While this does not guarantee that the memory will be used on that NUMA node, 885it is no worse than a scheme where the memory is always allocated on a fixed 886or random node. 887 888The key fields of the heap structure and their function are described below 889(see also diagram above): 890 891* lock - the lock field is needed to synchronize access to the heap. 892 Given that the free space in the heap is tracked using a linked list, 893 we need a lock to prevent two threads manipulating the list at the same time. 894 895* free_head - this points to the first element in the list of free nodes for 896 this malloc heap. 897 898* first - this points to the first element in the heap. 899 900* last - this points to the last element in the heap. 901 902.. _figure_malloc_heap: 903 904.. figure:: img/malloc_heap.* 905 906 Example of a malloc heap and malloc elements within the malloc library 907 908 909.. _malloc_elem: 910 911Structure: malloc_elem 912"""""""""""""""""""""" 913 914The malloc_elem structure is used as a generic header structure for various 915blocks of memory. 916It is used in two different ways - all shown in the diagram above: 917 918#. As a header on a block of free or allocated memory - normal case 919 920#. As a padding header inside a block of memory 921 922The most important fields in the structure and how they are used are described below. 923 924Malloc heap is a doubly-linked list, where each element keeps track of its 925previous and next elements. Due to the fact that hugepage memory can come and 926go, neighboring malloc elements may not necessarily be adjacent in memory. 927Also, since a malloc element may span multiple pages, its contents may not 928necessarily be IOVA-contiguous either - each malloc element is only guaranteed 929to be virtually contiguous. 930 931.. note:: 932 933 If the usage of a particular field in one of the above three usages is not 934 described, the field can be assumed to have an undefined value in that 935 situation, for example, for padding headers only the "state" and "pad" 936 fields have valid values. 937 938* heap - this pointer is a reference back to the heap structure from which 939 this block was allocated. 940 It is used for normal memory blocks when they are being freed, to add the 941 newly-freed block to the heap's free-list. 942 943* prev - this pointer points to previous header element/block in memory. When 944 freeing a block, this pointer is used to reference the previous block to 945 check if that block is also free. If so, and the two blocks are immediately 946 adjacent to each other, then the two free blocks are merged to form a single 947 larger block. 948 949* next - this pointer points to next header element/block in memory. When 950 freeing a block, this pointer is used to reference the next block to check 951 if that block is also free. If so, and the two blocks are immediately 952 adjacent to each other, then the two free blocks are merged to form a single 953 larger block. 954 955* free_list - this is a structure pointing to previous and next elements in 956 this heap's free list. 957 It is only used in normal memory blocks; on ``malloc()`` to find a suitable 958 free block to allocate and on ``free()`` to add the newly freed element to 959 the free-list. 960 961* state - This field can have one of three values: ``FREE``, ``BUSY`` or 962 ``PAD``. 963 The former two are to indicate the allocation state of a normal memory block 964 and the latter is to indicate that the element structure is a dummy structure 965 at the end of the start-of-block padding, i.e. where the start of the data 966 within a block is not at the start of the block itself, due to alignment 967 constraints. 968 In that case, the pad header is used to locate the actual malloc element 969 header for the block. 970 971* dirty - this flag is only meaningful when ``state`` is ``FREE``. 972 It indicates that the content of the element is not fully zero-filled. 973 Memory from such blocks must be cleared when requested via ``rte_zmalloc*()``. 974 Dirty elements only appear with ``--huge-unlink=never``. 975 976* pad - this holds the length of the padding present at the start of the block. 977 In the case of a normal block header, it is added to the address of the end 978 of the header to give the address of the start of the data area, i.e. the 979 value passed back to the application on a malloc. 980 Within a dummy header inside the padding, this same value is stored, and is 981 subtracted from the address of the dummy header to yield the address of the 982 actual block header. 983 984* size - the size of the data block, including the header itself. 985 986Memory Allocation 987^^^^^^^^^^^^^^^^^ 988 989On EAL initialization, all preallocated memory segments are setup as part of the 990malloc heap. This setup involves placing an :ref:`element header<malloc_elem>` 991with ``FREE`` at the start of each virtually contiguous segment of memory. 992The ``FREE`` element is then added to the ``free_list`` for the malloc heap. 993 994This setup also happens whenever memory is allocated at runtime (if supported), 995in which case newly allocated pages are also added to the heap, merging with any 996adjacent free segments if there are any. 997 998When an application makes a call to a malloc-like function, the malloc function 999will first index the ``lcore_config`` structure for the calling thread, and 1000determine the NUMA node of that thread. 1001The NUMA node is used to index the array of ``malloc_heap`` structures which is 1002passed as a parameter to the ``heap_alloc()`` function, along with the 1003requested size, type, alignment and boundary parameters. 1004 1005The ``heap_alloc()`` function will scan the free_list of the heap, and attempt 1006to find a free block suitable for storing data of the requested size, with the 1007requested alignment and boundary constraints. 1008 1009When a suitable free element has been identified, the pointer to be returned 1010to the user is calculated. 1011The cache-line of memory immediately preceding this pointer is filled with a 1012struct malloc_elem header. 1013Because of alignment and boundary constraints, there could be free space at 1014the start and/or end of the element, resulting in the following behavior: 1015 1016#. Check for trailing space. 1017 If the trailing space is big enough, i.e. > 128 bytes, then the free element 1018 is split. 1019 If it is not, then we just ignore it (wasted space). 1020 1021#. Check for space at the start of the element. 1022 If the space at the start is small, i.e. <=128 bytes, then a pad header is 1023 used, and the remaining space is wasted. 1024 If, however, the remaining space is greater, then the free element is split. 1025 1026The advantage of allocating the memory from the end of the existing element is 1027that no adjustment of the free list needs to take place - the existing element 1028on the free list just has its size value adjusted, and the next/previous elements 1029have their "prev"/"next" pointers redirected to the newly created element. 1030 1031In case when there is not enough memory in the heap to satisfy allocation 1032request, EAL will attempt to allocate more memory from the system (if supported) 1033and, following successful allocation, will retry reserving the memory again. In 1034a multiprocessing scenario, all primary and secondary processes will synchronize 1035their memory maps to ensure that any valid pointer to DPDK memory is guaranteed 1036to be valid at all times in all currently running processes. 1037 1038Failure to synchronize memory maps in one of the processes will cause allocation 1039to fail, even though some of the processes may have allocated the memory 1040successfully. The memory is not added to the malloc heap unless primary process 1041has ensured that all other processes have mapped this memory successfully. 1042 1043Any successful allocation event will trigger a callback, for which user 1044applications and other DPDK subsystems can register. Additionally, validation 1045callbacks will be triggered before allocation if the newly allocated memory will 1046exceed threshold set by the user, giving a chance to allow or deny allocation. 1047 1048.. note:: 1049 1050 Any allocation of new pages has to go through primary process. If the 1051 primary process is not active, no memory will be allocated even if it was 1052 theoretically possible to do so. This is because primary's process map acts 1053 as an authority on what should or should not be mapped, while each secondary 1054 process has its own, local memory map. Secondary processes do not update the 1055 shared memory map, they only copy its contents to their local memory map. 1056 1057Freeing Memory 1058^^^^^^^^^^^^^^ 1059 1060To free an area of memory, the pointer to the start of the data area is passed 1061to the free function. 1062The size of the ``malloc_elem`` structure is subtracted from this pointer to get 1063the element header for the block. 1064If this header is of type ``PAD`` then the pad length is further subtracted from 1065the pointer to get the proper element header for the entire block. 1066 1067From this element header, we get pointers to the heap from which the block was 1068allocated and to where it must be freed, as well as the pointer to the previous 1069and next elements. These next and previous elements are then checked to see if 1070they are also ``FREE`` and are immediately adjacent to the current one, and if 1071so, they are merged with the current element. This means that we can never have 1072two ``FREE`` memory blocks adjacent to one another, as they are always merged 1073into a single block. 1074 1075If deallocating pages at runtime is supported, and the free element encloses 1076one or more pages, those pages can be deallocated and be removed from the heap. 1077If DPDK was started with command-line parameters for preallocating memory 1078(``-m`` or ``--socket-mem``), then those pages that were allocated at startup 1079will not be deallocated. 1080 1081Any successful deallocation event will trigger a callback, for which user 1082applications and other DPDK subsystems can register. 1083