2 Copyright(c) 2010-2015 Intel Corporation. All rights reserved.
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions
9 * Redistributions of source code must retain the above copyright
10 notice, this list of conditions and the following disclaimer.
11 * Redistributions in binary form must reproduce the above copyright
12 notice, this list of conditions and the following disclaimer in
13 the documentation and/or other materials provided with the
15 * Neither the name of Intel Corporation nor the names of its
16 contributors may be used to endorse or promote products derived
17 from this software without specific prior written permission.
19 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
22 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
23 OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25 LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26 DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27 THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29 OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
36 The DPDK includes 1 Gigabit, 10 Gigabit and 40 Gigabit and para virtualized virtio Poll Mode Drivers.
38 A Poll Mode Driver (PMD) consists of APIs, provided through the BSD driver running in user space,
39 to configure the devices and their respective queues.
40 In addition, a PMD accesses the RX and TX descriptors directly without any interrupts
41 (with the exception of Link Status Change interrupts) to quickly receive,
42 process and deliver packets in the user's application.
43 This section describes the requirements of the PMDs,
44 their global design principles and proposes a high-level architecture and a generic external API for the Ethernet PMDs.
46 Requirements and Assumptions
47 ----------------------------
49 The DPDK environment for packet processing applications allows for two models, run-to-completion and pipe-line:
51 * In the *run-to-completion* model, a specific port's RX descriptor ring is polled for packets through an API.
52 Packets are then processed on the same core and placed on a port's TX descriptor ring through an API for transmission.
54 * In the *pipe-line* model, one core polls one or more port's RX descriptor ring through an API.
55 Packets are received and passed to another core via a ring.
56 The other core continues to process the packet which then may be placed on a port's TX descriptor ring through an API for transmission.
58 In a synchronous run-to-completion model,
59 each logical core assigned to the DPDK executes a packet processing loop that includes the following steps:
61 * Retrieve input packets through the PMD receive API
63 * Process each received packet one at a time, up to its forwarding
65 * Send pending output packets through the PMD transmit API
67 Conversely, in an asynchronous pipe-line model, some logical cores may be dedicated to the retrieval of received packets and
68 other logical cores to the processing of previously received packets.
69 Received packets are exchanged between logical cores through rings.
70 The loop for packet retrieval includes the following steps:
72 * Retrieve input packets through the PMD receive API
74 * Provide received packets to processing lcores through packet queues
76 The loop for packet processing includes the following steps:
78 * Retrieve the received packet from the packet queue
80 * Process the received packet, up to its retransmission if forwarded
82 To avoid any unnecessary interrupt processing overhead, the execution environment must not use any asynchronous notification mechanisms.
83 Whenever needed and appropriate, asynchronous communication should be introduced as much as possible through the use of rings.
85 Avoiding lock contention is a key issue in a multi-core environment.
86 To address this issue, PMDs are designed to work with per-core private resources as much as possible.
87 For example, a PMD maintains a separate transmit queue per-core, per-port, if the PMD is not ``DEV_TX_OFFLOAD_MT_LOCKFREE`` capable.
88 In the same way, every receive queue of a port is assigned to and polled by a single logical core (lcore).
90 To comply with Non-Uniform Memory Access (NUMA), memory management is designed to assign to each logical core
91 a private buffer pool in local memory to minimize remote memory access.
92 The configuration of packet buffer pools should take into account the underlying physical memory architecture in terms of DIMMS,
94 The application must ensure that appropriate parameters are given at memory pool creation time.
95 See :ref:`Mempool Library <Mempool_Library>`.
100 The API and architecture of the Ethernet* PMDs are designed with the following guidelines in mind.
102 PMDs must help global policy-oriented decisions to be enforced at the upper application level.
103 Conversely, NIC PMD functions should not impede the benefits expected by upper-level global policies,
104 or worse prevent such policies from being applied.
106 For instance, both the receive and transmit functions of a PMD have a maximum number of packets/descriptors to poll.
107 This allows a run-to-completion processing stack to statically fix or
108 to dynamically adapt its overall behavior through different global loop policies, such as:
110 * Receive, process immediately and transmit packets one at a time in a piecemeal fashion.
112 * Receive as many packets as possible, then process all received packets, transmitting them immediately.
114 * Receive a given maximum number of packets, process the received packets, accumulate them and finally send all accumulated packets to transmit.
116 To achieve optimal performance, overall software design choices and pure software optimization techniques must be considered and
117 balanced against available low-level hardware-based optimization features (CPU cache properties, bus speed, NIC PCI bandwidth, and so on).
118 The case of packet transmission is an example of this software/hardware tradeoff issue when optimizing burst-oriented network packet processing engines.
119 In the initial case, the PMD could export only an rte_eth_tx_one function to transmit one packet at a time on a given queue.
120 On top of that, one can easily build an rte_eth_tx_burst function that loops invoking the rte_eth_tx_one function to transmit several packets at a time.
121 However, an rte_eth_tx_burst function is effectively implemented by the PMD to minimize the driver-level transmit cost per packet through the following optimizations:
123 * Share among multiple packets the un-amortized cost of invoking the rte_eth_tx_one function.
125 * Enable the rte_eth_tx_burst function to take advantage of burst-oriented hardware features (prefetch data in cache, use of NIC head/tail registers)
126 to minimize the number of CPU cycles per packet, for example by avoiding unnecessary read memory accesses to ring transmit descriptors,
127 or by systematically using arrays of pointers that exactly fit cache line boundaries and sizes.
129 * Apply burst-oriented software optimization techniques to remove operations that would otherwise be unavoidable, such as ring index wrap back management.
131 Burst-oriented functions are also introduced via the API for services that are intensively used by the PMD.
132 This applies in particular to buffer allocators used to populate NIC rings, which provide functions to allocate/free several buffers at a time.
133 For example, an mbuf_multiple_alloc function returning an array of pointers to rte_mbuf buffers which speeds up the receive poll function of the PMD when
134 replenishing multiple descriptors of the receive ring.
136 Logical Cores, Memory and NIC Queues Relationships
137 --------------------------------------------------
139 The DPDK supports NUMA allowing for better performance when a processor's logical cores and interfaces utilize its local memory.
140 Therefore, mbuf allocation associated with local PCIe* interfaces should be allocated from memory pools created in the local memory.
141 The buffers should, if possible, remain on the local processor to obtain the best performance results and RX and TX buffer descriptors
142 should be populated with mbufs allocated from a mempool allocated from local memory.
144 The run-to-completion model also performs better if packet or data manipulation is in local memory instead of a remote processors memory.
145 This is also true for the pipe-line model provided all logical cores used are located on the same processor.
147 Multiple logical cores should never share receive or transmit queues for interfaces since this would require global locks and hinder performance.
149 If the PMD is ``DEV_TX_OFFLOAD_MT_LOCKFREE`` capable, multiple threads can invoke ``rte_eth_tx_burst()``
150 concurrently on the same tx queue without SW lock. This PMD feature found in some NICs and useful in the following use cases:
152 * Remove explicit spinlock in some applications where lcores are not mapped to Tx queues with 1:1 relation.
154 * In the eventdev use case, avoid dedicating a separate TX core for transmitting and thus
155 enables more scaling as all workers can send the packets.
157 See `Hardware Offload`_ for ``DEV_TX_OFFLOAD_MT_LOCKFREE`` capability probing details.
159 Device Identification and Configuration
160 ---------------------------------------
162 Device Identification
163 ~~~~~~~~~~~~~~~~~~~~~
165 Each NIC port is uniquely designated by its (bus/bridge, device, function) PCI
166 identifiers assigned by the PCI probing/enumeration function executed at DPDK initialization.
167 Based on their PCI identifier, NIC ports are assigned two other identifiers:
169 * A port index used to designate the NIC port in all functions exported by the PMD API.
171 * A port name used to designate the port in console messages, for administration or debugging purposes.
172 For ease of use, the port name includes the port index.
177 The configuration of each NIC port includes the following operations:
179 * Allocate PCI resources
181 * Reset the hardware (issue a Global Reset) to a well-known default state
183 * Set up the PHY and the link
185 * Initialize statistics counters
187 The PMD API must also export functions to start/stop the all-multicast feature of a port and functions to set/unset the port in promiscuous mode.
189 Some hardware offload features must be individually configured at port initialization through specific configuration parameters.
190 This is the case for the Receive Side Scaling (RSS) and Data Center Bridging (DCB) features for example.
192 On-the-Fly Configuration
193 ~~~~~~~~~~~~~~~~~~~~~~~~
195 All device features that can be started or stopped "on the fly" (that is, without stopping the device) do not require the PMD API to export dedicated functions for this purpose.
197 All that is required is the mapping address of the device PCI registers to implement the configuration of these features in specific functions outside of the drivers.
200 the PMD API exports a function that provides all the information associated with a device that can be used to set up a given device feature outside of the driver.
201 This includes the PCI vendor identifier, the PCI device identifier, the mapping address of the PCI device registers, and the name of the driver.
203 The main advantage of this approach is that it gives complete freedom on the choice of the API used to configure, to start, and to stop such features.
205 As an example, refer to the configuration of the IEEE1588 feature for the Intel® 82576 Gigabit Ethernet Controller and
206 the Intel® 82599 10 Gigabit Ethernet Controller controllers in the testpmd application.
208 Other features such as the L3/L4 5-Tuple packet filtering feature of a port can be configured in the same way.
209 Ethernet* flow control (pause frame) can be configured on the individual port.
210 Refer to the testpmd source code for details.
211 Also, L4 (UDP/TCP/ SCTP) checksum offload by the NIC can be enabled for an individual packet as long as the packet mbuf is set up correctly. See `Hardware Offload`_ for details.
213 Configuration of Transmit Queues
214 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
216 Each transmit queue is independently configured with the following information:
218 * The number of descriptors of the transmit ring
220 * The socket identifier used to identify the appropriate DMA memory zone from which to allocate the transmit ring in NUMA architectures
222 * The values of the Prefetch, Host and Write-Back threshold registers of the transmit queue
224 * The *minimum* transmit packets to free threshold (tx_free_thresh).
225 When the number of descriptors used to transmit packets exceeds this threshold, the network adaptor should be checked to see if it has written back descriptors.
226 A value of 0 can be passed during the TX queue configuration to indicate the default value should be used.
227 The default value for tx_free_thresh is 32.
228 This ensures that the PMD does not search for completed descriptors until at least 32 have been processed by the NIC for this queue.
230 * The *minimum* RS bit threshold. The minimum number of transmit descriptors to use before setting the Report Status (RS) bit in the transmit descriptor.
231 Note that this parameter may only be valid for Intel 10 GbE network adapters.
232 The RS bit is set on the last descriptor used to transmit a packet if the number of descriptors used since the last RS bit setting,
233 up to the first descriptor used to transmit the packet, exceeds the transmit RS bit threshold (tx_rs_thresh).
234 In short, this parameter controls which transmit descriptors are written back to host memory by the network adapter.
235 A value of 0 can be passed during the TX queue configuration to indicate that the default value should be used.
236 The default value for tx_rs_thresh is 32.
237 This ensures that at least 32 descriptors are used before the network adapter writes back the most recently used descriptor.
238 This saves upstream PCIe* bandwidth resulting from TX descriptor write-backs.
239 It is important to note that the TX Write-back threshold (TX wthresh) should be set to 0 when tx_rs_thresh is greater than 1.
240 Refer to the Intel® 82599 10 Gigabit Ethernet Controller Datasheet for more details.
242 The following constraints must be satisfied for tx_free_thresh and tx_rs_thresh:
244 * tx_rs_thresh must be greater than 0.
246 * tx_rs_thresh must be less than the size of the ring minus 2.
248 * tx_rs_thresh must be less than or equal to tx_free_thresh.
250 * tx_free_thresh must be greater than 0.
252 * tx_free_thresh must be less than the size of the ring minus 3.
254 * For optimal performance, TX wthresh should be set to 0 when tx_rs_thresh is greater than 1.
256 One descriptor in the TX ring is used as a sentinel to avoid a hardware race condition, hence the maximum threshold constraints.
260 When configuring for DCB operation, at port initialization, both the number of transmit queues and the number of receive queues must be set to 128.
262 Free Tx mbuf on Demand
263 ~~~~~~~~~~~~~~~~~~~~~~
265 Many of the drivers do not release the mbuf back to the mempool, or local cache,
266 immediately after the packet has been transmitted.
267 Instead, they leave the mbuf in their Tx ring and
268 either perform a bulk release when the ``tx_rs_thresh`` has been crossed
269 or free the mbuf when a slot in the Tx ring is needed.
271 An application can request the driver to release used mbufs with the ``rte_eth_tx_done_cleanup()`` API.
272 This API requests the driver to release mbufs that are no longer in use,
273 independent of whether or not the ``tx_rs_thresh`` has been crossed.
274 There are two scenarios when an application may want the mbuf released immediately:
276 * When a given packet needs to be sent to multiple destination interfaces
277 (either for Layer 2 flooding or Layer 3 multi-cast).
278 One option is to make a copy of the packet or a copy of the header portion that needs to be manipulated.
279 A second option is to transmit the packet and then poll the ``rte_eth_tx_done_cleanup()`` API
280 until the reference count on the packet is decremented.
281 Then the same packet can be transmitted to the next destination interface.
282 The application is still responsible for managing any packet manipulations needed
283 between the different destination interfaces, but a packet copy can be avoided.
284 This API is independent of whether the packet was transmitted or dropped,
285 only that the mbuf is no longer in use by the interface.
287 * Some applications are designed to make multiple runs, like a packet generator.
288 For performance reasons and consistency between runs,
289 the application may want to reset back to an initial state
290 between each run, where all mbufs are returned to the mempool.
291 In this case, it can call the ``rte_eth_tx_done_cleanup()`` API
292 for each destination interface it has been using
293 to request it to release of all its used mbufs.
295 To determine if a driver supports this API, check for the *Free Tx mbuf on demand* feature
296 in the *Network Interface Controller Drivers* document.
301 Depending on driver capabilities advertised by
302 ``rte_eth_dev_info_get()``, the PMD may support hardware offloading
303 feature like checksumming, TCP segmentation, VLAN insertion or
304 lockfree multithreaded TX burst on the same TX queue.
306 The support of these offload features implies the addition of dedicated
307 status bit(s) and value field(s) into the rte_mbuf data structure, along
308 with their appropriate handling by the receive/transmit functions
309 exported by each PMD. The list of flags and their precise meaning is
310 described in the mbuf API documentation and in the in :ref:`Mbuf Library
311 <Mbuf_Library>`, section "Meta Information".
313 Per-Port and Per-Queue Offloads
314 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
316 In the DPDK offload API, offloads are divided into per-port and per-queue offloads.
317 The different offloads capabilities can be queried using ``rte_eth_dev_info_get()``.
318 Supported offloads can be either per-port or per-queue.
320 Offloads are enabled using the existing ``DEV_TX_OFFLOAD_*`` or ``DEV_RX_OFFLOAD_*`` flags.
321 Per-port offload configuration is set using ``rte_eth_dev_configure``.
322 Per-queue offload configuration is set using ``rte_eth_rx_queue_setup`` and ``rte_eth_tx_queue_setup``.
323 To enable per-port offload, the offload should be set on both device configuration and queue setup.
324 In case of a mixed configuration the queue setup shall return with an error.
325 To enable per-queue offload, the offload can be set only on the queue setup.
326 Offloads which are not enabled are disabled by default.
328 For an application to use the Tx offloads API it should set the ``ETH_TXQ_FLAGS_IGNORE`` flag in the ``txq_flags`` field located in ``rte_eth_txconf`` struct.
329 In such cases it is not required to set other flags in ``txq_flags``.
330 For an application to use the Rx offloads API it should set the ``ignore_offload_bitfield`` bit in the ``rte_eth_rxmode`` struct.
331 In such cases it is not required to set other bitfield offloads in the ``rxmode`` struct.
339 By default, all functions exported by a PMD are lock-free functions that are assumed
340 not to be invoked in parallel on different logical cores to work on the same target object.
341 For instance, a PMD receive function cannot be invoked in parallel on two logical cores to poll the same RX queue of the same port.
342 Of course, this function can be invoked in parallel by different logical cores on different RX queues.
343 It is the responsibility of the upper-level application to enforce this rule.
345 If needed, parallel accesses by multiple logical cores to shared queues can be explicitly protected by dedicated inline lock-aware functions
346 built on top of their corresponding lock-free functions of the PMD API.
348 Generic Packet Representation
349 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
351 A packet is represented by an rte_mbuf structure, which is a generic metadata structure containing all necessary housekeeping information.
352 This includes fields and status bits corresponding to offload hardware features, such as checksum computation of IP headers or VLAN tags.
354 The rte_mbuf data structure includes specific fields to represent, in a generic way, the offload features provided by network controllers.
355 For an input packet, most fields of the rte_mbuf structure are filled in by the PMD receive function with the information contained in the receive descriptor.
356 Conversely, for output packets, most fields of rte_mbuf structures are used by the PMD transmit function to initialize transmit descriptors.
358 The mbuf structure is fully described in the :ref:`Mbuf Library <Mbuf_Library>` chapter.
363 The Ethernet device API exported by the Ethernet PMDs is described in the *DPDK API Reference*.
365 Extended Statistics API
366 ~~~~~~~~~~~~~~~~~~~~~~~
368 The extended statistics API allows a PMD to expose all statistics that are
369 available to it, including statistics that are unique to the device.
370 Each statistic has three properties ``name``, ``id`` and ``value``:
372 * ``name``: A human readable string formatted by the scheme detailed below.
373 * ``id``: An integer that represents only that statistic.
374 * ``value``: A unsigned 64-bit integer that is the value of the statistic.
376 Note that extended statistic identifiers are
377 driver-specific, and hence might not be the same for different ports.
378 The API consists of various ``rte_eth_xstats_*()`` functions, and allows an
379 application to be flexible in how it retrieves statistics.
381 Scheme for Human Readable Names
382 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
384 A naming scheme exists for the strings exposed to clients of the API. This is
385 to allow scraping of the API for statistics of interest. The naming scheme uses
386 strings split by a single underscore ``_``. The scheme is as follows:
394 Examples of common statistics xstats strings, formatted to comply to the scheme
399 * ``tx_multicast_packets``
401 The scheme, although quite simple, allows flexibility in presenting and reading
402 information from the statistic strings. The following example illustrates the
403 naming scheme:``rx_packets``. In this example, the string is split into two
404 components. The first component ``rx`` indicates that the statistic is
405 associated with the receive side of the NIC. The second component ``packets``
406 indicates that the unit of measure is packets.
408 A more complicated example: ``tx_size_128_to_255_packets``. In this example,
409 ``tx`` indicates transmission, ``size`` is the first detail, ``128`` etc are
410 more details, and ``packets`` indicates that this is a packet counter.
412 Some additions in the metadata scheme are as follows:
414 * If the first part does not match ``rx`` or ``tx``, the statistic does not
415 have an affinity with either receive of transmit.
417 * If the first letter of the second part is ``q`` and this ``q`` is followed
418 by a number, this statistic is part of a specific queue.
420 An example where queue numbers are used is as follows: ``tx_q7_bytes`` which
421 indicates this statistic applies to queue number 7, and represents the number
422 of transmitted bytes on that queue.
427 The xstats API uses the ``name``, ``id``, and ``value`` to allow performant
428 lookup of specific statistics. Performant lookup means two things;
430 * No string comparisons with the ``name`` of the statistic in fast-path
431 * Allow requesting of only the statistics of interest
433 The API ensures these requirements are met by mapping the ``name`` of the
434 statistic to a unique ``id``, which is used as a key for lookup in the fast-path.
435 The API allows applications to request an array of ``id`` values, so that the
436 PMD only performs the required calculations. Expected usage is that the
437 application scans the ``name`` of each statistic, and caches the ``id``
438 if it has an interest in that statistic. On the fast-path, the integer can be used
439 to retrieve the actual ``value`` of the statistic that the ``id`` represents.
444 The API is built out of a small number of functions, which can be used to
445 retrieve the number of statistics and the names, IDs and values of those
448 * ``rte_eth_xstats_get_names_by_id()``: returns the names of the statistics. When given a
449 ``NULL`` parameter the function returns the number of statistics that are available.
451 * ``rte_eth_xstats_get_id_by_name()``: Searches for the statistic ID that matches
452 ``xstat_name``. If found, the ``id`` integer is set.
454 * ``rte_eth_xstats_get_by_id()``: Fills in an array of ``uint64_t`` values
455 with matching the provided ``ids`` array. If the ``ids`` array is NULL, it
456 returns all statistics that are available.
462 Imagine an application that wants to view the dropped packet count. If no
463 packets are dropped, the application does not read any other metrics for
464 performance reasons. If packets are dropped, the application has a particular
465 set of statistics that it requests. This "set" of statistics allows the app to
466 decide what next steps to perform. The following code-snippets show how the
467 xstats API can be used to achieve this goal.
469 First step is to get all statistics names and list them:
473 struct rte_eth_xstat_name *xstats_names;
477 /* Get number of stats */
478 len = rte_eth_xstats_get_names_by_id(port_id, NULL, NULL, 0);
480 printf("Cannot get xstats count\n");
484 xstats_names = malloc(sizeof(struct rte_eth_xstat_name) * len);
485 if (xstats_names == NULL) {
486 printf("Cannot allocate memory for xstat names\n");
490 /* Retrieve xstats names, passing NULL for IDs to return all statistics */
491 if (len != rte_eth_xstats_get_names_by_id(port_id, xstats_names, NULL, len)) {
492 printf("Cannot get xstat names\n");
496 values = malloc(sizeof(values) * len);
497 if (values == NULL) {
498 printf("Cannot allocate memory for xstats\n");
502 /* Getting xstats values */
503 if (len != rte_eth_xstats_get_by_id(port_id, NULL, values, len)) {
504 printf("Cannot get xstat values\n");
508 /* Print all xstats names and values */
509 for (i = 0; i < len; i++) {
510 printf("%s: %"PRIu64"\n", xstats_names[i].name, values[i]);
513 The application has access to the names of all of the statistics that the PMD
514 exposes. The application can decide which statistics are of interest, cache the
515 ids of those statistics by looking up the name as follows:
521 const char *xstat_name = "rx_errors";
523 if(!rte_eth_xstats_get_id_by_name(port_id, xstat_name, &id)) {
524 rte_eth_xstats_get_by_id(port_id, &id, &value, 1);
525 printf("%s: %"PRIu64"\n", xstat_name, value);
528 printf("Cannot find xstats with a given name\n");
532 The API provides flexibility to the application so that it can look up multiple
533 statistics using an array containing multiple ``id`` numbers. This reduces the
534 function call overhead of retrieving statistics, and makes lookup of multiple
535 statistics simpler for the application.
539 #define APP_NUM_STATS 4
540 /* application cached these ids previously; see above */
541 uint64_t ids_array[APP_NUM_STATS] = {3,4,7,21};
542 uint64_t value_array[APP_NUM_STATS];
544 /* Getting multiple xstats values from array of IDs */
545 rte_eth_xstats_get_by_id(port_id, ids_array, value_array, APP_NUM_STATS);
548 for(i = 0; i < APP_NUM_STATS; i++) {
549 printf("%d: %"PRIu64"\n", ids_array[i], value_array[i]);
553 This array lookup API for xstats allows the application create multiple
554 "groups" of statistics, and look up the values of those IDs using a single API
555 call. As an end result, the application is able to achieve its goal of
556 monitoring a single statistic ("rx_errors" in this case), and if that shows
557 packets being dropped, it can easily retrieve a "set" of statistics using the
558 IDs array parameter to ``rte_eth_xstats_get_by_id`` function.
565 int rte_eth_dev_reset(uint8_t port_id);
567 Sometimes a port has to be reset passively. For example when a PF is
568 reset, all its VFs should also be reset by the application to make them
569 consistent with the PF. A DPDK application also can call this function
570 to trigger a port reset. Normally, a DPDK application would invokes this
571 function when an RTE_ETH_EVENT_INTR_RESET event is detected.
573 It is the duty of the PMD to trigger RTE_ETH_EVENT_INTR_RESET events and
574 the application should register a callback function to handle these
575 events. When a PMD needs to trigger a reset, it can trigger an
576 RTE_ETH_EVENT_INTR_RESET event. On receiving an RTE_ETH_EVENT_INTR_RESET
577 event, applications can handle it as follows: Stop working queues, stop
578 calling Rx and Tx functions, and then call rte_eth_dev_reset(). For
579 thread safety all these operations should be called from the same thread.
581 For example when PF is reset, the PF sends a message to notify VFs of
582 this event and also trigger an interrupt to VFs. Then in the interrupt
583 service routine the VFs detects this notification message and calls
584 _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_RESET, NULL,
585 NULL). This means that a PF reset triggers an RTE_ETH_EVENT_INTR_RESET
586 event within VFs. The function _rte_eth_dev_callback_process() will
587 call the registered callback function. The callback function can trigger
588 the application to handle all operations the VF reset requires including
589 stopping Rx/Tx queues and calling rte_eth_dev_reset().
591 The rte_eth_dev_reset() itself is a generic function which only does
592 some hardware reset operations through calling dev_unint() and
593 dev_init(), and itself does not handle synchronization, which is handled
596 The PMD itself should not call rte_eth_dev_reset(). The PMD can trigger
597 the application to handle reset event. It is duty of application to
598 handle all synchronization before it calls rte_eth_dev_reset().