1.. SPDX-License-Identifier: BSD-3-Clause 2 Copyright(c) 2010-2016 Intel Corporation. 3 4Vhost Library 5============= 6 7The vhost library implements a user space virtio net server allowing the user 8to manipulate the virtio ring directly. In another words, it allows the user 9to fetch/put packets from/to the VM virtio net device. To achieve this, a 10vhost library should be able to: 11 12* Access the guest memory: 13 14 For QEMU, this is done by using the ``-object memory-backend-file,share=on,...`` 15 option. Which means QEMU will create a file to serve as the guest RAM. 16 The ``share=on`` option allows another process to map that file, which 17 means it can access the guest RAM. 18 19* Know all the necessary information about the vring: 20 21 Information such as where the available ring is stored. Vhost defines some 22 messages (passed through a Unix domain socket file) to tell the backend all 23 the information it needs to know how to manipulate the vring. 24 25 26Vhost API Overview 27------------------ 28 29The following is an overview of some key Vhost API functions: 30 31* ``rte_vhost_driver_register(path, flags)`` 32 33 This function registers a vhost driver into the system. ``path`` specifies 34 the Unix domain socket file path. 35 36 Currently supported flags are: 37 38 - ``RTE_VHOST_USER_CLIENT`` 39 40 DPDK vhost-user will act as the client when this flag is given. See below 41 for an explanation. 42 43 - ``RTE_VHOST_USER_NO_RECONNECT`` 44 45 When DPDK vhost-user acts as the client it will keep trying to reconnect 46 to the server (QEMU) until it succeeds. This is useful in two cases: 47 48 * When QEMU is not started yet. 49 * When QEMU restarts (for example due to a guest OS reboot). 50 51 This reconnect option is enabled by default. However, it can be turned off 52 by setting this flag. 53 54 - ``RTE_VHOST_USER_IOMMU_SUPPORT`` 55 56 IOMMU support will be enabled when this flag is set. It is disabled by 57 default. 58 59 Enabling this flag makes possible to use guest vIOMMU to protect vhost 60 from accessing memory the virtio device isn't allowed to, when the feature 61 is negotiated and an IOMMU device is declared. 62 63 - ``RTE_VHOST_USER_POSTCOPY_SUPPORT`` 64 65 Postcopy live-migration support will be enabled when this flag is set. 66 It is disabled by default. 67 68 Enabling this flag should only be done when the calling application does 69 not pre-fault the guest shared memory, otherwise migration would fail. 70 71 - ``RTE_VHOST_USER_LINEARBUF_SUPPORT`` 72 73 Enabling this flag forces vhost dequeue function to only provide linear 74 pktmbuf (no multi-segmented pktmbuf). 75 76 The vhost library by default provides a single pktmbuf for given a 77 packet, but if for some reason the data doesn't fit into a single 78 pktmbuf (e.g., TSO is enabled), the library will allocate additional 79 pktmbufs from the same mempool and chain them together to create a 80 multi-segmented pktmbuf. 81 82 However, the vhost application needs to support multi-segmented format. 83 If the vhost application does not support that format and requires large 84 buffers to be dequeue, this flag should be enabled to force only linear 85 buffers (see RTE_VHOST_USER_EXTBUF_SUPPORT) or drop the packet. 86 87 It is disabled by default. 88 89 - ``RTE_VHOST_USER_EXTBUF_SUPPORT`` 90 91 Enabling this flag allows vhost dequeue function to allocate and attach 92 an external buffer to a pktmbuf if the pkmbuf doesn't provide enough 93 space to store all data. 94 95 This is useful when the vhost application wants to support large packets 96 but doesn't want to increase the default mempool object size nor to 97 support multi-segmented mbufs (non-linear). In this case, a fresh buffer 98 is allocated using rte_malloc() which gets attached to a pktmbuf using 99 rte_pktmbuf_attach_extbuf(). 100 101 See RTE_VHOST_USER_LINEARBUF_SUPPORT as well to disable multi-segmented 102 mbufs for application that doesn't support chained mbufs. 103 104 It is disabled by default. 105 106 - ``RTE_VHOST_USER_ASYNC_COPY`` 107 108 Asynchronous data path will be enabled when this flag is set. Async 109 data path allows applications to enable DMA acceleration for vhost 110 queues. Vhost leverages the registered DMA channels to free CPU from 111 memory copy operations in data path. A set of async data path APIs are 112 defined for DPDK applications to make use of the async capability. Only 113 packets enqueued/dequeued by async APIs are processed through the async 114 data path. 115 116 Currently this feature is only implemented on split ring enqueue data 117 path. 118 119 It is disabled by default. 120 121 - ``RTE_VHOST_USER_NET_COMPLIANT_OL_FLAGS`` 122 123 Since v16.04, the vhost library forwards checksum and gso requests for 124 packets received from a virtio driver by filling Tx offload metadata in 125 the mbuf. This behavior is inconsistent with other drivers but it is left 126 untouched for existing applications that might rely on it. 127 128 This flag disables the legacy behavior and instead ask vhost to simply 129 populate Rx offload metadata in the mbuf. 130 131 It is disabled by default. 132 133* ``rte_vhost_driver_set_features(path, features)`` 134 135 This function sets the feature bits the vhost-user driver supports. The 136 vhost-user driver could be vhost-user net, yet it could be something else, 137 say, vhost-user SCSI. 138 139* ``rte_vhost_driver_callback_register(path, vhost_device_ops)`` 140 141 This function registers a set of callbacks, to let DPDK applications take 142 the appropriate action when some events happen. The following events are 143 currently supported: 144 145 * ``new_device(int vid)`` 146 147 This callback is invoked when a virtio device becomes ready. ``vid`` 148 is the vhost device ID. 149 150 * ``destroy_device(int vid)`` 151 152 This callback is invoked when a virtio device is paused or shut down. 153 154 * ``vring_state_changed(int vid, uint16_t queue_id, int enable)`` 155 156 This callback is invoked when a specific queue's state is changed, for 157 example to enabled or disabled. 158 159 * ``features_changed(int vid, uint64_t features)`` 160 161 This callback is invoked when the features is changed. For example, 162 ``VHOST_F_LOG_ALL`` will be set/cleared at the start/end of live 163 migration, respectively. 164 165 * ``new_connection(int vid)`` 166 167 This callback is invoked on new vhost-user socket connection. If DPDK 168 acts as the server the device should not be deleted before 169 ``destroy_connection`` callback is received. 170 171 * ``destroy_connection(int vid)`` 172 173 This callback is invoked when vhost-user socket connection is closed. 174 It indicates that device with id ``vid`` is no longer in use and can be 175 safely deleted. 176 177* ``rte_vhost_driver_disable/enable_features(path, features))`` 178 179 This function disables/enables some features. For example, it can be used to 180 disable mergeable buffers and TSO features, which both are enabled by 181 default. 182 183* ``rte_vhost_driver_start(path)`` 184 185 This function triggers the vhost-user negotiation. It should be invoked at 186 the end of initializing a vhost-user driver. 187 188* ``rte_vhost_enqueue_burst(vid, queue_id, pkts, count)`` 189 190 Transmits (enqueues) ``count`` packets from host to guest. 191 192* ``rte_vhost_dequeue_burst(vid, queue_id, mbuf_pool, pkts, count)`` 193 194 Receives (dequeues) ``count`` packets from guest, and stored them at ``pkts``. 195 196* ``rte_vhost_crypto_create(vid, cryptodev_id, sess_mempool, socket_id)`` 197 198 As an extension of new_device(), this function adds virtio-crypto workload 199 acceleration capability to the device. All crypto workload is processed by 200 DPDK cryptodev with the device ID of ``cryptodev_id``. 201 202* ``rte_vhost_crypto_free(vid)`` 203 204 Frees the memory and vhost-user message handlers created in 205 rte_vhost_crypto_create(). 206 207* ``rte_vhost_crypto_fetch_requests(vid, queue_id, ops, nb_ops)`` 208 209 Receives (dequeues) ``nb_ops`` virtio-crypto requests from guest, parses 210 them to DPDK Crypto Operations, and fills the ``ops`` with parsing results. 211 212* ``rte_vhost_crypto_finalize_requests(queue_id, ops, nb_ops)`` 213 214 After the ``ops`` are dequeued from Cryptodev, finalizes the jobs and 215 notifies the guest(s). 216 217* ``rte_vhost_crypto_set_zero_copy(vid, option)`` 218 219 Enable or disable zero copy feature of the vhost crypto backend. 220 221* ``rte_vhost_async_dma_configure(dma_id, vchan_id)`` 222 223 Tell vhost which DMA vChannel is going to use. This function needs to 224 be called before register async data-path for vring. 225 226* ``rte_vhost_async_channel_register(vid, queue_id)`` 227 228 Register async DMA acceleration for a vhost queue after vring is enabled. 229 230* ``rte_vhost_async_channel_register_thread_unsafe(vid, queue_id)`` 231 232 Register async DMA acceleration for a vhost queue without performing 233 any locking. 234 235 This function is only safe to call in vhost callback functions 236 (i.e., struct rte_vhost_device_ops). 237 238* ``rte_vhost_async_channel_unregister(vid, queue_id)`` 239 240 Unregister the async DMA acceleration from a vhost queue. 241 Unregistration will fail, if the vhost queue has in-flight 242 packets that are not completed. 243 244 Unregister async DMA acceleration in vring_state_changed() may 245 fail, as this API tries to acquire the spinlock of vhost 246 queue. The recommended way is to unregister async copy 247 devices for all vhost queues in destroy_device(), when a 248 virtio device is paused or shut down. 249 250* ``rte_vhost_async_channel_unregister_thread_unsafe(vid, queue_id)`` 251 252 Unregister async DMA acceleration for a vhost queue without performing 253 any locking. 254 255 This function is only safe to call in vhost callback functions 256 (i.e., struct rte_vhost_device_ops). 257 258* ``rte_vhost_submit_enqueue_burst(vid, queue_id, pkts, count, dma_id, vchan_id)`` 259 260 Submit an enqueue request to transmit ``count`` packets from host to guest 261 by async data path. Applications must not free the packets submitted for 262 enqueue until the packets are completed. 263 264* ``rte_vhost_poll_enqueue_completed(vid, queue_id, pkts, count, dma_id, vchan_id)`` 265 266 Poll enqueue completion status from async data path. Completed packets 267 are returned to applications through ``pkts``. 268 269* ``rte_vhost_async_get_inflight(vid, queue_id)`` 270 271 This function returns the amount of in-flight packets for the vhost 272 queue using async acceleration. 273 274* ``rte_vhost_clear_queue_thread_unsafe(vid, queue_id, **pkts, count, dma_id, vchan_id)`` 275 276 Clear inflight packets which are submitted to DMA engine in vhost async data 277 path. Completed packets are returned to applications through ``pkts``. 278 279Vhost-user Implementations 280-------------------------- 281 282Vhost-user uses Unix domain sockets for passing messages. This means the DPDK 283vhost-user implementation has two options: 284 285* DPDK vhost-user acts as the server. 286 287 DPDK will create a Unix domain socket server file and listen for 288 connections from the frontend. 289 290 Note, this is the default mode, and the only mode before DPDK v16.07. 291 292 293* DPDK vhost-user acts as the client. 294 295 Unlike the server mode, this mode doesn't create the socket file; 296 it just tries to connect to the server (which responses to create the 297 file instead). 298 299 When the DPDK vhost-user application restarts, DPDK vhost-user will try to 300 connect to the server again. This is how the "reconnect" feature works. 301 302 .. Note:: 303 * The "reconnect" feature requires **QEMU v2.7** (or above). 304 305 * The vhost supported features must be exactly the same before and 306 after the restart. For example, if TSO is disabled and then enabled, 307 nothing will work and issues undefined might happen. 308 309No matter which mode is used, once a connection is established, DPDK 310vhost-user will start receiving and processing vhost messages from QEMU. 311 312For messages with a file descriptor, the file descriptor can be used directly 313in the vhost process as it is already installed by the Unix domain socket. 314 315The supported vhost messages are: 316 317* ``VHOST_SET_MEM_TABLE`` 318* ``VHOST_SET_VRING_KICK`` 319* ``VHOST_SET_VRING_CALL`` 320* ``VHOST_SET_LOG_FD`` 321* ``VHOST_SET_VRING_ERR`` 322 323For ``VHOST_SET_MEM_TABLE`` message, QEMU will send information for each 324memory region and its file descriptor in the ancillary data of the message. 325The file descriptor is used to map that region. 326 327``VHOST_SET_VRING_KICK`` is used as the signal to put the vhost device into 328the data plane, and ``VHOST_GET_VRING_BASE`` is used as the signal to remove 329the vhost device from the data plane. 330 331When the socket connection is closed, vhost will destroy the device. 332 333Guest memory requirement 334------------------------ 335 336* Memory pre-allocation 337 338 For non-async data path, guest memory pre-allocation is not a 339 must. This can help save of memory. If users really want the guest memory 340 to be pre-allocated (e.g., for performance reason), we can add option 341 ``-mem-prealloc`` when starting QEMU. Or, we can lock all memory at vhost 342 side which will force memory to be allocated when mmap at vhost side; 343 option --mlockall in ovs-dpdk is an example in hand. 344 345 For async data path, we force the VM memory to be pre-allocated at vhost 346 lib when mapping the guest memory; and also we need to lock the memory to 347 prevent pages being swapped out to disk. 348 349* Memory sharing 350 351 Make sure ``share=on`` QEMU option is given. vhost-user will not work with 352 a QEMU version without shared memory mapping. 353 354Vhost supported vSwitch reference 355--------------------------------- 356 357For more vhost details and how to support vhost in vSwitch, please refer to 358the vhost example in the DPDK Sample Applications Guide. 359 360Vhost data path acceleration (vDPA) 361----------------------------------- 362 363vDPA supports selective datapath in vhost-user lib by enabling virtio ring 364compatible devices to serve virtio driver directly for datapath acceleration. 365 366``rte_vhost_driver_attach_vdpa_device`` is used to configure the vhost device 367with accelerated backend. 368 369Also vhost device capabilities are made configurable to adopt various devices. 370Such capabilities include supported features, protocol features, queue number. 371 372Finally, a set of device ops is defined for device specific operations: 373 374* ``get_queue_num`` 375 376 Called to get supported queue number of the device. 377 378* ``get_features`` 379 380 Called to get supported features of the device. 381 382* ``get_protocol_features`` 383 384 Called to get supported protocol features of the device. 385 386* ``dev_conf`` 387 388 Called to configure the actual device when the virtio device becomes ready. 389 390* ``dev_close`` 391 392 Called to close the actual device when the virtio device is stopped. 393 394* ``set_vring_state`` 395 396 Called to change the state of the vring in the actual device when vring state 397 changes. 398 399* ``set_features`` 400 401 Called to set the negotiated features to device. 402 403* ``migration_done`` 404 405 Called to allow the device to response to RARP sending. 406 407* ``get_vfio_group_fd`` 408 409 Called to get the VFIO group fd of the device. 410 411* ``get_vfio_device_fd`` 412 413 Called to get the VFIO device fd of the device. 414 415* ``get_notify_area`` 416 417 Called to get the notify area info of the queue. 418 419Vhost asynchronous data path 420---------------------------- 421 422Vhost asynchronous data path leverages DMA devices to offload memory 423copies from the CPU and it is implemented in an asynchronous way. It 424enables applications, like OVS, to save CPU cycles and hide memory copy 425overhead, thus achieving higher throughput. 426 427Vhost doesn't manage DMA devices and applications, like OVS, need to 428manage and configure DMA devices. Applications need to tell vhost what 429DMA devices to use in every data path function call. This design enables 430the flexibility for applications to dynamically use DMA channels in 431different function modules, not limited in vhost. 432 433In addition, vhost supports M:N mapping between vrings and DMA virtual 434channels. Specifically, one vring can use multiple different DMA channels 435and one DMA channel can be shared by multiple vrings at the same time. 436The reason of enabling one vring to use multiple DMA channels is that 437it's possible that more than one dataplane threads enqueue packets to 438the same vring with their own DMA virtual channels. Besides, the number 439of DMA devices is limited. For the purpose of scaling, it's necessary to 440support sharing DMA channels among vrings. 441 442Recommended IOVA mode in async datapath 443--------------------------------------- 444 445When DMA devices are bound to VFIO driver, VA mode is recommended. 446For PA mode, page by page mapping may exceed IOMMU's max capability, 447better to use 1G guest hugepage. 448 449For UIO driver, any VFIO related error message can be ignored. 450