4 * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
11 * * Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * * Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in
15 * the documentation and/or other materials provided with the
17 * * Neither the name of Intel Corporation nor the names of its
18 * contributors may be used to endorse or promote products derived
19 * from this software without specific prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
22 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
23 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
24 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
25 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
26 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
27 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
28 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
29 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
30 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
31 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
34 #ifndef _VIRTIO_NET_H_
35 #define _VIRTIO_NET_H_
39 * Interface to vhost net
43 #include <linux/vhost.h>
44 #include <linux/virtio_ring.h>
45 #include <linux/virtio_net.h>
46 #include <sys/eventfd.h>
47 #include <sys/socket.h>
50 #include <rte_memory.h>
51 #include <rte_mempool.h>
55 #define VHOST_MEMORY_MAX_NREGIONS 8
57 /* Used to indicate that the device is running on a data core */
58 #define VIRTIO_DEV_RUNNING 1
60 /* Backend value set by guest. */
61 #define VIRTIO_DEV_STOPPED -1
64 /* Enum for virtqueue management. */
65 enum {VIRTIO_RXQ, VIRTIO_TXQ, VIRTIO_QNUM};
67 #define BUF_VECTOR_MAX 256
70 * Structure contains buffer address, length and descriptor index
71 * from vring to do scatter RX.
80 * Structure contains variables relevant to RX/TX virtqueues.
82 struct vhost_virtqueue {
83 struct vring_desc *desc; /**< Virtqueue descriptor ring. */
84 struct vring_avail *avail; /**< Virtqueue available ring. */
85 struct vring_used *used; /**< Virtqueue used ring. */
86 uint32_t size; /**< Size of descriptor ring. */
87 uint32_t backend; /**< Backend value to determine if device should started/stopped. */
88 uint16_t vhost_hlen; /**< Vhost header length (varies depending on RX merge buffers. */
89 volatile uint16_t last_used_idx; /**< Last index used on the available ring */
90 volatile uint16_t last_used_idx_res; /**< Used for multiple devices reserving buffers. */
91 int callfd; /**< Used to notify the guest (trigger interrupt). */
92 int kickfd; /**< Currently unused as polling mode is enabled. */
94 uint64_t log_guest_addr; /**< Physical address of used ring, for logging */
95 uint64_t reserved[15]; /**< Reserve some spaces for future extension. */
96 struct buf_vector buf_vec[BUF_VECTOR_MAX]; /**< for scatter RX. */
97 } __rte_cache_aligned;
101 * Make an extra wrapper for VIRTIO_NET_F_MQ and
102 * VIRTIO_NET_CTRL_MQ_VQ_PAIRS_MAX as they are
103 * introduced since kernel v3.8. This makes our
104 * code buildable for older kernel.
106 #ifdef VIRTIO_NET_F_MQ
107 #define VHOST_MAX_QUEUE_PAIRS VIRTIO_NET_CTRL_MQ_VQ_PAIRS_MAX
108 #define VHOST_SUPPORTS_MQ (1ULL << VIRTIO_NET_F_MQ)
110 #define VHOST_MAX_QUEUE_PAIRS 1
111 #define VHOST_SUPPORTS_MQ 0
115 * Define virtio 1.0 for older kernels
117 #ifndef VIRTIO_F_VERSION_1
118 #define VIRTIO_F_VERSION_1 32
122 * Device structure contains all configuration information relating to the device.
125 struct virtio_memory *mem; /**< QEMU memory and memory region information. */
126 uint64_t features; /**< Negotiated feature set. */
127 uint64_t protocol_features; /**< Negotiated protocol feature set. */
128 uint64_t device_fh; /**< device identifier. */
129 uint32_t flags; /**< Device flags. Only used to check if device is running on data core. */
130 #define IF_NAME_SZ (PATH_MAX > IFNAMSIZ ? PATH_MAX : IFNAMSIZ)
131 char ifname[IF_NAME_SZ]; /**< Name of the tap device or socket path. */
132 uint32_t virt_qp_nb; /**< number of queue pair we have allocated */
133 void *priv; /**< private context */
134 uint64_t log_size; /**< Size of log area */
135 uint64_t log_base; /**< Where dirty pages are logged */
136 uint64_t reserved[62]; /**< Reserve some spaces for future extension. */
137 struct vhost_virtqueue *virtqueue[VHOST_MAX_QUEUE_PAIRS * 2]; /**< Contains all virtqueue information. */
138 } __rte_cache_aligned;
141 * Information relating to memory regions including offsets to addresses in QEMUs memory file.
143 struct virtio_memory_regions {
144 uint64_t guest_phys_address; /**< Base guest physical address of region. */
145 uint64_t guest_phys_address_end; /**< End guest physical address of region. */
146 uint64_t memory_size; /**< Size of region. */
147 uint64_t userspace_address; /**< Base userspace address of region. */
148 uint64_t address_offset; /**< Offset of region for address translation. */
153 * Memory structure includes region and mapping information.
155 struct virtio_memory {
156 uint64_t base_address; /**< Base QEMU userspace address of the memory file. */
157 uint64_t mapped_address; /**< Mapped address of memory file base in our applications memory space. */
158 uint64_t mapped_size; /**< Total size of memory file. */
159 uint32_t nregions; /**< Number of memory regions. */
160 struct virtio_memory_regions regions[0]; /**< Memory region information. */
164 * Device and vring operations.
166 * Make sure to set VIRTIO_DEV_RUNNING to the device flags in new_device and
167 * remove it in destroy_device.
170 struct virtio_net_device_ops {
171 int (*new_device)(struct virtio_net *); /**< Add device. */
172 void (*destroy_device)(volatile struct virtio_net *); /**< Remove device. */
174 int (*vring_state_changed)(struct virtio_net *dev, uint16_t queue_id, int enable); /**< triggered when a vring is enabled or disabled */
177 static inline uint16_t __attribute__((always_inline))
178 rte_vring_available_entries(struct virtio_net *dev, uint16_t queue_id)
180 struct vhost_virtqueue *vq = dev->virtqueue[queue_id];
185 return *(volatile uint16_t *)&vq->avail->idx - vq->last_used_idx_res;
189 * Function to convert guest physical addresses to vhost virtual addresses.
190 * This is used to convert guest virtio buffer addresses.
192 static inline uint64_t __attribute__((always_inline))
193 gpa_to_vva(struct virtio_net *dev, uint64_t guest_pa)
195 struct virtio_memory_regions *region;
197 uint64_t vhost_va = 0;
199 for (regionidx = 0; regionidx < dev->mem->nregions; regionidx++) {
200 region = &dev->mem->regions[regionidx];
201 if ((guest_pa >= region->guest_phys_address) &&
202 (guest_pa <= region->guest_phys_address_end)) {
203 vhost_va = region->address_offset + guest_pa;
212 * Disable features in feature_mask. Returns 0 on success.
214 int rte_vhost_feature_disable(uint64_t feature_mask);
217 * Enable features in feature_mask. Returns 0 on success.
219 int rte_vhost_feature_enable(uint64_t feature_mask);
221 /* Returns currently supported vhost features */
222 uint64_t rte_vhost_feature_get(void);
224 int rte_vhost_enable_guest_notification(struct virtio_net *dev, uint16_t queue_id, int enable);
226 /* Register vhost driver. dev_name could be different for multiple instance support. */
227 int rte_vhost_driver_register(const char *dev_name);
229 /* Unregister vhost driver. This is only meaningful to vhost user. */
230 int rte_vhost_driver_unregister(const char *dev_name);
232 /* Register callbacks. */
233 int rte_vhost_driver_callback_register(struct virtio_net_device_ops const * const);
234 /* Start vhost driver session blocking loop. */
235 int rte_vhost_driver_session_start(void);
238 * This function adds buffers to the virtio devices RX virtqueue. Buffers can
239 * be received from the physical port or from another virtual device. A packet
240 * count is returned to indicate the number of packets that were succesfully
241 * added to the RX queue.
245 * virtio queue index in mq case
247 * array to contain packets to be enqueued
249 * packets num to be enqueued
251 * num of packets enqueued
253 uint16_t rte_vhost_enqueue_burst(struct virtio_net *dev, uint16_t queue_id,
254 struct rte_mbuf **pkts, uint16_t count);
257 * This function gets guest buffers from the virtio device TX virtqueue,
258 * construct host mbufs, copies guest buffer content to host mbufs and
259 * store them in pkts to be processed.
263 * virtio queue index in mq case
265 * mbuf_pool where host mbuf is allocated.
267 * array to contain packets to be dequeued
269 * packets num to be dequeued
271 * num of packets dequeued
273 uint16_t rte_vhost_dequeue_burst(struct virtio_net *dev, uint16_t queue_id,
274 struct rte_mempool *mbuf_pool, struct rte_mbuf **pkts, uint16_t count);
276 #endif /* _VIRTIO_NET_H_ */