ceph/src/spdk/lib/vhost/rte_vhost/vhost.h

   1 /*-
   2  *   BSD LICENSE
   3  *
   4  *   Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
   5  *   All rights reserved.
   6  *
   7  *   Redistribution and use in source and binary forms, with or without
   8  *   modification, are permitted provided that the following conditions
   9  *   are met:
  10  *
  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
  16  *       distribution.
  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.
  20  *
  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.
  32  */
  33
  34 #ifndef _VHOST_NET_CDEV_H_
  35 #define _VHOST_NET_CDEV_H_
  36 #include <stdint.h>
  37 #include <stdio.h>
  38 #include <sys/types.h>
  39 #include <sys/queue.h>
  40 #include <unistd.h>
  41 #include <linux/vhost.h>
  42
  43 #include <rte_log.h>
  44
  45 #include "rte_virtio_net.h"
  46 #include "vhost_user.h"
  47
  48 /* Used to indicate that the device is running on a data core */
  49 #define VIRTIO_DEV_RUNNING 1
  50
  51 /* Backend value set by guest. */
  52 #define VIRTIO_DEV_STOPPED -1
  53
  54 #define BUF_VECTOR_MAX 256
  55
  56 /**
  57  * Structure contains buffer address, length and descriptor index
  58  * from vring to do scatter RX.
  59  */
  60 struct buf_vector {
  61         uint64_t buf_addr;
  62         uint32_t buf_len;
  63         uint32_t desc_idx;
  64 };
  65
  66 /*
  67  * A structure to hold some fields needed in zero copy code path,
  68  * mainly for associating an mbuf with the right desc_idx.
  69  */
  70 struct zcopy_mbuf {
  71         struct rte_mbuf *mbuf;
  72         uint32_t desc_idx;
  73         uint16_t in_use;
  74
  75         TAILQ_ENTRY(zcopy_mbuf) next;
  76 };
  77 TAILQ_HEAD(zcopy_mbuf_list, zcopy_mbuf);
  78
  79 /**
  80  * Structure contains variables relevant to RX/TX virtqueues.
  81  */
  82 struct vhost_virtqueue {
  83         struct vring_desc       *desc;
  84         struct vring_avail      *avail;
  85         struct vring_used       *used;
  86         uint32_t                size;
  87
  88         uint16_t                last_avail_idx;
  89         uint16_t                last_used_idx;
  90 #define VIRTIO_INVALID_EVENTFD          (-1)
  91 #define VIRTIO_UNINITIALIZED_EVENTFD    (-2)
  92
  93         /* Backend value to determine if device should started/stopped */
  94         int                     backend;
  95         /* Used to notify the guest (trigger interrupt) */
  96         int                     callfd;
  97         /* Currently unused as polling mode is enabled */
  98         int                     kickfd;
  99         int                     enabled;
 100
 101         /* Physical address of used ring, for logging */
 102         uint64_t                log_guest_addr;
 103
 104         uint16_t                nr_zmbuf;
 105         uint16_t                zmbuf_size;
 106         uint16_t                last_zmbuf_idx;
 107         struct zcopy_mbuf       *zmbufs;
 108         struct zcopy_mbuf_list  zmbuf_list;
 109
 110         struct vring_used_elem  *shadow_used_ring;
 111         uint16_t                shadow_used_idx;
 112 } __rte_cache_aligned;
 113
 114 /* Old kernels have no such macro defined */
 115 #ifndef VIRTIO_NET_F_GUEST_ANNOUNCE
 116  #define VIRTIO_NET_F_GUEST_ANNOUNCE 21
 117 #endif
 118
 119
 120 /*
 121  * Make an extra wrapper for VIRTIO_NET_F_MQ and
 122  * VIRTIO_NET_CTRL_MQ_VQ_PAIRS_MAX as they are
 123  * introduced since kernel v3.8. This makes our
 124  * code buildable for older kernel.
 125  */
 126 #ifdef VIRTIO_NET_F_MQ
 127  #define VHOST_MAX_QUEUE_PAIRS  VIRTIO_NET_CTRL_MQ_VQ_PAIRS_MAX
 128  #define VHOST_SUPPORTS_MQ      (1ULL << VIRTIO_NET_F_MQ)
 129 #else
 130  #define VHOST_MAX_QUEUE_PAIRS  1
 131  #define VHOST_SUPPORTS_MQ      0
 132 #endif
 133
 134 /*
 135  * Define virtio 1.0 for older kernels
 136  */
 137 #ifndef VIRTIO_F_VERSION_1
 138  #define VIRTIO_F_VERSION_1 32
 139 #endif
 140
 141 struct guest_page {
 142         uint64_t guest_phys_addr;
 143         uint64_t host_phys_addr;
 144         uint64_t size;
 145 };
 146
 147 /**
 148  * Device structure contains all configuration information relating
 149  * to the device.
 150  */
 151 struct virtio_net {
 152         /* Frontend (QEMU) memory and memory region information */
 153         struct virtio_memory    *mem;
 154         uint64_t                features;
 155         uint64_t                protocol_features;
 156         int                     vid;
 157         uint32_t                flags;
 158         uint16_t                vhost_hlen;
 159         /* to tell if we need broadcast rarp packet */
 160         rte_atomic16_t          broadcast_rarp;
 161         uint32_t                virt_qp_nb;
 162         uint32_t                num_queues;
 163         int                     dequeue_zero_copy;
 164         struct vhost_virtqueue  *virtqueue[VHOST_MAX_QUEUE_PAIRS * 2];
 165 #define IF_NAME_SZ (PATH_MAX > IFNAMSIZ ? PATH_MAX : IFNAMSIZ)
 166         char                    ifname[IF_NAME_SZ];
 167         uint64_t                log_size;
 168         uint64_t                log_base;
 169         uint64_t                log_addr;
 170         struct ether_addr       mac;
 171
 172         uint32_t                nr_guest_pages;
 173         uint32_t                max_guest_pages;
 174         struct guest_page       *guest_pages;
 175         int                     has_new_mem_table;
 176         struct VhostUserMemory  mem_table;
 177         int                     mem_table_fds[VHOST_MEMORY_MAX_NREGIONS];
 178 } __rte_cache_aligned;
 179
 180 /**
 181  * Information relating to memory regions including offsets to
 182  * addresses in QEMUs memory file.
 183  */
 184 struct virtio_memory_region {
 185         uint64_t guest_phys_addr;
 186         uint64_t guest_user_addr;
 187         uint64_t host_user_addr;
 188         uint64_t size;
 189         void     *mmap_addr;
 190         uint64_t mmap_size;
 191         int fd;
 192 };
 193
 194
 195 /**
 196  * Memory structure includes region and mapping information.
 197  */
 198 struct virtio_memory {
 199         uint32_t nregions;
 200         struct virtio_memory_region regions[0];
 201 };
 202
 203
 204 /* Macros for printing using RTE_LOG */
 205 #define RTE_LOGTYPE_VHOST_CONFIG RTE_LOGTYPE_USER1
 206 #define RTE_LOGTYPE_VHOST_DATA   RTE_LOGTYPE_USER1
 207
 208 #ifdef RTE_LIBRTE_VHOST_DEBUG
 209 #define VHOST_MAX_PRINT_BUFF 6072
 210 #define LOG_LEVEL RTE_LOG_DEBUG
 211 #define LOG_DEBUG(log_type, fmt, args...) RTE_LOG(DEBUG, log_type, fmt, ##args)
 212 #define PRINT_PACKET(device, addr, size, header) do { \
 213         char *pkt_addr = (char *)(addr); \
 214         unsigned int index; \
 215         char packet[VHOST_MAX_PRINT_BUFF]; \
 216         \
 217         if ((header)) \
 218                 snprintf(packet, VHOST_MAX_PRINT_BUFF, "(%d) Header size %d: ", (device->vid), (size)); \
 219         else \
 220                 snprintf(packet, VHOST_MAX_PRINT_BUFF, "(%d) Packet size %d: ", (device->vid), (size)); \
 221         for (index = 0; index < (size); index++) { \
 222                 snprintf(packet + strnlen(packet, VHOST_MAX_PRINT_BUFF), VHOST_MAX_PRINT_BUFF - strnlen(packet, VHOST_MAX_PRINT_BUFF), \
 223                         "%02hhx ", pkt_addr[index]); \
 224         } \
 225         snprintf(packet + strnlen(packet, VHOST_MAX_PRINT_BUFF), VHOST_MAX_PRINT_BUFF - strnlen(packet, VHOST_MAX_PRINT_BUFF), "\n"); \
 226         \
 227         LOG_DEBUG(VHOST_DATA, "%s", packet); \
 228 } while (0)
 229 #else
 230 #define LOG_LEVEL RTE_LOG_INFO
 231 #define LOG_DEBUG(log_type, fmt, args...) do {} while (0)
 232 #define PRINT_PACKET(device, addr, size, header) do {} while (0)
 233 #endif
 234
 235 extern uint64_t VHOST_FEATURES;
 236 #define MAX_VHOST_DEVICE        1024
 237 extern struct virtio_net *vhost_devices[MAX_VHOST_DEVICE];
 238
 239 /* Convert guest physical Address to host virtual address */
 240 static inline uint64_t __attribute__((always_inline))
 241 gpa_to_vva(struct virtio_net *dev, uint64_t gpa)
 242 {
 243         struct virtio_memory_region *reg;
 244         uint32_t i;
 245
 246         for (i = 0; i < dev->mem->nregions; i++) {
 247                 reg = &dev->mem->regions[i];
 248                 if (gpa >= reg->guest_phys_addr &&
 249                     gpa <  reg->guest_phys_addr + reg->size) {
 250                         return gpa - reg->guest_phys_addr +
 251                                reg->host_user_addr;
 252                 }
 253         }
 254
 255         return 0;
 256 }
 257
 258 /* Convert guest physical address to host physical address */
 259 static inline phys_addr_t __attribute__((always_inline))
 260 gpa_to_hpa(struct virtio_net *dev, uint64_t gpa, uint64_t size)
 261 {
 262         uint32_t i;
 263         struct guest_page *page;
 264
 265         for (i = 0; i < dev->nr_guest_pages; i++) {
 266                 page = &dev->guest_pages[i];
 267
 268                 if (gpa >= page->guest_phys_addr &&
 269                     gpa + size < page->guest_phys_addr + page->size) {
 270                         return gpa - page->guest_phys_addr +
 271                                page->host_phys_addr;
 272                 }
 273         }
 274
 275         return 0;
 276 }
 277
 278 extern struct virtio_net_device_ops const *notify_ops;
 279 struct virtio_net *get_device(int vid);
 280
 281 int vhost_new_device(void);
 282 void cleanup_device(struct virtio_net *dev, int destroy);
 283 void reset_device(struct virtio_net *dev);
 284 void vhost_destroy_device(int);
 285
 286 int alloc_vring_queue_pair(struct virtio_net *dev, uint32_t qp_idx);
 287
 288 void vhost_set_ifname(int, const char *if_name, unsigned int if_len);
 289 void vhost_enable_dequeue_zero_copy(int vid);
 290
 291 /*
 292  * Backend-specific cleanup.
 293  *
 294  * TODO: fix it; we have one backend now
 295  */
 296 void vhost_backend_cleanup(struct virtio_net *dev);
 297
 298 #endif /* _VHOST_NET_CDEV_H_ */