1 /* 2 * Copyright (c) 2013 Chris Torek <torek @ torek net> 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 17 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 24 * SUCH DAMAGE. 25 */ 26 /* 27 * This file and its contents are supplied under the terms of the 28 * Common Development and Distribution License ("CDDL"), version 1.0. 29 * You may only use this file in accordance with the terms of version 30 * 1.0 of the CDDL. 31 * 32 * A full copy of the text of the CDDL should have accompanied this 33 * source. A copy of the CDDL is also available via the Internet at 34 * http://www.illumos.org/license/CDDL. 35 * 36 * Copyright 2015 Pluribus Networks Inc. 37 * Copyright 2019 Joyent, Inc. 38 * Copyright 2022 OmniOS Community Edition (OmniOSce) Association. 39 * Copyright 2024 Oxide Computer Company 40 */ 41 42 #ifndef _VIONA_IMPL_H 43 #define _VIONA_IMPL_H 44 45 #include <sys/ddi.h> 46 #include <sys/list.h> 47 #include <sys/sunddi.h> 48 #include <sys/sunndi.h> 49 #include <sys/strsun.h> 50 #include <sys/sysmacros.h> 51 #include <sys/uio.h> 52 53 #include <sys/mac_client.h> 54 #include <sys/mac_provider.h> 55 #include <sys/mac_client_priv.h> 56 #include <sys/neti.h> 57 #include <inet/ip.h> 58 #include <inet/tcp.h> 59 60 #include <sys/vmm_drv.h> 61 #include <sys/viona_io.h> 62 63 struct viona_link; 64 typedef struct viona_link viona_link_t; 65 struct viona_desb; 66 typedef struct viona_desb viona_desb_t; 67 struct viona_net; 68 typedef struct viona_neti viona_neti_t; 69 70 typedef struct viona_transfer_stats { 71 /* Packets transferred successfully */ 72 uint64_t vts_packets; 73 /* Bytes transferred successfully */ 74 uint64_t vts_bytes; 75 /* 76 * Count of transfers which encountered errors, not including 77 * insufficient space in ring. 78 */ 79 uint64_t vts_errors; 80 /* 81 * Count of packets dropped due to insufficient space in the ring or by 82 * order of associated hook. 83 */ 84 uint64_t vts_drops; 85 } viona_transfer_stats_t; 86 87 enum viona_ring_state { 88 VRS_RESET = 0x0, /* just allocated or reset */ 89 VRS_SETUP = 0x1, /* addrs setup and starting worker thread */ 90 VRS_INIT = 0x2, /* worker thread started & waiting to run */ 91 VRS_RUN = 0x3, /* running work routine */ 92 VRS_STOP = 0x4, /* worker is exiting */ 93 }; 94 enum viona_ring_state_flags { 95 VRSF_REQ_START = 0x1, /* start running from INIT state */ 96 VRSF_REQ_STOP = 0x2, /* stop running, clean up, goto RESET state */ 97 VRSF_REQ_PAUSE = 0x4, /* stop running, goto INIT state */ 98 VRSF_RENEW = 0x8, /* ring renewing lease */ 99 }; 100 101 typedef struct viona_vring { 102 viona_link_t *vr_link; 103 104 kmutex_t vr_lock; 105 kcondvar_t vr_cv; 106 uint16_t vr_state; 107 uint16_t vr_state_flags; 108 uint_t vr_xfer_outstanding; 109 kthread_t *vr_worker_thread; 110 vmm_lease_t *vr_lease; 111 112 /* ring-sized resources for TX activity */ 113 viona_desb_t *vr_txdesb; 114 struct iovec *vr_txiov; 115 116 uint_t vr_intr_enabled; 117 uint64_t vr_msi_addr; 118 uint64_t vr_msi_msg; 119 120 /* Internal ring-related state */ 121 kmutex_t vr_a_mutex; /* sync consumers of 'avail' */ 122 kmutex_t vr_u_mutex; /* sync consumers of 'used' */ 123 uint64_t vr_pa; 124 uint16_t vr_size; 125 uint16_t vr_mask; /* cached from vr_size */ 126 uint16_t vr_cur_aidx; /* trails behind 'avail_idx' */ 127 uint16_t vr_cur_uidx; /* drives 'used_idx' */ 128 129 /* Reference to guest pages holding virtqueue */ 130 void **vr_map_pages; 131 vmm_page_t *vr_map_hold; 132 133 /* Per-ring general statistics */ 134 struct viona_transfer_stats vr_stats; 135 136 /* Per-ring error condition statistics */ 137 struct viona_ring_err_stats { 138 uint64_t rs_ndesc_too_high; 139 uint64_t rs_bad_idx; 140 uint64_t rs_indir_bad_len; 141 uint64_t rs_indir_bad_nest; 142 uint64_t rs_indir_bad_next; 143 uint64_t rs_no_space; 144 uint64_t rs_too_many_desc; 145 uint64_t rs_desc_bad_len; 146 147 uint64_t rs_bad_ring_addr; 148 149 uint64_t rs_fail_hcksum; 150 uint64_t rs_fail_hcksum6; 151 uint64_t rs_fail_hcksum_proto; 152 153 uint64_t rs_bad_rx_frame; 154 uint64_t rs_rx_merge_overrun; 155 uint64_t rs_rx_merge_underrun; 156 uint64_t rs_rx_pad_short; 157 uint64_t rs_rx_mcast_check; 158 uint64_t rs_too_short; 159 uint64_t rs_tx_absent; 160 161 uint64_t rs_rx_hookdrop; 162 uint64_t rs_tx_hookdrop; 163 } vr_err_stats; 164 } viona_vring_t; 165 166 struct viona_link { 167 vmm_hold_t *l_vm_hold; 168 boolean_t l_destroyed; 169 170 viona_vring_t l_vrings[VIONA_VQ_MAX]; 171 172 uint32_t l_features; 173 uint32_t l_features_hw; 174 uint32_t l_cap_csum; 175 176 uint16_t l_notify_ioport; 177 void *l_notify_cookie; 178 179 datalink_id_t l_linkid; 180 mac_handle_t l_mh; 181 mac_client_handle_t l_mch; 182 mac_promisc_handle_t l_mph; 183 mac_unicast_handle_t l_muh; 184 viona_promisc_t l_promisc; 185 186 pollhead_t l_pollhead; 187 188 viona_neti_t *l_neti; 189 190 kmutex_t l_stats_lock; 191 struct viona_link_stats { 192 struct viona_transfer_stats vls_rx; 193 struct viona_transfer_stats vls_tx; 194 } l_stats; 195 }; 196 197 typedef struct viona_nethook { 198 net_handle_t vnh_neti; 199 hook_family_t vnh_family; 200 hook_event_t vnh_event_in; 201 hook_event_t vnh_event_out; 202 hook_event_token_t vnh_token_in; 203 hook_event_token_t vnh_token_out; 204 boolean_t vnh_hooked; 205 } viona_nethook_t; 206 207 struct viona_neti { 208 list_node_t vni_node; 209 210 netid_t vni_netid; 211 zoneid_t vni_zid; 212 213 viona_nethook_t vni_nethook; 214 215 kmutex_t vni_lock; /* Protects remaining members */ 216 kcondvar_t vni_ref_change; /* Protected by vni_lock */ 217 uint_t vni_ref; /* Protected by vni_lock */ 218 list_t vni_dev_list; /* Protected by vni_lock */ 219 }; 220 221 typedef struct viona_kstats { 222 kstat_named_t vk_rx_packets; 223 kstat_named_t vk_rx_bytes; 224 kstat_named_t vk_rx_errors; 225 kstat_named_t vk_rx_drops; 226 kstat_named_t vk_tx_packets; 227 kstat_named_t vk_tx_bytes; 228 kstat_named_t vk_tx_errors; 229 kstat_named_t vk_tx_drops; 230 } viona_kstats_t; 231 232 typedef struct used_elem { 233 uint16_t id; 234 uint32_t len; 235 } used_elem_t; 236 237 typedef struct viona_soft_state { 238 kmutex_t ss_lock; 239 viona_link_t *ss_link; 240 list_node_t ss_node; 241 kstat_t *ss_kstat; 242 minor_t ss_minor; 243 } viona_soft_state_t; 244 245 #pragma pack(1) 246 struct virtio_desc { 247 uint64_t vd_addr; 248 uint32_t vd_len; 249 uint16_t vd_flags; 250 uint16_t vd_next; 251 }; 252 253 struct virtio_used { 254 uint32_t vu_idx; 255 uint32_t vu_tlen; 256 }; 257 258 struct virtio_net_mrgrxhdr { 259 uint8_t vrh_flags; 260 uint8_t vrh_gso_type; 261 uint16_t vrh_hdr_len; 262 uint16_t vrh_gso_size; 263 uint16_t vrh_csum_start; 264 uint16_t vrh_csum_offset; 265 uint16_t vrh_bufs; 266 }; 267 268 struct virtio_net_hdr { 269 uint8_t vrh_flags; 270 uint8_t vrh_gso_type; 271 uint16_t vrh_hdr_len; 272 uint16_t vrh_gso_size; 273 uint16_t vrh_csum_start; 274 uint16_t vrh_csum_offset; 275 }; 276 #pragma pack() 277 278 #define VNETHOOK_INTERESTED_IN(neti) \ 279 (neti)->vni_nethook.vnh_event_in.he_interested 280 #define VNETHOOK_INTERESTED_OUT(neti) \ 281 (neti)->vni_nethook.vnh_event_out.he_interested 282 283 284 #define VIONA_PROBE(name) DTRACE_PROBE(viona__##name) 285 #define VIONA_PROBE1(name, arg1, arg2) \ 286 DTRACE_PROBE1(viona__##name, arg1, arg2) 287 #define VIONA_PROBE2(name, arg1, arg2, arg3, arg4) \ 288 DTRACE_PROBE2(viona__##name, arg1, arg2, arg3, arg4) 289 #define VIONA_PROBE3(name, arg1, arg2, arg3, arg4, arg5, arg6) \ 290 DTRACE_PROBE3(viona__##name, arg1, arg2, arg3, arg4, arg5, arg6) 291 #define VIONA_PROBE4(name, arg1, arg2, arg3, arg4, arg5, arg6, arg7, arg8) \ 292 DTRACE_PROBE4(viona__##name, arg1, arg2, arg3, arg4, arg5, arg6, arg7, \ 293 arg8) 294 #define VIONA_PROBE5(name, arg1, arg2, arg3, arg4, arg5, arg6, arg7, arg8, \ 295 arg9, arg10) \ 296 DTRACE_PROBE5(viona__##name, arg1, arg2, arg3, arg4, arg5, arg6, arg7, \ 297 arg8, arg9, arg10) 298 #define VIONA_PROBE_BAD_RING_ADDR(r, a) \ 299 VIONA_PROBE2(bad_ring_addr, viona_vring_t *, r, void *, (void *)(a)) 300 301 /* Increment one of the named ring error stats */ 302 #define VIONA_RING_STAT_INCR(r, name) \ 303 (((r)->vr_err_stats.rs_ ## name)++) 304 305 #define VIONA_MAX_HDRS_LEN (sizeof (struct ether_vlan_header) + \ 306 IP_MAX_HDR_LENGTH + TCP_MAX_HDR_LENGTH) 307 308 #define VRING_AVAIL_F_NO_INTERRUPT 1 309 #define VRING_USED_F_NO_NOTIFY 1 310 311 #define VRING_DESC_F_NEXT (1 << 0) 312 #define VRING_DESC_F_WRITE (1 << 1) 313 #define VRING_DESC_F_INDIRECT (1 << 2) 314 315 #define VIRTIO_NET_HDR_F_NEEDS_CSUM (1 << 0) 316 #define VIRTIO_NET_HDR_F_DATA_VALID (1 << 1) 317 318 #define VIRTIO_NET_HDR_GSO_NONE 0 319 #define VIRTIO_NET_HDR_GSO_TCPV4 1 320 321 #define VIRTIO_NET_F_CSUM (1 << 0) 322 #define VIRTIO_NET_F_GUEST_CSUM (1 << 1) 323 #define VIRTIO_NET_F_MAC (1 << 5) /* host supplies MAC */ 324 #define VIRTIO_NET_F_GUEST_TSO4 (1 << 7) /* guest can accept TSO */ 325 #define VIRTIO_NET_F_HOST_TSO4 (1 << 11) /* host can accept TSO */ 326 #define VIRTIO_NET_F_MRG_RXBUF (1 << 15) /* host can merge RX bufs */ 327 #define VIRTIO_NET_F_STATUS (1 << 16) /* cfg status field present */ 328 #define VIRTIO_F_RING_NOTIFY_ON_EMPTY (1 << 24) 329 #define VIRTIO_F_RING_INDIRECT_DESC (1 << 28) 330 #define VIRTIO_F_RING_EVENT_IDX (1 << 29) 331 332 struct viona_ring_params { 333 uint64_t vrp_pa; 334 uint16_t vrp_size; 335 uint16_t vrp_avail_idx; 336 uint16_t vrp_used_idx; 337 }; 338 339 void viona_ring_alloc(viona_link_t *, viona_vring_t *); 340 void viona_ring_free(viona_vring_t *); 341 int viona_ring_get_state(viona_link_t *, uint16_t, struct viona_ring_params *); 342 int viona_ring_set_state(viona_link_t *, uint16_t, 343 const struct viona_ring_params *); 344 int viona_ring_reset(viona_vring_t *, boolean_t); 345 int viona_ring_init(viona_link_t *, uint16_t, const struct viona_ring_params *); 346 boolean_t viona_ring_lease_renew(viona_vring_t *); 347 bool vring_need_bail(const viona_vring_t *); 348 int viona_ring_pause(viona_vring_t *); 349 350 int vq_popchain(viona_vring_t *, struct iovec *, uint_t, uint16_t *, 351 vmm_page_t **); 352 void vq_pushchain(viona_vring_t *, uint32_t, uint16_t); 353 void vq_pushchain_many(viona_vring_t *, uint_t, used_elem_t *); 354 355 void viona_intr_ring(viona_vring_t *ring, boolean_t); 356 void viona_ring_set_no_notify(viona_vring_t *, boolean_t); 357 void viona_ring_disable_notify(viona_vring_t *); 358 void viona_ring_enable_notify(viona_vring_t *); 359 uint16_t viona_ring_num_avail(viona_vring_t *); 360 361 void viona_ring_stat_accept(viona_vring_t *, uint32_t); 362 void viona_ring_stat_drop(viona_vring_t *); 363 void viona_ring_stat_error(viona_vring_t *); 364 365 366 void viona_rx_init(void); 367 void viona_rx_fini(void); 368 int viona_rx_set(viona_link_t *, viona_promisc_t); 369 void viona_rx_clear(viona_link_t *); 370 void viona_worker_rx(viona_vring_t *, viona_link_t *); 371 372 extern kmutex_t viona_force_copy_lock; 373 void viona_worker_tx(viona_vring_t *, viona_link_t *); 374 void viona_tx_ring_alloc(viona_vring_t *, const uint16_t); 375 void viona_tx_ring_free(viona_vring_t *, const uint16_t); 376 377 void viona_neti_attach(void); 378 void viona_neti_detach(void); 379 viona_neti_t *viona_neti_lookup_by_zid(zoneid_t); 380 void viona_neti_rele(viona_neti_t *); 381 int viona_hook(viona_link_t *, viona_vring_t *, mblk_t **, boolean_t); 382 383 #endif /* _VIONA_IMPL_H */ 384