1 // SPDX-License-Identifier: GPL-2.0-only 2 /* 3 * Hyper-V transport for vsock 4 * 5 * Hyper-V Sockets supplies a byte-stream based communication mechanism 6 * between the host and the VM. This driver implements the necessary 7 * support in the VM by introducing the new vsock transport. 8 * 9 * Copyright (c) 2017, Microsoft Corporation. 10 */ 11 #include <linux/module.h> 12 #include <linux/vmalloc.h> 13 #include <linux/hyperv.h> 14 #include <net/sock.h> 15 #include <net/af_vsock.h> 16 17 /* The host side's design of the feature requires 6 exact 4KB pages for 18 * recv/send rings respectively -- this is suboptimal considering memory 19 * consumption, however unluckily we have to live with it, before the 20 * host comes up with a better design in the future. 21 */ 22 #define PAGE_SIZE_4K 4096 23 #define RINGBUFFER_HVS_RCV_SIZE (PAGE_SIZE_4K * 6) 24 #define RINGBUFFER_HVS_SND_SIZE (PAGE_SIZE_4K * 6) 25 26 /* The MTU is 16KB per the host side's design */ 27 #define HVS_MTU_SIZE (1024 * 16) 28 29 /* How long to wait for graceful shutdown of a connection */ 30 #define HVS_CLOSE_TIMEOUT (8 * HZ) 31 32 struct vmpipe_proto_header { 33 u32 pkt_type; 34 u32 data_size; 35 }; 36 37 /* For recv, we use the VMBus in-place packet iterator APIs to directly copy 38 * data from the ringbuffer into the userspace buffer. 39 */ 40 struct hvs_recv_buf { 41 /* The header before the payload data */ 42 struct vmpipe_proto_header hdr; 43 44 /* The payload */ 45 u8 data[HVS_MTU_SIZE]; 46 }; 47 48 /* We can send up to HVS_MTU_SIZE bytes of payload to the host, but let's use 49 * a small size, i.e. HVS_SEND_BUF_SIZE, to minimize the dynamically-allocated 50 * buffer, because tests show there is no significant performance difference. 51 * 52 * Note: the buffer can be eliminated in the future when we add new VMBus 53 * ringbuffer APIs that allow us to directly copy data from userspace buffer 54 * to VMBus ringbuffer. 55 */ 56 #define HVS_SEND_BUF_SIZE (PAGE_SIZE_4K - sizeof(struct vmpipe_proto_header)) 57 58 struct hvs_send_buf { 59 /* The header before the payload data */ 60 struct vmpipe_proto_header hdr; 61 62 /* The payload */ 63 u8 data[HVS_SEND_BUF_SIZE]; 64 }; 65 66 #define HVS_HEADER_LEN (sizeof(struct vmpacket_descriptor) + \ 67 sizeof(struct vmpipe_proto_header)) 68 69 /* See 'prev_indices' in hv_ringbuffer_read(), hv_ringbuffer_write(), and 70 * __hv_pkt_iter_next(). 71 */ 72 #define VMBUS_PKT_TRAILER_SIZE (sizeof(u64)) 73 74 #define HVS_PKT_LEN(payload_len) (HVS_HEADER_LEN + \ 75 ALIGN((payload_len), 8) + \ 76 VMBUS_PKT_TRAILER_SIZE) 77 78 union hvs_service_id { 79 uuid_le srv_id; 80 81 struct { 82 unsigned int svm_port; 83 unsigned char b[sizeof(uuid_le) - sizeof(unsigned int)]; 84 }; 85 }; 86 87 /* Per-socket state (accessed via vsk->trans) */ 88 struct hvsock { 89 struct vsock_sock *vsk; 90 91 uuid_le vm_srv_id; 92 uuid_le host_srv_id; 93 94 struct vmbus_channel *chan; 95 struct vmpacket_descriptor *recv_desc; 96 97 /* The length of the payload not delivered to userland yet */ 98 u32 recv_data_len; 99 /* The offset of the payload */ 100 u32 recv_data_off; 101 102 /* Have we sent the zero-length packet (FIN)? */ 103 bool fin_sent; 104 }; 105 106 /* In the VM, we support Hyper-V Sockets with AF_VSOCK, and the endpoint is 107 * <cid, port> (see struct sockaddr_vm). Note: cid is not really used here: 108 * when we write apps to connect to the host, we can only use VMADDR_CID_ANY 109 * or VMADDR_CID_HOST (both are equivalent) as the remote cid, and when we 110 * write apps to bind() & listen() in the VM, we can only use VMADDR_CID_ANY 111 * as the local cid. 112 * 113 * On the host, Hyper-V Sockets are supported by Winsock AF_HYPERV: 114 * https://docs.microsoft.com/en-us/virtualization/hyper-v-on-windows/user- 115 * guide/make-integration-service, and the endpoint is <VmID, ServiceId> with 116 * the below sockaddr: 117 * 118 * struct SOCKADDR_HV 119 * { 120 * ADDRESS_FAMILY Family; 121 * USHORT Reserved; 122 * GUID VmId; 123 * GUID ServiceId; 124 * }; 125 * Note: VmID is not used by Linux VM and actually it isn't transmitted via 126 * VMBus, because here it's obvious the host and the VM can easily identify 127 * each other. Though the VmID is useful on the host, especially in the case 128 * of Windows container, Linux VM doesn't need it at all. 129 * 130 * To make use of the AF_VSOCK infrastructure in Linux VM, we have to limit 131 * the available GUID space of SOCKADDR_HV so that we can create a mapping 132 * between AF_VSOCK port and SOCKADDR_HV Service GUID. The rule of writing 133 * Hyper-V Sockets apps on the host and in Linux VM is: 134 * 135 **************************************************************************** 136 * The only valid Service GUIDs, from the perspectives of both the host and * 137 * Linux VM, that can be connected by the other end, must conform to this * 138 * format: <port>-facb-11e6-bd58-64006a7986d3, and the "port" must be in * 139 * this range [0, 0x7FFFFFFF]. * 140 **************************************************************************** 141 * 142 * When we write apps on the host to connect(), the GUID ServiceID is used. 143 * When we write apps in Linux VM to connect(), we only need to specify the 144 * port and the driver will form the GUID and use that to request the host. 145 * 146 * From the perspective of Linux VM: 147 * 1. the local ephemeral port (i.e. the local auto-bound port when we call 148 * connect() without explicit bind()) is generated by __vsock_bind_stream(), 149 * and the range is [1024, 0xFFFFFFFF). 150 * 2. the remote ephemeral port (i.e. the auto-generated remote port for 151 * a connect request initiated by the host's connect()) is generated by 152 * hvs_remote_addr_init() and the range is [0x80000000, 0xFFFFFFFF). 153 */ 154 155 #define MAX_LISTEN_PORT ((u32)0x7FFFFFFF) 156 #define MAX_VM_LISTEN_PORT MAX_LISTEN_PORT 157 #define MAX_HOST_LISTEN_PORT MAX_LISTEN_PORT 158 #define MIN_HOST_EPHEMERAL_PORT (MAX_HOST_LISTEN_PORT + 1) 159 160 /* 00000000-facb-11e6-bd58-64006a7986d3 */ 161 static const uuid_le srv_id_template = 162 UUID_LE(0x00000000, 0xfacb, 0x11e6, 0xbd, 0x58, 163 0x64, 0x00, 0x6a, 0x79, 0x86, 0xd3); 164 165 static bool is_valid_srv_id(const uuid_le *id) 166 { 167 return !memcmp(&id->b[4], &srv_id_template.b[4], sizeof(uuid_le) - 4); 168 } 169 170 static unsigned int get_port_by_srv_id(const uuid_le *svr_id) 171 { 172 return *((unsigned int *)svr_id); 173 } 174 175 static void hvs_addr_init(struct sockaddr_vm *addr, const uuid_le *svr_id) 176 { 177 unsigned int port = get_port_by_srv_id(svr_id); 178 179 vsock_addr_init(addr, VMADDR_CID_ANY, port); 180 } 181 182 static void hvs_remote_addr_init(struct sockaddr_vm *remote, 183 struct sockaddr_vm *local) 184 { 185 static u32 host_ephemeral_port = MIN_HOST_EPHEMERAL_PORT; 186 struct sock *sk; 187 188 vsock_addr_init(remote, VMADDR_CID_ANY, VMADDR_PORT_ANY); 189 190 while (1) { 191 /* Wrap around ? */ 192 if (host_ephemeral_port < MIN_HOST_EPHEMERAL_PORT || 193 host_ephemeral_port == VMADDR_PORT_ANY) 194 host_ephemeral_port = MIN_HOST_EPHEMERAL_PORT; 195 196 remote->svm_port = host_ephemeral_port++; 197 198 sk = vsock_find_connected_socket(remote, local); 199 if (!sk) { 200 /* Found an available ephemeral port */ 201 return; 202 } 203 204 /* Release refcnt got in vsock_find_connected_socket */ 205 sock_put(sk); 206 } 207 } 208 209 static void hvs_set_channel_pending_send_size(struct vmbus_channel *chan) 210 { 211 set_channel_pending_send_size(chan, 212 HVS_PKT_LEN(HVS_SEND_BUF_SIZE)); 213 214 virt_mb(); 215 } 216 217 static bool hvs_channel_readable(struct vmbus_channel *chan) 218 { 219 u32 readable = hv_get_bytes_to_read(&chan->inbound); 220 221 /* 0-size payload means FIN */ 222 return readable >= HVS_PKT_LEN(0); 223 } 224 225 static int hvs_channel_readable_payload(struct vmbus_channel *chan) 226 { 227 u32 readable = hv_get_bytes_to_read(&chan->inbound); 228 229 if (readable > HVS_PKT_LEN(0)) { 230 /* At least we have 1 byte to read. We don't need to return 231 * the exact readable bytes: see vsock_stream_recvmsg() -> 232 * vsock_stream_has_data(). 233 */ 234 return 1; 235 } 236 237 if (readable == HVS_PKT_LEN(0)) { 238 /* 0-size payload means FIN */ 239 return 0; 240 } 241 242 /* No payload or FIN */ 243 return -1; 244 } 245 246 static size_t hvs_channel_writable_bytes(struct vmbus_channel *chan) 247 { 248 u32 writeable = hv_get_bytes_to_write(&chan->outbound); 249 size_t ret; 250 251 /* The ringbuffer mustn't be 100% full, and we should reserve a 252 * zero-length-payload packet for the FIN: see hv_ringbuffer_write() 253 * and hvs_shutdown(). 254 */ 255 if (writeable <= HVS_PKT_LEN(1) + HVS_PKT_LEN(0)) 256 return 0; 257 258 ret = writeable - HVS_PKT_LEN(1) - HVS_PKT_LEN(0); 259 260 return round_down(ret, 8); 261 } 262 263 static int hvs_send_data(struct vmbus_channel *chan, 264 struct hvs_send_buf *send_buf, size_t to_write) 265 { 266 send_buf->hdr.pkt_type = 1; 267 send_buf->hdr.data_size = to_write; 268 return vmbus_sendpacket(chan, &send_buf->hdr, 269 sizeof(send_buf->hdr) + to_write, 270 0, VM_PKT_DATA_INBAND, 0); 271 } 272 273 static void hvs_channel_cb(void *ctx) 274 { 275 struct sock *sk = (struct sock *)ctx; 276 struct vsock_sock *vsk = vsock_sk(sk); 277 struct hvsock *hvs = vsk->trans; 278 struct vmbus_channel *chan = hvs->chan; 279 280 if (hvs_channel_readable(chan)) 281 sk->sk_data_ready(sk); 282 283 if (hv_get_bytes_to_write(&chan->outbound) > 0) 284 sk->sk_write_space(sk); 285 } 286 287 static void hvs_do_close_lock_held(struct vsock_sock *vsk, 288 bool cancel_timeout) 289 { 290 struct sock *sk = sk_vsock(vsk); 291 292 sock_set_flag(sk, SOCK_DONE); 293 vsk->peer_shutdown = SHUTDOWN_MASK; 294 if (vsock_stream_has_data(vsk) <= 0) 295 sk->sk_state = TCP_CLOSING; 296 sk->sk_state_change(sk); 297 if (vsk->close_work_scheduled && 298 (!cancel_timeout || cancel_delayed_work(&vsk->close_work))) { 299 vsk->close_work_scheduled = false; 300 vsock_remove_sock(vsk); 301 302 /* Release the reference taken while scheduling the timeout */ 303 sock_put(sk); 304 } 305 } 306 307 static void hvs_close_connection(struct vmbus_channel *chan) 308 { 309 struct sock *sk = get_per_channel_state(chan); 310 311 lock_sock(sk); 312 hvs_do_close_lock_held(vsock_sk(sk), true); 313 release_sock(sk); 314 } 315 316 static void hvs_open_connection(struct vmbus_channel *chan) 317 { 318 uuid_le *if_instance, *if_type; 319 unsigned char conn_from_host; 320 321 struct sockaddr_vm addr; 322 struct sock *sk, *new = NULL; 323 struct vsock_sock *vnew = NULL; 324 struct hvsock *hvs, *hvs_new = NULL; 325 int ret; 326 327 if_type = &chan->offermsg.offer.if_type; 328 if_instance = &chan->offermsg.offer.if_instance; 329 conn_from_host = chan->offermsg.offer.u.pipe.user_def[0]; 330 331 /* The host or the VM should only listen on a port in 332 * [0, MAX_LISTEN_PORT] 333 */ 334 if (!is_valid_srv_id(if_type) || 335 get_port_by_srv_id(if_type) > MAX_LISTEN_PORT) 336 return; 337 338 hvs_addr_init(&addr, conn_from_host ? if_type : if_instance); 339 sk = vsock_find_bound_socket(&addr); 340 if (!sk) 341 return; 342 343 lock_sock(sk); 344 if ((conn_from_host && sk->sk_state != TCP_LISTEN) || 345 (!conn_from_host && sk->sk_state != TCP_SYN_SENT)) 346 goto out; 347 348 if (conn_from_host) { 349 if (sk->sk_ack_backlog >= sk->sk_max_ack_backlog) 350 goto out; 351 352 new = __vsock_create(sock_net(sk), NULL, sk, GFP_KERNEL, 353 sk->sk_type, 0); 354 if (!new) 355 goto out; 356 357 new->sk_state = TCP_SYN_SENT; 358 vnew = vsock_sk(new); 359 hvs_new = vnew->trans; 360 hvs_new->chan = chan; 361 } else { 362 hvs = vsock_sk(sk)->trans; 363 hvs->chan = chan; 364 } 365 366 set_channel_read_mode(chan, HV_CALL_DIRECT); 367 ret = vmbus_open(chan, RINGBUFFER_HVS_SND_SIZE, 368 RINGBUFFER_HVS_RCV_SIZE, NULL, 0, 369 hvs_channel_cb, conn_from_host ? new : sk); 370 if (ret != 0) { 371 if (conn_from_host) { 372 hvs_new->chan = NULL; 373 sock_put(new); 374 } else { 375 hvs->chan = NULL; 376 } 377 goto out; 378 } 379 380 set_per_channel_state(chan, conn_from_host ? new : sk); 381 vmbus_set_chn_rescind_callback(chan, hvs_close_connection); 382 383 /* Set the pending send size to max packet size to always get 384 * notifications from the host when there is enough writable space. 385 * The host is optimized to send notifications only when the pending 386 * size boundary is crossed, and not always. 387 */ 388 hvs_set_channel_pending_send_size(chan); 389 390 if (conn_from_host) { 391 new->sk_state = TCP_ESTABLISHED; 392 sk->sk_ack_backlog++; 393 394 hvs_addr_init(&vnew->local_addr, if_type); 395 hvs_remote_addr_init(&vnew->remote_addr, &vnew->local_addr); 396 397 hvs_new->vm_srv_id = *if_type; 398 hvs_new->host_srv_id = *if_instance; 399 400 vsock_insert_connected(vnew); 401 402 vsock_enqueue_accept(sk, new); 403 } else { 404 sk->sk_state = TCP_ESTABLISHED; 405 sk->sk_socket->state = SS_CONNECTED; 406 407 vsock_insert_connected(vsock_sk(sk)); 408 } 409 410 sk->sk_state_change(sk); 411 412 out: 413 /* Release refcnt obtained when we called vsock_find_bound_socket() */ 414 sock_put(sk); 415 416 release_sock(sk); 417 } 418 419 static u32 hvs_get_local_cid(void) 420 { 421 return VMADDR_CID_ANY; 422 } 423 424 static int hvs_sock_init(struct vsock_sock *vsk, struct vsock_sock *psk) 425 { 426 struct hvsock *hvs; 427 428 hvs = kzalloc(sizeof(*hvs), GFP_KERNEL); 429 if (!hvs) 430 return -ENOMEM; 431 432 vsk->trans = hvs; 433 hvs->vsk = vsk; 434 435 return 0; 436 } 437 438 static int hvs_connect(struct vsock_sock *vsk) 439 { 440 union hvs_service_id vm, host; 441 struct hvsock *h = vsk->trans; 442 443 vm.srv_id = srv_id_template; 444 vm.svm_port = vsk->local_addr.svm_port; 445 h->vm_srv_id = vm.srv_id; 446 447 host.srv_id = srv_id_template; 448 host.svm_port = vsk->remote_addr.svm_port; 449 h->host_srv_id = host.srv_id; 450 451 return vmbus_send_tl_connect_request(&h->vm_srv_id, &h->host_srv_id); 452 } 453 454 static void hvs_shutdown_lock_held(struct hvsock *hvs, int mode) 455 { 456 struct vmpipe_proto_header hdr; 457 458 if (hvs->fin_sent || !hvs->chan) 459 return; 460 461 /* It can't fail: see hvs_channel_writable_bytes(). */ 462 (void)hvs_send_data(hvs->chan, (struct hvs_send_buf *)&hdr, 0); 463 hvs->fin_sent = true; 464 } 465 466 static int hvs_shutdown(struct vsock_sock *vsk, int mode) 467 { 468 struct sock *sk = sk_vsock(vsk); 469 470 if (!(mode & SEND_SHUTDOWN)) 471 return 0; 472 473 lock_sock(sk); 474 hvs_shutdown_lock_held(vsk->trans, mode); 475 release_sock(sk); 476 return 0; 477 } 478 479 static void hvs_close_timeout(struct work_struct *work) 480 { 481 struct vsock_sock *vsk = 482 container_of(work, struct vsock_sock, close_work.work); 483 struct sock *sk = sk_vsock(vsk); 484 485 sock_hold(sk); 486 lock_sock(sk); 487 if (!sock_flag(sk, SOCK_DONE)) 488 hvs_do_close_lock_held(vsk, false); 489 490 vsk->close_work_scheduled = false; 491 release_sock(sk); 492 sock_put(sk); 493 } 494 495 /* Returns true, if it is safe to remove socket; false otherwise */ 496 static bool hvs_close_lock_held(struct vsock_sock *vsk) 497 { 498 struct sock *sk = sk_vsock(vsk); 499 500 if (!(sk->sk_state == TCP_ESTABLISHED || 501 sk->sk_state == TCP_CLOSING)) 502 return true; 503 504 if ((sk->sk_shutdown & SHUTDOWN_MASK) != SHUTDOWN_MASK) 505 hvs_shutdown_lock_held(vsk->trans, SHUTDOWN_MASK); 506 507 if (sock_flag(sk, SOCK_DONE)) 508 return true; 509 510 /* This reference will be dropped by the delayed close routine */ 511 sock_hold(sk); 512 INIT_DELAYED_WORK(&vsk->close_work, hvs_close_timeout); 513 vsk->close_work_scheduled = true; 514 schedule_delayed_work(&vsk->close_work, HVS_CLOSE_TIMEOUT); 515 return false; 516 } 517 518 static void hvs_release(struct vsock_sock *vsk) 519 { 520 struct sock *sk = sk_vsock(vsk); 521 bool remove_sock; 522 523 lock_sock(sk); 524 remove_sock = hvs_close_lock_held(vsk); 525 release_sock(sk); 526 if (remove_sock) 527 vsock_remove_sock(vsk); 528 } 529 530 static void hvs_destruct(struct vsock_sock *vsk) 531 { 532 struct hvsock *hvs = vsk->trans; 533 struct vmbus_channel *chan = hvs->chan; 534 535 if (chan) 536 vmbus_hvsock_device_unregister(chan); 537 538 kfree(hvs); 539 } 540 541 static int hvs_dgram_bind(struct vsock_sock *vsk, struct sockaddr_vm *addr) 542 { 543 return -EOPNOTSUPP; 544 } 545 546 static int hvs_dgram_dequeue(struct vsock_sock *vsk, struct msghdr *msg, 547 size_t len, int flags) 548 { 549 return -EOPNOTSUPP; 550 } 551 552 static int hvs_dgram_enqueue(struct vsock_sock *vsk, 553 struct sockaddr_vm *remote, struct msghdr *msg, 554 size_t dgram_len) 555 { 556 return -EOPNOTSUPP; 557 } 558 559 static bool hvs_dgram_allow(u32 cid, u32 port) 560 { 561 return false; 562 } 563 564 static int hvs_update_recv_data(struct hvsock *hvs) 565 { 566 struct hvs_recv_buf *recv_buf; 567 u32 payload_len; 568 569 recv_buf = (struct hvs_recv_buf *)(hvs->recv_desc + 1); 570 payload_len = recv_buf->hdr.data_size; 571 572 if (payload_len > HVS_MTU_SIZE) 573 return -EIO; 574 575 if (payload_len == 0) 576 hvs->vsk->peer_shutdown |= SEND_SHUTDOWN; 577 578 hvs->recv_data_len = payload_len; 579 hvs->recv_data_off = 0; 580 581 return 0; 582 } 583 584 static ssize_t hvs_stream_dequeue(struct vsock_sock *vsk, struct msghdr *msg, 585 size_t len, int flags) 586 { 587 struct hvsock *hvs = vsk->trans; 588 bool need_refill = !hvs->recv_desc; 589 struct hvs_recv_buf *recv_buf; 590 u32 to_read; 591 int ret; 592 593 if (flags & MSG_PEEK) 594 return -EOPNOTSUPP; 595 596 if (need_refill) { 597 hvs->recv_desc = hv_pkt_iter_first(hvs->chan); 598 ret = hvs_update_recv_data(hvs); 599 if (ret) 600 return ret; 601 } 602 603 recv_buf = (struct hvs_recv_buf *)(hvs->recv_desc + 1); 604 to_read = min_t(u32, len, hvs->recv_data_len); 605 ret = memcpy_to_msg(msg, recv_buf->data + hvs->recv_data_off, to_read); 606 if (ret != 0) 607 return ret; 608 609 hvs->recv_data_len -= to_read; 610 if (hvs->recv_data_len == 0) { 611 hvs->recv_desc = hv_pkt_iter_next(hvs->chan, hvs->recv_desc); 612 if (hvs->recv_desc) { 613 ret = hvs_update_recv_data(hvs); 614 if (ret) 615 return ret; 616 } 617 } else { 618 hvs->recv_data_off += to_read; 619 } 620 621 return to_read; 622 } 623 624 static ssize_t hvs_stream_enqueue(struct vsock_sock *vsk, struct msghdr *msg, 625 size_t len) 626 { 627 struct hvsock *hvs = vsk->trans; 628 struct vmbus_channel *chan = hvs->chan; 629 struct hvs_send_buf *send_buf; 630 ssize_t to_write, max_writable, ret; 631 632 BUILD_BUG_ON(sizeof(*send_buf) != PAGE_SIZE_4K); 633 634 send_buf = kmalloc(sizeof(*send_buf), GFP_KERNEL); 635 if (!send_buf) 636 return -ENOMEM; 637 638 max_writable = hvs_channel_writable_bytes(chan); 639 to_write = min_t(ssize_t, len, max_writable); 640 to_write = min_t(ssize_t, to_write, HVS_SEND_BUF_SIZE); 641 642 ret = memcpy_from_msg(send_buf->data, msg, to_write); 643 if (ret < 0) 644 goto out; 645 646 ret = hvs_send_data(hvs->chan, send_buf, to_write); 647 if (ret < 0) 648 goto out; 649 650 ret = to_write; 651 out: 652 kfree(send_buf); 653 return ret; 654 } 655 656 static s64 hvs_stream_has_data(struct vsock_sock *vsk) 657 { 658 struct hvsock *hvs = vsk->trans; 659 s64 ret; 660 661 if (hvs->recv_data_len > 0) 662 return 1; 663 664 switch (hvs_channel_readable_payload(hvs->chan)) { 665 case 1: 666 ret = 1; 667 break; 668 case 0: 669 vsk->peer_shutdown |= SEND_SHUTDOWN; 670 ret = 0; 671 break; 672 default: /* -1 */ 673 ret = 0; 674 break; 675 } 676 677 return ret; 678 } 679 680 static s64 hvs_stream_has_space(struct vsock_sock *vsk) 681 { 682 struct hvsock *hvs = vsk->trans; 683 684 return hvs_channel_writable_bytes(hvs->chan); 685 } 686 687 static u64 hvs_stream_rcvhiwat(struct vsock_sock *vsk) 688 { 689 return HVS_MTU_SIZE + 1; 690 } 691 692 static bool hvs_stream_is_active(struct vsock_sock *vsk) 693 { 694 struct hvsock *hvs = vsk->trans; 695 696 return hvs->chan != NULL; 697 } 698 699 static bool hvs_stream_allow(u32 cid, u32 port) 700 { 701 /* The host's port range [MIN_HOST_EPHEMERAL_PORT, 0xFFFFFFFF) is 702 * reserved as ephemeral ports, which are used as the host's ports 703 * when the host initiates connections. 704 * 705 * Perform this check in the guest so an immediate error is produced 706 * instead of a timeout. 707 */ 708 if (port > MAX_HOST_LISTEN_PORT) 709 return false; 710 711 if (cid == VMADDR_CID_HOST) 712 return true; 713 714 return false; 715 } 716 717 static 718 int hvs_notify_poll_in(struct vsock_sock *vsk, size_t target, bool *readable) 719 { 720 struct hvsock *hvs = vsk->trans; 721 722 *readable = hvs_channel_readable(hvs->chan); 723 return 0; 724 } 725 726 static 727 int hvs_notify_poll_out(struct vsock_sock *vsk, size_t target, bool *writable) 728 { 729 *writable = hvs_stream_has_space(vsk) > 0; 730 731 return 0; 732 } 733 734 static 735 int hvs_notify_recv_init(struct vsock_sock *vsk, size_t target, 736 struct vsock_transport_recv_notify_data *d) 737 { 738 return 0; 739 } 740 741 static 742 int hvs_notify_recv_pre_block(struct vsock_sock *vsk, size_t target, 743 struct vsock_transport_recv_notify_data *d) 744 { 745 return 0; 746 } 747 748 static 749 int hvs_notify_recv_pre_dequeue(struct vsock_sock *vsk, size_t target, 750 struct vsock_transport_recv_notify_data *d) 751 { 752 return 0; 753 } 754 755 static 756 int hvs_notify_recv_post_dequeue(struct vsock_sock *vsk, size_t target, 757 ssize_t copied, bool data_read, 758 struct vsock_transport_recv_notify_data *d) 759 { 760 return 0; 761 } 762 763 static 764 int hvs_notify_send_init(struct vsock_sock *vsk, 765 struct vsock_transport_send_notify_data *d) 766 { 767 return 0; 768 } 769 770 static 771 int hvs_notify_send_pre_block(struct vsock_sock *vsk, 772 struct vsock_transport_send_notify_data *d) 773 { 774 return 0; 775 } 776 777 static 778 int hvs_notify_send_pre_enqueue(struct vsock_sock *vsk, 779 struct vsock_transport_send_notify_data *d) 780 { 781 return 0; 782 } 783 784 static 785 int hvs_notify_send_post_enqueue(struct vsock_sock *vsk, ssize_t written, 786 struct vsock_transport_send_notify_data *d) 787 { 788 return 0; 789 } 790 791 static void hvs_set_buffer_size(struct vsock_sock *vsk, u64 val) 792 { 793 /* Ignored. */ 794 } 795 796 static void hvs_set_min_buffer_size(struct vsock_sock *vsk, u64 val) 797 { 798 /* Ignored. */ 799 } 800 801 static void hvs_set_max_buffer_size(struct vsock_sock *vsk, u64 val) 802 { 803 /* Ignored. */ 804 } 805 806 static u64 hvs_get_buffer_size(struct vsock_sock *vsk) 807 { 808 return -ENOPROTOOPT; 809 } 810 811 static u64 hvs_get_min_buffer_size(struct vsock_sock *vsk) 812 { 813 return -ENOPROTOOPT; 814 } 815 816 static u64 hvs_get_max_buffer_size(struct vsock_sock *vsk) 817 { 818 return -ENOPROTOOPT; 819 } 820 821 static struct vsock_transport hvs_transport = { 822 .get_local_cid = hvs_get_local_cid, 823 824 .init = hvs_sock_init, 825 .destruct = hvs_destruct, 826 .release = hvs_release, 827 .connect = hvs_connect, 828 .shutdown = hvs_shutdown, 829 830 .dgram_bind = hvs_dgram_bind, 831 .dgram_dequeue = hvs_dgram_dequeue, 832 .dgram_enqueue = hvs_dgram_enqueue, 833 .dgram_allow = hvs_dgram_allow, 834 835 .stream_dequeue = hvs_stream_dequeue, 836 .stream_enqueue = hvs_stream_enqueue, 837 .stream_has_data = hvs_stream_has_data, 838 .stream_has_space = hvs_stream_has_space, 839 .stream_rcvhiwat = hvs_stream_rcvhiwat, 840 .stream_is_active = hvs_stream_is_active, 841 .stream_allow = hvs_stream_allow, 842 843 .notify_poll_in = hvs_notify_poll_in, 844 .notify_poll_out = hvs_notify_poll_out, 845 .notify_recv_init = hvs_notify_recv_init, 846 .notify_recv_pre_block = hvs_notify_recv_pre_block, 847 .notify_recv_pre_dequeue = hvs_notify_recv_pre_dequeue, 848 .notify_recv_post_dequeue = hvs_notify_recv_post_dequeue, 849 .notify_send_init = hvs_notify_send_init, 850 .notify_send_pre_block = hvs_notify_send_pre_block, 851 .notify_send_pre_enqueue = hvs_notify_send_pre_enqueue, 852 .notify_send_post_enqueue = hvs_notify_send_post_enqueue, 853 854 .set_buffer_size = hvs_set_buffer_size, 855 .set_min_buffer_size = hvs_set_min_buffer_size, 856 .set_max_buffer_size = hvs_set_max_buffer_size, 857 .get_buffer_size = hvs_get_buffer_size, 858 .get_min_buffer_size = hvs_get_min_buffer_size, 859 .get_max_buffer_size = hvs_get_max_buffer_size, 860 }; 861 862 static int hvs_probe(struct hv_device *hdev, 863 const struct hv_vmbus_device_id *dev_id) 864 { 865 struct vmbus_channel *chan = hdev->channel; 866 867 hvs_open_connection(chan); 868 869 /* Always return success to suppress the unnecessary error message 870 * in vmbus_probe(): on error the host will rescind the device in 871 * 30 seconds and we can do cleanup at that time in 872 * vmbus_onoffer_rescind(). 873 */ 874 return 0; 875 } 876 877 static int hvs_remove(struct hv_device *hdev) 878 { 879 struct vmbus_channel *chan = hdev->channel; 880 881 vmbus_close(chan); 882 883 return 0; 884 } 885 886 /* This isn't really used. See vmbus_match() and vmbus_probe() */ 887 static const struct hv_vmbus_device_id id_table[] = { 888 {}, 889 }; 890 891 static struct hv_driver hvs_drv = { 892 .name = "hv_sock", 893 .hvsock = true, 894 .id_table = id_table, 895 .probe = hvs_probe, 896 .remove = hvs_remove, 897 }; 898 899 static int __init hvs_init(void) 900 { 901 int ret; 902 903 if (vmbus_proto_version < VERSION_WIN10) 904 return -ENODEV; 905 906 ret = vmbus_driver_register(&hvs_drv); 907 if (ret != 0) 908 return ret; 909 910 ret = vsock_core_init(&hvs_transport); 911 if (ret) { 912 vmbus_driver_unregister(&hvs_drv); 913 return ret; 914 } 915 916 return 0; 917 } 918 919 static void __exit hvs_exit(void) 920 { 921 vsock_core_exit(); 922 vmbus_driver_unregister(&hvs_drv); 923 } 924 925 module_init(hvs_init); 926 module_exit(hvs_exit); 927 928 MODULE_DESCRIPTION("Hyper-V Sockets"); 929 MODULE_VERSION("1.0.0"); 930 MODULE_LICENSE("GPL"); 931 MODULE_ALIAS_NETPROTO(PF_VSOCK); 932