1 /* 2 * VMware vSockets Driver 3 * 4 * Copyright (C) 2007-2013 VMware, Inc. All rights reserved. 5 * 6 * This program is free software; you can redistribute it and/or modify it 7 * under the terms of the GNU General Public License as published by the Free 8 * Software Foundation version 2 and no later version. 9 * 10 * This program is distributed in the hope that it will be useful, but WITHOUT 11 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or 12 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for 13 * more details. 14 */ 15 16 #include <linux/types.h> 17 #include <linux/bitops.h> 18 #include <linux/cred.h> 19 #include <linux/init.h> 20 #include <linux/io.h> 21 #include <linux/kernel.h> 22 #include <linux/kmod.h> 23 #include <linux/list.h> 24 #include <linux/miscdevice.h> 25 #include <linux/module.h> 26 #include <linux/mutex.h> 27 #include <linux/net.h> 28 #include <linux/poll.h> 29 #include <linux/skbuff.h> 30 #include <linux/smp.h> 31 #include <linux/socket.h> 32 #include <linux/stddef.h> 33 #include <linux/unistd.h> 34 #include <linux/wait.h> 35 #include <linux/workqueue.h> 36 #include <net/sock.h> 37 #include <net/af_vsock.h> 38 39 #include "vmci_transport_notify.h" 40 41 static int vmci_transport_recv_dgram_cb(void *data, struct vmci_datagram *dg); 42 static int vmci_transport_recv_stream_cb(void *data, struct vmci_datagram *dg); 43 static void vmci_transport_peer_detach_cb(u32 sub_id, 44 const struct vmci_event_data *ed, 45 void *client_data); 46 static void vmci_transport_recv_pkt_work(struct work_struct *work); 47 static void vmci_transport_cleanup(struct work_struct *work); 48 static int vmci_transport_recv_listen(struct sock *sk, 49 struct vmci_transport_packet *pkt); 50 static int vmci_transport_recv_connecting_server( 51 struct sock *sk, 52 struct sock *pending, 53 struct vmci_transport_packet *pkt); 54 static int vmci_transport_recv_connecting_client( 55 struct sock *sk, 56 struct vmci_transport_packet *pkt); 57 static int vmci_transport_recv_connecting_client_negotiate( 58 struct sock *sk, 59 struct vmci_transport_packet *pkt); 60 static int vmci_transport_recv_connecting_client_invalid( 61 struct sock *sk, 62 struct vmci_transport_packet *pkt); 63 static int vmci_transport_recv_connected(struct sock *sk, 64 struct vmci_transport_packet *pkt); 65 static bool vmci_transport_old_proto_override(bool *old_pkt_proto); 66 static u16 vmci_transport_new_proto_supported_versions(void); 67 static bool vmci_transport_proto_to_notify_struct(struct sock *sk, u16 *proto, 68 bool old_pkt_proto); 69 70 struct vmci_transport_recv_pkt_info { 71 struct work_struct work; 72 struct sock *sk; 73 struct vmci_transport_packet pkt; 74 }; 75 76 static LIST_HEAD(vmci_transport_cleanup_list); 77 static DEFINE_SPINLOCK(vmci_transport_cleanup_lock); 78 static DECLARE_WORK(vmci_transport_cleanup_work, vmci_transport_cleanup); 79 80 static struct vmci_handle vmci_transport_stream_handle = { VMCI_INVALID_ID, 81 VMCI_INVALID_ID }; 82 static u32 vmci_transport_qp_resumed_sub_id = VMCI_INVALID_ID; 83 84 static int PROTOCOL_OVERRIDE = -1; 85 86 #define VMCI_TRANSPORT_DEFAULT_QP_SIZE_MIN 128 87 #define VMCI_TRANSPORT_DEFAULT_QP_SIZE 262144 88 #define VMCI_TRANSPORT_DEFAULT_QP_SIZE_MAX 262144 89 90 /* The default peer timeout indicates how long we will wait for a peer response 91 * to a control message. 92 */ 93 #define VSOCK_DEFAULT_CONNECT_TIMEOUT (2 * HZ) 94 95 /* Helper function to convert from a VMCI error code to a VSock error code. */ 96 97 static s32 vmci_transport_error_to_vsock_error(s32 vmci_error) 98 { 99 int err; 100 101 switch (vmci_error) { 102 case VMCI_ERROR_NO_MEM: 103 err = ENOMEM; 104 break; 105 case VMCI_ERROR_DUPLICATE_ENTRY: 106 case VMCI_ERROR_ALREADY_EXISTS: 107 err = EADDRINUSE; 108 break; 109 case VMCI_ERROR_NO_ACCESS: 110 err = EPERM; 111 break; 112 case VMCI_ERROR_NO_RESOURCES: 113 err = ENOBUFS; 114 break; 115 case VMCI_ERROR_INVALID_RESOURCE: 116 err = EHOSTUNREACH; 117 break; 118 case VMCI_ERROR_INVALID_ARGS: 119 default: 120 err = EINVAL; 121 } 122 123 return err > 0 ? -err : err; 124 } 125 126 static u32 vmci_transport_peer_rid(u32 peer_cid) 127 { 128 if (VMADDR_CID_HYPERVISOR == peer_cid) 129 return VMCI_TRANSPORT_HYPERVISOR_PACKET_RID; 130 131 return VMCI_TRANSPORT_PACKET_RID; 132 } 133 134 static inline void 135 vmci_transport_packet_init(struct vmci_transport_packet *pkt, 136 struct sockaddr_vm *src, 137 struct sockaddr_vm *dst, 138 u8 type, 139 u64 size, 140 u64 mode, 141 struct vmci_transport_waiting_info *wait, 142 u16 proto, 143 struct vmci_handle handle) 144 { 145 /* We register the stream control handler as an any cid handle so we 146 * must always send from a source address of VMADDR_CID_ANY 147 */ 148 pkt->dg.src = vmci_make_handle(VMADDR_CID_ANY, 149 VMCI_TRANSPORT_PACKET_RID); 150 pkt->dg.dst = vmci_make_handle(dst->svm_cid, 151 vmci_transport_peer_rid(dst->svm_cid)); 152 pkt->dg.payload_size = sizeof(*pkt) - sizeof(pkt->dg); 153 pkt->version = VMCI_TRANSPORT_PACKET_VERSION; 154 pkt->type = type; 155 pkt->src_port = src->svm_port; 156 pkt->dst_port = dst->svm_port; 157 memset(&pkt->proto, 0, sizeof(pkt->proto)); 158 memset(&pkt->_reserved2, 0, sizeof(pkt->_reserved2)); 159 160 switch (pkt->type) { 161 case VMCI_TRANSPORT_PACKET_TYPE_INVALID: 162 pkt->u.size = 0; 163 break; 164 165 case VMCI_TRANSPORT_PACKET_TYPE_REQUEST: 166 case VMCI_TRANSPORT_PACKET_TYPE_NEGOTIATE: 167 pkt->u.size = size; 168 break; 169 170 case VMCI_TRANSPORT_PACKET_TYPE_OFFER: 171 case VMCI_TRANSPORT_PACKET_TYPE_ATTACH: 172 pkt->u.handle = handle; 173 break; 174 175 case VMCI_TRANSPORT_PACKET_TYPE_WROTE: 176 case VMCI_TRANSPORT_PACKET_TYPE_READ: 177 case VMCI_TRANSPORT_PACKET_TYPE_RST: 178 pkt->u.size = 0; 179 break; 180 181 case VMCI_TRANSPORT_PACKET_TYPE_SHUTDOWN: 182 pkt->u.mode = mode; 183 break; 184 185 case VMCI_TRANSPORT_PACKET_TYPE_WAITING_READ: 186 case VMCI_TRANSPORT_PACKET_TYPE_WAITING_WRITE: 187 memcpy(&pkt->u.wait, wait, sizeof(pkt->u.wait)); 188 break; 189 190 case VMCI_TRANSPORT_PACKET_TYPE_REQUEST2: 191 case VMCI_TRANSPORT_PACKET_TYPE_NEGOTIATE2: 192 pkt->u.size = size; 193 pkt->proto = proto; 194 break; 195 } 196 } 197 198 static inline void 199 vmci_transport_packet_get_addresses(struct vmci_transport_packet *pkt, 200 struct sockaddr_vm *local, 201 struct sockaddr_vm *remote) 202 { 203 vsock_addr_init(local, pkt->dg.dst.context, pkt->dst_port); 204 vsock_addr_init(remote, pkt->dg.src.context, pkt->src_port); 205 } 206 207 static int 208 __vmci_transport_send_control_pkt(struct vmci_transport_packet *pkt, 209 struct sockaddr_vm *src, 210 struct sockaddr_vm *dst, 211 enum vmci_transport_packet_type type, 212 u64 size, 213 u64 mode, 214 struct vmci_transport_waiting_info *wait, 215 u16 proto, 216 struct vmci_handle handle, 217 bool convert_error) 218 { 219 int err; 220 221 vmci_transport_packet_init(pkt, src, dst, type, size, mode, wait, 222 proto, handle); 223 err = vmci_datagram_send(&pkt->dg); 224 if (convert_error && (err < 0)) 225 return vmci_transport_error_to_vsock_error(err); 226 227 return err; 228 } 229 230 static int 231 vmci_transport_reply_control_pkt_fast(struct vmci_transport_packet *pkt, 232 enum vmci_transport_packet_type type, 233 u64 size, 234 u64 mode, 235 struct vmci_transport_waiting_info *wait, 236 struct vmci_handle handle) 237 { 238 struct vmci_transport_packet reply; 239 struct sockaddr_vm src, dst; 240 241 if (pkt->type == VMCI_TRANSPORT_PACKET_TYPE_RST) { 242 return 0; 243 } else { 244 vmci_transport_packet_get_addresses(pkt, &src, &dst); 245 return __vmci_transport_send_control_pkt(&reply, &src, &dst, 246 type, 247 size, mode, wait, 248 VSOCK_PROTO_INVALID, 249 handle, true); 250 } 251 } 252 253 static int 254 vmci_transport_send_control_pkt_bh(struct sockaddr_vm *src, 255 struct sockaddr_vm *dst, 256 enum vmci_transport_packet_type type, 257 u64 size, 258 u64 mode, 259 struct vmci_transport_waiting_info *wait, 260 struct vmci_handle handle) 261 { 262 /* Note that it is safe to use a single packet across all CPUs since 263 * two tasklets of the same type are guaranteed to not ever run 264 * simultaneously. If that ever changes, or VMCI stops using tasklets, 265 * we can use per-cpu packets. 266 */ 267 static struct vmci_transport_packet pkt; 268 269 return __vmci_transport_send_control_pkt(&pkt, src, dst, type, 270 size, mode, wait, 271 VSOCK_PROTO_INVALID, handle, 272 false); 273 } 274 275 static int 276 vmci_transport_send_control_pkt(struct sock *sk, 277 enum vmci_transport_packet_type type, 278 u64 size, 279 u64 mode, 280 struct vmci_transport_waiting_info *wait, 281 u16 proto, 282 struct vmci_handle handle) 283 { 284 struct vmci_transport_packet *pkt; 285 struct vsock_sock *vsk; 286 int err; 287 288 vsk = vsock_sk(sk); 289 290 if (!vsock_addr_bound(&vsk->local_addr)) 291 return -EINVAL; 292 293 if (!vsock_addr_bound(&vsk->remote_addr)) 294 return -EINVAL; 295 296 pkt = kmalloc(sizeof(*pkt), GFP_KERNEL); 297 if (!pkt) 298 return -ENOMEM; 299 300 err = __vmci_transport_send_control_pkt(pkt, &vsk->local_addr, 301 &vsk->remote_addr, type, size, 302 mode, wait, proto, handle, 303 true); 304 kfree(pkt); 305 306 return err; 307 } 308 309 static int vmci_transport_send_reset_bh(struct sockaddr_vm *dst, 310 struct sockaddr_vm *src, 311 struct vmci_transport_packet *pkt) 312 { 313 if (pkt->type == VMCI_TRANSPORT_PACKET_TYPE_RST) 314 return 0; 315 return vmci_transport_send_control_pkt_bh( 316 dst, src, 317 VMCI_TRANSPORT_PACKET_TYPE_RST, 0, 318 0, NULL, VMCI_INVALID_HANDLE); 319 } 320 321 static int vmci_transport_send_reset(struct sock *sk, 322 struct vmci_transport_packet *pkt) 323 { 324 if (pkt->type == VMCI_TRANSPORT_PACKET_TYPE_RST) 325 return 0; 326 return vmci_transport_send_control_pkt(sk, 327 VMCI_TRANSPORT_PACKET_TYPE_RST, 328 0, 0, NULL, VSOCK_PROTO_INVALID, 329 VMCI_INVALID_HANDLE); 330 } 331 332 static int vmci_transport_send_negotiate(struct sock *sk, size_t size) 333 { 334 return vmci_transport_send_control_pkt( 335 sk, 336 VMCI_TRANSPORT_PACKET_TYPE_NEGOTIATE, 337 size, 0, NULL, 338 VSOCK_PROTO_INVALID, 339 VMCI_INVALID_HANDLE); 340 } 341 342 static int vmci_transport_send_negotiate2(struct sock *sk, size_t size, 343 u16 version) 344 { 345 return vmci_transport_send_control_pkt( 346 sk, 347 VMCI_TRANSPORT_PACKET_TYPE_NEGOTIATE2, 348 size, 0, NULL, version, 349 VMCI_INVALID_HANDLE); 350 } 351 352 static int vmci_transport_send_qp_offer(struct sock *sk, 353 struct vmci_handle handle) 354 { 355 return vmci_transport_send_control_pkt( 356 sk, VMCI_TRANSPORT_PACKET_TYPE_OFFER, 0, 357 0, NULL, 358 VSOCK_PROTO_INVALID, handle); 359 } 360 361 static int vmci_transport_send_attach(struct sock *sk, 362 struct vmci_handle handle) 363 { 364 return vmci_transport_send_control_pkt( 365 sk, VMCI_TRANSPORT_PACKET_TYPE_ATTACH, 366 0, 0, NULL, VSOCK_PROTO_INVALID, 367 handle); 368 } 369 370 static int vmci_transport_reply_reset(struct vmci_transport_packet *pkt) 371 { 372 return vmci_transport_reply_control_pkt_fast( 373 pkt, 374 VMCI_TRANSPORT_PACKET_TYPE_RST, 375 0, 0, NULL, 376 VMCI_INVALID_HANDLE); 377 } 378 379 static int vmci_transport_send_invalid_bh(struct sockaddr_vm *dst, 380 struct sockaddr_vm *src) 381 { 382 return vmci_transport_send_control_pkt_bh( 383 dst, src, 384 VMCI_TRANSPORT_PACKET_TYPE_INVALID, 385 0, 0, NULL, VMCI_INVALID_HANDLE); 386 } 387 388 int vmci_transport_send_wrote_bh(struct sockaddr_vm *dst, 389 struct sockaddr_vm *src) 390 { 391 return vmci_transport_send_control_pkt_bh( 392 dst, src, 393 VMCI_TRANSPORT_PACKET_TYPE_WROTE, 0, 394 0, NULL, VMCI_INVALID_HANDLE); 395 } 396 397 int vmci_transport_send_read_bh(struct sockaddr_vm *dst, 398 struct sockaddr_vm *src) 399 { 400 return vmci_transport_send_control_pkt_bh( 401 dst, src, 402 VMCI_TRANSPORT_PACKET_TYPE_READ, 0, 403 0, NULL, VMCI_INVALID_HANDLE); 404 } 405 406 int vmci_transport_send_wrote(struct sock *sk) 407 { 408 return vmci_transport_send_control_pkt( 409 sk, VMCI_TRANSPORT_PACKET_TYPE_WROTE, 0, 410 0, NULL, VSOCK_PROTO_INVALID, 411 VMCI_INVALID_HANDLE); 412 } 413 414 int vmci_transport_send_read(struct sock *sk) 415 { 416 return vmci_transport_send_control_pkt( 417 sk, VMCI_TRANSPORT_PACKET_TYPE_READ, 0, 418 0, NULL, VSOCK_PROTO_INVALID, 419 VMCI_INVALID_HANDLE); 420 } 421 422 int vmci_transport_send_waiting_write(struct sock *sk, 423 struct vmci_transport_waiting_info *wait) 424 { 425 return vmci_transport_send_control_pkt( 426 sk, VMCI_TRANSPORT_PACKET_TYPE_WAITING_WRITE, 427 0, 0, wait, VSOCK_PROTO_INVALID, 428 VMCI_INVALID_HANDLE); 429 } 430 431 int vmci_transport_send_waiting_read(struct sock *sk, 432 struct vmci_transport_waiting_info *wait) 433 { 434 return vmci_transport_send_control_pkt( 435 sk, VMCI_TRANSPORT_PACKET_TYPE_WAITING_READ, 436 0, 0, wait, VSOCK_PROTO_INVALID, 437 VMCI_INVALID_HANDLE); 438 } 439 440 static int vmci_transport_shutdown(struct vsock_sock *vsk, int mode) 441 { 442 return vmci_transport_send_control_pkt( 443 &vsk->sk, 444 VMCI_TRANSPORT_PACKET_TYPE_SHUTDOWN, 445 0, mode, NULL, 446 VSOCK_PROTO_INVALID, 447 VMCI_INVALID_HANDLE); 448 } 449 450 static int vmci_transport_send_conn_request(struct sock *sk, size_t size) 451 { 452 return vmci_transport_send_control_pkt(sk, 453 VMCI_TRANSPORT_PACKET_TYPE_REQUEST, 454 size, 0, NULL, 455 VSOCK_PROTO_INVALID, 456 VMCI_INVALID_HANDLE); 457 } 458 459 static int vmci_transport_send_conn_request2(struct sock *sk, size_t size, 460 u16 version) 461 { 462 return vmci_transport_send_control_pkt( 463 sk, VMCI_TRANSPORT_PACKET_TYPE_REQUEST2, 464 size, 0, NULL, version, 465 VMCI_INVALID_HANDLE); 466 } 467 468 static struct sock *vmci_transport_get_pending( 469 struct sock *listener, 470 struct vmci_transport_packet *pkt) 471 { 472 struct vsock_sock *vlistener; 473 struct vsock_sock *vpending; 474 struct sock *pending; 475 struct sockaddr_vm src; 476 477 vsock_addr_init(&src, pkt->dg.src.context, pkt->src_port); 478 479 vlistener = vsock_sk(listener); 480 481 list_for_each_entry(vpending, &vlistener->pending_links, 482 pending_links) { 483 if (vsock_addr_equals_addr(&src, &vpending->remote_addr) && 484 pkt->dst_port == vpending->local_addr.svm_port) { 485 pending = sk_vsock(vpending); 486 sock_hold(pending); 487 goto found; 488 } 489 } 490 491 pending = NULL; 492 found: 493 return pending; 494 495 } 496 497 static void vmci_transport_release_pending(struct sock *pending) 498 { 499 sock_put(pending); 500 } 501 502 /* We allow two kinds of sockets to communicate with a restricted VM: 1) 503 * trusted sockets 2) sockets from applications running as the same user as the 504 * VM (this is only true for the host side and only when using hosted products) 505 */ 506 507 static bool vmci_transport_is_trusted(struct vsock_sock *vsock, u32 peer_cid) 508 { 509 return vsock->trusted || 510 vmci_is_context_owner(peer_cid, vsock->owner->uid); 511 } 512 513 /* We allow sending datagrams to and receiving datagrams from a restricted VM 514 * only if it is trusted as described in vmci_transport_is_trusted. 515 */ 516 517 static bool vmci_transport_allow_dgram(struct vsock_sock *vsock, u32 peer_cid) 518 { 519 if (VMADDR_CID_HYPERVISOR == peer_cid) 520 return true; 521 522 if (vsock->cached_peer != peer_cid) { 523 vsock->cached_peer = peer_cid; 524 if (!vmci_transport_is_trusted(vsock, peer_cid) && 525 (vmci_context_get_priv_flags(peer_cid) & 526 VMCI_PRIVILEGE_FLAG_RESTRICTED)) { 527 vsock->cached_peer_allow_dgram = false; 528 } else { 529 vsock->cached_peer_allow_dgram = true; 530 } 531 } 532 533 return vsock->cached_peer_allow_dgram; 534 } 535 536 static int 537 vmci_transport_queue_pair_alloc(struct vmci_qp **qpair, 538 struct vmci_handle *handle, 539 u64 produce_size, 540 u64 consume_size, 541 u32 peer, u32 flags, bool trusted) 542 { 543 int err = 0; 544 545 if (trusted) { 546 /* Try to allocate our queue pair as trusted. This will only 547 * work if vsock is running in the host. 548 */ 549 550 err = vmci_qpair_alloc(qpair, handle, produce_size, 551 consume_size, 552 peer, flags, 553 VMCI_PRIVILEGE_FLAG_TRUSTED); 554 if (err != VMCI_ERROR_NO_ACCESS) 555 goto out; 556 557 } 558 559 err = vmci_qpair_alloc(qpair, handle, produce_size, consume_size, 560 peer, flags, VMCI_NO_PRIVILEGE_FLAGS); 561 out: 562 if (err < 0) { 563 pr_err("Could not attach to queue pair with %d\n", 564 err); 565 err = vmci_transport_error_to_vsock_error(err); 566 } 567 568 return err; 569 } 570 571 static int 572 vmci_transport_datagram_create_hnd(u32 resource_id, 573 u32 flags, 574 vmci_datagram_recv_cb recv_cb, 575 void *client_data, 576 struct vmci_handle *out_handle) 577 { 578 int err = 0; 579 580 /* Try to allocate our datagram handler as trusted. This will only work 581 * if vsock is running in the host. 582 */ 583 584 err = vmci_datagram_create_handle_priv(resource_id, flags, 585 VMCI_PRIVILEGE_FLAG_TRUSTED, 586 recv_cb, 587 client_data, out_handle); 588 589 if (err == VMCI_ERROR_NO_ACCESS) 590 err = vmci_datagram_create_handle(resource_id, flags, 591 recv_cb, client_data, 592 out_handle); 593 594 return err; 595 } 596 597 /* This is invoked as part of a tasklet that's scheduled when the VMCI 598 * interrupt fires. This is run in bottom-half context and if it ever needs to 599 * sleep it should defer that work to a work queue. 600 */ 601 602 static int vmci_transport_recv_dgram_cb(void *data, struct vmci_datagram *dg) 603 { 604 struct sock *sk; 605 size_t size; 606 struct sk_buff *skb; 607 struct vsock_sock *vsk; 608 609 sk = (struct sock *)data; 610 611 /* This handler is privileged when this module is running on the host. 612 * We will get datagrams from all endpoints (even VMs that are in a 613 * restricted context). If we get one from a restricted context then 614 * the destination socket must be trusted. 615 * 616 * NOTE: We access the socket struct without holding the lock here. 617 * This is ok because the field we are interested is never modified 618 * outside of the create and destruct socket functions. 619 */ 620 vsk = vsock_sk(sk); 621 if (!vmci_transport_allow_dgram(vsk, dg->src.context)) 622 return VMCI_ERROR_NO_ACCESS; 623 624 size = VMCI_DG_SIZE(dg); 625 626 /* Attach the packet to the socket's receive queue as an sk_buff. */ 627 skb = alloc_skb(size, GFP_ATOMIC); 628 if (!skb) 629 return VMCI_ERROR_NO_MEM; 630 631 /* sk_receive_skb() will do a sock_put(), so hold here. */ 632 sock_hold(sk); 633 skb_put(skb, size); 634 memcpy(skb->data, dg, size); 635 sk_receive_skb(sk, skb, 0); 636 637 return VMCI_SUCCESS; 638 } 639 640 static bool vmci_transport_stream_allow(u32 cid, u32 port) 641 { 642 static const u32 non_socket_contexts[] = { 643 VMADDR_CID_RESERVED, 644 }; 645 int i; 646 647 BUILD_BUG_ON(sizeof(cid) != sizeof(*non_socket_contexts)); 648 649 for (i = 0; i < ARRAY_SIZE(non_socket_contexts); i++) { 650 if (cid == non_socket_contexts[i]) 651 return false; 652 } 653 654 return true; 655 } 656 657 /* This is invoked as part of a tasklet that's scheduled when the VMCI 658 * interrupt fires. This is run in bottom-half context but it defers most of 659 * its work to the packet handling work queue. 660 */ 661 662 static int vmci_transport_recv_stream_cb(void *data, struct vmci_datagram *dg) 663 { 664 struct sock *sk; 665 struct sockaddr_vm dst; 666 struct sockaddr_vm src; 667 struct vmci_transport_packet *pkt; 668 struct vsock_sock *vsk; 669 bool bh_process_pkt; 670 int err; 671 672 sk = NULL; 673 err = VMCI_SUCCESS; 674 bh_process_pkt = false; 675 676 /* Ignore incoming packets from contexts without sockets, or resources 677 * that aren't vsock implementations. 678 */ 679 680 if (!vmci_transport_stream_allow(dg->src.context, -1) 681 || vmci_transport_peer_rid(dg->src.context) != dg->src.resource) 682 return VMCI_ERROR_NO_ACCESS; 683 684 if (VMCI_DG_SIZE(dg) < sizeof(*pkt)) 685 /* Drop datagrams that do not contain full VSock packets. */ 686 return VMCI_ERROR_INVALID_ARGS; 687 688 pkt = (struct vmci_transport_packet *)dg; 689 690 /* Find the socket that should handle this packet. First we look for a 691 * connected socket and if there is none we look for a socket bound to 692 * the destintation address. 693 */ 694 vsock_addr_init(&src, pkt->dg.src.context, pkt->src_port); 695 vsock_addr_init(&dst, pkt->dg.dst.context, pkt->dst_port); 696 697 sk = vsock_find_connected_socket(&src, &dst); 698 if (!sk) { 699 sk = vsock_find_bound_socket(&dst); 700 if (!sk) { 701 /* We could not find a socket for this specified 702 * address. If this packet is a RST, we just drop it. 703 * If it is another packet, we send a RST. Note that 704 * we do not send a RST reply to RSTs so that we do not 705 * continually send RSTs between two endpoints. 706 * 707 * Note that since this is a reply, dst is src and src 708 * is dst. 709 */ 710 if (vmci_transport_send_reset_bh(&dst, &src, pkt) < 0) 711 pr_err("unable to send reset\n"); 712 713 err = VMCI_ERROR_NOT_FOUND; 714 goto out; 715 } 716 } 717 718 /* If the received packet type is beyond all types known to this 719 * implementation, reply with an invalid message. Hopefully this will 720 * help when implementing backwards compatibility in the future. 721 */ 722 if (pkt->type >= VMCI_TRANSPORT_PACKET_TYPE_MAX) { 723 vmci_transport_send_invalid_bh(&dst, &src); 724 err = VMCI_ERROR_INVALID_ARGS; 725 goto out; 726 } 727 728 /* This handler is privileged when this module is running on the host. 729 * We will get datagram connect requests from all endpoints (even VMs 730 * that are in a restricted context). If we get one from a restricted 731 * context then the destination socket must be trusted. 732 * 733 * NOTE: We access the socket struct without holding the lock here. 734 * This is ok because the field we are interested is never modified 735 * outside of the create and destruct socket functions. 736 */ 737 vsk = vsock_sk(sk); 738 if (!vmci_transport_allow_dgram(vsk, pkt->dg.src.context)) { 739 err = VMCI_ERROR_NO_ACCESS; 740 goto out; 741 } 742 743 /* We do most everything in a work queue, but let's fast path the 744 * notification of reads and writes to help data transfer performance. 745 * We can only do this if there is no process context code executing 746 * for this socket since that may change the state. 747 */ 748 bh_lock_sock(sk); 749 750 if (!sock_owned_by_user(sk)) { 751 /* The local context ID may be out of date, update it. */ 752 vsk->local_addr.svm_cid = dst.svm_cid; 753 754 if (sk->sk_state == SS_CONNECTED) 755 vmci_trans(vsk)->notify_ops->handle_notify_pkt( 756 sk, pkt, true, &dst, &src, 757 &bh_process_pkt); 758 } 759 760 bh_unlock_sock(sk); 761 762 if (!bh_process_pkt) { 763 struct vmci_transport_recv_pkt_info *recv_pkt_info; 764 765 recv_pkt_info = kmalloc(sizeof(*recv_pkt_info), GFP_ATOMIC); 766 if (!recv_pkt_info) { 767 if (vmci_transport_send_reset_bh(&dst, &src, pkt) < 0) 768 pr_err("unable to send reset\n"); 769 770 err = VMCI_ERROR_NO_MEM; 771 goto out; 772 } 773 774 recv_pkt_info->sk = sk; 775 memcpy(&recv_pkt_info->pkt, pkt, sizeof(recv_pkt_info->pkt)); 776 INIT_WORK(&recv_pkt_info->work, vmci_transport_recv_pkt_work); 777 778 schedule_work(&recv_pkt_info->work); 779 /* Clear sk so that the reference count incremented by one of 780 * the Find functions above is not decremented below. We need 781 * that reference count for the packet handler we've scheduled 782 * to run. 783 */ 784 sk = NULL; 785 } 786 787 out: 788 if (sk) 789 sock_put(sk); 790 791 return err; 792 } 793 794 static void vmci_transport_handle_detach(struct sock *sk) 795 { 796 struct vsock_sock *vsk; 797 798 vsk = vsock_sk(sk); 799 if (!vmci_handle_is_invalid(vmci_trans(vsk)->qp_handle)) { 800 sock_set_flag(sk, SOCK_DONE); 801 802 /* On a detach the peer will not be sending or receiving 803 * anymore. 804 */ 805 vsk->peer_shutdown = SHUTDOWN_MASK; 806 807 /* We should not be sending anymore since the peer won't be 808 * there to receive, but we can still receive if there is data 809 * left in our consume queue. 810 */ 811 if (vsock_stream_has_data(vsk) <= 0) { 812 if (sk->sk_state == SS_CONNECTING) { 813 /* The peer may detach from a queue pair while 814 * we are still in the connecting state, i.e., 815 * if the peer VM is killed after attaching to 816 * a queue pair, but before we complete the 817 * handshake. In that case, we treat the detach 818 * event like a reset. 819 */ 820 821 sk->sk_state = SS_UNCONNECTED; 822 sk->sk_err = ECONNRESET; 823 sk->sk_error_report(sk); 824 return; 825 } 826 sk->sk_state = SS_UNCONNECTED; 827 } 828 sk->sk_state_change(sk); 829 } 830 } 831 832 static void vmci_transport_peer_detach_cb(u32 sub_id, 833 const struct vmci_event_data *e_data, 834 void *client_data) 835 { 836 struct vmci_transport *trans = client_data; 837 const struct vmci_event_payload_qp *e_payload; 838 839 e_payload = vmci_event_data_const_payload(e_data); 840 841 /* XXX This is lame, we should provide a way to lookup sockets by 842 * qp_handle. 843 */ 844 if (vmci_handle_is_invalid(e_payload->handle) || 845 vmci_handle_is_equal(trans->qp_handle, e_payload->handle)) 846 return; 847 848 /* We don't ask for delayed CBs when we subscribe to this event (we 849 * pass 0 as flags to vmci_event_subscribe()). VMCI makes no 850 * guarantees in that case about what context we might be running in, 851 * so it could be BH or process, blockable or non-blockable. So we 852 * need to account for all possible contexts here. 853 */ 854 spin_lock_bh(&trans->lock); 855 if (!trans->sk) 856 goto out; 857 858 /* Apart from here, trans->lock is only grabbed as part of sk destruct, 859 * where trans->sk isn't locked. 860 */ 861 bh_lock_sock(trans->sk); 862 863 vmci_transport_handle_detach(trans->sk); 864 865 bh_unlock_sock(trans->sk); 866 out: 867 spin_unlock_bh(&trans->lock); 868 } 869 870 static void vmci_transport_qp_resumed_cb(u32 sub_id, 871 const struct vmci_event_data *e_data, 872 void *client_data) 873 { 874 vsock_for_each_connected_socket(vmci_transport_handle_detach); 875 } 876 877 static void vmci_transport_recv_pkt_work(struct work_struct *work) 878 { 879 struct vmci_transport_recv_pkt_info *recv_pkt_info; 880 struct vmci_transport_packet *pkt; 881 struct sock *sk; 882 883 recv_pkt_info = 884 container_of(work, struct vmci_transport_recv_pkt_info, work); 885 sk = recv_pkt_info->sk; 886 pkt = &recv_pkt_info->pkt; 887 888 lock_sock(sk); 889 890 /* The local context ID may be out of date. */ 891 vsock_sk(sk)->local_addr.svm_cid = pkt->dg.dst.context; 892 893 switch (sk->sk_state) { 894 case VSOCK_SS_LISTEN: 895 vmci_transport_recv_listen(sk, pkt); 896 break; 897 case SS_CONNECTING: 898 /* Processing of pending connections for servers goes through 899 * the listening socket, so see vmci_transport_recv_listen() 900 * for that path. 901 */ 902 vmci_transport_recv_connecting_client(sk, pkt); 903 break; 904 case SS_CONNECTED: 905 vmci_transport_recv_connected(sk, pkt); 906 break; 907 default: 908 /* Because this function does not run in the same context as 909 * vmci_transport_recv_stream_cb it is possible that the 910 * socket has closed. We need to let the other side know or it 911 * could be sitting in a connect and hang forever. Send a 912 * reset to prevent that. 913 */ 914 vmci_transport_send_reset(sk, pkt); 915 break; 916 } 917 918 release_sock(sk); 919 kfree(recv_pkt_info); 920 /* Release reference obtained in the stream callback when we fetched 921 * this socket out of the bound or connected list. 922 */ 923 sock_put(sk); 924 } 925 926 static int vmci_transport_recv_listen(struct sock *sk, 927 struct vmci_transport_packet *pkt) 928 { 929 struct sock *pending; 930 struct vsock_sock *vpending; 931 int err; 932 u64 qp_size; 933 bool old_request = false; 934 bool old_pkt_proto = false; 935 936 err = 0; 937 938 /* Because we are in the listen state, we could be receiving a packet 939 * for ourself or any previous connection requests that we received. 940 * If it's the latter, we try to find a socket in our list of pending 941 * connections and, if we do, call the appropriate handler for the 942 * state that that socket is in. Otherwise we try to service the 943 * connection request. 944 */ 945 pending = vmci_transport_get_pending(sk, pkt); 946 if (pending) { 947 lock_sock(pending); 948 949 /* The local context ID may be out of date. */ 950 vsock_sk(pending)->local_addr.svm_cid = pkt->dg.dst.context; 951 952 switch (pending->sk_state) { 953 case SS_CONNECTING: 954 err = vmci_transport_recv_connecting_server(sk, 955 pending, 956 pkt); 957 break; 958 default: 959 vmci_transport_send_reset(pending, pkt); 960 err = -EINVAL; 961 } 962 963 if (err < 0) 964 vsock_remove_pending(sk, pending); 965 966 release_sock(pending); 967 vmci_transport_release_pending(pending); 968 969 return err; 970 } 971 972 /* The listen state only accepts connection requests. Reply with a 973 * reset unless we received a reset. 974 */ 975 976 if (!(pkt->type == VMCI_TRANSPORT_PACKET_TYPE_REQUEST || 977 pkt->type == VMCI_TRANSPORT_PACKET_TYPE_REQUEST2)) { 978 vmci_transport_reply_reset(pkt); 979 return -EINVAL; 980 } 981 982 if (pkt->u.size == 0) { 983 vmci_transport_reply_reset(pkt); 984 return -EINVAL; 985 } 986 987 /* If this socket can't accommodate this connection request, we send a 988 * reset. Otherwise we create and initialize a child socket and reply 989 * with a connection negotiation. 990 */ 991 if (sk->sk_ack_backlog >= sk->sk_max_ack_backlog) { 992 vmci_transport_reply_reset(pkt); 993 return -ECONNREFUSED; 994 } 995 996 pending = __vsock_create(sock_net(sk), NULL, sk, GFP_KERNEL, 997 sk->sk_type, 0); 998 if (!pending) { 999 vmci_transport_send_reset(sk, pkt); 1000 return -ENOMEM; 1001 } 1002 1003 vpending = vsock_sk(pending); 1004 1005 vsock_addr_init(&vpending->local_addr, pkt->dg.dst.context, 1006 pkt->dst_port); 1007 vsock_addr_init(&vpending->remote_addr, pkt->dg.src.context, 1008 pkt->src_port); 1009 1010 /* If the proposed size fits within our min/max, accept it. Otherwise 1011 * propose our own size. 1012 */ 1013 if (pkt->u.size >= vmci_trans(vpending)->queue_pair_min_size && 1014 pkt->u.size <= vmci_trans(vpending)->queue_pair_max_size) { 1015 qp_size = pkt->u.size; 1016 } else { 1017 qp_size = vmci_trans(vpending)->queue_pair_size; 1018 } 1019 1020 /* Figure out if we are using old or new requests based on the 1021 * overrides pkt types sent by our peer. 1022 */ 1023 if (vmci_transport_old_proto_override(&old_pkt_proto)) { 1024 old_request = old_pkt_proto; 1025 } else { 1026 if (pkt->type == VMCI_TRANSPORT_PACKET_TYPE_REQUEST) 1027 old_request = true; 1028 else if (pkt->type == VMCI_TRANSPORT_PACKET_TYPE_REQUEST2) 1029 old_request = false; 1030 1031 } 1032 1033 if (old_request) { 1034 /* Handle a REQUEST (or override) */ 1035 u16 version = VSOCK_PROTO_INVALID; 1036 if (vmci_transport_proto_to_notify_struct( 1037 pending, &version, true)) 1038 err = vmci_transport_send_negotiate(pending, qp_size); 1039 else 1040 err = -EINVAL; 1041 1042 } else { 1043 /* Handle a REQUEST2 (or override) */ 1044 int proto_int = pkt->proto; 1045 int pos; 1046 u16 active_proto_version = 0; 1047 1048 /* The list of possible protocols is the intersection of all 1049 * protocols the client supports ... plus all the protocols we 1050 * support. 1051 */ 1052 proto_int &= vmci_transport_new_proto_supported_versions(); 1053 1054 /* We choose the highest possible protocol version and use that 1055 * one. 1056 */ 1057 pos = fls(proto_int); 1058 if (pos) { 1059 active_proto_version = (1 << (pos - 1)); 1060 if (vmci_transport_proto_to_notify_struct( 1061 pending, &active_proto_version, false)) 1062 err = vmci_transport_send_negotiate2(pending, 1063 qp_size, 1064 active_proto_version); 1065 else 1066 err = -EINVAL; 1067 1068 } else { 1069 err = -EINVAL; 1070 } 1071 } 1072 1073 if (err < 0) { 1074 vmci_transport_send_reset(sk, pkt); 1075 sock_put(pending); 1076 err = vmci_transport_error_to_vsock_error(err); 1077 goto out; 1078 } 1079 1080 vsock_add_pending(sk, pending); 1081 sk->sk_ack_backlog++; 1082 1083 pending->sk_state = SS_CONNECTING; 1084 vmci_trans(vpending)->produce_size = 1085 vmci_trans(vpending)->consume_size = qp_size; 1086 vmci_trans(vpending)->queue_pair_size = qp_size; 1087 1088 vmci_trans(vpending)->notify_ops->process_request(pending); 1089 1090 /* We might never receive another message for this socket and it's not 1091 * connected to any process, so we have to ensure it gets cleaned up 1092 * ourself. Our delayed work function will take care of that. Note 1093 * that we do not ever cancel this function since we have few 1094 * guarantees about its state when calling cancel_delayed_work(). 1095 * Instead we hold a reference on the socket for that function and make 1096 * it capable of handling cases where it needs to do nothing but 1097 * release that reference. 1098 */ 1099 vpending->listener = sk; 1100 sock_hold(sk); 1101 sock_hold(pending); 1102 INIT_DELAYED_WORK(&vpending->dwork, vsock_pending_work); 1103 schedule_delayed_work(&vpending->dwork, HZ); 1104 1105 out: 1106 return err; 1107 } 1108 1109 static int 1110 vmci_transport_recv_connecting_server(struct sock *listener, 1111 struct sock *pending, 1112 struct vmci_transport_packet *pkt) 1113 { 1114 struct vsock_sock *vpending; 1115 struct vmci_handle handle; 1116 struct vmci_qp *qpair; 1117 bool is_local; 1118 u32 flags; 1119 u32 detach_sub_id; 1120 int err; 1121 int skerr; 1122 1123 vpending = vsock_sk(pending); 1124 detach_sub_id = VMCI_INVALID_ID; 1125 1126 switch (pkt->type) { 1127 case VMCI_TRANSPORT_PACKET_TYPE_OFFER: 1128 if (vmci_handle_is_invalid(pkt->u.handle)) { 1129 vmci_transport_send_reset(pending, pkt); 1130 skerr = EPROTO; 1131 err = -EINVAL; 1132 goto destroy; 1133 } 1134 break; 1135 default: 1136 /* Close and cleanup the connection. */ 1137 vmci_transport_send_reset(pending, pkt); 1138 skerr = EPROTO; 1139 err = pkt->type == VMCI_TRANSPORT_PACKET_TYPE_RST ? 0 : -EINVAL; 1140 goto destroy; 1141 } 1142 1143 /* In order to complete the connection we need to attach to the offered 1144 * queue pair and send an attach notification. We also subscribe to the 1145 * detach event so we know when our peer goes away, and we do that 1146 * before attaching so we don't miss an event. If all this succeeds, 1147 * we update our state and wakeup anything waiting in accept() for a 1148 * connection. 1149 */ 1150 1151 /* We don't care about attach since we ensure the other side has 1152 * attached by specifying the ATTACH_ONLY flag below. 1153 */ 1154 err = vmci_event_subscribe(VMCI_EVENT_QP_PEER_DETACH, 1155 vmci_transport_peer_detach_cb, 1156 vmci_trans(vpending), &detach_sub_id); 1157 if (err < VMCI_SUCCESS) { 1158 vmci_transport_send_reset(pending, pkt); 1159 err = vmci_transport_error_to_vsock_error(err); 1160 skerr = -err; 1161 goto destroy; 1162 } 1163 1164 vmci_trans(vpending)->detach_sub_id = detach_sub_id; 1165 1166 /* Now attach to the queue pair the client created. */ 1167 handle = pkt->u.handle; 1168 1169 /* vpending->local_addr always has a context id so we do not need to 1170 * worry about VMADDR_CID_ANY in this case. 1171 */ 1172 is_local = 1173 vpending->remote_addr.svm_cid == vpending->local_addr.svm_cid; 1174 flags = VMCI_QPFLAG_ATTACH_ONLY; 1175 flags |= is_local ? VMCI_QPFLAG_LOCAL : 0; 1176 1177 err = vmci_transport_queue_pair_alloc( 1178 &qpair, 1179 &handle, 1180 vmci_trans(vpending)->produce_size, 1181 vmci_trans(vpending)->consume_size, 1182 pkt->dg.src.context, 1183 flags, 1184 vmci_transport_is_trusted( 1185 vpending, 1186 vpending->remote_addr.svm_cid)); 1187 if (err < 0) { 1188 vmci_transport_send_reset(pending, pkt); 1189 skerr = -err; 1190 goto destroy; 1191 } 1192 1193 vmci_trans(vpending)->qp_handle = handle; 1194 vmci_trans(vpending)->qpair = qpair; 1195 1196 /* When we send the attach message, we must be ready to handle incoming 1197 * control messages on the newly connected socket. So we move the 1198 * pending socket to the connected state before sending the attach 1199 * message. Otherwise, an incoming packet triggered by the attach being 1200 * received by the peer may be processed concurrently with what happens 1201 * below after sending the attach message, and that incoming packet 1202 * will find the listening socket instead of the (currently) pending 1203 * socket. Note that enqueueing the socket increments the reference 1204 * count, so even if a reset comes before the connection is accepted, 1205 * the socket will be valid until it is removed from the queue. 1206 * 1207 * If we fail sending the attach below, we remove the socket from the 1208 * connected list and move the socket to SS_UNCONNECTED before 1209 * releasing the lock, so a pending slow path processing of an incoming 1210 * packet will not see the socket in the connected state in that case. 1211 */ 1212 pending->sk_state = SS_CONNECTED; 1213 1214 vsock_insert_connected(vpending); 1215 1216 /* Notify our peer of our attach. */ 1217 err = vmci_transport_send_attach(pending, handle); 1218 if (err < 0) { 1219 vsock_remove_connected(vpending); 1220 pr_err("Could not send attach\n"); 1221 vmci_transport_send_reset(pending, pkt); 1222 err = vmci_transport_error_to_vsock_error(err); 1223 skerr = -err; 1224 goto destroy; 1225 } 1226 1227 /* We have a connection. Move the now connected socket from the 1228 * listener's pending list to the accept queue so callers of accept() 1229 * can find it. 1230 */ 1231 vsock_remove_pending(listener, pending); 1232 vsock_enqueue_accept(listener, pending); 1233 1234 /* Callers of accept() will be be waiting on the listening socket, not 1235 * the pending socket. 1236 */ 1237 listener->sk_data_ready(listener); 1238 1239 return 0; 1240 1241 destroy: 1242 pending->sk_err = skerr; 1243 pending->sk_state = SS_UNCONNECTED; 1244 /* As long as we drop our reference, all necessary cleanup will handle 1245 * when the cleanup function drops its reference and our destruct 1246 * implementation is called. Note that since the listen handler will 1247 * remove pending from the pending list upon our failure, the cleanup 1248 * function won't drop the additional reference, which is why we do it 1249 * here. 1250 */ 1251 sock_put(pending); 1252 1253 return err; 1254 } 1255 1256 static int 1257 vmci_transport_recv_connecting_client(struct sock *sk, 1258 struct vmci_transport_packet *pkt) 1259 { 1260 struct vsock_sock *vsk; 1261 int err; 1262 int skerr; 1263 1264 vsk = vsock_sk(sk); 1265 1266 switch (pkt->type) { 1267 case VMCI_TRANSPORT_PACKET_TYPE_ATTACH: 1268 if (vmci_handle_is_invalid(pkt->u.handle) || 1269 !vmci_handle_is_equal(pkt->u.handle, 1270 vmci_trans(vsk)->qp_handle)) { 1271 skerr = EPROTO; 1272 err = -EINVAL; 1273 goto destroy; 1274 } 1275 1276 /* Signify the socket is connected and wakeup the waiter in 1277 * connect(). Also place the socket in the connected table for 1278 * accounting (it can already be found since it's in the bound 1279 * table). 1280 */ 1281 sk->sk_state = SS_CONNECTED; 1282 sk->sk_socket->state = SS_CONNECTED; 1283 vsock_insert_connected(vsk); 1284 sk->sk_state_change(sk); 1285 1286 break; 1287 case VMCI_TRANSPORT_PACKET_TYPE_NEGOTIATE: 1288 case VMCI_TRANSPORT_PACKET_TYPE_NEGOTIATE2: 1289 if (pkt->u.size == 0 1290 || pkt->dg.src.context != vsk->remote_addr.svm_cid 1291 || pkt->src_port != vsk->remote_addr.svm_port 1292 || !vmci_handle_is_invalid(vmci_trans(vsk)->qp_handle) 1293 || vmci_trans(vsk)->qpair 1294 || vmci_trans(vsk)->produce_size != 0 1295 || vmci_trans(vsk)->consume_size != 0 1296 || vmci_trans(vsk)->detach_sub_id != VMCI_INVALID_ID) { 1297 skerr = EPROTO; 1298 err = -EINVAL; 1299 1300 goto destroy; 1301 } 1302 1303 err = vmci_transport_recv_connecting_client_negotiate(sk, pkt); 1304 if (err) { 1305 skerr = -err; 1306 goto destroy; 1307 } 1308 1309 break; 1310 case VMCI_TRANSPORT_PACKET_TYPE_INVALID: 1311 err = vmci_transport_recv_connecting_client_invalid(sk, pkt); 1312 if (err) { 1313 skerr = -err; 1314 goto destroy; 1315 } 1316 1317 break; 1318 case VMCI_TRANSPORT_PACKET_TYPE_RST: 1319 /* Older versions of the linux code (WS 6.5 / ESX 4.0) used to 1320 * continue processing here after they sent an INVALID packet. 1321 * This meant that we got a RST after the INVALID. We ignore a 1322 * RST after an INVALID. The common code doesn't send the RST 1323 * ... so we can hang if an old version of the common code 1324 * fails between getting a REQUEST and sending an OFFER back. 1325 * Not much we can do about it... except hope that it doesn't 1326 * happen. 1327 */ 1328 if (vsk->ignore_connecting_rst) { 1329 vsk->ignore_connecting_rst = false; 1330 } else { 1331 skerr = ECONNRESET; 1332 err = 0; 1333 goto destroy; 1334 } 1335 1336 break; 1337 default: 1338 /* Close and cleanup the connection. */ 1339 skerr = EPROTO; 1340 err = -EINVAL; 1341 goto destroy; 1342 } 1343 1344 return 0; 1345 1346 destroy: 1347 vmci_transport_send_reset(sk, pkt); 1348 1349 sk->sk_state = SS_UNCONNECTED; 1350 sk->sk_err = skerr; 1351 sk->sk_error_report(sk); 1352 return err; 1353 } 1354 1355 static int vmci_transport_recv_connecting_client_negotiate( 1356 struct sock *sk, 1357 struct vmci_transport_packet *pkt) 1358 { 1359 int err; 1360 struct vsock_sock *vsk; 1361 struct vmci_handle handle; 1362 struct vmci_qp *qpair; 1363 u32 detach_sub_id; 1364 bool is_local; 1365 u32 flags; 1366 bool old_proto = true; 1367 bool old_pkt_proto; 1368 u16 version; 1369 1370 vsk = vsock_sk(sk); 1371 handle = VMCI_INVALID_HANDLE; 1372 detach_sub_id = VMCI_INVALID_ID; 1373 1374 /* If we have gotten here then we should be past the point where old 1375 * linux vsock could have sent the bogus rst. 1376 */ 1377 vsk->sent_request = false; 1378 vsk->ignore_connecting_rst = false; 1379 1380 /* Verify that we're OK with the proposed queue pair size */ 1381 if (pkt->u.size < vmci_trans(vsk)->queue_pair_min_size || 1382 pkt->u.size > vmci_trans(vsk)->queue_pair_max_size) { 1383 err = -EINVAL; 1384 goto destroy; 1385 } 1386 1387 /* At this point we know the CID the peer is using to talk to us. */ 1388 1389 if (vsk->local_addr.svm_cid == VMADDR_CID_ANY) 1390 vsk->local_addr.svm_cid = pkt->dg.dst.context; 1391 1392 /* Setup the notify ops to be the highest supported version that both 1393 * the server and the client support. 1394 */ 1395 1396 if (vmci_transport_old_proto_override(&old_pkt_proto)) { 1397 old_proto = old_pkt_proto; 1398 } else { 1399 if (pkt->type == VMCI_TRANSPORT_PACKET_TYPE_NEGOTIATE) 1400 old_proto = true; 1401 else if (pkt->type == VMCI_TRANSPORT_PACKET_TYPE_NEGOTIATE2) 1402 old_proto = false; 1403 1404 } 1405 1406 if (old_proto) 1407 version = VSOCK_PROTO_INVALID; 1408 else 1409 version = pkt->proto; 1410 1411 if (!vmci_transport_proto_to_notify_struct(sk, &version, old_proto)) { 1412 err = -EINVAL; 1413 goto destroy; 1414 } 1415 1416 /* Subscribe to detach events first. 1417 * 1418 * XXX We attach once for each queue pair created for now so it is easy 1419 * to find the socket (it's provided), but later we should only 1420 * subscribe once and add a way to lookup sockets by queue pair handle. 1421 */ 1422 err = vmci_event_subscribe(VMCI_EVENT_QP_PEER_DETACH, 1423 vmci_transport_peer_detach_cb, 1424 vmci_trans(vsk), &detach_sub_id); 1425 if (err < VMCI_SUCCESS) { 1426 err = vmci_transport_error_to_vsock_error(err); 1427 goto destroy; 1428 } 1429 1430 /* Make VMCI select the handle for us. */ 1431 handle = VMCI_INVALID_HANDLE; 1432 is_local = vsk->remote_addr.svm_cid == vsk->local_addr.svm_cid; 1433 flags = is_local ? VMCI_QPFLAG_LOCAL : 0; 1434 1435 err = vmci_transport_queue_pair_alloc(&qpair, 1436 &handle, 1437 pkt->u.size, 1438 pkt->u.size, 1439 vsk->remote_addr.svm_cid, 1440 flags, 1441 vmci_transport_is_trusted( 1442 vsk, 1443 vsk-> 1444 remote_addr.svm_cid)); 1445 if (err < 0) 1446 goto destroy; 1447 1448 err = vmci_transport_send_qp_offer(sk, handle); 1449 if (err < 0) { 1450 err = vmci_transport_error_to_vsock_error(err); 1451 goto destroy; 1452 } 1453 1454 vmci_trans(vsk)->qp_handle = handle; 1455 vmci_trans(vsk)->qpair = qpair; 1456 1457 vmci_trans(vsk)->produce_size = vmci_trans(vsk)->consume_size = 1458 pkt->u.size; 1459 1460 vmci_trans(vsk)->detach_sub_id = detach_sub_id; 1461 1462 vmci_trans(vsk)->notify_ops->process_negotiate(sk); 1463 1464 return 0; 1465 1466 destroy: 1467 if (detach_sub_id != VMCI_INVALID_ID) 1468 vmci_event_unsubscribe(detach_sub_id); 1469 1470 if (!vmci_handle_is_invalid(handle)) 1471 vmci_qpair_detach(&qpair); 1472 1473 return err; 1474 } 1475 1476 static int 1477 vmci_transport_recv_connecting_client_invalid(struct sock *sk, 1478 struct vmci_transport_packet *pkt) 1479 { 1480 int err = 0; 1481 struct vsock_sock *vsk = vsock_sk(sk); 1482 1483 if (vsk->sent_request) { 1484 vsk->sent_request = false; 1485 vsk->ignore_connecting_rst = true; 1486 1487 err = vmci_transport_send_conn_request( 1488 sk, vmci_trans(vsk)->queue_pair_size); 1489 if (err < 0) 1490 err = vmci_transport_error_to_vsock_error(err); 1491 else 1492 err = 0; 1493 1494 } 1495 1496 return err; 1497 } 1498 1499 static int vmci_transport_recv_connected(struct sock *sk, 1500 struct vmci_transport_packet *pkt) 1501 { 1502 struct vsock_sock *vsk; 1503 bool pkt_processed = false; 1504 1505 /* In cases where we are closing the connection, it's sufficient to 1506 * mark the state change (and maybe error) and wake up any waiting 1507 * threads. Since this is a connected socket, it's owned by a user 1508 * process and will be cleaned up when the failure is passed back on 1509 * the current or next system call. Our system call implementations 1510 * must therefore check for error and state changes on entry and when 1511 * being awoken. 1512 */ 1513 switch (pkt->type) { 1514 case VMCI_TRANSPORT_PACKET_TYPE_SHUTDOWN: 1515 if (pkt->u.mode) { 1516 vsk = vsock_sk(sk); 1517 1518 vsk->peer_shutdown |= pkt->u.mode; 1519 sk->sk_state_change(sk); 1520 } 1521 break; 1522 1523 case VMCI_TRANSPORT_PACKET_TYPE_RST: 1524 vsk = vsock_sk(sk); 1525 /* It is possible that we sent our peer a message (e.g a 1526 * WAITING_READ) right before we got notified that the peer had 1527 * detached. If that happens then we can get a RST pkt back 1528 * from our peer even though there is data available for us to 1529 * read. In that case, don't shutdown the socket completely but 1530 * instead allow the local client to finish reading data off 1531 * the queuepair. Always treat a RST pkt in connected mode like 1532 * a clean shutdown. 1533 */ 1534 sock_set_flag(sk, SOCK_DONE); 1535 vsk->peer_shutdown = SHUTDOWN_MASK; 1536 if (vsock_stream_has_data(vsk) <= 0) 1537 sk->sk_state = SS_DISCONNECTING; 1538 1539 sk->sk_state_change(sk); 1540 break; 1541 1542 default: 1543 vsk = vsock_sk(sk); 1544 vmci_trans(vsk)->notify_ops->handle_notify_pkt( 1545 sk, pkt, false, NULL, NULL, 1546 &pkt_processed); 1547 if (!pkt_processed) 1548 return -EINVAL; 1549 1550 break; 1551 } 1552 1553 return 0; 1554 } 1555 1556 static int vmci_transport_socket_init(struct vsock_sock *vsk, 1557 struct vsock_sock *psk) 1558 { 1559 vsk->trans = kmalloc(sizeof(struct vmci_transport), GFP_KERNEL); 1560 if (!vsk->trans) 1561 return -ENOMEM; 1562 1563 vmci_trans(vsk)->dg_handle = VMCI_INVALID_HANDLE; 1564 vmci_trans(vsk)->qp_handle = VMCI_INVALID_HANDLE; 1565 vmci_trans(vsk)->qpair = NULL; 1566 vmci_trans(vsk)->produce_size = vmci_trans(vsk)->consume_size = 0; 1567 vmci_trans(vsk)->detach_sub_id = VMCI_INVALID_ID; 1568 vmci_trans(vsk)->notify_ops = NULL; 1569 INIT_LIST_HEAD(&vmci_trans(vsk)->elem); 1570 vmci_trans(vsk)->sk = &vsk->sk; 1571 spin_lock_init(&vmci_trans(vsk)->lock); 1572 if (psk) { 1573 vmci_trans(vsk)->queue_pair_size = 1574 vmci_trans(psk)->queue_pair_size; 1575 vmci_trans(vsk)->queue_pair_min_size = 1576 vmci_trans(psk)->queue_pair_min_size; 1577 vmci_trans(vsk)->queue_pair_max_size = 1578 vmci_trans(psk)->queue_pair_max_size; 1579 } else { 1580 vmci_trans(vsk)->queue_pair_size = 1581 VMCI_TRANSPORT_DEFAULT_QP_SIZE; 1582 vmci_trans(vsk)->queue_pair_min_size = 1583 VMCI_TRANSPORT_DEFAULT_QP_SIZE_MIN; 1584 vmci_trans(vsk)->queue_pair_max_size = 1585 VMCI_TRANSPORT_DEFAULT_QP_SIZE_MAX; 1586 } 1587 1588 return 0; 1589 } 1590 1591 static void vmci_transport_free_resources(struct list_head *transport_list) 1592 { 1593 while (!list_empty(transport_list)) { 1594 struct vmci_transport *transport = 1595 list_first_entry(transport_list, struct vmci_transport, 1596 elem); 1597 list_del(&transport->elem); 1598 1599 if (transport->detach_sub_id != VMCI_INVALID_ID) { 1600 vmci_event_unsubscribe(transport->detach_sub_id); 1601 transport->detach_sub_id = VMCI_INVALID_ID; 1602 } 1603 1604 if (!vmci_handle_is_invalid(transport->qp_handle)) { 1605 vmci_qpair_detach(&transport->qpair); 1606 transport->qp_handle = VMCI_INVALID_HANDLE; 1607 transport->produce_size = 0; 1608 transport->consume_size = 0; 1609 } 1610 1611 kfree(transport); 1612 } 1613 } 1614 1615 static void vmci_transport_cleanup(struct work_struct *work) 1616 { 1617 LIST_HEAD(pending); 1618 1619 spin_lock_bh(&vmci_transport_cleanup_lock); 1620 list_replace_init(&vmci_transport_cleanup_list, &pending); 1621 spin_unlock_bh(&vmci_transport_cleanup_lock); 1622 vmci_transport_free_resources(&pending); 1623 } 1624 1625 static void vmci_transport_destruct(struct vsock_sock *vsk) 1626 { 1627 /* Ensure that the detach callback doesn't use the sk/vsk 1628 * we are about to destruct. 1629 */ 1630 spin_lock_bh(&vmci_trans(vsk)->lock); 1631 vmci_trans(vsk)->sk = NULL; 1632 spin_unlock_bh(&vmci_trans(vsk)->lock); 1633 1634 if (vmci_trans(vsk)->notify_ops) 1635 vmci_trans(vsk)->notify_ops->socket_destruct(vsk); 1636 1637 spin_lock_bh(&vmci_transport_cleanup_lock); 1638 list_add(&vmci_trans(vsk)->elem, &vmci_transport_cleanup_list); 1639 spin_unlock_bh(&vmci_transport_cleanup_lock); 1640 schedule_work(&vmci_transport_cleanup_work); 1641 1642 vsk->trans = NULL; 1643 } 1644 1645 static void vmci_transport_release(struct vsock_sock *vsk) 1646 { 1647 if (!vmci_handle_is_invalid(vmci_trans(vsk)->dg_handle)) { 1648 vmci_datagram_destroy_handle(vmci_trans(vsk)->dg_handle); 1649 vmci_trans(vsk)->dg_handle = VMCI_INVALID_HANDLE; 1650 } 1651 } 1652 1653 static int vmci_transport_dgram_bind(struct vsock_sock *vsk, 1654 struct sockaddr_vm *addr) 1655 { 1656 u32 port; 1657 u32 flags; 1658 int err; 1659 1660 /* VMCI will select a resource ID for us if we provide 1661 * VMCI_INVALID_ID. 1662 */ 1663 port = addr->svm_port == VMADDR_PORT_ANY ? 1664 VMCI_INVALID_ID : addr->svm_port; 1665 1666 if (port <= LAST_RESERVED_PORT && !capable(CAP_NET_BIND_SERVICE)) 1667 return -EACCES; 1668 1669 flags = addr->svm_cid == VMADDR_CID_ANY ? 1670 VMCI_FLAG_ANYCID_DG_HND : 0; 1671 1672 err = vmci_transport_datagram_create_hnd(port, flags, 1673 vmci_transport_recv_dgram_cb, 1674 &vsk->sk, 1675 &vmci_trans(vsk)->dg_handle); 1676 if (err < VMCI_SUCCESS) 1677 return vmci_transport_error_to_vsock_error(err); 1678 vsock_addr_init(&vsk->local_addr, addr->svm_cid, 1679 vmci_trans(vsk)->dg_handle.resource); 1680 1681 return 0; 1682 } 1683 1684 static int vmci_transport_dgram_enqueue( 1685 struct vsock_sock *vsk, 1686 struct sockaddr_vm *remote_addr, 1687 struct msghdr *msg, 1688 size_t len) 1689 { 1690 int err; 1691 struct vmci_datagram *dg; 1692 1693 if (len > VMCI_MAX_DG_PAYLOAD_SIZE) 1694 return -EMSGSIZE; 1695 1696 if (!vmci_transport_allow_dgram(vsk, remote_addr->svm_cid)) 1697 return -EPERM; 1698 1699 /* Allocate a buffer for the user's message and our packet header. */ 1700 dg = kmalloc(len + sizeof(*dg), GFP_KERNEL); 1701 if (!dg) 1702 return -ENOMEM; 1703 1704 memcpy_from_msg(VMCI_DG_PAYLOAD(dg), msg, len); 1705 1706 dg->dst = vmci_make_handle(remote_addr->svm_cid, 1707 remote_addr->svm_port); 1708 dg->src = vmci_make_handle(vsk->local_addr.svm_cid, 1709 vsk->local_addr.svm_port); 1710 dg->payload_size = len; 1711 1712 err = vmci_datagram_send(dg); 1713 kfree(dg); 1714 if (err < 0) 1715 return vmci_transport_error_to_vsock_error(err); 1716 1717 return err - sizeof(*dg); 1718 } 1719 1720 static int vmci_transport_dgram_dequeue(struct vsock_sock *vsk, 1721 struct msghdr *msg, size_t len, 1722 int flags) 1723 { 1724 int err; 1725 int noblock; 1726 struct vmci_datagram *dg; 1727 size_t payload_len; 1728 struct sk_buff *skb; 1729 1730 noblock = flags & MSG_DONTWAIT; 1731 1732 if (flags & MSG_OOB || flags & MSG_ERRQUEUE) 1733 return -EOPNOTSUPP; 1734 1735 /* Retrieve the head sk_buff from the socket's receive queue. */ 1736 err = 0; 1737 skb = skb_recv_datagram(&vsk->sk, flags, noblock, &err); 1738 if (err) 1739 return err; 1740 1741 if (!skb) 1742 return -EAGAIN; 1743 1744 dg = (struct vmci_datagram *)skb->data; 1745 if (!dg) 1746 /* err is 0, meaning we read zero bytes. */ 1747 goto out; 1748 1749 payload_len = dg->payload_size; 1750 /* Ensure the sk_buff matches the payload size claimed in the packet. */ 1751 if (payload_len != skb->len - sizeof(*dg)) { 1752 err = -EINVAL; 1753 goto out; 1754 } 1755 1756 if (payload_len > len) { 1757 payload_len = len; 1758 msg->msg_flags |= MSG_TRUNC; 1759 } 1760 1761 /* Place the datagram payload in the user's iovec. */ 1762 err = skb_copy_datagram_msg(skb, sizeof(*dg), msg, payload_len); 1763 if (err) 1764 goto out; 1765 1766 if (msg->msg_name) { 1767 /* Provide the address of the sender. */ 1768 DECLARE_SOCKADDR(struct sockaddr_vm *, vm_addr, msg->msg_name); 1769 vsock_addr_init(vm_addr, dg->src.context, dg->src.resource); 1770 msg->msg_namelen = sizeof(*vm_addr); 1771 } 1772 err = payload_len; 1773 1774 out: 1775 skb_free_datagram(&vsk->sk, skb); 1776 return err; 1777 } 1778 1779 static bool vmci_transport_dgram_allow(u32 cid, u32 port) 1780 { 1781 if (cid == VMADDR_CID_HYPERVISOR) { 1782 /* Registrations of PBRPC Servers do not modify VMX/Hypervisor 1783 * state and are allowed. 1784 */ 1785 return port == VMCI_UNITY_PBRPC_REGISTER; 1786 } 1787 1788 return true; 1789 } 1790 1791 static int vmci_transport_connect(struct vsock_sock *vsk) 1792 { 1793 int err; 1794 bool old_pkt_proto = false; 1795 struct sock *sk = &vsk->sk; 1796 1797 if (vmci_transport_old_proto_override(&old_pkt_proto) && 1798 old_pkt_proto) { 1799 err = vmci_transport_send_conn_request( 1800 sk, vmci_trans(vsk)->queue_pair_size); 1801 if (err < 0) { 1802 sk->sk_state = SS_UNCONNECTED; 1803 return err; 1804 } 1805 } else { 1806 int supported_proto_versions = 1807 vmci_transport_new_proto_supported_versions(); 1808 err = vmci_transport_send_conn_request2( 1809 sk, vmci_trans(vsk)->queue_pair_size, 1810 supported_proto_versions); 1811 if (err < 0) { 1812 sk->sk_state = SS_UNCONNECTED; 1813 return err; 1814 } 1815 1816 vsk->sent_request = true; 1817 } 1818 1819 return err; 1820 } 1821 1822 static ssize_t vmci_transport_stream_dequeue( 1823 struct vsock_sock *vsk, 1824 struct msghdr *msg, 1825 size_t len, 1826 int flags) 1827 { 1828 if (flags & MSG_PEEK) 1829 return vmci_qpair_peekv(vmci_trans(vsk)->qpair, msg, len, 0); 1830 else 1831 return vmci_qpair_dequev(vmci_trans(vsk)->qpair, msg, len, 0); 1832 } 1833 1834 static ssize_t vmci_transport_stream_enqueue( 1835 struct vsock_sock *vsk, 1836 struct msghdr *msg, 1837 size_t len) 1838 { 1839 return vmci_qpair_enquev(vmci_trans(vsk)->qpair, msg, len, 0); 1840 } 1841 1842 static s64 vmci_transport_stream_has_data(struct vsock_sock *vsk) 1843 { 1844 return vmci_qpair_consume_buf_ready(vmci_trans(vsk)->qpair); 1845 } 1846 1847 static s64 vmci_transport_stream_has_space(struct vsock_sock *vsk) 1848 { 1849 return vmci_qpair_produce_free_space(vmci_trans(vsk)->qpair); 1850 } 1851 1852 static u64 vmci_transport_stream_rcvhiwat(struct vsock_sock *vsk) 1853 { 1854 return vmci_trans(vsk)->consume_size; 1855 } 1856 1857 static bool vmci_transport_stream_is_active(struct vsock_sock *vsk) 1858 { 1859 return !vmci_handle_is_invalid(vmci_trans(vsk)->qp_handle); 1860 } 1861 1862 static u64 vmci_transport_get_buffer_size(struct vsock_sock *vsk) 1863 { 1864 return vmci_trans(vsk)->queue_pair_size; 1865 } 1866 1867 static u64 vmci_transport_get_min_buffer_size(struct vsock_sock *vsk) 1868 { 1869 return vmci_trans(vsk)->queue_pair_min_size; 1870 } 1871 1872 static u64 vmci_transport_get_max_buffer_size(struct vsock_sock *vsk) 1873 { 1874 return vmci_trans(vsk)->queue_pair_max_size; 1875 } 1876 1877 static void vmci_transport_set_buffer_size(struct vsock_sock *vsk, u64 val) 1878 { 1879 if (val < vmci_trans(vsk)->queue_pair_min_size) 1880 vmci_trans(vsk)->queue_pair_min_size = val; 1881 if (val > vmci_trans(vsk)->queue_pair_max_size) 1882 vmci_trans(vsk)->queue_pair_max_size = val; 1883 vmci_trans(vsk)->queue_pair_size = val; 1884 } 1885 1886 static void vmci_transport_set_min_buffer_size(struct vsock_sock *vsk, 1887 u64 val) 1888 { 1889 if (val > vmci_trans(vsk)->queue_pair_size) 1890 vmci_trans(vsk)->queue_pair_size = val; 1891 vmci_trans(vsk)->queue_pair_min_size = val; 1892 } 1893 1894 static void vmci_transport_set_max_buffer_size(struct vsock_sock *vsk, 1895 u64 val) 1896 { 1897 if (val < vmci_trans(vsk)->queue_pair_size) 1898 vmci_trans(vsk)->queue_pair_size = val; 1899 vmci_trans(vsk)->queue_pair_max_size = val; 1900 } 1901 1902 static int vmci_transport_notify_poll_in( 1903 struct vsock_sock *vsk, 1904 size_t target, 1905 bool *data_ready_now) 1906 { 1907 return vmci_trans(vsk)->notify_ops->poll_in( 1908 &vsk->sk, target, data_ready_now); 1909 } 1910 1911 static int vmci_transport_notify_poll_out( 1912 struct vsock_sock *vsk, 1913 size_t target, 1914 bool *space_available_now) 1915 { 1916 return vmci_trans(vsk)->notify_ops->poll_out( 1917 &vsk->sk, target, space_available_now); 1918 } 1919 1920 static int vmci_transport_notify_recv_init( 1921 struct vsock_sock *vsk, 1922 size_t target, 1923 struct vsock_transport_recv_notify_data *data) 1924 { 1925 return vmci_trans(vsk)->notify_ops->recv_init( 1926 &vsk->sk, target, 1927 (struct vmci_transport_recv_notify_data *)data); 1928 } 1929 1930 static int vmci_transport_notify_recv_pre_block( 1931 struct vsock_sock *vsk, 1932 size_t target, 1933 struct vsock_transport_recv_notify_data *data) 1934 { 1935 return vmci_trans(vsk)->notify_ops->recv_pre_block( 1936 &vsk->sk, target, 1937 (struct vmci_transport_recv_notify_data *)data); 1938 } 1939 1940 static int vmci_transport_notify_recv_pre_dequeue( 1941 struct vsock_sock *vsk, 1942 size_t target, 1943 struct vsock_transport_recv_notify_data *data) 1944 { 1945 return vmci_trans(vsk)->notify_ops->recv_pre_dequeue( 1946 &vsk->sk, target, 1947 (struct vmci_transport_recv_notify_data *)data); 1948 } 1949 1950 static int vmci_transport_notify_recv_post_dequeue( 1951 struct vsock_sock *vsk, 1952 size_t target, 1953 ssize_t copied, 1954 bool data_read, 1955 struct vsock_transport_recv_notify_data *data) 1956 { 1957 return vmci_trans(vsk)->notify_ops->recv_post_dequeue( 1958 &vsk->sk, target, copied, data_read, 1959 (struct vmci_transport_recv_notify_data *)data); 1960 } 1961 1962 static int vmci_transport_notify_send_init( 1963 struct vsock_sock *vsk, 1964 struct vsock_transport_send_notify_data *data) 1965 { 1966 return vmci_trans(vsk)->notify_ops->send_init( 1967 &vsk->sk, 1968 (struct vmci_transport_send_notify_data *)data); 1969 } 1970 1971 static int vmci_transport_notify_send_pre_block( 1972 struct vsock_sock *vsk, 1973 struct vsock_transport_send_notify_data *data) 1974 { 1975 return vmci_trans(vsk)->notify_ops->send_pre_block( 1976 &vsk->sk, 1977 (struct vmci_transport_send_notify_data *)data); 1978 } 1979 1980 static int vmci_transport_notify_send_pre_enqueue( 1981 struct vsock_sock *vsk, 1982 struct vsock_transport_send_notify_data *data) 1983 { 1984 return vmci_trans(vsk)->notify_ops->send_pre_enqueue( 1985 &vsk->sk, 1986 (struct vmci_transport_send_notify_data *)data); 1987 } 1988 1989 static int vmci_transport_notify_send_post_enqueue( 1990 struct vsock_sock *vsk, 1991 ssize_t written, 1992 struct vsock_transport_send_notify_data *data) 1993 { 1994 return vmci_trans(vsk)->notify_ops->send_post_enqueue( 1995 &vsk->sk, written, 1996 (struct vmci_transport_send_notify_data *)data); 1997 } 1998 1999 static bool vmci_transport_old_proto_override(bool *old_pkt_proto) 2000 { 2001 if (PROTOCOL_OVERRIDE != -1) { 2002 if (PROTOCOL_OVERRIDE == 0) 2003 *old_pkt_proto = true; 2004 else 2005 *old_pkt_proto = false; 2006 2007 pr_info("Proto override in use\n"); 2008 return true; 2009 } 2010 2011 return false; 2012 } 2013 2014 static bool vmci_transport_proto_to_notify_struct(struct sock *sk, 2015 u16 *proto, 2016 bool old_pkt_proto) 2017 { 2018 struct vsock_sock *vsk = vsock_sk(sk); 2019 2020 if (old_pkt_proto) { 2021 if (*proto != VSOCK_PROTO_INVALID) { 2022 pr_err("Can't set both an old and new protocol\n"); 2023 return false; 2024 } 2025 vmci_trans(vsk)->notify_ops = &vmci_transport_notify_pkt_ops; 2026 goto exit; 2027 } 2028 2029 switch (*proto) { 2030 case VSOCK_PROTO_PKT_ON_NOTIFY: 2031 vmci_trans(vsk)->notify_ops = 2032 &vmci_transport_notify_pkt_q_state_ops; 2033 break; 2034 default: 2035 pr_err("Unknown notify protocol version\n"); 2036 return false; 2037 } 2038 2039 exit: 2040 vmci_trans(vsk)->notify_ops->socket_init(sk); 2041 return true; 2042 } 2043 2044 static u16 vmci_transport_new_proto_supported_versions(void) 2045 { 2046 if (PROTOCOL_OVERRIDE != -1) 2047 return PROTOCOL_OVERRIDE; 2048 2049 return VSOCK_PROTO_ALL_SUPPORTED; 2050 } 2051 2052 static u32 vmci_transport_get_local_cid(void) 2053 { 2054 return vmci_get_context_id(); 2055 } 2056 2057 static struct vsock_transport vmci_transport = { 2058 .init = vmci_transport_socket_init, 2059 .destruct = vmci_transport_destruct, 2060 .release = vmci_transport_release, 2061 .connect = vmci_transport_connect, 2062 .dgram_bind = vmci_transport_dgram_bind, 2063 .dgram_dequeue = vmci_transport_dgram_dequeue, 2064 .dgram_enqueue = vmci_transport_dgram_enqueue, 2065 .dgram_allow = vmci_transport_dgram_allow, 2066 .stream_dequeue = vmci_transport_stream_dequeue, 2067 .stream_enqueue = vmci_transport_stream_enqueue, 2068 .stream_has_data = vmci_transport_stream_has_data, 2069 .stream_has_space = vmci_transport_stream_has_space, 2070 .stream_rcvhiwat = vmci_transport_stream_rcvhiwat, 2071 .stream_is_active = vmci_transport_stream_is_active, 2072 .stream_allow = vmci_transport_stream_allow, 2073 .notify_poll_in = vmci_transport_notify_poll_in, 2074 .notify_poll_out = vmci_transport_notify_poll_out, 2075 .notify_recv_init = vmci_transport_notify_recv_init, 2076 .notify_recv_pre_block = vmci_transport_notify_recv_pre_block, 2077 .notify_recv_pre_dequeue = vmci_transport_notify_recv_pre_dequeue, 2078 .notify_recv_post_dequeue = vmci_transport_notify_recv_post_dequeue, 2079 .notify_send_init = vmci_transport_notify_send_init, 2080 .notify_send_pre_block = vmci_transport_notify_send_pre_block, 2081 .notify_send_pre_enqueue = vmci_transport_notify_send_pre_enqueue, 2082 .notify_send_post_enqueue = vmci_transport_notify_send_post_enqueue, 2083 .shutdown = vmci_transport_shutdown, 2084 .set_buffer_size = vmci_transport_set_buffer_size, 2085 .set_min_buffer_size = vmci_transport_set_min_buffer_size, 2086 .set_max_buffer_size = vmci_transport_set_max_buffer_size, 2087 .get_buffer_size = vmci_transport_get_buffer_size, 2088 .get_min_buffer_size = vmci_transport_get_min_buffer_size, 2089 .get_max_buffer_size = vmci_transport_get_max_buffer_size, 2090 .get_local_cid = vmci_transport_get_local_cid, 2091 }; 2092 2093 static int __init vmci_transport_init(void) 2094 { 2095 int err; 2096 2097 /* Create the datagram handle that we will use to send and receive all 2098 * VSocket control messages for this context. 2099 */ 2100 err = vmci_transport_datagram_create_hnd(VMCI_TRANSPORT_PACKET_RID, 2101 VMCI_FLAG_ANYCID_DG_HND, 2102 vmci_transport_recv_stream_cb, 2103 NULL, 2104 &vmci_transport_stream_handle); 2105 if (err < VMCI_SUCCESS) { 2106 pr_err("Unable to create datagram handle. (%d)\n", err); 2107 return vmci_transport_error_to_vsock_error(err); 2108 } 2109 2110 err = vmci_event_subscribe(VMCI_EVENT_QP_RESUMED, 2111 vmci_transport_qp_resumed_cb, 2112 NULL, &vmci_transport_qp_resumed_sub_id); 2113 if (err < VMCI_SUCCESS) { 2114 pr_err("Unable to subscribe to resumed event. (%d)\n", err); 2115 err = vmci_transport_error_to_vsock_error(err); 2116 vmci_transport_qp_resumed_sub_id = VMCI_INVALID_ID; 2117 goto err_destroy_stream_handle; 2118 } 2119 2120 err = vsock_core_init(&vmci_transport); 2121 if (err < 0) 2122 goto err_unsubscribe; 2123 2124 return 0; 2125 2126 err_unsubscribe: 2127 vmci_event_unsubscribe(vmci_transport_qp_resumed_sub_id); 2128 err_destroy_stream_handle: 2129 vmci_datagram_destroy_handle(vmci_transport_stream_handle); 2130 return err; 2131 } 2132 module_init(vmci_transport_init); 2133 2134 static void __exit vmci_transport_exit(void) 2135 { 2136 cancel_work_sync(&vmci_transport_cleanup_work); 2137 vmci_transport_free_resources(&vmci_transport_cleanup_list); 2138 2139 if (!vmci_handle_is_invalid(vmci_transport_stream_handle)) { 2140 if (vmci_datagram_destroy_handle( 2141 vmci_transport_stream_handle) != VMCI_SUCCESS) 2142 pr_err("Couldn't destroy datagram handle\n"); 2143 vmci_transport_stream_handle = VMCI_INVALID_HANDLE; 2144 } 2145 2146 if (vmci_transport_qp_resumed_sub_id != VMCI_INVALID_ID) { 2147 vmci_event_unsubscribe(vmci_transport_qp_resumed_sub_id); 2148 vmci_transport_qp_resumed_sub_id = VMCI_INVALID_ID; 2149 } 2150 2151 vsock_core_exit(); 2152 } 2153 module_exit(vmci_transport_exit); 2154 2155 MODULE_AUTHOR("VMware, Inc."); 2156 MODULE_DESCRIPTION("VMCI transport for Virtual Sockets"); 2157 MODULE_VERSION("1.0.2.0-k"); 2158 MODULE_LICENSE("GPL v2"); 2159 MODULE_ALIAS("vmware_vsock"); 2160 MODULE_ALIAS_NETPROTO(PF_VSOCK); 2161