1 /* $NetBSD: uipc_socket.c,v 1.209 2012/02/01 02:27:23 matt Exp $ */ 2 3 /*- 4 * Copyright (c) 2002, 2007, 2008, 2009 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Jason R. Thorpe of Wasabi Systems, Inc, and by Andrew Doran. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 20 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 21 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 22 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 23 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 24 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 25 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 26 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 27 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 28 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 29 * POSSIBILITY OF SUCH DAMAGE. 30 */ 31 32 /* 33 * Copyright (c) 2004 The FreeBSD Foundation 34 * Copyright (c) 2004 Robert Watson 35 * Copyright (c) 1982, 1986, 1988, 1990, 1993 36 * The Regents of the University of California. All rights reserved. 37 * 38 * Redistribution and use in source and binary forms, with or without 39 * modification, are permitted provided that the following conditions 40 * are met: 41 * 1. Redistributions of source code must retain the above copyright 42 * notice, this list of conditions and the following disclaimer. 43 * 2. Redistributions in binary form must reproduce the above copyright 44 * notice, this list of conditions and the following disclaimer in the 45 * documentation and/or other materials provided with the distribution. 46 * 3. Neither the name of the University nor the names of its contributors 47 * may be used to endorse or promote products derived from this software 48 * without specific prior written permission. 49 * 50 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 51 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 52 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 53 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 54 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 55 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 56 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 57 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 58 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 59 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 60 * SUCH DAMAGE. 61 * 62 * @(#)uipc_socket.c 8.6 (Berkeley) 5/2/95 63 */ 64 65 #include <sys/cdefs.h> 66 __KERNEL_RCSID(0, "$NetBSD: uipc_socket.c,v 1.209 2012/02/01 02:27:23 matt Exp $"); 67 68 #include "opt_compat_netbsd.h" 69 #include "opt_sock_counters.h" 70 #include "opt_sosend_loan.h" 71 #include "opt_mbuftrace.h" 72 #include "opt_somaxkva.h" 73 #include "opt_multiprocessor.h" /* XXX */ 74 75 #include <sys/param.h> 76 #include <sys/systm.h> 77 #include <sys/proc.h> 78 #include <sys/file.h> 79 #include <sys/filedesc.h> 80 #include <sys/kmem.h> 81 #include <sys/mbuf.h> 82 #include <sys/domain.h> 83 #include <sys/kernel.h> 84 #include <sys/protosw.h> 85 #include <sys/socket.h> 86 #include <sys/socketvar.h> 87 #include <sys/signalvar.h> 88 #include <sys/resourcevar.h> 89 #include <sys/uidinfo.h> 90 #include <sys/event.h> 91 #include <sys/poll.h> 92 #include <sys/kauth.h> 93 #include <sys/mutex.h> 94 #include <sys/condvar.h> 95 #include <sys/kthread.h> 96 97 #ifdef COMPAT_50 98 #include <compat/sys/time.h> 99 #include <compat/sys/socket.h> 100 #endif 101 102 #include <uvm/uvm_extern.h> 103 #include <uvm/uvm_loan.h> 104 #include <uvm/uvm_page.h> 105 106 MALLOC_DEFINE(M_SOOPTS, "soopts", "socket options"); 107 MALLOC_DEFINE(M_SONAME, "soname", "socket name"); 108 109 extern const struct fileops socketops; 110 111 extern int somaxconn; /* patchable (XXX sysctl) */ 112 int somaxconn = SOMAXCONN; 113 kmutex_t *softnet_lock; 114 115 #ifdef SOSEND_COUNTERS 116 #include <sys/device.h> 117 118 static struct evcnt sosend_loan_big = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 119 NULL, "sosend", "loan big"); 120 static struct evcnt sosend_copy_big = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 121 NULL, "sosend", "copy big"); 122 static struct evcnt sosend_copy_small = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 123 NULL, "sosend", "copy small"); 124 static struct evcnt sosend_kvalimit = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, 125 NULL, "sosend", "kva limit"); 126 127 #define SOSEND_COUNTER_INCR(ev) (ev)->ev_count++ 128 129 EVCNT_ATTACH_STATIC(sosend_loan_big); 130 EVCNT_ATTACH_STATIC(sosend_copy_big); 131 EVCNT_ATTACH_STATIC(sosend_copy_small); 132 EVCNT_ATTACH_STATIC(sosend_kvalimit); 133 #else 134 135 #define SOSEND_COUNTER_INCR(ev) /* nothing */ 136 137 #endif /* SOSEND_COUNTERS */ 138 139 #if defined(SOSEND_NO_LOAN) || defined(MULTIPROCESSOR) 140 int sock_loan_thresh = -1; 141 #else 142 int sock_loan_thresh = 4096; 143 #endif 144 145 static kmutex_t so_pendfree_lock; 146 static struct mbuf *so_pendfree = NULL; 147 148 #ifndef SOMAXKVA 149 #define SOMAXKVA (16 * 1024 * 1024) 150 #endif 151 int somaxkva = SOMAXKVA; 152 static int socurkva; 153 static kcondvar_t socurkva_cv; 154 155 static kauth_listener_t socket_listener; 156 157 #define SOCK_LOAN_CHUNK 65536 158 159 static void sopendfree_thread(void *); 160 static kcondvar_t pendfree_thread_cv; 161 static lwp_t *sopendfree_lwp; 162 163 static void sysctl_kern_somaxkva_setup(void); 164 static struct sysctllog *socket_sysctllog; 165 166 static vsize_t 167 sokvareserve(struct socket *so, vsize_t len) 168 { 169 int error; 170 171 mutex_enter(&so_pendfree_lock); 172 while (socurkva + len > somaxkva) { 173 SOSEND_COUNTER_INCR(&sosend_kvalimit); 174 error = cv_wait_sig(&socurkva_cv, &so_pendfree_lock); 175 if (error) { 176 len = 0; 177 break; 178 } 179 } 180 socurkva += len; 181 mutex_exit(&so_pendfree_lock); 182 return len; 183 } 184 185 static void 186 sokvaunreserve(vsize_t len) 187 { 188 189 mutex_enter(&so_pendfree_lock); 190 socurkva -= len; 191 cv_broadcast(&socurkva_cv); 192 mutex_exit(&so_pendfree_lock); 193 } 194 195 /* 196 * sokvaalloc: allocate kva for loan. 197 */ 198 199 vaddr_t 200 sokvaalloc(vaddr_t sva, vsize_t len, struct socket *so) 201 { 202 vaddr_t lva; 203 204 /* 205 * reserve kva. 206 */ 207 208 if (sokvareserve(so, len) == 0) 209 return 0; 210 211 /* 212 * allocate kva. 213 */ 214 215 lva = uvm_km_alloc(kernel_map, len, atop(sva) & uvmexp.colormask, 216 UVM_KMF_COLORMATCH | UVM_KMF_VAONLY | UVM_KMF_WAITVA); 217 if (lva == 0) { 218 sokvaunreserve(len); 219 return (0); 220 } 221 222 return lva; 223 } 224 225 /* 226 * sokvafree: free kva for loan. 227 */ 228 229 void 230 sokvafree(vaddr_t sva, vsize_t len) 231 { 232 233 /* 234 * free kva. 235 */ 236 237 uvm_km_free(kernel_map, sva, len, UVM_KMF_VAONLY); 238 239 /* 240 * unreserve kva. 241 */ 242 243 sokvaunreserve(len); 244 } 245 246 static void 247 sodoloanfree(struct vm_page **pgs, void *buf, size_t size) 248 { 249 vaddr_t sva, eva; 250 vsize_t len; 251 int npgs; 252 253 KASSERT(pgs != NULL); 254 255 eva = round_page((vaddr_t) buf + size); 256 sva = trunc_page((vaddr_t) buf); 257 len = eva - sva; 258 npgs = len >> PAGE_SHIFT; 259 260 pmap_kremove(sva, len); 261 pmap_update(pmap_kernel()); 262 uvm_unloan(pgs, npgs, UVM_LOAN_TOPAGE); 263 sokvafree(sva, len); 264 } 265 266 /* 267 * sopendfree_thread: free mbufs on "pendfree" list. 268 * unlock and relock so_pendfree_lock when freeing mbufs. 269 */ 270 271 static void 272 sopendfree_thread(void *v) 273 { 274 struct mbuf *m, *next; 275 size_t rv; 276 277 mutex_enter(&so_pendfree_lock); 278 279 for (;;) { 280 rv = 0; 281 while (so_pendfree != NULL) { 282 m = so_pendfree; 283 so_pendfree = NULL; 284 mutex_exit(&so_pendfree_lock); 285 286 for (; m != NULL; m = next) { 287 next = m->m_next; 288 KASSERT((~m->m_flags & (M_EXT|M_EXT_PAGES)) == 0); 289 KASSERT(m->m_ext.ext_refcnt == 0); 290 291 rv += m->m_ext.ext_size; 292 sodoloanfree(m->m_ext.ext_pgs, m->m_ext.ext_buf, 293 m->m_ext.ext_size); 294 pool_cache_put(mb_cache, m); 295 } 296 297 mutex_enter(&so_pendfree_lock); 298 } 299 if (rv) 300 cv_broadcast(&socurkva_cv); 301 cv_wait(&pendfree_thread_cv, &so_pendfree_lock); 302 } 303 panic("sopendfree_thread"); 304 /* NOTREACHED */ 305 } 306 307 void 308 soloanfree(struct mbuf *m, void *buf, size_t size, void *arg) 309 { 310 311 KASSERT(m != NULL); 312 313 /* 314 * postpone freeing mbuf. 315 * 316 * we can't do it in interrupt context 317 * because we need to put kva back to kernel_map. 318 */ 319 320 mutex_enter(&so_pendfree_lock); 321 m->m_next = so_pendfree; 322 so_pendfree = m; 323 cv_signal(&pendfree_thread_cv); 324 mutex_exit(&so_pendfree_lock); 325 } 326 327 static long 328 sosend_loan(struct socket *so, struct uio *uio, struct mbuf *m, long space) 329 { 330 struct iovec *iov = uio->uio_iov; 331 vaddr_t sva, eva; 332 vsize_t len; 333 vaddr_t lva; 334 int npgs, error; 335 vaddr_t va; 336 int i; 337 338 if (VMSPACE_IS_KERNEL_P(uio->uio_vmspace)) 339 return (0); 340 341 if (iov->iov_len < (size_t) space) 342 space = iov->iov_len; 343 if (space > SOCK_LOAN_CHUNK) 344 space = SOCK_LOAN_CHUNK; 345 346 eva = round_page((vaddr_t) iov->iov_base + space); 347 sva = trunc_page((vaddr_t) iov->iov_base); 348 len = eva - sva; 349 npgs = len >> PAGE_SHIFT; 350 351 KASSERT(npgs <= M_EXT_MAXPAGES); 352 353 lva = sokvaalloc(sva, len, so); 354 if (lva == 0) 355 return 0; 356 357 error = uvm_loan(&uio->uio_vmspace->vm_map, sva, len, 358 m->m_ext.ext_pgs, UVM_LOAN_TOPAGE); 359 if (error) { 360 sokvafree(lva, len); 361 return (0); 362 } 363 364 for (i = 0, va = lva; i < npgs; i++, va += PAGE_SIZE) 365 pmap_kenter_pa(va, VM_PAGE_TO_PHYS(m->m_ext.ext_pgs[i]), 366 VM_PROT_READ, 0); 367 pmap_update(pmap_kernel()); 368 369 lva += (vaddr_t) iov->iov_base & PAGE_MASK; 370 371 MEXTADD(m, (void *) lva, space, M_MBUF, soloanfree, so); 372 m->m_flags |= M_EXT_PAGES | M_EXT_ROMAP; 373 374 uio->uio_resid -= space; 375 /* uio_offset not updated, not set/used for write(2) */ 376 uio->uio_iov->iov_base = (char *)uio->uio_iov->iov_base + space; 377 uio->uio_iov->iov_len -= space; 378 if (uio->uio_iov->iov_len == 0) { 379 uio->uio_iov++; 380 uio->uio_iovcnt--; 381 } 382 383 return (space); 384 } 385 386 struct mbuf * 387 getsombuf(struct socket *so, int type) 388 { 389 struct mbuf *m; 390 391 m = m_get(M_WAIT, type); 392 MCLAIM(m, so->so_mowner); 393 return m; 394 } 395 396 static int 397 socket_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie, 398 void *arg0, void *arg1, void *arg2, void *arg3) 399 { 400 int result; 401 enum kauth_network_req req; 402 403 result = KAUTH_RESULT_DEFER; 404 req = (enum kauth_network_req)arg0; 405 406 if ((action != KAUTH_NETWORK_SOCKET) && 407 (action != KAUTH_NETWORK_BIND)) 408 return result; 409 410 switch (req) { 411 case KAUTH_REQ_NETWORK_BIND_PORT: 412 result = KAUTH_RESULT_ALLOW; 413 break; 414 415 case KAUTH_REQ_NETWORK_SOCKET_DROP: { 416 /* Normal users can only drop their own connections. */ 417 struct socket *so = (struct socket *)arg1; 418 419 if (proc_uidmatch(cred, so->so_cred)) 420 result = KAUTH_RESULT_ALLOW; 421 422 break; 423 } 424 425 case KAUTH_REQ_NETWORK_SOCKET_OPEN: 426 /* We allow "raw" routing/bluetooth sockets to anyone. */ 427 if ((u_long)arg1 == PF_ROUTE || (u_long)arg1 == PF_OROUTE 428 || (u_long)arg1 == PF_BLUETOOTH) { 429 result = KAUTH_RESULT_ALLOW; 430 } else { 431 /* Privileged, let secmodel handle this. */ 432 if ((u_long)arg2 == SOCK_RAW) 433 break; 434 } 435 436 result = KAUTH_RESULT_ALLOW; 437 438 break; 439 440 case KAUTH_REQ_NETWORK_SOCKET_CANSEE: 441 result = KAUTH_RESULT_ALLOW; 442 443 break; 444 445 default: 446 break; 447 } 448 449 return result; 450 } 451 452 void 453 soinit(void) 454 { 455 456 sysctl_kern_somaxkva_setup(); 457 458 mutex_init(&so_pendfree_lock, MUTEX_DEFAULT, IPL_VM); 459 softnet_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 460 cv_init(&socurkva_cv, "sokva"); 461 cv_init(&pendfree_thread_cv, "sopendfr"); 462 soinit2(); 463 464 /* Set the initial adjusted socket buffer size. */ 465 if (sb_max_set(sb_max)) 466 panic("bad initial sb_max value: %lu", sb_max); 467 468 socket_listener = kauth_listen_scope(KAUTH_SCOPE_NETWORK, 469 socket_listener_cb, NULL); 470 } 471 472 void 473 soinit1(void) 474 { 475 int error = kthread_create(PRI_NONE, KTHREAD_MPSAFE, NULL, 476 sopendfree_thread, NULL, &sopendfree_lwp, "sopendfree"); 477 if (error) 478 panic("soinit1 %d", error); 479 } 480 481 /* 482 * Socket operation routines. 483 * These routines are called by the routines in 484 * sys_socket.c or from a system process, and 485 * implement the semantics of socket operations by 486 * switching out to the protocol specific routines. 487 */ 488 /*ARGSUSED*/ 489 int 490 socreate(int dom, struct socket **aso, int type, int proto, struct lwp *l, 491 struct socket *lockso) 492 { 493 const struct protosw *prp; 494 struct socket *so; 495 uid_t uid; 496 int error; 497 kmutex_t *lock; 498 499 error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_SOCKET, 500 KAUTH_REQ_NETWORK_SOCKET_OPEN, KAUTH_ARG(dom), KAUTH_ARG(type), 501 KAUTH_ARG(proto)); 502 if (error != 0) 503 return error; 504 505 if (proto) 506 prp = pffindproto(dom, proto, type); 507 else 508 prp = pffindtype(dom, type); 509 if (prp == NULL) { 510 /* no support for domain */ 511 if (pffinddomain(dom) == 0) 512 return EAFNOSUPPORT; 513 /* no support for socket type */ 514 if (proto == 0 && type != 0) 515 return EPROTOTYPE; 516 return EPROTONOSUPPORT; 517 } 518 if (prp->pr_usrreq == NULL) 519 return EPROTONOSUPPORT; 520 if (prp->pr_type != type) 521 return EPROTOTYPE; 522 523 so = soget(true); 524 so->so_type = type; 525 so->so_proto = prp; 526 so->so_send = sosend; 527 so->so_receive = soreceive; 528 #ifdef MBUFTRACE 529 so->so_rcv.sb_mowner = &prp->pr_domain->dom_mowner; 530 so->so_snd.sb_mowner = &prp->pr_domain->dom_mowner; 531 so->so_mowner = &prp->pr_domain->dom_mowner; 532 #endif 533 uid = kauth_cred_geteuid(l->l_cred); 534 so->so_uidinfo = uid_find(uid); 535 so->so_cpid = l->l_proc->p_pid; 536 if (lockso != NULL) { 537 /* Caller wants us to share a lock. */ 538 lock = lockso->so_lock; 539 so->so_lock = lock; 540 mutex_obj_hold(lock); 541 mutex_enter(lock); 542 } else { 543 /* Lock assigned and taken during PRU_ATTACH. */ 544 } 545 error = (*prp->pr_usrreq)(so, PRU_ATTACH, NULL, 546 (struct mbuf *)(long)proto, NULL, l); 547 KASSERT(solocked(so)); 548 if (error != 0) { 549 so->so_state |= SS_NOFDREF; 550 sofree(so); 551 return error; 552 } 553 so->so_cred = kauth_cred_dup(l->l_cred); 554 sounlock(so); 555 *aso = so; 556 return 0; 557 } 558 559 /* On success, write file descriptor to fdout and return zero. On 560 * failure, return non-zero; *fdout will be undefined. 561 */ 562 int 563 fsocreate(int domain, struct socket **sop, int type, int protocol, 564 struct lwp *l, int *fdout) 565 { 566 struct socket *so; 567 struct file *fp; 568 int fd, error; 569 int flags = type & SOCK_FLAGS_MASK; 570 571 type &= ~SOCK_FLAGS_MASK; 572 if ((error = fd_allocfile(&fp, &fd)) != 0) 573 return error; 574 fd_set_exclose(l, fd, (flags & SOCK_CLOEXEC) != 0); 575 fp->f_flag = FREAD|FWRITE|((flags & SOCK_NONBLOCK) ? FNONBLOCK : 0)| 576 ((flags & SOCK_NOSIGPIPE) ? FNOSIGPIPE : 0); 577 fp->f_type = DTYPE_SOCKET; 578 fp->f_ops = &socketops; 579 error = socreate(domain, &so, type, protocol, l, NULL); 580 if (error != 0) { 581 fd_abort(curproc, fp, fd); 582 } else { 583 if (sop != NULL) 584 *sop = so; 585 fp->f_data = so; 586 fd_affix(curproc, fp, fd); 587 *fdout = fd; 588 } 589 return error; 590 } 591 592 int 593 sofamily(const struct socket *so) 594 { 595 const struct protosw *pr; 596 const struct domain *dom; 597 598 if ((pr = so->so_proto) == NULL) 599 return AF_UNSPEC; 600 if ((dom = pr->pr_domain) == NULL) 601 return AF_UNSPEC; 602 return dom->dom_family; 603 } 604 605 int 606 sobind(struct socket *so, struct mbuf *nam, struct lwp *l) 607 { 608 int error; 609 610 solock(so); 611 error = (*so->so_proto->pr_usrreq)(so, PRU_BIND, NULL, nam, NULL, l); 612 sounlock(so); 613 return error; 614 } 615 616 int 617 solisten(struct socket *so, int backlog, struct lwp *l) 618 { 619 int error; 620 621 solock(so); 622 if ((so->so_state & (SS_ISCONNECTED | SS_ISCONNECTING | 623 SS_ISDISCONNECTING)) != 0) { 624 sounlock(so); 625 return (EOPNOTSUPP); 626 } 627 error = (*so->so_proto->pr_usrreq)(so, PRU_LISTEN, NULL, 628 NULL, NULL, l); 629 if (error != 0) { 630 sounlock(so); 631 return error; 632 } 633 if (TAILQ_EMPTY(&so->so_q)) 634 so->so_options |= SO_ACCEPTCONN; 635 if (backlog < 0) 636 backlog = 0; 637 so->so_qlimit = min(backlog, somaxconn); 638 sounlock(so); 639 return 0; 640 } 641 642 void 643 sofree(struct socket *so) 644 { 645 u_int refs; 646 647 KASSERT(solocked(so)); 648 649 if (so->so_pcb || (so->so_state & SS_NOFDREF) == 0) { 650 sounlock(so); 651 return; 652 } 653 if (so->so_head) { 654 /* 655 * We must not decommission a socket that's on the accept(2) 656 * queue. If we do, then accept(2) may hang after select(2) 657 * indicated that the listening socket was ready. 658 */ 659 if (!soqremque(so, 0)) { 660 sounlock(so); 661 return; 662 } 663 } 664 if (so->so_rcv.sb_hiwat) 665 (void)chgsbsize(so->so_uidinfo, &so->so_rcv.sb_hiwat, 0, 666 RLIM_INFINITY); 667 if (so->so_snd.sb_hiwat) 668 (void)chgsbsize(so->so_uidinfo, &so->so_snd.sb_hiwat, 0, 669 RLIM_INFINITY); 670 sbrelease(&so->so_snd, so); 671 KASSERT(!cv_has_waiters(&so->so_cv)); 672 KASSERT(!cv_has_waiters(&so->so_rcv.sb_cv)); 673 KASSERT(!cv_has_waiters(&so->so_snd.sb_cv)); 674 sorflush(so); 675 refs = so->so_aborting; /* XXX */ 676 /* Remove acccept filter if one is present. */ 677 if (so->so_accf != NULL) 678 (void)accept_filt_clear(so); 679 sounlock(so); 680 if (refs == 0) /* XXX */ 681 soput(so); 682 } 683 684 /* 685 * Close a socket on last file table reference removal. 686 * Initiate disconnect if connected. 687 * Free socket when disconnect complete. 688 */ 689 int 690 soclose(struct socket *so) 691 { 692 struct socket *so2; 693 int error; 694 int error2; 695 696 error = 0; 697 solock(so); 698 if (so->so_options & SO_ACCEPTCONN) { 699 for (;;) { 700 if ((so2 = TAILQ_FIRST(&so->so_q0)) != 0) { 701 KASSERT(solocked2(so, so2)); 702 (void) soqremque(so2, 0); 703 /* soabort drops the lock. */ 704 (void) soabort(so2); 705 solock(so); 706 continue; 707 } 708 if ((so2 = TAILQ_FIRST(&so->so_q)) != 0) { 709 KASSERT(solocked2(so, so2)); 710 (void) soqremque(so2, 1); 711 /* soabort drops the lock. */ 712 (void) soabort(so2); 713 solock(so); 714 continue; 715 } 716 break; 717 } 718 } 719 if (so->so_pcb == 0) 720 goto discard; 721 if (so->so_state & SS_ISCONNECTED) { 722 if ((so->so_state & SS_ISDISCONNECTING) == 0) { 723 error = sodisconnect(so); 724 if (error) 725 goto drop; 726 } 727 if (so->so_options & SO_LINGER) { 728 if ((so->so_state & (SS_ISDISCONNECTING|SS_NBIO)) == 729 (SS_ISDISCONNECTING|SS_NBIO)) 730 goto drop; 731 while (so->so_state & SS_ISCONNECTED) { 732 error = sowait(so, true, so->so_linger * hz); 733 if (error) 734 break; 735 } 736 } 737 } 738 drop: 739 if (so->so_pcb) { 740 error2 = (*so->so_proto->pr_usrreq)(so, PRU_DETACH, 741 NULL, NULL, NULL, NULL); 742 if (error == 0) 743 error = error2; 744 } 745 discard: 746 if (so->so_state & SS_NOFDREF) 747 panic("soclose: NOFDREF"); 748 kauth_cred_free(so->so_cred); 749 so->so_state |= SS_NOFDREF; 750 sofree(so); 751 return (error); 752 } 753 754 /* 755 * Must be called with the socket locked.. Will return with it unlocked. 756 */ 757 int 758 soabort(struct socket *so) 759 { 760 u_int refs; 761 int error; 762 763 KASSERT(solocked(so)); 764 KASSERT(so->so_head == NULL); 765 766 so->so_aborting++; /* XXX */ 767 error = (*so->so_proto->pr_usrreq)(so, PRU_ABORT, NULL, 768 NULL, NULL, NULL); 769 refs = --so->so_aborting; /* XXX */ 770 if (error || (refs == 0)) { 771 sofree(so); 772 } else { 773 sounlock(so); 774 } 775 return error; 776 } 777 778 int 779 soaccept(struct socket *so, struct mbuf *nam) 780 { 781 int error; 782 783 KASSERT(solocked(so)); 784 785 error = 0; 786 if ((so->so_state & SS_NOFDREF) == 0) 787 panic("soaccept: !NOFDREF"); 788 so->so_state &= ~SS_NOFDREF; 789 if ((so->so_state & SS_ISDISCONNECTED) == 0 || 790 (so->so_proto->pr_flags & PR_ABRTACPTDIS) == 0) 791 error = (*so->so_proto->pr_usrreq)(so, PRU_ACCEPT, 792 NULL, nam, NULL, NULL); 793 else 794 error = ECONNABORTED; 795 796 return (error); 797 } 798 799 int 800 soconnect(struct socket *so, struct mbuf *nam, struct lwp *l) 801 { 802 int error; 803 804 KASSERT(solocked(so)); 805 806 if (so->so_options & SO_ACCEPTCONN) 807 return (EOPNOTSUPP); 808 /* 809 * If protocol is connection-based, can only connect once. 810 * Otherwise, if connected, try to disconnect first. 811 * This allows user to disconnect by connecting to, e.g., 812 * a null address. 813 */ 814 if (so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING) && 815 ((so->so_proto->pr_flags & PR_CONNREQUIRED) || 816 (error = sodisconnect(so)))) 817 error = EISCONN; 818 else 819 error = (*so->so_proto->pr_usrreq)(so, PRU_CONNECT, 820 NULL, nam, NULL, l); 821 return (error); 822 } 823 824 int 825 soconnect2(struct socket *so1, struct socket *so2) 826 { 827 int error; 828 829 KASSERT(solocked2(so1, so2)); 830 831 error = (*so1->so_proto->pr_usrreq)(so1, PRU_CONNECT2, 832 NULL, (struct mbuf *)so2, NULL, NULL); 833 return (error); 834 } 835 836 int 837 sodisconnect(struct socket *so) 838 { 839 int error; 840 841 KASSERT(solocked(so)); 842 843 if ((so->so_state & SS_ISCONNECTED) == 0) { 844 error = ENOTCONN; 845 } else if (so->so_state & SS_ISDISCONNECTING) { 846 error = EALREADY; 847 } else { 848 error = (*so->so_proto->pr_usrreq)(so, PRU_DISCONNECT, 849 NULL, NULL, NULL, NULL); 850 } 851 return (error); 852 } 853 854 #define SBLOCKWAIT(f) (((f) & MSG_DONTWAIT) ? M_NOWAIT : M_WAITOK) 855 /* 856 * Send on a socket. 857 * If send must go all at once and message is larger than 858 * send buffering, then hard error. 859 * Lock against other senders. 860 * If must go all at once and not enough room now, then 861 * inform user that this would block and do nothing. 862 * Otherwise, if nonblocking, send as much as possible. 863 * The data to be sent is described by "uio" if nonzero, 864 * otherwise by the mbuf chain "top" (which must be null 865 * if uio is not). Data provided in mbuf chain must be small 866 * enough to send all at once. 867 * 868 * Returns nonzero on error, timeout or signal; callers 869 * must check for short counts if EINTR/ERESTART are returned. 870 * Data and control buffers are freed on return. 871 */ 872 int 873 sosend(struct socket *so, struct mbuf *addr, struct uio *uio, struct mbuf *top, 874 struct mbuf *control, int flags, struct lwp *l) 875 { 876 struct mbuf **mp, *m; 877 struct proc *p; 878 long space, len, resid, clen, mlen; 879 int error, s, dontroute, atomic; 880 short wakeup_state = 0; 881 882 p = l->l_proc; 883 clen = 0; 884 885 /* 886 * solock() provides atomicity of access. splsoftnet() prevents 887 * protocol processing soft interrupts from interrupting us and 888 * blocking (expensive). 889 */ 890 s = splsoftnet(); 891 solock(so); 892 atomic = sosendallatonce(so) || top; 893 if (uio) 894 resid = uio->uio_resid; 895 else 896 resid = top->m_pkthdr.len; 897 /* 898 * In theory resid should be unsigned. 899 * However, space must be signed, as it might be less than 0 900 * if we over-committed, and we must use a signed comparison 901 * of space and resid. On the other hand, a negative resid 902 * causes us to loop sending 0-length segments to the protocol. 903 */ 904 if (resid < 0) { 905 error = EINVAL; 906 goto out; 907 } 908 dontroute = 909 (flags & MSG_DONTROUTE) && (so->so_options & SO_DONTROUTE) == 0 && 910 (so->so_proto->pr_flags & PR_ATOMIC); 911 l->l_ru.ru_msgsnd++; 912 if (control) 913 clen = control->m_len; 914 restart: 915 if ((error = sblock(&so->so_snd, SBLOCKWAIT(flags))) != 0) 916 goto out; 917 do { 918 if (so->so_state & SS_CANTSENDMORE) { 919 error = EPIPE; 920 goto release; 921 } 922 if (so->so_error) { 923 error = so->so_error; 924 so->so_error = 0; 925 goto release; 926 } 927 if ((so->so_state & SS_ISCONNECTED) == 0) { 928 if (so->so_proto->pr_flags & PR_CONNREQUIRED) { 929 if ((so->so_state & SS_ISCONFIRMING) == 0 && 930 !(resid == 0 && clen != 0)) { 931 error = ENOTCONN; 932 goto release; 933 } 934 } else if (addr == 0) { 935 error = EDESTADDRREQ; 936 goto release; 937 } 938 } 939 space = sbspace(&so->so_snd); 940 if (flags & MSG_OOB) 941 space += 1024; 942 if ((atomic && resid > so->so_snd.sb_hiwat) || 943 clen > so->so_snd.sb_hiwat) { 944 error = EMSGSIZE; 945 goto release; 946 } 947 if (space < resid + clen && 948 (atomic || space < so->so_snd.sb_lowat || space < clen)) { 949 if ((so->so_state & SS_NBIO) || (flags & MSG_NBIO)) { 950 error = EWOULDBLOCK; 951 goto release; 952 } 953 sbunlock(&so->so_snd); 954 if (wakeup_state & SS_RESTARTSYS) { 955 error = ERESTART; 956 goto out; 957 } 958 error = sbwait(&so->so_snd); 959 if (error) 960 goto out; 961 wakeup_state = so->so_state; 962 goto restart; 963 } 964 wakeup_state = 0; 965 mp = ⊤ 966 space -= clen; 967 do { 968 if (uio == NULL) { 969 /* 970 * Data is prepackaged in "top". 971 */ 972 resid = 0; 973 if (flags & MSG_EOR) 974 top->m_flags |= M_EOR; 975 } else do { 976 sounlock(so); 977 splx(s); 978 if (top == NULL) { 979 m = m_gethdr(M_WAIT, MT_DATA); 980 mlen = MHLEN; 981 m->m_pkthdr.len = 0; 982 m->m_pkthdr.rcvif = NULL; 983 } else { 984 m = m_get(M_WAIT, MT_DATA); 985 mlen = MLEN; 986 } 987 MCLAIM(m, so->so_snd.sb_mowner); 988 if (sock_loan_thresh >= 0 && 989 uio->uio_iov->iov_len >= sock_loan_thresh && 990 space >= sock_loan_thresh && 991 (len = sosend_loan(so, uio, m, 992 space)) != 0) { 993 SOSEND_COUNTER_INCR(&sosend_loan_big); 994 space -= len; 995 goto have_data; 996 } 997 if (resid >= MINCLSIZE && space >= MCLBYTES) { 998 SOSEND_COUNTER_INCR(&sosend_copy_big); 999 m_clget(m, M_DONTWAIT); 1000 if ((m->m_flags & M_EXT) == 0) 1001 goto nopages; 1002 mlen = MCLBYTES; 1003 if (atomic && top == 0) { 1004 len = lmin(MCLBYTES - max_hdr, 1005 resid); 1006 m->m_data += max_hdr; 1007 } else 1008 len = lmin(MCLBYTES, resid); 1009 space -= len; 1010 } else { 1011 nopages: 1012 SOSEND_COUNTER_INCR(&sosend_copy_small); 1013 len = lmin(lmin(mlen, resid), space); 1014 space -= len; 1015 /* 1016 * For datagram protocols, leave room 1017 * for protocol headers in first mbuf. 1018 */ 1019 if (atomic && top == 0 && len < mlen) 1020 MH_ALIGN(m, len); 1021 } 1022 error = uiomove(mtod(m, void *), (int)len, uio); 1023 have_data: 1024 resid = uio->uio_resid; 1025 m->m_len = len; 1026 *mp = m; 1027 top->m_pkthdr.len += len; 1028 s = splsoftnet(); 1029 solock(so); 1030 if (error != 0) 1031 goto release; 1032 mp = &m->m_next; 1033 if (resid <= 0) { 1034 if (flags & MSG_EOR) 1035 top->m_flags |= M_EOR; 1036 break; 1037 } 1038 } while (space > 0 && atomic); 1039 1040 if (so->so_state & SS_CANTSENDMORE) { 1041 error = EPIPE; 1042 goto release; 1043 } 1044 if (dontroute) 1045 so->so_options |= SO_DONTROUTE; 1046 if (resid > 0) 1047 so->so_state |= SS_MORETOCOME; 1048 error = (*so->so_proto->pr_usrreq)(so, 1049 (flags & MSG_OOB) ? PRU_SENDOOB : PRU_SEND, 1050 top, addr, control, curlwp); 1051 if (dontroute) 1052 so->so_options &= ~SO_DONTROUTE; 1053 if (resid > 0) 1054 so->so_state &= ~SS_MORETOCOME; 1055 clen = 0; 1056 control = NULL; 1057 top = NULL; 1058 mp = ⊤ 1059 if (error != 0) 1060 goto release; 1061 } while (resid && space > 0); 1062 } while (resid); 1063 1064 release: 1065 sbunlock(&so->so_snd); 1066 out: 1067 sounlock(so); 1068 splx(s); 1069 if (top) 1070 m_freem(top); 1071 if (control) 1072 m_freem(control); 1073 return (error); 1074 } 1075 1076 /* 1077 * Following replacement or removal of the first mbuf on the first 1078 * mbuf chain of a socket buffer, push necessary state changes back 1079 * into the socket buffer so that other consumers see the values 1080 * consistently. 'nextrecord' is the callers locally stored value of 1081 * the original value of sb->sb_mb->m_nextpkt which must be restored 1082 * when the lead mbuf changes. NOTE: 'nextrecord' may be NULL. 1083 */ 1084 static void 1085 sbsync(struct sockbuf *sb, struct mbuf *nextrecord) 1086 { 1087 1088 KASSERT(solocked(sb->sb_so)); 1089 1090 /* 1091 * First, update for the new value of nextrecord. If necessary, 1092 * make it the first record. 1093 */ 1094 if (sb->sb_mb != NULL) 1095 sb->sb_mb->m_nextpkt = nextrecord; 1096 else 1097 sb->sb_mb = nextrecord; 1098 1099 /* 1100 * Now update any dependent socket buffer fields to reflect 1101 * the new state. This is an inline of SB_EMPTY_FIXUP, with 1102 * the addition of a second clause that takes care of the 1103 * case where sb_mb has been updated, but remains the last 1104 * record. 1105 */ 1106 if (sb->sb_mb == NULL) { 1107 sb->sb_mbtail = NULL; 1108 sb->sb_lastrecord = NULL; 1109 } else if (sb->sb_mb->m_nextpkt == NULL) 1110 sb->sb_lastrecord = sb->sb_mb; 1111 } 1112 1113 /* 1114 * Implement receive operations on a socket. 1115 * We depend on the way that records are added to the sockbuf 1116 * by sbappend*. In particular, each record (mbufs linked through m_next) 1117 * must begin with an address if the protocol so specifies, 1118 * followed by an optional mbuf or mbufs containing ancillary data, 1119 * and then zero or more mbufs of data. 1120 * In order to avoid blocking network interrupts for the entire time here, 1121 * we splx() while doing the actual copy to user space. 1122 * Although the sockbuf is locked, new data may still be appended, 1123 * and thus we must maintain consistency of the sockbuf during that time. 1124 * 1125 * The caller may receive the data as a single mbuf chain by supplying 1126 * an mbuf **mp0 for use in returning the chain. The uio is then used 1127 * only for the count in uio_resid. 1128 */ 1129 int 1130 soreceive(struct socket *so, struct mbuf **paddr, struct uio *uio, 1131 struct mbuf **mp0, struct mbuf **controlp, int *flagsp) 1132 { 1133 struct lwp *l = curlwp; 1134 struct mbuf *m, **mp, *mt; 1135 int atomic, flags, len, error, s, offset, moff, type, orig_resid; 1136 const struct protosw *pr; 1137 struct mbuf *nextrecord; 1138 int mbuf_removed = 0; 1139 const struct domain *dom; 1140 short wakeup_state = 0; 1141 1142 pr = so->so_proto; 1143 atomic = pr->pr_flags & PR_ATOMIC; 1144 dom = pr->pr_domain; 1145 mp = mp0; 1146 type = 0; 1147 orig_resid = uio->uio_resid; 1148 1149 if (paddr != NULL) 1150 *paddr = NULL; 1151 if (controlp != NULL) 1152 *controlp = NULL; 1153 if (flagsp != NULL) 1154 flags = *flagsp &~ MSG_EOR; 1155 else 1156 flags = 0; 1157 1158 if (flags & MSG_OOB) { 1159 m = m_get(M_WAIT, MT_DATA); 1160 solock(so); 1161 error = (*pr->pr_usrreq)(so, PRU_RCVOOB, m, 1162 (struct mbuf *)(long)(flags & MSG_PEEK), NULL, l); 1163 sounlock(so); 1164 if (error) 1165 goto bad; 1166 do { 1167 error = uiomove(mtod(m, void *), 1168 (int) min(uio->uio_resid, m->m_len), uio); 1169 m = m_free(m); 1170 } while (uio->uio_resid > 0 && error == 0 && m); 1171 bad: 1172 if (m != NULL) 1173 m_freem(m); 1174 return error; 1175 } 1176 if (mp != NULL) 1177 *mp = NULL; 1178 1179 /* 1180 * solock() provides atomicity of access. splsoftnet() prevents 1181 * protocol processing soft interrupts from interrupting us and 1182 * blocking (expensive). 1183 */ 1184 s = splsoftnet(); 1185 solock(so); 1186 if (so->so_state & SS_ISCONFIRMING && uio->uio_resid) 1187 (*pr->pr_usrreq)(so, PRU_RCVD, NULL, NULL, NULL, l); 1188 1189 restart: 1190 if ((error = sblock(&so->so_rcv, SBLOCKWAIT(flags))) != 0) { 1191 sounlock(so); 1192 splx(s); 1193 return error; 1194 } 1195 1196 m = so->so_rcv.sb_mb; 1197 /* 1198 * If we have less data than requested, block awaiting more 1199 * (subject to any timeout) if: 1200 * 1. the current count is less than the low water mark, 1201 * 2. MSG_WAITALL is set, and it is possible to do the entire 1202 * receive operation at once if we block (resid <= hiwat), or 1203 * 3. MSG_DONTWAIT is not set. 1204 * If MSG_WAITALL is set but resid is larger than the receive buffer, 1205 * we have to do the receive in sections, and thus risk returning 1206 * a short count if a timeout or signal occurs after we start. 1207 */ 1208 if (m == NULL || 1209 ((flags & MSG_DONTWAIT) == 0 && 1210 so->so_rcv.sb_cc < uio->uio_resid && 1211 (so->so_rcv.sb_cc < so->so_rcv.sb_lowat || 1212 ((flags & MSG_WAITALL) && 1213 uio->uio_resid <= so->so_rcv.sb_hiwat)) && 1214 m->m_nextpkt == NULL && !atomic)) { 1215 #ifdef DIAGNOSTIC 1216 if (m == NULL && so->so_rcv.sb_cc) 1217 panic("receive 1"); 1218 #endif 1219 if (so->so_error) { 1220 if (m != NULL) 1221 goto dontblock; 1222 error = so->so_error; 1223 if ((flags & MSG_PEEK) == 0) 1224 so->so_error = 0; 1225 goto release; 1226 } 1227 if (so->so_state & SS_CANTRCVMORE) { 1228 if (m != NULL) 1229 goto dontblock; 1230 else 1231 goto release; 1232 } 1233 for (; m != NULL; m = m->m_next) 1234 if (m->m_type == MT_OOBDATA || (m->m_flags & M_EOR)) { 1235 m = so->so_rcv.sb_mb; 1236 goto dontblock; 1237 } 1238 if ((so->so_state & (SS_ISCONNECTED|SS_ISCONNECTING)) == 0 && 1239 (so->so_proto->pr_flags & PR_CONNREQUIRED)) { 1240 error = ENOTCONN; 1241 goto release; 1242 } 1243 if (uio->uio_resid == 0) 1244 goto release; 1245 if ((so->so_state & SS_NBIO) || 1246 (flags & (MSG_DONTWAIT|MSG_NBIO))) { 1247 error = EWOULDBLOCK; 1248 goto release; 1249 } 1250 SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 1"); 1251 SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 1"); 1252 sbunlock(&so->so_rcv); 1253 if (wakeup_state & SS_RESTARTSYS) 1254 error = ERESTART; 1255 else 1256 error = sbwait(&so->so_rcv); 1257 if (error != 0) { 1258 sounlock(so); 1259 splx(s); 1260 return error; 1261 } 1262 wakeup_state = so->so_state; 1263 goto restart; 1264 } 1265 dontblock: 1266 /* 1267 * On entry here, m points to the first record of the socket buffer. 1268 * From this point onward, we maintain 'nextrecord' as a cache of the 1269 * pointer to the next record in the socket buffer. We must keep the 1270 * various socket buffer pointers and local stack versions of the 1271 * pointers in sync, pushing out modifications before dropping the 1272 * socket lock, and re-reading them when picking it up. 1273 * 1274 * Otherwise, we will race with the network stack appending new data 1275 * or records onto the socket buffer by using inconsistent/stale 1276 * versions of the field, possibly resulting in socket buffer 1277 * corruption. 1278 * 1279 * By holding the high-level sblock(), we prevent simultaneous 1280 * readers from pulling off the front of the socket buffer. 1281 */ 1282 if (l != NULL) 1283 l->l_ru.ru_msgrcv++; 1284 KASSERT(m == so->so_rcv.sb_mb); 1285 SBLASTRECORDCHK(&so->so_rcv, "soreceive 1"); 1286 SBLASTMBUFCHK(&so->so_rcv, "soreceive 1"); 1287 nextrecord = m->m_nextpkt; 1288 if (pr->pr_flags & PR_ADDR) { 1289 #ifdef DIAGNOSTIC 1290 if (m->m_type != MT_SONAME) 1291 panic("receive 1a"); 1292 #endif 1293 orig_resid = 0; 1294 if (flags & MSG_PEEK) { 1295 if (paddr) 1296 *paddr = m_copy(m, 0, m->m_len); 1297 m = m->m_next; 1298 } else { 1299 sbfree(&so->so_rcv, m); 1300 mbuf_removed = 1; 1301 if (paddr != NULL) { 1302 *paddr = m; 1303 so->so_rcv.sb_mb = m->m_next; 1304 m->m_next = NULL; 1305 m = so->so_rcv.sb_mb; 1306 } else { 1307 MFREE(m, so->so_rcv.sb_mb); 1308 m = so->so_rcv.sb_mb; 1309 } 1310 sbsync(&so->so_rcv, nextrecord); 1311 } 1312 } 1313 1314 /* 1315 * Process one or more MT_CONTROL mbufs present before any data mbufs 1316 * in the first mbuf chain on the socket buffer. If MSG_PEEK, we 1317 * just copy the data; if !MSG_PEEK, we call into the protocol to 1318 * perform externalization (or freeing if controlp == NULL). 1319 */ 1320 if (__predict_false(m != NULL && m->m_type == MT_CONTROL)) { 1321 struct mbuf *cm = NULL, *cmn; 1322 struct mbuf **cme = &cm; 1323 1324 do { 1325 if (flags & MSG_PEEK) { 1326 if (controlp != NULL) { 1327 *controlp = m_copy(m, 0, m->m_len); 1328 controlp = &(*controlp)->m_next; 1329 } 1330 m = m->m_next; 1331 } else { 1332 sbfree(&so->so_rcv, m); 1333 so->so_rcv.sb_mb = m->m_next; 1334 m->m_next = NULL; 1335 *cme = m; 1336 cme = &(*cme)->m_next; 1337 m = so->so_rcv.sb_mb; 1338 } 1339 } while (m != NULL && m->m_type == MT_CONTROL); 1340 if ((flags & MSG_PEEK) == 0) 1341 sbsync(&so->so_rcv, nextrecord); 1342 for (; cm != NULL; cm = cmn) { 1343 cmn = cm->m_next; 1344 cm->m_next = NULL; 1345 type = mtod(cm, struct cmsghdr *)->cmsg_type; 1346 if (controlp != NULL) { 1347 if (dom->dom_externalize != NULL && 1348 type == SCM_RIGHTS) { 1349 sounlock(so); 1350 splx(s); 1351 error = (*dom->dom_externalize)(cm, l, 1352 (flags & MSG_CMSG_CLOEXEC) ? 1353 O_CLOEXEC : 0); 1354 s = splsoftnet(); 1355 solock(so); 1356 } 1357 *controlp = cm; 1358 while (*controlp != NULL) 1359 controlp = &(*controlp)->m_next; 1360 } else { 1361 /* 1362 * Dispose of any SCM_RIGHTS message that went 1363 * through the read path rather than recv. 1364 */ 1365 if (dom->dom_dispose != NULL && 1366 type == SCM_RIGHTS) { 1367 sounlock(so); 1368 (*dom->dom_dispose)(cm); 1369 solock(so); 1370 } 1371 m_freem(cm); 1372 } 1373 } 1374 if (m != NULL) 1375 nextrecord = so->so_rcv.sb_mb->m_nextpkt; 1376 else 1377 nextrecord = so->so_rcv.sb_mb; 1378 orig_resid = 0; 1379 } 1380 1381 /* If m is non-NULL, we have some data to read. */ 1382 if (__predict_true(m != NULL)) { 1383 type = m->m_type; 1384 if (type == MT_OOBDATA) 1385 flags |= MSG_OOB; 1386 } 1387 SBLASTRECORDCHK(&so->so_rcv, "soreceive 2"); 1388 SBLASTMBUFCHK(&so->so_rcv, "soreceive 2"); 1389 1390 moff = 0; 1391 offset = 0; 1392 while (m != NULL && uio->uio_resid > 0 && error == 0) { 1393 if (m->m_type == MT_OOBDATA) { 1394 if (type != MT_OOBDATA) 1395 break; 1396 } else if (type == MT_OOBDATA) 1397 break; 1398 #ifdef DIAGNOSTIC 1399 else if (m->m_type != MT_DATA && m->m_type != MT_HEADER) 1400 panic("receive 3"); 1401 #endif 1402 so->so_state &= ~SS_RCVATMARK; 1403 wakeup_state = 0; 1404 len = uio->uio_resid; 1405 if (so->so_oobmark && len > so->so_oobmark - offset) 1406 len = so->so_oobmark - offset; 1407 if (len > m->m_len - moff) 1408 len = m->m_len - moff; 1409 /* 1410 * If mp is set, just pass back the mbufs. 1411 * Otherwise copy them out via the uio, then free. 1412 * Sockbuf must be consistent here (points to current mbuf, 1413 * it points to next record) when we drop priority; 1414 * we must note any additions to the sockbuf when we 1415 * block interrupts again. 1416 */ 1417 if (mp == NULL) { 1418 SBLASTRECORDCHK(&so->so_rcv, "soreceive uiomove"); 1419 SBLASTMBUFCHK(&so->so_rcv, "soreceive uiomove"); 1420 sounlock(so); 1421 splx(s); 1422 error = uiomove(mtod(m, char *) + moff, (int)len, uio); 1423 s = splsoftnet(); 1424 solock(so); 1425 if (error != 0) { 1426 /* 1427 * If any part of the record has been removed 1428 * (such as the MT_SONAME mbuf, which will 1429 * happen when PR_ADDR, and thus also 1430 * PR_ATOMIC, is set), then drop the entire 1431 * record to maintain the atomicity of the 1432 * receive operation. 1433 * 1434 * This avoids a later panic("receive 1a") 1435 * when compiled with DIAGNOSTIC. 1436 */ 1437 if (m && mbuf_removed && atomic) 1438 (void) sbdroprecord(&so->so_rcv); 1439 1440 goto release; 1441 } 1442 } else 1443 uio->uio_resid -= len; 1444 if (len == m->m_len - moff) { 1445 if (m->m_flags & M_EOR) 1446 flags |= MSG_EOR; 1447 if (flags & MSG_PEEK) { 1448 m = m->m_next; 1449 moff = 0; 1450 } else { 1451 nextrecord = m->m_nextpkt; 1452 sbfree(&so->so_rcv, m); 1453 if (mp) { 1454 *mp = m; 1455 mp = &m->m_next; 1456 so->so_rcv.sb_mb = m = m->m_next; 1457 *mp = NULL; 1458 } else { 1459 MFREE(m, so->so_rcv.sb_mb); 1460 m = so->so_rcv.sb_mb; 1461 } 1462 /* 1463 * If m != NULL, we also know that 1464 * so->so_rcv.sb_mb != NULL. 1465 */ 1466 KASSERT(so->so_rcv.sb_mb == m); 1467 if (m) { 1468 m->m_nextpkt = nextrecord; 1469 if (nextrecord == NULL) 1470 so->so_rcv.sb_lastrecord = m; 1471 } else { 1472 so->so_rcv.sb_mb = nextrecord; 1473 SB_EMPTY_FIXUP(&so->so_rcv); 1474 } 1475 SBLASTRECORDCHK(&so->so_rcv, "soreceive 3"); 1476 SBLASTMBUFCHK(&so->so_rcv, "soreceive 3"); 1477 } 1478 } else if (flags & MSG_PEEK) 1479 moff += len; 1480 else { 1481 if (mp != NULL) { 1482 mt = m_copym(m, 0, len, M_NOWAIT); 1483 if (__predict_false(mt == NULL)) { 1484 sounlock(so); 1485 mt = m_copym(m, 0, len, M_WAIT); 1486 solock(so); 1487 } 1488 *mp = mt; 1489 } 1490 m->m_data += len; 1491 m->m_len -= len; 1492 so->so_rcv.sb_cc -= len; 1493 } 1494 if (so->so_oobmark) { 1495 if ((flags & MSG_PEEK) == 0) { 1496 so->so_oobmark -= len; 1497 if (so->so_oobmark == 0) { 1498 so->so_state |= SS_RCVATMARK; 1499 break; 1500 } 1501 } else { 1502 offset += len; 1503 if (offset == so->so_oobmark) 1504 break; 1505 } 1506 } 1507 if (flags & MSG_EOR) 1508 break; 1509 /* 1510 * If the MSG_WAITALL flag is set (for non-atomic socket), 1511 * we must not quit until "uio->uio_resid == 0" or an error 1512 * termination. If a signal/timeout occurs, return 1513 * with a short count but without error. 1514 * Keep sockbuf locked against other readers. 1515 */ 1516 while (flags & MSG_WAITALL && m == NULL && uio->uio_resid > 0 && 1517 !sosendallatonce(so) && !nextrecord) { 1518 if (so->so_error || so->so_state & SS_CANTRCVMORE) 1519 break; 1520 /* 1521 * If we are peeking and the socket receive buffer is 1522 * full, stop since we can't get more data to peek at. 1523 */ 1524 if ((flags & MSG_PEEK) && sbspace(&so->so_rcv) <= 0) 1525 break; 1526 /* 1527 * If we've drained the socket buffer, tell the 1528 * protocol in case it needs to do something to 1529 * get it filled again. 1530 */ 1531 if ((pr->pr_flags & PR_WANTRCVD) && so->so_pcb) 1532 (*pr->pr_usrreq)(so, PRU_RCVD, 1533 NULL, (struct mbuf *)(long)flags, NULL, l); 1534 SBLASTRECORDCHK(&so->so_rcv, "soreceive sbwait 2"); 1535 SBLASTMBUFCHK(&so->so_rcv, "soreceive sbwait 2"); 1536 if (wakeup_state & SS_RESTARTSYS) 1537 error = ERESTART; 1538 else 1539 error = sbwait(&so->so_rcv); 1540 if (error != 0) { 1541 sbunlock(&so->so_rcv); 1542 sounlock(so); 1543 splx(s); 1544 return 0; 1545 } 1546 if ((m = so->so_rcv.sb_mb) != NULL) 1547 nextrecord = m->m_nextpkt; 1548 wakeup_state = so->so_state; 1549 } 1550 } 1551 1552 if (m && atomic) { 1553 flags |= MSG_TRUNC; 1554 if ((flags & MSG_PEEK) == 0) 1555 (void) sbdroprecord(&so->so_rcv); 1556 } 1557 if ((flags & MSG_PEEK) == 0) { 1558 if (m == NULL) { 1559 /* 1560 * First part is an inline SB_EMPTY_FIXUP(). Second 1561 * part makes sure sb_lastrecord is up-to-date if 1562 * there is still data in the socket buffer. 1563 */ 1564 so->so_rcv.sb_mb = nextrecord; 1565 if (so->so_rcv.sb_mb == NULL) { 1566 so->so_rcv.sb_mbtail = NULL; 1567 so->so_rcv.sb_lastrecord = NULL; 1568 } else if (nextrecord->m_nextpkt == NULL) 1569 so->so_rcv.sb_lastrecord = nextrecord; 1570 } 1571 SBLASTRECORDCHK(&so->so_rcv, "soreceive 4"); 1572 SBLASTMBUFCHK(&so->so_rcv, "soreceive 4"); 1573 if (pr->pr_flags & PR_WANTRCVD && so->so_pcb) 1574 (*pr->pr_usrreq)(so, PRU_RCVD, NULL, 1575 (struct mbuf *)(long)flags, NULL, l); 1576 } 1577 if (orig_resid == uio->uio_resid && orig_resid && 1578 (flags & MSG_EOR) == 0 && (so->so_state & SS_CANTRCVMORE) == 0) { 1579 sbunlock(&so->so_rcv); 1580 goto restart; 1581 } 1582 1583 if (flagsp != NULL) 1584 *flagsp |= flags; 1585 release: 1586 sbunlock(&so->so_rcv); 1587 sounlock(so); 1588 splx(s); 1589 return error; 1590 } 1591 1592 int 1593 soshutdown(struct socket *so, int how) 1594 { 1595 const struct protosw *pr; 1596 int error; 1597 1598 KASSERT(solocked(so)); 1599 1600 pr = so->so_proto; 1601 if (!(how == SHUT_RD || how == SHUT_WR || how == SHUT_RDWR)) 1602 return (EINVAL); 1603 1604 if (how == SHUT_RD || how == SHUT_RDWR) { 1605 sorflush(so); 1606 error = 0; 1607 } 1608 if (how == SHUT_WR || how == SHUT_RDWR) 1609 error = (*pr->pr_usrreq)(so, PRU_SHUTDOWN, NULL, 1610 NULL, NULL, NULL); 1611 1612 return error; 1613 } 1614 1615 void 1616 sorestart(struct socket *so) 1617 { 1618 /* 1619 * An application has called close() on an fd on which another 1620 * of its threads has called a socket system call. 1621 * Mark this and wake everyone up, and code that would block again 1622 * instead returns ERESTART. 1623 * On system call re-entry the fd is validated and EBADF returned. 1624 * Any other fd will block again on the 2nd syscall. 1625 */ 1626 solock(so); 1627 so->so_state |= SS_RESTARTSYS; 1628 cv_broadcast(&so->so_cv); 1629 cv_broadcast(&so->so_snd.sb_cv); 1630 cv_broadcast(&so->so_rcv.sb_cv); 1631 sounlock(so); 1632 } 1633 1634 void 1635 sorflush(struct socket *so) 1636 { 1637 struct sockbuf *sb, asb; 1638 const struct protosw *pr; 1639 1640 KASSERT(solocked(so)); 1641 1642 sb = &so->so_rcv; 1643 pr = so->so_proto; 1644 socantrcvmore(so); 1645 sb->sb_flags |= SB_NOINTR; 1646 (void )sblock(sb, M_WAITOK); 1647 sbunlock(sb); 1648 asb = *sb; 1649 /* 1650 * Clear most of the sockbuf structure, but leave some of the 1651 * fields valid. 1652 */ 1653 memset(&sb->sb_startzero, 0, 1654 sizeof(*sb) - offsetof(struct sockbuf, sb_startzero)); 1655 if (pr->pr_flags & PR_RIGHTS && pr->pr_domain->dom_dispose) { 1656 sounlock(so); 1657 (*pr->pr_domain->dom_dispose)(asb.sb_mb); 1658 solock(so); 1659 } 1660 sbrelease(&asb, so); 1661 } 1662 1663 /* 1664 * internal set SOL_SOCKET options 1665 */ 1666 static int 1667 sosetopt1(struct socket *so, const struct sockopt *sopt) 1668 { 1669 int error = EINVAL, optval, opt; 1670 struct linger l; 1671 struct timeval tv; 1672 1673 switch ((opt = sopt->sopt_name)) { 1674 1675 case SO_ACCEPTFILTER: 1676 error = accept_filt_setopt(so, sopt); 1677 KASSERT(solocked(so)); 1678 break; 1679 1680 case SO_LINGER: 1681 error = sockopt_get(sopt, &l, sizeof(l)); 1682 solock(so); 1683 if (error) 1684 break; 1685 if (l.l_linger < 0 || l.l_linger > USHRT_MAX || 1686 l.l_linger > (INT_MAX / hz)) { 1687 error = EDOM; 1688 break; 1689 } 1690 so->so_linger = l.l_linger; 1691 if (l.l_onoff) 1692 so->so_options |= SO_LINGER; 1693 else 1694 so->so_options &= ~SO_LINGER; 1695 break; 1696 1697 case SO_DEBUG: 1698 case SO_KEEPALIVE: 1699 case SO_DONTROUTE: 1700 case SO_USELOOPBACK: 1701 case SO_BROADCAST: 1702 case SO_REUSEADDR: 1703 case SO_REUSEPORT: 1704 case SO_OOBINLINE: 1705 case SO_TIMESTAMP: 1706 case SO_NOSIGPIPE: 1707 #ifdef SO_OTIMESTAMP 1708 case SO_OTIMESTAMP: 1709 #endif 1710 error = sockopt_getint(sopt, &optval); 1711 solock(so); 1712 if (error) 1713 break; 1714 if (optval) 1715 so->so_options |= opt; 1716 else 1717 so->so_options &= ~opt; 1718 break; 1719 1720 case SO_SNDBUF: 1721 case SO_RCVBUF: 1722 case SO_SNDLOWAT: 1723 case SO_RCVLOWAT: 1724 error = sockopt_getint(sopt, &optval); 1725 solock(so); 1726 if (error) 1727 break; 1728 1729 /* 1730 * Values < 1 make no sense for any of these 1731 * options, so disallow them. 1732 */ 1733 if (optval < 1) { 1734 error = EINVAL; 1735 break; 1736 } 1737 1738 switch (opt) { 1739 case SO_SNDBUF: 1740 if (sbreserve(&so->so_snd, (u_long)optval, so) == 0) { 1741 error = ENOBUFS; 1742 break; 1743 } 1744 so->so_snd.sb_flags &= ~SB_AUTOSIZE; 1745 break; 1746 1747 case SO_RCVBUF: 1748 if (sbreserve(&so->so_rcv, (u_long)optval, so) == 0) { 1749 error = ENOBUFS; 1750 break; 1751 } 1752 so->so_rcv.sb_flags &= ~SB_AUTOSIZE; 1753 break; 1754 1755 /* 1756 * Make sure the low-water is never greater than 1757 * the high-water. 1758 */ 1759 case SO_SNDLOWAT: 1760 if (optval > so->so_snd.sb_hiwat) 1761 optval = so->so_snd.sb_hiwat; 1762 1763 so->so_snd.sb_lowat = optval; 1764 break; 1765 1766 case SO_RCVLOWAT: 1767 if (optval > so->so_rcv.sb_hiwat) 1768 optval = so->so_rcv.sb_hiwat; 1769 1770 so->so_rcv.sb_lowat = optval; 1771 break; 1772 } 1773 break; 1774 1775 #ifdef COMPAT_50 1776 case SO_OSNDTIMEO: 1777 case SO_ORCVTIMEO: { 1778 struct timeval50 otv; 1779 error = sockopt_get(sopt, &otv, sizeof(otv)); 1780 if (error) { 1781 solock(so); 1782 break; 1783 } 1784 timeval50_to_timeval(&otv, &tv); 1785 opt = opt == SO_OSNDTIMEO ? SO_SNDTIMEO : SO_RCVTIMEO; 1786 error = 0; 1787 /*FALLTHROUGH*/ 1788 } 1789 #endif /* COMPAT_50 */ 1790 1791 case SO_SNDTIMEO: 1792 case SO_RCVTIMEO: 1793 if (error) 1794 error = sockopt_get(sopt, &tv, sizeof(tv)); 1795 solock(so); 1796 if (error) 1797 break; 1798 1799 if (tv.tv_sec > (INT_MAX - tv.tv_usec / tick) / hz) { 1800 error = EDOM; 1801 break; 1802 } 1803 1804 optval = tv.tv_sec * hz + tv.tv_usec / tick; 1805 if (optval == 0 && tv.tv_usec != 0) 1806 optval = 1; 1807 1808 switch (opt) { 1809 case SO_SNDTIMEO: 1810 so->so_snd.sb_timeo = optval; 1811 break; 1812 case SO_RCVTIMEO: 1813 so->so_rcv.sb_timeo = optval; 1814 break; 1815 } 1816 break; 1817 1818 default: 1819 solock(so); 1820 error = ENOPROTOOPT; 1821 break; 1822 } 1823 KASSERT(solocked(so)); 1824 return error; 1825 } 1826 1827 int 1828 sosetopt(struct socket *so, struct sockopt *sopt) 1829 { 1830 int error, prerr; 1831 1832 if (sopt->sopt_level == SOL_SOCKET) { 1833 error = sosetopt1(so, sopt); 1834 KASSERT(solocked(so)); 1835 } else { 1836 error = ENOPROTOOPT; 1837 solock(so); 1838 } 1839 1840 if ((error == 0 || error == ENOPROTOOPT) && 1841 so->so_proto != NULL && so->so_proto->pr_ctloutput != NULL) { 1842 /* give the protocol stack a shot */ 1843 prerr = (*so->so_proto->pr_ctloutput)(PRCO_SETOPT, so, sopt); 1844 if (prerr == 0) 1845 error = 0; 1846 else if (prerr != ENOPROTOOPT) 1847 error = prerr; 1848 } 1849 sounlock(so); 1850 return error; 1851 } 1852 1853 /* 1854 * so_setsockopt() is a wrapper providing a sockopt structure for sosetopt() 1855 */ 1856 int 1857 so_setsockopt(struct lwp *l, struct socket *so, int level, int name, 1858 const void *val, size_t valsize) 1859 { 1860 struct sockopt sopt; 1861 int error; 1862 1863 KASSERT(valsize == 0 || val != NULL); 1864 1865 sockopt_init(&sopt, level, name, valsize); 1866 sockopt_set(&sopt, val, valsize); 1867 1868 error = sosetopt(so, &sopt); 1869 1870 sockopt_destroy(&sopt); 1871 1872 return error; 1873 } 1874 1875 /* 1876 * internal get SOL_SOCKET options 1877 */ 1878 static int 1879 sogetopt1(struct socket *so, struct sockopt *sopt) 1880 { 1881 int error, optval, opt; 1882 struct linger l; 1883 struct timeval tv; 1884 1885 switch ((opt = sopt->sopt_name)) { 1886 1887 case SO_ACCEPTFILTER: 1888 error = accept_filt_getopt(so, sopt); 1889 break; 1890 1891 case SO_LINGER: 1892 l.l_onoff = (so->so_options & SO_LINGER) ? 1 : 0; 1893 l.l_linger = so->so_linger; 1894 1895 error = sockopt_set(sopt, &l, sizeof(l)); 1896 break; 1897 1898 case SO_USELOOPBACK: 1899 case SO_DONTROUTE: 1900 case SO_DEBUG: 1901 case SO_KEEPALIVE: 1902 case SO_REUSEADDR: 1903 case SO_REUSEPORT: 1904 case SO_BROADCAST: 1905 case SO_OOBINLINE: 1906 case SO_TIMESTAMP: 1907 case SO_NOSIGPIPE: 1908 #ifdef SO_OTIMESTAMP 1909 case SO_OTIMESTAMP: 1910 #endif 1911 error = sockopt_setint(sopt, (so->so_options & opt) ? 1 : 0); 1912 break; 1913 1914 case SO_TYPE: 1915 error = sockopt_setint(sopt, so->so_type); 1916 break; 1917 1918 case SO_ERROR: 1919 error = sockopt_setint(sopt, so->so_error); 1920 so->so_error = 0; 1921 break; 1922 1923 case SO_SNDBUF: 1924 error = sockopt_setint(sopt, so->so_snd.sb_hiwat); 1925 break; 1926 1927 case SO_RCVBUF: 1928 error = sockopt_setint(sopt, so->so_rcv.sb_hiwat); 1929 break; 1930 1931 case SO_SNDLOWAT: 1932 error = sockopt_setint(sopt, so->so_snd.sb_lowat); 1933 break; 1934 1935 case SO_RCVLOWAT: 1936 error = sockopt_setint(sopt, so->so_rcv.sb_lowat); 1937 break; 1938 1939 #ifdef COMPAT_50 1940 case SO_OSNDTIMEO: 1941 case SO_ORCVTIMEO: { 1942 struct timeval50 otv; 1943 1944 optval = (opt == SO_OSNDTIMEO ? 1945 so->so_snd.sb_timeo : so->so_rcv.sb_timeo); 1946 1947 otv.tv_sec = optval / hz; 1948 otv.tv_usec = (optval % hz) * tick; 1949 1950 error = sockopt_set(sopt, &otv, sizeof(otv)); 1951 break; 1952 } 1953 #endif /* COMPAT_50 */ 1954 1955 case SO_SNDTIMEO: 1956 case SO_RCVTIMEO: 1957 optval = (opt == SO_SNDTIMEO ? 1958 so->so_snd.sb_timeo : so->so_rcv.sb_timeo); 1959 1960 tv.tv_sec = optval / hz; 1961 tv.tv_usec = (optval % hz) * tick; 1962 1963 error = sockopt_set(sopt, &tv, sizeof(tv)); 1964 break; 1965 1966 case SO_OVERFLOWED: 1967 error = sockopt_setint(sopt, so->so_rcv.sb_overflowed); 1968 break; 1969 1970 default: 1971 error = ENOPROTOOPT; 1972 break; 1973 } 1974 1975 return (error); 1976 } 1977 1978 int 1979 sogetopt(struct socket *so, struct sockopt *sopt) 1980 { 1981 int error; 1982 1983 solock(so); 1984 if (sopt->sopt_level != SOL_SOCKET) { 1985 if (so->so_proto && so->so_proto->pr_ctloutput) { 1986 error = ((*so->so_proto->pr_ctloutput) 1987 (PRCO_GETOPT, so, sopt)); 1988 } else 1989 error = (ENOPROTOOPT); 1990 } else { 1991 error = sogetopt1(so, sopt); 1992 } 1993 sounlock(so); 1994 return (error); 1995 } 1996 1997 /* 1998 * alloc sockopt data buffer buffer 1999 * - will be released at destroy 2000 */ 2001 static int 2002 sockopt_alloc(struct sockopt *sopt, size_t len, km_flag_t kmflag) 2003 { 2004 2005 KASSERT(sopt->sopt_size == 0); 2006 2007 if (len > sizeof(sopt->sopt_buf)) { 2008 sopt->sopt_data = kmem_zalloc(len, kmflag); 2009 if (sopt->sopt_data == NULL) 2010 return ENOMEM; 2011 } else 2012 sopt->sopt_data = sopt->sopt_buf; 2013 2014 sopt->sopt_size = len; 2015 return 0; 2016 } 2017 2018 /* 2019 * initialise sockopt storage 2020 * - MAY sleep during allocation 2021 */ 2022 void 2023 sockopt_init(struct sockopt *sopt, int level, int name, size_t size) 2024 { 2025 2026 memset(sopt, 0, sizeof(*sopt)); 2027 2028 sopt->sopt_level = level; 2029 sopt->sopt_name = name; 2030 (void)sockopt_alloc(sopt, size, KM_SLEEP); 2031 } 2032 2033 /* 2034 * destroy sockopt storage 2035 * - will release any held memory references 2036 */ 2037 void 2038 sockopt_destroy(struct sockopt *sopt) 2039 { 2040 2041 if (sopt->sopt_data != sopt->sopt_buf) 2042 kmem_free(sopt->sopt_data, sopt->sopt_size); 2043 2044 memset(sopt, 0, sizeof(*sopt)); 2045 } 2046 2047 /* 2048 * set sockopt value 2049 * - value is copied into sockopt 2050 * - memory is allocated when necessary, will not sleep 2051 */ 2052 int 2053 sockopt_set(struct sockopt *sopt, const void *buf, size_t len) 2054 { 2055 int error; 2056 2057 if (sopt->sopt_size == 0) { 2058 error = sockopt_alloc(sopt, len, KM_NOSLEEP); 2059 if (error) 2060 return error; 2061 } 2062 2063 KASSERT(sopt->sopt_size == len); 2064 memcpy(sopt->sopt_data, buf, len); 2065 return 0; 2066 } 2067 2068 /* 2069 * common case of set sockopt integer value 2070 */ 2071 int 2072 sockopt_setint(struct sockopt *sopt, int val) 2073 { 2074 2075 return sockopt_set(sopt, &val, sizeof(int)); 2076 } 2077 2078 /* 2079 * get sockopt value 2080 * - correct size must be given 2081 */ 2082 int 2083 sockopt_get(const struct sockopt *sopt, void *buf, size_t len) 2084 { 2085 2086 if (sopt->sopt_size != len) 2087 return EINVAL; 2088 2089 memcpy(buf, sopt->sopt_data, len); 2090 return 0; 2091 } 2092 2093 /* 2094 * common case of get sockopt integer value 2095 */ 2096 int 2097 sockopt_getint(const struct sockopt *sopt, int *valp) 2098 { 2099 2100 return sockopt_get(sopt, valp, sizeof(int)); 2101 } 2102 2103 /* 2104 * set sockopt value from mbuf 2105 * - ONLY for legacy code 2106 * - mbuf is released by sockopt 2107 * - will not sleep 2108 */ 2109 int 2110 sockopt_setmbuf(struct sockopt *sopt, struct mbuf *m) 2111 { 2112 size_t len; 2113 int error; 2114 2115 len = m_length(m); 2116 2117 if (sopt->sopt_size == 0) { 2118 error = sockopt_alloc(sopt, len, KM_NOSLEEP); 2119 if (error) 2120 return error; 2121 } 2122 2123 KASSERT(sopt->sopt_size == len); 2124 m_copydata(m, 0, len, sopt->sopt_data); 2125 m_freem(m); 2126 2127 return 0; 2128 } 2129 2130 /* 2131 * get sockopt value into mbuf 2132 * - ONLY for legacy code 2133 * - mbuf to be released by the caller 2134 * - will not sleep 2135 */ 2136 struct mbuf * 2137 sockopt_getmbuf(const struct sockopt *sopt) 2138 { 2139 struct mbuf *m; 2140 2141 if (sopt->sopt_size > MCLBYTES) 2142 return NULL; 2143 2144 m = m_get(M_DONTWAIT, MT_SOOPTS); 2145 if (m == NULL) 2146 return NULL; 2147 2148 if (sopt->sopt_size > MLEN) { 2149 MCLGET(m, M_DONTWAIT); 2150 if ((m->m_flags & M_EXT) == 0) { 2151 m_free(m); 2152 return NULL; 2153 } 2154 } 2155 2156 memcpy(mtod(m, void *), sopt->sopt_data, sopt->sopt_size); 2157 m->m_len = sopt->sopt_size; 2158 2159 return m; 2160 } 2161 2162 void 2163 sohasoutofband(struct socket *so) 2164 { 2165 2166 fownsignal(so->so_pgid, SIGURG, POLL_PRI, POLLPRI|POLLRDBAND, so); 2167 selnotify(&so->so_rcv.sb_sel, POLLPRI | POLLRDBAND, NOTE_SUBMIT); 2168 } 2169 2170 static void 2171 filt_sordetach(struct knote *kn) 2172 { 2173 struct socket *so; 2174 2175 so = ((file_t *)kn->kn_obj)->f_data; 2176 solock(so); 2177 SLIST_REMOVE(&so->so_rcv.sb_sel.sel_klist, kn, knote, kn_selnext); 2178 if (SLIST_EMPTY(&so->so_rcv.sb_sel.sel_klist)) 2179 so->so_rcv.sb_flags &= ~SB_KNOTE; 2180 sounlock(so); 2181 } 2182 2183 /*ARGSUSED*/ 2184 static int 2185 filt_soread(struct knote *kn, long hint) 2186 { 2187 struct socket *so; 2188 int rv; 2189 2190 so = ((file_t *)kn->kn_obj)->f_data; 2191 if (hint != NOTE_SUBMIT) 2192 solock(so); 2193 kn->kn_data = so->so_rcv.sb_cc; 2194 if (so->so_state & SS_CANTRCVMORE) { 2195 kn->kn_flags |= EV_EOF; 2196 kn->kn_fflags = so->so_error; 2197 rv = 1; 2198 } else if (so->so_error) /* temporary udp error */ 2199 rv = 1; 2200 else if (kn->kn_sfflags & NOTE_LOWAT) 2201 rv = (kn->kn_data >= kn->kn_sdata); 2202 else 2203 rv = (kn->kn_data >= so->so_rcv.sb_lowat); 2204 if (hint != NOTE_SUBMIT) 2205 sounlock(so); 2206 return rv; 2207 } 2208 2209 static void 2210 filt_sowdetach(struct knote *kn) 2211 { 2212 struct socket *so; 2213 2214 so = ((file_t *)kn->kn_obj)->f_data; 2215 solock(so); 2216 SLIST_REMOVE(&so->so_snd.sb_sel.sel_klist, kn, knote, kn_selnext); 2217 if (SLIST_EMPTY(&so->so_snd.sb_sel.sel_klist)) 2218 so->so_snd.sb_flags &= ~SB_KNOTE; 2219 sounlock(so); 2220 } 2221 2222 /*ARGSUSED*/ 2223 static int 2224 filt_sowrite(struct knote *kn, long hint) 2225 { 2226 struct socket *so; 2227 int rv; 2228 2229 so = ((file_t *)kn->kn_obj)->f_data; 2230 if (hint != NOTE_SUBMIT) 2231 solock(so); 2232 kn->kn_data = sbspace(&so->so_snd); 2233 if (so->so_state & SS_CANTSENDMORE) { 2234 kn->kn_flags |= EV_EOF; 2235 kn->kn_fflags = so->so_error; 2236 rv = 1; 2237 } else if (so->so_error) /* temporary udp error */ 2238 rv = 1; 2239 else if (((so->so_state & SS_ISCONNECTED) == 0) && 2240 (so->so_proto->pr_flags & PR_CONNREQUIRED)) 2241 rv = 0; 2242 else if (kn->kn_sfflags & NOTE_LOWAT) 2243 rv = (kn->kn_data >= kn->kn_sdata); 2244 else 2245 rv = (kn->kn_data >= so->so_snd.sb_lowat); 2246 if (hint != NOTE_SUBMIT) 2247 sounlock(so); 2248 return rv; 2249 } 2250 2251 /*ARGSUSED*/ 2252 static int 2253 filt_solisten(struct knote *kn, long hint) 2254 { 2255 struct socket *so; 2256 int rv; 2257 2258 so = ((file_t *)kn->kn_obj)->f_data; 2259 2260 /* 2261 * Set kn_data to number of incoming connections, not 2262 * counting partial (incomplete) connections. 2263 */ 2264 if (hint != NOTE_SUBMIT) 2265 solock(so); 2266 kn->kn_data = so->so_qlen; 2267 rv = (kn->kn_data > 0); 2268 if (hint != NOTE_SUBMIT) 2269 sounlock(so); 2270 return rv; 2271 } 2272 2273 static const struct filterops solisten_filtops = 2274 { 1, NULL, filt_sordetach, filt_solisten }; 2275 static const struct filterops soread_filtops = 2276 { 1, NULL, filt_sordetach, filt_soread }; 2277 static const struct filterops sowrite_filtops = 2278 { 1, NULL, filt_sowdetach, filt_sowrite }; 2279 2280 int 2281 soo_kqfilter(struct file *fp, struct knote *kn) 2282 { 2283 struct socket *so; 2284 struct sockbuf *sb; 2285 2286 so = ((file_t *)kn->kn_obj)->f_data; 2287 solock(so); 2288 switch (kn->kn_filter) { 2289 case EVFILT_READ: 2290 if (so->so_options & SO_ACCEPTCONN) 2291 kn->kn_fop = &solisten_filtops; 2292 else 2293 kn->kn_fop = &soread_filtops; 2294 sb = &so->so_rcv; 2295 break; 2296 case EVFILT_WRITE: 2297 kn->kn_fop = &sowrite_filtops; 2298 sb = &so->so_snd; 2299 break; 2300 default: 2301 sounlock(so); 2302 return (EINVAL); 2303 } 2304 SLIST_INSERT_HEAD(&sb->sb_sel.sel_klist, kn, kn_selnext); 2305 sb->sb_flags |= SB_KNOTE; 2306 sounlock(so); 2307 return (0); 2308 } 2309 2310 static int 2311 sodopoll(struct socket *so, int events) 2312 { 2313 int revents; 2314 2315 revents = 0; 2316 2317 if (events & (POLLIN | POLLRDNORM)) 2318 if (soreadable(so)) 2319 revents |= events & (POLLIN | POLLRDNORM); 2320 2321 if (events & (POLLOUT | POLLWRNORM)) 2322 if (sowritable(so)) 2323 revents |= events & (POLLOUT | POLLWRNORM); 2324 2325 if (events & (POLLPRI | POLLRDBAND)) 2326 if (so->so_oobmark || (so->so_state & SS_RCVATMARK)) 2327 revents |= events & (POLLPRI | POLLRDBAND); 2328 2329 return revents; 2330 } 2331 2332 int 2333 sopoll(struct socket *so, int events) 2334 { 2335 int revents = 0; 2336 2337 #ifndef DIAGNOSTIC 2338 /* 2339 * Do a quick, unlocked check in expectation that the socket 2340 * will be ready for I/O. Don't do this check if DIAGNOSTIC, 2341 * as the solocked() assertions will fail. 2342 */ 2343 if ((revents = sodopoll(so, events)) != 0) 2344 return revents; 2345 #endif 2346 2347 solock(so); 2348 if ((revents = sodopoll(so, events)) == 0) { 2349 if (events & (POLLIN | POLLPRI | POLLRDNORM | POLLRDBAND)) { 2350 selrecord(curlwp, &so->so_rcv.sb_sel); 2351 so->so_rcv.sb_flags |= SB_NOTIFY; 2352 } 2353 2354 if (events & (POLLOUT | POLLWRNORM)) { 2355 selrecord(curlwp, &so->so_snd.sb_sel); 2356 so->so_snd.sb_flags |= SB_NOTIFY; 2357 } 2358 } 2359 sounlock(so); 2360 2361 return revents; 2362 } 2363 2364 2365 #include <sys/sysctl.h> 2366 2367 static int sysctl_kern_somaxkva(SYSCTLFN_PROTO); 2368 2369 /* 2370 * sysctl helper routine for kern.somaxkva. ensures that the given 2371 * value is not too small. 2372 * (XXX should we maybe make sure it's not too large as well?) 2373 */ 2374 static int 2375 sysctl_kern_somaxkva(SYSCTLFN_ARGS) 2376 { 2377 int error, new_somaxkva; 2378 struct sysctlnode node; 2379 2380 new_somaxkva = somaxkva; 2381 node = *rnode; 2382 node.sysctl_data = &new_somaxkva; 2383 error = sysctl_lookup(SYSCTLFN_CALL(&node)); 2384 if (error || newp == NULL) 2385 return (error); 2386 2387 if (new_somaxkva < (16 * 1024 * 1024)) /* sanity */ 2388 return (EINVAL); 2389 2390 mutex_enter(&so_pendfree_lock); 2391 somaxkva = new_somaxkva; 2392 cv_broadcast(&socurkva_cv); 2393 mutex_exit(&so_pendfree_lock); 2394 2395 return (error); 2396 } 2397 2398 static void 2399 sysctl_kern_somaxkva_setup(void) 2400 { 2401 2402 KASSERT(socket_sysctllog == NULL); 2403 sysctl_createv(&socket_sysctllog, 0, NULL, NULL, 2404 CTLFLAG_PERMANENT, 2405 CTLTYPE_NODE, "kern", NULL, 2406 NULL, 0, NULL, 0, 2407 CTL_KERN, CTL_EOL); 2408 2409 sysctl_createv(&socket_sysctllog, 0, NULL, NULL, 2410 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 2411 CTLTYPE_INT, "somaxkva", 2412 SYSCTL_DESCR("Maximum amount of kernel memory to be " 2413 "used for socket buffers"), 2414 sysctl_kern_somaxkva, 0, NULL, 0, 2415 CTL_KERN, KERN_SOMAXKVA, CTL_EOL); 2416 } 2417