uipc_usrreq.c revision 1.121
1 /* $NetBSD: uipc_usrreq.c,v 1.121 2009/03/11 06:05:29 mrg Exp $ */ 2 3 /*- 4 * Copyright (c) 1998, 2000, 2004, 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 the Numerical Aerospace Simulation Facility, 9 * NASA Ames Research Center, and by Andrew Doran. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 22 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 23 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 24 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 25 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 26 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 27 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 29 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 30 * POSSIBILITY OF SUCH DAMAGE. 31 */ 32 33 /* 34 * Copyright (c) 1982, 1986, 1989, 1991, 1993 35 * The Regents of the University of California. All rights reserved. 36 * 37 * Redistribution and use in source and binary forms, with or without 38 * modification, are permitted provided that the following conditions 39 * are met: 40 * 1. Redistributions of source code must retain the above copyright 41 * notice, this list of conditions and the following disclaimer. 42 * 2. Redistributions in binary form must reproduce the above copyright 43 * notice, this list of conditions and the following disclaimer in the 44 * documentation and/or other materials provided with the distribution. 45 * 3. Neither the name of the University nor the names of its contributors 46 * may be used to endorse or promote products derived from this software 47 * without specific prior written permission. 48 * 49 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 50 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 51 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 52 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 53 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 54 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 55 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 56 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 57 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 58 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 59 * SUCH DAMAGE. 60 * 61 * @(#)uipc_usrreq.c 8.9 (Berkeley) 5/14/95 62 */ 63 64 /* 65 * Copyright (c) 1997 Christopher G. Demetriou. All rights reserved. 66 * 67 * Redistribution and use in source and binary forms, with or without 68 * modification, are permitted provided that the following conditions 69 * are met: 70 * 1. Redistributions of source code must retain the above copyright 71 * notice, this list of conditions and the following disclaimer. 72 * 2. Redistributions in binary form must reproduce the above copyright 73 * notice, this list of conditions and the following disclaimer in the 74 * documentation and/or other materials provided with the distribution. 75 * 3. All advertising materials mentioning features or use of this software 76 * must display the following acknowledgement: 77 * This product includes software developed by the University of 78 * California, Berkeley and its contributors. 79 * 4. Neither the name of the University nor the names of its contributors 80 * may be used to endorse or promote products derived from this software 81 * without specific prior written permission. 82 * 83 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 84 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 85 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 86 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 87 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 88 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 89 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 90 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 91 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 92 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 93 * SUCH DAMAGE. 94 * 95 * @(#)uipc_usrreq.c 8.9 (Berkeley) 5/14/95 96 */ 97 98 #include <sys/cdefs.h> 99 __KERNEL_RCSID(0, "$NetBSD: uipc_usrreq.c,v 1.121 2009/03/11 06:05:29 mrg Exp $"); 100 101 #include <sys/param.h> 102 #include <sys/systm.h> 103 #include <sys/proc.h> 104 #include <sys/filedesc.h> 105 #include <sys/domain.h> 106 #include <sys/protosw.h> 107 #include <sys/socket.h> 108 #include <sys/socketvar.h> 109 #include <sys/unpcb.h> 110 #include <sys/un.h> 111 #include <sys/namei.h> 112 #include <sys/vnode.h> 113 #include <sys/file.h> 114 #include <sys/stat.h> 115 #include <sys/mbuf.h> 116 #include <sys/kauth.h> 117 #include <sys/kmem.h> 118 #include <sys/atomic.h> 119 #include <sys/uidinfo.h> 120 #include <sys/kernel.h> 121 #include <sys/kthread.h> 122 123 /* 124 * Unix communications domain. 125 * 126 * TODO: 127 * SEQPACKET, RDM 128 * rethink name space problems 129 * need a proper out-of-band 130 * 131 * Notes on locking: 132 * 133 * The generic rules noted in uipc_socket2.c apply. In addition: 134 * 135 * o We have a global lock, uipc_lock. 136 * 137 * o All datagram sockets are locked by uipc_lock. 138 * 139 * o For stream socketpairs, the two endpoints are created sharing the same 140 * independent lock. Sockets presented to PRU_CONNECT2 must already have 141 * matching locks. 142 * 143 * o Stream sockets created via socket() start life with their own 144 * independent lock. 145 * 146 * o Stream connections to a named endpoint are slightly more complicated. 147 * Sockets that have called listen() have their lock pointer mutated to 148 * the global uipc_lock. When establishing a connection, the connecting 149 * socket also has its lock mutated to uipc_lock, which matches the head 150 * (listening socket). We create a new socket for accept() to return, and 151 * that also shares the head's lock. Until the connection is completely 152 * done on both ends, all three sockets are locked by uipc_lock. Once the 153 * connection is complete, the association with the head's lock is broken. 154 * The connecting socket and the socket returned from accept() have their 155 * lock pointers mutated away from uipc_lock, and back to the connecting 156 * socket's original, independent lock. The head continues to be locked 157 * by uipc_lock. 158 * 159 * o If uipc_lock is determined to be a significant source of contention, 160 * it could easily be hashed out. It is difficult to simply make it an 161 * independent lock because of visibility / garbage collection issues: 162 * if a socket has been associated with a lock at any point, that lock 163 * must remain valid until the socket is no longer visible in the system. 164 * The lock must not be freed or otherwise destroyed until any sockets 165 * that had referenced it have also been destroyed. 166 */ 167 const struct sockaddr_un sun_noname = { 168 .sun_len = sizeof(sun_noname), 169 .sun_family = AF_LOCAL, 170 }; 171 ino_t unp_ino; /* prototype for fake inode numbers */ 172 173 struct mbuf *unp_addsockcred(struct lwp *, struct mbuf *); 174 static void unp_mark(file_t *); 175 static void unp_scan(struct mbuf *, void (*)(file_t *), int); 176 static void unp_discard_now(file_t *); 177 static void unp_discard_later(file_t *); 178 static void unp_thread(void *); 179 static void unp_thread_kick(void); 180 static kmutex_t *uipc_lock; 181 182 static kcondvar_t unp_thread_cv; 183 static lwp_t *unp_thread_lwp; 184 static SLIST_HEAD(,file) unp_thread_discard; 185 static int unp_defer; 186 187 /* 188 * Initialize Unix protocols. 189 */ 190 void 191 uipc_init(void) 192 { 193 int error; 194 195 uipc_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 196 cv_init(&unp_thread_cv, "unpgc"); 197 198 error = kthread_create(PRI_NONE, KTHREAD_MPSAFE, NULL, unp_thread, 199 NULL, &unp_thread_lwp, "unpgc"); 200 if (error != 0) 201 panic("uipc_init %d", error); 202 } 203 204 /* 205 * A connection succeeded: disassociate both endpoints from the head's 206 * lock, and make them share their own lock. There is a race here: for 207 * a very brief time one endpoint will be locked by a different lock 208 * than the other end. However, since the current thread holds the old 209 * lock (the listening socket's lock, the head) access can still only be 210 * made to one side of the connection. 211 */ 212 static void 213 unp_setpeerlocks(struct socket *so, struct socket *so2) 214 { 215 struct unpcb *unp; 216 kmutex_t *lock; 217 218 KASSERT(solocked2(so, so2)); 219 220 /* 221 * Bail out if either end of the socket is not yet fully 222 * connected or accepted. We only break the lock association 223 * with the head when the pair of sockets stand completely 224 * on their own. 225 */ 226 if (so->so_head != NULL || so2->so_head != NULL) 227 return; 228 229 /* 230 * Drop references to old lock. A third reference (from the 231 * queue head) must be held as we still hold its lock. Bonus: 232 * we don't need to worry about garbage collecting the lock. 233 */ 234 lock = so->so_lock; 235 KASSERT(lock == uipc_lock); 236 mutex_obj_free(lock); 237 mutex_obj_free(lock); 238 239 /* 240 * Grab stream lock from the initiator and share between the two 241 * endpoints. Issue memory barrier to ensure all modifications 242 * become globally visible before the lock change. so2 is 243 * assumed not to have a stream lock, because it was created 244 * purely for the server side to accept this connection and 245 * started out life using the domain-wide lock. 246 */ 247 unp = sotounpcb(so); 248 KASSERT(unp->unp_streamlock != NULL); 249 KASSERT(sotounpcb(so2)->unp_streamlock == NULL); 250 lock = unp->unp_streamlock; 251 unp->unp_streamlock = NULL; 252 mutex_obj_hold(lock); 253 membar_exit(); 254 solockreset(so, lock); 255 solockreset(so2, lock); 256 } 257 258 /* 259 * Reset a socket's lock back to the domain-wide lock. 260 */ 261 static void 262 unp_resetlock(struct socket *so) 263 { 264 kmutex_t *olock, *nlock; 265 struct unpcb *unp; 266 267 KASSERT(solocked(so)); 268 269 olock = so->so_lock; 270 nlock = uipc_lock; 271 if (olock == nlock) 272 return; 273 unp = sotounpcb(so); 274 KASSERT(unp->unp_streamlock == NULL); 275 unp->unp_streamlock = olock; 276 mutex_obj_hold(nlock); 277 mutex_enter(nlock); 278 solockreset(so, nlock); 279 mutex_exit(olock); 280 } 281 282 static void 283 unp_free(struct unpcb *unp) 284 { 285 286 if (unp->unp_addr) 287 free(unp->unp_addr, M_SONAME); 288 if (unp->unp_streamlock != NULL) 289 mutex_obj_free(unp->unp_streamlock); 290 free(unp, M_PCB); 291 } 292 293 int 294 unp_output(struct mbuf *m, struct mbuf *control, struct unpcb *unp, 295 struct lwp *l) 296 { 297 struct socket *so2; 298 const struct sockaddr_un *sun; 299 300 so2 = unp->unp_conn->unp_socket; 301 302 KASSERT(solocked(so2)); 303 304 if (unp->unp_addr) 305 sun = unp->unp_addr; 306 else 307 sun = &sun_noname; 308 if (unp->unp_conn->unp_flags & UNP_WANTCRED) 309 control = unp_addsockcred(l, control); 310 if (sbappendaddr(&so2->so_rcv, (const struct sockaddr *)sun, m, 311 control) == 0) { 312 so2->so_rcv.sb_overflowed++; 313 unp_dispose(control); 314 m_freem(control); 315 m_freem(m); 316 return (ENOBUFS); 317 } else { 318 sorwakeup(so2); 319 return (0); 320 } 321 } 322 323 void 324 unp_setaddr(struct socket *so, struct mbuf *nam, bool peeraddr) 325 { 326 const struct sockaddr_un *sun; 327 struct unpcb *unp; 328 bool ext; 329 330 unp = sotounpcb(so); 331 ext = false; 332 333 for (;;) { 334 sun = NULL; 335 if (peeraddr) { 336 if (unp->unp_conn && unp->unp_conn->unp_addr) 337 sun = unp->unp_conn->unp_addr; 338 } else { 339 if (unp->unp_addr) 340 sun = unp->unp_addr; 341 } 342 if (sun == NULL) 343 sun = &sun_noname; 344 nam->m_len = sun->sun_len; 345 if (nam->m_len > MLEN && !ext) { 346 sounlock(so); 347 MEXTMALLOC(nam, MAXPATHLEN * 2, M_WAITOK); 348 solock(so); 349 ext = true; 350 } else { 351 KASSERT(nam->m_len <= MAXPATHLEN * 2); 352 memcpy(mtod(nam, void *), sun, (size_t)nam->m_len); 353 break; 354 } 355 } 356 } 357 358 /*ARGSUSED*/ 359 int 360 uipc_usrreq(struct socket *so, int req, struct mbuf *m, struct mbuf *nam, 361 struct mbuf *control, struct lwp *l) 362 { 363 struct unpcb *unp = sotounpcb(so); 364 struct socket *so2; 365 struct proc *p; 366 u_int newhiwat; 367 int error = 0; 368 369 if (req == PRU_CONTROL) 370 return (EOPNOTSUPP); 371 372 #ifdef DIAGNOSTIC 373 if (req != PRU_SEND && req != PRU_SENDOOB && control) 374 panic("uipc_usrreq: unexpected control mbuf"); 375 #endif 376 p = l ? l->l_proc : NULL; 377 if (req != PRU_ATTACH) { 378 if (unp == 0) { 379 error = EINVAL; 380 goto release; 381 } 382 KASSERT(solocked(so)); 383 } 384 385 switch (req) { 386 387 case PRU_ATTACH: 388 if (unp != 0) { 389 error = EISCONN; 390 break; 391 } 392 error = unp_attach(so); 393 break; 394 395 case PRU_DETACH: 396 unp_detach(unp); 397 break; 398 399 case PRU_BIND: 400 KASSERT(l != NULL); 401 error = unp_bind(so, nam, l); 402 break; 403 404 case PRU_LISTEN: 405 /* 406 * If the socket can accept a connection, it must be 407 * locked by uipc_lock. 408 */ 409 unp_resetlock(so); 410 if (unp->unp_vnode == 0) 411 error = EINVAL; 412 break; 413 414 case PRU_CONNECT: 415 KASSERT(l != NULL); 416 error = unp_connect(so, nam, l); 417 break; 418 419 case PRU_CONNECT2: 420 error = unp_connect2(so, (struct socket *)nam, PRU_CONNECT2); 421 break; 422 423 case PRU_DISCONNECT: 424 unp_disconnect(unp); 425 break; 426 427 case PRU_ACCEPT: 428 KASSERT(so->so_lock == uipc_lock); 429 /* 430 * Mark the initiating STREAM socket as connected *ONLY* 431 * after it's been accepted. This prevents a client from 432 * overrunning a server and receiving ECONNREFUSED. 433 */ 434 if (unp->unp_conn == NULL) 435 break; 436 so2 = unp->unp_conn->unp_socket; 437 if (so2->so_state & SS_ISCONNECTING) { 438 KASSERT(solocked2(so, so->so_head)); 439 KASSERT(solocked2(so2, so->so_head)); 440 soisconnected(so2); 441 } 442 /* 443 * If the connection is fully established, break the 444 * association with uipc_lock and give the connected 445 * pair a seperate lock to share. 446 */ 447 unp_setpeerlocks(so2, so); 448 /* 449 * Only now return peer's address, as we may need to 450 * block in order to allocate memory. 451 * 452 * XXX Minor race: connection can be broken while 453 * lock is dropped in unp_setaddr(). We will return 454 * error == 0 and sun_noname as the peer address. 455 */ 456 unp_setaddr(so, nam, true); 457 break; 458 459 case PRU_SHUTDOWN: 460 socantsendmore(so); 461 unp_shutdown(unp); 462 break; 463 464 case PRU_RCVD: 465 switch (so->so_type) { 466 467 case SOCK_DGRAM: 468 panic("uipc 1"); 469 /*NOTREACHED*/ 470 471 case SOCK_STREAM: 472 #define rcv (&so->so_rcv) 473 #define snd (&so2->so_snd) 474 if (unp->unp_conn == 0) 475 break; 476 so2 = unp->unp_conn->unp_socket; 477 KASSERT(solocked2(so, so2)); 478 /* 479 * Adjust backpressure on sender 480 * and wakeup any waiting to write. 481 */ 482 snd->sb_mbmax += unp->unp_mbcnt - rcv->sb_mbcnt; 483 unp->unp_mbcnt = rcv->sb_mbcnt; 484 newhiwat = snd->sb_hiwat + unp->unp_cc - rcv->sb_cc; 485 (void)chgsbsize(so2->so_uidinfo, 486 &snd->sb_hiwat, newhiwat, RLIM_INFINITY); 487 unp->unp_cc = rcv->sb_cc; 488 sowwakeup(so2); 489 #undef snd 490 #undef rcv 491 break; 492 493 default: 494 panic("uipc 2"); 495 } 496 break; 497 498 case PRU_SEND: 499 /* 500 * Note: unp_internalize() rejects any control message 501 * other than SCM_RIGHTS, and only allows one. This 502 * has the side-effect of preventing a caller from 503 * forging SCM_CREDS. 504 */ 505 if (control) { 506 sounlock(so); 507 error = unp_internalize(&control); 508 solock(so); 509 if (error != 0) { 510 m_freem(control); 511 m_freem(m); 512 break; 513 } 514 } 515 switch (so->so_type) { 516 517 case SOCK_DGRAM: { 518 KASSERT(so->so_lock == uipc_lock); 519 if (nam) { 520 if ((so->so_state & SS_ISCONNECTED) != 0) 521 error = EISCONN; 522 else { 523 /* 524 * Note: once connected, the 525 * socket's lock must not be 526 * dropped until we have sent 527 * the message and disconnected. 528 * This is necessary to prevent 529 * intervening control ops, like 530 * another connection. 531 */ 532 error = unp_connect(so, nam, l); 533 } 534 } else { 535 if ((so->so_state & SS_ISCONNECTED) == 0) 536 error = ENOTCONN; 537 } 538 if (error) { 539 unp_dispose(control); 540 m_freem(control); 541 m_freem(m); 542 break; 543 } 544 KASSERT(p != NULL); 545 error = unp_output(m, control, unp, l); 546 if (nam) 547 unp_disconnect(unp); 548 break; 549 } 550 551 case SOCK_STREAM: 552 #define rcv (&so2->so_rcv) 553 #define snd (&so->so_snd) 554 if (unp->unp_conn == NULL) { 555 error = ENOTCONN; 556 break; 557 } 558 so2 = unp->unp_conn->unp_socket; 559 KASSERT(solocked2(so, so2)); 560 if (unp->unp_conn->unp_flags & UNP_WANTCRED) { 561 /* 562 * Credentials are passed only once on 563 * SOCK_STREAM. 564 */ 565 unp->unp_conn->unp_flags &= ~UNP_WANTCRED; 566 control = unp_addsockcred(l, control); 567 } 568 /* 569 * Send to paired receive port, and then reduce 570 * send buffer hiwater marks to maintain backpressure. 571 * Wake up readers. 572 */ 573 if (control) { 574 if (sbappendcontrol(rcv, m, control) != 0) 575 control = NULL; 576 } else 577 sbappend(rcv, m); 578 snd->sb_mbmax -= 579 rcv->sb_mbcnt - unp->unp_conn->unp_mbcnt; 580 unp->unp_conn->unp_mbcnt = rcv->sb_mbcnt; 581 newhiwat = snd->sb_hiwat - 582 (rcv->sb_cc - unp->unp_conn->unp_cc); 583 (void)chgsbsize(so->so_uidinfo, 584 &snd->sb_hiwat, newhiwat, RLIM_INFINITY); 585 unp->unp_conn->unp_cc = rcv->sb_cc; 586 sorwakeup(so2); 587 #undef snd 588 #undef rcv 589 if (control != NULL) { 590 unp_dispose(control); 591 m_freem(control); 592 } 593 break; 594 595 default: 596 panic("uipc 4"); 597 } 598 break; 599 600 case PRU_ABORT: 601 (void)unp_drop(unp, ECONNABORTED); 602 603 KASSERT(so->so_head == NULL); 604 #ifdef DIAGNOSTIC 605 if (so->so_pcb == 0) 606 panic("uipc 5: drop killed pcb"); 607 #endif 608 unp_detach(unp); 609 break; 610 611 case PRU_SENSE: 612 ((struct stat *) m)->st_blksize = so->so_snd.sb_hiwat; 613 if (so->so_type == SOCK_STREAM && unp->unp_conn != 0) { 614 so2 = unp->unp_conn->unp_socket; 615 KASSERT(solocked2(so, so2)); 616 ((struct stat *) m)->st_blksize += so2->so_rcv.sb_cc; 617 } 618 ((struct stat *) m)->st_dev = NODEV; 619 if (unp->unp_ino == 0) 620 unp->unp_ino = unp_ino++; 621 ((struct stat *) m)->st_atimespec = 622 ((struct stat *) m)->st_mtimespec = 623 ((struct stat *) m)->st_ctimespec = unp->unp_ctime; 624 ((struct stat *) m)->st_ino = unp->unp_ino; 625 return (0); 626 627 case PRU_RCVOOB: 628 error = EOPNOTSUPP; 629 break; 630 631 case PRU_SENDOOB: 632 m_freem(control); 633 m_freem(m); 634 error = EOPNOTSUPP; 635 break; 636 637 case PRU_SOCKADDR: 638 unp_setaddr(so, nam, false); 639 break; 640 641 case PRU_PEERADDR: 642 unp_setaddr(so, nam, true); 643 break; 644 645 default: 646 panic("piusrreq"); 647 } 648 649 release: 650 return (error); 651 } 652 653 /* 654 * Unix domain socket option processing. 655 */ 656 int 657 uipc_ctloutput(int op, struct socket *so, struct sockopt *sopt) 658 { 659 struct unpcb *unp = sotounpcb(so); 660 int optval = 0, error = 0; 661 662 KASSERT(solocked(so)); 663 664 if (sopt->sopt_level != 0) { 665 error = ENOPROTOOPT; 666 } else switch (op) { 667 668 case PRCO_SETOPT: 669 switch (sopt->sopt_name) { 670 case LOCAL_CREDS: 671 case LOCAL_CONNWAIT: 672 error = sockopt_getint(sopt, &optval); 673 if (error) 674 break; 675 switch (sopt->sopt_name) { 676 #define OPTSET(bit) \ 677 if (optval) \ 678 unp->unp_flags |= (bit); \ 679 else \ 680 unp->unp_flags &= ~(bit); 681 682 case LOCAL_CREDS: 683 OPTSET(UNP_WANTCRED); 684 break; 685 case LOCAL_CONNWAIT: 686 OPTSET(UNP_CONNWAIT); 687 break; 688 } 689 break; 690 #undef OPTSET 691 692 default: 693 error = ENOPROTOOPT; 694 break; 695 } 696 break; 697 698 case PRCO_GETOPT: 699 sounlock(so); 700 switch (sopt->sopt_name) { 701 case LOCAL_PEEREID: 702 if (unp->unp_flags & UNP_EIDSVALID) { 703 error = sockopt_set(sopt, 704 &unp->unp_connid, sizeof(unp->unp_connid)); 705 } else { 706 error = EINVAL; 707 } 708 break; 709 case LOCAL_CREDS: 710 #define OPTBIT(bit) (unp->unp_flags & (bit) ? 1 : 0) 711 712 optval = OPTBIT(UNP_WANTCRED); 713 error = sockopt_setint(sopt, optval); 714 break; 715 #undef OPTBIT 716 717 default: 718 error = ENOPROTOOPT; 719 break; 720 } 721 solock(so); 722 break; 723 } 724 return (error); 725 } 726 727 /* 728 * Both send and receive buffers are allocated PIPSIZ bytes of buffering 729 * for stream sockets, although the total for sender and receiver is 730 * actually only PIPSIZ. 731 * Datagram sockets really use the sendspace as the maximum datagram size, 732 * and don't really want to reserve the sendspace. Their recvspace should 733 * be large enough for at least one max-size datagram plus address. 734 */ 735 #define PIPSIZ 4096 736 u_long unpst_sendspace = PIPSIZ; 737 u_long unpst_recvspace = PIPSIZ; 738 u_long unpdg_sendspace = 2*1024; /* really max datagram size */ 739 u_long unpdg_recvspace = 4*1024; 740 741 u_int unp_rights; /* files in flight */ 742 u_int unp_rights_ratio = 2; /* limit, fraction of maxfiles */ 743 744 int 745 unp_attach(struct socket *so) 746 { 747 struct unpcb *unp; 748 int error; 749 750 switch (so->so_type) { 751 case SOCK_STREAM: 752 if (so->so_lock == NULL) { 753 /* 754 * XXX Assuming that no socket locks are held, 755 * as this call may sleep. 756 */ 757 so->so_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 758 solock(so); 759 } 760 if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) { 761 error = soreserve(so, unpst_sendspace, unpst_recvspace); 762 if (error != 0) 763 return (error); 764 } 765 break; 766 767 case SOCK_DGRAM: 768 if (so->so_lock == NULL) { 769 mutex_obj_hold(uipc_lock); 770 so->so_lock = uipc_lock; 771 solock(so); 772 } 773 if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) { 774 error = soreserve(so, unpdg_sendspace, unpdg_recvspace); 775 if (error != 0) 776 return (error); 777 } 778 break; 779 780 default: 781 panic("unp_attach"); 782 } 783 KASSERT(solocked(so)); 784 unp = malloc(sizeof(*unp), M_PCB, M_NOWAIT); 785 if (unp == NULL) 786 return (ENOBUFS); 787 memset((void *)unp, 0, sizeof(*unp)); 788 unp->unp_socket = so; 789 so->so_pcb = unp; 790 nanotime(&unp->unp_ctime); 791 return (0); 792 } 793 794 void 795 unp_detach(struct unpcb *unp) 796 { 797 struct socket *so; 798 vnode_t *vp; 799 800 so = unp->unp_socket; 801 802 retry: 803 if ((vp = unp->unp_vnode) != NULL) { 804 sounlock(so); 805 /* Acquire v_interlock to protect against unp_connect(). */ 806 /* XXXAD racy */ 807 mutex_enter(&vp->v_interlock); 808 vp->v_socket = NULL; 809 vrelel(vp, 0); 810 solock(so); 811 unp->unp_vnode = NULL; 812 } 813 if (unp->unp_conn) 814 unp_disconnect(unp); 815 while (unp->unp_refs) { 816 KASSERT(solocked2(so, unp->unp_refs->unp_socket)); 817 if (unp_drop(unp->unp_refs, ECONNRESET)) { 818 solock(so); 819 goto retry; 820 } 821 } 822 soisdisconnected(so); 823 so->so_pcb = NULL; 824 if (unp_rights) { 825 /* 826 * Normally the receive buffer is flushed later, in sofree, 827 * but if our receive buffer holds references to files that 828 * are now garbage, we will enqueue those file references to 829 * the garbage collector and kick it into action. 830 */ 831 sorflush(so); 832 unp_free(unp); 833 unp_thread_kick(); 834 } else 835 unp_free(unp); 836 } 837 838 int 839 unp_bind(struct socket *so, struct mbuf *nam, struct lwp *l) 840 { 841 struct sockaddr_un *sun; 842 struct unpcb *unp; 843 vnode_t *vp; 844 struct vattr vattr; 845 size_t addrlen; 846 int error; 847 struct nameidata nd; 848 proc_t *p; 849 850 unp = sotounpcb(so); 851 if (unp->unp_vnode != NULL) 852 return (EINVAL); 853 if ((unp->unp_flags & UNP_BUSY) != 0) { 854 /* 855 * EALREADY may not be strictly accurate, but since this 856 * is a major application error it's hardly a big deal. 857 */ 858 return (EALREADY); 859 } 860 unp->unp_flags |= UNP_BUSY; 861 sounlock(so); 862 863 /* 864 * Allocate the new sockaddr. We have to allocate one 865 * extra byte so that we can ensure that the pathname 866 * is nul-terminated. 867 */ 868 p = l->l_proc; 869 addrlen = nam->m_len + 1; 870 sun = malloc(addrlen, M_SONAME, M_WAITOK); 871 m_copydata(nam, 0, nam->m_len, (void *)sun); 872 *(((char *)sun) + nam->m_len) = '\0'; 873 874 NDINIT(&nd, CREATE, FOLLOW | LOCKPARENT | TRYEMULROOT, UIO_SYSSPACE, 875 sun->sun_path); 876 877 /* SHOULD BE ABLE TO ADOPT EXISTING AND wakeup() ALA FIFO's */ 878 if ((error = namei(&nd)) != 0) 879 goto bad; 880 vp = nd.ni_vp; 881 if (vp != NULL) { 882 VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd); 883 if (nd.ni_dvp == vp) 884 vrele(nd.ni_dvp); 885 else 886 vput(nd.ni_dvp); 887 vrele(vp); 888 error = EADDRINUSE; 889 goto bad; 890 } 891 VATTR_NULL(&vattr); 892 vattr.va_type = VSOCK; 893 vattr.va_mode = ACCESSPERMS & ~(p->p_cwdi->cwdi_cmask); 894 error = VOP_CREATE(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr); 895 if (error) 896 goto bad; 897 vp = nd.ni_vp; 898 solock(so); 899 vp->v_socket = unp->unp_socket; 900 unp->unp_vnode = vp; 901 unp->unp_addrlen = addrlen; 902 unp->unp_addr = sun; 903 unp->unp_connid.unp_pid = p->p_pid; 904 unp->unp_connid.unp_euid = kauth_cred_geteuid(l->l_cred); 905 unp->unp_connid.unp_egid = kauth_cred_getegid(l->l_cred); 906 unp->unp_flags |= UNP_EIDSBIND; 907 VOP_UNLOCK(vp, 0); 908 unp->unp_flags &= ~UNP_BUSY; 909 return (0); 910 911 bad: 912 free(sun, M_SONAME); 913 solock(so); 914 unp->unp_flags &= ~UNP_BUSY; 915 return (error); 916 } 917 918 int 919 unp_connect(struct socket *so, struct mbuf *nam, struct lwp *l) 920 { 921 struct sockaddr_un *sun; 922 vnode_t *vp; 923 struct socket *so2, *so3; 924 struct unpcb *unp, *unp2, *unp3; 925 size_t addrlen; 926 int error; 927 struct nameidata nd; 928 929 unp = sotounpcb(so); 930 if ((unp->unp_flags & UNP_BUSY) != 0) { 931 /* 932 * EALREADY may not be strictly accurate, but since this 933 * is a major application error it's hardly a big deal. 934 */ 935 return (EALREADY); 936 } 937 unp->unp_flags |= UNP_BUSY; 938 sounlock(so); 939 940 /* 941 * Allocate a temporary sockaddr. We have to allocate one extra 942 * byte so that we can ensure that the pathname is nul-terminated. 943 * When we establish the connection, we copy the other PCB's 944 * sockaddr to our own. 945 */ 946 addrlen = nam->m_len + 1; 947 sun = malloc(addrlen, M_SONAME, M_WAITOK); 948 m_copydata(nam, 0, nam->m_len, (void *)sun); 949 *(((char *)sun) + nam->m_len) = '\0'; 950 951 NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | TRYEMULROOT, UIO_SYSSPACE, 952 sun->sun_path); 953 954 if ((error = namei(&nd)) != 0) 955 goto bad2; 956 vp = nd.ni_vp; 957 if (vp->v_type != VSOCK) { 958 error = ENOTSOCK; 959 goto bad; 960 } 961 if ((error = VOP_ACCESS(vp, VWRITE, l->l_cred)) != 0) 962 goto bad; 963 /* Acquire v_interlock to protect against unp_detach(). */ 964 mutex_enter(&vp->v_interlock); 965 so2 = vp->v_socket; 966 if (so2 == NULL) { 967 mutex_exit(&vp->v_interlock); 968 error = ECONNREFUSED; 969 goto bad; 970 } 971 if (so->so_type != so2->so_type) { 972 mutex_exit(&vp->v_interlock); 973 error = EPROTOTYPE; 974 goto bad; 975 } 976 solock(so); 977 unp_resetlock(so); 978 mutex_exit(&vp->v_interlock); 979 if ((so->so_proto->pr_flags & PR_CONNREQUIRED) != 0) { 980 /* 981 * This may seem somewhat fragile but is OK: if we can 982 * see SO_ACCEPTCONN set on the endpoint, then it must 983 * be locked by the domain-wide uipc_lock. 984 */ 985 KASSERT((so->so_options & SO_ACCEPTCONN) == 0 || 986 so2->so_lock == uipc_lock); 987 if ((so2->so_options & SO_ACCEPTCONN) == 0 || 988 (so3 = sonewconn(so2, 0)) == 0) { 989 error = ECONNREFUSED; 990 sounlock(so); 991 goto bad; 992 } 993 unp2 = sotounpcb(so2); 994 unp3 = sotounpcb(so3); 995 if (unp2->unp_addr) { 996 unp3->unp_addr = malloc(unp2->unp_addrlen, 997 M_SONAME, M_WAITOK); 998 memcpy(unp3->unp_addr, unp2->unp_addr, 999 unp2->unp_addrlen); 1000 unp3->unp_addrlen = unp2->unp_addrlen; 1001 } 1002 unp3->unp_flags = unp2->unp_flags; 1003 unp3->unp_connid.unp_pid = l->l_proc->p_pid; 1004 unp3->unp_connid.unp_euid = kauth_cred_geteuid(l->l_cred); 1005 unp3->unp_connid.unp_egid = kauth_cred_getegid(l->l_cred); 1006 unp3->unp_flags |= UNP_EIDSVALID; 1007 if (unp2->unp_flags & UNP_EIDSBIND) { 1008 unp->unp_connid = unp2->unp_connid; 1009 unp->unp_flags |= UNP_EIDSVALID; 1010 } 1011 so2 = so3; 1012 } 1013 error = unp_connect2(so, so2, PRU_CONNECT); 1014 sounlock(so); 1015 bad: 1016 vput(vp); 1017 bad2: 1018 free(sun, M_SONAME); 1019 solock(so); 1020 unp->unp_flags &= ~UNP_BUSY; 1021 return (error); 1022 } 1023 1024 int 1025 unp_connect2(struct socket *so, struct socket *so2, int req) 1026 { 1027 struct unpcb *unp = sotounpcb(so); 1028 struct unpcb *unp2; 1029 1030 if (so2->so_type != so->so_type) 1031 return (EPROTOTYPE); 1032 1033 /* 1034 * All three sockets involved must be locked by same lock: 1035 * 1036 * local endpoint (so) 1037 * remote endpoint (so2) 1038 * queue head (so->so_head, only if PR_CONNREQUIRED) 1039 */ 1040 KASSERT(solocked2(so, so2)); 1041 if (so->so_head != NULL) { 1042 KASSERT(so->so_lock == uipc_lock); 1043 KASSERT(solocked2(so, so->so_head)); 1044 } 1045 1046 unp2 = sotounpcb(so2); 1047 unp->unp_conn = unp2; 1048 switch (so->so_type) { 1049 1050 case SOCK_DGRAM: 1051 unp->unp_nextref = unp2->unp_refs; 1052 unp2->unp_refs = unp; 1053 soisconnected(so); 1054 break; 1055 1056 case SOCK_STREAM: 1057 unp2->unp_conn = unp; 1058 if (req == PRU_CONNECT && 1059 ((unp->unp_flags | unp2->unp_flags) & UNP_CONNWAIT)) 1060 soisconnecting(so); 1061 else 1062 soisconnected(so); 1063 soisconnected(so2); 1064 /* 1065 * If the connection is fully established, break the 1066 * association with uipc_lock and give the connected 1067 * pair a seperate lock to share. For CONNECT2, we 1068 * require that the locks already match (the sockets 1069 * are created that way). 1070 */ 1071 if (req == PRU_CONNECT) 1072 unp_setpeerlocks(so, so2); 1073 break; 1074 1075 default: 1076 panic("unp_connect2"); 1077 } 1078 return (0); 1079 } 1080 1081 void 1082 unp_disconnect(struct unpcb *unp) 1083 { 1084 struct unpcb *unp2 = unp->unp_conn; 1085 struct socket *so; 1086 1087 if (unp2 == 0) 1088 return; 1089 unp->unp_conn = 0; 1090 so = unp->unp_socket; 1091 switch (so->so_type) { 1092 case SOCK_DGRAM: 1093 if (unp2->unp_refs == unp) 1094 unp2->unp_refs = unp->unp_nextref; 1095 else { 1096 unp2 = unp2->unp_refs; 1097 for (;;) { 1098 KASSERT(solocked2(so, unp2->unp_socket)); 1099 if (unp2 == 0) 1100 panic("unp_disconnect"); 1101 if (unp2->unp_nextref == unp) 1102 break; 1103 unp2 = unp2->unp_nextref; 1104 } 1105 unp2->unp_nextref = unp->unp_nextref; 1106 } 1107 unp->unp_nextref = 0; 1108 so->so_state &= ~SS_ISCONNECTED; 1109 break; 1110 1111 case SOCK_STREAM: 1112 KASSERT(solocked2(so, unp2->unp_socket)); 1113 soisdisconnected(so); 1114 unp2->unp_conn = 0; 1115 soisdisconnected(unp2->unp_socket); 1116 break; 1117 } 1118 } 1119 1120 #ifdef notdef 1121 unp_abort(struct unpcb *unp) 1122 { 1123 unp_detach(unp); 1124 } 1125 #endif 1126 1127 void 1128 unp_shutdown(struct unpcb *unp) 1129 { 1130 struct socket *so; 1131 1132 if (unp->unp_socket->so_type == SOCK_STREAM && unp->unp_conn && 1133 (so = unp->unp_conn->unp_socket)) 1134 socantrcvmore(so); 1135 } 1136 1137 bool 1138 unp_drop(struct unpcb *unp, int errno) 1139 { 1140 struct socket *so = unp->unp_socket; 1141 1142 KASSERT(solocked(so)); 1143 1144 so->so_error = errno; 1145 unp_disconnect(unp); 1146 if (so->so_head) { 1147 so->so_pcb = NULL; 1148 /* sofree() drops the socket lock */ 1149 sofree(so); 1150 unp_free(unp); 1151 return true; 1152 } 1153 return false; 1154 } 1155 1156 #ifdef notdef 1157 unp_drain(void) 1158 { 1159 1160 } 1161 #endif 1162 1163 int 1164 unp_externalize(struct mbuf *rights, struct lwp *l) 1165 { 1166 struct cmsghdr *cm = mtod(rights, struct cmsghdr *); 1167 struct proc *p = l->l_proc; 1168 int i, *fdp; 1169 file_t **rp; 1170 file_t *fp; 1171 int nfds, error = 0; 1172 1173 nfds = (cm->cmsg_len - CMSG_ALIGN(sizeof(*cm))) / 1174 sizeof(file_t *); 1175 rp = (file_t **)CMSG_DATA(cm); 1176 1177 fdp = malloc(nfds * sizeof(int), M_TEMP, M_WAITOK); 1178 rw_enter(&p->p_cwdi->cwdi_lock, RW_READER); 1179 1180 /* Make sure the recipient should be able to see the files.. */ 1181 if (p->p_cwdi->cwdi_rdir != NULL) { 1182 rp = (file_t **)CMSG_DATA(cm); 1183 for (i = 0; i < nfds; i++) { 1184 fp = *rp++; 1185 /* 1186 * If we are in a chroot'ed directory, and 1187 * someone wants to pass us a directory, make 1188 * sure it's inside the subtree we're allowed 1189 * to access. 1190 */ 1191 if (fp->f_type == DTYPE_VNODE) { 1192 vnode_t *vp = (vnode_t *)fp->f_data; 1193 if ((vp->v_type == VDIR) && 1194 !vn_isunder(vp, p->p_cwdi->cwdi_rdir, l)) { 1195 error = EPERM; 1196 break; 1197 } 1198 } 1199 } 1200 } 1201 1202 restart: 1203 rp = (file_t **)CMSG_DATA(cm); 1204 if (error != 0) { 1205 for (i = 0; i < nfds; i++) { 1206 fp = *rp; 1207 *rp++ = 0; 1208 unp_discard_now(fp); 1209 } 1210 goto out; 1211 } 1212 1213 /* 1214 * First loop -- allocate file descriptor table slots for the 1215 * new files. 1216 */ 1217 for (i = 0; i < nfds; i++) { 1218 fp = *rp++; 1219 if ((error = fd_alloc(p, 0, &fdp[i])) != 0) { 1220 /* 1221 * Back out what we've done so far. 1222 */ 1223 for (--i; i >= 0; i--) { 1224 fd_abort(p, NULL, fdp[i]); 1225 } 1226 if (error == ENOSPC) { 1227 fd_tryexpand(p); 1228 error = 0; 1229 } else { 1230 /* 1231 * This is the error that has historically 1232 * been returned, and some callers may 1233 * expect it. 1234 */ 1235 error = EMSGSIZE; 1236 } 1237 goto restart; 1238 } 1239 } 1240 1241 /* 1242 * Now that adding them has succeeded, update all of the 1243 * file passing state and affix the descriptors. 1244 */ 1245 rp = (file_t **)CMSG_DATA(cm); 1246 for (i = 0; i < nfds; i++) { 1247 fp = *rp++; 1248 atomic_dec_uint(&unp_rights); 1249 fd_affix(p, fp, fdp[i]); 1250 mutex_enter(&fp->f_lock); 1251 fp->f_msgcount--; 1252 mutex_exit(&fp->f_lock); 1253 /* 1254 * Note that fd_affix() adds a reference to the file. 1255 * The file may already have been closed by another 1256 * LWP in the process, so we must drop the reference 1257 * added by unp_internalize() with closef(). 1258 */ 1259 closef(fp); 1260 } 1261 1262 /* 1263 * Copy temporary array to message and adjust length, in case of 1264 * transition from large file_t pointers to ints. 1265 */ 1266 memcpy(CMSG_DATA(cm), fdp, nfds * sizeof(int)); 1267 cm->cmsg_len = CMSG_LEN(nfds * sizeof(int)); 1268 rights->m_len = CMSG_SPACE(nfds * sizeof(int)); 1269 out: 1270 rw_exit(&p->p_cwdi->cwdi_lock); 1271 free(fdp, M_TEMP); 1272 return (error); 1273 } 1274 1275 int 1276 unp_internalize(struct mbuf **controlp) 1277 { 1278 filedesc_t *fdescp = curlwp->l_fd; 1279 struct mbuf *control = *controlp; 1280 struct cmsghdr *newcm, *cm = mtod(control, struct cmsghdr *); 1281 file_t **rp, **files; 1282 file_t *fp; 1283 int i, fd, *fdp; 1284 int nfds, error; 1285 u_int maxmsg; 1286 1287 error = 0; 1288 newcm = NULL; 1289 1290 /* Sanity check the control message header. */ 1291 if (cm->cmsg_type != SCM_RIGHTS || cm->cmsg_level != SOL_SOCKET || 1292 cm->cmsg_len > control->m_len || 1293 cm->cmsg_len < CMSG_ALIGN(sizeof(*cm))) 1294 return (EINVAL); 1295 1296 /* 1297 * Verify that the file descriptors are valid, and acquire 1298 * a reference to each. 1299 */ 1300 nfds = (cm->cmsg_len - CMSG_ALIGN(sizeof(*cm))) / sizeof(int); 1301 fdp = (int *)CMSG_DATA(cm); 1302 maxmsg = maxfiles / unp_rights_ratio; 1303 for (i = 0; i < nfds; i++) { 1304 fd = *fdp++; 1305 if (atomic_inc_uint_nv(&unp_rights) > maxmsg) { 1306 atomic_dec_uint(&unp_rights); 1307 nfds = i; 1308 error = EAGAIN; 1309 goto out; 1310 } 1311 if ((fp = fd_getfile(fd)) == NULL) { 1312 atomic_dec_uint(&unp_rights); 1313 nfds = i; 1314 error = EBADF; 1315 goto out; 1316 } 1317 } 1318 1319 /* Allocate new space and copy header into it. */ 1320 newcm = malloc(CMSG_SPACE(nfds * sizeof(file_t *)), M_MBUF, M_WAITOK); 1321 if (newcm == NULL) { 1322 error = E2BIG; 1323 goto out; 1324 } 1325 memcpy(newcm, cm, sizeof(struct cmsghdr)); 1326 files = (file_t **)CMSG_DATA(newcm); 1327 1328 /* 1329 * Transform the file descriptors into file_t pointers, in 1330 * reverse order so that if pointers are bigger than ints, the 1331 * int won't get until we're done. No need to lock, as we have 1332 * already validated the descriptors with fd_getfile(). 1333 */ 1334 fdp = (int *)CMSG_DATA(cm) + nfds; 1335 rp = files + nfds; 1336 for (i = 0; i < nfds; i++) { 1337 fp = fdescp->fd_ofiles[*--fdp]->ff_file; 1338 KASSERT(fp != NULL); 1339 mutex_enter(&fp->f_lock); 1340 *--rp = fp; 1341 fp->f_count++; 1342 fp->f_msgcount++; 1343 mutex_exit(&fp->f_lock); 1344 } 1345 1346 out: 1347 /* Release descriptor references. */ 1348 fdp = (int *)CMSG_DATA(cm); 1349 for (i = 0; i < nfds; i++) { 1350 fd_putfile(*fdp++); 1351 if (error != 0) { 1352 atomic_dec_uint(&unp_rights); 1353 } 1354 } 1355 1356 if (error == 0) { 1357 if (control->m_flags & M_EXT) { 1358 m_freem(control); 1359 *controlp = control = m_get(M_WAIT, MT_CONTROL); 1360 } 1361 MEXTADD(control, newcm, CMSG_SPACE(nfds * sizeof(file_t *)), 1362 M_MBUF, NULL, NULL); 1363 cm = newcm; 1364 /* 1365 * Adjust message & mbuf to note amount of space 1366 * actually used. 1367 */ 1368 cm->cmsg_len = CMSG_LEN(nfds * sizeof(file_t *)); 1369 control->m_len = CMSG_SPACE(nfds * sizeof(file_t *)); 1370 } 1371 1372 return error; 1373 } 1374 1375 struct mbuf * 1376 unp_addsockcred(struct lwp *l, struct mbuf *control) 1377 { 1378 struct cmsghdr *cmp; 1379 struct sockcred *sc; 1380 struct mbuf *m, *n; 1381 int len, space, i; 1382 1383 len = CMSG_LEN(SOCKCREDSIZE(kauth_cred_ngroups(l->l_cred))); 1384 space = CMSG_SPACE(SOCKCREDSIZE(kauth_cred_ngroups(l->l_cred))); 1385 1386 m = m_get(M_WAIT, MT_CONTROL); 1387 if (space > MLEN) { 1388 if (space > MCLBYTES) 1389 MEXTMALLOC(m, space, M_WAITOK); 1390 else 1391 m_clget(m, M_WAIT); 1392 if ((m->m_flags & M_EXT) == 0) { 1393 m_free(m); 1394 return (control); 1395 } 1396 } 1397 1398 m->m_len = space; 1399 m->m_next = NULL; 1400 cmp = mtod(m, struct cmsghdr *); 1401 sc = (struct sockcred *)CMSG_DATA(cmp); 1402 cmp->cmsg_len = len; 1403 cmp->cmsg_level = SOL_SOCKET; 1404 cmp->cmsg_type = SCM_CREDS; 1405 sc->sc_uid = kauth_cred_getuid(l->l_cred); 1406 sc->sc_euid = kauth_cred_geteuid(l->l_cred); 1407 sc->sc_gid = kauth_cred_getgid(l->l_cred); 1408 sc->sc_egid = kauth_cred_getegid(l->l_cred); 1409 sc->sc_ngroups = kauth_cred_ngroups(l->l_cred); 1410 for (i = 0; i < sc->sc_ngroups; i++) 1411 sc->sc_groups[i] = kauth_cred_group(l->l_cred, i); 1412 1413 /* 1414 * If a control message already exists, append us to the end. 1415 */ 1416 if (control != NULL) { 1417 for (n = control; n->m_next != NULL; n = n->m_next) 1418 ; 1419 n->m_next = m; 1420 } else 1421 control = m; 1422 1423 return (control); 1424 } 1425 1426 /* 1427 * Do a mark-sweep GC of files in the system, to free up any which are 1428 * caught in flight to an about-to-be-closed socket. Additionally, 1429 * process deferred file closures. 1430 */ 1431 static void 1432 unp_gc(file_t *dp) 1433 { 1434 extern struct domain unixdomain; 1435 file_t *fp, *np; 1436 struct socket *so, *so1; 1437 u_int i, old, new; 1438 bool didwork; 1439 1440 KASSERT(curlwp == unp_thread_lwp); 1441 KASSERT(mutex_owned(&filelist_lock)); 1442 1443 /* 1444 * First, process deferred file closures. 1445 */ 1446 while (!SLIST_EMPTY(&unp_thread_discard)) { 1447 fp = SLIST_FIRST(&unp_thread_discard); 1448 KASSERT(fp->f_unpcount > 0); 1449 KASSERT(fp->f_count > 0); 1450 KASSERT(fp->f_msgcount > 0); 1451 KASSERT(fp->f_count >= fp->f_unpcount); 1452 KASSERT(fp->f_count >= fp->f_msgcount); 1453 KASSERT(fp->f_msgcount >= fp->f_unpcount); 1454 SLIST_REMOVE_HEAD(&unp_thread_discard, f_unplist); 1455 i = fp->f_unpcount; 1456 fp->f_unpcount = 0; 1457 mutex_exit(&filelist_lock); 1458 for (; i != 0; i--) { 1459 unp_discard_now(fp); 1460 } 1461 mutex_enter(&filelist_lock); 1462 } 1463 1464 /* 1465 * Clear mark bits. Ensure that we don't consider new files 1466 * entering the file table during this loop (they will not have 1467 * FSCAN set). 1468 */ 1469 unp_defer = 0; 1470 LIST_FOREACH(fp, &filehead, f_list) { 1471 for (old = fp->f_flag;; old = new) { 1472 new = atomic_cas_uint(&fp->f_flag, old, 1473 (old | FSCAN) & ~(FMARK|FDEFER)); 1474 if (__predict_true(old == new)) { 1475 break; 1476 } 1477 } 1478 } 1479 1480 /* 1481 * Iterate over the set of sockets, marking ones believed (based on 1482 * refcount) to be referenced from a process, and marking for rescan 1483 * sockets which are queued on a socket. Recan continues descending 1484 * and searching for sockets referenced by sockets (FDEFER), until 1485 * there are no more socket->socket references to be discovered. 1486 */ 1487 do { 1488 didwork = false; 1489 for (fp = LIST_FIRST(&filehead); fp != NULL; fp = np) { 1490 KASSERT(mutex_owned(&filelist_lock)); 1491 np = LIST_NEXT(fp, f_list); 1492 mutex_enter(&fp->f_lock); 1493 if ((fp->f_flag & FDEFER) != 0) { 1494 atomic_and_uint(&fp->f_flag, ~FDEFER); 1495 unp_defer--; 1496 KASSERT(fp->f_count != 0); 1497 } else { 1498 if (fp->f_count == 0 || 1499 (fp->f_flag & FMARK) != 0 || 1500 fp->f_count == fp->f_msgcount || 1501 fp->f_unpcount != 0) { 1502 mutex_exit(&fp->f_lock); 1503 continue; 1504 } 1505 } 1506 atomic_or_uint(&fp->f_flag, FMARK); 1507 1508 if (fp->f_type != DTYPE_SOCKET || 1509 (so = fp->f_data) == NULL || 1510 so->so_proto->pr_domain != &unixdomain || 1511 (so->so_proto->pr_flags & PR_RIGHTS) == 0) { 1512 mutex_exit(&fp->f_lock); 1513 continue; 1514 } 1515 1516 /* Gain file ref, mark our position, and unlock. */ 1517 didwork = true; 1518 LIST_INSERT_AFTER(fp, dp, f_list); 1519 fp->f_count++; 1520 mutex_exit(&fp->f_lock); 1521 mutex_exit(&filelist_lock); 1522 1523 /* 1524 * Mark files referenced from sockets queued on the 1525 * accept queue as well. 1526 */ 1527 solock(so); 1528 unp_scan(so->so_rcv.sb_mb, unp_mark, 0); 1529 if ((so->so_options & SO_ACCEPTCONN) != 0) { 1530 TAILQ_FOREACH(so1, &so->so_q0, so_qe) { 1531 unp_scan(so1->so_rcv.sb_mb, unp_mark, 0); 1532 } 1533 TAILQ_FOREACH(so1, &so->so_q, so_qe) { 1534 unp_scan(so1->so_rcv.sb_mb, unp_mark, 0); 1535 } 1536 } 1537 sounlock(so); 1538 1539 /* Re-lock and restart from where we left off. */ 1540 closef(fp); 1541 mutex_enter(&filelist_lock); 1542 np = LIST_NEXT(dp, f_list); 1543 LIST_REMOVE(dp, f_list); 1544 } 1545 /* 1546 * Bail early if we did nothing in the loop above. Could 1547 * happen because of concurrent activity causing unp_defer 1548 * to get out of sync. 1549 */ 1550 } while (unp_defer != 0 && didwork); 1551 1552 /* 1553 * Sweep pass. 1554 * 1555 * We grab an extra reference to each of the files that are 1556 * not otherwise accessible and then free the rights that are 1557 * stored in messages on them. 1558 */ 1559 for (fp = LIST_FIRST(&filehead); fp != NULL; fp = np) { 1560 KASSERT(mutex_owned(&filelist_lock)); 1561 np = LIST_NEXT(fp, f_list); 1562 mutex_enter(&fp->f_lock); 1563 1564 /* 1565 * Ignore non-sockets. 1566 * Ignore dead sockets, or sockets with pending close. 1567 * Ignore sockets obviously referenced elsewhere. 1568 * Ignore sockets marked as referenced by our scan. 1569 * Ignore new sockets that did not exist during the scan. 1570 */ 1571 if (fp->f_type != DTYPE_SOCKET || 1572 fp->f_count == 0 || fp->f_unpcount != 0 || 1573 fp->f_count != fp->f_msgcount || 1574 (fp->f_flag & (FMARK | FSCAN)) != FSCAN) { 1575 mutex_exit(&fp->f_lock); 1576 continue; 1577 } 1578 1579 /* Gain file ref, mark our position, and unlock. */ 1580 LIST_INSERT_AFTER(fp, dp, f_list); 1581 fp->f_count++; 1582 mutex_exit(&fp->f_lock); 1583 mutex_exit(&filelist_lock); 1584 1585 /* 1586 * Flush all data from the socket's receive buffer. 1587 * This will cause files referenced only by the 1588 * socket to be queued for close. 1589 */ 1590 so = fp->f_data; 1591 solock(so); 1592 sorflush(so); 1593 sounlock(so); 1594 1595 /* Re-lock and restart from where we left off. */ 1596 closef(fp); 1597 mutex_enter(&filelist_lock); 1598 np = LIST_NEXT(dp, f_list); 1599 LIST_REMOVE(dp, f_list); 1600 } 1601 } 1602 1603 /* 1604 * Garbage collector thread. While SCM_RIGHTS messages are in transit, 1605 * wake once per second to garbage collect. Run continually while we 1606 * have deferred closes to process. 1607 */ 1608 static void 1609 unp_thread(void *cookie) 1610 { 1611 file_t *dp; 1612 1613 /* Allocate a dummy file for our scans. */ 1614 if ((dp = fgetdummy()) == NULL) { 1615 panic("unp_thread"); 1616 } 1617 1618 mutex_enter(&filelist_lock); 1619 for (;;) { 1620 KASSERT(mutex_owned(&filelist_lock)); 1621 if (SLIST_EMPTY(&unp_thread_discard)) { 1622 if (unp_rights != 0) { 1623 (void)cv_timedwait(&unp_thread_cv, 1624 &filelist_lock, hz); 1625 } else { 1626 cv_wait(&unp_thread_cv, &filelist_lock); 1627 } 1628 } 1629 unp_gc(dp); 1630 } 1631 /* NOTREACHED */ 1632 } 1633 1634 /* 1635 * Kick the garbage collector into action if there is something for 1636 * it to process. 1637 */ 1638 static void 1639 unp_thread_kick(void) 1640 { 1641 1642 if (!SLIST_EMPTY(&unp_thread_discard) || unp_rights != 0) { 1643 mutex_enter(&filelist_lock); 1644 cv_signal(&unp_thread_cv); 1645 mutex_exit(&filelist_lock); 1646 } 1647 } 1648 1649 void 1650 unp_dispose(struct mbuf *m) 1651 { 1652 1653 if (m) 1654 unp_scan(m, unp_discard_later, 1); 1655 } 1656 1657 void 1658 unp_scan(struct mbuf *m0, void (*op)(file_t *), int discard) 1659 { 1660 struct mbuf *m; 1661 file_t **rp, *fp; 1662 struct cmsghdr *cm; 1663 int i, qfds; 1664 1665 while (m0) { 1666 for (m = m0; m; m = m->m_next) { 1667 if (m->m_type != MT_CONTROL || 1668 m->m_len < sizeof(*cm)) { 1669 continue; 1670 } 1671 cm = mtod(m, struct cmsghdr *); 1672 if (cm->cmsg_level != SOL_SOCKET || 1673 cm->cmsg_type != SCM_RIGHTS) 1674 continue; 1675 qfds = (cm->cmsg_len - CMSG_ALIGN(sizeof(*cm))) 1676 / sizeof(file_t *); 1677 rp = (file_t **)CMSG_DATA(cm); 1678 for (i = 0; i < qfds; i++) { 1679 fp = *rp; 1680 if (discard) { 1681 *rp = 0; 1682 } 1683 (*op)(fp); 1684 rp++; 1685 } 1686 } 1687 m0 = m0->m_nextpkt; 1688 } 1689 } 1690 1691 void 1692 unp_mark(file_t *fp) 1693 { 1694 1695 if (fp == NULL) 1696 return; 1697 1698 /* If we're already deferred, don't screw up the defer count */ 1699 mutex_enter(&fp->f_lock); 1700 if (fp->f_flag & (FMARK | FDEFER)) { 1701 mutex_exit(&fp->f_lock); 1702 return; 1703 } 1704 1705 /* 1706 * Minimize the number of deferrals... Sockets are the only type of 1707 * file which can hold references to another file, so just mark 1708 * other files, and defer unmarked sockets for the next pass. 1709 */ 1710 if (fp->f_type == DTYPE_SOCKET) { 1711 unp_defer++; 1712 KASSERT(fp->f_count != 0); 1713 atomic_or_uint(&fp->f_flag, FDEFER); 1714 } else { 1715 atomic_or_uint(&fp->f_flag, FMARK); 1716 } 1717 mutex_exit(&fp->f_lock); 1718 } 1719 1720 static void 1721 unp_discard_now(file_t *fp) 1722 { 1723 1724 if (fp == NULL) 1725 return; 1726 1727 KASSERT(fp->f_count > 0); 1728 KASSERT(fp->f_msgcount > 0); 1729 1730 mutex_enter(&fp->f_lock); 1731 fp->f_msgcount--; 1732 mutex_exit(&fp->f_lock); 1733 atomic_dec_uint(&unp_rights); 1734 (void)closef(fp); 1735 } 1736 1737 static void 1738 unp_discard_later(file_t *fp) 1739 { 1740 1741 if (fp == NULL) 1742 return; 1743 1744 KASSERT(fp->f_count > 0); 1745 KASSERT(fp->f_msgcount > 0); 1746 1747 mutex_enter(&filelist_lock); 1748 if (fp->f_unpcount++ == 0) { 1749 SLIST_INSERT_HEAD(&unp_thread_discard, fp, f_unplist); 1750 } 1751 mutex_exit(&filelist_lock); 1752 } 1753
Indexes created Sun Sep 21 20:09:37 GMT 2025