1 /* $NetBSD: uipc_usrreq.c,v 1.105 2008/01/25 14:32:14 ad Exp $ */ 2 3 /*- 4 * Copyright (c) 1998, 2000, 2004 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. 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 * 3. All advertising materials mentioning features or use of this software 20 * must display the following acknowledgement: 21 * This product includes software developed by the NetBSD 22 * Foundation, Inc. and its contributors. 23 * 4. Neither the name of The NetBSD Foundation nor the names of its 24 * contributors may be used to endorse or promote products derived 25 * from this software without specific prior written permission. 26 * 27 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 28 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 29 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 30 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 31 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 32 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 33 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 34 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 35 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 36 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 37 * POSSIBILITY OF SUCH DAMAGE. 38 */ 39 40 /* 41 * Copyright (c) 1982, 1986, 1989, 1991, 1993 42 * The Regents of the University of California. All rights reserved. 43 * 44 * Redistribution and use in source and binary forms, with or without 45 * modification, are permitted provided that the following conditions 46 * are met: 47 * 1. Redistributions of source code must retain the above copyright 48 * notice, this list of conditions and the following disclaimer. 49 * 2. Redistributions in binary form must reproduce the above copyright 50 * notice, this list of conditions and the following disclaimer in the 51 * documentation and/or other materials provided with the distribution. 52 * 3. Neither the name of the University nor the names of its contributors 53 * may be used to endorse or promote products derived from this software 54 * without specific prior written permission. 55 * 56 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 57 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 58 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 59 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 60 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 61 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 62 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 63 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 64 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 65 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 66 * SUCH DAMAGE. 67 * 68 * @(#)uipc_usrreq.c 8.9 (Berkeley) 5/14/95 69 */ 70 71 /* 72 * Copyright (c) 1997 Christopher G. Demetriou. All rights reserved. 73 * 74 * Redistribution and use in source and binary forms, with or without 75 * modification, are permitted provided that the following conditions 76 * are met: 77 * 1. Redistributions of source code must retain the above copyright 78 * notice, this list of conditions and the following disclaimer. 79 * 2. Redistributions in binary form must reproduce the above copyright 80 * notice, this list of conditions and the following disclaimer in the 81 * documentation and/or other materials provided with the distribution. 82 * 3. All advertising materials mentioning features or use of this software 83 * must display the following acknowledgement: 84 * This product includes software developed by the University of 85 * California, Berkeley and its contributors. 86 * 4. Neither the name of the University nor the names of its contributors 87 * may be used to endorse or promote products derived from this software 88 * without specific prior written permission. 89 * 90 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 91 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 92 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 93 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 94 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 95 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 96 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 97 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 98 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 99 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 100 * SUCH DAMAGE. 101 * 102 * @(#)uipc_usrreq.c 8.9 (Berkeley) 5/14/95 103 */ 104 105 #include <sys/cdefs.h> 106 __KERNEL_RCSID(0, "$NetBSD: uipc_usrreq.c,v 1.105 2008/01/25 14:32:14 ad Exp $"); 107 108 #include <sys/param.h> 109 #include <sys/systm.h> 110 #include <sys/proc.h> 111 #include <sys/filedesc.h> 112 #include <sys/domain.h> 113 #include <sys/protosw.h> 114 #include <sys/socket.h> 115 #include <sys/socketvar.h> 116 #include <sys/unpcb.h> 117 #include <sys/un.h> 118 #include <sys/namei.h> 119 #include <sys/vnode.h> 120 #include <sys/file.h> 121 #include <sys/stat.h> 122 #include <sys/mbuf.h> 123 #include <sys/kauth.h> 124 #include <sys/kmem.h> 125 126 /* 127 * Unix communications domain. 128 * 129 * TODO: 130 * SEQPACKET, RDM 131 * rethink name space problems 132 * need a proper out-of-band 133 */ 134 const struct sockaddr_un sun_noname = { 135 .sun_len = sizeof(sun_noname), 136 .sun_family = AF_LOCAL, 137 }; 138 ino_t unp_ino; /* prototype for fake inode numbers */ 139 140 struct mbuf *unp_addsockcred(struct lwp *, struct mbuf *); 141 142 int 143 unp_output(struct mbuf *m, struct mbuf *control, struct unpcb *unp, 144 struct lwp *l) 145 { 146 struct socket *so2; 147 const struct sockaddr_un *sun; 148 149 so2 = unp->unp_conn->unp_socket; 150 if (unp->unp_addr) 151 sun = unp->unp_addr; 152 else 153 sun = &sun_noname; 154 if (unp->unp_conn->unp_flags & UNP_WANTCRED) 155 control = unp_addsockcred(l, control); 156 if (sbappendaddr(&so2->so_rcv, (const struct sockaddr *)sun, m, 157 control) == 0) { 158 unp_dispose(control); 159 m_freem(control); 160 m_freem(m); 161 so2->so_rcv.sb_overflowed++; 162 return (ENOBUFS); 163 } else { 164 sorwakeup(so2); 165 return (0); 166 } 167 } 168 169 void 170 unp_setsockaddr(struct unpcb *unp, struct mbuf *nam) 171 { 172 const struct sockaddr_un *sun; 173 174 if (unp->unp_addr) 175 sun = unp->unp_addr; 176 else 177 sun = &sun_noname; 178 nam->m_len = sun->sun_len; 179 if (nam->m_len > MLEN) 180 MEXTMALLOC(nam, nam->m_len, M_WAITOK); 181 memcpy(mtod(nam, void *), sun, (size_t)nam->m_len); 182 } 183 184 void 185 unp_setpeeraddr(struct unpcb *unp, struct mbuf *nam) 186 { 187 const struct sockaddr_un *sun; 188 189 if (unp->unp_conn && unp->unp_conn->unp_addr) 190 sun = unp->unp_conn->unp_addr; 191 else 192 sun = &sun_noname; 193 nam->m_len = sun->sun_len; 194 if (nam->m_len > MLEN) 195 MEXTMALLOC(nam, nam->m_len, M_WAITOK); 196 memcpy(mtod(nam, void *), sun, (size_t)nam->m_len); 197 } 198 199 /*ARGSUSED*/ 200 int 201 uipc_usrreq(struct socket *so, int req, struct mbuf *m, struct mbuf *nam, 202 struct mbuf *control, struct lwp *l) 203 { 204 struct unpcb *unp = sotounpcb(so); 205 struct socket *so2; 206 struct proc *p; 207 u_int newhiwat; 208 int error = 0; 209 210 if (req == PRU_CONTROL) 211 return (EOPNOTSUPP); 212 213 #ifdef DIAGNOSTIC 214 if (req != PRU_SEND && req != PRU_SENDOOB && control) 215 panic("uipc_usrreq: unexpected control mbuf"); 216 #endif 217 p = l ? l->l_proc : NULL; 218 if (unp == 0 && req != PRU_ATTACH) { 219 error = EINVAL; 220 goto release; 221 } 222 223 switch (req) { 224 225 case PRU_ATTACH: 226 if (unp != 0) { 227 error = EISCONN; 228 break; 229 } 230 error = unp_attach(so); 231 break; 232 233 case PRU_DETACH: 234 unp_detach(unp); 235 break; 236 237 case PRU_BIND: 238 KASSERT(l != NULL); 239 error = unp_bind(unp, nam, l); 240 break; 241 242 case PRU_LISTEN: 243 if (unp->unp_vnode == 0) 244 error = EINVAL; 245 break; 246 247 case PRU_CONNECT: 248 KASSERT(l != NULL); 249 error = unp_connect(so, nam, l); 250 break; 251 252 case PRU_CONNECT2: 253 error = unp_connect2(so, (struct socket *)nam, PRU_CONNECT2); 254 break; 255 256 case PRU_DISCONNECT: 257 unp_disconnect(unp); 258 break; 259 260 case PRU_ACCEPT: 261 unp_setpeeraddr(unp, nam); 262 /* 263 * Mark the initiating STREAM socket as connected *ONLY* 264 * after it's been accepted. This prevents a client from 265 * overrunning a server and receiving ECONNREFUSED. 266 */ 267 if (unp->unp_conn != NULL && 268 (unp->unp_conn->unp_socket->so_state & SS_ISCONNECTING)) 269 soisconnected(unp->unp_conn->unp_socket); 270 break; 271 272 case PRU_SHUTDOWN: 273 socantsendmore(so); 274 unp_shutdown(unp); 275 break; 276 277 case PRU_RCVD: 278 switch (so->so_type) { 279 280 case SOCK_DGRAM: 281 panic("uipc 1"); 282 /*NOTREACHED*/ 283 284 case SOCK_STREAM: 285 #define rcv (&so->so_rcv) 286 #define snd (&so2->so_snd) 287 if (unp->unp_conn == 0) 288 break; 289 so2 = unp->unp_conn->unp_socket; 290 /* 291 * Adjust backpressure on sender 292 * and wakeup any waiting to write. 293 */ 294 snd->sb_mbmax += unp->unp_mbcnt - rcv->sb_mbcnt; 295 unp->unp_mbcnt = rcv->sb_mbcnt; 296 newhiwat = snd->sb_hiwat + unp->unp_cc - rcv->sb_cc; 297 (void)chgsbsize(so2->so_uidinfo, 298 &snd->sb_hiwat, newhiwat, RLIM_INFINITY); 299 unp->unp_cc = rcv->sb_cc; 300 sowwakeup(so2); 301 #undef snd 302 #undef rcv 303 break; 304 305 default: 306 panic("uipc 2"); 307 } 308 break; 309 310 case PRU_SEND: 311 /* 312 * Note: unp_internalize() rejects any control message 313 * other than SCM_RIGHTS, and only allows one. This 314 * has the side-effect of preventing a caller from 315 * forging SCM_CREDS. 316 */ 317 if (control) { 318 KASSERT(l != NULL); 319 if ((error = unp_internalize(control, l)) != 0) 320 goto die; 321 } 322 switch (so->so_type) { 323 324 case SOCK_DGRAM: { 325 if (nam) { 326 if ((so->so_state & SS_ISCONNECTED) != 0) { 327 error = EISCONN; 328 goto die; 329 } 330 KASSERT(l != NULL); 331 error = unp_connect(so, nam, l); 332 if (error) { 333 die: 334 unp_dispose(control); 335 m_freem(control); 336 m_freem(m); 337 break; 338 } 339 } else { 340 if ((so->so_state & SS_ISCONNECTED) == 0) { 341 error = ENOTCONN; 342 goto die; 343 } 344 } 345 KASSERT(p != NULL); 346 error = unp_output(m, control, unp, l); 347 if (nam) 348 unp_disconnect(unp); 349 break; 350 } 351 352 case SOCK_STREAM: 353 #define rcv (&so2->so_rcv) 354 #define snd (&so->so_snd) 355 if (unp->unp_conn == NULL) { 356 error = ENOTCONN; 357 break; 358 } 359 so2 = unp->unp_conn->unp_socket; 360 if (unp->unp_conn->unp_flags & UNP_WANTCRED) { 361 /* 362 * Credentials are passed only once on 363 * SOCK_STREAM. 364 */ 365 unp->unp_conn->unp_flags &= ~UNP_WANTCRED; 366 control = unp_addsockcred(l, control); 367 } 368 /* 369 * Send to paired receive port, and then reduce 370 * send buffer hiwater marks to maintain backpressure. 371 * Wake up readers. 372 */ 373 if (control) { 374 if (sbappendcontrol(rcv, m, control) == 0) { 375 unp_dispose(control); 376 m_freem(control); 377 } 378 } else 379 sbappend(rcv, m); 380 snd->sb_mbmax -= 381 rcv->sb_mbcnt - unp->unp_conn->unp_mbcnt; 382 unp->unp_conn->unp_mbcnt = rcv->sb_mbcnt; 383 newhiwat = snd->sb_hiwat - 384 (rcv->sb_cc - unp->unp_conn->unp_cc); 385 (void)chgsbsize(so->so_uidinfo, 386 &snd->sb_hiwat, newhiwat, RLIM_INFINITY); 387 unp->unp_conn->unp_cc = rcv->sb_cc; 388 sorwakeup(so2); 389 #undef snd 390 #undef rcv 391 break; 392 393 default: 394 panic("uipc 4"); 395 } 396 break; 397 398 case PRU_ABORT: 399 unp_drop(unp, ECONNABORTED); 400 401 KASSERT(so->so_head == NULL); 402 #ifdef DIAGNOSTIC 403 if (so->so_pcb == 0) 404 panic("uipc 5: drop killed pcb"); 405 #endif 406 unp_detach(unp); 407 break; 408 409 case PRU_SENSE: 410 ((struct stat *) m)->st_blksize = so->so_snd.sb_hiwat; 411 if (so->so_type == SOCK_STREAM && unp->unp_conn != 0) { 412 so2 = unp->unp_conn->unp_socket; 413 ((struct stat *) m)->st_blksize += so2->so_rcv.sb_cc; 414 } 415 ((struct stat *) m)->st_dev = NODEV; 416 if (unp->unp_ino == 0) 417 unp->unp_ino = unp_ino++; 418 ((struct stat *) m)->st_atimespec = 419 ((struct stat *) m)->st_mtimespec = 420 ((struct stat *) m)->st_ctimespec = unp->unp_ctime; 421 ((struct stat *) m)->st_ino = unp->unp_ino; 422 return (0); 423 424 case PRU_RCVOOB: 425 error = EOPNOTSUPP; 426 break; 427 428 case PRU_SENDOOB: 429 m_freem(control); 430 m_freem(m); 431 error = EOPNOTSUPP; 432 break; 433 434 case PRU_SOCKADDR: 435 unp_setsockaddr(unp, nam); 436 break; 437 438 case PRU_PEERADDR: 439 unp_setpeeraddr(unp, nam); 440 break; 441 442 default: 443 panic("piusrreq"); 444 } 445 446 release: 447 return (error); 448 } 449 450 /* 451 * Unix domain socket option processing. 452 */ 453 int 454 uipc_ctloutput(int op, struct socket *so, int level, int optname, 455 struct mbuf **mp) 456 { 457 struct unpcb *unp = sotounpcb(so); 458 struct mbuf *m = *mp; 459 int optval = 0, error = 0; 460 461 if (level != 0) { 462 error = ENOPROTOOPT; 463 if (op == PRCO_SETOPT && m) 464 (void) m_free(m); 465 } else switch (op) { 466 467 case PRCO_SETOPT: 468 switch (optname) { 469 case LOCAL_CREDS: 470 case LOCAL_CONNWAIT: 471 if (m == NULL || m->m_len != sizeof(int)) 472 error = EINVAL; 473 else { 474 optval = *mtod(m, int *); 475 switch (optname) { 476 #define OPTSET(bit) \ 477 if (optval) \ 478 unp->unp_flags |= (bit); \ 479 else \ 480 unp->unp_flags &= ~(bit); 481 482 case LOCAL_CREDS: 483 OPTSET(UNP_WANTCRED); 484 break; 485 case LOCAL_CONNWAIT: 486 OPTSET(UNP_CONNWAIT); 487 break; 488 } 489 } 490 break; 491 #undef OPTSET 492 493 default: 494 error = ENOPROTOOPT; 495 break; 496 } 497 if (m) 498 (void) m_free(m); 499 break; 500 501 case PRCO_GETOPT: 502 switch (optname) { 503 case LOCAL_PEEREID: 504 if (unp->unp_flags & UNP_EIDSVALID) { 505 *mp = m = m_get(M_WAIT, MT_SOOPTS); 506 m->m_len = sizeof(struct unpcbid); 507 *mtod(m, struct unpcbid *) = unp->unp_connid; 508 } else { 509 error = EINVAL; 510 } 511 break; 512 case LOCAL_CREDS: 513 *mp = m = m_get(M_WAIT, MT_SOOPTS); 514 m->m_len = sizeof(int); 515 516 #define OPTBIT(bit) (unp->unp_flags & (bit) ? 1 : 0) 517 518 optval = OPTBIT(UNP_WANTCRED); 519 *mtod(m, int *) = optval; 520 break; 521 #undef OPTBIT 522 523 default: 524 error = ENOPROTOOPT; 525 break; 526 } 527 break; 528 } 529 return (error); 530 } 531 532 /* 533 * Both send and receive buffers are allocated PIPSIZ bytes of buffering 534 * for stream sockets, although the total for sender and receiver is 535 * actually only PIPSIZ. 536 * Datagram sockets really use the sendspace as the maximum datagram size, 537 * and don't really want to reserve the sendspace. Their recvspace should 538 * be large enough for at least one max-size datagram plus address. 539 */ 540 #define PIPSIZ 4096 541 u_long unpst_sendspace = PIPSIZ; 542 u_long unpst_recvspace = PIPSIZ; 543 u_long unpdg_sendspace = 2*1024; /* really max datagram size */ 544 u_long unpdg_recvspace = 4*1024; 545 546 int unp_rights; /* file descriptors in flight */ 547 548 int 549 unp_attach(struct socket *so) 550 { 551 struct unpcb *unp; 552 int error; 553 554 if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) { 555 switch (so->so_type) { 556 557 case SOCK_STREAM: 558 error = soreserve(so, unpst_sendspace, unpst_recvspace); 559 break; 560 561 case SOCK_DGRAM: 562 error = soreserve(so, unpdg_sendspace, unpdg_recvspace); 563 break; 564 565 default: 566 panic("unp_attach"); 567 } 568 if (error) 569 return (error); 570 } 571 unp = malloc(sizeof(*unp), M_PCB, M_NOWAIT); 572 if (unp == NULL) 573 return (ENOBUFS); 574 memset((void *)unp, 0, sizeof(*unp)); 575 unp->unp_socket = so; 576 so->so_pcb = unp; 577 nanotime(&unp->unp_ctime); 578 return (0); 579 } 580 581 void 582 unp_detach(struct unpcb *unp) 583 { 584 585 if (unp->unp_vnode) { 586 unp->unp_vnode->v_socket = 0; 587 vrele(unp->unp_vnode); 588 unp->unp_vnode = 0; 589 } 590 if (unp->unp_conn) 591 unp_disconnect(unp); 592 while (unp->unp_refs) 593 unp_drop(unp->unp_refs, ECONNRESET); 594 soisdisconnected(unp->unp_socket); 595 unp->unp_socket->so_pcb = 0; 596 if (unp->unp_addr) 597 free(unp->unp_addr, M_SONAME); 598 if (unp_rights) { 599 /* 600 * Normally the receive buffer is flushed later, 601 * in sofree, but if our receive buffer holds references 602 * to descriptors that are now garbage, we will dispose 603 * of those descriptor references after the garbage collector 604 * gets them (resulting in a "panic: closef: count < 0"). 605 */ 606 sorflush(unp->unp_socket); 607 free(unp, M_PCB); 608 unp_gc(); 609 } else 610 free(unp, M_PCB); 611 } 612 613 int 614 unp_bind(struct unpcb *unp, struct mbuf *nam, struct lwp *l) 615 { 616 struct sockaddr_un *sun; 617 struct vnode *vp; 618 struct vattr vattr; 619 size_t addrlen; 620 struct proc *p; 621 int error; 622 struct nameidata nd; 623 624 if (unp->unp_vnode != 0) 625 return (EINVAL); 626 627 p = l->l_proc; 628 /* 629 * Allocate the new sockaddr. We have to allocate one 630 * extra byte so that we can ensure that the pathname 631 * is nul-terminated. 632 */ 633 addrlen = nam->m_len + 1; 634 sun = malloc(addrlen, M_SONAME, M_WAITOK); 635 m_copydata(nam, 0, nam->m_len, (void *)sun); 636 *(((char *)sun) + nam->m_len) = '\0'; 637 638 NDINIT(&nd, CREATE, FOLLOW | LOCKPARENT | TRYEMULROOT, UIO_SYSSPACE, 639 sun->sun_path); 640 641 /* SHOULD BE ABLE TO ADOPT EXISTING AND wakeup() ALA FIFO's */ 642 if ((error = namei(&nd)) != 0) 643 goto bad; 644 vp = nd.ni_vp; 645 if (vp != NULL) { 646 VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd); 647 if (nd.ni_dvp == vp) 648 vrele(nd.ni_dvp); 649 else 650 vput(nd.ni_dvp); 651 vrele(vp); 652 error = EADDRINUSE; 653 goto bad; 654 } 655 VATTR_NULL(&vattr); 656 vattr.va_type = VSOCK; 657 vattr.va_mode = ACCESSPERMS & ~(p->p_cwdi->cwdi_cmask); 658 error = VOP_CREATE(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr); 659 if (error) 660 goto bad; 661 vp = nd.ni_vp; 662 vp->v_socket = unp->unp_socket; 663 unp->unp_vnode = vp; 664 unp->unp_addrlen = addrlen; 665 unp->unp_addr = sun; 666 unp->unp_connid.unp_pid = p->p_pid; 667 unp->unp_connid.unp_euid = kauth_cred_geteuid(p->p_cred); 668 unp->unp_connid.unp_egid = kauth_cred_getegid(p->p_cred); 669 unp->unp_flags |= UNP_EIDSBIND; 670 VOP_UNLOCK(vp, 0); 671 return (0); 672 673 bad: 674 free(sun, M_SONAME); 675 return (error); 676 } 677 678 int 679 unp_connect(struct socket *so, struct mbuf *nam, struct lwp *l) 680 { 681 struct sockaddr_un *sun; 682 struct vnode *vp; 683 struct socket *so2, *so3; 684 struct unpcb *unp, *unp2, *unp3; 685 size_t addrlen; 686 struct proc *p; 687 int error; 688 struct nameidata nd; 689 690 p = l->l_proc; 691 /* 692 * Allocate a temporary sockaddr. We have to allocate one extra 693 * byte so that we can ensure that the pathname is nul-terminated. 694 * When we establish the connection, we copy the other PCB's 695 * sockaddr to our own. 696 */ 697 addrlen = nam->m_len + 1; 698 sun = malloc(addrlen, M_SONAME, M_WAITOK); 699 m_copydata(nam, 0, nam->m_len, (void *)sun); 700 *(((char *)sun) + nam->m_len) = '\0'; 701 702 NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | TRYEMULROOT, UIO_SYSSPACE, 703 sun->sun_path); 704 705 if ((error = namei(&nd)) != 0) 706 goto bad2; 707 vp = nd.ni_vp; 708 if (vp->v_type != VSOCK) { 709 error = ENOTSOCK; 710 goto bad; 711 } 712 if ((error = VOP_ACCESS(vp, VWRITE, l->l_cred)) != 0) 713 goto bad; 714 so2 = vp->v_socket; 715 if (so2 == 0) { 716 error = ECONNREFUSED; 717 goto bad; 718 } 719 if (so->so_type != so2->so_type) { 720 error = EPROTOTYPE; 721 goto bad; 722 } 723 if (so->so_proto->pr_flags & PR_CONNREQUIRED) { 724 if ((so2->so_options & SO_ACCEPTCONN) == 0 || 725 (so3 = sonewconn(so2, 0)) == 0) { 726 error = ECONNREFUSED; 727 goto bad; 728 } 729 unp = sotounpcb(so); 730 unp2 = sotounpcb(so2); 731 unp3 = sotounpcb(so3); 732 if (unp2->unp_addr) { 733 unp3->unp_addr = malloc(unp2->unp_addrlen, 734 M_SONAME, M_WAITOK); 735 memcpy(unp3->unp_addr, unp2->unp_addr, 736 unp2->unp_addrlen); 737 unp3->unp_addrlen = unp2->unp_addrlen; 738 } 739 unp3->unp_flags = unp2->unp_flags; 740 unp3->unp_connid.unp_pid = p->p_pid; 741 unp3->unp_connid.unp_euid = kauth_cred_geteuid(p->p_cred); 742 unp3->unp_connid.unp_egid = kauth_cred_getegid(p->p_cred); 743 unp3->unp_flags |= UNP_EIDSVALID; 744 so2 = so3; 745 if (unp2->unp_flags & UNP_EIDSBIND) { 746 unp->unp_connid = unp2->unp_connid; 747 unp->unp_flags |= UNP_EIDSVALID; 748 } 749 } 750 error = unp_connect2(so, so2, PRU_CONNECT); 751 bad: 752 vput(vp); 753 bad2: 754 free(sun, M_SONAME); 755 return (error); 756 } 757 758 int 759 unp_connect2(struct socket *so, struct socket *so2, int req) 760 { 761 struct unpcb *unp = sotounpcb(so); 762 struct unpcb *unp2; 763 764 if (so2->so_type != so->so_type) 765 return (EPROTOTYPE); 766 unp2 = sotounpcb(so2); 767 unp->unp_conn = unp2; 768 switch (so->so_type) { 769 770 case SOCK_DGRAM: 771 unp->unp_nextref = unp2->unp_refs; 772 unp2->unp_refs = unp; 773 soisconnected(so); 774 break; 775 776 case SOCK_STREAM: 777 unp2->unp_conn = unp; 778 if (req == PRU_CONNECT && 779 ((unp->unp_flags | unp2->unp_flags) & UNP_CONNWAIT)) 780 soisconnecting(so); 781 else 782 soisconnected(so); 783 soisconnected(so2); 784 break; 785 786 default: 787 panic("unp_connect2"); 788 } 789 return (0); 790 } 791 792 void 793 unp_disconnect(struct unpcb *unp) 794 { 795 struct unpcb *unp2 = unp->unp_conn; 796 797 if (unp2 == 0) 798 return; 799 unp->unp_conn = 0; 800 switch (unp->unp_socket->so_type) { 801 802 case SOCK_DGRAM: 803 if (unp2->unp_refs == unp) 804 unp2->unp_refs = unp->unp_nextref; 805 else { 806 unp2 = unp2->unp_refs; 807 for (;;) { 808 if (unp2 == 0) 809 panic("unp_disconnect"); 810 if (unp2->unp_nextref == unp) 811 break; 812 unp2 = unp2->unp_nextref; 813 } 814 unp2->unp_nextref = unp->unp_nextref; 815 } 816 unp->unp_nextref = 0; 817 unp->unp_socket->so_state &= ~SS_ISCONNECTED; 818 break; 819 820 case SOCK_STREAM: 821 soisdisconnected(unp->unp_socket); 822 unp2->unp_conn = 0; 823 soisdisconnected(unp2->unp_socket); 824 break; 825 } 826 } 827 828 #ifdef notdef 829 unp_abort(struct unpcb *unp) 830 { 831 unp_detach(unp); 832 } 833 #endif 834 835 void 836 unp_shutdown(struct unpcb *unp) 837 { 838 struct socket *so; 839 840 if (unp->unp_socket->so_type == SOCK_STREAM && unp->unp_conn && 841 (so = unp->unp_conn->unp_socket)) 842 socantrcvmore(so); 843 } 844 845 void 846 unp_drop(struct unpcb *unp, int errno) 847 { 848 struct socket *so = unp->unp_socket; 849 850 so->so_error = errno; 851 unp_disconnect(unp); 852 if (so->so_head) { 853 so->so_pcb = 0; 854 sofree(so); 855 if (unp->unp_addr) 856 free(unp->unp_addr, M_SONAME); 857 free(unp, M_PCB); 858 } 859 } 860 861 #ifdef notdef 862 unp_drain(void) 863 { 864 865 } 866 #endif 867 868 int 869 unp_externalize(struct mbuf *rights, struct lwp *l) 870 { 871 struct cmsghdr *cm = mtod(rights, struct cmsghdr *); 872 struct proc *p = l->l_proc; 873 int i, *fdp; 874 struct file **rp; 875 struct file *fp; 876 int nfds, error = 0; 877 878 nfds = (cm->cmsg_len - CMSG_ALIGN(sizeof(*cm))) / 879 sizeof(struct file *); 880 rp = (struct file **)CMSG_DATA(cm); 881 882 fdp = malloc(nfds * sizeof(int), M_TEMP, M_WAITOK); 883 rw_enter(&p->p_cwdi->cwdi_lock, RW_READER); 884 885 /* Make sure the recipient should be able to see the descriptors.. */ 886 if (p->p_cwdi->cwdi_rdir != NULL) { 887 rp = (struct file **)CMSG_DATA(cm); 888 for (i = 0; i < nfds; i++) { 889 fp = *rp++; 890 /* 891 * If we are in a chroot'ed directory, and 892 * someone wants to pass us a directory, make 893 * sure it's inside the subtree we're allowed 894 * to access. 895 */ 896 if (fp->f_type == DTYPE_VNODE) { 897 struct vnode *vp = (struct vnode *)fp->f_data; 898 if ((vp->v_type == VDIR) && 899 !vn_isunder(vp, p->p_cwdi->cwdi_rdir, l)) { 900 error = EPERM; 901 break; 902 } 903 } 904 } 905 } 906 907 restart: 908 rp = (struct file **)CMSG_DATA(cm); 909 if (error != 0) { 910 for (i = 0; i < nfds; i++) { 911 fp = *rp; 912 /* 913 * zero the pointer before calling unp_discard, 914 * since it may end up in unp_gc().. 915 */ 916 *rp++ = 0; 917 unp_discard(fp); 918 } 919 goto out; 920 } 921 922 /* 923 * First loop -- allocate file descriptor table slots for the 924 * new descriptors. 925 */ 926 for (i = 0; i < nfds; i++) { 927 fp = *rp++; 928 if ((error = fdalloc(p, 0, &fdp[i])) != 0) { 929 /* 930 * Back out what we've done so far. 931 */ 932 for (--i; i >= 0; i--) 933 fdremove(p->p_fd, fdp[i]); 934 935 if (error == ENOSPC) { 936 fdexpand(p); 937 error = 0; 938 } else { 939 /* 940 * This is the error that has historically 941 * been returned, and some callers may 942 * expect it. 943 */ 944 error = EMSGSIZE; 945 } 946 goto restart; 947 } 948 949 /* 950 * Make the slot reference the descriptor so that 951 * fdalloc() works properly.. We finalize it all 952 * in the loop below. 953 */ 954 rw_enter(&p->p_fd->fd_lock, RW_WRITER); 955 p->p_fd->fd_ofiles[fdp[i]] = fp; 956 rw_exit(&p->p_fd->fd_lock); 957 } 958 959 /* 960 * Now that adding them has succeeded, update all of the 961 * descriptor passing state. 962 */ 963 rp = (struct file **)CMSG_DATA(cm); 964 for (i = 0; i < nfds; i++) { 965 fp = *rp++; 966 fp->f_msgcount--; 967 unp_rights--; 968 } 969 970 /* 971 * Copy temporary array to message and adjust length, in case of 972 * transition from large struct file pointers to ints. 973 */ 974 memcpy(CMSG_DATA(cm), fdp, nfds * sizeof(int)); 975 cm->cmsg_len = CMSG_LEN(nfds * sizeof(int)); 976 rights->m_len = CMSG_SPACE(nfds * sizeof(int)); 977 out: 978 rw_exit(&p->p_cwdi->cwdi_lock); 979 free(fdp, M_TEMP); 980 return (error); 981 } 982 983 int 984 unp_internalize(struct mbuf *control, struct lwp *l) 985 { 986 struct proc *p = l->l_proc; 987 struct filedesc *fdescp = p->p_fd; 988 struct cmsghdr *newcm, *cm = mtod(control, struct cmsghdr *); 989 struct file **rp, **files; 990 struct file *fp; 991 int i, fd, *fdp; 992 int nfds; 993 u_int neededspace; 994 995 /* Sanity check the control message header */ 996 if (cm->cmsg_type != SCM_RIGHTS || cm->cmsg_level != SOL_SOCKET || 997 cm->cmsg_len != control->m_len) 998 return (EINVAL); 999 1000 /* Verify that the file descriptors are valid */ 1001 nfds = (cm->cmsg_len - CMSG_ALIGN(sizeof(*cm))) / sizeof(int); 1002 fdp = (int *)CMSG_DATA(cm); 1003 for (i = 0; i < nfds; i++) { 1004 fd = *fdp++; 1005 if ((fp = fd_getfile(fdescp, fd)) == NULL) 1006 return (EBADF); 1007 /* XXXSMP grab reference to file */ 1008 FILE_UNLOCK(fp); 1009 } 1010 1011 /* Make sure we have room for the struct file pointers */ 1012 neededspace = CMSG_SPACE(nfds * sizeof(struct file *)) - 1013 control->m_len; 1014 if (neededspace > M_TRAILINGSPACE(control)) { 1015 1016 /* allocate new space and copy header into it */ 1017 newcm = malloc( 1018 CMSG_SPACE(nfds * sizeof(struct file *)), 1019 M_MBUF, M_WAITOK); 1020 if (newcm == NULL) { 1021 /* XXXSMP drop references to files */ 1022 return (E2BIG); 1023 } 1024 memcpy(newcm, cm, sizeof(struct cmsghdr)); 1025 files = (struct file **)CMSG_DATA(newcm); 1026 } else { 1027 /* we can convert in-place */ 1028 newcm = NULL; 1029 files = (struct file **)CMSG_DATA(cm); 1030 } 1031 1032 /* 1033 * Transform the file descriptors into struct file pointers, in 1034 * reverse order so that if pointers are bigger than ints, the 1035 * int won't get until we're done. 1036 */ 1037 rw_enter(&fdescp->fd_lock, RW_READER); 1038 fdp = (int *)CMSG_DATA(cm) + nfds; 1039 rp = files + nfds; 1040 for (i = 0; i < nfds; i++) { 1041 fp = fdescp->fd_ofiles[*--fdp]; 1042 FILE_LOCK(fp); 1043 #ifdef DIAGNOSTIC 1044 if (fp->f_iflags & FIF_WANTCLOSE) 1045 panic("unp_internalize: file already closed"); 1046 #endif 1047 *--rp = fp; 1048 fp->f_count++; 1049 fp->f_msgcount++; 1050 FILE_UNLOCK(fp); 1051 unp_rights++; 1052 } 1053 rw_exit(&fdescp->fd_lock); 1054 1055 if (newcm) { 1056 if (control->m_flags & M_EXT) 1057 MEXTREMOVE(control); 1058 MEXTADD(control, newcm, 1059 CMSG_SPACE(nfds * sizeof(struct file *)), 1060 M_MBUF, NULL, NULL); 1061 cm = newcm; 1062 } 1063 1064 /* adjust message & mbuf to note amount of space actually used. */ 1065 cm->cmsg_len = CMSG_LEN(nfds * sizeof(struct file *)); 1066 control->m_len = CMSG_SPACE(nfds * sizeof(struct file *)); 1067 1068 return (0); 1069 } 1070 1071 struct mbuf * 1072 unp_addsockcred(struct lwp *l, struct mbuf *control) 1073 { 1074 struct cmsghdr *cmp; 1075 struct sockcred *sc; 1076 struct mbuf *m, *n; 1077 int len, space, i; 1078 1079 len = CMSG_LEN(SOCKCREDSIZE(kauth_cred_ngroups(l->l_cred))); 1080 space = CMSG_SPACE(SOCKCREDSIZE(kauth_cred_ngroups(l->l_cred))); 1081 1082 m = m_get(M_WAIT, MT_CONTROL); 1083 if (space > MLEN) { 1084 if (space > MCLBYTES) 1085 MEXTMALLOC(m, space, M_WAITOK); 1086 else 1087 m_clget(m, M_WAIT); 1088 if ((m->m_flags & M_EXT) == 0) { 1089 m_free(m); 1090 return (control); 1091 } 1092 } 1093 1094 m->m_len = space; 1095 m->m_next = NULL; 1096 cmp = mtod(m, struct cmsghdr *); 1097 sc = (struct sockcred *)CMSG_DATA(cmp); 1098 cmp->cmsg_len = len; 1099 cmp->cmsg_level = SOL_SOCKET; 1100 cmp->cmsg_type = SCM_CREDS; 1101 sc->sc_uid = kauth_cred_getuid(l->l_cred); 1102 sc->sc_euid = kauth_cred_geteuid(l->l_cred); 1103 sc->sc_gid = kauth_cred_getgid(l->l_cred); 1104 sc->sc_egid = kauth_cred_getegid(l->l_cred); 1105 sc->sc_ngroups = kauth_cred_ngroups(l->l_cred); 1106 for (i = 0; i < sc->sc_ngroups; i++) 1107 sc->sc_groups[i] = kauth_cred_group(l->l_cred, i); 1108 1109 /* 1110 * If a control message already exists, append us to the end. 1111 */ 1112 if (control != NULL) { 1113 for (n = control; n->m_next != NULL; n = n->m_next) 1114 ; 1115 n->m_next = m; 1116 } else 1117 control = m; 1118 1119 return (control); 1120 } 1121 1122 int unp_defer, unp_gcing; 1123 extern struct domain unixdomain; 1124 1125 /* 1126 * Comment added long after the fact explaining what's going on here. 1127 * Do a mark-sweep GC of file descriptors on the system, to free up 1128 * any which are caught in flight to an about-to-be-closed socket. 1129 * 1130 * Traditional mark-sweep gc's start at the "root", and mark 1131 * everything reachable from the root (which, in our case would be the 1132 * process table). The mark bits are cleared during the sweep. 1133 * 1134 * XXX For some inexplicable reason (perhaps because the file 1135 * descriptor tables used to live in the u area which could be swapped 1136 * out and thus hard to reach), we do multiple scans over the set of 1137 * descriptors, using use *two* mark bits per object (DEFER and MARK). 1138 * Whenever we find a descriptor which references other descriptors, 1139 * the ones it references are marked with both bits, and we iterate 1140 * over the whole file table until there are no more DEFER bits set. 1141 * We also make an extra pass *before* the GC to clear the mark bits, 1142 * which could have been cleared at almost no cost during the previous 1143 * sweep. 1144 * 1145 * XXX MP: this needs to run with locks such that no other thread of 1146 * control can create or destroy references to file descriptors. it 1147 * may be necessary to defer the GC until later (when the locking 1148 * situation is more hospitable); it may be necessary to push this 1149 * into a separate thread. 1150 */ 1151 void 1152 unp_gc(void) 1153 { 1154 struct file *fp, *nextfp; 1155 struct socket *so, *so1; 1156 struct file **extra_ref, **fpp; 1157 int nunref, i; 1158 1159 if (unp_gcing) 1160 return; 1161 unp_gcing = 1; 1162 unp_defer = 0; 1163 1164 mutex_enter(&filelist_lock); 1165 1166 /* Clear mark bits */ 1167 LIST_FOREACH(fp, &filehead, f_list) 1168 fp->f_flag &= ~(FMARK|FDEFER); 1169 1170 /* 1171 * Iterate over the set of descriptors, marking ones believed 1172 * (based on refcount) to be referenced from a process, and 1173 * marking for rescan descriptors which are queued on a socket. 1174 */ 1175 do { 1176 LIST_FOREACH(fp, &filehead, f_list) { 1177 FILE_LOCK(fp); 1178 if (fp->f_flag & FDEFER) { 1179 fp->f_flag &= ~FDEFER; 1180 unp_defer--; 1181 #ifdef DIAGNOSTIC 1182 if (fp->f_count == 0) 1183 panic("unp_gc: deferred unreferenced socket"); 1184 #endif 1185 } else { 1186 if (fp->f_count == 0 || 1187 (fp->f_flag & FMARK) || 1188 fp->f_count == fp->f_msgcount) { 1189 FILE_UNLOCK(fp); 1190 continue; 1191 } 1192 } 1193 fp->f_flag |= FMARK; 1194 1195 if (fp->f_type != DTYPE_SOCKET || 1196 (so = (struct socket *)fp->f_data) == 0 || 1197 so->so_proto->pr_domain != &unixdomain || 1198 (so->so_proto->pr_flags&PR_RIGHTS) == 0) { 1199 FILE_UNLOCK(fp); 1200 continue; 1201 } 1202 #ifdef notdef 1203 if (so->so_rcv.sb_flags & SB_LOCK) { 1204 /* 1205 * This is problematical; it's not clear 1206 * we need to wait for the sockbuf to be 1207 * unlocked (on a uniprocessor, at least), 1208 * and it's also not clear what to do 1209 * if sbwait returns an error due to receipt 1210 * of a signal. If sbwait does return 1211 * an error, we'll go into an infinite 1212 * loop. Delete all of this for now. 1213 */ 1214 (void) sbwait(&so->so_rcv); 1215 goto restart; 1216 } 1217 #endif 1218 FILE_UNLOCK(fp); 1219 1220 unp_scan(so->so_rcv.sb_mb, unp_mark, 0); 1221 /* 1222 * mark descriptors referenced from sockets queued on the accept queue as well. 1223 */ 1224 if (so->so_options & SO_ACCEPTCONN) { 1225 TAILQ_FOREACH(so1, &so->so_q0, so_qe) { 1226 unp_scan(so1->so_rcv.sb_mb, unp_mark, 0); 1227 } 1228 TAILQ_FOREACH(so1, &so->so_q, so_qe) { 1229 unp_scan(so1->so_rcv.sb_mb, unp_mark, 0); 1230 } 1231 } 1232 } 1233 } while (unp_defer); 1234 1235 mutex_exit(&filelist_lock); 1236 1237 /* 1238 * Sweep pass. Find unmarked descriptors, and free them. 1239 * 1240 * We grab an extra reference to each of the file table entries 1241 * that are not otherwise accessible and then free the rights 1242 * that are stored in messages on them. 1243 * 1244 * The bug in the original code is a little tricky, so I'll describe 1245 * what's wrong with it here. 1246 * 1247 * It is incorrect to simply unp_discard each entry for f_msgcount 1248 * times -- consider the case of sockets A and B that contain 1249 * references to each other. On a last close of some other socket, 1250 * we trigger a gc since the number of outstanding rights (unp_rights) 1251 * is non-zero. If during the sweep phase the gc code un_discards, 1252 * we end up doing a (full) closef on the descriptor. A closef on A 1253 * results in the following chain. Closef calls soo_close, which 1254 * calls soclose. Soclose calls first (through the switch 1255 * uipc_usrreq) unp_detach, which re-invokes unp_gc. Unp_gc simply 1256 * returns because the previous instance had set unp_gcing, and 1257 * we return all the way back to soclose, which marks the socket 1258 * with SS_NOFDREF, and then calls sofree. Sofree calls sorflush 1259 * to free up the rights that are queued in messages on the socket A, 1260 * i.e., the reference on B. The sorflush calls via the dom_dispose 1261 * switch unp_dispose, which unp_scans with unp_discard. This second 1262 * instance of unp_discard just calls closef on B. 1263 * 1264 * Well, a similar chain occurs on B, resulting in a sorflush on B, 1265 * which results in another closef on A. Unfortunately, A is already 1266 * being closed, and the descriptor has already been marked with 1267 * SS_NOFDREF, and soclose panics at this point. 1268 * 1269 * Here, we first take an extra reference to each inaccessible 1270 * descriptor. Then, if the inaccessible descriptor is a 1271 * socket, we call sorflush in case it is a Unix domain 1272 * socket. After we destroy all the rights carried in 1273 * messages, we do a last closef to get rid of our extra 1274 * reference. This is the last close, and the unp_detach etc 1275 * will shut down the socket. 1276 * 1277 * 91/09/19, bsy@cs.cmu.edu 1278 */ 1279 extra_ref = kmem_alloc(nfiles * sizeof(struct file *), KM_SLEEP); 1280 1281 mutex_enter(&filelist_lock); 1282 for (nunref = 0, fp = LIST_FIRST(&filehead), fpp = extra_ref; fp != 0; 1283 fp = nextfp) { 1284 nextfp = LIST_NEXT(fp, f_list); 1285 FILE_LOCK(fp); 1286 if (fp->f_count != 0 && 1287 fp->f_count == fp->f_msgcount && !(fp->f_flag & FMARK)) { 1288 *fpp++ = fp; 1289 nunref++; 1290 fp->f_count++; 1291 } 1292 FILE_UNLOCK(fp); 1293 } 1294 mutex_exit(&filelist_lock); 1295 1296 for (i = nunref, fpp = extra_ref; --i >= 0; ++fpp) { 1297 fp = *fpp; 1298 FILE_LOCK(fp); 1299 FILE_USE(fp); 1300 if (fp->f_type == DTYPE_SOCKET) 1301 sorflush((struct socket *)fp->f_data); 1302 FILE_UNUSE(fp, NULL); 1303 } 1304 for (i = nunref, fpp = extra_ref; --i >= 0; ++fpp) { 1305 fp = *fpp; 1306 FILE_LOCK(fp); 1307 FILE_USE(fp); 1308 (void) closef(fp, (struct lwp *)0); 1309 } 1310 kmem_free(extra_ref, nfiles * sizeof(struct file *)); 1311 unp_gcing = 0; 1312 } 1313 1314 void 1315 unp_dispose(struct mbuf *m) 1316 { 1317 1318 if (m) 1319 unp_scan(m, unp_discard, 1); 1320 } 1321 1322 void 1323 unp_scan(struct mbuf *m0, void (*op)(struct file *), int discard) 1324 { 1325 struct mbuf *m; 1326 struct file **rp; 1327 struct cmsghdr *cm; 1328 int i; 1329 int qfds; 1330 1331 while (m0) { 1332 for (m = m0; m; m = m->m_next) { 1333 if (m->m_type == MT_CONTROL && 1334 m->m_len >= sizeof(*cm)) { 1335 cm = mtod(m, struct cmsghdr *); 1336 if (cm->cmsg_level != SOL_SOCKET || 1337 cm->cmsg_type != SCM_RIGHTS) 1338 continue; 1339 qfds = (cm->cmsg_len - CMSG_ALIGN(sizeof(*cm))) 1340 / sizeof(struct file *); 1341 rp = (struct file **)CMSG_DATA(cm); 1342 for (i = 0; i < qfds; i++) { 1343 struct file *fp = *rp; 1344 if (discard) 1345 *rp = 0; 1346 (*op)(fp); 1347 rp++; 1348 } 1349 break; /* XXX, but saves time */ 1350 } 1351 } 1352 m0 = m0->m_nextpkt; 1353 } 1354 } 1355 1356 void 1357 unp_mark(struct file *fp) 1358 { 1359 1360 if (fp == NULL) 1361 return; 1362 1363 /* If we're already deferred, don't screw up the defer count */ 1364 FILE_LOCK(fp); 1365 if (fp->f_flag & (FMARK | FDEFER)) { 1366 FILE_UNLOCK(fp); 1367 return; 1368 } 1369 1370 /* 1371 * Minimize the number of deferrals... Sockets are the only 1372 * type of descriptor which can hold references to another 1373 * descriptor, so just mark other descriptors, and defer 1374 * unmarked sockets for the next pass. 1375 */ 1376 if (fp->f_type == DTYPE_SOCKET) { 1377 unp_defer++; 1378 if (fp->f_count == 0) 1379 panic("unp_mark: queued unref"); 1380 fp->f_flag |= FDEFER; 1381 } else { 1382 fp->f_flag |= FMARK; 1383 } 1384 FILE_UNLOCK(fp); 1385 return; 1386 } 1387 1388 void 1389 unp_discard(struct file *fp) 1390 { 1391 if (fp == NULL) 1392 return; 1393 FILE_LOCK(fp); 1394 fp->f_usecount++; /* i.e. FILE_USE(fp) sans locking */ 1395 fp->f_msgcount--; 1396 FILE_UNLOCK(fp); 1397 unp_rights--; 1398 (void) closef(fp, (struct lwp *)0); 1399 } 1400