1 /* $NetBSD: linux_socket.c,v 1.115 2013/01/11 19:01:36 christos Exp $ */ 2 3 /*- 4 * Copyright (c) 1995, 1998, 2008 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Frank van der Linden and Eric Haszlakiewicz. 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 * Functions in multiarch: 34 * linux_sys_socketcall : linux_socketcall.c 35 */ 36 37 #include <sys/cdefs.h> 38 __KERNEL_RCSID(0, "$NetBSD: linux_socket.c,v 1.115 2013/01/11 19:01:36 christos Exp $"); 39 40 #if defined(_KERNEL_OPT) 41 #include "opt_inet.h" 42 #endif /* defined(_KERNEL_OPT) */ 43 44 #include <sys/param.h> 45 #include <sys/kernel.h> 46 #include <sys/systm.h> 47 #include <sys/buf.h> 48 #include <sys/malloc.h> 49 #include <sys/ioctl.h> 50 #include <sys/tty.h> 51 #include <sys/file.h> 52 #include <sys/filedesc.h> 53 #include <sys/select.h> 54 #include <sys/socket.h> 55 #include <sys/socketvar.h> 56 #include <sys/domain.h> 57 #include <net/if.h> 58 #include <net/if_dl.h> 59 #include <net/if_types.h> 60 #include <netinet/in.h> 61 #include <netinet/tcp.h> 62 #include <sys/mount.h> 63 #include <sys/proc.h> 64 #include <sys/vnode.h> 65 #include <sys/device.h> 66 #include <sys/protosw.h> 67 #include <sys/mbuf.h> 68 #include <sys/syslog.h> 69 #include <sys/exec.h> 70 #include <sys/kauth.h> 71 #include <sys/syscallargs.h> 72 #include <sys/ktrace.h> 73 #include <sys/fcntl.h> 74 75 #include <lib/libkern/libkern.h> 76 77 #include <netinet/ip6.h> 78 #include <netinet6/ip6_var.h> 79 80 #include <compat/sys/socket.h> 81 #include <compat/sys/sockio.h> 82 83 #include <compat/linux/common/linux_types.h> 84 #include <compat/linux/common/linux_util.h> 85 #include <compat/linux/common/linux_signal.h> 86 #include <compat/linux/common/linux_ioctl.h> 87 #include <compat/linux/common/linux_socket.h> 88 #include <compat/linux/common/linux_fcntl.h> 89 #if !defined(__alpha__) && !defined(__amd64__) 90 #include <compat/linux/common/linux_socketcall.h> 91 #endif 92 #include <compat/linux/common/linux_sockio.h> 93 #include <compat/linux/common/linux_ipc.h> 94 #include <compat/linux/common/linux_sem.h> 95 96 #include <compat/linux/linux_syscallargs.h> 97 98 #ifdef DEBUG_LINUX 99 #define DPRINTF(a) uprintf a 100 #else 101 #define DPRINTF(a) 102 #endif 103 104 /* 105 * The calls in this file are entered either via the linux_socketcall() 106 * interface or, on the Alpha, as individual syscalls. The 107 * linux_socketcall function does any massaging of arguments so that all 108 * the calls in here need not think that they are anything other 109 * than a normal syscall. 110 */ 111 112 static int linux_to_bsd_domain(int); 113 static int bsd_to_linux_domain(int); 114 int linux_to_bsd_sopt_level(int); 115 int linux_to_bsd_so_sockopt(int); 116 int linux_to_bsd_ip_sockopt(int); 117 int linux_to_bsd_tcp_sockopt(int); 118 int linux_to_bsd_udp_sockopt(int); 119 int linux_getifname(struct lwp *, register_t *, void *); 120 int linux_getifconf(struct lwp *, register_t *, void *); 121 int linux_getifhwaddr(struct lwp *, register_t *, u_int, void *); 122 static int linux_get_sa(struct lwp *, int, struct mbuf **, 123 const struct osockaddr *, unsigned int); 124 static int linux_sa_put(struct osockaddr *osa); 125 static int linux_to_bsd_msg_flags(int); 126 static int bsd_to_linux_msg_flags(int); 127 static void linux_to_bsd_msghdr(struct linux_msghdr *, struct msghdr *); 128 static void bsd_to_linux_msghdr(struct msghdr *, struct linux_msghdr *); 129 130 static const int linux_to_bsd_domain_[LINUX_AF_MAX] = { 131 AF_UNSPEC, 132 AF_UNIX, 133 AF_INET, 134 AF_CCITT, /* LINUX_AF_AX25 */ 135 AF_IPX, 136 AF_APPLETALK, 137 -1, /* LINUX_AF_NETROM */ 138 -1, /* LINUX_AF_BRIDGE */ 139 -1, /* LINUX_AF_ATMPVC */ 140 AF_CCITT, /* LINUX_AF_X25 */ 141 AF_INET6, 142 -1, /* LINUX_AF_ROSE */ 143 AF_DECnet, 144 -1, /* LINUX_AF_NETBEUI */ 145 -1, /* LINUX_AF_SECURITY */ 146 pseudo_AF_KEY, 147 AF_ROUTE, /* LINUX_AF_NETLINK */ 148 -1, /* LINUX_AF_PACKET */ 149 -1, /* LINUX_AF_ASH */ 150 -1, /* LINUX_AF_ECONET */ 151 -1, /* LINUX_AF_ATMSVC */ 152 AF_SNA, 153 /* rest up to LINUX_AF_MAX-1 is not allocated */ 154 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 155 }; 156 157 static const int bsd_to_linux_domain_[AF_MAX] = { 158 LINUX_AF_UNSPEC, 159 LINUX_AF_UNIX, 160 LINUX_AF_INET, 161 -1, /* AF_IMPLINK */ 162 -1, /* AF_PUP */ 163 -1, /* AF_CHAOS */ 164 -1, /* AF_NS */ 165 -1, /* AF_ISO */ 166 -1, /* AF_ECMA */ 167 -1, /* AF_DATAKIT */ 168 LINUX_AF_AX25, /* AF_CCITT */ 169 LINUX_AF_SNA, 170 LINUX_AF_DECnet, 171 -1, /* AF_DLI */ 172 -1, /* AF_LAT */ 173 -1, /* AF_HYLINK */ 174 LINUX_AF_APPLETALK, 175 LINUX_AF_NETLINK, 176 -1, /* AF_LINK */ 177 -1, /* AF_XTP */ 178 -1, /* AF_COIP */ 179 -1, /* AF_CNT */ 180 -1, /* pseudo_AF_RTIP */ 181 LINUX_AF_IPX, 182 LINUX_AF_INET6, 183 -1, /* pseudo_AF_PIP */ 184 -1, /* AF_ISDN */ 185 -1, /* AF_NATM */ 186 -1, /* AF_ARP */ 187 LINUX_pseudo_AF_KEY, 188 -1, /* pseudo_AF_HDRCMPLT */ 189 }; 190 191 static const struct { 192 int bfl; 193 int lfl; 194 } bsd_to_linux_msg_flags_[] = { 195 {MSG_OOB, LINUX_MSG_OOB}, 196 {MSG_PEEK, LINUX_MSG_PEEK}, 197 {MSG_DONTROUTE, LINUX_MSG_DONTROUTE}, 198 {MSG_EOR, LINUX_MSG_EOR}, 199 {MSG_TRUNC, LINUX_MSG_TRUNC}, 200 {MSG_CTRUNC, LINUX_MSG_CTRUNC}, 201 {MSG_WAITALL, LINUX_MSG_WAITALL}, 202 {MSG_DONTWAIT, LINUX_MSG_DONTWAIT}, 203 {MSG_BCAST, 0}, /* not supported, clear */ 204 {MSG_MCAST, 0}, /* not supported, clear */ 205 {MSG_NOSIGNAL, LINUX_MSG_NOSIGNAL}, 206 {-1, /* not supp */ LINUX_MSG_PROBE}, 207 {-1, /* not supp */ LINUX_MSG_FIN}, 208 {-1, /* not supp */ LINUX_MSG_SYN}, 209 {-1, /* not supp */ LINUX_MSG_CONFIRM}, 210 {-1, /* not supp */ LINUX_MSG_RST}, 211 {-1, /* not supp */ LINUX_MSG_ERRQUEUE}, 212 {-1, /* not supp */ LINUX_MSG_MORE}, 213 }; 214 215 /* 216 * Convert between Linux and BSD socket domain values 217 */ 218 static int 219 linux_to_bsd_domain(int ldom) 220 { 221 if (ldom < 0 || ldom >= LINUX_AF_MAX) 222 return (-1); 223 224 return linux_to_bsd_domain_[ldom]; 225 } 226 227 /* 228 * Convert between BSD and Linux socket domain values 229 */ 230 static int 231 bsd_to_linux_domain(int bdom) 232 { 233 if (bdom < 0 || bdom >= AF_MAX) 234 return (-1); 235 236 return bsd_to_linux_domain_[bdom]; 237 } 238 239 static int 240 linux_to_bsd_msg_flags(int lflag) 241 { 242 int i, lfl, bfl; 243 int bflag = 0; 244 245 if (lflag == 0) 246 return (0); 247 248 for(i = 0; i < __arraycount(bsd_to_linux_msg_flags_); i++) { 249 bfl = bsd_to_linux_msg_flags_[i].bfl; 250 lfl = bsd_to_linux_msg_flags_[i].lfl; 251 252 if (lfl == 0) 253 continue; 254 255 if (lflag & lfl) { 256 if (bfl < 0) 257 return (-1); 258 259 bflag |= bfl; 260 } 261 } 262 263 return (bflag); 264 } 265 266 static int 267 bsd_to_linux_msg_flags(int bflag) 268 { 269 int i, lfl, bfl; 270 int lflag = 0; 271 272 if (bflag == 0) 273 return (0); 274 275 for(i = 0; i < __arraycount(bsd_to_linux_msg_flags_); i++) { 276 bfl = bsd_to_linux_msg_flags_[i].bfl; 277 lfl = bsd_to_linux_msg_flags_[i].lfl; 278 279 if (bfl <= 0) 280 continue; 281 282 if (bflag & bfl) { 283 if (lfl < 0) 284 return (-1); 285 286 lflag |= lfl; 287 } 288 } 289 290 return (lflag); 291 } 292 293 int 294 linux_sys_socket(struct lwp *l, const struct linux_sys_socket_args *uap, register_t *retval) 295 { 296 /* { 297 syscallarg(int) domain; 298 syscallarg(int) type; 299 syscallarg(int) protocol; 300 } */ 301 struct sys___socket30_args bsa; 302 struct sys_fcntl_args fsa; 303 register_t fretval[2]; 304 int error, flags; 305 306 307 SCARG(&bsa, protocol) = SCARG(uap, protocol); 308 SCARG(&bsa, type) = SCARG(uap, type) & LINUX_SOCK_TYPE_MASK; 309 SCARG(&bsa, domain) = linux_to_bsd_domain(SCARG(uap, domain)); 310 if (SCARG(&bsa, domain) == -1) 311 return EINVAL; 312 /* 313 * Apparently linux uses this to talk to ISDN sockets. If we fail 314 * now programs seems to handle it, but if we don't we are going 315 * to fail when we bind and programs don't handle this well. 316 */ 317 if (SCARG(&bsa, domain) == AF_ROUTE && SCARG(&bsa, type) == SOCK_RAW) 318 return ENOTSUP; 319 flags = SCARG(uap, type) & ~LINUX_SOCK_TYPE_MASK; 320 if (flags & ~(LINUX_SOCK_CLOEXEC | LINUX_SOCK_NONBLOCK)) 321 return EINVAL; 322 error = sys___socket30(l, &bsa, retval); 323 324 /* 325 * Linux overloads the "type" parameter to include some 326 * fcntl flags to be set on the file descriptor. 327 * Process those if creating the socket succeeded. 328 */ 329 330 if (!error && flags & LINUX_SOCK_CLOEXEC) { 331 SCARG(&fsa, fd) = *retval; 332 SCARG(&fsa, cmd) = F_SETFD; 333 SCARG(&fsa, arg) = (void *)(uintptr_t)FD_CLOEXEC; 334 (void) sys_fcntl(l, &fsa, fretval); 335 } 336 if (!error && flags & LINUX_SOCK_NONBLOCK) { 337 SCARG(&fsa, fd) = *retval; 338 SCARG(&fsa, cmd) = F_SETFL; 339 SCARG(&fsa, arg) = (void *)(uintptr_t)O_NONBLOCK; 340 error = sys_fcntl(l, &fsa, fretval); 341 if (error) { 342 struct sys_close_args csa; 343 344 SCARG(&csa, fd) = *retval; 345 (void) sys_close(l, &csa, fretval); 346 } 347 } 348 349 #ifdef INET6 350 /* 351 * Linux AF_INET6 socket has IPV6_V6ONLY setsockopt set to 0 by 352 * default and some apps depend on this. So, set V6ONLY to 0 353 * for Linux apps if the sysctl value is set to 1. 354 */ 355 if (!error && ip6_v6only && SCARG(&bsa, domain) == PF_INET6) { 356 struct socket *so; 357 358 if (fd_getsock(*retval, &so) == 0) { 359 int val = 0; 360 361 /* ignore error */ 362 (void)so_setsockopt(l, so, IPPROTO_IPV6, IPV6_V6ONLY, 363 &val, sizeof(val)); 364 365 fd_putfile(*retval); 366 } 367 } 368 #endif 369 370 return (error); 371 } 372 373 int 374 linux_sys_socketpair(struct lwp *l, const struct linux_sys_socketpair_args *uap, register_t *retval) 375 { 376 /* { 377 syscallarg(int) domain; 378 syscallarg(int) type; 379 syscallarg(int) protocol; 380 syscallarg(int *) rsv; 381 } */ 382 struct sys_socketpair_args bsa; 383 384 SCARG(&bsa, domain) = linux_to_bsd_domain(SCARG(uap, domain)); 385 if (SCARG(&bsa, domain) == -1) 386 return EINVAL; 387 SCARG(&bsa, type) = SCARG(uap, type); 388 SCARG(&bsa, protocol) = SCARG(uap, protocol); 389 SCARG(&bsa, rsv) = SCARG(uap, rsv); 390 391 return sys_socketpair(l, &bsa, retval); 392 } 393 394 int 395 linux_sys_sendto(struct lwp *l, const struct linux_sys_sendto_args *uap, register_t *retval) 396 { 397 /* { 398 syscallarg(int) s; 399 syscallarg(void *) msg; 400 syscallarg(int) len; 401 syscallarg(int) flags; 402 syscallarg(struct osockaddr *) to; 403 syscallarg(int) tolen; 404 } */ 405 struct msghdr msg; 406 struct iovec aiov; 407 struct mbuf *nam; 408 int bflags; 409 int error; 410 411 /* Translate message flags. */ 412 bflags = linux_to_bsd_msg_flags(SCARG(uap, flags)); 413 if (bflags < 0) 414 /* Some supported flag */ 415 return EINVAL; 416 417 msg.msg_flags = 0; 418 msg.msg_name = NULL; 419 msg.msg_control = NULL; 420 421 if (SCARG(uap, tolen)) { 422 /* Read in and convert the sockaddr */ 423 error = linux_get_sa(l, SCARG(uap, s), &nam, SCARG(uap, to), 424 SCARG(uap, tolen)); 425 if (error) 426 return (error); 427 msg.msg_flags |= MSG_NAMEMBUF; 428 msg.msg_name = nam; 429 msg.msg_namelen = SCARG(uap, tolen); 430 } 431 432 msg.msg_iov = &aiov; 433 msg.msg_iovlen = 1; 434 aiov.iov_base = __UNCONST(SCARG(uap, msg)); 435 aiov.iov_len = SCARG(uap, len); 436 437 return do_sys_sendmsg(l, SCARG(uap, s), &msg, bflags, retval); 438 } 439 440 static void 441 linux_to_bsd_msghdr(struct linux_msghdr *lmsg, struct msghdr *bmsg) 442 { 443 bmsg->msg_name = lmsg->msg_name; 444 bmsg->msg_namelen = lmsg->msg_namelen; 445 bmsg->msg_iov = lmsg->msg_iov; 446 bmsg->msg_iovlen = lmsg->msg_iovlen; 447 bmsg->msg_control = lmsg->msg_control; 448 bmsg->msg_controllen = lmsg->msg_controllen; 449 bmsg->msg_flags = lmsg->msg_flags; 450 } 451 452 static void 453 bsd_to_linux_msghdr(struct msghdr *bmsg, struct linux_msghdr *lmsg) 454 { 455 lmsg->msg_name = bmsg->msg_name; 456 lmsg->msg_namelen = bmsg->msg_namelen; 457 lmsg->msg_iov = bmsg->msg_iov; 458 lmsg->msg_iovlen = bmsg->msg_iovlen; 459 lmsg->msg_control = bmsg->msg_control; 460 lmsg->msg_controllen = bmsg->msg_controllen; 461 lmsg->msg_flags = bmsg->msg_flags; 462 } 463 464 int 465 linux_sys_sendmsg(struct lwp *l, const struct linux_sys_sendmsg_args *uap, register_t *retval) 466 { 467 /* { 468 syscallarg(int) s; 469 syscallarg(struct linux_msghdr *) msg; 470 syscallarg(u_int) flags; 471 } */ 472 struct msghdr msg; 473 struct linux_msghdr lmsg; 474 int error; 475 int bflags; 476 struct mbuf *nam; 477 u_int8_t *control; 478 struct mbuf *ctl_mbuf = NULL; 479 480 error = copyin(SCARG(uap, msg), &lmsg, sizeof(lmsg)); 481 if (error) 482 return error; 483 linux_to_bsd_msghdr(&lmsg, &msg); 484 485 msg.msg_flags = MSG_IOVUSRSPACE; 486 487 /* 488 * Translate message flags. 489 */ 490 bflags = linux_to_bsd_msg_flags(SCARG(uap, flags)); 491 if (bflags < 0) 492 /* Some supported flag */ 493 return EINVAL; 494 495 if (lmsg.msg_name) { 496 /* Read in and convert the sockaddr */ 497 error = linux_get_sa(l, SCARG(uap, s), &nam, msg.msg_name, 498 msg.msg_namelen); 499 if (error) 500 return (error); 501 msg.msg_flags |= MSG_NAMEMBUF; 502 msg.msg_name = nam; 503 } 504 505 /* 506 * Handle cmsg if there is any. 507 */ 508 if (LINUX_CMSG_FIRSTHDR(&lmsg)) { 509 struct linux_cmsghdr l_cmsg, *l_cc; 510 struct cmsghdr *cmsg; 511 ssize_t resid = msg.msg_controllen; 512 size_t clen, cidx = 0, cspace; 513 514 ctl_mbuf = m_get(M_WAIT, MT_CONTROL); 515 clen = MLEN; 516 control = mtod(ctl_mbuf, void *); 517 518 l_cc = LINUX_CMSG_FIRSTHDR(&lmsg); 519 do { 520 error = copyin(l_cc, &l_cmsg, sizeof(l_cmsg)); 521 if (error) 522 goto done; 523 524 /* 525 * Sanity check the control message length. 526 */ 527 if (l_cmsg.cmsg_len > resid 528 || l_cmsg.cmsg_len < sizeof l_cmsg) { 529 error = EINVAL; 530 goto done; 531 } 532 533 /* 534 * Refuse unsupported control messages, and 535 * translate fields as appropriate. 536 */ 537 switch (l_cmsg.cmsg_level) { 538 case LINUX_SOL_SOCKET: 539 /* It only differs on some archs */ 540 if (LINUX_SOL_SOCKET != SOL_SOCKET) 541 l_cmsg.cmsg_level = SOL_SOCKET; 542 543 switch(l_cmsg.cmsg_type) { 544 case LINUX_SCM_RIGHTS: 545 /* Linux SCM_RIGHTS is same as NetBSD */ 546 break; 547 548 case LINUX_SCM_CREDENTIALS: 549 /* no native equivalent, just drop it */ 550 m_free(ctl_mbuf); 551 ctl_mbuf = NULL; 552 msg.msg_control = NULL; 553 msg.msg_controllen = 0; 554 goto skipcmsg; 555 556 default: 557 /* other types not supported */ 558 error = EINVAL; 559 goto done; 560 } 561 break; 562 default: 563 /* pray and leave intact */ 564 break; 565 } 566 567 cspace = CMSG_SPACE(l_cmsg.cmsg_len - sizeof(l_cmsg)); 568 569 /* Check the buffer is big enough */ 570 if (__predict_false(cidx + cspace > clen)) { 571 u_int8_t *nc; 572 573 clen = cidx + cspace; 574 if (clen >= PAGE_SIZE) { 575 error = EINVAL; 576 goto done; 577 } 578 nc = realloc(clen <= MLEN ? NULL : control, 579 clen, M_TEMP, M_WAITOK); 580 if (!nc) { 581 error = ENOMEM; 582 goto done; 583 } 584 if (cidx <= MLEN) 585 /* Old buffer was in mbuf... */ 586 memcpy(nc, control, cidx); 587 control = nc; 588 } 589 590 /* Copy header */ 591 cmsg = (void *)&control[cidx]; 592 cmsg->cmsg_len = l_cmsg.cmsg_len + LINUX_CMSG_ALIGN_DELTA; 593 cmsg->cmsg_level = l_cmsg.cmsg_level; 594 cmsg->cmsg_type = l_cmsg.cmsg_type; 595 596 /* Zero area between header and data */ 597 memset(cmsg + 1, 0, 598 CMSG_ALIGN(sizeof(*cmsg)) - sizeof(*cmsg)); 599 600 /* Copyin the data */ 601 error = copyin(LINUX_CMSG_DATA(l_cc), 602 CMSG_DATA(cmsg), 603 l_cmsg.cmsg_len - sizeof(l_cmsg)); 604 if (error) 605 goto done; 606 607 resid -= LINUX_CMSG_ALIGN(l_cmsg.cmsg_len); 608 cidx += cspace; 609 } while ((l_cc = LINUX_CMSG_NXTHDR(&msg, l_cc)) && resid > 0); 610 611 /* If we allocated a buffer, attach to mbuf */ 612 if (cidx > MLEN) { 613 MEXTADD(ctl_mbuf, control, clen, M_MBUF, NULL, NULL); 614 ctl_mbuf->m_flags |= M_EXT_RW; 615 } 616 control = NULL; 617 ctl_mbuf->m_len = cidx; 618 619 msg.msg_control = ctl_mbuf; 620 msg.msg_flags |= MSG_CONTROLMBUF; 621 622 ktrkuser("mbcontrol", mtod(ctl_mbuf, void *), 623 msg.msg_controllen); 624 } 625 626 skipcmsg: 627 error = do_sys_sendmsg(l, SCARG(uap, s), &msg, bflags, retval); 628 /* Freed internally */ 629 ctl_mbuf = NULL; 630 631 done: 632 if (ctl_mbuf != NULL) { 633 if (control != NULL && control != mtod(ctl_mbuf, void *)) 634 free(control, M_MBUF); 635 m_free(ctl_mbuf); 636 } 637 return (error); 638 } 639 640 int 641 linux_sys_recvfrom(struct lwp *l, const struct linux_sys_recvfrom_args *uap, register_t *retval) 642 { 643 /* { 644 syscallarg(int) s; 645 syscallarg(void *) buf; 646 syscallarg(int) len; 647 syscallarg(int) flags; 648 syscallarg(struct osockaddr *) from; 649 syscallarg(int *) fromlenaddr; 650 } */ 651 int error; 652 struct sys_recvfrom_args bra; 653 654 SCARG(&bra, s) = SCARG(uap, s); 655 SCARG(&bra, buf) = SCARG(uap, buf); 656 SCARG(&bra, len) = SCARG(uap, len); 657 SCARG(&bra, flags) = SCARG(uap, flags); 658 SCARG(&bra, from) = (struct sockaddr *) SCARG(uap, from); 659 SCARG(&bra, fromlenaddr) = (socklen_t *)SCARG(uap, fromlenaddr); 660 661 if ((error = sys_recvfrom(l, &bra, retval))) 662 return (error); 663 664 if (SCARG(uap, from) && (error = linux_sa_put(SCARG(uap, from)))) 665 return (error); 666 667 return (0); 668 } 669 670 static int 671 linux_copyout_msg_control(struct lwp *l, struct msghdr *mp, struct mbuf *control) 672 { 673 int dlen, error = 0; 674 struct cmsghdr *cmsg; 675 struct linux_cmsghdr linux_cmsg; 676 struct mbuf *m; 677 char *q, *q_end; 678 679 if (mp->msg_controllen <= 0 || control == 0) { 680 mp->msg_controllen = 0; 681 free_control_mbuf(l, control, control); 682 return 0; 683 } 684 685 ktrkuser("msgcontrol", mtod(control, void *), mp->msg_controllen); 686 687 q = (char *)mp->msg_control; 688 q_end = q + mp->msg_controllen; 689 690 for (m = control; m != NULL; ) { 691 cmsg = mtod(m, struct cmsghdr *); 692 693 /* 694 * Fixup cmsg. We handle two things: 695 * 0. different sizeof cmsg_len. 696 * 1. different values for level/type on some archs 697 * 2. different alignment of CMSG_DATA on some archs 698 */ 699 linux_cmsg.cmsg_len = cmsg->cmsg_len - LINUX_CMSG_ALIGN_DELTA; 700 linux_cmsg.cmsg_level = cmsg->cmsg_level; 701 linux_cmsg.cmsg_type = cmsg->cmsg_type; 702 703 dlen = q_end - q; 704 if (linux_cmsg.cmsg_len > dlen) { 705 /* Not enough room for the parameter */ 706 dlen -= sizeof linux_cmsg; 707 if (dlen <= 0) 708 /* Discard if header wont fit */ 709 break; 710 mp->msg_flags |= MSG_CTRUNC; 711 if (linux_cmsg.cmsg_level == SOL_SOCKET 712 && linux_cmsg.cmsg_type == SCM_RIGHTS) 713 /* Do not truncate me ... */ 714 break; 715 } else 716 dlen = linux_cmsg.cmsg_len - sizeof linux_cmsg; 717 718 switch (linux_cmsg.cmsg_level) { 719 case SOL_SOCKET: 720 linux_cmsg.cmsg_level = LINUX_SOL_SOCKET; 721 switch (linux_cmsg.cmsg_type) { 722 case SCM_RIGHTS: 723 /* Linux SCM_RIGHTS is same as NetBSD */ 724 break; 725 726 default: 727 /* other types not supported */ 728 error = EINVAL; 729 goto done; 730 } 731 /* machine dependent ! */ 732 break; 733 default: 734 /* pray and leave intact */ 735 break; 736 } 737 738 /* There can be padding between the header and data... */ 739 error = copyout(&linux_cmsg, q, sizeof linux_cmsg); 740 if (error != 0) { 741 error = copyout(CCMSG_DATA(cmsg), q + sizeof linux_cmsg, 742 dlen); 743 } 744 if (error != 0) { 745 /* We must free all the SCM_RIGHTS */ 746 m = control; 747 break; 748 } 749 m = m->m_next; 750 if (m == NULL || q + LINUX_CMSG_SPACE(dlen) > q_end) { 751 q += LINUX_CMSG_LEN(dlen); 752 break; 753 } 754 q += LINUX_CMSG_SPACE(dlen); 755 } 756 757 done: 758 free_control_mbuf(l, control, m); 759 760 mp->msg_controllen = q - (char *)mp->msg_control; 761 return error; 762 } 763 764 int 765 linux_sys_recvmsg(struct lwp *l, const struct linux_sys_recvmsg_args *uap, register_t *retval) 766 { 767 /* { 768 syscallarg(int) s; 769 syscallarg(struct linux_msghdr *) msg; 770 syscallarg(u_int) flags; 771 } */ 772 struct msghdr msg; 773 struct linux_msghdr lmsg; 774 int error; 775 struct mbuf *from, *control; 776 777 error = copyin(SCARG(uap, msg), &lmsg, sizeof(lmsg)); 778 if (error) 779 return (error); 780 linux_to_bsd_msghdr(&lmsg, &msg); 781 782 msg.msg_flags = linux_to_bsd_msg_flags(SCARG(uap, flags)); 783 if (msg.msg_flags < 0) { 784 /* Some unsupported flag */ 785 return (EINVAL); 786 } 787 msg.msg_flags |= MSG_IOVUSRSPACE; 788 789 error = do_sys_recvmsg(l, SCARG(uap, s), &msg, &from, 790 msg.msg_control != NULL ? &control : NULL, retval); 791 if (error != 0) 792 return error; 793 794 if (msg.msg_control != NULL) 795 error = linux_copyout_msg_control(l, &msg, control); 796 797 if (error == 0 && from != 0) { 798 mtod(from, struct osockaddr *)->sa_family = 799 bsd_to_linux_domain(mtod(from, struct sockaddr *)->sa_family); 800 error = copyout_sockname(msg.msg_name, &msg.msg_namelen, 0, 801 from); 802 } else 803 msg.msg_namelen = 0; 804 805 if (from != NULL) 806 m_free(from); 807 808 if (error == 0) { 809 msg.msg_flags = bsd_to_linux_msg_flags(msg.msg_flags); 810 if (msg.msg_flags < 0) 811 /* Some flag unsupported by Linux */ 812 error = EINVAL; 813 else { 814 ktrkuser("msghdr", &msg, sizeof(msg)); 815 bsd_to_linux_msghdr(&msg, &lmsg); 816 error = copyout(&lmsg, SCARG(uap, msg), sizeof(lmsg)); 817 } 818 } 819 820 return (error); 821 } 822 823 /* 824 * Convert socket option level from Linux to NetBSD value. Only SOL_SOCKET 825 * is different, the rest matches IPPROTO_* on both systems. 826 */ 827 int 828 linux_to_bsd_sopt_level(int llevel) 829 { 830 831 switch (llevel) { 832 case LINUX_SOL_SOCKET: 833 return SOL_SOCKET; 834 case LINUX_SOL_IP: 835 return IPPROTO_IP; 836 case LINUX_SOL_TCP: 837 return IPPROTO_TCP; 838 case LINUX_SOL_UDP: 839 return IPPROTO_UDP; 840 default: 841 return -1; 842 } 843 } 844 845 /* 846 * Convert Linux socket level socket option numbers to NetBSD values. 847 */ 848 int 849 linux_to_bsd_so_sockopt(int lopt) 850 { 851 852 switch (lopt) { 853 case LINUX_SO_DEBUG: 854 return SO_DEBUG; 855 case LINUX_SO_REUSEADDR: 856 /* 857 * Linux does not implement SO_REUSEPORT, but allows reuse of a 858 * host:port pair through SO_REUSEADDR even if the address is not a 859 * multicast-address. Effectively, this means that we should use 860 * SO_REUSEPORT to allow Linux applications to not exit with 861 * EADDRINUSE 862 */ 863 return SO_REUSEPORT; 864 case LINUX_SO_TYPE: 865 return SO_TYPE; 866 case LINUX_SO_ERROR: 867 return SO_ERROR; 868 case LINUX_SO_DONTROUTE: 869 return SO_DONTROUTE; 870 case LINUX_SO_BROADCAST: 871 return SO_BROADCAST; 872 case LINUX_SO_SNDBUF: 873 return SO_SNDBUF; 874 case LINUX_SO_RCVBUF: 875 return SO_RCVBUF; 876 case LINUX_SO_KEEPALIVE: 877 return SO_KEEPALIVE; 878 case LINUX_SO_OOBINLINE: 879 return SO_OOBINLINE; 880 case LINUX_SO_LINGER: 881 return SO_LINGER; 882 case LINUX_SO_PRIORITY: 883 case LINUX_SO_NO_CHECK: 884 default: 885 return -1; 886 } 887 } 888 889 /* 890 * Convert Linux IP level socket option number to NetBSD values. 891 */ 892 int 893 linux_to_bsd_ip_sockopt(int lopt) 894 { 895 896 switch (lopt) { 897 case LINUX_IP_TOS: 898 return IP_TOS; 899 case LINUX_IP_TTL: 900 return IP_TTL; 901 case LINUX_IP_HDRINCL: 902 return IP_HDRINCL; 903 case LINUX_IP_MULTICAST_TTL: 904 return IP_MULTICAST_TTL; 905 case LINUX_IP_MULTICAST_LOOP: 906 return IP_MULTICAST_LOOP; 907 case LINUX_IP_MULTICAST_IF: 908 return IP_MULTICAST_IF; 909 case LINUX_IP_ADD_MEMBERSHIP: 910 return IP_ADD_MEMBERSHIP; 911 case LINUX_IP_DROP_MEMBERSHIP: 912 return IP_DROP_MEMBERSHIP; 913 default: 914 return -1; 915 } 916 } 917 918 /* 919 * Convert Linux TCP level socket option number to NetBSD values. 920 */ 921 int 922 linux_to_bsd_tcp_sockopt(int lopt) 923 { 924 925 switch (lopt) { 926 case LINUX_TCP_NODELAY: 927 return TCP_NODELAY; 928 case LINUX_TCP_MAXSEG: 929 return TCP_MAXSEG; 930 default: 931 return -1; 932 } 933 } 934 935 /* 936 * Convert Linux UDP level socket option number to NetBSD values. 937 */ 938 int 939 linux_to_bsd_udp_sockopt(int lopt) 940 { 941 942 switch (lopt) { 943 default: 944 return -1; 945 } 946 } 947 948 /* 949 * Another reasonably straightforward function: setsockopt(2). 950 * The level and option numbers are converted; the values passed 951 * are not (yet) converted, the ones currently implemented don't 952 * need conversion, as they are the same on both systems. 953 */ 954 int 955 linux_sys_setsockopt(struct lwp *l, const struct linux_sys_setsockopt_args *uap, register_t *retval) 956 { 957 /* { 958 syscallarg(int) s; 959 syscallarg(int) level; 960 syscallarg(int) optname; 961 syscallarg(void *) optval; 962 syscallarg(int) optlen; 963 } */ 964 struct sys_setsockopt_args bsa; 965 int name; 966 967 SCARG(&bsa, s) = SCARG(uap, s); 968 SCARG(&bsa, level) = linux_to_bsd_sopt_level(SCARG(uap, level)); 969 SCARG(&bsa, val) = SCARG(uap, optval); 970 SCARG(&bsa, valsize) = SCARG(uap, optlen); 971 972 /* 973 * Linux supports only SOL_SOCKET for AF_LOCAL domain sockets 974 * and returns EOPNOTSUPP for other levels 975 */ 976 if (SCARG(&bsa, level) != SOL_SOCKET) { 977 struct socket *so; 978 int error, family; 979 980 /* fd_getsock() will use the descriptor for us */ 981 if ((error = fd_getsock(SCARG(&bsa, s), &so)) != 0) 982 return error; 983 family = so->so_proto->pr_domain->dom_family; 984 fd_putfile(SCARG(&bsa, s)); 985 986 if (family == AF_LOCAL) 987 return EOPNOTSUPP; 988 } 989 990 switch (SCARG(&bsa, level)) { 991 case SOL_SOCKET: 992 name = linux_to_bsd_so_sockopt(SCARG(uap, optname)); 993 break; 994 case IPPROTO_IP: 995 name = linux_to_bsd_ip_sockopt(SCARG(uap, optname)); 996 break; 997 case IPPROTO_TCP: 998 name = linux_to_bsd_tcp_sockopt(SCARG(uap, optname)); 999 break; 1000 case IPPROTO_UDP: 1001 name = linux_to_bsd_udp_sockopt(SCARG(uap, optname)); 1002 break; 1003 default: 1004 return EINVAL; 1005 } 1006 1007 if (name == -1) 1008 return EINVAL; 1009 SCARG(&bsa, name) = name; 1010 1011 return sys_setsockopt(l, &bsa, retval); 1012 } 1013 1014 /* 1015 * getsockopt(2) is very much the same as setsockopt(2) (see above) 1016 */ 1017 int 1018 linux_sys_getsockopt(struct lwp *l, const struct linux_sys_getsockopt_args *uap, register_t *retval) 1019 { 1020 /* { 1021 syscallarg(int) s; 1022 syscallarg(int) level; 1023 syscallarg(int) optname; 1024 syscallarg(void *) optval; 1025 syscallarg(int *) optlen; 1026 } */ 1027 struct sys_getsockopt_args bga; 1028 int name; 1029 1030 SCARG(&bga, s) = SCARG(uap, s); 1031 SCARG(&bga, level) = linux_to_bsd_sopt_level(SCARG(uap, level)); 1032 SCARG(&bga, val) = SCARG(uap, optval); 1033 SCARG(&bga, avalsize) = (socklen_t *)SCARG(uap, optlen); 1034 1035 switch (SCARG(&bga, level)) { 1036 case SOL_SOCKET: 1037 name = linux_to_bsd_so_sockopt(SCARG(uap, optname)); 1038 break; 1039 case IPPROTO_IP: 1040 name = linux_to_bsd_ip_sockopt(SCARG(uap, optname)); 1041 break; 1042 case IPPROTO_TCP: 1043 name = linux_to_bsd_tcp_sockopt(SCARG(uap, optname)); 1044 break; 1045 case IPPROTO_UDP: 1046 name = linux_to_bsd_udp_sockopt(SCARG(uap, optname)); 1047 break; 1048 default: 1049 return EINVAL; 1050 } 1051 1052 if (name == -1) 1053 return EINVAL; 1054 SCARG(&bga, name) = name; 1055 1056 return sys_getsockopt(l, &bga, retval); 1057 } 1058 1059 int 1060 linux_getifname(struct lwp *l, register_t *retval, void *data) 1061 { 1062 struct ifnet *ifp; 1063 struct linux_ifreq ifr; 1064 int error; 1065 1066 error = copyin(data, &ifr, sizeof(ifr)); 1067 if (error) 1068 return error; 1069 1070 if (ifr.ifr_ifru.ifru_ifindex >= if_indexlim) 1071 return ENODEV; 1072 1073 ifp = ifindex2ifnet[ifr.ifr_ifru.ifru_ifindex]; 1074 if (ifp == NULL) 1075 return ENODEV; 1076 1077 strncpy(ifr.ifr_name, ifp->if_xname, sizeof(ifr.ifr_name)); 1078 1079 return copyout(&ifr, data, sizeof(ifr)); 1080 } 1081 1082 int 1083 linux_getifconf(struct lwp *l, register_t *retval, void *data) 1084 { 1085 struct linux_ifreq ifr, *ifrp; 1086 struct linux_ifconf ifc; 1087 struct ifnet *ifp; 1088 struct ifaddr *ifa; 1089 struct sockaddr *sa; 1090 struct osockaddr *osa; 1091 int space, error = 0; 1092 const int sz = (int)sizeof(ifr); 1093 1094 error = copyin(data, &ifc, sizeof(ifc)); 1095 if (error) 1096 return error; 1097 1098 ifrp = ifc.ifc_req; 1099 if (ifrp == NULL) 1100 space = 0; 1101 else 1102 space = ifc.ifc_len; 1103 1104 IFNET_FOREACH(ifp) { 1105 (void)strncpy(ifr.ifr_name, ifp->if_xname, 1106 sizeof(ifr.ifr_name)); 1107 if (ifr.ifr_name[sizeof(ifr.ifr_name) - 1] != '\0') 1108 return ENAMETOOLONG; 1109 if (IFADDR_EMPTY(ifp)) 1110 continue; 1111 IFADDR_FOREACH(ifa, ifp) { 1112 sa = ifa->ifa_addr; 1113 if (sa->sa_family != AF_INET || 1114 sa->sa_len > sizeof(*osa)) 1115 continue; 1116 memcpy(&ifr.ifr_addr, sa, sa->sa_len); 1117 osa = (struct osockaddr *)&ifr.ifr_addr; 1118 osa->sa_family = sa->sa_family; 1119 if (space >= sz) { 1120 error = copyout(&ifr, ifrp, sz); 1121 if (error != 0) 1122 return error; 1123 ifrp++; 1124 } 1125 space -= sz; 1126 } 1127 } 1128 1129 if (ifrp != NULL) 1130 ifc.ifc_len -= space; 1131 else 1132 ifc.ifc_len = -space; 1133 1134 return copyout(&ifc, data, sizeof(ifc)); 1135 } 1136 1137 int 1138 linux_getifhwaddr(struct lwp *l, register_t *retval, u_int fd, 1139 void *data) 1140 { 1141 /* Not the full structure, just enough to map what we do here */ 1142 struct linux_ifreq lreq; 1143 file_t *fp; 1144 struct ifaddr *ifa; 1145 struct ifnet *ifp; 1146 struct sockaddr_dl *sadl; 1147 int error, found; 1148 int index, ifnum; 1149 1150 /* 1151 * We can't emulate this ioctl by calling sys_ioctl() to run 1152 * SIOCGIFCONF, because the user buffer is not of the right 1153 * type to take those results. We can't use kernel buffers to 1154 * receive the results, as the implementation of sys_ioctl() 1155 * and ifconf() [which implements SIOCGIFCONF] use 1156 * copyin()/copyout() which will fail on kernel addresses. 1157 * 1158 * So, we must duplicate code from sys_ioctl() and ifconf(). Ugh. 1159 */ 1160 1161 if ((fp = fd_getfile(fd)) == NULL) 1162 return (EBADF); 1163 1164 KERNEL_LOCK(1, NULL); 1165 1166 if ((fp->f_flag & (FREAD | FWRITE)) == 0) { 1167 error = EBADF; 1168 goto out; 1169 } 1170 1171 error = copyin(data, &lreq, sizeof(lreq)); 1172 if (error) 1173 goto out; 1174 lreq.ifr_name[LINUX_IFNAMSIZ-1] = '\0'; /* just in case */ 1175 1176 /* 1177 * Try real interface name first, then fake "ethX" 1178 */ 1179 found = 0; 1180 IFNET_FOREACH(ifp) { 1181 if (found) 1182 break; 1183 if (strcmp(lreq.ifr_name, ifp->if_xname)) 1184 /* not this interface */ 1185 continue; 1186 found=1; 1187 if (IFADDR_EMPTY(ifp)) { 1188 error = ENODEV; 1189 goto out; 1190 } 1191 IFADDR_FOREACH(ifa, ifp) { 1192 sadl = satosdl(ifa->ifa_addr); 1193 /* only return ethernet addresses */ 1194 /* XXX what about FDDI, etc. ? */ 1195 if (sadl->sdl_family != AF_LINK || 1196 sadl->sdl_type != IFT_ETHER) 1197 continue; 1198 memcpy(&lreq.ifr_hwaddr.sa_data, CLLADDR(sadl), 1199 MIN(sadl->sdl_alen, 1200 sizeof(lreq.ifr_hwaddr.sa_data))); 1201 lreq.ifr_hwaddr.sa_family = 1202 sadl->sdl_family; 1203 error = copyout(&lreq, data, sizeof(lreq)); 1204 goto out; 1205 } 1206 } 1207 1208 if (strncmp(lreq.ifr_name, "eth", 3) == 0) { 1209 for (ifnum = 0, index = 3; 1210 index < LINUX_IFNAMSIZ && lreq.ifr_name[index] != '\0'; 1211 index++) { 1212 ifnum *= 10; 1213 ifnum += lreq.ifr_name[index] - '0'; 1214 } 1215 1216 error = EINVAL; /* in case we don't find one */ 1217 found = 0; 1218 IFNET_FOREACH(ifp) { 1219 if (found) 1220 break; 1221 memcpy(lreq.ifr_name, ifp->if_xname, 1222 MIN(LINUX_IFNAMSIZ, IFNAMSIZ)); 1223 IFADDR_FOREACH(ifa, ifp) { 1224 sadl = satosdl(ifa->ifa_addr); 1225 /* only return ethernet addresses */ 1226 /* XXX what about FDDI, etc. ? */ 1227 if (sadl->sdl_family != AF_LINK || 1228 sadl->sdl_type != IFT_ETHER) 1229 continue; 1230 if (ifnum--) 1231 /* not the reqested iface */ 1232 continue; 1233 memcpy(&lreq.ifr_hwaddr.sa_data, 1234 CLLADDR(sadl), 1235 MIN(sadl->sdl_alen, 1236 sizeof(lreq.ifr_hwaddr.sa_data))); 1237 lreq.ifr_hwaddr.sa_family = 1238 sadl->sdl_family; 1239 error = copyout(&lreq, data, sizeof(lreq)); 1240 found = 1; 1241 break; 1242 } 1243 } 1244 } else { 1245 /* unknown interface, not even an "eth*" name */ 1246 error = ENODEV; 1247 } 1248 1249 out: 1250 KERNEL_UNLOCK_ONE(NULL); 1251 fd_putfile(fd); 1252 return error; 1253 } 1254 1255 int 1256 linux_ioctl_socket(struct lwp *l, const struct linux_sys_ioctl_args *uap, register_t *retval) 1257 { 1258 /* { 1259 syscallarg(int) fd; 1260 syscallarg(u_long) com; 1261 syscallarg(void *) data; 1262 } */ 1263 u_long com; 1264 int error = 0, isdev = 0, dosys = 1; 1265 struct sys_ioctl_args ia; 1266 file_t *fp; 1267 struct vnode *vp; 1268 int (*ioctlf)(file_t *, u_long, void *); 1269 struct ioctl_pt pt; 1270 1271 if ((fp = fd_getfile(SCARG(uap, fd))) == NULL) 1272 return (EBADF); 1273 1274 if (fp->f_type == DTYPE_VNODE) { 1275 vp = (struct vnode *)fp->f_data; 1276 isdev = vp->v_type == VCHR; 1277 } 1278 1279 /* 1280 * Don't try to interpret socket ioctl calls that are done 1281 * on a device filedescriptor, just pass them through, to 1282 * emulate Linux behaviour. Use PTIOCLINUX so that the 1283 * device will only handle these if it's prepared to do 1284 * so, to avoid unexpected things from happening. 1285 */ 1286 if (isdev) { 1287 dosys = 0; 1288 ioctlf = fp->f_ops->fo_ioctl; 1289 pt.com = SCARG(uap, com); 1290 pt.data = SCARG(uap, data); 1291 error = ioctlf(fp, PTIOCLINUX, &pt); 1292 /* 1293 * XXX hack: if the function returns EJUSTRETURN, 1294 * it has stuffed a sysctl return value in pt.data. 1295 */ 1296 if (error == EJUSTRETURN) { 1297 retval[0] = (register_t)pt.data; 1298 error = 0; 1299 } 1300 goto out; 1301 } 1302 1303 com = SCARG(uap, com); 1304 retval[0] = 0; 1305 1306 switch (com) { 1307 case LINUX_SIOCGIFNAME: 1308 error = linux_getifname(l, retval, SCARG(uap, data)); 1309 dosys = 0; 1310 break; 1311 case LINUX_SIOCGIFCONF: 1312 error = linux_getifconf(l, retval, SCARG(uap, data)); 1313 dosys = 0; 1314 break; 1315 case LINUX_SIOCGIFFLAGS: 1316 SCARG(&ia, com) = OSIOCGIFFLAGS; 1317 break; 1318 case LINUX_SIOCSIFFLAGS: 1319 SCARG(&ia, com) = OSIOCSIFFLAGS; 1320 break; 1321 case LINUX_SIOCGIFADDR: 1322 SCARG(&ia, com) = OOSIOCGIFADDR; 1323 break; 1324 case LINUX_SIOCGIFDSTADDR: 1325 SCARG(&ia, com) = OOSIOCGIFDSTADDR; 1326 break; 1327 case LINUX_SIOCGIFBRDADDR: 1328 SCARG(&ia, com) = OOSIOCGIFBRDADDR; 1329 break; 1330 case LINUX_SIOCGIFNETMASK: 1331 SCARG(&ia, com) = OOSIOCGIFNETMASK; 1332 break; 1333 case LINUX_SIOCGIFMTU: 1334 SCARG(&ia, com) = OSIOCGIFMTU; 1335 break; 1336 case LINUX_SIOCADDMULTI: 1337 SCARG(&ia, com) = OSIOCADDMULTI; 1338 break; 1339 case LINUX_SIOCDELMULTI: 1340 SCARG(&ia, com) = OSIOCDELMULTI; 1341 break; 1342 case LINUX_SIOCGIFHWADDR: 1343 error = linux_getifhwaddr(l, retval, SCARG(uap, fd), 1344 SCARG(uap, data)); 1345 dosys = 0; 1346 break; 1347 default: 1348 error = EINVAL; 1349 } 1350 1351 out: 1352 fd_putfile(SCARG(uap, fd)); 1353 1354 if (error ==0 && dosys) { 1355 SCARG(&ia, fd) = SCARG(uap, fd); 1356 SCARG(&ia, data) = SCARG(uap, data); 1357 error = sys_ioctl(curlwp, &ia, retval); 1358 } 1359 1360 return error; 1361 } 1362 1363 int 1364 linux_sys_connect(struct lwp *l, const struct linux_sys_connect_args *uap, register_t *retval) 1365 { 1366 /* { 1367 syscallarg(int) s; 1368 syscallarg(const struct sockaddr *) name; 1369 syscallarg(int) namelen; 1370 } */ 1371 int error; 1372 struct mbuf *nam; 1373 1374 error = linux_get_sa(l, SCARG(uap, s), &nam, SCARG(uap, name), 1375 SCARG(uap, namelen)); 1376 if (error) 1377 return (error); 1378 1379 error = do_sys_connect(l, SCARG(uap, s), nam); 1380 1381 if (error == EISCONN) { 1382 struct socket *so; 1383 int state, prflags; 1384 1385 /* fd_getsock() will use the descriptor for us */ 1386 if (fd_getsock(SCARG(uap, s), &so) != 0) 1387 return EISCONN; 1388 1389 solock(so); 1390 state = so->so_state; 1391 prflags = so->so_proto->pr_flags; 1392 sounlock(so); 1393 fd_putfile(SCARG(uap, s)); 1394 /* 1395 * We should only let this call succeed once per 1396 * non-blocking connect; however we don't have 1397 * a convenient place to keep that state.. 1398 */ 1399 if ((state & (SS_ISCONNECTED|SS_NBIO)) == 1400 (SS_ISCONNECTED|SS_NBIO) && 1401 (prflags & PR_CONNREQUIRED)) 1402 return 0; 1403 } 1404 1405 return (error); 1406 } 1407 1408 int 1409 linux_sys_bind(struct lwp *l, const struct linux_sys_bind_args *uap, register_t *retval) 1410 { 1411 /* { 1412 syscallarg(int) s; 1413 syscallarg(const struct osockaddr *) name; 1414 syscallarg(int) namelen; 1415 } */ 1416 int error; 1417 struct mbuf *nam; 1418 1419 error = linux_get_sa(l, SCARG(uap, s), &nam, SCARG(uap, name), 1420 SCARG(uap, namelen)); 1421 if (error) 1422 return (error); 1423 1424 return do_sys_bind(l, SCARG(uap, s), nam); 1425 } 1426 1427 int 1428 linux_sys_getsockname(struct lwp *l, const struct linux_sys_getsockname_args *uap, register_t *retval) 1429 { 1430 /* { 1431 syscallarg(int) fdes; 1432 syscallarg(void *) asa; 1433 syscallarg(int *) alen; 1434 } */ 1435 int error; 1436 1437 if ((error = sys_getsockname(l, (const void *)uap, retval)) != 0) 1438 return (error); 1439 1440 if ((error = linux_sa_put((struct osockaddr *)SCARG(uap, asa)))) 1441 return (error); 1442 1443 return (0); 1444 } 1445 1446 int 1447 linux_sys_getpeername(struct lwp *l, const struct linux_sys_getpeername_args *uap, register_t *retval) 1448 { 1449 /* { 1450 syscallarg(int) fdes; 1451 syscallarg(void *) asa; 1452 syscallarg(int *) alen; 1453 } */ 1454 int error; 1455 1456 if ((error = sys_getpeername(l, (const void *)uap, retval)) != 0) 1457 return (error); 1458 1459 if ((error = linux_sa_put((struct osockaddr *)SCARG(uap, asa)))) 1460 return (error); 1461 1462 return (0); 1463 } 1464 1465 /* 1466 * Copy the osockaddr structure pointed to by osa to mbuf, adjust 1467 * family and convert to sockaddr. 1468 */ 1469 static int 1470 linux_get_sa(struct lwp *l, int s, struct mbuf **mp, 1471 const struct osockaddr *osa, unsigned int salen) 1472 { 1473 int error, bdom; 1474 struct sockaddr *sa; 1475 struct osockaddr *kosa; 1476 struct mbuf *m; 1477 1478 if (salen == 1 || salen > UCHAR_MAX) { 1479 DPRINTF(("bad osa=%p salen=%d\n", osa, salen)); 1480 return EINVAL; 1481 } 1482 1483 /* We'll need the address in an mbuf later, so copy into one here */ 1484 m = m_get(M_WAIT, MT_SONAME); 1485 if (salen > MLEN) 1486 MEXTMALLOC(m, salen, M_WAITOK); 1487 1488 m->m_len = salen; 1489 1490 if (salen == 0) { 1491 *mp = m; 1492 return 0; 1493 } 1494 1495 kosa = mtod(m, void *); 1496 if ((error = copyin(osa, kosa, salen))) { 1497 DPRINTF(("error %d copying osa %p len %d\n", 1498 error, osa, salen)); 1499 goto bad; 1500 } 1501 1502 ktrkuser("linux/sockaddr", kosa, salen); 1503 1504 bdom = linux_to_bsd_domain(kosa->sa_family); 1505 if (bdom == -1) { 1506 DPRINTF(("bad linux family=%d\n", kosa->sa_family)); 1507 error = EINVAL; 1508 goto bad; 1509 } 1510 1511 /* 1512 * If the family is unspecified, use address family of the socket. 1513 * This avoid triggering strict family checks in netinet/in_pcb.c et.al. 1514 */ 1515 if (bdom == AF_UNSPEC) { 1516 struct socket *so; 1517 1518 /* fd_getsock() will use the descriptor for us */ 1519 if ((error = fd_getsock(s, &so)) != 0) 1520 goto bad; 1521 1522 bdom = so->so_proto->pr_domain->dom_family; 1523 fd_putfile(s); 1524 1525 DPRINTF(("AF_UNSPEC family adjusted to %d\n", bdom)); 1526 } 1527 1528 /* 1529 * Older Linux IPv6 code uses obsolete RFC2133 struct sockaddr_in6, 1530 * which lacks the scope id compared with RFC2553 one. If we detect 1531 * the situation, reject the address and write a message to system log. 1532 * 1533 * Still accept addresses for which the scope id is not used. 1534 */ 1535 if (bdom == AF_INET6 && salen == sizeof (struct sockaddr_in6) - sizeof (u_int32_t)) { 1536 struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)kosa; 1537 if (!IN6_IS_ADDR_V4MAPPED(&sin6->sin6_addr) && 1538 (IN6_IS_ADDR_LINKLOCAL(&sin6->sin6_addr) || 1539 IN6_IS_ADDR_SITELOCAL(&sin6->sin6_addr) || 1540 IN6_IS_ADDR_V4COMPAT(&sin6->sin6_addr) || 1541 IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr) || 1542 IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr))) { 1543 struct proc *p = l->l_proc; 1544 int uid = l->l_cred ? kauth_cred_geteuid(l->l_cred) : -1; 1545 1546 log(LOG_DEBUG, 1547 "pid %d (%s), uid %d: obsolete pre-RFC2553 " 1548 "sockaddr_in6 rejected", 1549 p->p_pid, p->p_comm, uid); 1550 error = EINVAL; 1551 goto bad; 1552 } 1553 salen = sizeof (struct sockaddr_in6); 1554 sin6->sin6_scope_id = 0; 1555 } 1556 1557 if (bdom == AF_INET) 1558 salen = sizeof(struct sockaddr_in); 1559 1560 sa = (struct sockaddr *) kosa; 1561 sa->sa_family = bdom; 1562 sa->sa_len = salen; 1563 m->m_len = salen; 1564 ktrkuser("mbsoname", kosa, salen); 1565 1566 #ifdef DEBUG_LINUX 1567 DPRINTF(("family %d, len = %d [ ", sa->sa_family, sa->sa_len)); 1568 for (bdom = 0; bdom < sizeof(sa->sa_data); bdom++) 1569 DPRINTF(("%02x ", (unsigned char) sa->sa_data[bdom])); 1570 DPRINTF(("\n")); 1571 #endif 1572 1573 *mp = m; 1574 return 0; 1575 1576 bad: 1577 m_free(m); 1578 return error; 1579 } 1580 1581 static int 1582 linux_sa_put(struct osockaddr *osa) 1583 { 1584 struct sockaddr sa; 1585 struct osockaddr *kosa; 1586 int error, bdom, len; 1587 1588 /* 1589 * Only read/write the sockaddr family and length part, the rest is 1590 * not changed. 1591 */ 1592 len = sizeof(sa.sa_len) + sizeof(sa.sa_family); 1593 1594 error = copyin(osa, &sa, len); 1595 if (error) 1596 return (error); 1597 1598 bdom = bsd_to_linux_domain(sa.sa_family); 1599 if (bdom == -1) 1600 return (EINVAL); 1601 1602 /* Note: we convert from sockaddr to osockaddr here, too */ 1603 kosa = (struct osockaddr *) &sa; 1604 kosa->sa_family = bdom; 1605 error = copyout(kosa, osa, len); 1606 if (error) 1607 return (error); 1608 1609 return (0); 1610 } 1611 1612 #ifndef __amd64__ 1613 int 1614 linux_sys_recv(struct lwp *l, const struct linux_sys_recv_args *uap, register_t *retval) 1615 { 1616 /* { 1617 syscallarg(int) s; 1618 syscallarg(void *) buf; 1619 syscallarg(int) len; 1620 syscallarg(int) flags; 1621 } */ 1622 struct sys_recvfrom_args bra; 1623 1624 1625 SCARG(&bra, s) = SCARG(uap, s); 1626 SCARG(&bra, buf) = SCARG(uap, buf); 1627 SCARG(&bra, len) = (size_t) SCARG(uap, len); 1628 SCARG(&bra, flags) = SCARG(uap, flags); 1629 SCARG(&bra, from) = NULL; 1630 SCARG(&bra, fromlenaddr) = NULL; 1631 1632 return (sys_recvfrom(l, &bra, retval)); 1633 } 1634 1635 int 1636 linux_sys_send(struct lwp *l, const struct linux_sys_send_args *uap, register_t *retval) 1637 { 1638 /* { 1639 syscallarg(int) s; 1640 syscallarg(void *) buf; 1641 syscallarg(int) len; 1642 syscallarg(int) flags; 1643 } */ 1644 struct sys_sendto_args bsa; 1645 1646 SCARG(&bsa, s) = SCARG(uap, s); 1647 SCARG(&bsa, buf) = SCARG(uap, buf); 1648 SCARG(&bsa, len) = SCARG(uap, len); 1649 SCARG(&bsa, flags) = SCARG(uap, flags); 1650 SCARG(&bsa, to) = NULL; 1651 SCARG(&bsa, tolen) = 0; 1652 1653 return (sys_sendto(l, &bsa, retval)); 1654 } 1655 #endif 1656 1657 int 1658 linux_sys_accept(struct lwp *l, const struct linux_sys_accept_args *uap, register_t *retval) 1659 { 1660 /* { 1661 syscallarg(int) s; 1662 syscallarg(struct osockaddr *) name; 1663 syscallarg(int *) anamelen; 1664 } */ 1665 int error; 1666 struct sys_accept_args baa; 1667 1668 SCARG(&baa, s) = SCARG(uap, s); 1669 SCARG(&baa, name) = (struct sockaddr *) SCARG(uap, name); 1670 SCARG(&baa, anamelen) = (unsigned int *) SCARG(uap, anamelen); 1671 1672 if ((error = sys_accept(l, &baa, retval))) 1673 return (error); 1674 1675 if (SCARG(uap, name) && (error = linux_sa_put(SCARG(uap, name)))) 1676 return (error); 1677 1678 return (0); 1679 } 1680