sys_descrip.c revision 1.44
1 /* $NetBSD: sys_descrip.c,v 1.44 2023/04/22 14:23:32 riastradh Exp $ */ 2 3 /*- 4 * Copyright (c) 2008, 2020 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 16 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 17 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 18 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 19 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 20 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 21 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 22 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 23 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 24 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 25 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 26 * POSSIBILITY OF SUCH DAMAGE. 27 */ 28 29 /* 30 * Copyright (c) 1982, 1986, 1989, 1991, 1993 31 * The Regents of the University of California. All rights reserved. 32 * (c) UNIX System Laboratories, Inc. 33 * All or some portions of this file are derived from material licensed 34 * to the University of California by American Telephone and Telegraph 35 * Co. or Unix System Laboratories, Inc. and are reproduced herein with 36 * the permission of UNIX System Laboratories, Inc. 37 * 38 * Redistribution and use in source and binary forms, with or without 39 * modification, are permitted provided that the following conditions 40 * are met: 41 * 1. Redistributions of source code must retain the above copyright 42 * notice, this list of conditions and the following disclaimer. 43 * 2. Redistributions in binary form must reproduce the above copyright 44 * notice, this list of conditions and the following disclaimer in the 45 * documentation and/or other materials provided with the distribution. 46 * 3. Neither the name of the University nor the names of its contributors 47 * may be used to endorse or promote products derived from this software 48 * without specific prior written permission. 49 * 50 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 51 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 52 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 53 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 54 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 55 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 56 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 57 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 58 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 59 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 60 * SUCH DAMAGE. 61 * 62 * @(#)kern_descrip.c 8.8 (Berkeley) 2/14/95 63 */ 64 65 /* 66 * System calls on descriptors. 67 */ 68 69 #include <sys/cdefs.h> 70 __KERNEL_RCSID(0, "$NetBSD: sys_descrip.c,v 1.44 2023/04/22 14:23:32 riastradh Exp $"); 71 72 #include <sys/param.h> 73 #include <sys/systm.h> 74 #include <sys/filedesc.h> 75 #include <sys/kernel.h> 76 #include <sys/vnode.h> 77 #include <sys/proc.h> 78 #include <sys/file.h> 79 #include <sys/namei.h> 80 #include <sys/socket.h> 81 #include <sys/socketvar.h> 82 #include <sys/stat.h> 83 #include <sys/ioctl.h> 84 #include <sys/fcntl.h> 85 #include <sys/kmem.h> 86 #include <sys/pool.h> 87 #include <sys/syslog.h> 88 #include <sys/unistd.h> 89 #include <sys/resourcevar.h> 90 #include <sys/conf.h> 91 #include <sys/event.h> 92 #include <sys/kauth.h> 93 #include <sys/atomic.h> 94 #include <sys/mount.h> 95 #include <sys/syscallargs.h> 96 97 #include <uvm/uvm_readahead.h> 98 99 /* 100 * Duplicate a file descriptor. 101 */ 102 int 103 sys_dup(struct lwp *l, const struct sys_dup_args *uap, register_t *retval) 104 { 105 /* { 106 syscallarg(int) fd; 107 } */ 108 int error, newfd, oldfd; 109 file_t *fp; 110 111 oldfd = SCARG(uap, fd); 112 113 if ((fp = fd_getfile(oldfd)) == NULL) { 114 return EBADF; 115 } 116 error = fd_dup(fp, 0, &newfd, false); 117 fd_putfile(oldfd); 118 *retval = newfd; 119 return error; 120 } 121 122 /* 123 * Duplicate a file descriptor to a particular value. 124 */ 125 int 126 dodup(struct lwp *l, int from, int to, int flags, register_t *retval) 127 { 128 int error; 129 file_t *fp; 130 131 if ((fp = fd_getfile(from)) == NULL) 132 return EBADF; 133 mutex_enter(&fp->f_lock); 134 fp->f_count++; 135 mutex_exit(&fp->f_lock); 136 fd_putfile(from); 137 138 if ((u_int)to >= curproc->p_rlimit[RLIMIT_NOFILE].rlim_cur || 139 (u_int)to >= maxfiles) 140 error = EBADF; 141 else if (from == to) 142 error = 0; 143 else 144 error = fd_dup2(fp, to, flags); 145 closef(fp); 146 *retval = to; 147 148 return error; 149 } 150 151 int 152 sys_dup3(struct lwp *l, const struct sys_dup3_args *uap, register_t *retval) 153 { 154 /* { 155 syscallarg(int) from; 156 syscallarg(int) to; 157 syscallarg(int) flags; 158 } */ 159 return dodup(l, SCARG(uap, from), SCARG(uap, to), SCARG(uap, flags), 160 retval); 161 } 162 163 int 164 sys_dup2(struct lwp *l, const struct sys_dup2_args *uap, register_t *retval) 165 { 166 /* { 167 syscallarg(int) from; 168 syscallarg(int) to; 169 } */ 170 return dodup(l, SCARG(uap, from), SCARG(uap, to), 0, retval); 171 } 172 173 /* 174 * fcntl call which is being passed to the file's fs. 175 */ 176 static int 177 fcntl_forfs(int fd, file_t *fp, int cmd, void *arg) 178 { 179 int error; 180 u_int size; 181 void *data, *memp; 182 #define STK_PARAMS 128 183 char stkbuf[STK_PARAMS]; 184 185 if ((fp->f_flag & (FREAD | FWRITE)) == 0) 186 return (EBADF); 187 188 /* 189 * Interpret high order word to find amount of data to be 190 * copied to/from the user's address space. 191 */ 192 size = (size_t)F_PARAM_LEN(cmd); 193 if (size > F_PARAM_MAX) 194 return (EINVAL); 195 memp = NULL; 196 if (size > sizeof(stkbuf)) { 197 memp = kmem_alloc(size, KM_SLEEP); 198 data = memp; 199 } else 200 data = stkbuf; 201 if (cmd & F_FSIN) { 202 if (size) { 203 error = copyin(arg, data, size); 204 if (error) { 205 if (memp) 206 kmem_free(memp, size); 207 return (error); 208 } 209 } else 210 *(void **)data = arg; 211 } else if ((cmd & F_FSOUT) != 0 && size != 0) { 212 /* 213 * Zero the buffer so the user always 214 * gets back something deterministic. 215 */ 216 memset(data, 0, size); 217 } else if (cmd & F_FSVOID) 218 *(void **)data = arg; 219 220 221 error = (*fp->f_ops->fo_fcntl)(fp, cmd, data); 222 223 /* 224 * Copy any data to user, size was 225 * already set and checked above. 226 */ 227 if (error == 0 && (cmd & F_FSOUT) && size) 228 error = copyout(data, arg, size); 229 if (memp) 230 kmem_free(memp, size); 231 return (error); 232 } 233 234 int 235 do_fcntl_lock(int fd, int cmd, struct flock *fl) 236 { 237 file_t *fp; 238 proc_t *p; 239 int (*fo_advlock)(struct file *, void *, int, struct flock *, int); 240 int error, flg; 241 242 if ((fp = fd_getfile(fd)) == NULL) 243 return EBADF; 244 if ((fo_advlock = fp->f_ops->fo_advlock) == NULL) { 245 error = EINVAL; 246 goto out; 247 } 248 249 flg = F_POSIX; 250 p = curproc; 251 252 switch (cmd) { 253 case F_SETLKW: 254 flg |= F_WAIT; 255 /* Fall into F_SETLK */ 256 257 /* FALLTHROUGH */ 258 case F_SETLK: 259 switch (fl->l_type) { 260 case F_RDLCK: 261 if ((fp->f_flag & FREAD) == 0) { 262 error = EBADF; 263 break; 264 } 265 if ((p->p_flag & PK_ADVLOCK) == 0) { 266 mutex_enter(p->p_lock); 267 p->p_flag |= PK_ADVLOCK; 268 mutex_exit(p->p_lock); 269 } 270 error = (*fo_advlock)(fp, p, F_SETLK, fl, flg); 271 break; 272 273 case F_WRLCK: 274 if ((fp->f_flag & FWRITE) == 0) { 275 error = EBADF; 276 break; 277 } 278 if ((p->p_flag & PK_ADVLOCK) == 0) { 279 mutex_enter(p->p_lock); 280 p->p_flag |= PK_ADVLOCK; 281 mutex_exit(p->p_lock); 282 } 283 error = (*fo_advlock)(fp, p, F_SETLK, fl, flg); 284 break; 285 286 case F_UNLCK: 287 error = (*fo_advlock)(fp, p, F_UNLCK, fl, F_POSIX); 288 break; 289 290 default: 291 error = EINVAL; 292 break; 293 } 294 break; 295 296 case F_GETLK: 297 if (fl->l_type != F_RDLCK && 298 fl->l_type != F_WRLCK && 299 fl->l_type != F_UNLCK) { 300 error = EINVAL; 301 break; 302 } 303 error = (*fo_advlock)(fp, p, F_GETLK, fl, F_POSIX); 304 break; 305 306 default: 307 error = EINVAL; 308 break; 309 } 310 311 out: fd_putfile(fd); 312 return error; 313 } 314 315 static int 316 do_fcntl_getpath(struct lwp *l, file_t *fp, char *upath) 317 { 318 char *kpath; 319 int error; 320 321 if (fp->f_type != DTYPE_VNODE) 322 return EOPNOTSUPP; 323 324 kpath = PNBUF_GET(); 325 326 error = vnode_to_path(kpath, MAXPATHLEN, fp->f_vnode, l, l->l_proc); 327 if (!error) 328 error = copyoutstr(kpath, upath, MAXPATHLEN, NULL); 329 330 PNBUF_PUT(kpath); 331 332 return error; 333 } 334 335 /* 336 * The file control system call. 337 */ 338 int 339 sys_fcntl(struct lwp *l, const struct sys_fcntl_args *uap, register_t *retval) 340 { 341 /* { 342 syscallarg(int) fd; 343 syscallarg(int) cmd; 344 syscallarg(void *) arg; 345 } */ 346 int fd, i, tmp, error, cmd, newmin; 347 filedesc_t *fdp; 348 fdtab_t *dt; 349 file_t *fp; 350 struct flock fl; 351 bool cloexec = false; 352 353 fd = SCARG(uap, fd); 354 cmd = SCARG(uap, cmd); 355 fdp = l->l_fd; 356 error = 0; 357 358 switch (cmd) { 359 case F_CLOSEM: 360 if (fd < 0) 361 return EBADF; 362 while ((i = fdp->fd_lastfile) >= fd) { 363 if (fd_getfile(i) == NULL) { 364 /* Another thread has updated. */ 365 continue; 366 } 367 fd_close(i); 368 } 369 return 0; 370 371 case F_MAXFD: 372 *retval = fdp->fd_lastfile; 373 return 0; 374 375 case F_SETLKW: 376 case F_SETLK: 377 case F_GETLK: 378 error = copyin(SCARG(uap, arg), &fl, sizeof(fl)); 379 if (error) 380 return error; 381 error = do_fcntl_lock(fd, cmd, &fl); 382 if (cmd == F_GETLK && error == 0) 383 error = copyout(&fl, SCARG(uap, arg), sizeof(fl)); 384 return error; 385 386 default: 387 /* Handled below */ 388 break; 389 } 390 391 if ((fp = fd_getfile(fd)) == NULL) 392 return EBADF; 393 394 if ((cmd & F_FSCTL)) { 395 error = fcntl_forfs(fd, fp, cmd, SCARG(uap, arg)); 396 fd_putfile(fd); 397 return error; 398 } 399 400 switch (cmd) { 401 case F_DUPFD_CLOEXEC: 402 cloexec = true; 403 /*FALLTHROUGH*/ 404 case F_DUPFD: 405 newmin = (long)SCARG(uap, arg); 406 if ((u_int)newmin >= 407 l->l_proc->p_rlimit[RLIMIT_NOFILE].rlim_cur || 408 (u_int)newmin >= maxfiles) { 409 fd_putfile(fd); 410 return EINVAL; 411 } 412 error = fd_dup(fp, newmin, &i, cloexec); 413 *retval = i; 414 break; 415 416 case F_GETFD: 417 dt = atomic_load_consume(&fdp->fd_dt); 418 *retval = dt->dt_ff[fd]->ff_exclose; 419 break; 420 421 case F_SETFD: 422 fd_set_exclose(l, fd, 423 ((long)SCARG(uap, arg) & FD_CLOEXEC) != 0); 424 break; 425 426 case F_GETNOSIGPIPE: 427 *retval = (fp->f_flag & FNOSIGPIPE) != 0; 428 break; 429 430 case F_SETNOSIGPIPE: 431 if (SCARG(uap, arg)) 432 atomic_or_uint(&fp->f_flag, FNOSIGPIPE); 433 else 434 atomic_and_uint(&fp->f_flag, ~FNOSIGPIPE); 435 *retval = 0; 436 break; 437 438 case F_GETFL: 439 *retval = OFLAGS(fp->f_flag); 440 break; 441 442 case F_SETFL: 443 /* XXX not guaranteed to be atomic. */ 444 tmp = FFLAGS((long)SCARG(uap, arg)) & FCNTLFLAGS; 445 error = (*fp->f_ops->fo_fcntl)(fp, F_SETFL, &tmp); 446 if (error) 447 break; 448 i = tmp ^ fp->f_flag; 449 if (i & FNONBLOCK) { 450 int flgs = tmp & FNONBLOCK; 451 error = (*fp->f_ops->fo_ioctl)(fp, FIONBIO, &flgs); 452 if (error) { 453 (*fp->f_ops->fo_fcntl)(fp, F_SETFL, 454 &fp->f_flag); 455 break; 456 } 457 } 458 if (i & FASYNC) { 459 int flgs = tmp & FASYNC; 460 error = (*fp->f_ops->fo_ioctl)(fp, FIOASYNC, &flgs); 461 if (error) { 462 if (i & FNONBLOCK) { 463 tmp = fp->f_flag & FNONBLOCK; 464 (void)(*fp->f_ops->fo_ioctl)(fp, 465 FIONBIO, &tmp); 466 } 467 (*fp->f_ops->fo_fcntl)(fp, F_SETFL, 468 &fp->f_flag); 469 break; 470 } 471 } 472 fp->f_flag = (fp->f_flag & ~FCNTLFLAGS) | tmp; 473 break; 474 475 case F_GETOWN: 476 error = (*fp->f_ops->fo_ioctl)(fp, FIOGETOWN, &tmp); 477 *retval = tmp; 478 break; 479 480 case F_SETOWN: 481 tmp = (int)(uintptr_t) SCARG(uap, arg); 482 error = (*fp->f_ops->fo_ioctl)(fp, FIOSETOWN, &tmp); 483 break; 484 485 case F_GETPATH: 486 error = do_fcntl_getpath(l, fp, SCARG(uap, arg)); 487 break; 488 489 default: 490 error = EINVAL; 491 } 492 493 fd_putfile(fd); 494 return (error); 495 } 496 497 /* 498 * Close a file descriptor. 499 */ 500 int 501 sys_close(struct lwp *l, const struct sys_close_args *uap, register_t *retval) 502 { 503 /* { 504 syscallarg(int) fd; 505 } */ 506 int error; 507 int fd = SCARG(uap, fd); 508 509 if (fd_getfile(fd) == NULL) { 510 return EBADF; 511 } 512 513 error = fd_close(fd); 514 if (error == ERESTART) { 515 #ifdef DIAGNOSTIC 516 printf("%s[%d]: close(%d) returned ERESTART\n", 517 l->l_proc->p_comm, (int)l->l_proc->p_pid, fd); 518 #endif 519 error = EINTR; 520 } 521 522 return error; 523 } 524 525 /* 526 * Return status information about a file descriptor. 527 * Common function for compat code. 528 */ 529 int 530 do_sys_fstat(int fd, struct stat *sb) 531 { 532 file_t *fp; 533 int error; 534 535 if ((fp = fd_getfile(fd)) == NULL) { 536 return EBADF; 537 } 538 error = (*fp->f_ops->fo_stat)(fp, sb); 539 fd_putfile(fd); 540 541 return error; 542 } 543 544 /* 545 * Return status information about a file descriptor. 546 */ 547 int 548 sys___fstat50(struct lwp *l, const struct sys___fstat50_args *uap, 549 register_t *retval) 550 { 551 /* { 552 syscallarg(int) fd; 553 syscallarg(struct stat *) sb; 554 } */ 555 struct stat sb; 556 int error; 557 558 error = do_sys_fstat(SCARG(uap, fd), &sb); 559 if (error == 0) { 560 error = copyout(&sb, SCARG(uap, sb), sizeof(sb)); 561 } 562 return error; 563 } 564 565 /* 566 * Return pathconf information about a file descriptor. 567 */ 568 int 569 sys_fpathconf(struct lwp *l, const struct sys_fpathconf_args *uap, 570 register_t *retval) 571 { 572 /* { 573 syscallarg(int) fd; 574 syscallarg(int) name; 575 } */ 576 int fd, name, error; 577 file_t *fp; 578 579 fd = SCARG(uap, fd); 580 name = SCARG(uap, name); 581 error = 0; 582 583 if ((fp = fd_getfile(fd)) == NULL) 584 return EBADF; 585 if (fp->f_ops->fo_fpathconf == NULL) 586 error = EOPNOTSUPP; 587 else 588 error = (*fp->f_ops->fo_fpathconf)(fp, name, retval); 589 fd_putfile(fd); 590 return error; 591 } 592 593 /* 594 * Apply an advisory lock on a file descriptor. 595 * 596 * Just attempt to get a record lock of the requested type on 597 * the entire file (l_whence = SEEK_SET, l_start = 0, l_len = 0). 598 */ 599 /* ARGSUSED */ 600 int 601 sys_flock(struct lwp *l, const struct sys_flock_args *uap, register_t *retval) 602 { 603 /* { 604 syscallarg(int) fd; 605 syscallarg(int) how; 606 } */ 607 int fd, how, error; 608 file_t *fp; 609 int (*fo_advlock)(struct file *, void *, int, struct flock *, int); 610 struct flock lf; 611 612 fd = SCARG(uap, fd); 613 how = SCARG(uap, how); 614 615 if ((fp = fd_getfile(fd)) == NULL) 616 return EBADF; 617 if ((fo_advlock = fp->f_ops->fo_advlock) == NULL) { 618 error = EOPNOTSUPP; 619 goto out; 620 } 621 622 lf.l_whence = SEEK_SET; 623 lf.l_start = 0; 624 lf.l_len = 0; 625 626 switch (how & ~LOCK_NB) { 627 case LOCK_UN: 628 lf.l_type = F_UNLCK; 629 atomic_and_uint(&fp->f_flag, ~FHASLOCK); 630 error = (*fo_advlock)(fp, fp, F_UNLCK, &lf, F_FLOCK); 631 fd_putfile(fd); 632 return error; 633 case LOCK_EX: 634 lf.l_type = F_WRLCK; 635 break; 636 case LOCK_SH: 637 lf.l_type = F_RDLCK; 638 break; 639 default: 640 fd_putfile(fd); 641 return EINVAL; 642 } 643 644 atomic_or_uint(&fp->f_flag, FHASLOCK); 645 if (how & LOCK_NB) { 646 error = (*fo_advlock)(fp, fp, F_SETLK, &lf, F_FLOCK); 647 } else { 648 error = (*fo_advlock)(fp, fp, F_SETLK, &lf, F_FLOCK|F_WAIT); 649 } 650 out: fd_putfile(fd); 651 return error; 652 } 653 654 int 655 do_posix_fadvise(int fd, off_t offset, off_t len, int advice) 656 { 657 file_t *fp; 658 int error; 659 660 if ((fp = fd_getfile(fd)) == NULL) 661 return EBADF; 662 if (fp->f_ops->fo_posix_fadvise == NULL) { 663 error = EOPNOTSUPP; 664 } else { 665 error = (*fp->f_ops->fo_posix_fadvise)(fp, offset, len, 666 advice); 667 } 668 fd_putfile(fd); 669 return error; 670 } 671 672 int 673 sys___posix_fadvise50(struct lwp *l, 674 const struct sys___posix_fadvise50_args *uap, 675 register_t *retval) 676 { 677 /* { 678 syscallarg(int) fd; 679 syscallarg(int) pad; 680 syscallarg(off_t) offset; 681 syscallarg(off_t) len; 682 syscallarg(int) advice; 683 } */ 684 685 *retval = do_posix_fadvise(SCARG(uap, fd), SCARG(uap, offset), 686 SCARG(uap, len), SCARG(uap, advice)); 687 688 return 0; 689 } 690 691 int 692 sys_pipe(struct lwp *l, const void *v, register_t *retval) 693 { 694 int fd[2], error; 695 696 if ((error = pipe1(l, fd, 0)) != 0) 697 return error; 698 699 retval[0] = fd[0]; 700 retval[1] = fd[1]; 701 702 return 0; 703 } 704 705 int 706 sys_pipe2(struct lwp *l, const struct sys_pipe2_args *uap, register_t *retval) 707 { 708 /* { 709 syscallarg(int[2]) fildes; 710 syscallarg(int) flags; 711 } */ 712 int fd[2], error; 713 714 if ((error = pipe1(l, fd, SCARG(uap, flags))) != 0) 715 return error; 716 717 if ((error = copyout(fd, SCARG(uap, fildes), sizeof(fd))) != 0) 718 return error; 719 retval[0] = 0; 720 return 0; 721 } 722
Indexes created Tue Sep 23 07:09:52 GMT 2025