kern_descrip.c revision 1.259
1 /* $NetBSD: kern_descrip.c,v 1.259 2023/09/10 14:45:52 ad Exp $ */ 2 3 /*- 4 * Copyright (c) 2008, 2009, 2023 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Andrew Doran. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 20 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 21 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 22 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 23 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 24 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 25 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 26 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 27 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 28 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 29 * POSSIBILITY OF SUCH DAMAGE. 30 */ 31 32 /* 33 * Copyright (c) 1982, 1986, 1989, 1991, 1993 34 * The Regents of the University of California. All rights reserved. 35 * (c) UNIX System Laboratories, Inc. 36 * All or some portions of this file are derived from material licensed 37 * to the University of California by American Telephone and Telegraph 38 * Co. or Unix System Laboratories, Inc. and are reproduced herein with 39 * the permission of UNIX System Laboratories, Inc. 40 * 41 * Redistribution and use in source and binary forms, with or without 42 * modification, are permitted provided that the following conditions 43 * are met: 44 * 1. Redistributions of source code must retain the above copyright 45 * notice, this list of conditions and the following disclaimer. 46 * 2. Redistributions in binary form must reproduce the above copyright 47 * notice, this list of conditions and the following disclaimer in the 48 * documentation and/or other materials provided with the distribution. 49 * 3. Neither the name of the University nor the names of its contributors 50 * may be used to endorse or promote products derived from this software 51 * without specific prior written permission. 52 * 53 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 54 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 55 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 56 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 57 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 58 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 59 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 60 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 61 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 62 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 63 * SUCH DAMAGE. 64 * 65 * @(#)kern_descrip.c 8.8 (Berkeley) 2/14/95 66 */ 67 68 /* 69 * File descriptor management. 70 */ 71 72 #include <sys/cdefs.h> 73 __KERNEL_RCSID(0, "$NetBSD: kern_descrip.c,v 1.259 2023/09/10 14:45:52 ad Exp $"); 74 75 #include <sys/param.h> 76 #include <sys/systm.h> 77 #include <sys/filedesc.h> 78 #include <sys/kernel.h> 79 #include <sys/proc.h> 80 #include <sys/file.h> 81 #include <sys/socket.h> 82 #include <sys/socketvar.h> 83 #include <sys/stat.h> 84 #include <sys/ioctl.h> 85 #include <sys/fcntl.h> 86 #include <sys/pool.h> 87 #include <sys/unistd.h> 88 #include <sys/resourcevar.h> 89 #include <sys/conf.h> 90 #include <sys/event.h> 91 #include <sys/kauth.h> 92 #include <sys/atomic.h> 93 #include <sys/syscallargs.h> 94 #include <sys/cpu.h> 95 #include <sys/kmem.h> 96 #include <sys/vnode.h> 97 #include <sys/sysctl.h> 98 #include <sys/ktrace.h> 99 100 /* 101 * A list (head) of open files, counter, and lock protecting them. 102 */ 103 struct filelist filehead __cacheline_aligned; 104 static u_int nfiles __cacheline_aligned; 105 kmutex_t filelist_lock __cacheline_aligned; 106 107 static pool_cache_t filedesc_cache __read_mostly; 108 static pool_cache_t file_cache __read_mostly; 109 110 static int file_ctor(void *, void *, int); 111 static void file_dtor(void *, void *); 112 static void fdfile_ctor(fdfile_t *); 113 static void fdfile_dtor(fdfile_t *); 114 static int filedesc_ctor(void *, void *, int); 115 static void filedesc_dtor(void *, void *); 116 static int filedescopen(dev_t, int, int, lwp_t *); 117 118 static int sysctl_kern_file(SYSCTLFN_PROTO); 119 static int sysctl_kern_file2(SYSCTLFN_PROTO); 120 static void fill_file(struct file *, const struct file *); 121 static void fill_file2(struct kinfo_file *, const file_t *, const fdfile_t *, 122 int, pid_t); 123 124 const struct cdevsw filedesc_cdevsw = { 125 .d_open = filedescopen, 126 .d_close = noclose, 127 .d_read = noread, 128 .d_write = nowrite, 129 .d_ioctl = noioctl, 130 .d_stop = nostop, 131 .d_tty = notty, 132 .d_poll = nopoll, 133 .d_mmap = nommap, 134 .d_kqfilter = nokqfilter, 135 .d_discard = nodiscard, 136 .d_flag = D_OTHER | D_MPSAFE 137 }; 138 139 /* For ease of reading. */ 140 __strong_alias(fd_putvnode,fd_putfile) 141 __strong_alias(fd_putsock,fd_putfile) 142 143 /* 144 * Initialize the descriptor system. 145 */ 146 void 147 fd_sys_init(void) 148 { 149 static struct sysctllog *clog; 150 151 mutex_init(&filelist_lock, MUTEX_DEFAULT, IPL_NONE); 152 153 LIST_INIT(&filehead); 154 155 file_cache = pool_cache_init(sizeof(file_t), coherency_unit, 0, 156 0, "file", NULL, IPL_NONE, file_ctor, file_dtor, NULL); 157 KASSERT(file_cache != NULL); 158 159 filedesc_cache = pool_cache_init(sizeof(filedesc_t), coherency_unit, 160 0, 0, "filedesc", NULL, IPL_NONE, filedesc_ctor, filedesc_dtor, 161 NULL); 162 KASSERT(filedesc_cache != NULL); 163 164 sysctl_createv(&clog, 0, NULL, NULL, 165 CTLFLAG_PERMANENT, 166 CTLTYPE_STRUCT, "file", 167 SYSCTL_DESCR("System open file table"), 168 sysctl_kern_file, 0, NULL, 0, 169 CTL_KERN, KERN_FILE, CTL_EOL); 170 sysctl_createv(&clog, 0, NULL, NULL, 171 CTLFLAG_PERMANENT, 172 CTLTYPE_STRUCT, "file2", 173 SYSCTL_DESCR("System open file table"), 174 sysctl_kern_file2, 0, NULL, 0, 175 CTL_KERN, KERN_FILE2, CTL_EOL); 176 } 177 178 static bool 179 fd_isused(filedesc_t *fdp, unsigned fd) 180 { 181 u_int off = fd >> NDENTRYSHIFT; 182 183 KASSERT(fd < atomic_load_consume(&fdp->fd_dt)->dt_nfiles); 184 185 return (fdp->fd_lomap[off] & (1U << (fd & NDENTRYMASK))) != 0; 186 } 187 188 /* 189 * Verify that the bitmaps match the descriptor table. 190 */ 191 static inline void 192 fd_checkmaps(filedesc_t *fdp) 193 { 194 #ifdef DEBUG 195 fdtab_t *dt; 196 u_int fd; 197 198 KASSERT(fdp->fd_refcnt <= 1 || mutex_owned(&fdp->fd_lock)); 199 200 dt = fdp->fd_dt; 201 if (fdp->fd_refcnt == -1) { 202 /* 203 * fd_free tears down the table without maintaining its bitmap. 204 */ 205 return; 206 } 207 for (fd = 0; fd < dt->dt_nfiles; fd++) { 208 if (fd < NDFDFILE) { 209 KASSERT(dt->dt_ff[fd] == 210 (fdfile_t *)fdp->fd_dfdfile[fd]); 211 } 212 if (dt->dt_ff[fd] == NULL) { 213 KASSERT(!fd_isused(fdp, fd)); 214 } else if (dt->dt_ff[fd]->ff_file != NULL) { 215 KASSERT(fd_isused(fdp, fd)); 216 } 217 } 218 #endif 219 } 220 221 static int 222 fd_next_zero(filedesc_t *fdp, uint32_t *bitmap, int want, u_int bits) 223 { 224 int i, off, maxoff; 225 uint32_t sub; 226 227 KASSERT(mutex_owned(&fdp->fd_lock)); 228 229 fd_checkmaps(fdp); 230 231 if (want > bits) 232 return -1; 233 234 off = want >> NDENTRYSHIFT; 235 i = want & NDENTRYMASK; 236 if (i) { 237 sub = bitmap[off] | ((u_int)~0 >> (NDENTRIES - i)); 238 if (sub != ~0) 239 goto found; 240 off++; 241 } 242 243 maxoff = NDLOSLOTS(bits); 244 while (off < maxoff) { 245 if ((sub = bitmap[off]) != ~0) 246 goto found; 247 off++; 248 } 249 250 return -1; 251 252 found: 253 return (off << NDENTRYSHIFT) + ffs(~sub) - 1; 254 } 255 256 static int 257 fd_last_set(filedesc_t *fd, int last) 258 { 259 int off, i; 260 fdfile_t **ff = fd->fd_dt->dt_ff; 261 uint32_t *bitmap = fd->fd_lomap; 262 263 KASSERT(mutex_owned(&fd->fd_lock)); 264 265 fd_checkmaps(fd); 266 267 off = (last - 1) >> NDENTRYSHIFT; 268 269 while (off >= 0 && !bitmap[off]) 270 off--; 271 272 if (off < 0) 273 return -1; 274 275 i = ((off + 1) << NDENTRYSHIFT) - 1; 276 if (i >= last) 277 i = last - 1; 278 279 /* XXX should use bitmap */ 280 while (i > 0 && (ff[i] == NULL || !ff[i]->ff_allocated)) 281 i--; 282 283 return i; 284 } 285 286 static inline void 287 fd_used(filedesc_t *fdp, unsigned fd) 288 { 289 u_int off = fd >> NDENTRYSHIFT; 290 fdfile_t *ff; 291 292 ff = fdp->fd_dt->dt_ff[fd]; 293 294 KASSERT(mutex_owned(&fdp->fd_lock)); 295 KASSERT((fdp->fd_lomap[off] & (1U << (fd & NDENTRYMASK))) == 0); 296 KASSERT(ff != NULL); 297 KASSERT(ff->ff_file == NULL); 298 KASSERT(!ff->ff_allocated); 299 300 ff->ff_allocated = true; 301 fdp->fd_lomap[off] |= 1U << (fd & NDENTRYMASK); 302 if (__predict_false(fdp->fd_lomap[off] == ~0)) { 303 KASSERT((fdp->fd_himap[off >> NDENTRYSHIFT] & 304 (1U << (off & NDENTRYMASK))) == 0); 305 fdp->fd_himap[off >> NDENTRYSHIFT] |= 1U << (off & NDENTRYMASK); 306 } 307 308 if ((int)fd > fdp->fd_lastfile) { 309 fdp->fd_lastfile = fd; 310 } 311 312 fd_checkmaps(fdp); 313 } 314 315 static inline void 316 fd_unused(filedesc_t *fdp, unsigned fd) 317 { 318 u_int off = fd >> NDENTRYSHIFT; 319 fdfile_t *ff; 320 321 ff = fdp->fd_dt->dt_ff[fd]; 322 323 KASSERT(mutex_owned(&fdp->fd_lock)); 324 KASSERT(ff != NULL); 325 KASSERT(ff->ff_file == NULL); 326 KASSERT(ff->ff_allocated); 327 328 if (fd < fdp->fd_freefile) { 329 fdp->fd_freefile = fd; 330 } 331 332 if (fdp->fd_lomap[off] == ~0) { 333 KASSERT((fdp->fd_himap[off >> NDENTRYSHIFT] & 334 (1U << (off & NDENTRYMASK))) != 0); 335 fdp->fd_himap[off >> NDENTRYSHIFT] &= 336 ~(1U << (off & NDENTRYMASK)); 337 } 338 KASSERT((fdp->fd_lomap[off] & (1U << (fd & NDENTRYMASK))) != 0); 339 fdp->fd_lomap[off] &= ~(1U << (fd & NDENTRYMASK)); 340 ff->ff_allocated = false; 341 342 KASSERT(fd <= fdp->fd_lastfile); 343 if (fd == fdp->fd_lastfile) { 344 fdp->fd_lastfile = fd_last_set(fdp, fd); 345 } 346 fd_checkmaps(fdp); 347 } 348 349 /* 350 * Look up the file structure corresponding to a file descriptor 351 * and return the file, holding a reference on the descriptor. 352 */ 353 file_t * 354 fd_getfile(unsigned fd) 355 { 356 filedesc_t *fdp; 357 fdfile_t *ff; 358 file_t *fp; 359 fdtab_t *dt; 360 361 /* 362 * Look up the fdfile structure representing this descriptor. 363 * We are doing this unlocked. See fd_tryexpand(). 364 */ 365 fdp = curlwp->l_fd; 366 dt = atomic_load_consume(&fdp->fd_dt); 367 if (__predict_false(fd >= dt->dt_nfiles)) { 368 return NULL; 369 } 370 ff = dt->dt_ff[fd]; 371 KASSERT(fd >= NDFDFILE || ff == (fdfile_t *)fdp->fd_dfdfile[fd]); 372 if (__predict_false(ff == NULL)) { 373 return NULL; 374 } 375 376 /* Now get a reference to the descriptor. */ 377 if (fdp->fd_refcnt == 1) { 378 /* 379 * Single threaded: don't need to worry about concurrent 380 * access (other than earlier calls to kqueue, which may 381 * hold a reference to the descriptor). 382 */ 383 ff->ff_refcnt++; 384 } else { 385 /* 386 * Multi threaded: issue a memory barrier to ensure that we 387 * acquire the file pointer _after_ adding a reference. If 388 * no memory barrier, we could fetch a stale pointer. 389 * 390 * In particular, we must coordinate the following four 391 * memory operations: 392 * 393 * A. fd_close store ff->ff_file = NULL 394 * B. fd_close refcnt = atomic_dec_uint_nv(&ff->ff_refcnt) 395 * C. fd_getfile atomic_inc_uint(&ff->ff_refcnt) 396 * D. fd_getfile load fp = ff->ff_file 397 * 398 * If the order is D;A;B;C: 399 * 400 * 1. D: fp = ff->ff_file 401 * 2. A: ff->ff_file = NULL 402 * 3. B: refcnt = atomic_dec_uint_nv(&ff->ff_refcnt) 403 * 4. C: atomic_inc_uint(&ff->ff_refcnt) 404 * 405 * then fd_close determines that there are no more 406 * references and decides to free fp immediately, at 407 * the same that fd_getfile ends up with an fp that's 408 * about to be freed. *boom* 409 * 410 * By making B a release operation in fd_close, and by 411 * making C an acquire operation in fd_getfile, since 412 * they are atomic operations on the same object, which 413 * has a total modification order, we guarantee either: 414 * 415 * - B happens before C. Then since A is 416 * sequenced before B in fd_close, and C is 417 * sequenced before D in fd_getfile, we 418 * guarantee A happens before D, so fd_getfile 419 * reads a null fp and safely fails. 420 * 421 * - C happens before B. Then fd_getfile may read 422 * null or nonnull, but either way, fd_close 423 * will safely wait for references to drain. 424 */ 425 atomic_inc_uint(&ff->ff_refcnt); 426 membar_acquire(); 427 } 428 429 /* 430 * If the file is not open or is being closed then put the 431 * reference back. 432 */ 433 fp = atomic_load_consume(&ff->ff_file); 434 if (__predict_true(fp != NULL)) { 435 return fp; 436 } 437 fd_putfile(fd); 438 return NULL; 439 } 440 441 /* 442 * Release a reference to a file descriptor acquired with fd_getfile(). 443 */ 444 void 445 fd_putfile(unsigned fd) 446 { 447 filedesc_t *fdp; 448 fdfile_t *ff; 449 u_int u, v; 450 451 fdp = curlwp->l_fd; 452 KASSERT(fd < atomic_load_consume(&fdp->fd_dt)->dt_nfiles); 453 ff = atomic_load_consume(&fdp->fd_dt)->dt_ff[fd]; 454 455 KASSERT(ff != NULL); 456 KASSERT((ff->ff_refcnt & FR_MASK) > 0); 457 KASSERT(fd >= NDFDFILE || ff == (fdfile_t *)fdp->fd_dfdfile[fd]); 458 459 if (fdp->fd_refcnt == 1) { 460 /* 461 * Single threaded: don't need to worry about concurrent 462 * access (other than earlier calls to kqueue, which may 463 * hold a reference to the descriptor). 464 */ 465 if (__predict_false((ff->ff_refcnt & FR_CLOSING) != 0)) { 466 fd_close(fd); 467 return; 468 } 469 ff->ff_refcnt--; 470 return; 471 } 472 473 /* 474 * Ensure that any use of the file is complete and globally 475 * visible before dropping the final reference. If no membar, 476 * the current CPU could still access memory associated with 477 * the file after it has been freed or recycled by another 478 * CPU. 479 */ 480 membar_release(); 481 482 /* 483 * Be optimistic and start out with the assumption that no other 484 * threads are trying to close the descriptor. If the CAS fails, 485 * we lost a race and/or it's being closed. 486 */ 487 for (u = ff->ff_refcnt & FR_MASK;; u = v) { 488 v = atomic_cas_uint(&ff->ff_refcnt, u, u - 1); 489 if (__predict_true(u == v)) { 490 return; 491 } 492 if (__predict_false((v & FR_CLOSING) != 0)) { 493 break; 494 } 495 } 496 497 /* Another thread is waiting to close the file: join it. */ 498 (void)fd_close(fd); 499 } 500 501 /* 502 * Convenience wrapper around fd_getfile() that returns reference 503 * to a vnode. 504 */ 505 int 506 fd_getvnode(unsigned fd, file_t **fpp) 507 { 508 vnode_t *vp; 509 file_t *fp; 510 511 fp = fd_getfile(fd); 512 if (__predict_false(fp == NULL)) { 513 return EBADF; 514 } 515 if (__predict_false(fp->f_type != DTYPE_VNODE)) { 516 fd_putfile(fd); 517 return EINVAL; 518 } 519 vp = fp->f_vnode; 520 if (__predict_false(vp->v_type == VBAD)) { 521 /* XXX Is this case really necessary? */ 522 fd_putfile(fd); 523 return EBADF; 524 } 525 *fpp = fp; 526 return 0; 527 } 528 529 /* 530 * Convenience wrapper around fd_getfile() that returns reference 531 * to a socket. 532 */ 533 int 534 fd_getsock1(unsigned fd, struct socket **sop, file_t **fp) 535 { 536 *fp = fd_getfile(fd); 537 if (__predict_false(*fp == NULL)) { 538 return EBADF; 539 } 540 if (__predict_false((*fp)->f_type != DTYPE_SOCKET)) { 541 fd_putfile(fd); 542 return ENOTSOCK; 543 } 544 *sop = (*fp)->f_socket; 545 return 0; 546 } 547 548 int 549 fd_getsock(unsigned fd, struct socket **sop) 550 { 551 file_t *fp; 552 return fd_getsock1(fd, sop, &fp); 553 } 554 555 /* 556 * Look up the file structure corresponding to a file descriptor 557 * and return it with a reference held on the file, not the 558 * descriptor. 559 * 560 * This is heavyweight and only used when accessing descriptors 561 * from a foreign process. The caller must ensure that `p' does 562 * not exit or fork across this call. 563 * 564 * To release the file (not descriptor) reference, use closef(). 565 */ 566 file_t * 567 fd_getfile2(proc_t *p, unsigned fd) 568 { 569 filedesc_t *fdp; 570 fdfile_t *ff; 571 file_t *fp; 572 fdtab_t *dt; 573 574 fdp = p->p_fd; 575 mutex_enter(&fdp->fd_lock); 576 dt = fdp->fd_dt; 577 if (fd >= dt->dt_nfiles) { 578 mutex_exit(&fdp->fd_lock); 579 return NULL; 580 } 581 if ((ff = dt->dt_ff[fd]) == NULL) { 582 mutex_exit(&fdp->fd_lock); 583 return NULL; 584 } 585 if ((fp = atomic_load_consume(&ff->ff_file)) == NULL) { 586 mutex_exit(&fdp->fd_lock); 587 return NULL; 588 } 589 mutex_enter(&fp->f_lock); 590 fp->f_count++; 591 mutex_exit(&fp->f_lock); 592 mutex_exit(&fdp->fd_lock); 593 594 return fp; 595 } 596 597 /* 598 * Internal form of close. Must be called with a reference to the 599 * descriptor, and will drop the reference. When all descriptor 600 * references are dropped, releases the descriptor slot and a single 601 * reference to the file structure. 602 */ 603 int 604 fd_close(unsigned fd) 605 { 606 struct flock lf; 607 filedesc_t *fdp; 608 fdfile_t *ff; 609 file_t *fp; 610 proc_t *p; 611 lwp_t *l; 612 u_int refcnt; 613 614 l = curlwp; 615 p = l->l_proc; 616 fdp = l->l_fd; 617 ff = atomic_load_consume(&fdp->fd_dt)->dt_ff[fd]; 618 619 KASSERT(fd >= NDFDFILE || ff == (fdfile_t *)fdp->fd_dfdfile[fd]); 620 621 mutex_enter(&fdp->fd_lock); 622 KASSERT((ff->ff_refcnt & FR_MASK) > 0); 623 fp = atomic_load_consume(&ff->ff_file); 624 if (__predict_false(fp == NULL)) { 625 /* 626 * Another user of the file is already closing, and is 627 * waiting for other users of the file to drain. Release 628 * our reference, and wake up the closer. 629 */ 630 membar_release(); 631 atomic_dec_uint(&ff->ff_refcnt); 632 cv_broadcast(&ff->ff_closing); 633 mutex_exit(&fdp->fd_lock); 634 635 /* 636 * An application error, so pretend that the descriptor 637 * was already closed. We can't safely wait for it to 638 * be closed without potentially deadlocking. 639 */ 640 return (EBADF); 641 } 642 KASSERT((ff->ff_refcnt & FR_CLOSING) == 0); 643 644 /* 645 * There may be multiple users of this file within the process. 646 * Notify existing and new users that the file is closing. This 647 * will prevent them from adding additional uses to this file 648 * while we are closing it. 649 */ 650 atomic_store_relaxed(&ff->ff_file, NULL); 651 ff->ff_exclose = false; 652 653 /* 654 * We expect the caller to hold a descriptor reference - drop it. 655 * The reference count may increase beyond zero at this point due 656 * to an erroneous descriptor reference by an application, but 657 * fd_getfile() will notice that the file is being closed and drop 658 * the reference again. 659 */ 660 if (fdp->fd_refcnt == 1) { 661 /* Single threaded. */ 662 refcnt = --(ff->ff_refcnt); 663 } else { 664 /* Multi threaded. */ 665 membar_release(); 666 refcnt = atomic_dec_uint_nv(&ff->ff_refcnt); 667 membar_acquire(); 668 } 669 if (__predict_false(refcnt != 0)) { 670 /* 671 * Wait for other references to drain. This is typically 672 * an application error - the descriptor is being closed 673 * while still in use. 674 * (Or just a threaded application trying to unblock its 675 * thread that sleeps in (say) accept()). 676 */ 677 atomic_or_uint(&ff->ff_refcnt, FR_CLOSING); 678 679 /* 680 * Remove any knotes attached to the file. A knote 681 * attached to the descriptor can hold references on it. 682 */ 683 mutex_exit(&fdp->fd_lock); 684 if (!SLIST_EMPTY(&ff->ff_knlist)) { 685 knote_fdclose(fd); 686 } 687 688 /* 689 * Since the file system code doesn't know which fd 690 * each request came from (think dup()), we have to 691 * ask it to return ERESTART for any long-term blocks. 692 * The re-entry through read/write/etc will detect the 693 * closed fd and return EBAFD. 694 * Blocked partial writes may return a short length. 695 */ 696 (*fp->f_ops->fo_restart)(fp); 697 mutex_enter(&fdp->fd_lock); 698 699 /* 700 * We need to see the count drop to zero at least once, 701 * in order to ensure that all pre-existing references 702 * have been drained. New references past this point are 703 * of no interest. 704 * XXX (dsl) this may need to call fo_restart() after a 705 * timeout to guarantee that all the system calls exit. 706 */ 707 while ((ff->ff_refcnt & FR_MASK) != 0) { 708 cv_wait(&ff->ff_closing, &fdp->fd_lock); 709 } 710 atomic_and_uint(&ff->ff_refcnt, ~FR_CLOSING); 711 } else { 712 /* If no references, there must be no knotes. */ 713 KASSERT(SLIST_EMPTY(&ff->ff_knlist)); 714 } 715 716 /* 717 * POSIX record locking dictates that any close releases ALL 718 * locks owned by this process. This is handled by setting 719 * a flag in the unlock to free ONLY locks obeying POSIX 720 * semantics, and not to free BSD-style file locks. 721 * If the descriptor was in a message, POSIX-style locks 722 * aren't passed with the descriptor. 723 */ 724 if (__predict_false((p->p_flag & PK_ADVLOCK) != 0) && 725 fp->f_ops->fo_advlock != NULL) { 726 lf.l_whence = SEEK_SET; 727 lf.l_start = 0; 728 lf.l_len = 0; 729 lf.l_type = F_UNLCK; 730 mutex_exit(&fdp->fd_lock); 731 (void)(*fp->f_ops->fo_advlock)(fp, p, F_UNLCK, &lf, F_POSIX); 732 mutex_enter(&fdp->fd_lock); 733 } 734 735 /* Free descriptor slot. */ 736 fd_unused(fdp, fd); 737 mutex_exit(&fdp->fd_lock); 738 739 /* Now drop reference to the file itself. */ 740 return closef(fp); 741 } 742 743 /* 744 * Duplicate a file descriptor. 745 */ 746 int 747 fd_dup(file_t *fp, int minfd, int *newp, bool exclose) 748 { 749 proc_t *p = curproc; 750 fdtab_t *dt; 751 int error; 752 753 while ((error = fd_alloc(p, minfd, newp)) != 0) { 754 if (error != ENOSPC) { 755 return error; 756 } 757 fd_tryexpand(p); 758 } 759 760 dt = atomic_load_consume(&curlwp->l_fd->fd_dt); 761 dt->dt_ff[*newp]->ff_exclose = exclose; 762 fd_affix(p, fp, *newp); 763 return 0; 764 } 765 766 /* 767 * dup2 operation. 768 */ 769 int 770 fd_dup2(file_t *fp, unsigned newfd, int flags) 771 { 772 filedesc_t *fdp = curlwp->l_fd; 773 fdfile_t *ff; 774 fdtab_t *dt; 775 776 if (flags & ~(O_CLOEXEC|O_NONBLOCK|O_NOSIGPIPE)) 777 return EINVAL; 778 /* 779 * Ensure there are enough slots in the descriptor table, 780 * and allocate an fdfile_t up front in case we need it. 781 */ 782 while (newfd >= atomic_load_consume(&fdp->fd_dt)->dt_nfiles) { 783 fd_tryexpand(curproc); 784 } 785 ff = kmem_alloc(sizeof(*ff), KM_SLEEP); 786 fdfile_ctor(ff); 787 788 /* 789 * If there is already a file open, close it. If the file is 790 * half open, wait for it to be constructed before closing it. 791 * XXX Potential for deadlock here? 792 */ 793 mutex_enter(&fdp->fd_lock); 794 while (fd_isused(fdp, newfd)) { 795 mutex_exit(&fdp->fd_lock); 796 if (fd_getfile(newfd) != NULL) { 797 (void)fd_close(newfd); 798 } else { 799 /* 800 * Crummy, but unlikely to happen. 801 * Can occur if we interrupt another 802 * thread while it is opening a file. 803 */ 804 kpause("dup2", false, 1, NULL); 805 } 806 mutex_enter(&fdp->fd_lock); 807 } 808 dt = fdp->fd_dt; 809 if (dt->dt_ff[newfd] == NULL) { 810 KASSERT(newfd >= NDFDFILE); 811 dt->dt_ff[newfd] = ff; 812 ff = NULL; 813 } 814 fd_used(fdp, newfd); 815 mutex_exit(&fdp->fd_lock); 816 817 dt->dt_ff[newfd]->ff_exclose = (flags & O_CLOEXEC) != 0; 818 fp->f_flag |= flags & (FNONBLOCK|FNOSIGPIPE); 819 /* Slot is now allocated. Insert copy of the file. */ 820 fd_affix(curproc, fp, newfd); 821 if (ff != NULL) { 822 cv_destroy(&ff->ff_closing); 823 kmem_free(ff, sizeof(*ff)); 824 } 825 return 0; 826 } 827 828 /* 829 * Drop reference to a file structure. 830 */ 831 int 832 closef(file_t *fp) 833 { 834 struct flock lf; 835 int error; 836 837 /* 838 * Drop reference. If referenced elsewhere it's still open 839 * and we have nothing more to do. 840 */ 841 mutex_enter(&fp->f_lock); 842 KASSERT(fp->f_count > 0); 843 if (--fp->f_count > 0) { 844 mutex_exit(&fp->f_lock); 845 return 0; 846 } 847 KASSERT(fp->f_count == 0); 848 mutex_exit(&fp->f_lock); 849 850 /* We held the last reference - release locks, close and free. */ 851 if (fp->f_ops->fo_advlock == NULL) { 852 KASSERT((fp->f_flag & FHASLOCK) == 0); 853 } else if (fp->f_flag & FHASLOCK) { 854 lf.l_whence = SEEK_SET; 855 lf.l_start = 0; 856 lf.l_len = 0; 857 lf.l_type = F_UNLCK; 858 (void)(*fp->f_ops->fo_advlock)(fp, fp, F_UNLCK, &lf, F_FLOCK); 859 } 860 if (fp->f_ops != NULL) { 861 error = (*fp->f_ops->fo_close)(fp); 862 } else { 863 error = 0; 864 } 865 KASSERT(fp->f_count == 0); 866 KASSERT(fp->f_cred != NULL); 867 pool_cache_put(file_cache, fp); 868 869 return error; 870 } 871 872 /* 873 * Allocate a file descriptor for the process. 874 * 875 * Future idea for experimentation: replace all of this with radixtree. 876 */ 877 int 878 fd_alloc(proc_t *p, int want, int *result) 879 { 880 filedesc_t *fdp = p->p_fd; 881 int i, lim, last, error, hi; 882 u_int off; 883 fdtab_t *dt; 884 885 KASSERT(p == curproc || p == &proc0); 886 887 /* 888 * Search for a free descriptor starting at the higher 889 * of want or fd_freefile. 890 */ 891 mutex_enter(&fdp->fd_lock); 892 fd_checkmaps(fdp); 893 dt = fdp->fd_dt; 894 KASSERT(dt->dt_ff[0] == (fdfile_t *)fdp->fd_dfdfile[0]); 895 lim = uimin((int)p->p_rlimit[RLIMIT_NOFILE].rlim_cur, maxfiles); 896 last = uimin(dt->dt_nfiles, lim); 897 898 for (;;) { 899 if ((i = want) < fdp->fd_freefile) 900 i = fdp->fd_freefile; 901 off = i >> NDENTRYSHIFT; 902 hi = fd_next_zero(fdp, fdp->fd_himap, off, 903 (last + NDENTRIES - 1) >> NDENTRYSHIFT); 904 if (hi == -1) 905 break; 906 i = fd_next_zero(fdp, &fdp->fd_lomap[hi], 907 hi > off ? 0 : i & NDENTRYMASK, NDENTRIES); 908 if (i == -1) { 909 /* 910 * Free file descriptor in this block was 911 * below want, try again with higher want. 912 */ 913 want = (hi + 1) << NDENTRYSHIFT; 914 continue; 915 } 916 i += (hi << NDENTRYSHIFT); 917 if (i >= last) { 918 break; 919 } 920 if (dt->dt_ff[i] == NULL) { 921 KASSERT(i >= NDFDFILE); 922 dt->dt_ff[i] = kmem_alloc(sizeof(fdfile_t), KM_SLEEP); 923 fdfile_ctor(dt->dt_ff[i]); 924 } 925 KASSERT(dt->dt_ff[i]->ff_file == NULL); 926 fd_used(fdp, i); 927 if (want <= fdp->fd_freefile) { 928 fdp->fd_freefile = i; 929 } 930 *result = i; 931 KASSERT(i >= NDFDFILE || 932 dt->dt_ff[i] == (fdfile_t *)fdp->fd_dfdfile[i]); 933 fd_checkmaps(fdp); 934 mutex_exit(&fdp->fd_lock); 935 return 0; 936 } 937 938 /* No space in current array. Let the caller expand and retry. */ 939 error = (dt->dt_nfiles >= lim) ? EMFILE : ENOSPC; 940 mutex_exit(&fdp->fd_lock); 941 return error; 942 } 943 944 /* 945 * Allocate memory for a descriptor table. 946 */ 947 static fdtab_t * 948 fd_dtab_alloc(int n) 949 { 950 fdtab_t *dt; 951 size_t sz; 952 953 KASSERT(n > NDFILE); 954 955 sz = sizeof(*dt) + (n - NDFILE) * sizeof(dt->dt_ff[0]); 956 dt = kmem_alloc(sz, KM_SLEEP); 957 #ifdef DIAGNOSTIC 958 memset(dt, 0xff, sz); 959 #endif 960 dt->dt_nfiles = n; 961 dt->dt_link = NULL; 962 return dt; 963 } 964 965 /* 966 * Free a descriptor table, and all tables linked for deferred free. 967 */ 968 static void 969 fd_dtab_free(fdtab_t *dt) 970 { 971 fdtab_t *next; 972 size_t sz; 973 974 do { 975 next = dt->dt_link; 976 KASSERT(dt->dt_nfiles > NDFILE); 977 sz = sizeof(*dt) + 978 (dt->dt_nfiles - NDFILE) * sizeof(dt->dt_ff[0]); 979 #ifdef DIAGNOSTIC 980 memset(dt, 0xff, sz); 981 #endif 982 kmem_free(dt, sz); 983 dt = next; 984 } while (dt != NULL); 985 } 986 987 /* 988 * Allocate descriptor bitmap. 989 */ 990 static void 991 fd_map_alloc(int n, uint32_t **lo, uint32_t **hi) 992 { 993 uint8_t *ptr; 994 size_t szlo, szhi; 995 996 KASSERT(n > NDENTRIES); 997 998 szlo = NDLOSLOTS(n) * sizeof(uint32_t); 999 szhi = NDHISLOTS(n) * sizeof(uint32_t); 1000 ptr = kmem_alloc(szlo + szhi, KM_SLEEP); 1001 *lo = (uint32_t *)ptr; 1002 *hi = (uint32_t *)(ptr + szlo); 1003 } 1004 1005 /* 1006 * Free descriptor bitmap. 1007 */ 1008 static void 1009 fd_map_free(int n, uint32_t *lo, uint32_t *hi) 1010 { 1011 size_t szlo, szhi; 1012 1013 KASSERT(n > NDENTRIES); 1014 1015 szlo = NDLOSLOTS(n) * sizeof(uint32_t); 1016 szhi = NDHISLOTS(n) * sizeof(uint32_t); 1017 KASSERT(hi == (uint32_t *)((uint8_t *)lo + szlo)); 1018 kmem_free(lo, szlo + szhi); 1019 } 1020 1021 /* 1022 * Expand a process' descriptor table. 1023 */ 1024 void 1025 fd_tryexpand(proc_t *p) 1026 { 1027 filedesc_t *fdp; 1028 int i, numfiles, oldnfiles; 1029 fdtab_t *newdt, *dt; 1030 uint32_t *newhimap, *newlomap; 1031 1032 KASSERT(p == curproc || p == &proc0); 1033 1034 fdp = p->p_fd; 1035 newhimap = NULL; 1036 newlomap = NULL; 1037 oldnfiles = atomic_load_consume(&fdp->fd_dt)->dt_nfiles; 1038 1039 if (oldnfiles < NDEXTENT) 1040 numfiles = NDEXTENT; 1041 else 1042 numfiles = 2 * oldnfiles; 1043 1044 newdt = fd_dtab_alloc(numfiles); 1045 if (NDHISLOTS(numfiles) > NDHISLOTS(oldnfiles)) { 1046 fd_map_alloc(numfiles, &newlomap, &newhimap); 1047 } 1048 1049 mutex_enter(&fdp->fd_lock); 1050 dt = fdp->fd_dt; 1051 KASSERT(dt->dt_ff[0] == (fdfile_t *)fdp->fd_dfdfile[0]); 1052 if (dt->dt_nfiles != oldnfiles) { 1053 /* fdp changed; caller must retry */ 1054 mutex_exit(&fdp->fd_lock); 1055 fd_dtab_free(newdt); 1056 if (NDHISLOTS(numfiles) > NDHISLOTS(oldnfiles)) { 1057 fd_map_free(numfiles, newlomap, newhimap); 1058 } 1059 return; 1060 } 1061 1062 /* Copy the existing descriptor table and zero the new portion. */ 1063 i = sizeof(fdfile_t *) * oldnfiles; 1064 memcpy(newdt->dt_ff, dt->dt_ff, i); 1065 memset((uint8_t *)newdt->dt_ff + i, 0, 1066 numfiles * sizeof(fdfile_t *) - i); 1067 1068 /* 1069 * Link old descriptor array into list to be discarded. We defer 1070 * freeing until the last reference to the descriptor table goes 1071 * away (usually process exit). This allows us to do lockless 1072 * lookups in fd_getfile(). 1073 */ 1074 if (oldnfiles > NDFILE) { 1075 if (fdp->fd_refcnt > 1) { 1076 newdt->dt_link = dt; 1077 } else { 1078 fd_dtab_free(dt); 1079 } 1080 } 1081 1082 if (NDHISLOTS(numfiles) > NDHISLOTS(oldnfiles)) { 1083 i = NDHISLOTS(oldnfiles) * sizeof(uint32_t); 1084 memcpy(newhimap, fdp->fd_himap, i); 1085 memset((uint8_t *)newhimap + i, 0, 1086 NDHISLOTS(numfiles) * sizeof(uint32_t) - i); 1087 1088 i = NDLOSLOTS(oldnfiles) * sizeof(uint32_t); 1089 memcpy(newlomap, fdp->fd_lomap, i); 1090 memset((uint8_t *)newlomap + i, 0, 1091 NDLOSLOTS(numfiles) * sizeof(uint32_t) - i); 1092 1093 if (NDHISLOTS(oldnfiles) > NDHISLOTS(NDFILE)) { 1094 fd_map_free(oldnfiles, fdp->fd_lomap, fdp->fd_himap); 1095 } 1096 fdp->fd_himap = newhimap; 1097 fdp->fd_lomap = newlomap; 1098 } 1099 1100 /* 1101 * All other modifications must become globally visible before 1102 * the change to fd_dt. See fd_getfile(). 1103 */ 1104 atomic_store_release(&fdp->fd_dt, newdt); 1105 KASSERT(newdt->dt_ff[0] == (fdfile_t *)fdp->fd_dfdfile[0]); 1106 fd_checkmaps(fdp); 1107 mutex_exit(&fdp->fd_lock); 1108 } 1109 1110 /* 1111 * Create a new open file structure and allocate a file descriptor 1112 * for the current process. 1113 */ 1114 int 1115 fd_allocfile(file_t **resultfp, int *resultfd) 1116 { 1117 proc_t *p = curproc; 1118 kauth_cred_t cred; 1119 file_t *fp; 1120 int error; 1121 1122 while ((error = fd_alloc(p, 0, resultfd)) != 0) { 1123 if (error != ENOSPC) { 1124 return error; 1125 } 1126 fd_tryexpand(p); 1127 } 1128 1129 fp = pool_cache_get(file_cache, PR_WAITOK); 1130 if (fp == NULL) { 1131 fd_abort(p, NULL, *resultfd); 1132 return ENFILE; 1133 } 1134 KASSERT(fp->f_count == 0); 1135 KASSERT(fp->f_msgcount == 0); 1136 KASSERT(fp->f_unpcount == 0); 1137 1138 /* Replace cached credentials if not what we need. */ 1139 cred = curlwp->l_cred; 1140 if (__predict_false(cred != fp->f_cred)) { 1141 kauth_cred_free(fp->f_cred); 1142 kauth_cred_hold(cred); 1143 fp->f_cred = cred; 1144 } 1145 1146 /* 1147 * Don't allow recycled files to be scanned. 1148 * See uipc_usrreq.c. 1149 */ 1150 if (__predict_false((fp->f_flag & FSCAN) != 0)) { 1151 mutex_enter(&fp->f_lock); 1152 atomic_and_uint(&fp->f_flag, ~FSCAN); 1153 mutex_exit(&fp->f_lock); 1154 } 1155 1156 fp->f_advice = 0; 1157 fp->f_offset = 0; 1158 *resultfp = fp; 1159 1160 return 0; 1161 } 1162 1163 /* 1164 * Successful creation of a new descriptor: make visible to the process. 1165 */ 1166 void 1167 fd_affix(proc_t *p, file_t *fp, unsigned fd) 1168 { 1169 fdfile_t *ff; 1170 filedesc_t *fdp; 1171 fdtab_t *dt; 1172 1173 KASSERT(p == curproc || p == &proc0); 1174 1175 /* Add a reference to the file structure. */ 1176 mutex_enter(&fp->f_lock); 1177 fp->f_count++; 1178 mutex_exit(&fp->f_lock); 1179 1180 /* 1181 * Insert the new file into the descriptor slot. 1182 */ 1183 fdp = p->p_fd; 1184 dt = atomic_load_consume(&fdp->fd_dt); 1185 ff = dt->dt_ff[fd]; 1186 1187 KASSERT(ff != NULL); 1188 KASSERT(ff->ff_file == NULL); 1189 KASSERT(ff->ff_allocated); 1190 KASSERT(fd_isused(fdp, fd)); 1191 KASSERT(fd >= NDFDFILE || ff == (fdfile_t *)fdp->fd_dfdfile[fd]); 1192 1193 /* No need to lock in order to make file initially visible. */ 1194 atomic_store_release(&ff->ff_file, fp); 1195 } 1196 1197 /* 1198 * Abort creation of a new descriptor: free descriptor slot and file. 1199 */ 1200 void 1201 fd_abort(proc_t *p, file_t *fp, unsigned fd) 1202 { 1203 filedesc_t *fdp; 1204 fdfile_t *ff; 1205 1206 KASSERT(p == curproc || p == &proc0); 1207 1208 fdp = p->p_fd; 1209 ff = atomic_load_consume(&fdp->fd_dt)->dt_ff[fd]; 1210 ff->ff_exclose = false; 1211 1212 KASSERT(fd >= NDFDFILE || ff == (fdfile_t *)fdp->fd_dfdfile[fd]); 1213 1214 mutex_enter(&fdp->fd_lock); 1215 KASSERT(fd_isused(fdp, fd)); 1216 fd_unused(fdp, fd); 1217 mutex_exit(&fdp->fd_lock); 1218 1219 if (fp != NULL) { 1220 KASSERT(fp->f_count == 0); 1221 KASSERT(fp->f_cred != NULL); 1222 pool_cache_put(file_cache, fp); 1223 } 1224 } 1225 1226 static int 1227 file_ctor(void *arg, void *obj, int flags) 1228 { 1229 /* 1230 * It's easy to exhaust the open file limit on a system with many 1231 * CPUs due to caching. Allow a bit of leeway to reduce the element 1232 * of surprise. 1233 */ 1234 u_int slop = PCG_NOBJECTS_NORMAL * (ncpu - 1); 1235 file_t *fp = obj; 1236 1237 memset(fp, 0, sizeof(*fp)); 1238 1239 mutex_enter(&filelist_lock); 1240 if (__predict_false(nfiles >= slop + maxfiles)) { 1241 mutex_exit(&filelist_lock); 1242 tablefull("file", "increase kern.maxfiles or MAXFILES"); 1243 return ENFILE; 1244 } 1245 nfiles++; 1246 LIST_INSERT_HEAD(&filehead, fp, f_list); 1247 mutex_init(&fp->f_lock, MUTEX_DEFAULT, IPL_NONE); 1248 fp->f_cred = curlwp->l_cred; 1249 kauth_cred_hold(fp->f_cred); 1250 mutex_exit(&filelist_lock); 1251 1252 return 0; 1253 } 1254 1255 static void 1256 file_dtor(void *arg, void *obj) 1257 { 1258 file_t *fp = obj; 1259 1260 mutex_enter(&filelist_lock); 1261 nfiles--; 1262 LIST_REMOVE(fp, f_list); 1263 mutex_exit(&filelist_lock); 1264 1265 KASSERT(fp->f_count == 0); 1266 kauth_cred_free(fp->f_cred); 1267 mutex_destroy(&fp->f_lock); 1268 } 1269 1270 static void 1271 fdfile_ctor(fdfile_t *ff) 1272 { 1273 1274 memset(ff, 0, sizeof(*ff)); 1275 cv_init(&ff->ff_closing, "fdclose"); 1276 } 1277 1278 static void 1279 fdfile_dtor(fdfile_t *ff) 1280 { 1281 1282 cv_destroy(&ff->ff_closing); 1283 } 1284 1285 file_t * 1286 fgetdummy(void) 1287 { 1288 file_t *fp; 1289 1290 fp = kmem_zalloc(sizeof(*fp), KM_SLEEP); 1291 mutex_init(&fp->f_lock, MUTEX_DEFAULT, IPL_NONE); 1292 return fp; 1293 } 1294 1295 void 1296 fputdummy(file_t *fp) 1297 { 1298 1299 mutex_destroy(&fp->f_lock); 1300 kmem_free(fp, sizeof(*fp)); 1301 } 1302 1303 /* 1304 * Create an initial filedesc structure. 1305 */ 1306 filedesc_t * 1307 fd_init(filedesc_t *fdp) 1308 { 1309 #ifdef DIAGNOSTIC 1310 unsigned fd; 1311 #endif 1312 1313 if (__predict_true(fdp == NULL)) { 1314 fdp = pool_cache_get(filedesc_cache, PR_WAITOK); 1315 } else { 1316 KASSERT(fdp == &filedesc0); 1317 filedesc_ctor(NULL, fdp, PR_WAITOK); 1318 } 1319 1320 #ifdef DIAGNOSTIC 1321 KASSERT(fdp->fd_lastfile == -1); 1322 KASSERT(fdp->fd_lastkqfile == -1); 1323 KASSERT(fdp->fd_knhash == NULL); 1324 KASSERT(fdp->fd_freefile == 0); 1325 KASSERT(fdp->fd_exclose == false); 1326 KASSERT(fdp->fd_dt == &fdp->fd_dtbuiltin); 1327 KASSERT(fdp->fd_dtbuiltin.dt_nfiles == NDFILE); 1328 for (fd = 0; fd < NDFDFILE; fd++) { 1329 KASSERT(fdp->fd_dtbuiltin.dt_ff[fd] == 1330 (fdfile_t *)fdp->fd_dfdfile[fd]); 1331 } 1332 for (fd = NDFDFILE; fd < NDFILE; fd++) { 1333 KASSERT(fdp->fd_dtbuiltin.dt_ff[fd] == NULL); 1334 } 1335 KASSERT(fdp->fd_himap == fdp->fd_dhimap); 1336 KASSERT(fdp->fd_lomap == fdp->fd_dlomap); 1337 #endif /* DIAGNOSTIC */ 1338 1339 fdp->fd_refcnt = 1; 1340 fd_checkmaps(fdp); 1341 1342 return fdp; 1343 } 1344 1345 /* 1346 * Initialize a file descriptor table. 1347 */ 1348 static int 1349 filedesc_ctor(void *arg, void *obj, int flag) 1350 { 1351 filedesc_t *fdp = obj; 1352 fdfile_t **ffp; 1353 int i; 1354 1355 memset(fdp, 0, sizeof(*fdp)); 1356 mutex_init(&fdp->fd_lock, MUTEX_DEFAULT, IPL_NONE); 1357 fdp->fd_lastfile = -1; 1358 fdp->fd_lastkqfile = -1; 1359 fdp->fd_dt = &fdp->fd_dtbuiltin; 1360 fdp->fd_dtbuiltin.dt_nfiles = NDFILE; 1361 fdp->fd_himap = fdp->fd_dhimap; 1362 fdp->fd_lomap = fdp->fd_dlomap; 1363 1364 CTASSERT(sizeof(fdp->fd_dfdfile[0]) >= sizeof(fdfile_t)); 1365 for (i = 0, ffp = fdp->fd_dt->dt_ff; i < NDFDFILE; i++, ffp++) { 1366 fdfile_ctor(*ffp = (fdfile_t *)fdp->fd_dfdfile[i]); 1367 } 1368 1369 return 0; 1370 } 1371 1372 static void 1373 filedesc_dtor(void *arg, void *obj) 1374 { 1375 filedesc_t *fdp = obj; 1376 int i; 1377 1378 for (i = 0; i < NDFDFILE; i++) { 1379 fdfile_dtor((fdfile_t *)fdp->fd_dfdfile[i]); 1380 } 1381 1382 mutex_destroy(&fdp->fd_lock); 1383 } 1384 1385 /* 1386 * Make p share curproc's filedesc structure. 1387 */ 1388 void 1389 fd_share(struct proc *p) 1390 { 1391 filedesc_t *fdp; 1392 1393 fdp = curlwp->l_fd; 1394 p->p_fd = fdp; 1395 atomic_inc_uint(&fdp->fd_refcnt); 1396 } 1397 1398 /* 1399 * Acquire a hold on a filedesc structure. 1400 */ 1401 void 1402 fd_hold(lwp_t *l) 1403 { 1404 filedesc_t *fdp = l->l_fd; 1405 1406 atomic_inc_uint(&fdp->fd_refcnt); 1407 } 1408 1409 /* 1410 * Copy a filedesc structure. 1411 */ 1412 filedesc_t * 1413 fd_copy(void) 1414 { 1415 filedesc_t *newfdp, *fdp; 1416 fdfile_t *ff, **ffp, **nffp, *ff2; 1417 int i, j, numfiles, lastfile, newlast; 1418 file_t *fp; 1419 fdtab_t *newdt; 1420 1421 fdp = curproc->p_fd; 1422 newfdp = pool_cache_get(filedesc_cache, PR_WAITOK); 1423 newfdp->fd_refcnt = 1; 1424 1425 #ifdef DIAGNOSTIC 1426 KASSERT(newfdp->fd_lastfile == -1); 1427 KASSERT(newfdp->fd_lastkqfile == -1); 1428 KASSERT(newfdp->fd_knhash == NULL); 1429 KASSERT(newfdp->fd_freefile == 0); 1430 KASSERT(newfdp->fd_exclose == false); 1431 KASSERT(newfdp->fd_dt == &newfdp->fd_dtbuiltin); 1432 KASSERT(newfdp->fd_dtbuiltin.dt_nfiles == NDFILE); 1433 for (i = 0; i < NDFDFILE; i++) { 1434 KASSERT(newfdp->fd_dtbuiltin.dt_ff[i] == 1435 (fdfile_t *)&newfdp->fd_dfdfile[i]); 1436 } 1437 for (i = NDFDFILE; i < NDFILE; i++) { 1438 KASSERT(newfdp->fd_dtbuiltin.dt_ff[i] == NULL); 1439 } 1440 #endif /* DIAGNOSTIC */ 1441 1442 mutex_enter(&fdp->fd_lock); 1443 fd_checkmaps(fdp); 1444 numfiles = fdp->fd_dt->dt_nfiles; 1445 lastfile = fdp->fd_lastfile; 1446 1447 /* 1448 * If the number of open files fits in the internal arrays 1449 * of the open file structure, use them, otherwise allocate 1450 * additional memory for the number of descriptors currently 1451 * in use. 1452 */ 1453 if (lastfile < NDFILE) { 1454 i = NDFILE; 1455 newdt = newfdp->fd_dt; 1456 KASSERT(newfdp->fd_dt == &newfdp->fd_dtbuiltin); 1457 } else { 1458 /* 1459 * Compute the smallest multiple of NDEXTENT needed 1460 * for the file descriptors currently in use, 1461 * allowing the table to shrink. 1462 */ 1463 i = numfiles; 1464 while (i >= 2 * NDEXTENT && i > lastfile * 2) { 1465 i /= 2; 1466 } 1467 KASSERT(i > NDFILE); 1468 newdt = fd_dtab_alloc(i); 1469 newfdp->fd_dt = newdt; 1470 memcpy(newdt->dt_ff, newfdp->fd_dtbuiltin.dt_ff, 1471 NDFDFILE * sizeof(fdfile_t **)); 1472 memset(newdt->dt_ff + NDFDFILE, 0, 1473 (i - NDFDFILE) * sizeof(fdfile_t **)); 1474 } 1475 if (NDHISLOTS(i) <= NDHISLOTS(NDFILE)) { 1476 newfdp->fd_himap = newfdp->fd_dhimap; 1477 newfdp->fd_lomap = newfdp->fd_dlomap; 1478 } else { 1479 fd_map_alloc(i, &newfdp->fd_lomap, &newfdp->fd_himap); 1480 KASSERT(i >= NDENTRIES * NDENTRIES); 1481 memset(newfdp->fd_himap, 0, NDHISLOTS(i)*sizeof(uint32_t)); 1482 memset(newfdp->fd_lomap, 0, NDLOSLOTS(i)*sizeof(uint32_t)); 1483 } 1484 newfdp->fd_freefile = fdp->fd_freefile; 1485 newfdp->fd_exclose = fdp->fd_exclose; 1486 1487 ffp = fdp->fd_dt->dt_ff; 1488 nffp = newdt->dt_ff; 1489 newlast = -1; 1490 for (i = 0; i <= lastfile; i++, ffp++, nffp++) { 1491 KASSERT(i >= NDFDFILE || 1492 *nffp == (fdfile_t *)newfdp->fd_dfdfile[i]); 1493 ff = *ffp; 1494 if (ff == NULL || 1495 (fp = atomic_load_consume(&ff->ff_file)) == NULL) { 1496 /* Descriptor unused, or descriptor half open. */ 1497 KASSERT(!fd_isused(newfdp, i)); 1498 continue; 1499 } 1500 if (__predict_false(fp->f_type == DTYPE_KQUEUE)) { 1501 /* kqueue descriptors cannot be copied. */ 1502 if (i < newfdp->fd_freefile) { 1503 newfdp->fd_freefile = i; 1504 } 1505 continue; 1506 } 1507 /* It's active: add a reference to the file. */ 1508 mutex_enter(&fp->f_lock); 1509 fp->f_count++; 1510 mutex_exit(&fp->f_lock); 1511 1512 /* Allocate an fdfile_t to represent it. */ 1513 if (i >= NDFDFILE) { 1514 ff2 = kmem_alloc(sizeof(*ff2), KM_SLEEP); 1515 fdfile_ctor(ff2); 1516 *nffp = ff2; 1517 } else { 1518 ff2 = newdt->dt_ff[i]; 1519 } 1520 ff2->ff_file = fp; 1521 ff2->ff_exclose = ff->ff_exclose; 1522 ff2->ff_allocated = true; 1523 1524 /* Fix up bitmaps. */ 1525 j = i >> NDENTRYSHIFT; 1526 KASSERT((newfdp->fd_lomap[j] & (1U << (i & NDENTRYMASK))) == 0); 1527 newfdp->fd_lomap[j] |= 1U << (i & NDENTRYMASK); 1528 if (__predict_false(newfdp->fd_lomap[j] == ~0)) { 1529 KASSERT((newfdp->fd_himap[j >> NDENTRYSHIFT] & 1530 (1U << (j & NDENTRYMASK))) == 0); 1531 newfdp->fd_himap[j >> NDENTRYSHIFT] |= 1532 1U << (j & NDENTRYMASK); 1533 } 1534 newlast = i; 1535 } 1536 KASSERT(newdt->dt_ff[0] == (fdfile_t *)newfdp->fd_dfdfile[0]); 1537 newfdp->fd_lastfile = newlast; 1538 fd_checkmaps(newfdp); 1539 mutex_exit(&fdp->fd_lock); 1540 1541 return newfdp; 1542 } 1543 1544 /* 1545 * Release a filedesc structure. 1546 */ 1547 void 1548 fd_free(void) 1549 { 1550 fdfile_t *ff; 1551 file_t *fp; 1552 int fd, nf; 1553 fdtab_t *dt; 1554 lwp_t * const l = curlwp; 1555 filedesc_t * const fdp = l->l_fd; 1556 const bool noadvlock = (l->l_proc->p_flag & PK_ADVLOCK) == 0; 1557 1558 KASSERT(atomic_load_consume(&fdp->fd_dt)->dt_ff[0] == 1559 (fdfile_t *)fdp->fd_dfdfile[0]); 1560 KASSERT(fdp->fd_dtbuiltin.dt_nfiles == NDFILE); 1561 KASSERT(fdp->fd_dtbuiltin.dt_link == NULL); 1562 1563 membar_release(); 1564 if (atomic_dec_uint_nv(&fdp->fd_refcnt) > 0) 1565 return; 1566 membar_acquire(); 1567 1568 /* 1569 * Close any files that the process holds open. 1570 */ 1571 dt = fdp->fd_dt; 1572 fd_checkmaps(fdp); 1573 #ifdef DEBUG 1574 fdp->fd_refcnt = -1; /* see fd_checkmaps */ 1575 #endif 1576 for (fd = 0, nf = dt->dt_nfiles; fd < nf; fd++) { 1577 ff = dt->dt_ff[fd]; 1578 KASSERT(fd >= NDFDFILE || 1579 ff == (fdfile_t *)fdp->fd_dfdfile[fd]); 1580 if (ff == NULL) 1581 continue; 1582 if ((fp = atomic_load_consume(&ff->ff_file)) != NULL) { 1583 /* 1584 * Must use fd_close() here if there is 1585 * a reference from kqueue or we might have posix 1586 * advisory locks. 1587 */ 1588 if (__predict_true(ff->ff_refcnt == 0) && 1589 (noadvlock || fp->f_type != DTYPE_VNODE)) { 1590 ff->ff_file = NULL; 1591 ff->ff_exclose = false; 1592 ff->ff_allocated = false; 1593 closef(fp); 1594 } else { 1595 ff->ff_refcnt++; 1596 fd_close(fd); 1597 } 1598 } 1599 KASSERT(ff->ff_refcnt == 0); 1600 KASSERT(ff->ff_file == NULL); 1601 KASSERT(!ff->ff_exclose); 1602 KASSERT(!ff->ff_allocated); 1603 if (fd >= NDFDFILE) { 1604 cv_destroy(&ff->ff_closing); 1605 kmem_free(ff, sizeof(*ff)); 1606 dt->dt_ff[fd] = NULL; 1607 } 1608 } 1609 1610 /* 1611 * Clean out the descriptor table for the next user and return 1612 * to the cache. 1613 */ 1614 if (__predict_false(dt != &fdp->fd_dtbuiltin)) { 1615 fd_dtab_free(fdp->fd_dt); 1616 /* Otherwise, done above. */ 1617 memset(&fdp->fd_dtbuiltin.dt_ff[NDFDFILE], 0, 1618 (NDFILE - NDFDFILE) * sizeof(fdp->fd_dtbuiltin.dt_ff[0])); 1619 fdp->fd_dt = &fdp->fd_dtbuiltin; 1620 } 1621 if (__predict_false(NDHISLOTS(nf) > NDHISLOTS(NDFILE))) { 1622 KASSERT(fdp->fd_himap != fdp->fd_dhimap); 1623 KASSERT(fdp->fd_lomap != fdp->fd_dlomap); 1624 fd_map_free(nf, fdp->fd_lomap, fdp->fd_himap); 1625 } 1626 if (__predict_false(fdp->fd_knhash != NULL)) { 1627 hashdone(fdp->fd_knhash, HASH_LIST, fdp->fd_knhashmask); 1628 fdp->fd_knhash = NULL; 1629 fdp->fd_knhashmask = 0; 1630 } else { 1631 KASSERT(fdp->fd_knhashmask == 0); 1632 } 1633 fdp->fd_dt = &fdp->fd_dtbuiltin; 1634 fdp->fd_lastkqfile = -1; 1635 fdp->fd_lastfile = -1; 1636 fdp->fd_freefile = 0; 1637 fdp->fd_exclose = false; 1638 memset(&fdp->fd_startzero, 0, sizeof(*fdp) - 1639 offsetof(filedesc_t, fd_startzero)); 1640 fdp->fd_himap = fdp->fd_dhimap; 1641 fdp->fd_lomap = fdp->fd_dlomap; 1642 KASSERT(fdp->fd_dtbuiltin.dt_nfiles == NDFILE); 1643 KASSERT(fdp->fd_dtbuiltin.dt_link == NULL); 1644 KASSERT(fdp->fd_dt == &fdp->fd_dtbuiltin); 1645 #ifdef DEBUG 1646 fdp->fd_refcnt = 0; /* see fd_checkmaps */ 1647 #endif 1648 fd_checkmaps(fdp); 1649 pool_cache_put(filedesc_cache, fdp); 1650 } 1651 1652 /* 1653 * File Descriptor pseudo-device driver (/dev/fd/). 1654 * 1655 * Opening minor device N dup()s the file (if any) connected to file 1656 * descriptor N belonging to the calling process. Note that this driver 1657 * consists of only the ``open()'' routine, because all subsequent 1658 * references to this file will be direct to the other driver. 1659 */ 1660 static int 1661 filedescopen(dev_t dev, int mode, int type, lwp_t *l) 1662 { 1663 1664 /* 1665 * XXX Kludge: set dupfd to contain the value of the 1666 * the file descriptor being sought for duplication. The error 1667 * return ensures that the vnode for this device will be released 1668 * by vn_open. Open will detect this special error and take the 1669 * actions in fd_dupopen below. Other callers of vn_open or VOP_OPEN 1670 * will simply report the error. 1671 */ 1672 l->l_dupfd = minor(dev); /* XXX */ 1673 return EDUPFD; 1674 } 1675 1676 /* 1677 * Duplicate the specified descriptor to a free descriptor. 1678 * 1679 * old is the original fd. 1680 * moveit is true if we should move rather than duplicate. 1681 * flags are the open flags (converted from O_* to F*). 1682 * newp returns the new fd on success. 1683 * 1684 * These two cases are produced by the EDUPFD and EMOVEFD magic 1685 * errnos, but in the interest of removing that regrettable interface, 1686 * vn_open has been changed to intercept them. Now vn_open returns 1687 * either a vnode or a filehandle, and the filehandle is accompanied 1688 * by a boolean that says whether we should dup (moveit == false) or 1689 * move (moveit == true) the fd. 1690 * 1691 * The dup case is used by /dev/stderr, /proc/self/fd, and such. The 1692 * move case is used by cloner devices that allocate a fd of their 1693 * own (a layering violation that should go away eventually) that 1694 * then needs to be put in the place open() expects it. 1695 */ 1696 int 1697 fd_dupopen(int old, bool moveit, int flags, int *newp) 1698 { 1699 filedesc_t *fdp; 1700 fdfile_t *ff; 1701 file_t *fp; 1702 fdtab_t *dt; 1703 int error; 1704 1705 if ((fp = fd_getfile(old)) == NULL) { 1706 return EBADF; 1707 } 1708 fdp = curlwp->l_fd; 1709 dt = atomic_load_consume(&fdp->fd_dt); 1710 ff = dt->dt_ff[old]; 1711 1712 /* 1713 * There are two cases of interest here. 1714 * 1715 * 1. moveit == false (used to be the EDUPFD magic errno): 1716 * simply dup (old) to file descriptor (new) and return. 1717 * 1718 * 2. moveit == true (used to be the EMOVEFD magic errno): 1719 * steal away the file structure from (old) and store it in 1720 * (new). (old) is effectively closed by this operation. 1721 */ 1722 if (moveit == false) { 1723 /* 1724 * Check that the mode the file is being opened for is a 1725 * subset of the mode of the existing descriptor. 1726 */ 1727 if (((flags & (FREAD|FWRITE)) | fp->f_flag) != fp->f_flag) { 1728 error = EACCES; 1729 goto out; 1730 } 1731 1732 /* Copy it. */ 1733 error = fd_dup(fp, 0, newp, ff->ff_exclose); 1734 } else { 1735 /* Copy it. */ 1736 error = fd_dup(fp, 0, newp, ff->ff_exclose); 1737 if (error != 0) { 1738 goto out; 1739 } 1740 1741 /* Steal away the file pointer from 'old'. */ 1742 (void)fd_close(old); 1743 return 0; 1744 } 1745 1746 out: 1747 fd_putfile(old); 1748 return error; 1749 } 1750 1751 /* 1752 * Close open files on exec. 1753 */ 1754 void 1755 fd_closeexec(void) 1756 { 1757 proc_t *p; 1758 filedesc_t *fdp; 1759 fdfile_t *ff; 1760 lwp_t *l; 1761 fdtab_t *dt; 1762 int fd; 1763 1764 l = curlwp; 1765 p = l->l_proc; 1766 fdp = p->p_fd; 1767 1768 if (fdp->fd_refcnt > 1) { 1769 fdp = fd_copy(); 1770 fd_free(); 1771 p->p_fd = fdp; 1772 l->l_fd = fdp; 1773 } 1774 if (!fdp->fd_exclose) { 1775 return; 1776 } 1777 fdp->fd_exclose = false; 1778 dt = atomic_load_consume(&fdp->fd_dt); 1779 1780 for (fd = 0; fd <= fdp->fd_lastfile; fd++) { 1781 if ((ff = dt->dt_ff[fd]) == NULL) { 1782 KASSERT(fd >= NDFDFILE); 1783 continue; 1784 } 1785 KASSERT(fd >= NDFDFILE || 1786 ff == (fdfile_t *)fdp->fd_dfdfile[fd]); 1787 if (ff->ff_file == NULL) 1788 continue; 1789 if (ff->ff_exclose) { 1790 /* 1791 * We need a reference to close the file. 1792 * No other threads can see the fdfile_t at 1793 * this point, so don't bother locking. 1794 */ 1795 KASSERT((ff->ff_refcnt & FR_CLOSING) == 0); 1796 ff->ff_refcnt++; 1797 fd_close(fd); 1798 } 1799 } 1800 } 1801 1802 /* 1803 * Sets descriptor owner. If the owner is a process, 'pgid' 1804 * is set to positive value, process ID. If the owner is process group, 1805 * 'pgid' is set to -pg_id. 1806 */ 1807 int 1808 fsetown(pid_t *pgid, u_long cmd, const void *data) 1809 { 1810 pid_t id = *(const pid_t *)data; 1811 int error; 1812 1813 if (id == INT_MIN) 1814 return EINVAL; 1815 1816 switch (cmd) { 1817 case TIOCSPGRP: 1818 if (id < 0) 1819 return EINVAL; 1820 id = -id; 1821 break; 1822 default: 1823 break; 1824 } 1825 if (id > 0) { 1826 mutex_enter(&proc_lock); 1827 error = proc_find(id) ? 0 : ESRCH; 1828 mutex_exit(&proc_lock); 1829 } else if (id < 0) { 1830 error = pgid_in_session(curproc, -id); 1831 } else { 1832 error = 0; 1833 } 1834 if (!error) { 1835 *pgid = id; 1836 } 1837 return error; 1838 } 1839 1840 void 1841 fd_set_exclose(struct lwp *l, int fd, bool exclose) 1842 { 1843 filedesc_t *fdp = l->l_fd; 1844 fdfile_t *ff = atomic_load_consume(&fdp->fd_dt)->dt_ff[fd]; 1845 1846 ff->ff_exclose = exclose; 1847 if (exclose) 1848 fdp->fd_exclose = true; 1849 } 1850 1851 /* 1852 * Return descriptor owner information. If the value is positive, 1853 * it's process ID. If it's negative, it's process group ID and 1854 * needs the sign removed before use. 1855 */ 1856 int 1857 fgetown(pid_t pgid, u_long cmd, void *data) 1858 { 1859 1860 switch (cmd) { 1861 case TIOCGPGRP: 1862 *(int *)data = -pgid; 1863 break; 1864 default: 1865 *(int *)data = pgid; 1866 break; 1867 } 1868 return 0; 1869 } 1870 1871 /* 1872 * Send signal to descriptor owner, either process or process group. 1873 */ 1874 void 1875 fownsignal(pid_t pgid, int signo, int code, int band, void *fdescdata) 1876 { 1877 ksiginfo_t ksi; 1878 1879 KASSERT(!cpu_intr_p()); 1880 1881 if (pgid == 0) { 1882 return; 1883 } 1884 1885 KSI_INIT(&ksi); 1886 ksi.ksi_signo = signo; 1887 ksi.ksi_code = code; 1888 ksi.ksi_band = band; 1889 1890 mutex_enter(&proc_lock); 1891 if (pgid > 0) { 1892 struct proc *p1; 1893 1894 p1 = proc_find(pgid); 1895 if (p1 != NULL) { 1896 kpsignal(p1, &ksi, fdescdata); 1897 } 1898 } else { 1899 struct pgrp *pgrp; 1900 1901 KASSERT(pgid < 0); 1902 pgrp = pgrp_find(-pgid); 1903 if (pgrp != NULL) { 1904 kpgsignal(pgrp, &ksi, fdescdata, 0); 1905 } 1906 } 1907 mutex_exit(&proc_lock); 1908 } 1909 1910 int 1911 fd_clone(file_t *fp, unsigned fd, int flag, const struct fileops *fops, 1912 void *data) 1913 { 1914 fdfile_t *ff; 1915 filedesc_t *fdp; 1916 1917 fp->f_flag = flag & FMASK; 1918 fdp = curproc->p_fd; 1919 ff = atomic_load_consume(&fdp->fd_dt)->dt_ff[fd]; 1920 KASSERT(ff != NULL); 1921 ff->ff_exclose = (flag & O_CLOEXEC) != 0; 1922 fp->f_type = DTYPE_MISC; 1923 fp->f_ops = fops; 1924 fp->f_data = data; 1925 curlwp->l_dupfd = fd; 1926 fd_affix(curproc, fp, fd); 1927 1928 return EMOVEFD; 1929 } 1930 1931 int 1932 fnullop_fcntl(file_t *fp, u_int cmd, void *data) 1933 { 1934 1935 if (cmd == F_SETFL) 1936 return 0; 1937 1938 return EOPNOTSUPP; 1939 } 1940 1941 int 1942 fnullop_poll(file_t *fp, int which) 1943 { 1944 1945 return 0; 1946 } 1947 1948 int 1949 fnullop_kqfilter(file_t *fp, struct knote *kn) 1950 { 1951 1952 return EOPNOTSUPP; 1953 } 1954 1955 void 1956 fnullop_restart(file_t *fp) 1957 { 1958 1959 } 1960 1961 int 1962 fbadop_read(file_t *fp, off_t *offset, struct uio *uio, 1963 kauth_cred_t cred, int flags) 1964 { 1965 1966 return EOPNOTSUPP; 1967 } 1968 1969 int 1970 fbadop_write(file_t *fp, off_t *offset, struct uio *uio, 1971 kauth_cred_t cred, int flags) 1972 { 1973 1974 return EOPNOTSUPP; 1975 } 1976 1977 int 1978 fbadop_ioctl(file_t *fp, u_long com, void *data) 1979 { 1980 1981 return EOPNOTSUPP; 1982 } 1983 1984 int 1985 fbadop_stat(file_t *fp, struct stat *sb) 1986 { 1987 1988 return EOPNOTSUPP; 1989 } 1990 1991 int 1992 fbadop_close(file_t *fp) 1993 { 1994 1995 return EOPNOTSUPP; 1996 } 1997 1998 /* 1999 * sysctl routines pertaining to file descriptors 2000 */ 2001 2002 /* Initialized in sysctl_init() for now... */ 2003 extern kmutex_t sysctl_file_marker_lock; 2004 static u_int sysctl_file_marker = 1; 2005 2006 /* 2007 * Expects to be called with proc_lock and sysctl_file_marker_lock locked. 2008 */ 2009 static void 2010 sysctl_file_marker_reset(void) 2011 { 2012 struct proc *p; 2013 2014 PROCLIST_FOREACH(p, &allproc) { 2015 struct filedesc *fd = p->p_fd; 2016 fdtab_t *dt; 2017 u_int i; 2018 2019 mutex_enter(&fd->fd_lock); 2020 dt = fd->fd_dt; 2021 for (i = 0; i < dt->dt_nfiles; i++) { 2022 struct file *fp; 2023 fdfile_t *ff; 2024 2025 if ((ff = dt->dt_ff[i]) == NULL) { 2026 continue; 2027 } 2028 if ((fp = atomic_load_consume(&ff->ff_file)) == NULL) { 2029 continue; 2030 } 2031 fp->f_marker = 0; 2032 } 2033 mutex_exit(&fd->fd_lock); 2034 } 2035 } 2036 2037 /* 2038 * sysctl helper routine for kern.file pseudo-subtree. 2039 */ 2040 static int 2041 sysctl_kern_file(SYSCTLFN_ARGS) 2042 { 2043 const bool allowaddr = get_expose_address(curproc); 2044 struct filelist flist; 2045 int error; 2046 size_t buflen; 2047 struct file *fp, fbuf; 2048 char *start, *where; 2049 struct proc *p; 2050 2051 start = where = oldp; 2052 buflen = *oldlenp; 2053 2054 if (where == NULL) { 2055 /* 2056 * overestimate by 10 files 2057 */ 2058 *oldlenp = sizeof(filehead) + (nfiles + 10) * 2059 sizeof(struct file); 2060 return 0; 2061 } 2062 2063 /* 2064 * first sysctl_copyout filehead 2065 */ 2066 if (buflen < sizeof(filehead)) { 2067 *oldlenp = 0; 2068 return 0; 2069 } 2070 sysctl_unlock(); 2071 if (allowaddr) { 2072 memcpy(&flist, &filehead, sizeof(flist)); 2073 } else { 2074 memset(&flist, 0, sizeof(flist)); 2075 } 2076 error = sysctl_copyout(l, &flist, where, sizeof(flist)); 2077 if (error) { 2078 sysctl_relock(); 2079 return error; 2080 } 2081 buflen -= sizeof(flist); 2082 where += sizeof(flist); 2083 2084 /* 2085 * followed by an array of file structures 2086 */ 2087 mutex_enter(&sysctl_file_marker_lock); 2088 mutex_enter(&proc_lock); 2089 PROCLIST_FOREACH(p, &allproc) { 2090 struct filedesc *fd; 2091 fdtab_t *dt; 2092 u_int i; 2093 2094 if (p->p_stat == SIDL) { 2095 /* skip embryonic processes */ 2096 continue; 2097 } 2098 mutex_enter(p->p_lock); 2099 error = kauth_authorize_process(l->l_cred, 2100 KAUTH_PROCESS_CANSEE, p, 2101 KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_OPENFILES), 2102 NULL, NULL); 2103 mutex_exit(p->p_lock); 2104 if (error != 0) { 2105 /* 2106 * Don't leak kauth retval if we're silently 2107 * skipping this entry. 2108 */ 2109 error = 0; 2110 continue; 2111 } 2112 2113 /* 2114 * Grab a hold on the process. 2115 */ 2116 if (!rw_tryenter(&p->p_reflock, RW_READER)) { 2117 continue; 2118 } 2119 mutex_exit(&proc_lock); 2120 2121 fd = p->p_fd; 2122 mutex_enter(&fd->fd_lock); 2123 dt = fd->fd_dt; 2124 for (i = 0; i < dt->dt_nfiles; i++) { 2125 fdfile_t *ff; 2126 2127 if ((ff = dt->dt_ff[i]) == NULL) { 2128 continue; 2129 } 2130 if ((fp = atomic_load_consume(&ff->ff_file)) == NULL) { 2131 continue; 2132 } 2133 2134 mutex_enter(&fp->f_lock); 2135 2136 if ((fp->f_count == 0) || 2137 (fp->f_marker == sysctl_file_marker)) { 2138 mutex_exit(&fp->f_lock); 2139 continue; 2140 } 2141 2142 /* Check that we have enough space. */ 2143 if (buflen < sizeof(struct file)) { 2144 *oldlenp = where - start; 2145 mutex_exit(&fp->f_lock); 2146 error = ENOMEM; 2147 break; 2148 } 2149 2150 fill_file(&fbuf, fp); 2151 mutex_exit(&fp->f_lock); 2152 error = sysctl_copyout(l, &fbuf, where, sizeof(fbuf)); 2153 if (error) { 2154 break; 2155 } 2156 buflen -= sizeof(struct file); 2157 where += sizeof(struct file); 2158 2159 fp->f_marker = sysctl_file_marker; 2160 } 2161 mutex_exit(&fd->fd_lock); 2162 2163 /* 2164 * Release reference to process. 2165 */ 2166 mutex_enter(&proc_lock); 2167 rw_exit(&p->p_reflock); 2168 2169 if (error) 2170 break; 2171 } 2172 2173 sysctl_file_marker++; 2174 /* Reset all markers if wrapped. */ 2175 if (sysctl_file_marker == 0) { 2176 sysctl_file_marker_reset(); 2177 sysctl_file_marker++; 2178 } 2179 2180 mutex_exit(&proc_lock); 2181 mutex_exit(&sysctl_file_marker_lock); 2182 2183 *oldlenp = where - start; 2184 sysctl_relock(); 2185 return error; 2186 } 2187 2188 /* 2189 * sysctl helper function for kern.file2 2190 */ 2191 static int 2192 sysctl_kern_file2(SYSCTLFN_ARGS) 2193 { 2194 struct proc *p; 2195 struct file *fp; 2196 struct filedesc *fd; 2197 struct kinfo_file kf; 2198 char *dp; 2199 u_int i, op; 2200 size_t len, needed, elem_size, out_size; 2201 int error, arg, elem_count; 2202 fdfile_t *ff; 2203 fdtab_t *dt; 2204 2205 if (namelen == 1 && name[0] == CTL_QUERY) 2206 return sysctl_query(SYSCTLFN_CALL(rnode)); 2207 2208 if (namelen != 4) 2209 return EINVAL; 2210 2211 error = 0; 2212 dp = oldp; 2213 len = (oldp != NULL) ? *oldlenp : 0; 2214 op = name[0]; 2215 arg = name[1]; 2216 elem_size = name[2]; 2217 elem_count = name[3]; 2218 out_size = MIN(sizeof(kf), elem_size); 2219 needed = 0; 2220 2221 if (elem_size < 1 || elem_count < 0) 2222 return EINVAL; 2223 2224 switch (op) { 2225 case KERN_FILE_BYFILE: 2226 case KERN_FILE_BYPID: 2227 /* 2228 * We're traversing the process list in both cases; the BYFILE 2229 * case does additional work of keeping track of files already 2230 * looked at. 2231 */ 2232 2233 /* doesn't use arg so it must be zero */ 2234 if ((op == KERN_FILE_BYFILE) && (arg != 0)) 2235 return EINVAL; 2236 2237 if ((op == KERN_FILE_BYPID) && (arg < -1)) 2238 /* -1 means all processes */ 2239 return EINVAL; 2240 2241 sysctl_unlock(); 2242 if (op == KERN_FILE_BYFILE) 2243 mutex_enter(&sysctl_file_marker_lock); 2244 mutex_enter(&proc_lock); 2245 PROCLIST_FOREACH(p, &allproc) { 2246 if (p->p_stat == SIDL) { 2247 /* skip embryonic processes */ 2248 continue; 2249 } 2250 if (arg > 0 && p->p_pid != arg) { 2251 /* pick only the one we want */ 2252 /* XXX want 0 to mean "kernel files" */ 2253 continue; 2254 } 2255 mutex_enter(p->p_lock); 2256 error = kauth_authorize_process(l->l_cred, 2257 KAUTH_PROCESS_CANSEE, p, 2258 KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_OPENFILES), 2259 NULL, NULL); 2260 mutex_exit(p->p_lock); 2261 if (error != 0) { 2262 /* 2263 * Don't leak kauth retval if we're silently 2264 * skipping this entry. 2265 */ 2266 error = 0; 2267 continue; 2268 } 2269 2270 /* 2271 * Grab a hold on the process. 2272 */ 2273 if (!rw_tryenter(&p->p_reflock, RW_READER)) { 2274 continue; 2275 } 2276 mutex_exit(&proc_lock); 2277 2278 fd = p->p_fd; 2279 mutex_enter(&fd->fd_lock); 2280 dt = fd->fd_dt; 2281 for (i = 0; i < dt->dt_nfiles; i++) { 2282 if ((ff = dt->dt_ff[i]) == NULL) { 2283 continue; 2284 } 2285 if ((fp = atomic_load_consume(&ff->ff_file)) == 2286 NULL) { 2287 continue; 2288 } 2289 2290 if ((op == KERN_FILE_BYFILE) && 2291 (fp->f_marker == sysctl_file_marker)) { 2292 continue; 2293 } 2294 if (len >= elem_size && elem_count > 0) { 2295 mutex_enter(&fp->f_lock); 2296 fill_file2(&kf, fp, ff, i, p->p_pid); 2297 mutex_exit(&fp->f_lock); 2298 mutex_exit(&fd->fd_lock); 2299 error = sysctl_copyout(l, 2300 &kf, dp, out_size); 2301 mutex_enter(&fd->fd_lock); 2302 if (error) 2303 break; 2304 dp += elem_size; 2305 len -= elem_size; 2306 } 2307 if (op == KERN_FILE_BYFILE) 2308 fp->f_marker = sysctl_file_marker; 2309 needed += elem_size; 2310 if (elem_count > 0 && elem_count != INT_MAX) 2311 elem_count--; 2312 } 2313 mutex_exit(&fd->fd_lock); 2314 2315 /* 2316 * Release reference to process. 2317 */ 2318 mutex_enter(&proc_lock); 2319 rw_exit(&p->p_reflock); 2320 } 2321 if (op == KERN_FILE_BYFILE) { 2322 sysctl_file_marker++; 2323 2324 /* Reset all markers if wrapped. */ 2325 if (sysctl_file_marker == 0) { 2326 sysctl_file_marker_reset(); 2327 sysctl_file_marker++; 2328 } 2329 } 2330 mutex_exit(&proc_lock); 2331 if (op == KERN_FILE_BYFILE) 2332 mutex_exit(&sysctl_file_marker_lock); 2333 sysctl_relock(); 2334 break; 2335 default: 2336 return EINVAL; 2337 } 2338 2339 if (oldp == NULL) 2340 needed += KERN_FILESLOP * elem_size; 2341 *oldlenp = needed; 2342 2343 return error; 2344 } 2345 2346 static void 2347 fill_file(struct file *fp, const struct file *fpsrc) 2348 { 2349 const bool allowaddr = get_expose_address(curproc); 2350 2351 memset(fp, 0, sizeof(*fp)); 2352 2353 fp->f_offset = fpsrc->f_offset; 2354 COND_SET_PTR(fp->f_cred, fpsrc->f_cred, allowaddr); 2355 COND_SET_CPTR(fp->f_ops, fpsrc->f_ops, allowaddr); 2356 COND_SET_STRUCT(fp->f_undata, fpsrc->f_undata, allowaddr); 2357 COND_SET_STRUCT(fp->f_list, fpsrc->f_list, allowaddr); 2358 fp->f_flag = fpsrc->f_flag; 2359 fp->f_marker = fpsrc->f_marker; 2360 fp->f_type = fpsrc->f_type; 2361 fp->f_advice = fpsrc->f_advice; 2362 fp->f_count = fpsrc->f_count; 2363 fp->f_msgcount = fpsrc->f_msgcount; 2364 fp->f_unpcount = fpsrc->f_unpcount; 2365 COND_SET_STRUCT(fp->f_unplist, fpsrc->f_unplist, allowaddr); 2366 } 2367 2368 static void 2369 fill_file2(struct kinfo_file *kp, const file_t *fp, const fdfile_t *ff, 2370 int i, pid_t pid) 2371 { 2372 const bool allowaddr = get_expose_address(curproc); 2373 2374 memset(kp, 0, sizeof(*kp)); 2375 2376 COND_SET_VALUE(kp->ki_fileaddr, PTRTOUINT64(fp), allowaddr); 2377 kp->ki_flag = fp->f_flag; 2378 kp->ki_iflags = 0; 2379 kp->ki_ftype = fp->f_type; 2380 kp->ki_count = fp->f_count; 2381 kp->ki_msgcount = fp->f_msgcount; 2382 COND_SET_VALUE(kp->ki_fucred, PTRTOUINT64(fp->f_cred), allowaddr); 2383 kp->ki_fuid = kauth_cred_geteuid(fp->f_cred); 2384 kp->ki_fgid = kauth_cred_getegid(fp->f_cred); 2385 COND_SET_VALUE(kp->ki_fops, PTRTOUINT64(fp->f_ops), allowaddr); 2386 kp->ki_foffset = fp->f_offset; 2387 COND_SET_VALUE(kp->ki_fdata, PTRTOUINT64(fp->f_data), allowaddr); 2388 2389 /* vnode information to glue this file to something */ 2390 if (fp->f_type == DTYPE_VNODE) { 2391 struct vnode *vp = fp->f_vnode; 2392 2393 COND_SET_VALUE(kp->ki_vun, PTRTOUINT64(vp->v_un.vu_socket), 2394 allowaddr); 2395 kp->ki_vsize = vp->v_size; 2396 kp->ki_vtype = vp->v_type; 2397 kp->ki_vtag = vp->v_tag; 2398 COND_SET_VALUE(kp->ki_vdata, PTRTOUINT64(vp->v_data), 2399 allowaddr); 2400 } 2401 2402 /* process information when retrieved via KERN_FILE_BYPID */ 2403 if (ff != NULL) { 2404 kp->ki_pid = pid; 2405 kp->ki_fd = i; 2406 kp->ki_ofileflags = ff->ff_exclose; 2407 kp->ki_usecount = ff->ff_refcnt; 2408 } 2409 } 2410
Indexes created Sat Oct 11 14:09:46 GMT 2025