1 /* $NetBSD: uipc_mbuf.c,v 1.232 2019/01/17 02:47:15 knakahara Exp $ */ 2 3 /* 4 * Copyright (c) 1999, 2001, 2018 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility, 9 * NASA Ames Research Center, and Maxime Villard. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 22 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 23 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 24 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 25 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 26 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 27 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 29 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 30 * POSSIBILITY OF SUCH DAMAGE. 31 */ 32 33 /* 34 * Copyright (c) 1982, 1986, 1988, 1991, 1993 35 * The Regents of the University of California. All rights reserved. 36 * 37 * Redistribution and use in source and binary forms, with or without 38 * modification, are permitted provided that the following conditions 39 * are met: 40 * 1. Redistributions of source code must retain the above copyright 41 * notice, this list of conditions and the following disclaimer. 42 * 2. Redistributions in binary form must reproduce the above copyright 43 * notice, this list of conditions and the following disclaimer in the 44 * documentation and/or other materials provided with the distribution. 45 * 3. Neither the name of the University nor the names of its contributors 46 * may be used to endorse or promote products derived from this software 47 * without specific prior written permission. 48 * 49 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 50 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 51 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 52 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 53 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 54 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 55 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 56 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 57 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 58 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 59 * SUCH DAMAGE. 60 * 61 * @(#)uipc_mbuf.c 8.4 (Berkeley) 2/14/95 62 */ 63 64 #include <sys/cdefs.h> 65 __KERNEL_RCSID(0, "$NetBSD: uipc_mbuf.c,v 1.232 2019/01/17 02:47:15 knakahara Exp $"); 66 67 #ifdef _KERNEL_OPT 68 #include "opt_mbuftrace.h" 69 #include "opt_nmbclusters.h" 70 #include "opt_ddb.h" 71 #include "ether.h" 72 #endif 73 74 #include <sys/param.h> 75 #include <sys/systm.h> 76 #include <sys/atomic.h> 77 #include <sys/cpu.h> 78 #include <sys/proc.h> 79 #include <sys/mbuf.h> 80 #include <sys/kernel.h> 81 #include <sys/syslog.h> 82 #include <sys/domain.h> 83 #include <sys/protosw.h> 84 #include <sys/percpu.h> 85 #include <sys/pool.h> 86 #include <sys/socket.h> 87 #include <sys/sysctl.h> 88 89 #include <net/if.h> 90 91 pool_cache_t mb_cache; /* mbuf cache */ 92 static pool_cache_t mcl_cache; /* mbuf cluster cache */ 93 94 struct mbstat mbstat; 95 int max_linkhdr; 96 int max_protohdr; 97 int max_hdr; 98 int max_datalen; 99 100 static void mb_drain(void *, int); 101 static int mb_ctor(void *, void *, int); 102 103 static void sysctl_kern_mbuf_setup(void); 104 105 static struct sysctllog *mbuf_sysctllog; 106 107 static struct mbuf *m_copy_internal(struct mbuf *, int, int, int, bool); 108 static struct mbuf *m_split_internal(struct mbuf *, int, int, bool); 109 static int m_copyback_internal(struct mbuf **, int, int, const void *, 110 int, int); 111 112 /* Flags for m_copyback_internal. */ 113 #define CB_COPYBACK 0x0001 /* copyback from cp */ 114 #define CB_PRESERVE 0x0002 /* preserve original data */ 115 #define CB_COW 0x0004 /* do copy-on-write */ 116 #define CB_EXTEND 0x0008 /* extend chain */ 117 118 static const char mclpool_warnmsg[] = 119 "WARNING: mclpool limit reached; increase kern.mbuf.nmbclusters"; 120 121 MALLOC_DEFINE(M_MBUF, "mbuf", "mbuf"); 122 123 static percpu_t *mbstat_percpu; 124 125 #ifdef MBUFTRACE 126 struct mownerhead mowners = LIST_HEAD_INITIALIZER(mowners); 127 struct mowner unknown_mowners[] = { 128 MOWNER_INIT("unknown", "free"), 129 MOWNER_INIT("unknown", "data"), 130 MOWNER_INIT("unknown", "header"), 131 MOWNER_INIT("unknown", "soname"), 132 MOWNER_INIT("unknown", "soopts"), 133 MOWNER_INIT("unknown", "ftable"), 134 MOWNER_INIT("unknown", "control"), 135 MOWNER_INIT("unknown", "oobdata"), 136 }; 137 struct mowner revoked_mowner = MOWNER_INIT("revoked", ""); 138 #endif 139 140 #define MEXT_ISEMBEDDED(m) ((m)->m_ext_ref == (m)) 141 142 #define MCLADDREFERENCE(o, n) \ 143 do { \ 144 KASSERT(((o)->m_flags & M_EXT) != 0); \ 145 KASSERT(((n)->m_flags & M_EXT) == 0); \ 146 KASSERT((o)->m_ext.ext_refcnt >= 1); \ 147 (n)->m_flags |= ((o)->m_flags & M_EXTCOPYFLAGS); \ 148 atomic_inc_uint(&(o)->m_ext.ext_refcnt); \ 149 (n)->m_ext_ref = (o)->m_ext_ref; \ 150 mowner_ref((n), (n)->m_flags); \ 151 } while (/* CONSTCOND */ 0) 152 153 static int 154 nmbclusters_limit(void) 155 { 156 #if defined(PMAP_MAP_POOLPAGE) 157 /* direct mapping, doesn't use space in kmem_arena */ 158 vsize_t max_size = physmem / 4; 159 #else 160 vsize_t max_size = MIN(physmem / 4, nkmempages / 4); 161 #endif 162 163 max_size = max_size * PAGE_SIZE / MCLBYTES; 164 #ifdef NMBCLUSTERS_MAX 165 max_size = MIN(max_size, NMBCLUSTERS_MAX); 166 #endif 167 168 #ifdef NMBCLUSTERS 169 return MIN(max_size, NMBCLUSTERS); 170 #else 171 return max_size; 172 #endif 173 } 174 175 /* 176 * Initialize the mbuf allocator. 177 */ 178 void 179 mbinit(void) 180 { 181 182 CTASSERT(sizeof(struct _m_ext) <= MHLEN); 183 CTASSERT(sizeof(struct mbuf) == MSIZE); 184 185 sysctl_kern_mbuf_setup(); 186 187 mb_cache = pool_cache_init(msize, 0, 0, 0, "mbpl", 188 NULL, IPL_VM, mb_ctor, NULL, NULL); 189 KASSERT(mb_cache != NULL); 190 191 mcl_cache = pool_cache_init(mclbytes, 0, 0, 0, "mclpl", NULL, 192 IPL_VM, NULL, NULL, NULL); 193 KASSERT(mcl_cache != NULL); 194 195 pool_cache_set_drain_hook(mb_cache, mb_drain, NULL); 196 pool_cache_set_drain_hook(mcl_cache, mb_drain, NULL); 197 198 /* 199 * Set an arbitrary default limit on the number of mbuf clusters. 200 */ 201 #ifdef NMBCLUSTERS 202 nmbclusters = nmbclusters_limit(); 203 #else 204 nmbclusters = MAX(1024, 205 (vsize_t)physmem * PAGE_SIZE / MCLBYTES / 16); 206 nmbclusters = MIN(nmbclusters, nmbclusters_limit()); 207 #endif 208 209 /* 210 * Set the hard limit on the mclpool to the number of 211 * mbuf clusters the kernel is to support. Log the limit 212 * reached message max once a minute. 213 */ 214 pool_cache_sethardlimit(mcl_cache, nmbclusters, mclpool_warnmsg, 60); 215 216 mbstat_percpu = percpu_alloc(sizeof(struct mbstat_cpu)); 217 218 /* 219 * Set a low water mark for both mbufs and clusters. This should 220 * help ensure that they can be allocated in a memory starvation 221 * situation. This is important for e.g. diskless systems which 222 * must allocate mbufs in order for the pagedaemon to clean pages. 223 */ 224 pool_cache_setlowat(mb_cache, mblowat); 225 pool_cache_setlowat(mcl_cache, mcllowat); 226 227 #ifdef MBUFTRACE 228 { 229 /* 230 * Attach the unknown mowners. 231 */ 232 int i; 233 MOWNER_ATTACH(&revoked_mowner); 234 for (i = sizeof(unknown_mowners)/sizeof(unknown_mowners[0]); 235 i-- > 0; ) 236 MOWNER_ATTACH(&unknown_mowners[i]); 237 } 238 #endif 239 } 240 241 static void 242 mb_drain(void *arg, int flags) 243 { 244 struct domain *dp; 245 const struct protosw *pr; 246 struct ifnet *ifp; 247 int s; 248 249 KERNEL_LOCK(1, NULL); 250 s = splvm(); 251 DOMAIN_FOREACH(dp) { 252 for (pr = dp->dom_protosw; 253 pr < dp->dom_protoswNPROTOSW; pr++) 254 if (pr->pr_drain) 255 (*pr->pr_drain)(); 256 } 257 /* XXX we cannot use psref in H/W interrupt */ 258 if (!cpu_intr_p()) { 259 int bound = curlwp_bind(); 260 IFNET_READER_FOREACH(ifp) { 261 struct psref psref; 262 263 if_acquire(ifp, &psref); 264 265 if (ifp->if_drain) 266 (*ifp->if_drain)(ifp); 267 268 if_release(ifp, &psref); 269 } 270 curlwp_bindx(bound); 271 } 272 splx(s); 273 mbstat.m_drain++; 274 KERNEL_UNLOCK_ONE(NULL); 275 } 276 277 /* 278 * sysctl helper routine for the kern.mbuf subtree. 279 * nmbclusters, mblowat and mcllowat need range 280 * checking and pool tweaking after being reset. 281 */ 282 static int 283 sysctl_kern_mbuf(SYSCTLFN_ARGS) 284 { 285 int error, newval; 286 struct sysctlnode node; 287 288 node = *rnode; 289 node.sysctl_data = &newval; 290 switch (rnode->sysctl_num) { 291 case MBUF_NMBCLUSTERS: 292 case MBUF_MBLOWAT: 293 case MBUF_MCLLOWAT: 294 newval = *(int*)rnode->sysctl_data; 295 break; 296 default: 297 return EOPNOTSUPP; 298 } 299 300 error = sysctl_lookup(SYSCTLFN_CALL(&node)); 301 if (error || newp == NULL) 302 return error; 303 if (newval < 0) 304 return EINVAL; 305 306 switch (node.sysctl_num) { 307 case MBUF_NMBCLUSTERS: 308 if (newval < nmbclusters) 309 return EINVAL; 310 if (newval > nmbclusters_limit()) 311 return EINVAL; 312 nmbclusters = newval; 313 pool_cache_sethardlimit(mcl_cache, nmbclusters, 314 mclpool_warnmsg, 60); 315 break; 316 case MBUF_MBLOWAT: 317 mblowat = newval; 318 pool_cache_setlowat(mb_cache, mblowat); 319 break; 320 case MBUF_MCLLOWAT: 321 mcllowat = newval; 322 pool_cache_setlowat(mcl_cache, mcllowat); 323 break; 324 } 325 326 return 0; 327 } 328 329 #ifdef MBUFTRACE 330 static void 331 mowner_convert_to_user_cb(void *v1, void *v2, struct cpu_info *ci) 332 { 333 struct mowner_counter *mc = v1; 334 struct mowner_user *mo_user = v2; 335 int i; 336 337 for (i = 0; i < MOWNER_COUNTER_NCOUNTERS; i++) { 338 mo_user->mo_counter[i] += mc->mc_counter[i]; 339 } 340 } 341 342 static void 343 mowner_convert_to_user(struct mowner *mo, struct mowner_user *mo_user) 344 { 345 346 memset(mo_user, 0, sizeof(*mo_user)); 347 CTASSERT(sizeof(mo_user->mo_name) == sizeof(mo->mo_name)); 348 CTASSERT(sizeof(mo_user->mo_descr) == sizeof(mo->mo_descr)); 349 memcpy(mo_user->mo_name, mo->mo_name, sizeof(mo->mo_name)); 350 memcpy(mo_user->mo_descr, mo->mo_descr, sizeof(mo->mo_descr)); 351 percpu_foreach(mo->mo_counters, mowner_convert_to_user_cb, mo_user); 352 } 353 354 static int 355 sysctl_kern_mbuf_mowners(SYSCTLFN_ARGS) 356 { 357 struct mowner *mo; 358 size_t len = 0; 359 int error = 0; 360 361 if (namelen != 0) 362 return EINVAL; 363 if (newp != NULL) 364 return EPERM; 365 366 LIST_FOREACH(mo, &mowners, mo_link) { 367 struct mowner_user mo_user; 368 369 mowner_convert_to_user(mo, &mo_user); 370 371 if (oldp != NULL) { 372 if (*oldlenp - len < sizeof(mo_user)) { 373 error = ENOMEM; 374 break; 375 } 376 error = copyout(&mo_user, (char *)oldp + len, 377 sizeof(mo_user)); 378 if (error) 379 break; 380 } 381 len += sizeof(mo_user); 382 } 383 384 if (error == 0) 385 *oldlenp = len; 386 387 return error; 388 } 389 #endif /* MBUFTRACE */ 390 391 void 392 mbstat_type_add(int type, int diff) 393 { 394 struct mbstat_cpu *mb; 395 int s; 396 397 s = splvm(); 398 mb = percpu_getref(mbstat_percpu); 399 mb->m_mtypes[type] += diff; 400 percpu_putref(mbstat_percpu); 401 splx(s); 402 } 403 404 static void 405 mbstat_conver_to_user_cb(void *v1, void *v2, struct cpu_info *ci) 406 { 407 struct mbstat_cpu *mbsc = v1; 408 struct mbstat *mbs = v2; 409 int i; 410 411 for (i = 0; i < __arraycount(mbs->m_mtypes); i++) { 412 mbs->m_mtypes[i] += mbsc->m_mtypes[i]; 413 } 414 } 415 416 static void 417 mbstat_convert_to_user(struct mbstat *mbs) 418 { 419 420 memset(mbs, 0, sizeof(*mbs)); 421 mbs->m_drain = mbstat.m_drain; 422 percpu_foreach(mbstat_percpu, mbstat_conver_to_user_cb, mbs); 423 } 424 425 static int 426 sysctl_kern_mbuf_stats(SYSCTLFN_ARGS) 427 { 428 struct sysctlnode node; 429 struct mbstat mbs; 430 431 mbstat_convert_to_user(&mbs); 432 node = *rnode; 433 node.sysctl_data = &mbs; 434 node.sysctl_size = sizeof(mbs); 435 return sysctl_lookup(SYSCTLFN_CALL(&node)); 436 } 437 438 static void 439 sysctl_kern_mbuf_setup(void) 440 { 441 442 KASSERT(mbuf_sysctllog == NULL); 443 sysctl_createv(&mbuf_sysctllog, 0, NULL, NULL, 444 CTLFLAG_PERMANENT, 445 CTLTYPE_NODE, "mbuf", 446 SYSCTL_DESCR("mbuf control variables"), 447 NULL, 0, NULL, 0, 448 CTL_KERN, KERN_MBUF, CTL_EOL); 449 450 sysctl_createv(&mbuf_sysctllog, 0, NULL, NULL, 451 CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE, 452 CTLTYPE_INT, "msize", 453 SYSCTL_DESCR("mbuf base size"), 454 NULL, msize, NULL, 0, 455 CTL_KERN, KERN_MBUF, MBUF_MSIZE, CTL_EOL); 456 sysctl_createv(&mbuf_sysctllog, 0, NULL, NULL, 457 CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE, 458 CTLTYPE_INT, "mclbytes", 459 SYSCTL_DESCR("mbuf cluster size"), 460 NULL, mclbytes, NULL, 0, 461 CTL_KERN, KERN_MBUF, MBUF_MCLBYTES, CTL_EOL); 462 sysctl_createv(&mbuf_sysctllog, 0, NULL, NULL, 463 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 464 CTLTYPE_INT, "nmbclusters", 465 SYSCTL_DESCR("Limit on the number of mbuf clusters"), 466 sysctl_kern_mbuf, 0, &nmbclusters, 0, 467 CTL_KERN, KERN_MBUF, MBUF_NMBCLUSTERS, CTL_EOL); 468 sysctl_createv(&mbuf_sysctllog, 0, NULL, NULL, 469 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 470 CTLTYPE_INT, "mblowat", 471 SYSCTL_DESCR("mbuf low water mark"), 472 sysctl_kern_mbuf, 0, &mblowat, 0, 473 CTL_KERN, KERN_MBUF, MBUF_MBLOWAT, CTL_EOL); 474 sysctl_createv(&mbuf_sysctllog, 0, NULL, NULL, 475 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 476 CTLTYPE_INT, "mcllowat", 477 SYSCTL_DESCR("mbuf cluster low water mark"), 478 sysctl_kern_mbuf, 0, &mcllowat, 0, 479 CTL_KERN, KERN_MBUF, MBUF_MCLLOWAT, CTL_EOL); 480 sysctl_createv(&mbuf_sysctllog, 0, NULL, NULL, 481 CTLFLAG_PERMANENT, 482 CTLTYPE_STRUCT, "stats", 483 SYSCTL_DESCR("mbuf allocation statistics"), 484 sysctl_kern_mbuf_stats, 0, NULL, 0, 485 CTL_KERN, KERN_MBUF, MBUF_STATS, CTL_EOL); 486 #ifdef MBUFTRACE 487 sysctl_createv(&mbuf_sysctllog, 0, NULL, NULL, 488 CTLFLAG_PERMANENT, 489 CTLTYPE_STRUCT, "mowners", 490 SYSCTL_DESCR("Information about mbuf owners"), 491 sysctl_kern_mbuf_mowners, 0, NULL, 0, 492 CTL_KERN, KERN_MBUF, MBUF_MOWNERS, CTL_EOL); 493 #endif 494 } 495 496 static int 497 mb_ctor(void *arg, void *object, int flags) 498 { 499 struct mbuf *m = object; 500 501 #ifdef POOL_VTOPHYS 502 m->m_paddr = POOL_VTOPHYS(m); 503 #else 504 m->m_paddr = M_PADDR_INVALID; 505 #endif 506 return 0; 507 } 508 509 /* 510 * Add mbuf to the end of a chain 511 */ 512 struct mbuf * 513 m_add(struct mbuf *c, struct mbuf *m) 514 { 515 struct mbuf *n; 516 517 if (c == NULL) 518 return m; 519 520 for (n = c; n->m_next != NULL; n = n->m_next) 521 continue; 522 n->m_next = m; 523 return c; 524 } 525 526 struct mbuf * 527 m_get(int how, int type) 528 { 529 struct mbuf *m; 530 531 KASSERT(type != MT_FREE); 532 533 m = pool_cache_get(mb_cache, 534 how == M_WAIT ? PR_WAITOK|PR_LIMITFAIL : PR_NOWAIT); 535 if (m == NULL) 536 return NULL; 537 538 mbstat_type_add(type, 1); 539 540 mowner_init(m, type); 541 m->m_ext_ref = m; /* default */ 542 m->m_type = type; 543 m->m_len = 0; 544 m->m_next = NULL; 545 m->m_nextpkt = NULL; /* default */ 546 m->m_data = m->m_dat; 547 m->m_flags = 0; /* default */ 548 549 return m; 550 } 551 552 struct mbuf * 553 m_gethdr(int how, int type) 554 { 555 struct mbuf *m; 556 557 m = m_get(how, type); 558 if (m == NULL) 559 return NULL; 560 561 m->m_data = m->m_pktdat; 562 m->m_flags = M_PKTHDR; 563 564 m_reset_rcvif(m); 565 m->m_pkthdr.len = 0; 566 m->m_pkthdr.csum_flags = 0; 567 m->m_pkthdr.csum_data = 0; 568 m->m_pkthdr.segsz = 0; 569 m->m_pkthdr.ether_vtag = 0; 570 m->m_pkthdr.pkthdr_flags = 0; 571 SLIST_INIT(&m->m_pkthdr.tags); 572 573 m->m_pkthdr.pattr_class = NULL; 574 m->m_pkthdr.pattr_af = AF_UNSPEC; 575 m->m_pkthdr.pattr_hdr = NULL; 576 577 return m; 578 } 579 580 void 581 m_clget(struct mbuf *m, int how) 582 { 583 m->m_ext_storage.ext_buf = (char *)pool_cache_get_paddr(mcl_cache, 584 how == M_WAIT ? (PR_WAITOK|PR_LIMITFAIL) : PR_NOWAIT, 585 &m->m_ext_storage.ext_paddr); 586 587 if (m->m_ext_storage.ext_buf == NULL) 588 return; 589 590 MCLINITREFERENCE(m); 591 m->m_data = m->m_ext.ext_buf; 592 m->m_flags = (m->m_flags & ~M_EXTCOPYFLAGS) | 593 M_EXT|M_EXT_CLUSTER|M_EXT_RW; 594 m->m_ext.ext_size = MCLBYTES; 595 m->m_ext.ext_free = NULL; 596 m->m_ext.ext_arg = NULL; 597 /* ext_paddr initialized above */ 598 599 mowner_ref(m, M_EXT|M_EXT_CLUSTER); 600 } 601 602 struct mbuf * 603 m_getcl(int how, int type, int flags) 604 { 605 struct mbuf *mp; 606 607 if ((flags & M_PKTHDR) != 0) 608 mp = m_gethdr(how, type); 609 else 610 mp = m_get(how, type); 611 612 if (mp == NULL) 613 return NULL; 614 615 MCLGET(mp, how); 616 if ((mp->m_flags & M_EXT) != 0) 617 return mp; 618 619 m_free(mp); 620 return NULL; 621 } 622 623 /* 624 * Utility function for M_PREPEND. Do *NOT* use it directly. 625 */ 626 struct mbuf * 627 m_prepend(struct mbuf *m, int len, int how) 628 { 629 struct mbuf *mn; 630 631 if (__predict_false(len > MHLEN)) { 632 panic("%s: len > MHLEN", __func__); 633 } 634 635 KASSERT(len != M_COPYALL); 636 mn = m_get(how, m->m_type); 637 if (mn == NULL) { 638 m_freem(m); 639 return NULL; 640 } 641 642 if (m->m_flags & M_PKTHDR) { 643 m_move_pkthdr(mn, m); 644 } else { 645 MCLAIM(mn, m->m_owner); 646 } 647 mn->m_next = m; 648 m = mn; 649 650 if (m->m_flags & M_PKTHDR) { 651 if (len < MHLEN) 652 m_align(m, len); 653 } else { 654 if (len < MLEN) 655 m_align(m, len); 656 } 657 658 m->m_len = len; 659 return m; 660 } 661 662 struct mbuf * 663 m_copym(struct mbuf *m, int off, int len, int wait) 664 { 665 /* Shallow copy on M_EXT. */ 666 return m_copy_internal(m, off, len, wait, false); 667 } 668 669 struct mbuf * 670 m_dup(struct mbuf *m, int off, int len, int wait) 671 { 672 /* Deep copy. */ 673 return m_copy_internal(m, off, len, wait, true); 674 } 675 676 static inline int 677 m_copylen(int len, int copylen) 678 { 679 return (len == M_COPYALL) ? copylen : uimin(len, copylen); 680 } 681 682 static struct mbuf * 683 m_copy_internal(struct mbuf *m, int off0, int len, int wait, bool deep) 684 { 685 struct mbuf *n, **np; 686 int off = off0; 687 struct mbuf *top; 688 int copyhdr = 0; 689 690 if (off < 0 || (len != M_COPYALL && len < 0)) 691 panic("%s: off %d, len %d", __func__, off, len); 692 if (off == 0 && m->m_flags & M_PKTHDR) 693 copyhdr = 1; 694 while (off > 0) { 695 if (m == NULL) 696 panic("%s: m == NULL, off %d", __func__, off); 697 if (off < m->m_len) 698 break; 699 off -= m->m_len; 700 m = m->m_next; 701 } 702 703 np = ⊤ 704 top = NULL; 705 while (len == M_COPYALL || len > 0) { 706 if (m == NULL) { 707 if (len != M_COPYALL) 708 panic("%s: m == NULL, len %d [!COPYALL]", 709 __func__, len); 710 break; 711 } 712 713 n = m_get(wait, m->m_type); 714 *np = n; 715 if (n == NULL) 716 goto nospace; 717 MCLAIM(n, m->m_owner); 718 719 if (copyhdr) { 720 m_copy_pkthdr(n, m); 721 if (len == M_COPYALL) 722 n->m_pkthdr.len -= off0; 723 else 724 n->m_pkthdr.len = len; 725 copyhdr = 0; 726 } 727 n->m_len = m_copylen(len, m->m_len - off); 728 729 if (m->m_flags & M_EXT) { 730 if (!deep) { 731 n->m_data = m->m_data + off; 732 MCLADDREFERENCE(m, n); 733 } else { 734 /* 735 * We don't care if MCLGET fails. n->m_len is 736 * recomputed and handles that. 737 */ 738 MCLGET(n, wait); 739 n->m_len = 0; 740 n->m_len = M_TRAILINGSPACE(n); 741 n->m_len = m_copylen(len, n->m_len); 742 n->m_len = uimin(n->m_len, m->m_len - off); 743 memcpy(mtod(n, void *), mtod(m, char *) + off, 744 (unsigned)n->m_len); 745 } 746 } else { 747 memcpy(mtod(n, void *), mtod(m, char *) + off, 748 (unsigned)n->m_len); 749 } 750 751 if (len != M_COPYALL) 752 len -= n->m_len; 753 off += n->m_len; 754 755 KASSERT(off <= m->m_len); 756 757 if (off == m->m_len) { 758 m = m->m_next; 759 off = 0; 760 } 761 np = &n->m_next; 762 } 763 764 return top; 765 766 nospace: 767 m_freem(top); 768 return NULL; 769 } 770 771 /* 772 * Copy an entire packet, including header (which must be present). 773 * An optimization of the common case 'm_copym(m, 0, M_COPYALL, how)'. 774 */ 775 struct mbuf * 776 m_copypacket(struct mbuf *m, int how) 777 { 778 struct mbuf *top, *n, *o; 779 780 if (__predict_false((m->m_flags & M_PKTHDR) == 0)) { 781 panic("%s: no header (m = %p)", __func__, m); 782 } 783 784 n = m_get(how, m->m_type); 785 top = n; 786 if (!n) 787 goto nospace; 788 789 MCLAIM(n, m->m_owner); 790 m_copy_pkthdr(n, m); 791 n->m_len = m->m_len; 792 if (m->m_flags & M_EXT) { 793 n->m_data = m->m_data; 794 MCLADDREFERENCE(m, n); 795 } else { 796 memcpy(mtod(n, char *), mtod(m, char *), n->m_len); 797 } 798 799 m = m->m_next; 800 while (m) { 801 o = m_get(how, m->m_type); 802 if (!o) 803 goto nospace; 804 805 MCLAIM(o, m->m_owner); 806 n->m_next = o; 807 n = n->m_next; 808 809 n->m_len = m->m_len; 810 if (m->m_flags & M_EXT) { 811 n->m_data = m->m_data; 812 MCLADDREFERENCE(m, n); 813 } else { 814 memcpy(mtod(n, char *), mtod(m, char *), n->m_len); 815 } 816 817 m = m->m_next; 818 } 819 return top; 820 821 nospace: 822 m_freem(top); 823 return NULL; 824 } 825 826 void 827 m_copydata(struct mbuf *m, int off, int len, void *cp) 828 { 829 unsigned int count; 830 struct mbuf *m0 = m; 831 int len0 = len; 832 int off0 = off; 833 void *cp0 = cp; 834 835 KASSERT(len != M_COPYALL); 836 if (off < 0 || len < 0) 837 panic("m_copydata: off %d, len %d", off, len); 838 while (off > 0) { 839 if (m == NULL) 840 panic("m_copydata(%p,%d,%d,%p): m=NULL, off=%d (%d)", 841 m0, len0, off0, cp0, off, off0 - off); 842 if (off < m->m_len) 843 break; 844 off -= m->m_len; 845 m = m->m_next; 846 } 847 while (len > 0) { 848 if (m == NULL) 849 panic("m_copydata(%p,%d,%d,%p): " 850 "m=NULL, off=%d (%d), len=%d (%d)", 851 m0, len0, off0, cp0, 852 off, off0 - off, len, len0 - len); 853 count = uimin(m->m_len - off, len); 854 memcpy(cp, mtod(m, char *) + off, count); 855 len -= count; 856 cp = (char *)cp + count; 857 off = 0; 858 m = m->m_next; 859 } 860 } 861 862 /* 863 * Concatenate mbuf chain n to m. 864 * n might be copied into m (when n->m_len is small), therefore data portion of 865 * n could be copied into an mbuf of different mbuf type. 866 * Any m_pkthdr is not updated. 867 */ 868 void 869 m_cat(struct mbuf *m, struct mbuf *n) 870 { 871 872 while (m->m_next) 873 m = m->m_next; 874 while (n) { 875 if (M_READONLY(m) || n->m_len > M_TRAILINGSPACE(m)) { 876 /* just join the two chains */ 877 m->m_next = n; 878 return; 879 } 880 /* splat the data from one into the other */ 881 memcpy(mtod(m, char *) + m->m_len, mtod(n, void *), 882 (u_int)n->m_len); 883 m->m_len += n->m_len; 884 n = m_free(n); 885 } 886 } 887 888 void 889 m_adj(struct mbuf *mp, int req_len) 890 { 891 int len = req_len; 892 struct mbuf *m; 893 int count; 894 895 if ((m = mp) == NULL) 896 return; 897 if (len >= 0) { 898 /* 899 * Trim from head. 900 */ 901 while (m != NULL && len > 0) { 902 if (m->m_len <= len) { 903 len -= m->m_len; 904 m->m_len = 0; 905 m = m->m_next; 906 } else { 907 m->m_len -= len; 908 m->m_data += len; 909 len = 0; 910 } 911 } 912 if (mp->m_flags & M_PKTHDR) 913 mp->m_pkthdr.len -= (req_len - len); 914 } else { 915 /* 916 * Trim from tail. Scan the mbuf chain, 917 * calculating its length and finding the last mbuf. 918 * If the adjustment only affects this mbuf, then just 919 * adjust and return. Otherwise, rescan and truncate 920 * after the remaining size. 921 */ 922 len = -len; 923 count = 0; 924 for (;;) { 925 count += m->m_len; 926 if (m->m_next == NULL) 927 break; 928 m = m->m_next; 929 } 930 if (m->m_len >= len) { 931 m->m_len -= len; 932 if (mp->m_flags & M_PKTHDR) 933 mp->m_pkthdr.len -= len; 934 return; 935 } 936 937 count -= len; 938 if (count < 0) 939 count = 0; 940 941 /* 942 * Correct length for chain is "count". 943 * Find the mbuf with last data, adjust its length, 944 * and toss data from remaining mbufs on chain. 945 */ 946 m = mp; 947 if (m->m_flags & M_PKTHDR) 948 m->m_pkthdr.len = count; 949 for (; m; m = m->m_next) { 950 if (m->m_len >= count) { 951 m->m_len = count; 952 break; 953 } 954 count -= m->m_len; 955 } 956 if (m) { 957 while (m->m_next) 958 (m = m->m_next)->m_len = 0; 959 } 960 } 961 } 962 963 /* 964 * m_ensure_contig: rearrange an mbuf chain that given length of bytes 965 * would be contiguous and in the data area of an mbuf (therefore, mtod() 966 * would work for a structure of given length). 967 * 968 * => On success, returns true and the resulting mbuf chain; false otherwise. 969 * => The mbuf chain may change, but is always preserved valid. 970 */ 971 bool 972 m_ensure_contig(struct mbuf **m0, int len) 973 { 974 struct mbuf *n = *m0, *m; 975 size_t count, space; 976 977 KASSERT(len != M_COPYALL); 978 /* 979 * If first mbuf has no cluster, and has room for len bytes 980 * without shifting current data, pullup into it, 981 * otherwise allocate a new mbuf to prepend to the chain. 982 */ 983 if ((n->m_flags & M_EXT) == 0 && 984 n->m_data + len < &n->m_dat[MLEN] && n->m_next) { 985 if (n->m_len >= len) { 986 return true; 987 } 988 m = n; 989 n = n->m_next; 990 len -= m->m_len; 991 } else { 992 if (len > MHLEN) { 993 return false; 994 } 995 m = m_get(M_DONTWAIT, n->m_type); 996 if (m == NULL) { 997 return false; 998 } 999 MCLAIM(m, n->m_owner); 1000 if (n->m_flags & M_PKTHDR) { 1001 m_move_pkthdr(m, n); 1002 } 1003 } 1004 space = &m->m_dat[MLEN] - (m->m_data + m->m_len); 1005 do { 1006 count = MIN(MIN(MAX(len, max_protohdr), space), n->m_len); 1007 memcpy(mtod(m, char *) + m->m_len, mtod(n, void *), 1008 (unsigned)count); 1009 len -= count; 1010 m->m_len += count; 1011 n->m_len -= count; 1012 space -= count; 1013 if (n->m_len) 1014 n->m_data += count; 1015 else 1016 n = m_free(n); 1017 } while (len > 0 && n); 1018 1019 m->m_next = n; 1020 *m0 = m; 1021 1022 return len <= 0; 1023 } 1024 1025 /* 1026 * m_pullup: same as m_ensure_contig(), but destroys mbuf chain on error. 1027 */ 1028 struct mbuf * 1029 m_pullup(struct mbuf *n, int len) 1030 { 1031 struct mbuf *m = n; 1032 1033 KASSERT(len != M_COPYALL); 1034 if (!m_ensure_contig(&m, len)) { 1035 KASSERT(m != NULL); 1036 m_freem(m); 1037 m = NULL; 1038 } 1039 return m; 1040 } 1041 1042 /* 1043 * ensure that [off, off + len) is contiguous on the mbuf chain "m". 1044 * packet chain before "off" is kept untouched. 1045 * if offp == NULL, the target will start at <retval, 0> on resulting chain. 1046 * if offp != NULL, the target will start at <retval, *offp> on resulting chain. 1047 * 1048 * on error return (NULL return value), original "m" will be freed. 1049 * 1050 * XXX M_TRAILINGSPACE/M_LEADINGSPACE on shared cluster (sharedcluster) 1051 */ 1052 struct mbuf * 1053 m_pulldown(struct mbuf *m, int off, int len, int *offp) 1054 { 1055 struct mbuf *n, *o; 1056 int hlen, tlen, olen; 1057 int sharedcluster; 1058 1059 /* Check invalid arguments. */ 1060 if (m == NULL) 1061 panic("%s: m == NULL", __func__); 1062 if (len > MCLBYTES) { 1063 m_freem(m); 1064 return NULL; 1065 } 1066 1067 n = m; 1068 while (n != NULL && off > 0) { 1069 if (n->m_len > off) 1070 break; 1071 off -= n->m_len; 1072 n = n->m_next; 1073 } 1074 /* Be sure to point non-empty mbuf. */ 1075 while (n != NULL && n->m_len == 0) 1076 n = n->m_next; 1077 if (!n) { 1078 m_freem(m); 1079 return NULL; /* mbuf chain too short */ 1080 } 1081 1082 sharedcluster = M_READONLY(n); 1083 1084 /* 1085 * The target data is on <n, off>. If we got enough data on the mbuf 1086 * "n", we're done. 1087 */ 1088 #ifdef __NO_STRICT_ALIGNMENT 1089 if ((off == 0 || offp) && len <= n->m_len - off && !sharedcluster) 1090 #else 1091 if ((off == 0 || offp) && len <= n->m_len - off && !sharedcluster && 1092 ALIGNED_POINTER((mtod(n, char *) + off), uint32_t)) 1093 #endif 1094 goto ok; 1095 1096 /* 1097 * When (len <= n->m_len - off) and (off != 0), it is a special case. 1098 * Len bytes from <n, off> sit in single mbuf, but the caller does 1099 * not like the starting position (off). 1100 * 1101 * Chop the current mbuf into two pieces, set off to 0. 1102 */ 1103 if (len <= n->m_len - off) { 1104 struct mbuf *mlast; 1105 1106 o = m_dup(n, off, n->m_len - off, M_DONTWAIT); 1107 if (o == NULL) { 1108 m_freem(m); 1109 return NULL; /* ENOBUFS */ 1110 } 1111 KASSERT(o->m_len >= len); 1112 for (mlast = o; mlast->m_next != NULL; mlast = mlast->m_next) 1113 ; 1114 n->m_len = off; 1115 mlast->m_next = n->m_next; 1116 n->m_next = o; 1117 n = o; 1118 off = 0; 1119 goto ok; 1120 } 1121 1122 /* 1123 * We need to take hlen from <n, off> and tlen from <n->m_next, 0>, 1124 * and construct contiguous mbuf with m_len == len. 1125 * 1126 * Note that hlen + tlen == len, and tlen > 0. 1127 */ 1128 hlen = n->m_len - off; 1129 tlen = len - hlen; 1130 1131 /* 1132 * Ensure that we have enough trailing data on mbuf chain. If not, 1133 * we can do nothing about the chain. 1134 */ 1135 olen = 0; 1136 for (o = n->m_next; o != NULL; o = o->m_next) 1137 olen += o->m_len; 1138 if (hlen + olen < len) { 1139 m_freem(m); 1140 return NULL; /* mbuf chain too short */ 1141 } 1142 1143 /* 1144 * Easy cases first. We need to use m_copydata() to get data from 1145 * <n->m_next, 0>. 1146 */ 1147 if ((off == 0 || offp) && M_TRAILINGSPACE(n) >= tlen && 1148 !sharedcluster) { 1149 m_copydata(n->m_next, 0, tlen, mtod(n, char *) + n->m_len); 1150 n->m_len += tlen; 1151 m_adj(n->m_next, tlen); 1152 goto ok; 1153 } 1154 if ((off == 0 || offp) && M_LEADINGSPACE(n->m_next) >= hlen && 1155 #ifndef __NO_STRICT_ALIGNMENT 1156 ALIGNED_POINTER((n->m_next->m_data - hlen), uint32_t) && 1157 #endif 1158 !sharedcluster && n->m_next->m_len >= tlen) { 1159 n->m_next->m_data -= hlen; 1160 n->m_next->m_len += hlen; 1161 memcpy(mtod(n->m_next, void *), mtod(n, char *) + off, hlen); 1162 n->m_len -= hlen; 1163 n = n->m_next; 1164 off = 0; 1165 goto ok; 1166 } 1167 1168 /* 1169 * Now, we need to do the hard way. Don't copy as there's no room 1170 * on both ends. 1171 */ 1172 o = m_get(M_DONTWAIT, m->m_type); 1173 if (o && len > MLEN) { 1174 MCLGET(o, M_DONTWAIT); 1175 if ((o->m_flags & M_EXT) == 0) { 1176 m_free(o); 1177 o = NULL; 1178 } 1179 } 1180 if (!o) { 1181 m_freem(m); 1182 return NULL; /* ENOBUFS */ 1183 } 1184 /* get hlen from <n, off> into <o, 0> */ 1185 o->m_len = hlen; 1186 memcpy(mtod(o, void *), mtod(n, char *) + off, hlen); 1187 n->m_len -= hlen; 1188 /* get tlen from <n->m_next, 0> into <o, hlen> */ 1189 m_copydata(n->m_next, 0, tlen, mtod(o, char *) + o->m_len); 1190 o->m_len += tlen; 1191 m_adj(n->m_next, tlen); 1192 o->m_next = n->m_next; 1193 n->m_next = o; 1194 n = o; 1195 off = 0; 1196 1197 ok: 1198 if (offp) 1199 *offp = off; 1200 return n; 1201 } 1202 1203 /* 1204 * Like m_pullup(), except a new mbuf is always allocated, and we allow 1205 * the amount of empty space before the data in the new mbuf to be specified 1206 * (in the event that the caller expects to prepend later). 1207 */ 1208 struct mbuf * 1209 m_copyup(struct mbuf *n, int len, int dstoff) 1210 { 1211 struct mbuf *m; 1212 int count, space; 1213 1214 KASSERT(len != M_COPYALL); 1215 if (len > ((int)MHLEN - dstoff)) 1216 goto bad; 1217 m = m_get(M_DONTWAIT, n->m_type); 1218 if (m == NULL) 1219 goto bad; 1220 MCLAIM(m, n->m_owner); 1221 if (n->m_flags & M_PKTHDR) { 1222 m_move_pkthdr(m, n); 1223 } 1224 m->m_data += dstoff; 1225 space = &m->m_dat[MLEN] - (m->m_data + m->m_len); 1226 do { 1227 count = uimin(uimin(uimax(len, max_protohdr), space), n->m_len); 1228 memcpy(mtod(m, char *) + m->m_len, mtod(n, void *), 1229 (unsigned)count); 1230 len -= count; 1231 m->m_len += count; 1232 n->m_len -= count; 1233 space -= count; 1234 if (n->m_len) 1235 n->m_data += count; 1236 else 1237 n = m_free(n); 1238 } while (len > 0 && n); 1239 if (len > 0) { 1240 (void) m_free(m); 1241 goto bad; 1242 } 1243 m->m_next = n; 1244 return m; 1245 bad: 1246 m_freem(n); 1247 return NULL; 1248 } 1249 1250 struct mbuf * 1251 m_split(struct mbuf *m0, int len, int wait) 1252 { 1253 return m_split_internal(m0, len, wait, true); 1254 } 1255 1256 static struct mbuf * 1257 m_split_internal(struct mbuf *m0, int len0, int wait, bool copyhdr) 1258 { 1259 struct mbuf *m, *n; 1260 unsigned len = len0, remain, len_save; 1261 1262 KASSERT(len0 != M_COPYALL); 1263 for (m = m0; m && len > m->m_len; m = m->m_next) 1264 len -= m->m_len; 1265 if (m == NULL) 1266 return NULL; 1267 1268 remain = m->m_len - len; 1269 if (copyhdr && (m0->m_flags & M_PKTHDR)) { 1270 n = m_gethdr(wait, m0->m_type); 1271 if (n == NULL) 1272 return NULL; 1273 1274 MCLAIM(n, m0->m_owner); 1275 m_copy_rcvif(n, m0); 1276 n->m_pkthdr.len = m0->m_pkthdr.len - len0; 1277 len_save = m0->m_pkthdr.len; 1278 m0->m_pkthdr.len = len0; 1279 1280 if (m->m_flags & M_EXT) 1281 goto extpacket; 1282 1283 if (remain > MHLEN) { 1284 /* m can't be the lead packet */ 1285 m_align(n, 0); 1286 n->m_len = 0; 1287 n->m_next = m_split(m, len, wait); 1288 if (n->m_next == NULL) { 1289 (void)m_free(n); 1290 m0->m_pkthdr.len = len_save; 1291 return NULL; 1292 } 1293 return n; 1294 } else { 1295 m_align(n, remain); 1296 } 1297 } else if (remain == 0) { 1298 n = m->m_next; 1299 m->m_next = NULL; 1300 return n; 1301 } else { 1302 n = m_get(wait, m->m_type); 1303 if (n == NULL) 1304 return NULL; 1305 MCLAIM(n, m->m_owner); 1306 m_align(n, remain); 1307 } 1308 1309 extpacket: 1310 if (m->m_flags & M_EXT) { 1311 n->m_data = m->m_data + len; 1312 MCLADDREFERENCE(m, n); 1313 } else { 1314 memcpy(mtod(n, void *), mtod(m, char *) + len, remain); 1315 } 1316 1317 n->m_len = remain; 1318 m->m_len = len; 1319 n->m_next = m->m_next; 1320 m->m_next = NULL; 1321 return n; 1322 } 1323 1324 /* 1325 * Routine to copy from device local memory into mbufs. 1326 */ 1327 struct mbuf * 1328 m_devget(char *buf, int totlen, int off, struct ifnet *ifp) 1329 { 1330 struct mbuf *m; 1331 struct mbuf *top = NULL, **mp = ⊤ 1332 char *cp, *epkt; 1333 int len; 1334 1335 cp = buf; 1336 epkt = cp + totlen; 1337 if (off) { 1338 /* 1339 * If 'off' is non-zero, packet is trailer-encapsulated, 1340 * so we have to skip the type and length fields. 1341 */ 1342 cp += off + 2 * sizeof(uint16_t); 1343 totlen -= 2 * sizeof(uint16_t); 1344 } 1345 1346 m = m_gethdr(M_DONTWAIT, MT_DATA); 1347 if (m == NULL) 1348 return NULL; 1349 m_set_rcvif(m, ifp); 1350 m->m_pkthdr.len = totlen; 1351 m->m_len = MHLEN; 1352 1353 while (totlen > 0) { 1354 if (top) { 1355 m = m_get(M_DONTWAIT, MT_DATA); 1356 if (m == NULL) { 1357 m_freem(top); 1358 return NULL; 1359 } 1360 m->m_len = MLEN; 1361 } 1362 1363 len = uimin(totlen, epkt - cp); 1364 1365 if (len >= MINCLSIZE) { 1366 MCLGET(m, M_DONTWAIT); 1367 if ((m->m_flags & M_EXT) == 0) { 1368 m_free(m); 1369 m_freem(top); 1370 return NULL; 1371 } 1372 m->m_len = len = uimin(len, MCLBYTES); 1373 } else { 1374 /* 1375 * Place initial small packet/header at end of mbuf. 1376 */ 1377 if (len < m->m_len) { 1378 if (top == 0 && len + max_linkhdr <= m->m_len) 1379 m->m_data += max_linkhdr; 1380 m->m_len = len; 1381 } else 1382 len = m->m_len; 1383 } 1384 1385 memcpy(mtod(m, void *), cp, (size_t)len); 1386 1387 cp += len; 1388 *mp = m; 1389 mp = &m->m_next; 1390 totlen -= len; 1391 if (cp == epkt) 1392 cp = buf; 1393 } 1394 1395 return top; 1396 } 1397 1398 /* 1399 * Copy data from a buffer back into the indicated mbuf chain, 1400 * starting "off" bytes from the beginning, extending the mbuf 1401 * chain if necessary. 1402 */ 1403 void 1404 m_copyback(struct mbuf *m0, int off, int len, const void *cp) 1405 { 1406 #if defined(DEBUG) 1407 struct mbuf *origm = m0; 1408 int error; 1409 #endif 1410 1411 if (m0 == NULL) 1412 return; 1413 1414 #if defined(DEBUG) 1415 error = 1416 #endif 1417 m_copyback_internal(&m0, off, len, cp, CB_COPYBACK|CB_EXTEND, 1418 M_DONTWAIT); 1419 1420 #if defined(DEBUG) 1421 if (error != 0 || (m0 != NULL && origm != m0)) 1422 panic("m_copyback"); 1423 #endif 1424 } 1425 1426 struct mbuf * 1427 m_copyback_cow(struct mbuf *m0, int off, int len, const void *cp, int how) 1428 { 1429 int error; 1430 1431 /* don't support chain expansion */ 1432 KASSERT(len != M_COPYALL); 1433 KDASSERT(off + len <= m_length(m0)); 1434 1435 error = m_copyback_internal(&m0, off, len, cp, CB_COPYBACK|CB_COW, 1436 how); 1437 if (error) { 1438 /* 1439 * no way to recover from partial success. 1440 * just free the chain. 1441 */ 1442 m_freem(m0); 1443 return NULL; 1444 } 1445 return m0; 1446 } 1447 1448 int 1449 m_makewritable(struct mbuf **mp, int off, int len, int how) 1450 { 1451 int error; 1452 #if defined(DEBUG) 1453 int origlen = m_length(*mp); 1454 #endif 1455 1456 error = m_copyback_internal(mp, off, len, NULL, CB_PRESERVE|CB_COW, 1457 how); 1458 if (error) 1459 return error; 1460 1461 #if defined(DEBUG) 1462 int reslen = 0; 1463 for (struct mbuf *n = *mp; n; n = n->m_next) 1464 reslen += n->m_len; 1465 if (origlen != reslen) 1466 panic("m_makewritable: length changed"); 1467 if (((*mp)->m_flags & M_PKTHDR) != 0 && reslen != (*mp)->m_pkthdr.len) 1468 panic("m_makewritable: inconsist"); 1469 #endif 1470 1471 return 0; 1472 } 1473 1474 static int 1475 m_copyback_internal(struct mbuf **mp0, int off, int len, const void *vp, 1476 int flags, int how) 1477 { 1478 int mlen; 1479 struct mbuf *m, *n; 1480 struct mbuf **mp; 1481 int totlen = 0; 1482 const char *cp = vp; 1483 1484 KASSERT(mp0 != NULL); 1485 KASSERT(*mp0 != NULL); 1486 KASSERT((flags & CB_PRESERVE) == 0 || cp == NULL); 1487 KASSERT((flags & CB_COPYBACK) == 0 || cp != NULL); 1488 1489 if (len == M_COPYALL) 1490 len = m_length(*mp0) - off; 1491 1492 /* 1493 * we don't bother to update "totlen" in the case of CB_COW, 1494 * assuming that CB_EXTEND and CB_COW are exclusive. 1495 */ 1496 1497 KASSERT((~flags & (CB_EXTEND|CB_COW)) != 0); 1498 1499 mp = mp0; 1500 m = *mp; 1501 while (off > (mlen = m->m_len)) { 1502 off -= mlen; 1503 totlen += mlen; 1504 if (m->m_next == NULL) { 1505 int tspace; 1506 extend: 1507 if ((flags & CB_EXTEND) == 0) 1508 goto out; 1509 1510 /* 1511 * try to make some space at the end of "m". 1512 */ 1513 1514 mlen = m->m_len; 1515 if (off + len >= MINCLSIZE && 1516 (m->m_flags & M_EXT) == 0 && m->m_len == 0) { 1517 MCLGET(m, how); 1518 } 1519 tspace = M_TRAILINGSPACE(m); 1520 if (tspace > 0) { 1521 tspace = uimin(tspace, off + len); 1522 KASSERT(tspace > 0); 1523 memset(mtod(m, char *) + m->m_len, 0, 1524 uimin(off, tspace)); 1525 m->m_len += tspace; 1526 off += mlen; 1527 totlen -= mlen; 1528 continue; 1529 } 1530 1531 /* 1532 * need to allocate an mbuf. 1533 */ 1534 1535 if (off + len >= MINCLSIZE) { 1536 n = m_getcl(how, m->m_type, 0); 1537 } else { 1538 n = m_get(how, m->m_type); 1539 } 1540 if (n == NULL) { 1541 goto out; 1542 } 1543 n->m_len = uimin(M_TRAILINGSPACE(n), off + len); 1544 memset(mtod(n, char *), 0, uimin(n->m_len, off)); 1545 m->m_next = n; 1546 } 1547 mp = &m->m_next; 1548 m = m->m_next; 1549 } 1550 while (len > 0) { 1551 mlen = m->m_len - off; 1552 if (mlen != 0 && M_READONLY(m)) { 1553 /* 1554 * This mbuf is read-only. Allocate a new writable 1555 * mbuf and try again. 1556 */ 1557 char *datap; 1558 int eatlen; 1559 1560 KASSERT((flags & CB_COW) != 0); 1561 1562 /* 1563 * if we're going to write into the middle of 1564 * a mbuf, split it first. 1565 */ 1566 if (off > 0) { 1567 n = m_split_internal(m, off, how, false); 1568 if (n == NULL) 1569 goto enobufs; 1570 m->m_next = n; 1571 mp = &m->m_next; 1572 m = n; 1573 off = 0; 1574 continue; 1575 } 1576 1577 /* 1578 * XXX TODO coalesce into the trailingspace of 1579 * the previous mbuf when possible. 1580 */ 1581 1582 /* 1583 * allocate a new mbuf. copy packet header if needed. 1584 */ 1585 n = m_get(how, m->m_type); 1586 if (n == NULL) 1587 goto enobufs; 1588 MCLAIM(n, m->m_owner); 1589 if (off == 0 && (m->m_flags & M_PKTHDR) != 0) { 1590 m_move_pkthdr(n, m); 1591 n->m_len = MHLEN; 1592 } else { 1593 if (len >= MINCLSIZE) 1594 MCLGET(n, M_DONTWAIT); 1595 n->m_len = 1596 (n->m_flags & M_EXT) ? MCLBYTES : MLEN; 1597 } 1598 if (n->m_len > len) 1599 n->m_len = len; 1600 1601 /* 1602 * free the region which has been overwritten. 1603 * copying data from old mbufs if requested. 1604 */ 1605 if (flags & CB_PRESERVE) 1606 datap = mtod(n, char *); 1607 else 1608 datap = NULL; 1609 eatlen = n->m_len; 1610 while (m != NULL && M_READONLY(m) && 1611 n->m_type == m->m_type && eatlen > 0) { 1612 mlen = uimin(eatlen, m->m_len); 1613 if (datap) { 1614 m_copydata(m, 0, mlen, datap); 1615 datap += mlen; 1616 } 1617 m->m_data += mlen; 1618 m->m_len -= mlen; 1619 eatlen -= mlen; 1620 if (m->m_len == 0) 1621 *mp = m = m_free(m); 1622 } 1623 if (eatlen > 0) 1624 n->m_len -= eatlen; 1625 n->m_next = m; 1626 *mp = m = n; 1627 continue; 1628 } 1629 mlen = uimin(mlen, len); 1630 if (flags & CB_COPYBACK) { 1631 memcpy(mtod(m, char *) + off, cp, (unsigned)mlen); 1632 cp += mlen; 1633 } 1634 len -= mlen; 1635 mlen += off; 1636 off = 0; 1637 totlen += mlen; 1638 if (len == 0) 1639 break; 1640 if (m->m_next == NULL) { 1641 goto extend; 1642 } 1643 mp = &m->m_next; 1644 m = m->m_next; 1645 } 1646 1647 out: 1648 if (((m = *mp0)->m_flags & M_PKTHDR) && (m->m_pkthdr.len < totlen)) { 1649 KASSERT((flags & CB_EXTEND) != 0); 1650 m->m_pkthdr.len = totlen; 1651 } 1652 1653 return 0; 1654 1655 enobufs: 1656 return ENOBUFS; 1657 } 1658 1659 /* 1660 * Compress the mbuf chain. Return the new mbuf chain on success, NULL on 1661 * failure. The first mbuf is preserved, and on success the pointer returned 1662 * is the same as the one passed. 1663 */ 1664 struct mbuf * 1665 m_defrag(struct mbuf *m, int how) 1666 { 1667 struct mbuf *m0, *mn, *n; 1668 int sz; 1669 1670 KASSERT((m->m_flags & M_PKTHDR) != 0); 1671 1672 if (m->m_next == NULL) 1673 return m; 1674 1675 m0 = m_get(how, MT_DATA); 1676 if (m0 == NULL) 1677 return NULL; 1678 mn = m0; 1679 1680 sz = m->m_pkthdr.len - m->m_len; 1681 KASSERT(sz >= 0); 1682 1683 do { 1684 if (sz > MLEN) { 1685 MCLGET(mn, how); 1686 if ((mn->m_flags & M_EXT) == 0) { 1687 m_freem(m0); 1688 return NULL; 1689 } 1690 } 1691 1692 mn->m_len = MIN(sz, MCLBYTES); 1693 1694 m_copydata(m, m->m_pkthdr.len - sz, mn->m_len, 1695 mtod(mn, void *)); 1696 1697 sz -= mn->m_len; 1698 1699 if (sz > 0) { 1700 /* need more mbufs */ 1701 n = m_get(how, MT_DATA); 1702 if (n == NULL) { 1703 m_freem(m0); 1704 return NULL; 1705 } 1706 1707 mn->m_next = n; 1708 mn = n; 1709 } 1710 } while (sz > 0); 1711 1712 m_freem(m->m_next); 1713 m->m_next = m0; 1714 1715 return m; 1716 } 1717 1718 void 1719 m_remove_pkthdr(struct mbuf *m) 1720 { 1721 KASSERT(m->m_flags & M_PKTHDR); 1722 1723 m_tag_delete_chain(m); 1724 m->m_flags &= ~M_PKTHDR; 1725 memset(&m->m_pkthdr, 0, sizeof(m->m_pkthdr)); 1726 } 1727 1728 void 1729 m_copy_pkthdr(struct mbuf *to, struct mbuf *from) 1730 { 1731 KASSERT((to->m_flags & M_EXT) == 0); 1732 KASSERT((to->m_flags & M_PKTHDR) == 0 || 1733 SLIST_FIRST(&to->m_pkthdr.tags) == NULL); 1734 KASSERT((from->m_flags & M_PKTHDR) != 0); 1735 1736 to->m_pkthdr = from->m_pkthdr; 1737 to->m_flags = from->m_flags & M_COPYFLAGS; 1738 to->m_data = to->m_pktdat; 1739 1740 SLIST_INIT(&to->m_pkthdr.tags); 1741 m_tag_copy_chain(to, from); 1742 } 1743 1744 void 1745 m_move_pkthdr(struct mbuf *to, struct mbuf *from) 1746 { 1747 KASSERT((to->m_flags & M_EXT) == 0); 1748 KASSERT((to->m_flags & M_PKTHDR) == 0 || 1749 SLIST_FIRST(&to->m_pkthdr.tags) == NULL); 1750 KASSERT((from->m_flags & M_PKTHDR) != 0); 1751 1752 to->m_pkthdr = from->m_pkthdr; 1753 to->m_flags = from->m_flags & M_COPYFLAGS; 1754 to->m_data = to->m_pktdat; 1755 1756 from->m_flags &= ~M_PKTHDR; 1757 } 1758 1759 /* 1760 * Set the m_data pointer of a newly-allocated mbuf to place an object of the 1761 * specified size at the end of the mbuf, longword aligned. 1762 */ 1763 void 1764 m_align(struct mbuf *m, int len) 1765 { 1766 int buflen, adjust; 1767 1768 KASSERT(len != M_COPYALL); 1769 KASSERT(M_LEADINGSPACE(m) == 0); 1770 1771 if (m->m_flags & M_EXT) 1772 buflen = m->m_ext.ext_size; 1773 else if (m->m_flags & M_PKTHDR) 1774 buflen = MHLEN; 1775 else 1776 buflen = MLEN; 1777 1778 KASSERT(len <= buflen); 1779 adjust = buflen - len; 1780 m->m_data += adjust &~ (sizeof(long)-1); 1781 } 1782 1783 /* 1784 * Apply function f to the data in an mbuf chain starting "off" bytes from the 1785 * beginning, continuing for "len" bytes. 1786 */ 1787 int 1788 m_apply(struct mbuf *m, int off, int len, 1789 int (*f)(void *, void *, unsigned int), void *arg) 1790 { 1791 unsigned int count; 1792 int rval; 1793 1794 KASSERT(len != M_COPYALL); 1795 KASSERT(len >= 0); 1796 KASSERT(off >= 0); 1797 1798 while (off > 0) { 1799 KASSERT(m != NULL); 1800 if (off < m->m_len) 1801 break; 1802 off -= m->m_len; 1803 m = m->m_next; 1804 } 1805 while (len > 0) { 1806 KASSERT(m != NULL); 1807 count = uimin(m->m_len - off, len); 1808 1809 rval = (*f)(arg, mtod(m, char *) + off, count); 1810 if (rval) 1811 return rval; 1812 1813 len -= count; 1814 off = 0; 1815 m = m->m_next; 1816 } 1817 1818 return 0; 1819 } 1820 1821 /* 1822 * Return a pointer to mbuf/offset of location in mbuf chain. 1823 */ 1824 struct mbuf * 1825 m_getptr(struct mbuf *m, int loc, int *off) 1826 { 1827 1828 while (loc >= 0) { 1829 /* Normal end of search */ 1830 if (m->m_len > loc) { 1831 *off = loc; 1832 return m; 1833 } 1834 1835 loc -= m->m_len; 1836 1837 if (m->m_next == NULL) { 1838 if (loc == 0) { 1839 /* Point at the end of valid data */ 1840 *off = m->m_len; 1841 return m; 1842 } 1843 return NULL; 1844 } else { 1845 m = m->m_next; 1846 } 1847 } 1848 1849 return NULL; 1850 } 1851 1852 /* 1853 * Release a reference to the mbuf external storage. 1854 * 1855 * => free the mbuf m itself as well. 1856 */ 1857 static void 1858 m_ext_free(struct mbuf *m) 1859 { 1860 const bool embedded = MEXT_ISEMBEDDED(m); 1861 bool dofree = true; 1862 u_int refcnt; 1863 1864 KASSERT((m->m_flags & M_EXT) != 0); 1865 KASSERT(MEXT_ISEMBEDDED(m->m_ext_ref)); 1866 KASSERT((m->m_ext_ref->m_flags & M_EXT) != 0); 1867 KASSERT((m->m_flags & M_EXT_CLUSTER) == 1868 (m->m_ext_ref->m_flags & M_EXT_CLUSTER)); 1869 1870 if (__predict_false(m->m_type == MT_FREE)) { 1871 panic("mbuf %p already freed", m); 1872 } 1873 1874 if (__predict_true(m->m_ext.ext_refcnt == 1)) { 1875 refcnt = m->m_ext.ext_refcnt = 0; 1876 } else { 1877 refcnt = atomic_dec_uint_nv(&m->m_ext.ext_refcnt); 1878 } 1879 1880 if (refcnt > 0) { 1881 if (embedded) { 1882 /* 1883 * other mbuf's m_ext_ref still points to us. 1884 */ 1885 dofree = false; 1886 } else { 1887 m->m_ext_ref = m; 1888 } 1889 } else { 1890 /* 1891 * dropping the last reference 1892 */ 1893 if (!embedded) { 1894 m->m_ext.ext_refcnt++; /* XXX */ 1895 m_ext_free(m->m_ext_ref); 1896 m->m_ext_ref = m; 1897 } else if ((m->m_flags & M_EXT_CLUSTER) != 0) { 1898 pool_cache_put_paddr(mcl_cache, 1899 m->m_ext.ext_buf, m->m_ext.ext_paddr); 1900 } else if (m->m_ext.ext_free) { 1901 (*m->m_ext.ext_free)(m, 1902 m->m_ext.ext_buf, m->m_ext.ext_size, 1903 m->m_ext.ext_arg); 1904 /* 1905 * 'm' is already freed by the ext_free callback. 1906 */ 1907 dofree = false; 1908 } else { 1909 free(m->m_ext.ext_buf, 0); 1910 } 1911 } 1912 1913 if (dofree) { 1914 m->m_type = MT_FREE; 1915 m->m_data = NULL; 1916 pool_cache_put(mb_cache, m); 1917 } 1918 } 1919 1920 /* 1921 * Free a single mbuf and associated external storage. Return the 1922 * successor, if any. 1923 */ 1924 struct mbuf * 1925 m_free(struct mbuf *m) 1926 { 1927 struct mbuf *n; 1928 1929 mowner_revoke(m, 1, m->m_flags); 1930 mbstat_type_add(m->m_type, -1); 1931 1932 if (m->m_flags & M_PKTHDR) 1933 m_tag_delete_chain(m); 1934 1935 n = m->m_next; 1936 1937 if (m->m_flags & M_EXT) { 1938 m_ext_free(m); 1939 } else { 1940 if (__predict_false(m->m_type == MT_FREE)) { 1941 panic("mbuf %p already freed", m); 1942 } 1943 m->m_type = MT_FREE; 1944 m->m_data = NULL; 1945 pool_cache_put(mb_cache, m); 1946 } 1947 1948 return n; 1949 } 1950 1951 void 1952 m_freem(struct mbuf *m) 1953 { 1954 if (m == NULL) 1955 return; 1956 do { 1957 m = m_free(m); 1958 } while (m); 1959 } 1960 1961 #if defined(DDB) 1962 void 1963 m_print(const struct mbuf *m, const char *modif, void (*pr)(const char *, ...)) 1964 { 1965 char ch; 1966 bool opt_c = false; 1967 bool opt_d = false; 1968 #if NETHER > 0 1969 bool opt_v = false; 1970 const struct mbuf *m0 = NULL; 1971 #endif 1972 int no = 0; 1973 char buf[512]; 1974 1975 while ((ch = *(modif++)) != '\0') { 1976 switch (ch) { 1977 case 'c': 1978 opt_c = true; 1979 break; 1980 case 'd': 1981 opt_d = true; 1982 break; 1983 #if NETHER > 0 1984 case 'v': 1985 opt_v = true; 1986 m0 = m; 1987 break; 1988 #endif 1989 default: 1990 break; 1991 } 1992 } 1993 1994 nextchain: 1995 (*pr)("MBUF(%d) %p\n", no, m); 1996 snprintb(buf, sizeof(buf), M_FLAGS_BITS, (u_int)m->m_flags); 1997 (*pr)(" data=%p, len=%d, type=%d, flags=%s\n", 1998 m->m_data, m->m_len, m->m_type, buf); 1999 if (opt_d) { 2000 int i; 2001 unsigned char *p = m->m_data; 2002 2003 (*pr)(" data:"); 2004 2005 for (i = 0; i < m->m_len; i++) { 2006 if (i % 16 == 0) 2007 (*pr)("\n"); 2008 (*pr)(" %02x", p[i]); 2009 } 2010 2011 (*pr)("\n"); 2012 } 2013 (*pr)(" owner=%p, next=%p, nextpkt=%p\n", m->m_owner, m->m_next, 2014 m->m_nextpkt); 2015 (*pr)(" leadingspace=%u, trailingspace=%u, readonly=%u\n", 2016 (int)M_LEADINGSPACE(m), (int)M_TRAILINGSPACE(m), 2017 (int)M_READONLY(m)); 2018 if ((m->m_flags & M_PKTHDR) != 0) { 2019 snprintb(buf, sizeof(buf), M_CSUM_BITS, m->m_pkthdr.csum_flags); 2020 (*pr)(" pktlen=%d, rcvif=%p, csum_flags=%s, csum_data=0x%" 2021 PRIx32 ", segsz=%u\n", 2022 m->m_pkthdr.len, m_get_rcvif_NOMPSAFE(m), 2023 buf, m->m_pkthdr.csum_data, m->m_pkthdr.segsz); 2024 } 2025 if ((m->m_flags & M_EXT)) { 2026 (*pr)(" ext_refcnt=%u, ext_buf=%p, ext_size=%zd, " 2027 "ext_free=%p, ext_arg=%p\n", 2028 m->m_ext.ext_refcnt, 2029 m->m_ext.ext_buf, m->m_ext.ext_size, 2030 m->m_ext.ext_free, m->m_ext.ext_arg); 2031 } 2032 if ((~m->m_flags & (M_EXT|M_EXT_PAGES)) == 0) { 2033 vaddr_t sva = (vaddr_t)m->m_ext.ext_buf; 2034 vaddr_t eva = sva + m->m_ext.ext_size; 2035 int n = (round_page(eva) - trunc_page(sva)) >> PAGE_SHIFT; 2036 int i; 2037 2038 (*pr)(" pages:"); 2039 for (i = 0; i < n; i ++) { 2040 (*pr)(" %p", m->m_ext.ext_pgs[i]); 2041 } 2042 (*pr)("\n"); 2043 } 2044 2045 if (opt_c) { 2046 m = m->m_next; 2047 if (m != NULL) { 2048 no++; 2049 goto nextchain; 2050 } 2051 } 2052 2053 #if NETHER > 0 2054 if (opt_v && m0) 2055 m_examine(m0, AF_ETHER, modif, pr); 2056 #endif 2057 } 2058 #endif /* defined(DDB) */ 2059 2060 #if defined(MBUFTRACE) 2061 void 2062 mowner_attach(struct mowner *mo) 2063 { 2064 2065 KASSERT(mo->mo_counters == NULL); 2066 mo->mo_counters = percpu_alloc(sizeof(struct mowner_counter)); 2067 2068 /* XXX lock */ 2069 LIST_INSERT_HEAD(&mowners, mo, mo_link); 2070 } 2071 2072 void 2073 mowner_detach(struct mowner *mo) 2074 { 2075 2076 KASSERT(mo->mo_counters != NULL); 2077 2078 /* XXX lock */ 2079 LIST_REMOVE(mo, mo_link); 2080 2081 percpu_free(mo->mo_counters, sizeof(struct mowner_counter)); 2082 mo->mo_counters = NULL; 2083 } 2084 2085 void 2086 mowner_init(struct mbuf *m, int type) 2087 { 2088 struct mowner_counter *mc; 2089 struct mowner *mo; 2090 int s; 2091 2092 m->m_owner = mo = &unknown_mowners[type]; 2093 s = splvm(); 2094 mc = percpu_getref(mo->mo_counters); 2095 mc->mc_counter[MOWNER_COUNTER_CLAIMS]++; 2096 percpu_putref(mo->mo_counters); 2097 splx(s); 2098 } 2099 2100 void 2101 mowner_ref(struct mbuf *m, int flags) 2102 { 2103 struct mowner *mo = m->m_owner; 2104 struct mowner_counter *mc; 2105 int s; 2106 2107 s = splvm(); 2108 mc = percpu_getref(mo->mo_counters); 2109 if ((flags & M_EXT) != 0) 2110 mc->mc_counter[MOWNER_COUNTER_EXT_CLAIMS]++; 2111 if ((flags & M_EXT_CLUSTER) != 0) 2112 mc->mc_counter[MOWNER_COUNTER_CLUSTER_CLAIMS]++; 2113 percpu_putref(mo->mo_counters); 2114 splx(s); 2115 } 2116 2117 void 2118 mowner_revoke(struct mbuf *m, bool all, int flags) 2119 { 2120 struct mowner *mo = m->m_owner; 2121 struct mowner_counter *mc; 2122 int s; 2123 2124 s = splvm(); 2125 mc = percpu_getref(mo->mo_counters); 2126 if ((flags & M_EXT) != 0) 2127 mc->mc_counter[MOWNER_COUNTER_EXT_RELEASES]++; 2128 if ((flags & M_EXT_CLUSTER) != 0) 2129 mc->mc_counter[MOWNER_COUNTER_CLUSTER_RELEASES]++; 2130 if (all) 2131 mc->mc_counter[MOWNER_COUNTER_RELEASES]++; 2132 percpu_putref(mo->mo_counters); 2133 splx(s); 2134 if (all) 2135 m->m_owner = &revoked_mowner; 2136 } 2137 2138 static void 2139 mowner_claim(struct mbuf *m, struct mowner *mo) 2140 { 2141 struct mowner_counter *mc; 2142 int flags = m->m_flags; 2143 int s; 2144 2145 s = splvm(); 2146 mc = percpu_getref(mo->mo_counters); 2147 mc->mc_counter[MOWNER_COUNTER_CLAIMS]++; 2148 if ((flags & M_EXT) != 0) 2149 mc->mc_counter[MOWNER_COUNTER_EXT_CLAIMS]++; 2150 if ((flags & M_EXT_CLUSTER) != 0) 2151 mc->mc_counter[MOWNER_COUNTER_CLUSTER_CLAIMS]++; 2152 percpu_putref(mo->mo_counters); 2153 splx(s); 2154 m->m_owner = mo; 2155 } 2156 2157 void 2158 m_claim(struct mbuf *m, struct mowner *mo) 2159 { 2160 2161 if (m->m_owner == mo || mo == NULL) 2162 return; 2163 2164 mowner_revoke(m, true, m->m_flags); 2165 mowner_claim(m, mo); 2166 } 2167 2168 void 2169 m_claimm(struct mbuf *m, struct mowner *mo) 2170 { 2171 2172 for (; m != NULL; m = m->m_next) 2173 m_claim(m, mo); 2174 } 2175 #endif /* defined(MBUFTRACE) */ 2176 2177 #ifdef DIAGNOSTIC 2178 /* 2179 * Verify that the mbuf chain is not malformed. Used only for diagnostic. 2180 * Panics on error. 2181 */ 2182 void 2183 m_verify_packet(struct mbuf *m) 2184 { 2185 struct mbuf *n = m; 2186 char *low, *high, *dat; 2187 int totlen = 0, len; 2188 2189 if (__predict_false((m->m_flags & M_PKTHDR) == 0)) { 2190 panic("%s: mbuf doesn't have M_PKTHDR", __func__); 2191 } 2192 2193 while (n != NULL) { 2194 if (__predict_false(n->m_type == MT_FREE)) { 2195 panic("%s: mbuf already freed (n = %p)", __func__, n); 2196 } 2197 #if 0 2198 /* 2199 * This ought to be a rule of the mbuf API. Unfortunately, 2200 * many places don't respect that rule. 2201 */ 2202 if (__predict_false((n != m) && (n->m_flags & M_PKTHDR) != 0)) { 2203 panic("%s: M_PKTHDR set on secondary mbuf", __func__); 2204 } 2205 #endif 2206 if (__predict_false(n->m_nextpkt != NULL)) { 2207 panic("%s: m_nextpkt not null (m_nextpkt = %p)", 2208 __func__, n->m_nextpkt); 2209 } 2210 2211 dat = n->m_data; 2212 len = n->m_len; 2213 2214 if (n->m_flags & M_EXT) { 2215 low = n->m_ext.ext_buf; 2216 high = low + n->m_ext.ext_size; 2217 } else if (n->m_flags & M_PKTHDR) { 2218 low = n->m_pktdat; 2219 high = low + MHLEN; 2220 } else { 2221 low = n->m_dat; 2222 high = low + MLEN; 2223 } 2224 if (__predict_false(dat + len < dat)) { 2225 panic("%s: incorrect length (len = %d)", __func__, len); 2226 } 2227 if (__predict_false((dat < low) || (dat + len > high))) { 2228 panic("%s: m_data not in packet" 2229 "(dat = %p, len = %d, low = %p, high = %p)", 2230 __func__, dat, len, low, high); 2231 } 2232 2233 totlen += len; 2234 n = n->m_next; 2235 } 2236 2237 if (__predict_false(totlen != m->m_pkthdr.len)) { 2238 panic("%s: inconsistent mbuf length (%d != %d)", __func__, 2239 totlen, m->m_pkthdr.len); 2240 } 2241 } 2242 #endif 2243 2244 struct m_tag * 2245 m_tag_get(int type, int len, int wait) 2246 { 2247 struct m_tag *t; 2248 2249 if (len < 0) 2250 return NULL; 2251 t = malloc(len + sizeof(struct m_tag), M_PACKET_TAGS, wait); 2252 if (t == NULL) 2253 return NULL; 2254 t->m_tag_id = type; 2255 t->m_tag_len = len; 2256 return t; 2257 } 2258 2259 void 2260 m_tag_free(struct m_tag *t) 2261 { 2262 free(t, M_PACKET_TAGS); 2263 } 2264 2265 void 2266 m_tag_prepend(struct mbuf *m, struct m_tag *t) 2267 { 2268 KASSERT((m->m_flags & M_PKTHDR) != 0); 2269 SLIST_INSERT_HEAD(&m->m_pkthdr.tags, t, m_tag_link); 2270 } 2271 2272 void 2273 m_tag_unlink(struct mbuf *m, struct m_tag *t) 2274 { 2275 KASSERT((m->m_flags & M_PKTHDR) != 0); 2276 SLIST_REMOVE(&m->m_pkthdr.tags, t, m_tag, m_tag_link); 2277 } 2278 2279 void 2280 m_tag_delete(struct mbuf *m, struct m_tag *t) 2281 { 2282 m_tag_unlink(m, t); 2283 m_tag_free(t); 2284 } 2285 2286 void 2287 m_tag_delete_chain(struct mbuf *m) 2288 { 2289 struct m_tag *p, *q; 2290 2291 KASSERT((m->m_flags & M_PKTHDR) != 0); 2292 2293 p = SLIST_FIRST(&m->m_pkthdr.tags); 2294 if (p == NULL) 2295 return; 2296 while ((q = SLIST_NEXT(p, m_tag_link)) != NULL) 2297 m_tag_delete(m, q); 2298 m_tag_delete(m, p); 2299 } 2300 2301 struct m_tag * 2302 m_tag_find(const struct mbuf *m, int type) 2303 { 2304 struct m_tag *p; 2305 2306 KASSERT((m->m_flags & M_PKTHDR) != 0); 2307 2308 p = SLIST_FIRST(&m->m_pkthdr.tags); 2309 while (p != NULL) { 2310 if (p->m_tag_id == type) 2311 return p; 2312 p = SLIST_NEXT(p, m_tag_link); 2313 } 2314 return NULL; 2315 } 2316 2317 struct m_tag * 2318 m_tag_copy(struct m_tag *t) 2319 { 2320 struct m_tag *p; 2321 2322 p = m_tag_get(t->m_tag_id, t->m_tag_len, M_NOWAIT); 2323 if (p == NULL) 2324 return NULL; 2325 memcpy(p + 1, t + 1, t->m_tag_len); 2326 return p; 2327 } 2328 2329 /* 2330 * Copy two tag chains. The destination mbuf (to) loses any attached 2331 * tags even if the operation fails. This should not be a problem, as 2332 * m_tag_copy_chain() is typically called with a newly-allocated 2333 * destination mbuf. 2334 */ 2335 int 2336 m_tag_copy_chain(struct mbuf *to, struct mbuf *from) 2337 { 2338 struct m_tag *p, *t, *tprev = NULL; 2339 2340 KASSERT((from->m_flags & M_PKTHDR) != 0); 2341 2342 m_tag_delete_chain(to); 2343 SLIST_FOREACH(p, &from->m_pkthdr.tags, m_tag_link) { 2344 t = m_tag_copy(p); 2345 if (t == NULL) { 2346 m_tag_delete_chain(to); 2347 return 0; 2348 } 2349 if (tprev == NULL) 2350 SLIST_INSERT_HEAD(&to->m_pkthdr.tags, t, m_tag_link); 2351 else 2352 SLIST_INSERT_AFTER(tprev, t, m_tag_link); 2353 tprev = t; 2354 } 2355 return 1; 2356 } 2357