1 /*- 2 * Copyright (c) 1982, 1986, 1988, 1991, 1993 3 * The Regents of the University of California. All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 4. Neither the name of the University nor the names of its contributors 14 * may be used to endorse or promote products derived from this software 15 * without specific prior written permission. 16 * 17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 20 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 27 * SUCH DAMAGE. 28 * 29 * @(#)uipc_mbuf.c 8.2 (Berkeley) 1/4/94 30 */ 31 32 #include <sys/cdefs.h> 33 __FBSDID("$FreeBSD$"); 34 35 #include "opt_param.h" 36 #include "opt_mbuf_stress_test.h" 37 #include "opt_mbuf_profiling.h" 38 39 #include <sys/param.h> 40 #include <sys/systm.h> 41 #include <sys/kernel.h> 42 #include <sys/limits.h> 43 #include <sys/lock.h> 44 #include <sys/malloc.h> 45 #include <sys/mbuf.h> 46 #include <sys/sysctl.h> 47 #include <sys/domain.h> 48 #include <sys/protosw.h> 49 #include <sys/uio.h> 50 51 int max_linkhdr; 52 int max_protohdr; 53 int max_hdr; 54 int max_datalen; 55 #ifdef MBUF_STRESS_TEST 56 int m_defragpackets; 57 int m_defragbytes; 58 int m_defraguseless; 59 int m_defragfailure; 60 int m_defragrandomfailures; 61 #endif 62 63 /* 64 * sysctl(8) exported objects 65 */ 66 SYSCTL_INT(_kern_ipc, KIPC_MAX_LINKHDR, max_linkhdr, CTLFLAG_RD, 67 &max_linkhdr, 0, "Size of largest link layer header"); 68 SYSCTL_INT(_kern_ipc, KIPC_MAX_PROTOHDR, max_protohdr, CTLFLAG_RD, 69 &max_protohdr, 0, "Size of largest protocol layer header"); 70 SYSCTL_INT(_kern_ipc, KIPC_MAX_HDR, max_hdr, CTLFLAG_RD, 71 &max_hdr, 0, "Size of largest link plus protocol header"); 72 SYSCTL_INT(_kern_ipc, KIPC_MAX_DATALEN, max_datalen, CTLFLAG_RD, 73 &max_datalen, 0, "Minimum space left in mbuf after max_hdr"); 74 #ifdef MBUF_STRESS_TEST 75 SYSCTL_INT(_kern_ipc, OID_AUTO, m_defragpackets, CTLFLAG_RD, 76 &m_defragpackets, 0, ""); 77 SYSCTL_INT(_kern_ipc, OID_AUTO, m_defragbytes, CTLFLAG_RD, 78 &m_defragbytes, 0, ""); 79 SYSCTL_INT(_kern_ipc, OID_AUTO, m_defraguseless, CTLFLAG_RD, 80 &m_defraguseless, 0, ""); 81 SYSCTL_INT(_kern_ipc, OID_AUTO, m_defragfailure, CTLFLAG_RD, 82 &m_defragfailure, 0, ""); 83 SYSCTL_INT(_kern_ipc, OID_AUTO, m_defragrandomfailures, CTLFLAG_RW, 84 &m_defragrandomfailures, 0, ""); 85 #endif 86 87 /* 88 * Ensure the correct size of various mbuf parameters. It could be off due 89 * to compiler-induced padding and alignment artifacts. 90 */ 91 CTASSERT(sizeof(struct mbuf) == MSIZE); 92 CTASSERT(MSIZE - offsetof(struct mbuf, m_dat) == MLEN); 93 CTASSERT(MSIZE - offsetof(struct mbuf, m_pktdat) == MHLEN); 94 95 /* 96 * m_get2() allocates minimum mbuf that would fit "size" argument. 97 */ 98 struct mbuf * 99 m_get2(int size, int how, short type, int flags) 100 { 101 struct mb_args args; 102 struct mbuf *m, *n; 103 104 args.flags = flags; 105 args.type = type; 106 107 if (size <= MHLEN || (size <= MLEN && (flags & M_PKTHDR) == 0)) 108 return (uma_zalloc_arg(zone_mbuf, &args, how)); 109 if (size <= MCLBYTES) 110 return (uma_zalloc_arg(zone_pack, &args, how)); 111 112 if (size > MJUMPAGESIZE) 113 return (NULL); 114 115 m = uma_zalloc_arg(zone_mbuf, &args, how); 116 if (m == NULL) 117 return (NULL); 118 119 n = uma_zalloc_arg(zone_jumbop, m, how); 120 if (n == NULL) { 121 uma_zfree(zone_mbuf, m); 122 return (NULL); 123 } 124 125 return (m); 126 } 127 128 /* 129 * m_getjcl() returns an mbuf with a cluster of the specified size attached. 130 * For size it takes MCLBYTES, MJUMPAGESIZE, MJUM9BYTES, MJUM16BYTES. 131 */ 132 struct mbuf * 133 m_getjcl(int how, short type, int flags, int size) 134 { 135 struct mb_args args; 136 struct mbuf *m, *n; 137 uma_zone_t zone; 138 139 if (size == MCLBYTES) 140 return m_getcl(how, type, flags); 141 142 args.flags = flags; 143 args.type = type; 144 145 m = uma_zalloc_arg(zone_mbuf, &args, how); 146 if (m == NULL) 147 return (NULL); 148 149 zone = m_getzone(size); 150 n = uma_zalloc_arg(zone, m, how); 151 if (n == NULL) { 152 uma_zfree(zone_mbuf, m); 153 return (NULL); 154 } 155 return (m); 156 } 157 158 /* 159 * Allocate a given length worth of mbufs and/or clusters (whatever fits 160 * best) and return a pointer to the top of the allocated chain. If an 161 * existing mbuf chain is provided, then we will append the new chain 162 * to the existing one but still return the top of the newly allocated 163 * chain. 164 */ 165 struct mbuf * 166 m_getm2(struct mbuf *m, int len, int how, short type, int flags) 167 { 168 struct mbuf *mb, *nm = NULL, *mtail = NULL; 169 170 KASSERT(len >= 0, ("%s: len is < 0", __func__)); 171 172 /* Validate flags. */ 173 flags &= (M_PKTHDR | M_EOR); 174 175 /* Packet header mbuf must be first in chain. */ 176 if ((flags & M_PKTHDR) && m != NULL) 177 flags &= ~M_PKTHDR; 178 179 /* Loop and append maximum sized mbufs to the chain tail. */ 180 while (len > 0) { 181 if (len > MCLBYTES) 182 mb = m_getjcl(how, type, (flags & M_PKTHDR), 183 MJUMPAGESIZE); 184 else if (len >= MINCLSIZE) 185 mb = m_getcl(how, type, (flags & M_PKTHDR)); 186 else if (flags & M_PKTHDR) 187 mb = m_gethdr(how, type); 188 else 189 mb = m_get(how, type); 190 191 /* Fail the whole operation if one mbuf can't be allocated. */ 192 if (mb == NULL) { 193 if (nm != NULL) 194 m_freem(nm); 195 return (NULL); 196 } 197 198 /* Book keeping. */ 199 len -= (mb->m_flags & M_EXT) ? mb->m_ext.ext_size : 200 ((mb->m_flags & M_PKTHDR) ? MHLEN : MLEN); 201 if (mtail != NULL) 202 mtail->m_next = mb; 203 else 204 nm = mb; 205 mtail = mb; 206 flags &= ~M_PKTHDR; /* Only valid on the first mbuf. */ 207 } 208 if (flags & M_EOR) 209 mtail->m_flags |= M_EOR; /* Only valid on the last mbuf. */ 210 211 /* If mbuf was supplied, append new chain to the end of it. */ 212 if (m != NULL) { 213 for (mtail = m; mtail->m_next != NULL; mtail = mtail->m_next) 214 ; 215 mtail->m_next = nm; 216 mtail->m_flags &= ~M_EOR; 217 } else 218 m = nm; 219 220 return (m); 221 } 222 223 /* 224 * Free an entire chain of mbufs and associated external buffers, if 225 * applicable. 226 */ 227 void 228 m_freem(struct mbuf *mb) 229 { 230 231 while (mb != NULL) 232 mb = m_free(mb); 233 } 234 235 /*- 236 * Configure a provided mbuf to refer to the provided external storage 237 * buffer and setup a reference count for said buffer. If the setting 238 * up of the reference count fails, the M_EXT bit will not be set. If 239 * successfull, the M_EXT bit is set in the mbuf's flags. 240 * 241 * Arguments: 242 * mb The existing mbuf to which to attach the provided buffer. 243 * buf The address of the provided external storage buffer. 244 * size The size of the provided buffer. 245 * freef A pointer to a routine that is responsible for freeing the 246 * provided external storage buffer. 247 * args A pointer to an argument structure (of any type) to be passed 248 * to the provided freef routine (may be NULL). 249 * flags Any other flags to be passed to the provided mbuf. 250 * type The type that the external storage buffer should be 251 * labeled with. 252 * 253 * Returns: 254 * Nothing. 255 */ 256 int 257 m_extadd(struct mbuf *mb, caddr_t buf, u_int size, 258 int (*freef)(struct mbuf *, void *, void *), void *arg1, void *arg2, 259 int flags, int type, int wait) 260 { 261 KASSERT(type != EXT_CLUSTER, ("%s: EXT_CLUSTER not allowed", __func__)); 262 263 if (type != EXT_EXTREF) 264 mb->m_ext.ref_cnt = uma_zalloc(zone_ext_refcnt, wait); 265 266 if (mb->m_ext.ref_cnt == NULL) 267 return (ENOMEM); 268 269 *(mb->m_ext.ref_cnt) = 1; 270 mb->m_flags |= (M_EXT | flags); 271 mb->m_ext.ext_buf = buf; 272 mb->m_data = mb->m_ext.ext_buf; 273 mb->m_ext.ext_size = size; 274 mb->m_ext.ext_free = freef; 275 mb->m_ext.ext_arg1 = arg1; 276 mb->m_ext.ext_arg2 = arg2; 277 mb->m_ext.ext_type = type; 278 mb->m_ext.ext_flags = 0; 279 280 return (0); 281 } 282 283 /* 284 * Non-directly-exported function to clean up after mbufs with M_EXT 285 * storage attached to them if the reference count hits 1. 286 */ 287 void 288 mb_free_ext(struct mbuf *m) 289 { 290 int skipmbuf; 291 292 KASSERT((m->m_flags & M_EXT) == M_EXT, ("%s: M_EXT not set", __func__)); 293 KASSERT(m->m_ext.ref_cnt != NULL, ("%s: ref_cnt not set", __func__)); 294 295 /* 296 * check if the header is embedded in the cluster 297 */ 298 skipmbuf = (m->m_flags & M_NOFREE); 299 300 /* Free attached storage if this mbuf is the only reference to it. */ 301 if (*(m->m_ext.ref_cnt) == 1 || 302 atomic_fetchadd_int(m->m_ext.ref_cnt, -1) == 1) { 303 switch (m->m_ext.ext_type) { 304 case EXT_PACKET: /* The packet zone is special. */ 305 if (*(m->m_ext.ref_cnt) == 0) 306 *(m->m_ext.ref_cnt) = 1; 307 uma_zfree(zone_pack, m); 308 return; /* Job done. */ 309 case EXT_CLUSTER: 310 uma_zfree(zone_clust, m->m_ext.ext_buf); 311 break; 312 case EXT_JUMBOP: 313 uma_zfree(zone_jumbop, m->m_ext.ext_buf); 314 break; 315 case EXT_JUMBO9: 316 uma_zfree(zone_jumbo9, m->m_ext.ext_buf); 317 break; 318 case EXT_JUMBO16: 319 uma_zfree(zone_jumbo16, m->m_ext.ext_buf); 320 break; 321 case EXT_SFBUF: 322 case EXT_NET_DRV: 323 case EXT_MOD_TYPE: 324 case EXT_DISPOSABLE: 325 *(m->m_ext.ref_cnt) = 0; 326 uma_zfree(zone_ext_refcnt, __DEVOLATILE(u_int *, 327 m->m_ext.ref_cnt)); 328 /* FALLTHROUGH */ 329 case EXT_EXTREF: 330 KASSERT(m->m_ext.ext_free != NULL, 331 ("%s: ext_free not set", __func__)); 332 (void)(*(m->m_ext.ext_free))(m, m->m_ext.ext_arg1, 333 m->m_ext.ext_arg2); 334 break; 335 default: 336 KASSERT(m->m_ext.ext_type == 0, 337 ("%s: unknown ext_type", __func__)); 338 } 339 } 340 if (skipmbuf) 341 return; 342 343 /* 344 * Free this mbuf back to the mbuf zone with all m_ext 345 * information purged. 346 */ 347 m->m_ext.ext_buf = NULL; 348 m->m_ext.ext_free = NULL; 349 m->m_ext.ext_arg1 = NULL; 350 m->m_ext.ext_arg2 = NULL; 351 m->m_ext.ref_cnt = NULL; 352 m->m_ext.ext_size = 0; 353 m->m_ext.ext_type = 0; 354 m->m_ext.ext_flags = 0; 355 m->m_flags &= ~M_EXT; 356 uma_zfree(zone_mbuf, m); 357 } 358 359 /* 360 * Attach the cluster from *m to *n, set up m_ext in *n 361 * and bump the refcount of the cluster. 362 */ 363 static void 364 mb_dupcl(struct mbuf *n, struct mbuf *m) 365 { 366 KASSERT((m->m_flags & M_EXT) == M_EXT, ("%s: M_EXT not set", __func__)); 367 KASSERT(m->m_ext.ref_cnt != NULL, ("%s: ref_cnt not set", __func__)); 368 KASSERT((n->m_flags & M_EXT) == 0, ("%s: M_EXT set", __func__)); 369 370 if (*(m->m_ext.ref_cnt) == 1) 371 *(m->m_ext.ref_cnt) += 1; 372 else 373 atomic_add_int(m->m_ext.ref_cnt, 1); 374 n->m_ext.ext_buf = m->m_ext.ext_buf; 375 n->m_ext.ext_free = m->m_ext.ext_free; 376 n->m_ext.ext_arg1 = m->m_ext.ext_arg1; 377 n->m_ext.ext_arg2 = m->m_ext.ext_arg2; 378 n->m_ext.ext_size = m->m_ext.ext_size; 379 n->m_ext.ref_cnt = m->m_ext.ref_cnt; 380 n->m_ext.ext_type = m->m_ext.ext_type; 381 n->m_ext.ext_flags = m->m_ext.ext_flags; 382 n->m_flags |= M_EXT; 383 n->m_flags |= m->m_flags & M_RDONLY; 384 } 385 386 /* 387 * Clean up mbuf (chain) from any tags and packet headers. 388 * If "all" is set then the first mbuf in the chain will be 389 * cleaned too. 390 */ 391 void 392 m_demote(struct mbuf *m0, int all) 393 { 394 struct mbuf *m; 395 396 for (m = all ? m0 : m0->m_next; m != NULL; m = m->m_next) { 397 if (m->m_flags & M_PKTHDR) { 398 m_tag_delete_chain(m, NULL); 399 m->m_flags &= ~M_PKTHDR; 400 bzero(&m->m_pkthdr, sizeof(struct pkthdr)); 401 } 402 if (m != m0 && m->m_nextpkt != NULL) { 403 KASSERT(m->m_nextpkt == NULL, 404 ("%s: m_nextpkt not NULL", __func__)); 405 m_freem(m->m_nextpkt); 406 m->m_nextpkt = NULL; 407 } 408 m->m_flags = m->m_flags & (M_EXT|M_RDONLY|M_NOFREE); 409 } 410 } 411 412 /* 413 * Sanity checks on mbuf (chain) for use in KASSERT() and general 414 * debugging. 415 * Returns 0 or panics when bad and 1 on all tests passed. 416 * Sanitize, 0 to run M_SANITY_ACTION, 1 to garble things so they 417 * blow up later. 418 */ 419 int 420 m_sanity(struct mbuf *m0, int sanitize) 421 { 422 struct mbuf *m; 423 caddr_t a, b; 424 int pktlen = 0; 425 426 #ifdef INVARIANTS 427 #define M_SANITY_ACTION(s) panic("mbuf %p: " s, m) 428 #else 429 #define M_SANITY_ACTION(s) printf("mbuf %p: " s, m) 430 #endif 431 432 for (m = m0; m != NULL; m = m->m_next) { 433 /* 434 * Basic pointer checks. If any of these fails then some 435 * unrelated kernel memory before or after us is trashed. 436 * No way to recover from that. 437 */ 438 a = ((m->m_flags & M_EXT) ? m->m_ext.ext_buf : 439 ((m->m_flags & M_PKTHDR) ? (caddr_t)(&m->m_pktdat) : 440 (caddr_t)(&m->m_dat)) ); 441 b = (caddr_t)(a + (m->m_flags & M_EXT ? m->m_ext.ext_size : 442 ((m->m_flags & M_PKTHDR) ? MHLEN : MLEN))); 443 if ((caddr_t)m->m_data < a) 444 M_SANITY_ACTION("m_data outside mbuf data range left"); 445 if ((caddr_t)m->m_data > b) 446 M_SANITY_ACTION("m_data outside mbuf data range right"); 447 if ((caddr_t)m->m_data + m->m_len > b) 448 M_SANITY_ACTION("m_data + m_len exeeds mbuf space"); 449 450 /* m->m_nextpkt may only be set on first mbuf in chain. */ 451 if (m != m0 && m->m_nextpkt != NULL) { 452 if (sanitize) { 453 m_freem(m->m_nextpkt); 454 m->m_nextpkt = (struct mbuf *)0xDEADC0DE; 455 } else 456 M_SANITY_ACTION("m->m_nextpkt on in-chain mbuf"); 457 } 458 459 /* packet length (not mbuf length!) calculation */ 460 if (m0->m_flags & M_PKTHDR) 461 pktlen += m->m_len; 462 463 /* m_tags may only be attached to first mbuf in chain. */ 464 if (m != m0 && m->m_flags & M_PKTHDR && 465 !SLIST_EMPTY(&m->m_pkthdr.tags)) { 466 if (sanitize) { 467 m_tag_delete_chain(m, NULL); 468 /* put in 0xDEADC0DE perhaps? */ 469 } else 470 M_SANITY_ACTION("m_tags on in-chain mbuf"); 471 } 472 473 /* M_PKTHDR may only be set on first mbuf in chain */ 474 if (m != m0 && m->m_flags & M_PKTHDR) { 475 if (sanitize) { 476 bzero(&m->m_pkthdr, sizeof(m->m_pkthdr)); 477 m->m_flags &= ~M_PKTHDR; 478 /* put in 0xDEADCODE and leave hdr flag in */ 479 } else 480 M_SANITY_ACTION("M_PKTHDR on in-chain mbuf"); 481 } 482 } 483 m = m0; 484 if (pktlen && pktlen != m->m_pkthdr.len) { 485 if (sanitize) 486 m->m_pkthdr.len = 0; 487 else 488 M_SANITY_ACTION("m_pkthdr.len != mbuf chain length"); 489 } 490 return 1; 491 492 #undef M_SANITY_ACTION 493 } 494 495 496 /* 497 * "Move" mbuf pkthdr from "from" to "to". 498 * "from" must have M_PKTHDR set, and "to" must be empty. 499 */ 500 void 501 m_move_pkthdr(struct mbuf *to, struct mbuf *from) 502 { 503 504 #if 0 505 /* see below for why these are not enabled */ 506 M_ASSERTPKTHDR(to); 507 /* Note: with MAC, this may not be a good assertion. */ 508 KASSERT(SLIST_EMPTY(&to->m_pkthdr.tags), 509 ("m_move_pkthdr: to has tags")); 510 #endif 511 #ifdef MAC 512 /* 513 * XXXMAC: It could be this should also occur for non-MAC? 514 */ 515 if (to->m_flags & M_PKTHDR) 516 m_tag_delete_chain(to, NULL); 517 #endif 518 to->m_flags = (from->m_flags & M_COPYFLAGS) | (to->m_flags & M_EXT); 519 if ((to->m_flags & M_EXT) == 0) 520 to->m_data = to->m_pktdat; 521 to->m_pkthdr = from->m_pkthdr; /* especially tags */ 522 SLIST_INIT(&from->m_pkthdr.tags); /* purge tags from src */ 523 from->m_flags &= ~M_PKTHDR; 524 } 525 526 /* 527 * Duplicate "from"'s mbuf pkthdr in "to". 528 * "from" must have M_PKTHDR set, and "to" must be empty. 529 * In particular, this does a deep copy of the packet tags. 530 */ 531 int 532 m_dup_pkthdr(struct mbuf *to, struct mbuf *from, int how) 533 { 534 535 #if 0 536 /* 537 * The mbuf allocator only initializes the pkthdr 538 * when the mbuf is allocated with m_gethdr(). Many users 539 * (e.g. m_copy*, m_prepend) use m_get() and then 540 * smash the pkthdr as needed causing these 541 * assertions to trip. For now just disable them. 542 */ 543 M_ASSERTPKTHDR(to); 544 /* Note: with MAC, this may not be a good assertion. */ 545 KASSERT(SLIST_EMPTY(&to->m_pkthdr.tags), ("m_dup_pkthdr: to has tags")); 546 #endif 547 MBUF_CHECKSLEEP(how); 548 #ifdef MAC 549 if (to->m_flags & M_PKTHDR) 550 m_tag_delete_chain(to, NULL); 551 #endif 552 to->m_flags = (from->m_flags & M_COPYFLAGS) | (to->m_flags & M_EXT); 553 if ((to->m_flags & M_EXT) == 0) 554 to->m_data = to->m_pktdat; 555 to->m_pkthdr = from->m_pkthdr; 556 SLIST_INIT(&to->m_pkthdr.tags); 557 return (m_tag_copy_chain(to, from, MBTOM(how))); 558 } 559 560 /* 561 * Lesser-used path for M_PREPEND: 562 * allocate new mbuf to prepend to chain, 563 * copy junk along. 564 */ 565 struct mbuf * 566 m_prepend(struct mbuf *m, int len, int how) 567 { 568 struct mbuf *mn; 569 570 if (m->m_flags & M_PKTHDR) 571 mn = m_gethdr(how, m->m_type); 572 else 573 mn = m_get(how, m->m_type); 574 if (mn == NULL) { 575 m_freem(m); 576 return (NULL); 577 } 578 if (m->m_flags & M_PKTHDR) 579 m_move_pkthdr(mn, m); 580 mn->m_next = m; 581 m = mn; 582 if(m->m_flags & M_PKTHDR) { 583 if (len < MHLEN) 584 MH_ALIGN(m, len); 585 } else { 586 if (len < MLEN) 587 M_ALIGN(m, len); 588 } 589 m->m_len = len; 590 return (m); 591 } 592 593 /* 594 * Make a copy of an mbuf chain starting "off0" bytes from the beginning, 595 * continuing for "len" bytes. If len is M_COPYALL, copy to end of mbuf. 596 * The wait parameter is a choice of M_WAITOK/M_NOWAIT from caller. 597 * Note that the copy is read-only, because clusters are not copied, 598 * only their reference counts are incremented. 599 */ 600 struct mbuf * 601 m_copym(struct mbuf *m, int off0, int len, int wait) 602 { 603 struct mbuf *n, **np; 604 int off = off0; 605 struct mbuf *top; 606 int copyhdr = 0; 607 608 KASSERT(off >= 0, ("m_copym, negative off %d", off)); 609 KASSERT(len >= 0, ("m_copym, negative len %d", len)); 610 MBUF_CHECKSLEEP(wait); 611 if (off == 0 && m->m_flags & M_PKTHDR) 612 copyhdr = 1; 613 while (off > 0) { 614 KASSERT(m != NULL, ("m_copym, offset > size of mbuf chain")); 615 if (off < m->m_len) 616 break; 617 off -= m->m_len; 618 m = m->m_next; 619 } 620 np = ⊤ 621 top = 0; 622 while (len > 0) { 623 if (m == NULL) { 624 KASSERT(len == M_COPYALL, 625 ("m_copym, length > size of mbuf chain")); 626 break; 627 } 628 if (copyhdr) 629 n = m_gethdr(wait, m->m_type); 630 else 631 n = m_get(wait, m->m_type); 632 *np = n; 633 if (n == NULL) 634 goto nospace; 635 if (copyhdr) { 636 if (!m_dup_pkthdr(n, m, wait)) 637 goto nospace; 638 if (len == M_COPYALL) 639 n->m_pkthdr.len -= off0; 640 else 641 n->m_pkthdr.len = len; 642 copyhdr = 0; 643 } 644 n->m_len = min(len, m->m_len - off); 645 if (m->m_flags & M_EXT) { 646 n->m_data = m->m_data + off; 647 mb_dupcl(n, m); 648 } else 649 bcopy(mtod(m, caddr_t)+off, mtod(n, caddr_t), 650 (u_int)n->m_len); 651 if (len != M_COPYALL) 652 len -= n->m_len; 653 off = 0; 654 m = m->m_next; 655 np = &n->m_next; 656 } 657 658 return (top); 659 nospace: 660 m_freem(top); 661 return (NULL); 662 } 663 664 /* 665 * Returns mbuf chain with new head for the prepending case. 666 * Copies from mbuf (chain) n from off for len to mbuf (chain) m 667 * either prepending or appending the data. 668 * The resulting mbuf (chain) m is fully writeable. 669 * m is destination (is made writeable) 670 * n is source, off is offset in source, len is len from offset 671 * dir, 0 append, 1 prepend 672 * how, wait or nowait 673 */ 674 675 static int 676 m_bcopyxxx(void *s, void *t, u_int len) 677 { 678 bcopy(s, t, (size_t)len); 679 return 0; 680 } 681 682 struct mbuf * 683 m_copymdata(struct mbuf *m, struct mbuf *n, int off, int len, 684 int prep, int how) 685 { 686 struct mbuf *mm, *x, *z, *prev = NULL; 687 caddr_t p; 688 int i, nlen = 0; 689 caddr_t buf[MLEN]; 690 691 KASSERT(m != NULL && n != NULL, ("m_copymdata, no target or source")); 692 KASSERT(off >= 0, ("m_copymdata, negative off %d", off)); 693 KASSERT(len >= 0, ("m_copymdata, negative len %d", len)); 694 KASSERT(prep == 0 || prep == 1, ("m_copymdata, unknown direction %d", prep)); 695 696 mm = m; 697 if (!prep) { 698 while(mm->m_next) { 699 prev = mm; 700 mm = mm->m_next; 701 } 702 } 703 for (z = n; z != NULL; z = z->m_next) 704 nlen += z->m_len; 705 if (len == M_COPYALL) 706 len = nlen - off; 707 if (off + len > nlen || len < 1) 708 return NULL; 709 710 if (!M_WRITABLE(mm)) { 711 /* XXX: Use proper m_xxx function instead. */ 712 x = m_getcl(how, MT_DATA, mm->m_flags); 713 if (x == NULL) 714 return NULL; 715 bcopy(mm->m_ext.ext_buf, x->m_ext.ext_buf, x->m_ext.ext_size); 716 p = x->m_ext.ext_buf + (mm->m_data - mm->m_ext.ext_buf); 717 x->m_data = p; 718 mm->m_next = NULL; 719 if (mm != m) 720 prev->m_next = x; 721 m_free(mm); 722 mm = x; 723 } 724 725 /* 726 * Append/prepend the data. Allocating mbufs as necessary. 727 */ 728 /* Shortcut if enough free space in first/last mbuf. */ 729 if (!prep && M_TRAILINGSPACE(mm) >= len) { 730 m_apply(n, off, len, m_bcopyxxx, mtod(mm, caddr_t) + 731 mm->m_len); 732 mm->m_len += len; 733 mm->m_pkthdr.len += len; 734 return m; 735 } 736 if (prep && M_LEADINGSPACE(mm) >= len) { 737 mm->m_data = mtod(mm, caddr_t) - len; 738 m_apply(n, off, len, m_bcopyxxx, mtod(mm, caddr_t)); 739 mm->m_len += len; 740 mm->m_pkthdr.len += len; 741 return mm; 742 } 743 744 /* Expand first/last mbuf to cluster if possible. */ 745 if (!prep && !(mm->m_flags & M_EXT) && len > M_TRAILINGSPACE(mm)) { 746 bcopy(mm->m_data, &buf, mm->m_len); 747 m_clget(mm, how); 748 if (!(mm->m_flags & M_EXT)) 749 return NULL; 750 bcopy(&buf, mm->m_ext.ext_buf, mm->m_len); 751 mm->m_data = mm->m_ext.ext_buf; 752 } 753 if (prep && !(mm->m_flags & M_EXT) && len > M_LEADINGSPACE(mm)) { 754 bcopy(mm->m_data, &buf, mm->m_len); 755 m_clget(mm, how); 756 if (!(mm->m_flags & M_EXT)) 757 return NULL; 758 bcopy(&buf, (caddr_t *)mm->m_ext.ext_buf + 759 mm->m_ext.ext_size - mm->m_len, mm->m_len); 760 mm->m_data = (caddr_t)mm->m_ext.ext_buf + 761 mm->m_ext.ext_size - mm->m_len; 762 } 763 764 /* Append/prepend as many mbuf (clusters) as necessary to fit len. */ 765 if (!prep && len > M_TRAILINGSPACE(mm)) { 766 if (!m_getm(mm, len - M_TRAILINGSPACE(mm), how, MT_DATA)) 767 return NULL; 768 } 769 if (prep && len > M_LEADINGSPACE(mm)) { 770 if (!(z = m_getm(NULL, len - M_LEADINGSPACE(mm), how, MT_DATA))) 771 return NULL; 772 i = 0; 773 for (x = z; x != NULL; x = x->m_next) { 774 i += x->m_flags & M_EXT ? x->m_ext.ext_size : 775 (x->m_flags & M_PKTHDR ? MHLEN : MLEN); 776 if (!x->m_next) 777 break; 778 } 779 z->m_data += i - len; 780 m_move_pkthdr(mm, z); 781 x->m_next = mm; 782 mm = z; 783 } 784 785 /* Seek to start position in source mbuf. Optimization for long chains. */ 786 while (off > 0) { 787 if (off < n->m_len) 788 break; 789 off -= n->m_len; 790 n = n->m_next; 791 } 792 793 /* Copy data into target mbuf. */ 794 z = mm; 795 while (len > 0) { 796 KASSERT(z != NULL, ("m_copymdata, falling off target edge")); 797 i = M_TRAILINGSPACE(z); 798 m_apply(n, off, i, m_bcopyxxx, mtod(z, caddr_t) + z->m_len); 799 z->m_len += i; 800 /* fixup pkthdr.len if necessary */ 801 if ((prep ? mm : m)->m_flags & M_PKTHDR) 802 (prep ? mm : m)->m_pkthdr.len += i; 803 off += i; 804 len -= i; 805 z = z->m_next; 806 } 807 return (prep ? mm : m); 808 } 809 810 /* 811 * Copy an entire packet, including header (which must be present). 812 * An optimization of the common case `m_copym(m, 0, M_COPYALL, how)'. 813 * Note that the copy is read-only, because clusters are not copied, 814 * only their reference counts are incremented. 815 * Preserve alignment of the first mbuf so if the creator has left 816 * some room at the beginning (e.g. for inserting protocol headers) 817 * the copies still have the room available. 818 */ 819 struct mbuf * 820 m_copypacket(struct mbuf *m, int how) 821 { 822 struct mbuf *top, *n, *o; 823 824 MBUF_CHECKSLEEP(how); 825 n = m_get(how, m->m_type); 826 top = n; 827 if (n == NULL) 828 goto nospace; 829 830 if (!m_dup_pkthdr(n, m, how)) 831 goto nospace; 832 n->m_len = m->m_len; 833 if (m->m_flags & M_EXT) { 834 n->m_data = m->m_data; 835 mb_dupcl(n, m); 836 } else { 837 n->m_data = n->m_pktdat + (m->m_data - m->m_pktdat ); 838 bcopy(mtod(m, char *), mtod(n, char *), n->m_len); 839 } 840 841 m = m->m_next; 842 while (m) { 843 o = m_get(how, m->m_type); 844 if (o == NULL) 845 goto nospace; 846 847 n->m_next = o; 848 n = n->m_next; 849 850 n->m_len = m->m_len; 851 if (m->m_flags & M_EXT) { 852 n->m_data = m->m_data; 853 mb_dupcl(n, m); 854 } else { 855 bcopy(mtod(m, char *), mtod(n, char *), n->m_len); 856 } 857 858 m = m->m_next; 859 } 860 return top; 861 nospace: 862 m_freem(top); 863 return (NULL); 864 } 865 866 /* 867 * Copy data from an mbuf chain starting "off" bytes from the beginning, 868 * continuing for "len" bytes, into the indicated buffer. 869 */ 870 void 871 m_copydata(const struct mbuf *m, int off, int len, caddr_t cp) 872 { 873 u_int count; 874 875 KASSERT(off >= 0, ("m_copydata, negative off %d", off)); 876 KASSERT(len >= 0, ("m_copydata, negative len %d", len)); 877 while (off > 0) { 878 KASSERT(m != NULL, ("m_copydata, offset > size of mbuf chain")); 879 if (off < m->m_len) 880 break; 881 off -= m->m_len; 882 m = m->m_next; 883 } 884 while (len > 0) { 885 KASSERT(m != NULL, ("m_copydata, length > size of mbuf chain")); 886 count = min(m->m_len - off, len); 887 bcopy(mtod(m, caddr_t) + off, cp, count); 888 len -= count; 889 cp += count; 890 off = 0; 891 m = m->m_next; 892 } 893 } 894 895 /* 896 * Copy a packet header mbuf chain into a completely new chain, including 897 * copying any mbuf clusters. Use this instead of m_copypacket() when 898 * you need a writable copy of an mbuf chain. 899 */ 900 struct mbuf * 901 m_dup(struct mbuf *m, int how) 902 { 903 struct mbuf **p, *top = NULL; 904 int remain, moff, nsize; 905 906 MBUF_CHECKSLEEP(how); 907 /* Sanity check */ 908 if (m == NULL) 909 return (NULL); 910 M_ASSERTPKTHDR(m); 911 912 /* While there's more data, get a new mbuf, tack it on, and fill it */ 913 remain = m->m_pkthdr.len; 914 moff = 0; 915 p = ⊤ 916 while (remain > 0 || top == NULL) { /* allow m->m_pkthdr.len == 0 */ 917 struct mbuf *n; 918 919 /* Get the next new mbuf */ 920 if (remain >= MINCLSIZE) { 921 n = m_getcl(how, m->m_type, 0); 922 nsize = MCLBYTES; 923 } else { 924 n = m_get(how, m->m_type); 925 nsize = MLEN; 926 } 927 if (n == NULL) 928 goto nospace; 929 930 if (top == NULL) { /* First one, must be PKTHDR */ 931 if (!m_dup_pkthdr(n, m, how)) { 932 m_free(n); 933 goto nospace; 934 } 935 if ((n->m_flags & M_EXT) == 0) 936 nsize = MHLEN; 937 } 938 n->m_len = 0; 939 940 /* Link it into the new chain */ 941 *p = n; 942 p = &n->m_next; 943 944 /* Copy data from original mbuf(s) into new mbuf */ 945 while (n->m_len < nsize && m != NULL) { 946 int chunk = min(nsize - n->m_len, m->m_len - moff); 947 948 bcopy(m->m_data + moff, n->m_data + n->m_len, chunk); 949 moff += chunk; 950 n->m_len += chunk; 951 remain -= chunk; 952 if (moff == m->m_len) { 953 m = m->m_next; 954 moff = 0; 955 } 956 } 957 958 /* Check correct total mbuf length */ 959 KASSERT((remain > 0 && m != NULL) || (remain == 0 && m == NULL), 960 ("%s: bogus m_pkthdr.len", __func__)); 961 } 962 return (top); 963 964 nospace: 965 m_freem(top); 966 return (NULL); 967 } 968 969 /* 970 * Concatenate mbuf chain n to m. 971 * Both chains must be of the same type (e.g. MT_DATA). 972 * Any m_pkthdr is not updated. 973 */ 974 void 975 m_cat(struct mbuf *m, struct mbuf *n) 976 { 977 while (m->m_next) 978 m = m->m_next; 979 while (n) { 980 if (!M_WRITABLE(m) || 981 M_TRAILINGSPACE(m) < n->m_len) { 982 /* just join the two chains */ 983 m->m_next = n; 984 return; 985 } 986 /* splat the data from one into the other */ 987 bcopy(mtod(n, caddr_t), mtod(m, caddr_t) + m->m_len, 988 (u_int)n->m_len); 989 m->m_len += n->m_len; 990 n = m_free(n); 991 } 992 } 993 994 void 995 m_adj(struct mbuf *mp, int req_len) 996 { 997 int len = req_len; 998 struct mbuf *m; 999 int count; 1000 1001 if ((m = mp) == NULL) 1002 return; 1003 if (len >= 0) { 1004 /* 1005 * Trim from head. 1006 */ 1007 while (m != NULL && len > 0) { 1008 if (m->m_len <= len) { 1009 len -= m->m_len; 1010 m->m_len = 0; 1011 m = m->m_next; 1012 } else { 1013 m->m_len -= len; 1014 m->m_data += len; 1015 len = 0; 1016 } 1017 } 1018 if (mp->m_flags & M_PKTHDR) 1019 mp->m_pkthdr.len -= (req_len - len); 1020 } else { 1021 /* 1022 * Trim from tail. Scan the mbuf chain, 1023 * calculating its length and finding the last mbuf. 1024 * If the adjustment only affects this mbuf, then just 1025 * adjust and return. Otherwise, rescan and truncate 1026 * after the remaining size. 1027 */ 1028 len = -len; 1029 count = 0; 1030 for (;;) { 1031 count += m->m_len; 1032 if (m->m_next == (struct mbuf *)0) 1033 break; 1034 m = m->m_next; 1035 } 1036 if (m->m_len >= len) { 1037 m->m_len -= len; 1038 if (mp->m_flags & M_PKTHDR) 1039 mp->m_pkthdr.len -= len; 1040 return; 1041 } 1042 count -= len; 1043 if (count < 0) 1044 count = 0; 1045 /* 1046 * Correct length for chain is "count". 1047 * Find the mbuf with last data, adjust its length, 1048 * and toss data from remaining mbufs on chain. 1049 */ 1050 m = mp; 1051 if (m->m_flags & M_PKTHDR) 1052 m->m_pkthdr.len = count; 1053 for (; m; m = m->m_next) { 1054 if (m->m_len >= count) { 1055 m->m_len = count; 1056 if (m->m_next != NULL) { 1057 m_freem(m->m_next); 1058 m->m_next = NULL; 1059 } 1060 break; 1061 } 1062 count -= m->m_len; 1063 } 1064 } 1065 } 1066 1067 /* 1068 * Rearange an mbuf chain so that len bytes are contiguous 1069 * and in the data area of an mbuf (so that mtod will work 1070 * for a structure of size len). Returns the resulting 1071 * mbuf chain on success, frees it and returns null on failure. 1072 * If there is room, it will add up to max_protohdr-len extra bytes to the 1073 * contiguous region in an attempt to avoid being called next time. 1074 */ 1075 struct mbuf * 1076 m_pullup(struct mbuf *n, int len) 1077 { 1078 struct mbuf *m; 1079 int count; 1080 int space; 1081 1082 /* 1083 * If first mbuf has no cluster, and has room for len bytes 1084 * without shifting current data, pullup into it, 1085 * otherwise allocate a new mbuf to prepend to the chain. 1086 */ 1087 if ((n->m_flags & M_EXT) == 0 && 1088 n->m_data + len < &n->m_dat[MLEN] && n->m_next) { 1089 if (n->m_len >= len) 1090 return (n); 1091 m = n; 1092 n = n->m_next; 1093 len -= m->m_len; 1094 } else { 1095 if (len > MHLEN) 1096 goto bad; 1097 m = m_get(M_NOWAIT, n->m_type); 1098 if (m == NULL) 1099 goto bad; 1100 if (n->m_flags & M_PKTHDR) 1101 m_move_pkthdr(m, n); 1102 } 1103 space = &m->m_dat[MLEN] - (m->m_data + m->m_len); 1104 do { 1105 count = min(min(max(len, max_protohdr), space), n->m_len); 1106 bcopy(mtod(n, caddr_t), mtod(m, caddr_t) + m->m_len, 1107 (u_int)count); 1108 len -= count; 1109 m->m_len += count; 1110 n->m_len -= count; 1111 space -= count; 1112 if (n->m_len) 1113 n->m_data += count; 1114 else 1115 n = m_free(n); 1116 } while (len > 0 && n); 1117 if (len > 0) { 1118 (void) m_free(m); 1119 goto bad; 1120 } 1121 m->m_next = n; 1122 return (m); 1123 bad: 1124 m_freem(n); 1125 return (NULL); 1126 } 1127 1128 /* 1129 * Like m_pullup(), except a new mbuf is always allocated, and we allow 1130 * the amount of empty space before the data in the new mbuf to be specified 1131 * (in the event that the caller expects to prepend later). 1132 */ 1133 int MSFail; 1134 1135 struct mbuf * 1136 m_copyup(struct mbuf *n, int len, int dstoff) 1137 { 1138 struct mbuf *m; 1139 int count, space; 1140 1141 if (len > (MHLEN - dstoff)) 1142 goto bad; 1143 m = m_get(M_NOWAIT, n->m_type); 1144 if (m == NULL) 1145 goto bad; 1146 if (n->m_flags & M_PKTHDR) 1147 m_move_pkthdr(m, n); 1148 m->m_data += dstoff; 1149 space = &m->m_dat[MLEN] - (m->m_data + m->m_len); 1150 do { 1151 count = min(min(max(len, max_protohdr), space), n->m_len); 1152 memcpy(mtod(m, caddr_t) + m->m_len, mtod(n, caddr_t), 1153 (unsigned)count); 1154 len -= count; 1155 m->m_len += count; 1156 n->m_len -= count; 1157 space -= count; 1158 if (n->m_len) 1159 n->m_data += count; 1160 else 1161 n = m_free(n); 1162 } while (len > 0 && n); 1163 if (len > 0) { 1164 (void) m_free(m); 1165 goto bad; 1166 } 1167 m->m_next = n; 1168 return (m); 1169 bad: 1170 m_freem(n); 1171 MSFail++; 1172 return (NULL); 1173 } 1174 1175 /* 1176 * Partition an mbuf chain in two pieces, returning the tail -- 1177 * all but the first len0 bytes. In case of failure, it returns NULL and 1178 * attempts to restore the chain to its original state. 1179 * 1180 * Note that the resulting mbufs might be read-only, because the new 1181 * mbuf can end up sharing an mbuf cluster with the original mbuf if 1182 * the "breaking point" happens to lie within a cluster mbuf. Use the 1183 * M_WRITABLE() macro to check for this case. 1184 */ 1185 struct mbuf * 1186 m_split(struct mbuf *m0, int len0, int wait) 1187 { 1188 struct mbuf *m, *n; 1189 u_int len = len0, remain; 1190 1191 MBUF_CHECKSLEEP(wait); 1192 for (m = m0; m && len > m->m_len; m = m->m_next) 1193 len -= m->m_len; 1194 if (m == NULL) 1195 return (NULL); 1196 remain = m->m_len - len; 1197 if (m0->m_flags & M_PKTHDR && remain == 0) { 1198 n = m_gethdr(wait, m0->m_type); 1199 if (n == NULL) 1200 return (NULL); 1201 n->m_next = m->m_next; 1202 m->m_next = NULL; 1203 n->m_pkthdr.rcvif = m0->m_pkthdr.rcvif; 1204 n->m_pkthdr.len = m0->m_pkthdr.len - len0; 1205 m0->m_pkthdr.len = len0; 1206 return (n); 1207 } else if (m0->m_flags & M_PKTHDR) { 1208 n = m_gethdr(wait, m0->m_type); 1209 if (n == NULL) 1210 return (NULL); 1211 n->m_pkthdr.rcvif = m0->m_pkthdr.rcvif; 1212 n->m_pkthdr.len = m0->m_pkthdr.len - len0; 1213 m0->m_pkthdr.len = len0; 1214 if (m->m_flags & M_EXT) 1215 goto extpacket; 1216 if (remain > MHLEN) { 1217 /* m can't be the lead packet */ 1218 MH_ALIGN(n, 0); 1219 n->m_next = m_split(m, len, wait); 1220 if (n->m_next == NULL) { 1221 (void) m_free(n); 1222 return (NULL); 1223 } else { 1224 n->m_len = 0; 1225 return (n); 1226 } 1227 } else 1228 MH_ALIGN(n, remain); 1229 } else if (remain == 0) { 1230 n = m->m_next; 1231 m->m_next = NULL; 1232 return (n); 1233 } else { 1234 n = m_get(wait, m->m_type); 1235 if (n == NULL) 1236 return (NULL); 1237 M_ALIGN(n, remain); 1238 } 1239 extpacket: 1240 if (m->m_flags & M_EXT) { 1241 n->m_data = m->m_data + len; 1242 mb_dupcl(n, m); 1243 } else { 1244 bcopy(mtod(m, caddr_t) + len, mtod(n, caddr_t), remain); 1245 } 1246 n->m_len = remain; 1247 m->m_len = len; 1248 n->m_next = m->m_next; 1249 m->m_next = NULL; 1250 return (n); 1251 } 1252 /* 1253 * Routine to copy from device local memory into mbufs. 1254 * Note that `off' argument is offset into first mbuf of target chain from 1255 * which to begin copying the data to. 1256 */ 1257 struct mbuf * 1258 m_devget(char *buf, int totlen, int off, struct ifnet *ifp, 1259 void (*copy)(char *from, caddr_t to, u_int len)) 1260 { 1261 struct mbuf *m; 1262 struct mbuf *top = NULL, **mp = ⊤ 1263 int len; 1264 1265 if (off < 0 || off > MHLEN) 1266 return (NULL); 1267 1268 while (totlen > 0) { 1269 if (top == NULL) { /* First one, must be PKTHDR */ 1270 if (totlen + off >= MINCLSIZE) { 1271 m = m_getcl(M_NOWAIT, MT_DATA, M_PKTHDR); 1272 len = MCLBYTES; 1273 } else { 1274 m = m_gethdr(M_NOWAIT, MT_DATA); 1275 len = MHLEN; 1276 1277 /* Place initial small packet/header at end of mbuf */ 1278 if (m && totlen + off + max_linkhdr <= MLEN) { 1279 m->m_data += max_linkhdr; 1280 len -= max_linkhdr; 1281 } 1282 } 1283 if (m == NULL) 1284 return NULL; 1285 m->m_pkthdr.rcvif = ifp; 1286 m->m_pkthdr.len = totlen; 1287 } else { 1288 if (totlen + off >= MINCLSIZE) { 1289 m = m_getcl(M_NOWAIT, MT_DATA, 0); 1290 len = MCLBYTES; 1291 } else { 1292 m = m_get(M_NOWAIT, MT_DATA); 1293 len = MLEN; 1294 } 1295 if (m == NULL) { 1296 m_freem(top); 1297 return NULL; 1298 } 1299 } 1300 if (off) { 1301 m->m_data += off; 1302 len -= off; 1303 off = 0; 1304 } 1305 m->m_len = len = min(totlen, len); 1306 if (copy) 1307 copy(buf, mtod(m, caddr_t), (u_int)len); 1308 else 1309 bcopy(buf, mtod(m, caddr_t), (u_int)len); 1310 buf += len; 1311 *mp = m; 1312 mp = &m->m_next; 1313 totlen -= len; 1314 } 1315 return (top); 1316 } 1317 1318 /* 1319 * Copy data from a buffer back into the indicated mbuf chain, 1320 * starting "off" bytes from the beginning, extending the mbuf 1321 * chain if necessary. 1322 */ 1323 void 1324 m_copyback(struct mbuf *m0, int off, int len, c_caddr_t cp) 1325 { 1326 int mlen; 1327 struct mbuf *m = m0, *n; 1328 int totlen = 0; 1329 1330 if (m0 == NULL) 1331 return; 1332 while (off > (mlen = m->m_len)) { 1333 off -= mlen; 1334 totlen += mlen; 1335 if (m->m_next == NULL) { 1336 n = m_get(M_NOWAIT, m->m_type); 1337 if (n == NULL) 1338 goto out; 1339 bzero(mtod(n, caddr_t), MLEN); 1340 n->m_len = min(MLEN, len + off); 1341 m->m_next = n; 1342 } 1343 m = m->m_next; 1344 } 1345 while (len > 0) { 1346 if (m->m_next == NULL && (len > m->m_len - off)) { 1347 m->m_len += min(len - (m->m_len - off), 1348 M_TRAILINGSPACE(m)); 1349 } 1350 mlen = min (m->m_len - off, len); 1351 bcopy(cp, off + mtod(m, caddr_t), (u_int)mlen); 1352 cp += mlen; 1353 len -= mlen; 1354 mlen += off; 1355 off = 0; 1356 totlen += mlen; 1357 if (len == 0) 1358 break; 1359 if (m->m_next == NULL) { 1360 n = m_get(M_NOWAIT, m->m_type); 1361 if (n == NULL) 1362 break; 1363 n->m_len = min(MLEN, len); 1364 m->m_next = n; 1365 } 1366 m = m->m_next; 1367 } 1368 out: if (((m = m0)->m_flags & M_PKTHDR) && (m->m_pkthdr.len < totlen)) 1369 m->m_pkthdr.len = totlen; 1370 } 1371 1372 /* 1373 * Append the specified data to the indicated mbuf chain, 1374 * Extend the mbuf chain if the new data does not fit in 1375 * existing space. 1376 * 1377 * Return 1 if able to complete the job; otherwise 0. 1378 */ 1379 int 1380 m_append(struct mbuf *m0, int len, c_caddr_t cp) 1381 { 1382 struct mbuf *m, *n; 1383 int remainder, space; 1384 1385 for (m = m0; m->m_next != NULL; m = m->m_next) 1386 ; 1387 remainder = len; 1388 space = M_TRAILINGSPACE(m); 1389 if (space > 0) { 1390 /* 1391 * Copy into available space. 1392 */ 1393 if (space > remainder) 1394 space = remainder; 1395 bcopy(cp, mtod(m, caddr_t) + m->m_len, space); 1396 m->m_len += space; 1397 cp += space, remainder -= space; 1398 } 1399 while (remainder > 0) { 1400 /* 1401 * Allocate a new mbuf; could check space 1402 * and allocate a cluster instead. 1403 */ 1404 n = m_get(M_NOWAIT, m->m_type); 1405 if (n == NULL) 1406 break; 1407 n->m_len = min(MLEN, remainder); 1408 bcopy(cp, mtod(n, caddr_t), n->m_len); 1409 cp += n->m_len, remainder -= n->m_len; 1410 m->m_next = n; 1411 m = n; 1412 } 1413 if (m0->m_flags & M_PKTHDR) 1414 m0->m_pkthdr.len += len - remainder; 1415 return (remainder == 0); 1416 } 1417 1418 /* 1419 * Apply function f to the data in an mbuf chain starting "off" bytes from 1420 * the beginning, continuing for "len" bytes. 1421 */ 1422 int 1423 m_apply(struct mbuf *m, int off, int len, 1424 int (*f)(void *, void *, u_int), void *arg) 1425 { 1426 u_int count; 1427 int rval; 1428 1429 KASSERT(off >= 0, ("m_apply, negative off %d", off)); 1430 KASSERT(len >= 0, ("m_apply, negative len %d", len)); 1431 while (off > 0) { 1432 KASSERT(m != NULL, ("m_apply, offset > size of mbuf chain")); 1433 if (off < m->m_len) 1434 break; 1435 off -= m->m_len; 1436 m = m->m_next; 1437 } 1438 while (len > 0) { 1439 KASSERT(m != NULL, ("m_apply, offset > size of mbuf chain")); 1440 count = min(m->m_len - off, len); 1441 rval = (*f)(arg, mtod(m, caddr_t) + off, count); 1442 if (rval) 1443 return (rval); 1444 len -= count; 1445 off = 0; 1446 m = m->m_next; 1447 } 1448 return (0); 1449 } 1450 1451 /* 1452 * Return a pointer to mbuf/offset of location in mbuf chain. 1453 */ 1454 struct mbuf * 1455 m_getptr(struct mbuf *m, int loc, int *off) 1456 { 1457 1458 while (loc >= 0) { 1459 /* Normal end of search. */ 1460 if (m->m_len > loc) { 1461 *off = loc; 1462 return (m); 1463 } else { 1464 loc -= m->m_len; 1465 if (m->m_next == NULL) { 1466 if (loc == 0) { 1467 /* Point at the end of valid data. */ 1468 *off = m->m_len; 1469 return (m); 1470 } 1471 return (NULL); 1472 } 1473 m = m->m_next; 1474 } 1475 } 1476 return (NULL); 1477 } 1478 1479 void 1480 m_print(const struct mbuf *m, int maxlen) 1481 { 1482 int len; 1483 int pdata; 1484 const struct mbuf *m2; 1485 1486 if (m == NULL) { 1487 printf("mbuf: %p\n", m); 1488 return; 1489 } 1490 1491 if (m->m_flags & M_PKTHDR) 1492 len = m->m_pkthdr.len; 1493 else 1494 len = -1; 1495 m2 = m; 1496 while (m2 != NULL && (len == -1 || len)) { 1497 pdata = m2->m_len; 1498 if (maxlen != -1 && pdata > maxlen) 1499 pdata = maxlen; 1500 printf("mbuf: %p len: %d, next: %p, %b%s", m2, m2->m_len, 1501 m2->m_next, m2->m_flags, "\20\20freelist\17skipfw" 1502 "\11proto5\10proto4\7proto3\6proto2\5proto1\4rdonly" 1503 "\3eor\2pkthdr\1ext", pdata ? "" : "\n"); 1504 if (pdata) 1505 printf(", %*D\n", pdata, (u_char *)m2->m_data, "-"); 1506 if (len != -1) 1507 len -= m2->m_len; 1508 m2 = m2->m_next; 1509 } 1510 if (len > 0) 1511 printf("%d bytes unaccounted for.\n", len); 1512 return; 1513 } 1514 1515 u_int 1516 m_fixhdr(struct mbuf *m0) 1517 { 1518 u_int len; 1519 1520 len = m_length(m0, NULL); 1521 m0->m_pkthdr.len = len; 1522 return (len); 1523 } 1524 1525 u_int 1526 m_length(struct mbuf *m0, struct mbuf **last) 1527 { 1528 struct mbuf *m; 1529 u_int len; 1530 1531 len = 0; 1532 for (m = m0; m != NULL; m = m->m_next) { 1533 len += m->m_len; 1534 if (m->m_next == NULL) 1535 break; 1536 } 1537 if (last != NULL) 1538 *last = m; 1539 return (len); 1540 } 1541 1542 /* 1543 * Defragment a mbuf chain, returning the shortest possible 1544 * chain of mbufs and clusters. If allocation fails and 1545 * this cannot be completed, NULL will be returned, but 1546 * the passed in chain will be unchanged. Upon success, 1547 * the original chain will be freed, and the new chain 1548 * will be returned. 1549 * 1550 * If a non-packet header is passed in, the original 1551 * mbuf (chain?) will be returned unharmed. 1552 */ 1553 struct mbuf * 1554 m_defrag(struct mbuf *m0, int how) 1555 { 1556 struct mbuf *m_new = NULL, *m_final = NULL; 1557 int progress = 0, length; 1558 1559 MBUF_CHECKSLEEP(how); 1560 if (!(m0->m_flags & M_PKTHDR)) 1561 return (m0); 1562 1563 m_fixhdr(m0); /* Needed sanity check */ 1564 1565 #ifdef MBUF_STRESS_TEST 1566 if (m_defragrandomfailures) { 1567 int temp = arc4random() & 0xff; 1568 if (temp == 0xba) 1569 goto nospace; 1570 } 1571 #endif 1572 1573 if (m0->m_pkthdr.len > MHLEN) 1574 m_final = m_getcl(how, MT_DATA, M_PKTHDR); 1575 else 1576 m_final = m_gethdr(how, MT_DATA); 1577 1578 if (m_final == NULL) 1579 goto nospace; 1580 1581 if (m_dup_pkthdr(m_final, m0, how) == 0) 1582 goto nospace; 1583 1584 m_new = m_final; 1585 1586 while (progress < m0->m_pkthdr.len) { 1587 length = m0->m_pkthdr.len - progress; 1588 if (length > MCLBYTES) 1589 length = MCLBYTES; 1590 1591 if (m_new == NULL) { 1592 if (length > MLEN) 1593 m_new = m_getcl(how, MT_DATA, 0); 1594 else 1595 m_new = m_get(how, MT_DATA); 1596 if (m_new == NULL) 1597 goto nospace; 1598 } 1599 1600 m_copydata(m0, progress, length, mtod(m_new, caddr_t)); 1601 progress += length; 1602 m_new->m_len = length; 1603 if (m_new != m_final) 1604 m_cat(m_final, m_new); 1605 m_new = NULL; 1606 } 1607 #ifdef MBUF_STRESS_TEST 1608 if (m0->m_next == NULL) 1609 m_defraguseless++; 1610 #endif 1611 m_freem(m0); 1612 m0 = m_final; 1613 #ifdef MBUF_STRESS_TEST 1614 m_defragpackets++; 1615 m_defragbytes += m0->m_pkthdr.len; 1616 #endif 1617 return (m0); 1618 nospace: 1619 #ifdef MBUF_STRESS_TEST 1620 m_defragfailure++; 1621 #endif 1622 if (m_final) 1623 m_freem(m_final); 1624 return (NULL); 1625 } 1626 1627 /* 1628 * Defragment an mbuf chain, returning at most maxfrags separate 1629 * mbufs+clusters. If this is not possible NULL is returned and 1630 * the original mbuf chain is left in it's present (potentially 1631 * modified) state. We use two techniques: collapsing consecutive 1632 * mbufs and replacing consecutive mbufs by a cluster. 1633 * 1634 * NB: this should really be named m_defrag but that name is taken 1635 */ 1636 struct mbuf * 1637 m_collapse(struct mbuf *m0, int how, int maxfrags) 1638 { 1639 struct mbuf *m, *n, *n2, **prev; 1640 u_int curfrags; 1641 1642 /* 1643 * Calculate the current number of frags. 1644 */ 1645 curfrags = 0; 1646 for (m = m0; m != NULL; m = m->m_next) 1647 curfrags++; 1648 /* 1649 * First, try to collapse mbufs. Note that we always collapse 1650 * towards the front so we don't need to deal with moving the 1651 * pkthdr. This may be suboptimal if the first mbuf has much 1652 * less data than the following. 1653 */ 1654 m = m0; 1655 again: 1656 for (;;) { 1657 n = m->m_next; 1658 if (n == NULL) 1659 break; 1660 if (M_WRITABLE(m) && 1661 n->m_len < M_TRAILINGSPACE(m)) { 1662 bcopy(mtod(n, void *), mtod(m, char *) + m->m_len, 1663 n->m_len); 1664 m->m_len += n->m_len; 1665 m->m_next = n->m_next; 1666 m_free(n); 1667 if (--curfrags <= maxfrags) 1668 return m0; 1669 } else 1670 m = n; 1671 } 1672 KASSERT(maxfrags > 1, 1673 ("maxfrags %u, but normal collapse failed", maxfrags)); 1674 /* 1675 * Collapse consecutive mbufs to a cluster. 1676 */ 1677 prev = &m0->m_next; /* NB: not the first mbuf */ 1678 while ((n = *prev) != NULL) { 1679 if ((n2 = n->m_next) != NULL && 1680 n->m_len + n2->m_len < MCLBYTES) { 1681 m = m_getcl(how, MT_DATA, 0); 1682 if (m == NULL) 1683 goto bad; 1684 bcopy(mtod(n, void *), mtod(m, void *), n->m_len); 1685 bcopy(mtod(n2, void *), mtod(m, char *) + n->m_len, 1686 n2->m_len); 1687 m->m_len = n->m_len + n2->m_len; 1688 m->m_next = n2->m_next; 1689 *prev = m; 1690 m_free(n); 1691 m_free(n2); 1692 if (--curfrags <= maxfrags) /* +1 cl -2 mbufs */ 1693 return m0; 1694 /* 1695 * Still not there, try the normal collapse 1696 * again before we allocate another cluster. 1697 */ 1698 goto again; 1699 } 1700 prev = &n->m_next; 1701 } 1702 /* 1703 * No place where we can collapse to a cluster; punt. 1704 * This can occur if, for example, you request 2 frags 1705 * but the packet requires that both be clusters (we 1706 * never reallocate the first mbuf to avoid moving the 1707 * packet header). 1708 */ 1709 bad: 1710 return NULL; 1711 } 1712 1713 #ifdef MBUF_STRESS_TEST 1714 1715 /* 1716 * Fragment an mbuf chain. There's no reason you'd ever want to do 1717 * this in normal usage, but it's great for stress testing various 1718 * mbuf consumers. 1719 * 1720 * If fragmentation is not possible, the original chain will be 1721 * returned. 1722 * 1723 * Possible length values: 1724 * 0 no fragmentation will occur 1725 * > 0 each fragment will be of the specified length 1726 * -1 each fragment will be the same random value in length 1727 * -2 each fragment's length will be entirely random 1728 * (Random values range from 1 to 256) 1729 */ 1730 struct mbuf * 1731 m_fragment(struct mbuf *m0, int how, int length) 1732 { 1733 struct mbuf *m_new = NULL, *m_final = NULL; 1734 int progress = 0; 1735 1736 if (!(m0->m_flags & M_PKTHDR)) 1737 return (m0); 1738 1739 if ((length == 0) || (length < -2)) 1740 return (m0); 1741 1742 m_fixhdr(m0); /* Needed sanity check */ 1743 1744 m_final = m_getcl(how, MT_DATA, M_PKTHDR); 1745 1746 if (m_final == NULL) 1747 goto nospace; 1748 1749 if (m_dup_pkthdr(m_final, m0, how) == 0) 1750 goto nospace; 1751 1752 m_new = m_final; 1753 1754 if (length == -1) 1755 length = 1 + (arc4random() & 255); 1756 1757 while (progress < m0->m_pkthdr.len) { 1758 int fraglen; 1759 1760 if (length > 0) 1761 fraglen = length; 1762 else 1763 fraglen = 1 + (arc4random() & 255); 1764 if (fraglen > m0->m_pkthdr.len - progress) 1765 fraglen = m0->m_pkthdr.len - progress; 1766 1767 if (fraglen > MCLBYTES) 1768 fraglen = MCLBYTES; 1769 1770 if (m_new == NULL) { 1771 m_new = m_getcl(how, MT_DATA, 0); 1772 if (m_new == NULL) 1773 goto nospace; 1774 } 1775 1776 m_copydata(m0, progress, fraglen, mtod(m_new, caddr_t)); 1777 progress += fraglen; 1778 m_new->m_len = fraglen; 1779 if (m_new != m_final) 1780 m_cat(m_final, m_new); 1781 m_new = NULL; 1782 } 1783 m_freem(m0); 1784 m0 = m_final; 1785 return (m0); 1786 nospace: 1787 if (m_final) 1788 m_freem(m_final); 1789 /* Return the original chain on failure */ 1790 return (m0); 1791 } 1792 1793 #endif 1794 1795 /* 1796 * Copy the contents of uio into a properly sized mbuf chain. 1797 */ 1798 struct mbuf * 1799 m_uiotombuf(struct uio *uio, int how, int len, int align, int flags) 1800 { 1801 struct mbuf *m, *mb; 1802 int error, length; 1803 ssize_t total; 1804 int progress = 0; 1805 1806 /* 1807 * len can be zero or an arbitrary large value bound by 1808 * the total data supplied by the uio. 1809 */ 1810 if (len > 0) 1811 total = min(uio->uio_resid, len); 1812 else 1813 total = uio->uio_resid; 1814 1815 /* 1816 * The smallest unit returned by m_getm2() is a single mbuf 1817 * with pkthdr. We can't align past it. 1818 */ 1819 if (align >= MHLEN) 1820 return (NULL); 1821 1822 /* 1823 * Give us the full allocation or nothing. 1824 * If len is zero return the smallest empty mbuf. 1825 */ 1826 m = m_getm2(NULL, max(total + align, 1), how, MT_DATA, flags); 1827 if (m == NULL) 1828 return (NULL); 1829 m->m_data += align; 1830 1831 /* Fill all mbufs with uio data and update header information. */ 1832 for (mb = m; mb != NULL; mb = mb->m_next) { 1833 length = min(M_TRAILINGSPACE(mb), total - progress); 1834 1835 error = uiomove(mtod(mb, void *), length, uio); 1836 if (error) { 1837 m_freem(m); 1838 return (NULL); 1839 } 1840 1841 mb->m_len = length; 1842 progress += length; 1843 if (flags & M_PKTHDR) 1844 m->m_pkthdr.len += length; 1845 } 1846 KASSERT(progress == total, ("%s: progress != total", __func__)); 1847 1848 return (m); 1849 } 1850 1851 /* 1852 * Copy an mbuf chain into a uio limited by len if set. 1853 */ 1854 int 1855 m_mbuftouio(struct uio *uio, struct mbuf *m, int len) 1856 { 1857 int error, length, total; 1858 int progress = 0; 1859 1860 if (len > 0) 1861 total = min(uio->uio_resid, len); 1862 else 1863 total = uio->uio_resid; 1864 1865 /* Fill the uio with data from the mbufs. */ 1866 for (; m != NULL; m = m->m_next) { 1867 length = min(m->m_len, total - progress); 1868 1869 error = uiomove(mtod(m, void *), length, uio); 1870 if (error) 1871 return (error); 1872 1873 progress += length; 1874 } 1875 1876 return (0); 1877 } 1878 1879 /* 1880 * Set the m_data pointer of a newly-allocated mbuf 1881 * to place an object of the specified size at the 1882 * end of the mbuf, longword aligned. 1883 */ 1884 void 1885 m_align(struct mbuf *m, int len) 1886 { 1887 #ifdef INVARIANTS 1888 const char *msg = "%s: not a virgin mbuf"; 1889 #endif 1890 int adjust; 1891 1892 if (m->m_flags & M_EXT) { 1893 KASSERT(m->m_data == m->m_ext.ext_buf, (msg, __func__)); 1894 adjust = m->m_ext.ext_size - len; 1895 } else if (m->m_flags & M_PKTHDR) { 1896 KASSERT(m->m_data == m->m_pktdat, (msg, __func__)); 1897 adjust = MHLEN - len; 1898 } else { 1899 KASSERT(m->m_data == m->m_dat, (msg, __func__)); 1900 adjust = MLEN - len; 1901 } 1902 1903 m->m_data += adjust &~ (sizeof(long)-1); 1904 } 1905 1906 /* 1907 * Create a writable copy of the mbuf chain. While doing this 1908 * we compact the chain with a goal of producing a chain with 1909 * at most two mbufs. The second mbuf in this chain is likely 1910 * to be a cluster. The primary purpose of this work is to create 1911 * a writable packet for encryption, compression, etc. The 1912 * secondary goal is to linearize the data so the data can be 1913 * passed to crypto hardware in the most efficient manner possible. 1914 */ 1915 struct mbuf * 1916 m_unshare(struct mbuf *m0, int how) 1917 { 1918 struct mbuf *m, *mprev; 1919 struct mbuf *n, *mfirst, *mlast; 1920 int len, off; 1921 1922 mprev = NULL; 1923 for (m = m0; m != NULL; m = mprev->m_next) { 1924 /* 1925 * Regular mbufs are ignored unless there's a cluster 1926 * in front of it that we can use to coalesce. We do 1927 * the latter mainly so later clusters can be coalesced 1928 * also w/o having to handle them specially (i.e. convert 1929 * mbuf+cluster -> cluster). This optimization is heavily 1930 * influenced by the assumption that we're running over 1931 * Ethernet where MCLBYTES is large enough that the max 1932 * packet size will permit lots of coalescing into a 1933 * single cluster. This in turn permits efficient 1934 * crypto operations, especially when using hardware. 1935 */ 1936 if ((m->m_flags & M_EXT) == 0) { 1937 if (mprev && (mprev->m_flags & M_EXT) && 1938 m->m_len <= M_TRAILINGSPACE(mprev)) { 1939 /* XXX: this ignores mbuf types */ 1940 memcpy(mtod(mprev, caddr_t) + mprev->m_len, 1941 mtod(m, caddr_t), m->m_len); 1942 mprev->m_len += m->m_len; 1943 mprev->m_next = m->m_next; /* unlink from chain */ 1944 m_free(m); /* reclaim mbuf */ 1945 #if 0 1946 newipsecstat.ips_mbcoalesced++; 1947 #endif 1948 } else { 1949 mprev = m; 1950 } 1951 continue; 1952 } 1953 /* 1954 * Writable mbufs are left alone (for now). 1955 */ 1956 if (M_WRITABLE(m)) { 1957 mprev = m; 1958 continue; 1959 } 1960 1961 /* 1962 * Not writable, replace with a copy or coalesce with 1963 * the previous mbuf if possible (since we have to copy 1964 * it anyway, we try to reduce the number of mbufs and 1965 * clusters so that future work is easier). 1966 */ 1967 KASSERT(m->m_flags & M_EXT, ("m_flags 0x%x", m->m_flags)); 1968 /* NB: we only coalesce into a cluster or larger */ 1969 if (mprev != NULL && (mprev->m_flags & M_EXT) && 1970 m->m_len <= M_TRAILINGSPACE(mprev)) { 1971 /* XXX: this ignores mbuf types */ 1972 memcpy(mtod(mprev, caddr_t) + mprev->m_len, 1973 mtod(m, caddr_t), m->m_len); 1974 mprev->m_len += m->m_len; 1975 mprev->m_next = m->m_next; /* unlink from chain */ 1976 m_free(m); /* reclaim mbuf */ 1977 #if 0 1978 newipsecstat.ips_clcoalesced++; 1979 #endif 1980 continue; 1981 } 1982 1983 /* 1984 * Allocate new space to hold the copy and copy the data. 1985 * We deal with jumbo mbufs (i.e. m_len > MCLBYTES) by 1986 * splitting them into clusters. We could just malloc a 1987 * buffer and make it external but too many device drivers 1988 * don't know how to break up the non-contiguous memory when 1989 * doing DMA. 1990 */ 1991 n = m_getcl(how, m->m_type, m->m_flags); 1992 if (n == NULL) { 1993 m_freem(m0); 1994 return (NULL); 1995 } 1996 len = m->m_len; 1997 off = 0; 1998 mfirst = n; 1999 mlast = NULL; 2000 for (;;) { 2001 int cc = min(len, MCLBYTES); 2002 memcpy(mtod(n, caddr_t), mtod(m, caddr_t) + off, cc); 2003 n->m_len = cc; 2004 if (mlast != NULL) 2005 mlast->m_next = n; 2006 mlast = n; 2007 #if 0 2008 newipsecstat.ips_clcopied++; 2009 #endif 2010 2011 len -= cc; 2012 if (len <= 0) 2013 break; 2014 off += cc; 2015 2016 n = m_getcl(how, m->m_type, m->m_flags); 2017 if (n == NULL) { 2018 m_freem(mfirst); 2019 m_freem(m0); 2020 return (NULL); 2021 } 2022 } 2023 n->m_next = m->m_next; 2024 if (mprev == NULL) 2025 m0 = mfirst; /* new head of chain */ 2026 else 2027 mprev->m_next = mfirst; /* replace old mbuf */ 2028 m_free(m); /* release old mbuf */ 2029 mprev = mfirst; 2030 } 2031 return (m0); 2032 } 2033 2034 #ifdef MBUF_PROFILING 2035 2036 #define MP_BUCKETS 32 /* don't just change this as things may overflow.*/ 2037 struct mbufprofile { 2038 uintmax_t wasted[MP_BUCKETS]; 2039 uintmax_t used[MP_BUCKETS]; 2040 uintmax_t segments[MP_BUCKETS]; 2041 } mbprof; 2042 2043 #define MP_MAXDIGITS 21 /* strlen("16,000,000,000,000,000,000") == 21 */ 2044 #define MP_NUMLINES 6 2045 #define MP_NUMSPERLINE 16 2046 #define MP_EXTRABYTES 64 /* > strlen("used:\nwasted:\nsegments:\n") */ 2047 /* work out max space needed and add a bit of spare space too */ 2048 #define MP_MAXLINE ((MP_MAXDIGITS+1) * MP_NUMSPERLINE) 2049 #define MP_BUFSIZE ((MP_MAXLINE * MP_NUMLINES) + 1 + MP_EXTRABYTES) 2050 2051 char mbprofbuf[MP_BUFSIZE]; 2052 2053 void 2054 m_profile(struct mbuf *m) 2055 { 2056 int segments = 0; 2057 int used = 0; 2058 int wasted = 0; 2059 2060 while (m) { 2061 segments++; 2062 used += m->m_len; 2063 if (m->m_flags & M_EXT) { 2064 wasted += MHLEN - sizeof(m->m_ext) + 2065 m->m_ext.ext_size - m->m_len; 2066 } else { 2067 if (m->m_flags & M_PKTHDR) 2068 wasted += MHLEN - m->m_len; 2069 else 2070 wasted += MLEN - m->m_len; 2071 } 2072 m = m->m_next; 2073 } 2074 /* be paranoid.. it helps */ 2075 if (segments > MP_BUCKETS - 1) 2076 segments = MP_BUCKETS - 1; 2077 if (used > 100000) 2078 used = 100000; 2079 if (wasted > 100000) 2080 wasted = 100000; 2081 /* store in the appropriate bucket */ 2082 /* don't bother locking. if it's slightly off, so what? */ 2083 mbprof.segments[segments]++; 2084 mbprof.used[fls(used)]++; 2085 mbprof.wasted[fls(wasted)]++; 2086 } 2087 2088 static void 2089 mbprof_textify(void) 2090 { 2091 int offset; 2092 char *c; 2093 uint64_t *p; 2094 2095 2096 p = &mbprof.wasted[0]; 2097 c = mbprofbuf; 2098 offset = snprintf(c, MP_MAXLINE + 10, 2099 "wasted:\n" 2100 "%ju %ju %ju %ju %ju %ju %ju %ju " 2101 "%ju %ju %ju %ju %ju %ju %ju %ju\n", 2102 p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], 2103 p[8], p[9], p[10], p[11], p[12], p[13], p[14], p[15]); 2104 #ifdef BIG_ARRAY 2105 p = &mbprof.wasted[16]; 2106 c += offset; 2107 offset = snprintf(c, MP_MAXLINE, 2108 "%ju %ju %ju %ju %ju %ju %ju %ju " 2109 "%ju %ju %ju %ju %ju %ju %ju %ju\n", 2110 p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], 2111 p[8], p[9], p[10], p[11], p[12], p[13], p[14], p[15]); 2112 #endif 2113 p = &mbprof.used[0]; 2114 c += offset; 2115 offset = snprintf(c, MP_MAXLINE + 10, 2116 "used:\n" 2117 "%ju %ju %ju %ju %ju %ju %ju %ju " 2118 "%ju %ju %ju %ju %ju %ju %ju %ju\n", 2119 p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], 2120 p[8], p[9], p[10], p[11], p[12], p[13], p[14], p[15]); 2121 #ifdef BIG_ARRAY 2122 p = &mbprof.used[16]; 2123 c += offset; 2124 offset = snprintf(c, MP_MAXLINE, 2125 "%ju %ju %ju %ju %ju %ju %ju %ju " 2126 "%ju %ju %ju %ju %ju %ju %ju %ju\n", 2127 p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], 2128 p[8], p[9], p[10], p[11], p[12], p[13], p[14], p[15]); 2129 #endif 2130 p = &mbprof.segments[0]; 2131 c += offset; 2132 offset = snprintf(c, MP_MAXLINE + 10, 2133 "segments:\n" 2134 "%ju %ju %ju %ju %ju %ju %ju %ju " 2135 "%ju %ju %ju %ju %ju %ju %ju %ju\n", 2136 p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], 2137 p[8], p[9], p[10], p[11], p[12], p[13], p[14], p[15]); 2138 #ifdef BIG_ARRAY 2139 p = &mbprof.segments[16]; 2140 c += offset; 2141 offset = snprintf(c, MP_MAXLINE, 2142 "%ju %ju %ju %ju %ju %ju %ju %ju " 2143 "%ju %ju %ju %ju %ju %ju %ju %jju", 2144 p[0], p[1], p[2], p[3], p[4], p[5], p[6], p[7], 2145 p[8], p[9], p[10], p[11], p[12], p[13], p[14], p[15]); 2146 #endif 2147 } 2148 2149 static int 2150 mbprof_handler(SYSCTL_HANDLER_ARGS) 2151 { 2152 int error; 2153 2154 mbprof_textify(); 2155 error = SYSCTL_OUT(req, mbprofbuf, strlen(mbprofbuf) + 1); 2156 return (error); 2157 } 2158 2159 static int 2160 mbprof_clr_handler(SYSCTL_HANDLER_ARGS) 2161 { 2162 int clear, error; 2163 2164 clear = 0; 2165 error = sysctl_handle_int(oidp, &clear, 0, req); 2166 if (error || !req->newptr) 2167 return (error); 2168 2169 if (clear) { 2170 bzero(&mbprof, sizeof(mbprof)); 2171 } 2172 2173 return (error); 2174 } 2175 2176 2177 SYSCTL_PROC(_kern_ipc, OID_AUTO, mbufprofile, CTLTYPE_STRING|CTLFLAG_RD, 2178 NULL, 0, mbprof_handler, "A", "mbuf profiling statistics"); 2179 2180 SYSCTL_PROC(_kern_ipc, OID_AUTO, mbufprofileclr, CTLTYPE_INT|CTLFLAG_RW, 2181 NULL, 0, mbprof_clr_handler, "I", "clear mbuf profiling statistics"); 2182 #endif 2183 2184