1 /* $NetBSD: sysv_shm.c,v 1.23 1994/07/04 23:25:12 glass Exp $ */ 2 /*- 3 * Copyright (c) 1994 Adam Glass and Charles Hannum. 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 * 3. All advertising materials mentioning features or use of this software 14 * must display the following acknowledgement: 15 * This product includes software developed by Adam Glass and Charles 16 * Hannum. 17 * 4. The names of the authors may not be used to endorse or promote products 18 * derived from this software without specific prior written permission. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE AUTHORS ``AS IS'' AND ANY EXPRESS OR 21 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 22 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 23 * IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY DIRECT, INDIRECT, 24 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 25 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 26 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 27 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 28 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 29 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 30 */ 31 /*- 32 * Copyright (c) 2003-2005 McAfee, Inc. 33 * All rights reserved. 34 * 35 * This software was developed for the FreeBSD Project in part by McAfee 36 * Research, the Security Research Division of McAfee, Inc under DARPA/SPAWAR 37 * contract N66001-01-C-8035 ("CBOSS"), as part of the DARPA CHATS research 38 * program. 39 * 40 * Redistribution and use in source and binary forms, with or without 41 * modification, are permitted provided that the following conditions 42 * are met: 43 * 1. Redistributions of source code must retain the above copyright 44 * notice, this list of conditions and the following disclaimer. 45 * 2. Redistributions in binary form must reproduce the above copyright 46 * notice, this list of conditions and the following disclaimer in the 47 * documentation and/or other materials provided with the distribution. 48 * 49 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR 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 AUTHOR 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 62 #include <sys/cdefs.h> 63 __FBSDID("$FreeBSD$"); 64 65 #include "opt_compat.h" 66 #include "opt_sysvipc.h" 67 68 #include <sys/param.h> 69 #include <sys/systm.h> 70 #include <sys/kernel.h> 71 #include <sys/limits.h> 72 #include <sys/lock.h> 73 #include <sys/sysctl.h> 74 #include <sys/shm.h> 75 #include <sys/proc.h> 76 #include <sys/malloc.h> 77 #include <sys/mman.h> 78 #include <sys/module.h> 79 #include <sys/mutex.h> 80 #include <sys/resourcevar.h> 81 #include <sys/stat.h> 82 #include <sys/syscall.h> 83 #include <sys/syscallsubr.h> 84 #include <sys/sysent.h> 85 #include <sys/sysproto.h> 86 #include <sys/jail.h> 87 88 #include <security/mac/mac_framework.h> 89 90 #include <vm/vm.h> 91 #include <vm/vm_param.h> 92 #include <vm/pmap.h> 93 #include <vm/vm_object.h> 94 #include <vm/vm_map.h> 95 #include <vm/vm_page.h> 96 #include <vm/vm_pager.h> 97 98 static MALLOC_DEFINE(M_SHM, "shm", "SVID compatible shared memory segments"); 99 100 static int shmget_allocate_segment(struct thread *td, 101 struct shmget_args *uap, int mode); 102 static int shmget_existing(struct thread *td, struct shmget_args *uap, 103 int mode, int segnum); 104 105 #define SHMSEG_FREE 0x0200 106 #define SHMSEG_REMOVED 0x0400 107 #define SHMSEG_ALLOCATED 0x0800 108 #define SHMSEG_WANTED 0x1000 109 110 static int shm_last_free, shm_nused, shmalloced; 111 vm_size_t shm_committed; 112 static struct shmid_kernel *shmsegs; 113 114 struct shmmap_state { 115 vm_offset_t va; 116 int shmid; 117 }; 118 119 static void shm_deallocate_segment(struct shmid_kernel *); 120 static int shm_find_segment_by_key(key_t); 121 static struct shmid_kernel *shm_find_segment_by_shmid(int); 122 static struct shmid_kernel *shm_find_segment_by_shmidx(int); 123 static int shm_delete_mapping(struct vmspace *vm, struct shmmap_state *); 124 static void shmrealloc(void); 125 static void shminit(void); 126 static int sysvshm_modload(struct module *, int, void *); 127 static int shmunload(void); 128 static void shmexit_myhook(struct vmspace *vm); 129 static void shmfork_myhook(struct proc *p1, struct proc *p2); 130 static int sysctl_shmsegs(SYSCTL_HANDLER_ARGS); 131 132 /* 133 * Tuneable values. 134 */ 135 #ifndef SHMMAXPGS 136 #define SHMMAXPGS 8192 /* Note: sysv shared memory is swap backed. */ 137 #endif 138 #ifndef SHMMAX 139 #define SHMMAX (SHMMAXPGS*PAGE_SIZE) 140 #endif 141 #ifndef SHMMIN 142 #define SHMMIN 1 143 #endif 144 #ifndef SHMMNI 145 #define SHMMNI 192 146 #endif 147 #ifndef SHMSEG 148 #define SHMSEG 128 149 #endif 150 #ifndef SHMALL 151 #define SHMALL (SHMMAXPGS) 152 #endif 153 154 struct shminfo shminfo = { 155 SHMMAX, 156 SHMMIN, 157 SHMMNI, 158 SHMSEG, 159 SHMALL 160 }; 161 162 static int shm_use_phys; 163 static int shm_allow_removed; 164 165 SYSCTL_ULONG(_kern_ipc, OID_AUTO, shmmax, CTLFLAG_RW, &shminfo.shmmax, 0, 166 "Maximum shared memory segment size"); 167 SYSCTL_ULONG(_kern_ipc, OID_AUTO, shmmin, CTLFLAG_RW, &shminfo.shmmin, 0, 168 "Minimum shared memory segment size"); 169 SYSCTL_ULONG(_kern_ipc, OID_AUTO, shmmni, CTLFLAG_RDTUN, &shminfo.shmmni, 0, 170 "Number of shared memory identifiers"); 171 SYSCTL_ULONG(_kern_ipc, OID_AUTO, shmseg, CTLFLAG_RDTUN, &shminfo.shmseg, 0, 172 "Number of segments per process"); 173 SYSCTL_ULONG(_kern_ipc, OID_AUTO, shmall, CTLFLAG_RW, &shminfo.shmall, 0, 174 "Maximum number of pages available for shared memory"); 175 SYSCTL_INT(_kern_ipc, OID_AUTO, shm_use_phys, CTLFLAG_RW, 176 &shm_use_phys, 0, "Enable/Disable locking of shared memory pages in core"); 177 SYSCTL_INT(_kern_ipc, OID_AUTO, shm_allow_removed, CTLFLAG_RW, 178 &shm_allow_removed, 0, 179 "Enable/Disable attachment to attached segments marked for removal"); 180 SYSCTL_PROC(_kern_ipc, OID_AUTO, shmsegs, CTLFLAG_RD, 181 NULL, 0, sysctl_shmsegs, "", 182 "Current number of shared memory segments allocated"); 183 184 static int 185 shm_find_segment_by_key(key) 186 key_t key; 187 { 188 int i; 189 190 for (i = 0; i < shmalloced; i++) 191 if ((shmsegs[i].u.shm_perm.mode & SHMSEG_ALLOCATED) && 192 shmsegs[i].u.shm_perm.key == key) 193 return (i); 194 return (-1); 195 } 196 197 static struct shmid_kernel * 198 shm_find_segment_by_shmid(int shmid) 199 { 200 int segnum; 201 struct shmid_kernel *shmseg; 202 203 segnum = IPCID_TO_IX(shmid); 204 if (segnum < 0 || segnum >= shmalloced) 205 return (NULL); 206 shmseg = &shmsegs[segnum]; 207 if ((shmseg->u.shm_perm.mode & SHMSEG_ALLOCATED) == 0 || 208 (!shm_allow_removed && 209 (shmseg->u.shm_perm.mode & SHMSEG_REMOVED) != 0) || 210 shmseg->u.shm_perm.seq != IPCID_TO_SEQ(shmid)) 211 return (NULL); 212 return (shmseg); 213 } 214 215 static struct shmid_kernel * 216 shm_find_segment_by_shmidx(int segnum) 217 { 218 struct shmid_kernel *shmseg; 219 220 if (segnum < 0 || segnum >= shmalloced) 221 return (NULL); 222 shmseg = &shmsegs[segnum]; 223 if ((shmseg->u.shm_perm.mode & SHMSEG_ALLOCATED) == 0 || 224 (!shm_allow_removed && 225 (shmseg->u.shm_perm.mode & SHMSEG_REMOVED) != 0)) 226 return (NULL); 227 return (shmseg); 228 } 229 230 static void 231 shm_deallocate_segment(shmseg) 232 struct shmid_kernel *shmseg; 233 { 234 vm_size_t size; 235 236 GIANT_REQUIRED; 237 238 vm_object_deallocate(shmseg->object); 239 shmseg->object = NULL; 240 size = round_page(shmseg->u.shm_segsz); 241 shm_committed -= btoc(size); 242 shm_nused--; 243 shmseg->u.shm_perm.mode = SHMSEG_FREE; 244 #ifdef MAC 245 mac_sysvshm_cleanup(shmseg); 246 #endif 247 } 248 249 static int 250 shm_delete_mapping(struct vmspace *vm, struct shmmap_state *shmmap_s) 251 { 252 struct shmid_kernel *shmseg; 253 int segnum, result; 254 vm_size_t size; 255 256 GIANT_REQUIRED; 257 258 segnum = IPCID_TO_IX(shmmap_s->shmid); 259 shmseg = &shmsegs[segnum]; 260 size = round_page(shmseg->u.shm_segsz); 261 result = vm_map_remove(&vm->vm_map, shmmap_s->va, shmmap_s->va + size); 262 if (result != KERN_SUCCESS) 263 return (EINVAL); 264 shmmap_s->shmid = -1; 265 shmseg->u.shm_dtime = time_second; 266 if ((--shmseg->u.shm_nattch <= 0) && 267 (shmseg->u.shm_perm.mode & SHMSEG_REMOVED)) { 268 shm_deallocate_segment(shmseg); 269 shm_last_free = segnum; 270 } 271 return (0); 272 } 273 274 #ifndef _SYS_SYSPROTO_H_ 275 struct shmdt_args { 276 const void *shmaddr; 277 }; 278 #endif 279 int 280 shmdt(td, uap) 281 struct thread *td; 282 struct shmdt_args *uap; 283 { 284 struct proc *p = td->td_proc; 285 struct shmmap_state *shmmap_s; 286 #ifdef MAC 287 struct shmid_kernel *shmsegptr; 288 #endif 289 int i; 290 int error = 0; 291 292 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC)) 293 return (ENOSYS); 294 mtx_lock(&Giant); 295 shmmap_s = p->p_vmspace->vm_shm; 296 if (shmmap_s == NULL) { 297 error = EINVAL; 298 goto done2; 299 } 300 for (i = 0; i < shminfo.shmseg; i++, shmmap_s++) { 301 if (shmmap_s->shmid != -1 && 302 shmmap_s->va == (vm_offset_t)uap->shmaddr) { 303 break; 304 } 305 } 306 if (i == shminfo.shmseg) { 307 error = EINVAL; 308 goto done2; 309 } 310 #ifdef MAC 311 shmsegptr = &shmsegs[IPCID_TO_IX(shmmap_s->shmid)]; 312 error = mac_sysvshm_check_shmdt(td->td_ucred, shmsegptr); 313 if (error != 0) 314 goto done2; 315 #endif 316 error = shm_delete_mapping(p->p_vmspace, shmmap_s); 317 done2: 318 mtx_unlock(&Giant); 319 return (error); 320 } 321 322 #ifndef _SYS_SYSPROTO_H_ 323 struct shmat_args { 324 int shmid; 325 const void *shmaddr; 326 int shmflg; 327 }; 328 #endif 329 int 330 kern_shmat(td, shmid, shmaddr, shmflg) 331 struct thread *td; 332 int shmid; 333 const void *shmaddr; 334 int shmflg; 335 { 336 struct proc *p = td->td_proc; 337 int i, flags; 338 struct shmid_kernel *shmseg; 339 struct shmmap_state *shmmap_s = NULL; 340 vm_offset_t attach_va; 341 vm_prot_t prot; 342 vm_size_t size; 343 int rv; 344 int error = 0; 345 346 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC)) 347 return (ENOSYS); 348 mtx_lock(&Giant); 349 shmmap_s = p->p_vmspace->vm_shm; 350 if (shmmap_s == NULL) { 351 shmmap_s = malloc(shminfo.shmseg * sizeof(struct shmmap_state), 352 M_SHM, M_WAITOK); 353 for (i = 0; i < shminfo.shmseg; i++) 354 shmmap_s[i].shmid = -1; 355 p->p_vmspace->vm_shm = shmmap_s; 356 } 357 shmseg = shm_find_segment_by_shmid(shmid); 358 if (shmseg == NULL) { 359 error = EINVAL; 360 goto done2; 361 } 362 error = ipcperm(td, &shmseg->u.shm_perm, 363 (shmflg & SHM_RDONLY) ? IPC_R : IPC_R|IPC_W); 364 if (error) 365 goto done2; 366 #ifdef MAC 367 error = mac_sysvshm_check_shmat(td->td_ucred, shmseg, shmflg); 368 if (error != 0) 369 goto done2; 370 #endif 371 for (i = 0; i < shminfo.shmseg; i++) { 372 if (shmmap_s->shmid == -1) 373 break; 374 shmmap_s++; 375 } 376 if (i >= shminfo.shmseg) { 377 error = EMFILE; 378 goto done2; 379 } 380 size = round_page(shmseg->u.shm_segsz); 381 prot = VM_PROT_READ; 382 if ((shmflg & SHM_RDONLY) == 0) 383 prot |= VM_PROT_WRITE; 384 flags = MAP_ANON | MAP_SHARED; 385 if (shmaddr) { 386 flags |= MAP_FIXED; 387 if (shmflg & SHM_RND) { 388 attach_va = (vm_offset_t)shmaddr & ~(SHMLBA-1); 389 } else if (((vm_offset_t)shmaddr & (SHMLBA-1)) == 0) { 390 attach_va = (vm_offset_t)shmaddr; 391 } else { 392 error = EINVAL; 393 goto done2; 394 } 395 } else { 396 /* 397 * This is just a hint to vm_map_find() about where to 398 * put it. 399 */ 400 PROC_LOCK(p); 401 attach_va = round_page((vm_offset_t)p->p_vmspace->vm_daddr + 402 lim_max(p, RLIMIT_DATA)); 403 PROC_UNLOCK(p); 404 } 405 406 vm_object_reference(shmseg->object); 407 rv = vm_map_find(&p->p_vmspace->vm_map, shmseg->object, 408 0, &attach_va, size, (flags & MAP_FIXED) ? VMFS_NO_SPACE : 409 VMFS_ANY_SPACE, prot, prot, 0); 410 if (rv != KERN_SUCCESS) { 411 vm_object_deallocate(shmseg->object); 412 error = ENOMEM; 413 goto done2; 414 } 415 vm_map_inherit(&p->p_vmspace->vm_map, 416 attach_va, attach_va + size, VM_INHERIT_SHARE); 417 418 shmmap_s->va = attach_va; 419 shmmap_s->shmid = shmid; 420 shmseg->u.shm_lpid = p->p_pid; 421 shmseg->u.shm_atime = time_second; 422 shmseg->u.shm_nattch++; 423 td->td_retval[0] = attach_va; 424 done2: 425 mtx_unlock(&Giant); 426 return (error); 427 } 428 429 int 430 shmat(td, uap) 431 struct thread *td; 432 struct shmat_args *uap; 433 { 434 return kern_shmat(td, uap->shmid, uap->shmaddr, uap->shmflg); 435 } 436 437 int 438 kern_shmctl(td, shmid, cmd, buf, bufsz) 439 struct thread *td; 440 int shmid; 441 int cmd; 442 void *buf; 443 size_t *bufsz; 444 { 445 int error = 0; 446 struct shmid_kernel *shmseg; 447 448 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC)) 449 return (ENOSYS); 450 451 mtx_lock(&Giant); 452 switch (cmd) { 453 /* 454 * It is possible that kern_shmctl is being called from the Linux ABI 455 * layer, in which case, we will need to implement IPC_INFO. It should 456 * be noted that other shmctl calls will be funneled through here for 457 * Linix binaries as well. 458 * 459 * NB: The Linux ABI layer will convert this data to structure(s) more 460 * consistent with the Linux ABI. 461 */ 462 case IPC_INFO: 463 memcpy(buf, &shminfo, sizeof(shminfo)); 464 if (bufsz) 465 *bufsz = sizeof(shminfo); 466 td->td_retval[0] = shmalloced; 467 goto done2; 468 case SHM_INFO: { 469 struct shm_info shm_info; 470 shm_info.used_ids = shm_nused; 471 shm_info.shm_rss = 0; /*XXX where to get from ? */ 472 shm_info.shm_tot = 0; /*XXX where to get from ? */ 473 shm_info.shm_swp = 0; /*XXX where to get from ? */ 474 shm_info.swap_attempts = 0; /*XXX where to get from ? */ 475 shm_info.swap_successes = 0; /*XXX where to get from ? */ 476 memcpy(buf, &shm_info, sizeof(shm_info)); 477 if (bufsz) 478 *bufsz = sizeof(shm_info); 479 td->td_retval[0] = shmalloced; 480 goto done2; 481 } 482 } 483 if (cmd == SHM_STAT) 484 shmseg = shm_find_segment_by_shmidx(shmid); 485 else 486 shmseg = shm_find_segment_by_shmid(shmid); 487 if (shmseg == NULL) { 488 error = EINVAL; 489 goto done2; 490 } 491 #ifdef MAC 492 error = mac_sysvshm_check_shmctl(td->td_ucred, shmseg, cmd); 493 if (error != 0) 494 goto done2; 495 #endif 496 switch (cmd) { 497 case SHM_STAT: 498 case IPC_STAT: 499 error = ipcperm(td, &shmseg->u.shm_perm, IPC_R); 500 if (error) 501 goto done2; 502 memcpy(buf, &shmseg->u, sizeof(struct shmid_ds)); 503 if (bufsz) 504 *bufsz = sizeof(struct shmid_ds); 505 if (cmd == SHM_STAT) 506 td->td_retval[0] = IXSEQ_TO_IPCID(shmid, shmseg->u.shm_perm); 507 break; 508 case IPC_SET: { 509 struct shmid_ds *shmid; 510 511 shmid = (struct shmid_ds *)buf; 512 error = ipcperm(td, &shmseg->u.shm_perm, IPC_M); 513 if (error) 514 goto done2; 515 shmseg->u.shm_perm.uid = shmid->shm_perm.uid; 516 shmseg->u.shm_perm.gid = shmid->shm_perm.gid; 517 shmseg->u.shm_perm.mode = 518 (shmseg->u.shm_perm.mode & ~ACCESSPERMS) | 519 (shmid->shm_perm.mode & ACCESSPERMS); 520 shmseg->u.shm_ctime = time_second; 521 break; 522 } 523 case IPC_RMID: 524 error = ipcperm(td, &shmseg->u.shm_perm, IPC_M); 525 if (error) 526 goto done2; 527 shmseg->u.shm_perm.key = IPC_PRIVATE; 528 shmseg->u.shm_perm.mode |= SHMSEG_REMOVED; 529 if (shmseg->u.shm_nattch <= 0) { 530 shm_deallocate_segment(shmseg); 531 shm_last_free = IPCID_TO_IX(shmid); 532 } 533 break; 534 #if 0 535 case SHM_LOCK: 536 case SHM_UNLOCK: 537 #endif 538 default: 539 error = EINVAL; 540 break; 541 } 542 done2: 543 mtx_unlock(&Giant); 544 return (error); 545 } 546 547 #ifndef _SYS_SYSPROTO_H_ 548 struct shmctl_args { 549 int shmid; 550 int cmd; 551 struct shmid_ds *buf; 552 }; 553 #endif 554 int 555 shmctl(td, uap) 556 struct thread *td; 557 struct shmctl_args *uap; 558 { 559 int error = 0; 560 struct shmid_ds buf; 561 size_t bufsz; 562 563 /* 564 * The only reason IPC_INFO, SHM_INFO, SHM_STAT exists is to support 565 * Linux binaries. If we see the call come through the FreeBSD ABI, 566 * return an error back to the user since we do not to support this. 567 */ 568 if (uap->cmd == IPC_INFO || uap->cmd == SHM_INFO || 569 uap->cmd == SHM_STAT) 570 return (EINVAL); 571 572 /* IPC_SET needs to copyin the buffer before calling kern_shmctl */ 573 if (uap->cmd == IPC_SET) { 574 if ((error = copyin(uap->buf, &buf, sizeof(struct shmid_ds)))) 575 goto done; 576 } 577 578 error = kern_shmctl(td, uap->shmid, uap->cmd, (void *)&buf, &bufsz); 579 if (error) 580 goto done; 581 582 /* Cases in which we need to copyout */ 583 switch (uap->cmd) { 584 case IPC_STAT: 585 error = copyout(&buf, uap->buf, bufsz); 586 break; 587 } 588 589 done: 590 if (error) { 591 /* Invalidate the return value */ 592 td->td_retval[0] = -1; 593 } 594 return (error); 595 } 596 597 598 static int 599 shmget_existing(td, uap, mode, segnum) 600 struct thread *td; 601 struct shmget_args *uap; 602 int mode; 603 int segnum; 604 { 605 struct shmid_kernel *shmseg; 606 int error; 607 608 shmseg = &shmsegs[segnum]; 609 if (shmseg->u.shm_perm.mode & SHMSEG_REMOVED) { 610 /* 611 * This segment is in the process of being allocated. Wait 612 * until it's done, and look the key up again (in case the 613 * allocation failed or it was freed). 614 */ 615 shmseg->u.shm_perm.mode |= SHMSEG_WANTED; 616 error = tsleep(shmseg, PLOCK | PCATCH, "shmget", 0); 617 if (error) 618 return (error); 619 return (EAGAIN); 620 } 621 if ((uap->shmflg & (IPC_CREAT | IPC_EXCL)) == (IPC_CREAT | IPC_EXCL)) 622 return (EEXIST); 623 #ifdef MAC 624 error = mac_sysvshm_check_shmget(td->td_ucred, shmseg, uap->shmflg); 625 if (error != 0) 626 return (error); 627 #endif 628 if (uap->size != 0 && uap->size > shmseg->u.shm_segsz) 629 return (EINVAL); 630 td->td_retval[0] = IXSEQ_TO_IPCID(segnum, shmseg->u.shm_perm); 631 return (0); 632 } 633 634 static int 635 shmget_allocate_segment(td, uap, mode) 636 struct thread *td; 637 struct shmget_args *uap; 638 int mode; 639 { 640 int i, segnum, shmid; 641 size_t size; 642 struct ucred *cred = td->td_ucred; 643 struct shmid_kernel *shmseg; 644 vm_object_t shm_object; 645 646 GIANT_REQUIRED; 647 648 if (uap->size < shminfo.shmmin || uap->size > shminfo.shmmax) 649 return (EINVAL); 650 if (shm_nused >= shminfo.shmmni) /* Any shmids left? */ 651 return (ENOSPC); 652 size = round_page(uap->size); 653 if (shm_committed + btoc(size) > shminfo.shmall) 654 return (ENOMEM); 655 if (shm_last_free < 0) { 656 shmrealloc(); /* Maybe expand the shmsegs[] array. */ 657 for (i = 0; i < shmalloced; i++) 658 if (shmsegs[i].u.shm_perm.mode & SHMSEG_FREE) 659 break; 660 if (i == shmalloced) 661 return (ENOSPC); 662 segnum = i; 663 } else { 664 segnum = shm_last_free; 665 shm_last_free = -1; 666 } 667 shmseg = &shmsegs[segnum]; 668 /* 669 * In case we sleep in malloc(), mark the segment present but deleted 670 * so that noone else tries to create the same key. 671 */ 672 shmseg->u.shm_perm.mode = SHMSEG_ALLOCATED | SHMSEG_REMOVED; 673 shmseg->u.shm_perm.key = uap->key; 674 shmseg->u.shm_perm.seq = (shmseg->u.shm_perm.seq + 1) & 0x7fff; 675 shmid = IXSEQ_TO_IPCID(segnum, shmseg->u.shm_perm); 676 677 /* 678 * We make sure that we have allocated a pager before we need 679 * to. 680 */ 681 shm_object = vm_pager_allocate(shm_use_phys ? OBJT_PHYS : OBJT_SWAP, 682 0, size, VM_PROT_DEFAULT, 0, cred); 683 if (shm_object == NULL) 684 return (ENOMEM); 685 VM_OBJECT_LOCK(shm_object); 686 vm_object_clear_flag(shm_object, OBJ_ONEMAPPING); 687 vm_object_set_flag(shm_object, OBJ_NOSPLIT); 688 VM_OBJECT_UNLOCK(shm_object); 689 690 shmseg->object = shm_object; 691 shmseg->u.shm_perm.cuid = shmseg->u.shm_perm.uid = cred->cr_uid; 692 shmseg->u.shm_perm.cgid = shmseg->u.shm_perm.gid = cred->cr_gid; 693 shmseg->u.shm_perm.mode = (shmseg->u.shm_perm.mode & SHMSEG_WANTED) | 694 (mode & ACCESSPERMS) | SHMSEG_ALLOCATED; 695 shmseg->u.shm_segsz = uap->size; 696 shmseg->u.shm_cpid = td->td_proc->p_pid; 697 shmseg->u.shm_lpid = shmseg->u.shm_nattch = 0; 698 shmseg->u.shm_atime = shmseg->u.shm_dtime = 0; 699 #ifdef MAC 700 mac_sysvshm_create(cred, shmseg); 701 #endif 702 shmseg->u.shm_ctime = time_second; 703 shm_committed += btoc(size); 704 shm_nused++; 705 if (shmseg->u.shm_perm.mode & SHMSEG_WANTED) { 706 /* 707 * Somebody else wanted this key while we were asleep. Wake 708 * them up now. 709 */ 710 shmseg->u.shm_perm.mode &= ~SHMSEG_WANTED; 711 wakeup(shmseg); 712 } 713 td->td_retval[0] = shmid; 714 return (0); 715 } 716 717 #ifndef _SYS_SYSPROTO_H_ 718 struct shmget_args { 719 key_t key; 720 size_t size; 721 int shmflg; 722 }; 723 #endif 724 int 725 shmget(td, uap) 726 struct thread *td; 727 struct shmget_args *uap; 728 { 729 int segnum, mode; 730 int error; 731 732 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC)) 733 return (ENOSYS); 734 mtx_lock(&Giant); 735 mode = uap->shmflg & ACCESSPERMS; 736 if (uap->key != IPC_PRIVATE) { 737 again: 738 segnum = shm_find_segment_by_key(uap->key); 739 if (segnum >= 0) { 740 error = shmget_existing(td, uap, mode, segnum); 741 if (error == EAGAIN) 742 goto again; 743 goto done2; 744 } 745 if ((uap->shmflg & IPC_CREAT) == 0) { 746 error = ENOENT; 747 goto done2; 748 } 749 } 750 error = shmget_allocate_segment(td, uap, mode); 751 done2: 752 mtx_unlock(&Giant); 753 return (error); 754 } 755 756 static void 757 shmfork_myhook(p1, p2) 758 struct proc *p1, *p2; 759 { 760 struct shmmap_state *shmmap_s; 761 size_t size; 762 int i; 763 764 mtx_lock(&Giant); 765 size = shminfo.shmseg * sizeof(struct shmmap_state); 766 shmmap_s = malloc(size, M_SHM, M_WAITOK); 767 bcopy(p1->p_vmspace->vm_shm, shmmap_s, size); 768 p2->p_vmspace->vm_shm = shmmap_s; 769 for (i = 0; i < shminfo.shmseg; i++, shmmap_s++) 770 if (shmmap_s->shmid != -1) 771 shmsegs[IPCID_TO_IX(shmmap_s->shmid)].u.shm_nattch++; 772 mtx_unlock(&Giant); 773 } 774 775 static void 776 shmexit_myhook(struct vmspace *vm) 777 { 778 struct shmmap_state *base, *shm; 779 int i; 780 781 if ((base = vm->vm_shm) != NULL) { 782 vm->vm_shm = NULL; 783 mtx_lock(&Giant); 784 for (i = 0, shm = base; i < shminfo.shmseg; i++, shm++) { 785 if (shm->shmid != -1) 786 shm_delete_mapping(vm, shm); 787 } 788 mtx_unlock(&Giant); 789 free(base, M_SHM); 790 } 791 } 792 793 static void 794 shmrealloc(void) 795 { 796 int i; 797 struct shmid_kernel *newsegs; 798 799 if (shmalloced >= shminfo.shmmni) 800 return; 801 802 newsegs = malloc(shminfo.shmmni * sizeof(*newsegs), M_SHM, M_WAITOK); 803 if (newsegs == NULL) 804 return; 805 for (i = 0; i < shmalloced; i++) 806 bcopy(&shmsegs[i], &newsegs[i], sizeof(newsegs[0])); 807 for (; i < shminfo.shmmni; i++) { 808 shmsegs[i].u.shm_perm.mode = SHMSEG_FREE; 809 shmsegs[i].u.shm_perm.seq = 0; 810 #ifdef MAC 811 mac_sysvshm_init(&shmsegs[i]); 812 #endif 813 } 814 free(shmsegs, M_SHM); 815 shmsegs = newsegs; 816 shmalloced = shminfo.shmmni; 817 } 818 819 static void 820 shminit() 821 { 822 int i; 823 824 TUNABLE_ULONG_FETCH("kern.ipc.shmmaxpgs", &shminfo.shmall); 825 for (i = PAGE_SIZE; i > 0; i--) { 826 shminfo.shmmax = shminfo.shmall * i; 827 if (shminfo.shmmax >= shminfo.shmall) 828 break; 829 } 830 TUNABLE_ULONG_FETCH("kern.ipc.shmmin", &shminfo.shmmin); 831 TUNABLE_ULONG_FETCH("kern.ipc.shmmni", &shminfo.shmmni); 832 TUNABLE_ULONG_FETCH("kern.ipc.shmseg", &shminfo.shmseg); 833 TUNABLE_INT_FETCH("kern.ipc.shm_use_phys", &shm_use_phys); 834 835 shmalloced = shminfo.shmmni; 836 shmsegs = malloc(shmalloced * sizeof(shmsegs[0]), M_SHM, M_WAITOK); 837 if (shmsegs == NULL) 838 panic("cannot allocate initial memory for sysvshm"); 839 for (i = 0; i < shmalloced; i++) { 840 shmsegs[i].u.shm_perm.mode = SHMSEG_FREE; 841 shmsegs[i].u.shm_perm.seq = 0; 842 #ifdef MAC 843 mac_sysvshm_init(&shmsegs[i]); 844 #endif 845 } 846 shm_last_free = 0; 847 shm_nused = 0; 848 shm_committed = 0; 849 shmexit_hook = &shmexit_myhook; 850 shmfork_hook = &shmfork_myhook; 851 } 852 853 static int 854 shmunload() 855 { 856 #ifdef MAC 857 int i; 858 #endif 859 860 if (shm_nused > 0) 861 return (EBUSY); 862 863 #ifdef MAC 864 for (i = 0; i < shmalloced; i++) 865 mac_sysvshm_destroy(&shmsegs[i]); 866 #endif 867 free(shmsegs, M_SHM); 868 shmexit_hook = NULL; 869 shmfork_hook = NULL; 870 return (0); 871 } 872 873 static int 874 sysctl_shmsegs(SYSCTL_HANDLER_ARGS) 875 { 876 877 return (SYSCTL_OUT(req, shmsegs, shmalloced * sizeof(shmsegs[0]))); 878 } 879 880 #if defined(__i386__) && (defined(COMPAT_FREEBSD4) || defined(COMPAT_43)) 881 struct oshmid_ds { 882 struct ipc_perm_old shm_perm; /* operation perms */ 883 int shm_segsz; /* size of segment (bytes) */ 884 u_short shm_cpid; /* pid, creator */ 885 u_short shm_lpid; /* pid, last operation */ 886 short shm_nattch; /* no. of current attaches */ 887 time_t shm_atime; /* last attach time */ 888 time_t shm_dtime; /* last detach time */ 889 time_t shm_ctime; /* last change time */ 890 void *shm_handle; /* internal handle for shm segment */ 891 }; 892 893 struct oshmctl_args { 894 int shmid; 895 int cmd; 896 struct oshmid_ds *ubuf; 897 }; 898 899 static int 900 oshmctl(struct thread *td, struct oshmctl_args *uap) 901 { 902 #ifdef COMPAT_43 903 int error = 0; 904 struct shmid_kernel *shmseg; 905 struct oshmid_ds outbuf; 906 907 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC)) 908 return (ENOSYS); 909 mtx_lock(&Giant); 910 shmseg = shm_find_segment_by_shmid(uap->shmid); 911 if (shmseg == NULL) { 912 error = EINVAL; 913 goto done2; 914 } 915 switch (uap->cmd) { 916 case IPC_STAT: 917 error = ipcperm(td, &shmseg->u.shm_perm, IPC_R); 918 if (error) 919 goto done2; 920 #ifdef MAC 921 error = mac_sysvshm_check_shmctl(td->td_ucred, shmseg, uap->cmd); 922 if (error != 0) 923 goto done2; 924 #endif 925 ipcperm_new2old(&shmseg->u.shm_perm, &outbuf.shm_perm); 926 outbuf.shm_segsz = shmseg->u.shm_segsz; 927 outbuf.shm_cpid = shmseg->u.shm_cpid; 928 outbuf.shm_lpid = shmseg->u.shm_lpid; 929 outbuf.shm_nattch = shmseg->u.shm_nattch; 930 outbuf.shm_atime = shmseg->u.shm_atime; 931 outbuf.shm_dtime = shmseg->u.shm_dtime; 932 outbuf.shm_ctime = shmseg->u.shm_ctime; 933 outbuf.shm_handle = shmseg->object; 934 error = copyout(&outbuf, uap->ubuf, sizeof(outbuf)); 935 if (error) 936 goto done2; 937 break; 938 default: 939 error = freebsd7_shmctl(td, (struct freebsd7_shmctl_args *)uap); 940 break; 941 } 942 done2: 943 mtx_unlock(&Giant); 944 return (error); 945 #else 946 return (EINVAL); 947 #endif 948 } 949 950 /* XXX casting to (sy_call_t *) is bogus, as usual. */ 951 static sy_call_t *shmcalls[] = { 952 (sy_call_t *)shmat, (sy_call_t *)oshmctl, 953 (sy_call_t *)shmdt, (sy_call_t *)shmget, 954 (sy_call_t *)freebsd7_shmctl 955 }; 956 957 int 958 shmsys(td, uap) 959 struct thread *td; 960 /* XXX actually varargs. */ 961 struct shmsys_args /* { 962 int which; 963 int a2; 964 int a3; 965 int a4; 966 } */ *uap; 967 { 968 int error; 969 970 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC)) 971 return (ENOSYS); 972 if (uap->which < 0 || 973 uap->which >= sizeof(shmcalls)/sizeof(shmcalls[0])) 974 return (EINVAL); 975 mtx_lock(&Giant); 976 error = (*shmcalls[uap->which])(td, &uap->a2); 977 mtx_unlock(&Giant); 978 return (error); 979 } 980 981 SYSCALL_MODULE_HELPER(shmsys); 982 #endif /* i386 && (COMPAT_FREEBSD4 || COMPAT_43) */ 983 984 #if defined(COMPAT_FREEBSD4) || defined(COMPAT_FREEBSD5) || \ 985 defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD7) 986 987 #define CP(src, dst, fld) do { (dst).fld = (src).fld; } while (0) 988 989 990 #ifndef _SYS_SYSPROTO_H_ 991 struct freebsd7_shmctl_args { 992 int shmid; 993 int cmd; 994 struct shmid_ds_old *buf; 995 }; 996 #endif 997 int 998 freebsd7_shmctl(td, uap) 999 struct thread *td; 1000 struct freebsd7_shmctl_args *uap; 1001 { 1002 int error = 0; 1003 struct shmid_ds_old old; 1004 struct shmid_ds buf; 1005 size_t bufsz; 1006 1007 /* 1008 * The only reason IPC_INFO, SHM_INFO, SHM_STAT exists is to support 1009 * Linux binaries. If we see the call come through the FreeBSD ABI, 1010 * return an error back to the user since we do not to support this. 1011 */ 1012 if (uap->cmd == IPC_INFO || uap->cmd == SHM_INFO || 1013 uap->cmd == SHM_STAT) 1014 return (EINVAL); 1015 1016 /* IPC_SET needs to copyin the buffer before calling kern_shmctl */ 1017 if (uap->cmd == IPC_SET) { 1018 if ((error = copyin(uap->buf, &old, sizeof(old)))) 1019 goto done; 1020 ipcperm_old2new(&old.shm_perm, &buf.shm_perm); 1021 CP(old, buf, shm_segsz); 1022 CP(old, buf, shm_lpid); 1023 CP(old, buf, shm_cpid); 1024 CP(old, buf, shm_nattch); 1025 CP(old, buf, shm_atime); 1026 CP(old, buf, shm_dtime); 1027 CP(old, buf, shm_ctime); 1028 } 1029 1030 error = kern_shmctl(td, uap->shmid, uap->cmd, (void *)&buf, &bufsz); 1031 if (error) 1032 goto done; 1033 1034 /* Cases in which we need to copyout */ 1035 switch (uap->cmd) { 1036 case IPC_STAT: 1037 ipcperm_new2old(&buf.shm_perm, &old.shm_perm); 1038 if (buf.shm_segsz > INT_MAX) 1039 old.shm_segsz = INT_MAX; 1040 else 1041 CP(buf, old, shm_segsz); 1042 CP(buf, old, shm_lpid); 1043 CP(buf, old, shm_cpid); 1044 if (buf.shm_nattch > SHRT_MAX) 1045 old.shm_nattch = SHRT_MAX; 1046 else 1047 CP(buf, old, shm_nattch); 1048 CP(buf, old, shm_atime); 1049 CP(buf, old, shm_dtime); 1050 CP(buf, old, shm_ctime); 1051 old.shm_internal = NULL; 1052 error = copyout(&old, uap->buf, sizeof(old)); 1053 break; 1054 } 1055 1056 done: 1057 if (error) { 1058 /* Invalidate the return value */ 1059 td->td_retval[0] = -1; 1060 } 1061 return (error); 1062 } 1063 1064 SYSCALL_MODULE_HELPER(freebsd7_shmctl); 1065 1066 #undef CP 1067 1068 #endif /* COMPAT_FREEBSD4 || COMPAT_FREEBSD5 || COMPAT_FREEBSD6 || 1069 COMPAT_FREEBSD7 */ 1070 1071 static int 1072 sysvshm_modload(struct module *module, int cmd, void *arg) 1073 { 1074 int error = 0; 1075 1076 switch (cmd) { 1077 case MOD_LOAD: 1078 shminit(); 1079 break; 1080 case MOD_UNLOAD: 1081 error = shmunload(); 1082 break; 1083 case MOD_SHUTDOWN: 1084 break; 1085 default: 1086 error = EINVAL; 1087 break; 1088 } 1089 return (error); 1090 } 1091 1092 static moduledata_t sysvshm_mod = { 1093 "sysvshm", 1094 &sysvshm_modload, 1095 NULL 1096 }; 1097 1098 SYSCALL_MODULE_HELPER(shmat); 1099 SYSCALL_MODULE_HELPER(shmctl); 1100 SYSCALL_MODULE_HELPER(shmdt); 1101 SYSCALL_MODULE_HELPER(shmget); 1102 1103 DECLARE_MODULE(sysvshm, sysvshm_mod, SI_SUB_SYSV_SHM, SI_ORDER_FIRST); 1104 MODULE_VERSION(sysvshm, 1); 1105