1 /* $NetBSD: sysv_shm.c,v 1.23 1994/07/04 23:25:12 glass Exp $ */
3 * Copyright (c) 1994 Adam Glass and Charles Hannum. All rights reserved.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
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
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.
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.
32 * Copyright (c) 2003-2005 McAfee, Inc.
33 * All rights reserved.
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
40 * Redistribution and use in source and binary forms, with or without
41 * modification, are permitted provided that the following conditions
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.
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
62 #include <sys/cdefs.h>
63 __FBSDID("$FreeBSD$");
65 #include "opt_compat.h"
66 #include "opt_sysvipc.h"
68 #include <sys/param.h>
69 #include <sys/systm.h>
70 #include <sys/kernel.h>
71 #include <sys/limits.h>
73 #include <sys/sysctl.h>
76 #include <sys/malloc.h>
78 #include <sys/module.h>
79 #include <sys/mutex.h>
80 #include <sys/racct.h>
81 #include <sys/resourcevar.h>
82 #include <sys/rwlock.h>
84 #include <sys/syscall.h>
85 #include <sys/syscallsubr.h>
86 #include <sys/sysent.h>
87 #include <sys/sysproto.h>
90 #include <security/mac/mac_framework.h>
93 #include <vm/vm_param.h>
95 #include <vm/vm_object.h>
96 #include <vm/vm_map.h>
97 #include <vm/vm_page.h>
98 #include <vm/vm_pager.h>
100 FEATURE(sysv_shm, "System V shared memory segments support");
102 static MALLOC_DEFINE(M_SHM, "shm", "SVID compatible shared memory segments");
104 static int shmget_allocate_segment(struct thread *td,
105 struct shmget_args *uap, int mode);
106 static int shmget_existing(struct thread *td, struct shmget_args *uap,
107 int mode, int segnum);
109 #define SHMSEG_FREE 0x0200
110 #define SHMSEG_REMOVED 0x0400
111 #define SHMSEG_ALLOCATED 0x0800
113 static int shm_last_free, shm_nused, shmalloced;
114 vm_size_t shm_committed;
115 static struct shmid_kernel *shmsegs;
117 struct shmmap_state {
122 static void shm_deallocate_segment(struct shmid_kernel *);
123 static int shm_find_segment_by_key(key_t);
124 static struct shmid_kernel *shm_find_segment(int, bool);
125 static int shm_delete_mapping(struct vmspace *vm, struct shmmap_state *);
126 static void shmrealloc(void);
127 static int shminit(void);
128 static int sysvshm_modload(struct module *, int, void *);
129 static int shmunload(void);
130 static void shmexit_myhook(struct vmspace *vm);
131 static void shmfork_myhook(struct proc *p1, struct proc *p2);
132 static int sysctl_shmsegs(SYSCTL_HANDLER_ARGS);
138 #define SHMMAXPGS 131072 /* Note: sysv shared memory is swap backed. */
141 #define SHMMAX (SHMMAXPGS*PAGE_SIZE)
153 #define SHMALL (SHMMAXPGS)
156 struct shminfo shminfo = {
164 static int shm_use_phys;
165 static int shm_allow_removed;
167 SYSCTL_ULONG(_kern_ipc, OID_AUTO, shmmax, CTLFLAG_RWTUN, &shminfo.shmmax, 0,
168 "Maximum shared memory segment size");
169 SYSCTL_ULONG(_kern_ipc, OID_AUTO, shmmin, CTLFLAG_RWTUN, &shminfo.shmmin, 0,
170 "Minimum shared memory segment size");
171 SYSCTL_ULONG(_kern_ipc, OID_AUTO, shmmni, CTLFLAG_RDTUN, &shminfo.shmmni, 0,
172 "Number of shared memory identifiers");
173 SYSCTL_ULONG(_kern_ipc, OID_AUTO, shmseg, CTLFLAG_RDTUN, &shminfo.shmseg, 0,
174 "Number of segments per process");
175 SYSCTL_ULONG(_kern_ipc, OID_AUTO, shmall, CTLFLAG_RWTUN, &shminfo.shmall, 0,
176 "Maximum number of pages available for shared memory");
177 SYSCTL_INT(_kern_ipc, OID_AUTO, shm_use_phys, CTLFLAG_RWTUN,
178 &shm_use_phys, 0, "Enable/Disable locking of shared memory pages in core");
179 SYSCTL_INT(_kern_ipc, OID_AUTO, shm_allow_removed, CTLFLAG_RWTUN,
180 &shm_allow_removed, 0,
181 "Enable/Disable attachment to attached segments marked for removal");
182 SYSCTL_PROC(_kern_ipc, OID_AUTO, shmsegs, CTLTYPE_OPAQUE | CTLFLAG_RD |
183 CTLFLAG_MPSAFE, NULL, 0, sysctl_shmsegs, "",
184 "Current number of shared memory segments allocated");
186 static struct sx sysvshmsx;
187 #define SYSVSHM_LOCK() sx_xlock(&sysvshmsx)
188 #define SYSVSHM_UNLOCK() sx_xunlock(&sysvshmsx)
189 #define SYSVSHM_ASSERT_LOCKED() sx_assert(&sysvshmsx, SA_XLOCKED)
192 shm_find_segment_by_key(key_t key)
196 for (i = 0; i < shmalloced; i++)
197 if ((shmsegs[i].u.shm_perm.mode & SHMSEG_ALLOCATED) &&
198 shmsegs[i].u.shm_perm.key == key)
204 * Finds segment either by shmid if is_shmid is true, or by segnum if
207 static struct shmid_kernel *
208 shm_find_segment(int arg, bool is_shmid)
210 struct shmid_kernel *shmseg;
213 segnum = is_shmid ? IPCID_TO_IX(arg) : arg;
214 if (segnum < 0 || segnum >= shmalloced)
216 shmseg = &shmsegs[segnum];
217 if ((shmseg->u.shm_perm.mode & SHMSEG_ALLOCATED) == 0 ||
218 (!shm_allow_removed &&
219 (shmseg->u.shm_perm.mode & SHMSEG_REMOVED) != 0) ||
220 (is_shmid && shmseg->u.shm_perm.seq != IPCID_TO_SEQ(arg)))
226 shm_deallocate_segment(struct shmid_kernel *shmseg)
230 SYSVSHM_ASSERT_LOCKED();
232 vm_object_deallocate(shmseg->object);
233 shmseg->object = NULL;
234 size = round_page(shmseg->u.shm_segsz);
235 shm_committed -= btoc(size);
237 shmseg->u.shm_perm.mode = SHMSEG_FREE;
239 mac_sysvshm_cleanup(shmseg);
241 racct_sub_cred(shmseg->cred, RACCT_NSHM, 1);
242 racct_sub_cred(shmseg->cred, RACCT_SHMSIZE, size);
243 crfree(shmseg->cred);
248 shm_delete_mapping(struct vmspace *vm, struct shmmap_state *shmmap_s)
250 struct shmid_kernel *shmseg;
254 SYSVSHM_ASSERT_LOCKED();
255 segnum = IPCID_TO_IX(shmmap_s->shmid);
256 KASSERT(segnum >= 0 && segnum < shmalloced,
257 ("segnum %d shmalloced %d", segnum, shmalloced));
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)
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;
275 kern_shmdt_locked(struct thread *td, const void *shmaddr)
277 struct proc *p = td->td_proc;
278 struct shmmap_state *shmmap_s;
280 struct shmid_kernel *shmsegptr;
285 SYSVSHM_ASSERT_LOCKED();
286 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC))
288 shmmap_s = p->p_vmspace->vm_shm;
289 if (shmmap_s == NULL)
291 for (i = 0; i < shminfo.shmseg; i++, shmmap_s++) {
292 if (shmmap_s->shmid != -1 &&
293 shmmap_s->va == (vm_offset_t)shmaddr) {
297 if (i == shminfo.shmseg)
300 shmsegptr = &shmsegs[IPCID_TO_IX(shmmap_s->shmid)];
301 error = mac_sysvshm_check_shmdt(td->td_ucred, shmsegptr);
305 return (shm_delete_mapping(p->p_vmspace, shmmap_s));
308 #ifndef _SYS_SYSPROTO_H_
314 sys_shmdt(struct thread *td, struct shmdt_args *uap)
319 error = kern_shmdt_locked(td, uap->shmaddr);
325 kern_shmat_locked(struct thread *td, int shmid, const void *shmaddr,
328 struct proc *p = td->td_proc;
329 struct shmid_kernel *shmseg;
330 struct shmmap_state *shmmap_s;
331 vm_offset_t attach_va;
336 SYSVSHM_ASSERT_LOCKED();
337 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC))
339 shmmap_s = p->p_vmspace->vm_shm;
340 if (shmmap_s == NULL) {
341 shmmap_s = malloc(shminfo.shmseg * sizeof(struct shmmap_state),
343 for (i = 0; i < shminfo.shmseg; i++)
344 shmmap_s[i].shmid = -1;
345 KASSERT(p->p_vmspace->vm_shm == NULL, ("raced"));
346 p->p_vmspace->vm_shm = shmmap_s;
348 shmseg = shm_find_segment(shmid, true);
351 error = ipcperm(td, &shmseg->u.shm_perm,
352 (shmflg & SHM_RDONLY) ? IPC_R : IPC_R|IPC_W);
356 error = mac_sysvshm_check_shmat(td->td_ucred, shmseg, shmflg);
360 for (i = 0; i < shminfo.shmseg; i++) {
361 if (shmmap_s->shmid == -1)
365 if (i >= shminfo.shmseg)
367 size = round_page(shmseg->u.shm_segsz);
369 if ((shmflg & SHM_RDONLY) == 0)
370 prot |= VM_PROT_WRITE;
371 if (shmaddr != NULL) {
372 if ((shmflg & SHM_RND) != 0)
373 attach_va = (vm_offset_t)shmaddr & ~(SHMLBA-1);
374 else if (((vm_offset_t)shmaddr & (SHMLBA-1)) == 0)
375 attach_va = (vm_offset_t)shmaddr;
380 * This is just a hint to vm_map_find() about where to
383 attach_va = round_page((vm_offset_t)p->p_vmspace->vm_daddr +
384 lim_max(td, RLIMIT_DATA));
387 vm_object_reference(shmseg->object);
388 rv = vm_map_find(&p->p_vmspace->vm_map, shmseg->object, 0, &attach_va,
389 size, 0, shmaddr != NULL ? VMFS_NO_SPACE : VMFS_OPTIMAL_SPACE,
390 prot, prot, MAP_INHERIT_SHARE | MAP_PREFAULT_PARTIAL);
391 if (rv != KERN_SUCCESS) {
392 vm_object_deallocate(shmseg->object);
396 shmmap_s->va = attach_va;
397 shmmap_s->shmid = shmid;
398 shmseg->u.shm_lpid = p->p_pid;
399 shmseg->u.shm_atime = time_second;
400 shmseg->u.shm_nattch++;
401 td->td_retval[0] = attach_va;
406 kern_shmat(struct thread *td, int shmid, const void *shmaddr, int shmflg)
411 error = kern_shmat_locked(td, shmid, shmaddr, shmflg);
416 #ifndef _SYS_SYSPROTO_H_
424 sys_shmat(struct thread *td, struct shmat_args *uap)
427 return (kern_shmat(td, uap->shmid, uap->shmaddr, uap->shmflg));
431 kern_shmctl_locked(struct thread *td, int shmid, int cmd, void *buf,
434 struct shmid_kernel *shmseg;
435 struct shmid_ds *shmidp;
436 struct shm_info shm_info;
439 SYSVSHM_ASSERT_LOCKED();
441 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC))
446 * It is possible that kern_shmctl is being called from the Linux ABI
447 * layer, in which case, we will need to implement IPC_INFO. It should
448 * be noted that other shmctl calls will be funneled through here for
449 * Linix binaries as well.
451 * NB: The Linux ABI layer will convert this data to structure(s) more
452 * consistent with the Linux ABI.
455 memcpy(buf, &shminfo, sizeof(shminfo));
457 *bufsz = sizeof(shminfo);
458 td->td_retval[0] = shmalloced;
461 shm_info.used_ids = shm_nused;
462 shm_info.shm_rss = 0; /*XXX where to get from ? */
463 shm_info.shm_tot = 0; /*XXX where to get from ? */
464 shm_info.shm_swp = 0; /*XXX where to get from ? */
465 shm_info.swap_attempts = 0; /*XXX where to get from ? */
466 shm_info.swap_successes = 0; /*XXX where to get from ? */
467 memcpy(buf, &shm_info, sizeof(shm_info));
469 *bufsz = sizeof(shm_info);
470 td->td_retval[0] = shmalloced;
474 shmseg = shm_find_segment(shmid, cmd != SHM_STAT);
478 error = mac_sysvshm_check_shmctl(td->td_ucred, shmseg, cmd);
485 error = ipcperm(td, &shmseg->u.shm_perm, IPC_R);
488 memcpy(buf, &shmseg->u, sizeof(struct shmid_ds));
490 *bufsz = sizeof(struct shmid_ds);
491 if (cmd == SHM_STAT) {
492 td->td_retval[0] = IXSEQ_TO_IPCID(shmid,
497 shmidp = (struct shmid_ds *)buf;
498 error = ipcperm(td, &shmseg->u.shm_perm, IPC_M);
501 shmseg->u.shm_perm.uid = shmidp->shm_perm.uid;
502 shmseg->u.shm_perm.gid = shmidp->shm_perm.gid;
503 shmseg->u.shm_perm.mode =
504 (shmseg->u.shm_perm.mode & ~ACCESSPERMS) |
505 (shmidp->shm_perm.mode & ACCESSPERMS);
506 shmseg->u.shm_ctime = time_second;
509 error = ipcperm(td, &shmseg->u.shm_perm, IPC_M);
512 shmseg->u.shm_perm.key = IPC_PRIVATE;
513 shmseg->u.shm_perm.mode |= SHMSEG_REMOVED;
514 if (shmseg->u.shm_nattch <= 0) {
515 shm_deallocate_segment(shmseg);
516 shm_last_free = IPCID_TO_IX(shmid);
531 kern_shmctl(struct thread *td, int shmid, int cmd, void *buf, size_t *bufsz)
536 error = kern_shmctl_locked(td, shmid, cmd, buf, bufsz);
542 #ifndef _SYS_SYSPROTO_H_
546 struct shmid_ds *buf;
550 sys_shmctl(struct thread *td, struct shmctl_args *uap)
557 * The only reason IPC_INFO, SHM_INFO, SHM_STAT exists is to support
558 * Linux binaries. If we see the call come through the FreeBSD ABI,
559 * return an error back to the user since we do not to support this.
561 if (uap->cmd == IPC_INFO || uap->cmd == SHM_INFO ||
562 uap->cmd == SHM_STAT)
565 /* IPC_SET needs to copyin the buffer before calling kern_shmctl */
566 if (uap->cmd == IPC_SET) {
567 if ((error = copyin(uap->buf, &buf, sizeof(struct shmid_ds))))
571 error = kern_shmctl(td, uap->shmid, uap->cmd, (void *)&buf, &bufsz);
575 /* Cases in which we need to copyout */
578 error = copyout(&buf, uap->buf, bufsz);
584 /* Invalidate the return value */
585 td->td_retval[0] = -1;
592 shmget_existing(struct thread *td, struct shmget_args *uap, int mode,
595 struct shmid_kernel *shmseg;
600 SYSVSHM_ASSERT_LOCKED();
601 KASSERT(segnum >= 0 && segnum < shmalloced,
602 ("segnum %d shmalloced %d", segnum, shmalloced));
603 shmseg = &shmsegs[segnum];
604 if ((uap->shmflg & (IPC_CREAT | IPC_EXCL)) == (IPC_CREAT | IPC_EXCL))
607 error = mac_sysvshm_check_shmget(td->td_ucred, shmseg, uap->shmflg);
611 if (uap->size != 0 && uap->size > shmseg->u.shm_segsz)
613 td->td_retval[0] = IXSEQ_TO_IPCID(segnum, shmseg->u.shm_perm);
618 shmget_allocate_segment(struct thread *td, struct shmget_args *uap, int mode)
620 struct ucred *cred = td->td_ucred;
621 struct shmid_kernel *shmseg;
622 vm_object_t shm_object;
626 SYSVSHM_ASSERT_LOCKED();
628 if (uap->size < shminfo.shmmin || uap->size > shminfo.shmmax)
630 if (shm_nused >= shminfo.shmmni) /* Any shmids left? */
632 size = round_page(uap->size);
633 if (shm_committed + btoc(size) > shminfo.shmall)
635 if (shm_last_free < 0) {
636 shmrealloc(); /* Maybe expand the shmsegs[] array. */
637 for (i = 0; i < shmalloced; i++)
638 if (shmsegs[i].u.shm_perm.mode & SHMSEG_FREE)
644 segnum = shm_last_free;
647 KASSERT(segnum >= 0 && segnum < shmalloced,
648 ("segnum %d shmalloced %d", segnum, shmalloced));
649 shmseg = &shmsegs[segnum];
652 PROC_LOCK(td->td_proc);
653 if (racct_add(td->td_proc, RACCT_NSHM, 1)) {
654 PROC_UNLOCK(td->td_proc);
657 if (racct_add(td->td_proc, RACCT_SHMSIZE, size)) {
658 racct_sub(td->td_proc, RACCT_NSHM, 1);
659 PROC_UNLOCK(td->td_proc);
662 PROC_UNLOCK(td->td_proc);
667 * We make sure that we have allocated a pager before we need
670 shm_object = vm_pager_allocate(shm_use_phys ? OBJT_PHYS : OBJT_SWAP,
671 0, size, VM_PROT_DEFAULT, 0, cred);
672 if (shm_object == NULL) {
675 PROC_LOCK(td->td_proc);
676 racct_sub(td->td_proc, RACCT_NSHM, 1);
677 racct_sub(td->td_proc, RACCT_SHMSIZE, size);
678 PROC_UNLOCK(td->td_proc);
683 shm_object->pg_color = 0;
684 VM_OBJECT_WLOCK(shm_object);
685 vm_object_clear_flag(shm_object, OBJ_ONEMAPPING);
686 vm_object_set_flag(shm_object, OBJ_COLORED | OBJ_NOSPLIT);
687 VM_OBJECT_WUNLOCK(shm_object);
689 shmseg->object = shm_object;
690 shmseg->u.shm_perm.cuid = shmseg->u.shm_perm.uid = cred->cr_uid;
691 shmseg->u.shm_perm.cgid = shmseg->u.shm_perm.gid = cred->cr_gid;
692 shmseg->u.shm_perm.mode = (mode & ACCESSPERMS) | SHMSEG_ALLOCATED;
693 shmseg->u.shm_perm.key = uap->key;
694 shmseg->u.shm_perm.seq = (shmseg->u.shm_perm.seq + 1) & 0x7fff;
695 shmseg->cred = crhold(cred);
696 shmseg->u.shm_segsz = uap->size;
697 shmseg->u.shm_cpid = td->td_proc->p_pid;
698 shmseg->u.shm_lpid = shmseg->u.shm_nattch = 0;
699 shmseg->u.shm_atime = shmseg->u.shm_dtime = 0;
701 mac_sysvshm_create(cred, shmseg);
703 shmseg->u.shm_ctime = time_second;
704 shm_committed += btoc(size);
706 td->td_retval[0] = IXSEQ_TO_IPCID(segnum, shmseg->u.shm_perm);
711 #ifndef _SYS_SYSPROTO_H_
719 sys_shmget(struct thread *td, struct shmget_args *uap)
724 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC))
726 mode = uap->shmflg & ACCESSPERMS;
728 if (uap->key == IPC_PRIVATE) {
729 error = shmget_allocate_segment(td, uap, mode);
731 segnum = shm_find_segment_by_key(uap->key);
733 error = shmget_existing(td, uap, mode, segnum);
734 else if ((uap->shmflg & IPC_CREAT) == 0)
737 error = shmget_allocate_segment(td, uap, mode);
744 shmfork_myhook(struct proc *p1, struct proc *p2)
746 struct shmmap_state *shmmap_s;
751 size = shminfo.shmseg * sizeof(struct shmmap_state);
752 shmmap_s = malloc(size, M_SHM, M_WAITOK);
753 bcopy(p1->p_vmspace->vm_shm, shmmap_s, size);
754 p2->p_vmspace->vm_shm = shmmap_s;
755 for (i = 0; i < shminfo.shmseg; i++, shmmap_s++) {
756 if (shmmap_s->shmid != -1) {
757 KASSERT(IPCID_TO_IX(shmmap_s->shmid) >= 0 &&
758 IPCID_TO_IX(shmmap_s->shmid) < shmalloced,
759 ("segnum %d shmalloced %d",
760 IPCID_TO_IX(shmmap_s->shmid), shmalloced));
761 shmsegs[IPCID_TO_IX(shmmap_s->shmid)].u.shm_nattch++;
768 shmexit_myhook(struct vmspace *vm)
770 struct shmmap_state *base, *shm;
777 for (i = 0, shm = base; i < shminfo.shmseg; i++, shm++) {
778 if (shm->shmid != -1)
779 shm_delete_mapping(vm, shm);
789 struct shmid_kernel *newsegs;
792 SYSVSHM_ASSERT_LOCKED();
794 if (shmalloced >= shminfo.shmmni)
797 newsegs = malloc(shminfo.shmmni * sizeof(*newsegs), M_SHM, M_WAITOK);
798 for (i = 0; i < shmalloced; i++)
799 bcopy(&shmsegs[i], &newsegs[i], sizeof(newsegs[0]));
800 for (; i < shminfo.shmmni; i++) {
801 newsegs[i].u.shm_perm.mode = SHMSEG_FREE;
802 newsegs[i].u.shm_perm.seq = 0;
804 mac_sysvshm_init(&newsegs[i]);
807 free(shmsegs, M_SHM);
809 shmalloced = shminfo.shmmni;
812 static struct syscall_helper_data shm_syscalls[] = {
813 SYSCALL_INIT_HELPER(shmat),
814 SYSCALL_INIT_HELPER(shmctl),
815 SYSCALL_INIT_HELPER(shmdt),
816 SYSCALL_INIT_HELPER(shmget),
817 #if defined(COMPAT_FREEBSD4) || defined(COMPAT_FREEBSD5) || \
818 defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD7)
819 SYSCALL_INIT_HELPER_COMPAT(freebsd7_shmctl),
821 #if defined(__i386__) && (defined(COMPAT_FREEBSD4) || defined(COMPAT_43))
822 SYSCALL_INIT_HELPER(shmsys),
827 #ifdef COMPAT_FREEBSD32
828 #include <compat/freebsd32/freebsd32.h>
829 #include <compat/freebsd32/freebsd32_ipc.h>
830 #include <compat/freebsd32/freebsd32_proto.h>
831 #include <compat/freebsd32/freebsd32_signal.h>
832 #include <compat/freebsd32/freebsd32_syscall.h>
833 #include <compat/freebsd32/freebsd32_util.h>
835 static struct syscall_helper_data shm32_syscalls[] = {
836 SYSCALL32_INIT_HELPER_COMPAT(shmat),
837 SYSCALL32_INIT_HELPER_COMPAT(shmdt),
838 SYSCALL32_INIT_HELPER_COMPAT(shmget),
839 SYSCALL32_INIT_HELPER(freebsd32_shmsys),
840 SYSCALL32_INIT_HELPER(freebsd32_shmctl),
841 #if defined(COMPAT_FREEBSD4) || defined(COMPAT_FREEBSD5) || \
842 defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD7)
843 SYSCALL32_INIT_HELPER(freebsd7_freebsd32_shmctl),
855 if (TUNABLE_ULONG_FETCH("kern.ipc.shmmaxpgs", &shminfo.shmall) != 0)
856 printf("kern.ipc.shmmaxpgs is now called kern.ipc.shmall!\n");
858 if (shminfo.shmmax == SHMMAX) {
859 /* Initialize shmmax dealing with possible overflow. */
860 for (i = PAGE_SIZE; i != 0; i--) {
861 shminfo.shmmax = shminfo.shmall * i;
862 if ((shminfo.shmmax / shminfo.shmall) == (u_long)i)
866 shmalloced = shminfo.shmmni;
867 shmsegs = malloc(shmalloced * sizeof(shmsegs[0]), M_SHM, M_WAITOK);
868 for (i = 0; i < shmalloced; i++) {
869 shmsegs[i].u.shm_perm.mode = SHMSEG_FREE;
870 shmsegs[i].u.shm_perm.seq = 0;
872 mac_sysvshm_init(&shmsegs[i]);
878 sx_init(&sysvshmsx, "sysvshmsx");
879 shmexit_hook = &shmexit_myhook;
880 shmfork_hook = &shmfork_myhook;
882 error = syscall_helper_register(shm_syscalls, SY_THR_STATIC_KLD);
885 #ifdef COMPAT_FREEBSD32
886 error = syscall32_helper_register(shm32_syscalls, SY_THR_STATIC_KLD);
901 #ifdef COMPAT_FREEBSD32
902 syscall32_helper_unregister(shm32_syscalls);
904 syscall_helper_unregister(shm_syscalls);
906 for (i = 0; i < shmalloced; i++) {
908 mac_sysvshm_destroy(&shmsegs[i]);
911 * Objects might be still mapped into the processes
912 * address spaces. Actual free would happen on the
913 * last mapping destruction.
915 if (shmsegs[i].u.shm_perm.mode != SHMSEG_FREE)
916 vm_object_deallocate(shmsegs[i].object);
918 free(shmsegs, M_SHM);
921 sx_destroy(&sysvshmsx);
926 sysctl_shmsegs(SYSCTL_HANDLER_ARGS)
931 error = SYSCTL_OUT(req, shmsegs, shmalloced * sizeof(shmsegs[0]));
936 #if defined(__i386__) && (defined(COMPAT_FREEBSD4) || defined(COMPAT_43))
938 struct ipc_perm_old shm_perm; /* operation perms */
939 int shm_segsz; /* size of segment (bytes) */
940 u_short shm_cpid; /* pid, creator */
941 u_short shm_lpid; /* pid, last operation */
942 short shm_nattch; /* no. of current attaches */
943 time_t shm_atime; /* last attach time */
944 time_t shm_dtime; /* last detach time */
945 time_t shm_ctime; /* last change time */
946 void *shm_handle; /* internal handle for shm segment */
949 struct oshmctl_args {
952 struct oshmid_ds *ubuf;
956 oshmctl(struct thread *td, struct oshmctl_args *uap)
960 struct shmid_kernel *shmseg;
961 struct oshmid_ds outbuf;
963 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC))
965 if (uap->cmd != IPC_STAT) {
966 return (freebsd7_shmctl(td,
967 (struct freebsd7_shmctl_args *)uap));
970 shmseg = shm_find_segment(uap->shmid, true);
971 if (shmseg == NULL) {
975 error = ipcperm(td, &shmseg->u.shm_perm, IPC_R);
981 error = mac_sysvshm_check_shmctl(td->td_ucred, shmseg, uap->cmd);
987 ipcperm_new2old(&shmseg->u.shm_perm, &outbuf.shm_perm);
988 outbuf.shm_segsz = shmseg->u.shm_segsz;
989 outbuf.shm_cpid = shmseg->u.shm_cpid;
990 outbuf.shm_lpid = shmseg->u.shm_lpid;
991 outbuf.shm_nattch = shmseg->u.shm_nattch;
992 outbuf.shm_atime = shmseg->u.shm_atime;
993 outbuf.shm_dtime = shmseg->u.shm_dtime;
994 outbuf.shm_ctime = shmseg->u.shm_ctime;
995 outbuf.shm_handle = shmseg->object;
997 return (copyout(&outbuf, uap->ubuf, sizeof(outbuf)));
1003 /* XXX casting to (sy_call_t *) is bogus, as usual. */
1004 static sy_call_t *shmcalls[] = {
1005 (sy_call_t *)sys_shmat, (sy_call_t *)oshmctl,
1006 (sy_call_t *)sys_shmdt, (sy_call_t *)sys_shmget,
1007 (sy_call_t *)freebsd7_shmctl
1010 #ifndef _SYS_SYSPROTO_H_
1011 /* XXX actually varargs. */
1012 struct shmsys_args {
1020 sys_shmsys(struct thread *td, struct shmsys_args *uap)
1023 if (!prison_allow(td->td_ucred, PR_ALLOW_SYSVIPC))
1025 if (uap->which < 0 || uap->which >= nitems(shmcalls))
1027 return ((*shmcalls[uap->which])(td, &uap->a2));
1030 #endif /* i386 && (COMPAT_FREEBSD4 || COMPAT_43) */
1032 #ifdef COMPAT_FREEBSD32
1035 freebsd32_shmsys(struct thread *td, struct freebsd32_shmsys_args *uap)
1038 #if defined(COMPAT_FREEBSD4) || defined(COMPAT_FREEBSD5) || \
1039 defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD7)
1040 switch (uap->which) {
1041 case 0: { /* shmat */
1042 struct shmat_args ap;
1045 ap.shmaddr = PTRIN(uap->a3);
1046 ap.shmflg = uap->a4;
1047 return (sysent[SYS_shmat].sy_call(td, &ap));
1049 case 2: { /* shmdt */
1050 struct shmdt_args ap;
1052 ap.shmaddr = PTRIN(uap->a2);
1053 return (sysent[SYS_shmdt].sy_call(td, &ap));
1055 case 3: { /* shmget */
1056 struct shmget_args ap;
1060 ap.shmflg = uap->a4;
1061 return (sysent[SYS_shmget].sy_call(td, &ap));
1063 case 4: { /* shmctl */
1064 struct freebsd7_freebsd32_shmctl_args ap;
1068 ap.buf = PTRIN(uap->a4);
1069 return (freebsd7_freebsd32_shmctl(td, &ap));
1071 case 1: /* oshmctl */
1076 return (nosys(td, NULL));
1080 #if defined(COMPAT_FREEBSD4) || defined(COMPAT_FREEBSD5) || \
1081 defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD7)
1083 freebsd7_freebsd32_shmctl(struct thread *td,
1084 struct freebsd7_freebsd32_shmctl_args *uap)
1088 struct shmid_ds shmid_ds;
1089 struct shm_info shm_info;
1090 struct shminfo shminfo;
1093 struct shmid_ds32_old shmid_ds32;
1094 struct shm_info32 shm_info32;
1095 struct shminfo32 shminfo32;
1099 if (uap->cmd == IPC_SET) {
1100 if ((error = copyin(uap->buf, &u32.shmid_ds32,
1101 sizeof(u32.shmid_ds32))))
1103 freebsd32_ipcperm_old_in(&u32.shmid_ds32.shm_perm,
1104 &u.shmid_ds.shm_perm);
1105 CP(u32.shmid_ds32, u.shmid_ds, shm_segsz);
1106 CP(u32.shmid_ds32, u.shmid_ds, shm_lpid);
1107 CP(u32.shmid_ds32, u.shmid_ds, shm_cpid);
1108 CP(u32.shmid_ds32, u.shmid_ds, shm_nattch);
1109 CP(u32.shmid_ds32, u.shmid_ds, shm_atime);
1110 CP(u32.shmid_ds32, u.shmid_ds, shm_dtime);
1111 CP(u32.shmid_ds32, u.shmid_ds, shm_ctime);
1114 error = kern_shmctl(td, uap->shmid, uap->cmd, (void *)&u, &sz);
1118 /* Cases in which we need to copyout */
1121 CP(u.shminfo, u32.shminfo32, shmmax);
1122 CP(u.shminfo, u32.shminfo32, shmmin);
1123 CP(u.shminfo, u32.shminfo32, shmmni);
1124 CP(u.shminfo, u32.shminfo32, shmseg);
1125 CP(u.shminfo, u32.shminfo32, shmall);
1126 error = copyout(&u32.shminfo32, uap->buf,
1127 sizeof(u32.shminfo32));
1130 CP(u.shm_info, u32.shm_info32, used_ids);
1131 CP(u.shm_info, u32.shm_info32, shm_rss);
1132 CP(u.shm_info, u32.shm_info32, shm_tot);
1133 CP(u.shm_info, u32.shm_info32, shm_swp);
1134 CP(u.shm_info, u32.shm_info32, swap_attempts);
1135 CP(u.shm_info, u32.shm_info32, swap_successes);
1136 error = copyout(&u32.shm_info32, uap->buf,
1137 sizeof(u32.shm_info32));
1141 freebsd32_ipcperm_old_out(&u.shmid_ds.shm_perm,
1142 &u32.shmid_ds32.shm_perm);
1143 if (u.shmid_ds.shm_segsz > INT32_MAX)
1144 u32.shmid_ds32.shm_segsz = INT32_MAX;
1146 CP(u.shmid_ds, u32.shmid_ds32, shm_segsz);
1147 CP(u.shmid_ds, u32.shmid_ds32, shm_lpid);
1148 CP(u.shmid_ds, u32.shmid_ds32, shm_cpid);
1149 CP(u.shmid_ds, u32.shmid_ds32, shm_nattch);
1150 CP(u.shmid_ds, u32.shmid_ds32, shm_atime);
1151 CP(u.shmid_ds, u32.shmid_ds32, shm_dtime);
1152 CP(u.shmid_ds, u32.shmid_ds32, shm_ctime);
1153 u32.shmid_ds32.shm_internal = 0;
1154 error = copyout(&u32.shmid_ds32, uap->buf,
1155 sizeof(u32.shmid_ds32));
1161 /* Invalidate the return value */
1162 td->td_retval[0] = -1;
1169 freebsd32_shmctl(struct thread *td, struct freebsd32_shmctl_args *uap)
1173 struct shmid_ds shmid_ds;
1174 struct shm_info shm_info;
1175 struct shminfo shminfo;
1178 struct shmid_ds32 shmid_ds32;
1179 struct shm_info32 shm_info32;
1180 struct shminfo32 shminfo32;
1184 if (uap->cmd == IPC_SET) {
1185 if ((error = copyin(uap->buf, &u32.shmid_ds32,
1186 sizeof(u32.shmid_ds32))))
1188 freebsd32_ipcperm_in(&u32.shmid_ds32.shm_perm,
1189 &u.shmid_ds.shm_perm);
1190 CP(u32.shmid_ds32, u.shmid_ds, shm_segsz);
1191 CP(u32.shmid_ds32, u.shmid_ds, shm_lpid);
1192 CP(u32.shmid_ds32, u.shmid_ds, shm_cpid);
1193 CP(u32.shmid_ds32, u.shmid_ds, shm_nattch);
1194 CP(u32.shmid_ds32, u.shmid_ds, shm_atime);
1195 CP(u32.shmid_ds32, u.shmid_ds, shm_dtime);
1196 CP(u32.shmid_ds32, u.shmid_ds, shm_ctime);
1199 error = kern_shmctl(td, uap->shmid, uap->cmd, (void *)&u, &sz);
1203 /* Cases in which we need to copyout */
1206 CP(u.shminfo, u32.shminfo32, shmmax);
1207 CP(u.shminfo, u32.shminfo32, shmmin);
1208 CP(u.shminfo, u32.shminfo32, shmmni);
1209 CP(u.shminfo, u32.shminfo32, shmseg);
1210 CP(u.shminfo, u32.shminfo32, shmall);
1211 error = copyout(&u32.shminfo32, uap->buf,
1212 sizeof(u32.shminfo32));
1215 CP(u.shm_info, u32.shm_info32, used_ids);
1216 CP(u.shm_info, u32.shm_info32, shm_rss);
1217 CP(u.shm_info, u32.shm_info32, shm_tot);
1218 CP(u.shm_info, u32.shm_info32, shm_swp);
1219 CP(u.shm_info, u32.shm_info32, swap_attempts);
1220 CP(u.shm_info, u32.shm_info32, swap_successes);
1221 error = copyout(&u32.shm_info32, uap->buf,
1222 sizeof(u32.shm_info32));
1226 freebsd32_ipcperm_out(&u.shmid_ds.shm_perm,
1227 &u32.shmid_ds32.shm_perm);
1228 if (u.shmid_ds.shm_segsz > INT32_MAX)
1229 u32.shmid_ds32.shm_segsz = INT32_MAX;
1231 CP(u.shmid_ds, u32.shmid_ds32, shm_segsz);
1232 CP(u.shmid_ds, u32.shmid_ds32, shm_lpid);
1233 CP(u.shmid_ds, u32.shmid_ds32, shm_cpid);
1234 CP(u.shmid_ds, u32.shmid_ds32, shm_nattch);
1235 CP(u.shmid_ds, u32.shmid_ds32, shm_atime);
1236 CP(u.shmid_ds, u32.shmid_ds32, shm_dtime);
1237 CP(u.shmid_ds, u32.shmid_ds32, shm_ctime);
1238 error = copyout(&u32.shmid_ds32, uap->buf,
1239 sizeof(u32.shmid_ds32));
1245 /* Invalidate the return value */
1246 td->td_retval[0] = -1;
1252 #if defined(COMPAT_FREEBSD4) || defined(COMPAT_FREEBSD5) || \
1253 defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD7)
1256 #define CP(src, dst, fld) do { (dst).fld = (src).fld; } while (0)
1259 #ifndef _SYS_SYSPROTO_H_
1260 struct freebsd7_shmctl_args {
1263 struct shmid_ds_old *buf;
1267 freebsd7_shmctl(struct thread *td, struct freebsd7_shmctl_args *uap)
1270 struct shmid_ds_old old;
1271 struct shmid_ds buf;
1275 * The only reason IPC_INFO, SHM_INFO, SHM_STAT exists is to support
1276 * Linux binaries. If we see the call come through the FreeBSD ABI,
1277 * return an error back to the user since we do not to support this.
1279 if (uap->cmd == IPC_INFO || uap->cmd == SHM_INFO ||
1280 uap->cmd == SHM_STAT)
1283 /* IPC_SET needs to copyin the buffer before calling kern_shmctl */
1284 if (uap->cmd == IPC_SET) {
1285 if ((error = copyin(uap->buf, &old, sizeof(old))))
1287 ipcperm_old2new(&old.shm_perm, &buf.shm_perm);
1288 CP(old, buf, shm_segsz);
1289 CP(old, buf, shm_lpid);
1290 CP(old, buf, shm_cpid);
1291 CP(old, buf, shm_nattch);
1292 CP(old, buf, shm_atime);
1293 CP(old, buf, shm_dtime);
1294 CP(old, buf, shm_ctime);
1297 error = kern_shmctl(td, uap->shmid, uap->cmd, (void *)&buf, &bufsz);
1301 /* Cases in which we need to copyout */
1304 ipcperm_new2old(&buf.shm_perm, &old.shm_perm);
1305 if (buf.shm_segsz > INT_MAX)
1306 old.shm_segsz = INT_MAX;
1308 CP(buf, old, shm_segsz);
1309 CP(buf, old, shm_lpid);
1310 CP(buf, old, shm_cpid);
1311 if (buf.shm_nattch > SHRT_MAX)
1312 old.shm_nattch = SHRT_MAX;
1314 CP(buf, old, shm_nattch);
1315 CP(buf, old, shm_atime);
1316 CP(buf, old, shm_dtime);
1317 CP(buf, old, shm_ctime);
1318 old.shm_internal = NULL;
1319 error = copyout(&old, uap->buf, sizeof(old));
1325 /* Invalidate the return value */
1326 td->td_retval[0] = -1;
1331 #endif /* COMPAT_FREEBSD4 || COMPAT_FREEBSD5 || COMPAT_FREEBSD6 ||
1335 sysvshm_modload(struct module *module, int cmd, void *arg)
1346 error = shmunload();
1357 static moduledata_t sysvshm_mod = {
1363 DECLARE_MODULE(sysvshm, sysvshm_mod, SI_SUB_SYSV_SHM, SI_ORDER_FIRST);
1364 MODULE_VERSION(sysvshm, 1);