3 * Copyright (C) 1992 Krishna Balasubramanian
4 * Copyright (C) 1995 Eric Schenk, Bruno Haible
6 * /proc/sysvipc/sem support (c) 1999 Dragos Acostachioaie <dragos@iname.com>
8 * SMP-threaded, sysctl's added
9 * (c) 1999 Manfred Spraul <manfred@colorfullife.com>
10 * Enforced range limit on SEM_UNDO
11 * (c) 2001 Red Hat Inc
13 * (c) 2003 Manfred Spraul <manfred@colorfullife.com>
14 * Further wakeup optimizations, documentation
15 * (c) 2010 Manfred Spraul <manfred@colorfullife.com>
17 * support for audit of ipc object properties and permission changes
18 * Dustin Kirkland <dustin.kirkland@us.ibm.com>
22 * Pavel Emelianov <xemul@openvz.org>
24 * Implementation notes: (May 2010)
25 * This file implements System V semaphores.
27 * User space visible behavior:
28 * - FIFO ordering for semop() operations (just FIFO, not starvation
30 * - multiple semaphore operations that alter the same semaphore in
31 * one semop() are handled.
32 * - sem_ctime (time of last semctl()) is updated in the IPC_SET, SETVAL and
34 * - two Linux specific semctl() commands: SEM_STAT, SEM_INFO.
35 * - undo adjustments at process exit are limited to 0..SEMVMX.
36 * - namespace are supported.
37 * - SEMMSL, SEMMNS, SEMOPM and SEMMNI can be configured at runtine by writing
38 * to /proc/sys/kernel/sem.
39 * - statistics about the usage are reported in /proc/sysvipc/sem.
43 * - all global variables are read-mostly.
44 * - semop() calls and semctl(RMID) are synchronized by RCU.
45 * - most operations do write operations (actually: spin_lock calls) to
46 * the per-semaphore array structure.
47 * Thus: Perfect SMP scaling between independent semaphore arrays.
48 * If multiple semaphores in one array are used, then cache line
49 * trashing on the semaphore array spinlock will limit the scaling.
50 * - semncnt and semzcnt are calculated on demand in count_semncnt() and
52 * - the task that performs a successful semop() scans the list of all
53 * sleeping tasks and completes any pending operations that can be fulfilled.
54 * Semaphores are actively given to waiting tasks (necessary for FIFO).
55 * (see update_queue())
56 * - To improve the scalability, the actual wake-up calls are performed after
57 * dropping all locks. (see wake_up_sem_queue_prepare(),
58 * wake_up_sem_queue_do())
59 * - All work is done by the waker, the woken up task does not have to do
60 * anything - not even acquiring a lock or dropping a refcount.
61 * - A woken up task may not even touch the semaphore array anymore, it may
62 * have been destroyed already by a semctl(RMID).
63 * - The synchronizations between wake-ups due to a timeout/signal and a
64 * wake-up due to a completed semaphore operation is achieved by using an
65 * intermediate state (IN_WAKEUP).
66 * - UNDO values are stored in an array (one per process and per
67 * semaphore array, lazily allocated). For backwards compatibility, multiple
68 * modes for the UNDO variables are supported (per process, per thread)
69 * (see copy_semundo, CLONE_SYSVSEM)
70 * - There are two lists of the pending operations: a per-array list
71 * and per-semaphore list (stored in the array). This allows to achieve FIFO
72 * ordering without always scanning all pending operations.
73 * The worst-case behavior is nevertheless O(N^2) for N wakeups.
76 #include <linux/slab.h>
77 #include <linux/spinlock.h>
78 #include <linux/init.h>
79 #include <linux/proc_fs.h>
80 #include <linux/time.h>
81 #include <linux/security.h>
82 #include <linux/syscalls.h>
83 #include <linux/audit.h>
84 #include <linux/capability.h>
85 #include <linux/seq_file.h>
86 #include <linux/rwsem.h>
87 #include <linux/nsproxy.h>
88 #include <linux/ipc_namespace.h>
90 #include <asm/uaccess.h>
93 /* One semaphore structure for each semaphore in the system. */
95 int semval
; /* current value */
96 int sempid
; /* pid of last operation */
97 spinlock_t lock
; /* spinlock for fine-grained semtimedop */
98 struct list_head pending_alter
; /* pending single-sop operations */
99 /* that alter the semaphore */
100 struct list_head pending_const
; /* pending single-sop operations */
101 /* that do not alter the semaphore*/
102 time_t sem_otime
; /* candidate for sem_otime */
103 } ____cacheline_aligned_in_smp
;
105 /* One queue for each sleeping process in the system. */
107 struct list_head list
; /* queue of pending operations */
108 struct task_struct
*sleeper
; /* this process */
109 struct sem_undo
*undo
; /* undo structure */
110 int pid
; /* process id of requesting process */
111 int status
; /* completion status of operation */
112 struct sembuf
*sops
; /* array of pending operations */
113 int nsops
; /* number of operations */
114 int alter
; /* does *sops alter the array? */
117 /* Each task has a list of undo requests. They are executed automatically
118 * when the process exits.
121 struct list_head list_proc
; /* per-process list: *
122 * all undos from one process
124 struct rcu_head rcu
; /* rcu struct for sem_undo */
125 struct sem_undo_list
*ulp
; /* back ptr to sem_undo_list */
126 struct list_head list_id
; /* per semaphore array list:
127 * all undos for one array */
128 int semid
; /* semaphore set identifier */
129 short *semadj
; /* array of adjustments */
130 /* one per semaphore */
133 /* sem_undo_list controls shared access to the list of sem_undo structures
134 * that may be shared among all a CLONE_SYSVSEM task group.
136 struct sem_undo_list
{
139 struct list_head list_proc
;
143 #define sem_ids(ns) ((ns)->ids[IPC_SEM_IDS])
145 #define sem_checkid(sma, semid) ipc_checkid(&sma->sem_perm, semid)
147 static int newary(struct ipc_namespace
*, struct ipc_params
*);
148 static void freeary(struct ipc_namespace
*, struct kern_ipc_perm
*);
149 #ifdef CONFIG_PROC_FS
150 static int sysvipc_sem_proc_show(struct seq_file
*s
, void *it
);
153 #define SEMMSL_FAST 256 /* 512 bytes on stack */
154 #define SEMOPM_FAST 64 /* ~ 372 bytes on stack */
159 * sem_array.complex_count,
160 * sem_array.pending{_alter,_cont},
161 * sem_array.sem_undo: global sem_lock() for read/write
162 * sem_undo.proc_next: only "current" is allowed to read/write that field.
164 * sem_array.sem_base[i].pending_{const,alter}:
165 * global or semaphore sem_lock() for read/write
168 #define sc_semmsl sem_ctls[0]
169 #define sc_semmns sem_ctls[1]
170 #define sc_semopm sem_ctls[2]
171 #define sc_semmni sem_ctls[3]
173 void sem_init_ns(struct ipc_namespace
*ns
)
175 ns
->sc_semmsl
= SEMMSL
;
176 ns
->sc_semmns
= SEMMNS
;
177 ns
->sc_semopm
= SEMOPM
;
178 ns
->sc_semmni
= SEMMNI
;
180 ipc_init_ids(&ns
->ids
[IPC_SEM_IDS
]);
184 void sem_exit_ns(struct ipc_namespace
*ns
)
186 free_ipcs(ns
, &sem_ids(ns
), freeary
);
187 idr_destroy(&ns
->ids
[IPC_SEM_IDS
].ipcs_idr
);
191 void __init
sem_init (void)
193 sem_init_ns(&init_ipc_ns
);
194 ipc_init_proc_interface("sysvipc/sem",
195 " key semid perms nsems uid gid cuid cgid otime ctime\n",
196 IPC_SEM_IDS
, sysvipc_sem_proc_show
);
200 * unmerge_queues - unmerge queues, if possible.
201 * @sma: semaphore array
203 * The function unmerges the wait queues if complex_count is 0.
204 * It must be called prior to dropping the global semaphore array lock.
206 static void unmerge_queues(struct sem_array
*sma
)
208 struct sem_queue
*q
, *tq
;
210 /* complex operations still around? */
211 if (sma
->complex_count
)
214 * We will switch back to simple mode.
215 * Move all pending operation back into the per-semaphore
218 list_for_each_entry_safe(q
, tq
, &sma
->pending_alter
, list
) {
220 curr
= &sma
->sem_base
[q
->sops
[0].sem_num
];
222 list_add_tail(&q
->list
, &curr
->pending_alter
);
224 INIT_LIST_HEAD(&sma
->pending_alter
);
228 * merge_queues - Merge single semop queues into global queue
229 * @sma: semaphore array
231 * This function merges all per-semaphore queues into the global queue.
232 * It is necessary to achieve FIFO ordering for the pending single-sop
233 * operations when a multi-semop operation must sleep.
234 * Only the alter operations must be moved, the const operations can stay.
236 static void merge_queues(struct sem_array
*sma
)
239 for (i
= 0; i
< sma
->sem_nsems
; i
++) {
240 struct sem
*sem
= sma
->sem_base
+ i
;
242 list_splice_init(&sem
->pending_alter
, &sma
->pending_alter
);
246 static void sem_rcu_free(struct rcu_head
*head
)
248 struct ipc_rcu
*p
= container_of(head
, struct ipc_rcu
, rcu
);
249 struct sem_array
*sma
= ipc_rcu_to_struct(p
);
251 security_sem_free(sma
);
256 * If the request contains only one semaphore operation, and there are
257 * no complex transactions pending, lock only the semaphore involved.
258 * Otherwise, lock the entire semaphore array, since we either have
259 * multiple semaphores in our own semops, or we need to look at
260 * semaphores from other pending complex operations.
262 * Carefully guard against sma->complex_count changing between zero
263 * and non-zero while we are spinning for the lock. The value of
264 * sma->complex_count cannot change while we are holding the lock,
265 * so sem_unlock should be fine.
267 * The global lock path checks that all the local locks have been released,
268 * checking each local lock once. This means that the local lock paths
269 * cannot start their critical sections while the global lock is held.
271 static inline int sem_lock(struct sem_array
*sma
, struct sembuf
*sops
,
276 if (nsops
== 1 && !sma
->complex_count
) {
277 struct sem
*sem
= sma
->sem_base
+ sops
->sem_num
;
279 /* Lock just the semaphore we are interested in. */
280 spin_lock(&sem
->lock
);
283 * If sma->complex_count was set while we were spinning,
284 * we may need to look at things we did not lock here.
286 if (unlikely(sma
->complex_count
)) {
287 spin_unlock(&sem
->lock
);
292 * Another process is holding the global lock on the
293 * sem_array; we cannot enter our critical section,
294 * but have to wait for the global lock to be released.
296 if (unlikely(spin_is_locked(&sma
->sem_perm
.lock
))) {
297 spin_unlock(&sem
->lock
);
298 spin_unlock_wait(&sma
->sem_perm
.lock
);
302 locknum
= sops
->sem_num
;
306 * Lock the semaphore array, and wait for all of the
307 * individual semaphore locks to go away. The code
308 * above ensures no new single-lock holders will enter
309 * their critical section while the array lock is held.
312 ipc_lock_object(&sma
->sem_perm
);
313 for (i
= 0; i
< sma
->sem_nsems
; i
++) {
314 struct sem
*sem
= sma
->sem_base
+ i
;
315 spin_unlock_wait(&sem
->lock
);
322 static inline void sem_unlock(struct sem_array
*sma
, int locknum
)
326 ipc_unlock_object(&sma
->sem_perm
);
328 struct sem
*sem
= sma
->sem_base
+ locknum
;
329 spin_unlock(&sem
->lock
);
334 * sem_lock_(check_) routines are called in the paths where the rwsem
337 * The caller holds the RCU read lock.
339 static inline struct sem_array
*sem_obtain_lock(struct ipc_namespace
*ns
,
340 int id
, struct sembuf
*sops
, int nsops
, int *locknum
)
342 struct kern_ipc_perm
*ipcp
;
343 struct sem_array
*sma
;
345 ipcp
= ipc_obtain_object(&sem_ids(ns
), id
);
347 return ERR_CAST(ipcp
);
349 sma
= container_of(ipcp
, struct sem_array
, sem_perm
);
350 *locknum
= sem_lock(sma
, sops
, nsops
);
352 /* ipc_rmid() may have already freed the ID while sem_lock
353 * was spinning: verify that the structure is still valid
356 return container_of(ipcp
, struct sem_array
, sem_perm
);
358 sem_unlock(sma
, *locknum
);
359 return ERR_PTR(-EINVAL
);
362 static inline struct sem_array
*sem_obtain_object(struct ipc_namespace
*ns
, int id
)
364 struct kern_ipc_perm
*ipcp
= ipc_obtain_object(&sem_ids(ns
), id
);
367 return ERR_CAST(ipcp
);
369 return container_of(ipcp
, struct sem_array
, sem_perm
);
372 static inline struct sem_array
*sem_obtain_object_check(struct ipc_namespace
*ns
,
375 struct kern_ipc_perm
*ipcp
= ipc_obtain_object_check(&sem_ids(ns
), id
);
378 return ERR_CAST(ipcp
);
380 return container_of(ipcp
, struct sem_array
, sem_perm
);
383 static inline void sem_lock_and_putref(struct sem_array
*sma
)
385 sem_lock(sma
, NULL
, -1);
386 ipc_rcu_putref(sma
, ipc_rcu_free
);
389 static inline void sem_rmid(struct ipc_namespace
*ns
, struct sem_array
*s
)
391 ipc_rmid(&sem_ids(ns
), &s
->sem_perm
);
395 * Lockless wakeup algorithm:
396 * Without the check/retry algorithm a lockless wakeup is possible:
397 * - queue.status is initialized to -EINTR before blocking.
398 * - wakeup is performed by
399 * * unlinking the queue entry from the pending list
400 * * setting queue.status to IN_WAKEUP
401 * This is the notification for the blocked thread that a
402 * result value is imminent.
403 * * call wake_up_process
404 * * set queue.status to the final value.
405 * - the previously blocked thread checks queue.status:
406 * * if it's IN_WAKEUP, then it must wait until the value changes
407 * * if it's not -EINTR, then the operation was completed by
408 * update_queue. semtimedop can return queue.status without
409 * performing any operation on the sem array.
410 * * otherwise it must acquire the spinlock and check what's up.
412 * The two-stage algorithm is necessary to protect against the following
414 * - if queue.status is set after wake_up_process, then the woken up idle
415 * thread could race forward and try (and fail) to acquire sma->lock
416 * before update_queue had a chance to set queue.status
417 * - if queue.status is written before wake_up_process and if the
418 * blocked process is woken up by a signal between writing
419 * queue.status and the wake_up_process, then the woken up
420 * process could return from semtimedop and die by calling
421 * sys_exit before wake_up_process is called. Then wake_up_process
422 * will oops, because the task structure is already invalid.
423 * (yes, this happened on s390 with sysv msg).
429 * newary - Create a new semaphore set
431 * @params: ptr to the structure that contains key, semflg and nsems
433 * Called with sem_ids.rwsem held (as a writer)
436 static int newary(struct ipc_namespace
*ns
, struct ipc_params
*params
)
440 struct sem_array
*sma
;
442 key_t key
= params
->key
;
443 int nsems
= params
->u
.nsems
;
444 int semflg
= params
->flg
;
449 if (ns
->used_sems
+ nsems
> ns
->sc_semmns
)
452 size
= sizeof (*sma
) + nsems
* sizeof (struct sem
);
453 sma
= ipc_rcu_alloc(size
);
457 memset (sma
, 0, size
);
459 sma
->sem_perm
.mode
= (semflg
& S_IRWXUGO
);
460 sma
->sem_perm
.key
= key
;
462 sma
->sem_perm
.security
= NULL
;
463 retval
= security_sem_alloc(sma
);
465 ipc_rcu_putref(sma
, ipc_rcu_free
);
469 id
= ipc_addid(&sem_ids(ns
), &sma
->sem_perm
, ns
->sc_semmni
);
471 ipc_rcu_putref(sma
, sem_rcu_free
);
474 ns
->used_sems
+= nsems
;
476 sma
->sem_base
= (struct sem
*) &sma
[1];
478 for (i
= 0; i
< nsems
; i
++) {
479 INIT_LIST_HEAD(&sma
->sem_base
[i
].pending_alter
);
480 INIT_LIST_HEAD(&sma
->sem_base
[i
].pending_const
);
481 spin_lock_init(&sma
->sem_base
[i
].lock
);
484 sma
->complex_count
= 0;
485 INIT_LIST_HEAD(&sma
->pending_alter
);
486 INIT_LIST_HEAD(&sma
->pending_const
);
487 INIT_LIST_HEAD(&sma
->list_id
);
488 sma
->sem_nsems
= nsems
;
489 sma
->sem_ctime
= get_seconds();
493 return sma
->sem_perm
.id
;
498 * Called with sem_ids.rwsem and ipcp locked.
500 static inline int sem_security(struct kern_ipc_perm
*ipcp
, int semflg
)
502 struct sem_array
*sma
;
504 sma
= container_of(ipcp
, struct sem_array
, sem_perm
);
505 return security_sem_associate(sma
, semflg
);
509 * Called with sem_ids.rwsem and ipcp locked.
511 static inline int sem_more_checks(struct kern_ipc_perm
*ipcp
,
512 struct ipc_params
*params
)
514 struct sem_array
*sma
;
516 sma
= container_of(ipcp
, struct sem_array
, sem_perm
);
517 if (params
->u
.nsems
> sma
->sem_nsems
)
523 SYSCALL_DEFINE3(semget
, key_t
, key
, int, nsems
, int, semflg
)
525 struct ipc_namespace
*ns
;
526 struct ipc_ops sem_ops
;
527 struct ipc_params sem_params
;
529 ns
= current
->nsproxy
->ipc_ns
;
531 if (nsems
< 0 || nsems
> ns
->sc_semmsl
)
534 sem_ops
.getnew
= newary
;
535 sem_ops
.associate
= sem_security
;
536 sem_ops
.more_checks
= sem_more_checks
;
538 sem_params
.key
= key
;
539 sem_params
.flg
= semflg
;
540 sem_params
.u
.nsems
= nsems
;
542 return ipcget(ns
, &sem_ids(ns
), &sem_ops
, &sem_params
);
545 /** perform_atomic_semop - Perform (if possible) a semaphore operation
546 * @sma: semaphore array
547 * @sops: array with operations that should be checked
548 * @nsems: number of sops
550 * @pid: pid that did the change
552 * Returns 0 if the operation was possible.
553 * Returns 1 if the operation is impossible, the caller must sleep.
554 * Negative values are error codes.
557 static int perform_atomic_semop(struct sem_array
*sma
, struct sembuf
*sops
,
558 int nsops
, struct sem_undo
*un
, int pid
)
564 for (sop
= sops
; sop
< sops
+ nsops
; sop
++) {
565 curr
= sma
->sem_base
+ sop
->sem_num
;
566 sem_op
= sop
->sem_op
;
567 result
= curr
->semval
;
569 if (!sem_op
&& result
)
577 if (sop
->sem_flg
& SEM_UNDO
) {
578 int undo
= un
->semadj
[sop
->sem_num
] - sem_op
;
580 * Exceeding the undo range is an error.
582 if (undo
< (-SEMAEM
- 1) || undo
> SEMAEM
)
585 curr
->semval
= result
;
589 while (sop
>= sops
) {
590 sma
->sem_base
[sop
->sem_num
].sempid
= pid
;
591 if (sop
->sem_flg
& SEM_UNDO
)
592 un
->semadj
[sop
->sem_num
] -= sop
->sem_op
;
603 if (sop
->sem_flg
& IPC_NOWAIT
)
610 while (sop
>= sops
) {
611 sma
->sem_base
[sop
->sem_num
].semval
-= sop
->sem_op
;
618 /** wake_up_sem_queue_prepare(q, error): Prepare wake-up
619 * @q: queue entry that must be signaled
620 * @error: Error value for the signal
622 * Prepare the wake-up of the queue entry q.
624 static void wake_up_sem_queue_prepare(struct list_head
*pt
,
625 struct sem_queue
*q
, int error
)
627 if (list_empty(pt
)) {
629 * Hold preempt off so that we don't get preempted and have the
630 * wakee busy-wait until we're scheduled back on.
634 q
->status
= IN_WAKEUP
;
637 list_add_tail(&q
->list
, pt
);
641 * wake_up_sem_queue_do(pt) - do the actual wake-up
642 * @pt: list of tasks to be woken up
644 * Do the actual wake-up.
645 * The function is called without any locks held, thus the semaphore array
646 * could be destroyed already and the tasks can disappear as soon as the
647 * status is set to the actual return code.
649 static void wake_up_sem_queue_do(struct list_head
*pt
)
651 struct sem_queue
*q
, *t
;
654 did_something
= !list_empty(pt
);
655 list_for_each_entry_safe(q
, t
, pt
, list
) {
656 wake_up_process(q
->sleeper
);
657 /* q can disappear immediately after writing q->status. */
665 static void unlink_queue(struct sem_array
*sma
, struct sem_queue
*q
)
669 sma
->complex_count
--;
672 /** check_restart(sma, q)
673 * @sma: semaphore array
674 * @q: the operation that just completed
676 * update_queue is O(N^2) when it restarts scanning the whole queue of
677 * waiting operations. Therefore this function checks if the restart is
678 * really necessary. It is called after a previously waiting operation
679 * modified the array.
680 * Note that wait-for-zero operations are handled without restart.
682 static int check_restart(struct sem_array
*sma
, struct sem_queue
*q
)
684 /* pending complex alter operations are too difficult to analyse */
685 if (!list_empty(&sma
->pending_alter
))
688 /* we were a sleeping complex operation. Too difficult */
692 /* It is impossible that someone waits for the new value:
693 * - complex operations always restart.
694 * - wait-for-zero are handled seperately.
695 * - q is a previously sleeping simple operation that
696 * altered the array. It must be a decrement, because
697 * simple increments never sleep.
698 * - If there are older (higher priority) decrements
699 * in the queue, then they have observed the original
700 * semval value and couldn't proceed. The operation
701 * decremented to value - thus they won't proceed either.
707 * wake_const_ops(sma, semnum, pt) - Wake up non-alter tasks
708 * @sma: semaphore array.
709 * @semnum: semaphore that was modified.
710 * @pt: list head for the tasks that must be woken up.
712 * wake_const_ops must be called after a semaphore in a semaphore array
713 * was set to 0. If complex const operations are pending, wake_const_ops must
714 * be called with semnum = -1, as well as with the number of each modified
716 * The tasks that must be woken up are added to @pt. The return code
717 * is stored in q->pid.
718 * The function returns 1 if at least one operation was completed successfully.
720 static int wake_const_ops(struct sem_array
*sma
, int semnum
,
721 struct list_head
*pt
)
724 struct list_head
*walk
;
725 struct list_head
*pending_list
;
726 int semop_completed
= 0;
729 pending_list
= &sma
->pending_const
;
731 pending_list
= &sma
->sem_base
[semnum
].pending_const
;
733 walk
= pending_list
->next
;
734 while (walk
!= pending_list
) {
737 q
= container_of(walk
, struct sem_queue
, list
);
740 error
= perform_atomic_semop(sma
, q
->sops
, q
->nsops
,
744 /* operation completed, remove from queue & wakeup */
746 unlink_queue(sma
, q
);
748 wake_up_sem_queue_prepare(pt
, q
, error
);
753 return semop_completed
;
757 * do_smart_wakeup_zero(sma, sops, nsops, pt) - wakeup all wait for zero tasks
758 * @sma: semaphore array
759 * @sops: operations that were performed
760 * @nsops: number of operations
761 * @pt: list head of the tasks that must be woken up.
763 * do_smart_wakeup_zero() checks all required queue for wait-for-zero
764 * operations, based on the actual changes that were performed on the
766 * The function returns 1 if at least one operation was completed successfully.
768 static int do_smart_wakeup_zero(struct sem_array
*sma
, struct sembuf
*sops
,
769 int nsops
, struct list_head
*pt
)
772 int semop_completed
= 0;
775 /* first: the per-semaphore queues, if known */
777 for (i
= 0; i
< nsops
; i
++) {
778 int num
= sops
[i
].sem_num
;
780 if (sma
->sem_base
[num
].semval
== 0) {
782 semop_completed
|= wake_const_ops(sma
, num
, pt
);
787 * No sops means modified semaphores not known.
788 * Assume all were changed.
790 for (i
= 0; i
< sma
->sem_nsems
; i
++) {
791 if (sma
->sem_base
[i
].semval
== 0) {
793 semop_completed
|= wake_const_ops(sma
, i
, pt
);
798 * If one of the modified semaphores got 0,
799 * then check the global queue, too.
802 semop_completed
|= wake_const_ops(sma
, -1, pt
);
804 return semop_completed
;
809 * update_queue(sma, semnum): Look for tasks that can be completed.
810 * @sma: semaphore array.
811 * @semnum: semaphore that was modified.
812 * @pt: list head for the tasks that must be woken up.
814 * update_queue must be called after a semaphore in a semaphore array
815 * was modified. If multiple semaphores were modified, update_queue must
816 * be called with semnum = -1, as well as with the number of each modified
818 * The tasks that must be woken up are added to @pt. The return code
819 * is stored in q->pid.
820 * The function internally checks if const operations can now succeed.
822 * The function return 1 if at least one semop was completed successfully.
824 static int update_queue(struct sem_array
*sma
, int semnum
, struct list_head
*pt
)
827 struct list_head
*walk
;
828 struct list_head
*pending_list
;
829 int semop_completed
= 0;
832 pending_list
= &sma
->pending_alter
;
834 pending_list
= &sma
->sem_base
[semnum
].pending_alter
;
837 walk
= pending_list
->next
;
838 while (walk
!= pending_list
) {
841 q
= container_of(walk
, struct sem_queue
, list
);
844 /* If we are scanning the single sop, per-semaphore list of
845 * one semaphore and that semaphore is 0, then it is not
846 * necessary to scan further: simple increments
847 * that affect only one entry succeed immediately and cannot
848 * be in the per semaphore pending queue, and decrements
849 * cannot be successful if the value is already 0.
851 if (semnum
!= -1 && sma
->sem_base
[semnum
].semval
== 0)
854 error
= perform_atomic_semop(sma
, q
->sops
, q
->nsops
,
857 /* Does q->sleeper still need to sleep? */
861 unlink_queue(sma
, q
);
867 do_smart_wakeup_zero(sma
, q
->sops
, q
->nsops
, pt
);
868 restart
= check_restart(sma
, q
);
871 wake_up_sem_queue_prepare(pt
, q
, error
);
875 return semop_completed
;
879 * do_smart_update(sma, sops, nsops, otime, pt) - optimized update_queue
880 * @sma: semaphore array
881 * @sops: operations that were performed
882 * @nsops: number of operations
883 * @otime: force setting otime
884 * @pt: list head of the tasks that must be woken up.
886 * do_smart_update() does the required calls to update_queue and wakeup_zero,
887 * based on the actual changes that were performed on the semaphore array.
888 * Note that the function does not do the actual wake-up: the caller is
889 * responsible for calling wake_up_sem_queue_do(@pt).
890 * It is safe to perform this call after dropping all locks.
892 static void do_smart_update(struct sem_array
*sma
, struct sembuf
*sops
, int nsops
,
893 int otime
, struct list_head
*pt
)
897 otime
|= do_smart_wakeup_zero(sma
, sops
, nsops
, pt
);
899 if (!list_empty(&sma
->pending_alter
)) {
900 /* semaphore array uses the global queue - just process it. */
901 otime
|= update_queue(sma
, -1, pt
);
905 * No sops, thus the modified semaphores are not
908 for (i
= 0; i
< sma
->sem_nsems
; i
++)
909 otime
|= update_queue(sma
, i
, pt
);
912 * Check the semaphores that were increased:
913 * - No complex ops, thus all sleeping ops are
915 * - if we decreased the value, then any sleeping
916 * semaphore ops wont be able to run: If the
917 * previous value was too small, then the new
918 * value will be too small, too.
920 for (i
= 0; i
< nsops
; i
++) {
921 if (sops
[i
].sem_op
> 0) {
922 otime
|= update_queue(sma
,
923 sops
[i
].sem_num
, pt
);
930 sma
->sem_base
[0].sem_otime
= get_seconds();
932 sma
->sem_base
[sops
[0].sem_num
].sem_otime
=
939 /* The following counts are associated to each semaphore:
940 * semncnt number of tasks waiting on semval being nonzero
941 * semzcnt number of tasks waiting on semval being zero
942 * This model assumes that a task waits on exactly one semaphore.
943 * Since semaphore operations are to be performed atomically, tasks actually
944 * wait on a whole sequence of semaphores simultaneously.
945 * The counts we return here are a rough approximation, but still
946 * warrant that semncnt+semzcnt>0 if the task is on the pending queue.
948 static int count_semncnt (struct sem_array
* sma
, ushort semnum
)
951 struct sem_queue
* q
;
954 list_for_each_entry(q
, &sma
->sem_base
[semnum
].pending_alter
, list
) {
955 struct sembuf
* sops
= q
->sops
;
956 BUG_ON(sops
->sem_num
!= semnum
);
957 if ((sops
->sem_op
< 0) && !(sops
->sem_flg
& IPC_NOWAIT
))
961 list_for_each_entry(q
, &sma
->pending_alter
, list
) {
962 struct sembuf
* sops
= q
->sops
;
963 int nsops
= q
->nsops
;
965 for (i
= 0; i
< nsops
; i
++)
966 if (sops
[i
].sem_num
== semnum
967 && (sops
[i
].sem_op
< 0)
968 && !(sops
[i
].sem_flg
& IPC_NOWAIT
))
974 static int count_semzcnt (struct sem_array
* sma
, ushort semnum
)
977 struct sem_queue
* q
;
980 list_for_each_entry(q
, &sma
->sem_base
[semnum
].pending_const
, list
) {
981 struct sembuf
* sops
= q
->sops
;
982 BUG_ON(sops
->sem_num
!= semnum
);
983 if ((sops
->sem_op
== 0) && !(sops
->sem_flg
& IPC_NOWAIT
))
987 list_for_each_entry(q
, &sma
->pending_const
, list
) {
988 struct sembuf
* sops
= q
->sops
;
989 int nsops
= q
->nsops
;
991 for (i
= 0; i
< nsops
; i
++)
992 if (sops
[i
].sem_num
== semnum
993 && (sops
[i
].sem_op
== 0)
994 && !(sops
[i
].sem_flg
& IPC_NOWAIT
))
1000 /* Free a semaphore set. freeary() is called with sem_ids.rwsem locked
1001 * as a writer and the spinlock for this semaphore set hold. sem_ids.rwsem
1002 * remains locked on exit.
1004 static void freeary(struct ipc_namespace
*ns
, struct kern_ipc_perm
*ipcp
)
1006 struct sem_undo
*un
, *tu
;
1007 struct sem_queue
*q
, *tq
;
1008 struct sem_array
*sma
= container_of(ipcp
, struct sem_array
, sem_perm
);
1009 struct list_head tasks
;
1012 /* Free the existing undo structures for this semaphore set. */
1013 ipc_assert_locked_object(&sma
->sem_perm
);
1014 list_for_each_entry_safe(un
, tu
, &sma
->list_id
, list_id
) {
1015 list_del(&un
->list_id
);
1016 spin_lock(&un
->ulp
->lock
);
1018 list_del_rcu(&un
->list_proc
);
1019 spin_unlock(&un
->ulp
->lock
);
1023 /* Wake up all pending processes and let them fail with EIDRM. */
1024 INIT_LIST_HEAD(&tasks
);
1025 list_for_each_entry_safe(q
, tq
, &sma
->pending_const
, list
) {
1026 unlink_queue(sma
, q
);
1027 wake_up_sem_queue_prepare(&tasks
, q
, -EIDRM
);
1030 list_for_each_entry_safe(q
, tq
, &sma
->pending_alter
, list
) {
1031 unlink_queue(sma
, q
);
1032 wake_up_sem_queue_prepare(&tasks
, q
, -EIDRM
);
1034 for (i
= 0; i
< sma
->sem_nsems
; i
++) {
1035 struct sem
*sem
= sma
->sem_base
+ i
;
1036 list_for_each_entry_safe(q
, tq
, &sem
->pending_const
, list
) {
1037 unlink_queue(sma
, q
);
1038 wake_up_sem_queue_prepare(&tasks
, q
, -EIDRM
);
1040 list_for_each_entry_safe(q
, tq
, &sem
->pending_alter
, list
) {
1041 unlink_queue(sma
, q
);
1042 wake_up_sem_queue_prepare(&tasks
, q
, -EIDRM
);
1046 /* Remove the semaphore set from the IDR */
1048 sem_unlock(sma
, -1);
1051 wake_up_sem_queue_do(&tasks
);
1052 ns
->used_sems
-= sma
->sem_nsems
;
1053 ipc_rcu_putref(sma
, sem_rcu_free
);
1056 static unsigned long copy_semid_to_user(void __user
*buf
, struct semid64_ds
*in
, int version
)
1060 return copy_to_user(buf
, in
, sizeof(*in
));
1063 struct semid_ds out
;
1065 memset(&out
, 0, sizeof(out
));
1067 ipc64_perm_to_ipc_perm(&in
->sem_perm
, &out
.sem_perm
);
1069 out
.sem_otime
= in
->sem_otime
;
1070 out
.sem_ctime
= in
->sem_ctime
;
1071 out
.sem_nsems
= in
->sem_nsems
;
1073 return copy_to_user(buf
, &out
, sizeof(out
));
1080 static time_t get_semotime(struct sem_array
*sma
)
1085 res
= sma
->sem_base
[0].sem_otime
;
1086 for (i
= 1; i
< sma
->sem_nsems
; i
++) {
1087 time_t to
= sma
->sem_base
[i
].sem_otime
;
1095 static int semctl_nolock(struct ipc_namespace
*ns
, int semid
,
1096 int cmd
, int version
, void __user
*p
)
1099 struct sem_array
*sma
;
1105 struct seminfo seminfo
;
1108 err
= security_sem_semctl(NULL
, cmd
);
1112 memset(&seminfo
,0,sizeof(seminfo
));
1113 seminfo
.semmni
= ns
->sc_semmni
;
1114 seminfo
.semmns
= ns
->sc_semmns
;
1115 seminfo
.semmsl
= ns
->sc_semmsl
;
1116 seminfo
.semopm
= ns
->sc_semopm
;
1117 seminfo
.semvmx
= SEMVMX
;
1118 seminfo
.semmnu
= SEMMNU
;
1119 seminfo
.semmap
= SEMMAP
;
1120 seminfo
.semume
= SEMUME
;
1121 down_read(&sem_ids(ns
).rwsem
);
1122 if (cmd
== SEM_INFO
) {
1123 seminfo
.semusz
= sem_ids(ns
).in_use
;
1124 seminfo
.semaem
= ns
->used_sems
;
1126 seminfo
.semusz
= SEMUSZ
;
1127 seminfo
.semaem
= SEMAEM
;
1129 max_id
= ipc_get_maxid(&sem_ids(ns
));
1130 up_read(&sem_ids(ns
).rwsem
);
1131 if (copy_to_user(p
, &seminfo
, sizeof(struct seminfo
)))
1133 return (max_id
< 0) ? 0: max_id
;
1138 struct semid64_ds tbuf
;
1141 memset(&tbuf
, 0, sizeof(tbuf
));
1144 if (cmd
== SEM_STAT
) {
1145 sma
= sem_obtain_object(ns
, semid
);
1150 id
= sma
->sem_perm
.id
;
1152 sma
= sem_obtain_object_check(ns
, semid
);
1160 if (ipcperms(ns
, &sma
->sem_perm
, S_IRUGO
))
1163 err
= security_sem_semctl(sma
, cmd
);
1167 kernel_to_ipc64_perm(&sma
->sem_perm
, &tbuf
.sem_perm
);
1168 tbuf
.sem_otime
= get_semotime(sma
);
1169 tbuf
.sem_ctime
= sma
->sem_ctime
;
1170 tbuf
.sem_nsems
= sma
->sem_nsems
;
1172 if (copy_semid_to_user(p
, &tbuf
, version
))
1184 static int semctl_setval(struct ipc_namespace
*ns
, int semid
, int semnum
,
1187 struct sem_undo
*un
;
1188 struct sem_array
*sma
;
1191 struct list_head tasks
;
1193 #if defined(CONFIG_64BIT) && defined(__BIG_ENDIAN)
1194 /* big-endian 64bit */
1197 /* 32bit or little-endian 64bit */
1201 if (val
> SEMVMX
|| val
< 0)
1204 INIT_LIST_HEAD(&tasks
);
1207 sma
= sem_obtain_object_check(ns
, semid
);
1210 return PTR_ERR(sma
);
1213 if (semnum
< 0 || semnum
>= sma
->sem_nsems
) {
1219 if (ipcperms(ns
, &sma
->sem_perm
, S_IWUGO
)) {
1224 err
= security_sem_semctl(sma
, SETVAL
);
1230 sem_lock(sma
, NULL
, -1);
1232 curr
= &sma
->sem_base
[semnum
];
1234 ipc_assert_locked_object(&sma
->sem_perm
);
1235 list_for_each_entry(un
, &sma
->list_id
, list_id
)
1236 un
->semadj
[semnum
] = 0;
1239 curr
->sempid
= task_tgid_vnr(current
);
1240 sma
->sem_ctime
= get_seconds();
1241 /* maybe some queued-up processes were waiting for this */
1242 do_smart_update(sma
, NULL
, 0, 0, &tasks
);
1243 sem_unlock(sma
, -1);
1245 wake_up_sem_queue_do(&tasks
);
1249 static int semctl_main(struct ipc_namespace
*ns
, int semid
, int semnum
,
1250 int cmd
, void __user
*p
)
1252 struct sem_array
*sma
;
1255 ushort fast_sem_io
[SEMMSL_FAST
];
1256 ushort
* sem_io
= fast_sem_io
;
1257 struct list_head tasks
;
1259 INIT_LIST_HEAD(&tasks
);
1262 sma
= sem_obtain_object_check(ns
, semid
);
1265 return PTR_ERR(sma
);
1268 nsems
= sma
->sem_nsems
;
1271 if (ipcperms(ns
, &sma
->sem_perm
, cmd
== SETALL
? S_IWUGO
: S_IRUGO
))
1272 goto out_rcu_wakeup
;
1274 err
= security_sem_semctl(sma
, cmd
);
1276 goto out_rcu_wakeup
;
1282 ushort __user
*array
= p
;
1285 sem_lock(sma
, NULL
, -1);
1286 if(nsems
> SEMMSL_FAST
) {
1287 if (!ipc_rcu_getref(sma
)) {
1288 sem_unlock(sma
, -1);
1293 sem_unlock(sma
, -1);
1295 sem_io
= ipc_alloc(sizeof(ushort
)*nsems
);
1296 if(sem_io
== NULL
) {
1297 ipc_rcu_putref(sma
, ipc_rcu_free
);
1302 sem_lock_and_putref(sma
);
1303 if (sma
->sem_perm
.deleted
) {
1304 sem_unlock(sma
, -1);
1310 for (i
= 0; i
< sma
->sem_nsems
; i
++)
1311 sem_io
[i
] = sma
->sem_base
[i
].semval
;
1312 sem_unlock(sma
, -1);
1315 if(copy_to_user(array
, sem_io
, nsems
*sizeof(ushort
)))
1322 struct sem_undo
*un
;
1324 if (!ipc_rcu_getref(sma
)) {
1330 if(nsems
> SEMMSL_FAST
) {
1331 sem_io
= ipc_alloc(sizeof(ushort
)*nsems
);
1332 if(sem_io
== NULL
) {
1333 ipc_rcu_putref(sma
, ipc_rcu_free
);
1338 if (copy_from_user (sem_io
, p
, nsems
*sizeof(ushort
))) {
1339 ipc_rcu_putref(sma
, ipc_rcu_free
);
1344 for (i
= 0; i
< nsems
; i
++) {
1345 if (sem_io
[i
] > SEMVMX
) {
1346 ipc_rcu_putref(sma
, ipc_rcu_free
);
1352 sem_lock_and_putref(sma
);
1353 if (sma
->sem_perm
.deleted
) {
1354 sem_unlock(sma
, -1);
1360 for (i
= 0; i
< nsems
; i
++)
1361 sma
->sem_base
[i
].semval
= sem_io
[i
];
1363 ipc_assert_locked_object(&sma
->sem_perm
);
1364 list_for_each_entry(un
, &sma
->list_id
, list_id
) {
1365 for (i
= 0; i
< nsems
; i
++)
1368 sma
->sem_ctime
= get_seconds();
1369 /* maybe some queued-up processes were waiting for this */
1370 do_smart_update(sma
, NULL
, 0, 0, &tasks
);
1374 /* GETVAL, GETPID, GETNCTN, GETZCNT: fall-through */
1377 if (semnum
< 0 || semnum
>= nsems
)
1378 goto out_rcu_wakeup
;
1380 sem_lock(sma
, NULL
, -1);
1381 curr
= &sma
->sem_base
[semnum
];
1391 err
= count_semncnt(sma
,semnum
);
1394 err
= count_semzcnt(sma
,semnum
);
1399 sem_unlock(sma
, -1);
1402 wake_up_sem_queue_do(&tasks
);
1404 if(sem_io
!= fast_sem_io
)
1405 ipc_free(sem_io
, sizeof(ushort
)*nsems
);
1409 static inline unsigned long
1410 copy_semid_from_user(struct semid64_ds
*out
, void __user
*buf
, int version
)
1414 if (copy_from_user(out
, buf
, sizeof(*out
)))
1419 struct semid_ds tbuf_old
;
1421 if(copy_from_user(&tbuf_old
, buf
, sizeof(tbuf_old
)))
1424 out
->sem_perm
.uid
= tbuf_old
.sem_perm
.uid
;
1425 out
->sem_perm
.gid
= tbuf_old
.sem_perm
.gid
;
1426 out
->sem_perm
.mode
= tbuf_old
.sem_perm
.mode
;
1436 * This function handles some semctl commands which require the rwsem
1437 * to be held in write mode.
1438 * NOTE: no locks must be held, the rwsem is taken inside this function.
1440 static int semctl_down(struct ipc_namespace
*ns
, int semid
,
1441 int cmd
, int version
, void __user
*p
)
1443 struct sem_array
*sma
;
1445 struct semid64_ds semid64
;
1446 struct kern_ipc_perm
*ipcp
;
1448 if(cmd
== IPC_SET
) {
1449 if (copy_semid_from_user(&semid64
, p
, version
))
1453 down_write(&sem_ids(ns
).rwsem
);
1456 ipcp
= ipcctl_pre_down_nolock(ns
, &sem_ids(ns
), semid
, cmd
,
1457 &semid64
.sem_perm
, 0);
1459 err
= PTR_ERR(ipcp
);
1463 sma
= container_of(ipcp
, struct sem_array
, sem_perm
);
1465 err
= security_sem_semctl(sma
, cmd
);
1471 sem_lock(sma
, NULL
, -1);
1472 /* freeary unlocks the ipc object and rcu */
1476 sem_lock(sma
, NULL
, -1);
1477 err
= ipc_update_perm(&semid64
.sem_perm
, ipcp
);
1480 sma
->sem_ctime
= get_seconds();
1488 sem_unlock(sma
, -1);
1492 up_write(&sem_ids(ns
).rwsem
);
1496 SYSCALL_DEFINE4(semctl
, int, semid
, int, semnum
, int, cmd
, unsigned long, arg
)
1499 struct ipc_namespace
*ns
;
1500 void __user
*p
= (void __user
*)arg
;
1505 version
= ipc_parse_version(&cmd
);
1506 ns
= current
->nsproxy
->ipc_ns
;
1513 return semctl_nolock(ns
, semid
, cmd
, version
, p
);
1520 return semctl_main(ns
, semid
, semnum
, cmd
, p
);
1522 return semctl_setval(ns
, semid
, semnum
, arg
);
1525 return semctl_down(ns
, semid
, cmd
, version
, p
);
1531 /* If the task doesn't already have a undo_list, then allocate one
1532 * here. We guarantee there is only one thread using this undo list,
1533 * and current is THE ONE
1535 * If this allocation and assignment succeeds, but later
1536 * portions of this code fail, there is no need to free the sem_undo_list.
1537 * Just let it stay associated with the task, and it'll be freed later
1540 * This can block, so callers must hold no locks.
1542 static inline int get_undo_list(struct sem_undo_list
**undo_listp
)
1544 struct sem_undo_list
*undo_list
;
1546 undo_list
= current
->sysvsem
.undo_list
;
1548 undo_list
= kzalloc(sizeof(*undo_list
), GFP_KERNEL
);
1549 if (undo_list
== NULL
)
1551 spin_lock_init(&undo_list
->lock
);
1552 atomic_set(&undo_list
->refcnt
, 1);
1553 INIT_LIST_HEAD(&undo_list
->list_proc
);
1555 current
->sysvsem
.undo_list
= undo_list
;
1557 *undo_listp
= undo_list
;
1561 static struct sem_undo
*__lookup_undo(struct sem_undo_list
*ulp
, int semid
)
1563 struct sem_undo
*un
;
1565 list_for_each_entry_rcu(un
, &ulp
->list_proc
, list_proc
) {
1566 if (un
->semid
== semid
)
1572 static struct sem_undo
*lookup_undo(struct sem_undo_list
*ulp
, int semid
)
1574 struct sem_undo
*un
;
1576 assert_spin_locked(&ulp
->lock
);
1578 un
= __lookup_undo(ulp
, semid
);
1580 list_del_rcu(&un
->list_proc
);
1581 list_add_rcu(&un
->list_proc
, &ulp
->list_proc
);
1587 * find_alloc_undo - Lookup (and if not present create) undo array
1589 * @semid: semaphore array id
1591 * The function looks up (and if not present creates) the undo structure.
1592 * The size of the undo structure depends on the size of the semaphore
1593 * array, thus the alloc path is not that straightforward.
1594 * Lifetime-rules: sem_undo is rcu-protected, on success, the function
1595 * performs a rcu_read_lock().
1597 static struct sem_undo
*find_alloc_undo(struct ipc_namespace
*ns
, int semid
)
1599 struct sem_array
*sma
;
1600 struct sem_undo_list
*ulp
;
1601 struct sem_undo
*un
, *new;
1604 error
= get_undo_list(&ulp
);
1606 return ERR_PTR(error
);
1609 spin_lock(&ulp
->lock
);
1610 un
= lookup_undo(ulp
, semid
);
1611 spin_unlock(&ulp
->lock
);
1612 if (likely(un
!=NULL
))
1615 /* no undo structure around - allocate one. */
1616 /* step 1: figure out the size of the semaphore array */
1617 sma
= sem_obtain_object_check(ns
, semid
);
1620 return ERR_CAST(sma
);
1623 nsems
= sma
->sem_nsems
;
1624 if (!ipc_rcu_getref(sma
)) {
1626 un
= ERR_PTR(-EIDRM
);
1631 /* step 2: allocate new undo structure */
1632 new = kzalloc(sizeof(struct sem_undo
) + sizeof(short)*nsems
, GFP_KERNEL
);
1634 ipc_rcu_putref(sma
, ipc_rcu_free
);
1635 return ERR_PTR(-ENOMEM
);
1638 /* step 3: Acquire the lock on semaphore array */
1640 sem_lock_and_putref(sma
);
1641 if (sma
->sem_perm
.deleted
) {
1642 sem_unlock(sma
, -1);
1645 un
= ERR_PTR(-EIDRM
);
1648 spin_lock(&ulp
->lock
);
1651 * step 4: check for races: did someone else allocate the undo struct?
1653 un
= lookup_undo(ulp
, semid
);
1658 /* step 5: initialize & link new undo structure */
1659 new->semadj
= (short *) &new[1];
1662 assert_spin_locked(&ulp
->lock
);
1663 list_add_rcu(&new->list_proc
, &ulp
->list_proc
);
1664 ipc_assert_locked_object(&sma
->sem_perm
);
1665 list_add(&new->list_id
, &sma
->list_id
);
1669 spin_unlock(&ulp
->lock
);
1670 sem_unlock(sma
, -1);
1677 * get_queue_result - Retrieve the result code from sem_queue
1678 * @q: Pointer to queue structure
1680 * Retrieve the return code from the pending queue. If IN_WAKEUP is found in
1681 * q->status, then we must loop until the value is replaced with the final
1682 * value: This may happen if a task is woken up by an unrelated event (e.g.
1683 * signal) and in parallel the task is woken up by another task because it got
1684 * the requested semaphores.
1686 * The function can be called with or without holding the semaphore spinlock.
1688 static int get_queue_result(struct sem_queue
*q
)
1693 while (unlikely(error
== IN_WAKEUP
)) {
1701 SYSCALL_DEFINE4(semtimedop
, int, semid
, struct sembuf __user
*, tsops
,
1702 unsigned, nsops
, const struct timespec __user
*, timeout
)
1704 int error
= -EINVAL
;
1705 struct sem_array
*sma
;
1706 struct sembuf fast_sops
[SEMOPM_FAST
];
1707 struct sembuf
* sops
= fast_sops
, *sop
;
1708 struct sem_undo
*un
;
1709 int undos
= 0, alter
= 0, max
, locknum
;
1710 struct sem_queue queue
;
1711 unsigned long jiffies_left
= 0;
1712 struct ipc_namespace
*ns
;
1713 struct list_head tasks
;
1715 ns
= current
->nsproxy
->ipc_ns
;
1717 if (nsops
< 1 || semid
< 0)
1719 if (nsops
> ns
->sc_semopm
)
1721 if(nsops
> SEMOPM_FAST
) {
1722 sops
= kmalloc(sizeof(*sops
)*nsops
,GFP_KERNEL
);
1726 if (copy_from_user (sops
, tsops
, nsops
* sizeof(*tsops
))) {
1731 struct timespec _timeout
;
1732 if (copy_from_user(&_timeout
, timeout
, sizeof(*timeout
))) {
1736 if (_timeout
.tv_sec
< 0 || _timeout
.tv_nsec
< 0 ||
1737 _timeout
.tv_nsec
>= 1000000000L) {
1741 jiffies_left
= timespec_to_jiffies(&_timeout
);
1744 for (sop
= sops
; sop
< sops
+ nsops
; sop
++) {
1745 if (sop
->sem_num
>= max
)
1747 if (sop
->sem_flg
& SEM_UNDO
)
1749 if (sop
->sem_op
!= 0)
1753 INIT_LIST_HEAD(&tasks
);
1756 /* On success, find_alloc_undo takes the rcu_read_lock */
1757 un
= find_alloc_undo(ns
, semid
);
1759 error
= PTR_ERR(un
);
1767 sma
= sem_obtain_object_check(ns
, semid
);
1770 error
= PTR_ERR(sma
);
1775 if (max
>= sma
->sem_nsems
)
1776 goto out_rcu_wakeup
;
1779 if (ipcperms(ns
, &sma
->sem_perm
, alter
? S_IWUGO
: S_IRUGO
))
1780 goto out_rcu_wakeup
;
1782 error
= security_sem_semop(sma
, sops
, nsops
, alter
);
1784 goto out_rcu_wakeup
;
1787 * semid identifiers are not unique - find_alloc_undo may have
1788 * allocated an undo structure, it was invalidated by an RMID
1789 * and now a new array with received the same id. Check and fail.
1790 * This case can be detected checking un->semid. The existence of
1791 * "un" itself is guaranteed by rcu.
1794 locknum
= sem_lock(sma
, sops
, nsops
);
1795 if (un
&& un
->semid
== -1)
1796 goto out_unlock_free
;
1798 error
= perform_atomic_semop(sma
, sops
, nsops
, un
,
1799 task_tgid_vnr(current
));
1801 if (alter
&& error
== 0)
1802 do_smart_update(sma
, sops
, nsops
, 1, &tasks
);
1804 goto out_unlock_free
;
1807 /* We need to sleep on this operation, so we put the current
1808 * task into the pending queue and go to sleep.
1812 queue
.nsops
= nsops
;
1814 queue
.pid
= task_tgid_vnr(current
);
1815 queue
.alter
= alter
;
1819 curr
= &sma
->sem_base
[sops
->sem_num
];
1822 if (sma
->complex_count
) {
1823 list_add_tail(&queue
.list
,
1824 &sma
->pending_alter
);
1827 list_add_tail(&queue
.list
,
1828 &curr
->pending_alter
);
1831 list_add_tail(&queue
.list
, &curr
->pending_const
);
1834 if (!sma
->complex_count
)
1838 list_add_tail(&queue
.list
, &sma
->pending_alter
);
1840 list_add_tail(&queue
.list
, &sma
->pending_const
);
1842 sma
->complex_count
++;
1845 queue
.status
= -EINTR
;
1846 queue
.sleeper
= current
;
1849 current
->state
= TASK_INTERRUPTIBLE
;
1850 sem_unlock(sma
, locknum
);
1854 jiffies_left
= schedule_timeout(jiffies_left
);
1858 error
= get_queue_result(&queue
);
1860 if (error
!= -EINTR
) {
1861 /* fast path: update_queue already obtained all requested
1863 * Perform a smp_mb(): User space could assume that semop()
1864 * is a memory barrier: Without the mb(), the cpu could
1865 * speculatively read in user space stale data that was
1866 * overwritten by the previous owner of the semaphore.
1874 sma
= sem_obtain_lock(ns
, semid
, sops
, nsops
, &locknum
);
1877 * Wait until it's guaranteed that no wakeup_sem_queue_do() is ongoing.
1879 error
= get_queue_result(&queue
);
1882 * Array removed? If yes, leave without sem_unlock().
1891 * If queue.status != -EINTR we are woken up by another process.
1892 * Leave without unlink_queue(), but with sem_unlock().
1895 if (error
!= -EINTR
) {
1896 goto out_unlock_free
;
1900 * If an interrupt occurred we have to clean up the queue
1902 if (timeout
&& jiffies_left
== 0)
1906 * If the wakeup was spurious, just retry
1908 if (error
== -EINTR
&& !signal_pending(current
))
1911 unlink_queue(sma
, &queue
);
1914 sem_unlock(sma
, locknum
);
1917 wake_up_sem_queue_do(&tasks
);
1919 if(sops
!= fast_sops
)
1924 SYSCALL_DEFINE3(semop
, int, semid
, struct sembuf __user
*, tsops
,
1927 return sys_semtimedop(semid
, tsops
, nsops
, NULL
);
1930 /* If CLONE_SYSVSEM is set, establish sharing of SEM_UNDO state between
1931 * parent and child tasks.
1934 int copy_semundo(unsigned long clone_flags
, struct task_struct
*tsk
)
1936 struct sem_undo_list
*undo_list
;
1939 if (clone_flags
& CLONE_SYSVSEM
) {
1940 error
= get_undo_list(&undo_list
);
1943 atomic_inc(&undo_list
->refcnt
);
1944 tsk
->sysvsem
.undo_list
= undo_list
;
1946 tsk
->sysvsem
.undo_list
= NULL
;
1952 * add semadj values to semaphores, free undo structures.
1953 * undo structures are not freed when semaphore arrays are destroyed
1954 * so some of them may be out of date.
1955 * IMPLEMENTATION NOTE: There is some confusion over whether the
1956 * set of adjustments that needs to be done should be done in an atomic
1957 * manner or not. That is, if we are attempting to decrement the semval
1958 * should we queue up and wait until we can do so legally?
1959 * The original implementation attempted to do this (queue and wait).
1960 * The current implementation does not do so. The POSIX standard
1961 * and SVID should be consulted to determine what behavior is mandated.
1963 void exit_sem(struct task_struct
*tsk
)
1965 struct sem_undo_list
*ulp
;
1967 ulp
= tsk
->sysvsem
.undo_list
;
1970 tsk
->sysvsem
.undo_list
= NULL
;
1972 if (!atomic_dec_and_test(&ulp
->refcnt
))
1976 struct sem_array
*sma
;
1977 struct sem_undo
*un
;
1978 struct list_head tasks
;
1982 un
= list_entry_rcu(ulp
->list_proc
.next
,
1983 struct sem_undo
, list_proc
);
1984 if (&un
->list_proc
== &ulp
->list_proc
)
1994 sma
= sem_obtain_object_check(tsk
->nsproxy
->ipc_ns
, un
->semid
);
1995 /* exit_sem raced with IPC_RMID, nothing to do */
2001 sem_lock(sma
, NULL
, -1);
2002 un
= __lookup_undo(ulp
, semid
);
2004 /* exit_sem raced with IPC_RMID+semget() that created
2005 * exactly the same semid. Nothing to do.
2007 sem_unlock(sma
, -1);
2012 /* remove un from the linked lists */
2013 ipc_assert_locked_object(&sma
->sem_perm
);
2014 list_del(&un
->list_id
);
2016 spin_lock(&ulp
->lock
);
2017 list_del_rcu(&un
->list_proc
);
2018 spin_unlock(&ulp
->lock
);
2020 /* perform adjustments registered in un */
2021 for (i
= 0; i
< sma
->sem_nsems
; i
++) {
2022 struct sem
* semaphore
= &sma
->sem_base
[i
];
2023 if (un
->semadj
[i
]) {
2024 semaphore
->semval
+= un
->semadj
[i
];
2026 * Range checks of the new semaphore value,
2027 * not defined by sus:
2028 * - Some unices ignore the undo entirely
2029 * (e.g. HP UX 11i 11.22, Tru64 V5.1)
2030 * - some cap the value (e.g. FreeBSD caps
2031 * at 0, but doesn't enforce SEMVMX)
2033 * Linux caps the semaphore value, both at 0
2036 * Manfred <manfred@colorfullife.com>
2038 if (semaphore
->semval
< 0)
2039 semaphore
->semval
= 0;
2040 if (semaphore
->semval
> SEMVMX
)
2041 semaphore
->semval
= SEMVMX
;
2042 semaphore
->sempid
= task_tgid_vnr(current
);
2045 /* maybe some queued-up processes were waiting for this */
2046 INIT_LIST_HEAD(&tasks
);
2047 do_smart_update(sma
, NULL
, 0, 1, &tasks
);
2048 sem_unlock(sma
, -1);
2050 wake_up_sem_queue_do(&tasks
);
2057 #ifdef CONFIG_PROC_FS
2058 static int sysvipc_sem_proc_show(struct seq_file
*s
, void *it
)
2060 struct user_namespace
*user_ns
= seq_user_ns(s
);
2061 struct sem_array
*sma
= it
;
2064 sem_otime
= get_semotime(sma
);
2066 return seq_printf(s
,
2067 "%10d %10d %4o %10u %5u %5u %5u %5u %10lu %10lu\n",
2072 from_kuid_munged(user_ns
, sma
->sem_perm
.uid
),
2073 from_kgid_munged(user_ns
, sma
->sem_perm
.gid
),
2074 from_kuid_munged(user_ns
, sma
->sem_perm
.cuid
),
2075 from_kgid_munged(user_ns
, sma
->sem_perm
.cgid
),