sem.c 53.7 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8
/*
 * linux/ipc/sem.c
 * Copyright (C) 1992 Krishna Balasubramanian
 * Copyright (C) 1995 Eric Schenk, Bruno Haible
 *
 * /proc/sysvipc/sem support (c) 1999 Dragos Acostachioaie <dragos@iname.com>
 *
 * SMP-threaded, sysctl's added
9
 * (c) 1999 Manfred Spraul <manfred@colorfullife.com>
L
Linus Torvalds 已提交
10
 * Enforced range limit on SEM_UNDO
A
Alan Cox 已提交
11
 * (c) 2001 Red Hat Inc
L
Linus Torvalds 已提交
12 13
 * Lockless wakeup
 * (c) 2003 Manfred Spraul <manfred@colorfullife.com>
14 15
 * Further wakeup optimizations, documentation
 * (c) 2010 Manfred Spraul <manfred@colorfullife.com>
S
Steve Grubb 已提交
16 17 18
 *
 * support for audit of ipc object properties and permission changes
 * Dustin Kirkland <dustin.kirkland@us.ibm.com>
K
Kirill Korotaev 已提交
19 20 21 22
 *
 * namespaces support
 * OpenVZ, SWsoft Inc.
 * Pavel Emelianov <xemul@openvz.org>
23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73
 *
 * Implementation notes: (May 2010)
 * This file implements System V semaphores.
 *
 * User space visible behavior:
 * - FIFO ordering for semop() operations (just FIFO, not starvation
 *   protection)
 * - multiple semaphore operations that alter the same semaphore in
 *   one semop() are handled.
 * - sem_ctime (time of last semctl()) is updated in the IPC_SET, SETVAL and
 *   SETALL calls.
 * - two Linux specific semctl() commands: SEM_STAT, SEM_INFO.
 * - undo adjustments at process exit are limited to 0..SEMVMX.
 * - namespace are supported.
 * - SEMMSL, SEMMNS, SEMOPM and SEMMNI can be configured at runtine by writing
 *   to /proc/sys/kernel/sem.
 * - statistics about the usage are reported in /proc/sysvipc/sem.
 *
 * Internals:
 * - scalability:
 *   - all global variables are read-mostly.
 *   - semop() calls and semctl(RMID) are synchronized by RCU.
 *   - most operations do write operations (actually: spin_lock calls) to
 *     the per-semaphore array structure.
 *   Thus: Perfect SMP scaling between independent semaphore arrays.
 *         If multiple semaphores in one array are used, then cache line
 *         trashing on the semaphore array spinlock will limit the scaling.
 * - semncnt and semzcnt are calculated on demand in count_semncnt() and
 *   count_semzcnt()
 * - the task that performs a successful semop() scans the list of all
 *   sleeping tasks and completes any pending operations that can be fulfilled.
 *   Semaphores are actively given to waiting tasks (necessary for FIFO).
 *   (see update_queue())
 * - To improve the scalability, the actual wake-up calls are performed after
 *   dropping all locks. (see wake_up_sem_queue_prepare(),
 *   wake_up_sem_queue_do())
 * - All work is done by the waker, the woken up task does not have to do
 *   anything - not even acquiring a lock or dropping a refcount.
 * - A woken up task may not even touch the semaphore array anymore, it may
 *   have been destroyed already by a semctl(RMID).
 * - The synchronizations between wake-ups due to a timeout/signal and a
 *   wake-up due to a completed semaphore operation is achieved by using an
 *   intermediate state (IN_WAKEUP).
 * - UNDO values are stored in an array (one per process and per
 *   semaphore array, lazily allocated). For backwards compatibility, multiple
 *   modes for the UNDO variables are supported (per process, per thread)
 *   (see copy_semundo, CLONE_SYSVSEM)
 * - There are two lists of the pending operations: a per-array list
 *   and per-semaphore list (stored in the array). This allows to achieve FIFO
 *   ordering without always scanning all pending operations.
 *   The worst-case behavior is nevertheless O(N^2) for N wakeups.
L
Linus Torvalds 已提交
74 75 76 77 78 79 80 81 82 83
 */

#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/init.h>
#include <linux/proc_fs.h>
#include <linux/time.h>
#include <linux/security.h>
#include <linux/syscalls.h>
#include <linux/audit.h>
84
#include <linux/capability.h>
85
#include <linux/seq_file.h>
N
Nadia Derbey 已提交
86
#include <linux/rwsem.h>
K
Kirill Korotaev 已提交
87
#include <linux/nsproxy.h>
88
#include <linux/ipc_namespace.h>
I
Ingo Molnar 已提交
89

L
Linus Torvalds 已提交
90 91 92
#include <asm/uaccess.h>
#include "util.h"

93 94 95 96
/* One semaphore structure for each semaphore in the system. */
struct sem {
	int	semval;		/* current value */
	int	sempid;		/* pid of last operation */
97
	spinlock_t	lock;	/* spinlock for fine-grained semtimedop */
98 99 100 101
	struct list_head pending_alter; /* pending single-sop operations */
					/* that alter the semaphore */
	struct list_head pending_const; /* pending single-sop operations */
					/* that do not alter the semaphore*/
102
	time_t	sem_otime;	/* candidate for sem_otime */
103
} ____cacheline_aligned_in_smp;
104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142

/* One queue for each sleeping process in the system. */
struct sem_queue {
	struct list_head	list;	 /* queue of pending operations */
	struct task_struct	*sleeper; /* this process */
	struct sem_undo		*undo;	 /* undo structure */
	int			pid;	 /* process id of requesting process */
	int			status;	 /* completion status of operation */
	struct sembuf		*sops;	 /* array of pending operations */
	int			nsops;	 /* number of operations */
	int			alter;	 /* does *sops alter the array? */
};

/* Each task has a list of undo requests. They are executed automatically
 * when the process exits.
 */
struct sem_undo {
	struct list_head	list_proc;	/* per-process list: *
						 * all undos from one process
						 * rcu protected */
	struct rcu_head		rcu;		/* rcu struct for sem_undo */
	struct sem_undo_list	*ulp;		/* back ptr to sem_undo_list */
	struct list_head	list_id;	/* per semaphore array list:
						 * all undos for one array */
	int			semid;		/* semaphore set identifier */
	short			*semadj;	/* array of adjustments */
						/* one per semaphore */
};

/* sem_undo_list controls shared access to the list of sem_undo structures
 * that may be shared among all a CLONE_SYSVSEM task group.
 */
struct sem_undo_list {
	atomic_t		refcnt;
	spinlock_t		lock;
	struct list_head	list_proc;
};


143
#define sem_ids(ns)	((ns)->ids[IPC_SEM_IDS])
K
Kirill Korotaev 已提交
144

N
Nadia Derbey 已提交
145
#define sem_checkid(sma, semid)	ipc_checkid(&sma->sem_perm, semid)
L
Linus Torvalds 已提交
146

N
Nadia Derbey 已提交
147
static int newary(struct ipc_namespace *, struct ipc_params *);
148
static void freeary(struct ipc_namespace *, struct kern_ipc_perm *);
L
Linus Torvalds 已提交
149
#ifdef CONFIG_PROC_FS
150
static int sysvipc_sem_proc_show(struct seq_file *s, void *it);
L
Linus Torvalds 已提交
151 152 153 154 155 156
#endif

#define SEMMSL_FAST	256 /* 512 bytes on stack */
#define SEMOPM_FAST	64  /* ~ 372 bytes on stack */

/*
157
 * Locking:
L
Linus Torvalds 已提交
158
 *	sem_undo.id_next,
159
 *	sem_array.complex_count,
160
 *	sem_array.pending{_alter,_cont},
161
 *	sem_array.sem_undo: global sem_lock() for read/write
L
Linus Torvalds 已提交
162 163
 *	sem_undo.proc_next: only "current" is allowed to read/write that field.
 *	
164 165
 *	sem_array.sem_base[i].pending_{const,alter}:
 *		global or semaphore sem_lock() for read/write
L
Linus Torvalds 已提交
166 167
 */

K
Kirill Korotaev 已提交
168 169 170 171 172
#define sc_semmsl	sem_ctls[0]
#define sc_semmns	sem_ctls[1]
#define sc_semopm	sem_ctls[2]
#define sc_semmni	sem_ctls[3]

173
void sem_init_ns(struct ipc_namespace *ns)
K
Kirill Korotaev 已提交
174 175 176 177 178 179
{
	ns->sc_semmsl = SEMMSL;
	ns->sc_semmns = SEMMNS;
	ns->sc_semopm = SEMOPM;
	ns->sc_semmni = SEMMNI;
	ns->used_sems = 0;
180
	ipc_init_ids(&ns->ids[IPC_SEM_IDS]);
K
Kirill Korotaev 已提交
181 182
}

183
#ifdef CONFIG_IPC_NS
K
Kirill Korotaev 已提交
184 185
void sem_exit_ns(struct ipc_namespace *ns)
{
186
	free_ipcs(ns, &sem_ids(ns), freeary);
S
Serge E. Hallyn 已提交
187
	idr_destroy(&ns->ids[IPC_SEM_IDS].ipcs_idr);
K
Kirill Korotaev 已提交
188
}
189
#endif
L
Linus Torvalds 已提交
190 191 192

void __init sem_init (void)
{
193
	sem_init_ns(&init_ipc_ns);
194 195
	ipc_init_proc_interface("sysvipc/sem",
				"       key      semid perms      nsems   uid   gid  cuid  cgid      otime      ctime\n",
K
Kirill Korotaev 已提交
196
				IPC_SEM_IDS, sysvipc_sem_proc_show);
L
Linus Torvalds 已提交
197 198
}

199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245
/**
 * unmerge_queues - unmerge queues, if possible.
 * @sma: semaphore array
 *
 * The function unmerges the wait queues if complex_count is 0.
 * It must be called prior to dropping the global semaphore array lock.
 */
static void unmerge_queues(struct sem_array *sma)
{
	struct sem_queue *q, *tq;

	/* complex operations still around? */
	if (sma->complex_count)
		return;
	/*
	 * We will switch back to simple mode.
	 * Move all pending operation back into the per-semaphore
	 * queues.
	 */
	list_for_each_entry_safe(q, tq, &sma->pending_alter, list) {
		struct sem *curr;
		curr = &sma->sem_base[q->sops[0].sem_num];

		list_add_tail(&q->list, &curr->pending_alter);
	}
	INIT_LIST_HEAD(&sma->pending_alter);
}

/**
 * merge_queues - Merge single semop queues into global queue
 * @sma: semaphore array
 *
 * This function merges all per-semaphore queues into the global queue.
 * It is necessary to achieve FIFO ordering for the pending single-sop
 * operations when a multi-semop operation must sleep.
 * Only the alter operations must be moved, the const operations can stay.
 */
static void merge_queues(struct sem_array *sma)
{
	int i;
	for (i = 0; i < sma->sem_nsems; i++) {
		struct sem *sem = sma->sem_base + i;

		list_splice_init(&sem->pending_alter, &sma->pending_alter);
	}
}

D
Davidlohr Bueso 已提交
246 247 248 249 250 251 252 253 254
static void sem_rcu_free(struct rcu_head *head)
{
	struct ipc_rcu *p = container_of(head, struct ipc_rcu, rcu);
	struct sem_array *sma = ipc_rcu_to_struct(p);

	security_sem_free(sma);
	ipc_rcu_free(head);
}

255 256 257 258 259
/*
 * Wait until all currently ongoing simple ops have completed.
 * Caller must own sem_perm.lock.
 * New simple ops cannot start, because simple ops first check
 * that sem_perm.lock is free.
M
Manfred Spraul 已提交
260
 * that a) sem_perm.lock is free and b) complex_count is 0.
261 262 263 264 265 266
 */
static void sem_wait_array(struct sem_array *sma)
{
	int i;
	struct sem *sem;

M
Manfred Spraul 已提交
267 268 269 270 271 272 273
	if (sma->complex_count)  {
		/* The thread that increased sma->complex_count waited on
		 * all sem->lock locks. Thus we don't need to wait again.
		 */
		return;
	}

274 275 276 277 278 279
	for (i = 0; i < sma->sem_nsems; i++) {
		sem = sma->sem_base + i;
		spin_unlock_wait(&sem->lock);
	}
}

280 281 282 283 284 285 286 287 288 289
/*
 * If the request contains only one semaphore operation, and there are
 * no complex transactions pending, lock only the semaphore involved.
 * Otherwise, lock the entire semaphore array, since we either have
 * multiple semaphores in our own semops, or we need to look at
 * semaphores from other pending complex operations.
 */
static inline int sem_lock(struct sem_array *sma, struct sembuf *sops,
			      int nsops)
{
290
	struct sem *sem;
291

292 293 294
	if (nsops != 1) {
		/* Complex operation - acquire a full lock */
		ipc_lock_object(&sma->sem_perm);
295

296 297
		/* And wait until all simple ops that are processed
		 * right now have dropped their locks.
298
		 */
299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318
		sem_wait_array(sma);
		return -1;
	}

	/*
	 * Only one semaphore affected - try to optimize locking.
	 * The rules are:
	 * - optimized locking is possible if no complex operation
	 *   is either enqueued or processed right now.
	 * - The test for enqueued complex ops is simple:
	 *      sma->complex_count != 0
	 * - Testing for complex ops that are processed right now is
	 *   a bit more difficult. Complex ops acquire the full lock
	 *   and first wait that the running simple ops have completed.
	 *   (see above)
	 *   Thus: If we own a simple lock and the global lock is free
	 *	and complex_count is now 0, then it will stay 0 and
	 *	thus just locking sem->lock is sufficient.
	 */
	sem = sma->sem_base + sops->sem_num;
319

320
	if (sma->complex_count == 0) {
321
		/*
322 323
		 * It appears that no complex operation is around.
		 * Acquire the per-semaphore lock.
324
		 */
325 326 327 328 329 330 331 332 333 334 335 336 337 338 339
		spin_lock(&sem->lock);

		/* Then check that the global lock is free */
		if (!spin_is_locked(&sma->sem_perm.lock)) {
			/* spin_is_locked() is not a memory barrier */
			smp_mb();

			/* Now repeat the test of complex_count:
			 * It can't change anymore until we drop sem->lock.
			 * Thus: if is now 0, then it will stay 0.
			 */
			if (sma->complex_count == 0) {
				/* fast path successful! */
				return sops->sem_num;
			}
340
		}
341 342 343 344 345
		spin_unlock(&sem->lock);
	}

	/* slow path: acquire the full lock */
	ipc_lock_object(&sma->sem_perm);
346

347 348 349 350 351 352 353 354
	if (sma->complex_count == 0) {
		/* False alarm:
		 * There is no complex operation, thus we can switch
		 * back to the fast path.
		 */
		spin_lock(&sem->lock);
		ipc_unlock_object(&sma->sem_perm);
		return sops->sem_num;
355
	} else {
356 357
		/* Not a false alarm, thus complete the sequence for a
		 * full lock.
358
		 */
359 360
		sem_wait_array(sma);
		return -1;
361 362 363 364 365 366
	}
}

static inline void sem_unlock(struct sem_array *sma, int locknum)
{
	if (locknum == -1) {
367
		unmerge_queues(sma);
368
		ipc_unlock_object(&sma->sem_perm);
369 370 371 372 373 374
	} else {
		struct sem *sem = sma->sem_base + locknum;
		spin_unlock(&sem->lock);
	}
}

N
Nadia Derbey 已提交
375
/*
D
Davidlohr Bueso 已提交
376
 * sem_lock_(check_) routines are called in the paths where the rwsem
N
Nadia Derbey 已提交
377
 * is not held.
378 379
 *
 * The caller holds the RCU read lock.
N
Nadia Derbey 已提交
380
 */
381 382
static inline struct sem_array *sem_obtain_lock(struct ipc_namespace *ns,
			int id, struct sembuf *sops, int nsops, int *locknum)
383
{
384 385
	struct kern_ipc_perm *ipcp;
	struct sem_array *sma;
N
Nadia Derbey 已提交
386

387
	ipcp = ipc_obtain_object(&sem_ids(ns), id);
388 389
	if (IS_ERR(ipcp))
		return ERR_CAST(ipcp);
390

391 392
	sma = container_of(ipcp, struct sem_array, sem_perm);
	*locknum = sem_lock(sma, sops, nsops);
393 394 395 396 397 398 399

	/* ipc_rmid() may have already freed the ID while sem_lock
	 * was spinning: verify that the structure is still valid
	 */
	if (!ipcp->deleted)
		return container_of(ipcp, struct sem_array, sem_perm);

400
	sem_unlock(sma, *locknum);
401
	return ERR_PTR(-EINVAL);
402 403
}

404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420
static inline struct sem_array *sem_obtain_object(struct ipc_namespace *ns, int id)
{
	struct kern_ipc_perm *ipcp = ipc_obtain_object(&sem_ids(ns), id);

	if (IS_ERR(ipcp))
		return ERR_CAST(ipcp);

	return container_of(ipcp, struct sem_array, sem_perm);
}

static inline struct sem_array *sem_obtain_object_check(struct ipc_namespace *ns,
							int id)
{
	struct kern_ipc_perm *ipcp = ipc_obtain_object_check(&sem_ids(ns), id);

	if (IS_ERR(ipcp))
		return ERR_CAST(ipcp);
421

N
Nadia Derbey 已提交
422
	return container_of(ipcp, struct sem_array, sem_perm);
423 424
}

425 426
static inline void sem_lock_and_putref(struct sem_array *sma)
{
427
	sem_lock(sma, NULL, -1);
D
Davidlohr Bueso 已提交
428
	ipc_rcu_putref(sma, ipc_rcu_free);
429 430
}

N
Nadia Derbey 已提交
431 432 433 434 435
static inline void sem_rmid(struct ipc_namespace *ns, struct sem_array *s)
{
	ipc_rmid(&sem_ids(ns), &s->sem_perm);
}

L
Linus Torvalds 已提交
436 437 438 439 440
/*
 * Lockless wakeup algorithm:
 * Without the check/retry algorithm a lockless wakeup is possible:
 * - queue.status is initialized to -EINTR before blocking.
 * - wakeup is performed by
441
 *	* unlinking the queue entry from the pending list
L
Linus Torvalds 已提交
442 443 444 445 446 447 448 449 450
 *	* setting queue.status to IN_WAKEUP
 *	  This is the notification for the blocked thread that a
 *	  result value is imminent.
 *	* call wake_up_process
 *	* set queue.status to the final value.
 * - the previously blocked thread checks queue.status:
 *   	* if it's IN_WAKEUP, then it must wait until the value changes
 *   	* if it's not -EINTR, then the operation was completed by
 *   	  update_queue. semtimedop can return queue.status without
I
Ingo Molnar 已提交
451
 *   	  performing any operation on the sem array.
L
Linus Torvalds 已提交
452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469
 *   	* otherwise it must acquire the spinlock and check what's up.
 *
 * The two-stage algorithm is necessary to protect against the following
 * races:
 * - if queue.status is set after wake_up_process, then the woken up idle
 *   thread could race forward and try (and fail) to acquire sma->lock
 *   before update_queue had a chance to set queue.status
 * - if queue.status is written before wake_up_process and if the
 *   blocked process is woken up by a signal between writing
 *   queue.status and the wake_up_process, then the woken up
 *   process could return from semtimedop and die by calling
 *   sys_exit before wake_up_process is called. Then wake_up_process
 *   will oops, because the task structure is already invalid.
 *   (yes, this happened on s390 with sysv msg).
 *
 */
#define IN_WAKEUP	1

N
Nadia Derbey 已提交
470 471 472 473 474
/**
 * newary - Create a new semaphore set
 * @ns: namespace
 * @params: ptr to the structure that contains key, semflg and nsems
 *
D
Davidlohr Bueso 已提交
475
 * Called with sem_ids.rwsem held (as a writer)
N
Nadia Derbey 已提交
476 477
 */

N
Nadia Derbey 已提交
478
static int newary(struct ipc_namespace *ns, struct ipc_params *params)
L
Linus Torvalds 已提交
479 480 481 482 483
{
	int id;
	int retval;
	struct sem_array *sma;
	int size;
N
Nadia Derbey 已提交
484 485 486
	key_t key = params->key;
	int nsems = params->u.nsems;
	int semflg = params->flg;
487
	int i;
L
Linus Torvalds 已提交
488 489 490

	if (!nsems)
		return -EINVAL;
K
Kirill Korotaev 已提交
491
	if (ns->used_sems + nsems > ns->sc_semmns)
L
Linus Torvalds 已提交
492 493 494 495 496 497 498 499 500 501 502 503 504 505 506
		return -ENOSPC;

	size = sizeof (*sma) + nsems * sizeof (struct sem);
	sma = ipc_rcu_alloc(size);
	if (!sma) {
		return -ENOMEM;
	}
	memset (sma, 0, size);

	sma->sem_perm.mode = (semflg & S_IRWXUGO);
	sma->sem_perm.key = key;

	sma->sem_perm.security = NULL;
	retval = security_sem_alloc(sma);
	if (retval) {
D
Davidlohr Bueso 已提交
507
		ipc_rcu_putref(sma, ipc_rcu_free);
L
Linus Torvalds 已提交
508 509 510
		return retval;
	}

K
Kirill Korotaev 已提交
511
	id = ipc_addid(&sem_ids(ns), &sma->sem_perm, ns->sc_semmni);
512
	if (id < 0) {
D
Davidlohr Bueso 已提交
513
		ipc_rcu_putref(sma, sem_rcu_free);
514
		return id;
L
Linus Torvalds 已提交
515
	}
K
Kirill Korotaev 已提交
516
	ns->used_sems += nsems;
L
Linus Torvalds 已提交
517 518

	sma->sem_base = (struct sem *) &sma[1];
519

520
	for (i = 0; i < nsems; i++) {
521 522
		INIT_LIST_HEAD(&sma->sem_base[i].pending_alter);
		INIT_LIST_HEAD(&sma->sem_base[i].pending_const);
523 524
		spin_lock_init(&sma->sem_base[i].lock);
	}
525 526

	sma->complex_count = 0;
527 528
	INIT_LIST_HEAD(&sma->pending_alter);
	INIT_LIST_HEAD(&sma->pending_const);
529
	INIT_LIST_HEAD(&sma->list_id);
L
Linus Torvalds 已提交
530 531
	sma->sem_nsems = nsems;
	sma->sem_ctime = get_seconds();
532
	sem_unlock(sma, -1);
533
	rcu_read_unlock();
L
Linus Torvalds 已提交
534

N
Nadia Derbey 已提交
535
	return sma->sem_perm.id;
L
Linus Torvalds 已提交
536 537
}

N
Nadia Derbey 已提交
538

N
Nadia Derbey 已提交
539
/*
D
Davidlohr Bueso 已提交
540
 * Called with sem_ids.rwsem and ipcp locked.
N
Nadia Derbey 已提交
541
 */
N
Nadia Derbey 已提交
542
static inline int sem_security(struct kern_ipc_perm *ipcp, int semflg)
N
Nadia Derbey 已提交
543
{
N
Nadia Derbey 已提交
544 545 546 547
	struct sem_array *sma;

	sma = container_of(ipcp, struct sem_array, sem_perm);
	return security_sem_associate(sma, semflg);
N
Nadia Derbey 已提交
548 549
}

N
Nadia Derbey 已提交
550
/*
D
Davidlohr Bueso 已提交
551
 * Called with sem_ids.rwsem and ipcp locked.
N
Nadia Derbey 已提交
552
 */
N
Nadia Derbey 已提交
553 554
static inline int sem_more_checks(struct kern_ipc_perm *ipcp,
				struct ipc_params *params)
N
Nadia Derbey 已提交
555
{
N
Nadia Derbey 已提交
556 557 558 559
	struct sem_array *sma;

	sma = container_of(ipcp, struct sem_array, sem_perm);
	if (params->u.nsems > sma->sem_nsems)
N
Nadia Derbey 已提交
560 561 562 563 564
		return -EINVAL;

	return 0;
}

565
SYSCALL_DEFINE3(semget, key_t, key, int, nsems, int, semflg)
L
Linus Torvalds 已提交
566
{
K
Kirill Korotaev 已提交
567
	struct ipc_namespace *ns;
N
Nadia Derbey 已提交
568 569
	struct ipc_ops sem_ops;
	struct ipc_params sem_params;
K
Kirill Korotaev 已提交
570 571

	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
572

K
Kirill Korotaev 已提交
573
	if (nsems < 0 || nsems > ns->sc_semmsl)
L
Linus Torvalds 已提交
574
		return -EINVAL;
N
Nadia Derbey 已提交
575

N
Nadia Derbey 已提交
576 577 578 579 580 581 582
	sem_ops.getnew = newary;
	sem_ops.associate = sem_security;
	sem_ops.more_checks = sem_more_checks;

	sem_params.key = key;
	sem_params.flg = semflg;
	sem_params.u.nsems = nsems;
L
Linus Torvalds 已提交
583

N
Nadia Derbey 已提交
584
	return ipcget(ns, &sem_ids(ns), &sem_ops, &sem_params);
L
Linus Torvalds 已提交
585 586
}

587 588 589 590 591 592 593 594 595 596
/** perform_atomic_semop - Perform (if possible) a semaphore operation
 * @sma: semaphore array
 * @sops: array with operations that should be checked
 * @nsems: number of sops
 * @un: undo array
 * @pid: pid that did the change
 *
 * Returns 0 if the operation was possible.
 * Returns 1 if the operation is impossible, the caller must sleep.
 * Negative values are error codes.
L
Linus Torvalds 已提交
597 598
 */

599
static int perform_atomic_semop(struct sem_array *sma, struct sembuf *sops,
L
Linus Torvalds 已提交
600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659
			     int nsops, struct sem_undo *un, int pid)
{
	int result, sem_op;
	struct sembuf *sop;
	struct sem * curr;

	for (sop = sops; sop < sops + nsops; sop++) {
		curr = sma->sem_base + sop->sem_num;
		sem_op = sop->sem_op;
		result = curr->semval;
  
		if (!sem_op && result)
			goto would_block;

		result += sem_op;
		if (result < 0)
			goto would_block;
		if (result > SEMVMX)
			goto out_of_range;
		if (sop->sem_flg & SEM_UNDO) {
			int undo = un->semadj[sop->sem_num] - sem_op;
			/*
	 		 *	Exceeding the undo range is an error.
			 */
			if (undo < (-SEMAEM - 1) || undo > SEMAEM)
				goto out_of_range;
		}
		curr->semval = result;
	}

	sop--;
	while (sop >= sops) {
		sma->sem_base[sop->sem_num].sempid = pid;
		if (sop->sem_flg & SEM_UNDO)
			un->semadj[sop->sem_num] -= sop->sem_op;
		sop--;
	}
	
	return 0;

out_of_range:
	result = -ERANGE;
	goto undo;

would_block:
	if (sop->sem_flg & IPC_NOWAIT)
		result = -EAGAIN;
	else
		result = 1;

undo:
	sop--;
	while (sop >= sops) {
		sma->sem_base[sop->sem_num].semval -= sop->sem_op;
		sop--;
	}

	return result;
}

660 661 662 663 664
/** wake_up_sem_queue_prepare(q, error): Prepare wake-up
 * @q: queue entry that must be signaled
 * @error: Error value for the signal
 *
 * Prepare the wake-up of the queue entry q.
N
Nick Piggin 已提交
665
 */
666 667
static void wake_up_sem_queue_prepare(struct list_head *pt,
				struct sem_queue *q, int error)
N
Nick Piggin 已提交
668
{
669 670 671 672 673 674 675
	if (list_empty(pt)) {
		/*
		 * Hold preempt off so that we don't get preempted and have the
		 * wakee busy-wait until we're scheduled back on.
		 */
		preempt_disable();
	}
N
Nick Piggin 已提交
676
	q->status = IN_WAKEUP;
677 678
	q->pid = error;

679
	list_add_tail(&q->list, pt);
680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696
}

/**
 * wake_up_sem_queue_do(pt) - do the actual wake-up
 * @pt: list of tasks to be woken up
 *
 * Do the actual wake-up.
 * The function is called without any locks held, thus the semaphore array
 * could be destroyed already and the tasks can disappear as soon as the
 * status is set to the actual return code.
 */
static void wake_up_sem_queue_do(struct list_head *pt)
{
	struct sem_queue *q, *t;
	int did_something;

	did_something = !list_empty(pt);
697
	list_for_each_entry_safe(q, t, pt, list) {
698 699 700 701 702 703 704
		wake_up_process(q->sleeper);
		/* q can disappear immediately after writing q->status. */
		smp_wmb();
		q->status = q->pid;
	}
	if (did_something)
		preempt_enable();
N
Nick Piggin 已提交
705 706
}

707 708 709
static void unlink_queue(struct sem_array *sma, struct sem_queue *q)
{
	list_del(&q->list);
710
	if (q->nsops > 1)
711 712 713
		sma->complex_count--;
}

714 715 716 717 718 719 720
/** check_restart(sma, q)
 * @sma: semaphore array
 * @q: the operation that just completed
 *
 * update_queue is O(N^2) when it restarts scanning the whole queue of
 * waiting operations. Therefore this function checks if the restart is
 * really necessary. It is called after a previously waiting operation
721 722
 * modified the array.
 * Note that wait-for-zero operations are handled without restart.
723 724 725
 */
static int check_restart(struct sem_array *sma, struct sem_queue *q)
{
726 727
	/* pending complex alter operations are too difficult to analyse */
	if (!list_empty(&sma->pending_alter))
728 729 730 731 732 733
		return 1;

	/* we were a sleeping complex operation. Too difficult */
	if (q->nsops > 1)
		return 1;

734 735 736 737 738 739 740 741 742 743 744 745 746
	/* It is impossible that someone waits for the new value:
	 * - complex operations always restart.
	 * - wait-for-zero are handled seperately.
	 * - q is a previously sleeping simple operation that
	 *   altered the array. It must be a decrement, because
	 *   simple increments never sleep.
	 * - If there are older (higher priority) decrements
	 *   in the queue, then they have observed the original
	 *   semval value and couldn't proceed. The operation
	 *   decremented to value - thus they won't proceed either.
	 */
	return 0;
}
747

748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773
/**
 * wake_const_ops(sma, semnum, pt) - Wake up non-alter tasks
 * @sma: semaphore array.
 * @semnum: semaphore that was modified.
 * @pt: list head for the tasks that must be woken up.
 *
 * wake_const_ops must be called after a semaphore in a semaphore array
 * was set to 0. If complex const operations are pending, wake_const_ops must
 * be called with semnum = -1, as well as with the number of each modified
 * semaphore.
 * The tasks that must be woken up are added to @pt. The return code
 * is stored in q->pid.
 * The function returns 1 if at least one operation was completed successfully.
 */
static int wake_const_ops(struct sem_array *sma, int semnum,
				struct list_head *pt)
{
	struct sem_queue *q;
	struct list_head *walk;
	struct list_head *pending_list;
	int semop_completed = 0;

	if (semnum == -1)
		pending_list = &sma->pending_const;
	else
		pending_list = &sma->sem_base[semnum].pending_const;
774

775 776 777 778 779 780 781
	walk = pending_list->next;
	while (walk != pending_list) {
		int error;

		q = container_of(walk, struct sem_queue, list);
		walk = walk->next;

782 783
		error = perform_atomic_semop(sma, q->sops, q->nsops,
						 q->undo, q->pid);
784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830

		if (error <= 0) {
			/* operation completed, remove from queue & wakeup */

			unlink_queue(sma, q);

			wake_up_sem_queue_prepare(pt, q, error);
			if (error == 0)
				semop_completed = 1;
		}
	}
	return semop_completed;
}

/**
 * do_smart_wakeup_zero(sma, sops, nsops, pt) - wakeup all wait for zero tasks
 * @sma: semaphore array
 * @sops: operations that were performed
 * @nsops: number of operations
 * @pt: list head of the tasks that must be woken up.
 *
 * do_smart_wakeup_zero() checks all required queue for wait-for-zero
 * operations, based on the actual changes that were performed on the
 * semaphore array.
 * The function returns 1 if at least one operation was completed successfully.
 */
static int do_smart_wakeup_zero(struct sem_array *sma, struct sembuf *sops,
					int nsops, struct list_head *pt)
{
	int i;
	int semop_completed = 0;
	int got_zero = 0;

	/* first: the per-semaphore queues, if known */
	if (sops) {
		for (i = 0; i < nsops; i++) {
			int num = sops[i].sem_num;

			if (sma->sem_base[num].semval == 0) {
				got_zero = 1;
				semop_completed |= wake_const_ops(sma, num, pt);
			}
		}
	} else {
		/*
		 * No sops means modified semaphores not known.
		 * Assume all were changed.
831
		 */
832 833 834 835 836 837
		for (i = 0; i < sma->sem_nsems; i++) {
			if (sma->sem_base[i].semval == 0) {
				got_zero = 1;
				semop_completed |= wake_const_ops(sma, i, pt);
			}
		}
838 839
	}
	/*
840 841
	 * If one of the modified semaphores got 0,
	 * then check the global queue, too.
842
	 */
843 844
	if (got_zero)
		semop_completed |= wake_const_ops(sma, -1, pt);
845

846
	return semop_completed;
847 848
}

849 850 851 852 853

/**
 * update_queue(sma, semnum): Look for tasks that can be completed.
 * @sma: semaphore array.
 * @semnum: semaphore that was modified.
854
 * @pt: list head for the tasks that must be woken up.
855 856
 *
 * update_queue must be called after a semaphore in a semaphore array
857 858 859
 * was modified. If multiple semaphores were modified, update_queue must
 * be called with semnum = -1, as well as with the number of each modified
 * semaphore.
860 861
 * The tasks that must be woken up are added to @pt. The return code
 * is stored in q->pid.
862 863
 * The function internally checks if const operations can now succeed.
 *
864
 * The function return 1 if at least one semop was completed successfully.
L
Linus Torvalds 已提交
865
 */
866
static int update_queue(struct sem_array *sma, int semnum, struct list_head *pt)
L
Linus Torvalds 已提交
867
{
868 869 870
	struct sem_queue *q;
	struct list_head *walk;
	struct list_head *pending_list;
871
	int semop_completed = 0;
872

873
	if (semnum == -1)
874
		pending_list = &sma->pending_alter;
875
	else
876
		pending_list = &sma->sem_base[semnum].pending_alter;
N
Nick Piggin 已提交
877 878

again:
879 880
	walk = pending_list->next;
	while (walk != pending_list) {
881
		int error, restart;
882

883
		q = container_of(walk, struct sem_queue, list);
884
		walk = walk->next;
L
Linus Torvalds 已提交
885

886 887
		/* If we are scanning the single sop, per-semaphore list of
		 * one semaphore and that semaphore is 0, then it is not
888
		 * necessary to scan further: simple increments
889 890 891 892
		 * that affect only one entry succeed immediately and cannot
		 * be in the  per semaphore pending queue, and decrements
		 * cannot be successful if the value is already 0.
		 */
893
		if (semnum != -1 && sma->sem_base[semnum].semval == 0)
894 895
			break;

896
		error = perform_atomic_semop(sma, q->sops, q->nsops,
L
Linus Torvalds 已提交
897 898 899
					 q->undo, q->pid);

		/* Does q->sleeper still need to sleep? */
N
Nick Piggin 已提交
900 901 902
		if (error > 0)
			continue;

903
		unlink_queue(sma, q);
N
Nick Piggin 已提交
904

905
		if (error) {
906
			restart = 0;
907 908
		} else {
			semop_completed = 1;
909
			do_smart_wakeup_zero(sma, q->sops, q->nsops, pt);
910
			restart = check_restart(sma, q);
911
		}
912

913
		wake_up_sem_queue_prepare(pt, q, error);
914
		if (restart)
N
Nick Piggin 已提交
915
			goto again;
L
Linus Torvalds 已提交
916
	}
917
	return semop_completed;
L
Linus Torvalds 已提交
918 919
}

920 921
/**
 * do_smart_update(sma, sops, nsops, otime, pt) - optimized update_queue
922 923 924
 * @sma: semaphore array
 * @sops: operations that were performed
 * @nsops: number of operations
925 926
 * @otime: force setting otime
 * @pt: list head of the tasks that must be woken up.
927
 *
928 929
 * do_smart_update() does the required calls to update_queue and wakeup_zero,
 * based on the actual changes that were performed on the semaphore array.
930 931 932
 * Note that the function does not do the actual wake-up: the caller is
 * responsible for calling wake_up_sem_queue_do(@pt).
 * It is safe to perform this call after dropping all locks.
933
 */
934 935
static void do_smart_update(struct sem_array *sma, struct sembuf *sops, int nsops,
			int otime, struct list_head *pt)
936 937 938
{
	int i;

939 940
	otime |= do_smart_wakeup_zero(sma, sops, nsops, pt);

941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966
	if (!list_empty(&sma->pending_alter)) {
		/* semaphore array uses the global queue - just process it. */
		otime |= update_queue(sma, -1, pt);
	} else {
		if (!sops) {
			/*
			 * No sops, thus the modified semaphores are not
			 * known. Check all.
			 */
			for (i = 0; i < sma->sem_nsems; i++)
				otime |= update_queue(sma, i, pt);
		} else {
			/*
			 * Check the semaphores that were increased:
			 * - No complex ops, thus all sleeping ops are
			 *   decrease.
			 * - if we decreased the value, then any sleeping
			 *   semaphore ops wont be able to run: If the
			 *   previous value was too small, then the new
			 *   value will be too small, too.
			 */
			for (i = 0; i < nsops; i++) {
				if (sops[i].sem_op > 0) {
					otime |= update_queue(sma,
							sops[i].sem_num, pt);
				}
967
			}
968
		}
969
	}
970 971 972 973 974 975 976 977
	if (otime) {
		if (sops == NULL) {
			sma->sem_base[0].sem_otime = get_seconds();
		} else {
			sma->sem_base[sops[0].sem_num].sem_otime =
								get_seconds();
		}
	}
978 979 980
}


L
Linus Torvalds 已提交
981 982 983 984 985 986 987 988 989 990 991 992 993 994 995
/* The following counts are associated to each semaphore:
 *   semncnt        number of tasks waiting on semval being nonzero
 *   semzcnt        number of tasks waiting on semval being zero
 * This model assumes that a task waits on exactly one semaphore.
 * Since semaphore operations are to be performed atomically, tasks actually
 * wait on a whole sequence of semaphores simultaneously.
 * The counts we return here are a rough approximation, but still
 * warrant that semncnt+semzcnt>0 if the task is on the pending queue.
 */
static int count_semncnt (struct sem_array * sma, ushort semnum)
{
	int semncnt;
	struct sem_queue * q;

	semncnt = 0;
996
	list_for_each_entry(q, &sma->sem_base[semnum].pending_alter, list) {
R
Rik van Riel 已提交
997 998 999 1000 1001 1002
		struct sembuf * sops = q->sops;
		BUG_ON(sops->sem_num != semnum);
		if ((sops->sem_op < 0) && !(sops->sem_flg & IPC_NOWAIT))
			semncnt++;
	}

1003
	list_for_each_entry(q, &sma->pending_alter, list) {
L
Linus Torvalds 已提交
1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014
		struct sembuf * sops = q->sops;
		int nsops = q->nsops;
		int i;
		for (i = 0; i < nsops; i++)
			if (sops[i].sem_num == semnum
			    && (sops[i].sem_op < 0)
			    && !(sops[i].sem_flg & IPC_NOWAIT))
				semncnt++;
	}
	return semncnt;
}
1015

L
Linus Torvalds 已提交
1016 1017 1018 1019 1020 1021
static int count_semzcnt (struct sem_array * sma, ushort semnum)
{
	int semzcnt;
	struct sem_queue * q;

	semzcnt = 0;
1022
	list_for_each_entry(q, &sma->sem_base[semnum].pending_const, list) {
R
Rik van Riel 已提交
1023 1024 1025 1026 1027 1028
		struct sembuf * sops = q->sops;
		BUG_ON(sops->sem_num != semnum);
		if ((sops->sem_op == 0) && !(sops->sem_flg & IPC_NOWAIT))
			semzcnt++;
	}

1029
	list_for_each_entry(q, &sma->pending_const, list) {
L
Linus Torvalds 已提交
1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
		struct sembuf * sops = q->sops;
		int nsops = q->nsops;
		int i;
		for (i = 0; i < nsops; i++)
			if (sops[i].sem_num == semnum
			    && (sops[i].sem_op == 0)
			    && !(sops[i].sem_flg & IPC_NOWAIT))
				semzcnt++;
	}
	return semzcnt;
}

D
Davidlohr Bueso 已提交
1042 1043
/* Free a semaphore set. freeary() is called with sem_ids.rwsem locked
 * as a writer and the spinlock for this semaphore set hold. sem_ids.rwsem
N
Nadia Derbey 已提交
1044
 * remains locked on exit.
L
Linus Torvalds 已提交
1045
 */
1046
static void freeary(struct ipc_namespace *ns, struct kern_ipc_perm *ipcp)
L
Linus Torvalds 已提交
1047
{
1048 1049
	struct sem_undo *un, *tu;
	struct sem_queue *q, *tq;
1050
	struct sem_array *sma = container_of(ipcp, struct sem_array, sem_perm);
1051
	struct list_head tasks;
1052
	int i;
L
Linus Torvalds 已提交
1053

1054
	/* Free the existing undo structures for this semaphore set.  */
1055
	ipc_assert_locked_object(&sma->sem_perm);
1056 1057 1058
	list_for_each_entry_safe(un, tu, &sma->list_id, list_id) {
		list_del(&un->list_id);
		spin_lock(&un->ulp->lock);
L
Linus Torvalds 已提交
1059
		un->semid = -1;
1060 1061
		list_del_rcu(&un->list_proc);
		spin_unlock(&un->ulp->lock);
1062
		kfree_rcu(un, rcu);
1063
	}
L
Linus Torvalds 已提交
1064 1065

	/* Wake up all pending processes and let them fail with EIDRM. */
1066
	INIT_LIST_HEAD(&tasks);
1067 1068 1069 1070 1071 1072
	list_for_each_entry_safe(q, tq, &sma->pending_const, list) {
		unlink_queue(sma, q);
		wake_up_sem_queue_prepare(&tasks, q, -EIDRM);
	}

	list_for_each_entry_safe(q, tq, &sma->pending_alter, list) {
1073
		unlink_queue(sma, q);
1074
		wake_up_sem_queue_prepare(&tasks, q, -EIDRM);
L
Linus Torvalds 已提交
1075
	}
1076 1077
	for (i = 0; i < sma->sem_nsems; i++) {
		struct sem *sem = sma->sem_base + i;
1078 1079 1080 1081 1082
		list_for_each_entry_safe(q, tq, &sem->pending_const, list) {
			unlink_queue(sma, q);
			wake_up_sem_queue_prepare(&tasks, q, -EIDRM);
		}
		list_for_each_entry_safe(q, tq, &sem->pending_alter, list) {
1083 1084 1085 1086
			unlink_queue(sma, q);
			wake_up_sem_queue_prepare(&tasks, q, -EIDRM);
		}
	}
L
Linus Torvalds 已提交
1087

N
Nadia Derbey 已提交
1088 1089
	/* Remove the semaphore set from the IDR */
	sem_rmid(ns, sma);
1090
	sem_unlock(sma, -1);
1091
	rcu_read_unlock();
L
Linus Torvalds 已提交
1092

1093
	wake_up_sem_queue_do(&tasks);
K
Kirill Korotaev 已提交
1094
	ns->used_sems -= sma->sem_nsems;
D
Davidlohr Bueso 已提交
1095
	ipc_rcu_putref(sma, sem_rcu_free);
L
Linus Torvalds 已提交
1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106
}

static unsigned long copy_semid_to_user(void __user *buf, struct semid64_ds *in, int version)
{
	switch(version) {
	case IPC_64:
		return copy_to_user(buf, in, sizeof(*in));
	case IPC_OLD:
	    {
		struct semid_ds out;

1107 1108
		memset(&out, 0, sizeof(out));

L
Linus Torvalds 已提交
1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121
		ipc64_perm_to_ipc_perm(&in->sem_perm, &out.sem_perm);

		out.sem_otime	= in->sem_otime;
		out.sem_ctime	= in->sem_ctime;
		out.sem_nsems	= in->sem_nsems;

		return copy_to_user(buf, &out, sizeof(out));
	    }
	default:
		return -EINVAL;
	}
}

1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136
static time_t get_semotime(struct sem_array *sma)
{
	int i;
	time_t res;

	res = sma->sem_base[0].sem_otime;
	for (i = 1; i < sma->sem_nsems; i++) {
		time_t to = sma->sem_base[i].sem_otime;

		if (to > res)
			res = to;
	}
	return res;
}

1137
static int semctl_nolock(struct ipc_namespace *ns, int semid,
1138
			 int cmd, int version, void __user *p)
L
Linus Torvalds 已提交
1139
{
1140
	int err;
L
Linus Torvalds 已提交
1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154
	struct sem_array *sma;

	switch(cmd) {
	case IPC_INFO:
	case SEM_INFO:
	{
		struct seminfo seminfo;
		int max_id;

		err = security_sem_semctl(NULL, cmd);
		if (err)
			return err;
		
		memset(&seminfo,0,sizeof(seminfo));
K
Kirill Korotaev 已提交
1155 1156 1157 1158
		seminfo.semmni = ns->sc_semmni;
		seminfo.semmns = ns->sc_semmns;
		seminfo.semmsl = ns->sc_semmsl;
		seminfo.semopm = ns->sc_semopm;
L
Linus Torvalds 已提交
1159 1160 1161 1162
		seminfo.semvmx = SEMVMX;
		seminfo.semmnu = SEMMNU;
		seminfo.semmap = SEMMAP;
		seminfo.semume = SEMUME;
D
Davidlohr Bueso 已提交
1163
		down_read(&sem_ids(ns).rwsem);
L
Linus Torvalds 已提交
1164
		if (cmd == SEM_INFO) {
K
Kirill Korotaev 已提交
1165 1166
			seminfo.semusz = sem_ids(ns).in_use;
			seminfo.semaem = ns->used_sems;
L
Linus Torvalds 已提交
1167 1168 1169 1170
		} else {
			seminfo.semusz = SEMUSZ;
			seminfo.semaem = SEMAEM;
		}
N
Nadia Derbey 已提交
1171
		max_id = ipc_get_maxid(&sem_ids(ns));
D
Davidlohr Bueso 已提交
1172
		up_read(&sem_ids(ns).rwsem);
1173
		if (copy_to_user(p, &seminfo, sizeof(struct seminfo))) 
L
Linus Torvalds 已提交
1174 1175 1176
			return -EFAULT;
		return (max_id < 0) ? 0: max_id;
	}
1177
	case IPC_STAT:
L
Linus Torvalds 已提交
1178 1179 1180
	case SEM_STAT:
	{
		struct semid64_ds tbuf;
1181 1182 1183
		int id = 0;

		memset(&tbuf, 0, sizeof(tbuf));
L
Linus Torvalds 已提交
1184

1185
		rcu_read_lock();
1186
		if (cmd == SEM_STAT) {
1187 1188 1189 1190 1191
			sma = sem_obtain_object(ns, semid);
			if (IS_ERR(sma)) {
				err = PTR_ERR(sma);
				goto out_unlock;
			}
1192 1193
			id = sma->sem_perm.id;
		} else {
1194 1195 1196 1197 1198
			sma = sem_obtain_object_check(ns, semid);
			if (IS_ERR(sma)) {
				err = PTR_ERR(sma);
				goto out_unlock;
			}
1199
		}
L
Linus Torvalds 已提交
1200 1201

		err = -EACCES;
1202
		if (ipcperms(ns, &sma->sem_perm, S_IRUGO))
L
Linus Torvalds 已提交
1203 1204 1205 1206 1207 1208 1209
			goto out_unlock;

		err = security_sem_semctl(sma, cmd);
		if (err)
			goto out_unlock;

		kernel_to_ipc64_perm(&sma->sem_perm, &tbuf.sem_perm);
1210 1211 1212
		tbuf.sem_otime = get_semotime(sma);
		tbuf.sem_ctime = sma->sem_ctime;
		tbuf.sem_nsems = sma->sem_nsems;
1213
		rcu_read_unlock();
1214
		if (copy_semid_to_user(p, &tbuf, version))
L
Linus Torvalds 已提交
1215 1216 1217 1218 1219 1220 1221
			return -EFAULT;
		return id;
	}
	default:
		return -EINVAL;
	}
out_unlock:
1222
	rcu_read_unlock();
L
Linus Torvalds 已提交
1223 1224 1225
	return err;
}

1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242
static int semctl_setval(struct ipc_namespace *ns, int semid, int semnum,
		unsigned long arg)
{
	struct sem_undo *un;
	struct sem_array *sma;
	struct sem* curr;
	int err;
	struct list_head tasks;
	int val;
#if defined(CONFIG_64BIT) && defined(__BIG_ENDIAN)
	/* big-endian 64bit */
	val = arg >> 32;
#else
	/* 32bit or little-endian 64bit */
	val = arg;
#endif

1243 1244
	if (val > SEMVMX || val < 0)
		return -ERANGE;
1245 1246 1247

	INIT_LIST_HEAD(&tasks);

1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264
	rcu_read_lock();
	sma = sem_obtain_object_check(ns, semid);
	if (IS_ERR(sma)) {
		rcu_read_unlock();
		return PTR_ERR(sma);
	}

	if (semnum < 0 || semnum >= sma->sem_nsems) {
		rcu_read_unlock();
		return -EINVAL;
	}


	if (ipcperms(ns, &sma->sem_perm, S_IWUGO)) {
		rcu_read_unlock();
		return -EACCES;
	}
1265 1266

	err = security_sem_semctl(sma, SETVAL);
1267 1268 1269 1270
	if (err) {
		rcu_read_unlock();
		return -EACCES;
	}
1271

1272
	sem_lock(sma, NULL, -1);
1273 1274 1275

	curr = &sma->sem_base[semnum];

1276
	ipc_assert_locked_object(&sma->sem_perm);
1277 1278 1279 1280 1281 1282 1283 1284
	list_for_each_entry(un, &sma->list_id, list_id)
		un->semadj[semnum] = 0;

	curr->semval = val;
	curr->sempid = task_tgid_vnr(current);
	sma->sem_ctime = get_seconds();
	/* maybe some queued-up processes were waiting for this */
	do_smart_update(sma, NULL, 0, 0, &tasks);
1285
	sem_unlock(sma, -1);
1286
	rcu_read_unlock();
1287
	wake_up_sem_queue_do(&tasks);
1288
	return 0;
1289 1290
}

K
Kirill Korotaev 已提交
1291
static int semctl_main(struct ipc_namespace *ns, int semid, int semnum,
1292
		int cmd, void __user *p)
L
Linus Torvalds 已提交
1293 1294 1295
{
	struct sem_array *sma;
	struct sem* curr;
1296
	int err, nsems;
L
Linus Torvalds 已提交
1297 1298
	ushort fast_sem_io[SEMMSL_FAST];
	ushort* sem_io = fast_sem_io;
1299
	struct list_head tasks;
L
Linus Torvalds 已提交
1300

1301 1302 1303 1304 1305 1306
	INIT_LIST_HEAD(&tasks);

	rcu_read_lock();
	sma = sem_obtain_object_check(ns, semid);
	if (IS_ERR(sma)) {
		rcu_read_unlock();
1307
		return PTR_ERR(sma);
1308
	}
L
Linus Torvalds 已提交
1309 1310 1311 1312

	nsems = sma->sem_nsems;

	err = -EACCES;
1313 1314
	if (ipcperms(ns, &sma->sem_perm, cmd == SETALL ? S_IWUGO : S_IRUGO))
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1315 1316

	err = security_sem_semctl(sma, cmd);
1317 1318
	if (err)
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1319 1320 1321 1322 1323

	err = -EACCES;
	switch (cmd) {
	case GETALL:
	{
1324
		ushort __user *array = p;
L
Linus Torvalds 已提交
1325 1326
		int i;

1327
		sem_lock(sma, NULL, -1);
L
Linus Torvalds 已提交
1328
		if(nsems > SEMMSL_FAST) {
1329 1330
			if (!ipc_rcu_getref(sma)) {
				sem_unlock(sma, -1);
1331
				rcu_read_unlock();
1332 1333 1334 1335
				err = -EIDRM;
				goto out_free;
			}
			sem_unlock(sma, -1);
1336
			rcu_read_unlock();
L
Linus Torvalds 已提交
1337 1338
			sem_io = ipc_alloc(sizeof(ushort)*nsems);
			if(sem_io == NULL) {
D
Davidlohr Bueso 已提交
1339
				ipc_rcu_putref(sma, ipc_rcu_free);
L
Linus Torvalds 已提交
1340 1341 1342
				return -ENOMEM;
			}

1343
			rcu_read_lock();
1344
			sem_lock_and_putref(sma);
L
Linus Torvalds 已提交
1345
			if (sma->sem_perm.deleted) {
1346
				sem_unlock(sma, -1);
1347
				rcu_read_unlock();
L
Linus Torvalds 已提交
1348 1349 1350
				err = -EIDRM;
				goto out_free;
			}
1351
		}
L
Linus Torvalds 已提交
1352 1353
		for (i = 0; i < sma->sem_nsems; i++)
			sem_io[i] = sma->sem_base[i].semval;
1354
		sem_unlock(sma, -1);
1355
		rcu_read_unlock();
L
Linus Torvalds 已提交
1356 1357 1358 1359 1360 1361 1362 1363 1364 1365
		err = 0;
		if(copy_to_user(array, sem_io, nsems*sizeof(ushort)))
			err = -EFAULT;
		goto out_free;
	}
	case SETALL:
	{
		int i;
		struct sem_undo *un;

1366 1367 1368 1369
		if (!ipc_rcu_getref(sma)) {
			rcu_read_unlock();
			return -EIDRM;
		}
1370
		rcu_read_unlock();
L
Linus Torvalds 已提交
1371 1372 1373 1374

		if(nsems > SEMMSL_FAST) {
			sem_io = ipc_alloc(sizeof(ushort)*nsems);
			if(sem_io == NULL) {
D
Davidlohr Bueso 已提交
1375
				ipc_rcu_putref(sma, ipc_rcu_free);
L
Linus Torvalds 已提交
1376 1377 1378 1379
				return -ENOMEM;
			}
		}

1380
		if (copy_from_user (sem_io, p, nsems*sizeof(ushort))) {
D
Davidlohr Bueso 已提交
1381
			ipc_rcu_putref(sma, ipc_rcu_free);
L
Linus Torvalds 已提交
1382 1383 1384 1385 1386 1387
			err = -EFAULT;
			goto out_free;
		}

		for (i = 0; i < nsems; i++) {
			if (sem_io[i] > SEMVMX) {
D
Davidlohr Bueso 已提交
1388
				ipc_rcu_putref(sma, ipc_rcu_free);
L
Linus Torvalds 已提交
1389 1390 1391 1392
				err = -ERANGE;
				goto out_free;
			}
		}
1393
		rcu_read_lock();
1394
		sem_lock_and_putref(sma);
L
Linus Torvalds 已提交
1395
		if (sma->sem_perm.deleted) {
1396
			sem_unlock(sma, -1);
1397
			rcu_read_unlock();
L
Linus Torvalds 已提交
1398 1399 1400 1401 1402 1403
			err = -EIDRM;
			goto out_free;
		}

		for (i = 0; i < nsems; i++)
			sma->sem_base[i].semval = sem_io[i];
1404

1405
		ipc_assert_locked_object(&sma->sem_perm);
1406
		list_for_each_entry(un, &sma->list_id, list_id) {
L
Linus Torvalds 已提交
1407 1408
			for (i = 0; i < nsems; i++)
				un->semadj[i] = 0;
1409
		}
L
Linus Torvalds 已提交
1410 1411
		sma->sem_ctime = get_seconds();
		/* maybe some queued-up processes were waiting for this */
1412
		do_smart_update(sma, NULL, 0, 0, &tasks);
L
Linus Torvalds 已提交
1413 1414 1415
		err = 0;
		goto out_unlock;
	}
1416
	/* GETVAL, GETPID, GETNCTN, GETZCNT: fall-through */
L
Linus Torvalds 已提交
1417 1418
	}
	err = -EINVAL;
1419 1420
	if (semnum < 0 || semnum >= nsems)
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1421

1422
	sem_lock(sma, NULL, -1);
L
Linus Torvalds 已提交
1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438
	curr = &sma->sem_base[semnum];

	switch (cmd) {
	case GETVAL:
		err = curr->semval;
		goto out_unlock;
	case GETPID:
		err = curr->sempid;
		goto out_unlock;
	case GETNCNT:
		err = count_semncnt(sma,semnum);
		goto out_unlock;
	case GETZCNT:
		err = count_semzcnt(sma,semnum);
		goto out_unlock;
	}
1439

L
Linus Torvalds 已提交
1440
out_unlock:
1441
	sem_unlock(sma, -1);
1442
out_rcu_wakeup:
1443
	rcu_read_unlock();
1444
	wake_up_sem_queue_do(&tasks);
L
Linus Torvalds 已提交
1445 1446 1447 1448 1449 1450
out_free:
	if(sem_io != fast_sem_io)
		ipc_free(sem_io, sizeof(ushort)*nsems);
	return err;
}

1451 1452
static inline unsigned long
copy_semid_from_user(struct semid64_ds *out, void __user *buf, int version)
L
Linus Torvalds 已提交
1453 1454 1455
{
	switch(version) {
	case IPC_64:
1456
		if (copy_from_user(out, buf, sizeof(*out)))
L
Linus Torvalds 已提交
1457 1458 1459 1460 1461 1462 1463 1464 1465
			return -EFAULT;
		return 0;
	case IPC_OLD:
	    {
		struct semid_ds tbuf_old;

		if(copy_from_user(&tbuf_old, buf, sizeof(tbuf_old)))
			return -EFAULT;

1466 1467 1468
		out->sem_perm.uid	= tbuf_old.sem_perm.uid;
		out->sem_perm.gid	= tbuf_old.sem_perm.gid;
		out->sem_perm.mode	= tbuf_old.sem_perm.mode;
L
Linus Torvalds 已提交
1469 1470 1471 1472 1473 1474 1475 1476

		return 0;
	    }
	default:
		return -EINVAL;
	}
}

1477
/*
D
Davidlohr Bueso 已提交
1478
 * This function handles some semctl commands which require the rwsem
1479
 * to be held in write mode.
D
Davidlohr Bueso 已提交
1480
 * NOTE: no locks must be held, the rwsem is taken inside this function.
1481
 */
1482
static int semctl_down(struct ipc_namespace *ns, int semid,
1483
		       int cmd, int version, void __user *p)
L
Linus Torvalds 已提交
1484 1485 1486
{
	struct sem_array *sma;
	int err;
1487
	struct semid64_ds semid64;
L
Linus Torvalds 已提交
1488 1489 1490
	struct kern_ipc_perm *ipcp;

	if(cmd == IPC_SET) {
1491
		if (copy_semid_from_user(&semid64, p, version))
L
Linus Torvalds 已提交
1492 1493
			return -EFAULT;
	}
S
Steve Grubb 已提交
1494

D
Davidlohr Bueso 已提交
1495
	down_write(&sem_ids(ns).rwsem);
1496 1497
	rcu_read_lock();

1498 1499
	ipcp = ipcctl_pre_down_nolock(ns, &sem_ids(ns), semid, cmd,
				      &semid64.sem_perm, 0);
1500 1501 1502 1503
	if (IS_ERR(ipcp)) {
		err = PTR_ERR(ipcp);
		goto out_unlock1;
	}
S
Steve Grubb 已提交
1504

1505
	sma = container_of(ipcp, struct sem_array, sem_perm);
L
Linus Torvalds 已提交
1506 1507

	err = security_sem_semctl(sma, cmd);
1508 1509
	if (err)
		goto out_unlock1;
L
Linus Torvalds 已提交
1510

1511
	switch (cmd) {
L
Linus Torvalds 已提交
1512
	case IPC_RMID:
1513
		sem_lock(sma, NULL, -1);
1514
		/* freeary unlocks the ipc object and rcu */
1515
		freeary(ns, ipcp);
1516
		goto out_up;
L
Linus Torvalds 已提交
1517
	case IPC_SET:
1518
		sem_lock(sma, NULL, -1);
1519 1520
		err = ipc_update_perm(&semid64.sem_perm, ipcp);
		if (err)
1521
			goto out_unlock0;
L
Linus Torvalds 已提交
1522 1523 1524 1525
		sma->sem_ctime = get_seconds();
		break;
	default:
		err = -EINVAL;
1526
		goto out_unlock1;
L
Linus Torvalds 已提交
1527 1528
	}

1529
out_unlock0:
1530
	sem_unlock(sma, -1);
1531
out_unlock1:
1532
	rcu_read_unlock();
1533
out_up:
D
Davidlohr Bueso 已提交
1534
	up_write(&sem_ids(ns).rwsem);
L
Linus Torvalds 已提交
1535 1536 1537
	return err;
}

1538
SYSCALL_DEFINE4(semctl, int, semid, int, semnum, int, cmd, unsigned long, arg)
L
Linus Torvalds 已提交
1539 1540
{
	int version;
K
Kirill Korotaev 已提交
1541
	struct ipc_namespace *ns;
1542
	void __user *p = (void __user *)arg;
L
Linus Torvalds 已提交
1543 1544 1545 1546 1547

	if (semid < 0)
		return -EINVAL;

	version = ipc_parse_version(&cmd);
K
Kirill Korotaev 已提交
1548
	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
1549 1550 1551 1552

	switch(cmd) {
	case IPC_INFO:
	case SEM_INFO:
1553
	case IPC_STAT:
L
Linus Torvalds 已提交
1554
	case SEM_STAT:
1555
		return semctl_nolock(ns, semid, cmd, version, p);
L
Linus Torvalds 已提交
1556 1557 1558 1559 1560 1561
	case GETALL:
	case GETVAL:
	case GETPID:
	case GETNCNT:
	case GETZCNT:
	case SETALL:
1562 1563 1564
		return semctl_main(ns, semid, semnum, cmd, p);
	case SETVAL:
		return semctl_setval(ns, semid, semnum, arg);
L
Linus Torvalds 已提交
1565 1566
	case IPC_RMID:
	case IPC_SET:
1567
		return semctl_down(ns, semid, cmd, version, p);
L
Linus Torvalds 已提交
1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589
	default:
		return -EINVAL;
	}
}

/* If the task doesn't already have a undo_list, then allocate one
 * here.  We guarantee there is only one thread using this undo list,
 * and current is THE ONE
 *
 * If this allocation and assignment succeeds, but later
 * portions of this code fail, there is no need to free the sem_undo_list.
 * Just let it stay associated with the task, and it'll be freed later
 * at exit time.
 *
 * This can block, so callers must hold no locks.
 */
static inline int get_undo_list(struct sem_undo_list **undo_listp)
{
	struct sem_undo_list *undo_list;

	undo_list = current->sysvsem.undo_list;
	if (!undo_list) {
1590
		undo_list = kzalloc(sizeof(*undo_list), GFP_KERNEL);
L
Linus Torvalds 已提交
1591 1592
		if (undo_list == NULL)
			return -ENOMEM;
I
Ingo Molnar 已提交
1593
		spin_lock_init(&undo_list->lock);
L
Linus Torvalds 已提交
1594
		atomic_set(&undo_list->refcnt, 1);
1595 1596
		INIT_LIST_HEAD(&undo_list->list_proc);

L
Linus Torvalds 已提交
1597 1598 1599 1600 1601 1602
		current->sysvsem.undo_list = undo_list;
	}
	*undo_listp = undo_list;
	return 0;
}

1603
static struct sem_undo *__lookup_undo(struct sem_undo_list *ulp, int semid)
L
Linus Torvalds 已提交
1604
{
1605
	struct sem_undo *un;
1606

1607 1608 1609
	list_for_each_entry_rcu(un, &ulp->list_proc, list_proc) {
		if (un->semid == semid)
			return un;
L
Linus Torvalds 已提交
1610
	}
1611
	return NULL;
L
Linus Torvalds 已提交
1612 1613
}

1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627
static struct sem_undo *lookup_undo(struct sem_undo_list *ulp, int semid)
{
	struct sem_undo *un;

  	assert_spin_locked(&ulp->lock);

	un = __lookup_undo(ulp, semid);
	if (un) {
		list_del_rcu(&un->list_proc);
		list_add_rcu(&un->list_proc, &ulp->list_proc);
	}
	return un;
}

1628 1629 1630 1631 1632 1633 1634 1635
/**
 * find_alloc_undo - Lookup (and if not present create) undo array
 * @ns: namespace
 * @semid: semaphore array id
 *
 * The function looks up (and if not present creates) the undo structure.
 * The size of the undo structure depends on the size of the semaphore
 * array, thus the alloc path is not that straightforward.
1636 1637
 * Lifetime-rules: sem_undo is rcu-protected, on success, the function
 * performs a rcu_read_lock().
1638 1639
 */
static struct sem_undo *find_alloc_undo(struct ipc_namespace *ns, int semid)
L
Linus Torvalds 已提交
1640 1641 1642 1643
{
	struct sem_array *sma;
	struct sem_undo_list *ulp;
	struct sem_undo *un, *new;
1644
	int nsems, error;
L
Linus Torvalds 已提交
1645 1646 1647 1648 1649

	error = get_undo_list(&ulp);
	if (error)
		return ERR_PTR(error);

1650
	rcu_read_lock();
1651
	spin_lock(&ulp->lock);
L
Linus Torvalds 已提交
1652
	un = lookup_undo(ulp, semid);
1653
	spin_unlock(&ulp->lock);
L
Linus Torvalds 已提交
1654 1655 1656 1657
	if (likely(un!=NULL))
		goto out;

	/* no undo structure around - allocate one. */
1658
	/* step 1: figure out the size of the semaphore array */
1659 1660 1661
	sma = sem_obtain_object_check(ns, semid);
	if (IS_ERR(sma)) {
		rcu_read_unlock();
J
Julia Lawall 已提交
1662
		return ERR_CAST(sma);
1663
	}
1664

L
Linus Torvalds 已提交
1665
	nsems = sma->sem_nsems;
1666 1667 1668 1669 1670
	if (!ipc_rcu_getref(sma)) {
		rcu_read_unlock();
		un = ERR_PTR(-EIDRM);
		goto out;
	}
1671
	rcu_read_unlock();
L
Linus Torvalds 已提交
1672

1673
	/* step 2: allocate new undo structure */
1674
	new = kzalloc(sizeof(struct sem_undo) + sizeof(short)*nsems, GFP_KERNEL);
L
Linus Torvalds 已提交
1675
	if (!new) {
D
Davidlohr Bueso 已提交
1676
		ipc_rcu_putref(sma, ipc_rcu_free);
L
Linus Torvalds 已提交
1677 1678 1679
		return ERR_PTR(-ENOMEM);
	}

1680
	/* step 3: Acquire the lock on semaphore array */
1681
	rcu_read_lock();
1682
	sem_lock_and_putref(sma);
L
Linus Torvalds 已提交
1683
	if (sma->sem_perm.deleted) {
1684
		sem_unlock(sma, -1);
1685
		rcu_read_unlock();
L
Linus Torvalds 已提交
1686 1687 1688 1689
		kfree(new);
		un = ERR_PTR(-EIDRM);
		goto out;
	}
1690 1691 1692 1693 1694 1695 1696 1697 1698 1699
	spin_lock(&ulp->lock);

	/*
	 * step 4: check for races: did someone else allocate the undo struct?
	 */
	un = lookup_undo(ulp, semid);
	if (un) {
		kfree(new);
		goto success;
	}
1700 1701
	/* step 5: initialize & link new undo structure */
	new->semadj = (short *) &new[1];
1702
	new->ulp = ulp;
1703 1704
	new->semid = semid;
	assert_spin_locked(&ulp->lock);
1705
	list_add_rcu(&new->list_proc, &ulp->list_proc);
1706
	ipc_assert_locked_object(&sma->sem_perm);
1707
	list_add(&new->list_id, &sma->list_id);
1708
	un = new;
1709

1710
success:
1711
	spin_unlock(&ulp->lock);
1712
	sem_unlock(sma, -1);
L
Linus Torvalds 已提交
1713 1714 1715 1716
out:
	return un;
}

1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742

/**
 * get_queue_result - Retrieve the result code from sem_queue
 * @q: Pointer to queue structure
 *
 * Retrieve the return code from the pending queue. If IN_WAKEUP is found in
 * q->status, then we must loop until the value is replaced with the final
 * value: This may happen if a task is woken up by an unrelated event (e.g.
 * signal) and in parallel the task is woken up by another task because it got
 * the requested semaphores.
 *
 * The function can be called with or without holding the semaphore spinlock.
 */
static int get_queue_result(struct sem_queue *q)
{
	int error;

	error = q->status;
	while (unlikely(error == IN_WAKEUP)) {
		cpu_relax();
		error = q->status;
	}

	return error;
}

1743 1744
SYSCALL_DEFINE4(semtimedop, int, semid, struct sembuf __user *, tsops,
		unsigned, nsops, const struct timespec __user *, timeout)
L
Linus Torvalds 已提交
1745 1746 1747 1748 1749 1750
{
	int error = -EINVAL;
	struct sem_array *sma;
	struct sembuf fast_sops[SEMOPM_FAST];
	struct sembuf* sops = fast_sops, *sop;
	struct sem_undo *un;
1751
	int undos = 0, alter = 0, max, locknum;
L
Linus Torvalds 已提交
1752 1753
	struct sem_queue queue;
	unsigned long jiffies_left = 0;
K
Kirill Korotaev 已提交
1754
	struct ipc_namespace *ns;
1755
	struct list_head tasks;
K
Kirill Korotaev 已提交
1756 1757

	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
1758 1759 1760

	if (nsops < 1 || semid < 0)
		return -EINVAL;
K
Kirill Korotaev 已提交
1761
	if (nsops > ns->sc_semopm)
L
Linus Torvalds 已提交
1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789
		return -E2BIG;
	if(nsops > SEMOPM_FAST) {
		sops = kmalloc(sizeof(*sops)*nsops,GFP_KERNEL);
		if(sops==NULL)
			return -ENOMEM;
	}
	if (copy_from_user (sops, tsops, nsops * sizeof(*tsops))) {
		error=-EFAULT;
		goto out_free;
	}
	if (timeout) {
		struct timespec _timeout;
		if (copy_from_user(&_timeout, timeout, sizeof(*timeout))) {
			error = -EFAULT;
			goto out_free;
		}
		if (_timeout.tv_sec < 0 || _timeout.tv_nsec < 0 ||
			_timeout.tv_nsec >= 1000000000L) {
			error = -EINVAL;
			goto out_free;
		}
		jiffies_left = timespec_to_jiffies(&_timeout);
	}
	max = 0;
	for (sop = sops; sop < sops + nsops; sop++) {
		if (sop->sem_num >= max)
			max = sop->sem_num;
		if (sop->sem_flg & SEM_UNDO)
1790 1791
			undos = 1;
		if (sop->sem_op != 0)
L
Linus Torvalds 已提交
1792 1793 1794
			alter = 1;
	}

1795 1796
	INIT_LIST_HEAD(&tasks);

L
Linus Torvalds 已提交
1797
	if (undos) {
1798
		/* On success, find_alloc_undo takes the rcu_read_lock */
1799
		un = find_alloc_undo(ns, semid);
L
Linus Torvalds 已提交
1800 1801 1802 1803
		if (IS_ERR(un)) {
			error = PTR_ERR(un);
			goto out_free;
		}
1804
	} else {
L
Linus Torvalds 已提交
1805
		un = NULL;
1806 1807
		rcu_read_lock();
	}
L
Linus Torvalds 已提交
1808

1809
	sma = sem_obtain_object_check(ns, semid);
1810
	if (IS_ERR(sma)) {
1811
		rcu_read_unlock();
1812
		error = PTR_ERR(sma);
L
Linus Torvalds 已提交
1813
		goto out_free;
1814 1815
	}

1816
	error = -EFBIG;
1817 1818
	if (max >= sma->sem_nsems)
		goto out_rcu_wakeup;
1819 1820

	error = -EACCES;
1821 1822
	if (ipcperms(ns, &sma->sem_perm, alter ? S_IWUGO : S_IRUGO))
		goto out_rcu_wakeup;
1823 1824

	error = security_sem_semop(sma, sops, nsops, alter);
1825 1826
	if (error)
		goto out_rcu_wakeup;
1827

L
Linus Torvalds 已提交
1828
	/*
1829
	 * semid identifiers are not unique - find_alloc_undo may have
L
Linus Torvalds 已提交
1830
	 * allocated an undo structure, it was invalidated by an RMID
1831
	 * and now a new array with received the same id. Check and fail.
L
Lucas De Marchi 已提交
1832
	 * This case can be detected checking un->semid. The existence of
1833
	 * "un" itself is guaranteed by rcu.
L
Linus Torvalds 已提交
1834
	 */
1835
	error = -EIDRM;
1836 1837 1838
	locknum = sem_lock(sma, sops, nsops);
	if (un && un->semid == -1)
		goto out_unlock_free;
1839

1840 1841
	error = perform_atomic_semop(sma, sops, nsops, un,
					task_tgid_vnr(current));
L
Linus Torvalds 已提交
1842 1843
	if (error <= 0) {
		if (alter && error == 0)
1844
			do_smart_update(sma, sops, nsops, 1, &tasks);
1845

L
Linus Torvalds 已提交
1846 1847 1848 1849 1850 1851 1852 1853 1854 1855
		goto out_unlock_free;
	}

	/* We need to sleep on this operation, so we put the current
	 * task into the pending queue and go to sleep.
	 */
		
	queue.sops = sops;
	queue.nsops = nsops;
	queue.undo = un;
1856
	queue.pid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
1857 1858
	queue.alter = alter;

1859 1860 1861 1862
	if (nsops == 1) {
		struct sem *curr;
		curr = &sma->sem_base[sops->sem_num];

1863 1864 1865 1866 1867 1868 1869 1870 1871 1872
		if (alter) {
			if (sma->complex_count) {
				list_add_tail(&queue.list,
						&sma->pending_alter);
			} else {

				list_add_tail(&queue.list,
						&curr->pending_alter);
			}
		} else {
1873
			list_add_tail(&queue.list, &curr->pending_const);
1874
		}
1875
	} else {
1876 1877 1878
		if (!sma->complex_count)
			merge_queues(sma);

1879
		if (alter)
1880
			list_add_tail(&queue.list, &sma->pending_alter);
1881
		else
1882 1883
			list_add_tail(&queue.list, &sma->pending_const);

1884 1885 1886
		sma->complex_count++;
	}

L
Linus Torvalds 已提交
1887 1888
	queue.status = -EINTR;
	queue.sleeper = current;
1889 1890

sleep_again:
L
Linus Torvalds 已提交
1891
	current->state = TASK_INTERRUPTIBLE;
1892
	sem_unlock(sma, locknum);
1893
	rcu_read_unlock();
L
Linus Torvalds 已提交
1894 1895 1896 1897 1898 1899

	if (timeout)
		jiffies_left = schedule_timeout(jiffies_left);
	else
		schedule();

1900
	error = get_queue_result(&queue);
L
Linus Torvalds 已提交
1901 1902 1903

	if (error != -EINTR) {
		/* fast path: update_queue already obtained all requested
1904 1905 1906 1907 1908 1909 1910 1911
		 * resources.
		 * Perform a smp_mb(): User space could assume that semop()
		 * is a memory barrier: Without the mb(), the cpu could
		 * speculatively read in user space stale data that was
		 * overwritten by the previous owner of the semaphore.
		 */
		smp_mb();

L
Linus Torvalds 已提交
1912 1913 1914
		goto out_free;
	}

1915
	rcu_read_lock();
1916
	sma = sem_obtain_lock(ns, semid, sops, nsops, &locknum);
1917 1918 1919 1920 1921 1922 1923 1924 1925

	/*
	 * Wait until it's guaranteed that no wakeup_sem_queue_do() is ongoing.
	 */
	error = get_queue_result(&queue);

	/*
	 * Array removed? If yes, leave without sem_unlock().
	 */
1926
	if (IS_ERR(sma)) {
1927
		rcu_read_unlock();
L
Linus Torvalds 已提交
1928 1929 1930
		goto out_free;
	}

1931

L
Linus Torvalds 已提交
1932
	/*
1933 1934
	 * If queue.status != -EINTR we are woken up by another process.
	 * Leave without unlink_queue(), but with sem_unlock().
L
Linus Torvalds 已提交
1935
	 */
1936

L
Linus Torvalds 已提交
1937 1938 1939 1940 1941 1942 1943 1944 1945
	if (error != -EINTR) {
		goto out_unlock_free;
	}

	/*
	 * If an interrupt occurred we have to clean up the queue
	 */
	if (timeout && jiffies_left == 0)
		error = -EAGAIN;
1946 1947 1948 1949 1950 1951 1952

	/*
	 * If the wakeup was spurious, just retry
	 */
	if (error == -EINTR && !signal_pending(current))
		goto sleep_again;

1953
	unlink_queue(sma, &queue);
L
Linus Torvalds 已提交
1954 1955

out_unlock_free:
1956
	sem_unlock(sma, locknum);
1957
out_rcu_wakeup:
1958
	rcu_read_unlock();
1959
	wake_up_sem_queue_do(&tasks);
L
Linus Torvalds 已提交
1960 1961 1962 1963 1964 1965
out_free:
	if(sops != fast_sops)
		kfree(sops);
	return error;
}

1966 1967
SYSCALL_DEFINE3(semop, int, semid, struct sembuf __user *, tsops,
		unsigned, nsops)
L
Linus Torvalds 已提交
1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
{
	return sys_semtimedop(semid, tsops, nsops, NULL);
}

/* If CLONE_SYSVSEM is set, establish sharing of SEM_UNDO state between
 * parent and child tasks.
 */

int copy_semundo(unsigned long clone_flags, struct task_struct *tsk)
{
	struct sem_undo_list *undo_list;
	int error;

	if (clone_flags & CLONE_SYSVSEM) {
		error = get_undo_list(&undo_list);
		if (error)
			return error;
		atomic_inc(&undo_list->refcnt);
		tsk->sysvsem.undo_list = undo_list;
	} else 
		tsk->sysvsem.undo_list = NULL;

	return 0;
}

/*
 * add semadj values to semaphores, free undo structures.
 * undo structures are not freed when semaphore arrays are destroyed
 * so some of them may be out of date.
 * IMPLEMENTATION NOTE: There is some confusion over whether the
 * set of adjustments that needs to be done should be done in an atomic
 * manner or not. That is, if we are attempting to decrement the semval
 * should we queue up and wait until we can do so legally?
 * The original implementation attempted to do this (queue and wait).
 * The current implementation does not do so. The POSIX standard
 * and SVID should be consulted to determine what behavior is mandated.
 */
void exit_sem(struct task_struct *tsk)
{
2007
	struct sem_undo_list *ulp;
L
Linus Torvalds 已提交
2008

2009 2010
	ulp = tsk->sysvsem.undo_list;
	if (!ulp)
L
Linus Torvalds 已提交
2011
		return;
2012
	tsk->sysvsem.undo_list = NULL;
L
Linus Torvalds 已提交
2013

2014
	if (!atomic_dec_and_test(&ulp->refcnt))
L
Linus Torvalds 已提交
2015 2016
		return;

2017
	for (;;) {
L
Linus Torvalds 已提交
2018
		struct sem_array *sma;
2019
		struct sem_undo *un;
2020
		struct list_head tasks;
2021
		int semid, i;
2022

2023
		rcu_read_lock();
2024 2025
		un = list_entry_rcu(ulp->list_proc.next,
				    struct sem_undo, list_proc);
2026 2027 2028 2029
		if (&un->list_proc == &ulp->list_proc)
			semid = -1;
		 else
			semid = un->semid;
2030

2031 2032
		if (semid == -1) {
			rcu_read_unlock();
2033
			break;
2034
		}
L
Linus Torvalds 已提交
2035

2036
		sma = sem_obtain_object_check(tsk->nsproxy->ipc_ns, un->semid);
2037
		/* exit_sem raced with IPC_RMID, nothing to do */
2038 2039
		if (IS_ERR(sma)) {
			rcu_read_unlock();
2040
			continue;
2041
		}
L
Linus Torvalds 已提交
2042

2043
		sem_lock(sma, NULL, -1);
2044
		un = __lookup_undo(ulp, semid);
2045 2046 2047 2048
		if (un == NULL) {
			/* exit_sem raced with IPC_RMID+semget() that created
			 * exactly the same semid. Nothing to do.
			 */
2049
			sem_unlock(sma, -1);
2050
			rcu_read_unlock();
2051 2052 2053 2054
			continue;
		}

		/* remove un from the linked lists */
2055
		ipc_assert_locked_object(&sma->sem_perm);
2056 2057
		list_del(&un->list_id);

2058 2059 2060 2061
		spin_lock(&ulp->lock);
		list_del_rcu(&un->list_proc);
		spin_unlock(&ulp->lock);

2062 2063
		/* perform adjustments registered in un */
		for (i = 0; i < sma->sem_nsems; i++) {
I
Ingo Molnar 已提交
2064
			struct sem * semaphore = &sma->sem_base[i];
2065 2066
			if (un->semadj[i]) {
				semaphore->semval += un->semadj[i];
L
Linus Torvalds 已提交
2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079
				/*
				 * Range checks of the new semaphore value,
				 * not defined by sus:
				 * - Some unices ignore the undo entirely
				 *   (e.g. HP UX 11i 11.22, Tru64 V5.1)
				 * - some cap the value (e.g. FreeBSD caps
				 *   at 0, but doesn't enforce SEMVMX)
				 *
				 * Linux caps the semaphore value, both at 0
				 * and at SEMVMX.
				 *
				 * 	Manfred <manfred@colorfullife.com>
				 */
I
Ingo Molnar 已提交
2080 2081 2082 2083
				if (semaphore->semval < 0)
					semaphore->semval = 0;
				if (semaphore->semval > SEMVMX)
					semaphore->semval = SEMVMX;
2084
				semaphore->sempid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
2085 2086 2087
			}
		}
		/* maybe some queued-up processes were waiting for this */
2088 2089
		INIT_LIST_HEAD(&tasks);
		do_smart_update(sma, NULL, 0, 1, &tasks);
2090
		sem_unlock(sma, -1);
2091
		rcu_read_unlock();
2092
		wake_up_sem_queue_do(&tasks);
2093

2094
		kfree_rcu(un, rcu);
L
Linus Torvalds 已提交
2095
	}
2096
	kfree(ulp);
L
Linus Torvalds 已提交
2097 2098 2099
}

#ifdef CONFIG_PROC_FS
2100
static int sysvipc_sem_proc_show(struct seq_file *s, void *it)
L
Linus Torvalds 已提交
2101
{
2102
	struct user_namespace *user_ns = seq_user_ns(s);
2103
	struct sem_array *sma = it;
2104 2105
	time_t sem_otime;

2106 2107 2108 2109 2110 2111 2112 2113
	/*
	 * The proc interface isn't aware of sem_lock(), it calls
	 * ipc_lock_object() directly (in sysvipc_find_ipc).
	 * In order to stay compatible with sem_lock(), we must wait until
	 * all simple semop() calls have left their critical regions.
	 */
	sem_wait_array(sma);

2114
	sem_otime = get_semotime(sma);
2115 2116

	return seq_printf(s,
2117
			  "%10d %10d  %4o %10u %5u %5u %5u %5u %10lu %10lu\n",
2118
			  sma->sem_perm.key,
N
Nadia Derbey 已提交
2119
			  sma->sem_perm.id,
2120 2121
			  sma->sem_perm.mode,
			  sma->sem_nsems,
2122 2123 2124 2125
			  from_kuid_munged(user_ns, sma->sem_perm.uid),
			  from_kgid_munged(user_ns, sma->sem_perm.gid),
			  from_kuid_munged(user_ns, sma->sem_perm.cuid),
			  from_kgid_munged(user_ns, sma->sem_perm.cgid),
2126
			  sem_otime,
2127
			  sma->sem_ctime);
L
Linus Torvalds 已提交
2128 2129
}
#endif