sem.c 50.0 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
} ____cacheline_aligned_in_smp;
103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141

/* 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;
};


142
#define sem_ids(ns)	((ns)->ids[IPC_SEM_IDS])
K
Kirill Korotaev 已提交
143

N
Nadia Derbey 已提交
144
#define sem_checkid(sma, semid)	ipc_checkid(&sma->sem_perm, semid)
L
Linus Torvalds 已提交
145

N
Nadia Derbey 已提交
146
static int newary(struct ipc_namespace *, struct ipc_params *);
147
static void freeary(struct ipc_namespace *, struct kern_ipc_perm *);
L
Linus Torvalds 已提交
148
#ifdef CONFIG_PROC_FS
149
static int sysvipc_sem_proc_show(struct seq_file *s, void *it);
L
Linus Torvalds 已提交
150 151 152 153 154 155 156 157
#endif

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

/*
 * linked list protection:
 *	sem_undo.id_next,
158
 *	sem_array.pending{_alter,_cont},
L
Linus Torvalds 已提交
159 160 161 162 163
 *	sem_array.sem_undo: sem_lock() for read/write
 *	sem_undo.proc_next: only "current" is allowed to read/write that field.
 *	
 */

K
Kirill Korotaev 已提交
164 165 166 167 168
#define sc_semmsl	sem_ctls[0]
#define sc_semmns	sem_ctls[1]
#define sc_semopm	sem_ctls[2]
#define sc_semmni	sem_ctls[3]

169
void sem_init_ns(struct ipc_namespace *ns)
K
Kirill Korotaev 已提交
170 171 172 173 174 175
{
	ns->sc_semmsl = SEMMSL;
	ns->sc_semmns = SEMMNS;
	ns->sc_semopm = SEMOPM;
	ns->sc_semmni = SEMMNI;
	ns->used_sems = 0;
176
	ipc_init_ids(&ns->ids[IPC_SEM_IDS]);
K
Kirill Korotaev 已提交
177 178
}

179
#ifdef CONFIG_IPC_NS
K
Kirill Korotaev 已提交
180 181
void sem_exit_ns(struct ipc_namespace *ns)
{
182
	free_ipcs(ns, &sem_ids(ns), freeary);
S
Serge E. Hallyn 已提交
183
	idr_destroy(&ns->ids[IPC_SEM_IDS].ipcs_idr);
K
Kirill Korotaev 已提交
184
}
185
#endif
L
Linus Torvalds 已提交
186 187 188

void __init sem_init (void)
{
189
	sem_init_ns(&init_ipc_ns);
190 191
	ipc_init_proc_interface("sysvipc/sem",
				"       key      semid perms      nsems   uid   gid  cuid  cgid      otime      ctime\n",
K
Kirill Korotaev 已提交
192
				IPC_SEM_IDS, sysvipc_sem_proc_show);
L
Linus Torvalds 已提交
193 194
}

195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251
/*
 * 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.
 *
 * Carefully guard against sma->complex_count changing between zero
 * and non-zero while we are spinning for the lock. The value of
 * sma->complex_count cannot change while we are holding the lock,
 * so sem_unlock should be fine.
 *
 * The global lock path checks that all the local locks have been released,
 * checking each local lock once. This means that the local lock paths
 * cannot start their critical sections while the global lock is held.
 */
static inline int sem_lock(struct sem_array *sma, struct sembuf *sops,
			      int nsops)
{
	int locknum;
 again:
	if (nsops == 1 && !sma->complex_count) {
		struct sem *sem = sma->sem_base + sops->sem_num;

		/* Lock just the semaphore we are interested in. */
		spin_lock(&sem->lock);

		/*
		 * If sma->complex_count was set while we were spinning,
		 * we may need to look at things we did not lock here.
		 */
		if (unlikely(sma->complex_count)) {
			spin_unlock(&sem->lock);
			goto lock_array;
		}

		/*
		 * Another process is holding the global lock on the
		 * sem_array; we cannot enter our critical section,
		 * but have to wait for the global lock to be released.
		 */
		if (unlikely(spin_is_locked(&sma->sem_perm.lock))) {
			spin_unlock(&sem->lock);
			spin_unlock_wait(&sma->sem_perm.lock);
			goto again;
		}

		locknum = sops->sem_num;
	} else {
		int i;
		/*
		 * Lock the semaphore array, and wait for all of the
		 * individual semaphore locks to go away.  The code
		 * above ensures no new single-lock holders will enter
		 * their critical section while the array lock is held.
		 */
 lock_array:
252
		ipc_lock_object(&sma->sem_perm);
253 254 255 256 257 258 259 260 261 262 263 264
		for (i = 0; i < sma->sem_nsems; i++) {
			struct sem *sem = sma->sem_base + i;
			spin_unlock_wait(&sem->lock);
		}
		locknum = -1;
	}
	return locknum;
}

static inline void sem_unlock(struct sem_array *sma, int locknum)
{
	if (locknum == -1) {
265
		ipc_unlock_object(&sma->sem_perm);
266 267 268 269 270 271
	} else {
		struct sem *sem = sma->sem_base + locknum;
		spin_unlock(&sem->lock);
	}
}

N
Nadia Derbey 已提交
272 273 274
/*
 * sem_lock_(check_) routines are called in the paths where the rw_mutex
 * is not held.
275 276
 *
 * The caller holds the RCU read lock.
N
Nadia Derbey 已提交
277
 */
278 279
static inline struct sem_array *sem_obtain_lock(struct ipc_namespace *ns,
			int id, struct sembuf *sops, int nsops, int *locknum)
280
{
281 282
	struct kern_ipc_perm *ipcp;
	struct sem_array *sma;
N
Nadia Derbey 已提交
283

284
	ipcp = ipc_obtain_object(&sem_ids(ns), id);
285 286
	if (IS_ERR(ipcp))
		return ERR_CAST(ipcp);
287

288 289
	sma = container_of(ipcp, struct sem_array, sem_perm);
	*locknum = sem_lock(sma, sops, nsops);
290 291 292 293 294 295 296

	/* 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);

297
	sem_unlock(sma, *locknum);
298
	return ERR_PTR(-EINVAL);
299 300
}

301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317
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);
318

N
Nadia Derbey 已提交
319
	return container_of(ipcp, struct sem_array, sem_perm);
320 321
}

322 323
static inline void sem_lock_and_putref(struct sem_array *sma)
{
324
	sem_lock(sma, NULL, -1);
325 326 327 328 329
	ipc_rcu_putref(sma);
}

static inline void sem_putref(struct sem_array *sma)
{
330
	ipc_rcu_putref(sma);
331 332
}

N
Nadia Derbey 已提交
333 334 335 336 337
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
338 339 340 341 342
/*
 * 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
343
 *	* unlinking the queue entry from the pending list
L
Linus Torvalds 已提交
344 345 346 347 348 349 350 351 352
 *	* 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
353
 *   	  performing any operation on the sem array.
L
Linus Torvalds 已提交
354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371
 *   	* 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
372 373 374 375 376
/**
 * newary - Create a new semaphore set
 * @ns: namespace
 * @params: ptr to the structure that contains key, semflg and nsems
 *
N
Nadia Derbey 已提交
377
 * Called with sem_ids.rw_mutex held (as a writer)
N
Nadia Derbey 已提交
378 379
 */

N
Nadia Derbey 已提交
380
static int newary(struct ipc_namespace *ns, struct ipc_params *params)
L
Linus Torvalds 已提交
381 382 383 384 385
{
	int id;
	int retval;
	struct sem_array *sma;
	int size;
N
Nadia Derbey 已提交
386 387 388
	key_t key = params->key;
	int nsems = params->u.nsems;
	int semflg = params->flg;
389
	int i;
L
Linus Torvalds 已提交
390 391 392

	if (!nsems)
		return -EINVAL;
K
Kirill Korotaev 已提交
393
	if (ns->used_sems + nsems > ns->sc_semmns)
L
Linus Torvalds 已提交
394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
		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) {
		ipc_rcu_putref(sma);
		return retval;
	}

K
Kirill Korotaev 已提交
413
	id = ipc_addid(&sem_ids(ns), &sma->sem_perm, ns->sc_semmni);
414
	if (id < 0) {
L
Linus Torvalds 已提交
415 416
		security_sem_free(sma);
		ipc_rcu_putref(sma);
417
		return id;
L
Linus Torvalds 已提交
418
	}
K
Kirill Korotaev 已提交
419
	ns->used_sems += nsems;
L
Linus Torvalds 已提交
420 421

	sma->sem_base = (struct sem *) &sma[1];
422

423
	for (i = 0; i < nsems; i++) {
424 425
		INIT_LIST_HEAD(&sma->sem_base[i].pending_alter);
		INIT_LIST_HEAD(&sma->sem_base[i].pending_const);
426 427
		spin_lock_init(&sma->sem_base[i].lock);
	}
428 429

	sma->complex_count = 0;
430 431
	INIT_LIST_HEAD(&sma->pending_alter);
	INIT_LIST_HEAD(&sma->pending_const);
432
	INIT_LIST_HEAD(&sma->list_id);
L
Linus Torvalds 已提交
433 434
	sma->sem_nsems = nsems;
	sma->sem_ctime = get_seconds();
435
	sem_unlock(sma, -1);
436
	rcu_read_unlock();
L
Linus Torvalds 已提交
437

N
Nadia Derbey 已提交
438
	return sma->sem_perm.id;
L
Linus Torvalds 已提交
439 440
}

N
Nadia Derbey 已提交
441

N
Nadia Derbey 已提交
442
/*
N
Nadia Derbey 已提交
443
 * Called with sem_ids.rw_mutex and ipcp locked.
N
Nadia Derbey 已提交
444
 */
N
Nadia Derbey 已提交
445
static inline int sem_security(struct kern_ipc_perm *ipcp, int semflg)
N
Nadia Derbey 已提交
446
{
N
Nadia Derbey 已提交
447 448 449 450
	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
451 452
}

N
Nadia Derbey 已提交
453
/*
N
Nadia Derbey 已提交
454
 * Called with sem_ids.rw_mutex and ipcp locked.
N
Nadia Derbey 已提交
455
 */
N
Nadia Derbey 已提交
456 457
static inline int sem_more_checks(struct kern_ipc_perm *ipcp,
				struct ipc_params *params)
N
Nadia Derbey 已提交
458
{
N
Nadia Derbey 已提交
459 460 461 462
	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
463 464 465 466 467
		return -EINVAL;

	return 0;
}

468
SYSCALL_DEFINE3(semget, key_t, key, int, nsems, int, semflg)
L
Linus Torvalds 已提交
469
{
K
Kirill Korotaev 已提交
470
	struct ipc_namespace *ns;
N
Nadia Derbey 已提交
471 472
	struct ipc_ops sem_ops;
	struct ipc_params sem_params;
K
Kirill Korotaev 已提交
473 474

	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
475

K
Kirill Korotaev 已提交
476
	if (nsems < 0 || nsems > ns->sc_semmsl)
L
Linus Torvalds 已提交
477
		return -EINVAL;
N
Nadia Derbey 已提交
478

N
Nadia Derbey 已提交
479 480 481 482 483 484 485
	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
486

N
Nadia Derbey 已提交
487
	return ipcget(ns, &sem_ids(ns), &sem_ops, &sem_params);
L
Linus Torvalds 已提交
488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555
}

/*
 * Determine whether a sequence of semaphore operations would succeed
 * all at once. Return 0 if yes, 1 if need to sleep, else return error code.
 */

static int try_atomic_semop (struct sem_array * sma, struct sembuf * sops,
			     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;
}

556 557 558 559 560
/** 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
561
 */
562 563
static void wake_up_sem_queue_prepare(struct list_head *pt,
				struct sem_queue *q, int error)
N
Nick Piggin 已提交
564
{
565 566 567 568 569 570 571
	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
572
	q->status = IN_WAKEUP;
573 574
	q->pid = error;

575
	list_add_tail(&q->list, pt);
576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592
}

/**
 * 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);
593
	list_for_each_entry_safe(q, t, pt, list) {
594 595 596 597 598 599 600
		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
601 602
}

603 604 605
static void unlink_queue(struct sem_array *sma, struct sem_queue *q)
{
	list_del(&q->list);
606
	if (q->nsops > 1)
607 608 609
		sma->complex_count--;
}

610 611 612 613 614 615 616
/** 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
617 618
 * modified the array.
 * Note that wait-for-zero operations are handled without restart.
619 620 621
 */
static int check_restart(struct sem_array *sma, struct sem_queue *q)
{
622 623
	/* pending complex alter operations are too difficult to analyse */
	if (!list_empty(&sma->pending_alter))
624 625 626 627 628 629
		return 1;

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

630 631 632 633 634 635 636 637 638 639 640 641 642
	/* 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;
}
643

644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669
/**
 * 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;
670

671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726
	walk = pending_list->next;
	while (walk != pending_list) {
		int error;

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

		error = try_atomic_semop(sma, q->sops, q->nsops,
						q->undo, q->pid);

		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.
727
		 */
728 729 730 731 732 733
		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);
			}
		}
734 735
	}
	/*
736 737
	 * If one of the modified semaphores got 0,
	 * then check the global queue, too.
738
	 */
739 740
	if (got_zero)
		semop_completed |= wake_const_ops(sma, -1, pt);
741

742
	return semop_completed;
743 744
}

745 746 747 748 749

/**
 * update_queue(sma, semnum): Look for tasks that can be completed.
 * @sma: semaphore array.
 * @semnum: semaphore that was modified.
750
 * @pt: list head for the tasks that must be woken up.
751 752
 *
 * update_queue must be called after a semaphore in a semaphore array
753 754 755
 * 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.
756 757
 * The tasks that must be woken up are added to @pt. The return code
 * is stored in q->pid.
758 759
 * The function internally checks if const operations can now succeed.
 *
760
 * The function return 1 if at least one semop was completed successfully.
L
Linus Torvalds 已提交
761
 */
762
static int update_queue(struct sem_array *sma, int semnum, struct list_head *pt)
L
Linus Torvalds 已提交
763
{
764 765 766
	struct sem_queue *q;
	struct list_head *walk;
	struct list_head *pending_list;
767
	int semop_completed = 0;
768

769
	if (semnum == -1)
770
		pending_list = &sma->pending_alter;
771
	else
772
		pending_list = &sma->sem_base[semnum].pending_alter;
N
Nick Piggin 已提交
773 774

again:
775 776
	walk = pending_list->next;
	while (walk != pending_list) {
777
		int error, restart;
778

779
		q = container_of(walk, struct sem_queue, list);
780
		walk = walk->next;
L
Linus Torvalds 已提交
781

782 783
		/* If we are scanning the single sop, per-semaphore list of
		 * one semaphore and that semaphore is 0, then it is not
784
		 * necessary to scan further: simple increments
785 786 787 788
		 * 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.
		 */
789
		if (semnum != -1 && sma->sem_base[semnum].semval == 0)
790 791
			break;

L
Linus Torvalds 已提交
792 793 794 795
		error = try_atomic_semop(sma, q->sops, q->nsops,
					 q->undo, q->pid);

		/* Does q->sleeper still need to sleep? */
N
Nick Piggin 已提交
796 797 798
		if (error > 0)
			continue;

799
		unlink_queue(sma, q);
N
Nick Piggin 已提交
800

801
		if (error) {
802
			restart = 0;
803 804
		} else {
			semop_completed = 1;
805
			do_smart_wakeup_zero(sma, q->sops, q->nsops, pt);
806
			restart = check_restart(sma, q);
807
		}
808

809
		wake_up_sem_queue_prepare(pt, q, error);
810
		if (restart)
N
Nick Piggin 已提交
811
			goto again;
L
Linus Torvalds 已提交
812
	}
813
	return semop_completed;
L
Linus Torvalds 已提交
814 815
}

816 817
/**
 * do_smart_update(sma, sops, nsops, otime, pt) - optimized update_queue
818 819 820
 * @sma: semaphore array
 * @sops: operations that were performed
 * @nsops: number of operations
821 822
 * @otime: force setting otime
 * @pt: list head of the tasks that must be woken up.
823
 *
824 825
 * 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.
826 827 828
 * 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.
829
 */
830 831
static void do_smart_update(struct sem_array *sma, struct sembuf *sops, int nsops,
			int otime, struct list_head *pt)
832 833
{
	int i;
834
	int progress;
835

836 837
	otime |= do_smart_wakeup_zero(sma, sops, nsops, pt);

838 839 840 841 842
	progress = 1;
retry_global:
	if (sma->complex_count) {
		if (update_queue(sma, -1, pt)) {
			progress = 1;
843
			otime = 1;
844 845
			sops = NULL;
		}
846
	}
847 848
	if (!progress)
		goto done;
849 850 851 852

	if (!sops) {
		/* No semops; something special is going on. */
		for (i = 0; i < sma->sem_nsems; i++) {
853
			if (update_queue(sma, i, pt)) {
854
				otime = 1;
855 856
				progress = 1;
			}
857
		}
858
		goto done_checkretry;
859 860
	}

861
	/* Check the semaphores that were modified. */
862 863 864 865
	for (i = 0; i < nsops; i++) {
		if (sops[i].sem_op > 0 ||
			(sops[i].sem_op < 0 &&
				sma->sem_base[sops[i].sem_num].semval == 0))
866
			if (update_queue(sma, sops[i].sem_num, pt)) {
867
				otime = 1;
868 869 870 871 872 873 874
				progress = 1;
			}
	}
done_checkretry:
	if (progress) {
		progress = 0;
		goto retry_global;
875
	}
876 877 878
done:
	if (otime)
		sma->sem_otime = get_seconds();
879 880 881
}


L
Linus Torvalds 已提交
882 883 884 885 886 887 888 889 890 891 892 893 894 895 896
/* 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;
897
	list_for_each_entry(q, &sma->sem_base[semnum].pending_alter, list) {
R
Rik van Riel 已提交
898 899 900 901 902 903
		struct sembuf * sops = q->sops;
		BUG_ON(sops->sem_num != semnum);
		if ((sops->sem_op < 0) && !(sops->sem_flg & IPC_NOWAIT))
			semncnt++;
	}

904
	list_for_each_entry(q, &sma->pending_alter, list) {
L
Linus Torvalds 已提交
905 906 907 908 909 910 911 912 913 914 915
		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;
}
916

L
Linus Torvalds 已提交
917 918 919 920 921 922
static int count_semzcnt (struct sem_array * sma, ushort semnum)
{
	int semzcnt;
	struct sem_queue * q;

	semzcnt = 0;
923
	list_for_each_entry(q, &sma->sem_base[semnum].pending_const, list) {
R
Rik van Riel 已提交
924 925 926 927 928 929
		struct sembuf * sops = q->sops;
		BUG_ON(sops->sem_num != semnum);
		if ((sops->sem_op == 0) && !(sops->sem_flg & IPC_NOWAIT))
			semzcnt++;
	}

930
	list_for_each_entry(q, &sma->pending_const, list) {
L
Linus Torvalds 已提交
931 932 933 934 935 936 937 938 939 940 941 942
		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;
}

N
Nadia Derbey 已提交
943 944 945
/* Free a semaphore set. freeary() is called with sem_ids.rw_mutex locked
 * as a writer and the spinlock for this semaphore set hold. sem_ids.rw_mutex
 * remains locked on exit.
L
Linus Torvalds 已提交
946
 */
947
static void freeary(struct ipc_namespace *ns, struct kern_ipc_perm *ipcp)
L
Linus Torvalds 已提交
948
{
949 950
	struct sem_undo *un, *tu;
	struct sem_queue *q, *tq;
951
	struct sem_array *sma = container_of(ipcp, struct sem_array, sem_perm);
952
	struct list_head tasks;
953
	int i;
L
Linus Torvalds 已提交
954

955
	/* Free the existing undo structures for this semaphore set.  */
956
	ipc_assert_locked_object(&sma->sem_perm);
957 958 959
	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
960
		un->semid = -1;
961 962
		list_del_rcu(&un->list_proc);
		spin_unlock(&un->ulp->lock);
963
		kfree_rcu(un, rcu);
964
	}
L
Linus Torvalds 已提交
965 966

	/* Wake up all pending processes and let them fail with EIDRM. */
967
	INIT_LIST_HEAD(&tasks);
968 969 970 971 972 973
	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) {
974
		unlink_queue(sma, q);
975
		wake_up_sem_queue_prepare(&tasks, q, -EIDRM);
L
Linus Torvalds 已提交
976
	}
977 978
	for (i = 0; i < sma->sem_nsems; i++) {
		struct sem *sem = sma->sem_base + i;
979 980 981 982 983
		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) {
984 985 986 987
			unlink_queue(sma, q);
			wake_up_sem_queue_prepare(&tasks, q, -EIDRM);
		}
	}
L
Linus Torvalds 已提交
988

N
Nadia Derbey 已提交
989 990
	/* Remove the semaphore set from the IDR */
	sem_rmid(ns, sma);
991
	sem_unlock(sma, -1);
992
	rcu_read_unlock();
L
Linus Torvalds 已提交
993

994
	wake_up_sem_queue_do(&tasks);
K
Kirill Korotaev 已提交
995
	ns->used_sems -= sma->sem_nsems;
L
Linus Torvalds 已提交
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008
	security_sem_free(sma);
	ipc_rcu_putref(sma);
}

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;

1009 1010
		memset(&out, 0, sizeof(out));

L
Linus Torvalds 已提交
1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023
		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;
	}
}

1024
static int semctl_nolock(struct ipc_namespace *ns, int semid,
1025
			 int cmd, int version, void __user *p)
L
Linus Torvalds 已提交
1026
{
1027
	int err;
L
Linus Torvalds 已提交
1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
	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
1042 1043 1044 1045
		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
1046 1047 1048 1049
		seminfo.semvmx = SEMVMX;
		seminfo.semmnu = SEMMNU;
		seminfo.semmap = SEMMAP;
		seminfo.semume = SEMUME;
N
Nadia Derbey 已提交
1050
		down_read(&sem_ids(ns).rw_mutex);
L
Linus Torvalds 已提交
1051
		if (cmd == SEM_INFO) {
K
Kirill Korotaev 已提交
1052 1053
			seminfo.semusz = sem_ids(ns).in_use;
			seminfo.semaem = ns->used_sems;
L
Linus Torvalds 已提交
1054 1055 1056 1057
		} else {
			seminfo.semusz = SEMUSZ;
			seminfo.semaem = SEMAEM;
		}
N
Nadia Derbey 已提交
1058
		max_id = ipc_get_maxid(&sem_ids(ns));
N
Nadia Derbey 已提交
1059
		up_read(&sem_ids(ns).rw_mutex);
1060
		if (copy_to_user(p, &seminfo, sizeof(struct seminfo))) 
L
Linus Torvalds 已提交
1061 1062 1063
			return -EFAULT;
		return (max_id < 0) ? 0: max_id;
	}
1064
	case IPC_STAT:
L
Linus Torvalds 已提交
1065 1066 1067
	case SEM_STAT:
	{
		struct semid64_ds tbuf;
1068 1069 1070
		int id = 0;

		memset(&tbuf, 0, sizeof(tbuf));
L
Linus Torvalds 已提交
1071

1072
		rcu_read_lock();
1073
		if (cmd == SEM_STAT) {
1074 1075 1076 1077 1078
			sma = sem_obtain_object(ns, semid);
			if (IS_ERR(sma)) {
				err = PTR_ERR(sma);
				goto out_unlock;
			}
1079 1080
			id = sma->sem_perm.id;
		} else {
1081 1082 1083 1084 1085
			sma = sem_obtain_object_check(ns, semid);
			if (IS_ERR(sma)) {
				err = PTR_ERR(sma);
				goto out_unlock;
			}
1086
		}
L
Linus Torvalds 已提交
1087 1088

		err = -EACCES;
1089
		if (ipcperms(ns, &sma->sem_perm, S_IRUGO))
L
Linus Torvalds 已提交
1090 1091 1092 1093 1094 1095 1096 1097 1098 1099
			goto out_unlock;

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

		kernel_to_ipc64_perm(&sma->sem_perm, &tbuf.sem_perm);
		tbuf.sem_otime  = sma->sem_otime;
		tbuf.sem_ctime  = sma->sem_ctime;
		tbuf.sem_nsems  = sma->sem_nsems;
1100
		rcu_read_unlock();
1101
		if (copy_semid_to_user(p, &tbuf, version))
L
Linus Torvalds 已提交
1102 1103 1104 1105 1106 1107 1108
			return -EFAULT;
		return id;
	}
	default:
		return -EINVAL;
	}
out_unlock:
1109
	rcu_read_unlock();
L
Linus Torvalds 已提交
1110 1111 1112
	return err;
}

1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129
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

1130 1131
	if (val > SEMVMX || val < 0)
		return -ERANGE;
1132 1133 1134

	INIT_LIST_HEAD(&tasks);

1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
	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;
	}
1152 1153

	err = security_sem_semctl(sma, SETVAL);
1154 1155 1156 1157
	if (err) {
		rcu_read_unlock();
		return -EACCES;
	}
1158

1159
	sem_lock(sma, NULL, -1);
1160 1161 1162

	curr = &sma->sem_base[semnum];

1163
	ipc_assert_locked_object(&sma->sem_perm);
1164 1165 1166 1167 1168 1169 1170 1171
	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);
1172
	sem_unlock(sma, -1);
1173
	rcu_read_unlock();
1174
	wake_up_sem_queue_do(&tasks);
1175
	return 0;
1176 1177
}

K
Kirill Korotaev 已提交
1178
static int semctl_main(struct ipc_namespace *ns, int semid, int semnum,
1179
		int cmd, void __user *p)
L
Linus Torvalds 已提交
1180 1181 1182
{
	struct sem_array *sma;
	struct sem* curr;
1183
	int err, nsems;
L
Linus Torvalds 已提交
1184 1185
	ushort fast_sem_io[SEMMSL_FAST];
	ushort* sem_io = fast_sem_io;
1186
	struct list_head tasks;
L
Linus Torvalds 已提交
1187

1188 1189 1190 1191 1192 1193
	INIT_LIST_HEAD(&tasks);

	rcu_read_lock();
	sma = sem_obtain_object_check(ns, semid);
	if (IS_ERR(sma)) {
		rcu_read_unlock();
1194
		return PTR_ERR(sma);
1195
	}
L
Linus Torvalds 已提交
1196 1197 1198 1199

	nsems = sma->sem_nsems;

	err = -EACCES;
1200 1201
	if (ipcperms(ns, &sma->sem_perm, cmd == SETALL ? S_IWUGO : S_IRUGO))
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1202 1203

	err = security_sem_semctl(sma, cmd);
1204 1205
	if (err)
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1206 1207 1208 1209 1210

	err = -EACCES;
	switch (cmd) {
	case GETALL:
	{
1211
		ushort __user *array = p;
L
Linus Torvalds 已提交
1212 1213
		int i;

1214
		sem_lock(sma, NULL, -1);
L
Linus Torvalds 已提交
1215
		if(nsems > SEMMSL_FAST) {
1216 1217
			if (!ipc_rcu_getref(sma)) {
				sem_unlock(sma, -1);
1218
				rcu_read_unlock();
1219 1220 1221 1222
				err = -EIDRM;
				goto out_free;
			}
			sem_unlock(sma, -1);
1223
			rcu_read_unlock();
L
Linus Torvalds 已提交
1224 1225
			sem_io = ipc_alloc(sizeof(ushort)*nsems);
			if(sem_io == NULL) {
1226
				sem_putref(sma);
L
Linus Torvalds 已提交
1227 1228 1229
				return -ENOMEM;
			}

1230
			rcu_read_lock();
1231
			sem_lock_and_putref(sma);
L
Linus Torvalds 已提交
1232
			if (sma->sem_perm.deleted) {
1233
				sem_unlock(sma, -1);
1234
				rcu_read_unlock();
L
Linus Torvalds 已提交
1235 1236 1237
				err = -EIDRM;
				goto out_free;
			}
1238
		}
L
Linus Torvalds 已提交
1239 1240
		for (i = 0; i < sma->sem_nsems; i++)
			sem_io[i] = sma->sem_base[i].semval;
1241
		sem_unlock(sma, -1);
1242
		rcu_read_unlock();
L
Linus Torvalds 已提交
1243 1244 1245 1246 1247 1248 1249 1250 1251 1252
		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;

1253 1254 1255 1256
		if (!ipc_rcu_getref(sma)) {
			rcu_read_unlock();
			return -EIDRM;
		}
1257
		rcu_read_unlock();
L
Linus Torvalds 已提交
1258 1259 1260 1261

		if(nsems > SEMMSL_FAST) {
			sem_io = ipc_alloc(sizeof(ushort)*nsems);
			if(sem_io == NULL) {
1262
				sem_putref(sma);
L
Linus Torvalds 已提交
1263 1264 1265 1266
				return -ENOMEM;
			}
		}

1267
		if (copy_from_user (sem_io, p, nsems*sizeof(ushort))) {
1268
			sem_putref(sma);
L
Linus Torvalds 已提交
1269 1270 1271 1272 1273 1274
			err = -EFAULT;
			goto out_free;
		}

		for (i = 0; i < nsems; i++) {
			if (sem_io[i] > SEMVMX) {
1275
				sem_putref(sma);
L
Linus Torvalds 已提交
1276 1277 1278 1279
				err = -ERANGE;
				goto out_free;
			}
		}
1280
		rcu_read_lock();
1281
		sem_lock_and_putref(sma);
L
Linus Torvalds 已提交
1282
		if (sma->sem_perm.deleted) {
1283
			sem_unlock(sma, -1);
1284
			rcu_read_unlock();
L
Linus Torvalds 已提交
1285 1286 1287 1288 1289 1290
			err = -EIDRM;
			goto out_free;
		}

		for (i = 0; i < nsems; i++)
			sma->sem_base[i].semval = sem_io[i];
1291

1292
		ipc_assert_locked_object(&sma->sem_perm);
1293
		list_for_each_entry(un, &sma->list_id, list_id) {
L
Linus Torvalds 已提交
1294 1295
			for (i = 0; i < nsems; i++)
				un->semadj[i] = 0;
1296
		}
L
Linus Torvalds 已提交
1297 1298
		sma->sem_ctime = get_seconds();
		/* maybe some queued-up processes were waiting for this */
1299
		do_smart_update(sma, NULL, 0, 0, &tasks);
L
Linus Torvalds 已提交
1300 1301 1302
		err = 0;
		goto out_unlock;
	}
1303
	/* GETVAL, GETPID, GETNCTN, GETZCNT: fall-through */
L
Linus Torvalds 已提交
1304 1305
	}
	err = -EINVAL;
1306 1307
	if (semnum < 0 || semnum >= nsems)
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1308

1309
	sem_lock(sma, NULL, -1);
L
Linus Torvalds 已提交
1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325
	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;
	}
1326

L
Linus Torvalds 已提交
1327
out_unlock:
1328
	sem_unlock(sma, -1);
1329
out_rcu_wakeup:
1330
	rcu_read_unlock();
1331
	wake_up_sem_queue_do(&tasks);
L
Linus Torvalds 已提交
1332 1333 1334 1335 1336 1337
out_free:
	if(sem_io != fast_sem_io)
		ipc_free(sem_io, sizeof(ushort)*nsems);
	return err;
}

1338 1339
static inline unsigned long
copy_semid_from_user(struct semid64_ds *out, void __user *buf, int version)
L
Linus Torvalds 已提交
1340 1341 1342
{
	switch(version) {
	case IPC_64:
1343
		if (copy_from_user(out, buf, sizeof(*out)))
L
Linus Torvalds 已提交
1344 1345 1346 1347 1348 1349 1350 1351 1352
			return -EFAULT;
		return 0;
	case IPC_OLD:
	    {
		struct semid_ds tbuf_old;

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

1353 1354 1355
		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
1356 1357 1358 1359 1360 1361 1362 1363

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

1364 1365 1366 1367 1368
/*
 * This function handles some semctl commands which require the rw_mutex
 * to be held in write mode.
 * NOTE: no locks must be held, the rw_mutex is taken inside this function.
 */
1369
static int semctl_down(struct ipc_namespace *ns, int semid,
1370
		       int cmd, int version, void __user *p)
L
Linus Torvalds 已提交
1371 1372 1373
{
	struct sem_array *sma;
	int err;
1374
	struct semid64_ds semid64;
L
Linus Torvalds 已提交
1375 1376 1377
	struct kern_ipc_perm *ipcp;

	if(cmd == IPC_SET) {
1378
		if (copy_semid_from_user(&semid64, p, version))
L
Linus Torvalds 已提交
1379 1380
			return -EFAULT;
	}
S
Steve Grubb 已提交
1381

1382 1383 1384
	down_write(&sem_ids(ns).rw_mutex);
	rcu_read_lock();

1385 1386
	ipcp = ipcctl_pre_down_nolock(ns, &sem_ids(ns), semid, cmd,
				      &semid64.sem_perm, 0);
1387 1388 1389 1390
	if (IS_ERR(ipcp)) {
		err = PTR_ERR(ipcp);
		goto out_unlock1;
	}
S
Steve Grubb 已提交
1391

1392
	sma = container_of(ipcp, struct sem_array, sem_perm);
L
Linus Torvalds 已提交
1393 1394

	err = security_sem_semctl(sma, cmd);
1395 1396
	if (err)
		goto out_unlock1;
L
Linus Torvalds 已提交
1397

1398
	switch (cmd) {
L
Linus Torvalds 已提交
1399
	case IPC_RMID:
1400
		sem_lock(sma, NULL, -1);
1401
		/* freeary unlocks the ipc object and rcu */
1402
		freeary(ns, ipcp);
1403
		goto out_up;
L
Linus Torvalds 已提交
1404
	case IPC_SET:
1405
		sem_lock(sma, NULL, -1);
1406 1407
		err = ipc_update_perm(&semid64.sem_perm, ipcp);
		if (err)
1408
			goto out_unlock0;
L
Linus Torvalds 已提交
1409 1410 1411 1412
		sma->sem_ctime = get_seconds();
		break;
	default:
		err = -EINVAL;
1413
		goto out_unlock1;
L
Linus Torvalds 已提交
1414 1415
	}

1416
out_unlock0:
1417
	sem_unlock(sma, -1);
1418
out_unlock1:
1419
	rcu_read_unlock();
1420 1421
out_up:
	up_write(&sem_ids(ns).rw_mutex);
L
Linus Torvalds 已提交
1422 1423 1424
	return err;
}

1425
SYSCALL_DEFINE4(semctl, int, semid, int, semnum, int, cmd, unsigned long, arg)
L
Linus Torvalds 已提交
1426 1427
{
	int version;
K
Kirill Korotaev 已提交
1428
	struct ipc_namespace *ns;
1429
	void __user *p = (void __user *)arg;
L
Linus Torvalds 已提交
1430 1431 1432 1433 1434

	if (semid < 0)
		return -EINVAL;

	version = ipc_parse_version(&cmd);
K
Kirill Korotaev 已提交
1435
	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
1436 1437 1438 1439

	switch(cmd) {
	case IPC_INFO:
	case SEM_INFO:
1440
	case IPC_STAT:
L
Linus Torvalds 已提交
1441
	case SEM_STAT:
1442
		return semctl_nolock(ns, semid, cmd, version, p);
L
Linus Torvalds 已提交
1443 1444 1445 1446 1447 1448
	case GETALL:
	case GETVAL:
	case GETPID:
	case GETNCNT:
	case GETZCNT:
	case SETALL:
1449 1450 1451
		return semctl_main(ns, semid, semnum, cmd, p);
	case SETVAL:
		return semctl_setval(ns, semid, semnum, arg);
L
Linus Torvalds 已提交
1452 1453
	case IPC_RMID:
	case IPC_SET:
1454
		return semctl_down(ns, semid, cmd, version, p);
L
Linus Torvalds 已提交
1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476
	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) {
1477
		undo_list = kzalloc(sizeof(*undo_list), GFP_KERNEL);
L
Linus Torvalds 已提交
1478 1479
		if (undo_list == NULL)
			return -ENOMEM;
I
Ingo Molnar 已提交
1480
		spin_lock_init(&undo_list->lock);
L
Linus Torvalds 已提交
1481
		atomic_set(&undo_list->refcnt, 1);
1482 1483
		INIT_LIST_HEAD(&undo_list->list_proc);

L
Linus Torvalds 已提交
1484 1485 1486 1487 1488 1489
		current->sysvsem.undo_list = undo_list;
	}
	*undo_listp = undo_list;
	return 0;
}

1490
static struct sem_undo *__lookup_undo(struct sem_undo_list *ulp, int semid)
L
Linus Torvalds 已提交
1491
{
1492
	struct sem_undo *un;
1493

1494 1495 1496
	list_for_each_entry_rcu(un, &ulp->list_proc, list_proc) {
		if (un->semid == semid)
			return un;
L
Linus Torvalds 已提交
1497
	}
1498
	return NULL;
L
Linus Torvalds 已提交
1499 1500
}

1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514
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;
}

1515 1516 1517 1518 1519 1520 1521 1522
/**
 * 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.
1523 1524
 * Lifetime-rules: sem_undo is rcu-protected, on success, the function
 * performs a rcu_read_lock().
1525 1526
 */
static struct sem_undo *find_alloc_undo(struct ipc_namespace *ns, int semid)
L
Linus Torvalds 已提交
1527 1528 1529 1530
{
	struct sem_array *sma;
	struct sem_undo_list *ulp;
	struct sem_undo *un, *new;
1531
	int nsems, error;
L
Linus Torvalds 已提交
1532 1533 1534 1535 1536

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

1537
	rcu_read_lock();
1538
	spin_lock(&ulp->lock);
L
Linus Torvalds 已提交
1539
	un = lookup_undo(ulp, semid);
1540
	spin_unlock(&ulp->lock);
L
Linus Torvalds 已提交
1541 1542 1543 1544
	if (likely(un!=NULL))
		goto out;

	/* no undo structure around - allocate one. */
1545
	/* step 1: figure out the size of the semaphore array */
1546 1547 1548
	sma = sem_obtain_object_check(ns, semid);
	if (IS_ERR(sma)) {
		rcu_read_unlock();
J
Julia Lawall 已提交
1549
		return ERR_CAST(sma);
1550
	}
1551

L
Linus Torvalds 已提交
1552
	nsems = sma->sem_nsems;
1553 1554 1555 1556 1557
	if (!ipc_rcu_getref(sma)) {
		rcu_read_unlock();
		un = ERR_PTR(-EIDRM);
		goto out;
	}
1558
	rcu_read_unlock();
L
Linus Torvalds 已提交
1559

1560
	/* step 2: allocate new undo structure */
1561
	new = kzalloc(sizeof(struct sem_undo) + sizeof(short)*nsems, GFP_KERNEL);
L
Linus Torvalds 已提交
1562
	if (!new) {
1563
		sem_putref(sma);
L
Linus Torvalds 已提交
1564 1565 1566
		return ERR_PTR(-ENOMEM);
	}

1567
	/* step 3: Acquire the lock on semaphore array */
1568
	rcu_read_lock();
1569
	sem_lock_and_putref(sma);
L
Linus Torvalds 已提交
1570
	if (sma->sem_perm.deleted) {
1571
		sem_unlock(sma, -1);
1572
		rcu_read_unlock();
L
Linus Torvalds 已提交
1573 1574 1575 1576
		kfree(new);
		un = ERR_PTR(-EIDRM);
		goto out;
	}
1577 1578 1579 1580 1581 1582 1583 1584 1585 1586
	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;
	}
1587 1588
	/* step 5: initialize & link new undo structure */
	new->semadj = (short *) &new[1];
1589
	new->ulp = ulp;
1590 1591
	new->semid = semid;
	assert_spin_locked(&ulp->lock);
1592
	list_add_rcu(&new->list_proc, &ulp->list_proc);
1593
	ipc_assert_locked_object(&sma->sem_perm);
1594
	list_add(&new->list_id, &sma->list_id);
1595
	un = new;
1596

1597
success:
1598
	spin_unlock(&ulp->lock);
1599
	sem_unlock(sma, -1);
L
Linus Torvalds 已提交
1600 1601 1602 1603
out:
	return un;
}

1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630

/**
 * 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;
}


1631 1632
SYSCALL_DEFINE4(semtimedop, int, semid, struct sembuf __user *, tsops,
		unsigned, nsops, const struct timespec __user *, timeout)
L
Linus Torvalds 已提交
1633 1634 1635 1636 1637 1638
{
	int error = -EINVAL;
	struct sem_array *sma;
	struct sembuf fast_sops[SEMOPM_FAST];
	struct sembuf* sops = fast_sops, *sop;
	struct sem_undo *un;
1639
	int undos = 0, alter = 0, max, locknum;
L
Linus Torvalds 已提交
1640 1641
	struct sem_queue queue;
	unsigned long jiffies_left = 0;
K
Kirill Korotaev 已提交
1642
	struct ipc_namespace *ns;
1643
	struct list_head tasks;
K
Kirill Korotaev 已提交
1644 1645

	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
1646 1647 1648

	if (nsops < 1 || semid < 0)
		return -EINVAL;
K
Kirill Korotaev 已提交
1649
	if (nsops > ns->sc_semopm)
L
Linus Torvalds 已提交
1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677
		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)
1678 1679
			undos = 1;
		if (sop->sem_op != 0)
L
Linus Torvalds 已提交
1680 1681 1682
			alter = 1;
	}

1683 1684
	INIT_LIST_HEAD(&tasks);

L
Linus Torvalds 已提交
1685
	if (undos) {
1686
		/* On success, find_alloc_undo takes the rcu_read_lock */
1687
		un = find_alloc_undo(ns, semid);
L
Linus Torvalds 已提交
1688 1689 1690 1691
		if (IS_ERR(un)) {
			error = PTR_ERR(un);
			goto out_free;
		}
1692
	} else {
L
Linus Torvalds 已提交
1693
		un = NULL;
1694 1695
		rcu_read_lock();
	}
L
Linus Torvalds 已提交
1696

1697
	sma = sem_obtain_object_check(ns, semid);
1698
	if (IS_ERR(sma)) {
1699
		rcu_read_unlock();
1700
		error = PTR_ERR(sma);
L
Linus Torvalds 已提交
1701
		goto out_free;
1702 1703
	}

1704
	error = -EFBIG;
1705 1706
	if (max >= sma->sem_nsems)
		goto out_rcu_wakeup;
1707 1708

	error = -EACCES;
1709 1710
	if (ipcperms(ns, &sma->sem_perm, alter ? S_IWUGO : S_IRUGO))
		goto out_rcu_wakeup;
1711 1712

	error = security_sem_semop(sma, sops, nsops, alter);
1713 1714
	if (error)
		goto out_rcu_wakeup;
1715

L
Linus Torvalds 已提交
1716
	/*
1717
	 * semid identifiers are not unique - find_alloc_undo may have
L
Linus Torvalds 已提交
1718
	 * allocated an undo structure, it was invalidated by an RMID
1719
	 * and now a new array with received the same id. Check and fail.
L
Lucas De Marchi 已提交
1720
	 * This case can be detected checking un->semid. The existence of
1721
	 * "un" itself is guaranteed by rcu.
L
Linus Torvalds 已提交
1722
	 */
1723
	error = -EIDRM;
1724 1725 1726
	locknum = sem_lock(sma, sops, nsops);
	if (un && un->semid == -1)
		goto out_unlock_free;
1727

1728
	error = try_atomic_semop (sma, sops, nsops, un, task_tgid_vnr(current));
L
Linus Torvalds 已提交
1729 1730
	if (error <= 0) {
		if (alter && error == 0)
1731
			do_smart_update(sma, sops, nsops, 1, &tasks);
1732

L
Linus Torvalds 已提交
1733 1734 1735 1736 1737 1738 1739 1740 1741 1742
		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;
1743
	queue.pid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
1744 1745
	queue.alter = alter;

1746 1747 1748 1749 1750
	if (nsops == 1) {
		struct sem *curr;
		curr = &sma->sem_base[sops->sem_num];

		if (alter)
1751
			list_add_tail(&queue.list, &curr->pending_alter);
1752
		else
1753
			list_add_tail(&queue.list, &curr->pending_const);
1754
	} else {
1755
		if (alter)
1756
			list_add_tail(&queue.list, &sma->pending_alter);
1757
		else
1758 1759
			list_add_tail(&queue.list, &sma->pending_const);

1760 1761 1762
		sma->complex_count++;
	}

L
Linus Torvalds 已提交
1763 1764
	queue.status = -EINTR;
	queue.sleeper = current;
1765 1766

sleep_again:
L
Linus Torvalds 已提交
1767
	current->state = TASK_INTERRUPTIBLE;
1768
	sem_unlock(sma, locknum);
1769
	rcu_read_unlock();
L
Linus Torvalds 已提交
1770 1771 1772 1773 1774 1775

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

1776
	error = get_queue_result(&queue);
L
Linus Torvalds 已提交
1777 1778 1779

	if (error != -EINTR) {
		/* fast path: update_queue already obtained all requested
1780 1781 1782 1783 1784 1785 1786 1787
		 * 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
1788 1789 1790
		goto out_free;
	}

1791
	rcu_read_lock();
1792
	sma = sem_obtain_lock(ns, semid, sops, nsops, &locknum);
1793 1794 1795 1796 1797 1798 1799 1800 1801

	/*
	 * 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().
	 */
1802
	if (IS_ERR(sma)) {
1803
		rcu_read_unlock();
L
Linus Torvalds 已提交
1804 1805 1806
		goto out_free;
	}

1807

L
Linus Torvalds 已提交
1808
	/*
1809 1810
	 * 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
1811
	 */
1812

L
Linus Torvalds 已提交
1813 1814 1815 1816 1817 1818 1819 1820 1821
	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;
1822 1823 1824 1825 1826 1827 1828

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

1829
	unlink_queue(sma, &queue);
L
Linus Torvalds 已提交
1830 1831

out_unlock_free:
1832
	sem_unlock(sma, locknum);
1833
out_rcu_wakeup:
1834
	rcu_read_unlock();
1835
	wake_up_sem_queue_do(&tasks);
L
Linus Torvalds 已提交
1836 1837 1838 1839 1840 1841
out_free:
	if(sops != fast_sops)
		kfree(sops);
	return error;
}

1842 1843
SYSCALL_DEFINE3(semop, int, semid, struct sembuf __user *, tsops,
		unsigned, nsops)
L
Linus Torvalds 已提交
1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882
{
	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)
{
1883
	struct sem_undo_list *ulp;
L
Linus Torvalds 已提交
1884

1885 1886
	ulp = tsk->sysvsem.undo_list;
	if (!ulp)
L
Linus Torvalds 已提交
1887
		return;
1888
	tsk->sysvsem.undo_list = NULL;
L
Linus Torvalds 已提交
1889

1890
	if (!atomic_dec_and_test(&ulp->refcnt))
L
Linus Torvalds 已提交
1891 1892
		return;

1893
	for (;;) {
L
Linus Torvalds 已提交
1894
		struct sem_array *sma;
1895
		struct sem_undo *un;
1896
		struct list_head tasks;
1897
		int semid, i;
1898

1899
		rcu_read_lock();
1900 1901
		un = list_entry_rcu(ulp->list_proc.next,
				    struct sem_undo, list_proc);
1902 1903 1904 1905
		if (&un->list_proc == &ulp->list_proc)
			semid = -1;
		 else
			semid = un->semid;
1906

1907 1908
		if (semid == -1) {
			rcu_read_unlock();
1909
			break;
1910
		}
L
Linus Torvalds 已提交
1911

1912
		sma = sem_obtain_object_check(tsk->nsproxy->ipc_ns, un->semid);
1913
		/* exit_sem raced with IPC_RMID, nothing to do */
1914 1915
		if (IS_ERR(sma)) {
			rcu_read_unlock();
1916
			continue;
1917
		}
L
Linus Torvalds 已提交
1918

1919
		sem_lock(sma, NULL, -1);
1920
		un = __lookup_undo(ulp, semid);
1921 1922 1923 1924
		if (un == NULL) {
			/* exit_sem raced with IPC_RMID+semget() that created
			 * exactly the same semid. Nothing to do.
			 */
1925
			sem_unlock(sma, -1);
1926
			rcu_read_unlock();
1927 1928 1929 1930
			continue;
		}

		/* remove un from the linked lists */
1931
		ipc_assert_locked_object(&sma->sem_perm);
1932 1933
		list_del(&un->list_id);

1934 1935 1936 1937
		spin_lock(&ulp->lock);
		list_del_rcu(&un->list_proc);
		spin_unlock(&ulp->lock);

1938 1939
		/* perform adjustments registered in un */
		for (i = 0; i < sma->sem_nsems; i++) {
I
Ingo Molnar 已提交
1940
			struct sem * semaphore = &sma->sem_base[i];
1941 1942
			if (un->semadj[i]) {
				semaphore->semval += un->semadj[i];
L
Linus Torvalds 已提交
1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955
				/*
				 * 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
1956 1957 1958 1959
				if (semaphore->semval < 0)
					semaphore->semval = 0;
				if (semaphore->semval > SEMVMX)
					semaphore->semval = SEMVMX;
1960
				semaphore->sempid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
1961 1962 1963
			}
		}
		/* maybe some queued-up processes were waiting for this */
1964 1965
		INIT_LIST_HEAD(&tasks);
		do_smart_update(sma, NULL, 0, 1, &tasks);
1966
		sem_unlock(sma, -1);
1967
		rcu_read_unlock();
1968
		wake_up_sem_queue_do(&tasks);
1969

1970
		kfree_rcu(un, rcu);
L
Linus Torvalds 已提交
1971
	}
1972
	kfree(ulp);
L
Linus Torvalds 已提交
1973 1974 1975
}

#ifdef CONFIG_PROC_FS
1976
static int sysvipc_sem_proc_show(struct seq_file *s, void *it)
L
Linus Torvalds 已提交
1977
{
1978
	struct user_namespace *user_ns = seq_user_ns(s);
1979 1980 1981
	struct sem_array *sma = it;

	return seq_printf(s,
1982
			  "%10d %10d  %4o %10u %5u %5u %5u %5u %10lu %10lu\n",
1983
			  sma->sem_perm.key,
N
Nadia Derbey 已提交
1984
			  sma->sem_perm.id,
1985 1986
			  sma->sem_perm.mode,
			  sma->sem_nsems,
1987 1988 1989 1990
			  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),
1991 1992
			  sma->sem_otime,
			  sma->sem_ctime);
L
Linus Torvalds 已提交
1993 1994
}
#endif