sem.c 47.3 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138
	struct list_head sem_pending; /* pending single-sop operations */
};

/* 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;
};


139
#define sem_ids(ns)	((ns)->ids[IPC_SEM_IDS])
K
Kirill Korotaev 已提交
140

N
Nadia Derbey 已提交
141
#define sem_checkid(sma, semid)	ipc_checkid(&sma->sem_perm, semid)
L
Linus Torvalds 已提交
142

N
Nadia Derbey 已提交
143
static int newary(struct ipc_namespace *, struct ipc_params *);
144
static void freeary(struct ipc_namespace *, struct kern_ipc_perm *);
L
Linus Torvalds 已提交
145
#ifdef CONFIG_PROC_FS
146
static int sysvipc_sem_proc_show(struct seq_file *s, void *it);
L
Linus Torvalds 已提交
147 148 149 150 151 152 153 154 155 156 157 158 159 160
#endif

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

/*
 * linked list protection:
 *	sem_undo.id_next,
 *	sem_array.sem_pending{,last},
 *	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
161 162 163 164 165
#define sc_semmsl	sem_ctls[0]
#define sc_semmns	sem_ctls[1]
#define sc_semopm	sem_ctls[2]
#define sc_semmni	sem_ctls[3]

166
void sem_init_ns(struct ipc_namespace *ns)
K
Kirill Korotaev 已提交
167 168 169 170 171 172
{
	ns->sc_semmsl = SEMMSL;
	ns->sc_semmns = SEMMNS;
	ns->sc_semopm = SEMOPM;
	ns->sc_semmni = SEMMNI;
	ns->used_sems = 0;
173
	ipc_init_ids(&ns->ids[IPC_SEM_IDS]);
K
Kirill Korotaev 已提交
174 175
}

176
#ifdef CONFIG_IPC_NS
K
Kirill Korotaev 已提交
177 178
void sem_exit_ns(struct ipc_namespace *ns)
{
179
	free_ipcs(ns, &sem_ids(ns), freeary);
S
Serge E. Hallyn 已提交
180
	idr_destroy(&ns->ids[IPC_SEM_IDS].ipcs_idr);
K
Kirill Korotaev 已提交
181
}
182
#endif
L
Linus Torvalds 已提交
183 184 185

void __init sem_init (void)
{
186
	sem_init_ns(&init_ipc_ns);
187 188
	ipc_init_proc_interface("sysvipc/sem",
				"       key      semid perms      nsems   uid   gid  cuid  cgid      otime      ctime\n",
K
Kirill Korotaev 已提交
189
				IPC_SEM_IDS, sysvipc_sem_proc_show);
L
Linus Torvalds 已提交
190 191
}

192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248
/*
 * 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:
249
		ipc_lock_object(&sma->sem_perm);
250 251 252 253 254 255 256 257 258 259 260 261
		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) {
262
		ipc_unlock_object(&sma->sem_perm);
263 264 265 266 267 268
	} else {
		struct sem *sem = sma->sem_base + locknum;
		spin_unlock(&sem->lock);
	}
}

N
Nadia Derbey 已提交
269 270 271
/*
 * sem_lock_(check_) routines are called in the paths where the rw_mutex
 * is not held.
272 273
 *
 * The caller holds the RCU read lock.
N
Nadia Derbey 已提交
274
 */
275 276
static inline struct sem_array *sem_obtain_lock(struct ipc_namespace *ns,
			int id, struct sembuf *sops, int nsops, int *locknum)
277
{
278 279
	struct kern_ipc_perm *ipcp;
	struct sem_array *sma;
N
Nadia Derbey 已提交
280

281
	ipcp = ipc_obtain_object(&sem_ids(ns), id);
282 283
	if (IS_ERR(ipcp))
		return ERR_CAST(ipcp);
284

285 286
	sma = container_of(ipcp, struct sem_array, sem_perm);
	*locknum = sem_lock(sma, sops, nsops);
287 288 289 290 291 292 293

	/* 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);

294
	sem_unlock(sma, *locknum);
295
	return ERR_PTR(-EINVAL);
296 297
}

298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314
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);
315

N
Nadia Derbey 已提交
316
	return container_of(ipcp, struct sem_array, sem_perm);
317 318
}

319 320
static inline void sem_lock_and_putref(struct sem_array *sma)
{
321
	sem_lock(sma, NULL, -1);
322 323 324 325 326
	ipc_rcu_putref(sma);
}

static inline void sem_putref(struct sem_array *sma)
{
327
	ipc_rcu_putref(sma);
328 329
}

N
Nadia Derbey 已提交
330 331 332 333 334
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
335 336 337 338 339 340 341 342 343 344 345 346 347 348 349
/*
 * 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
 *	* unlinking the queue entry from sma->sem_pending
 *	* 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
350
 *   	  performing any operation on the sem array.
L
Linus Torvalds 已提交
351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368
 *   	* 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
369 370 371 372 373
/**
 * 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
374
 * Called with sem_ids.rw_mutex held (as a writer)
N
Nadia Derbey 已提交
375 376
 */

N
Nadia Derbey 已提交
377
static int newary(struct ipc_namespace *ns, struct ipc_params *params)
L
Linus Torvalds 已提交
378 379 380 381 382
{
	int id;
	int retval;
	struct sem_array *sma;
	int size;
N
Nadia Derbey 已提交
383 384 385
	key_t key = params->key;
	int nsems = params->u.nsems;
	int semflg = params->flg;
386
	int i;
L
Linus Torvalds 已提交
387 388 389

	if (!nsems)
		return -EINVAL;
K
Kirill Korotaev 已提交
390
	if (ns->used_sems + nsems > ns->sc_semmns)
L
Linus Torvalds 已提交
391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409
		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
410
	id = ipc_addid(&sem_ids(ns), &sma->sem_perm, ns->sc_semmni);
411
	if (id < 0) {
L
Linus Torvalds 已提交
412 413
		security_sem_free(sma);
		ipc_rcu_putref(sma);
414
		return id;
L
Linus Torvalds 已提交
415
	}
K
Kirill Korotaev 已提交
416
	ns->used_sems += nsems;
L
Linus Torvalds 已提交
417 418

	sma->sem_base = (struct sem *) &sma[1];
419

420
	for (i = 0; i < nsems; i++) {
421
		INIT_LIST_HEAD(&sma->sem_base[i].sem_pending);
422 423
		spin_lock_init(&sma->sem_base[i].lock);
	}
424 425

	sma->complex_count = 0;
426
	INIT_LIST_HEAD(&sma->sem_pending);
427
	INIT_LIST_HEAD(&sma->list_id);
L
Linus Torvalds 已提交
428 429
	sma->sem_nsems = nsems;
	sma->sem_ctime = get_seconds();
430
	sem_unlock(sma, -1);
431
	rcu_read_unlock();
L
Linus Torvalds 已提交
432

N
Nadia Derbey 已提交
433
	return sma->sem_perm.id;
L
Linus Torvalds 已提交
434 435
}

N
Nadia Derbey 已提交
436

N
Nadia Derbey 已提交
437
/*
N
Nadia Derbey 已提交
438
 * Called with sem_ids.rw_mutex and ipcp locked.
N
Nadia Derbey 已提交
439
 */
N
Nadia Derbey 已提交
440
static inline int sem_security(struct kern_ipc_perm *ipcp, int semflg)
N
Nadia Derbey 已提交
441
{
N
Nadia Derbey 已提交
442 443 444 445
	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
446 447
}

N
Nadia Derbey 已提交
448
/*
N
Nadia Derbey 已提交
449
 * Called with sem_ids.rw_mutex and ipcp locked.
N
Nadia Derbey 已提交
450
 */
N
Nadia Derbey 已提交
451 452
static inline int sem_more_checks(struct kern_ipc_perm *ipcp,
				struct ipc_params *params)
N
Nadia Derbey 已提交
453
{
N
Nadia Derbey 已提交
454 455 456 457
	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
458 459 460 461 462
		return -EINVAL;

	return 0;
}

463
SYSCALL_DEFINE3(semget, key_t, key, int, nsems, int, semflg)
L
Linus Torvalds 已提交
464
{
K
Kirill Korotaev 已提交
465
	struct ipc_namespace *ns;
N
Nadia Derbey 已提交
466 467
	struct ipc_ops sem_ops;
	struct ipc_params sem_params;
K
Kirill Korotaev 已提交
468 469

	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
470

K
Kirill Korotaev 已提交
471
	if (nsems < 0 || nsems > ns->sc_semmsl)
L
Linus Torvalds 已提交
472
		return -EINVAL;
N
Nadia Derbey 已提交
473

N
Nadia Derbey 已提交
474 475 476 477 478 479 480
	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
481

N
Nadia Derbey 已提交
482
	return ipcget(ns, &sem_ids(ns), &sem_ops, &sem_params);
L
Linus Torvalds 已提交
483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550
}

/*
 * 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;
}

551 552 553 554 555
/** 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
556
 */
557 558
static void wake_up_sem_queue_prepare(struct list_head *pt,
				struct sem_queue *q, int error)
N
Nick Piggin 已提交
559
{
560 561 562 563 564 565 566
	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
567
	q->status = IN_WAKEUP;
568 569
	q->pid = error;

570
	list_add_tail(&q->list, pt);
571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587
}

/**
 * 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);
588
	list_for_each_entry_safe(q, t, pt, list) {
589 590 591 592 593 594 595
		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
596 597
}

598 599 600
static void unlink_queue(struct sem_array *sma, struct sem_queue *q)
{
	list_del(&q->list);
601
	if (q->nsops > 1)
602 603 604
		sma->complex_count--;
}

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
/** 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
 * was completed.
 */
static int check_restart(struct sem_array *sma, struct sem_queue *q)
{
	struct sem *curr;
	struct sem_queue *h;

	/* if the operation didn't modify the array, then no restart */
	if (q->alter == 0)
		return 0;

	/* pending complex operations are too difficult to analyse */
	if (sma->complex_count)
		return 1;

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

	curr = sma->sem_base + q->sops[0].sem_num;

	/* No-one waits on this queue */
	if (list_empty(&curr->sem_pending))
		return 0;

	/* the new semaphore value */
	if (curr->semval) {
		/* It is impossible that someone waits for the new value:
		 * - q is a previously sleeping simple operation that
		 *   altered the array. It must be a decrement, because
		 *   simple increments never sleep.
		 * - The value is not 0, thus wait-for-zero won't proceed.
		 * - 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.
		 */
		BUG_ON(q->sops[0].sem_op >= 0);
		return 0;
	}
	/*
	 * semval is 0. Check if there are wait-for-zero semops.
654
	 * They must be the first entries in the per-semaphore queue
655
	 */
656
	h = list_first_entry(&curr->sem_pending, struct sem_queue, list);
657 658 659 660 661 662 663 664 665 666 667
	BUG_ON(h->nsops != 1);
	BUG_ON(h->sops[0].sem_num != q->sops[0].sem_num);

	/* Yes, there is a wait-for-zero semop. Restart */
	if (h->sops[0].sem_op == 0)
		return 1;

	/* Again - no-one is waiting for the new value. */
	return 0;
}

668 669 670 671 672

/**
 * update_queue(sma, semnum): Look for tasks that can be completed.
 * @sma: semaphore array.
 * @semnum: semaphore that was modified.
673
 * @pt: list head for the tasks that must be woken up.
674 675
 *
 * update_queue must be called after a semaphore in a semaphore array
676 677 678
 * 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.
679 680 681
 * The tasks that must be woken up are added to @pt. The return code
 * is stored in q->pid.
 * The function return 1 if at least one semop was completed successfully.
L
Linus Torvalds 已提交
682
 */
683
static int update_queue(struct sem_array *sma, int semnum, struct list_head *pt)
L
Linus Torvalds 已提交
684
{
685 686 687
	struct sem_queue *q;
	struct list_head *walk;
	struct list_head *pending_list;
688
	int semop_completed = 0;
689

690
	if (semnum == -1)
691
		pending_list = &sma->sem_pending;
692
	else
693
		pending_list = &sma->sem_base[semnum].sem_pending;
N
Nick Piggin 已提交
694 695

again:
696 697
	walk = pending_list->next;
	while (walk != pending_list) {
698
		int error, restart;
699

700
		q = container_of(walk, struct sem_queue, list);
701
		walk = walk->next;
L
Linus Torvalds 已提交
702

703 704 705 706 707 708 709 710 711 712 713
		/* If we are scanning the single sop, per-semaphore list of
		 * one semaphore and that semaphore is 0, then it is not
		 * necessary to scan the "alter" entries: simple increments
		 * 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.
		 */
		if (semnum != -1 && sma->sem_base[semnum].semval == 0 &&
				q->alter)
			break;

L
Linus Torvalds 已提交
714 715 716 717
		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
718 719 720
		if (error > 0)
			continue;

721
		unlink_queue(sma, q);
N
Nick Piggin 已提交
722

723
		if (error) {
724
			restart = 0;
725 726
		} else {
			semop_completed = 1;
727
			restart = check_restart(sma, q);
728
		}
729

730
		wake_up_sem_queue_prepare(pt, q, error);
731
		if (restart)
N
Nick Piggin 已提交
732
			goto again;
L
Linus Torvalds 已提交
733
	}
734
	return semop_completed;
L
Linus Torvalds 已提交
735 736
}

737 738
/**
 * do_smart_update(sma, sops, nsops, otime, pt) - optimized update_queue
739 740 741
 * @sma: semaphore array
 * @sops: operations that were performed
 * @nsops: number of operations
742 743
 * @otime: force setting otime
 * @pt: list head of the tasks that must be woken up.
744 745 746
 *
 * do_smart_update() does the required called to update_queue, based on the
 * actual changes that were performed on the semaphore array.
747 748 749
 * 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.
750
 */
751 752
static void do_smart_update(struct sem_array *sma, struct sembuf *sops, int nsops,
			int otime, struct list_head *pt)
753 754
{
	int i;
755
	int progress;
756

757 758 759 760 761
	progress = 1;
retry_global:
	if (sma->complex_count) {
		if (update_queue(sma, -1, pt)) {
			progress = 1;
762
			otime = 1;
763 764
			sops = NULL;
		}
765
	}
766 767
	if (!progress)
		goto done;
768 769 770 771

	if (!sops) {
		/* No semops; something special is going on. */
		for (i = 0; i < sma->sem_nsems; i++) {
772
			if (update_queue(sma, i, pt)) {
773
				otime = 1;
774 775
				progress = 1;
			}
776
		}
777
		goto done_checkretry;
778 779
	}

780
	/* Check the semaphores that were modified. */
781 782 783 784
	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))
785
			if (update_queue(sma, sops[i].sem_num, pt)) {
786
				otime = 1;
787 788 789 790 791 792 793
				progress = 1;
			}
	}
done_checkretry:
	if (progress) {
		progress = 0;
		goto retry_global;
794
	}
795 796 797
done:
	if (otime)
		sma->sem_otime = get_seconds();
798 799 800
}


L
Linus Torvalds 已提交
801 802 803 804 805 806 807 808 809 810 811 812 813 814 815
/* 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;
R
Rik van Riel 已提交
816 817 818 819 820 821 822
	list_for_each_entry(q, &sma->sem_base[semnum].sem_pending, list) {
		struct sembuf * sops = q->sops;
		BUG_ON(sops->sem_num != semnum);
		if ((sops->sem_op < 0) && !(sops->sem_flg & IPC_NOWAIT))
			semncnt++;
	}

823
	list_for_each_entry(q, &sma->sem_pending, list) {
L
Linus Torvalds 已提交
824 825 826 827 828 829 830 831 832 833 834
		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;
}
835

L
Linus Torvalds 已提交
836 837 838 839 840 841
static int count_semzcnt (struct sem_array * sma, ushort semnum)
{
	int semzcnt;
	struct sem_queue * q;

	semzcnt = 0;
R
Rik van Riel 已提交
842 843 844 845 846 847 848
	list_for_each_entry(q, &sma->sem_base[semnum].sem_pending, list) {
		struct sembuf * sops = q->sops;
		BUG_ON(sops->sem_num != semnum);
		if ((sops->sem_op == 0) && !(sops->sem_flg & IPC_NOWAIT))
			semzcnt++;
	}

849
	list_for_each_entry(q, &sma->sem_pending, list) {
L
Linus Torvalds 已提交
850 851 852 853 854 855 856 857 858 859 860 861
		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
862 863 864
/* 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
865
 */
866
static void freeary(struct ipc_namespace *ns, struct kern_ipc_perm *ipcp)
L
Linus Torvalds 已提交
867
{
868 869
	struct sem_undo *un, *tu;
	struct sem_queue *q, *tq;
870
	struct sem_array *sma = container_of(ipcp, struct sem_array, sem_perm);
871
	struct list_head tasks;
872
	int i;
L
Linus Torvalds 已提交
873

874
	/* Free the existing undo structures for this semaphore set.  */
875
	ipc_assert_locked_object(&sma->sem_perm);
876 877 878
	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
879
		un->semid = -1;
880 881
		list_del_rcu(&un->list_proc);
		spin_unlock(&un->ulp->lock);
882
		kfree_rcu(un, rcu);
883
	}
L
Linus Torvalds 已提交
884 885

	/* Wake up all pending processes and let them fail with EIDRM. */
886
	INIT_LIST_HEAD(&tasks);
887
	list_for_each_entry_safe(q, tq, &sma->sem_pending, list) {
888
		unlink_queue(sma, q);
889
		wake_up_sem_queue_prepare(&tasks, q, -EIDRM);
L
Linus Torvalds 已提交
890
	}
891 892 893 894 895 896 897
	for (i = 0; i < sma->sem_nsems; i++) {
		struct sem *sem = sma->sem_base + i;
		list_for_each_entry_safe(q, tq, &sem->sem_pending, list) {
			unlink_queue(sma, q);
			wake_up_sem_queue_prepare(&tasks, q, -EIDRM);
		}
	}
L
Linus Torvalds 已提交
898

N
Nadia Derbey 已提交
899 900
	/* Remove the semaphore set from the IDR */
	sem_rmid(ns, sma);
901
	sem_unlock(sma, -1);
902
	rcu_read_unlock();
L
Linus Torvalds 已提交
903

904
	wake_up_sem_queue_do(&tasks);
K
Kirill Korotaev 已提交
905
	ns->used_sems -= sma->sem_nsems;
L
Linus Torvalds 已提交
906 907 908 909 910 911 912 913 914 915 916 917 918
	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;

919 920
		memset(&out, 0, sizeof(out));

L
Linus Torvalds 已提交
921 922 923 924 925 926 927 928 929 930 931 932 933
		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;
	}
}

934
static int semctl_nolock(struct ipc_namespace *ns, int semid,
935
			 int cmd, int version, void __user *p)
L
Linus Torvalds 已提交
936
{
937
	int err;
L
Linus Torvalds 已提交
938 939 940 941 942 943 944 945 946 947 948 949 950 951
	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
952 953 954 955
		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
956 957 958 959
		seminfo.semvmx = SEMVMX;
		seminfo.semmnu = SEMMNU;
		seminfo.semmap = SEMMAP;
		seminfo.semume = SEMUME;
N
Nadia Derbey 已提交
960
		down_read(&sem_ids(ns).rw_mutex);
L
Linus Torvalds 已提交
961
		if (cmd == SEM_INFO) {
K
Kirill Korotaev 已提交
962 963
			seminfo.semusz = sem_ids(ns).in_use;
			seminfo.semaem = ns->used_sems;
L
Linus Torvalds 已提交
964 965 966 967
		} else {
			seminfo.semusz = SEMUSZ;
			seminfo.semaem = SEMAEM;
		}
N
Nadia Derbey 已提交
968
		max_id = ipc_get_maxid(&sem_ids(ns));
N
Nadia Derbey 已提交
969
		up_read(&sem_ids(ns).rw_mutex);
970
		if (copy_to_user(p, &seminfo, sizeof(struct seminfo))) 
L
Linus Torvalds 已提交
971 972 973
			return -EFAULT;
		return (max_id < 0) ? 0: max_id;
	}
974
	case IPC_STAT:
L
Linus Torvalds 已提交
975 976 977
	case SEM_STAT:
	{
		struct semid64_ds tbuf;
978 979 980
		int id = 0;

		memset(&tbuf, 0, sizeof(tbuf));
L
Linus Torvalds 已提交
981

982
		rcu_read_lock();
983
		if (cmd == SEM_STAT) {
984 985 986 987 988
			sma = sem_obtain_object(ns, semid);
			if (IS_ERR(sma)) {
				err = PTR_ERR(sma);
				goto out_unlock;
			}
989 990
			id = sma->sem_perm.id;
		} else {
991 992 993 994 995
			sma = sem_obtain_object_check(ns, semid);
			if (IS_ERR(sma)) {
				err = PTR_ERR(sma);
				goto out_unlock;
			}
996
		}
L
Linus Torvalds 已提交
997 998

		err = -EACCES;
999
		if (ipcperms(ns, &sma->sem_perm, S_IRUGO))
L
Linus Torvalds 已提交
1000 1001 1002 1003 1004 1005 1006 1007 1008 1009
			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;
1010
		rcu_read_unlock();
1011
		if (copy_semid_to_user(p, &tbuf, version))
L
Linus Torvalds 已提交
1012 1013 1014 1015 1016 1017 1018
			return -EFAULT;
		return id;
	}
	default:
		return -EINVAL;
	}
out_unlock:
1019
	rcu_read_unlock();
L
Linus Torvalds 已提交
1020 1021 1022
	return err;
}

1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039
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

1040 1041
	if (val > SEMVMX || val < 0)
		return -ERANGE;
1042 1043 1044

	INIT_LIST_HEAD(&tasks);

1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061
	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;
	}
1062 1063

	err = security_sem_semctl(sma, SETVAL);
1064 1065 1066 1067
	if (err) {
		rcu_read_unlock();
		return -EACCES;
	}
1068

1069
	sem_lock(sma, NULL, -1);
1070 1071 1072

	curr = &sma->sem_base[semnum];

1073
	ipc_assert_locked_object(&sma->sem_perm);
1074 1075 1076 1077 1078 1079 1080 1081
	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);
1082
	sem_unlock(sma, -1);
1083
	rcu_read_unlock();
1084
	wake_up_sem_queue_do(&tasks);
1085
	return 0;
1086 1087
}

K
Kirill Korotaev 已提交
1088
static int semctl_main(struct ipc_namespace *ns, int semid, int semnum,
1089
		int cmd, void __user *p)
L
Linus Torvalds 已提交
1090 1091 1092
{
	struct sem_array *sma;
	struct sem* curr;
1093
	int err, nsems;
L
Linus Torvalds 已提交
1094 1095
	ushort fast_sem_io[SEMMSL_FAST];
	ushort* sem_io = fast_sem_io;
1096
	struct list_head tasks;
L
Linus Torvalds 已提交
1097

1098 1099 1100 1101 1102 1103
	INIT_LIST_HEAD(&tasks);

	rcu_read_lock();
	sma = sem_obtain_object_check(ns, semid);
	if (IS_ERR(sma)) {
		rcu_read_unlock();
1104
		return PTR_ERR(sma);
1105
	}
L
Linus Torvalds 已提交
1106 1107 1108 1109

	nsems = sma->sem_nsems;

	err = -EACCES;
1110 1111
	if (ipcperms(ns, &sma->sem_perm, cmd == SETALL ? S_IWUGO : S_IRUGO))
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1112 1113

	err = security_sem_semctl(sma, cmd);
1114 1115
	if (err)
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1116 1117 1118 1119 1120

	err = -EACCES;
	switch (cmd) {
	case GETALL:
	{
1121
		ushort __user *array = p;
L
Linus Torvalds 已提交
1122 1123
		int i;

1124
		sem_lock(sma, NULL, -1);
L
Linus Torvalds 已提交
1125
		if(nsems > SEMMSL_FAST) {
1126 1127
			if (!ipc_rcu_getref(sma)) {
				sem_unlock(sma, -1);
1128
				rcu_read_unlock();
1129 1130 1131 1132
				err = -EIDRM;
				goto out_free;
			}
			sem_unlock(sma, -1);
1133
			rcu_read_unlock();
L
Linus Torvalds 已提交
1134 1135
			sem_io = ipc_alloc(sizeof(ushort)*nsems);
			if(sem_io == NULL) {
1136
				sem_putref(sma);
L
Linus Torvalds 已提交
1137 1138 1139
				return -ENOMEM;
			}

1140
			rcu_read_lock();
1141
			sem_lock_and_putref(sma);
L
Linus Torvalds 已提交
1142
			if (sma->sem_perm.deleted) {
1143
				sem_unlock(sma, -1);
1144
				rcu_read_unlock();
L
Linus Torvalds 已提交
1145 1146 1147
				err = -EIDRM;
				goto out_free;
			}
1148
		}
L
Linus Torvalds 已提交
1149 1150
		for (i = 0; i < sma->sem_nsems; i++)
			sem_io[i] = sma->sem_base[i].semval;
1151
		sem_unlock(sma, -1);
1152
		rcu_read_unlock();
L
Linus Torvalds 已提交
1153 1154 1155 1156 1157 1158 1159 1160 1161 1162
		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;

1163 1164 1165 1166
		if (!ipc_rcu_getref(sma)) {
			rcu_read_unlock();
			return -EIDRM;
		}
1167
		rcu_read_unlock();
L
Linus Torvalds 已提交
1168 1169 1170 1171

		if(nsems > SEMMSL_FAST) {
			sem_io = ipc_alloc(sizeof(ushort)*nsems);
			if(sem_io == NULL) {
1172
				sem_putref(sma);
L
Linus Torvalds 已提交
1173 1174 1175 1176
				return -ENOMEM;
			}
		}

1177
		if (copy_from_user (sem_io, p, nsems*sizeof(ushort))) {
1178
			sem_putref(sma);
L
Linus Torvalds 已提交
1179 1180 1181 1182 1183 1184
			err = -EFAULT;
			goto out_free;
		}

		for (i = 0; i < nsems; i++) {
			if (sem_io[i] > SEMVMX) {
1185
				sem_putref(sma);
L
Linus Torvalds 已提交
1186 1187 1188 1189
				err = -ERANGE;
				goto out_free;
			}
		}
1190
		rcu_read_lock();
1191
		sem_lock_and_putref(sma);
L
Linus Torvalds 已提交
1192
		if (sma->sem_perm.deleted) {
1193
			sem_unlock(sma, -1);
1194
			rcu_read_unlock();
L
Linus Torvalds 已提交
1195 1196 1197 1198 1199 1200
			err = -EIDRM;
			goto out_free;
		}

		for (i = 0; i < nsems; i++)
			sma->sem_base[i].semval = sem_io[i];
1201

1202
		ipc_assert_locked_object(&sma->sem_perm);
1203
		list_for_each_entry(un, &sma->list_id, list_id) {
L
Linus Torvalds 已提交
1204 1205
			for (i = 0; i < nsems; i++)
				un->semadj[i] = 0;
1206
		}
L
Linus Torvalds 已提交
1207 1208
		sma->sem_ctime = get_seconds();
		/* maybe some queued-up processes were waiting for this */
1209
		do_smart_update(sma, NULL, 0, 0, &tasks);
L
Linus Torvalds 已提交
1210 1211 1212
		err = 0;
		goto out_unlock;
	}
1213
	/* GETVAL, GETPID, GETNCTN, GETZCNT: fall-through */
L
Linus Torvalds 已提交
1214 1215
	}
	err = -EINVAL;
1216 1217
	if (semnum < 0 || semnum >= nsems)
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1218

1219
	sem_lock(sma, NULL, -1);
L
Linus Torvalds 已提交
1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235
	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;
	}
1236

L
Linus Torvalds 已提交
1237
out_unlock:
1238
	sem_unlock(sma, -1);
1239
out_rcu_wakeup:
1240
	rcu_read_unlock();
1241
	wake_up_sem_queue_do(&tasks);
L
Linus Torvalds 已提交
1242 1243 1244 1245 1246 1247
out_free:
	if(sem_io != fast_sem_io)
		ipc_free(sem_io, sizeof(ushort)*nsems);
	return err;
}

1248 1249
static inline unsigned long
copy_semid_from_user(struct semid64_ds *out, void __user *buf, int version)
L
Linus Torvalds 已提交
1250 1251 1252
{
	switch(version) {
	case IPC_64:
1253
		if (copy_from_user(out, buf, sizeof(*out)))
L
Linus Torvalds 已提交
1254 1255 1256 1257 1258 1259 1260 1261 1262
			return -EFAULT;
		return 0;
	case IPC_OLD:
	    {
		struct semid_ds tbuf_old;

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

1263 1264 1265
		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
1266 1267 1268 1269 1270 1271 1272 1273

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

1274 1275 1276 1277 1278
/*
 * 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.
 */
1279
static int semctl_down(struct ipc_namespace *ns, int semid,
1280
		       int cmd, int version, void __user *p)
L
Linus Torvalds 已提交
1281 1282 1283
{
	struct sem_array *sma;
	int err;
1284
	struct semid64_ds semid64;
L
Linus Torvalds 已提交
1285 1286 1287
	struct kern_ipc_perm *ipcp;

	if(cmd == IPC_SET) {
1288
		if (copy_semid_from_user(&semid64, p, version))
L
Linus Torvalds 已提交
1289 1290
			return -EFAULT;
	}
S
Steve Grubb 已提交
1291

1292 1293 1294
	down_write(&sem_ids(ns).rw_mutex);
	rcu_read_lock();

1295 1296
	ipcp = ipcctl_pre_down_nolock(ns, &sem_ids(ns), semid, cmd,
				      &semid64.sem_perm, 0);
1297 1298 1299 1300
	if (IS_ERR(ipcp)) {
		err = PTR_ERR(ipcp);
		goto out_unlock1;
	}
S
Steve Grubb 已提交
1301

1302
	sma = container_of(ipcp, struct sem_array, sem_perm);
L
Linus Torvalds 已提交
1303 1304

	err = security_sem_semctl(sma, cmd);
1305 1306
	if (err)
		goto out_unlock1;
L
Linus Torvalds 已提交
1307

1308
	switch (cmd) {
L
Linus Torvalds 已提交
1309
	case IPC_RMID:
1310
		sem_lock(sma, NULL, -1);
1311
		/* freeary unlocks the ipc object and rcu */
1312
		freeary(ns, ipcp);
1313
		goto out_up;
L
Linus Torvalds 已提交
1314
	case IPC_SET:
1315
		sem_lock(sma, NULL, -1);
1316 1317
		err = ipc_update_perm(&semid64.sem_perm, ipcp);
		if (err)
1318
			goto out_unlock0;
L
Linus Torvalds 已提交
1319 1320 1321 1322
		sma->sem_ctime = get_seconds();
		break;
	default:
		err = -EINVAL;
1323
		goto out_unlock1;
L
Linus Torvalds 已提交
1324 1325
	}

1326
out_unlock0:
1327
	sem_unlock(sma, -1);
1328
out_unlock1:
1329
	rcu_read_unlock();
1330 1331
out_up:
	up_write(&sem_ids(ns).rw_mutex);
L
Linus Torvalds 已提交
1332 1333 1334
	return err;
}

1335
SYSCALL_DEFINE4(semctl, int, semid, int, semnum, int, cmd, unsigned long, arg)
L
Linus Torvalds 已提交
1336 1337
{
	int version;
K
Kirill Korotaev 已提交
1338
	struct ipc_namespace *ns;
1339
	void __user *p = (void __user *)arg;
L
Linus Torvalds 已提交
1340 1341 1342 1343 1344

	if (semid < 0)
		return -EINVAL;

	version = ipc_parse_version(&cmd);
K
Kirill Korotaev 已提交
1345
	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
1346 1347 1348 1349

	switch(cmd) {
	case IPC_INFO:
	case SEM_INFO:
1350
	case IPC_STAT:
L
Linus Torvalds 已提交
1351
	case SEM_STAT:
1352
		return semctl_nolock(ns, semid, cmd, version, p);
L
Linus Torvalds 已提交
1353 1354 1355 1356 1357 1358
	case GETALL:
	case GETVAL:
	case GETPID:
	case GETNCNT:
	case GETZCNT:
	case SETALL:
1359 1360 1361
		return semctl_main(ns, semid, semnum, cmd, p);
	case SETVAL:
		return semctl_setval(ns, semid, semnum, arg);
L
Linus Torvalds 已提交
1362 1363
	case IPC_RMID:
	case IPC_SET:
1364
		return semctl_down(ns, semid, cmd, version, p);
L
Linus Torvalds 已提交
1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386
	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) {
1387
		undo_list = kzalloc(sizeof(*undo_list), GFP_KERNEL);
L
Linus Torvalds 已提交
1388 1389
		if (undo_list == NULL)
			return -ENOMEM;
I
Ingo Molnar 已提交
1390
		spin_lock_init(&undo_list->lock);
L
Linus Torvalds 已提交
1391
		atomic_set(&undo_list->refcnt, 1);
1392 1393
		INIT_LIST_HEAD(&undo_list->list_proc);

L
Linus Torvalds 已提交
1394 1395 1396 1397 1398 1399
		current->sysvsem.undo_list = undo_list;
	}
	*undo_listp = undo_list;
	return 0;
}

1400
static struct sem_undo *__lookup_undo(struct sem_undo_list *ulp, int semid)
L
Linus Torvalds 已提交
1401
{
1402
	struct sem_undo *un;
1403

1404 1405 1406
	list_for_each_entry_rcu(un, &ulp->list_proc, list_proc) {
		if (un->semid == semid)
			return un;
L
Linus Torvalds 已提交
1407
	}
1408
	return NULL;
L
Linus Torvalds 已提交
1409 1410
}

1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424
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;
}

1425 1426 1427 1428 1429 1430 1431 1432
/**
 * 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.
1433 1434
 * Lifetime-rules: sem_undo is rcu-protected, on success, the function
 * performs a rcu_read_lock().
1435 1436
 */
static struct sem_undo *find_alloc_undo(struct ipc_namespace *ns, int semid)
L
Linus Torvalds 已提交
1437 1438 1439 1440
{
	struct sem_array *sma;
	struct sem_undo_list *ulp;
	struct sem_undo *un, *new;
1441
	int nsems, error;
L
Linus Torvalds 已提交
1442 1443 1444 1445 1446

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

1447
	rcu_read_lock();
1448
	spin_lock(&ulp->lock);
L
Linus Torvalds 已提交
1449
	un = lookup_undo(ulp, semid);
1450
	spin_unlock(&ulp->lock);
L
Linus Torvalds 已提交
1451 1452 1453 1454
	if (likely(un!=NULL))
		goto out;

	/* no undo structure around - allocate one. */
1455
	/* step 1: figure out the size of the semaphore array */
1456 1457 1458
	sma = sem_obtain_object_check(ns, semid);
	if (IS_ERR(sma)) {
		rcu_read_unlock();
J
Julia Lawall 已提交
1459
		return ERR_CAST(sma);
1460
	}
1461

L
Linus Torvalds 已提交
1462
	nsems = sma->sem_nsems;
1463 1464 1465 1466 1467
	if (!ipc_rcu_getref(sma)) {
		rcu_read_unlock();
		un = ERR_PTR(-EIDRM);
		goto out;
	}
1468
	rcu_read_unlock();
L
Linus Torvalds 已提交
1469

1470
	/* step 2: allocate new undo structure */
1471
	new = kzalloc(sizeof(struct sem_undo) + sizeof(short)*nsems, GFP_KERNEL);
L
Linus Torvalds 已提交
1472
	if (!new) {
1473
		sem_putref(sma);
L
Linus Torvalds 已提交
1474 1475 1476
		return ERR_PTR(-ENOMEM);
	}

1477
	/* step 3: Acquire the lock on semaphore array */
1478
	rcu_read_lock();
1479
	sem_lock_and_putref(sma);
L
Linus Torvalds 已提交
1480
	if (sma->sem_perm.deleted) {
1481
		sem_unlock(sma, -1);
1482
		rcu_read_unlock();
L
Linus Torvalds 已提交
1483 1484 1485 1486
		kfree(new);
		un = ERR_PTR(-EIDRM);
		goto out;
	}
1487 1488 1489 1490 1491 1492 1493 1494 1495 1496
	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;
	}
1497 1498
	/* step 5: initialize & link new undo structure */
	new->semadj = (short *) &new[1];
1499
	new->ulp = ulp;
1500 1501
	new->semid = semid;
	assert_spin_locked(&ulp->lock);
1502
	list_add_rcu(&new->list_proc, &ulp->list_proc);
1503
	ipc_assert_locked_object(&sma->sem_perm);
1504
	list_add(&new->list_id, &sma->list_id);
1505
	un = new;
1506

1507
success:
1508
	spin_unlock(&ulp->lock);
1509
	sem_unlock(sma, -1);
L
Linus Torvalds 已提交
1510 1511 1512 1513
out:
	return un;
}

1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540

/**
 * 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;
}


1541 1542
SYSCALL_DEFINE4(semtimedop, int, semid, struct sembuf __user *, tsops,
		unsigned, nsops, const struct timespec __user *, timeout)
L
Linus Torvalds 已提交
1543 1544 1545 1546 1547 1548
{
	int error = -EINVAL;
	struct sem_array *sma;
	struct sembuf fast_sops[SEMOPM_FAST];
	struct sembuf* sops = fast_sops, *sop;
	struct sem_undo *un;
1549
	int undos = 0, alter = 0, max, locknum;
L
Linus Torvalds 已提交
1550 1551
	struct sem_queue queue;
	unsigned long jiffies_left = 0;
K
Kirill Korotaev 已提交
1552
	struct ipc_namespace *ns;
1553
	struct list_head tasks;
K
Kirill Korotaev 已提交
1554 1555

	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
1556 1557 1558

	if (nsops < 1 || semid < 0)
		return -EINVAL;
K
Kirill Korotaev 已提交
1559
	if (nsops > ns->sc_semopm)
L
Linus Torvalds 已提交
1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587
		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)
1588 1589
			undos = 1;
		if (sop->sem_op != 0)
L
Linus Torvalds 已提交
1590 1591 1592
			alter = 1;
	}

1593 1594
	INIT_LIST_HEAD(&tasks);

L
Linus Torvalds 已提交
1595
	if (undos) {
1596
		/* On success, find_alloc_undo takes the rcu_read_lock */
1597
		un = find_alloc_undo(ns, semid);
L
Linus Torvalds 已提交
1598 1599 1600 1601
		if (IS_ERR(un)) {
			error = PTR_ERR(un);
			goto out_free;
		}
1602
	} else {
L
Linus Torvalds 已提交
1603
		un = NULL;
1604 1605
		rcu_read_lock();
	}
L
Linus Torvalds 已提交
1606

1607
	sma = sem_obtain_object_check(ns, semid);
1608
	if (IS_ERR(sma)) {
1609
		rcu_read_unlock();
1610
		error = PTR_ERR(sma);
L
Linus Torvalds 已提交
1611
		goto out_free;
1612 1613
	}

1614
	error = -EFBIG;
1615 1616
	if (max >= sma->sem_nsems)
		goto out_rcu_wakeup;
1617 1618

	error = -EACCES;
1619 1620
	if (ipcperms(ns, &sma->sem_perm, alter ? S_IWUGO : S_IRUGO))
		goto out_rcu_wakeup;
1621 1622

	error = security_sem_semop(sma, sops, nsops, alter);
1623 1624
	if (error)
		goto out_rcu_wakeup;
1625

L
Linus Torvalds 已提交
1626
	/*
1627
	 * semid identifiers are not unique - find_alloc_undo may have
L
Linus Torvalds 已提交
1628
	 * allocated an undo structure, it was invalidated by an RMID
1629
	 * and now a new array with received the same id. Check and fail.
L
Lucas De Marchi 已提交
1630
	 * This case can be detected checking un->semid. The existence of
1631
	 * "un" itself is guaranteed by rcu.
L
Linus Torvalds 已提交
1632
	 */
1633
	error = -EIDRM;
1634 1635 1636
	locknum = sem_lock(sma, sops, nsops);
	if (un && un->semid == -1)
		goto out_unlock_free;
1637

1638
	error = try_atomic_semop (sma, sops, nsops, un, task_tgid_vnr(current));
L
Linus Torvalds 已提交
1639 1640
	if (error <= 0) {
		if (alter && error == 0)
1641
			do_smart_update(sma, sops, nsops, 1, &tasks);
1642

L
Linus Torvalds 已提交
1643 1644 1645 1646 1647 1648 1649 1650 1651 1652
		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;
1653
	queue.pid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
1654 1655
	queue.alter = alter;

1656 1657 1658 1659 1660
	if (nsops == 1) {
		struct sem *curr;
		curr = &sma->sem_base[sops->sem_num];

		if (alter)
1661
			list_add_tail(&queue.list, &curr->sem_pending);
1662
		else
1663
			list_add(&queue.list, &curr->sem_pending);
1664
	} else {
1665 1666 1667 1668
		if (alter)
			list_add_tail(&queue.list, &sma->sem_pending);
		else
			list_add(&queue.list, &sma->sem_pending);
1669 1670 1671
		sma->complex_count++;
	}

L
Linus Torvalds 已提交
1672 1673
	queue.status = -EINTR;
	queue.sleeper = current;
1674 1675

sleep_again:
L
Linus Torvalds 已提交
1676
	current->state = TASK_INTERRUPTIBLE;
1677
	sem_unlock(sma, locknum);
1678
	rcu_read_unlock();
L
Linus Torvalds 已提交
1679 1680 1681 1682 1683 1684

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

1685
	error = get_queue_result(&queue);
L
Linus Torvalds 已提交
1686 1687 1688

	if (error != -EINTR) {
		/* fast path: update_queue already obtained all requested
1689 1690 1691 1692 1693 1694 1695 1696
		 * 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
1697 1698 1699
		goto out_free;
	}

1700
	rcu_read_lock();
1701
	sma = sem_obtain_lock(ns, semid, sops, nsops, &locknum);
1702 1703 1704 1705 1706 1707 1708 1709 1710

	/*
	 * 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().
	 */
1711
	if (IS_ERR(sma)) {
1712
		rcu_read_unlock();
L
Linus Torvalds 已提交
1713 1714 1715
		goto out_free;
	}

1716

L
Linus Torvalds 已提交
1717
	/*
1718 1719
	 * 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
1720
	 */
1721

L
Linus Torvalds 已提交
1722 1723 1724 1725 1726 1727 1728 1729 1730
	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;
1731 1732 1733 1734 1735 1736 1737

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

1738
	unlink_queue(sma, &queue);
L
Linus Torvalds 已提交
1739 1740

out_unlock_free:
1741
	sem_unlock(sma, locknum);
1742
out_rcu_wakeup:
1743
	rcu_read_unlock();
1744
	wake_up_sem_queue_do(&tasks);
L
Linus Torvalds 已提交
1745 1746 1747 1748 1749 1750
out_free:
	if(sops != fast_sops)
		kfree(sops);
	return error;
}

1751 1752
SYSCALL_DEFINE3(semop, int, semid, struct sembuf __user *, tsops,
		unsigned, nsops)
L
Linus Torvalds 已提交
1753 1754 1755 1756 1757 1758 1759 1760 1761 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 1790 1791
{
	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)
{
1792
	struct sem_undo_list *ulp;
L
Linus Torvalds 已提交
1793

1794 1795
	ulp = tsk->sysvsem.undo_list;
	if (!ulp)
L
Linus Torvalds 已提交
1796
		return;
1797
	tsk->sysvsem.undo_list = NULL;
L
Linus Torvalds 已提交
1798

1799
	if (!atomic_dec_and_test(&ulp->refcnt))
L
Linus Torvalds 已提交
1800 1801
		return;

1802
	for (;;) {
L
Linus Torvalds 已提交
1803
		struct sem_array *sma;
1804
		struct sem_undo *un;
1805
		struct list_head tasks;
1806
		int semid, i;
1807

1808
		rcu_read_lock();
1809 1810
		un = list_entry_rcu(ulp->list_proc.next,
				    struct sem_undo, list_proc);
1811 1812 1813 1814
		if (&un->list_proc == &ulp->list_proc)
			semid = -1;
		 else
			semid = un->semid;
1815

1816 1817
		if (semid == -1) {
			rcu_read_unlock();
1818
			break;
1819
		}
L
Linus Torvalds 已提交
1820

1821
		sma = sem_obtain_object_check(tsk->nsproxy->ipc_ns, un->semid);
1822
		/* exit_sem raced with IPC_RMID, nothing to do */
1823 1824
		if (IS_ERR(sma)) {
			rcu_read_unlock();
1825
			continue;
1826
		}
L
Linus Torvalds 已提交
1827

1828
		sem_lock(sma, NULL, -1);
1829
		un = __lookup_undo(ulp, semid);
1830 1831 1832 1833
		if (un == NULL) {
			/* exit_sem raced with IPC_RMID+semget() that created
			 * exactly the same semid. Nothing to do.
			 */
1834
			sem_unlock(sma, -1);
1835
			rcu_read_unlock();
1836 1837 1838 1839
			continue;
		}

		/* remove un from the linked lists */
1840
		ipc_assert_locked_object(&sma->sem_perm);
1841 1842
		list_del(&un->list_id);

1843 1844 1845 1846
		spin_lock(&ulp->lock);
		list_del_rcu(&un->list_proc);
		spin_unlock(&ulp->lock);

1847 1848
		/* perform adjustments registered in un */
		for (i = 0; i < sma->sem_nsems; i++) {
I
Ingo Molnar 已提交
1849
			struct sem * semaphore = &sma->sem_base[i];
1850 1851
			if (un->semadj[i]) {
				semaphore->semval += un->semadj[i];
L
Linus Torvalds 已提交
1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864
				/*
				 * 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
1865 1866 1867 1868
				if (semaphore->semval < 0)
					semaphore->semval = 0;
				if (semaphore->semval > SEMVMX)
					semaphore->semval = SEMVMX;
1869
				semaphore->sempid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
1870 1871 1872
			}
		}
		/* maybe some queued-up processes were waiting for this */
1873 1874
		INIT_LIST_HEAD(&tasks);
		do_smart_update(sma, NULL, 0, 1, &tasks);
1875
		sem_unlock(sma, -1);
1876
		rcu_read_unlock();
1877
		wake_up_sem_queue_do(&tasks);
1878

1879
		kfree_rcu(un, rcu);
L
Linus Torvalds 已提交
1880
	}
1881
	kfree(ulp);
L
Linus Torvalds 已提交
1882 1883 1884
}

#ifdef CONFIG_PROC_FS
1885
static int sysvipc_sem_proc_show(struct seq_file *s, void *it)
L
Linus Torvalds 已提交
1886
{
1887
	struct user_namespace *user_ns = seq_user_ns(s);
1888 1889 1890
	struct sem_array *sma = it;

	return seq_printf(s,
1891
			  "%10d %10d  %4o %10u %5u %5u %5u %5u %10lu %10lu\n",
1892
			  sma->sem_perm.key,
N
Nadia Derbey 已提交
1893
			  sma->sem_perm.id,
1894 1895
			  sma->sem_perm.mode,
			  sma->sem_nsems,
1896 1897 1898 1899
			  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),
1900 1901
			  sma->sem_otime,
			  sma->sem_ctime);
L
Linus Torvalds 已提交
1902 1903
}
#endif