sem.c 58.3 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
L
Linus Torvalds 已提交
2 3 4 5 6 7 8 9
/*
 * 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
10
 * (c) 1999 Manfred Spraul <manfred@colorfullife.com>
L
Linus Torvalds 已提交
11
 * Enforced range limit on SEM_UNDO
A
Alan Cox 已提交
12
 * (c) 2001 Red Hat Inc
L
Linus Torvalds 已提交
13 14
 * Lockless wakeup
 * (c) 2003 Manfred Spraul <manfred@colorfullife.com>
D
Davidlohr Bueso 已提交
15
 * (c) 2016 Davidlohr Bueso <dave@stgolabs.net>
16 17
 * Further wakeup optimizations, documentation
 * (c) 2010 Manfred Spraul <manfred@colorfullife.com>
S
Steve Grubb 已提交
18 19 20
 *
 * support for audit of ipc object properties and permission changes
 * Dustin Kirkland <dustin.kirkland@us.ibm.com>
K
Kirill Korotaev 已提交
21 22 23 24
 *
 * namespaces support
 * OpenVZ, SWsoft Inc.
 * Pavel Emelianov <xemul@openvz.org>
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
 *
 * 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.
52
 * - semncnt and semzcnt are calculated on demand in count_semcnt()
53 54 55 56 57
 * - 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
D
Davidlohr Bueso 已提交
58
 *   dropping all locks. (see wake_up_sem_queue_prepare())
59 60 61 62 63 64 65 66 67 68 69 70
 * - 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).
 * - 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
71 72 73 74 75 76 77 78 79 80
 */

#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>
81
#include <linux/capability.h>
82
#include <linux/seq_file.h>
N
Nadia Derbey 已提交
83
#include <linux/rwsem.h>
K
Kirill Korotaev 已提交
84
#include <linux/nsproxy.h>
85
#include <linux/ipc_namespace.h>
86
#include <linux/sched/wake_q.h>
I
Ingo Molnar 已提交
87

P
Paul McQuade 已提交
88
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
89 90
#include "util.h"

91 92 93 94 95 96 97 98 99

/* 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 */
100
	struct sembuf		*blocking; /* the operation that blocked */
101
	int			nsops;	 /* number of operations */
102 103
	bool			alter;	 /* does *sops alter the array? */
	bool                    dupsop;	 /* sops on more than one sem_num */
104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125
};

/* 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 {
126
	refcount_t		refcnt;
127 128 129 130 131
	spinlock_t		lock;
	struct list_head	list_proc;
};


132
#define sem_ids(ns)	((ns)->ids[IPC_SEM_IDS])
K
Kirill Korotaev 已提交
133

N
Nadia Derbey 已提交
134
static int newary(struct ipc_namespace *, struct ipc_params *);
135
static void freeary(struct ipc_namespace *, struct kern_ipc_perm *);
L
Linus Torvalds 已提交
136
#ifdef CONFIG_PROC_FS
137
static int sysvipc_sem_proc_show(struct seq_file *s, void *it);
L
Linus Torvalds 已提交
138 139 140 141 142
#endif

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

M
Manfred Spraul 已提交
143 144 145 146 147 148 149
/*
 * Switching from the mode suitable for simple ops
 * to the mode for complex ops is costly. Therefore:
 * use some hysteresis
 */
#define USE_GLOBAL_LOCK_HYSTERESIS	10

L
Linus Torvalds 已提交
150
/*
151
 * Locking:
152
 * a) global sem_lock() for read/write
L
Linus Torvalds 已提交
153
 *	sem_undo.id_next,
154
 *	sem_array.complex_count,
155 156
 *	sem_array.pending{_alter,_const},
 *	sem_array.sem_undo
P
Paul McQuade 已提交
157
 *
158
 * b) global or semaphore sem_lock() for read/write:
159
 *	sem_array.sems[i].pending_{const,alter}:
160 161 162 163 164
 *
 * c) special:
 *	sem_undo_list.list_proc:
 *	* undo_list->lock for write
 *	* rcu for read
M
Manfred Spraul 已提交
165 166 167 168 169 170 171 172 173 174 175 176 177 178 179
 *	use_global_lock:
 *	* global sem_lock() for write
 *	* either local or global sem_lock() for read.
 *
 * Memory ordering:
 * Most ordering is enforced by using spin_lock() and spin_unlock().
 * The special case is use_global_lock:
 * Setting it from non-zero to 0 is a RELEASE, this is ensured by
 * using smp_store_release().
 * Testing if it is non-zero is an ACQUIRE, this is ensured by using
 * smp_load_acquire().
 * Setting it from 0 to non-zero must be ordered with regards to
 * this smp_load_acquire(), this is guaranteed because the smp_load_acquire()
 * is inside a spin_lock() and after a write from 0 to non-zero a
 * spin_lock()+spin_unlock() is done.
L
Linus Torvalds 已提交
180 181
 */

K
Kirill Korotaev 已提交
182 183 184 185 186
#define sc_semmsl	sem_ctls[0]
#define sc_semmns	sem_ctls[1]
#define sc_semopm	sem_ctls[2]
#define sc_semmni	sem_ctls[3]

187
int sem_init_ns(struct ipc_namespace *ns)
K
Kirill Korotaev 已提交
188 189 190 191 192 193
{
	ns->sc_semmsl = SEMMSL;
	ns->sc_semmns = SEMMNS;
	ns->sc_semopm = SEMOPM;
	ns->sc_semmni = SEMMNI;
	ns->used_sems = 0;
194
	return ipc_init_ids(&ns->ids[IPC_SEM_IDS]);
K
Kirill Korotaev 已提交
195 196
}

197
#ifdef CONFIG_IPC_NS
K
Kirill Korotaev 已提交
198 199
void sem_exit_ns(struct ipc_namespace *ns)
{
200
	free_ipcs(ns, &sem_ids(ns), freeary);
S
Serge E. Hallyn 已提交
201
	idr_destroy(&ns->ids[IPC_SEM_IDS].ipcs_idr);
202
	rhashtable_destroy(&ns->ids[IPC_SEM_IDS].key_ht);
K
Kirill Korotaev 已提交
203
}
204
#endif
L
Linus Torvalds 已提交
205

206
int __init sem_init(void)
L
Linus Torvalds 已提交
207
{
208 209
	const int err = sem_init_ns(&init_ipc_ns);

210 211
	ipc_init_proc_interface("sysvipc/sem",
				"       key      semid perms      nsems   uid   gid  cuid  cgid      otime      ctime\n",
K
Kirill Korotaev 已提交
212
				IPC_SEM_IDS, sysvipc_sem_proc_show);
213
	return err;
L
Linus Torvalds 已提交
214 215
}

216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236
/**
 * unmerge_queues - unmerge queues, if possible.
 * @sma: semaphore array
 *
 * The function unmerges the wait queues if complex_count is 0.
 * It must be called prior to dropping the global semaphore array lock.
 */
static void unmerge_queues(struct sem_array *sma)
{
	struct sem_queue *q, *tq;

	/* complex operations still around? */
	if (sma->complex_count)
		return;
	/*
	 * We will switch back to simple mode.
	 * Move all pending operation back into the per-semaphore
	 * queues.
	 */
	list_for_each_entry_safe(q, tq, &sma->pending_alter, list) {
		struct sem *curr;
237
		curr = &sma->sems[q->sops[0].sem_num];
238 239 240 241 242 243 244

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

/**
D
Davidlohr Bueso 已提交
245
 * merge_queues - merge single semop queues into global queue
246 247 248 249 250 251 252 253 254 255 256
 * @sma: semaphore array
 *
 * This function merges all per-semaphore queues into the global queue.
 * It is necessary to achieve FIFO ordering for the pending single-sop
 * operations when a multi-semop operation must sleep.
 * Only the alter operations must be moved, the const operations can stay.
 */
static void merge_queues(struct sem_array *sma)
{
	int i;
	for (i = 0; i < sma->sem_nsems; i++) {
257
		struct sem *sem = &sma->sems[i];
258 259 260 261 262

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

D
Davidlohr Bueso 已提交
263 264
static void sem_rcu_free(struct rcu_head *head)
{
265 266
	struct kern_ipc_perm *p = container_of(head, struct kern_ipc_perm, rcu);
	struct sem_array *sma = container_of(p, struct sem_array, sem_perm);
D
Davidlohr Bueso 已提交
267 268

	security_sem_free(sma);
K
Kees Cook 已提交
269
	kvfree(sma);
D
Davidlohr Bueso 已提交
270 271
}

272
/*
273
 * Enter the mode suitable for non-simple operations:
274 275
 * Caller must own sem_perm.lock.
 */
276
static void complexmode_enter(struct sem_array *sma)
277 278 279 280
{
	int i;
	struct sem *sem;

M
Manfred Spraul 已提交
281 282 283 284 285 286 287
	if (sma->use_global_lock > 0)  {
		/*
		 * We are already in global lock mode.
		 * Nothing to do, just reset the
		 * counter until we return to simple mode.
		 */
		sma->use_global_lock = USE_GLOBAL_LOCK_HYSTERESIS;
M
Manfred Spraul 已提交
288 289
		return;
	}
M
Manfred Spraul 已提交
290
	sma->use_global_lock = USE_GLOBAL_LOCK_HYSTERESIS;
291

292
	for (i = 0; i < sma->sem_nsems; i++) {
293
		sem = &sma->sems[i];
294 295
		spin_lock(&sem->lock);
		spin_unlock(&sem->lock);
296
	}
297 298 299 300 301 302 303 304 305 306 307 308 309 310
}

/*
 * Try to leave the mode that disallows simple operations:
 * Caller must own sem_perm.lock.
 */
static void complexmode_tryleave(struct sem_array *sma)
{
	if (sma->complex_count)  {
		/* Complex ops are sleeping.
		 * We must stay in complex mode
		 */
		return;
	}
M
Manfred Spraul 已提交
311 312 313 314 315 316 317 318 319 320 321
	if (sma->use_global_lock == 1) {
		/*
		 * Immediately after setting use_global_lock to 0,
		 * a simple op can start. Thus: all memory writes
		 * performed by the current operation must be visible
		 * before we set use_global_lock to 0.
		 */
		smp_store_release(&sma->use_global_lock, 0);
	} else {
		sma->use_global_lock--;
	}
322 323
}

324
#define SEM_GLOBAL_LOCK	(-1)
325 326 327 328 329 330 331 332 333 334
/*
 * If the request contains only one semaphore operation, and there are
 * no complex transactions pending, lock only the semaphore involved.
 * Otherwise, lock the entire semaphore array, since we either have
 * multiple semaphores in our own semops, or we need to look at
 * semaphores from other pending complex operations.
 */
static inline int sem_lock(struct sem_array *sma, struct sembuf *sops,
			      int nsops)
{
335
	struct sem *sem;
336

337 338 339
	if (nsops != 1) {
		/* Complex operation - acquire a full lock */
		ipc_lock_object(&sma->sem_perm);
340

341 342 343
		/* Prevent parallel simple ops */
		complexmode_enter(sma);
		return SEM_GLOBAL_LOCK;
344 345 346 347
	}

	/*
	 * Only one semaphore affected - try to optimize locking.
348 349 350
	 * Optimized locking is possible if no complex operation
	 * is either enqueued or processed right now.
	 *
M
Manfred Spraul 已提交
351
	 * Both facts are tracked by use_global_mode.
352
	 */
353
	sem = &sma->sems[sops->sem_num];
354

355
	/*
M
Manfred Spraul 已提交
356
	 * Initial check for use_global_lock. Just an optimization,
357 358
	 * no locking, no memory barrier.
	 */
M
Manfred Spraul 已提交
359
	if (!sma->use_global_lock) {
360
		/*
361 362
		 * It appears that no complex operation is around.
		 * Acquire the per-semaphore lock.
363
		 */
364 365
		spin_lock(&sem->lock);

M
Manfred Spraul 已提交
366 367
		/* pairs with smp_store_release() */
		if (!smp_load_acquire(&sma->use_global_lock)) {
368 369
			/* fast path successful! */
			return sops->sem_num;
370
		}
371 372 373 374 375
		spin_unlock(&sem->lock);
	}

	/* slow path: acquire the full lock */
	ipc_lock_object(&sma->sem_perm);
376

M
Manfred Spraul 已提交
377 378 379 380 381 382 383 384 385
	if (sma->use_global_lock == 0) {
		/*
		 * The use_global_lock mode ended while we waited for
		 * sma->sem_perm.lock. Thus we must switch to locking
		 * with sem->lock.
		 * Unlike in the fast path, there is no need to recheck
		 * sma->use_global_lock after we have acquired sem->lock:
		 * We own sma->sem_perm.lock, thus use_global_lock cannot
		 * change.
386 387
		 */
		spin_lock(&sem->lock);
M
Manfred Spraul 已提交
388

389 390
		ipc_unlock_object(&sma->sem_perm);
		return sops->sem_num;
391
	} else {
M
Manfred Spraul 已提交
392 393 394 395
		/*
		 * Not a false alarm, thus continue to use the global lock
		 * mode. No need for complexmode_enter(), this was done by
		 * the caller that has set use_global_mode to non-zero.
396
		 */
397
		return SEM_GLOBAL_LOCK;
398 399 400 401 402
	}
}

static inline void sem_unlock(struct sem_array *sma, int locknum)
{
403
	if (locknum == SEM_GLOBAL_LOCK) {
404
		unmerge_queues(sma);
405
		complexmode_tryleave(sma);
406
		ipc_unlock_object(&sma->sem_perm);
407
	} else {
408
		struct sem *sem = &sma->sems[locknum];
409 410 411 412
		spin_unlock(&sem->lock);
	}
}

N
Nadia Derbey 已提交
413
/*
D
Davidlohr Bueso 已提交
414
 * sem_lock_(check_) routines are called in the paths where the rwsem
N
Nadia Derbey 已提交
415
 * is not held.
416 417
 *
 * The caller holds the RCU read lock.
N
Nadia Derbey 已提交
418
 */
419 420
static inline struct sem_array *sem_obtain_object(struct ipc_namespace *ns, int id)
{
D
Davidlohr Bueso 已提交
421
	struct kern_ipc_perm *ipcp = ipc_obtain_object_idr(&sem_ids(ns), id);
422 423 424 425 426 427 428 429 430 431 432 433 434 435

	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);
436

N
Nadia Derbey 已提交
437
	return container_of(ipcp, struct sem_array, sem_perm);
438 439
}

440 441
static inline void sem_lock_and_putref(struct sem_array *sma)
{
442
	sem_lock(sma, NULL, -1);
443
	ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
444 445
}

N
Nadia Derbey 已提交
446 447 448 449 450
static inline void sem_rmid(struct ipc_namespace *ns, struct sem_array *s)
{
	ipc_rmid(&sem_ids(ns), &s->sem_perm);
}

K
Kees Cook 已提交
451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468
static struct sem_array *sem_alloc(size_t nsems)
{
	struct sem_array *sma;
	size_t size;

	if (nsems > (INT_MAX - sizeof(*sma)) / sizeof(sma->sems[0]))
		return NULL;

	size = sizeof(*sma) + nsems * sizeof(sma->sems[0]);
	sma = kvmalloc(size, GFP_KERNEL);
	if (unlikely(!sma))
		return NULL;

	memset(sma, 0, size);

	return sma;
}

N
Nadia Derbey 已提交
469 470 471 472 473
/**
 * newary - Create a new semaphore set
 * @ns: namespace
 * @params: ptr to the structure that contains key, semflg and nsems
 *
D
Davidlohr Bueso 已提交
474
 * Called with sem_ids.rwsem held (as a writer)
N
Nadia Derbey 已提交
475
 */
N
Nadia Derbey 已提交
476
static int newary(struct ipc_namespace *ns, struct ipc_params *params)
L
Linus Torvalds 已提交
477 478 479
{
	int retval;
	struct sem_array *sma;
N
Nadia Derbey 已提交
480 481 482
	key_t key = params->key;
	int nsems = params->u.nsems;
	int semflg = params->flg;
483
	int i;
L
Linus Torvalds 已提交
484 485 486

	if (!nsems)
		return -EINVAL;
K
Kirill Korotaev 已提交
487
	if (ns->used_sems + nsems > ns->sc_semmns)
L
Linus Torvalds 已提交
488 489
		return -ENOSPC;

K
Kees Cook 已提交
490
	sma = sem_alloc(nsems);
491
	if (!sma)
L
Linus Torvalds 已提交
492
		return -ENOMEM;
493

L
Linus Torvalds 已提交
494 495 496 497 498 499
	sma->sem_perm.mode = (semflg & S_IRWXUGO);
	sma->sem_perm.key = key;

	sma->sem_perm.security = NULL;
	retval = security_sem_alloc(sma);
	if (retval) {
K
Kees Cook 已提交
500
		kvfree(sma);
L
Linus Torvalds 已提交
501 502 503
		return retval;
	}

504
	for (i = 0; i < nsems; i++) {
505 506 507
		INIT_LIST_HEAD(&sma->sems[i].pending_alter);
		INIT_LIST_HEAD(&sma->sems[i].pending_const);
		spin_lock_init(&sma->sems[i].lock);
508
	}
509 510

	sma->complex_count = 0;
M
Manfred Spraul 已提交
511
	sma->use_global_lock = USE_GLOBAL_LOCK_HYSTERESIS;
512 513
	INIT_LIST_HEAD(&sma->pending_alter);
	INIT_LIST_HEAD(&sma->pending_const);
514
	INIT_LIST_HEAD(&sma->list_id);
L
Linus Torvalds 已提交
515
	sma->sem_nsems = nsems;
516
	sma->sem_ctime = ktime_get_real_seconds();
517

518
	/* ipc_addid() locks sma upon success. */
519 520 521 522
	retval = ipc_addid(&sem_ids(ns), &sma->sem_perm, ns->sc_semmni);
	if (retval < 0) {
		call_rcu(&sma->sem_perm.rcu, sem_rcu_free);
		return retval;
523 524 525
	}
	ns->used_sems += nsems;

526
	sem_unlock(sma, -1);
527
	rcu_read_unlock();
L
Linus Torvalds 已提交
528

N
Nadia Derbey 已提交
529
	return sma->sem_perm.id;
L
Linus Torvalds 已提交
530 531
}

N
Nadia Derbey 已提交
532

N
Nadia Derbey 已提交
533
/*
D
Davidlohr Bueso 已提交
534
 * Called with sem_ids.rwsem and ipcp locked.
N
Nadia Derbey 已提交
535
 */
N
Nadia Derbey 已提交
536
static inline int sem_security(struct kern_ipc_perm *ipcp, int semflg)
N
Nadia Derbey 已提交
537
{
N
Nadia Derbey 已提交
538 539 540 541
	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
542 543
}

N
Nadia Derbey 已提交
544
/*
D
Davidlohr Bueso 已提交
545
 * Called with sem_ids.rwsem and ipcp locked.
N
Nadia Derbey 已提交
546
 */
N
Nadia Derbey 已提交
547 548
static inline int sem_more_checks(struct kern_ipc_perm *ipcp,
				struct ipc_params *params)
N
Nadia Derbey 已提交
549
{
N
Nadia Derbey 已提交
550 551 552 553
	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
554 555 556 557 558
		return -EINVAL;

	return 0;
}

559
SYSCALL_DEFINE3(semget, key_t, key, int, nsems, int, semflg)
L
Linus Torvalds 已提交
560
{
K
Kirill Korotaev 已提交
561
	struct ipc_namespace *ns;
M
Mathias Krause 已提交
562 563 564 565 566
	static const struct ipc_ops sem_ops = {
		.getnew = newary,
		.associate = sem_security,
		.more_checks = sem_more_checks,
	};
N
Nadia Derbey 已提交
567
	struct ipc_params sem_params;
K
Kirill Korotaev 已提交
568 569

	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
570

K
Kirill Korotaev 已提交
571
	if (nsems < 0 || nsems > ns->sc_semmsl)
L
Linus Torvalds 已提交
572
		return -EINVAL;
N
Nadia Derbey 已提交
573

N
Nadia Derbey 已提交
574 575 576
	sem_params.key = key;
	sem_params.flg = semflg;
	sem_params.u.nsems = nsems;
L
Linus Torvalds 已提交
577

N
Nadia Derbey 已提交
578
	return ipcget(ns, &sem_ids(ns), &sem_ops, &sem_params);
L
Linus Torvalds 已提交
579 580
}

581
/**
582 583
 * perform_atomic_semop[_slow] - Attempt to perform semaphore
 *                               operations on a given array.
584
 * @sma: semaphore array
585
 * @q: struct sem_queue that describes the operation
586
 *
587 588 589 590 591 592 593
 * Caller blocking are as follows, based the value
 * indicated by the semaphore operation (sem_op):
 *
 *  (1) >0 never blocks.
 *  (2)  0 (wait-for-zero operation): semval is non-zero.
 *  (3) <0 attempting to decrement semval to a value smaller than zero.
 *
594 595
 * Returns 0 if the operation was possible.
 * Returns 1 if the operation is impossible, the caller must sleep.
596
 * Returns <0 for error codes.
L
Linus Torvalds 已提交
597
 */
598
static int perform_atomic_semop_slow(struct sem_array *sma, struct sem_queue *q)
L
Linus Torvalds 已提交
599
{
600
	int result, sem_op, nsops, pid;
L
Linus Torvalds 已提交
601
	struct sembuf *sop;
M
Manfred Spraul 已提交
602
	struct sem *curr;
603 604 605 606 607 608
	struct sembuf *sops;
	struct sem_undo *un;

	sops = q->sops;
	nsops = q->nsops;
	un = q->undo;
L
Linus Torvalds 已提交
609 610

	for (sop = sops; sop < sops + nsops; sop++) {
611
		curr = &sma->sems[sop->sem_num];
L
Linus Torvalds 已提交
612 613
		sem_op = sop->sem_op;
		result = curr->semval;
614

L
Linus Torvalds 已提交
615 616 617 618 619 620 621 622
		if (!sem_op && result)
			goto would_block;

		result += sem_op;
		if (result < 0)
			goto would_block;
		if (result > SEMVMX)
			goto out_of_range;
623

L
Linus Torvalds 已提交
624 625
		if (sop->sem_flg & SEM_UNDO) {
			int undo = un->semadj[sop->sem_num] - sem_op;
626
			/* Exceeding the undo range is an error. */
L
Linus Torvalds 已提交
627 628
			if (undo < (-SEMAEM - 1) || undo > SEMAEM)
				goto out_of_range;
629
			un->semadj[sop->sem_num] = undo;
L
Linus Torvalds 已提交
630
		}
631

L
Linus Torvalds 已提交
632 633 634 635
		curr->semval = result;
	}

	sop--;
636
	pid = q->pid;
L
Linus Torvalds 已提交
637
	while (sop >= sops) {
638
		sma->sems[sop->sem_num].sempid = pid;
L
Linus Torvalds 已提交
639 640
		sop--;
	}
641

L
Linus Torvalds 已提交
642 643 644 645 646 647 648
	return 0;

out_of_range:
	result = -ERANGE;
	goto undo;

would_block:
649 650
	q->blocking = sop;

L
Linus Torvalds 已提交
651 652 653 654 655 656 657 658
	if (sop->sem_flg & IPC_NOWAIT)
		result = -EAGAIN;
	else
		result = 1;

undo:
	sop--;
	while (sop >= sops) {
659
		sem_op = sop->sem_op;
660
		sma->sems[sop->sem_num].semval -= sem_op;
661 662
		if (sop->sem_flg & SEM_UNDO)
			un->semadj[sop->sem_num] += sem_op;
L
Linus Torvalds 已提交
663 664 665 666 667 668
		sop--;
	}

	return result;
}

669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690
static int perform_atomic_semop(struct sem_array *sma, struct sem_queue *q)
{
	int result, sem_op, nsops;
	struct sembuf *sop;
	struct sem *curr;
	struct sembuf *sops;
	struct sem_undo *un;

	sops = q->sops;
	nsops = q->nsops;
	un = q->undo;

	if (unlikely(q->dupsop))
		return perform_atomic_semop_slow(sma, q);

	/*
	 * We scan the semaphore set twice, first to ensure that the entire
	 * operation can succeed, therefore avoiding any pointless writes
	 * to shared memory and having to undo such changes in order to block
	 * until the operations can go through.
	 */
	for (sop = sops; sop < sops + nsops; sop++) {
691
		curr = &sma->sems[sop->sem_num];
692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714
		sem_op = sop->sem_op;
		result = curr->semval;

		if (!sem_op && result)
			goto would_block; /* wait-for-zero */

		result += sem_op;
		if (result < 0)
			goto would_block;

		if (result > SEMVMX)
			return -ERANGE;

		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)
				return -ERANGE;
		}
	}

	for (sop = sops; sop < sops + nsops; sop++) {
715
		curr = &sma->sems[sop->sem_num];
716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734
		sem_op = sop->sem_op;
		result = curr->semval;

		if (sop->sem_flg & SEM_UNDO) {
			int undo = un->semadj[sop->sem_num] - sem_op;

			un->semadj[sop->sem_num] = undo;
		}
		curr->semval += sem_op;
		curr->sempid = q->pid;
	}

	return 0;

would_block:
	q->blocking = sop;
	return sop->sem_flg & IPC_NOWAIT ? -EAGAIN : 1;
}

D
Davidlohr Bueso 已提交
735 736
static inline void wake_up_sem_queue_prepare(struct sem_queue *q, int error,
					     struct wake_q_head *wake_q)
737
{
D
Davidlohr Bueso 已提交
738 739 740 741 742 743 744 745 746
	wake_q_add(wake_q, q->sleeper);
	/*
	 * Rely on the above implicit barrier, such that we can
	 * ensure that we hold reference to the task before setting
	 * q->status. Otherwise we could race with do_exit if the
	 * task is awoken by an external event before calling
	 * wake_up_process().
	 */
	WRITE_ONCE(q->status, error);
N
Nick Piggin 已提交
747 748
}

749 750 751
static void unlink_queue(struct sem_array *sma, struct sem_queue *q)
{
	list_del(&q->list);
752
	if (q->nsops > 1)
753 754 755
		sma->complex_count--;
}

756 757 758 759 760 761 762
/** 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
763 764
 * modified the array.
 * Note that wait-for-zero operations are handled without restart.
765
 */
766
static inline int check_restart(struct sem_array *sma, struct sem_queue *q)
767
{
768 769
	/* pending complex alter operations are too difficult to analyse */
	if (!list_empty(&sma->pending_alter))
770 771 772 773 774 775
		return 1;

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

776 777 778 779 780 781 782 783 784 785 786 787 788
	/* 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;
}
789

790
/**
D
Davidlohr Bueso 已提交
791
 * wake_const_ops - wake up non-alter tasks
792 793
 * @sma: semaphore array.
 * @semnum: semaphore that was modified.
D
Davidlohr Bueso 已提交
794
 * @wake_q: lockless wake-queue head.
795 796 797 798 799
 *
 * 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.
D
Davidlohr Bueso 已提交
800
 * The tasks that must be woken up are added to @wake_q. The return code
801 802 803 804
 * 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,
D
Davidlohr Bueso 已提交
805
			  struct wake_q_head *wake_q)
806
{
807
	struct sem_queue *q, *tmp;
808 809 810 811 812 813
	struct list_head *pending_list;
	int semop_completed = 0;

	if (semnum == -1)
		pending_list = &sma->pending_const;
	else
814
		pending_list = &sma->sems[semnum].pending_const;
815

816 817
	list_for_each_entry_safe(q, tmp, pending_list, list) {
		int error = perform_atomic_semop(sma, q);
818

819 820 821 822
		if (error > 0)
			continue;
		/* operation completed, remove from queue & wakeup */
		unlink_queue(sma, q);
823

824 825 826
		wake_up_sem_queue_prepare(q, error, wake_q);
		if (error == 0)
			semop_completed = 1;
827
	}
828

829 830 831 832
	return semop_completed;
}

/**
D
Davidlohr Bueso 已提交
833
 * do_smart_wakeup_zero - wakeup all wait for zero tasks
834 835 836
 * @sma: semaphore array
 * @sops: operations that were performed
 * @nsops: number of operations
D
Davidlohr Bueso 已提交
837
 * @wake_q: lockless wake-queue head
838
 *
D
Davidlohr Bueso 已提交
839 840
 * Checks all required queue for wait-for-zero operations, based
 * on the actual changes that were performed on the semaphore array.
841 842 843
 * 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,
D
Davidlohr Bueso 已提交
844
				int nsops, struct wake_q_head *wake_q)
845 846 847 848 849 850 851 852 853 854
{
	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;

855
			if (sma->sems[num].semval == 0) {
856
				got_zero = 1;
D
Davidlohr Bueso 已提交
857
				semop_completed |= wake_const_ops(sma, num, wake_q);
858 859 860 861 862 863
			}
		}
	} else {
		/*
		 * No sops means modified semaphores not known.
		 * Assume all were changed.
864
		 */
865
		for (i = 0; i < sma->sem_nsems; i++) {
866
			if (sma->sems[i].semval == 0) {
867
				got_zero = 1;
D
Davidlohr Bueso 已提交
868
				semop_completed |= wake_const_ops(sma, i, wake_q);
869 870
			}
		}
871 872
	}
	/*
873 874
	 * If one of the modified semaphores got 0,
	 * then check the global queue, too.
875
	 */
876
	if (got_zero)
D
Davidlohr Bueso 已提交
877
		semop_completed |= wake_const_ops(sma, -1, wake_q);
878

879
	return semop_completed;
880 881
}

882 883

/**
D
Davidlohr Bueso 已提交
884
 * update_queue - look for tasks that can be completed.
885 886
 * @sma: semaphore array.
 * @semnum: semaphore that was modified.
D
Davidlohr Bueso 已提交
887
 * @wake_q: lockless wake-queue head.
888 889
 *
 * update_queue must be called after a semaphore in a semaphore array
890 891 892
 * 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.
D
Davidlohr Bueso 已提交
893
 * The tasks that must be woken up are added to @wake_q. The return code
894
 * is stored in q->pid.
895 896
 * The function internally checks if const operations can now succeed.
 *
897
 * The function return 1 if at least one semop was completed successfully.
L
Linus Torvalds 已提交
898
 */
D
Davidlohr Bueso 已提交
899
static int update_queue(struct sem_array *sma, int semnum, struct wake_q_head *wake_q)
L
Linus Torvalds 已提交
900
{
901
	struct sem_queue *q, *tmp;
902
	struct list_head *pending_list;
903
	int semop_completed = 0;
904

905
	if (semnum == -1)
906
		pending_list = &sma->pending_alter;
907
	else
908
		pending_list = &sma->sems[semnum].pending_alter;
N
Nick Piggin 已提交
909 910

again:
911
	list_for_each_entry_safe(q, tmp, pending_list, list) {
912
		int error, restart;
913

914 915
		/* If we are scanning the single sop, per-semaphore list of
		 * one semaphore and that semaphore is 0, then it is not
916
		 * necessary to scan further: simple increments
917 918 919 920
		 * 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.
		 */
921
		if (semnum != -1 && sma->sems[semnum].semval == 0)
922 923
			break;

924
		error = perform_atomic_semop(sma, q);
L
Linus Torvalds 已提交
925 926

		/* Does q->sleeper still need to sleep? */
N
Nick Piggin 已提交
927 928 929
		if (error > 0)
			continue;

930
		unlink_queue(sma, q);
N
Nick Piggin 已提交
931

932
		if (error) {
933
			restart = 0;
934 935
		} else {
			semop_completed = 1;
D
Davidlohr Bueso 已提交
936
			do_smart_wakeup_zero(sma, q->sops, q->nsops, wake_q);
937
			restart = check_restart(sma, q);
938
		}
939

D
Davidlohr Bueso 已提交
940
		wake_up_sem_queue_prepare(q, error, wake_q);
941
		if (restart)
N
Nick Piggin 已提交
942
			goto again;
L
Linus Torvalds 已提交
943
	}
944
	return semop_completed;
L
Linus Torvalds 已提交
945 946
}

947
/**
D
Davidlohr Bueso 已提交
948
 * set_semotime - set sem_otime
949 950 951 952 953 954 955 956 957
 * @sma: semaphore array
 * @sops: operations that modified the array, may be NULL
 *
 * sem_otime is replicated to avoid cache line trashing.
 * This function sets one instance to the current time.
 */
static void set_semotime(struct sem_array *sma, struct sembuf *sops)
{
	if (sops == NULL) {
958
		sma->sems[0].sem_otime = get_seconds();
959
	} else {
960
		sma->sems[sops[0].sem_num].sem_otime =
961 962 963 964
							get_seconds();
	}
}

965
/**
D
Davidlohr Bueso 已提交
966
 * do_smart_update - optimized update_queue
967 968 969
 * @sma: semaphore array
 * @sops: operations that were performed
 * @nsops: number of operations
970
 * @otime: force setting otime
D
Davidlohr Bueso 已提交
971
 * @wake_q: lockless wake-queue head
972
 *
973 974
 * 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.
975
 * Note that the function does not do the actual wake-up: the caller is
D
Davidlohr Bueso 已提交
976
 * responsible for calling wake_up_q().
977
 * It is safe to perform this call after dropping all locks.
978
 */
979
static void do_smart_update(struct sem_array *sma, struct sembuf *sops, int nsops,
D
Davidlohr Bueso 已提交
980
			    int otime, struct wake_q_head *wake_q)
981 982 983
{
	int i;

D
Davidlohr Bueso 已提交
984
	otime |= do_smart_wakeup_zero(sma, sops, nsops, wake_q);
985

986 987
	if (!list_empty(&sma->pending_alter)) {
		/* semaphore array uses the global queue - just process it. */
D
Davidlohr Bueso 已提交
988
		otime |= update_queue(sma, -1, wake_q);
989 990 991 992 993 994 995
	} else {
		if (!sops) {
			/*
			 * No sops, thus the modified semaphores are not
			 * known. Check all.
			 */
			for (i = 0; i < sma->sem_nsems; i++)
D
Davidlohr Bueso 已提交
996
				otime |= update_queue(sma, i, wake_q);
997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009
		} else {
			/*
			 * Check the semaphores that were increased:
			 * - No complex ops, thus all sleeping ops are
			 *   decrease.
			 * - if we decreased the value, then any sleeping
			 *   semaphore ops wont be able to run: If the
			 *   previous value was too small, then the new
			 *   value will be too small, too.
			 */
			for (i = 0; i < nsops; i++) {
				if (sops[i].sem_op > 0) {
					otime |= update_queue(sma,
D
Davidlohr Bueso 已提交
1010
							      sops[i].sem_num, wake_q);
1011
				}
1012
			}
1013
		}
1014
	}
1015 1016
	if (otime)
		set_semotime(sma, sops);
1017 1018
}

1019
/*
1020
 * check_qop: Test if a queued operation sleeps on the semaphore semnum
1021 1022 1023 1024
 */
static int check_qop(struct sem_array *sma, int semnum, struct sem_queue *q,
			bool count_zero)
{
1025
	struct sembuf *sop = q->blocking;
1026

1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037
	/*
	 * Linux always (since 0.99.10) reported a task as sleeping on all
	 * semaphores. This violates SUS, therefore it was changed to the
	 * standard compliant behavior.
	 * Give the administrators a chance to notice that an application
	 * might misbehave because it relies on the Linux behavior.
	 */
	pr_info_once("semctl(GETNCNT/GETZCNT) is since 3.16 Single Unix Specification compliant.\n"
			"The task %s (%d) triggered the difference, watch for misbehavior.\n",
			current->comm, task_pid_nr(current));

1038 1039
	if (sop->sem_num != semnum)
		return 0;
1040

1041 1042 1043 1044 1045 1046
	if (count_zero && sop->sem_op == 0)
		return 1;
	if (!count_zero && sop->sem_op < 0)
		return 1;

	return 0;
1047 1048
}

L
Linus Torvalds 已提交
1049 1050 1051
/* 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
1052 1053 1054
 *
 * Per definition, a task waits only on the semaphore of the first semop
 * that cannot proceed, even if additional operation would block, too.
L
Linus Torvalds 已提交
1055
 */
1056 1057
static int count_semcnt(struct sem_array *sma, ushort semnum,
			bool count_zero)
L
Linus Torvalds 已提交
1058
{
1059
	struct list_head *l;
M
Manfred Spraul 已提交
1060
	struct sem_queue *q;
1061
	int semcnt;
L
Linus Torvalds 已提交
1062

1063 1064 1065
	semcnt = 0;
	/* First: check the simple operations. They are easy to evaluate */
	if (count_zero)
1066
		l = &sma->sems[semnum].pending_const;
1067
	else
1068
		l = &sma->sems[semnum].pending_alter;
L
Linus Torvalds 已提交
1069

1070 1071 1072 1073 1074
	list_for_each_entry(q, l, list) {
		/* all task on a per-semaphore list sleep on exactly
		 * that semaphore
		 */
		semcnt++;
R
Rik van Riel 已提交
1075 1076
	}

1077
	/* Then: check the complex operations. */
1078
	list_for_each_entry(q, &sma->pending_alter, list) {
1079 1080 1081 1082 1083 1084
		semcnt += check_qop(sma, semnum, q, count_zero);
	}
	if (count_zero) {
		list_for_each_entry(q, &sma->pending_const, list) {
			semcnt += check_qop(sma, semnum, q, count_zero);
		}
1085
	}
1086
	return semcnt;
L
Linus Torvalds 已提交
1087 1088
}

D
Davidlohr Bueso 已提交
1089 1090
/* Free a semaphore set. freeary() is called with sem_ids.rwsem locked
 * as a writer and the spinlock for this semaphore set hold. sem_ids.rwsem
N
Nadia Derbey 已提交
1091
 * remains locked on exit.
L
Linus Torvalds 已提交
1092
 */
1093
static void freeary(struct ipc_namespace *ns, struct kern_ipc_perm *ipcp)
L
Linus Torvalds 已提交
1094
{
1095 1096
	struct sem_undo *un, *tu;
	struct sem_queue *q, *tq;
1097
	struct sem_array *sma = container_of(ipcp, struct sem_array, sem_perm);
1098
	int i;
D
Davidlohr Bueso 已提交
1099
	DEFINE_WAKE_Q(wake_q);
L
Linus Torvalds 已提交
1100

1101
	/* Free the existing undo structures for this semaphore set.  */
1102
	ipc_assert_locked_object(&sma->sem_perm);
1103 1104 1105
	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
1106
		un->semid = -1;
1107 1108
		list_del_rcu(&un->list_proc);
		spin_unlock(&un->ulp->lock);
1109
		kfree_rcu(un, rcu);
1110
	}
L
Linus Torvalds 已提交
1111 1112

	/* Wake up all pending processes and let them fail with EIDRM. */
1113 1114
	list_for_each_entry_safe(q, tq, &sma->pending_const, list) {
		unlink_queue(sma, q);
D
Davidlohr Bueso 已提交
1115
		wake_up_sem_queue_prepare(q, -EIDRM, &wake_q);
1116 1117 1118
	}

	list_for_each_entry_safe(q, tq, &sma->pending_alter, list) {
1119
		unlink_queue(sma, q);
D
Davidlohr Bueso 已提交
1120
		wake_up_sem_queue_prepare(q, -EIDRM, &wake_q);
L
Linus Torvalds 已提交
1121
	}
1122
	for (i = 0; i < sma->sem_nsems; i++) {
1123
		struct sem *sem = &sma->sems[i];
1124 1125
		list_for_each_entry_safe(q, tq, &sem->pending_const, list) {
			unlink_queue(sma, q);
D
Davidlohr Bueso 已提交
1126
			wake_up_sem_queue_prepare(q, -EIDRM, &wake_q);
1127 1128
		}
		list_for_each_entry_safe(q, tq, &sem->pending_alter, list) {
1129
			unlink_queue(sma, q);
D
Davidlohr Bueso 已提交
1130
			wake_up_sem_queue_prepare(q, -EIDRM, &wake_q);
1131 1132
		}
	}
L
Linus Torvalds 已提交
1133

N
Nadia Derbey 已提交
1134 1135
	/* Remove the semaphore set from the IDR */
	sem_rmid(ns, sma);
1136
	sem_unlock(sma, -1);
1137
	rcu_read_unlock();
L
Linus Torvalds 已提交
1138

D
Davidlohr Bueso 已提交
1139
	wake_up_q(&wake_q);
K
Kirill Korotaev 已提交
1140
	ns->used_sems -= sma->sem_nsems;
1141
	ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
L
Linus Torvalds 已提交
1142 1143 1144 1145
}

static unsigned long copy_semid_to_user(void __user *buf, struct semid64_ds *in, int version)
{
M
Manfred Spraul 已提交
1146
	switch (version) {
L
Linus Torvalds 已提交
1147 1148 1149 1150 1151 1152
	case IPC_64:
		return copy_to_user(buf, in, sizeof(*in));
	case IPC_OLD:
	    {
		struct semid_ds out;

1153 1154
		memset(&out, 0, sizeof(out));

L
Linus Torvalds 已提交
1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167
		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;
	}
}

1168
static time64_t get_semotime(struct sem_array *sma)
1169 1170
{
	int i;
1171
	time64_t res;
1172

1173
	res = sma->sems[0].sem_otime;
1174
	for (i = 1; i < sma->sem_nsems; i++) {
1175
		time64_t to = sma->sems[i].sem_otime;
1176 1177 1178 1179 1180 1181 1182

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

1183 1184
static int semctl_stat(struct ipc_namespace *ns, int semid,
			 int cmd, struct semid64_ds *semid64)
L
Linus Torvalds 已提交
1185 1186
{
	struct sem_array *sma;
1187 1188
	int id = 0;
	int err;
L
Linus Torvalds 已提交
1189

1190
	memset(semid64, 0, sizeof(*semid64));
P
Paul McQuade 已提交
1191

1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204
	rcu_read_lock();
	if (cmd == SEM_STAT) {
		sma = sem_obtain_object(ns, semid);
		if (IS_ERR(sma)) {
			err = PTR_ERR(sma);
			goto out_unlock;
		}
		id = sma->sem_perm.id;
	} else {
		sma = sem_obtain_object_check(ns, semid);
		if (IS_ERR(sma)) {
			err = PTR_ERR(sma);
			goto out_unlock;
L
Linus Torvalds 已提交
1205 1206 1207
		}
	}

1208 1209 1210
	err = -EACCES;
	if (ipcperms(ns, &sma->sem_perm, S_IRUGO))
		goto out_unlock;
L
Linus Torvalds 已提交
1211

1212 1213 1214
	err = security_sem_semctl(sma, cmd);
	if (err)
		goto out_unlock;
L
Linus Torvalds 已提交
1215

1216 1217 1218 1219 1220 1221 1222 1223
	ipc_lock_object(&sma->sem_perm);

	if (!ipc_valid_object(&sma->sem_perm)) {
		ipc_unlock_object(&sma->sem_perm);
		err = -EIDRM;
		goto out_unlock;
	}

1224 1225 1226 1227
	kernel_to_ipc64_perm(&sma->sem_perm, &semid64->sem_perm);
	semid64->sem_otime = get_semotime(sma);
	semid64->sem_ctime = sma->sem_ctime;
	semid64->sem_nsems = sma->sem_nsems;
1228 1229

	ipc_unlock_object(&sma->sem_perm);
1230 1231
	rcu_read_unlock();
	return id;
L
Linus Torvalds 已提交
1232 1233

out_unlock:
1234
	rcu_read_unlock();
L
Linus Torvalds 已提交
1235 1236 1237
	return err;
}

1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272
static int semctl_info(struct ipc_namespace *ns, int semid,
			 int cmd, void __user *p)
{
	struct seminfo seminfo;
	int max_id;
	int err;

	err = security_sem_semctl(NULL, cmd);
	if (err)
		return err;

	memset(&seminfo, 0, sizeof(seminfo));
	seminfo.semmni = ns->sc_semmni;
	seminfo.semmns = ns->sc_semmns;
	seminfo.semmsl = ns->sc_semmsl;
	seminfo.semopm = ns->sc_semopm;
	seminfo.semvmx = SEMVMX;
	seminfo.semmnu = SEMMNU;
	seminfo.semmap = SEMMAP;
	seminfo.semume = SEMUME;
	down_read(&sem_ids(ns).rwsem);
	if (cmd == SEM_INFO) {
		seminfo.semusz = sem_ids(ns).in_use;
		seminfo.semaem = ns->used_sems;
	} else {
		seminfo.semusz = SEMUSZ;
		seminfo.semaem = SEMAEM;
	}
	max_id = ipc_get_maxid(&sem_ids(ns));
	up_read(&sem_ids(ns).rwsem);
	if (copy_to_user(p, &seminfo, sizeof(struct seminfo)))
		return -EFAULT;
	return (max_id < 0) ? 0 : max_id;
}

1273
static int semctl_setval(struct ipc_namespace *ns, int semid, int semnum,
1274
		int val)
1275 1276 1277
{
	struct sem_undo *un;
	struct sem_array *sma;
M
Manfred Spraul 已提交
1278
	struct sem *curr;
1279
	int err;
D
Davidlohr Bueso 已提交
1280 1281
	DEFINE_WAKE_Q(wake_q);

1282 1283
	if (val > SEMVMX || val < 0)
		return -ERANGE;
1284

1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301
	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;
	}
1302 1303

	err = security_sem_semctl(sma, SETVAL);
1304 1305 1306 1307
	if (err) {
		rcu_read_unlock();
		return -EACCES;
	}
1308

1309
	sem_lock(sma, NULL, -1);
1310

1311
	if (!ipc_valid_object(&sma->sem_perm)) {
1312 1313 1314 1315 1316
		sem_unlock(sma, -1);
		rcu_read_unlock();
		return -EIDRM;
	}

1317
	curr = &sma->sems[semnum];
1318

1319
	ipc_assert_locked_object(&sma->sem_perm);
1320 1321 1322 1323 1324
	list_for_each_entry(un, &sma->list_id, list_id)
		un->semadj[semnum] = 0;

	curr->semval = val;
	curr->sempid = task_tgid_vnr(current);
1325
	sma->sem_ctime = ktime_get_real_seconds();
1326
	/* maybe some queued-up processes were waiting for this */
D
Davidlohr Bueso 已提交
1327
	do_smart_update(sma, NULL, 0, 0, &wake_q);
1328
	sem_unlock(sma, -1);
1329
	rcu_read_unlock();
D
Davidlohr Bueso 已提交
1330
	wake_up_q(&wake_q);
1331
	return 0;
1332 1333
}

K
Kirill Korotaev 已提交
1334
static int semctl_main(struct ipc_namespace *ns, int semid, int semnum,
1335
		int cmd, void __user *p)
L
Linus Torvalds 已提交
1336 1337
{
	struct sem_array *sma;
M
Manfred Spraul 已提交
1338
	struct sem *curr;
1339
	int err, nsems;
L
Linus Torvalds 已提交
1340
	ushort fast_sem_io[SEMMSL_FAST];
M
Manfred Spraul 已提交
1341
	ushort *sem_io = fast_sem_io;
D
Davidlohr Bueso 已提交
1342
	DEFINE_WAKE_Q(wake_q);
1343 1344 1345 1346 1347

	rcu_read_lock();
	sma = sem_obtain_object_check(ns, semid);
	if (IS_ERR(sma)) {
		rcu_read_unlock();
1348
		return PTR_ERR(sma);
1349
	}
L
Linus Torvalds 已提交
1350 1351 1352 1353

	nsems = sma->sem_nsems;

	err = -EACCES;
1354 1355
	if (ipcperms(ns, &sma->sem_perm, cmd == SETALL ? S_IWUGO : S_IRUGO))
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1356 1357

	err = security_sem_semctl(sma, cmd);
1358 1359
	if (err)
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1360 1361 1362 1363 1364

	err = -EACCES;
	switch (cmd) {
	case GETALL:
	{
1365
		ushort __user *array = p;
L
Linus Torvalds 已提交
1366 1367
		int i;

1368
		sem_lock(sma, NULL, -1);
1369
		if (!ipc_valid_object(&sma->sem_perm)) {
1370 1371 1372
			err = -EIDRM;
			goto out_unlock;
		}
M
Manfred Spraul 已提交
1373
		if (nsems > SEMMSL_FAST) {
1374
			if (!ipc_rcu_getref(&sma->sem_perm)) {
1375
				err = -EIDRM;
1376
				goto out_unlock;
1377 1378
			}
			sem_unlock(sma, -1);
1379
			rcu_read_unlock();
K
Kees Cook 已提交
1380 1381
			sem_io = kvmalloc_array(nsems, sizeof(ushort),
						GFP_KERNEL);
M
Manfred Spraul 已提交
1382
			if (sem_io == NULL) {
1383
				ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
L
Linus Torvalds 已提交
1384 1385 1386
				return -ENOMEM;
			}

1387
			rcu_read_lock();
1388
			sem_lock_and_putref(sma);
1389
			if (!ipc_valid_object(&sma->sem_perm)) {
L
Linus Torvalds 已提交
1390
				err = -EIDRM;
1391
				goto out_unlock;
L
Linus Torvalds 已提交
1392
			}
1393
		}
L
Linus Torvalds 已提交
1394
		for (i = 0; i < sma->sem_nsems; i++)
1395
			sem_io[i] = sma->sems[i].semval;
1396
		sem_unlock(sma, -1);
1397
		rcu_read_unlock();
L
Linus Torvalds 已提交
1398
		err = 0;
M
Manfred Spraul 已提交
1399
		if (copy_to_user(array, sem_io, nsems*sizeof(ushort)))
L
Linus Torvalds 已提交
1400 1401 1402 1403 1404 1405 1406 1407
			err = -EFAULT;
		goto out_free;
	}
	case SETALL:
	{
		int i;
		struct sem_undo *un;

1408
		if (!ipc_rcu_getref(&sma->sem_perm)) {
1409 1410
			err = -EIDRM;
			goto out_rcu_wakeup;
1411
		}
1412
		rcu_read_unlock();
L
Linus Torvalds 已提交
1413

M
Manfred Spraul 已提交
1414
		if (nsems > SEMMSL_FAST) {
K
Kees Cook 已提交
1415 1416
			sem_io = kvmalloc_array(nsems, sizeof(ushort),
						GFP_KERNEL);
M
Manfred Spraul 已提交
1417
			if (sem_io == NULL) {
1418
				ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
L
Linus Torvalds 已提交
1419 1420 1421 1422
				return -ENOMEM;
			}
		}

M
Manfred Spraul 已提交
1423
		if (copy_from_user(sem_io, p, nsems*sizeof(ushort))) {
1424
			ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
L
Linus Torvalds 已提交
1425 1426 1427 1428 1429 1430
			err = -EFAULT;
			goto out_free;
		}

		for (i = 0; i < nsems; i++) {
			if (sem_io[i] > SEMVMX) {
1431
				ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
L
Linus Torvalds 已提交
1432 1433 1434 1435
				err = -ERANGE;
				goto out_free;
			}
		}
1436
		rcu_read_lock();
1437
		sem_lock_and_putref(sma);
1438
		if (!ipc_valid_object(&sma->sem_perm)) {
L
Linus Torvalds 已提交
1439
			err = -EIDRM;
1440
			goto out_unlock;
L
Linus Torvalds 已提交
1441 1442
		}

1443
		for (i = 0; i < nsems; i++) {
1444 1445
			sma->sems[i].semval = sem_io[i];
			sma->sems[i].sempid = task_tgid_vnr(current);
1446
		}
1447

1448
		ipc_assert_locked_object(&sma->sem_perm);
1449
		list_for_each_entry(un, &sma->list_id, list_id) {
L
Linus Torvalds 已提交
1450 1451
			for (i = 0; i < nsems; i++)
				un->semadj[i] = 0;
1452
		}
1453
		sma->sem_ctime = ktime_get_real_seconds();
L
Linus Torvalds 已提交
1454
		/* maybe some queued-up processes were waiting for this */
D
Davidlohr Bueso 已提交
1455
		do_smart_update(sma, NULL, 0, 0, &wake_q);
L
Linus Torvalds 已提交
1456 1457 1458
		err = 0;
		goto out_unlock;
	}
1459
	/* GETVAL, GETPID, GETNCTN, GETZCNT: fall-through */
L
Linus Torvalds 已提交
1460 1461
	}
	err = -EINVAL;
1462 1463
	if (semnum < 0 || semnum >= nsems)
		goto out_rcu_wakeup;
L
Linus Torvalds 已提交
1464

1465
	sem_lock(sma, NULL, -1);
1466
	if (!ipc_valid_object(&sma->sem_perm)) {
1467 1468 1469
		err = -EIDRM;
		goto out_unlock;
	}
1470
	curr = &sma->sems[semnum];
L
Linus Torvalds 已提交
1471 1472 1473 1474 1475 1476 1477 1478 1479

	switch (cmd) {
	case GETVAL:
		err = curr->semval;
		goto out_unlock;
	case GETPID:
		err = curr->sempid;
		goto out_unlock;
	case GETNCNT:
1480
		err = count_semcnt(sma, semnum, 0);
L
Linus Torvalds 已提交
1481 1482
		goto out_unlock;
	case GETZCNT:
1483
		err = count_semcnt(sma, semnum, 1);
L
Linus Torvalds 已提交
1484 1485
		goto out_unlock;
	}
1486

L
Linus Torvalds 已提交
1487
out_unlock:
1488
	sem_unlock(sma, -1);
1489
out_rcu_wakeup:
1490
	rcu_read_unlock();
D
Davidlohr Bueso 已提交
1491
	wake_up_q(&wake_q);
L
Linus Torvalds 已提交
1492
out_free:
M
Manfred Spraul 已提交
1493
	if (sem_io != fast_sem_io)
K
Kees Cook 已提交
1494
		kvfree(sem_io);
L
Linus Torvalds 已提交
1495 1496 1497
	return err;
}

1498 1499
static inline unsigned long
copy_semid_from_user(struct semid64_ds *out, void __user *buf, int version)
L
Linus Torvalds 已提交
1500
{
M
Manfred Spraul 已提交
1501
	switch (version) {
L
Linus Torvalds 已提交
1502
	case IPC_64:
1503
		if (copy_from_user(out, buf, sizeof(*out)))
L
Linus Torvalds 已提交
1504 1505 1506 1507 1508 1509
			return -EFAULT;
		return 0;
	case IPC_OLD:
	    {
		struct semid_ds tbuf_old;

M
Manfred Spraul 已提交
1510
		if (copy_from_user(&tbuf_old, buf, sizeof(tbuf_old)))
L
Linus Torvalds 已提交
1511 1512
			return -EFAULT;

1513 1514 1515
		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
1516 1517 1518 1519 1520 1521 1522 1523

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

1524
/*
D
Davidlohr Bueso 已提交
1525
 * This function handles some semctl commands which require the rwsem
1526
 * to be held in write mode.
D
Davidlohr Bueso 已提交
1527
 * NOTE: no locks must be held, the rwsem is taken inside this function.
1528
 */
1529
static int semctl_down(struct ipc_namespace *ns, int semid,
1530
		       int cmd, struct semid64_ds *semid64)
L
Linus Torvalds 已提交
1531 1532 1533 1534 1535
{
	struct sem_array *sma;
	int err;
	struct kern_ipc_perm *ipcp;

D
Davidlohr Bueso 已提交
1536
	down_write(&sem_ids(ns).rwsem);
1537 1538
	rcu_read_lock();

1539
	ipcp = ipcctl_pre_down_nolock(ns, &sem_ids(ns), semid, cmd,
1540
				      &semid64->sem_perm, 0);
1541 1542 1543 1544
	if (IS_ERR(ipcp)) {
		err = PTR_ERR(ipcp);
		goto out_unlock1;
	}
S
Steve Grubb 已提交
1545

1546
	sma = container_of(ipcp, struct sem_array, sem_perm);
L
Linus Torvalds 已提交
1547 1548

	err = security_sem_semctl(sma, cmd);
1549 1550
	if (err)
		goto out_unlock1;
L
Linus Torvalds 已提交
1551

1552
	switch (cmd) {
L
Linus Torvalds 已提交
1553
	case IPC_RMID:
1554
		sem_lock(sma, NULL, -1);
1555
		/* freeary unlocks the ipc object and rcu */
1556
		freeary(ns, ipcp);
1557
		goto out_up;
L
Linus Torvalds 已提交
1558
	case IPC_SET:
1559
		sem_lock(sma, NULL, -1);
1560
		err = ipc_update_perm(&semid64->sem_perm, ipcp);
1561
		if (err)
1562
			goto out_unlock0;
1563
		sma->sem_ctime = ktime_get_real_seconds();
L
Linus Torvalds 已提交
1564 1565 1566
		break;
	default:
		err = -EINVAL;
1567
		goto out_unlock1;
L
Linus Torvalds 已提交
1568 1569
	}

1570
out_unlock0:
1571
	sem_unlock(sma, -1);
1572
out_unlock1:
1573
	rcu_read_unlock();
1574
out_up:
D
Davidlohr Bueso 已提交
1575
	up_write(&sem_ids(ns).rwsem);
L
Linus Torvalds 已提交
1576 1577 1578
	return err;
}

1579
SYSCALL_DEFINE4(semctl, int, semid, int, semnum, int, cmd, unsigned long, arg)
L
Linus Torvalds 已提交
1580 1581
{
	int version;
K
Kirill Korotaev 已提交
1582
	struct ipc_namespace *ns;
1583
	void __user *p = (void __user *)arg;
1584 1585
	struct semid64_ds semid64;
	int err;
L
Linus Torvalds 已提交
1586 1587 1588 1589 1590

	if (semid < 0)
		return -EINVAL;

	version = ipc_parse_version(&cmd);
K
Kirill Korotaev 已提交
1591
	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
1592

M
Manfred Spraul 已提交
1593
	switch (cmd) {
L
Linus Torvalds 已提交
1594 1595
	case IPC_INFO:
	case SEM_INFO:
1596
		return semctl_info(ns, semid, cmd, p);
1597
	case IPC_STAT:
L
Linus Torvalds 已提交
1598
	case SEM_STAT:
1599 1600 1601 1602 1603 1604
		err = semctl_stat(ns, semid, cmd, &semid64);
		if (err < 0)
			return err;
		if (copy_semid_to_user(p, &semid64, version))
			err = -EFAULT;
		return err;
L
Linus Torvalds 已提交
1605 1606 1607 1608 1609 1610
	case GETALL:
	case GETVAL:
	case GETPID:
	case GETNCNT:
	case GETZCNT:
	case SETALL:
1611
		return semctl_main(ns, semid, semnum, cmd, p);
1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622
	case SETVAL: {
		int val;
#if defined(CONFIG_64BIT) && defined(__BIG_ENDIAN)
		/* big-endian 64bit */
		val = arg >> 32;
#else
		/* 32bit or little-endian 64bit */
		val = arg;
#endif
		return semctl_setval(ns, semid, semnum, val);
	}
L
Linus Torvalds 已提交
1623
	case IPC_SET:
1624 1625 1626 1627
		if (copy_semid_from_user(&semid64, p, version))
			return -EFAULT;
	case IPC_RMID:
		return semctl_down(ns, semid, cmd, &semid64);
L
Linus Torvalds 已提交
1628 1629 1630 1631 1632
	default:
		return -EINVAL;
	}
}

A
Al Viro 已提交
1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650
#ifdef CONFIG_COMPAT

struct compat_semid_ds {
	struct compat_ipc_perm sem_perm;
	compat_time_t sem_otime;
	compat_time_t sem_ctime;
	compat_uptr_t sem_base;
	compat_uptr_t sem_pending;
	compat_uptr_t sem_pending_last;
	compat_uptr_t undo;
	unsigned short sem_nsems;
};

static int copy_compat_semid_from_user(struct semid64_ds *out, void __user *buf,
					int version)
{
	memset(out, 0, sizeof(*out));
	if (version == IPC_64) {
1651
		struct compat_semid64_ds __user *p = buf;
A
Al Viro 已提交
1652 1653
		return get_compat_ipc64_perm(&out->sem_perm, &p->sem_perm);
	} else {
1654
		struct compat_semid_ds __user *p = buf;
A
Al Viro 已提交
1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710
		return get_compat_ipc_perm(&out->sem_perm, &p->sem_perm);
	}
}

static int copy_compat_semid_to_user(void __user *buf, struct semid64_ds *in,
					int version)
{
	if (version == IPC_64) {
		struct compat_semid64_ds v;
		memset(&v, 0, sizeof(v));
		to_compat_ipc64_perm(&v.sem_perm, &in->sem_perm);
		v.sem_otime = in->sem_otime;
		v.sem_ctime = in->sem_ctime;
		v.sem_nsems = in->sem_nsems;
		return copy_to_user(buf, &v, sizeof(v));
	} else {
		struct compat_semid_ds v;
		memset(&v, 0, sizeof(v));
		to_compat_ipc_perm(&v.sem_perm, &in->sem_perm);
		v.sem_otime = in->sem_otime;
		v.sem_ctime = in->sem_ctime;
		v.sem_nsems = in->sem_nsems;
		return copy_to_user(buf, &v, sizeof(v));
	}
}

COMPAT_SYSCALL_DEFINE4(semctl, int, semid, int, semnum, int, cmd, int, arg)
{
	void __user *p = compat_ptr(arg);
	struct ipc_namespace *ns;
	struct semid64_ds semid64;
	int version = compat_ipc_parse_version(&cmd);
	int err;

	ns = current->nsproxy->ipc_ns;

	if (semid < 0)
		return -EINVAL;

	switch (cmd & (~IPC_64)) {
	case IPC_INFO:
	case SEM_INFO:
		return semctl_info(ns, semid, cmd, p);
	case IPC_STAT:
	case SEM_STAT:
		err = semctl_stat(ns, semid, cmd, &semid64);
		if (err < 0)
			return err;
		if (copy_compat_semid_to_user(p, &semid64, version))
			err = -EFAULT;
		return err;
	case GETVAL:
	case GETPID:
	case GETNCNT:
	case GETZCNT:
	case GETALL:
L
Linus Torvalds 已提交
1711
	case SETALL:
1712 1713 1714
		return semctl_main(ns, semid, semnum, cmd, p);
	case SETVAL:
		return semctl_setval(ns, semid, semnum, arg);
L
Linus Torvalds 已提交
1715
	case IPC_SET:
A
Al Viro 已提交
1716 1717 1718 1719 1720
		if (copy_compat_semid_from_user(&semid64, p, version))
			return -EFAULT;
		/* fallthru */
	case IPC_RMID:
		return semctl_down(ns, semid, cmd, &semid64);
L
Linus Torvalds 已提交
1721 1722 1723 1724
	default:
		return -EINVAL;
	}
}
A
Al Viro 已提交
1725
#endif
L
Linus Torvalds 已提交
1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743

/* 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) {
1744
		undo_list = kzalloc(sizeof(*undo_list), GFP_KERNEL);
L
Linus Torvalds 已提交
1745 1746
		if (undo_list == NULL)
			return -ENOMEM;
I
Ingo Molnar 已提交
1747
		spin_lock_init(&undo_list->lock);
1748
		refcount_set(&undo_list->refcnt, 1);
1749 1750
		INIT_LIST_HEAD(&undo_list->list_proc);

L
Linus Torvalds 已提交
1751 1752 1753 1754 1755 1756
		current->sysvsem.undo_list = undo_list;
	}
	*undo_listp = undo_list;
	return 0;
}

1757
static struct sem_undo *__lookup_undo(struct sem_undo_list *ulp, int semid)
L
Linus Torvalds 已提交
1758
{
1759
	struct sem_undo *un;
1760

1761 1762 1763
	list_for_each_entry_rcu(un, &ulp->list_proc, list_proc) {
		if (un->semid == semid)
			return un;
L
Linus Torvalds 已提交
1764
	}
1765
	return NULL;
L
Linus Torvalds 已提交
1766 1767
}

1768 1769 1770 1771
static struct sem_undo *lookup_undo(struct sem_undo_list *ulp, int semid)
{
	struct sem_undo *un;

M
Manfred Spraul 已提交
1772
	assert_spin_locked(&ulp->lock);
1773 1774 1775 1776 1777 1778 1779 1780 1781

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

1782
/**
D
Davidlohr Bueso 已提交
1783
 * find_alloc_undo - lookup (and if not present create) undo array
1784 1785 1786 1787 1788 1789
 * @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.
1790 1791
 * Lifetime-rules: sem_undo is rcu-protected, on success, the function
 * performs a rcu_read_lock().
1792 1793
 */
static struct sem_undo *find_alloc_undo(struct ipc_namespace *ns, int semid)
L
Linus Torvalds 已提交
1794 1795 1796 1797
{
	struct sem_array *sma;
	struct sem_undo_list *ulp;
	struct sem_undo *un, *new;
1798
	int nsems, error;
L
Linus Torvalds 已提交
1799 1800 1801 1802 1803

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

1804
	rcu_read_lock();
1805
	spin_lock(&ulp->lock);
L
Linus Torvalds 已提交
1806
	un = lookup_undo(ulp, semid);
1807
	spin_unlock(&ulp->lock);
M
Manfred Spraul 已提交
1808
	if (likely(un != NULL))
L
Linus Torvalds 已提交
1809 1810 1811
		goto out;

	/* no undo structure around - allocate one. */
1812
	/* step 1: figure out the size of the semaphore array */
1813 1814 1815
	sma = sem_obtain_object_check(ns, semid);
	if (IS_ERR(sma)) {
		rcu_read_unlock();
J
Julia Lawall 已提交
1816
		return ERR_CAST(sma);
1817
	}
1818

L
Linus Torvalds 已提交
1819
	nsems = sma->sem_nsems;
1820
	if (!ipc_rcu_getref(&sma->sem_perm)) {
1821 1822 1823 1824
		rcu_read_unlock();
		un = ERR_PTR(-EIDRM);
		goto out;
	}
1825
	rcu_read_unlock();
L
Linus Torvalds 已提交
1826

1827
	/* step 2: allocate new undo structure */
1828
	new = kzalloc(sizeof(struct sem_undo) + sizeof(short)*nsems, GFP_KERNEL);
L
Linus Torvalds 已提交
1829
	if (!new) {
1830
		ipc_rcu_putref(&sma->sem_perm, sem_rcu_free);
L
Linus Torvalds 已提交
1831 1832 1833
		return ERR_PTR(-ENOMEM);
	}

1834
	/* step 3: Acquire the lock on semaphore array */
1835
	rcu_read_lock();
1836
	sem_lock_and_putref(sma);
1837
	if (!ipc_valid_object(&sma->sem_perm)) {
1838
		sem_unlock(sma, -1);
1839
		rcu_read_unlock();
L
Linus Torvalds 已提交
1840 1841 1842 1843
		kfree(new);
		un = ERR_PTR(-EIDRM);
		goto out;
	}
1844 1845 1846 1847 1848 1849 1850 1851 1852 1853
	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;
	}
1854 1855
	/* step 5: initialize & link new undo structure */
	new->semadj = (short *) &new[1];
1856
	new->ulp = ulp;
1857 1858
	new->semid = semid;
	assert_spin_locked(&ulp->lock);
1859
	list_add_rcu(&new->list_proc, &ulp->list_proc);
1860
	ipc_assert_locked_object(&sma->sem_perm);
1861
	list_add(&new->list_id, &sma->list_id);
1862
	un = new;
1863

1864
success:
1865
	spin_unlock(&ulp->lock);
1866
	sem_unlock(sma, -1);
L
Linus Torvalds 已提交
1867 1868 1869 1870
out:
	return un;
}

A
Al Viro 已提交
1871
static long do_semtimedop(int semid, struct sembuf __user *tsops,
1872
		unsigned nsops, const struct timespec64 *timeout)
L
Linus Torvalds 已提交
1873 1874 1875 1876
{
	int error = -EINVAL;
	struct sem_array *sma;
	struct sembuf fast_sops[SEMOPM_FAST];
M
Manfred Spraul 已提交
1877
	struct sembuf *sops = fast_sops, *sop;
L
Linus Torvalds 已提交
1878
	struct sem_undo *un;
1879 1880
	int max, locknum;
	bool undos = false, alter = false, dupsop = false;
L
Linus Torvalds 已提交
1881
	struct sem_queue queue;
1882
	unsigned long dup = 0, jiffies_left = 0;
K
Kirill Korotaev 已提交
1883 1884 1885
	struct ipc_namespace *ns;

	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
1886 1887 1888

	if (nsops < 1 || semid < 0)
		return -EINVAL;
K
Kirill Korotaev 已提交
1889
	if (nsops > ns->sc_semopm)
L
Linus Torvalds 已提交
1890
		return -E2BIG;
M
Manfred Spraul 已提交
1891
	if (nsops > SEMOPM_FAST) {
1892
		sops = kvmalloc(sizeof(*sops)*nsops, GFP_KERNEL);
M
Manfred Spraul 已提交
1893
		if (sops == NULL)
L
Linus Torvalds 已提交
1894 1895
			return -ENOMEM;
	}
1896

M
Manfred Spraul 已提交
1897 1898
	if (copy_from_user(sops, tsops, nsops * sizeof(*tsops))) {
		error =  -EFAULT;
L
Linus Torvalds 已提交
1899 1900
		goto out_free;
	}
1901

L
Linus Torvalds 已提交
1902
	if (timeout) {
A
Al Viro 已提交
1903 1904
		if (timeout->tv_sec < 0 || timeout->tv_nsec < 0 ||
			timeout->tv_nsec >= 1000000000L) {
L
Linus Torvalds 已提交
1905 1906 1907
			error = -EINVAL;
			goto out_free;
		}
1908
		jiffies_left = timespec64_to_jiffies(timeout);
L
Linus Torvalds 已提交
1909
	}
1910

L
Linus Torvalds 已提交
1911 1912
	max = 0;
	for (sop = sops; sop < sops + nsops; sop++) {
1913 1914
		unsigned long mask = 1ULL << ((sop->sem_num) % BITS_PER_LONG);

L
Linus Torvalds 已提交
1915 1916 1917
		if (sop->sem_num >= max)
			max = sop->sem_num;
		if (sop->sem_flg & SEM_UNDO)
1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931
			undos = true;
		if (dup & mask) {
			/*
			 * There was a previous alter access that appears
			 * to have accessed the same semaphore, thus use
			 * the dupsop logic. "appears", because the detection
			 * can only check % BITS_PER_LONG.
			 */
			dupsop = true;
		}
		if (sop->sem_op != 0) {
			alter = true;
			dup |= mask;
		}
L
Linus Torvalds 已提交
1932 1933 1934
	}

	if (undos) {
1935
		/* On success, find_alloc_undo takes the rcu_read_lock */
1936
		un = find_alloc_undo(ns, semid);
L
Linus Torvalds 已提交
1937 1938 1939 1940
		if (IS_ERR(un)) {
			error = PTR_ERR(un);
			goto out_free;
		}
1941
	} else {
L
Linus Torvalds 已提交
1942
		un = NULL;
1943 1944
		rcu_read_lock();
	}
L
Linus Torvalds 已提交
1945

1946
	sma = sem_obtain_object_check(ns, semid);
1947
	if (IS_ERR(sma)) {
1948
		rcu_read_unlock();
1949
		error = PTR_ERR(sma);
L
Linus Torvalds 已提交
1950
		goto out_free;
1951 1952
	}

1953
	error = -EFBIG;
1954 1955 1956 1957
	if (max >= sma->sem_nsems) {
		rcu_read_unlock();
		goto out_free;
	}
1958 1959

	error = -EACCES;
1960 1961 1962 1963
	if (ipcperms(ns, &sma->sem_perm, alter ? S_IWUGO : S_IRUGO)) {
		rcu_read_unlock();
		goto out_free;
	}
1964 1965

	error = security_sem_semop(sma, sops, nsops, alter);
1966 1967 1968 1969
	if (error) {
		rcu_read_unlock();
		goto out_free;
	}
1970

1971 1972
	error = -EIDRM;
	locknum = sem_lock(sma, sops, nsops);
1973 1974 1975 1976 1977 1978 1979 1980 1981
	/*
	 * We eventually might perform the following check in a lockless
	 * fashion, considering ipc_valid_object() locking constraints.
	 * If nsops == 1 and there is no contention for sem_perm.lock, then
	 * only a per-semaphore lock is held and it's OK to proceed with the
	 * check below. More details on the fine grained locking scheme
	 * entangled here and why it's RMID race safe on comments at sem_lock()
	 */
	if (!ipc_valid_object(&sma->sem_perm))
1982
		goto out_unlock_free;
L
Linus Torvalds 已提交
1983
	/*
1984
	 * semid identifiers are not unique - find_alloc_undo may have
L
Linus Torvalds 已提交
1985
	 * allocated an undo structure, it was invalidated by an RMID
1986
	 * and now a new array with received the same id. Check and fail.
L
Lucas De Marchi 已提交
1987
	 * This case can be detected checking un->semid. The existence of
1988
	 * "un" itself is guaranteed by rcu.
L
Linus Torvalds 已提交
1989
	 */
1990 1991
	if (un && un->semid == -1)
		goto out_unlock_free;
1992

1993 1994 1995 1996 1997
	queue.sops = sops;
	queue.nsops = nsops;
	queue.undo = un;
	queue.pid = task_tgid_vnr(current);
	queue.alter = alter;
1998
	queue.dupsop = dupsop;
1999 2000

	error = perform_atomic_semop(sma, &queue);
D
Davidlohr Bueso 已提交
2001 2002 2003 2004 2005
	if (error == 0) { /* non-blocking succesfull path */
		DEFINE_WAKE_Q(wake_q);

		/*
		 * If the operation was successful, then do
2006 2007 2008
		 * the required updates.
		 */
		if (alter)
D
Davidlohr Bueso 已提交
2009
			do_smart_update(sma, sops, nsops, 1, &wake_q);
2010 2011
		else
			set_semotime(sma, sops);
D
Davidlohr Bueso 已提交
2012 2013 2014 2015 2016 2017

		sem_unlock(sma, locknum);
		rcu_read_unlock();
		wake_up_q(&wake_q);

		goto out_free;
L
Linus Torvalds 已提交
2018
	}
D
Davidlohr Bueso 已提交
2019
	if (error < 0) /* non-blocking error path */
2020
		goto out_unlock_free;
L
Linus Torvalds 已提交
2021

D
Davidlohr Bueso 已提交
2022 2023
	/*
	 * We need to sleep on this operation, so we put the current
L
Linus Torvalds 已提交
2024 2025
	 * task into the pending queue and go to sleep.
	 */
2026 2027
	if (nsops == 1) {
		struct sem *curr;
2028
		curr = &sma->sems[sops->sem_num];
2029

2030 2031 2032 2033 2034 2035 2036 2037 2038 2039
		if (alter) {
			if (sma->complex_count) {
				list_add_tail(&queue.list,
						&sma->pending_alter);
			} else {

				list_add_tail(&queue.list,
						&curr->pending_alter);
			}
		} else {
2040
			list_add_tail(&queue.list, &curr->pending_const);
2041
		}
2042
	} else {
2043 2044 2045
		if (!sma->complex_count)
			merge_queues(sma);

2046
		if (alter)
2047
			list_add_tail(&queue.list, &sma->pending_alter);
2048
		else
2049 2050
			list_add_tail(&queue.list, &sma->pending_const);

2051 2052 2053
		sma->complex_count++;
	}

D
Davidlohr Bueso 已提交
2054 2055 2056
	do {
		queue.status = -EINTR;
		queue.sleeper = current;
2057

D
Davidlohr Bueso 已提交
2058 2059 2060
		__set_current_state(TASK_INTERRUPTIBLE);
		sem_unlock(sma, locknum);
		rcu_read_unlock();
L
Linus Torvalds 已提交
2061

D
Davidlohr Bueso 已提交
2062 2063 2064 2065
		if (timeout)
			jiffies_left = schedule_timeout(jiffies_left);
		else
			schedule();
L
Linus Torvalds 已提交
2066

D
Davidlohr Bueso 已提交
2067
		/*
D
Davidlohr Bueso 已提交
2068 2069 2070 2071 2072 2073 2074 2075 2076
		 * fastpath: the semop has completed, either successfully or
		 * not, from the syscall pov, is quite irrelevant to us at this
		 * point; we're done.
		 *
		 * We _do_ care, nonetheless, about being awoken by a signal or
		 * spuriously.  The queue.status is checked again in the
		 * slowpath (aka after taking sem_lock), such that we can detect
		 * scenarios where we were awakened externally, during the
		 * window between wake_q_add() and wake_up_q().
2077
		 */
D
Davidlohr Bueso 已提交
2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088
		error = READ_ONCE(queue.status);
		if (error != -EINTR) {
			/*
			 * User space could assume that semop() is a memory
			 * barrier: Without the mb(), the cpu could
			 * speculatively read in userspace stale data that was
			 * overwritten by the previous owner of the semaphore.
			 */
			smp_mb();
			goto out_free;
		}
2089

D
Davidlohr Bueso 已提交
2090
		rcu_read_lock();
2091
		locknum = sem_lock(sma, sops, nsops);
L
Linus Torvalds 已提交
2092

2093 2094 2095 2096
		if (!ipc_valid_object(&sma->sem_perm))
			goto out_unlock_free;

		error = READ_ONCE(queue.status);
L
Linus Torvalds 已提交
2097

D
Davidlohr Bueso 已提交
2098 2099 2100 2101 2102 2103
		/*
		 * If queue.status != -EINTR we are woken up by another process.
		 * Leave without unlink_queue(), but with sem_unlock().
		 */
		if (error != -EINTR)
			goto out_unlock_free;
2104

D
Davidlohr Bueso 已提交
2105 2106 2107 2108 2109 2110
		/*
		 * If an interrupt occurred we have to clean up the queue.
		 */
		if (timeout && jiffies_left == 0)
			error = -EAGAIN;
	} while (error == -EINTR && !signal_pending(current)); /* spurious */
2111

2112
	unlink_queue(sma, &queue);
L
Linus Torvalds 已提交
2113 2114

out_unlock_free:
2115
	sem_unlock(sma, locknum);
2116
	rcu_read_unlock();
L
Linus Torvalds 已提交
2117
out_free:
M
Manfred Spraul 已提交
2118
	if (sops != fast_sops)
2119
		kvfree(sops);
L
Linus Torvalds 已提交
2120 2121 2122
	return error;
}

A
Al Viro 已提交
2123 2124 2125 2126
SYSCALL_DEFINE4(semtimedop, int, semid, struct sembuf __user *, tsops,
		unsigned, nsops, const struct timespec __user *, timeout)
{
	if (timeout) {
2127 2128
		struct timespec64 ts;
		if (get_timespec64(&ts, timeout))
A
Al Viro 已提交
2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140
			return -EFAULT;
		return do_semtimedop(semid, tsops, nsops, &ts);
	}
	return do_semtimedop(semid, tsops, nsops, NULL);
}

#ifdef CONFIG_COMPAT
COMPAT_SYSCALL_DEFINE4(semtimedop, int, semid, struct sembuf __user *, tsems,
		       unsigned, nsops,
		       const struct compat_timespec __user *, timeout)
{
	if (timeout) {
2141 2142
		struct timespec64 ts;
		if (compat_get_timespec64(&ts, timeout))
A
Al Viro 已提交
2143 2144 2145 2146 2147 2148 2149
			return -EFAULT;
		return do_semtimedop(semid, tsems, nsops, &ts);
	}
	return do_semtimedop(semid, tsems, nsops, NULL);
}
#endif

2150 2151
SYSCALL_DEFINE3(semop, int, semid, struct sembuf __user *, tsops,
		unsigned, nsops)
L
Linus Torvalds 已提交
2152
{
A
Al Viro 已提交
2153
	return do_semtimedop(semid, tsops, nsops, NULL);
L
Linus Torvalds 已提交
2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168
}

/* 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;
2169
		refcount_inc(&undo_list->refcnt);
L
Linus Torvalds 已提交
2170
		tsk->sysvsem.undo_list = undo_list;
P
Paul McQuade 已提交
2171
	} else
L
Linus Torvalds 已提交
2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190
		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)
{
2191
	struct sem_undo_list *ulp;
L
Linus Torvalds 已提交
2192

2193 2194
	ulp = tsk->sysvsem.undo_list;
	if (!ulp)
L
Linus Torvalds 已提交
2195
		return;
2196
	tsk->sysvsem.undo_list = NULL;
L
Linus Torvalds 已提交
2197

2198
	if (!refcount_dec_and_test(&ulp->refcnt))
L
Linus Torvalds 已提交
2199 2200
		return;

2201
	for (;;) {
L
Linus Torvalds 已提交
2202
		struct sem_array *sma;
2203
		struct sem_undo *un;
2204
		int semid, i;
D
Davidlohr Bueso 已提交
2205
		DEFINE_WAKE_Q(wake_q);
2206

2207 2208
		cond_resched();

2209
		rcu_read_lock();
2210 2211
		un = list_entry_rcu(ulp->list_proc.next,
				    struct sem_undo, list_proc);
2212 2213 2214 2215 2216 2217 2218
		if (&un->list_proc == &ulp->list_proc) {
			/*
			 * We must wait for freeary() before freeing this ulp,
			 * in case we raced with last sem_undo. There is a small
			 * possibility where we exit while freeary() didn't
			 * finish unlocking sem_undo_list.
			 */
2219 2220
			spin_lock(&ulp->lock);
			spin_unlock(&ulp->lock);
2221 2222 2223 2224 2225 2226
			rcu_read_unlock();
			break;
		}
		spin_lock(&ulp->lock);
		semid = un->semid;
		spin_unlock(&ulp->lock);
2227

2228
		/* exit_sem raced with IPC_RMID, nothing to do */
2229 2230
		if (semid == -1) {
			rcu_read_unlock();
2231
			continue;
2232
		}
L
Linus Torvalds 已提交
2233

2234
		sma = sem_obtain_object_check(tsk->nsproxy->ipc_ns, semid);
2235
		/* exit_sem raced with IPC_RMID, nothing to do */
2236 2237
		if (IS_ERR(sma)) {
			rcu_read_unlock();
2238
			continue;
2239
		}
L
Linus Torvalds 已提交
2240

2241
		sem_lock(sma, NULL, -1);
2242
		/* exit_sem raced with IPC_RMID, nothing to do */
2243
		if (!ipc_valid_object(&sma->sem_perm)) {
2244 2245 2246 2247
			sem_unlock(sma, -1);
			rcu_read_unlock();
			continue;
		}
2248
		un = __lookup_undo(ulp, semid);
2249 2250 2251 2252
		if (un == NULL) {
			/* exit_sem raced with IPC_RMID+semget() that created
			 * exactly the same semid. Nothing to do.
			 */
2253
			sem_unlock(sma, -1);
2254
			rcu_read_unlock();
2255 2256 2257 2258
			continue;
		}

		/* remove un from the linked lists */
2259
		ipc_assert_locked_object(&sma->sem_perm);
2260 2261
		list_del(&un->list_id);

2262 2263 2264 2265
		/* we are the last process using this ulp, acquiring ulp->lock
		 * isn't required. Besides that, we are also protected against
		 * IPC_RMID as we hold sma->sem_perm lock now
		 */
2266 2267
		list_del_rcu(&un->list_proc);

2268 2269
		/* perform adjustments registered in un */
		for (i = 0; i < sma->sem_nsems; i++) {
2270
			struct sem *semaphore = &sma->sems[i];
2271 2272
			if (un->semadj[i]) {
				semaphore->semval += un->semadj[i];
L
Linus Torvalds 已提交
2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283
				/*
				 * 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.
				 *
M
Manfred Spraul 已提交
2284
				 *	Manfred <manfred@colorfullife.com>
L
Linus Torvalds 已提交
2285
				 */
I
Ingo Molnar 已提交
2286 2287 2288 2289
				if (semaphore->semval < 0)
					semaphore->semval = 0;
				if (semaphore->semval > SEMVMX)
					semaphore->semval = SEMVMX;
2290
				semaphore->sempid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
2291 2292 2293
			}
		}
		/* maybe some queued-up processes were waiting for this */
D
Davidlohr Bueso 已提交
2294
		do_smart_update(sma, NULL, 0, 1, &wake_q);
2295
		sem_unlock(sma, -1);
2296
		rcu_read_unlock();
D
Davidlohr Bueso 已提交
2297
		wake_up_q(&wake_q);
2298

2299
		kfree_rcu(un, rcu);
L
Linus Torvalds 已提交
2300
	}
2301
	kfree(ulp);
L
Linus Torvalds 已提交
2302 2303 2304
}

#ifdef CONFIG_PROC_FS
2305
static int sysvipc_sem_proc_show(struct seq_file *s, void *it)
L
Linus Torvalds 已提交
2306
{
2307
	struct user_namespace *user_ns = seq_user_ns(s);
2308 2309
	struct kern_ipc_perm *ipcp = it;
	struct sem_array *sma = container_of(ipcp, struct sem_array, sem_perm);
2310
	time64_t sem_otime;
2311

2312 2313 2314
	/*
	 * The proc interface isn't aware of sem_lock(), it calls
	 * ipc_lock_object() directly (in sysvipc_find_ipc).
2315 2316
	 * In order to stay compatible with sem_lock(), we must
	 * enter / leave complex_mode.
2317
	 */
2318
	complexmode_enter(sma);
2319

2320
	sem_otime = get_semotime(sma);
2321

2322
	seq_printf(s,
2323
		   "%10d %10d  %4o %10u %5u %5u %5u %5u %10llu %10llu\n",
2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334
		   sma->sem_perm.key,
		   sma->sem_perm.id,
		   sma->sem_perm.mode,
		   sma->sem_nsems,
		   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),
		   sem_otime,
		   sma->sem_ctime);

2335 2336
	complexmode_tryleave(sma);

2337
	return 0;
L
Linus Torvalds 已提交
2338 2339
}
#endif