shm.c 29.0 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
/*
 * linux/ipc/shm.c
 * Copyright (C) 1992, 1993 Krishna Balasubramanian
 *	 Many improvements/fixes by Bruno Haible.
 * Replaced `struct shm_desc' by `struct vm_area_struct', July 1994.
 * Fixed the shm swap deallocation (shm_unuse()), August 1998 Andrea Arcangeli.
 *
 * /proc/sysvipc/shm support (c) 1999 Dragos Acostachioaie <dragos@iname.com>
 * BIGMEM support, Andrea Arcangeli <andrea@suse.de>
 * SMP thread shm, Jean-Luc Boyard <jean-luc.boyard@siemens.fr>
 * HIGHMEM support, Ingo Molnar <mingo@redhat.com>
 * Make shmmax, shmall, shmmni sysctl'able, Christoph Rohland <cr@sap.com>
 * Shared /dev/zero support, Kanoj Sarcar <kanoj@sgi.com>
 * Move the mm functionality over to mm/shmem.c, Christoph Rohland <cr@sap.com>
 *
S
Steve Grubb 已提交
16 17
 * support for audit of ipc object properties and permission changes
 * Dustin Kirkland <dustin.kirkland@us.ibm.com>
K
Kirill Korotaev 已提交
18 19 20 21
 *
 * namespaces support
 * OpenVZ, SWsoft Inc.
 * Pavel Emelianov <xemul@openvz.org>
L
Linus Torvalds 已提交
22 23 24 25 26 27 28 29 30 31 32 33 34
 */

#include <linux/slab.h>
#include <linux/mm.h>
#include <linux/hugetlb.h>
#include <linux/shm.h>
#include <linux/init.h>
#include <linux/file.h>
#include <linux/mman.h>
#include <linux/shmem_fs.h>
#include <linux/security.h>
#include <linux/syscalls.h>
#include <linux/audit.h>
35
#include <linux/capability.h>
S
Stephen Rothwell 已提交
36
#include <linux/ptrace.h>
37
#include <linux/seq_file.h>
N
Nadia Derbey 已提交
38
#include <linux/rwsem.h>
K
Kirill Korotaev 已提交
39
#include <linux/nsproxy.h>
40
#include <linux/mount.h>
41
#include <linux/ipc_namespace.h>
S
Stephen Rothwell 已提交
42

L
Linus Torvalds 已提交
43 44 45 46
#include <asm/uaccess.h>

#include "util.h"

47 48 49 50 51 52 53 54 55
struct shm_file_data {
	int id;
	struct ipc_namespace *ns;
	struct file *file;
	const struct vm_operations_struct *vm_ops;
};

#define shm_file_data(file) (*((struct shm_file_data **)&(file)->private_data))

56
static const struct file_operations shm_file_operations;
57
static const struct vm_operations_struct shm_vm_ops;
L
Linus Torvalds 已提交
58

59
#define shm_ids(ns)	((ns)->ids[IPC_SHM_IDS])
L
Linus Torvalds 已提交
60

K
Kirill Korotaev 已提交
61 62
#define shm_unlock(shp)			\
	ipc_unlock(&(shp)->shm_perm)
L
Linus Torvalds 已提交
63

N
Nadia Derbey 已提交
64
static int newseg(struct ipc_namespace *, struct ipc_params *);
65 66
static void shm_open(struct vm_area_struct *vma);
static void shm_close(struct vm_area_struct *vma);
K
Kirill Korotaev 已提交
67
static void shm_destroy (struct ipc_namespace *ns, struct shmid_kernel *shp);
L
Linus Torvalds 已提交
68
#ifdef CONFIG_PROC_FS
69
static int sysvipc_shm_proc_show(struct seq_file *s, void *it);
L
Linus Torvalds 已提交
70 71
#endif

72
void shm_init_ns(struct ipc_namespace *ns)
K
Kirill Korotaev 已提交
73 74 75 76
{
	ns->shm_ctlmax = SHMMAX;
	ns->shm_ctlall = SHMALL;
	ns->shm_ctlmni = SHMMNI;
77
	ns->shm_rmid_forced = 0;
K
Kirill Korotaev 已提交
78
	ns->shm_tot = 0;
W
WANG Cong 已提交
79
	ipc_init_ids(&shm_ids(ns));
K
Kirill Korotaev 已提交
80 81
}

N
Nadia Derbey 已提交
82
/*
N
Nadia Derbey 已提交
83 84
 * Called with shm_ids.rw_mutex (writer) and the shp structure locked.
 * Only shm_ids.rw_mutex remains locked on exit.
N
Nadia Derbey 已提交
85
 */
86
static void do_shm_rmid(struct ipc_namespace *ns, struct kern_ipc_perm *ipcp)
K
Kirill Korotaev 已提交
87
{
88 89 90
	struct shmid_kernel *shp;
	shp = container_of(ipcp, struct shmid_kernel, shm_perm);

K
Kirill Korotaev 已提交
91 92 93 94 95 96 97 98 99
	if (shp->shm_nattch){
		shp->shm_perm.mode |= SHM_DEST;
		/* Do not find it any more */
		shp->shm_perm.key = IPC_PRIVATE;
		shm_unlock(shp);
	} else
		shm_destroy(ns, shp);
}

100
#ifdef CONFIG_IPC_NS
K
Kirill Korotaev 已提交
101 102
void shm_exit_ns(struct ipc_namespace *ns)
{
103
	free_ipcs(ns, &shm_ids(ns), do_shm_rmid);
S
Serge E. Hallyn 已提交
104
	idr_destroy(&ns->ids[IPC_SHM_IDS].ipcs_idr);
K
Kirill Korotaev 已提交
105
}
106
#endif
L
Linus Torvalds 已提交
107 108 109

void __init shm_init (void)
{
110
	shm_init_ns(&init_ipc_ns);
111
	ipc_init_proc_interface("sysvipc/shm",
112 113 114 115 116
#if BITS_PER_LONG <= 32
				"       key      shmid perms       size  cpid  lpid nattch   uid   gid  cuid  cgid      atime      dtime      ctime        rss       swap\n",
#else
				"       key      shmid perms                  size  cpid  lpid nattch   uid   gid  cuid  cgid      atime      dtime      ctime                   rss                  swap\n",
#endif
K
Kirill Korotaev 已提交
117
				IPC_SHM_IDS, sysvipc_shm_proc_show);
L
Linus Torvalds 已提交
118 119
}

N
Nadia Derbey 已提交
120 121
/*
 * shm_lock_(check_) routines are called in the paths where the rw_mutex
N
Nadia Derbey 已提交
122
 * is not necessarily held.
N
Nadia Derbey 已提交
123
 */
124
static inline struct shmid_kernel *shm_lock(struct ipc_namespace *ns, int id)
L
Linus Torvalds 已提交
125
{
N
Nadia Derbey 已提交
126 127
	struct kern_ipc_perm *ipcp = ipc_lock(&shm_ids(ns), id);

128 129 130
	if (IS_ERR(ipcp))
		return (struct shmid_kernel *)ipcp;

N
Nadia Derbey 已提交
131
	return container_of(ipcp, struct shmid_kernel, shm_perm);
132 133
}

134 135 136 137 138 139
static inline void shm_lock_by_ptr(struct shmid_kernel *ipcp)
{
	rcu_read_lock();
	spin_lock(&ipcp->shm_perm.lock);
}

140 141 142
static inline struct shmid_kernel *shm_lock_check(struct ipc_namespace *ns,
						int id)
{
N
Nadia Derbey 已提交
143 144
	struct kern_ipc_perm *ipcp = ipc_lock_check(&shm_ids(ns), id);

145 146 147
	if (IS_ERR(ipcp))
		return (struct shmid_kernel *)ipcp;

N
Nadia Derbey 已提交
148
	return container_of(ipcp, struct shmid_kernel, shm_perm);
L
Linus Torvalds 已提交
149 150
}

N
Nadia Derbey 已提交
151
static inline void shm_rmid(struct ipc_namespace *ns, struct shmid_kernel *s)
L
Linus Torvalds 已提交
152
{
N
Nadia Derbey 已提交
153
	ipc_rmid(&shm_ids(ns), &s->shm_perm);
L
Linus Torvalds 已提交
154 155 156
}


157 158
/* This is called by fork, once for every shm attach. */
static void shm_open(struct vm_area_struct *vma)
K
Kirill Korotaev 已提交
159
{
160 161
	struct file *file = vma->vm_file;
	struct shm_file_data *sfd = shm_file_data(file);
L
Linus Torvalds 已提交
162 163
	struct shmid_kernel *shp;

164
	shp = shm_lock(sfd->ns, sfd->id);
165
	BUG_ON(IS_ERR(shp));
L
Linus Torvalds 已提交
166
	shp->shm_atim = get_seconds();
167
	shp->shm_lprid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
168 169 170 171 172 173 174
	shp->shm_nattch++;
	shm_unlock(shp);
}

/*
 * shm_destroy - free the struct shmid_kernel
 *
N
Nadia Derbey 已提交
175
 * @ns: namespace
L
Linus Torvalds 已提交
176 177
 * @shp: struct to free
 *
N
Nadia Derbey 已提交
178
 * It has to be called with shp and shm_ids.rw_mutex (writer) locked,
L
Linus Torvalds 已提交
179 180
 * but returns with shp unlocked and freed.
 */
K
Kirill Korotaev 已提交
181
static void shm_destroy(struct ipc_namespace *ns, struct shmid_kernel *shp)
L
Linus Torvalds 已提交
182
{
K
Kirill Korotaev 已提交
183
	ns->shm_tot -= (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT;
N
Nadia Derbey 已提交
184
	shm_rmid(ns, shp);
L
Linus Torvalds 已提交
185 186 187
	shm_unlock(shp);
	if (!is_file_hugepages(shp->shm_file))
		shmem_lock(shp->shm_file, 0, shp->mlock_user);
188
	else if (shp->mlock_user)
J
Josef Sipek 已提交
189
		user_shm_unlock(shp->shm_file->f_path.dentry->d_inode->i_size,
L
Linus Torvalds 已提交
190 191 192 193 194 195
						shp->mlock_user);
	fput (shp->shm_file);
	security_shm_free(shp);
	ipc_rcu_putref(shp);
}

196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212
/*
 * shm_may_destroy - identifies whether shm segment should be destroyed now
 *
 * Returns true if and only if there are no active users of the segment and
 * one of the following is true:
 *
 * 1) shmctl(id, IPC_RMID, NULL) was called for this shp
 *
 * 2) sysctl kernel.shm_rmid_forced is set to 1.
 */
static bool shm_may_destroy(struct ipc_namespace *ns, struct shmid_kernel *shp)
{
	return (shp->shm_nattch == 0) &&
	       (ns->shm_rmid_forced ||
		(shp->shm_perm.mode & SHM_DEST));
}

L
Linus Torvalds 已提交
213
/*
214
 * remove the attach descriptor vma.
L
Linus Torvalds 已提交
215 216 217 218
 * free memory for segment if it is marked destroyed.
 * The descriptor has already been removed from the current->mm->mmap list
 * and will later be kfree()d.
 */
219
static void shm_close(struct vm_area_struct *vma)
L
Linus Torvalds 已提交
220
{
221 222
	struct file * file = vma->vm_file;
	struct shm_file_data *sfd = shm_file_data(file);
L
Linus Torvalds 已提交
223
	struct shmid_kernel *shp;
224
	struct ipc_namespace *ns = sfd->ns;
K
Kirill Korotaev 已提交
225

N
Nadia Derbey 已提交
226
	down_write(&shm_ids(ns).rw_mutex);
L
Linus Torvalds 已提交
227
	/* remove from the list of attaches of the shm segment */
N
Nadia Derbey 已提交
228
	shp = shm_lock(ns, sfd->id);
229
	BUG_ON(IS_ERR(shp));
230
	shp->shm_lprid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
231 232
	shp->shm_dtim = get_seconds();
	shp->shm_nattch--;
233 234 235 236 237 238 239
	if (shm_may_destroy(ns, shp))
		shm_destroy(ns, shp);
	else
		shm_unlock(shp);
	up_write(&shm_ids(ns).rw_mutex);
}

240
/* Called with ns->shm_ids(ns).rw_mutex locked */
241 242 243
static int shm_try_destroy_current(int id, void *p, void *data)
{
	struct ipc_namespace *ns = data;
244 245
	struct kern_ipc_perm *ipcp = p;
	struct shmid_kernel *shp = container_of(ipcp, struct shmid_kernel, shm_perm);
246

247
	if (shp->shm_creator != current)
248 249 250 251 252 253 254 255 256 257 258 259 260
		return 0;

	/*
	 * Mark it as orphaned to destroy the segment when
	 * kernel.shm_rmid_forced is changed.
	 * It is noop if the following shm_may_destroy() returns true.
	 */
	shp->shm_creator = NULL;

	/*
	 * Don't even try to destroy it.  If shm_rmid_forced=0 and IPC_RMID
	 * is not set, it shouldn't be deleted here.
	 */
261
	if (!ns->shm_rmid_forced)
262 263
		return 0;

264 265
	if (shm_may_destroy(ns, shp)) {
		shm_lock_by_ptr(shp);
266
		shm_destroy(ns, shp);
267
	}
268 269 270
	return 0;
}

271
/* Called with ns->shm_ids(ns).rw_mutex locked */
272 273 274
static int shm_try_destroy_orphaned(int id, void *p, void *data)
{
	struct ipc_namespace *ns = data;
275 276
	struct kern_ipc_perm *ipcp = p;
	struct shmid_kernel *shp = container_of(ipcp, struct shmid_kernel, shm_perm);
277 278 279 280

	/*
	 * We want to destroy segments without users and with already
	 * exit'ed originating process.
281 282
	 *
	 * As shp->* are changed under rw_mutex, it's safe to skip shp locking.
283
	 */
284
	if (shp->shm_creator != NULL)
285 286
		return 0;

287 288
	if (shm_may_destroy(ns, shp)) {
		shm_lock_by_ptr(shp);
K
Kirill Korotaev 已提交
289
		shm_destroy(ns, shp);
290
	}
291 292 293 294 295 296
	return 0;
}

void shm_destroy_orphaned(struct ipc_namespace *ns)
{
	down_write(&shm_ids(ns).rw_mutex);
V
Vasiliy Kulikov 已提交
297
	if (shm_ids(ns).in_use)
298
		idr_for_each(&shm_ids(ns).ipcs_idr, &shm_try_destroy_orphaned, ns);
299 300 301 302 303 304
	up_write(&shm_ids(ns).rw_mutex);
}


void exit_shm(struct task_struct *task)
{
305
	struct ipc_namespace *ns = task->nsproxy->ipc_ns;
306 307 308

	/* Destroy all already created segments, but not mapped yet */
	down_write(&shm_ids(ns).rw_mutex);
V
Vasiliy Kulikov 已提交
309
	if (shm_ids(ns).in_use)
310
		idr_for_each(&shm_ids(ns).ipcs_idr, &shm_try_destroy_current, ns);
N
Nadia Derbey 已提交
311
	up_write(&shm_ids(ns).rw_mutex);
L
Linus Torvalds 已提交
312 313
}

N
Nick Piggin 已提交
314
static int shm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
315 316 317 318
{
	struct file *file = vma->vm_file;
	struct shm_file_data *sfd = shm_file_data(file);

N
Nick Piggin 已提交
319
	return sfd->vm_ops->fault(vma, vmf);
320 321 322
}

#ifdef CONFIG_NUMA
A
Adrian Bunk 已提交
323
static int shm_set_policy(struct vm_area_struct *vma, struct mempolicy *new)
324 325 326 327 328 329 330 331 332
{
	struct file *file = vma->vm_file;
	struct shm_file_data *sfd = shm_file_data(file);
	int err = 0;
	if (sfd->vm_ops->set_policy)
		err = sfd->vm_ops->set_policy(vma, new);
	return err;
}

A
Adrian Bunk 已提交
333 334
static struct mempolicy *shm_get_policy(struct vm_area_struct *vma,
					unsigned long addr)
335 336 337 338 339 340 341
{
	struct file *file = vma->vm_file;
	struct shm_file_data *sfd = shm_file_data(file);
	struct mempolicy *pol = NULL;

	if (sfd->vm_ops->get_policy)
		pol = sfd->vm_ops->get_policy(vma, addr);
342
	else if (vma->vm_policy)
343
		pol = vma->vm_policy;
344

345 346 347 348
	return pol;
}
#endif

L
Linus Torvalds 已提交
349 350
static int shm_mmap(struct file * file, struct vm_area_struct * vma)
{
351
	struct shm_file_data *sfd = shm_file_data(file);
352 353
	int ret;

354 355 356 357
	ret = sfd->file->f_op->mmap(sfd->file, vma);
	if (ret != 0)
		return ret;
	sfd->vm_ops = vma->vm_ops;
D
David Howells 已提交
358
#ifdef CONFIG_MMU
359
	BUG_ON(!sfd->vm_ops->fault);
D
David Howells 已提交
360
#endif
361 362
	vma->vm_ops = &shm_vm_ops;
	shm_open(vma);
363 364

	return ret;
L
Linus Torvalds 已提交
365 366
}

K
Kirill Korotaev 已提交
367 368
static int shm_release(struct inode *ino, struct file *file)
{
369
	struct shm_file_data *sfd = shm_file_data(file);
K
Kirill Korotaev 已提交
370

371 372 373
	put_ipc_ns(sfd->ns);
	shm_file_data(file) = NULL;
	kfree(sfd);
K
Kirill Korotaev 已提交
374 375 376
	return 0;
}

377
static int shm_fsync(struct file *file, loff_t start, loff_t end, int datasync)
378 379 380
{
	struct shm_file_data *sfd = shm_file_data(file);

381 382
	if (!sfd->file->f_op->fsync)
		return -EINVAL;
383
	return sfd->file->f_op->fsync(sfd->file, start, end, datasync);
384 385
}

386 387 388 389 390
static unsigned long shm_get_unmapped_area(struct file *file,
	unsigned long addr, unsigned long len, unsigned long pgoff,
	unsigned long flags)
{
	struct shm_file_data *sfd = shm_file_data(file);
391 392
	return sfd->file->f_op->get_unmapped_area(sfd->file, addr, len,
						pgoff, flags);
393 394
}

395
static const struct file_operations shm_file_operations = {
K
Kirill Korotaev 已提交
396
	.mmap		= shm_mmap,
397
	.fsync		= shm_fsync,
K
Kirill Korotaev 已提交
398
	.release	= shm_release,
D
David Howells 已提交
399 400 401
#ifndef CONFIG_MMU
	.get_unmapped_area	= shm_get_unmapped_area,
#endif
402
	.llseek		= noop_llseek,
403 404 405 406 407 408
};

static const struct file_operations shm_file_operations_huge = {
	.mmap		= shm_mmap,
	.fsync		= shm_fsync,
	.release	= shm_release,
409
	.get_unmapped_area	= shm_get_unmapped_area,
410
	.llseek		= noop_llseek,
L
Linus Torvalds 已提交
411 412
};

413 414 415 416 417
int is_file_shm_hugepages(struct file *file)
{
	return file->f_op == &shm_file_operations_huge;
}

418
static const struct vm_operations_struct shm_vm_ops = {
L
Linus Torvalds 已提交
419 420
	.open	= shm_open,	/* callback for a new vm-area open */
	.close	= shm_close,	/* callback for when the vm-area is released */
421
	.fault	= shm_fault,
422 423 424
#if defined(CONFIG_NUMA)
	.set_policy = shm_set_policy,
	.get_policy = shm_get_policy,
L
Linus Torvalds 已提交
425 426 427
#endif
};

N
Nadia Derbey 已提交
428 429 430 431 432
/**
 * newseg - Create a new shared memory segment
 * @ns: namespace
 * @params: ptr to the structure that contains key, size and shmflg
 *
N
Nadia Derbey 已提交
433
 * Called with shm_ids.rw_mutex held as a writer.
N
Nadia Derbey 已提交
434 435
 */

N
Nadia Derbey 已提交
436
static int newseg(struct ipc_namespace *ns, struct ipc_params *params)
L
Linus Torvalds 已提交
437
{
N
Nadia Derbey 已提交
438 439 440
	key_t key = params->key;
	int shmflg = params->flg;
	size_t size = params->u.size;
L
Linus Torvalds 已提交
441 442 443 444 445 446
	int error;
	struct shmid_kernel *shp;
	int numpages = (size + PAGE_SIZE -1) >> PAGE_SHIFT;
	struct file * file;
	char name[13];
	int id;
447
	vm_flags_t acctflag = 0;
L
Linus Torvalds 已提交
448

K
Kirill Korotaev 已提交
449
	if (size < SHMMIN || size > ns->shm_ctlmax)
L
Linus Torvalds 已提交
450 451
		return -EINVAL;

452
	if (ns->shm_tot + numpages > ns->shm_ctlall)
L
Linus Torvalds 已提交
453 454 455 456 457 458 459
		return -ENOSPC;

	shp = ipc_rcu_alloc(sizeof(*shp));
	if (!shp)
		return -ENOMEM;

	shp->shm_perm.key = key;
A
Andrew Morton 已提交
460
	shp->shm_perm.mode = (shmflg & S_IRWXUGO);
L
Linus Torvalds 已提交
461 462 463 464 465 466 467 468 469
	shp->mlock_user = NULL;

	shp->shm_perm.security = NULL;
	error = security_shm_alloc(shp);
	if (error) {
		ipc_rcu_putref(shp);
		return error;
	}

470
	sprintf (name, "SYSV%08x", key);
L
Linus Torvalds 已提交
471
	if (shmflg & SHM_HUGETLB) {
472 473 474
		/* hugetlb_file_setup applies strict accounting */
		if (shmflg & SHM_NORESERVE)
			acctflag = VM_NORESERVE;
475
		file = hugetlb_file_setup(name, size, acctflag,
476
					&shp->mlock_user, HUGETLB_SHMFS_INODE);
L
Linus Torvalds 已提交
477
	} else {
478 479 480 481 482 483
		/*
		 * Do not allow no accounting for OVERCOMMIT_NEVER, even
	 	 * if it's asked for.
		 */
		if  ((shmflg & SHM_NORESERVE) &&
				sysctl_overcommit_memory != OVERCOMMIT_NEVER)
484
			acctflag = VM_NORESERVE;
485
		file = shmem_file_setup(name, size, acctflag);
L
Linus Torvalds 已提交
486 487 488 489 490
	}
	error = PTR_ERR(file);
	if (IS_ERR(file))
		goto no_file;

491
	id = ipc_addid(&shm_ids(ns), &shp->shm_perm, ns->shm_ctlmni);
492 493
	if (id < 0) {
		error = id;
L
Linus Torvalds 已提交
494
		goto no_id;
495
	}
L
Linus Torvalds 已提交
496

497
	shp->shm_cprid = task_tgid_vnr(current);
L
Linus Torvalds 已提交
498 499 500 501 502 503
	shp->shm_lprid = 0;
	shp->shm_atim = shp->shm_dtim = 0;
	shp->shm_ctim = get_seconds();
	shp->shm_segsz = size;
	shp->shm_nattch = 0;
	shp->shm_file = file;
504
	shp->shm_creator = current;
505 506 507 508
	/*
	 * shmid gets reported as "inode#" in /proc/pid/maps.
	 * proc-ps tools use this. Changing this will break them.
	 */
N
Nadia Derbey 已提交
509
	file->f_dentry->d_inode->i_ino = shp->shm_perm.id;
510

K
Kirill Korotaev 已提交
511
	ns->shm_tot += numpages;
N
Nadia Derbey 已提交
512
	error = shp->shm_perm.id;
L
Linus Torvalds 已提交
513
	shm_unlock(shp);
N
Nadia Derbey 已提交
514
	return error;
L
Linus Torvalds 已提交
515 516

no_id:
517
	if (is_file_hugepages(file) && shp->mlock_user)
518
		user_shm_unlock(size, shp->mlock_user);
L
Linus Torvalds 已提交
519 520 521 522 523 524 525
	fput(file);
no_file:
	security_shm_free(shp);
	ipc_rcu_putref(shp);
	return error;
}

N
Nadia Derbey 已提交
526
/*
N
Nadia Derbey 已提交
527
 * Called with shm_ids.rw_mutex and ipcp locked.
N
Nadia Derbey 已提交
528
 */
N
Nadia Derbey 已提交
529
static inline int shm_security(struct kern_ipc_perm *ipcp, int shmflg)
N
Nadia Derbey 已提交
530
{
N
Nadia Derbey 已提交
531 532 533 534
	struct shmid_kernel *shp;

	shp = container_of(ipcp, struct shmid_kernel, shm_perm);
	return security_shm_associate(shp, shmflg);
N
Nadia Derbey 已提交
535 536
}

N
Nadia Derbey 已提交
537
/*
N
Nadia Derbey 已提交
538
 * Called with shm_ids.rw_mutex and ipcp locked.
N
Nadia Derbey 已提交
539
 */
N
Nadia Derbey 已提交
540 541
static inline int shm_more_checks(struct kern_ipc_perm *ipcp,
				struct ipc_params *params)
N
Nadia Derbey 已提交
542
{
N
Nadia Derbey 已提交
543 544 545 546
	struct shmid_kernel *shp;

	shp = container_of(ipcp, struct shmid_kernel, shm_perm);
	if (shp->shm_segsz < params->u.size)
N
Nadia Derbey 已提交
547 548 549 550 551
		return -EINVAL;

	return 0;
}

552
SYSCALL_DEFINE3(shmget, key_t, key, size_t, size, int, shmflg)
L
Linus Torvalds 已提交
553
{
K
Kirill Korotaev 已提交
554
	struct ipc_namespace *ns;
N
Nadia Derbey 已提交
555 556
	struct ipc_ops shm_ops;
	struct ipc_params shm_params;
K
Kirill Korotaev 已提交
557 558

	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
559

N
Nadia Derbey 已提交
560 561 562
	shm_ops.getnew = newseg;
	shm_ops.associate = shm_security;
	shm_ops.more_checks = shm_more_checks;
N
Nadia Derbey 已提交
563

N
Nadia Derbey 已提交
564 565 566
	shm_params.key = key;
	shm_params.flg = shmflg;
	shm_params.u.size = size;
L
Linus Torvalds 已提交
567

N
Nadia Derbey 已提交
568
	return ipcget(ns, &shm_ids(ns), &shm_ops, &shm_params);
L
Linus Torvalds 已提交
569 570 571 572 573 574 575 576 577 578 579
}

static inline unsigned long copy_shmid_to_user(void __user *buf, struct shmid64_ds *in, int version)
{
	switch(version) {
	case IPC_64:
		return copy_to_user(buf, in, sizeof(*in));
	case IPC_OLD:
	    {
		struct shmid_ds out;

580
		memset(&out, 0, sizeof(out));
L
Linus Torvalds 已提交
581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596
		ipc64_perm_to_ipc_perm(&in->shm_perm, &out.shm_perm);
		out.shm_segsz	= in->shm_segsz;
		out.shm_atime	= in->shm_atime;
		out.shm_dtime	= in->shm_dtime;
		out.shm_ctime	= in->shm_ctime;
		out.shm_cpid	= in->shm_cpid;
		out.shm_lpid	= in->shm_lpid;
		out.shm_nattch	= in->shm_nattch;

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

597 598
static inline unsigned long
copy_shmid_from_user(struct shmid64_ds *out, void __user *buf, int version)
L
Linus Torvalds 已提交
599 600 601
{
	switch(version) {
	case IPC_64:
602
		if (copy_from_user(out, buf, sizeof(*out)))
L
Linus Torvalds 已提交
603 604 605 606 607 608 609 610 611
			return -EFAULT;
		return 0;
	case IPC_OLD:
	    {
		struct shmid_ds tbuf_old;

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

612 613 614
		out->shm_perm.uid	= tbuf_old.shm_perm.uid;
		out->shm_perm.gid	= tbuf_old.shm_perm.gid;
		out->shm_perm.mode	= tbuf_old.shm_perm.mode;
L
Linus Torvalds 已提交
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

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

static inline unsigned long copy_shminfo_to_user(void __user *buf, struct shminfo64 *in, int version)
{
	switch(version) {
	case IPC_64:
		return copy_to_user(buf, in, sizeof(*in));
	case IPC_OLD:
	    {
		struct shminfo out;

		if(in->shmmax > INT_MAX)
			out.shmmax = INT_MAX;
		else
			out.shmmax = (int)in->shmmax;

		out.shmmin	= in->shmmin;
		out.shmmni	= in->shmmni;
		out.shmseg	= in->shmseg;
		out.shmall	= in->shmall; 

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

649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676
/*
 * Calculate and add used RSS and swap pages of a shm.
 * Called with shm_ids.rw_mutex held as a reader
 */
static void shm_add_rss_swap(struct shmid_kernel *shp,
	unsigned long *rss_add, unsigned long *swp_add)
{
	struct inode *inode;

	inode = shp->shm_file->f_path.dentry->d_inode;

	if (is_file_hugepages(shp->shm_file)) {
		struct address_space *mapping = inode->i_mapping;
		struct hstate *h = hstate_file(shp->shm_file);
		*rss_add += pages_per_huge_page(h) * mapping->nrpages;
	} else {
#ifdef CONFIG_SHMEM
		struct shmem_inode_info *info = SHMEM_I(inode);
		spin_lock(&info->lock);
		*rss_add += inode->i_mapping->nrpages;
		*swp_add += info->swapped;
		spin_unlock(&info->lock);
#else
		*rss_add += inode->i_mapping->nrpages;
#endif
	}
}

N
Nadia Derbey 已提交
677
/*
N
Nadia Derbey 已提交
678
 * Called with shm_ids.rw_mutex held as a reader
N
Nadia Derbey 已提交
679
 */
K
Kirill Korotaev 已提交
680 681
static void shm_get_stat(struct ipc_namespace *ns, unsigned long *rss,
		unsigned long *swp)
L
Linus Torvalds 已提交
682
{
N
Nadia Derbey 已提交
683 684
	int next_id;
	int total, in_use;
L
Linus Torvalds 已提交
685 686 687 688

	*rss = 0;
	*swp = 0;

N
Nadia Derbey 已提交
689 690 691
	in_use = shm_ids(ns).in_use;

	for (total = 0, next_id = 0; total < in_use; next_id++) {
692
		struct kern_ipc_perm *ipc;
L
Linus Torvalds 已提交
693 694
		struct shmid_kernel *shp;

695 696
		ipc = idr_find(&shm_ids(ns).ipcs_idr, next_id);
		if (ipc == NULL)
L
Linus Torvalds 已提交
697
			continue;
698
		shp = container_of(ipc, struct shmid_kernel, shm_perm);
L
Linus Torvalds 已提交
699

700
		shm_add_rss_swap(shp, rss, swp);
N
Nadia Derbey 已提交
701 702

		total++;
L
Linus Torvalds 已提交
703 704 705
	}
}

706 707 708 709 710 711 712
/*
 * This function handles some shmctl 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.
 */
static int shmctl_down(struct ipc_namespace *ns, int shmid, int cmd,
		       struct shmid_ds __user *buf, int version)
L
Linus Torvalds 已提交
713
{
714
	struct kern_ipc_perm *ipcp;
715
	struct shmid64_ds shmid64;
716 717 718 719
	struct shmid_kernel *shp;
	int err;

	if (cmd == IPC_SET) {
720
		if (copy_shmid_from_user(&shmid64, buf, version))
721 722 723
			return -EFAULT;
	}

724 725
	ipcp = ipcctl_pre_down(ns, &shm_ids(ns), shmid, cmd,
			       &shmid64.shm_perm, 0);
726 727
	if (IS_ERR(ipcp))
		return PTR_ERR(ipcp);
728

729
	shp = container_of(ipcp, struct shmid_kernel, shm_perm);
730 731 732 733 734 735 736 737 738

	err = security_shm_shmctl(shp, cmd);
	if (err)
		goto out_unlock;
	switch (cmd) {
	case IPC_RMID:
		do_shm_rmid(ns, ipcp);
		goto out_up;
	case IPC_SET:
739
		ipc_update_perm(&shmid64.shm_perm, ipcp);
740 741 742 743 744 745 746 747 748 749 750 751
		shp->shm_ctim = get_seconds();
		break;
	default:
		err = -EINVAL;
	}
out_unlock:
	shm_unlock(shp);
out_up:
	up_write(&shm_ids(ns).rw_mutex);
	return err;
}

752
SYSCALL_DEFINE3(shmctl, int, shmid, int, cmd, struct shmid_ds __user *, buf)
753
{
L
Linus Torvalds 已提交
754 755
	struct shmid_kernel *shp;
	int err, version;
K
Kirill Korotaev 已提交
756
	struct ipc_namespace *ns;
L
Linus Torvalds 已提交
757 758 759 760 761 762 763

	if (cmd < 0 || shmid < 0) {
		err = -EINVAL;
		goto out;
	}

	version = ipc_parse_version(&cmd);
K
Kirill Korotaev 已提交
764
	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
765 766 767 768 769 770 771 772 773 774

	switch (cmd) { /* replace with proc interface ? */
	case IPC_INFO:
	{
		struct shminfo64 shminfo;

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

W
WANG Cong 已提交
775
		memset(&shminfo, 0, sizeof(shminfo));
K
Kirill Korotaev 已提交
776 777 778
		shminfo.shmmni = shminfo.shmseg = ns->shm_ctlmni;
		shminfo.shmmax = ns->shm_ctlmax;
		shminfo.shmall = ns->shm_ctlall;
L
Linus Torvalds 已提交
779 780 781 782

		shminfo.shmmin = SHMMIN;
		if(copy_shminfo_to_user (buf, &shminfo, version))
			return -EFAULT;
N
Nadia Derbey 已提交
783

N
Nadia Derbey 已提交
784
		down_read(&shm_ids(ns).rw_mutex);
N
Nadia Derbey 已提交
785
		err = ipc_get_maxid(&shm_ids(ns));
N
Nadia Derbey 已提交
786
		up_read(&shm_ids(ns).rw_mutex);
N
Nadia Derbey 已提交
787

L
Linus Torvalds 已提交
788 789 790 791 792 793 794 795 796 797 798 799
		if(err<0)
			err = 0;
		goto out;
	}
	case SHM_INFO:
	{
		struct shm_info shm_info;

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

W
WANG Cong 已提交
800
		memset(&shm_info, 0, sizeof(shm_info));
N
Nadia Derbey 已提交
801
		down_read(&shm_ids(ns).rw_mutex);
K
Kirill Korotaev 已提交
802 803 804
		shm_info.used_ids = shm_ids(ns).in_use;
		shm_get_stat (ns, &shm_info.shm_rss, &shm_info.shm_swp);
		shm_info.shm_tot = ns->shm_tot;
L
Linus Torvalds 已提交
805 806
		shm_info.swap_attempts = 0;
		shm_info.swap_successes = 0;
N
Nadia Derbey 已提交
807
		err = ipc_get_maxid(&shm_ids(ns));
N
Nadia Derbey 已提交
808
		up_read(&shm_ids(ns).rw_mutex);
W
WANG Cong 已提交
809
		if (copy_to_user(buf, &shm_info, sizeof(shm_info))) {
L
Linus Torvalds 已提交
810 811 812 813 814 815 816 817 818 819 820 821
			err = -EFAULT;
			goto out;
		}

		err = err < 0 ? 0 : err;
		goto out;
	}
	case SHM_STAT:
	case IPC_STAT:
	{
		struct shmid64_ds tbuf;
		int result;
822 823 824 825 826 827 828

		if (cmd == SHM_STAT) {
			shp = shm_lock(ns, shmid);
			if (IS_ERR(shp)) {
				err = PTR_ERR(shp);
				goto out;
			}
N
Nadia Derbey 已提交
829
			result = shp->shm_perm.id;
L
Linus Torvalds 已提交
830
		} else {
831 832 833 834 835
			shp = shm_lock_check(ns, shmid);
			if (IS_ERR(shp)) {
				err = PTR_ERR(shp);
				goto out;
			}
L
Linus Torvalds 已提交
836 837
			result = 0;
		}
W
WANG Cong 已提交
838
		err = -EACCES;
839
		if (ipcperms(ns, &shp->shm_perm, S_IRUGO))
L
Linus Torvalds 已提交
840 841 842 843
			goto out_unlock;
		err = security_shm_shmctl(shp, cmd);
		if (err)
			goto out_unlock;
844
		memset(&tbuf, 0, sizeof(tbuf));
L
Linus Torvalds 已提交
845 846 847 848 849 850 851
		kernel_to_ipc64_perm(&shp->shm_perm, &tbuf.shm_perm);
		tbuf.shm_segsz	= shp->shm_segsz;
		tbuf.shm_atime	= shp->shm_atim;
		tbuf.shm_dtime	= shp->shm_dtim;
		tbuf.shm_ctime	= shp->shm_ctim;
		tbuf.shm_cpid	= shp->shm_cprid;
		tbuf.shm_lpid	= shp->shm_lprid;
852
		tbuf.shm_nattch	= shp->shm_nattch;
L
Linus Torvalds 已提交
853 854 855 856 857 858 859 860 861 862
		shm_unlock(shp);
		if(copy_shmid_to_user (buf, &tbuf, version))
			err = -EFAULT;
		else
			err = result;
		goto out;
	}
	case SHM_LOCK:
	case SHM_UNLOCK:
	{
863 864 865 866
		struct file *uninitialized_var(shm_file);

		lru_add_drain_all();  /* drain pagevecs to lru lists */

867 868 869
		shp = shm_lock_check(ns, shmid);
		if (IS_ERR(shp)) {
			err = PTR_ERR(shp);
L
Linus Torvalds 已提交
870 871 872
			goto out;
		}

A
Al Viro 已提交
873
		audit_ipc_obj(&(shp->shm_perm));
S
Steve Grubb 已提交
874

875
		if (!ns_capable(ns->user_ns, CAP_IPC_LOCK)) {
876
			uid_t euid = current_euid();
L
Linus Torvalds 已提交
877
			err = -EPERM;
878 879
			if (euid != shp->shm_perm.uid &&
			    euid != shp->shm_perm.cuid)
L
Linus Torvalds 已提交
880
				goto out_unlock;
J
Jiri Slaby 已提交
881
			if (cmd == SHM_LOCK && !rlimit(RLIMIT_MEMLOCK))
L
Linus Torvalds 已提交
882 883 884 885 886 887 888 889
				goto out_unlock;
		}

		err = security_shm_shmctl(shp, cmd);
		if (err)
			goto out_unlock;
		
		if(cmd==SHM_LOCK) {
890
			struct user_struct *user = current_user();
L
Linus Torvalds 已提交
891 892
			if (!is_file_hugepages(shp->shm_file)) {
				err = shmem_lock(shp->shm_file, 1, user);
893
				if (!err && !(shp->shm_perm.mode & SHM_LOCKED)){
A
Andrew Morton 已提交
894
					shp->shm_perm.mode |= SHM_LOCKED;
L
Linus Torvalds 已提交
895 896 897 898 899
					shp->mlock_user = user;
				}
			}
		} else if (!is_file_hugepages(shp->shm_file)) {
			shmem_lock(shp->shm_file, 0, shp->mlock_user);
A
Andrew Morton 已提交
900
			shp->shm_perm.mode &= ~SHM_LOCKED;
L
Linus Torvalds 已提交
901 902 903 904 905 906 907
			shp->mlock_user = NULL;
		}
		shm_unlock(shp);
		goto out;
	}
	case IPC_RMID:
	case IPC_SET:
908 909
		err = shmctl_down(ns, shmid, cmd, buf, version);
		return err;
L
Linus Torvalds 已提交
910
	default:
911
		return -EINVAL;
L
Linus Torvalds 已提交
912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936
	}

out_unlock:
	shm_unlock(shp);
out:
	return err;
}

/*
 * Fix shmaddr, allocate descriptor, map shm, add attach descriptor to lists.
 *
 * NOTE! Despite the name, this is NOT a direct system call entrypoint. The
 * "raddr" thing points to kernel space, and there has to be a wrapper around
 * this.
 */
long do_shmat(int shmid, char __user *shmaddr, int shmflg, ulong *raddr)
{
	struct shmid_kernel *shp;
	unsigned long addr;
	unsigned long size;
	struct file * file;
	int    err;
	unsigned long flags;
	unsigned long prot;
	int acc_mode;
937
	unsigned long user_addr;
K
Kirill Korotaev 已提交
938
	struct ipc_namespace *ns;
939 940
	struct shm_file_data *sfd;
	struct path path;
941
	fmode_t f_mode;
L
Linus Torvalds 已提交
942

943 944
	err = -EINVAL;
	if (shmid < 0)
L
Linus Torvalds 已提交
945
		goto out;
946
	else if ((addr = (ulong)shmaddr)) {
L
Linus Torvalds 已提交
947 948 949 950 951 952 953
		if (addr & (SHMLBA-1)) {
			if (shmflg & SHM_RND)
				addr &= ~(SHMLBA-1);	   /* round down */
			else
#ifndef __ARCH_FORCE_SHMLBA
				if (addr & ~PAGE_MASK)
#endif
954
					goto out;
L
Linus Torvalds 已提交
955 956 957 958
		}
		flags = MAP_SHARED | MAP_FIXED;
	} else {
		if ((shmflg & SHM_REMAP))
959
			goto out;
L
Linus Torvalds 已提交
960 961 962 963 964 965 966

		flags = MAP_SHARED;
	}

	if (shmflg & SHM_RDONLY) {
		prot = PROT_READ;
		acc_mode = S_IRUGO;
967
		f_mode = FMODE_READ;
L
Linus Torvalds 已提交
968 969 970
	} else {
		prot = PROT_READ | PROT_WRITE;
		acc_mode = S_IRUGO | S_IWUGO;
971
		f_mode = FMODE_READ | FMODE_WRITE;
L
Linus Torvalds 已提交
972 973 974 975 976 977 978 979 980 981
	}
	if (shmflg & SHM_EXEC) {
		prot |= PROT_EXEC;
		acc_mode |= S_IXUGO;
	}

	/*
	 * We cannot rely on the fs check since SYSV IPC does have an
	 * additional creator id...
	 */
K
Kirill Korotaev 已提交
982
	ns = current->nsproxy->ipc_ns;
983 984 985
	shp = shm_lock_check(ns, shmid);
	if (IS_ERR(shp)) {
		err = PTR_ERR(shp);
L
Linus Torvalds 已提交
986
		goto out;
987
	}
988 989

	err = -EACCES;
990
	if (ipcperms(ns, &shp->shm_perm, acc_mode))
991
		goto out_unlock;
L
Linus Torvalds 已提交
992 993

	err = security_shm_shmat(shp, shmaddr, shmflg);
994 995 996
	if (err)
		goto out_unlock;

997 998
	path = shp->shm_file->f_path;
	path_get(&path);
L
Linus Torvalds 已提交
999
	shp->shm_nattch++;
1000
	size = i_size_read(path.dentry->d_inode);
L
Linus Torvalds 已提交
1001 1002
	shm_unlock(shp);

1003 1004 1005
	err = -ENOMEM;
	sfd = kzalloc(sizeof(*sfd), GFP_KERNEL);
	if (!sfd)
1006
		goto out_put_dentry;
1007

1008 1009
	file = alloc_file(&path, f_mode,
			  is_file_hugepages(shp->shm_file) ?
1010 1011
				&shm_file_operations_huge :
				&shm_file_operations);
1012 1013 1014 1015 1016
	if (!file)
		goto out_free;

	file->private_data = sfd;
	file->f_mapping = shp->shm_file->f_mapping;
N
Nadia Derbey 已提交
1017
	sfd->id = shp->shm_perm.id;
1018 1019 1020 1021
	sfd->ns = get_ipc_ns(ns);
	sfd->file = shp->shm_file;
	sfd->vm_ops = NULL;

L
Linus Torvalds 已提交
1022 1023
	down_write(&current->mm->mmap_sem);
	if (addr && !(shmflg & SHM_REMAP)) {
1024
		err = -EINVAL;
L
Linus Torvalds 已提交
1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
		if (find_vma_intersection(current->mm, addr, addr + size))
			goto invalid;
		/*
		 * If shm segment goes below stack, make sure there is some
		 * space left for the stack to grow (at least 4 pages).
		 */
		if (addr < current->mm->start_stack &&
		    addr > current->mm->start_stack - size - PAGE_SIZE * 5)
			goto invalid;
	}
		
1036 1037 1038 1039 1040
	user_addr = do_mmap (file, addr, size, prot, flags, 0);
	*raddr = user_addr;
	err = 0;
	if (IS_ERR_VALUE(user_addr))
		err = (long)user_addr;
L
Linus Torvalds 已提交
1041 1042 1043
invalid:
	up_write(&current->mm->mmap_sem);

1044 1045 1046
	fput(file);

out_nattch:
N
Nadia Derbey 已提交
1047
	down_write(&shm_ids(ns).rw_mutex);
N
Nadia Derbey 已提交
1048
	shp = shm_lock(ns, shmid);
1049
	BUG_ON(IS_ERR(shp));
L
Linus Torvalds 已提交
1050
	shp->shm_nattch--;
1051
	if (shm_may_destroy(ns, shp))
K
Kirill Korotaev 已提交
1052
		shm_destroy(ns, shp);
L
Linus Torvalds 已提交
1053 1054
	else
		shm_unlock(shp);
N
Nadia Derbey 已提交
1055
	up_write(&shm_ids(ns).rw_mutex);
L
Linus Torvalds 已提交
1056 1057 1058

out:
	return err;
1059 1060 1061 1062 1063 1064 1065

out_unlock:
	shm_unlock(shp);
	goto out;

out_free:
	kfree(sfd);
1066
out_put_dentry:
1067
	path_put(&path);
1068
	goto out_nattch;
L
Linus Torvalds 已提交
1069 1070
}

1071
SYSCALL_DEFINE3(shmat, int, shmid, char __user *, shmaddr, int, shmflg)
S
Stephen Rothwell 已提交
1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082
{
	unsigned long ret;
	long err;

	err = do_shmat(shmid, shmaddr, shmflg, &ret);
	if (err)
		return err;
	force_successful_syscall_return();
	return (long)ret;
}

L
Linus Torvalds 已提交
1083 1084 1085 1086
/*
 * detach and kill segment if marked destroyed.
 * The work is done in shm_close.
 */
1087
SYSCALL_DEFINE1(shmdt, char __user *, shmaddr)
L
Linus Torvalds 已提交
1088 1089
{
	struct mm_struct *mm = current->mm;
1090
	struct vm_area_struct *vma;
L
Linus Torvalds 已提交
1091 1092
	unsigned long addr = (unsigned long)shmaddr;
	int retval = -EINVAL;
1093 1094 1095 1096
#ifdef CONFIG_MMU
	loff_t size = 0;
	struct vm_area_struct *next;
#endif
L
Linus Torvalds 已提交
1097

1098 1099 1100
	if (addr & ~PAGE_MASK)
		return retval;

L
Linus Torvalds 已提交
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124
	down_write(&mm->mmap_sem);

	/*
	 * This function tries to be smart and unmap shm segments that
	 * were modified by partial mlock or munmap calls:
	 * - It first determines the size of the shm segment that should be
	 *   unmapped: It searches for a vma that is backed by shm and that
	 *   started at address shmaddr. It records it's size and then unmaps
	 *   it.
	 * - Then it unmaps all shm vmas that started at shmaddr and that
	 *   are within the initially determined size.
	 * Errors from do_munmap are ignored: the function only fails if
	 * it's called with invalid parameters or if it's called to unmap
	 * a part of a vma. Both calls in this function are for full vmas,
	 * the parameters are directly copied from the vma itself and always
	 * valid - therefore do_munmap cannot fail. (famous last words?)
	 */
	/*
	 * If it had been mremap()'d, the starting address would not
	 * match the usual checks anyway. So assume all vma's are
	 * above the starting address given.
	 */
	vma = find_vma(mm, addr);

1125
#ifdef CONFIG_MMU
L
Linus Torvalds 已提交
1126 1127 1128 1129 1130 1131 1132 1133
	while (vma) {
		next = vma->vm_next;

		/*
		 * Check if the starting address would match, i.e. it's
		 * a fragment created by mprotect() and/or munmap(), or it
		 * otherwise it starts at this address with no hassles.
		 */
1134
		if ((vma->vm_ops == &shm_vm_ops) &&
L
Linus Torvalds 已提交
1135 1136 1137
			(vma->vm_start - addr)/PAGE_SIZE == vma->vm_pgoff) {


J
Josef Sipek 已提交
1138
			size = vma->vm_file->f_path.dentry->d_inode->i_size;
L
Linus Torvalds 已提交
1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155
			do_munmap(mm, vma->vm_start, vma->vm_end - vma->vm_start);
			/*
			 * We discovered the size of the shm segment, so
			 * break out of here and fall through to the next
			 * loop that uses the size information to stop
			 * searching for matching vma's.
			 */
			retval = 0;
			vma = next;
			break;
		}
		vma = next;
	}

	/*
	 * We need look no further than the maximum address a fragment
	 * could possibly have landed at. Also cast things to loff_t to
L
Lucas De Marchi 已提交
1156
	 * prevent overflows and make comparisons vs. equal-width types.
L
Linus Torvalds 已提交
1157
	 */
1158
	size = PAGE_ALIGN(size);
L
Linus Torvalds 已提交
1159 1160 1161 1162
	while (vma && (loff_t)(vma->vm_end - addr) <= size) {
		next = vma->vm_next;

		/* finding a matching vma now does not alter retval */
1163
		if ((vma->vm_ops == &shm_vm_ops) &&
L
Linus Torvalds 已提交
1164 1165 1166 1167 1168 1169
			(vma->vm_start - addr)/PAGE_SIZE == vma->vm_pgoff)

			do_munmap(mm, vma->vm_start, vma->vm_end - vma->vm_start);
		vma = next;
	}

1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180
#else /* CONFIG_MMU */
	/* under NOMMU conditions, the exact address to be destroyed must be
	 * given */
	retval = -EINVAL;
	if (vma->vm_start == addr && vma->vm_ops == &shm_vm_ops) {
		do_munmap(mm, vma->vm_start, vma->vm_end - vma->vm_start);
		retval = 0;
	}

#endif

L
Linus Torvalds 已提交
1181 1182 1183 1184 1185
	up_write(&mm->mmap_sem);
	return retval;
}

#ifdef CONFIG_PROC_FS
1186
static int sysvipc_shm_proc_show(struct seq_file *s, void *it)
L
Linus Torvalds 已提交
1187
{
1188
	struct shmid_kernel *shp = it;
1189 1190 1191
	unsigned long rss = 0, swp = 0;

	shm_add_rss_swap(shp, &rss, &swp);
L
Linus Torvalds 已提交
1192

1193 1194 1195 1196 1197
#if BITS_PER_LONG <= 32
#define SIZE_SPEC "%10lu"
#else
#define SIZE_SPEC "%21lu"
#endif
L
Linus Torvalds 已提交
1198

1199 1200
	return seq_printf(s,
			  "%10d %10d  %4o " SIZE_SPEC " %5u %5u  "
1201 1202
			  "%5lu %5u %5u %5u %5u %10lu %10lu %10lu "
			  SIZE_SPEC " " SIZE_SPEC "\n",
1203
			  shp->shm_perm.key,
N
Nadia Derbey 已提交
1204
			  shp->shm_perm.id,
A
Andrew Morton 已提交
1205
			  shp->shm_perm.mode,
1206 1207 1208
			  shp->shm_segsz,
			  shp->shm_cprid,
			  shp->shm_lprid,
1209
			  shp->shm_nattch,
1210 1211 1212 1213 1214 1215
			  shp->shm_perm.uid,
			  shp->shm_perm.gid,
			  shp->shm_perm.cuid,
			  shp->shm_perm.cgid,
			  shp->shm_atim,
			  shp->shm_dtim,
1216 1217 1218
			  shp->shm_ctim,
			  rss * PAGE_SIZE,
			  swp * PAGE_SIZE);
L
Linus Torvalds 已提交
1219 1220
}
#endif