shm.c 25.8 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>
I
Ingo Molnar 已提交
38
#include <linux/mutex.h>
K
Kirill Korotaev 已提交
39
#include <linux/nsproxy.h>
40
#include <linux/mount.h>
S
Stephen Rothwell 已提交
41

L
Linus Torvalds 已提交
42 43 44 45
#include <asm/uaccess.h>

#include "util.h"

46 47 48 49 50 51 52 53 54
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))

55
static const struct file_operations shm_file_operations;
L
Linus Torvalds 已提交
56 57
static struct vm_operations_struct shm_vm_ops;

K
Kirill Korotaev 已提交
58 59 60
static struct ipc_ids init_shm_ids;

#define shm_ids(ns)	(*((ns)->ids[IPC_SHM_IDS]))
L
Linus Torvalds 已提交
61

K
Kirill Korotaev 已提交
62 63 64 65 66 67 68 69
#define shm_lock(ns, id)		\
	((struct shmid_kernel*)ipc_lock(&shm_ids(ns),id))
#define shm_unlock(shp)			\
	ipc_unlock(&(shp)->shm_perm)
#define shm_get(ns, id)			\
	((struct shmid_kernel*)ipc_get(&shm_ids(ns),id))
#define shm_buildid(ns, id, seq)	\
	ipc_buildid(&shm_ids(ns), id, seq)
L
Linus Torvalds 已提交
70

K
Kirill Korotaev 已提交
71 72
static int newseg (struct ipc_namespace *ns, key_t key,
		int shmflg, size_t size);
73 74
static void shm_open(struct vm_area_struct *vma);
static void shm_close(struct vm_area_struct *vma);
K
Kirill Korotaev 已提交
75
static void shm_destroy (struct ipc_namespace *ns, struct shmid_kernel *shp);
L
Linus Torvalds 已提交
76
#ifdef CONFIG_PROC_FS
77
static int sysvipc_shm_proc_show(struct seq_file *s, void *it);
L
Linus Torvalds 已提交
78 79
#endif

80
static void __shm_init_ns(struct ipc_namespace *ns, struct ipc_ids *ids)
K
Kirill Korotaev 已提交
81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107
{
	ns->ids[IPC_SHM_IDS] = ids;
	ns->shm_ctlmax = SHMMAX;
	ns->shm_ctlall = SHMALL;
	ns->shm_ctlmni = SHMMNI;
	ns->shm_tot = 0;
	ipc_init_ids(ids, 1);
}

static void do_shm_rmid(struct ipc_namespace *ns, struct shmid_kernel *shp)
{
	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);
}

int shm_init_ns(struct ipc_namespace *ns)
{
	struct ipc_ids *ids;

	ids = kmalloc(sizeof(struct ipc_ids), GFP_KERNEL);
	if (ids == NULL)
		return -ENOMEM;
L
Linus Torvalds 已提交
108

K
Kirill Korotaev 已提交
109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127
	__shm_init_ns(ns, ids);
	return 0;
}

void shm_exit_ns(struct ipc_namespace *ns)
{
	int i;
	struct shmid_kernel *shp;

	mutex_lock(&shm_ids(ns).mutex);
	for (i = 0; i <= shm_ids(ns).max_id; i++) {
		shp = shm_lock(ns, i);
		if (shp == NULL)
			continue;

		do_shm_rmid(ns, shp);
	}
	mutex_unlock(&shm_ids(ns).mutex);

P
Pavel Emelianov 已提交
128
	ipc_fini_ids(ns->ids[IPC_SHM_IDS]);
K
Kirill Korotaev 已提交
129 130 131
	kfree(ns->ids[IPC_SHM_IDS]);
	ns->ids[IPC_SHM_IDS] = NULL;
}
L
Linus Torvalds 已提交
132 133 134

void __init shm_init (void)
{
K
Kirill Korotaev 已提交
135
	__shm_init_ns(&init_ipc_ns, &init_shm_ids);
136 137
	ipc_init_proc_interface("sysvipc/shm",
				"       key      shmid perms       size  cpid  lpid nattch   uid   gid  cuid  cgid      atime      dtime      ctime\n",
K
Kirill Korotaev 已提交
138
				IPC_SHM_IDS, sysvipc_shm_proc_show);
L
Linus Torvalds 已提交
139 140
}

K
Kirill Korotaev 已提交
141 142
static inline int shm_checkid(struct ipc_namespace *ns,
		struct shmid_kernel *s, int id)
L
Linus Torvalds 已提交
143
{
K
Kirill Korotaev 已提交
144
	if (ipc_checkid(&shm_ids(ns), &s->shm_perm, id))
L
Linus Torvalds 已提交
145 146 147 148
		return -EIDRM;
	return 0;
}

K
Kirill Korotaev 已提交
149
static inline struct shmid_kernel *shm_rmid(struct ipc_namespace *ns, int id)
L
Linus Torvalds 已提交
150
{
K
Kirill Korotaev 已提交
151
	return (struct shmid_kernel *)ipc_rmid(&shm_ids(ns), id);
L
Linus Torvalds 已提交
152 153
}

K
Kirill Korotaev 已提交
154
static inline int shm_addid(struct ipc_namespace *ns, struct shmid_kernel *shp)
L
Linus Torvalds 已提交
155
{
K
Kirill Korotaev 已提交
156
	return ipc_addid(&shm_ids(ns), &shp->shm_perm, ns->shm_ctlmni);
L
Linus Torvalds 已提交
157 158 159 160
}



161 162
/* 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
163
{
164 165
	struct file *file = vma->vm_file;
	struct shm_file_data *sfd = shm_file_data(file);
L
Linus Torvalds 已提交
166 167
	struct shmid_kernel *shp;

168
	shp = shm_lock(sfd->ns, sfd->id);
169
	BUG_ON(!shp);
L
Linus Torvalds 已提交
170 171 172 173 174 175 176 177 178 179 180
	shp->shm_atim = get_seconds();
	shp->shm_lprid = current->tgid;
	shp->shm_nattch++;
	shm_unlock(shp);
}

/*
 * shm_destroy - free the struct shmid_kernel
 *
 * @shp: struct to free
 *
I
Ingo Molnar 已提交
181
 * It has to be called with shp and shm_ids.mutex locked,
L
Linus Torvalds 已提交
182 183
 * but returns with shp unlocked and freed.
 */
K
Kirill Korotaev 已提交
184
static void shm_destroy(struct ipc_namespace *ns, struct shmid_kernel *shp)
L
Linus Torvalds 已提交
185
{
K
Kirill Korotaev 已提交
186 187
	ns->shm_tot -= (shp->shm_segsz + PAGE_SIZE - 1) >> PAGE_SHIFT;
	shm_rmid(ns, shp->id);
L
Linus Torvalds 已提交
188 189 190 191
	shm_unlock(shp);
	if (!is_file_hugepages(shp->shm_file))
		shmem_lock(shp->shm_file, 0, shp->mlock_user);
	else
J
Josef Sipek 已提交
192
		user_shm_unlock(shp->shm_file->f_path.dentry->d_inode->i_size,
L
Linus Torvalds 已提交
193 194 195 196 197 198 199
						shp->mlock_user);
	fput (shp->shm_file);
	security_shm_free(shp);
	ipc_rcu_putref(shp);
}

/*
200
 * remove the attach descriptor vma.
L
Linus Torvalds 已提交
201 202 203 204
 * 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.
 */
205
static void shm_close(struct vm_area_struct *vma)
L
Linus Torvalds 已提交
206
{
207 208
	struct file * file = vma->vm_file;
	struct shm_file_data *sfd = shm_file_data(file);
L
Linus Torvalds 已提交
209
	struct shmid_kernel *shp;
210
	struct ipc_namespace *ns = sfd->ns;
K
Kirill Korotaev 已提交
211 212

	mutex_lock(&shm_ids(ns).mutex);
L
Linus Torvalds 已提交
213
	/* remove from the list of attaches of the shm segment */
214
	shp = shm_lock(ns, sfd->id);
215
	BUG_ON(!shp);
L
Linus Torvalds 已提交
216 217 218 219
	shp->shm_lprid = current->tgid;
	shp->shm_dtim = get_seconds();
	shp->shm_nattch--;
	if(shp->shm_nattch == 0 &&
A
Andrew Morton 已提交
220
	   shp->shm_perm.mode & SHM_DEST)
K
Kirill Korotaev 已提交
221
		shm_destroy(ns, shp);
L
Linus Torvalds 已提交
222 223
	else
		shm_unlock(shp);
K
Kirill Korotaev 已提交
224
	mutex_unlock(&shm_ids(ns).mutex);
L
Linus Torvalds 已提交
225 226
}

N
Nick Piggin 已提交
227
static int shm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
228 229 230 231
{
	struct file *file = vma->vm_file;
	struct shm_file_data *sfd = shm_file_data(file);

N
Nick Piggin 已提交
232
	return sfd->vm_ops->fault(vma, vmf);
233 234 235
}

#ifdef CONFIG_NUMA
A
Adrian Bunk 已提交
236
static int shm_set_policy(struct vm_area_struct *vma, struct mempolicy *new)
237 238 239 240 241 242 243 244 245
{
	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
246 247
static struct mempolicy *shm_get_policy(struct vm_area_struct *vma,
					unsigned long addr)
248 249 250 251 252 253 254
{
	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);
255
	else if (vma->vm_policy)
256
		pol = vma->vm_policy;
257 258
	else
		pol = current->mempolicy;
259 260 261 262
	return pol;
}
#endif

L
Linus Torvalds 已提交
263 264
static int shm_mmap(struct file * file, struct vm_area_struct * vma)
{
265
	struct shm_file_data *sfd = shm_file_data(file);
266 267
	int ret;

268 269 270 271
	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
272
#ifdef CONFIG_MMU
273
	BUG_ON(!sfd->vm_ops->fault);
D
David Howells 已提交
274
#endif
275 276
	vma->vm_ops = &shm_vm_ops;
	shm_open(vma);
277 278

	return ret;
L
Linus Torvalds 已提交
279 280
}

K
Kirill Korotaev 已提交
281 282
static int shm_release(struct inode *ino, struct file *file)
{
283
	struct shm_file_data *sfd = shm_file_data(file);
K
Kirill Korotaev 已提交
284

285 286 287
	put_ipc_ns(sfd->ns);
	shm_file_data(file) = NULL;
	kfree(sfd);
K
Kirill Korotaev 已提交
288 289 290
	return 0;
}

291 292 293 294 295 296 297 298 299 300 301 302
static int shm_fsync(struct file *file, struct dentry *dentry, int datasync)
{
	int (*fsync) (struct file *, struct dentry *, int datasync);
	struct shm_file_data *sfd = shm_file_data(file);
	int ret = -EINVAL;

	fsync = sfd->file->f_op->fsync;
	if (fsync)
		ret = fsync(sfd->file, sfd->file->f_path.dentry, datasync);
	return ret;
}

303 304 305 306 307
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);
308 309 310 311 312 313 314 315 316 317 318 319 320
	return get_unmapped_area(sfd->file, addr, len, pgoff, flags);
}

int is_file_shm_hugepages(struct file *file)
{
	int ret = 0;

	if (file->f_op == &shm_file_operations) {
		struct shm_file_data *sfd;
		sfd = shm_file_data(file);
		ret = is_file_hugepages(sfd->file);
	}
	return ret;
321 322
}

323
static const struct file_operations shm_file_operations = {
K
Kirill Korotaev 已提交
324
	.mmap		= shm_mmap,
325
	.fsync		= shm_fsync,
K
Kirill Korotaev 已提交
326
	.release	= shm_release,
327
	.get_unmapped_area	= shm_get_unmapped_area,
L
Linus Torvalds 已提交
328 329 330 331 332
};

static struct vm_operations_struct shm_vm_ops = {
	.open	= shm_open,	/* callback for a new vm-area open */
	.close	= shm_close,	/* callback for when the vm-area is released */
333
	.fault	= shm_fault,
334 335 336
#if defined(CONFIG_NUMA)
	.set_policy = shm_set_policy,
	.get_policy = shm_get_policy,
L
Linus Torvalds 已提交
337 338 339
#endif
};

K
Kirill Korotaev 已提交
340
static int newseg (struct ipc_namespace *ns, key_t key, int shmflg, size_t size)
L
Linus Torvalds 已提交
341 342 343 344 345 346 347 348
{
	int error;
	struct shmid_kernel *shp;
	int numpages = (size + PAGE_SIZE -1) >> PAGE_SHIFT;
	struct file * file;
	char name[13];
	int id;

K
Kirill Korotaev 已提交
349
	if (size < SHMMIN || size > ns->shm_ctlmax)
L
Linus Torvalds 已提交
350 351
		return -EINVAL;

352
	if (ns->shm_tot + numpages > ns->shm_ctlall)
L
Linus Torvalds 已提交
353 354 355 356 357 358 359
		return -ENOSPC;

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

	shp->shm_perm.key = key;
A
Andrew Morton 已提交
360
	shp->shm_perm.mode = (shmflg & S_IRWXUGO);
L
Linus Torvalds 已提交
361 362 363 364 365 366 367 368 369
	shp->mlock_user = NULL;

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

370
	sprintf (name, "SYSV%08x", key);
L
Linus Torvalds 已提交
371
	if (shmflg & SHM_HUGETLB) {
372 373
		/* hugetlb_file_setup takes care of mlock user accounting */
		file = hugetlb_file_setup(name, size);
L
Linus Torvalds 已提交
374 375
		shp->mlock_user = current->user;
	} else {
376 377 378 379 380 381 382 383 384
		int acctflag = VM_ACCOUNT;
		/*
		 * Do not allow no accounting for OVERCOMMIT_NEVER, even
	 	 * if it's asked for.
		 */
		if  ((shmflg & SHM_NORESERVE) &&
				sysctl_overcommit_memory != OVERCOMMIT_NEVER)
			acctflag = 0;
		file = shmem_file_setup(name, size, acctflag);
L
Linus Torvalds 已提交
385 386 387 388 389 390
	}
	error = PTR_ERR(file);
	if (IS_ERR(file))
		goto no_file;

	error = -ENOSPC;
K
Kirill Korotaev 已提交
391
	id = shm_addid(ns, shp);
L
Linus Torvalds 已提交
392 393 394 395 396 397 398 399 400
	if(id == -1) 
		goto no_id;

	shp->shm_cprid = current->tgid;
	shp->shm_lprid = 0;
	shp->shm_atim = shp->shm_dtim = 0;
	shp->shm_ctim = get_seconds();
	shp->shm_segsz = size;
	shp->shm_nattch = 0;
K
Kirill Korotaev 已提交
401
	shp->id = shm_buildid(ns, id, shp->shm_perm.seq);
L
Linus Torvalds 已提交
402
	shp->shm_file = file;
403 404 405 406 407
	/*
	 * shmid gets reported as "inode#" in /proc/pid/maps.
	 * proc-ps tools use this. Changing this will break them.
	 */
	file->f_dentry->d_inode->i_ino = shp->id;
408

K
Kirill Korotaev 已提交
409
	ns->shm_tot += numpages;
L
Linus Torvalds 已提交
410 411 412 413 414 415 416 417 418 419 420 421 422 423 424
	shm_unlock(shp);
	return shp->id;

no_id:
	fput(file);
no_file:
	security_shm_free(shp);
	ipc_rcu_putref(shp);
	return error;
}

asmlinkage long sys_shmget (key_t key, size_t size, int shmflg)
{
	struct shmid_kernel *shp;
	int err, id = 0;
K
Kirill Korotaev 已提交
425 426 427
	struct ipc_namespace *ns;

	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
428

K
Kirill Korotaev 已提交
429
	mutex_lock(&shm_ids(ns).mutex);
L
Linus Torvalds 已提交
430
	if (key == IPC_PRIVATE) {
K
Kirill Korotaev 已提交
431 432
		err = newseg(ns, key, shmflg, size);
	} else if ((id = ipc_findkey(&shm_ids(ns), key)) == -1) {
L
Linus Torvalds 已提交
433 434 435
		if (!(shmflg & IPC_CREAT))
			err = -ENOENT;
		else
K
Kirill Korotaev 已提交
436
			err = newseg(ns, key, shmflg, size);
L
Linus Torvalds 已提交
437 438 439
	} else if ((shmflg & IPC_CREAT) && (shmflg & IPC_EXCL)) {
		err = -EEXIST;
	} else {
K
Kirill Korotaev 已提交
440
		shp = shm_lock(ns, id);
441
		BUG_ON(shp==NULL);
L
Linus Torvalds 已提交
442 443 444 445 446
		if (shp->shm_segsz < size)
			err = -EINVAL;
		else if (ipcperms(&shp->shm_perm, shmflg))
			err = -EACCES;
		else {
K
Kirill Korotaev 已提交
447
			int shmid = shm_buildid(ns, id, shp->shm_perm.seq);
L
Linus Torvalds 已提交
448 449 450 451 452 453
			err = security_shm_associate(shp, shmflg);
			if (!err)
				err = shmid;
		}
		shm_unlock(shp);
	}
K
Kirill Korotaev 已提交
454
	mutex_unlock(&shm_ids(ns).mutex);
L
Linus Torvalds 已提交
455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501

	return err;
}

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;

		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;
	}
}

struct shm_setbuf {
	uid_t	uid;
	gid_t	gid;
	mode_t	mode;
};	

static inline unsigned long copy_shmid_from_user(struct shm_setbuf *out, void __user *buf, int version)
{
	switch(version) {
	case IPC_64:
	    {
		struct shmid64_ds tbuf;

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

		out->uid	= tbuf.shm_perm.uid;
		out->gid	= tbuf.shm_perm.gid;
A
Andrew Morton 已提交
502
		out->mode	= tbuf.shm_perm.mode;
L
Linus Torvalds 已提交
503 504 505 506 507 508 509 510 511 512 513 514

		return 0;
	    }
	case IPC_OLD:
	    {
		struct shmid_ds tbuf_old;

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

		out->uid	= tbuf_old.shm_perm.uid;
		out->gid	= tbuf_old.shm_perm.gid;
A
Andrew Morton 已提交
515
		out->mode	= tbuf_old.shm_perm.mode;
L
Linus Torvalds 已提交
516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549

		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;
	}
}

K
Kirill Korotaev 已提交
550 551
static void shm_get_stat(struct ipc_namespace *ns, unsigned long *rss,
		unsigned long *swp)
L
Linus Torvalds 已提交
552 553 554 555 556 557
{
	int i;

	*rss = 0;
	*swp = 0;

K
Kirill Korotaev 已提交
558
	for (i = 0; i <= shm_ids(ns).max_id; i++) {
L
Linus Torvalds 已提交
559 560 561
		struct shmid_kernel *shp;
		struct inode *inode;

K
Kirill Korotaev 已提交
562
		shp = shm_get(ns, i);
L
Linus Torvalds 已提交
563 564 565
		if(!shp)
			continue;

J
Josef Sipek 已提交
566
		inode = shp->shm_file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585

		if (is_file_hugepages(shp->shm_file)) {
			struct address_space *mapping = inode->i_mapping;
			*rss += (HPAGE_SIZE/PAGE_SIZE)*mapping->nrpages;
		} else {
			struct shmem_inode_info *info = SHMEM_I(inode);
			spin_lock(&info->lock);
			*rss += inode->i_mapping->nrpages;
			*swp += info->swapped;
			spin_unlock(&info->lock);
		}
	}
}

asmlinkage long sys_shmctl (int shmid, int cmd, struct shmid_ds __user *buf)
{
	struct shm_setbuf setbuf;
	struct shmid_kernel *shp;
	int err, version;
K
Kirill Korotaev 已提交
586
	struct ipc_namespace *ns;
L
Linus Torvalds 已提交
587 588 589 590 591 592 593

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

	version = ipc_parse_version(&cmd);
K
Kirill Korotaev 已提交
594
	ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
595 596 597 598 599 600 601 602 603 604 605

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

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

		memset(&shminfo,0,sizeof(shminfo));
K
Kirill Korotaev 已提交
606 607 608
		shminfo.shmmni = shminfo.shmseg = ns->shm_ctlmni;
		shminfo.shmmax = ns->shm_ctlmax;
		shminfo.shmall = ns->shm_ctlall;
L
Linus Torvalds 已提交
609 610 611 612 613

		shminfo.shmmin = SHMMIN;
		if(copy_shminfo_to_user (buf, &shminfo, version))
			return -EFAULT;
		/* reading a integer is always atomic */
K
Kirill Korotaev 已提交
614
		err= shm_ids(ns).max_id;
L
Linus Torvalds 已提交
615 616 617 618 619 620 621 622 623 624 625 626 627
		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;

		memset(&shm_info,0,sizeof(shm_info));
K
Kirill Korotaev 已提交
628 629 630 631
		mutex_lock(&shm_ids(ns).mutex);
		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
632 633
		shm_info.swap_attempts = 0;
		shm_info.swap_successes = 0;
K
Kirill Korotaev 已提交
634 635
		err = shm_ids(ns).max_id;
		mutex_unlock(&shm_ids(ns).mutex);
L
Linus Torvalds 已提交
636 637 638 639 640 641 642 643 644 645 646 647 648 649
		if(copy_to_user (buf, &shm_info, sizeof(shm_info))) {
			err = -EFAULT;
			goto out;
		}

		err = err < 0 ? 0 : err;
		goto out;
	}
	case SHM_STAT:
	case IPC_STAT:
	{
		struct shmid64_ds tbuf;
		int result;
		memset(&tbuf, 0, sizeof(tbuf));
K
Kirill Korotaev 已提交
650
		shp = shm_lock(ns, shmid);
L
Linus Torvalds 已提交
651 652 653 654 655
		if(shp==NULL) {
			err = -EINVAL;
			goto out;
		} else if(cmd==SHM_STAT) {
			err = -EINVAL;
K
Kirill Korotaev 已提交
656
			if (shmid > shm_ids(ns).max_id)
L
Linus Torvalds 已提交
657
				goto out_unlock;
K
Kirill Korotaev 已提交
658
			result = shm_buildid(ns, shmid, shp->shm_perm.seq);
L
Linus Torvalds 已提交
659
		} else {
K
Kirill Korotaev 已提交
660
			err = shm_checkid(ns, shp,shmid);
L
Linus Torvalds 已提交
661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677
			if(err)
				goto out_unlock;
			result = 0;
		}
		err=-EACCES;
		if (ipcperms (&shp->shm_perm, S_IRUGO))
			goto out_unlock;
		err = security_shm_shmctl(shp, cmd);
		if (err)
			goto out_unlock;
		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;
678
		tbuf.shm_nattch	= shp->shm_nattch;
L
Linus Torvalds 已提交
679 680 681 682 683 684 685 686 687 688
		shm_unlock(shp);
		if(copy_shmid_to_user (buf, &tbuf, version))
			err = -EFAULT;
		else
			err = result;
		goto out;
	}
	case SHM_LOCK:
	case SHM_UNLOCK:
	{
K
Kirill Korotaev 已提交
689
		shp = shm_lock(ns, shmid);
L
Linus Torvalds 已提交
690 691 692 693
		if(shp==NULL) {
			err = -EINVAL;
			goto out;
		}
K
Kirill Korotaev 已提交
694
		err = shm_checkid(ns, shp,shmid);
L
Linus Torvalds 已提交
695 696 697
		if(err)
			goto out_unlock;

S
Steve Grubb 已提交
698 699 700 701
		err = audit_ipc_obj(&(shp->shm_perm));
		if (err)
			goto out_unlock;

L
Linus Torvalds 已提交
702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719
		if (!capable(CAP_IPC_LOCK)) {
			err = -EPERM;
			if (current->euid != shp->shm_perm.uid &&
			    current->euid != shp->shm_perm.cuid)
				goto out_unlock;
			if (cmd == SHM_LOCK &&
			    !current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur)
				goto out_unlock;
		}

		err = security_shm_shmctl(shp, cmd);
		if (err)
			goto out_unlock;
		
		if(cmd==SHM_LOCK) {
			struct user_struct * user = current->user;
			if (!is_file_hugepages(shp->shm_file)) {
				err = shmem_lock(shp->shm_file, 1, user);
720
				if (!err && !(shp->shm_perm.mode & SHM_LOCKED)){
A
Andrew Morton 已提交
721
					shp->shm_perm.mode |= SHM_LOCKED;
L
Linus Torvalds 已提交
722 723 724 725 726
					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
727
			shp->shm_perm.mode &= ~SHM_LOCKED;
L
Linus Torvalds 已提交
728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744
			shp->mlock_user = NULL;
		}
		shm_unlock(shp);
		goto out;
	}
	case IPC_RMID:
	{
		/*
		 *	We cannot simply remove the file. The SVID states
		 *	that the block remains until the last person
		 *	detaches from it, then is deleted. A shmat() on
		 *	an RMID segment is legal in older Linux and if 
		 *	we change it apps break...
		 *
		 *	Instead we set a destroyed flag, and then blow
		 *	the name away when the usage hits zero.
		 */
K
Kirill Korotaev 已提交
745 746
		mutex_lock(&shm_ids(ns).mutex);
		shp = shm_lock(ns, shmid);
L
Linus Torvalds 已提交
747 748 749
		err = -EINVAL;
		if (shp == NULL) 
			goto out_up;
K
Kirill Korotaev 已提交
750
		err = shm_checkid(ns, shp, shmid);
L
Linus Torvalds 已提交
751 752 753
		if(err)
			goto out_unlock_up;

S
Steve Grubb 已提交
754 755 756 757
		err = audit_ipc_obj(&(shp->shm_perm));
		if (err)
			goto out_unlock_up;

L
Linus Torvalds 已提交
758 759 760 761 762 763 764 765 766 767 768
		if (current->euid != shp->shm_perm.uid &&
		    current->euid != shp->shm_perm.cuid && 
		    !capable(CAP_SYS_ADMIN)) {
			err=-EPERM;
			goto out_unlock_up;
		}

		err = security_shm_shmctl(shp, cmd);
		if (err)
			goto out_unlock_up;

K
Kirill Korotaev 已提交
769 770
		do_shm_rmid(ns, shp);
		mutex_unlock(&shm_ids(ns).mutex);
L
Linus Torvalds 已提交
771 772 773 774 775 776 777 778 779
		goto out;
	}

	case IPC_SET:
	{
		if (copy_shmid_from_user (&setbuf, buf, version)) {
			err = -EFAULT;
			goto out;
		}
K
Kirill Korotaev 已提交
780 781
		mutex_lock(&shm_ids(ns).mutex);
		shp = shm_lock(ns, shmid);
L
Linus Torvalds 已提交
782 783 784
		err=-EINVAL;
		if(shp==NULL)
			goto out_up;
K
Kirill Korotaev 已提交
785
		err = shm_checkid(ns, shp,shmid);
L
Linus Torvalds 已提交
786 787
		if(err)
			goto out_unlock_up;
S
Steve Grubb 已提交
788 789 790
		err = audit_ipc_obj(&(shp->shm_perm));
		if (err)
			goto out_unlock_up;
791
		err = audit_ipc_set_perm(0, setbuf.uid, setbuf.gid, setbuf.mode);
S
Steve Grubb 已提交
792 793
		if (err)
			goto out_unlock_up;
L
Linus Torvalds 已提交
794 795 796 797 798 799 800 801 802 803 804 805 806
		err=-EPERM;
		if (current->euid != shp->shm_perm.uid &&
		    current->euid != shp->shm_perm.cuid && 
		    !capable(CAP_SYS_ADMIN)) {
			goto out_unlock_up;
		}

		err = security_shm_shmctl(shp, cmd);
		if (err)
			goto out_unlock_up;
		
		shp->shm_perm.uid = setbuf.uid;
		shp->shm_perm.gid = setbuf.gid;
A
Andrew Morton 已提交
807
		shp->shm_perm.mode = (shp->shm_perm.mode & ~S_IRWXUGO)
L
Linus Torvalds 已提交
808 809 810 811 812 813 814 815 816 817 818 819 820 821
			| (setbuf.mode & S_IRWXUGO);
		shp->shm_ctim = get_seconds();
		break;
	}

	default:
		err = -EINVAL;
		goto out;
	}

	err = 0;
out_unlock_up:
	shm_unlock(shp);
out_up:
K
Kirill Korotaev 已提交
822
	mutex_unlock(&shm_ids(ns).mutex);
L
Linus Torvalds 已提交
823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846
	goto out;
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;
847
	unsigned long user_addr;
K
Kirill Korotaev 已提交
848
	struct ipc_namespace *ns;
849 850 851
	struct shm_file_data *sfd;
	struct path path;
	mode_t f_mode;
L
Linus Torvalds 已提交
852

853 854
	err = -EINVAL;
	if (shmid < 0)
L
Linus Torvalds 已提交
855
		goto out;
856
	else if ((addr = (ulong)shmaddr)) {
L
Linus Torvalds 已提交
857 858 859 860 861 862 863
		if (addr & (SHMLBA-1)) {
			if (shmflg & SHM_RND)
				addr &= ~(SHMLBA-1);	   /* round down */
			else
#ifndef __ARCH_FORCE_SHMLBA
				if (addr & ~PAGE_MASK)
#endif
864
					goto out;
L
Linus Torvalds 已提交
865 866 867 868
		}
		flags = MAP_SHARED | MAP_FIXED;
	} else {
		if ((shmflg & SHM_REMAP))
869
			goto out;
L
Linus Torvalds 已提交
870 871 872 873 874 875 876

		flags = MAP_SHARED;
	}

	if (shmflg & SHM_RDONLY) {
		prot = PROT_READ;
		acc_mode = S_IRUGO;
877
		f_mode = FMODE_READ;
L
Linus Torvalds 已提交
878 879 880
	} else {
		prot = PROT_READ | PROT_WRITE;
		acc_mode = S_IRUGO | S_IWUGO;
881
		f_mode = FMODE_READ | FMODE_WRITE;
L
Linus Torvalds 已提交
882 883 884 885 886 887 888 889 890 891
	}
	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
892 893
	ns = current->nsproxy->ipc_ns;
	shp = shm_lock(ns, shmid);
894
	if(shp == NULL)
L
Linus Torvalds 已提交
895
		goto out;
896

K
Kirill Korotaev 已提交
897
	err = shm_checkid(ns, shp,shmid);
898 899 900 901 902 903
	if (err)
		goto out_unlock;

	err = -EACCES;
	if (ipcperms(&shp->shm_perm, acc_mode))
		goto out_unlock;
L
Linus Torvalds 已提交
904 905

	err = security_shm_shmat(shp, shmaddr, shmflg);
906 907 908 909
	if (err)
		goto out_unlock;

	path.dentry = dget(shp->shm_file->f_path.dentry);
910
	path.mnt    = shp->shm_file->f_path.mnt;
L
Linus Torvalds 已提交
911
	shp->shm_nattch++;
912
	size = i_size_read(path.dentry->d_inode);
L
Linus Torvalds 已提交
913 914
	shm_unlock(shp);

915 916 917
	err = -ENOMEM;
	sfd = kzalloc(sizeof(*sfd), GFP_KERNEL);
	if (!sfd)
918
		goto out_put_dentry;
919 920

	err = -ENOMEM;
921 922

	file = alloc_file(path.mnt, path.dentry, f_mode, &shm_file_operations);
923 924 925 926 927 928 929 930 931 932
	if (!file)
		goto out_free;

	file->private_data = sfd;
	file->f_mapping = shp->shm_file->f_mapping;
	sfd->id = shp->id;
	sfd->ns = get_ipc_ns(ns);
	sfd->file = shp->shm_file;
	sfd->vm_ops = NULL;

L
Linus Torvalds 已提交
933 934
	down_write(&current->mm->mmap_sem);
	if (addr && !(shmflg & SHM_REMAP)) {
935
		err = -EINVAL;
L
Linus Torvalds 已提交
936 937 938 939 940 941 942 943 944 945 946
		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;
	}
		
947 948 949 950 951
	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
952 953 954
invalid:
	up_write(&current->mm->mmap_sem);

955 956 957
	fput(file);

out_nattch:
K
Kirill Korotaev 已提交
958 959
	mutex_lock(&shm_ids(ns).mutex);
	shp = shm_lock(ns, shmid);
960
	BUG_ON(!shp);
L
Linus Torvalds 已提交
961 962
	shp->shm_nattch--;
	if(shp->shm_nattch == 0 &&
A
Andrew Morton 已提交
963
	   shp->shm_perm.mode & SHM_DEST)
K
Kirill Korotaev 已提交
964
		shm_destroy(ns, shp);
L
Linus Torvalds 已提交
965 966
	else
		shm_unlock(shp);
K
Kirill Korotaev 已提交
967
	mutex_unlock(&shm_ids(ns).mutex);
L
Linus Torvalds 已提交
968 969 970

out:
	return err;
971 972 973 974 975 976 977

out_unlock:
	shm_unlock(shp);
	goto out;

out_free:
	kfree(sfd);
978
out_put_dentry:
979 980
	dput(path.dentry);
	goto out_nattch;
L
Linus Torvalds 已提交
981 982
}

S
Stephen Rothwell 已提交
983 984 985 986 987 988 989 990 991 992 993 994
asmlinkage long sys_shmat(int shmid, char __user *shmaddr, int shmflg)
{
	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
995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006
/*
 * detach and kill segment if marked destroyed.
 * The work is done in shm_close.
 */
asmlinkage long sys_shmdt(char __user *shmaddr)
{
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma, *next;
	unsigned long addr = (unsigned long)shmaddr;
	loff_t size = 0;
	int retval = -EINVAL;

1007 1008 1009
	if (addr & ~PAGE_MASK)
		return retval;

L
Linus Torvalds 已提交
1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
	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);

	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.
		 */
1042
		if ((vma->vm_ops == &shm_vm_ops) &&
L
Linus Torvalds 已提交
1043 1044 1045
			(vma->vm_start - addr)/PAGE_SIZE == vma->vm_pgoff) {


J
Josef Sipek 已提交
1046
			size = vma->vm_file->f_path.dentry->d_inode->i_size;
L
Linus Torvalds 已提交
1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065
			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
	 * prevent overflows and make comparisions vs. equal-width types.
	 */
1066
	size = PAGE_ALIGN(size);
L
Linus Torvalds 已提交
1067 1068 1069 1070
	while (vma && (loff_t)(vma->vm_end - addr) <= size) {
		next = vma->vm_next;

		/* finding a matching vma now does not alter retval */
1071
		if ((vma->vm_ops == &shm_vm_ops) &&
L
Linus Torvalds 已提交
1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082
			(vma->vm_start - addr)/PAGE_SIZE == vma->vm_pgoff)

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

	up_write(&mm->mmap_sem);
	return retval;
}

#ifdef CONFIG_PROC_FS
1083
static int sysvipc_shm_proc_show(struct seq_file *s, void *it)
L
Linus Torvalds 已提交
1084
{
1085 1086
	struct shmid_kernel *shp = it;
	char *format;
L
Linus Torvalds 已提交
1087 1088 1089 1090

#define SMALL_STRING "%10d %10d  %4o %10u %5u %5u  %5d %5u %5u %5u %5u %10lu %10lu %10lu\n"
#define BIG_STRING   "%10d %10d  %4o %21u %5u %5u  %5d %5u %5u %5u %5u %10lu %10lu %10lu\n"

1091 1092 1093 1094 1095 1096 1097
	if (sizeof(size_t) <= sizeof(int))
		format = SMALL_STRING;
	else
		format = BIG_STRING;
	return seq_printf(s, format,
			  shp->shm_perm.key,
			  shp->id,
A
Andrew Morton 已提交
1098
			  shp->shm_perm.mode,
1099 1100 1101
			  shp->shm_segsz,
			  shp->shm_cprid,
			  shp->shm_lprid,
1102
			  shp->shm_nattch,
1103 1104 1105 1106 1107 1108 1109
			  shp->shm_perm.uid,
			  shp->shm_perm.gid,
			  shp->shm_perm.cuid,
			  shp->shm_perm.cgid,
			  shp->shm_atim,
			  shp->shm_dtim,
			  shp->shm_ctim);
L
Linus Torvalds 已提交
1110 1111
}
#endif