mqueue.c 35.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4
/*
 * POSIX message queues filesystem for Linux.
 *
 * Copyright (C) 2003,2004  Krzysztof Benedyczak    (golbi@mat.uni.torun.pl)
5
 *                          Michal Wronski          (michal.wronski@gmail.com)
L
Linus Torvalds 已提交
6 7 8 9 10
 *
 * Spinlocks:               Mohamed Abbas           (abbas.mohamed@intel.com)
 * Lockless receive & send, fd based notify:
 * 			    Manfred Spraul	    (manfred@colorfullife.com)
 *
11 12
 * Audit:                   George Wilson           (ltcgcw@us.ibm.com)
 *
L
Linus Torvalds 已提交
13 14 15
 * This file is released under the GPL.
 */

16
#include <linux/capability.h>
L
Linus Torvalds 已提交
17 18 19 20 21 22 23 24 25 26
#include <linux/init.h>
#include <linux/pagemap.h>
#include <linux/file.h>
#include <linux/mount.h>
#include <linux/namei.h>
#include <linux/sysctl.h>
#include <linux/poll.h>
#include <linux/mqueue.h>
#include <linux/msg.h>
#include <linux/skbuff.h>
27
#include <linux/vmalloc.h>
L
Linus Torvalds 已提交
28 29
#include <linux/netlink.h>
#include <linux/syscalls.h>
30
#include <linux/audit.h>
31
#include <linux/signal.h>
I
Ingo Molnar 已提交
32
#include <linux/mutex.h>
33 34
#include <linux/nsproxy.h>
#include <linux/pid.h>
35
#include <linux/ipc_namespace.h>
36
#include <linux/user_namespace.h>
37
#include <linux/slab.h>
I
Ingo Molnar 已提交
38

L
Linus Torvalds 已提交
39 40 41 42 43 44 45 46 47 48 49 50 51 52
#include <net/sock.h>
#include "util.h"

#define MQUEUE_MAGIC	0x19800202
#define DIRENT_SIZE	20
#define FILENT_SIZE	80

#define SEND		0
#define RECV		1

#define STATE_NONE	0
#define STATE_PENDING	1
#define STATE_READY	2

53 54 55 56 57 58
struct posix_msg_tree_node {
	struct rb_node		rb_node;
	struct list_head	msg_list;
	int			priority;
};

L
Linus Torvalds 已提交
59 60 61 62 63 64 65 66 67 68 69 70
struct ext_wait_queue {		/* queue of sleeping tasks */
	struct task_struct *task;
	struct list_head list;
	struct msg_msg *msg;	/* ptr of loaded message */
	int state;		/* one of STATE_* values */
};

struct mqueue_inode_info {
	spinlock_t lock;
	struct inode vfs_inode;
	wait_queue_head_t wait_q;

71
	struct rb_root msg_tree;
72
	struct posix_msg_tree_node *node_cache;
L
Linus Torvalds 已提交
73 74 75
	struct mq_attr attr;

	struct sigevent notify;
76
	struct pid* notify_owner;
77
	struct user_namespace *notify_user_ns;
78
	struct user_struct *user;	/* user who created, for accounting */
L
Linus Torvalds 已提交
79 80 81 82 83 84 85 86 87
	struct sock *notify_sock;
	struct sk_buff *notify_cookie;

	/* for tasks waiting for free space and messages, respectively */
	struct ext_wait_queue e_wait_q[2];

	unsigned long qsize; /* size of queue in memory (sum of all msgs) */
};

88
static const struct inode_operations mqueue_dir_inode_operations;
89
static const struct file_operations mqueue_file_operations;
90
static const struct super_operations mqueue_super_ops;
L
Linus Torvalds 已提交
91 92
static void remove_notification(struct mqueue_inode_info *info);

93
static struct kmem_cache *mqueue_inode_cachep;
L
Linus Torvalds 已提交
94 95 96 97 98 99 100 101

static struct ctl_table_header * mq_sysctl_table;

static inline struct mqueue_inode_info *MQUEUE_I(struct inode *inode)
{
	return container_of(inode, struct mqueue_inode_info, vfs_inode);
}

102 103 104 105
/*
 * This routine should be called with the mq_lock held.
 */
static inline struct ipc_namespace *__get_ns_from_inode(struct inode *inode)
106
{
107
	return get_ipc_ns(inode->i_sb->s_fs_info);
108 109
}

110
static struct ipc_namespace *get_ns_from_inode(struct inode *inode)
111
{
112 113 114 115 116 117
	struct ipc_namespace *ns;

	spin_lock(&mq_lock);
	ns = __get_ns_from_inode(inode);
	spin_unlock(&mq_lock);
	return ns;
118 119
}

120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137
/* Auxiliary functions to manipulate messages' list */
static int msg_insert(struct msg_msg *msg, struct mqueue_inode_info *info)
{
	struct rb_node **p, *parent = NULL;
	struct posix_msg_tree_node *leaf;

	p = &info->msg_tree.rb_node;
	while (*p) {
		parent = *p;
		leaf = rb_entry(parent, struct posix_msg_tree_node, rb_node);

		if (likely(leaf->priority == msg->m_type))
			goto insert_msg;
		else if (msg->m_type < leaf->priority)
			p = &(*p)->rb_left;
		else
			p = &(*p)->rb_right;
	}
138 139 140 141 142 143 144 145 146 147 148
	if (info->node_cache) {
		leaf = info->node_cache;
		info->node_cache = NULL;
	} else {
		leaf = kmalloc(sizeof(*leaf), GFP_ATOMIC);
		if (!leaf)
			return -ENOMEM;
		rb_init_node(&leaf->rb_node);
		INIT_LIST_HEAD(&leaf->msg_list);
		info->qsize += sizeof(*leaf);
	}
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185
	leaf->priority = msg->m_type;
	rb_link_node(&leaf->rb_node, parent, p);
	rb_insert_color(&leaf->rb_node, &info->msg_tree);
insert_msg:
	info->attr.mq_curmsgs++;
	info->qsize += msg->m_ts;
	list_add_tail(&msg->m_list, &leaf->msg_list);
	return 0;
}

static inline struct msg_msg *msg_get(struct mqueue_inode_info *info)
{
	struct rb_node **p, *parent = NULL;
	struct posix_msg_tree_node *leaf;
	struct msg_msg *msg;

try_again:
	p = &info->msg_tree.rb_node;
	while (*p) {
		parent = *p;
		/*
		 * During insert, low priorities go to the left and high to the
		 * right.  On receive, we want the highest priorities first, so
		 * walk all the way to the right.
		 */
		p = &(*p)->rb_right;
	}
	if (!parent) {
		if (info->attr.mq_curmsgs) {
			pr_warn_once("Inconsistency in POSIX message queue, "
				     "no tree element, but supposedly messages "
				     "should exist!\n");
			info->attr.mq_curmsgs = 0;
		}
		return NULL;
	}
	leaf = rb_entry(parent, struct posix_msg_tree_node, rb_node);
186
	if (unlikely(list_empty(&leaf->msg_list))) {
187 188 189 190
		pr_warn_once("Inconsistency in POSIX message queue, "
			     "empty leaf node but we haven't implemented "
			     "lazy leaf delete!\n");
		rb_erase(&leaf->rb_node, &info->msg_tree);
191 192 193 194 195 196
		if (info->node_cache) {
			info->qsize -= sizeof(*leaf);
			kfree(leaf);
		} else {
			info->node_cache = leaf;
		}
197 198 199 200 201 202 203
		goto try_again;
	} else {
		msg = list_first_entry(&leaf->msg_list,
				       struct msg_msg, m_list);
		list_del(&msg->m_list);
		if (list_empty(&leaf->msg_list)) {
			rb_erase(&leaf->rb_node, &info->msg_tree);
204 205 206 207 208 209
			if (info->node_cache) {
				info->qsize -= sizeof(*leaf);
				kfree(leaf);
			} else {
				info->node_cache = leaf;
			}
210 211 212 213 214 215 216
		}
	}
	info->attr.mq_curmsgs--;
	info->qsize -= msg->m_ts;
	return msg;
}

217
static struct inode *mqueue_get_inode(struct super_block *sb,
A
Al Viro 已提交
218
		struct ipc_namespace *ipc_ns, umode_t mode,
219
		struct mq_attr *attr)
L
Linus Torvalds 已提交
220
{
221
	struct user_struct *u = current_user();
L
Linus Torvalds 已提交
222
	struct inode *inode;
223
	int ret = -ENOMEM;
L
Linus Torvalds 已提交
224 225

	inode = new_inode(sb);
226 227 228 229 230 231 232 233 234 235 236
	if (!inode)
		goto err;

	inode->i_ino = get_next_ino();
	inode->i_mode = mode;
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
	inode->i_mtime = inode->i_ctime = inode->i_atime = CURRENT_TIME;

	if (S_ISREG(mode)) {
		struct mqueue_inode_info *info;
237
		unsigned long mq_bytes, mq_treesize;
238 239 240 241 242 243 244 245 246 247

		inode->i_fop = &mqueue_file_operations;
		inode->i_size = FILENT_SIZE;
		/* mqueue specific info */
		info = MQUEUE_I(inode);
		spin_lock_init(&info->lock);
		init_waitqueue_head(&info->wait_q);
		INIT_LIST_HEAD(&info->e_wait_q[0].list);
		INIT_LIST_HEAD(&info->e_wait_q[1].list);
		info->notify_owner = NULL;
248
		info->notify_user_ns = NULL;
249 250
		info->qsize = 0;
		info->user = NULL;	/* set when all is ok */
251
		info->msg_tree = RB_ROOT;
252
		info->node_cache = NULL;
253
		memset(&info->attr, 0, sizeof(info->attr));
254 255 256 257
		info->attr.mq_maxmsg = min(ipc_ns->mq_msg_max,
					   ipc_ns->mq_msg_default);
		info->attr.mq_msgsize = min(ipc_ns->mq_msgsize_max,
					    ipc_ns->mq_msgsize_default);
258 259 260 261
		if (attr) {
			info->attr.mq_maxmsg = attr->mq_maxmsg;
			info->attr.mq_msgsize = attr->mq_msgsize;
		}
262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277
		/*
		 * We used to allocate a static array of pointers and account
		 * the size of that array as well as one msg_msg struct per
		 * possible message into the queue size. That's no longer
		 * accurate as the queue is now an rbtree and will grow and
		 * shrink depending on usage patterns.  We can, however, still
		 * account one msg_msg struct per message, but the nodes are
		 * allocated depending on priority usage, and most programs
		 * only use one, or a handful, of priorities.  However, since
		 * this is pinned memory, we need to assume worst case, so
		 * that means the min(mq_maxmsg, max_priorities) * struct
		 * posix_msg_tree_node.
		 */
		mq_treesize = info->attr.mq_maxmsg * sizeof(struct msg_msg) +
			min_t(unsigned int, info->attr.mq_maxmsg, MQ_PRIO_MAX) *
			sizeof(struct posix_msg_tree_node);
278

279 280
		mq_bytes = mq_treesize + (info->attr.mq_maxmsg *
					  info->attr.mq_msgsize);
L
Linus Torvalds 已提交
281

282 283
		spin_lock(&mq_lock);
		if (u->mq_bytes + mq_bytes < u->mq_bytes ||
284
		    u->mq_bytes + mq_bytes > rlimit(RLIMIT_MSGQUEUE)) {
285 286
			spin_unlock(&mq_lock);
			/* mqueue_evict_inode() releases info->messages */
287
			ret = -EMFILE;
288
			goto out_inode;
L
Linus Torvalds 已提交
289
		}
290 291 292 293 294 295 296 297 298 299 300
		u->mq_bytes += mq_bytes;
		spin_unlock(&mq_lock);

		/* all is ok */
		info->user = get_uid(u);
	} else if (S_ISDIR(mode)) {
		inc_nlink(inode);
		/* Some things misbehave if size == 0 on a directory */
		inode->i_size = 2 * DIRENT_SIZE;
		inode->i_op = &mqueue_dir_inode_operations;
		inode->i_fop = &simple_dir_operations;
L
Linus Torvalds 已提交
301
	}
302

L
Linus Torvalds 已提交
303 304 305
	return inode;
out_inode:
	iput(inode);
306
err:
307
	return ERR_PTR(ret);
L
Linus Torvalds 已提交
308 309 310 311 312
}

static int mqueue_fill_super(struct super_block *sb, void *data, int silent)
{
	struct inode *inode;
313
	struct ipc_namespace *ns = data;
L
Linus Torvalds 已提交
314 315 316 317 318 319

	sb->s_blocksize = PAGE_CACHE_SIZE;
	sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
	sb->s_magic = MQUEUE_MAGIC;
	sb->s_op = &mqueue_super_ops;

320 321 322
	inode = mqueue_get_inode(sb, ns, S_IFDIR | S_ISVTX | S_IRWXUGO, NULL);
	if (IS_ERR(inode))
		return PTR_ERR(inode);
L
Linus Torvalds 已提交
323

324 325 326 327
	sb->s_root = d_make_root(inode);
	if (!sb->s_root)
		return -ENOMEM;
	return 0;
L
Linus Torvalds 已提交
328 329
}

A
Al Viro 已提交
330
static struct dentry *mqueue_mount(struct file_system_type *fs_type,
331
			 int flags, const char *dev_name,
A
Al Viro 已提交
332
			 void *data)
L
Linus Torvalds 已提交
333
{
334 335
	if (!(flags & MS_KERNMOUNT))
		data = current->nsproxy->ipc_ns;
A
Al Viro 已提交
336
	return mount_ns(fs_type, flags, data, mqueue_fill_super);
L
Linus Torvalds 已提交
337 338
}

339
static void init_once(void *foo)
L
Linus Torvalds 已提交
340 341 342
{
	struct mqueue_inode_info *p = (struct mqueue_inode_info *) foo;

C
Christoph Lameter 已提交
343
	inode_init_once(&p->vfs_inode);
L
Linus Torvalds 已提交
344 345 346 347 348 349
}

static struct inode *mqueue_alloc_inode(struct super_block *sb)
{
	struct mqueue_inode_info *ei;

350
	ei = kmem_cache_alloc(mqueue_inode_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
351 352 353 354 355
	if (!ei)
		return NULL;
	return &ei->vfs_inode;
}

N
Nick Piggin 已提交
356
static void mqueue_i_callback(struct rcu_head *head)
L
Linus Torvalds 已提交
357
{
N
Nick Piggin 已提交
358
	struct inode *inode = container_of(head, struct inode, i_rcu);
L
Linus Torvalds 已提交
359 360 361
	kmem_cache_free(mqueue_inode_cachep, MQUEUE_I(inode));
}

N
Nick Piggin 已提交
362 363 364 365 366
static void mqueue_destroy_inode(struct inode *inode)
{
	call_rcu(&inode->i_rcu, mqueue_i_callback);
}

A
Al Viro 已提交
367
static void mqueue_evict_inode(struct inode *inode)
L
Linus Torvalds 已提交
368 369 370
{
	struct mqueue_inode_info *info;
	struct user_struct *user;
371
	unsigned long mq_bytes, mq_treesize;
372
	struct ipc_namespace *ipc_ns;
373
	struct msg_msg *msg;
L
Linus Torvalds 已提交
374

375
	clear_inode(inode);
A
Al Viro 已提交
376 377

	if (S_ISDIR(inode->i_mode))
L
Linus Torvalds 已提交
378
		return;
A
Al Viro 已提交
379

380
	ipc_ns = get_ns_from_inode(inode);
L
Linus Torvalds 已提交
381 382
	info = MQUEUE_I(inode);
	spin_lock(&info->lock);
383 384
	while ((msg = msg_get(info)) != NULL)
		free_msg(msg);
385
	kfree(info->node_cache);
L
Linus Torvalds 已提交
386 387
	spin_unlock(&info->lock);

388
	/* Total amount of bytes accounted for the mqueue */
389 390 391 392 393 394 395
	mq_treesize = info->attr.mq_maxmsg * sizeof(struct msg_msg) +
		min_t(unsigned int, info->attr.mq_maxmsg, MQ_PRIO_MAX) *
		sizeof(struct posix_msg_tree_node);

	mq_bytes = mq_treesize + (info->attr.mq_maxmsg *
				  info->attr.mq_msgsize);

L
Linus Torvalds 已提交
396 397 398 399
	user = info->user;
	if (user) {
		spin_lock(&mq_lock);
		user->mq_bytes -= mq_bytes;
400 401 402 403 404 405 406 407
		/*
		 * get_ns_from_inode() ensures that the
		 * (ipc_ns = sb->s_fs_info) is either a valid ipc_ns
		 * to which we now hold a reference, or it is NULL.
		 * We can't put it here under mq_lock, though.
		 */
		if (ipc_ns)
			ipc_ns->mq_queues_count--;
L
Linus Torvalds 已提交
408 409 410
		spin_unlock(&mq_lock);
		free_uid(user);
	}
411 412
	if (ipc_ns)
		put_ipc_ns(ipc_ns);
L
Linus Torvalds 已提交
413 414 415
}

static int mqueue_create(struct inode *dir, struct dentry *dentry,
A
Al Viro 已提交
416
				umode_t mode, struct nameidata *nd)
L
Linus Torvalds 已提交
417 418 419 420
{
	struct inode *inode;
	struct mq_attr *attr = dentry->d_fsdata;
	int error;
421
	struct ipc_namespace *ipc_ns;
L
Linus Torvalds 已提交
422 423

	spin_lock(&mq_lock);
424 425 426 427 428
	ipc_ns = __get_ns_from_inode(dir);
	if (!ipc_ns) {
		error = -EACCES;
		goto out_unlock;
	}
D
Doug Ledford 已提交
429 430 431
	if (ipc_ns->mq_queues_count >= HARD_QUEUESMAX ||
	    (ipc_ns->mq_queues_count >= ipc_ns->mq_queues_max &&
	     !capable(CAP_SYS_RESOURCE))) {
L
Linus Torvalds 已提交
432
		error = -ENOSPC;
433
		goto out_unlock;
L
Linus Torvalds 已提交
434
	}
435
	ipc_ns->mq_queues_count++;
L
Linus Torvalds 已提交
436 437
	spin_unlock(&mq_lock);

438
	inode = mqueue_get_inode(dir->i_sb, ipc_ns, mode, attr);
439 440
	if (IS_ERR(inode)) {
		error = PTR_ERR(inode);
L
Linus Torvalds 已提交
441
		spin_lock(&mq_lock);
442 443
		ipc_ns->mq_queues_count--;
		goto out_unlock;
L
Linus Torvalds 已提交
444 445
	}

446
	put_ipc_ns(ipc_ns);
L
Linus Torvalds 已提交
447 448 449 450 451 452
	dir->i_size += DIRENT_SIZE;
	dir->i_ctime = dir->i_mtime = dir->i_atime = CURRENT_TIME;

	d_instantiate(dentry, inode);
	dget(dentry);
	return 0;
453
out_unlock:
L
Linus Torvalds 已提交
454
	spin_unlock(&mq_lock);
455 456
	if (ipc_ns)
		put_ipc_ns(ipc_ns);
L
Linus Torvalds 已提交
457 458 459 460 461 462 463 464 465
	return error;
}

static int mqueue_unlink(struct inode *dir, struct dentry *dentry)
{
  	struct inode *inode = dentry->d_inode;

	dir->i_ctime = dir->i_mtime = dir->i_atime = CURRENT_TIME;
	dir->i_size -= DIRENT_SIZE;
466
  	drop_nlink(inode);
L
Linus Torvalds 已提交
467 468 469 470 471 472 473 474 475 476 477 478
  	dput(dentry);
  	return 0;
}

/*
*	This is routine for system read from queue file.
*	To avoid mess with doing here some sort of mq_receive we allow
*	to read only queue size & notification info (the only values
*	that are interesting from user point of view and aren't accessible
*	through std routines)
*/
static ssize_t mqueue_read_file(struct file *filp, char __user *u_data,
479
				size_t count, loff_t *off)
L
Linus Torvalds 已提交
480
{
J
Josef Sipek 已提交
481
	struct mqueue_inode_info *info = MQUEUE_I(filp->f_path.dentry->d_inode);
L
Linus Torvalds 已提交
482
	char buffer[FILENT_SIZE];
483
	ssize_t ret;
L
Linus Torvalds 已提交
484 485 486 487 488 489 490 491 492

	spin_lock(&info->lock);
	snprintf(buffer, sizeof(buffer),
			"QSIZE:%-10lu NOTIFY:%-5d SIGNO:%-5d NOTIFY_PID:%-6d\n",
			info->qsize,
			info->notify_owner ? info->notify.sigev_notify : 0,
			(info->notify_owner &&
			 info->notify.sigev_notify == SIGEV_SIGNAL) ?
				info->notify.sigev_signo : 0,
493
			pid_vnr(info->notify_owner));
L
Linus Torvalds 已提交
494 495 496
	spin_unlock(&info->lock);
	buffer[sizeof(buffer)-1] = '\0';

497 498 499 500
	ret = simple_read_from_buffer(u_data, count, off, buffer,
				strlen(buffer));
	if (ret <= 0)
		return ret;
L
Linus Torvalds 已提交
501

J
Josef Sipek 已提交
502
	filp->f_path.dentry->d_inode->i_atime = filp->f_path.dentry->d_inode->i_ctime = CURRENT_TIME;
503
	return ret;
L
Linus Torvalds 已提交
504 505
}

506
static int mqueue_flush_file(struct file *filp, fl_owner_t id)
L
Linus Torvalds 已提交
507
{
J
Josef Sipek 已提交
508
	struct mqueue_inode_info *info = MQUEUE_I(filp->f_path.dentry->d_inode);
L
Linus Torvalds 已提交
509 510

	spin_lock(&info->lock);
511
	if (task_tgid(current) == info->notify_owner)
L
Linus Torvalds 已提交
512 513 514 515 516 517 518 519
		remove_notification(info);

	spin_unlock(&info->lock);
	return 0;
}

static unsigned int mqueue_poll_file(struct file *filp, struct poll_table_struct *poll_tab)
{
J
Josef Sipek 已提交
520
	struct mqueue_inode_info *info = MQUEUE_I(filp->f_path.dentry->d_inode);
L
Linus Torvalds 已提交
521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558
	int retval = 0;

	poll_wait(filp, &info->wait_q, poll_tab);

	spin_lock(&info->lock);
	if (info->attr.mq_curmsgs)
		retval = POLLIN | POLLRDNORM;

	if (info->attr.mq_curmsgs < info->attr.mq_maxmsg)
		retval |= POLLOUT | POLLWRNORM;
	spin_unlock(&info->lock);

	return retval;
}

/* Adds current to info->e_wait_q[sr] before element with smaller prio */
static void wq_add(struct mqueue_inode_info *info, int sr,
			struct ext_wait_queue *ewp)
{
	struct ext_wait_queue *walk;

	ewp->task = current;

	list_for_each_entry(walk, &info->e_wait_q[sr].list, list) {
		if (walk->task->static_prio <= current->static_prio) {
			list_add_tail(&ewp->list, &walk->list);
			return;
		}
	}
	list_add_tail(&ewp->list, &info->e_wait_q[sr].list);
}

/*
 * Puts current task to sleep. Caller must hold queue lock. After return
 * lock isn't held.
 * sr: SEND or RECV
 */
static int wq_sleep(struct mqueue_inode_info *info, int sr,
559
		    ktime_t *timeout, struct ext_wait_queue *ewp)
L
Linus Torvalds 已提交
560 561 562 563 564 565 566 567 568 569
{
	int retval;
	signed long time;

	wq_add(info, sr, ewp);

	for (;;) {
		set_current_state(TASK_INTERRUPTIBLE);

		spin_unlock(&info->lock);
570 571
		time = schedule_hrtimeout_range_clock(timeout, 0,
			HRTIMER_MODE_ABS, CLOCK_REALTIME);
L
Linus Torvalds 已提交
572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643

		while (ewp->state == STATE_PENDING)
			cpu_relax();

		if (ewp->state == STATE_READY) {
			retval = 0;
			goto out;
		}
		spin_lock(&info->lock);
		if (ewp->state == STATE_READY) {
			retval = 0;
			goto out_unlock;
		}
		if (signal_pending(current)) {
			retval = -ERESTARTSYS;
			break;
		}
		if (time == 0) {
			retval = -ETIMEDOUT;
			break;
		}
	}
	list_del(&ewp->list);
out_unlock:
	spin_unlock(&info->lock);
out:
	return retval;
}

/*
 * Returns waiting task that should be serviced first or NULL if none exists
 */
static struct ext_wait_queue *wq_get_first_waiter(
		struct mqueue_inode_info *info, int sr)
{
	struct list_head *ptr;

	ptr = info->e_wait_q[sr].list.prev;
	if (ptr == &info->e_wait_q[sr].list)
		return NULL;
	return list_entry(ptr, struct ext_wait_queue, list);
}


static inline void set_cookie(struct sk_buff *skb, char code)
{
	((char*)skb->data)[NOTIFY_COOKIE_LEN-1] = code;
}

/*
 * The next function is only to split too long sys_mq_timedsend
 */
static void __do_notify(struct mqueue_inode_info *info)
{
	/* notification
	 * invoked when there is registered process and there isn't process
	 * waiting synchronously for message AND state of queue changed from
	 * empty to not empty. Here we are sure that no one is waiting
	 * synchronously. */
	if (info->notify_owner &&
	    info->attr.mq_curmsgs == 1) {
		struct siginfo sig_i;
		switch (info->notify.sigev_notify) {
		case SIGEV_NONE:
			break;
		case SIGEV_SIGNAL:
			/* sends signal */

			sig_i.si_signo = info->notify.sigev_signo;
			sig_i.si_errno = 0;
			sig_i.si_code = SI_MESGQ;
			sig_i.si_value = info->notify.sigev_value;
644 645
			/* map current pid/uid into info->owner's namespaces */
			rcu_read_lock();
646 647
			sig_i.si_pid = task_tgid_nr_ns(current,
						ns_of_pid(info->notify_owner));
648
			sig_i.si_uid = from_kuid_munged(info->notify_user_ns, current_uid());
649
			rcu_read_unlock();
L
Linus Torvalds 已提交
650

651 652
			kill_pid_info(info->notify.sigev_signo,
				      &sig_i, info->notify_owner);
L
Linus Torvalds 已提交
653 654 655
			break;
		case SIGEV_THREAD:
			set_cookie(info->notify_cookie, NOTIFY_WOKENUP);
656
			netlink_sendskb(info->notify_sock, info->notify_cookie);
L
Linus Torvalds 已提交
657 658 659
			break;
		}
		/* after notification unregisters process */
660
		put_pid(info->notify_owner);
661
		put_user_ns(info->notify_user_ns);
662
		info->notify_owner = NULL;
663
		info->notify_user_ns = NULL;
L
Linus Torvalds 已提交
664 665 666 667
	}
	wake_up(&info->wait_q);
}

668 669
static int prepare_timeout(const struct timespec __user *u_abs_timeout,
			   ktime_t *expires, struct timespec *ts)
L
Linus Torvalds 已提交
670
{
671 672 673 674
	if (copy_from_user(ts, u_abs_timeout, sizeof(struct timespec)))
		return -EFAULT;
	if (!timespec_valid(ts))
		return -EINVAL;
L
Linus Torvalds 已提交
675

676 677
	*expires = timespec_to_ktime(*ts);
	return 0;
L
Linus Torvalds 已提交
678 679 680 681
}

static void remove_notification(struct mqueue_inode_info *info)
{
682
	if (info->notify_owner != NULL &&
L
Linus Torvalds 已提交
683 684
	    info->notify.sigev_notify == SIGEV_THREAD) {
		set_cookie(info->notify_cookie, NOTIFY_REMOVED);
685
		netlink_sendskb(info->notify_sock, info->notify_cookie);
L
Linus Torvalds 已提交
686
	}
687
	put_pid(info->notify_owner);
688
	put_user_ns(info->notify_user_ns);
689
	info->notify_owner = NULL;
690
	info->notify_user_ns = NULL;
L
Linus Torvalds 已提交
691 692
}

693
static int mq_attr_ok(struct ipc_namespace *ipc_ns, struct mq_attr *attr)
L
Linus Torvalds 已提交
694
{
695 696 697
	int mq_treesize;
	unsigned long total_size;

L
Linus Torvalds 已提交
698
	if (attr->mq_maxmsg <= 0 || attr->mq_msgsize <= 0)
699
		return -EINVAL;
L
Linus Torvalds 已提交
700
	if (capable(CAP_SYS_RESOURCE)) {
D
Doug Ledford 已提交
701 702
		if (attr->mq_maxmsg > HARD_MSGMAX ||
		    attr->mq_msgsize > HARD_MSGSIZEMAX)
703
			return -EINVAL;
L
Linus Torvalds 已提交
704
	} else {
705 706
		if (attr->mq_maxmsg > ipc_ns->mq_msg_max ||
				attr->mq_msgsize > ipc_ns->mq_msgsize_max)
707
			return -EINVAL;
L
Linus Torvalds 已提交
708 709 710
	}
	/* check for overflow */
	if (attr->mq_msgsize > ULONG_MAX/attr->mq_maxmsg)
711
		return -EOVERFLOW;
712 713 714 715 716
	mq_treesize = attr->mq_maxmsg * sizeof(struct msg_msg) +
		min_t(unsigned int, attr->mq_maxmsg, MQ_PRIO_MAX) *
		sizeof(struct posix_msg_tree_node);
	total_size = attr->mq_maxmsg * attr->mq_msgsize;
	if (total_size + mq_treesize < total_size)
717 718
		return -EOVERFLOW;
	return 0;
L
Linus Torvalds 已提交
719 720 721 722 723
}

/*
 * Invoked when creating a new queue via sys_mq_open
 */
724
static struct file *do_create(struct ipc_namespace *ipc_ns, struct dentry *dir,
A
Al Viro 已提交
725
			struct dentry *dentry, int oflag, umode_t mode,
726
			struct mq_attr *attr)
L
Linus Torvalds 已提交
727
{
728
	const struct cred *cred = current_cred();
729
	struct file *result;
L
Linus Torvalds 已提交
730 731
	int ret;

A
Al Viro 已提交
732
	if (attr) {
733 734
		ret = mq_attr_ok(ipc_ns, attr);
		if (ret)
735
			goto out;
L
Linus Torvalds 已提交
736
		/* store for use during create */
A
Al Viro 已提交
737
		dentry->d_fsdata = attr;
738 739 740 741 742 743 744 745 746 747
	} else {
		struct mq_attr def_attr;

		def_attr.mq_maxmsg = min(ipc_ns->mq_msg_max,
					 ipc_ns->mq_msg_default);
		def_attr.mq_msgsize = min(ipc_ns->mq_msgsize_max,
					  ipc_ns->mq_msgsize_default);
		ret = mq_attr_ok(ipc_ns, &def_attr);
		if (ret)
			goto out;
L
Linus Torvalds 已提交
748 749
	}

A
Al Viro 已提交
750
	mode &= ~current_umask();
751
	ret = mnt_want_write(ipc_ns->mq_mnt);
752 753
	if (ret)
		goto out;
L
Linus Torvalds 已提交
754 755 756
	ret = vfs_create(dir->d_inode, dentry, mode, NULL);
	dentry->d_fsdata = NULL;
	if (ret)
757 758
		goto out_drop_write;

759
	result = dentry_open(dentry, ipc_ns->mq_mnt, oflag, cred);
760 761 762 763
	/*
	 * dentry_open() took a persistent mnt_want_write(),
	 * so we can now drop this one.
	 */
764
	mnt_drop_write(ipc_ns->mq_mnt);
765 766 767
	return result;

out_drop_write:
768
	mnt_drop_write(ipc_ns->mq_mnt);
769 770
out:
	dput(dentry);
771
	mntput(ipc_ns->mq_mnt);
772
	return ERR_PTR(ret);
L
Linus Torvalds 已提交
773 774 775
}

/* Opens existing queue */
776 777
static struct file *do_open(struct ipc_namespace *ipc_ns,
				struct dentry *dentry, int oflag)
L
Linus Torvalds 已提交
778
{
779
	int ret;
780 781 782 783
	const struct cred *cred = current_cred();

	static const int oflag2acc[O_ACCMODE] = { MAY_READ, MAY_WRITE,
						  MAY_READ | MAY_WRITE };
L
Linus Torvalds 已提交
784

785
	if ((oflag & O_ACCMODE) == (O_RDWR | O_WRONLY)) {
786 787
		ret = -EINVAL;
		goto err;
788
	}
L
Linus Torvalds 已提交
789

790
	if (inode_permission(dentry->d_inode, oflag2acc[oflag & O_ACCMODE])) {
791 792
		ret = -EACCES;
		goto err;
793
	}
L
Linus Torvalds 已提交
794

795
	return dentry_open(dentry, ipc_ns->mq_mnt, oflag, cred);
796 797 798 799 800

err:
	dput(dentry);
	mntput(ipc_ns->mq_mnt);
	return ERR_PTR(ret);
L
Linus Torvalds 已提交
801 802
}

A
Al Viro 已提交
803
SYSCALL_DEFINE4(mq_open, const char __user *, u_name, int, oflag, umode_t, mode,
804
		struct mq_attr __user *, u_attr)
L
Linus Torvalds 已提交
805 806 807 808
{
	struct dentry *dentry;
	struct file *filp;
	char *name;
A
Al Viro 已提交
809
	struct mq_attr attr;
L
Linus Torvalds 已提交
810
	int fd, error;
811
	struct ipc_namespace *ipc_ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
812

A
Al Viro 已提交
813 814 815 816
	if (u_attr && copy_from_user(&attr, u_attr, sizeof(struct mq_attr)))
		return -EFAULT;

	audit_mq_open(oflag, mode, u_attr ? &attr : NULL);
817

L
Linus Torvalds 已提交
818 819 820
	if (IS_ERR(name = getname(u_name)))
		return PTR_ERR(name);

U
Ulrich Drepper 已提交
821
	fd = get_unused_fd_flags(O_CLOEXEC);
L
Linus Torvalds 已提交
822 823 824
	if (fd < 0)
		goto out_putname;

825 826
	mutex_lock(&ipc_ns->mq_mnt->mnt_root->d_inode->i_mutex);
	dentry = lookup_one_len(name, ipc_ns->mq_mnt->mnt_root, strlen(name));
L
Linus Torvalds 已提交
827 828
	if (IS_ERR(dentry)) {
		error = PTR_ERR(dentry);
829
		goto out_putfd;
L
Linus Torvalds 已提交
830
	}
831
	mntget(ipc_ns->mq_mnt);
L
Linus Torvalds 已提交
832 833 834

	if (oflag & O_CREAT) {
		if (dentry->d_inode) {	/* entry already exists */
835
			audit_inode(name, dentry);
836 837
			if (oflag & O_EXCL) {
				error = -EEXIST;
838
				goto out;
839
			}
840
			filp = do_open(ipc_ns, dentry, oflag);
L
Linus Torvalds 已提交
841
		} else {
842 843
			filp = do_create(ipc_ns, ipc_ns->mq_mnt->mnt_root,
						dentry, oflag, mode,
A
Al Viro 已提交
844
						u_attr ? &attr : NULL);
L
Linus Torvalds 已提交
845
		}
846
	} else {
847 848
		if (!dentry->d_inode) {
			error = -ENOENT;
849
			goto out;
850
		}
851
		audit_inode(name, dentry);
852
		filp = do_open(ipc_ns, dentry, oflag);
853
	}
L
Linus Torvalds 已提交
854 855 856 857 858 859 860 861 862

	if (IS_ERR(filp)) {
		error = PTR_ERR(filp);
		goto out_putfd;
	}

	fd_install(fd, filp);
	goto out_upsem;

863 864
out:
	dput(dentry);
865
	mntput(ipc_ns->mq_mnt);
866
out_putfd:
L
Linus Torvalds 已提交
867 868 869
	put_unused_fd(fd);
	fd = error;
out_upsem:
870
	mutex_unlock(&ipc_ns->mq_mnt->mnt_root->d_inode->i_mutex);
L
Linus Torvalds 已提交
871 872 873 874 875
out_putname:
	putname(name);
	return fd;
}

876
SYSCALL_DEFINE1(mq_unlink, const char __user *, u_name)
L
Linus Torvalds 已提交
877 878 879 880 881
{
	int err;
	char *name;
	struct dentry *dentry;
	struct inode *inode = NULL;
882
	struct ipc_namespace *ipc_ns = current->nsproxy->ipc_ns;
L
Linus Torvalds 已提交
883 884 885 886 887

	name = getname(u_name);
	if (IS_ERR(name))
		return PTR_ERR(name);

888
	mutex_lock_nested(&ipc_ns->mq_mnt->mnt_root->d_inode->i_mutex,
889
			I_MUTEX_PARENT);
890
	dentry = lookup_one_len(name, ipc_ns->mq_mnt->mnt_root, strlen(name));
L
Linus Torvalds 已提交
891 892 893 894 895 896 897 898 899 900 901 902
	if (IS_ERR(dentry)) {
		err = PTR_ERR(dentry);
		goto out_unlock;
	}

	if (!dentry->d_inode) {
		err = -ENOENT;
		goto out_err;
	}

	inode = dentry->d_inode;
	if (inode)
A
Al Viro 已提交
903
		ihold(inode);
904
	err = mnt_want_write(ipc_ns->mq_mnt);
905 906
	if (err)
		goto out_err;
L
Linus Torvalds 已提交
907
	err = vfs_unlink(dentry->d_parent->d_inode, dentry);
908
	mnt_drop_write(ipc_ns->mq_mnt);
L
Linus Torvalds 已提交
909 910 911 912
out_err:
	dput(dentry);

out_unlock:
913
	mutex_unlock(&ipc_ns->mq_mnt->mnt_root->d_inode->i_mutex);
L
Linus Torvalds 已提交
914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930
	putname(name);
	if (inode)
		iput(inode);

	return err;
}

/* Pipelined send and receive functions.
 *
 * If a receiver finds no waiting message, then it registers itself in the
 * list of waiting receivers. A sender checks that list before adding the new
 * message into the message array. If there is a waiting receiver, then it
 * bypasses the message array and directly hands the message over to the
 * receiver.
 * The receiver accepts the message and returns without grabbing the queue
 * spinlock. Therefore an intermediate STATE_PENDING state and memory barriers
 * are necessary. The same algorithm is used for sysv semaphores, see
931
 * ipc/sem.c for more details.
L
Linus Torvalds 已提交
932 933 934 935 936 937 938 939 940 941 942 943 944 945 946
 *
 * The same algorithm is used for senders.
 */

/* pipelined_send() - send a message directly to the task waiting in
 * sys_mq_timedreceive() (without inserting message into a queue).
 */
static inline void pipelined_send(struct mqueue_inode_info *info,
				  struct msg_msg *message,
				  struct ext_wait_queue *receiver)
{
	receiver->msg = message;
	list_del(&receiver->list);
	receiver->state = STATE_PENDING;
	wake_up_process(receiver->task);
947
	smp_wmb();
L
Linus Torvalds 已提交
948 949 950 951 952 953 954 955 956 957 958 959 960 961
	receiver->state = STATE_READY;
}

/* pipelined_receive() - if there is task waiting in sys_mq_timedsend()
 * gets its message and put to the queue (we have one free place for sure). */
static inline void pipelined_receive(struct mqueue_inode_info *info)
{
	struct ext_wait_queue *sender = wq_get_first_waiter(info, SEND);

	if (!sender) {
		/* for poll */
		wake_up_interruptible(&info->wait_q);
		return;
	}
962 963
	if (msg_insert(sender->msg, info))
		return;
L
Linus Torvalds 已提交
964 965 966
	list_del(&sender->list);
	sender->state = STATE_PENDING;
	wake_up_process(sender->task);
967
	smp_wmb();
L
Linus Torvalds 已提交
968 969 970
	sender->state = STATE_READY;
}

971 972 973
SYSCALL_DEFINE5(mq_timedsend, mqd_t, mqdes, const char __user *, u_msg_ptr,
		size_t, msg_len, unsigned int, msg_prio,
		const struct timespec __user *, u_abs_timeout)
L
Linus Torvalds 已提交
974 975 976 977 978 979 980
{
	struct file *filp;
	struct inode *inode;
	struct ext_wait_queue wait;
	struct ext_wait_queue *receiver;
	struct msg_msg *msg_ptr;
	struct mqueue_inode_info *info;
981 982
	ktime_t expires, *timeout = NULL;
	struct timespec ts;
983 984
	struct posix_msg_tree_node *new_leaf = NULL;
	int ret = 0;
L
Linus Torvalds 已提交
985

A
Al Viro 已提交
986
	if (u_abs_timeout) {
987 988 989 990
		int res = prepare_timeout(u_abs_timeout, &expires, &ts);
		if (res)
			return res;
		timeout = &expires;
A
Al Viro 已提交
991
	}
992

L
Linus Torvalds 已提交
993 994 995
	if (unlikely(msg_prio >= (unsigned long) MQ_PRIO_MAX))
		return -EINVAL;

996
	audit_mq_sendrecv(mqdes, msg_len, msg_prio, timeout ? &ts : NULL);
L
Linus Torvalds 已提交
997 998

	filp = fget(mqdes);
999 1000
	if (unlikely(!filp)) {
		ret = -EBADF;
L
Linus Torvalds 已提交
1001
		goto out;
1002
	}
L
Linus Torvalds 已提交
1003

J
Josef Sipek 已提交
1004
	inode = filp->f_path.dentry->d_inode;
1005 1006
	if (unlikely(filp->f_op != &mqueue_file_operations)) {
		ret = -EBADF;
L
Linus Torvalds 已提交
1007
		goto out_fput;
1008
	}
L
Linus Torvalds 已提交
1009
	info = MQUEUE_I(inode);
1010
	audit_inode(NULL, filp->f_path.dentry);
L
Linus Torvalds 已提交
1011

1012 1013
	if (unlikely(!(filp->f_mode & FMODE_WRITE))) {
		ret = -EBADF;
L
Linus Torvalds 已提交
1014
		goto out_fput;
1015
	}
L
Linus Torvalds 已提交
1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031

	if (unlikely(msg_len > info->attr.mq_msgsize)) {
		ret = -EMSGSIZE;
		goto out_fput;
	}

	/* First try to allocate memory, before doing anything with
	 * existing queues. */
	msg_ptr = load_msg(u_msg_ptr, msg_len);
	if (IS_ERR(msg_ptr)) {
		ret = PTR_ERR(msg_ptr);
		goto out_fput;
	}
	msg_ptr->m_ts = msg_len;
	msg_ptr->m_type = msg_prio;

1032 1033 1034 1035 1036 1037 1038 1039
	/*
	 * msg_insert really wants us to have a valid, spare node struct so
	 * it doesn't have to kmalloc a GFP_ATOMIC allocation, but it will
	 * fall back to that if necessary.
	 */
	if (!info->node_cache)
		new_leaf = kmalloc(sizeof(*new_leaf), GFP_KERNEL);

L
Linus Torvalds 已提交
1040 1041
	spin_lock(&info->lock);

1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052
	if (!info->node_cache && new_leaf) {
		/* Save our speculative allocation into the cache */
		rb_init_node(&new_leaf->rb_node);
		INIT_LIST_HEAD(&new_leaf->msg_list);
		info->node_cache = new_leaf;
		info->qsize += sizeof(*new_leaf);
		new_leaf = NULL;
	} else {
		kfree(new_leaf);
	}

L
Linus Torvalds 已提交
1053 1054 1055 1056 1057 1058 1059 1060
	if (info->attr.mq_curmsgs == info->attr.mq_maxmsg) {
		if (filp->f_flags & O_NONBLOCK) {
			ret = -EAGAIN;
		} else {
			wait.task = current;
			wait.msg = (void *) msg_ptr;
			wait.state = STATE_NONE;
			ret = wq_sleep(info, SEND, timeout, &wait);
1061 1062 1063 1064 1065
			/*
			 * wq_sleep must be called with info->lock held, and
			 * returns with the lock released
			 */
			goto out_free;
L
Linus Torvalds 已提交
1066 1067 1068 1069 1070 1071 1072
		}
	} else {
		receiver = wq_get_first_waiter(info, RECV);
		if (receiver) {
			pipelined_send(info, msg_ptr, receiver);
		} else {
			/* adds message to the queue */
1073 1074 1075
			ret = msg_insert(msg_ptr, info);
			if (ret)
				goto out_unlock;
L
Linus Torvalds 已提交
1076 1077 1078 1079 1080
			__do_notify(info);
		}
		inode->i_atime = inode->i_mtime = inode->i_ctime =
				CURRENT_TIME;
	}
1081 1082 1083 1084 1085
out_unlock:
	spin_unlock(&info->lock);
out_free:
	if (ret)
		free_msg(msg_ptr);
L
Linus Torvalds 已提交
1086 1087 1088 1089 1090 1091
out_fput:
	fput(filp);
out:
	return ret;
}

1092 1093 1094
SYSCALL_DEFINE5(mq_timedreceive, mqd_t, mqdes, char __user *, u_msg_ptr,
		size_t, msg_len, unsigned int __user *, u_msg_prio,
		const struct timespec __user *, u_abs_timeout)
L
Linus Torvalds 已提交
1095 1096 1097 1098 1099 1100 1101
{
	ssize_t ret;
	struct msg_msg *msg_ptr;
	struct file *filp;
	struct inode *inode;
	struct mqueue_inode_info *info;
	struct ext_wait_queue wait;
1102 1103
	ktime_t expires, *timeout = NULL;
	struct timespec ts;
1104
	struct posix_msg_tree_node *new_leaf = NULL;
L
Linus Torvalds 已提交
1105

A
Al Viro 已提交
1106
	if (u_abs_timeout) {
1107 1108 1109 1110
		int res = prepare_timeout(u_abs_timeout, &expires, &ts);
		if (res)
			return res;
		timeout = &expires;
A
Al Viro 已提交
1111
	}
1112

1113
	audit_mq_sendrecv(mqdes, msg_len, 0, timeout ? &ts : NULL);
L
Linus Torvalds 已提交
1114 1115

	filp = fget(mqdes);
1116 1117
	if (unlikely(!filp)) {
		ret = -EBADF;
L
Linus Torvalds 已提交
1118
		goto out;
1119
	}
L
Linus Torvalds 已提交
1120

J
Josef Sipek 已提交
1121
	inode = filp->f_path.dentry->d_inode;
1122 1123
	if (unlikely(filp->f_op != &mqueue_file_operations)) {
		ret = -EBADF;
L
Linus Torvalds 已提交
1124
		goto out_fput;
1125
	}
L
Linus Torvalds 已提交
1126
	info = MQUEUE_I(inode);
1127
	audit_inode(NULL, filp->f_path.dentry);
L
Linus Torvalds 已提交
1128

1129 1130
	if (unlikely(!(filp->f_mode & FMODE_READ))) {
		ret = -EBADF;
L
Linus Torvalds 已提交
1131
		goto out_fput;
1132
	}
L
Linus Torvalds 已提交
1133 1134 1135 1136 1137 1138 1139

	/* checks if buffer is big enough */
	if (unlikely(msg_len < info->attr.mq_msgsize)) {
		ret = -EMSGSIZE;
		goto out_fput;
	}

1140 1141 1142 1143 1144 1145 1146 1147
	/*
	 * msg_insert really wants us to have a valid, spare node struct so
	 * it doesn't have to kmalloc a GFP_ATOMIC allocation, but it will
	 * fall back to that if necessary.
	 */
	if (!info->node_cache)
		new_leaf = kmalloc(sizeof(*new_leaf), GFP_KERNEL);

L
Linus Torvalds 已提交
1148
	spin_lock(&info->lock);
1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159

	if (!info->node_cache && new_leaf) {
		/* Save our speculative allocation into the cache */
		rb_init_node(&new_leaf->rb_node);
		INIT_LIST_HEAD(&new_leaf->msg_list);
		info->node_cache = new_leaf;
		info->qsize += sizeof(*new_leaf);
	} else {
		kfree(new_leaf);
	}

L
Linus Torvalds 已提交
1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200
	if (info->attr.mq_curmsgs == 0) {
		if (filp->f_flags & O_NONBLOCK) {
			spin_unlock(&info->lock);
			ret = -EAGAIN;
		} else {
			wait.task = current;
			wait.state = STATE_NONE;
			ret = wq_sleep(info, RECV, timeout, &wait);
			msg_ptr = wait.msg;
		}
	} else {
		msg_ptr = msg_get(info);

		inode->i_atime = inode->i_mtime = inode->i_ctime =
				CURRENT_TIME;

		/* There is now free space in queue. */
		pipelined_receive(info);
		spin_unlock(&info->lock);
		ret = 0;
	}
	if (ret == 0) {
		ret = msg_ptr->m_ts;

		if ((u_msg_prio && put_user(msg_ptr->m_type, u_msg_prio)) ||
			store_msg(u_msg_ptr, msg_ptr, msg_ptr->m_ts)) {
			ret = -EFAULT;
		}
		free_msg(msg_ptr);
	}
out_fput:
	fput(filp);
out:
	return ret;
}

/*
 * Notes: the case when user wants us to deregister (with NULL as pointer)
 * and he isn't currently owner of notification, will be silently discarded.
 * It isn't explicitly defined in the POSIX.
 */
1201 1202
SYSCALL_DEFINE2(mq_notify, mqd_t, mqdes,
		const struct sigevent __user *, u_notification)
L
Linus Torvalds 已提交
1203 1204 1205 1206 1207 1208 1209 1210 1211
{
	int ret;
	struct file *filp;
	struct sock *sock;
	struct inode *inode;
	struct sigevent notification;
	struct mqueue_inode_info *info;
	struct sk_buff *nc;

A
Al Viro 已提交
1212
	if (u_notification) {
L
Linus Torvalds 已提交
1213 1214 1215
		if (copy_from_user(&notification, u_notification,
					sizeof(struct sigevent)))
			return -EFAULT;
A
Al Viro 已提交
1216 1217 1218
	}

	audit_mq_notify(mqdes, u_notification ? &notification : NULL);
L
Linus Torvalds 已提交
1219

A
Al Viro 已提交
1220 1221 1222
	nc = NULL;
	sock = NULL;
	if (u_notification != NULL) {
L
Linus Torvalds 已提交
1223 1224 1225 1226 1227
		if (unlikely(notification.sigev_notify != SIGEV_NONE &&
			     notification.sigev_notify != SIGEV_SIGNAL &&
			     notification.sigev_notify != SIGEV_THREAD))
			return -EINVAL;
		if (notification.sigev_notify == SIGEV_SIGNAL &&
1228
			!valid_signal(notification.sigev_signo)) {
L
Linus Torvalds 已提交
1229 1230 1231
			return -EINVAL;
		}
		if (notification.sigev_notify == SIGEV_THREAD) {
P
Patrick McHardy 已提交
1232 1233
			long timeo;

L
Linus Torvalds 已提交
1234 1235
			/* create the notify skb */
			nc = alloc_skb(NOTIFY_COOKIE_LEN, GFP_KERNEL);
1236 1237
			if (!nc) {
				ret = -ENOMEM;
L
Linus Torvalds 已提交
1238
				goto out;
1239
			}
L
Linus Torvalds 已提交
1240 1241 1242
			if (copy_from_user(nc->data,
					notification.sigev_value.sival_ptr,
					NOTIFY_COOKIE_LEN)) {
1243
				ret = -EFAULT;
L
Linus Torvalds 已提交
1244 1245 1246 1247 1248 1249 1250 1251
				goto out;
			}

			/* TODO: add a header? */
			skb_put(nc, NOTIFY_COOKIE_LEN);
			/* and attach it to the socket */
retry:
			filp = fget(notification.sigev_signo);
1252 1253
			if (!filp) {
				ret = -EBADF;
L
Linus Torvalds 已提交
1254
				goto out;
1255
			}
L
Linus Torvalds 已提交
1256 1257 1258 1259 1260 1261 1262 1263
			sock = netlink_getsockbyfilp(filp);
			fput(filp);
			if (IS_ERR(sock)) {
				ret = PTR_ERR(sock);
				sock = NULL;
				goto out;
			}

P
Patrick McHardy 已提交
1264
			timeo = MAX_SCHEDULE_TIMEOUT;
1265
			ret = netlink_attachskb(sock, nc, &timeo, NULL);
L
Linus Torvalds 已提交
1266
			if (ret == 1)
1267
				goto retry;
L
Linus Torvalds 已提交
1268 1269 1270 1271 1272 1273 1274 1275 1276
			if (ret) {
				sock = NULL;
				nc = NULL;
				goto out;
			}
		}
	}

	filp = fget(mqdes);
1277 1278
	if (!filp) {
		ret = -EBADF;
L
Linus Torvalds 已提交
1279
		goto out;
1280
	}
L
Linus Torvalds 已提交
1281

J
Josef Sipek 已提交
1282
	inode = filp->f_path.dentry->d_inode;
1283 1284
	if (unlikely(filp->f_op != &mqueue_file_operations)) {
		ret = -EBADF;
L
Linus Torvalds 已提交
1285
		goto out_fput;
1286
	}
L
Linus Torvalds 已提交
1287 1288 1289 1290 1291
	info = MQUEUE_I(inode);

	ret = 0;
	spin_lock(&info->lock);
	if (u_notification == NULL) {
1292
		if (info->notify_owner == task_tgid(current)) {
L
Linus Torvalds 已提交
1293 1294 1295
			remove_notification(info);
			inode->i_atime = inode->i_ctime = CURRENT_TIME;
		}
1296
	} else if (info->notify_owner != NULL) {
L
Linus Torvalds 已提交
1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315
		ret = -EBUSY;
	} else {
		switch (notification.sigev_notify) {
		case SIGEV_NONE:
			info->notify.sigev_notify = SIGEV_NONE;
			break;
		case SIGEV_THREAD:
			info->notify_sock = sock;
			info->notify_cookie = nc;
			sock = NULL;
			nc = NULL;
			info->notify.sigev_notify = SIGEV_THREAD;
			break;
		case SIGEV_SIGNAL:
			info->notify.sigev_signo = notification.sigev_signo;
			info->notify.sigev_value = notification.sigev_value;
			info->notify.sigev_notify = SIGEV_SIGNAL;
			break;
		}
1316 1317

		info->notify_owner = get_pid(task_tgid(current));
1318
		info->notify_user_ns = get_user_ns(current_user_ns());
L
Linus Torvalds 已提交
1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332
		inode->i_atime = inode->i_ctime = CURRENT_TIME;
	}
	spin_unlock(&info->lock);
out_fput:
	fput(filp);
out:
	if (sock) {
		netlink_detachskb(sock, nc);
	} else if (nc) {
		dev_kfree_skb(nc);
	}
	return ret;
}

1333 1334 1335
SYSCALL_DEFINE3(mq_getsetattr, mqd_t, mqdes,
		const struct mq_attr __user *, u_mqstat,
		struct mq_attr __user *, u_omqstat)
L
Linus Torvalds 已提交
1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350
{
	int ret;
	struct mq_attr mqstat, omqstat;
	struct file *filp;
	struct inode *inode;
	struct mqueue_inode_info *info;

	if (u_mqstat != NULL) {
		if (copy_from_user(&mqstat, u_mqstat, sizeof(struct mq_attr)))
			return -EFAULT;
		if (mqstat.mq_flags & (~O_NONBLOCK))
			return -EINVAL;
	}

	filp = fget(mqdes);
1351 1352
	if (!filp) {
		ret = -EBADF;
L
Linus Torvalds 已提交
1353
		goto out;
1354
	}
L
Linus Torvalds 已提交
1355

J
Josef Sipek 已提交
1356
	inode = filp->f_path.dentry->d_inode;
1357 1358
	if (unlikely(filp->f_op != &mqueue_file_operations)) {
		ret = -EBADF;
L
Linus Torvalds 已提交
1359
		goto out_fput;
1360
	}
L
Linus Torvalds 已提交
1361 1362 1363 1364 1365 1366 1367
	info = MQUEUE_I(inode);

	spin_lock(&info->lock);

	omqstat = info->attr;
	omqstat.mq_flags = filp->f_flags & O_NONBLOCK;
	if (u_mqstat) {
A
Al Viro 已提交
1368
		audit_mq_getsetattr(mqdes, &mqstat);
J
Jonathan Corbet 已提交
1369
		spin_lock(&filp->f_lock);
L
Linus Torvalds 已提交
1370 1371 1372 1373
		if (mqstat.mq_flags & O_NONBLOCK)
			filp->f_flags |= O_NONBLOCK;
		else
			filp->f_flags &= ~O_NONBLOCK;
J
Jonathan Corbet 已提交
1374
		spin_unlock(&filp->f_lock);
L
Linus Torvalds 已提交
1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391

		inode->i_atime = inode->i_ctime = CURRENT_TIME;
	}

	spin_unlock(&info->lock);

	ret = 0;
	if (u_omqstat != NULL && copy_to_user(u_omqstat, &omqstat,
						sizeof(struct mq_attr)))
		ret = -EFAULT;

out_fput:
	fput(filp);
out:
	return ret;
}

1392
static const struct inode_operations mqueue_dir_inode_operations = {
L
Linus Torvalds 已提交
1393 1394 1395 1396 1397
	.lookup = simple_lookup,
	.create = mqueue_create,
	.unlink = mqueue_unlink,
};

1398
static const struct file_operations mqueue_file_operations = {
L
Linus Torvalds 已提交
1399 1400 1401
	.flush = mqueue_flush_file,
	.poll = mqueue_poll_file,
	.read = mqueue_read_file,
1402
	.llseek = default_llseek,
L
Linus Torvalds 已提交
1403 1404
};

1405
static const struct super_operations mqueue_super_ops = {
L
Linus Torvalds 已提交
1406 1407
	.alloc_inode = mqueue_alloc_inode,
	.destroy_inode = mqueue_destroy_inode,
A
Al Viro 已提交
1408
	.evict_inode = mqueue_evict_inode,
L
Linus Torvalds 已提交
1409 1410 1411 1412 1413
	.statfs = simple_statfs,
};

static struct file_system_type mqueue_fs_type = {
	.name = "mqueue",
A
Al Viro 已提交
1414
	.mount = mqueue_mount,
L
Linus Torvalds 已提交
1415 1416 1417
	.kill_sb = kill_litter_super,
};

1418 1419 1420 1421 1422 1423
int mq_init_ns(struct ipc_namespace *ns)
{
	ns->mq_queues_count  = 0;
	ns->mq_queues_max    = DFLT_QUEUESMAX;
	ns->mq_msg_max       = DFLT_MSGMAX;
	ns->mq_msgsize_max   = DFLT_MSGSIZEMAX;
1424 1425
	ns->mq_msg_default   = DFLT_MSG;
	ns->mq_msgsize_default  = DFLT_MSGSIZE;
1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442

	ns->mq_mnt = kern_mount_data(&mqueue_fs_type, ns);
	if (IS_ERR(ns->mq_mnt)) {
		int err = PTR_ERR(ns->mq_mnt);
		ns->mq_mnt = NULL;
		return err;
	}
	return 0;
}

void mq_clear_sbinfo(struct ipc_namespace *ns)
{
	ns->mq_mnt->mnt_sb->s_fs_info = NULL;
}

void mq_put_mnt(struct ipc_namespace *ns)
{
A
Al Viro 已提交
1443
	kern_unmount(ns->mq_mnt);
1444 1445
}

L
Linus Torvalds 已提交
1446 1447 1448 1449 1450 1451
static int __init init_mqueue_fs(void)
{
	int error;

	mqueue_inode_cachep = kmem_cache_create("mqueue_inode_cache",
				sizeof(struct mqueue_inode_info), 0,
1452
				SLAB_HWCACHE_ALIGN, init_once);
L
Linus Torvalds 已提交
1453 1454 1455
	if (mqueue_inode_cachep == NULL)
		return -ENOMEM;

1456
	/* ignore failures - they are not fatal */
1457
	mq_sysctl_table = mq_register_sysctl_table();
L
Linus Torvalds 已提交
1458 1459 1460 1461 1462

	error = register_filesystem(&mqueue_fs_type);
	if (error)
		goto out_sysctl;

1463 1464
	spin_lock_init(&mq_lock);

A
Al Viro 已提交
1465 1466
	error = mq_init_ns(&init_ipc_ns);
	if (error)
L
Linus Torvalds 已提交
1467 1468 1469 1470 1471 1472 1473 1474 1475
		goto out_filesystem;

	return 0;

out_filesystem:
	unregister_filesystem(&mqueue_fs_type);
out_sysctl:
	if (mq_sysctl_table)
		unregister_sysctl_table(mq_sysctl_table);
1476
	kmem_cache_destroy(mqueue_inode_cachep);
L
Linus Torvalds 已提交
1477 1478 1479 1480
	return error;
}

__initcall(init_mqueue_fs);