mqueue.c 39 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
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)
Linus Torvalds's avatar
Linus Torvalds committed
6 7 8
 *
 * Spinlocks:               Mohamed Abbas           (abbas.mohamed@intel.com)
 * Lockless receive & send, fd based notify:
Manfred Spraul's avatar
Manfred Spraul committed
9
 *			    Manfred Spraul	    (manfred@colorfullife.com)
Linus Torvalds's avatar
Linus Torvalds committed
10
 *
11 12
 * Audit:                   George Wilson           (ltcgcw@us.ibm.com)
 *
Linus Torvalds's avatar
Linus Torvalds committed
13 14 15
 * This file is released under the GPL.
 */

16
#include <linux/capability.h>
Linus Torvalds's avatar
Linus Torvalds committed
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>
Linus Torvalds's avatar
Linus Torvalds committed
28 29
#include <linux/netlink.h>
#include <linux/syscalls.h>
30
#include <linux/audit.h>
31
#include <linux/signal.h>
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>
38
#include <linux/sched/wake_q.h>
39
#include <linux/sched/signal.h>
40
#include <linux/sched/user.h>
41

Linus Torvalds's avatar
Linus Torvalds committed
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
53
#define STATE_READY	1
Linus Torvalds's avatar
Linus Torvalds committed
54

55 56 57 58 59 60
struct posix_msg_tree_node {
	struct rb_node		rb_node;
	struct list_head	msg_list;
	int			priority;
};

Linus Torvalds's avatar
Linus Torvalds committed
61 62 63 64 65 66 67 68 69 70 71 72
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;

73
	struct rb_root msg_tree;
74
	struct posix_msg_tree_node *node_cache;
Linus Torvalds's avatar
Linus Torvalds committed
75 76 77
	struct mq_attr attr;

	struct sigevent notify;
Manfred Spraul's avatar
Manfred Spraul committed
78
	struct pid *notify_owner;
79
	struct user_namespace *notify_user_ns;
80
	struct user_struct *user;	/* user who created, for accounting */
Linus Torvalds's avatar
Linus Torvalds committed
81 82 83 84 85 86 87 88 89
	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) */
};

90
static const struct inode_operations mqueue_dir_inode_operations;
91
static const struct file_operations mqueue_file_operations;
92
static const struct super_operations mqueue_super_ops;
Linus Torvalds's avatar
Linus Torvalds committed
93 94
static void remove_notification(struct mqueue_inode_info *info);

95
static struct kmem_cache *mqueue_inode_cachep;
Linus Torvalds's avatar
Linus Torvalds committed
96

Manfred Spraul's avatar
Manfred Spraul committed
97
static struct ctl_table_header *mq_sysctl_table;
Linus Torvalds's avatar
Linus Torvalds committed
98 99 100 101 102 103

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

104 105 106 107
/*
 * This routine should be called with the mq_lock held.
 */
static inline struct ipc_namespace *__get_ns_from_inode(struct inode *inode)
108
{
109
	return get_ipc_ns(inode->i_sb->s_fs_info);
110 111
}

112
static struct ipc_namespace *get_ns_from_inode(struct inode *inode)
113
{
114 115 116 117 118 119
	struct ipc_namespace *ns;

	spin_lock(&mq_lock);
	ns = __get_ns_from_inode(inode);
	spin_unlock(&mq_lock);
	return ns;
120 121
}

122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139
/* 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;
	}
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;
		INIT_LIST_HEAD(&leaf->msg_list);
	}
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
		if (info->node_cache) {
			kfree(leaf);
		} else {
			info->node_cache = leaf;
		}
196 197 198 199 200 201 202
		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);
203 204 205 206 207
			if (info->node_cache) {
				kfree(leaf);
			} else {
				info->node_cache = leaf;
			}
208 209 210 211 212 213 214
		}
	}
	info->attr.mq_curmsgs--;
	info->qsize -= msg->m_ts;
	return msg;
}

215
static struct inode *mqueue_get_inode(struct super_block *sb,
Al Viro's avatar
Al Viro committed
216
		struct ipc_namespace *ipc_ns, umode_t mode,
217
		struct mq_attr *attr)
Linus Torvalds's avatar
Linus Torvalds committed
218
{
219
	struct user_struct *u = current_user();
Linus Torvalds's avatar
Linus Torvalds committed
220
	struct inode *inode;
221
	int ret = -ENOMEM;
Linus Torvalds's avatar
Linus Torvalds committed
222 223

	inode = new_inode(sb);
224 225 226 227 228 229 230
	if (!inode)
		goto err;

	inode->i_ino = get_next_ino();
	inode->i_mode = mode;
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
231
	inode->i_mtime = inode->i_ctime = inode->i_atime = current_time(inode);
232 233 234

	if (S_ISREG(mode)) {
		struct mqueue_inode_info *info;
235
		unsigned long mq_bytes, mq_treesize;
236 237 238 239 240 241 242 243 244 245

		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;
246
		info->notify_user_ns = NULL;
247 248
		info->qsize = 0;
		info->user = NULL;	/* set when all is ok */
249
		info->msg_tree = RB_ROOT;
250
		info->node_cache = NULL;
251
		memset(&info->attr, 0, sizeof(info->attr));
252 253 254 255
		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);
256 257 258 259
		if (attr) {
			info->attr.mq_maxmsg = attr->mq_maxmsg;
			info->attr.mq_msgsize = attr->mq_msgsize;
		}
260 261 262 263 264 265 266 267 268 269 270 271 272
		/*
		 * 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.
		 */
273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289

		ret = -EINVAL;
		if (info->attr.mq_maxmsg <= 0 || info->attr.mq_msgsize <= 0)
			goto out_inode;
		if (capable(CAP_SYS_RESOURCE)) {
			if (info->attr.mq_maxmsg > HARD_MSGMAX ||
			    info->attr.mq_msgsize > HARD_MSGSIZEMAX)
				goto out_inode;
		} else {
			if (info->attr.mq_maxmsg > ipc_ns->mq_msg_max ||
					info->attr.mq_msgsize > ipc_ns->mq_msgsize_max)
				goto out_inode;
		}
		ret = -EOVERFLOW;
		/* check for overflow */
		if (info->attr.mq_msgsize > ULONG_MAX/info->attr.mq_maxmsg)
			goto out_inode;
290 291 292
		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);
293 294 295 296
		mq_bytes = info->attr.mq_maxmsg * info->attr.mq_msgsize;
		if (mq_bytes + mq_treesize < mq_bytes)
			goto out_inode;
		mq_bytes += mq_treesize;
297 298
		spin_lock(&mq_lock);
		if (u->mq_bytes + mq_bytes < u->mq_bytes ||
299
		    u->mq_bytes + mq_bytes > rlimit(RLIMIT_MSGQUEUE)) {
300 301
			spin_unlock(&mq_lock);
			/* mqueue_evict_inode() releases info->messages */
302
			ret = -EMFILE;
303
			goto out_inode;
Linus Torvalds's avatar
Linus Torvalds committed
304
		}
305 306 307 308 309 310 311 312 313 314 315
		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;
Linus Torvalds's avatar
Linus Torvalds committed
316
	}
317

Linus Torvalds's avatar
Linus Torvalds committed
318 319 320
	return inode;
out_inode:
	iput(inode);
321
err:
322
	return ERR_PTR(ret);
Linus Torvalds's avatar
Linus Torvalds committed
323 324 325 326 327
}

static int mqueue_fill_super(struct super_block *sb, void *data, int silent)
{
	struct inode *inode;
328
	struct ipc_namespace *ns = sb->s_fs_info;
Linus Torvalds's avatar
Linus Torvalds committed
329

330
	sb->s_iflags |= SB_I_NOEXEC | SB_I_NODEV;
331 332
	sb->s_blocksize = PAGE_SIZE;
	sb->s_blocksize_bits = PAGE_SHIFT;
Linus Torvalds's avatar
Linus Torvalds committed
333 334 335
	sb->s_magic = MQUEUE_MAGIC;
	sb->s_op = &mqueue_super_ops;

336 337 338
	inode = mqueue_get_inode(sb, ns, S_IFDIR | S_ISVTX | S_IRWXUGO, NULL);
	if (IS_ERR(inode))
		return PTR_ERR(inode);
Linus Torvalds's avatar
Linus Torvalds committed
339

340 341 342 343
	sb->s_root = d_make_root(inode);
	if (!sb->s_root)
		return -ENOMEM;
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
344 345
}

Al Viro's avatar
Al Viro committed
346
static struct dentry *mqueue_mount(struct file_system_type *fs_type,
347
			 int flags, const char *dev_name,
Al Viro's avatar
Al Viro committed
348
			 void *data)
Linus Torvalds's avatar
Linus Torvalds committed
349
{
350 351 352 353 354 355 356 357
	struct ipc_namespace *ns;
	if (flags & SB_KERNMOUNT) {
		ns = data;
		data = NULL;
	} else {
		ns = current->nsproxy->ipc_ns;
	}
	return mount_ns(fs_type, flags, data, ns, ns->user_ns, mqueue_fill_super);
Linus Torvalds's avatar
Linus Torvalds committed
358 359
}

360
static void init_once(void *foo)
Linus Torvalds's avatar
Linus Torvalds committed
361 362 363
{
	struct mqueue_inode_info *p = (struct mqueue_inode_info *) foo;

364
	inode_init_once(&p->vfs_inode);
Linus Torvalds's avatar
Linus Torvalds committed
365 366 367 368 369 370
}

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

371
	ei = kmem_cache_alloc(mqueue_inode_cachep, GFP_KERNEL);
Linus Torvalds's avatar
Linus Torvalds committed
372 373 374 375 376
	if (!ei)
		return NULL;
	return &ei->vfs_inode;
}

Nick Piggin's avatar
Nick Piggin committed
377
static void mqueue_i_callback(struct rcu_head *head)
Linus Torvalds's avatar
Linus Torvalds committed
378
{
Nick Piggin's avatar
Nick Piggin committed
379
	struct inode *inode = container_of(head, struct inode, i_rcu);
Linus Torvalds's avatar
Linus Torvalds committed
380 381 382
	kmem_cache_free(mqueue_inode_cachep, MQUEUE_I(inode));
}

Nick Piggin's avatar
Nick Piggin committed
383 384 385 386 387
static void mqueue_destroy_inode(struct inode *inode)
{
	call_rcu(&inode->i_rcu, mqueue_i_callback);
}

388
static void mqueue_evict_inode(struct inode *inode)
Linus Torvalds's avatar
Linus Torvalds committed
389 390 391
{
	struct mqueue_inode_info *info;
	struct user_struct *user;
392
	unsigned long mq_bytes, mq_treesize;
393
	struct ipc_namespace *ipc_ns;
394
	struct msg_msg *msg;
Linus Torvalds's avatar
Linus Torvalds committed
395

396
	clear_inode(inode);
397 398

	if (S_ISDIR(inode->i_mode))
Linus Torvalds's avatar
Linus Torvalds committed
399
		return;
400

401
	ipc_ns = get_ns_from_inode(inode);
Linus Torvalds's avatar
Linus Torvalds committed
402 403
	info = MQUEUE_I(inode);
	spin_lock(&info->lock);
404 405
	while ((msg = msg_get(info)) != NULL)
		free_msg(msg);
406
	kfree(info->node_cache);
Linus Torvalds's avatar
Linus Torvalds committed
407 408
	spin_unlock(&info->lock);

409
	/* Total amount of bytes accounted for the mqueue */
410 411 412 413 414 415 416
	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);

Linus Torvalds's avatar
Linus Torvalds committed
417 418 419 420
	user = info->user;
	if (user) {
		spin_lock(&mq_lock);
		user->mq_bytes -= mq_bytes;
421 422 423 424 425 426 427 428
		/*
		 * 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--;
Linus Torvalds's avatar
Linus Torvalds committed
429 430 431
		spin_unlock(&mq_lock);
		free_uid(user);
	}
432 433
	if (ipc_ns)
		put_ipc_ns(ipc_ns);
Linus Torvalds's avatar
Linus Torvalds committed
434 435
}

436
static int mqueue_create_attr(struct dentry *dentry, umode_t mode, void *arg)
Linus Torvalds's avatar
Linus Torvalds committed
437
{
438
	struct inode *dir = dentry->d_parent->d_inode;
Linus Torvalds's avatar
Linus Torvalds committed
439
	struct inode *inode;
440
	struct mq_attr *attr = arg;
Linus Torvalds's avatar
Linus Torvalds committed
441
	int error;
442
	struct ipc_namespace *ipc_ns;
Linus Torvalds's avatar
Linus Torvalds committed
443 444

	spin_lock(&mq_lock);
445 446 447 448 449
	ipc_ns = __get_ns_from_inode(dir);
	if (!ipc_ns) {
		error = -EACCES;
		goto out_unlock;
	}
450 451 452

	if (ipc_ns->mq_queues_count >= ipc_ns->mq_queues_max &&
	    !capable(CAP_SYS_RESOURCE)) {
Linus Torvalds's avatar
Linus Torvalds committed
453
		error = -ENOSPC;
454
		goto out_unlock;
Linus Torvalds's avatar
Linus Torvalds committed
455
	}
456
	ipc_ns->mq_queues_count++;
Linus Torvalds's avatar
Linus Torvalds committed
457 458
	spin_unlock(&mq_lock);

459
	inode = mqueue_get_inode(dir->i_sb, ipc_ns, mode, attr);
460 461
	if (IS_ERR(inode)) {
		error = PTR_ERR(inode);
Linus Torvalds's avatar
Linus Torvalds committed
462
		spin_lock(&mq_lock);
463 464
		ipc_ns->mq_queues_count--;
		goto out_unlock;
Linus Torvalds's avatar
Linus Torvalds committed
465 466
	}

467
	put_ipc_ns(ipc_ns);
Linus Torvalds's avatar
Linus Torvalds committed
468
	dir->i_size += DIRENT_SIZE;
469
	dir->i_ctime = dir->i_mtime = dir->i_atime = current_time(dir);
Linus Torvalds's avatar
Linus Torvalds committed
470 471 472 473

	d_instantiate(dentry, inode);
	dget(dentry);
	return 0;
474
out_unlock:
Linus Torvalds's avatar
Linus Torvalds committed
475
	spin_unlock(&mq_lock);
476 477
	if (ipc_ns)
		put_ipc_ns(ipc_ns);
Linus Torvalds's avatar
Linus Torvalds committed
478 479 480
	return error;
}

481 482 483 484 485 486
static int mqueue_create(struct inode *dir, struct dentry *dentry,
				umode_t mode, bool excl)
{
	return mqueue_create_attr(dentry, mode, NULL);
}

Linus Torvalds's avatar
Linus Torvalds committed
487 488
static int mqueue_unlink(struct inode *dir, struct dentry *dentry)
{
489
	struct inode *inode = d_inode(dentry);
Linus Torvalds's avatar
Linus Torvalds committed
490

491
	dir->i_ctime = dir->i_mtime = dir->i_atime = current_time(dir);
Linus Torvalds's avatar
Linus Torvalds committed
492
	dir->i_size -= DIRENT_SIZE;
Manfred Spraul's avatar
Manfred Spraul committed
493 494 495
	drop_nlink(inode);
	dput(dentry);
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
496 497 498 499 500 501 502 503 504 505
}

/*
*	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,
506
				size_t count, loff_t *off)
Linus Torvalds's avatar
Linus Torvalds committed
507
{
Al Viro's avatar
Al Viro committed
508
	struct mqueue_inode_info *info = MQUEUE_I(file_inode(filp));
Linus Torvalds's avatar
Linus Torvalds committed
509
	char buffer[FILENT_SIZE];
510
	ssize_t ret;
Linus Torvalds's avatar
Linus Torvalds committed
511 512 513 514 515 516 517 518 519

	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,
520
			pid_vnr(info->notify_owner));
Linus Torvalds's avatar
Linus Torvalds committed
521 522 523
	spin_unlock(&info->lock);
	buffer[sizeof(buffer)-1] = '\0';

524 525 526 527
	ret = simple_read_from_buffer(u_data, count, off, buffer,
				strlen(buffer));
	if (ret <= 0)
		return ret;
Linus Torvalds's avatar
Linus Torvalds committed
528

529
	file_inode(filp)->i_atime = file_inode(filp)->i_ctime = current_time(file_inode(filp));
530
	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
531 532
}

533
static int mqueue_flush_file(struct file *filp, fl_owner_t id)
Linus Torvalds's avatar
Linus Torvalds committed
534
{
Al Viro's avatar
Al Viro committed
535
	struct mqueue_inode_info *info = MQUEUE_I(file_inode(filp));
Linus Torvalds's avatar
Linus Torvalds committed
536 537

	spin_lock(&info->lock);
538
	if (task_tgid(current) == info->notify_owner)
Linus Torvalds's avatar
Linus Torvalds committed
539 540 541 542 543 544
		remove_notification(info);

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

545
static __poll_t mqueue_poll_file(struct file *filp, struct poll_table_struct *poll_tab)
Linus Torvalds's avatar
Linus Torvalds committed
546
{
Al Viro's avatar
Al Viro committed
547
	struct mqueue_inode_info *info = MQUEUE_I(file_inode(filp));
548
	__poll_t retval = 0;
Linus Torvalds's avatar
Linus Torvalds committed
549 550 551 552 553

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

	spin_lock(&info->lock);
	if (info->attr.mq_curmsgs)
554
		retval = EPOLLIN | EPOLLRDNORM;
Linus Torvalds's avatar
Linus Torvalds committed
555 556

	if (info->attr.mq_curmsgs < info->attr.mq_maxmsg)
557
		retval |= EPOLLOUT | EPOLLWRNORM;
Linus Torvalds's avatar
Linus Torvalds committed
558 559 560 561 562 563 564 565 566 567 568 569 570 571
	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) {
572
		if (walk->task->prio <= current->prio) {
Linus Torvalds's avatar
Linus Torvalds committed
573 574 575 576 577 578 579 580 581 582 583 584 585
			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,
586
		    ktime_t *timeout, struct ext_wait_queue *ewp)
587
	__releases(&info->lock)
Linus Torvalds's avatar
Linus Torvalds committed
588 589 590 591 592 593 594
{
	int retval;
	signed long time;

	wq_add(info, sr, ewp);

	for (;;) {
595
		__set_current_state(TASK_INTERRUPTIBLE);
Linus Torvalds's avatar
Linus Torvalds committed
596 597

		spin_unlock(&info->lock);
598 599
		time = schedule_hrtimeout_range_clock(timeout, 0,
			HRTIMER_MODE_ABS, CLOCK_REALTIME);
Linus Torvalds's avatar
Linus Torvalds committed
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

		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)
{
Manfred Spraul's avatar
Manfred Spraul committed
643
	((char *)skb->data)[NOTIFY_COOKIE_LEN-1] = code;
Linus Torvalds's avatar
Linus Torvalds committed
644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664
}

/*
 * 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 */

665
			clear_siginfo(&sig_i);
Linus Torvalds's avatar
Linus Torvalds committed
666 667 668 669
			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;
670 671
			/* map current pid/uid into info->owner's namespaces */
			rcu_read_lock();
672 673
			sig_i.si_pid = task_tgid_nr_ns(current,
						ns_of_pid(info->notify_owner));
674
			sig_i.si_uid = from_kuid_munged(info->notify_user_ns, current_uid());
675
			rcu_read_unlock();
Linus Torvalds's avatar
Linus Torvalds committed
676

677 678
			kill_pid_info(info->notify.sigev_signo,
				      &sig_i, info->notify_owner);
Linus Torvalds's avatar
Linus Torvalds committed
679 680 681
			break;
		case SIGEV_THREAD:
			set_cookie(info->notify_cookie, NOTIFY_WOKENUP);
682
			netlink_sendskb(info->notify_sock, info->notify_cookie);
Linus Torvalds's avatar
Linus Torvalds committed
683 684 685
			break;
		}
		/* after notification unregisters process */
686
		put_pid(info->notify_owner);
687
		put_user_ns(info->notify_user_ns);
688
		info->notify_owner = NULL;
689
		info->notify_user_ns = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
690 691 692 693
	}
	wake_up(&info->wait_q);
}

694
static int prepare_timeout(const struct __kernel_timespec __user *u_abs_timeout,
695
			   struct timespec64 *ts)
Linus Torvalds's avatar
Linus Torvalds committed
696
{
697
	if (get_timespec64(ts, u_abs_timeout))
698
		return -EFAULT;
699
	if (!timespec64_valid(ts))
700 701
		return -EINVAL;
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
702 703 704 705
}

static void remove_notification(struct mqueue_inode_info *info)
{
706
	if (info->notify_owner != NULL &&
Linus Torvalds's avatar
Linus Torvalds committed
707 708
	    info->notify.sigev_notify == SIGEV_THREAD) {
		set_cookie(info->notify_cookie, NOTIFY_REMOVED);
709
		netlink_sendskb(info->notify_sock, info->notify_cookie);
Linus Torvalds's avatar
Linus Torvalds committed
710
	}
711
	put_pid(info->notify_owner);
712
	put_user_ns(info->notify_user_ns);
713
	info->notify_owner = NULL;
714
	info->notify_user_ns = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
715 716
}

717 718
static int prepare_open(struct dentry *dentry, int oflag, int ro,
			umode_t mode, struct filename *name,
719
			struct mq_attr *attr)
Linus Torvalds's avatar
Linus Torvalds committed
720
{
721 722
	static const int oflag2acc[O_ACCMODE] = { MAY_READ, MAY_WRITE,
						  MAY_READ | MAY_WRITE };
723
	int acc;
724

725 726
	if (d_really_is_negative(dentry)) {
		if (!(oflag & O_CREAT))
727
			return -ENOENT;
728 729 730 731 732
		if (ro)
			return ro;
		audit_inode_parent_hidden(name, dentry->d_parent);
		return vfs_mkobj(dentry, mode & ~current_umask(),
				  mqueue_create_attr, attr);
733
	}
734 735 736 737
	/* it already existed */
	audit_inode(name, dentry, 0);
	if ((oflag & (O_CREAT|O_EXCL)) == (O_CREAT|O_EXCL))
		return -EEXIST;
738
	if ((oflag & O_ACCMODE) == (O_RDWR | O_WRONLY))
739
		return -EINVAL;
740
	acc = oflag2acc[oflag & O_ACCMODE];
741
	return inode_permission(d_inode(dentry), acc);
Linus Torvalds's avatar
Linus Torvalds committed
742 743
}

744 745
static int do_mq_open(const char __user *u_name, int oflag, umode_t mode,
		      struct mq_attr *attr)
Linus Torvalds's avatar
Linus Torvalds committed
746
{
747 748
	struct vfsmount *mnt = current->nsproxy->ipc_ns->mq_mnt;
	struct dentry *root = mnt->mnt_root;
749
	struct filename *name;
Al Viro's avatar
Al Viro committed
750
	struct path path;
Linus Torvalds's avatar
Linus Torvalds committed
751
	int fd, error;
752
	int ro;
Linus Torvalds's avatar
Linus Torvalds committed
753

754
	audit_mq_open(oflag, mode, attr);
755

Linus Torvalds's avatar
Linus Torvalds committed
756 757 758
	if (IS_ERR(name = getname(u_name)))
		return PTR_ERR(name);

759
	fd = get_unused_fd_flags(O_CLOEXEC);
Linus Torvalds's avatar
Linus Torvalds committed
760 761 762
	if (fd < 0)
		goto out_putname;

763
	ro = mnt_want_write(mnt);	/* we'll drop it in any case */
Al Viro's avatar
Al Viro committed
764
	inode_lock(d_inode(root));
765
	path.dentry = lookup_one_len(name->name, root, strlen(name->name));
766 767
	if (IS_ERR(path.dentry)) {
		error = PTR_ERR(path.dentry);
768
		goto out_putfd;
Linus Torvalds's avatar
Linus Torvalds committed
769
	}
770
	path.mnt = mntget(mnt);
771 772 773 774 775 776 777
	error = prepare_open(path.dentry, oflag, ro, mode, name, attr);
	if (!error) {
		struct file *file = dentry_open(&path, oflag, current_cred());
		if (!IS_ERR(file))
			fd_install(fd, file);
		else
			error = PTR_ERR(file);
778
	}
779
	path_put(&path);
780
out_putfd:
781 782 783 784
	if (error) {
		put_unused_fd(fd);
		fd = error;
	}
Al Viro's avatar
Al Viro committed
785
	inode_unlock(d_inode(root));
786 787
	if (!ro)
		mnt_drop_write(mnt);
Linus Torvalds's avatar
Linus Torvalds committed
788 789 790 791 792
out_putname:
	putname(name);
	return fd;
}

793 794 795 796 797 798 799 800 801 802
SYSCALL_DEFINE4(mq_open, const char __user *, u_name, int, oflag, umode_t, mode,
		struct mq_attr __user *, u_attr)
{
	struct mq_attr attr;
	if (u_attr && copy_from_user(&attr, u_attr, sizeof(struct mq_attr)))
		return -EFAULT;

	return do_mq_open(u_name, oflag, mode, u_attr ? &attr : NULL);
}

803
SYSCALL_DEFINE1(mq_unlink, const char __user *, u_name)
Linus Torvalds's avatar
Linus Torvalds committed
804 805
{
	int err;
806
	struct filename *name;
Linus Torvalds's avatar
Linus Torvalds committed
807 808
	struct dentry *dentry;
	struct inode *inode = NULL;
809
	struct ipc_namespace *ipc_ns = current->nsproxy->ipc_ns;
810
	struct vfsmount *mnt = ipc_ns->mq_mnt;
Linus Torvalds's avatar
Linus Torvalds committed
811 812 813 814 815

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

816
	audit_inode_parent_hidden(name, mnt->mnt_root);
817 818 819
	err = mnt_want_write(mnt);
	if (err)
		goto out_name;
Al Viro's avatar
Al Viro committed
820
	inode_lock_nested(d_inode(mnt->mnt_root), I_MUTEX_PARENT);
821 822
	dentry = lookup_one_len(name->name, mnt->mnt_root,
				strlen(name->name));
Linus Torvalds's avatar
Linus Torvalds committed
823 824 825 826 827
	if (IS_ERR(dentry)) {
		err = PTR_ERR(dentry);
		goto out_unlock;
	}

828
	inode = d_inode(dentry);
829 830 831
	if (!inode) {
		err = -ENOENT;
	} else {
Al Viro's avatar
Al Viro committed
832
		ihold(inode);
833
		err = vfs_unlink(d_inode(dentry->d_parent), dentry, NULL);
834
	}
Linus Torvalds's avatar
Linus Torvalds committed
835 836 837
	dput(dentry);

out_unlock:
Al Viro's avatar
Al Viro committed
838
	inode_unlock(d_inode(mnt->mnt_root));
Linus Torvalds's avatar
Linus Torvalds committed
839 840
	if (inode)
		iput(inode);
841 842 843
	mnt_drop_write(mnt);
out_name:
	putname(name);
Linus Torvalds's avatar
Linus Torvalds committed
844 845 846 847 848 849 850 851 852 853

	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
854 855 856 857 858 859 860 861 862
 * receiver. The receiver accepts the message and returns without grabbing the
 * queue spinlock:
 *
 * - Set pointer to message.
 * - Queue the receiver task for later wakeup (without the info->lock).
 * - Update its state to STATE_READY. Now the receiver can continue.
 * - Wake up the process after the lock is dropped. Should the process wake up
 *   before this wakeup (due to a timeout or a signal) it will either see
 *   STATE_READY and continue or acquire the lock to check the state again.
Linus Torvalds's avatar
Linus Torvalds committed
863 864 865 866 867 868 869
 *
 * 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).
 */
870 871
static inline void pipelined_send(struct wake_q_head *wake_q,
				  struct mqueue_inode_info *info,
Linus Torvalds's avatar
Linus Torvalds committed
872 873 874 875 876
				  struct msg_msg *message,
				  struct ext_wait_queue *receiver)
{
	receiver->msg = message;
	list_del(&receiver->list);
877 878 879 880 881 882 883 884 885
	wake_q_add(wake_q, receiver->task);
	/*
	 * Rely on the implicit cmpxchg barrier from wake_q_add such
	 * that we can ensure that updating receiver->state is the last
	 * write operation: As once set, the receiver can continue,
	 * and if we don't have the reference count from the wake_q,
	 * yet, at that point we can later have a use-after-free
	 * condition and bogus wakeup.
	 */
Linus Torvalds's avatar
Linus Torvalds committed
886 887 888 889 890
	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). */
891 892
static inline void pipelined_receive(struct wake_q_head *wake_q,
				     struct mqueue_inode_info *info)
Linus Torvalds's avatar
Linus Torvalds committed
893 894 895 896 897 898 899 900
{
	struct ext_wait_queue *sender = wq_get_first_waiter(info, SEND);

	if (!sender) {
		/* for poll */
		wake_up_interruptible(&info->wait_q);
		return;
	}
901 902
	if (msg_insert(sender->msg, info))
		return;
903

Linus Torvalds's avatar
Linus Torvalds committed
904
	list_del(&sender->list);
905
	wake_q_add(wake_q, sender->task);
Linus Torvalds's avatar
Linus Torvalds committed
906 907 908
	sender->state = STATE_READY;
}

909 910
static int do_mq_timedsend(mqd_t mqdes, const char __user *u_msg_ptr,
		size_t msg_len, unsigned int msg_prio,
911
		struct timespec64 *ts)
Linus Torvalds's avatar
Linus Torvalds committed
912
{
913
	struct fd f;
Linus Torvalds's avatar
Linus Torvalds committed
914 915 916 917 918
	struct inode *inode;
	struct ext_wait_queue wait;
	struct ext_wait_queue *receiver;
	struct msg_msg *msg_ptr;
	struct mqueue_inode_info *info;
919
	ktime_t expires, *timeout = NULL;
920
	struct posix_msg_tree_node *new_leaf = NULL;
921
	int ret = 0;
922
	DEFINE_WAKE_Q(wake_q);
Linus Torvalds's avatar
Linus Torvalds committed
923 924 925 926

	if (unlikely(msg_prio >= (unsigned long) MQ_PRIO_MAX))
		return -EINVAL;

927
	if (ts) {
928
		expires = timespec64_to_ktime(*ts);
929 930 931 932
		timeout = &expires;
	}

	audit_mq_sendrecv(mqdes, msg_len, msg_prio, ts);
Linus Torvalds's avatar
Linus Torvalds committed
933

934 935
	f = fdget(mqdes);
	if (unlikely(!f.file)) {
936
		ret = -EBADF;
Linus Torvalds's avatar
Linus Torvalds committed
937
		goto out;
938
	}
Linus Torvalds's avatar
Linus Torvalds committed
939

Al Viro's avatar
Al Viro committed
940
	inode = file_inode(f.file);
941
	if (unlikely(f.file->f_op != &mqueue_file_operations)) {
942
		ret = -EBADF;
Linus Torvalds's avatar
Linus Torvalds committed
943
		goto out_fput;
944
	}
Linus Torvalds's avatar
Linus Torvalds committed
945
	info = MQUEUE_I(inode);
Al Viro's avatar
Al Viro committed
946
	audit_file(f.file);
Linus Torvalds's avatar
Linus Torvalds committed
947

948
	if (unlikely(!(f.file->f_mode & FMODE_WRITE))) {
949
		ret = -EBADF;
Linus Torvalds's avatar
Linus Torvalds committed
950
		goto out_fput;
951
	}
Linus Torvalds's avatar
Linus Torvalds committed
952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967

	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;

968 969 970 971 972 973 974 975
	/*
	 * 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);

Linus Torvalds's avatar
Linus Torvalds committed
976 977
	spin_lock(&info->lock);

978 979 980 981 982 983 984 985 986
	if (!info->node_cache && new_leaf) {
		/* Save our speculative allocation into the cache */
		INIT_LIST_HEAD(&new_leaf->msg_list);
		info->node_cache = new_leaf;
		new_leaf = NULL;
	} else {
		kfree(new_leaf);
	}

Linus Torvalds's avatar
Linus Torvalds committed
987
	if (info->attr.mq_curmsgs == info->attr.mq_maxmsg) {
988
		if (f.file->f_flags & O_NONBLOCK) {
Linus Torvalds's avatar
Linus Torvalds committed
989 990 991 992 993 994
			ret = -EAGAIN;
		} else {
			wait.task = current;
			wait.msg = (void *) msg_ptr;
			wait.state = STATE_NONE;
			ret = wq_sleep(info, SEND, timeout, &wait);
995 996 997 998 999
			/*
			 * wq_sleep must be called with info->lock held, and
			 * returns with the lock released
			 */
			goto out_free;
Linus Torvalds's avatar
Linus Torvalds committed
1000 1001 1002 1003
		}
	} else {
		receiver = wq_get_first_waiter(info, RECV);
		if (receiver) {
1004
			pipelined_send(&wake_q, info, msg_ptr, receiver);
Linus Torvalds's avatar
Linus Torvalds committed
1005 1006
		} else {
			/* adds message to the queue */
1007 1008 1009
			ret = msg_insert(msg_ptr, info);
			if (ret)
				goto out_unlock;
Linus Torvalds's avatar
Linus Torvalds committed
1010 1011 1012
			__do_notify(info);
		}
		inode->i_atime = inode->i_mtime = inode->i_ctime =
1013
				current_time(inode);
Linus Torvalds's avatar
Linus Torvalds committed
1014
	}
1015 1016
out_unlock:
	spin_unlock(&info->lock);
1017
	wake_up_q(&wake_q);
1018 1019 1020
out_free:
	if (ret)
		free_msg(msg_ptr);
Linus Torvalds's avatar
Linus Torvalds committed
1021
out_fput:
1022
	fdput(f);
Linus Torvalds's avatar
Linus Torvalds committed
1023 1024 1025 1026
out:
	return ret;
}

1027 1028
static int do_mq_timedreceive(mqd_t mqdes, char __user *u_msg_ptr,
		size_t msg_len, unsigned int __user *u_msg_prio,
1029
		struct timespec64 *ts)
Linus Torvalds's avatar
Linus Torvalds committed
1030 1031 1032
{
	ssize_t ret;
	struct msg_msg *msg_ptr;
1033
	struct fd f;
Linus Torvalds's avatar
Linus Torvalds committed
1034 1035 1036
	struct inode *inode;
	struct mqueue_inode_info *info;
	struct ext_wait_queue wait;
1037
	ktime_t expires, *timeout = NULL;
1038
	struct posix_msg_tree_node *new_leaf = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
1039

1040
	if (ts) {
1041
		expires = timespec64_to_ktime(*ts);
1042
		timeout = &expires;
Al Viro's avatar
Al Viro committed
1043
	}
1044

1045
	audit_mq_sendrecv(mqdes, msg_len, 0, ts);
Linus Torvalds's avatar
Linus Torvalds committed
1046

1047 1048
	f = fdget(mqdes);
	if (unlikely(!f.file)) {
1049
		ret = -EBADF;
Linus Torvalds's avatar
Linus Torvalds committed
1050
		goto out;
1051
	}
Linus Torvalds's avatar
Linus Torvalds committed
1052

Al Viro's avatar
Al Viro committed
1053
	inode = file_inode(f.file);
1054
	if (unlikely(f.file->f_op != &mqueue_file_operations)) {
1055
		ret = -EBADF;
Linus Torvalds's avatar
Linus Torvalds committed
1056
		goto out_fput;
1057
	}
Linus Torvalds's avatar
Linus Torvalds committed
1058
	info = MQUEUE_I(inode);
Al Viro's avatar
Al Viro committed
1059
	audit_file(f.file);
Linus Torvalds's avatar
Linus Torvalds committed
1060

1061
	if (unlikely(!(f.file->f_mode & FMODE_READ))) {
1062
		ret = -EBADF;
Linus Torvalds's avatar
Linus Torvalds committed
1063
		goto out_fput;
1064
	}
Linus Torvalds's avatar
Linus Torvalds committed
1065 1066 1067 1068 1069 1070 1071

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

1072 1073 1074 1075 1076 1077 1078 1079
	/*
	 * 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);

Linus Torvalds's avatar
Linus Torvalds committed
1080
	spin_lock(&info->lock);
1081 1082 1083 1084 1085 1086 1087 1088 1089

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

Linus Torvalds's avatar
Linus Torvalds committed
1090
	if (info->attr.mq_curmsgs == 0) {
1091
		if (f.file->f_flags & O_NONBLOCK) {
Linus Torvalds's avatar
Linus Torvalds committed
1092 1093 1094 1095 1096 1097 1098 1099 1100
			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 {
1101
		DEFINE_WAKE_Q(wake_q);
1102

Linus Torvalds's avatar
Linus Torvalds committed
1103 1104 1105
		msg_ptr = msg_get(info);

		inode->i_atime = inode->i_mtime = inode->i_ctime =
1106
				current_time(inode);
Linus Torvalds's avatar
Linus Torvalds committed
1107 1108

		/* There is now free space in queue. */
1109
		pipelined_receive(&wake_q, info);
Linus Torvalds's avatar
Linus Torvalds committed
1110
		spin_unlock(&info->lock);
1111
		wake_up_q(&wake_q);
Linus Torvalds's avatar
Linus Torvalds committed
1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123
		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:
1124
	fdput(f);
Linus Torvalds's avatar
Linus Torvalds committed
1125 1126 1127 1128
out:
	return ret;
}

1129 1130
SYSCALL_DEFINE5(mq_timedsend, mqd_t, mqdes, const char __user *, u_msg_ptr,
		size_t, msg_len, unsigned int, msg_prio,
1131
		const struct __kernel_timespec __user *, u_abs_timeout)
1132
{
1133
	struct timespec64 ts, *p = NULL;
1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144
	if (u_abs_timeout) {
		int res = prepare_timeout(u_abs_timeout, &ts);
		if (res)
			return res;
		p = &ts;
	}
	return do_mq_timedsend(mqdes, u_msg_ptr, msg_len, msg_prio, p);
}

SYSCALL_DEFINE5(mq_timedreceive, mqd_t, mqdes, char __user *, u_msg_ptr,
		size_t, msg_len, unsigned int __user *, u_msg_prio,
1145
		const struct __kernel_timespec __user *, u_abs_timeout)
1146
{
1147
	struct timespec64 ts, *p = NULL;
1148 1149 1150 1151 1152 1153 1154 1155 1156
	if (u_abs_timeout) {
		int res = prepare_timeout(u_abs_timeout, &ts);
		if (res)
			return res;
		p = &ts;
	}
	return do_mq_timedreceive(mqdes, u_msg_ptr, msg_len, u_msg_prio, p);
}

Linus Torvalds's avatar
Linus Torvalds committed
1157 1158 1159 1160 1161
/*
 * 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.
 */
1162
static int do_mq_notify(mqd_t mqdes, const struct sigevent *notification)
Linus Torvalds's avatar
Linus Torvalds committed
1163
{
1164 1165
	int ret;
	struct fd f;
Linus Torvalds's avatar
Linus Torvalds committed
1166 1167 1168 1169 1170
	struct sock *sock;
	struct inode *inode;
	struct mqueue_inode_info *info;
	struct sk_buff *nc;

1171
	audit_mq_notify(mqdes, notification);
Linus Torvalds's avatar
Linus Torvalds committed
1172

Al Viro's avatar
Al Viro committed
1173 1174
	nc = NULL;
	sock = NULL;
1175 1176 1177 1178
	if (notification != NULL) {
		if (unlikely(notification->sigev_notify != SIGEV_NONE &&
			     notification->sigev_notify != SIGEV_SIGNAL &&
			     notification->sigev_notify != SIGEV_THREAD))
Linus Torvalds's avatar
Linus Torvalds committed
1179
			return -EINVAL;
1180 1181
		if (notification->sigev_notify == SIGEV_SIGNAL &&
			!valid_signal(notification->sigev_signo)) {
Linus Torvalds's avatar
Linus Torvalds committed
1182 1183
			return -EINVAL;
		}
1184
		if (notification->sigev_notify == SIGEV_THREAD) {
1185 1186
			long timeo;

Linus Torvalds's avatar
Linus Torvalds committed
1187 1188
			/* create the notify skb */
			nc = alloc_skb(NOTIFY_COOKIE_LEN, GFP_KERNEL);
1189 1190
			if (!nc) {
				ret = -ENOMEM;
Linus Torvalds's avatar
Linus Torvalds committed
1191
				goto out;
1192
			}
Linus Torvalds's avatar
Linus Torvalds committed
1193
			if (copy_from_user(nc->data,
1194
					notification->sigev_value.sival_ptr,
Linus Torvalds's avatar
Linus Torvalds committed
1195
					NOTIFY_COOKIE_LEN)) {
1196
				ret = -EFAULT;
Linus Torvalds's avatar
Linus Torvalds committed
1197 1198 1199 1200 1201 1202 1203
				goto out;
			}

			/* TODO: add a header? */
			skb_put(nc, NOTIFY_COOKIE_LEN);
			/* and attach it to the socket */
retry:
1204
			f = fdget(notification->sigev_signo);
1205
			if (!f.file) {
1206
				ret = -EBADF;
Linus Torvalds's avatar
Linus Torvalds committed
1207
				goto out;
1208
			}
1209 1210
			sock = netlink_getsockbyfilp(f.file);
			fdput(f);
Linus Torvalds's avatar
Linus Torvalds committed
1211 1212 1213 1214 1215 1216
			if (IS_ERR(sock)) {
				ret = PTR_ERR(sock);
				sock = NULL;
				goto out;
			}

1217
			timeo = MAX_SCHEDULE_TIMEOUT;
1218
			ret = netlink_attachskb(sock, nc, &timeo, NULL);
1219 1220
			if (ret == 1) {
				sock = NULL;
1221
				goto retry;
1222
			}
Linus Torvalds's avatar
Linus Torvalds committed
1223 1224 1225 1226 1227 1228 1229 1230
			if (ret) {
				sock = NULL;
				nc = NULL;
				goto out;
			}
		}
	}

1231 1232
	f = fdget(mqdes);
	if (!f.file) {
1233
		ret = -EBADF;
Linus Torvalds's avatar
Linus Torvalds committed
1234
		goto out;
1235
	}
Linus Torvalds's avatar
Linus Torvalds committed
1236

Al Viro's avatar
Al Viro committed
1237
	inode = file_inode(f.file);
1238
	if (unlikely(f.file->f_op != &mqueue_file_operations)) {
1239
		ret = -EBADF;
Linus Torvalds's avatar
Linus Torvalds committed
1240
		goto out_fput;
1241
	}
Linus Torvalds's avatar
Linus Torvalds committed
1242 1243 1244 1245
	info = MQUEUE_I(inode);

	ret = 0;
	spin_lock(&info->lock);
1246
	if (notification == NULL) {
1247
		if (info->notify_owner == task_tgid(current)) {
Linus Torvalds's avatar
Linus Torvalds committed
1248
			remove_notification(info);
1249
			inode->i_atime = inode->i_ctime = current_time(inode);
Linus Torvalds's avatar
Linus Torvalds committed
1250
		}
1251
	} else if (info->notify_owner != NULL) {
Linus Torvalds's avatar
Linus Torvalds committed
1252 1253
		ret = -EBUSY;
	} else {
1254
		switch (notification->sigev_notify) {
Linus Torvalds's avatar
Linus Torvalds committed
1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265
		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:
1266 1267
			info->notify.sigev_signo = notification->sigev_signo;
			info->notify.sigev_value = notification->sigev_value;
Linus Torvalds's avatar
Linus Torvalds committed
1268 1269 1270
			info->notify.sigev_notify = SIGEV_SIGNAL;
			break;
		}
1271 1272

		info->notify_owner = get_pid(task_tgid(current));
1273
		info->notify_user_ns = get_user_ns(current_user_ns());
1274
		inode->i_atime = inode->i_ctime = current_time(inode);
Linus Torvalds's avatar
Linus Torvalds committed
1275 1276 1277
	}
	spin_unlock(&info->lock);
out_fput:
1278
	fdput(f);
Linus Torvalds's avatar
Linus Torvalds committed
1279
out:
1280
	if (sock)
Linus Torvalds's avatar
Linus Torvalds committed
1281
		netlink_detachskb(sock, nc);
1282
	else if (nc)
Linus Torvalds's avatar
Linus Torvalds committed
1283
		dev_kfree_skb(nc);
1284

Linus Torvalds's avatar
Linus Torvalds committed
1285 1286 1287
	return ret;
}

1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300
SYSCALL_DEFINE2(mq_notify, mqd_t, mqdes,
		const struct sigevent __user *, u_notification)
{
	struct sigevent n, *p = NULL;
	if (u_notification) {
		if (copy_from_user(&n, u_notification, sizeof(struct sigevent)))
			return -EFAULT;
		p = &n;
	}
	return do_mq_notify(mqdes, p);
}

static int do_mq_getsetattr(int mqdes, struct mq_attr *new, struct mq_attr *old)
Linus Torvalds's avatar
Linus Torvalds committed
1301
{
1302
	struct fd f;
Linus Torvalds's avatar
Linus Torvalds committed
1303 1304 1305
	struct inode *inode;
	struct mqueue_inode_info *