audit.c 24.4 KB
Newer Older
1
/* audit.c -- Auditing support
L
Linus Torvalds 已提交
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
 * Gateway between the kernel (e.g., selinux) and the user-space audit daemon.
 * System-call specific features have moved to auditsc.c
 *
 * Copyright 2003-2004 Red Hat Inc., Durham, North Carolina.
 * All Rights Reserved.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 *
 * Written by Rickard E. (Rik) Faith <faith@redhat.com>
 *
 * Goals: 1) Integrate fully with SELinux.
 *	  2) Minimal run-time overhead:
 *	     a) Minimal when syscall auditing is disabled (audit_enable=0).
 *	     b) Small when syscall auditing is enabled and no audit record
 *		is generated (defer as much work as possible to record
 *		generation time):
 *		i) context is allocated,
 *		ii) names from getname are stored without a copy, and
 *		iii) inode information stored from path_lookup.
 *	  3) Ability to disable syscall auditing at boot time (audit=0).
 *	  4) Usable by other parts of the kernel (if audit_log* is called,
 *	     then a syscall record will be generated automatically for the
 *	     current syscall).
 *	  5) Netlink interface to user-space.
 *	  6) Support low-overhead kernel-based filtering to minimize the
 *	     information that must be passed to user-space.
 *
41
 * Example user-space utilities: http://people.redhat.com/sgrubb/audit/
L
Linus Torvalds 已提交
42 43 44 45 46 47 48
 */

#include <linux/init.h>
#include <asm/atomic.h>
#include <asm/types.h>
#include <linux/mm.h>
#include <linux/module.h>
49 50
#include <linux/err.h>
#include <linux/kthread.h>
L
Linus Torvalds 已提交
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72

#include <linux/audit.h>

#include <net/sock.h>
#include <linux/skbuff.h>
#include <linux/netlink.h>

/* No auditing will take place until audit_initialized != 0.
 * (Initialization happens after skb_init is called.) */
static int	audit_initialized;

/* No syscall auditing will take place unless audit_enabled != 0. */
int		audit_enabled;

/* Default state when kernel boots without any parameters. */
static int	audit_default;

/* If auditing cannot proceed, audit_failure selects what happens. */
static int	audit_failure = AUDIT_FAIL_PRINTK;

/* If audit records are to be written to the netlink socket, audit_pid
 * contains the (non-zero) pid. */
73
int		audit_pid;
L
Linus Torvalds 已提交
74 75 76 77 78 79 80 81 82

/* If audit_limit is non-zero, limit the rate of sending audit records
 * to that number per second.  This prevents DoS attacks, but results in
 * audit records being dropped. */
static int	audit_rate_limit;

/* Number of outstanding audit_buffers allowed. */
static int	audit_backlog_limit = 64;

83 84 85 86
/* The identity of the user shutting down the audit system. */
uid_t		audit_sig_uid = -1;
pid_t		audit_sig_pid = -1;

L
Linus Torvalds 已提交
87 88 89 90 91 92 93 94 95 96 97 98
/* Records can be lost in several ways:
   0) [suppressed in audit_alloc]
   1) out of memory in audit_log_start [kmalloc of struct audit_buffer]
   2) out of memory in audit_log_move [alloc_skb]
   3) suppressed due to audit_rate_limit
   4) suppressed due to audit_backlog_limit
*/
static atomic_t    audit_lost = ATOMIC_INIT(0);

/* The netlink socket. */
static struct sock *audit_sock;

99
/* The audit_freelist is a list of pre-allocated audit buffers (if more
L
Linus Torvalds 已提交
100 101 102 103 104 105
 * than AUDIT_MAXFREE are in use, the audit buffer is freed instead of
 * being placed on the freelist). */
static DEFINE_SPINLOCK(audit_freelist_lock);
static int	   audit_freelist_count = 0;
static LIST_HEAD(audit_freelist);

106 107 108
static struct sk_buff_head audit_skb_queue;
static struct task_struct *kauditd_task;
static DECLARE_WAIT_QUEUE_HEAD(kauditd_wait);
109
static DECLARE_WAIT_QUEUE_HEAD(audit_backlog_wait);
110

L
Linus Torvalds 已提交
111
/* The netlink socket is only to be read by 1 CPU, which lets us assume
S
Steve Grubb 已提交
112
 * that list additions and deletions never happen simultaneously in
L
Linus Torvalds 已提交
113
 * auditsc.c */
114
DECLARE_MUTEX(audit_netlink_sem);
L
Linus Torvalds 已提交
115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131

/* AUDIT_BUFSIZ is the size of the temporary buffer used for formatting
 * audit records.  Since printk uses a 1024 byte buffer, this buffer
 * should be at least that large. */
#define AUDIT_BUFSIZ 1024

/* AUDIT_MAXFREE is the number of empty audit_buffers we keep on the
 * audit_freelist.  Doing so eliminates many kmalloc/kfree calls. */
#define AUDIT_MAXFREE  (2*NR_CPUS)

/* The audit_buffer is used when formatting an audit record.  The caller
 * locks briefly to get the record off the freelist or to allocate the
 * buffer, and locks briefly to send the buffer to the netlink layer or
 * to place it on a transmit queue.  Multiple audit_buffers can be in
 * use simultaneously. */
struct audit_buffer {
	struct list_head     list;
132
	struct sk_buff       *skb;	/* formatted skb ready to send */
L
Linus Torvalds 已提交
133
	struct audit_context *ctx;	/* NULL or associated context */
134
	int		     gfp_mask;
L
Linus Torvalds 已提交
135 136
};

137 138 139 140 141 142
static void audit_set_pid(struct audit_buffer *ab, pid_t pid)
{
	struct nlmsghdr *nlh = (struct nlmsghdr *)ab->skb->data;
	nlh->nlmsg_pid = pid;
}

L
Linus Torvalds 已提交
143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
struct audit_entry {
	struct list_head  list;
	struct audit_rule rule;
};

static void audit_panic(const char *message)
{
	switch (audit_failure)
	{
	case AUDIT_FAIL_SILENT:
		break;
	case AUDIT_FAIL_PRINTK:
		printk(KERN_ERR "audit: %s\n", message);
		break;
	case AUDIT_FAIL_PANIC:
		panic("audit: %s\n", message);
		break;
	}
}

static inline int audit_rate_check(void)
{
	static unsigned long	last_check = 0;
	static int		messages   = 0;
	static DEFINE_SPINLOCK(lock);
	unsigned long		flags;
	unsigned long		now;
	unsigned long		elapsed;
	int			retval	   = 0;

	if (!audit_rate_limit) return 1;

	spin_lock_irqsave(&lock, flags);
	if (++messages < audit_rate_limit) {
		retval = 1;
	} else {
		now     = jiffies;
		elapsed = now - last_check;
		if (elapsed > HZ) {
			last_check = now;
			messages   = 0;
			retval     = 1;
		}
	}
	spin_unlock_irqrestore(&lock, flags);

	return retval;
}

/* Emit at least 1 message per second, even if audit_rate_check is
 * throttling. */
void audit_log_lost(const char *message)
{
	static unsigned long	last_msg = 0;
	static DEFINE_SPINLOCK(lock);
	unsigned long		flags;
	unsigned long		now;
	int			print;

	atomic_inc(&audit_lost);

	print = (audit_failure == AUDIT_FAIL_PANIC || !audit_rate_limit);

	if (!print) {
		spin_lock_irqsave(&lock, flags);
		now = jiffies;
		if (now - last_msg > HZ) {
			print = 1;
			last_msg = now;
		}
		spin_unlock_irqrestore(&lock, flags);
	}

	if (print) {
		printk(KERN_WARNING
218
		       "audit: audit_lost=%d audit_rate_limit=%d audit_backlog_limit=%d\n",
L
Linus Torvalds 已提交
219 220 221 222 223 224 225 226
		       atomic_read(&audit_lost),
		       audit_rate_limit,
		       audit_backlog_limit);
		audit_panic(message);
	}

}

227
static int audit_set_rate_limit(int limit, uid_t loginuid)
L
Linus Torvalds 已提交
228 229 230
{
	int old		 = audit_rate_limit;
	audit_rate_limit = limit;
231
	audit_log(NULL, GFP_KERNEL, AUDIT_CONFIG_CHANGE, 
232
			"audit_rate_limit=%d old=%d by auid=%u",
233
			audit_rate_limit, old, loginuid);
L
Linus Torvalds 已提交
234 235 236
	return old;
}

237
static int audit_set_backlog_limit(int limit, uid_t loginuid)
L
Linus Torvalds 已提交
238 239 240
{
	int old		 = audit_backlog_limit;
	audit_backlog_limit = limit;
241
	audit_log(NULL, GFP_KERNEL, AUDIT_CONFIG_CHANGE,
242
			"audit_backlog_limit=%d old=%d by auid=%u",
243
			audit_backlog_limit, old, loginuid);
L
Linus Torvalds 已提交
244 245 246
	return old;
}

247
static int audit_set_enabled(int state, uid_t loginuid)
L
Linus Torvalds 已提交
248 249 250 251 252
{
	int old		 = audit_enabled;
	if (state != 0 && state != 1)
		return -EINVAL;
	audit_enabled = state;
253
	audit_log(NULL, GFP_KERNEL, AUDIT_CONFIG_CHANGE,
254
			"audit_enabled=%d old=%d by auid=%u",
255
			audit_enabled, old, loginuid);
L
Linus Torvalds 已提交
256 257 258
	return old;
}

259
static int audit_set_failure(int state, uid_t loginuid)
L
Linus Torvalds 已提交
260 261 262 263 264 265 266
{
	int old		 = audit_failure;
	if (state != AUDIT_FAIL_SILENT
	    && state != AUDIT_FAIL_PRINTK
	    && state != AUDIT_FAIL_PANIC)
		return -EINVAL;
	audit_failure = state;
267
	audit_log(NULL, GFP_KERNEL, AUDIT_CONFIG_CHANGE,
268
			"audit_failure=%d old=%d by auid=%u",
269
			audit_failure, old, loginuid);
L
Linus Torvalds 已提交
270 271 272
	return old;
}

273 274 275 276 277 278
int kauditd_thread(void *dummy)
{
	struct sk_buff *skb;

	while (1) {
		skb = skb_dequeue(&audit_skb_queue);
279
		wake_up(&audit_backlog_wait);
280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305
		if (skb) {
			if (audit_pid) {
				int err = netlink_unicast(audit_sock, skb, audit_pid, 0);
				if (err < 0) {
					BUG_ON(err != -ECONNREFUSED); /* Shoudn't happen */
					printk(KERN_ERR "audit: *NO* daemon at audit_pid=%d\n", audit_pid);
					audit_pid = 0;
				}
			} else {
				printk(KERN_ERR "%s\n", skb->data + NLMSG_SPACE(0));
				kfree_skb(skb);
			}
		} else {
			DECLARE_WAITQUEUE(wait, current);
			set_current_state(TASK_INTERRUPTIBLE);
			add_wait_queue(&kauditd_wait, &wait);

			if (!skb_queue_len(&audit_skb_queue))
				schedule();

			__set_current_state(TASK_RUNNING);
			remove_wait_queue(&kauditd_wait, &wait);
		}
	}
}

L
Linus Torvalds 已提交
306 307 308 309 310 311 312 313 314 315 316 317
void audit_send_reply(int pid, int seq, int type, int done, int multi,
		      void *payload, int size)
{
	struct sk_buff	*skb;
	struct nlmsghdr	*nlh;
	int		len = NLMSG_SPACE(size);
	void		*data;
	int		flags = multi ? NLM_F_MULTI : 0;
	int		t     = done  ? NLMSG_DONE  : type;

	skb = alloc_skb(len, GFP_KERNEL);
	if (!skb)
318
		return;
L
Linus Torvalds 已提交
319

320
	nlh		 = NLMSG_PUT(skb, pid, seq, t, size);
L
Linus Torvalds 已提交
321 322 323
	nlh->nlmsg_flags = flags;
	data		 = NLMSG_DATA(nlh);
	memcpy(data, payload, size);
324 325 326 327

	/* Ignore failure. It'll only happen if the sender goes away,
	   because our timeout is set to infinite. */
	netlink_unicast(audit_sock, skb, pid, 0);
L
Linus Torvalds 已提交
328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348
	return;

nlmsg_failure:			/* Used by NLMSG_PUT */
	if (skb)
		kfree_skb(skb);
}

/*
 * Check for appropriate CAP_AUDIT_ capabilities on incoming audit
 * control messages.
 */
static int audit_netlink_ok(kernel_cap_t eff_cap, u16 msg_type)
{
	int err = 0;

	switch (msg_type) {
	case AUDIT_GET:
	case AUDIT_LIST:
	case AUDIT_SET:
	case AUDIT_ADD:
	case AUDIT_DEL:
349
	case AUDIT_SIGNAL_INFO:
L
Linus Torvalds 已提交
350 351 352
		if (!cap_raised(eff_cap, CAP_AUDIT_CONTROL))
			err = -EPERM;
		break;
353
	case AUDIT_USER:
354
	case AUDIT_FIRST_USER_MSG...AUDIT_LAST_USER_MSG:
L
Linus Torvalds 已提交
355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370
		if (!cap_raised(eff_cap, CAP_AUDIT_WRITE))
			err = -EPERM;
		break;
	default:  /* bad msg */
		err = -EINVAL;
	}

	return err;
}

static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh)
{
	u32			uid, pid, seq;
	void			*data;
	struct audit_status	*status_get, status_set;
	int			err;
371
	struct audit_buffer	*ab;
L
Linus Torvalds 已提交
372
	u16			msg_type = nlh->nlmsg_type;
373
	uid_t			loginuid; /* loginuid of sender */
374
	struct audit_sig_info   sig_data;
L
Linus Torvalds 已提交
375 376 377 378 379

	err = audit_netlink_ok(NETLINK_CB(skb).eff_cap, msg_type);
	if (err)
		return err;

380 381 382 383 384 385 386 387 388
	/* As soon as there's any sign of userspace auditd, start kauditd to talk to it */
	if (!kauditd_task)
		kauditd_task = kthread_run(kauditd_thread, NULL, "kauditd");
	if (IS_ERR(kauditd_task)) {
		err = PTR_ERR(kauditd_task);
		kauditd_task = NULL;
		return err;
	}

L
Linus Torvalds 已提交
389 390
	pid  = NETLINK_CREDS(skb)->pid;
	uid  = NETLINK_CREDS(skb)->uid;
391
	loginuid = NETLINK_CB(skb).loginuid;
L
Linus Torvalds 已提交
392 393 394 395 396 397 398 399 400 401 402
	seq  = nlh->nlmsg_seq;
	data = NLMSG_DATA(nlh);

	switch (msg_type) {
	case AUDIT_GET:
		status_set.enabled	 = audit_enabled;
		status_set.failure	 = audit_failure;
		status_set.pid		 = audit_pid;
		status_set.rate_limit	 = audit_rate_limit;
		status_set.backlog_limit = audit_backlog_limit;
		status_set.lost		 = atomic_read(&audit_lost);
403
		status_set.backlog	 = skb_queue_len(&audit_skb_queue);
L
Linus Torvalds 已提交
404 405 406 407 408 409 410 411
		audit_send_reply(NETLINK_CB(skb).pid, seq, AUDIT_GET, 0, 0,
				 &status_set, sizeof(status_set));
		break;
	case AUDIT_SET:
		if (nlh->nlmsg_len < sizeof(struct audit_status))
			return -EINVAL;
		status_get   = (struct audit_status *)data;
		if (status_get->mask & AUDIT_STATUS_ENABLED) {
412
			err = audit_set_enabled(status_get->enabled, loginuid);
L
Linus Torvalds 已提交
413 414 415
			if (err < 0) return err;
		}
		if (status_get->mask & AUDIT_STATUS_FAILURE) {
416
			err = audit_set_failure(status_get->failure, loginuid);
L
Linus Torvalds 已提交
417 418 419 420 421
			if (err < 0) return err;
		}
		if (status_get->mask & AUDIT_STATUS_PID) {
			int old   = audit_pid;
			audit_pid = status_get->pid;
422
			audit_log(NULL, GFP_KERNEL, AUDIT_CONFIG_CHANGE,
423
				"audit_pid=%d old=%d by auid=%u",
424
				  audit_pid, old, loginuid);
L
Linus Torvalds 已提交
425 426
		}
		if (status_get->mask & AUDIT_STATUS_RATE_LIMIT)
427
			audit_set_rate_limit(status_get->rate_limit, loginuid);
L
Linus Torvalds 已提交
428
		if (status_get->mask & AUDIT_STATUS_BACKLOG_LIMIT)
429 430
			audit_set_backlog_limit(status_get->backlog_limit,
							loginuid);
L
Linus Torvalds 已提交
431
		break;
432
	case AUDIT_USER:
433
	case AUDIT_FIRST_USER_MSG...AUDIT_LAST_USER_MSG:
434 435 436
		if (!audit_enabled && msg_type != AUDIT_USER_AVC)
			return 0;

437
		err = audit_filter_user(&NETLINK_CB(skb), msg_type);
438 439
		if (err == 1) {
			err = 0;
440
			ab = audit_log_start(NULL, GFP_KERNEL, msg_type);
441 442 443 444 445 446 447
			if (ab) {
				audit_log_format(ab,
						 "user pid=%d uid=%u auid=%u msg='%.1024s'",
						 pid, uid, loginuid, (char *)data);
				audit_set_pid(ab, pid);
				audit_log_end(ab);
			}
448
		}
L
Linus Torvalds 已提交
449 450 451 452 453 454 455 456
		break;
	case AUDIT_ADD:
	case AUDIT_DEL:
		if (nlh->nlmsg_len < sizeof(struct audit_rule))
			return -EINVAL;
		/* fallthrough */
	case AUDIT_LIST:
		err = audit_receive_filter(nlh->nlmsg_type, NETLINK_CB(skb).pid,
457
					   uid, seq, data, loginuid);
L
Linus Torvalds 已提交
458
		break;
459 460 461 462 463 464
	case AUDIT_SIGNAL_INFO:
		sig_data.uid = audit_sig_uid;
		sig_data.pid = audit_sig_pid;
		audit_send_reply(NETLINK_CB(skb).pid, seq, AUDIT_SIGNAL_INFO, 
				0, 0, &sig_data, sizeof(sig_data));
		break;
L
Linus Torvalds 已提交
465 466 467 468 469 470 471 472 473 474 475
	default:
		err = -EINVAL;
		break;
	}

	return err < 0 ? err : 0;
}

/* Get message from skb (based on rtnetlink_rcv_skb).  Each message is
 * processed by audit_receive_msg.  Malformed skbs with wrong length are
 * discarded silently.  */
476
static void audit_receive_skb(struct sk_buff *skb)
L
Linus Torvalds 已提交
477 478 479 480 481 482 483 484
{
	int		err;
	struct nlmsghdr	*nlh;
	u32		rlen;

	while (skb->len >= NLMSG_SPACE(0)) {
		nlh = (struct nlmsghdr *)skb->data;
		if (nlh->nlmsg_len < sizeof(*nlh) || skb->len < nlh->nlmsg_len)
485
			return;
L
Linus Torvalds 已提交
486 487 488 489 490 491 492 493 494 495 496 497 498 499 500
		rlen = NLMSG_ALIGN(nlh->nlmsg_len);
		if (rlen > skb->len)
			rlen = skb->len;
		if ((err = audit_receive_msg(skb, nlh))) {
			netlink_ack(skb, nlh, err);
		} else if (nlh->nlmsg_flags & NLM_F_ACK)
			netlink_ack(skb, nlh, 0);
		skb_pull(skb, rlen);
	}
}

/* Receive messages from netlink socket. */
static void audit_receive(struct sock *sk, int length)
{
	struct sk_buff  *skb;
501
	unsigned int qlen;
L
Linus Torvalds 已提交
502

503
	down(&audit_netlink_sem);
L
Linus Torvalds 已提交
504

505 506 507 508
	for (qlen = skb_queue_len(&sk->sk_receive_queue); qlen; qlen--) {
		skb = skb_dequeue(&sk->sk_receive_queue);
		audit_receive_skb(skb);
		kfree_skb(skb);
L
Linus Torvalds 已提交
509 510 511 512 513 514 515 516 517 518 519 520 521 522
	}
	up(&audit_netlink_sem);
}


/* Initialize audit support at boot time. */
static int __init audit_init(void)
{
	printk(KERN_INFO "audit: initializing netlink socket (%s)\n",
	       audit_default ? "enabled" : "disabled");
	audit_sock = netlink_kernel_create(NETLINK_AUDIT, audit_receive);
	if (!audit_sock)
		audit_panic("cannot initialize netlink socket");

523 524
	audit_sock->sk_sndtimeo = MAX_SCHEDULE_TIMEOUT;
	skb_queue_head_init(&audit_skb_queue);
L
Linus Torvalds 已提交
525 526
	audit_initialized = 1;
	audit_enabled = audit_default;
527
	audit_log(NULL, GFP_KERNEL, AUDIT_KERNEL, "initialized");
L
Linus Torvalds 已提交
528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545
	return 0;
}
__initcall(audit_init);

/* Process kernel command-line parameter at boot time.  audit=0 or audit=1. */
static int __init audit_enable(char *str)
{
	audit_default = !!simple_strtol(str, NULL, 0);
	printk(KERN_INFO "audit: %s%s\n",
	       audit_default ? "enabled" : "disabled",
	       audit_initialized ? "" : " (after initialization)");
	if (audit_initialized)
		audit_enabled = audit_default;
	return 0;
}

__setup("audit=", audit_enable);

546 547 548 549
static void audit_buffer_free(struct audit_buffer *ab)
{
	unsigned long flags;

550 551 552
	if (!ab)
		return;

553 554
	if (ab->skb)
		kfree_skb(ab->skb);
555

556 557 558 559 560 561 562 563
	spin_lock_irqsave(&audit_freelist_lock, flags);
	if (++audit_freelist_count > AUDIT_MAXFREE)
		kfree(ab);
	else
		list_add(&ab->list, &audit_freelist);
	spin_unlock_irqrestore(&audit_freelist_lock, flags);
}

564 565
static struct audit_buffer * audit_buffer_alloc(struct audit_context *ctx,
						int gfp_mask, int type)
566 567 568
{
	unsigned long flags;
	struct audit_buffer *ab = NULL;
569
	struct nlmsghdr *nlh;
570 571 572 573 574 575 576 577 578 579 580

	spin_lock_irqsave(&audit_freelist_lock, flags);
	if (!list_empty(&audit_freelist)) {
		ab = list_entry(audit_freelist.next,
				struct audit_buffer, list);
		list_del(&ab->list);
		--audit_freelist_count;
	}
	spin_unlock_irqrestore(&audit_freelist_lock, flags);

	if (!ab) {
581
		ab = kmalloc(sizeof(*ab), gfp_mask);
582
		if (!ab)
583
			goto err;
584
	}
585

586
	ab->skb = alloc_skb(AUDIT_BUFSIZ, gfp_mask);
587
	if (!ab->skb)
588 589
		goto err;

590
	ab->ctx = ctx;
591
	ab->gfp_mask = gfp_mask;
592 593 594 595 596
	nlh = (struct nlmsghdr *)skb_put(ab->skb, NLMSG_SPACE(0));
	nlh->nlmsg_type = type;
	nlh->nlmsg_flags = 0;
	nlh->nlmsg_pid = 0;
	nlh->nlmsg_seq = 0;
597
	return ab;
598 599 600
err:
	audit_buffer_free(ab);
	return NULL;
601
}
L
Linus Torvalds 已提交
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
/* Compute a serial number for the audit record.  Audit records are
 * written to user-space as soon as they are generated, so a complete
 * audit record may be written in several pieces.  The timestamp of the
 * record and this serial number are used by the user-space tools to
 * determine which pieces belong to the same audit record.  The
 * (timestamp,serial) tuple is unique for each syscall and is live from
 * syscall entry to syscall exit.
 *
 * Atomic values are only guaranteed to be 24-bit, so we count down.
 *
 * NOTE: Another possibility is to store the formatted records off the
 * audit context (for those records that have a context), and emit them
 * all at syscall exit.  However, this could delay the reporting of
 * significant errors until syscall exit (or never, if the system
 * halts). */
unsigned int audit_serial(void)
{
	static atomic_t serial = ATOMIC_INIT(0xffffff);
	unsigned int a, b;

	do {
		a = atomic_read(&serial);
		if (atomic_dec_and_test(&serial))
			atomic_set(&serial, 0xffffff);
		b = atomic_read(&serial);
	} while (b != a - 1);

	return 0xffffff - b;
}

static inline void audit_get_stamp(struct audit_context *ctx, 
				   struct timespec *t, unsigned int *serial)
{
	if (ctx)
		auditsc_get_stamp(ctx, t, serial);
	else {
		*t = CURRENT_TIME;
		*serial = audit_serial();
	}
}

L
Linus Torvalds 已提交
644 645 646 647 648 649
/* Obtain an audit buffer.  This routine does locking to obtain the
 * audit buffer, but then no locking is required for calls to
 * audit_log_*format.  If the tsk is a task that is currently in a
 * syscall, then the syscall is marked as auditable and an audit record
 * will be written at syscall exit.  If there is no associated task, tsk
 * should be NULL. */
650 651 652

struct audit_buffer *audit_log_start(struct audit_context *ctx, int gfp_mask,
				     int type)
L
Linus Torvalds 已提交
653 654 655
{
	struct audit_buffer	*ab	= NULL;
	struct timespec		t;
656
	unsigned int		serial;
657
	int reserve;
L
Linus Torvalds 已提交
658 659 660 661

	if (!audit_initialized)
		return NULL;

662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685
	if (gfp_mask & __GFP_WAIT)
		reserve = 0;
	else
		reserve = 5; /* Allow atomic callers to go up to five 
				entries over the normal backlog limit */

	while (audit_backlog_limit
	       && skb_queue_len(&audit_skb_queue) > audit_backlog_limit + reserve) {
		if (gfp_mask & __GFP_WAIT) {
			int ret = 1;
			/* Wait for auditd to drain the queue a little */
			DECLARE_WAITQUEUE(wait, current);
			set_current_state(TASK_INTERRUPTIBLE);
			add_wait_queue(&audit_backlog_wait, &wait);

			if (audit_backlog_limit &&
			    skb_queue_len(&audit_skb_queue) > audit_backlog_limit)
				ret = schedule_timeout(HZ * 60);

			__set_current_state(TASK_RUNNING);
			remove_wait_queue(&audit_backlog_wait, &wait);
			if (ret)
				continue;
		}
686 687 688 689 690 691 692 693 694 695
		if (audit_rate_check())
			printk(KERN_WARNING
			       "audit: audit_backlog=%d > "
			       "audit_backlog_limit=%d\n",
			       skb_queue_len(&audit_skb_queue),
			       audit_backlog_limit);
		audit_log_lost("backlog limit exceeded");
		return NULL;
	}

696
	ab = audit_buffer_alloc(ctx, gfp_mask, type);
L
Linus Torvalds 已提交
697 698 699 700 701
	if (!ab) {
		audit_log_lost("out of memory in audit_log_start");
		return NULL;
	}

702
	audit_get_stamp(ab->ctx, &t, &serial);
703

L
Linus Torvalds 已提交
704 705 706 707 708
	audit_log_format(ab, "audit(%lu.%03lu:%u): ",
			 t.tv_sec, t.tv_nsec/1000000, serial);
	return ab;
}

709
/**
710
 * audit_expand - expand skb in the audit buffer
711 712 713 714 715
 * @ab: audit_buffer
 *
 * Returns 0 (no space) on failed expansion, or available space if
 * successful.
 */
716
static inline int audit_expand(struct audit_buffer *ab, int extra)
717
{
718
	struct sk_buff *skb = ab->skb;
719
	int ret = pskb_expand_head(skb, skb_headroom(skb), extra,
720
				   ab->gfp_mask);
721 722
	if (ret < 0) {
		audit_log_lost("out of memory in audit_expand");
723
		return 0;
724 725
	}
	return skb_tailroom(skb);
726
}
L
Linus Torvalds 已提交
727 728 729 730 731 732 733 734 735

/* Format an audit message into the audit buffer.  If there isn't enough
 * room in the audit buffer, more room will be allocated and vsnprint
 * will be called a second time.  Currently, we assume that a printk
 * can't format message larger than 1024 bytes, so we don't either. */
static void audit_log_vformat(struct audit_buffer *ab, const char *fmt,
			      va_list args)
{
	int len, avail;
736
	struct sk_buff *skb;
D
David Woodhouse 已提交
737
	va_list args2;
L
Linus Torvalds 已提交
738 739 740 741

	if (!ab)
		return;

742 743 744 745
	BUG_ON(!ab->skb);
	skb = ab->skb;
	avail = skb_tailroom(skb);
	if (avail == 0) {
746
		avail = audit_expand(ab, AUDIT_BUFSIZ);
747 748
		if (!avail)
			goto out;
L
Linus Torvalds 已提交
749
	}
D
David Woodhouse 已提交
750
	va_copy(args2, args);
751
	len = vsnprintf(skb->tail, avail, fmt, args);
L
Linus Torvalds 已提交
752 753 754 755
	if (len >= avail) {
		/* The printk buffer is 1024 bytes long, so if we get
		 * here and AUDIT_BUFSIZ is at least 1024, then we can
		 * log everything that printk could have logged. */
D
David Woodhouse 已提交
756
		avail = audit_expand(ab, max_t(unsigned, AUDIT_BUFSIZ, 1+len-avail));
757 758
		if (!avail)
			goto out;
D
David Woodhouse 已提交
759
		len = vsnprintf(skb->tail, avail, fmt, args2);
L
Linus Torvalds 已提交
760
	}
761 762
	if (len > 0)
		skb_put(skb, len);
763 764
out:
	return;
L
Linus Torvalds 已提交
765 766 767 768 769 770 771 772 773 774 775 776 777 778 779
}

/* Format a message into the audit buffer.  All the work is done in
 * audit_log_vformat. */
void audit_log_format(struct audit_buffer *ab, const char *fmt, ...)
{
	va_list args;

	if (!ab)
		return;
	va_start(args, fmt);
	audit_log_vformat(ab, fmt, args);
	va_end(args);
}

780 781 782 783
/* This function will take the passed buf and convert it into a string of
 * ascii hex digits. The new string is placed onto the skb. */
void audit_log_hex(struct audit_buffer *ab, const unsigned char *buf, 
		size_t len)
784
{
785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800
	int i, avail, new_len;
	unsigned char *ptr;
	struct sk_buff *skb;
	static const unsigned char *hex = "0123456789ABCDEF";

	BUG_ON(!ab->skb);
	skb = ab->skb;
	avail = skb_tailroom(skb);
	new_len = len<<1;
	if (new_len >= avail) {
		/* Round the buffer request up to the next multiple */
		new_len = AUDIT_BUFSIZ*(((new_len-avail)/AUDIT_BUFSIZ) + 1);
		avail = audit_expand(ab, new_len);
		if (!avail)
			return;
	}
801

802 803 804 805 806 807 808
	ptr = skb->tail;
	for (i=0; i<len; i++) {
		*ptr++ = hex[(buf[i] & 0xF0)>>4]; /* Upper nibble */
		*ptr++ = hex[buf[i] & 0x0F];	  /* Lower nibble */
	}
	*ptr = 0;
	skb_put(skb, len << 1); /* new string is twice the old string */
809 810
}

811 812 813 814
/* This code will escape a string that is passed to it if the string
 * contains a control character, unprintable character, double quote mark, 
 * or a space. Unescaped strings will start and end with a double quote mark.
 * Strings that are escaped are printed in hex (2 digits per char). */
815 816
void audit_log_untrustedstring(struct audit_buffer *ab, const char *string)
{
817
	const unsigned char *p = string;
818 819

	while (*p) {
820
		if (*p == '"' || *p < 0x21 || *p > 0x7f) {
821 822 823 824 825 826 827 828
			audit_log_hex(ab, string, strlen(string));
			return;
		}
		p++;
	}
	audit_log_format(ab, "\"%s\"", string);
}

829
/* This is a helper-function to print the escaped d_path */
L
Linus Torvalds 已提交
830 831 832
void audit_log_d_path(struct audit_buffer *ab, const char *prefix,
		      struct dentry *dentry, struct vfsmount *vfsmnt)
{
833
	char *p, *path;
L
Linus Torvalds 已提交
834

835 836
	if (prefix)
		audit_log_format(ab, " %s", prefix);
L
Linus Torvalds 已提交
837

838
	/* We will allow 11 spaces for ' (deleted)' to be appended */
839
	path = kmalloc(PATH_MAX+11, ab->gfp_mask);
840 841 842
	if (!path) {
		audit_log_format(ab, "<no memory>");
		return;
L
Linus Torvalds 已提交
843
	}
844 845 846 847 848 849 850
	p = d_path(dentry, vfsmnt, path, PATH_MAX+11);
	if (IS_ERR(p)) { /* Should never happen since we send PATH_MAX */
		/* FIXME: can we save some information here? */
		audit_log_format(ab, "<too long>");
	} else 
		audit_log_untrustedstring(ab, p);
	kfree(path);
L
Linus Torvalds 已提交
851 852 853 854 855 856
}

/* The netlink_* functions cannot be called inside an irq context, so
 * the audit buffer is places on a queue and a tasklet is scheduled to
 * remove them from the queue outside the irq context.  May be called in
 * any context. */
857
void audit_log_end(struct audit_buffer *ab)
L
Linus Torvalds 已提交
858 859 860 861 862 863
{
	if (!ab)
		return;
	if (!audit_rate_check()) {
		audit_log_lost("rate limit exceeded");
	} else {
864 865 866 867 868 869 870 871 872
		if (audit_pid) {
			struct nlmsghdr *nlh = (struct nlmsghdr *)ab->skb->data;
			nlh->nlmsg_len = ab->skb->len - NLMSG_SPACE(0);
			skb_queue_tail(&audit_skb_queue, ab->skb);
			ab->skb = NULL;
			wake_up_interruptible(&kauditd_wait);
		} else {
			printk("%s\n", ab->skb->data + NLMSG_SPACE(0));
		}
L
Linus Torvalds 已提交
873
	}
874
	audit_buffer_free(ab);
L
Linus Torvalds 已提交
875 876 877 878 879
}

/* Log an audit record.  This is a convenience function that calls
 * audit_log_start, audit_log_vformat, and audit_log_end.  It may be
 * called in any context. */
880 881
void audit_log(struct audit_context *ctx, int gfp_mask, int type, 
	       const char *fmt, ...)
L
Linus Torvalds 已提交
882 883 884 885
{
	struct audit_buffer *ab;
	va_list args;

886
	ab = audit_log_start(ctx, gfp_mask, type);
L
Linus Torvalds 已提交
887 888 889 890 891 892 893
	if (ab) {
		va_start(args, fmt);
		audit_log_vformat(ab, fmt, args);
		va_end(args);
		audit_log_end(ab);
	}
}