xref: /freebsd/contrib/ntp/ntpd/ntp_monitor.c (revision d6b92ffa)
1 /*
2  * ntp_monitor - monitor ntpd statistics
3  */
4 #ifdef HAVE_CONFIG_H
5 # include <config.h>
6 #endif
7 
8 #include "ntpd.h"
9 #include "ntp_io.h"
10 #include "ntp_if.h"
11 #include "ntp_lists.h"
12 #include "ntp_stdlib.h"
13 #include <ntp_random.h>
14 
15 #include <stdio.h>
16 #include <signal.h>
17 #ifdef HAVE_SYS_IOCTL_H
18 # include <sys/ioctl.h>
19 #endif
20 
21 /*
22  * Record statistics based on source address, mode and version. The
23  * receive procedure calls us with the incoming rbufp before it does
24  * anything else. While at it, implement rate controls for inbound
25  * traffic.
26  *
27  * Each entry is doubly linked into two lists, a hash table and a most-
28  * recently-used (MRU) list. When a packet arrives it is looked up in
29  * the hash table. If found, the statistics are updated and the entry
30  * relinked at the head of the MRU list. If not found, a new entry is
31  * allocated, initialized and linked into both the hash table and at the
32  * head of the MRU list.
33  *
34  * Memory is usually allocated by grabbing a big chunk of new memory and
35  * cutting it up into littler pieces. The exception to this when we hit
36  * the memory limit. Then we free memory by grabbing entries off the
37  * tail for the MRU list, unlinking from the hash table, and
38  * reinitializing.
39  *
40  * INC_MONLIST is the default allocation granularity in entries.
41  * INIT_MONLIST is the default initial allocation in entries.
42  */
43 #ifdef MONMEMINC		/* old name */
44 # define	INC_MONLIST	MONMEMINC
45 #elif !defined(INC_MONLIST)
46 # define	INC_MONLIST	(4 * 1024 / sizeof(mon_entry))
47 #endif
48 #ifndef INIT_MONLIST
49 # define	INIT_MONLIST	(4 * 1024 / sizeof(mon_entry))
50 #endif
51 #ifndef MRU_MAXDEPTH_DEF
52 # define MRU_MAXDEPTH_DEF	(1024 * 1024 / sizeof(mon_entry))
53 #endif
54 
55 /*
56  * Hashing stuff
57  */
58 u_char	mon_hash_bits;
59 
60 /*
61  * Pointers to the hash table and the MRU list.  Memory for the hash
62  * table is allocated only if monitoring is enabled.
63  */
64 mon_entry **	mon_hash;	/* MRU hash table */
65 mon_entry	mon_mru_list;	/* mru listhead */
66 
67 /*
68  * List of free structures structures, and counters of in-use and total
69  * structures. The free structures are linked with the hash_next field.
70  */
71 static  mon_entry *mon_free;		/* free list or null if none */
72 	u_int mru_alloc;		/* mru list + free list count */
73 	u_int mru_entries;		/* mru list count */
74 	u_int mru_peakentries;		/* highest mru_entries seen */
75 	u_int mru_initalloc = INIT_MONLIST;/* entries to preallocate */
76 	u_int mru_incalloc = INC_MONLIST;/* allocation batch factor */
77 static	u_int mon_mem_increments;	/* times called malloc() */
78 
79 /*
80  * Parameters of the RES_LIMITED restriction option. We define headway
81  * as the idle time between packets. A packet is discarded if the
82  * headway is less than the minimum, as well as if the average headway
83  * is less than eight times the increment.
84  */
85 int	ntp_minpkt = NTP_MINPKT;	/* minimum (log 2 s) */
86 u_char	ntp_minpoll = NTP_MINPOLL;	/* increment (log 2 s) */
87 
88 /*
89  * Initialization state.  We may be monitoring, we may not.  If
90  * we aren't, we may not even have allocated any memory yet.
91  */
92 	u_int	mon_enabled;		/* enable switch */
93 	u_int	mru_mindepth = 600;	/* preempt above this */
94 	int	mru_maxage = 64;	/* for entries older than */
95 	u_int	mru_maxdepth = 		/* MRU count hard limit */
96 			MRU_MAXDEPTH_DEF;
97 	int	mon_age = 3000;		/* preemption limit */
98 
99 static	void		mon_getmoremem(void);
100 static	void		remove_from_hash(mon_entry *);
101 static	inline void	mon_free_entry(mon_entry *);
102 static	inline void	mon_reclaim_entry(mon_entry *);
103 
104 
105 /*
106  * init_mon - initialize monitoring global data
107  */
108 void
109 init_mon(void)
110 {
111 	/*
112 	 * Don't do much of anything here.  We don't allocate memory
113 	 * until mon_start().
114 	 */
115 	mon_enabled = MON_OFF;
116 	INIT_DLIST(mon_mru_list, mru);
117 }
118 
119 
120 /*
121  * remove_from_hash - removes an entry from the address hash table and
122  *		      decrements mru_entries.
123  */
124 static void
125 remove_from_hash(
126 	mon_entry *mon
127 	)
128 {
129 	u_int hash;
130 	mon_entry *punlinked;
131 
132 	mru_entries--;
133 	hash = MON_HASH(&mon->rmtadr);
134 	UNLINK_SLIST(punlinked, mon_hash[hash], mon, hash_next,
135 		     mon_entry);
136 	ENSURE(punlinked == mon);
137 }
138 
139 
140 static inline void
141 mon_free_entry(
142 	mon_entry *m
143 	)
144 {
145 	ZERO(*m);
146 	LINK_SLIST(mon_free, m, hash_next);
147 }
148 
149 
150 /*
151  * mon_reclaim_entry - Remove an entry from the MRU list and from the
152  *		       hash array, then zero-initialize it.  Indirectly
153  *		       decrements mru_entries.
154 
155  * The entry is prepared to be reused.  Before return, in
156  * remove_from_hash(), mru_entries is decremented.  It is the caller's
157  * responsibility to increment it again.
158  */
159 static inline void
160 mon_reclaim_entry(
161 	mon_entry *m
162 	)
163 {
164 	DEBUG_INSIST(NULL != m);
165 
166 	UNLINK_DLIST(m, mru);
167 	remove_from_hash(m);
168 	ZERO(*m);
169 }
170 
171 
172 /*
173  * mon_getmoremem - get more memory and put it on the free list
174  */
175 static void
176 mon_getmoremem(void)
177 {
178 	mon_entry *chunk;
179 	u_int entries;
180 
181 	entries = (0 == mon_mem_increments)
182 		      ? mru_initalloc
183 		      : mru_incalloc;
184 
185 	if (entries) {
186 		chunk = eallocarray(entries, sizeof(*chunk));
187 		mru_alloc += entries;
188 		for (chunk += entries; entries; entries--)
189 			mon_free_entry(--chunk);
190 
191 		mon_mem_increments++;
192 	}
193 }
194 
195 
196 /*
197  * mon_start - start up the monitoring software
198  */
199 void
200 mon_start(
201 	int mode
202 	)
203 {
204 	size_t octets;
205 	u_int min_hash_slots;
206 
207 	if (MON_OFF == mode)		/* MON_OFF is 0 */
208 		return;
209 	if (mon_enabled) {
210 		mon_enabled |= mode;
211 		return;
212 	}
213 	if (0 == mon_mem_increments)
214 		mon_getmoremem();
215 	/*
216 	 * Select the MRU hash table size to limit the average count
217 	 * per bucket at capacity (mru_maxdepth) to 8, if possible
218 	 * given our hash is limited to 16 bits.
219 	 */
220 	min_hash_slots = (mru_maxdepth / 8) + 1;
221 	mon_hash_bits = 0;
222 	while (min_hash_slots >>= 1)
223 		mon_hash_bits++;
224 	mon_hash_bits = max(4, mon_hash_bits);
225 	mon_hash_bits = min(16, mon_hash_bits);
226 	octets = sizeof(*mon_hash) * MON_HASH_SIZE;
227 	mon_hash = erealloc_zero(mon_hash, octets, 0);
228 
229 	mon_enabled = mode;
230 }
231 
232 
233 /*
234  * mon_stop - stop the monitoring software
235  */
236 void
237 mon_stop(
238 	int mode
239 	)
240 {
241 	mon_entry *mon;
242 
243 	if (MON_OFF == mon_enabled)
244 		return;
245 	if ((mon_enabled & mode) == 0 || mode == MON_OFF)
246 		return;
247 
248 	mon_enabled &= ~mode;
249 	if (mon_enabled != MON_OFF)
250 		return;
251 
252 	/*
253 	 * Move everything on the MRU list to the free list quickly,
254 	 * without bothering to remove each from either the MRU list or
255 	 * the hash table.
256 	 */
257 	ITER_DLIST_BEGIN(mon_mru_list, mon, mru, mon_entry)
258 		mon_free_entry(mon);
259 	ITER_DLIST_END()
260 
261 	/* empty the MRU list and hash table. */
262 	mru_entries = 0;
263 	INIT_DLIST(mon_mru_list, mru);
264 	zero_mem(mon_hash, sizeof(*mon_hash) * MON_HASH_SIZE);
265 }
266 
267 
268 /*
269  * mon_clearinterface -- remove mru entries referring to a local address
270  *			 which is going away.
271  */
272 void
273 mon_clearinterface(
274 	endpt *lcladr
275 	)
276 {
277 	mon_entry *mon;
278 
279 	/* iterate mon over mon_mru_list */
280 	ITER_DLIST_BEGIN(mon_mru_list, mon, mru, mon_entry)
281 		if (mon->lcladr == lcladr) {
282 			/* remove from mru list */
283 			UNLINK_DLIST(mon, mru);
284 			/* remove from hash list, adjust mru_entries */
285 			remove_from_hash(mon);
286 			/* put on free list */
287 			mon_free_entry(mon);
288 		}
289 	ITER_DLIST_END()
290 }
291 
292 
293 /*
294  * ntp_monitor - record stats about this packet
295  *
296  * Returns supplied restriction flags, with RES_LIMITED and RES_KOD
297  * cleared unless the packet should not be responded to normally
298  * (RES_LIMITED) and possibly should trigger a KoD response (RES_KOD).
299  * The returned flags are saved in the MRU entry, so that it reflects
300  * whether the last packet from that source triggered rate limiting,
301  * and if so, possible KoD response.  This implies you can not tell
302  * whether a given address is eligible for rate limiting/KoD from the
303  * monlist restrict bits, only whether or not the last packet triggered
304  * such responses.  ntpdc -c reslist lets you see whether RES_LIMITED
305  * or RES_KOD is lit for a particular address before ntp_monitor()'s
306  * typical dousing.
307  */
308 u_short
309 ntp_monitor(
310 	struct recvbuf *rbufp,
311 	u_short	flags
312 	)
313 {
314 	l_fp		interval_fp;
315 	struct pkt *	pkt;
316 	mon_entry *	mon;
317 	mon_entry *	oldest;
318 	int		oldest_age;
319 	u_int		hash;
320 	u_short		restrict_mask;
321 	u_char		mode;
322 	u_char		version;
323 	int		interval;
324 	int		head;		/* headway increment */
325 	int		leak;		/* new headway */
326 	int		limit;		/* average threshold */
327 
328 	REQUIRE(rbufp != NULL);
329 
330 	if (mon_enabled == MON_OFF)
331 		return ~(RES_LIMITED | RES_KOD) & flags;
332 
333 	pkt = &rbufp->recv_pkt;
334 	hash = MON_HASH(&rbufp->recv_srcadr);
335 	mode = PKT_MODE(pkt->li_vn_mode);
336 	version = PKT_VERSION(pkt->li_vn_mode);
337 	mon = mon_hash[hash];
338 
339 	/*
340 	 * We keep track of all traffic for a given IP in one entry,
341 	 * otherwise cron'ed ntpdate or similar evades RES_LIMITED.
342 	 */
343 
344 	for (; mon != NULL; mon = mon->hash_next)
345 		if (SOCK_EQ(&mon->rmtadr, &rbufp->recv_srcadr))
346 			break;
347 
348 	if (mon != NULL) {
349 		interval_fp = rbufp->recv_time;
350 		L_SUB(&interval_fp, &mon->last);
351 		/* add one-half second to round up */
352 		L_ADDUF(&interval_fp, 0x80000000);
353 		interval = interval_fp.l_i;
354 		mon->last = rbufp->recv_time;
355 		NSRCPORT(&mon->rmtadr) = NSRCPORT(&rbufp->recv_srcadr);
356 		mon->count++;
357 		restrict_mask = flags;
358 		mon->vn_mode = VN_MODE(version, mode);
359 
360 		/* Shuffle to the head of the MRU list. */
361 		UNLINK_DLIST(mon, mru);
362 		LINK_DLIST(mon_mru_list, mon, mru);
363 
364 		/*
365 		 * At this point the most recent arrival is first in the
366 		 * MRU list.  Decrease the counter by the headway, but
367 		 * not less than zero.
368 		 */
369 		mon->leak -= interval;
370 		mon->leak = max(0, mon->leak);
371 		head = 1 << ntp_minpoll;
372 		leak = mon->leak + head;
373 		limit = NTP_SHIFT * head;
374 
375 		DPRINTF(2, ("MRU: interval %d headway %d limit %d\n",
376 			    interval, leak, limit));
377 
378 		/*
379 		 * If the minimum and average thresholds are not
380 		 * exceeded, douse the RES_LIMITED and RES_KOD bits and
381 		 * increase the counter by the headway increment.  Note
382 		 * that we give a 1-s grace for the minimum threshold
383 		 * and a 2-s grace for the headway increment.  If one or
384 		 * both thresholds are exceeded and the old counter is
385 		 * less than the average threshold, set the counter to
386 		 * the average threshold plus the increment and leave
387 		 * the RES_LIMITED and RES_KOD bits lit. Otherwise,
388 		 * leave the counter alone and douse the RES_KOD bit.
389 		 * This rate-limits the KoDs to no less than the average
390 		 * headway.
391 		 */
392 		if (interval + 1 >= ntp_minpkt && leak < limit) {
393 			mon->leak = leak - 2;
394 			restrict_mask &= ~(RES_LIMITED | RES_KOD);
395 		} else if (mon->leak < limit)
396 			mon->leak = limit + head;
397 		else
398 			restrict_mask &= ~RES_KOD;
399 
400 		mon->flags = restrict_mask;
401 
402 		return mon->flags;
403 	}
404 
405 	/*
406 	 * If we got here, this is the first we've heard of this
407 	 * guy.  Get him some memory, either from the free list
408 	 * or from the tail of the MRU list.
409 	 *
410 	 * The following ntp.conf "mru" knobs come into play determining
411 	 * the depth (or count) of the MRU list:
412 	 * - mru_mindepth ("mru mindepth") is a floor beneath which
413 	 *   entries are kept without regard to their age.  The
414 	 *   default is 600 which matches the longtime implementation
415 	 *   limit on the total number of entries.
416 	 * - mru_maxage ("mru maxage") is a ceiling on the age in
417 	 *   seconds of entries.  Entries older than this are
418 	 *   reclaimed once mon_mindepth is exceeded.  64s default.
419 	 *   Note that entries older than this can easily survive
420 	 *   as they are reclaimed only as needed.
421 	 * - mru_maxdepth ("mru maxdepth") is a hard limit on the
422 	 *   number of entries.
423 	 * - "mru maxmem" sets mru_maxdepth to the number of entries
424 	 *   which fit in the given number of kilobytes.  The default is
425 	 *   1024, or 1 megabyte.
426 	 * - mru_initalloc ("mru initalloc" sets the count of the
427 	 *   initial allocation of MRU entries.
428 	 * - "mru initmem" sets mru_initalloc in units of kilobytes.
429 	 *   The default is 4.
430 	 * - mru_incalloc ("mru incalloc" sets the number of entries to
431 	 *   allocate on-demand each time the free list is empty.
432 	 * - "mru incmem" sets mru_incalloc in units of kilobytes.
433 	 *   The default is 4.
434 	 * Whichever of "mru maxmem" or "mru maxdepth" occurs last in
435 	 * ntp.conf controls.  Similarly for "mru initalloc" and "mru
436 	 * initmem", and for "mru incalloc" and "mru incmem".
437 	 */
438 	if (mru_entries < mru_mindepth) {
439 		if (NULL == mon_free)
440 			mon_getmoremem();
441 		UNLINK_HEAD_SLIST(mon, mon_free, hash_next);
442 	} else {
443 		oldest = TAIL_DLIST(mon_mru_list, mru);
444 		oldest_age = 0;		/* silence uninit warning */
445 		if (oldest != NULL) {
446 			interval_fp = rbufp->recv_time;
447 			L_SUB(&interval_fp, &oldest->last);
448 			/* add one-half second to round up */
449 			L_ADDUF(&interval_fp, 0x80000000);
450 			oldest_age = interval_fp.l_i;
451 		}
452 		/* note -1 is legal for mru_maxage (disables) */
453 		if (oldest != NULL && mru_maxage < oldest_age) {
454 			mon_reclaim_entry(oldest);
455 			mon = oldest;
456 		} else if (mon_free != NULL || mru_alloc <
457 			   mru_maxdepth) {
458 			if (NULL == mon_free)
459 				mon_getmoremem();
460 			UNLINK_HEAD_SLIST(mon, mon_free, hash_next);
461 		/* Preempt from the MRU list if old enough. */
462 		} else if (ntp_random() / (2. * FRAC) >
463 			   (double)oldest_age / mon_age) {
464 			return ~(RES_LIMITED | RES_KOD) & flags;
465 		} else {
466 			mon_reclaim_entry(oldest);
467 			mon = oldest;
468 		}
469 	}
470 
471 	INSIST(mon != NULL);
472 
473 	/*
474 	 * Got one, initialize it
475 	 */
476 	mru_entries++;
477 	mru_peakentries = max(mru_peakentries, mru_entries);
478 	mon->last = rbufp->recv_time;
479 	mon->first = mon->last;
480 	mon->count = 1;
481 	mon->flags = ~(RES_LIMITED | RES_KOD) & flags;
482 	mon->leak = 0;
483 	memcpy(&mon->rmtadr, &rbufp->recv_srcadr, sizeof(mon->rmtadr));
484 	mon->vn_mode = VN_MODE(version, mode);
485 	mon->lcladr = rbufp->dstadr;
486 	mon->cast_flags = (u_char)(((rbufp->dstadr->flags &
487 	    INT_MCASTOPEN) && rbufp->fd == mon->lcladr->fd) ? MDF_MCAST
488 	    : rbufp->fd == mon->lcladr->bfd ? MDF_BCAST : MDF_UCAST);
489 
490 	/*
491 	 * Drop him into front of the hash table. Also put him on top of
492 	 * the MRU list.
493 	 */
494 	LINK_SLIST(mon_hash[hash], mon, hash_next);
495 	LINK_DLIST(mon_mru_list, mon, mru);
496 
497 	return mon->flags;
498 }
499 
500 
501