xref: /illumos-gate/usr/src/lib/libpool/common/dict.c (revision 2aeafac3612e19716bf8164f89c3c9196342979c)
1 /*
2  * CDDL HEADER START
3  *
4  * The contents of this file are subject to the terms of the
5  * Common Development and Distribution License, Version 1.0 only
6  * (the "License").  You may not use this file except in compliance
7  * with the License.
8  *
9  * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
10  * or http://www.opensolaris.org/os/licensing.
11  * See the License for the specific language governing permissions
12  * and limitations under the License.
13  *
14  * When distributing Covered Code, include this CDDL HEADER in each
15  * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
16  * If applicable, add the following below this CDDL HEADER, with the
17  * fields enclosed by brackets "[]" replaced with your own identifying
18  * information: Portions Copyright [yyyy] [name of copyright owner]
19  *
20  * CDDL HEADER END
21  */
22 /*
23  * Copyright 2003 Sun Microsystems, Inc.  All rights reserved.
24  * Use is subject to license terms.
25  */
26 
27 #pragma ident	"%Z%%M%	%I%	%E% SMI"
28 
29 /*
30  * This file contains a basic dictionary implementation which stores
31  * arbitrary key-value mappings. It is used by libpool to store
32  * information about element pointers (pool_elem_t) in the kernel
33  * provider implementation.
34  */
35 
36 #include <stdio.h>
37 #include <stdlib.h>
38 #include <sys/types.h>
39 #include <unistd.h>
40 #include <string.h>
41 #include <assert.h>
42 
43 #include "dict.h"
44 
45 /*
46  * HASH_64_INIT is the same as the INIT value since it is the value
47  * used by FNV (FNV1_64_INIT). More details on FNV are available at:
48  *
49  * http://www.isthe.com/chongo/tech/comp/fnv/index.html
50  */
51 #define	HASH_64_INIT	(0xcbf29ce484222325ULL) /* Hash initializer */
52 
53 /*
54  * HASH_64_PRIME is a large prime number chosen to minimize hashing
55  * collisions.
56  */
57 #define	HASH_64_PRIME	(0x100000001b3ULL)	/* Large Prime */
58 
59 /*
60  * DICT_SIZE is chosen as it is the nearest prime to 2^9 (512). 512 is
61  * chosen as it is unlikely that this dictionary will contain more
62  * elements than this in normal operation. Of course overflow in each
63  * bucket is acceptable, but if there is too much overflow, then
64  * performance will degrade to that of a list.
65  */
66 #define	DICT_SIZE	509			/* Reasonable prime */
67 
68 /*
69  * Data Types
70  */
71 
72 /*
73  * A key bucket.
74  */
75 typedef struct dict_bucket
76 {
77 	const void		*db_key;	/* key */
78 	void			*db_value;	/* value */
79 	struct dict_bucket	*db_next;	/* next bucket */
80 } dict_bucket_t;
81 
82 /*
83  * A dictionary which holds a mapping between a key and a value.
84  *	dh_change	- detects changes to the dictionary.
85  *	dh_cmp		- comparison function
86  *	dh_hash		- hashing function
87  *	dh_buckets	- key storage
88  *	dh_size		- # of buckets
89  */
90 struct dict_hdl {
91 	uint64_t		dh_change;
92 	int			(*dh_cmp)(const void *, const void *);
93 	uint64_t		(*dh_hash)(const void *);
94 	uint64_t		dh_length;
95 	dict_bucket_t		**dh_buckets;
96 	uint64_t		dh_size;
97 };
98 
99 /*
100  * Utility functions. Mainly used for debugging
101  */
102 
103 #if defined(DEBUG)
104 
105 static void		bit_print_32(unsigned int);
106 static void		bit_print_64(unsigned long long);
107 
108 #endif /* DEBUG */
109 
110 /*
111  * Default functions for hashing and comparing if the user does not specify
112  * these values when creating the dictionary.
113  */
114 static int		cmp_addr(const void *, const void *);
115 static uint64_t		hash_addr(const void *);
116 
117 /*
118  * static functions
119  */
120 
121 #if defined(DEBUG)
122 
123 /*
124  * Print to standard out the bit representation of the supplied value
125  */
126 void
127 bit_print_32(unsigned int v)
128 {
129 #pragma	error_messages(off, E_INTEGER_OVERFLOW_DETECTED)
130 	int i, mask = 1 << 31;
131 #pragma	error_messages(on, E_INTEGER_OVERFLOW_DETECTED)
132 
133 	for (i = 1; i <= 32; i++) {
134 		(void) putchar(((v & mask) == 0) ? '0' : '1');
135 		v <<= 1;
136 		if (i % 8 == 0 && i != 32)
137 			(void) putchar(' ');
138 	}
139 	(void) putchar('\n');
140 }
141 
142 /*
143  * Print to standard out the bit representation of the supplied value
144  */
145 void
146 bit_print_64(unsigned long long v)
147 {
148 #pragma	error_messages(off, E_INTEGER_OVERFLOW_DETECTED)
149 	long long mask = 1ll << 63;
150 #pragma	error_messages(on, E_INTEGER_OVERFLOW_DETECTED)
151 	int i;
152 
153 	for (i = 1; i <= 64; i++) {
154 		(void) putchar(((v & mask) == 0) ? '0' : '1');
155 		v <<= 1;
156 		if (i % 8 == 0 && i != 64)
157 			(void) putchar(' ');
158 	}
159 	(void) putchar('\n');
160 }
161 
162 
163 
164 #endif /* DEBUG */
165 
166 /*
167  * Default comparison function which is used if no comparison function
168  * is supplied when the dictionary is created. The default behaviour
169  * is to compare memory address.
170  */
171 int
172 cmp_addr(const void *x, const void *y)
173 {
174 	return (x != y);
175 }
176 
177 
178 /*
179  * The default hashing function which is used if no hashing function
180  * is provided when the dictionary is created. The default behaviour
181  * is to use the hash_buf() function.
182  */
183 uint64_t
184 hash_addr(const void *key)
185 {
186 	return (hash_buf(&key, sizeof (key)));
187 }
188 
189 
190 /*
191  * public interface
192  */
193 
194 /*
195  * Return a hash which is built by manipulating each byte in the
196  * supplied data. The hash logic follows the approach suggested in the
197  * FNV hash.
198  */
199 uint64_t
200 hash_buf(const void *buf, size_t len)
201 {
202 	uchar_t *start = (uchar_t *)buf;
203 	uchar_t *end = start + len;
204 	uint64_t hash = HASH_64_INIT;
205 
206 	while (start < end) {
207 		hash *= HASH_64_PRIME;
208 		hash ^= (uint64_t)*start++;
209 	}
210 
211 	return (hash);
212 }
213 
214 
215 /*
216  * Return a hash which is built by manipulating each byte in the
217  * supplied string. The hash logic follows the approach suggested in
218  * the FNV hash.
219  */
220 uint64_t
221 hash_str(const char *str)
222 {
223 	uchar_t *p = (uchar_t *)str;
224 	uint64_t hash = HASH_64_INIT;
225 
226 	while (*p) {
227 		hash *= HASH_64_PRIME;
228 		hash ^= (uint64_t)*p++;
229 	}
230 
231 	return (hash);
232 }
233 
234 /*
235  * Return the number of keys held in the supplied dictionary.
236  */
237 uint64_t
238 dict_length(dict_hdl_t *hdl)
239 {
240 	return (hdl->dh_length);
241 }
242 
243 /*
244  * Free the supplied dictionary and all it's associated resource.
245  */
246 void
247 dict_free(dict_hdl_t **hdl)
248 {
249 	if ((*hdl)->dh_length > 0) {
250 		uint64_t i;
251 		for (i = 0; i < (*hdl)->dh_size; i++) {
252 			dict_bucket_t *this, *next;
253 			for (this = (*hdl)->dh_buckets[i]; this != NULL;
254 			    this = next) {
255 				next = this->db_next;
256 				free(this);
257 			}
258 		}
259 	}
260 	free((*hdl)->dh_buckets);
261 	free((*hdl));
262 	*hdl = NULL;
263 }
264 
265 /*
266  * Create a new dictionary using the supplied comparison and hashing
267  * functions. If none are supplied then the defaults are used.
268  */
269 dict_hdl_t *
270 dict_new(int (*cmp)(const void *, const void *),
271     uint64_t (*hash)(const void *))
272 {
273 	dict_hdl_t *hdl;
274 
275 	if ((hdl = calloc(1, sizeof (dict_hdl_t))) == NULL)
276 		return (NULL);
277 	hdl->dh_size = DICT_SIZE;
278 	if ((hdl->dh_buckets = calloc(hdl->dh_size, sizeof (dict_bucket_t *)))
279 	    == NULL) {
280 		free(hdl);
281 		return (NULL);
282 	}
283 	hdl->dh_cmp = cmp ? cmp : cmp_addr;
284 	hdl->dh_hash = hash ? hash : hash_addr;
285 	return (hdl);
286 }
287 
288 /*
289  * Get a value from the hash. Null is returned if the key cannot be
290  * found.
291  */
292 void *
293 dict_get(dict_hdl_t *hdl, const void *key)
294 {
295 	uint64_t i;
296 	dict_bucket_t *bucket;
297 
298 	i = (*hdl->dh_hash)(key)%hdl->dh_size;
299 	for (bucket = hdl->dh_buckets[i]; bucket != NULL;
300 	    bucket = bucket->db_next)
301 		if ((*hdl->dh_cmp)(key, bucket->db_key) == 0)
302 			break;
303 	return (bucket ? bucket->db_value : NULL);
304 }
305 
306 /*
307  * Put an entry into the hash. Null is returned if this key was not
308  * already present, otherwise the previous value is returned.
309  */
310 void *
311 dict_put(dict_hdl_t *hdl, const void *key, void *value)
312 {
313 	uint64_t i;
314 	dict_bucket_t *bucket;
315 	void *prev = NULL;
316 
317 	i = (*hdl->dh_hash)(key)%hdl->dh_size;
318 	for (bucket = hdl->dh_buckets[i]; bucket != NULL;
319 	    bucket = bucket->db_next)
320 		if ((*hdl->dh_cmp)(key, bucket->db_key) == 0)
321 			break;
322 	if (bucket) {
323 		prev = bucket->db_value;
324 	} else {
325 		bucket = malloc(sizeof (dict_bucket_t));
326 		bucket->db_key = key;
327 		bucket->db_next = hdl->dh_buckets[i];
328 		hdl->dh_buckets[i] = bucket;
329 		hdl->dh_length++;
330 	}
331 	hdl->dh_change++;
332 	bucket->db_value = value;
333 	return (prev);
334 }
335 
336 /*
337  * Remove the key/value from the dictionary. The value is returned if
338  * the key is found. NULL is returned if the key cannot be located.
339  */
340 void *
341 dict_remove(dict_hdl_t *hdl, const void *key)
342 {
343 	uint64_t i;
344 	dict_bucket_t	**pbucket;
345 
346 	hdl->dh_change++;
347 	i = (*hdl->dh_hash)(key)%hdl->dh_size;
348 
349 	for (pbucket = &hdl->dh_buckets[i]; *pbucket != NULL;
350 	    pbucket = &(*pbucket)->db_next) {
351 		if ((*hdl->dh_cmp)(key, (*pbucket)->db_key) == 0) {
352 			dict_bucket_t *bucket = *pbucket;
353 			void *value = bucket->db_value;
354 
355 			*pbucket = bucket->db_next;
356 			free(bucket);
357 			hdl->dh_length--;
358 			return (value);
359 		}
360 	}
361 	return (NULL);
362 }
363 
364 /*
365  * For all entries in the dictionary call the user supplied function
366  * (apply) with the key, value and user supplied data. If the
367  * dictionary is modifed while this function is executing, then the
368  * function will fail with an assertion about table modifcation.
369  */
370 void
371 dict_map(dict_hdl_t *hdl, void (*apply)(const void *, void **, void *),
372     void *cl)
373 {
374 	uint64_t i;
375 	dict_bucket_t *bucket = NULL;
376 	uint64_t change_stamp = hdl->dh_change;
377 
378 	for (i = 0; i < hdl->dh_size; i++) {
379 		for (bucket = hdl->dh_buckets[i]; bucket != NULL;
380 		    bucket = bucket->db_next) {
381 			apply(bucket->db_key, &bucket->db_value, cl);
382 			if (hdl->dh_change != change_stamp)
383 				assert(!"table modified illegally");
384 		}
385 	}
386 }
387