]>
Commit | Line | Data |
---|---|---|
62b83e6d MT |
1 | /* |
2 | libloc - A library to determine the location of someone on the Internet | |
3 | ||
4 | Copyright (C) 2017 IPFire Development Team <info@ipfire.org> | |
5 | ||
6 | This library is free software; you can redistribute it and/or | |
7 | modify it under the terms of the GNU Lesser General Public | |
8 | License as published by the Free Software Foundation; either | |
9 | version 2.1 of the License, or (at your option) any later version. | |
10 | ||
11 | This library is distributed in the hope that it will be useful, | |
12 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
13 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
14 | Lesser General Public License for more details. | |
15 | */ | |
16 | ||
17 | #include <errno.h> | |
18 | #include <stddef.h> | |
19 | #include <stdio.h> | |
20 | #include <stdlib.h> | |
21 | #include <string.h> | |
6809d5ac MT |
22 | #include <sys/mman.h> |
23 | #include <unistd.h> | |
62b83e6d | 24 | |
c12a1385 MT |
25 | #include <libloc/libloc.h> |
26 | #include <libloc/format.h> | |
27 | #include <libloc/private.h> | |
28 | #include <libloc/stringpool.h> | |
62b83e6d MT |
29 | |
30 | struct loc_stringpool { | |
31 | struct loc_ctx* ctx; | |
62b83e6d | 32 | int refcount; |
0e974d4b | 33 | |
e4b6ca42 MT |
34 | // A file descriptor when we open an existing stringpool |
35 | int fd; | |
0e974d4b | 36 | |
e4b6ca42 | 37 | off_t offset; |
0e974d4b | 38 | ssize_t length; |
62b83e6d | 39 | |
e4b6ca42 MT |
40 | // Mapped data (from mmap()) |
41 | char* mmapped_data; | |
42 | ||
43 | char* data; | |
0e974d4b | 44 | char* pos; |
e4b6ca42 MT |
45 | |
46 | char buffer[LOC_DATABASE_PAGE_SIZE]; | |
62b83e6d MT |
47 | }; |
48 | ||
010fe9b6 | 49 | static off_t loc_stringpool_get_offset(struct loc_stringpool* pool, const char* pos) { |
198e382c MT |
50 | if (pos < pool->data) { |
51 | errno = EFAULT; | |
52 | return -1; | |
53 | } | |
010fe9b6 | 54 | |
198e382c MT |
55 | if (pos > (pool->data + pool->length)) { |
56 | errno = EFAULT; | |
57 | return -1; | |
58 | } | |
010fe9b6 MT |
59 | |
60 | return pos - pool->data; | |
61 | } | |
62 | ||
63 | static char* __loc_stringpool_get(struct loc_stringpool* pool, off_t offset) { | |
e4b6ca42 MT |
64 | ssize_t bytes_read; |
65 | ||
66 | // Check boundaries | |
67 | if (offset < 0 || offset >= pool->length) { | |
68 | errno = ERANGE; | |
010fe9b6 | 69 | return NULL; |
e4b6ca42 MT |
70 | } |
71 | ||
72 | // Return any data that we have in memory | |
73 | if (pool->data) | |
74 | return pool->data + offset; | |
010fe9b6 | 75 | |
e4b6ca42 MT |
76 | // Otherwise read a block from file |
77 | bytes_read = pread(pool->fd, pool->buffer, sizeof(pool->buffer), | |
78 | pool->offset + offset); | |
79 | ||
80 | // Break on error | |
81 | if (bytes_read < 0) { | |
82 | ERROR(pool->ctx, "Could not read from string pool: %m\n"); | |
83 | return NULL; | |
84 | } | |
85 | ||
86 | // It is okay, if we did not read as much as we wanted, since we might be reading | |
87 | // the last block which might be of an unknown size. | |
88 | ||
89 | // Search for a complete string. If there is no NULL byte, the block is garbage. | |
90 | char* end = memchr(pool->buffer, bytes_read, '\0'); | |
91 | if (!end) | |
92 | return NULL; | |
93 | ||
94 | // Return what's in the buffer | |
95 | return pool->buffer; | |
010fe9b6 MT |
96 | } |
97 | ||
98 | static int loc_stringpool_grow(struct loc_stringpool* pool, size_t length) { | |
99 | DEBUG(pool->ctx, "Growing string pool to %zu bytes\n", length); | |
100 | ||
101 | // Save pos pointer | |
102 | off_t pos = loc_stringpool_get_offset(pool, pool->pos); | |
103 | ||
104 | // Reallocate data section | |
105 | pool->data = realloc(pool->data, length); | |
106 | if (!pool->data) | |
198e382c | 107 | return 1; |
010fe9b6 MT |
108 | |
109 | pool->length = length; | |
110 | ||
111 | // Restore pos | |
112 | pool->pos = __loc_stringpool_get(pool, pos); | |
113 | ||
114 | return 0; | |
115 | } | |
116 | ||
117 | static off_t loc_stringpool_append(struct loc_stringpool* pool, const char* string) { | |
198e382c MT |
118 | if (!string) { |
119 | errno = EINVAL; | |
120 | return -1; | |
121 | } | |
010fe9b6 MT |
122 | |
123 | DEBUG(pool->ctx, "Appending '%s' to string pool at %p\n", string, pool); | |
124 | ||
125 | // Make sure we have enough space | |
126 | int r = loc_stringpool_grow(pool, pool->length + strlen(string) + 1); | |
198e382c | 127 | if (r) |
010fe9b6 | 128 | return -1; |
010fe9b6 MT |
129 | |
130 | off_t offset = loc_stringpool_get_offset(pool, pool->pos); | |
131 | ||
132 | // Copy string byte by byte | |
133 | while (*string) | |
134 | *pool->pos++ = *string++; | |
135 | ||
136 | // Terminate the string | |
137 | *pool->pos++ = '\0'; | |
138 | ||
139 | return offset; | |
140 | } | |
141 | ||
acf44a73 MT |
142 | static void loc_stringpool_free(struct loc_stringpool* pool) { |
143 | DEBUG(pool->ctx, "Releasing string pool %p\n", pool); | |
144 | int r; | |
145 | ||
e4b6ca42 MT |
146 | // Close file |
147 | if (pool->fd > 0) | |
148 | close(pool->fd); | |
149 | ||
150 | // Unmap any mapped memory | |
151 | if (pool->mmapped_data) { | |
152 | r = munmap(pool->mmapped_data, pool->length); | |
153 | if (r) | |
154 | ERROR(pool->ctx, "Error unmapping string pool: %m\n"); | |
155 | ||
156 | if (pool->mmapped_data == pool->data) | |
157 | pool->data = NULL; | |
acf44a73 MT |
158 | } |
159 | ||
e4b6ca42 MT |
160 | // Free any data |
161 | if (pool->data) | |
162 | free(pool->data); | |
163 | ||
acf44a73 MT |
164 | loc_unref(pool->ctx); |
165 | free(pool); | |
166 | } | |
167 | ||
22d8abdc | 168 | int loc_stringpool_new(struct loc_ctx* ctx, struct loc_stringpool** pool) { |
0e974d4b MT |
169 | struct loc_stringpool* p = calloc(1, sizeof(*p)); |
170 | if (!p) | |
8d10d064 | 171 | return 1; |
0e974d4b MT |
172 | |
173 | p->ctx = loc_ref(ctx); | |
174 | p->refcount = 1; | |
6809d5ac | 175 | |
0e974d4b | 176 | *pool = p; |
6809d5ac MT |
177 | |
178 | return 0; | |
179 | } | |
180 | ||
e4b6ca42 MT |
181 | static int loc_stringpool_mmap(struct loc_stringpool* pool) { |
182 | // Try mmap() | |
183 | char* p = mmap(NULL, pool->length, PROT_READ, MAP_PRIVATE, pool->fd, pool->offset); | |
6809d5ac | 184 | |
e4b6ca42 MT |
185 | if (p == MAP_FAILED) { |
186 | // Ignore if data hasn't been aligned correctly | |
187 | if (errno == EINVAL) | |
188 | return 0; | |
6809d5ac | 189 | |
e4b6ca42 | 190 | ERROR(pool->ctx, "Could not mmap stringpool: %m\n"); |
acf44a73 | 191 | return 1; |
e4b6ca42 MT |
192 | } |
193 | ||
194 | // Store mapped memory area | |
195 | pool->data = pool->mmapped_data = pool->pos = p; | |
2601e83e MT |
196 | |
197 | return 0; | |
198 | } | |
199 | ||
22d8abdc | 200 | int loc_stringpool_open(struct loc_ctx* ctx, struct loc_stringpool** pool, |
0e974d4b | 201 | FILE* f, size_t length, off_t offset) { |
acf44a73 MT |
202 | struct loc_stringpool* p = NULL; |
203 | ||
204 | // Allocate a new stringpool | |
205 | int r = loc_stringpool_new(ctx, &p); | |
0e974d4b | 206 | if (r) |
e4b6ca42 MT |
207 | goto ERROR; |
208 | ||
209 | // Store offset and length | |
210 | p->offset = offset; | |
211 | p->length = length; | |
212 | ||
213 | DEBUG(p->ctx, "Reading string pool starting from %jd (%zu bytes)\n", | |
214 | (intmax_t)p->offset, p->length); | |
62b83e6d | 215 | |
e4b6ca42 MT |
216 | int fd = fileno(f); |
217 | ||
218 | // Copy the file descriptor | |
219 | p->fd = dup(fd); | |
220 | if (p->fd < 0) { | |
221 | ERROR(ctx, "Could not duplicate file the file descriptor: %m\n"); | |
222 | r = 1; | |
223 | goto ERROR; | |
224 | } | |
8d10d064 | 225 | |
0e974d4b | 226 | // Map data into memory |
e4b6ca42 MT |
227 | if (p->length > 0) { |
228 | r = loc_stringpool_mmap(p); | |
229 | if (r) | |
230 | goto ERROR; | |
2a30e4de | 231 | } |
62b83e6d | 232 | |
acf44a73 | 233 | *pool = p; |
62b83e6d | 234 | return 0; |
e4b6ca42 MT |
235 | |
236 | ERROR: | |
237 | if (p) | |
238 | loc_stringpool_free(p); | |
239 | ||
240 | return r; | |
62b83e6d MT |
241 | } |
242 | ||
22d8abdc | 243 | struct loc_stringpool* loc_stringpool_ref(struct loc_stringpool* pool) { |
62b83e6d MT |
244 | pool->refcount++; |
245 | ||
246 | return pool; | |
247 | } | |
248 | ||
22d8abdc | 249 | struct loc_stringpool* loc_stringpool_unref(struct loc_stringpool* pool) { |
62b83e6d MT |
250 | if (--pool->refcount > 0) |
251 | return NULL; | |
252 | ||
253 | loc_stringpool_free(pool); | |
254 | ||
255 | return NULL; | |
256 | } | |
257 | ||
62b83e6d MT |
258 | static off_t loc_stringpool_get_next_offset(struct loc_stringpool* pool, off_t offset) { |
259 | const char* string = loc_stringpool_get(pool, offset); | |
1afe5545 MT |
260 | if (!string) |
261 | return offset; | |
62b83e6d MT |
262 | |
263 | return offset + strlen(string) + 1; | |
264 | } | |
265 | ||
22d8abdc | 266 | const char* loc_stringpool_get(struct loc_stringpool* pool, off_t offset) { |
0e974d4b | 267 | return __loc_stringpool_get(pool, offset); |
62b83e6d MT |
268 | } |
269 | ||
22d8abdc | 270 | size_t loc_stringpool_get_size(struct loc_stringpool* pool) { |
2601e83e MT |
271 | return loc_stringpool_get_offset(pool, pool->pos); |
272 | } | |
273 | ||
62b83e6d | 274 | static off_t loc_stringpool_find(struct loc_stringpool* pool, const char* s) { |
198e382c MT |
275 | if (!s || !*s) { |
276 | errno = EINVAL; | |
277 | return -1; | |
278 | } | |
62b83e6d MT |
279 | |
280 | off_t offset = 0; | |
0e974d4b | 281 | while (offset < pool->length) { |
62b83e6d MT |
282 | const char* string = loc_stringpool_get(pool, offset); |
283 | if (!string) | |
284 | break; | |
285 | ||
286 | int r = strcmp(s, string); | |
287 | if (r == 0) | |
288 | return offset; | |
289 | ||
290 | offset = loc_stringpool_get_next_offset(pool, offset); | |
291 | } | |
292 | ||
198e382c MT |
293 | // Nothing found |
294 | errno = ENOENT; | |
295 | return -1; | |
62b83e6d MT |
296 | } |
297 | ||
22d8abdc | 298 | off_t loc_stringpool_add(struct loc_stringpool* pool, const char* string) { |
62b83e6d MT |
299 | off_t offset = loc_stringpool_find(pool, string); |
300 | if (offset >= 0) { | |
5c57de03 | 301 | DEBUG(pool->ctx, "Found '%s' at position %jd\n", string, (intmax_t)offset); |
62b83e6d MT |
302 | return offset; |
303 | } | |
304 | ||
305 | return loc_stringpool_append(pool, string); | |
306 | } | |
307 | ||
22d8abdc | 308 | void loc_stringpool_dump(struct loc_stringpool* pool) { |
62b83e6d MT |
309 | off_t offset = 0; |
310 | ||
0e974d4b | 311 | while (offset < pool->length) { |
62b83e6d MT |
312 | const char* string = loc_stringpool_get(pool, offset); |
313 | if (!string) | |
314 | break; | |
315 | ||
5c57de03 | 316 | printf("%jd (%zu): %s\n", (intmax_t)offset, strlen(string), string); |
62b83e6d MT |
317 | |
318 | offset = loc_stringpool_get_next_offset(pool, offset); | |
319 | } | |
320 | } | |
2601e83e | 321 | |
22d8abdc | 322 | size_t loc_stringpool_write(struct loc_stringpool* pool, FILE* f) { |
2601e83e MT |
323 | size_t size = loc_stringpool_get_size(pool); |
324 | ||
8f5b676a | 325 | return fwrite(pool->data, sizeof(*pool->data), size, f); |
2601e83e | 326 | } |