| 1 |
/* |
| 2 |
* collectd - src/utils/message_parser/message_parser.c |
| 3 |
* MIT License |
| 4 |
* |
| 5 |
* Copyright(c) 2017-2018 Intel Corporation. All rights reserved. |
| 6 |
* |
| 7 |
* Permission is hereby granted, free of charge, to any person obtaining a |
| 8 |
* copy of this software and associated documentation files (the "Software"), |
| 9 |
* to deal in the Software without restriction, including without limitation |
| 10 |
* the rights to use, copy, modify, merge, publish, distribute, sublicense, |
| 11 |
* and/or sell copies of the Software, and to permit persons to whom the |
| 12 |
* Software is furnished to do so, subject to the following conditions: |
| 13 |
* |
| 14 |
* The above copyright notice and this permission notice shall be included in |
| 15 |
* all copies or substantial portions of the Software. |
| 16 |
* |
| 17 |
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR |
| 18 |
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
| 19 |
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE |
| 20 |
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER |
| 21 |
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING |
| 22 |
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER |
| 23 |
* DEALINGS IN THE SOFTWARE. |
| 24 |
* |
| 25 |
* Authors: |
| 26 |
* Krzysztof Matczak <[email protected]> |
| 27 |
* Marcin Mozejko <[email protected]> |
| 28 |
*/ |
| 29 |
|
| 30 |
#include "collectd.h" |
| 31 |
|
| 32 |
#include "plugin.h" |
| 33 |
#include "utils/common/common.h" |
| 34 |
|
| 35 |
#include "utils/message_parser/message_parser.h" |
| 36 |
|
| 37 |
#define UTIL_NAME "utils_message_parser" |
| 38 |
|
| 39 |
#define MSG_STOR_INIT_LEN 64 |
| 40 |
#define MSG_STOR_INC_STEP 10 |
| 41 |
|
| 42 |
typedef struct checked_match_s { |
| 43 |
parser_job_data_t *parser_job; |
| 44 |
message_pattern_t msg_pattern; |
| 45 |
int msg_pattern_idx; |
| 46 |
} checked_match_t; |
| 47 |
|
| 48 |
struct parser_job_data_s { |
| 49 |
const char *filename; |
| 50 |
unsigned int start_idx; |
| 51 |
unsigned int stop_idx; |
| 52 |
cu_tail_match_t *tm; |
| 53 |
message_t *messages_storage; |
| 54 |
size_t messages_max_len; |
| 55 |
int message_idx; |
| 56 |
unsigned int message_item_idx; |
| 57 |
unsigned int messages_completed; |
| 58 |
message_pattern_t *message_patterns; |
| 59 |
size_t message_patterns_len; |
| 60 |
int (*resize_message_buffer)(parser_job_data_t *self, size_t); |
| 61 |
int (*start_message_assembly)(parser_job_data_t *self); |
| 62 |
void (*end_message_assembly)(parser_job_data_t *self); |
| 63 |
void (*message_item_assembly)(parser_job_data_t *self, checked_match_t *cm, |
| 64 |
char *const *matches); |
| 65 |
}; |
| 66 |
|
| 67 |
static void message_item_assembly(parser_job_data_t *self, checked_match_t *cm, |
| 68 |
char *const *matches) { |
| 69 |
message_item_t *msg_it = &self->messages_storage[self->message_idx] |
| 70 |
.message_items[self->message_item_idx]; |
| 71 |
sstrncpy(msg_it->name, cm->msg_pattern.name, sizeof(msg_it->name)); |
| 72 |
sstrncpy(msg_it->value, matches[cm->msg_pattern.submatch_idx], |
| 73 |
sizeof(msg_it->value)); |
| 74 |
msg_it->user_data = cm->msg_pattern.user_data; |
| 75 |
msg_it->free_user_data = cm->msg_pattern.free_user_data; |
| 76 |
++(self->message_item_idx); |
| 77 |
} |
| 78 |
|
| 79 |
static int start_message_assembly(parser_job_data_t *self) { |
| 80 |
/* Remove previous message assembly if unfinished */ |
| 81 |
if (self->message_idx >= 0 && |
| 82 |
self->messages_storage[self->message_idx].started == true && |
| 83 |
self->messages_storage[self->message_idx].completed == false) { |
| 84 |
DEBUG(UTIL_NAME ": Removing unfinished assembly of previous message"); |
| 85 |
self->messages_storage[self->message_idx] = (message_t){{{{0}}}}; |
| 86 |
self->message_item_idx = 0; |
| 87 |
} else |
| 88 |
++(self->message_idx); |
| 89 |
|
| 90 |
/* Resize messages buffer if needed */ |
| 91 |
if (self->message_idx >= self->messages_max_len) { |
| 92 |
INFO(UTIL_NAME ": Exceeded message buffer size: %zu", |
| 93 |
self->messages_max_len); |
| 94 |
if (self->resize_message_buffer(self, self->messages_max_len + |
| 95 |
MSG_STOR_INC_STEP) != 0) { |
| 96 |
ERROR(UTIL_NAME ": Insufficient message buffer size: %zu. Remaining " |
| 97 |
"messages for this read will be skipped", |
| 98 |
self->messages_max_len); |
| 99 |
self->message_idx = self->messages_max_len; |
| 100 |
return -1; |
| 101 |
} |
| 102 |
} |
| 103 |
self->messages_storage[self->message_idx] = (message_t){{{{0}}}}; |
| 104 |
self->message_item_idx = 0; |
| 105 |
self->messages_storage[self->message_idx].started = true; |
| 106 |
self->messages_storage[self->message_idx].completed = false; |
| 107 |
return 0; |
| 108 |
} |
| 109 |
|
| 110 |
static int resize_message_buffer(parser_job_data_t *self, size_t new_size) { |
| 111 |
INFO(UTIL_NAME ": Resizing message buffer size to %zu", new_size); |
| 112 |
void *new_storage = realloc(self->messages_storage, |
| 113 |
new_size * sizeof(*self->messages_storage)); |
| 114 |
if (new_storage == NULL) { |
| 115 |
ERROR(UTIL_NAME ": Error while reallocating message buffer"); |
| 116 |
return -1; |
| 117 |
} |
| 118 |
self->messages_storage = new_storage; |
| 119 |
self->messages_max_len = new_size; |
| 120 |
/* Fill unused memory with 0 */ |
| 121 |
unsigned int unused_idx = self->message_idx < 0 ? 0 : self->message_idx; |
| 122 |
memset(self->messages_storage + unused_idx, 0, |
| 123 |
(self->messages_max_len - unused_idx) * |
| 124 |
sizeof(*self->messages_storage)); |
| 125 |
return 0; |
| 126 |
} |
| 127 |
|
| 128 |
static void end_message_assembly(parser_job_data_t *self) { |
| 129 |
/* Checking mandatory items */ |
| 130 |
for (size_t i = 0; i < self->message_patterns_len; i++) { |
| 131 |
if (self->message_patterns[i].is_mandatory && |
| 132 |
!(self->messages_storage[self->message_idx] |
| 133 |
.matched_patterns_check[i])) { |
| 134 |
WARNING( |
| 135 |
UTIL_NAME |
| 136 |
": Mandatory message item pattern %s not found. Message discarded", |
| 137 |
self->message_patterns[i].regex); |
| 138 |
self->messages_storage[self->message_idx] = (message_t){{{{0}}}}; |
| 139 |
self->message_item_idx = 0; |
| 140 |
if (self->message_idx > 0) |
| 141 |
--(self->message_idx); |
| 142 |
return; |
| 143 |
} |
| 144 |
} |
| 145 |
self->messages_storage[self->message_idx].completed = true; |
| 146 |
++self->messages_completed; |
| 147 |
self->message_item_idx = 0; |
| 148 |
} |
| 149 |
|
| 150 |
static int message_assembler(const char *row, char *const *matches, |
| 151 |
size_t matches_num, void *user_data) { |
| 152 |
if (user_data == NULL) { |
| 153 |
ERROR(UTIL_NAME ": Invalid user_data pointer"); |
| 154 |
return -1; |
| 155 |
} |
| 156 |
checked_match_t *cm = (checked_match_t *)user_data; |
| 157 |
parser_job_data_t *parser_job = cm->parser_job; |
| 158 |
|
| 159 |
if (cm->msg_pattern.submatch_idx < -1 || |
| 160 |
cm->msg_pattern.submatch_idx >= (int)matches_num) { |
| 161 |
ERROR(UTIL_NAME ": Invalid target submatch index: %d", |
| 162 |
cm->msg_pattern.submatch_idx); |
| 163 |
return -1; |
| 164 |
} |
| 165 |
if (parser_job->message_item_idx >= |
| 166 |
STATIC_ARRAY_SIZE(parser_job->messages_storage[parser_job->message_idx] |
| 167 |
.message_items)) { |
| 168 |
ERROR(UTIL_NAME ": Message items number exceeded. Forced message end."); |
| 169 |
parser_job->end_message_assembly(parser_job); |
| 170 |
return -1; |
| 171 |
} |
| 172 |
|
| 173 |
/* Every matched start pattern resets current message items and starts |
| 174 |
* assembling new messages */ |
| 175 |
if (strcmp(cm->msg_pattern.regex, |
| 176 |
parser_job->message_patterns[parser_job->start_idx].regex) == 0) { |
| 177 |
DEBUG(UTIL_NAME ": Found beginning pattern"); |
| 178 |
if (parser_job->start_message_assembly(parser_job) != 0) |
| 179 |
return -1; |
| 180 |
} |
| 181 |
/* Ignoring message items without corresponding start item or |
| 182 |
* after completion */ |
| 183 |
if (parser_job->message_idx < 0 || |
| 184 |
parser_job->messages_storage[parser_job->message_idx].started == false || |
| 185 |
parser_job->messages_storage[parser_job->message_idx].completed == true) { |
| 186 |
DEBUG(UTIL_NAME ": Dropping item with no corresponding start element"); |
| 187 |
return 0; |
| 188 |
} |
| 189 |
/* Populate message items */ |
| 190 |
if (cm->msg_pattern.submatch_idx >= 0) |
| 191 |
parser_job->message_item_assembly(parser_job, cm, matches); |
| 192 |
|
| 193 |
/* Mark pattern as checked */ |
| 194 |
parser_job->messages_storage[parser_job->message_idx] |
| 195 |
.matched_patterns_check[cm->msg_pattern_idx] = 1; |
| 196 |
|
| 197 |
/* Handle message ending */ |
| 198 |
if (strcmp(cm->msg_pattern.regex, |
| 199 |
parser_job->message_patterns[parser_job->stop_idx].regex) == 0) { |
| 200 |
DEBUG(UTIL_NAME ": Found ending pattern"); |
| 201 |
parser_job->end_message_assembly(parser_job); |
| 202 |
} |
| 203 |
return 0; |
| 204 |
} |
| 205 |
|
| 206 |
parser_job_data_t *message_parser_init(const char *filename, |
| 207 |
unsigned int start_idx, |
| 208 |
unsigned int stop_idx, |
| 209 |
message_pattern_t message_patterns[], |
| 210 |
size_t message_patterns_len) { |
| 211 |
parser_job_data_t *parser_job = calloc(1, sizeof(*parser_job)); |
| 212 |
if (parser_job == NULL) { |
| 213 |
ERROR(UTIL_NAME ": Error allocating parser_job"); |
| 214 |
return NULL; |
| 215 |
} |
| 216 |
parser_job->resize_message_buffer = resize_message_buffer; |
| 217 |
parser_job->start_message_assembly = start_message_assembly; |
| 218 |
parser_job->end_message_assembly = end_message_assembly; |
| 219 |
parser_job->message_item_assembly = message_item_assembly; |
| 220 |
parser_job->messages_max_len = MSG_STOR_INIT_LEN; |
| 221 |
parser_job->filename = filename; |
| 222 |
parser_job->start_idx = start_idx; |
| 223 |
parser_job->stop_idx = stop_idx; |
| 224 |
parser_job->message_idx = -1; |
| 225 |
parser_job->messages_completed = 0; |
| 226 |
parser_job->message_patterns = |
| 227 |
calloc(message_patterns_len, sizeof(*parser_job->message_patterns)); |
| 228 |
if (parser_job->message_patterns == NULL) { |
| 229 |
ERROR(UTIL_NAME ": Error allocating message_patterns"); |
| 230 |
goto free_parser_job; |
| 231 |
} |
| 232 |
parser_job->messages_storage = calloc(parser_job->messages_max_len, |
| 233 |
sizeof(*parser_job->messages_storage)); |
| 234 |
if (parser_job->messages_storage == NULL) { |
| 235 |
ERROR(UTIL_NAME ": Error allocating messages_storage"); |
| 236 |
goto free_msg_patterns; |
| 237 |
} |
| 238 |
/* Crete own copy of regex patterns */ |
| 239 |
memcpy(parser_job->message_patterns, message_patterns, |
| 240 |
sizeof(*parser_job->message_patterns) * message_patterns_len); |
| 241 |
parser_job->message_patterns_len = message_patterns_len; |
| 242 |
/* Init tail match */ |
| 243 |
parser_job->tm = tail_match_create(parser_job->filename); |
| 244 |
if (parser_job->tm == NULL) { |
| 245 |
ERROR(UTIL_NAME ": Error creating tail match"); |
| 246 |
goto free_msg_storage; |
| 247 |
} |
| 248 |
|
| 249 |
for (size_t i = 0; i < message_patterns_len; i++) { |
| 250 |
/* Create current_match container for passing regex info |
| 251 |
* to callback function */ |
| 252 |
checked_match_t *current_match = calloc(1, sizeof(*current_match)); |
| 253 |
if (current_match == NULL) { |
| 254 |
ERROR(UTIL_NAME ": Error allocating current_match"); |
| 255 |
goto free_tail_match; |
| 256 |
} |
| 257 |
current_match->parser_job = parser_job; |
| 258 |
current_match->msg_pattern = message_patterns[i]; |
| 259 |
current_match->msg_pattern_idx = i; |
| 260 |
/* Create callback */ |
| 261 |
cu_match_t *m = match_create_callback( |
| 262 |
message_patterns[i].regex, message_patterns[i].excluderegex, |
| 263 |
message_assembler, current_match, free); |
| 264 |
if (m == NULL) { |
| 265 |
ERROR(UTIL_NAME ": Error creating match callback"); |
| 266 |
goto free_tail_match; |
| 267 |
} |
| 268 |
if (tail_match_add_match(parser_job->tm, m, 0, 0, 0) != 0) { |
| 269 |
ERROR(UTIL_NAME ": Error adding match callback"); |
| 270 |
goto free_tail_match; |
| 271 |
} |
| 272 |
} |
| 273 |
|
| 274 |
return parser_job; |
| 275 |
|
| 276 |
free_tail_match: |
| 277 |
tail_match_destroy(parser_job->tm); |
| 278 |
free_msg_storage: |
| 279 |
sfree(parser_job->messages_storage); |
| 280 |
free_msg_patterns: |
| 281 |
sfree(parser_job->message_patterns); |
| 282 |
free_parser_job: |
| 283 |
sfree(parser_job); |
| 284 |
|
| 285 |
return NULL; |
| 286 |
} |
| 287 |
|
| 288 |
int message_parser_read(parser_job_data_t *parser_job, |
| 289 |
message_t **messages_storage, bool force_rewind) { |
| 290 |
if (parser_job == NULL) { |
| 291 |
ERROR(UTIL_NAME ": Invalid parser_job pointer"); |
| 292 |
return -1; |
| 293 |
} |
| 294 |
parser_job->messages_completed = 0; |
| 295 |
bool incomplete_msg_found = false; |
| 296 |
/* Finish incomplete message assembly in this read */ |
| 297 |
if (parser_job->message_idx >= 0 && |
| 298 |
parser_job->messages_storage[parser_job->message_idx].started && |
| 299 |
!(parser_job->messages_storage[parser_job->message_idx].completed)) { |
| 300 |
INFO(UTIL_NAME ": Found incomplete message from previous read."); |
| 301 |
incomplete_msg_found = true; |
| 302 |
message_t tmp_message = |
| 303 |
parser_job->messages_storage[parser_job->message_idx]; |
| 304 |
int tmp_message_item_idx = parser_job->message_item_idx; |
| 305 |
memset(parser_job->messages_storage, 0, |
| 306 |
parser_job->messages_max_len * |
| 307 |
sizeof(*parser_job->messages_storage)); |
| 308 |
memcpy(parser_job->messages_storage, &tmp_message, |
| 309 |
sizeof(*parser_job->messages_storage)); |
| 310 |
parser_job->message_item_idx = tmp_message_item_idx; |
| 311 |
parser_job->message_idx = 0; |
| 312 |
} |
| 313 |
/* Reset message buffer after non empty read */ |
| 314 |
else if (parser_job->message_idx >= 0) { |
| 315 |
memset(parser_job->messages_storage, 0, |
| 316 |
parser_job->messages_max_len * |
| 317 |
sizeof(*parser_job->messages_storage)); |
| 318 |
parser_job->message_item_idx = 0; |
| 319 |
parser_job->message_idx = -1; |
| 320 |
} |
| 321 |
|
| 322 |
int status = tail_match_read(parser_job->tm, force_rewind); |
| 323 |
if (status != 0) { |
| 324 |
ERROR(UTIL_NAME ": Error while parser read. Status: %d", status); |
| 325 |
return -1; |
| 326 |
} |
| 327 |
|
| 328 |
/* Restore initial message buffer size if there's nothing to parse */ |
| 329 |
if (parser_job->messages_completed == 0 && |
| 330 |
parser_job->messages_max_len > MSG_STOR_INIT_LEN && !incomplete_msg_found) |
| 331 |
parser_job->resize_message_buffer(parser_job, MSG_STOR_INIT_LEN); |
| 332 |
|
| 333 |
*messages_storage = parser_job->messages_storage; |
| 334 |
return parser_job->messages_completed; |
| 335 |
} |
| 336 |
|
| 337 |
void message_parser_cleanup(parser_job_data_t *parser_job) { |
| 338 |
if (parser_job == NULL) { |
| 339 |
ERROR(UTIL_NAME ": Invalid parser_job pointer"); |
| 340 |
return; |
| 341 |
} |
| 342 |
sfree(parser_job->messages_storage); |
| 343 |
sfree(parser_job->message_patterns); |
| 344 |
if (parser_job->tm) |
| 345 |
tail_match_destroy(parser_job->tm); |
| 346 |
sfree(parser_job); |
| 347 |
} |