Loading...
1 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 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 | // SPDX-License-Identifier: GPL-2.0 /* * Copyright (C) 2021 VMware Inc, Steven Rostedt <rostedt@goodmis.org> */ #include <linux/spinlock.h> #include <linux/irq_work.h> #include <linux/slab.h> #include "trace.h" /* See pid_list.h for details */ static inline union lower_chunk *get_lower_chunk(struct trace_pid_list *pid_list) { union lower_chunk *chunk; lockdep_assert_held(&pid_list->lock); if (!pid_list->lower_list) return NULL; chunk = pid_list->lower_list; pid_list->lower_list = chunk->next; pid_list->free_lower_chunks--; WARN_ON_ONCE(pid_list->free_lower_chunks < 0); chunk->next = NULL; /* * If a refill needs to happen, it can not happen here * as the scheduler run queue locks are held. */ if (pid_list->free_lower_chunks <= CHUNK_REALLOC) irq_work_queue(&pid_list->refill_irqwork); return chunk; } static inline union upper_chunk *get_upper_chunk(struct trace_pid_list *pid_list) { union upper_chunk *chunk; lockdep_assert_held(&pid_list->lock); if (!pid_list->upper_list) return NULL; chunk = pid_list->upper_list; pid_list->upper_list = chunk->next; pid_list->free_upper_chunks--; WARN_ON_ONCE(pid_list->free_upper_chunks < 0); chunk->next = NULL; /* * If a refill needs to happen, it can not happen here * as the scheduler run queue locks are held. */ if (pid_list->free_upper_chunks <= CHUNK_REALLOC) irq_work_queue(&pid_list->refill_irqwork); return chunk; } static inline void put_lower_chunk(struct trace_pid_list *pid_list, union lower_chunk *chunk) { lockdep_assert_held(&pid_list->lock); chunk->next = pid_list->lower_list; pid_list->lower_list = chunk; pid_list->free_lower_chunks++; } static inline void put_upper_chunk(struct trace_pid_list *pid_list, union upper_chunk *chunk) { lockdep_assert_held(&pid_list->lock); chunk->next = pid_list->upper_list; pid_list->upper_list = chunk; pid_list->free_upper_chunks++; } static inline bool upper_empty(union upper_chunk *chunk) { /* * If chunk->data has no lower chunks, it will be the same * as a zeroed bitmask. Use find_first_bit() to test it * and if it doesn't find any bits set, then the array * is empty. */ int bit = find_first_bit((unsigned long *)chunk->data, sizeof(chunk->data) * 8); return bit >= sizeof(chunk->data) * 8; } static inline int pid_split(unsigned int pid, unsigned int *upper1, unsigned int *upper2, unsigned int *lower) { /* MAX_PID should cover all pids */ BUILD_BUG_ON(MAX_PID < PID_MAX_LIMIT); /* In case a bad pid is passed in, then fail */ if (unlikely(pid >= MAX_PID)) return -1; *upper1 = (pid >> UPPER1_SHIFT) & UPPER_MASK; *upper2 = (pid >> UPPER2_SHIFT) & UPPER_MASK; *lower = pid & LOWER_MASK; return 0; } static inline unsigned int pid_join(unsigned int upper1, unsigned int upper2, unsigned int lower) { return ((upper1 & UPPER_MASK) << UPPER1_SHIFT) | ((upper2 & UPPER_MASK) << UPPER2_SHIFT) | (lower & LOWER_MASK); } /** * trace_pid_list_is_set - test if the pid is set in the list * @pid_list: The pid list to test * @pid: The pid to see if set in the list. * * Tests if @pid is set in the @pid_list. This is usually called * from the scheduler when a task is scheduled. Its pid is checked * if it should be traced or not. * * Return true if the pid is in the list, false otherwise. */ bool trace_pid_list_is_set(struct trace_pid_list *pid_list, unsigned int pid) { union upper_chunk *upper_chunk; union lower_chunk *lower_chunk; unsigned long flags; unsigned int upper1; unsigned int upper2; unsigned int lower; bool ret = false; if (!pid_list) return false; if (pid_split(pid, &upper1, &upper2, &lower) < 0) return false; raw_spin_lock_irqsave(&pid_list->lock, flags); upper_chunk = pid_list->upper[upper1]; if (upper_chunk) { lower_chunk = upper_chunk->data[upper2]; if (lower_chunk) ret = test_bit(lower, lower_chunk->data); } raw_spin_unlock_irqrestore(&pid_list->lock, flags); return ret; } /** * trace_pid_list_set - add a pid to the list * @pid_list: The pid list to add the @pid to. * @pid: The pid to add. * * Adds @pid to @pid_list. This is usually done explicitly by a user * adding a task to be traced, or indirectly by the fork function * when children should be traced and a task's pid is in the list. * * Return 0 on success, negative otherwise. */ int trace_pid_list_set(struct trace_pid_list *pid_list, unsigned int pid) { union upper_chunk *upper_chunk; union lower_chunk *lower_chunk; unsigned long flags; unsigned int upper1; unsigned int upper2; unsigned int lower; int ret; if (!pid_list) return -ENODEV; if (pid_split(pid, &upper1, &upper2, &lower) < 0) return -EINVAL; raw_spin_lock_irqsave(&pid_list->lock, flags); upper_chunk = pid_list->upper[upper1]; if (!upper_chunk) { upper_chunk = get_upper_chunk(pid_list); if (!upper_chunk) { ret = -ENOMEM; goto out; } pid_list->upper[upper1] = upper_chunk; } lower_chunk = upper_chunk->data[upper2]; if (!lower_chunk) { lower_chunk = get_lower_chunk(pid_list); if (!lower_chunk) { ret = -ENOMEM; goto out; } upper_chunk->data[upper2] = lower_chunk; } set_bit(lower, lower_chunk->data); ret = 0; out: raw_spin_unlock_irqrestore(&pid_list->lock, flags); return ret; } /** * trace_pid_list_clear - remove a pid from the list * @pid_list: The pid list to remove the @pid from. * @pid: The pid to remove. * * Removes @pid from @pid_list. This is usually done explicitly by a user * removing tasks from tracing, or indirectly by the exit function * when a task that is set to be traced exits. * * Return 0 on success, negative otherwise. */ int trace_pid_list_clear(struct trace_pid_list *pid_list, unsigned int pid) { union upper_chunk *upper_chunk; union lower_chunk *lower_chunk; unsigned long flags; unsigned int upper1; unsigned int upper2; unsigned int lower; if (!pid_list) return -ENODEV; if (pid_split(pid, &upper1, &upper2, &lower) < 0) return -EINVAL; raw_spin_lock_irqsave(&pid_list->lock, flags); upper_chunk = pid_list->upper[upper1]; if (!upper_chunk) goto out; lower_chunk = upper_chunk->data[upper2]; if (!lower_chunk) goto out; clear_bit(lower, lower_chunk->data); /* if there's no more bits set, add it to the free list */ if (find_first_bit(lower_chunk->data, LOWER_MAX) >= LOWER_MAX) { put_lower_chunk(pid_list, lower_chunk); upper_chunk->data[upper2] = NULL; if (upper_empty(upper_chunk)) { put_upper_chunk(pid_list, upper_chunk); pid_list->upper[upper1] = NULL; } } out: raw_spin_unlock_irqrestore(&pid_list->lock, flags); return 0; } /** * trace_pid_list_next - return the next pid in the list * @pid_list: The pid list to examine. * @pid: The pid to start from * @next: The pointer to place the pid that is set starting from @pid. * * Looks for the next consecutive pid that is in @pid_list starting * at the pid specified by @pid. If one is set (including @pid), then * that pid is placed into @next. * * Return 0 when a pid is found, -1 if there are no more pids included. */ int trace_pid_list_next(struct trace_pid_list *pid_list, unsigned int pid, unsigned int *next) { union upper_chunk *upper_chunk; union lower_chunk *lower_chunk; unsigned long flags; unsigned int upper1; unsigned int upper2; unsigned int lower; if (!pid_list) return -ENODEV; if (pid_split(pid, &upper1, &upper2, &lower) < 0) return -EINVAL; raw_spin_lock_irqsave(&pid_list->lock, flags); for (; upper1 <= UPPER_MASK; upper1++, upper2 = 0) { upper_chunk = pid_list->upper[upper1]; if (!upper_chunk) continue; for (; upper2 <= UPPER_MASK; upper2++, lower = 0) { lower_chunk = upper_chunk->data[upper2]; if (!lower_chunk) continue; lower = find_next_bit(lower_chunk->data, LOWER_MAX, lower); if (lower < LOWER_MAX) goto found; } } found: raw_spin_unlock_irqrestore(&pid_list->lock, flags); if (upper1 > UPPER_MASK) return -1; *next = pid_join(upper1, upper2, lower); return 0; } /** * trace_pid_list_first - return the first pid in the list * @pid_list: The pid list to examine. * @pid: The pointer to place the pid first found pid that is set. * * Looks for the first pid that is set in @pid_list, and places it * into @pid if found. * * Return 0 when a pid is found, -1 if there are no pids set. */ int trace_pid_list_first(struct trace_pid_list *pid_list, unsigned int *pid) { return trace_pid_list_next(pid_list, 0, pid); } static void pid_list_refill_irq(struct irq_work *iwork) { struct trace_pid_list *pid_list = container_of(iwork, struct trace_pid_list, refill_irqwork); union upper_chunk *upper = NULL; union lower_chunk *lower = NULL; union upper_chunk **upper_next = &upper; union lower_chunk **lower_next = &lower; int upper_count; int lower_count; int ucnt = 0; int lcnt = 0; again: raw_spin_lock(&pid_list->lock); upper_count = CHUNK_ALLOC - pid_list->free_upper_chunks; lower_count = CHUNK_ALLOC - pid_list->free_lower_chunks; raw_spin_unlock(&pid_list->lock); if (upper_count <= 0 && lower_count <= 0) return; while (upper_count-- > 0) { union upper_chunk *chunk; chunk = kzalloc(sizeof(*chunk), GFP_KERNEL); if (!chunk) break; *upper_next = chunk; upper_next = &chunk->next; ucnt++; } while (lower_count-- > 0) { union lower_chunk *chunk; chunk = kzalloc(sizeof(*chunk), GFP_KERNEL); if (!chunk) break; *lower_next = chunk; lower_next = &chunk->next; lcnt++; } raw_spin_lock(&pid_list->lock); if (upper) { *upper_next = pid_list->upper_list; pid_list->upper_list = upper; pid_list->free_upper_chunks += ucnt; } if (lower) { *lower_next = pid_list->lower_list; pid_list->lower_list = lower; pid_list->free_lower_chunks += lcnt; } raw_spin_unlock(&pid_list->lock); /* * On success of allocating all the chunks, both counters * will be less than zero. If they are not, then an allocation * failed, and we should not try again. */ if (upper_count >= 0 || lower_count >= 0) return; /* * When the locks were released, free chunks could have * been used and allocation needs to be done again. Might as * well allocate it now. */ goto again; } /** * trace_pid_list_alloc - create a new pid_list * * Allocates a new pid_list to store pids into. * * Returns the pid_list on success, NULL otherwise. */ struct trace_pid_list *trace_pid_list_alloc(void) { struct trace_pid_list *pid_list; int i; /* According to linux/thread.h, pids can be no bigger that 30 bits */ WARN_ON_ONCE(pid_max > (1 << 30)); pid_list = kzalloc(sizeof(*pid_list), GFP_KERNEL); if (!pid_list) return NULL; init_irq_work(&pid_list->refill_irqwork, pid_list_refill_irq); raw_spin_lock_init(&pid_list->lock); for (i = 0; i < CHUNK_ALLOC; i++) { union upper_chunk *chunk; chunk = kzalloc(sizeof(*chunk), GFP_KERNEL); if (!chunk) break; chunk->next = pid_list->upper_list; pid_list->upper_list = chunk; pid_list->free_upper_chunks++; } for (i = 0; i < CHUNK_ALLOC; i++) { union lower_chunk *chunk; chunk = kzalloc(sizeof(*chunk), GFP_KERNEL); if (!chunk) break; chunk->next = pid_list->lower_list; pid_list->lower_list = chunk; pid_list->free_lower_chunks++; } return pid_list; } /** * trace_pid_list_free - Frees an allocated pid_list. * * Frees the memory for a pid_list that was allocated. */ void trace_pid_list_free(struct trace_pid_list *pid_list) { union upper_chunk *upper; union lower_chunk *lower; int i, j; if (!pid_list) return; irq_work_sync(&pid_list->refill_irqwork); while (pid_list->lower_list) { union lower_chunk *chunk; chunk = pid_list->lower_list; pid_list->lower_list = pid_list->lower_list->next; kfree(chunk); } while (pid_list->upper_list) { union upper_chunk *chunk; chunk = pid_list->upper_list; pid_list->upper_list = pid_list->upper_list->next; kfree(chunk); } for (i = 0; i < UPPER1_SIZE; i++) { upper = pid_list->upper[i]; if (upper) { for (j = 0; j < UPPER2_SIZE; j++) { lower = upper->data[j]; kfree(lower); } kfree(upper); } } kfree(pid_list); } |