| 1 | |
| 2 | * Generic ring buffer
|
| 3 | *
|
| 4 | * Copyright (C) 2008 Steven Rostedt <[email protected]>
|
| 5 |
|
| 6 | #include <linux/ring_buffer.h>
|
| 7 | #include <linux/trace_clock.h>
|
| 8 | #include <linux/ftrace_irq.h>
|
| 9 | #include <linux/spinlock.h>
|
| 10 | #include <linux/debugfs.h>
|
| 11 | #include <linux/uaccess.h>
|
| 12 | #include <linux/hardirq.h>
|
| 13 | #include <linux/kmemcheck.h>
|
| 14 | #include <linux/module.h>
|
| 15 | #include <linux/percpu.h>
|
| 16 | #include <linux/mutex.h>
|
| 17 | #include <linux/init.h>
|
| 18 | #include <linux/hash.h>
|
| 19 | #include <linux/list.h>
|
| 20 | #include <linux/cpu.h>
|
| 21 | #include <linux/fs.h>
|
| 22 |
|
| 23 | #include "trace.h"
|
| 24 |
|
| 25 | |
| 26 | * The ring buffer header is special. We must manually up keep it.
|
| 27 |
|
| 28 | int ring_buffer_print_entry_header(struct trace_seq *s)
|
| 29 | {
|
| 30 | int ret;
|
| 31 |
|
| 32 | ret = trace_seq_printf(s, "# compressed entry header\n");
|
| 33 | ret = trace_seq_printf(s, "\ttype_len : 5 bits\n");
|
| 34 | ret = trace_seq_printf(s, "\ttime_delta : 27 bits\n");
|
| 35 | ret = trace_seq_printf(s, "\tarray : 32 bits\n");
|
| 36 | ret = trace_seq_printf(s, "\n");
|
| 37 | ret = trace_seq_printf(s, "\tpadding : type == %d\n",
|
| 38 | RINGBUF_TYPE_PADDING);
|
| 39 | ret = trace_seq_printf(s, "\ttime_extend : type == %d\n",
|
| 40 | RINGBUF_TYPE_TIME_EXTEND);
|
| 41 | ret = trace_seq_printf(s, "\tdata max type_len == %d\n",
|
| 42 | RINGBUF_TYPE_DATA_TYPE_LEN_MAX);
|
| 43 |
|
| 44 | return ret;
|
| 45 | }
|
| 46 |
|
| 47 | |
| 48 | * The ring buffer is made up of a list of pages. A separate list of pages is
|
| 49 | * allocated for each CPU. A writer may only write to a buffer that is
|
| 50 | * associated with the CPU it is currently executing on. A reader may read
|
| 51 | * from any per cpu buffer.
|
| 52 | *
|
| 53 | * The reader is special. For each per cpu buffer, the reader has its own
|
| 54 | * reader page. When a reader has read the entire reader page, this reader
|
| 55 | * page is swapped with another page in the ring buffer.
|
| 56 | *
|
| 57 | * Now, as long as the writer is off the reader page, the reader can do what
|
| 58 | * ever it wants with that page. The writer will never write to that page
|
| 59 | * again (as long as it is out of the ring buffer).
|
| 60 | *
|
| 61 | * Here's some silly ASCII art.
|
| 62 | *
|
| 63 | * +------+
|
| 64 | * |reader| RING BUFFER
|
| 65 | * |page |
|
| 66 | * +------+ +---+ +---+ +---+
|
| 67 | * | |-->| |-->| |
|
| 68 | * +---+ +---+ +---+
|
| 69 | * ^ |
|
| 70 | * | |
|
| 71 | * +---------------+
|
| 72 | *
|
| 73 | *
|
| 74 | * +------+
|
| 75 | * |reader| RING BUFFER
|
| 76 | * |page |------------------v
|
| 77 | * +------+ +---+ +---+ +---+
|
| 78 | * | |-->| |-->| |
|
| 79 | * +---+ +---+ +---+
|
| 80 | * ^ |
|
| 81 | * | |
|
| 82 | * +---------------+
|
| 83 | *
|
| 84 | *
|
| 85 | * +------+
|
| 86 | * |reader| RING BUFFER
|
| 87 | * |page |------------------v
|
| 88 | * +------+ +---+ +---+ +---+
|
| 89 | * ^ | |-->| |-->| |
|
| 90 | * | +---+ +---+ +---+
|
| 91 | * | |
|
| 92 | * | |
|
| 93 | * +------------------------------+
|
| 94 | *
|
| 95 | *
|
| 96 | * +------+
|
| 97 | * |buffer| RING BUFFER
|
| 98 | * |page |------------------v
|
| 99 | * +------+ +---+ +---+ +---+
|
| 100 | * ^ | | | |-->| |
|
| 101 | * | New +---+ +---+ +---+
|
| 102 | * | Reader------^ |
|
| 103 | * | page |
|
| 104 | * +------------------------------+
|
| 105 | *
|
| 106 | *
|
| 107 | * After we make this swap, the reader can hand this page off to the splice
|
| 108 | * code and be done with it. It can even allocate a new page if it needs to
|
| 109 | * and swap that into the ring buffer.
|
| 110 | *
|
| 111 | * We will be using cmpxchg soon to make all this lockless.
|
| 112 | *
|
| 113 |
|
| 114 |
|
| 115 | |
| 116 | * A fast way to enable or disable all ring buffers is to
|
| 117 | * call tracing_on or tracing_off. Turning off the ring buffers
|
| 118 | * prevents all ring buffers from being recorded to.
|
| 119 | * Turning this switch on, makes it OK to write to the
|
| 120 | * ring buffer, if the ring buffer is enabled itself.
|
| 121 | *
|
| 122 | * There's three layers that must be on in order to write
|
| 123 | * to the ring buffer.
|
| 124 | *
|
| 125 | * 1) This global flag must be set.
|
| 126 | * 2) The ring buffer must be enabled for recording.
|
| 127 | * 3) The per cpu buffer must be enabled for recording.
|
| 128 | *
|
| 129 | * In case of an anomaly, this global flag has a bit set that
|
| 130 | * will permantly disable all ring buffers.
|
| 131 |
|
| 132 |
|
| 133 | |
| 134 | * Global flag to disable all recording to ring buffers
|
| 135 | * This has two bits: ON, DISABLED
|
| 136 | *
|
| 137 | * ON DISABLED
|
| 138 | * ---- ----------
|
| 139 | * 0 0 : ring buffers are off
|
| 140 | * 1 0 : ring buffers are on
|
| 141 | * X 1 : ring buffers are permanently disabled
|
| 142 |
|
| 143 |
|
| 144 | enum {
|
| 145 | RB_BUFFERS_ON_BIT = 0,
|
| 146 | RB_BUFFERS_DISABLED_BIT = 1,
|
| 147 | };
|
| 148 |
|
| 149 | enum {
|
| 150 | RB_BUFFERS_ON = 1 << RB_BUFFERS_ON_BIT,
|
| 151 | RB_BUFFERS_DISABLED = 1 << RB_BUFFERS_DISABLED_BIT,
|
| 152 | };
|
| 153 |
|
| 154 | static unsigned long ring_buffer_flags __read_mostly = RB_BUFFERS_ON;
|
| 155 |
|
| 156 | #define BUF_PAGE_HDR_SIZE offsetof(struct buffer_data_page, data)
|
| 157 |
|
| 158 | |
| 159 | * tracing_on - enable all tracing buffers
|
| 160 | *
|
| 161 | * This function enables all tracing buffers that may have been
|
| 162 | * disabled with tracing_off.
|
| 163 |
|
| 164 | void tracing_on(void)
|
| 165 | {
|
| 166 | set_bit(RB_BUFFERS_ON_BIT, &ring_buffer_flags);
|
| 167 | }
|
| 168 | EXPORT_SYMBOL_GPL(tracing_on);
|
| 169 |
|
| 170 | |
| 171 | * tracing_off - turn off all tracing buffers
|
| 172 | *
|
| 173 | * This function stops all tracing buffers from recording data.
|
| 174 | * It does not disable any overhead the tracers themselves may
|
| 175 | * be causing. This function simply causes all recording to
|
| 176 | * the ring buffers to fail.
|
| 177 |
|
| 178 | void tracing_off(void)
|
| 179 | {
|
| 180 | clear_bit(RB_BUFFERS_ON_BIT, &ring_buffer_flags);
|
| 181 | }
|
| 182 | EXPORT_SYMBOL_GPL(tracing_off);
|
| 183 |
|
| 184 | |
| 185 | * tracing_off_permanent - permanently disable ring buffers
|
| 186 | *
|
| 187 | * This function, once called, will disable all ring buffers
|
| 188 | * permanently.
|
| 189 |
|
| 190 | void tracing_off_permanent(void)
|
| 191 | {
|
| 192 | set_bit(RB_BUFFERS_DISABLED_BIT, &ring_buffer_flags);
|
| 193 | }
|
| 194 |
|
| 195 | |
| 196 | * tracing_is_on - show state of ring buffers enabled
|
| 197 |
|
| 198 | int tracing_is_on(void)
|
| 199 | {
|
| 200 | return ring_buffer_flags == RB_BUFFERS_ON;
|
| 201 | }
|
| 202 | EXPORT_SYMBOL_GPL(tracing_is_on);
|
| 203 |
|
| 204 | #include "trace.h"
|
| 205 |
|
| 206 | #define RB_EVNT_HDR_SIZE (offsetof(struct ring_buffer_event, array))
|
| 207 | #define RB_ALIGNMENT 4U
|
| 208 | #define RB_MAX_SMALL_DATA (RB_ALIGNMENT * RINGBUF_TYPE_DATA_TYPE_LEN_MAX)
|
| 209 | #define RB_EVNT_MIN_SIZE 8U
|
| 210 |
|
| 211 |
|
| 212 | #define RINGBUF_TYPE_DATA 0 ... RINGBUF_TYPE_DATA_TYPE_LEN_MAX
|
| 213 |
|
| 214 | enum {
|
| 215 | RB_LEN_TIME_EXTEND = 8,
|
| 216 | RB_LEN_TIME_STAMP = 16,
|
| 217 | };
|
| 218 |
|
| 219 | static inline int rb_null_event(struct ring_buffer_event *event)
|
| 220 | {
|
| 221 | return event->type_len == RINGBUF_TYPE_PADDING
|
| 222 | && event->time_delta == 0;
|
| 223 | }
|
| 224 |
|
| 225 | static inline int rb_discarded_event(struct ring_buffer_event *event)
|
| 226 | {
|
| 227 | return event->type_len == RINGBUF_TYPE_PADDING && event->time_delta;
|
| 228 | }
|
| 229 |
|
| 230 | static void rb_event_set_padding(struct ring_buffer_event *event)
|
| 231 | {
|
| 232 | event->type_len = RINGBUF_TYPE_PADDING;
|
| 233 | event->time_delta = 0;
|
| 234 | }
|
| 235 |
|
| 236 | static unsigned
|
| 237 | rb_event_data_length(struct ring_buffer_event *event)
|
| 238 | {
|
| 239 | unsigned length;
|
| 240 |
|
| 241 | if (event->type_len)
|
| 242 | length = event->type_len * RB_ALIGNMENT;
|
| 243 | else
|
| 244 | length = event->array[0];
|
| 245 | return length + RB_EVNT_HDR_SIZE;
|
| 246 | }
|
| 247 |
|
| 248 |
|
| 249 | static unsigned
|
| 250 | rb_event_length(struct ring_buffer_event *event)
|
| 251 | {
|
| 252 | switch (event->type_len) {
|
| 253 | case RINGBUF_TYPE_PADDING:
|
| 254 | if (rb_null_event(event))
|
| 255 |
|
| 256 | return -1;
|
| 257 | return event->array[0] + RB_EVNT_HDR_SIZE;
|
| 258 |
|
| 259 | case RINGBUF_TYPE_TIME_EXTEND:
|
| 260 | return RB_LEN_TIME_EXTEND;
|
| 261 |
|
| 262 | case RINGBUF_TYPE_TIME_STAMP:
|
| 263 | return RB_LEN_TIME_STAMP;
|
| 264 |
|
| 265 | case RINGBUF_TYPE_DATA:
|
| 266 | return rb_event_data_length(event);
|
| 267 | default:
|
| 268 | BUG();
|
| 269 | }
|
| 270 |
|
| 271 | return 0;
|
| 272 | }
|
| 273 |
|
| 274 | |
| 275 | * ring_buffer_event_length - return the length of the event
|
| 276 | * @event: the event to get the length of
|
| 277 |
|
| 278 | unsigned ring_buffer_event_length(struct ring_buffer_event *event)
|
| 279 | {
|
| 280 | unsigned length = rb_event_length(event);
|
| 281 | if (event->type_len > RINGBUF_TYPE_DATA_TYPE_LEN_MAX)
|
| 282 | return length;
|
| 283 | length -= RB_EVNT_HDR_SIZE;
|
| 284 | if (length > RB_MAX_SMALL_DATA + sizeof(event->array[0]))
|
| 285 | length -= sizeof(event->array[0]);
|
| 286 | return length;
|
| 287 | }
|
| 288 | EXPORT_SYMBOL_GPL(ring_buffer_event_length);
|
| 289 |
|
| 290 |
|
| 291 | static void *
|
| 292 | rb_event_data(struct ring_buffer_event *event)
|
| 293 | {
|
| 294 | BUG_ON(event->type_len > RINGBUF_TYPE_DATA_TYPE_LEN_MAX);
|
| 295 |
|
| 296 | if (event->type_len)
|
| 297 | return (void *)&event->array[0];
|
| 298 |
|
| 299 | return (void *)&event->array[1];
|
| 300 | }
|
| 301 |
|
| 302 | |
| 303 | * ring_buffer_event_data - return the data of the event
|
| 304 | * @event: the event to get the data from
|
| 305 |
|
| 306 | void *ring_buffer_event_data(struct ring_buffer_event *event)
|
| 307 | {
|
| 308 | return rb_event_data(event);
|
| 309 | }
|
| 310 | EXPORT_SYMBOL_GPL(ring_buffer_event_data);
|
| 311 |
|
| 312 | #define for_each_buffer_cpu(buffer, cpu) \
|
| 313 | for_each_cpu(cpu, buffer->cpumask)
|
| 314 |
|
| 315 | #define TS_SHIFT 27
|
| 316 | #define TS_MASK ((1ULL << TS_SHIFT) - 1)
|
| 317 | #define TS_DELTA_TEST (~TS_MASK)
|
| 318 |
|
| 319 | struct buffer_data_page {
|
| 320 | u64 time_stamp;
|
| 321 | local_t commit;
|
| 322 | unsigned char data[];
|
| 323 | };
|
| 324 |
|
| 325 | struct buffer_page {
|
| 326 | struct list_head list;
|
| 327 | local_t write;
|
| 328 | unsigned read;
|
| 329 | local_t entries;
|
| 330 | struct buffer_data_page *page;
|
| 331 | };
|
| 332 |
|
| 333 | static void rb_init_page(struct buffer_data_page *bpage)
|
| 334 | {
|
| 335 | local_set(&bpage->commit, 0);
|
| 336 | }
|
| 337 |
|
| 338 | |
| 339 | * ring_buffer_page_len - the size of data on the page.
|
| 340 | * @page: The page to read
|
| 341 | *
|
| 342 | * Returns the amount of data on the page, including buffer page header.
|
| 343 |
|
| 344 | size_t ring_buffer_page_len(void *page)
|
| 345 | {
|
| 346 | return local_read(&((struct buffer_data_page *)page)->commit)
|
| 347 | + BUF_PAGE_HDR_SIZE;
|
| 348 | }
|
| 349 |
|
| 350 | |
| 351 | * Also stolen from mm/slob.c. Thanks to Mathieu Desnoyers for pointing
|
| 352 | * this issue out.
|
| 353 |
|
| 354 | static void free_buffer_page(struct buffer_page *bpage)
|
| 355 | {
|
| 356 | free_page((unsigned long)bpage->page);
|
| 357 | kfree(bpage);
|
| 358 | }
|
| 359 |
|
| 360 | |
| 361 | * We need to fit the time_stamp delta into 27 bits.
|
| 362 |
|
| 363 | static inline int test_time_stamp(u64 delta)
|
| 364 | {
|
| 365 | if (delta & TS_DELTA_TEST)
|
| 366 | return 1;
|
| 367 | return 0;
|
| 368 | }
|
| 369 |
|
| 370 | #define BUF_PAGE_SIZE (PAGE_SIZE - BUF_PAGE_HDR_SIZE)
|
| 371 |
|
| 372 |
|
| 373 | #define BUF_MAX_DATA_SIZE (BUF_PAGE_SIZE - (sizeof(u32) * 2))
|
| 374 |
|
| 375 |
|
| 376 | #define RB_TIMESTAMPS_PER_PAGE (BUF_PAGE_SIZE / RB_LEN_TIME_STAMP)
|
| 377 |
|
| 378 | int ring_buffer_print_page_header(struct trace_seq *s)
|
| 379 | {
|
| 380 | struct buffer_data_page field;
|
| 381 | int ret;
|
| 382 |
|
| 383 | ret = trace_seq_printf(s, "\tfield: u64 timestamp;\t"
|
| 384 | "offset:0;\tsize:%u;\n",
|
| 385 | (unsigned int)sizeof(field.time_stamp));
|
| 386 |
|
| 387 | ret = trace_seq_printf(s, "\tfield: local_t commit;\t"
|
| 388 | "offset:%u;\tsize:%u;\n",
|
| 389 | (unsigned int)offsetof(typeof(field), commit),
|
| 390 | (unsigned int)sizeof(field.commit));
|
| 391 |
|
| 392 | ret = trace_seq_printf(s, "\tfield: char data;\t"
|
| 393 | "offset:%u;\tsize:%u;\n",
|
| 394 | (unsigned int)offsetof(typeof(field), data),
|
| 395 | (unsigned int)BUF_PAGE_SIZE);
|
| 396 |
|
| 397 | return ret;
|
| 398 | }
|
| 399 |
|
| 400 | |
| 401 | * head_page == tail_page && head == tail then buffer is empty.
|
| 402 |
|
| 403 | struct ring_buffer_per_cpu {
|
| 404 | int cpu;
|
| 405 | struct ring_buffer *buffer;
|
| 406 | spinlock_t reader_lock;
|
| 407 | raw_spinlock_t lock;
|
| 408 | struct lock_class_key lock_key;
|
| 409 | struct list_head pages;
|
| 410 | struct buffer_page *head_page;
|
| 411 | struct buffer_page *tail_page;
|
| 412 | struct buffer_page *commit_page;
|
| 413 | struct buffer_page *reader_page;
|
| 414 | unsigned long nmi_dropped;
|
| 415 | unsigned long commit_overrun;
|
| 416 | unsigned long overrun;
|
| 417 | unsigned long read;
|
| 418 | local_t entries;
|
| 419 | local_t committing;
|
| 420 | local_t commits;
|
| 421 | u64 write_stamp;
|
| 422 | u64 read_stamp;
|
| 423 | atomic_t record_disabled;
|
| 424 | };
|
| 425 |
|
| 426 | struct ring_buffer {
|
| 427 | unsigned pages;
|
| 428 | unsigned flags;
|
| 429 | int cpus;
|
| 430 | atomic_t record_disabled;
|
| 431 | cpumask_var_t cpumask;
|
| 432 |
|
| 433 | struct lock_class_key *reader_lock_key;
|
| 434 |
|
| 435 | struct mutex mutex;
|
| 436 |
|
| 437 | struct ring_buffer_per_cpu **buffers;
|
| 438 |
|
| 439 | #ifdef CONFIG_HOTPLUG_CPU
|
| 440 | struct notifier_block cpu_notify;
|
| 441 | #endif
|
| 442 | u64 (*clock)(void);
|
| 443 | };
|
| 444 |
|
| 445 | struct ring_buffer_iter {
|
| 446 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 447 | unsigned long head;
|
| 448 | struct buffer_page *head_page;
|
| 449 | u64 read_stamp;
|
| 450 | };
|
| 451 |
|
| 452 |
|
| 453 | #define RB_WARN_ON(buffer, cond) \
|
| 454 | ({ \
|
| 455 | int _____ret = unlikely(cond); \
|
| 456 | if (_____ret) { \
|
| 457 | atomic_inc(&buffer->record_disabled); \
|
| 458 | WARN_ON(1); \
|
| 459 | } \
|
| 460 | _____ret; \
|
| 461 | })
|
| 462 |
|
| 463 |
|
| 464 | #define DEBUG_SHIFT 0
|
| 465 |
|
| 466 | static inline u64 rb_time_stamp(struct ring_buffer *buffer, int cpu)
|
| 467 | {
|
| 468 |
|
| 469 | return buffer->clock() << DEBUG_SHIFT;
|
| 470 | }
|
| 471 |
|
| 472 | u64 ring_buffer_time_stamp(struct ring_buffer *buffer, int cpu)
|
| 473 | {
|
| 474 | u64 time;
|
| 475 |
|
| 476 | preempt_disable_notrace();
|
| 477 | time = rb_time_stamp(buffer, cpu);
|
| 478 | preempt_enable_no_resched_notrace();
|
| 479 |
|
| 480 | return time;
|
| 481 | }
|
| 482 | EXPORT_SYMBOL_GPL(ring_buffer_time_stamp);
|
| 483 |
|
| 484 | void ring_buffer_normalize_time_stamp(struct ring_buffer *buffer,
|
| 485 | int cpu, u64 *ts)
|
| 486 | {
|
| 487 |
|
| 488 | *ts >>= DEBUG_SHIFT;
|
| 489 | }
|
| 490 | EXPORT_SYMBOL_GPL(ring_buffer_normalize_time_stamp);
|
| 491 |
|
| 492 | |
| 493 | * check_pages - integrity check of buffer pages
|
| 494 | * @cpu_buffer: CPU buffer with pages to test
|
| 495 | *
|
| 496 | * As a safety measure we check to make sure the data pages have not
|
| 497 | * been corrupted.
|
| 498 |
|
| 499 | static int rb_check_pages(struct ring_buffer_per_cpu *cpu_buffer)
|
| 500 | {
|
| 501 | struct list_head *head = &cpu_buffer->pages;
|
| 502 | struct buffer_page *bpage, *tmp;
|
| 503 |
|
| 504 | if (RB_WARN_ON(cpu_buffer, head->next->prev != head))
|
| 505 | return -1;
|
| 506 | if (RB_WARN_ON(cpu_buffer, head->prev->next != head))
|
| 507 | return -1;
|
| 508 |
|
| 509 | list_for_each_entry_safe(bpage, tmp, head, list) {
|
| 510 | if (RB_WARN_ON(cpu_buffer,
|
| 511 | bpage->list.next->prev != &bpage->list))
|
| 512 | return -1;
|
| 513 | if (RB_WARN_ON(cpu_buffer,
|
| 514 | bpage->list.prev->next != &bpage->list))
|
| 515 | return -1;
|
| 516 | }
|
| 517 |
|
| 518 | return 0;
|
| 519 | }
|
| 520 |
|
| 521 | static int rb_allocate_pages(struct ring_buffer_per_cpu *cpu_buffer,
|
| 522 | unsigned nr_pages)
|
| 523 | {
|
| 524 | struct list_head *head = &cpu_buffer->pages;
|
| 525 | struct buffer_page *bpage, *tmp;
|
| 526 | unsigned long addr;
|
| 527 | LIST_HEAD(pages);
|
| 528 | unsigned i;
|
| 529 |
|
| 530 | for (i = 0; i < nr_pages; i++) {
|
| 531 | bpage = kzalloc_node(ALIGN(sizeof(*bpage), cache_line_size()),
|
| 532 | GFP_KERNEL, cpu_to_node(cpu_buffer->cpu));
|
| 533 | if (!bpage)
|
| 534 | goto free_pages;
|
| 535 | list_add(&bpage->list, &pages);
|
| 536 |
|
| 537 | addr = __get_free_page(GFP_KERNEL);
|
| 538 | if (!addr)
|
| 539 | goto free_pages;
|
| 540 | bpage->page = (void *)addr;
|
| 541 | rb_init_page(bpage->page);
|
| 542 | }
|
| 543 |
|
| 544 | list_splice(&pages, head);
|
| 545 |
|
| 546 | rb_check_pages(cpu_buffer);
|
| 547 |
|
| 548 | return 0;
|
| 549 |
|
| 550 | free_pages:
|
| 551 | list_for_each_entry_safe(bpage, tmp, &pages, list) {
|
| 552 | list_del_init(&bpage->list);
|
| 553 | free_buffer_page(bpage);
|
| 554 | }
|
| 555 | return -ENOMEM;
|
| 556 | }
|
| 557 |
|
| 558 | static struct ring_buffer_per_cpu *
|
| 559 | rb_allocate_cpu_buffer(struct ring_buffer *buffer, int cpu)
|
| 560 | {
|
| 561 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 562 | struct buffer_page *bpage;
|
| 563 | unsigned long addr;
|
| 564 | int ret;
|
| 565 |
|
| 566 | cpu_buffer = kzalloc_node(ALIGN(sizeof(*cpu_buffer), cache_line_size()),
|
| 567 | GFP_KERNEL, cpu_to_node(cpu));
|
| 568 | if (!cpu_buffer)
|
| 569 | return NULL;
|
| 570 |
|
| 571 | cpu_buffer->cpu = cpu;
|
| 572 | cpu_buffer->buffer = buffer;
|
| 573 | spin_lock_init(&cpu_buffer->reader_lock);
|
| 574 | lockdep_set_class(&cpu_buffer->reader_lock, buffer->reader_lock_key);
|
| 575 | cpu_buffer->lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
|
| 576 | INIT_LIST_HEAD(&cpu_buffer->pages);
|
| 577 |
|
| 578 | bpage = kzalloc_node(ALIGN(sizeof(*bpage), cache_line_size()),
|
| 579 | GFP_KERNEL, cpu_to_node(cpu));
|
| 580 | if (!bpage)
|
| 581 | goto fail_free_buffer;
|
| 582 |
|
| 583 | cpu_buffer->reader_page = bpage;
|
| 584 | addr = __get_free_page(GFP_KERNEL);
|
| 585 | if (!addr)
|
| 586 | goto fail_free_reader;
|
| 587 | bpage->page = (void *)addr;
|
| 588 | rb_init_page(bpage->page);
|
| 589 |
|
| 590 | INIT_LIST_HEAD(&cpu_buffer->reader_page->list);
|
| 591 |
|
| 592 | ret = rb_allocate_pages(cpu_buffer, buffer->pages);
|
| 593 | if (ret < 0)
|
| 594 | goto fail_free_reader;
|
| 595 |
|
| 596 | cpu_buffer->head_page
|
| 597 | = list_entry(cpu_buffer->pages.next, struct buffer_page, list);
|
| 598 | cpu_buffer->tail_page = cpu_buffer->commit_page = cpu_buffer->head_page;
|
| 599 |
|
| 600 | return cpu_buffer;
|
| 601 |
|
| 602 | fail_free_reader:
|
| 603 | free_buffer_page(cpu_buffer->reader_page);
|
| 604 |
|
| 605 | fail_free_buffer:
|
| 606 | kfree(cpu_buffer);
|
| 607 | return NULL;
|
| 608 | }
|
| 609 |
|
| 610 | static void rb_free_cpu_buffer(struct ring_buffer_per_cpu *cpu_buffer)
|
| 611 | {
|
| 612 | struct list_head *head = &cpu_buffer->pages;
|
| 613 | struct buffer_page *bpage, *tmp;
|
| 614 |
|
| 615 | free_buffer_page(cpu_buffer->reader_page);
|
| 616 |
|
| 617 | list_for_each_entry_safe(bpage, tmp, head, list) {
|
| 618 | list_del_init(&bpage->list);
|
| 619 | free_buffer_page(bpage);
|
| 620 | }
|
| 621 | kfree(cpu_buffer);
|
| 622 | }
|
| 623 |
|
| 624 | #ifdef CONFIG_HOTPLUG_CPU
|
| 625 | static int rb_cpu_notify(struct notifier_block *self,
|
| 626 | unsigned long action, void *hcpu);
|
| 627 | #endif
|
| 628 |
|
| 629 | |
| 630 | * ring_buffer_alloc - allocate a new ring_buffer
|
| 631 | * @size: the size in bytes per cpu that is needed.
|
| 632 | * @flags: attributes to set for the ring buffer.
|
| 633 | *
|
| 634 | * Currently the only flag that is available is the RB_FL_OVERWRITE
|
| 635 | * flag. This flag means that the buffer will overwrite old data
|
| 636 | * when the buffer wraps. If this flag is not set, the buffer will
|
| 637 | * drop data when the tail hits the head.
|
| 638 |
|
| 639 | struct ring_buffer *__ring_buffer_alloc(unsigned long size, unsigned flags,
|
| 640 | struct lock_class_key *key)
|
| 641 | {
|
| 642 | struct ring_buffer *buffer;
|
| 643 | int bsize;
|
| 644 | int cpu;
|
| 645 |
|
| 646 |
|
| 647 | buffer = kzalloc(ALIGN(sizeof(*buffer), cache_line_size()),
|
| 648 | GFP_KERNEL);
|
| 649 | if (!buffer)
|
| 650 | return NULL;
|
| 651 |
|
| 652 | if (!alloc_cpumask_var(&buffer->cpumask, GFP_KERNEL))
|
| 653 | goto fail_free_buffer;
|
| 654 |
|
| 655 | buffer->pages = DIV_ROUND_UP(size, BUF_PAGE_SIZE);
|
| 656 | buffer->flags = flags;
|
| 657 | buffer->clock = trace_clock_local;
|
| 658 | buffer->reader_lock_key = key;
|
| 659 |
|
| 660 |
|
| 661 | if (buffer->pages < 2)
|
| 662 | buffer->pages = 2;
|
| 663 |
|
| 664 | |
| 665 | * In case of non-hotplug cpu, if the ring-buffer is allocated
|
| 666 | * in early initcall, it will not be notified of secondary cpus.
|
| 667 | * In that off case, we need to allocate for all possible cpus.
|
| 668 |
|
| 669 | #ifdef CONFIG_HOTPLUG_CPU
|
| 670 | get_online_cpus();
|
| 671 | cpumask_copy(buffer->cpumask, cpu_online_mask);
|
| 672 | #else
|
| 673 | cpumask_copy(buffer->cpumask, cpu_possible_mask);
|
| 674 | #endif
|
| 675 | buffer->cpus = nr_cpu_ids;
|
| 676 |
|
| 677 | bsize = sizeof(void *) * nr_cpu_ids;
|
| 678 | buffer->buffers = kzalloc(ALIGN(bsize, cache_line_size()),
|
| 679 | GFP_KERNEL);
|
| 680 | if (!buffer->buffers)
|
| 681 | goto fail_free_cpumask;
|
| 682 |
|
| 683 | for_each_buffer_cpu(buffer, cpu) {
|
| 684 | buffer->buffers[cpu] =
|
| 685 | rb_allocate_cpu_buffer(buffer, cpu);
|
| 686 | if (!buffer->buffers[cpu])
|
| 687 | goto fail_free_buffers;
|
| 688 | }
|
| 689 |
|
| 690 | #ifdef CONFIG_HOTPLUG_CPU
|
| 691 | buffer->cpu_notify.notifier_call = rb_cpu_notify;
|
| 692 | buffer->cpu_notify.priority = 0;
|
| 693 | register_cpu_notifier(&buffer->cpu_notify);
|
| 694 | #endif
|
| 695 |
|
| 696 | put_online_cpus();
|
| 697 | mutex_init(&buffer->mutex);
|
| 698 |
|
| 699 | return buffer;
|
| 700 |
|
| 701 | fail_free_buffers:
|
| 702 | for_each_buffer_cpu(buffer, cpu) {
|
| 703 | if (buffer->buffers[cpu])
|
| 704 | rb_free_cpu_buffer(buffer->buffers[cpu]);
|
| 705 | }
|
| 706 | kfree(buffer->buffers);
|
| 707 |
|
| 708 | fail_free_cpumask:
|
| 709 | free_cpumask_var(buffer->cpumask);
|
| 710 | put_online_cpus();
|
| 711 |
|
| 712 | fail_free_buffer:
|
| 713 | kfree(buffer);
|
| 714 | return NULL;
|
| 715 | }
|
| 716 | EXPORT_SYMBOL_GPL(__ring_buffer_alloc);
|
| 717 |
|
| 718 | |
| 719 | * ring_buffer_free - free a ring buffer.
|
| 720 | * @buffer: the buffer to free.
|
| 721 |
|
| 722 | void
|
| 723 | ring_buffer_free(struct ring_buffer *buffer)
|
| 724 | {
|
| 725 | int cpu;
|
| 726 |
|
| 727 | get_online_cpus();
|
| 728 |
|
| 729 | #ifdef CONFIG_HOTPLUG_CPU
|
| 730 | unregister_cpu_notifier(&buffer->cpu_notify);
|
| 731 | #endif
|
| 732 |
|
| 733 | for_each_buffer_cpu(buffer, cpu)
|
| 734 | rb_free_cpu_buffer(buffer->buffers[cpu]);
|
| 735 |
|
| 736 | put_online_cpus();
|
| 737 |
|
| 738 | free_cpumask_var(buffer->cpumask);
|
| 739 |
|
| 740 | kfree(buffer);
|
| 741 | }
|
| 742 | EXPORT_SYMBOL_GPL(ring_buffer_free);
|
| 743 |
|
| 744 | void ring_buffer_set_clock(struct ring_buffer *buffer,
|
| 745 | u64 (*clock)(void))
|
| 746 | {
|
| 747 | buffer->clock = clock;
|
| 748 | }
|
| 749 |
|
| 750 | static void rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer);
|
| 751 |
|
| 752 | static void
|
| 753 | rb_remove_pages(struct ring_buffer_per_cpu *cpu_buffer, unsigned nr_pages)
|
| 754 | {
|
| 755 | struct buffer_page *bpage;
|
| 756 | struct list_head *p;
|
| 757 | unsigned i;
|
| 758 |
|
| 759 | atomic_inc(&cpu_buffer->record_disabled);
|
| 760 | synchronize_sched();
|
| 761 |
|
| 762 | for (i = 0; i < nr_pages; i++) {
|
| 763 | if (RB_WARN_ON(cpu_buffer, list_empty(&cpu_buffer->pages)))
|
| 764 | return;
|
| 765 | p = cpu_buffer->pages.next;
|
| 766 | bpage = list_entry(p, struct buffer_page, list);
|
| 767 | list_del_init(&bpage->list);
|
| 768 | free_buffer_page(bpage);
|
| 769 | }
|
| 770 | if (RB_WARN_ON(cpu_buffer, list_empty(&cpu_buffer->pages)))
|
| 771 | return;
|
| 772 |
|
| 773 | rb_reset_cpu(cpu_buffer);
|
| 774 |
|
| 775 | rb_check_pages(cpu_buffer);
|
| 776 |
|
| 777 | atomic_dec(&cpu_buffer->record_disabled);
|
| 778 |
|
| 779 | }
|
| 780 |
|
| 781 | static void
|
| 782 | rb_insert_pages(struct ring_buffer_per_cpu *cpu_buffer,
|
| 783 | struct list_head *pages, unsigned nr_pages)
|
| 784 | {
|
| 785 | struct buffer_page *bpage;
|
| 786 | struct list_head *p;
|
| 787 | unsigned i;
|
| 788 |
|
| 789 | atomic_inc(&cpu_buffer->record_disabled);
|
| 790 | synchronize_sched();
|
| 791 |
|
| 792 | for (i = 0; i < nr_pages; i++) {
|
| 793 | if (RB_WARN_ON(cpu_buffer, list_empty(pages)))
|
| 794 | return;
|
| 795 | p = pages->next;
|
| 796 | bpage = list_entry(p, struct buffer_page, list);
|
| 797 | list_del_init(&bpage->list);
|
| 798 | list_add_tail(&bpage->list, &cpu_buffer->pages);
|
| 799 | }
|
| 800 | rb_reset_cpu(cpu_buffer);
|
| 801 |
|
| 802 | rb_check_pages(cpu_buffer);
|
| 803 |
|
| 804 | atomic_dec(&cpu_buffer->record_disabled);
|
| 805 | }
|
| 806 |
|
| 807 | |
| 808 | * ring_buffer_resize - resize the ring buffer
|
| 809 | * @buffer: the buffer to resize.
|
| 810 | * @size: the new size.
|
| 811 | *
|
| 812 | * The tracer is responsible for making sure that the buffer is
|
| 813 | * not being used while changing the size.
|
| 814 | * Note: We may be able to change the above requirement by using
|
| 815 | * RCU synchronizations.
|
| 816 | *
|
| 817 | * Minimum size is 2 * BUF_PAGE_SIZE.
|
| 818 | *
|
| 819 | * Returns -1 on failure.
|
| 820 |
|
| 821 | int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size)
|
| 822 | {
|
| 823 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 824 | unsigned nr_pages, rm_pages, new_pages;
|
| 825 | struct buffer_page *bpage, *tmp;
|
| 826 | unsigned long buffer_size;
|
| 827 | unsigned long addr;
|
| 828 | LIST_HEAD(pages);
|
| 829 | int i, cpu;
|
| 830 |
|
| 831 | |
| 832 | * Always succeed at resizing a non-existent buffer:
|
| 833 |
|
| 834 | if (!buffer)
|
| 835 | return size;
|
| 836 |
|
| 837 | size = DIV_ROUND_UP(size, BUF_PAGE_SIZE);
|
| 838 | size *= BUF_PAGE_SIZE;
|
| 839 | buffer_size = buffer->pages * BUF_PAGE_SIZE;
|
| 840 |
|
| 841 |
|
| 842 | if (size < BUF_PAGE_SIZE * 2)
|
| 843 | size = BUF_PAGE_SIZE * 2;
|
| 844 |
|
| 845 | if (size == buffer_size)
|
| 846 | return size;
|
| 847 |
|
| 848 | mutex_lock(&buffer->mutex);
|
| 849 | get_online_cpus();
|
| 850 |
|
| 851 | nr_pages = DIV_ROUND_UP(size, BUF_PAGE_SIZE);
|
| 852 |
|
| 853 | if (size < buffer_size) {
|
| 854 |
|
| 855 |
|
| 856 | if (RB_WARN_ON(buffer, nr_pages >= buffer->pages))
|
| 857 | goto out_fail;
|
| 858 |
|
| 859 | rm_pages = buffer->pages - nr_pages;
|
| 860 |
|
| 861 | for_each_buffer_cpu(buffer, cpu) {
|
| 862 | cpu_buffer = buffer->buffers[cpu];
|
| 863 | rb_remove_pages(cpu_buffer, rm_pages);
|
| 864 | }
|
| 865 | goto out;
|
| 866 | }
|
| 867 |
|
| 868 | |
| 869 | * This is a bit more difficult. We only want to add pages
|
| 870 | * when we can allocate enough for all CPUs. We do this
|
| 871 | * by allocating all the pages and storing them on a local
|
| 872 | * link list. If we succeed in our allocation, then we
|
| 873 | * add these pages to the cpu_buffers. Otherwise we just free
|
| 874 | * them all and return -ENOMEM;
|
| 875 |
|
| 876 | if (RB_WARN_ON(buffer, nr_pages <= buffer->pages))
|
| 877 | goto out_fail;
|
| 878 |
|
| 879 | new_pages = nr_pages - buffer->pages;
|
| 880 |
|
| 881 | for_each_buffer_cpu(buffer, cpu) {
|
| 882 | for (i = 0; i < new_pages; i++) {
|
| 883 | bpage = kzalloc_node(ALIGN(sizeof(*bpage),
|
| 884 | cache_line_size()),
|
| 885 | GFP_KERNEL, cpu_to_node(cpu));
|
| 886 | if (!bpage)
|
| 887 | goto free_pages;
|
| 888 | list_add(&bpage->list, &pages);
|
| 889 | addr = __get_free_page(GFP_KERNEL);
|
| 890 | if (!addr)
|
| 891 | goto free_pages;
|
| 892 | bpage->page = (void *)addr;
|
| 893 | rb_init_page(bpage->page);
|
| 894 | }
|
| 895 | }
|
| 896 |
|
| 897 | for_each_buffer_cpu(buffer, cpu) {
|
| 898 | cpu_buffer = buffer->buffers[cpu];
|
| 899 | rb_insert_pages(cpu_buffer, &pages, new_pages);
|
| 900 | }
|
| 901 |
|
| 902 | if (RB_WARN_ON(buffer, !list_empty(&pages)))
|
| 903 | goto out_fail;
|
| 904 |
|
| 905 | out:
|
| 906 | buffer->pages = nr_pages;
|
| 907 | put_online_cpus();
|
| 908 | mutex_unlock(&buffer->mutex);
|
| 909 |
|
| 910 | return size;
|
| 911 |
|
| 912 | free_pages:
|
| 913 | list_for_each_entry_safe(bpage, tmp, &pages, list) {
|
| 914 | list_del_init(&bpage->list);
|
| 915 | free_buffer_page(bpage);
|
| 916 | }
|
| 917 | put_online_cpus();
|
| 918 | mutex_unlock(&buffer->mutex);
|
| 919 | return -ENOMEM;
|
| 920 |
|
| 921 | |
| 922 | * Something went totally wrong, and we are too paranoid
|
| 923 | * to even clean up the mess.
|
| 924 |
|
| 925 | out_fail:
|
| 926 | put_online_cpus();
|
| 927 | mutex_unlock(&buffer->mutex);
|
| 928 | return -1;
|
| 929 | }
|
| 930 | EXPORT_SYMBOL_GPL(ring_buffer_resize);
|
| 931 |
|
| 932 | static inline void *
|
| 933 | __rb_data_page_index(struct buffer_data_page *bpage, unsigned index)
|
| 934 | {
|
| 935 | return bpage->data + index;
|
| 936 | }
|
| 937 |
|
| 938 | static inline void *__rb_page_index(struct buffer_page *bpage, unsigned index)
|
| 939 | {
|
| 940 | return bpage->page->data + index;
|
| 941 | }
|
| 942 |
|
| 943 | static inline struct ring_buffer_event *
|
| 944 | rb_reader_event(struct ring_buffer_per_cpu *cpu_buffer)
|
| 945 | {
|
| 946 | return __rb_page_index(cpu_buffer->reader_page,
|
| 947 | cpu_buffer->reader_page->read);
|
| 948 | }
|
| 949 |
|
| 950 | static inline struct ring_buffer_event *
|
| 951 | rb_head_event(struct ring_buffer_per_cpu *cpu_buffer)
|
| 952 | {
|
| 953 | return __rb_page_index(cpu_buffer->head_page,
|
| 954 | cpu_buffer->head_page->read);
|
| 955 | }
|
| 956 |
|
| 957 | static inline struct ring_buffer_event *
|
| 958 | rb_iter_head_event(struct ring_buffer_iter *iter)
|
| 959 | {
|
| 960 | return __rb_page_index(iter->head_page, iter->head);
|
| 961 | }
|
| 962 |
|
| 963 | static inline unsigned rb_page_write(struct buffer_page *bpage)
|
| 964 | {
|
| 965 | return local_read(&bpage->write);
|
| 966 | }
|
| 967 |
|
| 968 | static inline unsigned rb_page_commit(struct buffer_page *bpage)
|
| 969 | {
|
| 970 | return local_read(&bpage->page->commit);
|
| 971 | }
|
| 972 |
|
| 973 |
|
| 974 | static inline unsigned rb_page_size(struct buffer_page *bpage)
|
| 975 | {
|
| 976 | return rb_page_commit(bpage);
|
| 977 | }
|
| 978 |
|
| 979 | static inline unsigned
|
| 980 | rb_commit_index(struct ring_buffer_per_cpu *cpu_buffer)
|
| 981 | {
|
| 982 | return rb_page_commit(cpu_buffer->commit_page);
|
| 983 | }
|
| 984 |
|
| 985 | static inline unsigned rb_head_size(struct ring_buffer_per_cpu *cpu_buffer)
|
| 986 | {
|
| 987 | return rb_page_commit(cpu_buffer->head_page);
|
| 988 | }
|
| 989 |
|
| 990 | static inline void rb_inc_page(struct ring_buffer_per_cpu *cpu_buffer,
|
| 991 | struct buffer_page **bpage)
|
| 992 | {
|
| 993 | struct list_head *p = (*bpage)->list.next;
|
| 994 |
|
| 995 | if (p == &cpu_buffer->pages)
|
| 996 | p = p->next;
|
| 997 |
|
| 998 | *bpage = list_entry(p, struct buffer_page, list);
|
| 999 | }
|
| 1000 |
|
| 1001 | static inline unsigned
|
| 1002 | rb_event_index(struct ring_buffer_event *event)
|
| 1003 | {
|
| 1004 | unsigned long addr = (unsigned long)event;
|
| 1005 |
|
| 1006 | return (addr & ~PAGE_MASK) - BUF_PAGE_HDR_SIZE;
|
| 1007 | }
|
| 1008 |
|
| 1009 | static inline int
|
| 1010 | rb_event_is_commit(struct ring_buffer_per_cpu *cpu_buffer,
|
| 1011 | struct ring_buffer_event *event)
|
| 1012 | {
|
| 1013 | unsigned long addr = (unsigned long)event;
|
| 1014 | unsigned long index;
|
| 1015 |
|
| 1016 | index = rb_event_index(event);
|
| 1017 | addr &= PAGE_MASK;
|
| 1018 |
|
| 1019 | return cpu_buffer->commit_page->page == (void *)addr &&
|
| 1020 | rb_commit_index(cpu_buffer) == index;
|
| 1021 | }
|
| 1022 |
|
| 1023 | static void
|
| 1024 | rb_set_commit_to_write(struct ring_buffer_per_cpu *cpu_buffer)
|
| 1025 | {
|
| 1026 | |
| 1027 | * We only race with interrupts and NMIs on this CPU.
|
| 1028 | * If we own the commit event, then we can commit
|
| 1029 | * all others that interrupted us, since the interruptions
|
| 1030 | * are in stack format (they finish before they come
|
| 1031 | * back to us). This allows us to do a simple loop to
|
| 1032 | * assign the commit to the tail.
|
| 1033 |
|
| 1034 | again:
|
| 1035 | while (cpu_buffer->commit_page != cpu_buffer->tail_page) {
|
| 1036 | cpu_buffer->commit_page->page->commit =
|
| 1037 | cpu_buffer->commit_page->write;
|
| 1038 | rb_inc_page(cpu_buffer, &cpu_buffer->commit_page);
|
| 1039 | cpu_buffer->write_stamp =
|
| 1040 | cpu_buffer->commit_page->page->time_stamp;
|
| 1041 |
|
| 1042 | barrier();
|
| 1043 | }
|
| 1044 | while (rb_commit_index(cpu_buffer) !=
|
| 1045 | rb_page_write(cpu_buffer->commit_page)) {
|
| 1046 | cpu_buffer->commit_page->page->commit =
|
| 1047 | cpu_buffer->commit_page->write;
|
| 1048 | barrier();
|
| 1049 | }
|
| 1050 |
|
| 1051 |
|
| 1052 | barrier();
|
| 1053 |
|
| 1054 | |
| 1055 | * If an interrupt came in just after the first while loop
|
| 1056 | * and pushed the tail page forward, we will be left with
|
| 1057 | * a dangling commit that will never go forward.
|
| 1058 |
|
| 1059 | if (unlikely(cpu_buffer->commit_page != cpu_buffer->tail_page))
|
| 1060 | goto again;
|
| 1061 | }
|
| 1062 |
|
| 1063 | static void rb_reset_reader_page(struct ring_buffer_per_cpu *cpu_buffer)
|
| 1064 | {
|
| 1065 | cpu_buffer->read_stamp = cpu_buffer->reader_page->page->time_stamp;
|
| 1066 | cpu_buffer->reader_page->read = 0;
|
| 1067 | }
|
| 1068 |
|
| 1069 | static void rb_inc_iter(struct ring_buffer_iter *iter)
|
| 1070 | {
|
| 1071 | struct ring_buffer_per_cpu *cpu_buffer = iter->cpu_buffer;
|
| 1072 |
|
| 1073 | |
| 1074 | * The iterator could be on the reader page (it starts there).
|
| 1075 | * But the head could have moved, since the reader was
|
| 1076 | * found. Check for this case and assign the iterator
|
| 1077 | * to the head page instead of next.
|
| 1078 |
|
| 1079 | if (iter->head_page == cpu_buffer->reader_page)
|
| 1080 | iter->head_page = cpu_buffer->head_page;
|
| 1081 | else
|
| 1082 | rb_inc_page(cpu_buffer, &iter->head_page);
|
| 1083 |
|
| 1084 | iter->read_stamp = iter->head_page->page->time_stamp;
|
| 1085 | iter->head = 0;
|
| 1086 | }
|
| 1087 |
|
| 1088 | |
| 1089 | * ring_buffer_update_event - update event type and data
|
| 1090 | * @event: the even to update
|
| 1091 | * @type: the type of event
|
| 1092 | * @length: the size of the event field in the ring buffer
|
| 1093 | *
|
| 1094 | * Update the type and data fields of the event. The length
|
| 1095 | * is the actual size that is written to the ring buffer,
|
| 1096 | * and with this, we can determine what to place into the
|
| 1097 | * data field.
|
| 1098 |
|
| 1099 | static void
|
| 1100 | rb_update_event(struct ring_buffer_event *event,
|
| 1101 | unsigned type, unsigned length)
|
| 1102 | {
|
| 1103 | event->type_len = type;
|
| 1104 |
|
| 1105 | switch (type) {
|
| 1106 |
|
| 1107 | case RINGBUF_TYPE_PADDING:
|
| 1108 | case RINGBUF_TYPE_TIME_EXTEND:
|
| 1109 | case RINGBUF_TYPE_TIME_STAMP:
|
| 1110 | break;
|
| 1111 |
|
| 1112 | case 0:
|
| 1113 | length -= RB_EVNT_HDR_SIZE;
|
| 1114 | if (length > RB_MAX_SMALL_DATA)
|
| 1115 | event->array[0] = length;
|
| 1116 | else
|
| 1117 | event->type_len = DIV_ROUND_UP(length, RB_ALIGNMENT);
|
| 1118 | break;
|
| 1119 | default:
|
| 1120 | BUG();
|
| 1121 | }
|
| 1122 | }
|
| 1123 |
|
| 1124 | static unsigned rb_calculate_event_length(unsigned length)
|
| 1125 | {
|
| 1126 | struct ring_buffer_event event;
|
| 1127 |
|
| 1128 |
|
| 1129 | if (!length)
|
| 1130 | length = 1;
|
| 1131 |
|
| 1132 | if (length > RB_MAX_SMALL_DATA)
|
| 1133 | length += sizeof(event.array[0]);
|
| 1134 |
|
| 1135 | length += RB_EVNT_HDR_SIZE;
|
| 1136 | length = ALIGN(length, RB_ALIGNMENT);
|
| 1137 |
|
| 1138 | return length;
|
| 1139 | }
|
| 1140 |
|
| 1141 | static inline void
|
| 1142 | rb_reset_tail(struct ring_buffer_per_cpu *cpu_buffer,
|
| 1143 | struct buffer_page *tail_page,
|
| 1144 | unsigned long tail, unsigned long length)
|
| 1145 | {
|
| 1146 | struct ring_buffer_event *event;
|
| 1147 |
|
| 1148 | |
| 1149 | * Only the event that crossed the page boundary
|
| 1150 | * must fill the old tail_page with padding.
|
| 1151 |
|
| 1152 | if (tail >= BUF_PAGE_SIZE) {
|
| 1153 | local_sub(length, &tail_page->write);
|
| 1154 | return;
|
| 1155 | }
|
| 1156 |
|
| 1157 | event = __rb_page_index(tail_page, tail);
|
| 1158 | kmemcheck_annotate_bitfield(event, bitfield);
|
| 1159 |
|
| 1160 | |
| 1161 | * If this event is bigger than the minimum size, then
|
| 1162 | * we need to be careful that we don't subtract the
|
| 1163 | * write counter enough to allow another writer to slip
|
| 1164 | * in on this page.
|
| 1165 | * We put in a discarded commit instead, to make sure
|
| 1166 | * that this space is not used again.
|
| 1167 | *
|
| 1168 | * If we are less than the minimum size, we don't need to
|
| 1169 | * worry about it.
|
| 1170 |
|
| 1171 | if (tail > (BUF_PAGE_SIZE - RB_EVNT_MIN_SIZE)) {
|
| 1172 |
|
| 1173 |
|
| 1174 |
|
| 1175 | rb_event_set_padding(event);
|
| 1176 |
|
| 1177 |
|
| 1178 | local_sub(length, &tail_page->write);
|
| 1179 | return;
|
| 1180 | }
|
| 1181 |
|
| 1182 |
|
| 1183 | event->array[0] = (BUF_PAGE_SIZE - tail) - RB_EVNT_HDR_SIZE;
|
| 1184 | event->type_len = RINGBUF_TYPE_PADDING;
|
| 1185 |
|
| 1186 | event->time_delta = 1;
|
| 1187 |
|
| 1188 | local_inc(&tail_page->entries);
|
| 1189 | local_inc(&cpu_buffer->entries);
|
| 1190 |
|
| 1191 |
|
| 1192 | length = (tail + length) - BUF_PAGE_SIZE;
|
| 1193 | local_sub(length, &tail_page->write);
|
| 1194 | }
|
| 1195 |
|
| 1196 | static struct ring_buffer_event *
|
| 1197 | rb_move_tail(struct ring_buffer_per_cpu *cpu_buffer,
|
| 1198 | unsigned long length, unsigned long tail,
|
| 1199 | struct buffer_page *commit_page,
|
| 1200 | struct buffer_page *tail_page, u64 *ts)
|
| 1201 | {
|
| 1202 | struct buffer_page *next_page, *head_page, *reader_page;
|
| 1203 | struct ring_buffer *buffer = cpu_buffer->buffer;
|
| 1204 | bool lock_taken = false;
|
| 1205 | unsigned long flags;
|
| 1206 |
|
| 1207 | next_page = tail_page;
|
| 1208 |
|
| 1209 | local_irq_save(flags);
|
| 1210 | |
| 1211 | * Since the write to the buffer is still not
|
| 1212 | * fully lockless, we must be careful with NMIs.
|
| 1213 | * The locks in the writers are taken when a write
|
| 1214 | * crosses to a new page. The locks protect against
|
| 1215 | * races with the readers (this will soon be fixed
|
| 1216 | * with a lockless solution).
|
| 1217 | *
|
| 1218 | * Because we can not protect against NMIs, and we
|
| 1219 | * want to keep traces reentrant, we need to manage
|
| 1220 | * what happens when we are in an NMI.
|
| 1221 | *
|
| 1222 | * NMIs can happen after we take the lock.
|
| 1223 | * If we are in an NMI, only take the lock
|
| 1224 | * if it is not already taken. Otherwise
|
| 1225 | * simply fail.
|
| 1226 |
|
| 1227 | if (unlikely(in_nmi())) {
|
| 1228 | if (!__raw_spin_trylock(&cpu_buffer->lock)) {
|
| 1229 | cpu_buffer->nmi_dropped++;
|
| 1230 | goto out_reset;
|
| 1231 | }
|
| 1232 | } else
|
| 1233 | __raw_spin_lock(&cpu_buffer->lock);
|
| 1234 |
|
| 1235 | lock_taken = true;
|
| 1236 |
|
| 1237 | rb_inc_page(cpu_buffer, &next_page);
|
| 1238 |
|
| 1239 | head_page = cpu_buffer->head_page;
|
| 1240 | reader_page = cpu_buffer->reader_page;
|
| 1241 |
|
| 1242 |
|
| 1243 | if (RB_WARN_ON(cpu_buffer, next_page == reader_page))
|
| 1244 | goto out_reset;
|
| 1245 |
|
| 1246 | |
| 1247 | * If for some reason, we had an interrupt storm that made
|
| 1248 | * it all the way around the buffer, bail, and warn
|
| 1249 | * about it.
|
| 1250 |
|
| 1251 | if (unlikely(next_page == commit_page)) {
|
| 1252 | cpu_buffer->commit_overrun++;
|
| 1253 | goto out_reset;
|
| 1254 | }
|
| 1255 |
|
| 1256 | if (next_page == head_page) {
|
| 1257 | if (!(buffer->flags & RB_FL_OVERWRITE))
|
| 1258 | goto out_reset;
|
| 1259 |
|
| 1260 |
|
| 1261 | if (tail_page == cpu_buffer->tail_page) {
|
| 1262 |
|
| 1263 | cpu_buffer->overrun +=
|
| 1264 | local_read(&head_page->entries);
|
| 1265 |
|
| 1266 | rb_inc_page(cpu_buffer, &head_page);
|
| 1267 | cpu_buffer->head_page = head_page;
|
| 1268 | cpu_buffer->head_page->read = 0;
|
| 1269 | }
|
| 1270 | }
|
| 1271 |
|
| 1272 | |
| 1273 | * If the tail page is still the same as what we think
|
| 1274 | * it is, then it is up to us to update the tail
|
| 1275 | * pointer.
|
| 1276 |
|
| 1277 | if (tail_page == cpu_buffer->tail_page) {
|
| 1278 | local_set(&next_page->write, 0);
|
| 1279 | local_set(&next_page->entries, 0);
|
| 1280 | local_set(&next_page->page->commit, 0);
|
| 1281 | cpu_buffer->tail_page = next_page;
|
| 1282 |
|
| 1283 |
|
| 1284 | *ts = rb_time_stamp(buffer, cpu_buffer->cpu);
|
| 1285 | cpu_buffer->tail_page->page->time_stamp = *ts;
|
| 1286 | }
|
| 1287 |
|
| 1288 | rb_reset_tail(cpu_buffer, tail_page, tail, length);
|
| 1289 |
|
| 1290 | __raw_spin_unlock(&cpu_buffer->lock);
|
| 1291 | local_irq_restore(flags);
|
| 1292 |
|
| 1293 |
|
| 1294 | return ERR_PTR(-EAGAIN);
|
| 1295 |
|
| 1296 | out_reset:
|
| 1297 |
|
| 1298 | rb_reset_tail(cpu_buffer, tail_page, tail, length);
|
| 1299 |
|
| 1300 | if (likely(lock_taken))
|
| 1301 | __raw_spin_unlock(&cpu_buffer->lock);
|
| 1302 | local_irq_restore(flags);
|
| 1303 | return NULL;
|
| 1304 | }
|
| 1305 |
|
| 1306 | static struct ring_buffer_event *
|
| 1307 | __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer,
|
| 1308 | unsigned type, unsigned long length, u64 *ts)
|
| 1309 | {
|
| 1310 | struct buffer_page *tail_page, *commit_page;
|
| 1311 | struct ring_buffer_event *event;
|
| 1312 | unsigned long tail, write;
|
| 1313 |
|
| 1314 | commit_page = cpu_buffer->commit_page;
|
| 1315 |
|
| 1316 | barrier();
|
| 1317 | tail_page = cpu_buffer->tail_page;
|
| 1318 | write = local_add_return(length, &tail_page->write);
|
| 1319 | tail = write - length;
|
| 1320 |
|
| 1321 |
|
| 1322 | if (write > BUF_PAGE_SIZE)
|
| 1323 | return rb_move_tail(cpu_buffer, length, tail,
|
| 1324 | commit_page, tail_page, ts);
|
| 1325 |
|
| 1326 |
|
| 1327 |
|
| 1328 | event = __rb_page_index(tail_page, tail);
|
| 1329 | kmemcheck_annotate_bitfield(event, bitfield);
|
| 1330 | rb_update_event(event, type, length);
|
| 1331 |
|
| 1332 |
|
| 1333 | if (likely(!type))
|
| 1334 | local_inc(&tail_page->entries);
|
| 1335 |
|
| 1336 | |
| 1337 | * If this is the first commit on the page, then update
|
| 1338 | * its timestamp.
|
| 1339 |
|
| 1340 | if (!tail)
|
| 1341 | tail_page->page->time_stamp = *ts;
|
| 1342 |
|
| 1343 | return event;
|
| 1344 | }
|
| 1345 |
|
| 1346 | static inline int
|
| 1347 | rb_try_to_discard(struct ring_buffer_per_cpu *cpu_buffer,
|
| 1348 | struct ring_buffer_event *event)
|
| 1349 | {
|
| 1350 | unsigned long new_index, old_index;
|
| 1351 | struct buffer_page *bpage;
|
| 1352 | unsigned long index;
|
| 1353 | unsigned long addr;
|
| 1354 |
|
| 1355 | new_index = rb_event_index(event);
|
| 1356 | old_index = new_index + rb_event_length(event);
|
| 1357 | addr = (unsigned long)event;
|
| 1358 | addr &= PAGE_MASK;
|
| 1359 |
|
| 1360 | bpage = cpu_buffer->tail_page;
|
| 1361 |
|
| 1362 | if (bpage->page == (void *)addr && rb_page_write(bpage) == old_index) {
|
| 1363 | |
| 1364 | * This is on the tail page. It is possible that
|
| 1365 | * a write could come in and move the tail page
|
| 1366 | * and write to the next page. That is fine
|
| 1367 | * because we just shorten what is on this page.
|
| 1368 |
|
| 1369 | index = local_cmpxchg(&bpage->write, old_index, new_index);
|
| 1370 | if (index == old_index)
|
| 1371 | return 1;
|
| 1372 | }
|
| 1373 |
|
| 1374 |
|
| 1375 | return 0;
|
| 1376 | }
|
| 1377 |
|
| 1378 | static int
|
| 1379 | rb_add_time_stamp(struct ring_buffer_per_cpu *cpu_buffer,
|
| 1380 | u64 *ts, u64 *delta)
|
| 1381 | {
|
| 1382 | struct ring_buffer_event *event;
|
| 1383 | static int once;
|
| 1384 | int ret;
|
| 1385 |
|
| 1386 | if (unlikely(*delta > (1ULL << 59) && !once++)) {
|
| 1387 | printk(KERN_WARNING "Delta way too big! %llu"
|
| 1388 | " ts=%llu write stamp = %llu\n",
|
| 1389 | (unsigned long long)*delta,
|
| 1390 | (unsigned long long)*ts,
|
| 1391 | (unsigned long long)cpu_buffer->write_stamp);
|
| 1392 | WARN_ON(1);
|
| 1393 | }
|
| 1394 |
|
| 1395 | |
| 1396 | * The delta is too big, we to add a
|
| 1397 | * new timestamp.
|
| 1398 |
|
| 1399 | event = __rb_reserve_next(cpu_buffer,
|
| 1400 | RINGBUF_TYPE_TIME_EXTEND,
|
| 1401 | RB_LEN_TIME_EXTEND,
|
| 1402 | ts);
|
| 1403 | if (!event)
|
| 1404 | return -EBUSY;
|
| 1405 |
|
| 1406 | if (PTR_ERR(event) == -EAGAIN)
|
| 1407 | return -EAGAIN;
|
| 1408 |
|
| 1409 |
|
| 1410 | if (rb_event_is_commit(cpu_buffer, event)) {
|
| 1411 | |
| 1412 | * If this is the first on the page, then it was
|
| 1413 | * updated with the page itself. Try to discard it
|
| 1414 | * and if we can't just make it zero.
|
| 1415 |
|
| 1416 | if (rb_event_index(event)) {
|
| 1417 | event->time_delta = *delta & TS_MASK;
|
| 1418 | event->array[0] = *delta >> TS_SHIFT;
|
| 1419 | } else {
|
| 1420 |
|
| 1421 | if (!rb_try_to_discard(cpu_buffer, event)) {
|
| 1422 |
|
| 1423 | event->time_delta = 0;
|
| 1424 | event->array[0] = 0;
|
| 1425 | }
|
| 1426 | }
|
| 1427 | cpu_buffer->write_stamp = *ts;
|
| 1428 |
|
| 1429 | ret = 1;
|
| 1430 | } else {
|
| 1431 |
|
| 1432 | if (!rb_try_to_discard(cpu_buffer, event)) {
|
| 1433 |
|
| 1434 | event->time_delta = 0;
|
| 1435 | event->array[0] = 0;
|
| 1436 | }
|
| 1437 | ret = 0;
|
| 1438 | }
|
| 1439 |
|
| 1440 | *delta = 0;
|
| 1441 |
|
| 1442 | return ret;
|
| 1443 | }
|
| 1444 |
|
| 1445 | static void rb_start_commit(struct ring_buffer_per_cpu *cpu_buffer)
|
| 1446 | {
|
| 1447 | local_inc(&cpu_buffer->committing);
|
| 1448 | local_inc(&cpu_buffer->commits);
|
| 1449 | }
|
| 1450 |
|
| 1451 | static void rb_end_commit(struct ring_buffer_per_cpu *cpu_buffer)
|
| 1452 | {
|
| 1453 | unsigned long commits;
|
| 1454 |
|
| 1455 | if (RB_WARN_ON(cpu_buffer,
|
| 1456 | !local_read(&cpu_buffer->committing)))
|
| 1457 | return;
|
| 1458 |
|
| 1459 | again:
|
| 1460 | commits = local_read(&cpu_buffer->commits);
|
| 1461 |
|
| 1462 | barrier();
|
| 1463 | if (local_read(&cpu_buffer->committing) == 1)
|
| 1464 | rb_set_commit_to_write(cpu_buffer);
|
| 1465 |
|
| 1466 | local_dec(&cpu_buffer->committing);
|
| 1467 |
|
| 1468 |
|
| 1469 | barrier();
|
| 1470 |
|
| 1471 | |
| 1472 | * Need to account for interrupts coming in between the
|
| 1473 | * updating of the commit page and the clearing of the
|
| 1474 | * committing counter.
|
| 1475 |
|
| 1476 | if (unlikely(local_read(&cpu_buffer->commits) != commits) &&
|
| 1477 | !local_read(&cpu_buffer->committing)) {
|
| 1478 | local_inc(&cpu_buffer->committing);
|
| 1479 | goto again;
|
| 1480 | }
|
| 1481 | }
|
| 1482 |
|
| 1483 | static struct ring_buffer_event *
|
| 1484 | rb_reserve_next_event(struct ring_buffer_per_cpu *cpu_buffer,
|
| 1485 | unsigned long length)
|
| 1486 | {
|
| 1487 | struct ring_buffer_event *event;
|
| 1488 | u64 ts, delta = 0;
|
| 1489 | int commit = 0;
|
| 1490 | int nr_loops = 0;
|
| 1491 |
|
| 1492 | rb_start_commit(cpu_buffer);
|
| 1493 |
|
| 1494 | length = rb_calculate_event_length(length);
|
| 1495 | again:
|
| 1496 | |
| 1497 | * We allow for interrupts to reenter here and do a trace.
|
| 1498 | * If one does, it will cause this original code to loop
|
| 1499 | * back here. Even with heavy interrupts happening, this
|
| 1500 | * should only happen a few times in a row. If this happens
|
| 1501 | * 1000 times in a row, there must be either an interrupt
|
| 1502 | * storm or we have something buggy.
|
| 1503 | * Bail!
|
| 1504 |
|
| 1505 | if (RB_WARN_ON(cpu_buffer, ++nr_loops > 1000))
|
| 1506 | goto out_fail;
|
| 1507 |
|
| 1508 | ts = rb_time_stamp(cpu_buffer->buffer, cpu_buffer->cpu);
|
| 1509 |
|
| 1510 | |
| 1511 | * Only the first commit can update the timestamp.
|
| 1512 | * Yes there is a race here. If an interrupt comes in
|
| 1513 | * just after the conditional and it traces too, then it
|
| 1514 | * will also check the deltas. More than one timestamp may
|
| 1515 | * also be made. But only the entry that did the actual
|
| 1516 | * commit will be something other than zero.
|
| 1517 |
|
| 1518 | if (likely(cpu_buffer->tail_page == cpu_buffer->commit_page &&
|
| 1519 | rb_page_write(cpu_buffer->tail_page) ==
|
| 1520 | rb_commit_index(cpu_buffer))) {
|
| 1521 | u64 diff;
|
| 1522 |
|
| 1523 | diff = ts - cpu_buffer->write_stamp;
|
| 1524 |
|
| 1525 |
|
| 1526 | barrier();
|
| 1527 |
|
| 1528 |
|
| 1529 | if (unlikely(ts < cpu_buffer->write_stamp))
|
| 1530 | goto get_event;
|
| 1531 |
|
| 1532 | delta = diff;
|
| 1533 | if (unlikely(test_time_stamp(delta))) {
|
| 1534 |
|
| 1535 | commit = rb_add_time_stamp(cpu_buffer, &ts, &delta);
|
| 1536 | if (commit == -EBUSY)
|
| 1537 | goto out_fail;
|
| 1538 |
|
| 1539 | if (commit == -EAGAIN)
|
| 1540 | goto again;
|
| 1541 |
|
| 1542 | RB_WARN_ON(cpu_buffer, commit < 0);
|
| 1543 | }
|
| 1544 | }
|
| 1545 |
|
| 1546 | get_event:
|
| 1547 | event = __rb_reserve_next(cpu_buffer, 0, length, &ts);
|
| 1548 | if (unlikely(PTR_ERR(event) == -EAGAIN))
|
| 1549 | goto again;
|
| 1550 |
|
| 1551 | if (!event)
|
| 1552 | goto out_fail;
|
| 1553 |
|
| 1554 | if (!rb_event_is_commit(cpu_buffer, event))
|
| 1555 | delta = 0;
|
| 1556 |
|
| 1557 | event->time_delta = delta;
|
| 1558 |
|
| 1559 | return event;
|
| 1560 |
|
| 1561 | out_fail:
|
| 1562 | rb_end_commit(cpu_buffer);
|
| 1563 | return NULL;
|
| 1564 | }
|
| 1565 |
|
| 1566 | #define TRACE_RECURSIVE_DEPTH 16
|
| 1567 |
|
| 1568 | static int trace_recursive_lock(void)
|
| 1569 | {
|
| 1570 | current->trace_recursion++;
|
| 1571 |
|
| 1572 | if (likely(current->trace_recursion < TRACE_RECURSIVE_DEPTH))
|
| 1573 | return 0;
|
| 1574 |
|
| 1575 |
|
| 1576 | tracing_off_permanent();
|
| 1577 |
|
| 1578 | printk_once(KERN_WARNING "Tracing recursion: depth[%ld]:"
|
| 1579 | "HC[%lu]:SC[%lu]:NMI[%lu]\n",
|
| 1580 | current->trace_recursion,
|
| 1581 | hardirq_count() >> HARDIRQ_SHIFT,
|
| 1582 | softirq_count() >> SOFTIRQ_SHIFT,
|
| 1583 | in_nmi());
|
| 1584 |
|
| 1585 | WARN_ON_ONCE(1);
|
| 1586 | return -1;
|
| 1587 | }
|
| 1588 |
|
| 1589 | static void trace_recursive_unlock(void)
|
| 1590 | {
|
| 1591 | WARN_ON_ONCE(!current->trace_recursion);
|
| 1592 |
|
| 1593 | current->trace_recursion--;
|
| 1594 | }
|
| 1595 |
|
| 1596 | static DEFINE_PER_CPU(int, rb_need_resched);
|
| 1597 |
|
| 1598 | |
| 1599 | * ring_buffer_lock_reserve - reserve a part of the buffer
|
| 1600 | * @buffer: the ring buffer to reserve from
|
| 1601 | * @length: the length of the data to reserve (excluding event header)
|
| 1602 | *
|
| 1603 | * Returns a reseverd event on the ring buffer to copy directly to.
|
| 1604 | * The user of this interface will need to get the body to write into
|
| 1605 | * and can use the ring_buffer_event_data() interface.
|
| 1606 | *
|
| 1607 | * The length is the length of the data needed, not the event length
|
| 1608 | * which also includes the event header.
|
| 1609 | *
|
| 1610 | * Must be paired with ring_buffer_unlock_commit, unless NULL is returned.
|
| 1611 | * If NULL is returned, then nothing has been allocated or locked.
|
| 1612 |
|
| 1613 | struct ring_buffer_event *
|
| 1614 | ring_buffer_lock_reserve(struct ring_buffer *buffer, unsigned long length)
|
| 1615 | {
|
| 1616 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 1617 | struct ring_buffer_event *event;
|
| 1618 | int cpu, resched;
|
| 1619 |
|
| 1620 | if (ring_buffer_flags != RB_BUFFERS_ON)
|
| 1621 | return NULL;
|
| 1622 |
|
| 1623 | if (atomic_read(&buffer->record_disabled))
|
| 1624 | return NULL;
|
| 1625 |
|
| 1626 |
|
| 1627 | resched = ftrace_preempt_disable();
|
| 1628 |
|
| 1629 | if (trace_recursive_lock())
|
| 1630 | goto out_nocheck;
|
| 1631 |
|
| 1632 | cpu = raw_smp_processor_id();
|
| 1633 |
|
| 1634 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 1635 | goto out;
|
| 1636 |
|
| 1637 | cpu_buffer = buffer->buffers[cpu];
|
| 1638 |
|
| 1639 | if (atomic_read(&cpu_buffer->record_disabled))
|
| 1640 | goto out;
|
| 1641 |
|
| 1642 | if (length > BUF_MAX_DATA_SIZE)
|
| 1643 | goto out;
|
| 1644 |
|
| 1645 | event = rb_reserve_next_event(cpu_buffer, length);
|
| 1646 | if (!event)
|
| 1647 | goto out;
|
| 1648 |
|
| 1649 | |
| 1650 | * Need to store resched state on this cpu.
|
| 1651 | * Only the first needs to.
|
| 1652 |
|
| 1653 |
|
| 1654 | if (preempt_count() == 1)
|
| 1655 | per_cpu(rb_need_resched, cpu) = resched;
|
| 1656 |
|
| 1657 | return event;
|
| 1658 |
|
| 1659 | out:
|
| 1660 | trace_recursive_unlock();
|
| 1661 |
|
| 1662 | out_nocheck:
|
| 1663 | ftrace_preempt_enable(resched);
|
| 1664 | return NULL;
|
| 1665 | }
|
| 1666 | EXPORT_SYMBOL_GPL(ring_buffer_lock_reserve);
|
| 1667 |
|
| 1668 | static void rb_commit(struct ring_buffer_per_cpu *cpu_buffer,
|
| 1669 | struct ring_buffer_event *event)
|
| 1670 | {
|
| 1671 | local_inc(&cpu_buffer->entries);
|
| 1672 |
|
| 1673 | |
| 1674 | * The event first in the commit queue updates the
|
| 1675 | * time stamp.
|
| 1676 |
|
| 1677 | if (rb_event_is_commit(cpu_buffer, event))
|
| 1678 | cpu_buffer->write_stamp += event->time_delta;
|
| 1679 |
|
| 1680 | rb_end_commit(cpu_buffer);
|
| 1681 | }
|
| 1682 |
|
| 1683 | |
| 1684 | * ring_buffer_unlock_commit - commit a reserved
|
| 1685 | * @buffer: The buffer to commit to
|
| 1686 | * @event: The event pointer to commit.
|
| 1687 | *
|
| 1688 | * This commits the data to the ring buffer, and releases any locks held.
|
| 1689 | *
|
| 1690 | * Must be paired with ring_buffer_lock_reserve.
|
| 1691 |
|
| 1692 | int ring_buffer_unlock_commit(struct ring_buffer *buffer,
|
| 1693 | struct ring_buffer_event *event)
|
| 1694 | {
|
| 1695 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 1696 | int cpu = raw_smp_processor_id();
|
| 1697 |
|
| 1698 | cpu_buffer = buffer->buffers[cpu];
|
| 1699 |
|
| 1700 | rb_commit(cpu_buffer, event);
|
| 1701 |
|
| 1702 | trace_recursive_unlock();
|
| 1703 |
|
| 1704 | |
| 1705 | * Only the last preempt count needs to restore preemption.
|
| 1706 |
|
| 1707 | if (preempt_count() == 1)
|
| 1708 | ftrace_preempt_enable(per_cpu(rb_need_resched, cpu));
|
| 1709 | else
|
| 1710 | preempt_enable_no_resched_notrace();
|
| 1711 |
|
| 1712 | return 0;
|
| 1713 | }
|
| 1714 | EXPORT_SYMBOL_GPL(ring_buffer_unlock_commit);
|
| 1715 |
|
| 1716 | static inline void rb_event_discard(struct ring_buffer_event *event)
|
| 1717 | {
|
| 1718 |
|
| 1719 | event->array[0] = rb_event_data_length(event) - RB_EVNT_HDR_SIZE;
|
| 1720 | event->type_len = RINGBUF_TYPE_PADDING;
|
| 1721 |
|
| 1722 | if (!event->time_delta)
|
| 1723 | event->time_delta = 1;
|
| 1724 | }
|
| 1725 |
|
| 1726 | |
| 1727 | * ring_buffer_event_discard - discard any event in the ring buffer
|
| 1728 | * @event: the event to discard
|
| 1729 | *
|
| 1730 | * Sometimes a event that is in the ring buffer needs to be ignored.
|
| 1731 | * This function lets the user discard an event in the ring buffer
|
| 1732 | * and then that event will not be read later.
|
| 1733 | *
|
| 1734 | * Note, it is up to the user to be careful with this, and protect
|
| 1735 | * against races. If the user discards an event that has been consumed
|
| 1736 | * it is possible that it could corrupt the ring buffer.
|
| 1737 |
|
| 1738 | void ring_buffer_event_discard(struct ring_buffer_event *event)
|
| 1739 | {
|
| 1740 | rb_event_discard(event);
|
| 1741 | }
|
| 1742 | EXPORT_SYMBOL_GPL(ring_buffer_event_discard);
|
| 1743 |
|
| 1744 | |
| 1745 | * ring_buffer_commit_discard - discard an event that has not been committed
|
| 1746 | * @buffer: the ring buffer
|
| 1747 | * @event: non committed event to discard
|
| 1748 | *
|
| 1749 | * This is similar to ring_buffer_event_discard but must only be
|
| 1750 | * performed on an event that has not been committed yet. The difference
|
| 1751 | * is that this will also try to free the event from the ring buffer
|
| 1752 | * if another event has not been added behind it.
|
| 1753 | *
|
| 1754 | * If another event has been added behind it, it will set the event
|
| 1755 | * up as discarded, and perform the commit.
|
| 1756 | *
|
| 1757 | * If this function is called, do not call ring_buffer_unlock_commit on
|
| 1758 | * the event.
|
| 1759 |
|
| 1760 | void ring_buffer_discard_commit(struct ring_buffer *buffer,
|
| 1761 | struct ring_buffer_event *event)
|
| 1762 | {
|
| 1763 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 1764 | int cpu;
|
| 1765 |
|
| 1766 |
|
| 1767 | rb_event_discard(event);
|
| 1768 |
|
| 1769 | cpu = smp_processor_id();
|
| 1770 | cpu_buffer = buffer->buffers[cpu];
|
| 1771 |
|
| 1772 | |
| 1773 | * This must only be called if the event has not been
|
| 1774 | * committed yet. Thus we can assume that preemption
|
| 1775 | * is still disabled.
|
| 1776 |
|
| 1777 | RB_WARN_ON(buffer, !local_read(&cpu_buffer->committing));
|
| 1778 |
|
| 1779 | if (!rb_try_to_discard(cpu_buffer, event))
|
| 1780 | goto out;
|
| 1781 |
|
| 1782 | |
| 1783 | * The commit is still visible by the reader, so we
|
| 1784 | * must increment entries.
|
| 1785 |
|
| 1786 | local_inc(&cpu_buffer->entries);
|
| 1787 | out:
|
| 1788 | rb_end_commit(cpu_buffer);
|
| 1789 |
|
| 1790 | trace_recursive_unlock();
|
| 1791 |
|
| 1792 | |
| 1793 | * Only the last preempt count needs to restore preemption.
|
| 1794 |
|
| 1795 | if (preempt_count() == 1)
|
| 1796 | ftrace_preempt_enable(per_cpu(rb_need_resched, cpu));
|
| 1797 | else
|
| 1798 | preempt_enable_no_resched_notrace();
|
| 1799 |
|
| 1800 | }
|
| 1801 | EXPORT_SYMBOL_GPL(ring_buffer_discard_commit);
|
| 1802 |
|
| 1803 | |
| 1804 | * ring_buffer_write - write data to the buffer without reserving
|
| 1805 | * @buffer: The ring buffer to write to.
|
| 1806 | * @length: The length of the data being written (excluding the event header)
|
| 1807 | * @data: The data to write to the buffer.
|
| 1808 | *
|
| 1809 | * This is like ring_buffer_lock_reserve and ring_buffer_unlock_commit as
|
| 1810 | * one function. If you already have the data to write to the buffer, it
|
| 1811 | * may be easier to simply call this function.
|
| 1812 | *
|
| 1813 | * Note, like ring_buffer_lock_reserve, the length is the length of the data
|
| 1814 | * and not the length of the event which would hold the header.
|
| 1815 |
|
| 1816 | int ring_buffer_write(struct ring_buffer *buffer,
|
| 1817 | unsigned long length,
|
| 1818 | void *data)
|
| 1819 | {
|
| 1820 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 1821 | struct ring_buffer_event *event;
|
| 1822 | void *body;
|
| 1823 | int ret = -EBUSY;
|
| 1824 | int cpu, resched;
|
| 1825 |
|
| 1826 | if (ring_buffer_flags != RB_BUFFERS_ON)
|
| 1827 | return -EBUSY;
|
| 1828 |
|
| 1829 | if (atomic_read(&buffer->record_disabled))
|
| 1830 | return -EBUSY;
|
| 1831 |
|
| 1832 | resched = ftrace_preempt_disable();
|
| 1833 |
|
| 1834 | cpu = raw_smp_processor_id();
|
| 1835 |
|
| 1836 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 1837 | goto out;
|
| 1838 |
|
| 1839 | cpu_buffer = buffer->buffers[cpu];
|
| 1840 |
|
| 1841 | if (atomic_read(&cpu_buffer->record_disabled))
|
| 1842 | goto out;
|
| 1843 |
|
| 1844 | if (length > BUF_MAX_DATA_SIZE)
|
| 1845 | goto out;
|
| 1846 |
|
| 1847 | event = rb_reserve_next_event(cpu_buffer, length);
|
| 1848 | if (!event)
|
| 1849 | goto out;
|
| 1850 |
|
| 1851 | body = rb_event_data(event);
|
| 1852 |
|
| 1853 | memcpy(body, data, length);
|
| 1854 |
|
| 1855 | rb_commit(cpu_buffer, event);
|
| 1856 |
|
| 1857 | ret = 0;
|
| 1858 | out:
|
| 1859 | ftrace_preempt_enable(resched);
|
| 1860 |
|
| 1861 | return ret;
|
| 1862 | }
|
| 1863 | EXPORT_SYMBOL_GPL(ring_buffer_write);
|
| 1864 |
|
| 1865 | static int rb_per_cpu_empty(struct ring_buffer_per_cpu *cpu_buffer)
|
| 1866 | {
|
| 1867 | struct buffer_page *reader = cpu_buffer->reader_page;
|
| 1868 | struct buffer_page *head = cpu_buffer->head_page;
|
| 1869 | struct buffer_page *commit = cpu_buffer->commit_page;
|
| 1870 |
|
| 1871 | return reader->read == rb_page_commit(reader) &&
|
| 1872 | (commit == reader ||
|
| 1873 | (commit == head &&
|
| 1874 | head->read == rb_page_commit(commit)));
|
| 1875 | }
|
| 1876 |
|
| 1877 | |
| 1878 | * ring_buffer_record_disable - stop all writes into the buffer
|
| 1879 | * @buffer: The ring buffer to stop writes to.
|
| 1880 | *
|
| 1881 | * This prevents all writes to the buffer. Any attempt to write
|
| 1882 | * to the buffer after this will fail and return NULL.
|
| 1883 | *
|
| 1884 | * The caller should call synchronize_sched() after this.
|
| 1885 |
|
| 1886 | void ring_buffer_record_disable(struct ring_buffer *buffer)
|
| 1887 | {
|
| 1888 | atomic_inc(&buffer->record_disabled);
|
| 1889 | }
|
| 1890 | EXPORT_SYMBOL_GPL(ring_buffer_record_disable);
|
| 1891 |
|
| 1892 | |
| 1893 | * ring_buffer_record_enable - enable writes to the buffer
|
| 1894 | * @buffer: The ring buffer to enable writes
|
| 1895 | *
|
| 1896 | * Note, multiple disables will need the same number of enables
|
| 1897 | * to truely enable the writing (much like preempt_disable).
|
| 1898 |
|
| 1899 | void ring_buffer_record_enable(struct ring_buffer *buffer)
|
| 1900 | {
|
| 1901 | atomic_dec(&buffer->record_disabled);
|
| 1902 | }
|
| 1903 | EXPORT_SYMBOL_GPL(ring_buffer_record_enable);
|
| 1904 |
|
| 1905 | |
| 1906 | * ring_buffer_record_disable_cpu - stop all writes into the cpu_buffer
|
| 1907 | * @buffer: The ring buffer to stop writes to.
|
| 1908 | * @cpu: The CPU buffer to stop
|
| 1909 | *
|
| 1910 | * This prevents all writes to the buffer. Any attempt to write
|
| 1911 | * to the buffer after this will fail and return NULL.
|
| 1912 | *
|
| 1913 | * The caller should call synchronize_sched() after this.
|
| 1914 |
|
| 1915 | void ring_buffer_record_disable_cpu(struct ring_buffer *buffer, int cpu)
|
| 1916 | {
|
| 1917 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 1918 |
|
| 1919 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 1920 | return;
|
| 1921 |
|
| 1922 | cpu_buffer = buffer->buffers[cpu];
|
| 1923 | atomic_inc(&cpu_buffer->record_disabled);
|
| 1924 | }
|
| 1925 | EXPORT_SYMBOL_GPL(ring_buffer_record_disable_cpu);
|
| 1926 |
|
| 1927 | |
| 1928 | * ring_buffer_record_enable_cpu - enable writes to the buffer
|
| 1929 | * @buffer: The ring buffer to enable writes
|
| 1930 | * @cpu: The CPU to enable.
|
| 1931 | *
|
| 1932 | * Note, multiple disables will need the same number of enables
|
| 1933 | * to truely enable the writing (much like preempt_disable).
|
| 1934 |
|
| 1935 | void ring_buffer_record_enable_cpu(struct ring_buffer *buffer, int cpu)
|
| 1936 | {
|
| 1937 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 1938 |
|
| 1939 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 1940 | return;
|
| 1941 |
|
| 1942 | cpu_buffer = buffer->buffers[cpu];
|
| 1943 | atomic_dec(&cpu_buffer->record_disabled);
|
| 1944 | }
|
| 1945 | EXPORT_SYMBOL_GPL(ring_buffer_record_enable_cpu);
|
| 1946 |
|
| 1947 | |
| 1948 | * ring_buffer_entries_cpu - get the number of entries in a cpu buffer
|
| 1949 | * @buffer: The ring buffer
|
| 1950 | * @cpu: The per CPU buffer to get the entries from.
|
| 1951 |
|
| 1952 | unsigned long ring_buffer_entries_cpu(struct ring_buffer *buffer, int cpu)
|
| 1953 | {
|
| 1954 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 1955 | unsigned long ret;
|
| 1956 |
|
| 1957 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 1958 | return 0;
|
| 1959 |
|
| 1960 | cpu_buffer = buffer->buffers[cpu];
|
| 1961 | ret = (local_read(&cpu_buffer->entries) - cpu_buffer->overrun)
|
| 1962 | - cpu_buffer->read;
|
| 1963 |
|
| 1964 | return ret;
|
| 1965 | }
|
| 1966 | EXPORT_SYMBOL_GPL(ring_buffer_entries_cpu);
|
| 1967 |
|
| 1968 | |
| 1969 | * ring_buffer_overrun_cpu - get the number of overruns in a cpu_buffer
|
| 1970 | * @buffer: The ring buffer
|
| 1971 | * @cpu: The per CPU buffer to get the number of overruns from
|
| 1972 |
|
| 1973 | unsigned long ring_buffer_overrun_cpu(struct ring_buffer *buffer, int cpu)
|
| 1974 | {
|
| 1975 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 1976 | unsigned long ret;
|
| 1977 |
|
| 1978 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 1979 | return 0;
|
| 1980 |
|
| 1981 | cpu_buffer = buffer->buffers[cpu];
|
| 1982 | ret = cpu_buffer->overrun;
|
| 1983 |
|
| 1984 | return ret;
|
| 1985 | }
|
| 1986 | EXPORT_SYMBOL_GPL(ring_buffer_overrun_cpu);
|
| 1987 |
|
| 1988 | |
| 1989 | * ring_buffer_nmi_dropped_cpu - get the number of nmis that were dropped
|
| 1990 | * @buffer: The ring buffer
|
| 1991 | * @cpu: The per CPU buffer to get the number of overruns from
|
| 1992 |
|
| 1993 | unsigned long ring_buffer_nmi_dropped_cpu(struct ring_buffer *buffer, int cpu)
|
| 1994 | {
|
| 1995 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 1996 | unsigned long ret;
|
| 1997 |
|
| 1998 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 1999 | return 0;
|
| 2000 |
|
| 2001 | cpu_buffer = buffer->buffers[cpu];
|
| 2002 | ret = cpu_buffer->nmi_dropped;
|
| 2003 |
|
| 2004 | return ret;
|
| 2005 | }
|
| 2006 | EXPORT_SYMBOL_GPL(ring_buffer_nmi_dropped_cpu);
|
| 2007 |
|
| 2008 | |
| 2009 | * ring_buffer_commit_overrun_cpu - get the number of overruns caused by commits
|
| 2010 | * @buffer: The ring buffer
|
| 2011 | * @cpu: The per CPU buffer to get the number of overruns from
|
| 2012 |
|
| 2013 | unsigned long
|
| 2014 | ring_buffer_commit_overrun_cpu(struct ring_buffer *buffer, int cpu)
|
| 2015 | {
|
| 2016 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2017 | unsigned long ret;
|
| 2018 |
|
| 2019 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 2020 | return 0;
|
| 2021 |
|
| 2022 | cpu_buffer = buffer->buffers[cpu];
|
| 2023 | ret = cpu_buffer->commit_overrun;
|
| 2024 |
|
| 2025 | return ret;
|
| 2026 | }
|
| 2027 | EXPORT_SYMBOL_GPL(ring_buffer_commit_overrun_cpu);
|
| 2028 |
|
| 2029 | |
| 2030 | * ring_buffer_entries - get the number of entries in a buffer
|
| 2031 | * @buffer: The ring buffer
|
| 2032 | *
|
| 2033 | * Returns the total number of entries in the ring buffer
|
| 2034 | * (all CPU entries)
|
| 2035 |
|
| 2036 | unsigned long ring_buffer_entries(struct ring_buffer *buffer)
|
| 2037 | {
|
| 2038 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2039 | unsigned long entries = 0;
|
| 2040 | int cpu;
|
| 2041 |
|
| 2042 |
|
| 2043 | for_each_buffer_cpu(buffer, cpu) {
|
| 2044 | cpu_buffer = buffer->buffers[cpu];
|
| 2045 | entries += (local_read(&cpu_buffer->entries) -
|
| 2046 | cpu_buffer->overrun) - cpu_buffer->read;
|
| 2047 | }
|
| 2048 |
|
| 2049 | return entries;
|
| 2050 | }
|
| 2051 | EXPORT_SYMBOL_GPL(ring_buffer_entries);
|
| 2052 |
|
| 2053 | |
| 2054 | * ring_buffer_overrun_cpu - get the number of overruns in buffer
|
| 2055 | * @buffer: The ring buffer
|
| 2056 | *
|
| 2057 | * Returns the total number of overruns in the ring buffer
|
| 2058 | * (all CPU entries)
|
| 2059 |
|
| 2060 | unsigned long ring_buffer_overruns(struct ring_buffer *buffer)
|
| 2061 | {
|
| 2062 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2063 | unsigned long overruns = 0;
|
| 2064 | int cpu;
|
| 2065 |
|
| 2066 |
|
| 2067 | for_each_buffer_cpu(buffer, cpu) {
|
| 2068 | cpu_buffer = buffer->buffers[cpu];
|
| 2069 | overruns += cpu_buffer->overrun;
|
| 2070 | }
|
| 2071 |
|
| 2072 | return overruns;
|
| 2073 | }
|
| 2074 | EXPORT_SYMBOL_GPL(ring_buffer_overruns);
|
| 2075 |
|
| 2076 | static void rb_iter_reset(struct ring_buffer_iter *iter)
|
| 2077 | {
|
| 2078 | struct ring_buffer_per_cpu *cpu_buffer = iter->cpu_buffer;
|
| 2079 |
|
| 2080 |
|
| 2081 | if (list_empty(&cpu_buffer->reader_page->list)) {
|
| 2082 | iter->head_page = cpu_buffer->head_page;
|
| 2083 | iter->head = cpu_buffer->head_page->read;
|
| 2084 | } else {
|
| 2085 | iter->head_page = cpu_buffer->reader_page;
|
| 2086 | iter->head = cpu_buffer->reader_page->read;
|
| 2087 | }
|
| 2088 | if (iter->head)
|
| 2089 | iter->read_stamp = cpu_buffer->read_stamp;
|
| 2090 | else
|
| 2091 | iter->read_stamp = iter->head_page->page->time_stamp;
|
| 2092 | }
|
| 2093 |
|
| 2094 | |
| 2095 | * ring_buffer_iter_reset - reset an iterator
|
| 2096 | * @iter: The iterator to reset
|
| 2097 | *
|
| 2098 | * Resets the iterator, so that it will start from the beginning
|
| 2099 | * again.
|
| 2100 |
|
| 2101 | void ring_buffer_iter_reset(struct ring_buffer_iter *iter)
|
| 2102 | {
|
| 2103 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2104 | unsigned long flags;
|
| 2105 |
|
| 2106 | if (!iter)
|
| 2107 | return;
|
| 2108 |
|
| 2109 | cpu_buffer = iter->cpu_buffer;
|
| 2110 |
|
| 2111 | spin_lock_irqsave(&cpu_buffer->reader_lock, flags);
|
| 2112 | rb_iter_reset(iter);
|
| 2113 | spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags);
|
| 2114 | }
|
| 2115 | EXPORT_SYMBOL_GPL(ring_buffer_iter_reset);
|
| 2116 |
|
| 2117 | |
| 2118 | * ring_buffer_iter_empty - check if an iterator has no more to read
|
| 2119 | * @iter: The iterator to check
|
| 2120 |
|
| 2121 | int ring_buffer_iter_empty(struct ring_buffer_iter *iter)
|
| 2122 | {
|
| 2123 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2124 |
|
| 2125 | cpu_buffer = iter->cpu_buffer;
|
| 2126 |
|
| 2127 | return iter->head_page == cpu_buffer->commit_page &&
|
| 2128 | iter->head == rb_commit_index(cpu_buffer);
|
| 2129 | }
|
| 2130 | EXPORT_SYMBOL_GPL(ring_buffer_iter_empty);
|
| 2131 |
|
| 2132 | static void
|
| 2133 | rb_update_read_stamp(struct ring_buffer_per_cpu *cpu_buffer,
|
| 2134 | struct ring_buffer_event *event)
|
| 2135 | {
|
| 2136 | u64 delta;
|
| 2137 |
|
| 2138 | switch (event->type_len) {
|
| 2139 | case RINGBUF_TYPE_PADDING:
|
| 2140 | return;
|
| 2141 |
|
| 2142 | case RINGBUF_TYPE_TIME_EXTEND:
|
| 2143 | delta = event->array[0];
|
| 2144 | delta <<= TS_SHIFT;
|
| 2145 | delta += event->time_delta;
|
| 2146 | cpu_buffer->read_stamp += delta;
|
| 2147 | return;
|
| 2148 |
|
| 2149 | case RINGBUF_TYPE_TIME_STAMP:
|
| 2150 |
|
| 2151 | return;
|
| 2152 |
|
| 2153 | case RINGBUF_TYPE_DATA:
|
| 2154 | cpu_buffer->read_stamp += event->time_delta;
|
| 2155 | return;
|
| 2156 |
|
| 2157 | default:
|
| 2158 | BUG();
|
| 2159 | }
|
| 2160 | return;
|
| 2161 | }
|
| 2162 |
|
| 2163 | static void
|
| 2164 | rb_update_iter_read_stamp(struct ring_buffer_iter *iter,
|
| 2165 | struct ring_buffer_event *event)
|
| 2166 | {
|
| 2167 | u64 delta;
|
| 2168 |
|
| 2169 | switch (event->type_len) {
|
| 2170 | case RINGBUF_TYPE_PADDING:
|
| 2171 | return;
|
| 2172 |
|
| 2173 | case RINGBUF_TYPE_TIME_EXTEND:
|
| 2174 | delta = event->array[0];
|
| 2175 | delta <<= TS_SHIFT;
|
| 2176 | delta += event->time_delta;
|
| 2177 | iter->read_stamp += delta;
|
| 2178 | return;
|
| 2179 |
|
| 2180 | case RINGBUF_TYPE_TIME_STAMP:
|
| 2181 |
|
| 2182 | return;
|
| 2183 |
|
| 2184 | case RINGBUF_TYPE_DATA:
|
| 2185 | iter->read_stamp += event->time_delta;
|
| 2186 | return;
|
| 2187 |
|
| 2188 | default:
|
| 2189 | BUG();
|
| 2190 | }
|
| 2191 | return;
|
| 2192 | }
|
| 2193 |
|
| 2194 | static struct buffer_page *
|
| 2195 | rb_get_reader_page(struct ring_buffer_per_cpu *cpu_buffer)
|
| 2196 | {
|
| 2197 | struct buffer_page *reader = NULL;
|
| 2198 | unsigned long flags;
|
| 2199 | int nr_loops = 0;
|
| 2200 |
|
| 2201 | local_irq_save(flags);
|
| 2202 | __raw_spin_lock(&cpu_buffer->lock);
|
| 2203 |
|
| 2204 | again:
|
| 2205 | |
| 2206 | * This should normally only loop twice. But because the
|
| 2207 | * start of the reader inserts an empty page, it causes
|
| 2208 | * a case where we will loop three times. There should be no
|
| 2209 | * reason to loop four times (that I know of).
|
| 2210 |
|
| 2211 | if (RB_WARN_ON(cpu_buffer, ++nr_loops > 3)) {
|
| 2212 | reader = NULL;
|
| 2213 | goto out;
|
| 2214 | }
|
| 2215 |
|
| 2216 | reader = cpu_buffer->reader_page;
|
| 2217 |
|
| 2218 |
|
| 2219 | if (cpu_buffer->reader_page->read < rb_page_size(reader))
|
| 2220 | goto out;
|
| 2221 |
|
| 2222 |
|
| 2223 | if (RB_WARN_ON(cpu_buffer,
|
| 2224 | cpu_buffer->reader_page->read > rb_page_size(reader)))
|
| 2225 | goto out;
|
| 2226 |
|
| 2227 |
|
| 2228 | reader = NULL;
|
| 2229 | if (cpu_buffer->commit_page == cpu_buffer->reader_page)
|
| 2230 | goto out;
|
| 2231 |
|
| 2232 | |
| 2233 | * Splice the empty reader page into the list around the head.
|
| 2234 | * Reset the reader page to size zero.
|
| 2235 |
|
| 2236 |
|
| 2237 | reader = cpu_buffer->head_page;
|
| 2238 | cpu_buffer->reader_page->list.next = reader->list.next;
|
| 2239 | cpu_buffer->reader_page->list.prev = reader->list.prev;
|
| 2240 |
|
| 2241 | local_set(&cpu_buffer->reader_page->write, 0);
|
| 2242 | local_set(&cpu_buffer->reader_page->entries, 0);
|
| 2243 | local_set(&cpu_buffer->reader_page->page->commit, 0);
|
| 2244 |
|
| 2245 |
|
| 2246 | reader->list.prev->next = &cpu_buffer->reader_page->list;
|
| 2247 | reader->list.next->prev = &cpu_buffer->reader_page->list;
|
| 2248 |
|
| 2249 | |
| 2250 | * If the tail is on the reader, then we must set the head
|
| 2251 | * to the inserted page, otherwise we set it one before.
|
| 2252 |
|
| 2253 | cpu_buffer->head_page = cpu_buffer->reader_page;
|
| 2254 |
|
| 2255 | if (cpu_buffer->commit_page != reader)
|
| 2256 | rb_inc_page(cpu_buffer, &cpu_buffer->head_page);
|
| 2257 |
|
| 2258 |
|
| 2259 | cpu_buffer->reader_page = reader;
|
| 2260 | rb_reset_reader_page(cpu_buffer);
|
| 2261 |
|
| 2262 | goto again;
|
| 2263 |
|
| 2264 | out:
|
| 2265 | __raw_spin_unlock(&cpu_buffer->lock);
|
| 2266 | local_irq_restore(flags);
|
| 2267 |
|
| 2268 | return reader;
|
| 2269 | }
|
| 2270 |
|
| 2271 | static void rb_advance_reader(struct ring_buffer_per_cpu *cpu_buffer)
|
| 2272 | {
|
| 2273 | struct ring_buffer_event *event;
|
| 2274 | struct buffer_page *reader;
|
| 2275 | unsigned length;
|
| 2276 |
|
| 2277 | reader = rb_get_reader_page(cpu_buffer);
|
| 2278 |
|
| 2279 |
|
| 2280 | if (RB_WARN_ON(cpu_buffer, !reader))
|
| 2281 | return;
|
| 2282 |
|
| 2283 | event = rb_reader_event(cpu_buffer);
|
| 2284 |
|
| 2285 | if (event->type_len <= RINGBUF_TYPE_DATA_TYPE_LEN_MAX
|
| 2286 | || rb_discarded_event(event))
|
| 2287 | cpu_buffer->read++;
|
| 2288 |
|
| 2289 | rb_update_read_stamp(cpu_buffer, event);
|
| 2290 |
|
| 2291 | length = rb_event_length(event);
|
| 2292 | cpu_buffer->reader_page->read += length;
|
| 2293 | }
|
| 2294 |
|
| 2295 | static void rb_advance_iter(struct ring_buffer_iter *iter)
|
| 2296 | {
|
| 2297 | struct ring_buffer *buffer;
|
| 2298 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2299 | struct ring_buffer_event *event;
|
| 2300 | unsigned length;
|
| 2301 |
|
| 2302 | cpu_buffer = iter->cpu_buffer;
|
| 2303 | buffer = cpu_buffer->buffer;
|
| 2304 |
|
| 2305 | |
| 2306 | * Check if we are at the end of the buffer.
|
| 2307 |
|
| 2308 | if (iter->head >= rb_page_size(iter->head_page)) {
|
| 2309 |
|
| 2310 | if (iter->head_page == cpu_buffer->commit_page)
|
| 2311 | return;
|
| 2312 | rb_inc_iter(iter);
|
| 2313 | return;
|
| 2314 | }
|
| 2315 |
|
| 2316 | event = rb_iter_head_event(iter);
|
| 2317 |
|
| 2318 | length = rb_event_length(event);
|
| 2319 |
|
| 2320 | |
| 2321 | * This should not be called to advance the header if we are
|
| 2322 | * at the tail of the buffer.
|
| 2323 |
|
| 2324 | if (RB_WARN_ON(cpu_buffer,
|
| 2325 | (iter->head_page == cpu_buffer->commit_page) &&
|
| 2326 | (iter->head + length > rb_commit_index(cpu_buffer))))
|
| 2327 | return;
|
| 2328 |
|
| 2329 | rb_update_iter_read_stamp(iter, event);
|
| 2330 |
|
| 2331 | iter->head += length;
|
| 2332 |
|
| 2333 |
|
| 2334 | if ((iter->head >= rb_page_size(iter->head_page)) &&
|
| 2335 | (iter->head_page != cpu_buffer->commit_page))
|
| 2336 | rb_advance_iter(iter);
|
| 2337 | }
|
| 2338 |
|
| 2339 | static struct ring_buffer_event *
|
| 2340 | rb_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts)
|
| 2341 | {
|
| 2342 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2343 | struct ring_buffer_event *event;
|
| 2344 | struct buffer_page *reader;
|
| 2345 | int nr_loops = 0;
|
| 2346 |
|
| 2347 | cpu_buffer = buffer->buffers[cpu];
|
| 2348 |
|
| 2349 | again:
|
| 2350 | |
| 2351 | * We repeat when a timestamp is encountered. It is possible
|
| 2352 | * to get multiple timestamps from an interrupt entering just
|
| 2353 | * as one timestamp is about to be written, or from discarded
|
| 2354 | * commits. The most that we can have is the number on a single page.
|
| 2355 |
|
| 2356 | if (RB_WARN_ON(cpu_buffer, ++nr_loops > RB_TIMESTAMPS_PER_PAGE))
|
| 2357 | return NULL;
|
| 2358 |
|
| 2359 | reader = rb_get_reader_page(cpu_buffer);
|
| 2360 | if (!reader)
|
| 2361 | return NULL;
|
| 2362 |
|
| 2363 | event = rb_reader_event(cpu_buffer);
|
| 2364 |
|
| 2365 | switch (event->type_len) {
|
| 2366 | case RINGBUF_TYPE_PADDING:
|
| 2367 | if (rb_null_event(event))
|
| 2368 | RB_WARN_ON(cpu_buffer, 1);
|
| 2369 | |
| 2370 | * Because the writer could be discarding every
|
| 2371 | * event it creates (which would probably be bad)
|
| 2372 | * if we were to go back to "again" then we may never
|
| 2373 | * catch up, and will trigger the warn on, or lock
|
| 2374 | * the box. Return the padding, and we will release
|
| 2375 | * the current locks, and try again.
|
| 2376 |
|
| 2377 | rb_advance_reader(cpu_buffer);
|
| 2378 | return event;
|
| 2379 |
|
| 2380 | case RINGBUF_TYPE_TIME_EXTEND:
|
| 2381 |
|
| 2382 | rb_advance_reader(cpu_buffer);
|
| 2383 | goto again;
|
| 2384 |
|
| 2385 | case RINGBUF_TYPE_TIME_STAMP:
|
| 2386 |
|
| 2387 | rb_advance_reader(cpu_buffer);
|
| 2388 | goto again;
|
| 2389 |
|
| 2390 | case RINGBUF_TYPE_DATA:
|
| 2391 | if (ts) {
|
| 2392 | *ts = cpu_buffer->read_stamp + event->time_delta;
|
| 2393 | ring_buffer_normalize_time_stamp(buffer,
|
| 2394 | cpu_buffer->cpu, ts);
|
| 2395 | }
|
| 2396 | return event;
|
| 2397 |
|
| 2398 | default:
|
| 2399 | BUG();
|
| 2400 | }
|
| 2401 |
|
| 2402 | return NULL;
|
| 2403 | }
|
| 2404 | EXPORT_SYMBOL_GPL(ring_buffer_peek);
|
| 2405 |
|
| 2406 | static struct ring_buffer_event *
|
| 2407 | rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts)
|
| 2408 | {
|
| 2409 | struct ring_buffer *buffer;
|
| 2410 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2411 | struct ring_buffer_event *event;
|
| 2412 | int nr_loops = 0;
|
| 2413 |
|
| 2414 | if (ring_buffer_iter_empty(iter))
|
| 2415 | return NULL;
|
| 2416 |
|
| 2417 | cpu_buffer = iter->cpu_buffer;
|
| 2418 | buffer = cpu_buffer->buffer;
|
| 2419 |
|
| 2420 | again:
|
| 2421 | |
| 2422 | * We repeat when a timestamp is encountered.
|
| 2423 | * We can get multiple timestamps by nested interrupts or also
|
| 2424 | * if filtering is on (discarding commits). Since discarding
|
| 2425 | * commits can be frequent we can get a lot of timestamps.
|
| 2426 | * But we limit them by not adding timestamps if they begin
|
| 2427 | * at the start of a page.
|
| 2428 |
|
| 2429 | if (RB_WARN_ON(cpu_buffer, ++nr_loops > RB_TIMESTAMPS_PER_PAGE))
|
| 2430 | return NULL;
|
| 2431 |
|
| 2432 | if (rb_per_cpu_empty(cpu_buffer))
|
| 2433 | return NULL;
|
| 2434 |
|
| 2435 | event = rb_iter_head_event(iter);
|
| 2436 |
|
| 2437 | switch (event->type_len) {
|
| 2438 | case RINGBUF_TYPE_PADDING:
|
| 2439 | if (rb_null_event(event)) {
|
| 2440 | rb_inc_iter(iter);
|
| 2441 | goto again;
|
| 2442 | }
|
| 2443 | rb_advance_iter(iter);
|
| 2444 | return event;
|
| 2445 |
|
| 2446 | case RINGBUF_TYPE_TIME_EXTEND:
|
| 2447 |
|
| 2448 | rb_advance_iter(iter);
|
| 2449 | goto again;
|
| 2450 |
|
| 2451 | case RINGBUF_TYPE_TIME_STAMP:
|
| 2452 |
|
| 2453 | rb_advance_iter(iter);
|
| 2454 | goto again;
|
| 2455 |
|
| 2456 | case RINGBUF_TYPE_DATA:
|
| 2457 | if (ts) {
|
| 2458 | *ts = iter->read_stamp + event->time_delta;
|
| 2459 | ring_buffer_normalize_time_stamp(buffer,
|
| 2460 | cpu_buffer->cpu, ts);
|
| 2461 | }
|
| 2462 | return event;
|
| 2463 |
|
| 2464 | default:
|
| 2465 | BUG();
|
| 2466 | }
|
| 2467 |
|
| 2468 | return NULL;
|
| 2469 | }
|
| 2470 | EXPORT_SYMBOL_GPL(ring_buffer_iter_peek);
|
| 2471 |
|
| 2472 | static inline int rb_ok_to_lock(void)
|
| 2473 | {
|
| 2474 | |
| 2475 | * If an NMI die dumps out the content of the ring buffer
|
| 2476 | * do not grab locks. We also permanently disable the ring
|
| 2477 | * buffer too. A one time deal is all you get from reading
|
| 2478 | * the ring buffer from an NMI.
|
| 2479 |
|
| 2480 | if (likely(!in_nmi() && !oops_in_progress))
|
| 2481 | return 1;
|
| 2482 |
|
| 2483 | tracing_off_permanent();
|
| 2484 | return 0;
|
| 2485 | }
|
| 2486 |
|
| 2487 | |
| 2488 | * ring_buffer_peek - peek at the next event to be read
|
| 2489 | * @buffer: The ring buffer to read
|
| 2490 | * @cpu: The cpu to peak at
|
| 2491 | * @ts: The timestamp counter of this event.
|
| 2492 | *
|
| 2493 | * This will return the event that will be read next, but does
|
| 2494 | * not consume the data.
|
| 2495 |
|
| 2496 | struct ring_buffer_event *
|
| 2497 | ring_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts)
|
| 2498 | {
|
| 2499 | struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu];
|
| 2500 | struct ring_buffer_event *event;
|
| 2501 | unsigned long flags;
|
| 2502 | int dolock;
|
| 2503 |
|
| 2504 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 2505 | return NULL;
|
| 2506 |
|
| 2507 | dolock = rb_ok_to_lock();
|
| 2508 | again:
|
| 2509 | local_irq_save(flags);
|
| 2510 | if (dolock)
|
| 2511 | spin_lock(&cpu_buffer->reader_lock);
|
| 2512 | event = rb_buffer_peek(buffer, cpu, ts);
|
| 2513 | if (dolock)
|
| 2514 | spin_unlock(&cpu_buffer->reader_lock);
|
| 2515 | local_irq_restore(flags);
|
| 2516 |
|
| 2517 | if (event && event->type_len == RINGBUF_TYPE_PADDING) {
|
| 2518 | cpu_relax();
|
| 2519 | goto again;
|
| 2520 | }
|
| 2521 |
|
| 2522 | return event;
|
| 2523 | }
|
| 2524 |
|
| 2525 | |
| 2526 | * ring_buffer_iter_peek - peek at the next event to be read
|
| 2527 | * @iter: The ring buffer iterator
|
| 2528 | * @ts: The timestamp counter of this event.
|
| 2529 | *
|
| 2530 | * This will return the event that will be read next, but does
|
| 2531 | * not increment the iterator.
|
| 2532 |
|
| 2533 | struct ring_buffer_event *
|
| 2534 | ring_buffer_iter_peek(struct ring_buffer_iter *iter, u64 *ts)
|
| 2535 | {
|
| 2536 | struct ring_buffer_per_cpu *cpu_buffer = iter->cpu_buffer;
|
| 2537 | struct ring_buffer_event *event;
|
| 2538 | unsigned long flags;
|
| 2539 |
|
| 2540 | again:
|
| 2541 | spin_lock_irqsave(&cpu_buffer->reader_lock, flags);
|
| 2542 | event = rb_iter_peek(iter, ts);
|
| 2543 | spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags);
|
| 2544 |
|
| 2545 | if (event && event->type_len == RINGBUF_TYPE_PADDING) {
|
| 2546 | cpu_relax();
|
| 2547 | goto again;
|
| 2548 | }
|
| 2549 |
|
| 2550 | return event;
|
| 2551 | }
|
| 2552 |
|
| 2553 | |
| 2554 | * ring_buffer_consume - return an event and consume it
|
| 2555 | * @buffer: The ring buffer to get the next event from
|
| 2556 | *
|
| 2557 | * Returns the next event in the ring buffer, and that event is consumed.
|
| 2558 | * Meaning, that sequential reads will keep returning a different event,
|
| 2559 | * and eventually empty the ring buffer if the producer is slower.
|
| 2560 |
|
| 2561 | struct ring_buffer_event *
|
| 2562 | ring_buffer_consume(struct ring_buffer *buffer, int cpu, u64 *ts)
|
| 2563 | {
|
| 2564 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2565 | struct ring_buffer_event *event = NULL;
|
| 2566 | unsigned long flags;
|
| 2567 | int dolock;
|
| 2568 |
|
| 2569 | dolock = rb_ok_to_lock();
|
| 2570 |
|
| 2571 | again:
|
| 2572 |
|
| 2573 | preempt_disable();
|
| 2574 |
|
| 2575 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 2576 | goto out;
|
| 2577 |
|
| 2578 | cpu_buffer = buffer->buffers[cpu];
|
| 2579 | local_irq_save(flags);
|
| 2580 | if (dolock)
|
| 2581 | spin_lock(&cpu_buffer->reader_lock);
|
| 2582 |
|
| 2583 | event = rb_buffer_peek(buffer, cpu, ts);
|
| 2584 | if (!event)
|
| 2585 | goto out_unlock;
|
| 2586 |
|
| 2587 | rb_advance_reader(cpu_buffer);
|
| 2588 |
|
| 2589 | out_unlock:
|
| 2590 | if (dolock)
|
| 2591 | spin_unlock(&cpu_buffer->reader_lock);
|
| 2592 | local_irq_restore(flags);
|
| 2593 |
|
| 2594 | out:
|
| 2595 | preempt_enable();
|
| 2596 |
|
| 2597 | if (event && event->type_len == RINGBUF_TYPE_PADDING) {
|
| 2598 | cpu_relax();
|
| 2599 | goto again;
|
| 2600 | }
|
| 2601 |
|
| 2602 | return event;
|
| 2603 | }
|
| 2604 | EXPORT_SYMBOL_GPL(ring_buffer_consume);
|
| 2605 |
|
| 2606 | |
| 2607 | * ring_buffer_read_start - start a non consuming read of the buffer
|
| 2608 | * @buffer: The ring buffer to read from
|
| 2609 | * @cpu: The cpu buffer to iterate over
|
| 2610 | *
|
| 2611 | * This starts up an iteration through the buffer. It also disables
|
| 2612 | * the recording to the buffer until the reading is finished.
|
| 2613 | * This prevents the reading from being corrupted. This is not
|
| 2614 | * a consuming read, so a producer is not expected.
|
| 2615 | *
|
| 2616 | * Must be paired with ring_buffer_finish.
|
| 2617 |
|
| 2618 | struct ring_buffer_iter *
|
| 2619 | ring_buffer_read_start(struct ring_buffer *buffer, int cpu)
|
| 2620 | {
|
| 2621 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2622 | struct ring_buffer_iter *iter;
|
| 2623 | unsigned long flags;
|
| 2624 |
|
| 2625 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 2626 | return NULL;
|
| 2627 |
|
| 2628 | iter = kmalloc(sizeof(*iter), GFP_KERNEL);
|
| 2629 | if (!iter)
|
| 2630 | return NULL;
|
| 2631 |
|
| 2632 | cpu_buffer = buffer->buffers[cpu];
|
| 2633 |
|
| 2634 | iter->cpu_buffer = cpu_buffer;
|
| 2635 |
|
| 2636 | atomic_inc(&cpu_buffer->record_disabled);
|
| 2637 | synchronize_sched();
|
| 2638 |
|
| 2639 | spin_lock_irqsave(&cpu_buffer->reader_lock, flags);
|
| 2640 | __raw_spin_lock(&cpu_buffer->lock);
|
| 2641 | rb_iter_reset(iter);
|
| 2642 | __raw_spin_unlock(&cpu_buffer->lock);
|
| 2643 | spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags);
|
| 2644 |
|
| 2645 | return iter;
|
| 2646 | }
|
| 2647 | EXPORT_SYMBOL_GPL(ring_buffer_read_start);
|
| 2648 |
|
| 2649 | |
| 2650 | * ring_buffer_finish - finish reading the iterator of the buffer
|
| 2651 | * @iter: The iterator retrieved by ring_buffer_start
|
| 2652 | *
|
| 2653 | * This re-enables the recording to the buffer, and frees the
|
| 2654 | * iterator.
|
| 2655 |
|
| 2656 | void
|
| 2657 | ring_buffer_read_finish(struct ring_buffer_iter *iter)
|
| 2658 | {
|
| 2659 | struct ring_buffer_per_cpu *cpu_buffer = iter->cpu_buffer;
|
| 2660 |
|
| 2661 | atomic_dec(&cpu_buffer->record_disabled);
|
| 2662 | kfree(iter);
|
| 2663 | }
|
| 2664 | EXPORT_SYMBOL_GPL(ring_buffer_read_finish);
|
| 2665 |
|
| 2666 | |
| 2667 | * ring_buffer_read - read the next item in the ring buffer by the iterator
|
| 2668 | * @iter: The ring buffer iterator
|
| 2669 | * @ts: The time stamp of the event read.
|
| 2670 | *
|
| 2671 | * This reads the next event in the ring buffer and increments the iterator.
|
| 2672 |
|
| 2673 | struct ring_buffer_event *
|
| 2674 | ring_buffer_read(struct ring_buffer_iter *iter, u64 *ts)
|
| 2675 | {
|
| 2676 | struct ring_buffer_event *event;
|
| 2677 | struct ring_buffer_per_cpu *cpu_buffer = iter->cpu_buffer;
|
| 2678 | unsigned long flags;
|
| 2679 |
|
| 2680 | again:
|
| 2681 | spin_lock_irqsave(&cpu_buffer->reader_lock, flags);
|
| 2682 | event = rb_iter_peek(iter, ts);
|
| 2683 | if (!event)
|
| 2684 | goto out;
|
| 2685 |
|
| 2686 | rb_advance_iter(iter);
|
| 2687 | out:
|
| 2688 | spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags);
|
| 2689 |
|
| 2690 | if (event && event->type_len == RINGBUF_TYPE_PADDING) {
|
| 2691 | cpu_relax();
|
| 2692 | goto again;
|
| 2693 | }
|
| 2694 |
|
| 2695 | return event;
|
| 2696 | }
|
| 2697 | EXPORT_SYMBOL_GPL(ring_buffer_read);
|
| 2698 |
|
| 2699 | |
| 2700 | * ring_buffer_size - return the size of the ring buffer (in bytes)
|
| 2701 | * @buffer: The ring buffer.
|
| 2702 |
|
| 2703 | unsigned long ring_buffer_size(struct ring_buffer *buffer)
|
| 2704 | {
|
| 2705 | return BUF_PAGE_SIZE * buffer->pages;
|
| 2706 | }
|
| 2707 | EXPORT_SYMBOL_GPL(ring_buffer_size);
|
| 2708 |
|
| 2709 | static void
|
| 2710 | rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer)
|
| 2711 | {
|
| 2712 | cpu_buffer->head_page
|
| 2713 | = list_entry(cpu_buffer->pages.next, struct buffer_page, list);
|
| 2714 | local_set(&cpu_buffer->head_page->write, 0);
|
| 2715 | local_set(&cpu_buffer->head_page->entries, 0);
|
| 2716 | local_set(&cpu_buffer->head_page->page->commit, 0);
|
| 2717 |
|
| 2718 | cpu_buffer->head_page->read = 0;
|
| 2719 |
|
| 2720 | cpu_buffer->tail_page = cpu_buffer->head_page;
|
| 2721 | cpu_buffer->commit_page = cpu_buffer->head_page;
|
| 2722 |
|
| 2723 | INIT_LIST_HEAD(&cpu_buffer->reader_page->list);
|
| 2724 | local_set(&cpu_buffer->reader_page->write, 0);
|
| 2725 | local_set(&cpu_buffer->reader_page->entries, 0);
|
| 2726 | local_set(&cpu_buffer->reader_page->page->commit, 0);
|
| 2727 | cpu_buffer->reader_page->read = 0;
|
| 2728 |
|
| 2729 | cpu_buffer->nmi_dropped = 0;
|
| 2730 | cpu_buffer->commit_overrun = 0;
|
| 2731 | cpu_buffer->overrun = 0;
|
| 2732 | cpu_buffer->read = 0;
|
| 2733 | local_set(&cpu_buffer->entries, 0);
|
| 2734 | local_set(&cpu_buffer->committing, 0);
|
| 2735 | local_set(&cpu_buffer->commits, 0);
|
| 2736 |
|
| 2737 | cpu_buffer->write_stamp = 0;
|
| 2738 | cpu_buffer->read_stamp = 0;
|
| 2739 | }
|
| 2740 |
|
| 2741 | |
| 2742 | * ring_buffer_reset_cpu - reset a ring buffer per CPU buffer
|
| 2743 | * @buffer: The ring buffer to reset a per cpu buffer of
|
| 2744 | * @cpu: The CPU buffer to be reset
|
| 2745 |
|
| 2746 | void ring_buffer_reset_cpu(struct ring_buffer *buffer, int cpu)
|
| 2747 | {
|
| 2748 | struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu];
|
| 2749 | unsigned long flags;
|
| 2750 |
|
| 2751 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 2752 | return;
|
| 2753 |
|
| 2754 | atomic_inc(&cpu_buffer->record_disabled);
|
| 2755 |
|
| 2756 | spin_lock_irqsave(&cpu_buffer->reader_lock, flags);
|
| 2757 |
|
| 2758 | __raw_spin_lock(&cpu_buffer->lock);
|
| 2759 |
|
| 2760 | rb_reset_cpu(cpu_buffer);
|
| 2761 |
|
| 2762 | __raw_spin_unlock(&cpu_buffer->lock);
|
| 2763 |
|
| 2764 | spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags);
|
| 2765 |
|
| 2766 | atomic_dec(&cpu_buffer->record_disabled);
|
| 2767 | }
|
| 2768 | EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu);
|
| 2769 |
|
| 2770 | |
| 2771 | * ring_buffer_reset - reset a ring buffer
|
| 2772 | * @buffer: The ring buffer to reset all cpu buffers
|
| 2773 |
|
| 2774 | void ring_buffer_reset(struct ring_buffer *buffer)
|
| 2775 | {
|
| 2776 | int cpu;
|
| 2777 |
|
| 2778 | for_each_buffer_cpu(buffer, cpu)
|
| 2779 | ring_buffer_reset_cpu(buffer, cpu);
|
| 2780 | }
|
| 2781 | EXPORT_SYMBOL_GPL(ring_buffer_reset);
|
| 2782 |
|
| 2783 | |
| 2784 | * rind_buffer_empty - is the ring buffer empty?
|
| 2785 | * @buffer: The ring buffer to test
|
| 2786 |
|
| 2787 | int ring_buffer_empty(struct ring_buffer *buffer)
|
| 2788 | {
|
| 2789 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2790 | unsigned long flags;
|
| 2791 | int dolock;
|
| 2792 | int cpu;
|
| 2793 | int ret;
|
| 2794 |
|
| 2795 | dolock = rb_ok_to_lock();
|
| 2796 |
|
| 2797 |
|
| 2798 | for_each_buffer_cpu(buffer, cpu) {
|
| 2799 | cpu_buffer = buffer->buffers[cpu];
|
| 2800 | local_irq_save(flags);
|
| 2801 | if (dolock)
|
| 2802 | spin_lock(&cpu_buffer->reader_lock);
|
| 2803 | ret = rb_per_cpu_empty(cpu_buffer);
|
| 2804 | if (dolock)
|
| 2805 | spin_unlock(&cpu_buffer->reader_lock);
|
| 2806 | local_irq_restore(flags);
|
| 2807 |
|
| 2808 | if (!ret)
|
| 2809 | return 0;
|
| 2810 | }
|
| 2811 |
|
| 2812 | return 1;
|
| 2813 | }
|
| 2814 | EXPORT_SYMBOL_GPL(ring_buffer_empty);
|
| 2815 |
|
| 2816 | |
| 2817 | * ring_buffer_empty_cpu - is a cpu buffer of a ring buffer empty?
|
| 2818 | * @buffer: The ring buffer
|
| 2819 | * @cpu: The CPU buffer to test
|
| 2820 |
|
| 2821 | int ring_buffer_empty_cpu(struct ring_buffer *buffer, int cpu)
|
| 2822 | {
|
| 2823 | struct ring_buffer_per_cpu *cpu_buffer;
|
| 2824 | unsigned long flags;
|
| 2825 | int dolock;
|
| 2826 | int ret;
|
| 2827 |
|
| 2828 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 2829 | return 1;
|
| 2830 |
|
| 2831 | dolock = rb_ok_to_lock();
|
| 2832 |
|
| 2833 | cpu_buffer = buffer->buffers[cpu];
|
| 2834 | local_irq_save(flags);
|
| 2835 | if (dolock)
|
| 2836 | spin_lock(&cpu_buffer->reader_lock);
|
| 2837 | ret = rb_per_cpu_empty(cpu_buffer);
|
| 2838 | if (dolock)
|
| 2839 | spin_unlock(&cpu_buffer->reader_lock);
|
| 2840 | local_irq_restore(flags);
|
| 2841 |
|
| 2842 | return ret;
|
| 2843 | }
|
| 2844 | EXPORT_SYMBOL_GPL(ring_buffer_empty_cpu);
|
| 2845 |
|
| 2846 | |
| 2847 | * ring_buffer_swap_cpu - swap a CPU buffer between two ring buffers
|
| 2848 | * @buffer_a: One buffer to swap with
|
| 2849 | * @buffer_b: The other buffer to swap with
|
| 2850 | *
|
| 2851 | * This function is useful for tracers that want to take a "snapshot"
|
| 2852 | * of a CPU buffer and has another back up buffer lying around.
|
| 2853 | * it is expected that the tracer handles the cpu buffer not being
|
| 2854 | * used at the moment.
|
| 2855 |
|
| 2856 | int ring_buffer_swap_cpu(struct ring_buffer *buffer_a,
|
| 2857 | struct ring_buffer *buffer_b, int cpu)
|
| 2858 | {
|
| 2859 | struct ring_buffer_per_cpu *cpu_buffer_a;
|
| 2860 | struct ring_buffer_per_cpu *cpu_buffer_b;
|
| 2861 | int ret = -EINVAL;
|
| 2862 |
|
| 2863 | if (!cpumask_test_cpu(cpu, buffer_a->cpumask) ||
|
| 2864 | !cpumask_test_cpu(cpu, buffer_b->cpumask))
|
| 2865 | goto out;
|
| 2866 |
|
| 2867 |
|
| 2868 | if (buffer_a->pages != buffer_b->pages)
|
| 2869 | goto out;
|
| 2870 |
|
| 2871 | ret = -EAGAIN;
|
| 2872 |
|
| 2873 | if (ring_buffer_flags != RB_BUFFERS_ON)
|
| 2874 | goto out;
|
| 2875 |
|
| 2876 | if (atomic_read(&buffer_a->record_disabled))
|
| 2877 | goto out;
|
| 2878 |
|
| 2879 | if (atomic_read(&buffer_b->record_disabled))
|
| 2880 | goto out;
|
| 2881 |
|
| 2882 | cpu_buffer_a = buffer_a->buffers[cpu];
|
| 2883 | cpu_buffer_b = buffer_b->buffers[cpu];
|
| 2884 |
|
| 2885 | if (atomic_read(&cpu_buffer_a->record_disabled))
|
| 2886 | goto out;
|
| 2887 |
|
| 2888 | if (atomic_read(&cpu_buffer_b->record_disabled))
|
| 2889 | goto out;
|
| 2890 |
|
| 2891 | |
| 2892 | * We can't do a synchronize_sched here because this
|
| 2893 | * function can be called in atomic context.
|
| 2894 | * Normally this will be called from the same CPU as cpu.
|
| 2895 | * If not it's up to the caller to protect this.
|
| 2896 |
|
| 2897 | atomic_inc(&cpu_buffer_a->record_disabled);
|
| 2898 | atomic_inc(&cpu_buffer_b->record_disabled);
|
| 2899 |
|
| 2900 | buffer_a->buffers[cpu] = cpu_buffer_b;
|
| 2901 | buffer_b->buffers[cpu] = cpu_buffer_a;
|
| 2902 |
|
| 2903 | cpu_buffer_b->buffer = buffer_a;
|
| 2904 | cpu_buffer_a->buffer = buffer_b;
|
| 2905 |
|
| 2906 | atomic_dec(&cpu_buffer_a->record_disabled);
|
| 2907 | atomic_dec(&cpu_buffer_b->record_disabled);
|
| 2908 |
|
| 2909 | ret = 0;
|
| 2910 | out:
|
| 2911 | return ret;
|
| 2912 | }
|
| 2913 | EXPORT_SYMBOL_GPL(ring_buffer_swap_cpu);
|
| 2914 |
|
| 2915 | |
| 2916 | * ring_buffer_alloc_read_page - allocate a page to read from buffer
|
| 2917 | * @buffer: the buffer to allocate for.
|
| 2918 | *
|
| 2919 | * This function is used in conjunction with ring_buffer_read_page.
|
| 2920 | * When reading a full page from the ring buffer, these functions
|
| 2921 | * can be used to speed up the process. The calling function should
|
| 2922 | * allocate a few pages first with this function. Then when it
|
| 2923 | * needs to get pages from the ring buffer, it passes the result
|
| 2924 | * of this function into ring_buffer_read_page, which will swap
|
| 2925 | * the page that was allocated, with the read page of the buffer.
|
| 2926 | *
|
| 2927 | * Returns:
|
| 2928 | * The page allocated, or NULL on error.
|
| 2929 |
|
| 2930 | void *ring_buffer_alloc_read_page(struct ring_buffer *buffer)
|
| 2931 | {
|
| 2932 | struct buffer_data_page *bpage;
|
| 2933 | unsigned long addr;
|
| 2934 |
|
| 2935 | addr = __get_free_page(GFP_KERNEL);
|
| 2936 | if (!addr)
|
| 2937 | return NULL;
|
| 2938 |
|
| 2939 | bpage = (void *)addr;
|
| 2940 |
|
| 2941 | rb_init_page(bpage);
|
| 2942 |
|
| 2943 | return bpage;
|
| 2944 | }
|
| 2945 | EXPORT_SYMBOL_GPL(ring_buffer_alloc_read_page);
|
| 2946 |
|
| 2947 | |
| 2948 | * ring_buffer_free_read_page - free an allocated read page
|
| 2949 | * @buffer: the buffer the page was allocate for
|
| 2950 | * @data: the page to free
|
| 2951 | *
|
| 2952 | * Free a page allocated from ring_buffer_alloc_read_page.
|
| 2953 |
|
| 2954 | void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data)
|
| 2955 | {
|
| 2956 | free_page((unsigned long)data);
|
| 2957 | }
|
| 2958 | EXPORT_SYMBOL_GPL(ring_buffer_free_read_page);
|
| 2959 |
|
| 2960 | |
| 2961 | * ring_buffer_read_page - extract a page from the ring buffer
|
| 2962 | * @buffer: buffer to extract from
|
| 2963 | * @data_page: the page to use allocated from ring_buffer_alloc_read_page
|
| 2964 | * @len: amount to extract
|
| 2965 | * @cpu: the cpu of the buffer to extract
|
| 2966 | * @full: should the extraction only happen when the page is full.
|
| 2967 | *
|
| 2968 | * This function will pull out a page from the ring buffer and consume it.
|
| 2969 | * @data_page must be the address of the variable that was returned
|
| 2970 | * from ring_buffer_alloc_read_page. This is because the page might be used
|
| 2971 | * to swap with a page in the ring buffer.
|
| 2972 | *
|
| 2973 | * for example:
|
| 2974 | * rpage = ring_buffer_alloc_read_page(buffer);
|
| 2975 | * if (!rpage)
|
| 2976 | * return error;
|
| 2977 | * ret = ring_buffer_read_page(buffer, &rpage, len, cpu, 0);
|
| 2978 | * if (ret >= 0)
|
| 2979 | * process_page(rpage, ret);
|
| 2980 | *
|
| 2981 | * When @full is set, the function will not return true unless
|
| 2982 | * the writer is off the reader page.
|
| 2983 | *
|
| 2984 | * Note: it is up to the calling functions to handle sleeps and wakeups.
|
| 2985 | * The ring buffer can be used anywhere in the kernel and can not
|
| 2986 | * blindly call wake_up. The layer that uses the ring buffer must be
|
| 2987 | * responsible for that.
|
| 2988 | *
|
| 2989 | * Returns:
|
| 2990 | * >=0 if data has been transferred, returns the offset of consumed data.
|
| 2991 | * <0 if no data has been transferred.
|
| 2992 |
|
| 2993 | int ring_buffer_read_page(struct ring_buffer *buffer,
|
| 2994 | void **data_page, size_t len, int cpu, int full)
|
| 2995 | {
|
| 2996 | struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu];
|
| 2997 | struct ring_buffer_event *event;
|
| 2998 | struct buffer_data_page *bpage;
|
| 2999 | struct buffer_page *reader;
|
| 3000 | unsigned long flags;
|
| 3001 | unsigned int commit;
|
| 3002 | unsigned int read;
|
| 3003 | u64 save_timestamp;
|
| 3004 | int ret = -1;
|
| 3005 |
|
| 3006 | if (!cpumask_test_cpu(cpu, buffer->cpumask))
|
| 3007 | goto out;
|
| 3008 |
|
| 3009 | |
| 3010 | * If len is not big enough to hold the page header, then
|
| 3011 | * we can not copy anything.
|
| 3012 |
|
| 3013 | if (len <= BUF_PAGE_HDR_SIZE)
|
| 3014 | goto out;
|
| 3015 |
|
| 3016 | len -= BUF_PAGE_HDR_SIZE;
|
| 3017 |
|
| 3018 | if (!data_page)
|
| 3019 | goto out;
|
| 3020 |
|
| 3021 | bpage = *data_page;
|
| 3022 | if (!bpage)
|
| 3023 | goto out;
|
| 3024 |
|
| 3025 | spin_lock_irqsave(&cpu_buffer->reader_lock, flags);
|
| 3026 |
|
| 3027 | reader = rb_get_reader_page(cpu_buffer);
|
| 3028 | if (!reader)
|
| 3029 | goto out_unlock;
|
| 3030 |
|
| 3031 | event = rb_reader_event(cpu_buffer);
|
| 3032 |
|
| 3033 | read = reader->read;
|
| 3034 | commit = rb_page_commit(reader);
|
| 3035 |
|
| 3036 | |
| 3037 | * If this page has been partially read or
|
| 3038 | * if len is not big enough to read the rest of the page or
|
| 3039 | * a writer is still on the page, then
|
| 3040 | * we must copy the data from the page to the buffer.
|
| 3041 | * Otherwise, we can simply swap the page with the one passed in.
|
| 3042 |
|
| 3043 | if (read || (len < (commit - read)) ||
|
| 3044 | cpu_buffer->reader_page == cpu_buffer->commit_page) {
|
| 3045 | struct buffer_data_page *rpage = cpu_buffer->reader_page->page;
|
| 3046 | unsigned int rpos = read;
|
| 3047 | unsigned int pos = 0;
|
| 3048 | unsigned int size;
|
| 3049 |
|
| 3050 | if (full)
|
| 3051 | goto out_unlock;
|
| 3052 |
|
| 3053 | if (len > (commit - read))
|
| 3054 | len = (commit - read);
|
| 3055 |
|
| 3056 | size = rb_event_length(event);
|
| 3057 |
|
| 3058 | if (len < size)
|
| 3059 | goto out_unlock;
|
| 3060 |
|
| 3061 |
|
| 3062 | save_timestamp = cpu_buffer->read_stamp;
|
| 3063 |
|
| 3064 |
|
| 3065 | do {
|
| 3066 | memcpy(bpage->data + pos, rpage->data + rpos, size);
|
| 3067 |
|
| 3068 | len -= size;
|
| 3069 |
|
| 3070 | rb_advance_reader(cpu_buffer);
|
| 3071 | rpos = reader->read;
|
| 3072 | pos += size;
|
| 3073 |
|
| 3074 | event = rb_reader_event(cpu_buffer);
|
| 3075 | size = rb_event_length(event);
|
| 3076 | } while (len > size);
|
| 3077 |
|
| 3078 |
|
| 3079 | local_set(&bpage->commit, pos);
|
| 3080 | bpage->time_stamp = save_timestamp;
|
| 3081 |
|
| 3082 |
|
| 3083 | read = 0;
|
| 3084 | } else {
|
| 3085 |
|
| 3086 | cpu_buffer->read += local_read(&reader->entries);
|
| 3087 |
|
| 3088 |
|
| 3089 | rb_init_page(bpage);
|
| 3090 | bpage = reader->page;
|
| 3091 | reader->page = *data_page;
|
| 3092 | local_set(&reader->write, 0);
|
| 3093 | local_set(&reader->entries, 0);
|
| 3094 | reader->read = 0;
|
| 3095 | *data_page = bpage;
|
| 3096 | }
|
| 3097 | ret = read;
|
| 3098 |
|
| 3099 | out_unlock:
|
| 3100 | spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags);
|
| 3101 |
|
| 3102 | out:
|
| 3103 | return ret;
|
| 3104 | }
|
| 3105 | EXPORT_SYMBOL_GPL(ring_buffer_read_page);
|
| 3106 |
|
| 3107 | static ssize_t
|
| 3108 | rb_simple_read(struct file *filp, char __user *ubuf,
|
| 3109 | size_t cnt, loff_t *ppos)
|
| 3110 | {
|
| 3111 | unsigned long *p = filp->private_data;
|
| 3112 | char buf[64];
|
| 3113 | int r;
|
| 3114 |
|
| 3115 | if (test_bit(RB_BUFFERS_DISABLED_BIT, p))
|
| 3116 | r = sprintf(buf, "permanently disabled\n");
|
| 3117 | else
|
| 3118 | r = sprintf(buf, "%d\n", test_bit(RB_BUFFERS_ON_BIT, p));
|
| 3119 |
|
| 3120 | return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
|
| 3121 | }
|
| 3122 |
|
| 3123 | static ssize_t
|
| 3124 | rb_simple_write(struct file *filp, const char __user *ubuf,
|
| 3125 | size_t cnt, loff_t *ppos)
|
| 3126 | {
|
| 3127 | unsigned long *p = filp->private_data;
|
| 3128 | char buf[64];
|
| 3129 | unsigned long val;
|
| 3130 | int ret;
|
| 3131 |
|
| 3132 | if (cnt >= sizeof(buf))
|
| 3133 | return -EINVAL;
|
| 3134 |
|
| 3135 | if (copy_from_user(&buf, ubuf, cnt))
|
| 3136 | return -EFAULT;
|
| 3137 |
|
| 3138 | buf[cnt] = 0;
|
| 3139 |
|
| 3140 | ret = strict_strtoul(buf, 10, &val);
|
| 3141 | if (ret < 0)
|
| 3142 | return ret;
|
| 3143 |
|
| 3144 | if (val)
|
| 3145 | set_bit(RB_BUFFERS_ON_BIT, p);
|
| 3146 | else
|
| 3147 | clear_bit(RB_BUFFERS_ON_BIT, p);
|
| 3148 |
|
| 3149 | (*ppos)++;
|
| 3150 |
|
| 3151 | return cnt;
|
| 3152 | }
|
| 3153 |
|
| 3154 | static const struct file_operations rb_simple_fops = {
|
| 3155 | .open = tracing_open_generic,
|
| 3156 | .read = rb_simple_read,
|
| 3157 | .write = rb_simple_write,
|
| 3158 | };
|
| 3159 |
|
| 3160 |
|
| 3161 | static __init int rb_init_debugfs(void)
|
| 3162 | {
|
| 3163 | struct dentry *d_tracer;
|
| 3164 |
|
| 3165 | d_tracer = tracing_init_dentry();
|
| 3166 |
|
| 3167 | trace_create_file("tracing_on", 0644, d_tracer,
|
| 3168 | &ring_buffer_flags, &rb_simple_fops);
|
| 3169 |
|
| 3170 | return 0;
|
| 3171 | }
|
| 3172 |
|
| 3173 | fs_initcall(rb_init_debugfs);
|
| 3174 |
|
| 3175 | #ifdef CONFIG_HOTPLUG_CPU
|
| 3176 | static int rb_cpu_notify(struct notifier_block *self,
|
| 3177 | unsigned long action, void *hcpu)
|
| 3178 | {
|
| 3179 | struct ring_buffer *buffer =
|
| 3180 | container_of(self, struct ring_buffer, cpu_notify);
|
| 3181 | long cpu = (long)hcpu;
|
| 3182 |
|
| 3183 | switch (action) {
|
| 3184 | case CPU_UP_PREPARE:
|
| 3185 | case CPU_UP_PREPARE_FROZEN:
|
| 3186 | if (cpumask_test_cpu(cpu, buffer->cpumask))
|
| 3187 | return NOTIFY_OK;
|
| 3188 |
|
| 3189 | buffer->buffers[cpu] =
|
| 3190 | rb_allocate_cpu_buffer(buffer, cpu);
|
| 3191 | if (!buffer->buffers[cpu]) {
|
| 3192 | WARN(1, "failed to allocate ring buffer on CPU %ld\n",
|
| 3193 | cpu);
|
| 3194 | return NOTIFY_OK;
|
| 3195 | }
|
| 3196 | smp_wmb();
|
| 3197 | cpumask_set_cpu(cpu, buffer->cpumask);
|
| 3198 | break;
|
| 3199 | case CPU_DOWN_PREPARE:
|
| 3200 | case CPU_DOWN_PREPARE_FROZEN:
|
| 3201 | |
| 3202 | * Do nothing.
|
| 3203 | * If we were to free the buffer, then the user would
|
| 3204 | * lose any trace that was in the buffer.
|
| 3205 |
|
| 3206 | break;
|
| 3207 | default:
|
| 3208 | break;
|
| 3209 | }
|
| 3210 | return NOTIFY_OK;
|
| 3211 | }
|
| 3212 | #endif
|