projects / blktrace.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
iowatcher: Fix path name handling when the trace files are in the current directory
[blktrace.git] / iowatcher / blkparse.c
CommitLineData
9e066e23
CM		 1/*
2 * Copyright (C) 2012 Fusion-io
3 *
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public
6 * License v2 as published by the Free Software Foundation.
7 *
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11 * GNU General Public License for more details.
12 *
13 * You should have received a copy of the GNU General Public License
14 * along with this program; if not, write to the Free Software
15 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
16 *
17 * Parts of this file were imported from Jens Axboe's blktrace sources (also GPL)
18 */
19#include <sys/types.h>
20#include <sys/stat.h>
21#include <fcntl.h>
22#include <unistd.h>
23#include <stdlib.h>
24#include <stdio.h>
25#include <math.h>
26#include <inttypes.h>
27#include <string.h>
28#include <asm/types.h>
29#include <errno.h>
30#include <sys/mman.h>
31#include <time.h>
32#include <math.h>
33
34#include "plot.h"
35#include "blkparse.h"
36#include "list.h"
37#include "tracers.h"
38
39#define IO_HASH_TABLE_BITS 11
40#define IO_HASH_TABLE_SIZE (1 << IO_HASH_TABLE_BITS)
41static struct list_head io_hash_table[IO_HASH_TABLE_SIZE];
42static u64 ios_in_flight = 0;
43
0a43b43f
JK		 44#define PROCESS_HASH_TABLE_BITS 7
45#define PROCESS_HASH_TABLE_SIZE (1 << PROCESS_HASH_TABLE_BITS)
46static struct list_head process_hash_table[PROCESS_HASH_TABLE_SIZE];
47
f2e40ddd 48extern int plot_io_action;
0a43b43f 49extern int io_per_process;
9e066e23
CM		 50
51/*
52 * Trace categories
53 */
54enum {
55 BLK_TC_READ = 1 << 0, /* reads */
56 BLK_TC_WRITE = 1 << 1, /* writes */
57 BLK_TC_FLUSH = 1 << 2, /* flush */
58 BLK_TC_SYNC = 1 << 3, /* sync */
59 BLK_TC_QUEUE = 1 << 4, /* queueing/merging */
60 BLK_TC_REQUEUE = 1 << 5, /* requeueing */
61 BLK_TC_ISSUE = 1 << 6, /* issue */
62 BLK_TC_COMPLETE = 1 << 7, /* completions */
63 BLK_TC_FS = 1 << 8, /* fs requests */
64 BLK_TC_PC = 1 << 9, /* pc requests */
65 BLK_TC_NOTIFY = 1 << 10, /* special message */
66 BLK_TC_AHEAD = 1 << 11, /* readahead */
67 BLK_TC_META = 1 << 12, /* metadata */
68 BLK_TC_DISCARD = 1 << 13, /* discard requests */
69 BLK_TC_DRV_DATA = 1 << 14, /* binary driver data */
70 BLK_TC_FUA = 1 << 15, /* fua requests */
71
72 BLK_TC_END = 1 << 15, /* we've run out of bits! */
73};
74
75#define BLK_TC_SHIFT (16)
76#define BLK_TC_ACT(act) ((act) << BLK_TC_SHIFT)
77#define BLK_DATADIR(a) (((a) >> BLK_TC_SHIFT) & (BLK_TC_READ | BLK_TC_WRITE))
78
79/*
80 * Basic trace actions
81 */
82enum {
83 __BLK_TA_QUEUE = 1, /* queued */
84 __BLK_TA_BACKMERGE, /* back merged to existing rq */
85 __BLK_TA_FRONTMERGE, /* front merge to existing rq */
86 __BLK_TA_GETRQ, /* allocated new request */
87 __BLK_TA_SLEEPRQ, /* sleeping on rq allocation */
88 __BLK_TA_REQUEUE, /* request requeued */
89 __BLK_TA_ISSUE, /* sent to driver */
90 __BLK_TA_COMPLETE, /* completed by driver */
91 __BLK_TA_PLUG, /* queue was plugged */
92 __BLK_TA_UNPLUG_IO, /* queue was unplugged by io */
93 __BLK_TA_UNPLUG_TIMER, /* queue was unplugged by timer */
94 __BLK_TA_INSERT, /* insert request */
95 __BLK_TA_SPLIT, /* bio was split */
96 __BLK_TA_BOUNCE, /* bio was bounced */
97 __BLK_TA_REMAP, /* bio was remapped */
98 __BLK_TA_ABORT, /* request aborted */
99 __BLK_TA_DRV_DATA, /* binary driver data */
100};
101
1582ecc9
JK		 102#define BLK_TA_MASK ((1 << BLK_TC_SHIFT) - 1)
103
9e066e23
CM		 104/*
105 * Notify events.
106 */
107enum blktrace_notify {
108 __BLK_TN_PROCESS = 0, /* establish pid/name mapping */
109 __BLK_TN_TIMESTAMP, /* include system clock */
110 __BLK_TN_MESSAGE, /* Character string message */
111};
112
113/*
114 * Trace actions in full. Additionally, read or write is masked
115 */
116#define BLK_TA_QUEUE (__BLK_TA_QUEUE | BLK_TC_ACT(BLK_TC_QUEUE))
117#define BLK_TA_BACKMERGE (__BLK_TA_BACKMERGE | BLK_TC_ACT(BLK_TC_QUEUE))
118#define BLK_TA_FRONTMERGE (__BLK_TA_FRONTMERGE | BLK_TC_ACT(BLK_TC_QUEUE))
119#define BLK_TA_GETRQ (__BLK_TA_GETRQ | BLK_TC_ACT(BLK_TC_QUEUE))
120#define BLK_TA_SLEEPRQ (__BLK_TA_SLEEPRQ | BLK_TC_ACT(BLK_TC_QUEUE))
121#define BLK_TA_REQUEUE (__BLK_TA_REQUEUE | BLK_TC_ACT(BLK_TC_REQUEUE))
122#define BLK_TA_ISSUE (__BLK_TA_ISSUE | BLK_TC_ACT(BLK_TC_ISSUE))
123#define BLK_TA_COMPLETE (__BLK_TA_COMPLETE| BLK_TC_ACT(BLK_TC_COMPLETE))
124#define BLK_TA_PLUG (__BLK_TA_PLUG | BLK_TC_ACT(BLK_TC_QUEUE))
125#define BLK_TA_UNPLUG_IO (__BLK_TA_UNPLUG_IO | BLK_TC_ACT(BLK_TC_QUEUE))
126#define BLK_TA_UNPLUG_TIMER (__BLK_TA_UNPLUG_TIMER | BLK_TC_ACT(BLK_TC_QUEUE))
127#define BLK_TA_INSERT (__BLK_TA_INSERT | BLK_TC_ACT(BLK_TC_QUEUE))
128#define BLK_TA_SPLIT (__BLK_TA_SPLIT)
129#define BLK_TA_BOUNCE (__BLK_TA_BOUNCE)
130#define BLK_TA_REMAP (__BLK_TA_REMAP | BLK_TC_ACT(BLK_TC_QUEUE))
131#define BLK_TA_ABORT (__BLK_TA_ABORT | BLK_TC_ACT(BLK_TC_QUEUE))
132#define BLK_TA_DRV_DATA (__BLK_TA_DRV_DATA | BLK_TC_ACT(BLK_TC_DRV_DATA))
133
134#define BLK_TN_PROCESS (__BLK_TN_PROCESS | BLK_TC_ACT(BLK_TC_NOTIFY))
135#define BLK_TN_TIMESTAMP (__BLK_TN_TIMESTAMP | BLK_TC_ACT(BLK_TC_NOTIFY))
136#define BLK_TN_MESSAGE (__BLK_TN_MESSAGE | BLK_TC_ACT(BLK_TC_NOTIFY))
137
138#define BLK_IO_TRACE_MAGIC 0x65617400
139#define BLK_IO_TRACE_VERSION 0x07
140/*
141 * The trace itself
142 */
143struct blk_io_trace {
144 __u32 magic; /* MAGIC << 8 | version */
145 __u32 sequence; /* event number */
146 __u64 time; /* in nanoseconds */
147 __u64 sector; /* disk offset */
148 __u32 bytes; /* transfer length */
149 __u32 action; /* what happened */
150 __u32 pid; /* who did it */
151 __u32 device; /* device identifier (dev_t) */
152 __u32 cpu; /* on what cpu did it happen */
153 __u16 error; /* completion error */
154 __u16 pdu_len; /* length of data after this trace */
155};
156
157struct pending_io {
158 /* sector offset of this IO */
159 u64 sector;
160
161 /* time this IO was dispatched */
162 u64 dispatch_time;
163 /* time this IO was finished */
164 u64 completion_time;
165 struct list_head hash_list;
0a43b43f
JK		 166 /* process which queued this IO */
167 u32 pid;
168};
169
170struct pid_map {
171 struct list_head hash_list;
172 u32 pid;
173 int index;
174 char name[0];
9e066e23
CM		 175};
176
177#define MINORBITS 20
178#define MINORMASK ((1 << MINORBITS) - 1)
179#define SECONDS(x) ((unsigned long long)(x) / 1000000000)
180#define NANO_SECONDS(x) ((unsigned long long)(x) % 1000000000)
181#define DOUBLE_TO_NANO_ULL(d) ((unsigned long long)((d) * 1000000000))
182#define CHECK_MAGIC(t) (((t)->magic & 0xffffff00) == BLK_IO_TRACE_MAGIC)
183
184void init_io_hash_table(void)
185{
186 int i;
187 struct list_head *head;
188
189 for (i = 0; i < IO_HASH_TABLE_SIZE; i++) {
190 head = io_hash_table + i;
191 INIT_LIST_HEAD(head);
192 }
193}
194
195/* taken from the kernel hash.h */
196static inline u64 hash_sector(u64 val)
197{
198 u64 hash = val;
199
200 /* Sigh, gcc can't optimise this alone like it does for 32 bits. */
201 u64 n = hash;
202 n <<= 18;
203 hash -= n;
204 n <<= 33;
205 hash -= n;
206 n <<= 3;
207 hash += n;
208 n <<= 3;
209 hash -= n;
210 n <<= 4;
211 hash += n;
212 n <<= 2;
213 hash += n;
214
215 /* High bits are more random, so use them. */
216 return hash >> (64 - IO_HASH_TABLE_BITS);
217}
218
0a43b43f 219static int io_hash_table_insert(struct pending_io *ins_pio)
9e066e23
CM		 220{
221 u64 sector = ins_pio->sector;
222 int slot = hash_sector(sector);
223 struct list_head *head;
224 struct pending_io *pio;
225
226 head = io_hash_table + slot;
227 list_for_each_entry(pio, head, hash_list) {
228 if (pio->sector == sector)
229 return -EEXIST;
230 }
231 list_add_tail(&ins_pio->hash_list, head);
232 return 0;
233}
234
0a43b43f 235static struct pending_io *io_hash_table_search(u64 sector)
9e066e23
CM		 236{
237 int slot = hash_sector(sector);
238 struct list_head *head;
239 struct pending_io *pio;
240
241 head = io_hash_table + slot;
242 list_for_each_entry(pio, head, hash_list) {
243 if (pio->sector == sector)
244 return pio;
245 }
246 return NULL;
247}
248
0a43b43f 249static int hash_queued_io(struct blk_io_trace *io)
9e066e23
CM		 250{
251 struct pending_io *pio;
252 int ret;
253
254 pio = calloc(1, sizeof(*pio));
255 pio->sector = io->sector;
0a43b43f 256 pio->pid = io->pid;
9e066e23 257
0a43b43f
JK		 258 ret = io_hash_table_insert(pio);
259 if (ret < 0) {
260 /* crud, the IO is there already */
9e066e23 261 free(pio);
0a43b43f 262 return ret;
9e066e23 263 }
0a43b43f
JK		 264 return 0;
265}
266
267static int hash_dispatched_io(struct blk_io_trace *io)
268{
269 struct pending_io *pio;
270
271 pio = io_hash_table_search(io->sector);
272 if (!pio) {
273 /* crud, the IO isn't here */
274 return -EEXIST;
275 }
276 pio->dispatch_time = io->time;
277 return 0;
9e066e23
CM		 278}
279
280static struct pending_io *hash_completed_io(struct blk_io_trace *io)
281{
282 struct pending_io *pio;
283
0a43b43f 284 pio = io_hash_table_search(io->sector);
9e066e23
CM		 285
286 if (!pio)
287 return NULL;
288 return pio;
289}
290
0a43b43f
JK		 291void init_process_hash_table(void)
292{
293 int i;
294 struct list_head *head;
295
296 for (i = 0; i < PROCESS_HASH_TABLE_SIZE; i++) {
297 head = process_hash_table + i;
298 INIT_LIST_HEAD(head);
299 }
300}
301
302static u32 hash_pid(u32 pid)
303{
304 u32 hash = pid;
305
306 hash ^= pid >> 3;
307 hash ^= pid >> 3;
308 hash ^= pid >> 4;
309 hash ^= pid >> 6;
310 return (hash & (PROCESS_HASH_TABLE_SIZE - 1));
311}
312
313static struct pid_map *process_hash_search(u32 pid)
314{
315 int slot = hash_pid(pid);
316 struct list_head *head;
317 struct pid_map *pm;
318
319 head = process_hash_table + slot;
320 list_for_each_entry(pm, head, hash_list) {
321 if (pm->pid == pid)
322 return pm;
323 }
324 return NULL;
325}
326
327static struct pid_map *process_hash_insert(u32 pid, char *name)
328{
329 int slot = hash_pid(pid);
330 struct pid_map *pm;
331 int old_index = 0;
332 char buf[16];
333
334 pm = process_hash_search(pid);
335 if (pm) {
336 /* Entry exists and name shouldn't be changed? */
337 if (!name || !strcmp(name, pm->name))
338 return pm;
339 list_del(&pm->hash_list);
340 old_index = pm->index;
341 free(pm);
342 }
343 if (!name) {
344 sprintf(buf, "[%u]", pid);
345 name = buf;
346 }
347 pm = malloc(sizeof(struct pid_map) + strlen(name) + 1);
348 pm->pid = pid;
349 pm->index = old_index;
350 strcpy(pm->name, name);
351 list_add_tail(&pm->hash_list, process_hash_table + slot);
352
353 return pm;
354}
355
9e066e23
CM		 356static void handle_notify(struct trace *trace)
357{
358 struct blk_io_trace *io = trace->io;
359 void *payload = (char *)io + sizeof(*io);
360 u32 two32[2];
361
0a43b43f
JK		 362 if (io->action == BLK_TN_PROCESS) {
363 if (io_per_process)
364 process_hash_insert(io->pid, payload);
365 return;
366 }
9e066e23
CM		 367
368 if (io->action != BLK_TN_TIMESTAMP)
369 return;
370
371 if (io->pdu_len != sizeof(two32))
372 return;
373
374 memcpy(two32, payload, sizeof(two32));
375 trace->start_timestamp = io->time;
376 trace->abs_start_time.tv_sec = two32[0];
377 trace->abs_start_time.tv_nsec = two32[1];
378 if (trace->abs_start_time.tv_nsec < 0) {
379 trace->abs_start_time.tv_sec--;
380 trace->abs_start_time.tv_nsec += 1000000000;
381 }
382}
383
384int next_record(struct trace *trace)
385{
386 int skip = trace->io->pdu_len;
387 u64 offset;
388
389 trace->cur += sizeof(*trace->io) + skip;
390 offset = trace->cur - trace->start;
391 if (offset >= trace->len)
392 return 1;
393
394 trace->io = (struct blk_io_trace *)trace->cur;
395 return 0;
396}
397
398void first_record(struct trace *trace)
399{
400 trace->cur = trace->start;
401 trace->io = (struct blk_io_trace *)trace->cur;
402}
403
bfb0e441
CM		 404int is_io_event(struct blk_io_trace *test)
405{
406 char *message;
407 if (!(test->action & BLK_TC_ACT(BLK_TC_NOTIFY)))
408 return 1;
409 if (test->action == BLK_TN_MESSAGE) {
410 int len = test->pdu_len;
411 if (len < 3)
412 return 0;
413 message = (char *)(test + 1);
414 if (strncmp(message, "fio ", 4) == 0) {
415 return 1;
416 }
417 }
418 return 0;
419}
420
9e066e23
CM		 421u64 find_last_time(struct trace *trace)
422{
423 char *p = trace->start + trace->len;
424 struct blk_io_trace *test;
425 int search_len = 0;
426 u64 found = 0;
427
428 if (trace->len < sizeof(*trace->io))
429 return 0;
430 p -= sizeof(*trace->io);
431 while (p >= trace->start) {
432 test = (struct blk_io_trace *)p;
bfb0e441 433 if (CHECK_MAGIC(test) && is_io_event(test)) {
9e066e23
CM		 434 u64 offset = p - trace->start;
435 if (offset + sizeof(*test) + test->pdu_len == trace->len) {
436 return test->time;
437 }
438 }
439 p--;
440 search_len++;
441 if (search_len > 8192) {
442 break;
443 }
444 }
445
446 /* searching backwards didn't work out, we'll have to scan the file */
447 first_record(trace);
448 while (1) {
bfb0e441 449 if (is_io_event(trace->io))
9e066e23
CM		 450 found = trace->io->time;
451 if (next_record(trace))
452 break;
453 }
454 first_record(trace);
455 return found;
456}
457
bfb0e441
CM		 458int parse_fio_bank_message(struct trace *trace, u64 *bank_ret, u64 *offset_ret,
459 u64 *num_banks_ret)
460{
461 char *s;
462 char *next;
463 char *message;
464 struct blk_io_trace *test = trace->io;
465 int len = test->pdu_len;
466 u64 bank;
467 u64 offset;
468 u64 num_banks;
469
470 if (!(test->action & BLK_TC_ACT(BLK_TC_NOTIFY)))
471 return -1;
472 if (test->action != BLK_TN_MESSAGE)
473 return -1;
474
475 /* the message is fio rw bank offset num_banks */
476 if (len < 3)
477 return -1;
478 message = (char *)(test + 1);
479 if (strncmp(message, "fio r ", 6) != 0)
480 return -1;
481
482 message = strndup(message, len);
483 s = strchr(message, ' ');
484 if (!s)
485 goto out;
486 s++;
487 s = strchr(s, ' ');
488 if (!s)
489 goto out;
490
491 bank = strtoll(s, &next, 10);
492 if (s == next)
493 goto out;
494 s = next;
495
496 offset = strtoll(s, &next, 10);
497 if (s == next)
498 goto out;
499 s = next;
500
501 num_banks = strtoll(s, &next, 10);
502 if (s == next)
503 goto out;
504
505 *bank_ret = bank;
506 *offset_ret = offset;
507 *num_banks_ret = num_banks;
508
509 return 0;
510out:
511 free(message);
512 return -1;
513}
514
9b9fa04b
JK		 515void find_extreme_offsets(struct trace *trace, u64 *min_ret, u64 *max_ret, u64 *max_bank_ret,
516 u64 *max_offset_ret)
9e066e23
CM		 517{
518 u64 found = 0;
9b9fa04b 519 u64 max = 0, min = ~(u64)0;
bfb0e441
CM		 520 u64 max_bank = 0;
521 u64 max_bank_offset = 0;
522 u64 num_banks = 0;
9e066e23
CM		 523 first_record(trace);
524 while (1) {
525 if (!(trace->io->action & BLK_TC_ACT(BLK_TC_NOTIFY))) {
526 found = trace->io->sector << 9;
9b9fa04b
JK		 527 if (found < min)
528 min = found;
9e066e23 529
9b9fa04b
JK		 530 found += trace->io->bytes;
531 if (max < found)
9e066e23 532 max = found;
bfb0e441
CM		 533 } else {
534 u64 bank;
535 u64 offset;
536 if (!parse_fio_bank_message(trace, &bank,
537 &offset, &num_banks)) {
538 if (bank > max_bank)
539 max_bank = bank;
540 if (offset > max_bank_offset)
541 max_bank_offset = offset;
542 }
9e066e23
CM		 543 }
544 if (next_record(trace))
545 break;
546 }
547 first_record(trace);
9b9fa04b 548 *min_ret = min;
bfb0e441
CM		 549 *max_ret = max;
550 *max_bank_ret = max_bank;
551 *max_offset_ret = max_bank_offset;
9e066e23
CM		 552}
553
9b9fa04b 554int filter_outliers(struct trace *trace, u64 min_offset, u64 max_offset,
9e066e23
CM		 555 u64 *yzoom_min, u64 *yzoom_max)
556{
557 int hits[11];
558 u64 max_per_bucket[11];
9b9fa04b
JK		 559 u64 min_per_bucket[11];
560 u64 bytes_per_bucket = (max_offset - min_offset + 1) / 10;
9e066e23
CM		 561 int slot;
562 int fat_count = 0;
563
564 memset(hits, 0, sizeof(int) * 11);
565 memset(max_per_bucket, 0, sizeof(u64) * 11);
9b9fa04b 566 memset(min_per_bucket, 0xff, sizeof(u64) * 11);
9e066e23
CM		 567 first_record(trace);
568 while (1) {
41fdf407
JK		 569 if (!(trace->io->action & BLK_TC_ACT(BLK_TC_NOTIFY)) &&
570 (trace->io->action & BLK_TA_MASK) == __BLK_TA_QUEUE) {
9b9fa04b
JK		 571 u64 off = (trace->io->sector << 9) - min_offset;
572
573 slot = (int)(off / bytes_per_bucket);
574 hits[slot]++;
575 if (off < min_per_bucket[slot])
576 min_per_bucket[slot] = off;
577
578 off += trace->io->bytes;
579 slot = (int)(off / bytes_per_bucket);
9e066e23 580 hits[slot]++;
9b9fa04b
JK		 581 if (off > max_per_bucket[slot])
582 max_per_bucket[slot] = off;
9e066e23
CM		 583 }
584 if (next_record(trace))
585 break;
586 }
587 first_record(trace);
588 for (slot = 0; slot < 11; slot++) {
589 if (hits[slot] > fat_count) {
590 fat_count = hits[slot];
591 }
592 }
593
594 *yzoom_max = max_offset;
595 for (slot = 10; slot >= 0; slot--) {
596 double d = hits[slot];
597
598 if (d >= (double)fat_count * .05) {
9b9fa04b 599 *yzoom_max = max_per_bucket[slot] + min_offset;
9e066e23
CM		 600 break;
601 }
602 }
603
9b9fa04b 604 *yzoom_min = min_offset;
9e066e23
CM		 605 for (slot = 0; slot < 10; slot++) {
606 double d = hits[slot];
607
608 if (d >= (double)fat_count * .05) {
9b9fa04b 609 *yzoom_min = min_per_bucket[slot] + min_offset;
9e066e23
CM		 610 break;
611 }
612 }
613 return 0;
614}
615
616static char *find_trace_file(char *filename)
617{
618 int ret;
619 struct stat st;
620 char line[1024];
621 char *dot;
622 char *try;
623
624 ret = stat(filename, &st);
625 if (ret == 0)
626 return strdup(filename);
627
628 snprintf(line, 1024, "%s.%s", filename, "dump");
e199d546 629 ret = stat(line, &st);
9e066e23
CM		 630 if (ret == 0)
631 return strdup(line);
632
633 try = strdup(filename);
634 dot = strrchr(try, '.');
635 if (!dot || strcmp(".dump", dot) != 0) {
e95ba659 636 if (dot && dot != try)
9e066e23
CM		 637 *dot = '\0';
638 snprintf(line, 1024, "%s%s", try, ".blktrace.0");
639 ret = stat(line, &st);
640 if (ret == 0) {
641 blktrace_to_dump(try);
642 snprintf(line, 1024, "%s.%s", try, "dump");
643 ret = stat(line, &st);
644 if (ret == 0) {
645 free(try);
646 return strdup(line);
647 }
648 }
649 }
650 free(try);
651 return NULL;
652}
653struct trace *open_trace(char *filename)
654{
655 int fd;
656 char *p;
657 struct stat st;
658 int ret;
659 struct trace *trace;
660 char *found_filename;
661
662 trace = calloc(1, sizeof(*trace));
663 if (!trace) {
664 fprintf(stderr, "unable to allocate memory for trace\n");
665 return NULL;
666 }
667
668 found_filename = find_trace_file(filename);
669 if (!found_filename) {
670 fprintf(stderr, "Unable to find trace file %s\n", filename);
671 goto fail;
672 }
9e066e23
CM		 673 filename = found_filename;
674
675 fd = open(filename, O_RDONLY);
676 if (fd < 0) {
677 fprintf(stderr, "Unable to open trace file %s err %s\n", filename, strerror(errno));
678 goto fail;
679 }
680 ret = fstat(fd, &st);
681 if (ret < 0) {
682 fprintf(stderr, "stat failed on %s err %s\n", filename, strerror(errno));
683 goto fail_fd;
684 }
685 p = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
686 if (p == MAP_FAILED) {
687 fprintf(stderr, "Unable to mmap trace file %s, err %s\n", filename, strerror(errno));
688 goto fail_fd;
689 }
690 trace->fd = fd;
691 trace->len = st.st_size;
692 trace->start = p;
693 trace->cur = p;
694 trace->io = (struct blk_io_trace *)p;
695 return trace;
696
697fail_fd:
698 close(fd);
699fail:
700 free(trace);
701 return NULL;
702}
703static inline int tput_event(struct trace *trace)
704{
705 if (trace->found_completion)
706 return __BLK_TA_COMPLETE;
707 if (trace->found_issue)
708 return __BLK_TA_ISSUE;
709 if (trace->found_queue)
710 return __BLK_TA_QUEUE;
711
712 return __BLK_TA_COMPLETE;
713}
714
f2e40ddd
JK		 715int action_char_to_num(char action)
716{
717 switch (action) {
718 case 'Q':
719 return __BLK_TA_QUEUE;
720 case 'D':
721 return __BLK_TA_ISSUE;
722 case 'C':
723 return __BLK_TA_COMPLETE;
724 }
725 return -1;
726}
727
9e066e23
CM		 728static inline int io_event(struct trace *trace)
729{
f2e40ddd
JK		 730 if (plot_io_action)
731 return plot_io_action;
9e066e23
CM		 732 if (trace->found_queue)
733 return __BLK_TA_QUEUE;
734 if (trace->found_issue)
735 return __BLK_TA_ISSUE;
736 if (trace->found_completion)
737 return __BLK_TA_COMPLETE;
738
739 return __BLK_TA_COMPLETE;
740}
741
742void add_tput(struct trace *trace, struct graph_line_data *gld)
743{
744 struct blk_io_trace *io = trace->io;
1582ecc9 745 int action = io->action & BLK_TA_MASK;
9e066e23
CM		 746 int seconds;
747
748 if (io->action & BLK_TC_ACT(BLK_TC_NOTIFY))
749 return;
750
751 if (action != tput_event(trace))
752 return;
753
754 seconds = SECONDS(io->time);
35686f9b
JK		 755 if (seconds > gld->max_seconds)
756 return;
9e066e23
CM		 757
758 gld->data[seconds].sum += io->bytes;
759 gld->data[seconds].count = 1;
760 if (gld->data[seconds].sum > gld->max)
761 gld->max = gld->data[seconds].sum;
762}
763
0a43b43f
JK		 764#define GDD_PTR_ALLOC_STEP 16
765
766static struct pid_map *get_pid_map(struct trace_file *tf, u32 pid)
767{
768 struct pid_map *pm;
769
770 if (!io_per_process) {
771 if (!tf->io_plots)
772 tf->io_plots = 1;
773 return NULL;
774 }
775
776 pm = process_hash_insert(pid, NULL);
777 /* New entry? */
778 if (!pm->index) {
779 if (tf->io_plots == tf->io_plots_allocated) {
780 tf->io_plots_allocated += GDD_PTR_ALLOC_STEP;
781 tf->gdd_reads = realloc(tf->gdd_reads, tf->io_plots_allocated * sizeof(struct graph_dot_data *));
782 if (!tf->gdd_reads)
783 abort();
784 tf->gdd_writes = realloc(tf->gdd_writes, tf->io_plots_allocated * sizeof(struct graph_dot_data *));
785 if (!tf->gdd_writes)
786 abort();
787 memset(tf->gdd_reads + tf->io_plots_allocated - GDD_PTR_ALLOC_STEP,
788 0, GDD_PTR_ALLOC_STEP * sizeof(struct graph_dot_data *));
789 memset(tf->gdd_writes + tf->io_plots_allocated - GDD_PTR_ALLOC_STEP,
790 0, GDD_PTR_ALLOC_STEP * sizeof(struct graph_dot_data *));
791 }
792 pm->index = tf->io_plots++;
793
794 return pm;
795 }
796 return pm;
797}
798
799void add_io(struct trace *trace, struct trace_file *tf)
9e066e23
CM		 800{
801 struct blk_io_trace *io = trace->io;
1582ecc9 802 int action = io->action & BLK_TA_MASK;
9e066e23 803 u64 offset;
0a43b43f
JK		 804 int index;
805 char *label;
806 struct pid_map *pm;
9e066e23
CM		 807
808 if (io->action & BLK_TC_ACT(BLK_TC_NOTIFY))
809 return;
810
811 if (action != io_event(trace))
812 return;
813
814 offset = io->sector << 9;
815
0a43b43f
JK		 816 pm = get_pid_map(tf, io->pid);
817 if (!pm) {
818 index = 0;
819 label = "";
820 } else {
821 index = pm->index;
822 label = pm->name;
823 }
824 if (BLK_DATADIR(io->action) & BLK_TC_READ) {
825 if (!tf->gdd_reads[index])
826 tf->gdd_reads[index] = alloc_dot_data(tf->min_seconds, tf->max_seconds, tf->min_offset, tf->max_offset, tf->stop_seconds, pick_color(), strdup(label));
827 set_gdd_bit(tf->gdd_reads[index], offset, io->bytes, io->time);
828 } else if (BLK_DATADIR(io->action) & BLK_TC_WRITE) {
829 if (!tf->gdd_writes[index])
830 tf->gdd_writes[index] = alloc_dot_data(tf->min_seconds, tf->max_seconds, tf->min_offset, tf->max_offset, tf->stop_seconds, pick_color(), strdup(label));
831 set_gdd_bit(tf->gdd_writes[index], offset, io->bytes, io->time);
832 }
9e066e23
CM		 833}
834
835void add_pending_io(struct trace *trace, struct graph_line_data *gld)
836{
837 int ret;
838 int seconds;
839 struct blk_io_trace *io = trace->io;
1582ecc9 840 int action = io->action & BLK_TA_MASK;
9e066e23
CM		 841 double avg;
842
843 if (io->action & BLK_TC_ACT(BLK_TC_NOTIFY))
844 return;
845
0a43b43f
JK		 846 if (action == __BLK_TA_QUEUE) {
847 hash_queued_io(trace->io);
848 return;
849 }
9e066e23
CM		 850 if (action != __BLK_TA_ISSUE)
851 return;
852
853 seconds = SECONDS(io->time);
35686f9b
JK		 854 if (seconds > gld->max_seconds)
855 return;
9e066e23
CM		 856
857 ret = hash_dispatched_io(trace->io);
858 if (ret)
859 return;
860
861 ios_in_flight++;
862
863 gld->data[seconds].sum += ios_in_flight;
864 gld->data[seconds].count++;
865
866 avg = (double)gld->data[seconds].sum / gld->data[seconds].count;
867 if (gld->max < (u64)avg) {
868 gld->max = avg;
869 }
870}
871
872void add_completed_io(struct trace *trace,
873 struct graph_line_data *latency_gld)
874{
875 struct blk_io_trace *io = trace->io;
876 int seconds;
1582ecc9 877 int action = io->action & BLK_TA_MASK;
9e066e23
CM		 878 struct pending_io *pio;
879 double avg;
880 u64 latency;
881
882 if (io->action & BLK_TC_ACT(BLK_TC_NOTIFY))
883 return;
884
885 if (action != __BLK_TA_COMPLETE)
886 return;
887
888 seconds = SECONDS(io->time);
889
890 pio = hash_completed_io(trace->io);
891 if (!pio)
892 return;
893
894 if (ios_in_flight > 0)
895 ios_in_flight--;
896 if (io->time >= pio->dispatch_time) {
897 latency = io->time - pio->dispatch_time;
898 latency_gld->data[seconds].sum += latency;
899 latency_gld->data[seconds].count++;
900 }
901
902 list_del(&pio->hash_list);
903 free(pio);
904
905 avg = (double)latency_gld->data[seconds].sum /
906 latency_gld->data[seconds].count;
907 if (latency_gld->max < (u64)avg) {
908 latency_gld->max = avg;
909 }
910}
911
912void add_iop(struct trace *trace, struct graph_line_data *gld)
913{
914 struct blk_io_trace *io = trace->io;
1582ecc9 915 int action = io->action & BLK_TA_MASK;
9e066e23
CM		 916 int seconds;
917
918 if (io->action & BLK_TC_ACT(BLK_TC_NOTIFY))
919 return;
920
921 /* iops and tput use the same events */
922 if (action != tput_event(trace))
923 return;
924
925 seconds = SECONDS(io->time);
35686f9b
JK		 926 if (seconds > gld->max_seconds)
927 return;
9e066e23
CM		 928
929 gld->data[seconds].sum += 1;
930 gld->data[seconds].count = 1;
931 if (gld->data[seconds].sum > gld->max)
932 gld->max = gld->data[seconds].sum;
933}
934
935void check_record(struct trace *trace)
936{
937 struct blk_io_trace *io = trace->io;
1582ecc9 938 int action = io->action & BLK_TA_MASK;
9e066e23
CM		 939
940 if (!(io->action & BLK_TC_ACT(BLK_TC_NOTIFY))) {
941 switch (action) {
942 case __BLK_TA_COMPLETE:
943 trace->found_completion = 1;
944 break;
945 case __BLK_TA_ISSUE:
946 trace->found_issue = 1;
947 break;
948 case __BLK_TA_QUEUE:
949 trace->found_queue = 1;
950 break;
951 };
952 }
953 handle_notify(trace);
954}
Block queue io tracer