projects / blktrace.git / blame_incremental
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (non-incremental) | history | HEAD
blkparse: Initialize and test for undefined request tracking timestamps
[blktrace.git] / blkparse.c
... / ...
CommitLineData
1/*
2 * block queue tracing parse application
3 *
4 * Copyright (C) 2005 Jens Axboe <axboe@suse.de>
5 * Copyright (C) 2006 Jens Axboe <axboe@kernel.dk>
6 *
7 * This program is free software; you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation; either version 2 of the License, or
10 * (at your option) any later version.
11 *
12 * This program is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 * GNU General Public License for more details.
16 *
17 * You should have received a copy of the GNU General Public License
18 * along with this program; if not, write to the Free Software
19 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20 *
21 */
22#include <sys/types.h>
23#include <sys/stat.h>
24#include <unistd.h>
25#include <stdio.h>
26#include <fcntl.h>
27#include <stdlib.h>
28#include <stdarg.h>
29#include <string.h>
30#include <getopt.h>
31#include <errno.h>
32#include <signal.h>
33#include <locale.h>
34#include <libgen.h>
35
36#include "blktrace.h"
37#include "rbtree.h"
38#include "jhash.h"
39
40static char blkparse_version[] = "1.2.0";
41
42struct skip_info {
43 unsigned long start, end;
44 struct skip_info *prev, *next;
45};
46
47struct per_dev_info {
48 dev_t dev;
49 char *name;
50
51 int backwards;
52 unsigned long long events;
53 unsigned long long first_reported_time;
54 unsigned long long last_reported_time;
55 unsigned long long last_read_time;
56 struct io_stats io_stats;
57 unsigned long skips;
58 unsigned long long seq_skips;
59 unsigned int max_depth[2];
60 unsigned int cur_depth[2];
61
62 struct rb_root rb_track;
63
64 int nfiles;
65 int ncpus;
66
67 unsigned long *cpu_map;
68 unsigned int cpu_map_max;
69
70 struct per_cpu_info *cpus;
71};
72
73/*
74 * some duplicated effort here, we can unify this hash and the ppi hash later
75 */
76struct process_pid_map {
77 pid_t pid;
78 char comm[16];
79 struct process_pid_map *hash_next, *list_next;
80};
81
82#define PPM_HASH_SHIFT (8)
83#define PPM_HASH_SIZE (1 << PPM_HASH_SHIFT)
84#define PPM_HASH_MASK (PPM_HASH_SIZE - 1)
85static struct process_pid_map *ppm_hash_table[PPM_HASH_SIZE];
86
87struct per_process_info {
88 struct process_pid_map *ppm;
89 struct io_stats io_stats;
90 struct per_process_info *hash_next, *list_next;
91 int more_than_one;
92
93 /*
94 * individual io stats
95 */
96 unsigned long long longest_allocation_wait[2];
97 unsigned long long longest_dispatch_wait[2];
98 unsigned long long longest_completion_wait[2];
99};
100
101#define PPI_HASH_SHIFT (8)
102#define PPI_HASH_SIZE (1 << PPI_HASH_SHIFT)
103#define PPI_HASH_MASK (PPI_HASH_SIZE - 1)
104
105enum {
106 SORT_PROG_EVENT_N, /* Program Name */
107 SORT_PROG_EVENT_QKB, /* KB: Queued read and write */
108 SORT_PROG_EVENT_RKB, /* KB: Queued Read */
109 SORT_PROG_EVENT_WKB, /* KB: Queued Write */
110 SORT_PROG_EVENT_CKB, /* KB: Complete */
111 SORT_PROG_EVENT_QIO, /* IO: Queued read and write */
112 SORT_PROG_EVENT_RIO, /* IO: Queued Read */
113 SORT_PROG_EVENT_WIO, /* IO: Queued Write */
114 SORT_PROG_EVENT_CIO, /* IO: Complete */
115};
116
117static struct per_process_info *ppi_hash_table[PPI_HASH_SIZE];
118static struct per_process_info *ppi_list;
119static int ppi_list_entries;
120
121static struct option l_opts[] = {
122 {
123 .name = "act-mask",
124 .has_arg = required_argument,
125 .flag = NULL,
126 .val = 'a'
127 },
128 {
129 .name = "set-mask",
130 .has_arg = required_argument,
131 .flag = NULL,
132 .val = 'A'
133 },
134 {
135 .name = "batch",
136 .has_arg = required_argument,
137 .flag = NULL,
138 .val = 'b'
139 },
140 {
141 .name = "input-directory",
142 .has_arg = required_argument,
143 .flag = NULL,
144 .val = 'D'
145 },
146 {
147 .name = "dump-binary",
148 .has_arg = required_argument,
149 .flag = NULL,
150 .val = 'd'
151 },
152 {
153 .name = "format",
154 .has_arg = required_argument,
155 .flag = NULL,
156 .val = 'f'
157 },
158 {
159 .name = "format-spec",
160 .has_arg = required_argument,
161 .flag = NULL,
162 .val = 'F'
163 },
164 {
165 .name = "hash-by-name",
166 .has_arg = no_argument,
167 .flag = NULL,
168 .val = 'h'
169 },
170 {
171 .name = "input",
172 .has_arg = required_argument,
173 .flag = NULL,
174 .val = 'i'
175 },
176 {
177 .name = "no-msgs",
178 .has_arg = no_argument,
179 .flag = NULL,
180 .val = 'M'
181 },
182 {
183 .name = "output",
184 .has_arg = required_argument,
185 .flag = NULL,
186 .val = 'o'
187 },
188 {
189 .name = "no-text-output",
190 .has_arg = no_argument,
191 .flag = NULL,
192 .val = 'O'
193 },
194 {
195 .name = "quiet",
196 .has_arg = no_argument,
197 .flag = NULL,
198 .val = 'q'
199 },
200 {
201 .name = "per-program-stats",
202 .has_arg = no_argument,
203 .flag = NULL,
204 .val = 's'
205 },
206 {
207 .name = "sort-program-stats",
208 .has_arg = required_argument,
209 .flag = NULL,
210 .val = 'S'
211 },
212 {
213 .name = "track-ios",
214 .has_arg = no_argument,
215 .flag = NULL,
216 .val = 't'
217 },
218 {
219 .name = "stopwatch",
220 .has_arg = required_argument,
221 .flag = NULL,
222 .val = 'w'
223 },
224 {
225 .name = "verbose",
226 .has_arg = no_argument,
227 .flag = NULL,
228 .val = 'v'
229 },
230 {
231 .name = "version",
232 .has_arg = no_argument,
233 .flag = NULL,
234 .val = 'V'
235 },
236 {
237 .name = NULL,
238 }
239};
240
241/*
242 * for sorting the displayed output
243 */
244struct trace {
245 struct blk_io_trace *bit;
246 struct rb_node rb_node;
247 struct trace *next;
248 unsigned long read_sequence;
249};
250
251static struct rb_root rb_sort_root;
252static unsigned long rb_sort_entries;
253
254static struct trace *trace_list;
255
256/*
257 * allocation cache
258 */
259static struct blk_io_trace *bit_alloc_list;
260static struct trace *t_alloc_list;
261
262/*
263 * for tracking individual ios
264 */
265struct io_track {
266 struct rb_node rb_node;
267
268 struct process_pid_map *ppm;
269 __u64 sector;
270 unsigned long long allocation_time;
271 unsigned long long queue_time;
272 unsigned long long dispatch_time;
273 unsigned long long completion_time;
274};
275
276static int ndevices;
277static struct per_dev_info *devices;
278static char *get_dev_name(struct per_dev_info *, char *, int);
279static int trace_rb_insert_last(struct per_dev_info *, struct trace *);
280
281FILE *ofp = NULL;
282static char *output_name;
283static char *input_dir;
284
285static unsigned long long genesis_time;
286static unsigned long long last_allowed_time;
287static unsigned long long stopwatch_start; /* start from zero by default */
288static unsigned long long stopwatch_end = -1ULL; /* "infinity" */
289static unsigned long read_sequence;
290
291static int per_process_stats;
292static int per_process_stats_event = SORT_PROG_EVENT_N;
293static int per_device_and_cpu_stats = 1;
294static int track_ios;
295static int ppi_hash_by_pid = 1;
296static int verbose;
297static unsigned int act_mask = -1U;
298static int stats_printed;
299static int bin_output_msgs = 1;
300int data_is_native = -1;
301
302static FILE *dump_fp;
303static char *dump_binary;
304
305static unsigned int t_alloc_cache;
306static unsigned int bit_alloc_cache;
307
308#define RB_BATCH_DEFAULT (512)
309static unsigned int rb_batch = RB_BATCH_DEFAULT;
310
311static int pipeline;
312static char *pipename;
313
314static int text_output = 1;
315
316#define is_done() (*(volatile int *)(&done))
317static volatile int done;
318
319struct timespec abs_start_time;
320static unsigned long long start_timestamp;
321
322static int have_drv_data = 0;
323
324#define JHASH_RANDOM (0x3af5f2ee)
325
326#define CPUS_PER_LONG (8 * sizeof(unsigned long))
327#define CPU_IDX(cpu) ((cpu) / CPUS_PER_LONG)
328#define CPU_BIT(cpu) ((cpu) & (CPUS_PER_LONG - 1))
329
330static void io_warn_unless(struct blk_io_trace *t, int condition,
331 const char *fmt, ...)
332{
333 va_list ap;
334
335 if (condition)
336 return;
337 va_start(ap, fmt);
338 printf("(%d,%d) request %llu + %u: ",
339 MAJOR(t->device), MINOR(t->device),
340 t->sector, t->bytes);
341 vfprintf(stderr, fmt, ap);
342 va_end(ap);
343}
344
345static void output_binary(void *buf, int len)
346{
347 if (dump_binary) {
348 size_t n = fwrite(buf, len, 1, dump_fp);
349 if (n != 1) {
350 perror(dump_binary);
351 fclose(dump_fp);
352 dump_binary = NULL;
353 }
354 }
355}
356
357static void resize_cpu_info(struct per_dev_info *pdi, int cpu)
358{
359 struct per_cpu_info *cpus = pdi->cpus;
360 int ncpus = pdi->ncpus;
361 int new_count = cpu + 1;
362 int new_space, size;
363 char *new_start;
364
365 size = new_count * sizeof(struct per_cpu_info);
366 cpus = realloc(cpus, size);
367 if (!cpus) {
368 char name[20];
369 fprintf(stderr, "Out of memory, CPU info for device %s (%d)\n",
370 get_dev_name(pdi, name, sizeof(name)), size);
371 exit(1);
372 }
373
374 new_start = (char *)cpus + (ncpus * sizeof(struct per_cpu_info));
375 new_space = (new_count - ncpus) * sizeof(struct per_cpu_info);
376 memset(new_start, 0, new_space);
377
378 pdi->ncpus = new_count;
379 pdi->cpus = cpus;
380
381 for (new_count = 0; new_count < pdi->ncpus; new_count++) {
382 struct per_cpu_info *pci = &pdi->cpus[new_count];
383
384 if (!pci->fd) {
385 pci->fd = -1;
386 memset(&pci->rb_last, 0, sizeof(pci->rb_last));
387 pci->rb_last_entries = 0;
388 pci->last_sequence = -1;
389 }
390 }
391}
392
393static struct per_cpu_info *get_cpu_info(struct per_dev_info *pdi, int cpu)
394{
395 struct per_cpu_info *pci;
396
397 if (cpu >= pdi->ncpus)
398 resize_cpu_info(pdi, cpu);
399
400 pci = &pdi->cpus[cpu];
401 pci->cpu = cpu;
402 return pci;
403}
404
405
406static int resize_devices(char *name)
407{
408 int size = (ndevices + 1) * sizeof(struct per_dev_info);
409
410 devices = realloc(devices, size);
411 if (!devices) {
412 fprintf(stderr, "Out of memory, device %s (%d)\n", name, size);
413 return 1;
414 }
415 memset(&devices[ndevices], 0, sizeof(struct per_dev_info));
416 devices[ndevices].name = name;
417 ndevices++;
418 return 0;
419}
420
421static struct per_dev_info *get_dev_info(dev_t dev)
422{
423 struct per_dev_info *pdi;
424 int i;
425
426 for (i = 0; i < ndevices; i++) {
427 if (!devices[i].dev)
428 devices[i].dev = dev;
429 if (devices[i].dev == dev)
430 return &devices[i];
431 }
432
433 if (resize_devices(NULL))
434 return NULL;
435
436 pdi = &devices[ndevices - 1];
437 pdi->dev = dev;
438 pdi->first_reported_time = 0;
439 pdi->last_read_time = 0;
440
441 return pdi;
442}
443
444static void insert_skip(struct per_cpu_info *pci, unsigned long start,
445 unsigned long end)
446{
447 struct skip_info *sip;
448
449 for (sip = pci->skips_tail; sip != NULL; sip = sip->prev) {
450 if (end == (sip->start - 1)) {
451 sip->start = start;
452 return;
453 } else if (start == (sip->end + 1)) {
454 sip->end = end;
455 return;
456 }
457 }
458
459 sip = malloc(sizeof(struct skip_info));
460 sip->start = start;
461 sip->end = end;
462 sip->prev = sip->next = NULL;
463 if (pci->skips_tail == NULL)
464 pci->skips_head = pci->skips_tail = sip;
465 else {
466 sip->prev = pci->skips_tail;
467 pci->skips_tail->next = sip;
468 pci->skips_tail = sip;
469 }
470}
471
472static void remove_sip(struct per_cpu_info *pci, struct skip_info *sip)
473{
474 if (sip->prev == NULL) {
475 if (sip->next == NULL)
476 pci->skips_head = pci->skips_tail = NULL;
477 else {
478 pci->skips_head = sip->next;
479 sip->next->prev = NULL;
480 }
481 } else if (sip->next == NULL) {
482 pci->skips_tail = sip->prev;
483 sip->prev->next = NULL;
484 } else {
485 sip->prev->next = sip->next;
486 sip->next->prev = sip->prev;
487 }
488
489 sip->prev = sip->next = NULL;
490 free(sip);
491}
492
493#define IN_SKIP(sip,seq) (((sip)->start <= (seq)) && ((seq) <= sip->end))
494static int check_current_skips(struct per_cpu_info *pci, unsigned long seq)
495{
496 struct skip_info *sip;
497
498 for (sip = pci->skips_tail; sip != NULL; sip = sip->prev) {
499 if (IN_SKIP(sip, seq)) {
500 if (sip->start == seq) {
501 if (sip->end == seq)
502 remove_sip(pci, sip);
503 else
504 sip->start += 1;
505 } else if (sip->end == seq)
506 sip->end -= 1;
507 else {
508 sip->end = seq - 1;
509 insert_skip(pci, seq + 1, sip->end);
510 }
511 return 1;
512 }
513 }
514
515 return 0;
516}
517
518static void collect_pdi_skips(struct per_dev_info *pdi)
519{
520 struct skip_info *sip;
521 int cpu;
522
523 pdi->skips = 0;
524 pdi->seq_skips = 0;
525
526 for (cpu = 0; cpu < pdi->ncpus; cpu++) {
527 struct per_cpu_info *pci = &pdi->cpus[cpu];
528
529 for (sip = pci->skips_head; sip != NULL; sip = sip->next) {
530 pdi->skips++;
531 pdi->seq_skips += (sip->end - sip->start + 1);
532 if (verbose)
533 fprintf(stderr,"(%d,%d): skipping %lu -> %lu\n",
534 MAJOR(pdi->dev), MINOR(pdi->dev),
535 sip->start, sip->end);
536 }
537 }
538}
539
540static void cpu_mark_online(struct per_dev_info *pdi, unsigned int cpu)
541{
542 if (cpu >= pdi->cpu_map_max || !pdi->cpu_map) {
543 int new_max = (cpu + CPUS_PER_LONG) & ~(CPUS_PER_LONG - 1);
544 unsigned long *map = malloc(new_max / sizeof(long));
545
546 memset(map, 0, new_max / sizeof(long));
547
548 if (pdi->cpu_map) {
549 memcpy(map, pdi->cpu_map, pdi->cpu_map_max / sizeof(long));
550 free(pdi->cpu_map);
551 }
552
553 pdi->cpu_map = map;
554 pdi->cpu_map_max = new_max;
555 }
556
557 pdi->cpu_map[CPU_IDX(cpu)] |= (1UL << CPU_BIT(cpu));
558}
559
560static inline void cpu_mark_offline(struct per_dev_info *pdi, int cpu)
561{
562 pdi->cpu_map[CPU_IDX(cpu)] &= ~(1UL << CPU_BIT(cpu));
563}
564
565static inline int cpu_is_online(struct per_dev_info *pdi, int cpu)
566{
567 return (pdi->cpu_map[CPU_IDX(cpu)] & (1UL << CPU_BIT(cpu))) != 0;
568}
569
570static inline int ppm_hash_pid(pid_t pid)
571{
572 return jhash_1word(pid, JHASH_RANDOM) & PPM_HASH_MASK;
573}
574
575static struct process_pid_map *find_ppm(pid_t pid)
576{
577 const int hash_idx = ppm_hash_pid(pid);
578 struct process_pid_map *ppm;
579
580 ppm = ppm_hash_table[hash_idx];
581 while (ppm) {
582 if (ppm->pid == pid)
583 return ppm;
584
585 ppm = ppm->hash_next;
586 }
587
588 return NULL;
589}
590
591static struct process_pid_map *add_ppm_hash(pid_t pid, const char *name)
592{
593 const int hash_idx = ppm_hash_pid(pid);
594 struct process_pid_map *ppm;
595
596 ppm = find_ppm(pid);
597 if (!ppm) {
598 ppm = malloc(sizeof(*ppm));
599 memset(ppm, 0, sizeof(*ppm));
600 ppm->pid = pid;
601 memset(ppm->comm, 0, sizeof(ppm->comm));
602 strncpy(ppm->comm, name, sizeof(ppm->comm));
603 ppm->comm[sizeof(ppm->comm) - 1] = '\0';
604 ppm->hash_next = ppm_hash_table[hash_idx];
605 ppm_hash_table[hash_idx] = ppm;
606 }
607
608 return ppm;
609}
610
611static void handle_notify(struct blk_io_trace *bit)
612{
613 void *payload = (caddr_t) bit + sizeof(*bit);
614 __u32 two32[2];
615
616 switch (bit->action) {
617 case BLK_TN_PROCESS:
618 add_ppm_hash(bit->pid, payload);
619 break;
620
621 case BLK_TN_TIMESTAMP:
622 if (bit->pdu_len != sizeof(two32))
623 return;
624 memcpy(two32, payload, sizeof(two32));
625 if (!data_is_native) {
626 two32[0] = be32_to_cpu(two32[0]);
627 two32[1] = be32_to_cpu(two32[1]);
628 }
629 start_timestamp = bit->time;
630 abs_start_time.tv_sec = two32[0];
631 abs_start_time.tv_nsec = two32[1];
632 if (abs_start_time.tv_nsec < 0) {
633 abs_start_time.tv_sec--;
634 abs_start_time.tv_nsec += 1000000000;
635 }
636
637 break;
638
639 case BLK_TN_MESSAGE:
640 if (bit->pdu_len > 0) {
641 char msg[bit->pdu_len+1];
642
643 memcpy(msg, (char *)payload, bit->pdu_len);
644 msg[bit->pdu_len] = '\0';
645
646 fprintf(ofp,
647 "%3d,%-3d %2d %8s %5d.%09lu %5u %2s %3s %s\n",
648 MAJOR(bit->device), MINOR(bit->device),
649 bit->cpu, "0", (int) SECONDS(bit->time),
650 (unsigned long) NANO_SECONDS(bit->time),
651 0, "m", "N", msg);
652 }
653 break;
654
655 default:
656 /* Ignore unknown notify events */
657 ;
658 }
659}
660
661char *find_process_name(pid_t pid)
662{
663 struct process_pid_map *ppm = find_ppm(pid);
664
665 if (ppm)
666 return ppm->comm;
667
668 return NULL;
669}
670
671static inline int ppi_hash_pid(pid_t pid)
672{
673 return jhash_1word(pid, JHASH_RANDOM) & PPI_HASH_MASK;
674}
675
676static inline int ppi_hash_name(const char *name)
677{
678 return jhash(name, 16, JHASH_RANDOM) & PPI_HASH_MASK;
679}
680
681static inline int ppi_hash(struct per_process_info *ppi)
682{
683 struct process_pid_map *ppm = ppi->ppm;
684
685 if (ppi_hash_by_pid)
686 return ppi_hash_pid(ppm->pid);
687
688 return ppi_hash_name(ppm->comm);
689}
690
691static inline void add_ppi_to_hash(struct per_process_info *ppi)
692{
693 const int hash_idx = ppi_hash(ppi);
694
695 ppi->hash_next = ppi_hash_table[hash_idx];
696 ppi_hash_table[hash_idx] = ppi;
697}
698
699static inline void add_ppi_to_list(struct per_process_info *ppi)
700{
701 ppi->list_next = ppi_list;
702 ppi_list = ppi;
703 ppi_list_entries++;
704}
705
706static struct per_process_info *find_ppi_by_name(char *name)
707{
708 const int hash_idx = ppi_hash_name(name);
709 struct per_process_info *ppi;
710
711 ppi = ppi_hash_table[hash_idx];
712 while (ppi) {
713 struct process_pid_map *ppm = ppi->ppm;
714
715 if (!strcmp(ppm->comm, name))
716 return ppi;
717
718 ppi = ppi->hash_next;
719 }
720
721 return NULL;
722}
723
724static struct per_process_info *find_ppi_by_pid(pid_t pid)
725{
726 const int hash_idx = ppi_hash_pid(pid);
727 struct per_process_info *ppi;
728
729 ppi = ppi_hash_table[hash_idx];
730 while (ppi) {
731 struct process_pid_map *ppm = ppi->ppm;
732
733 if (ppm->pid == pid)
734 return ppi;
735
736 ppi = ppi->hash_next;
737 }
738
739 return NULL;
740}
741
742static struct per_process_info *find_ppi(pid_t pid)
743{
744 struct per_process_info *ppi;
745 char *name;
746
747 if (ppi_hash_by_pid)
748 return find_ppi_by_pid(pid);
749
750 name = find_process_name(pid);
751 if (!name)
752 return NULL;
753
754 ppi = find_ppi_by_name(name);
755 if (ppi && ppi->ppm->pid != pid)
756 ppi->more_than_one = 1;
757
758 return ppi;
759}
760
761/*
762 * struct trace and blktrace allocation cache, we do potentially
763 * millions of mallocs for these structures while only using at most
764 * a few thousand at the time
765 */
766static inline void t_free(struct trace *t)
767{
768 if (t_alloc_cache < 1024) {
769 t->next = t_alloc_list;
770 t_alloc_list = t;
771 t_alloc_cache++;
772 } else
773 free(t);
774}
775
776static inline struct trace *t_alloc(void)
777{
778 struct trace *t = t_alloc_list;
779
780 if (t) {
781 t_alloc_list = t->next;
782 t_alloc_cache--;
783 return t;
784 }
785
786 return malloc(sizeof(*t));
787}
788
789static inline void bit_free(struct blk_io_trace *bit)
790{
791 if (bit_alloc_cache < 1024 && !bit->pdu_len) {
792 /*
793 * abuse a 64-bit field for a next pointer for the free item
794 */
795 bit->time = (__u64) (unsigned long) bit_alloc_list;
796 bit_alloc_list = (struct blk_io_trace *) bit;
797 bit_alloc_cache++;
798 } else
799 free(bit);
800}
801
802static inline struct blk_io_trace *bit_alloc(void)
803{
804 struct blk_io_trace *bit = bit_alloc_list;
805
806 if (bit) {
807 bit_alloc_list = (struct blk_io_trace *) (unsigned long) \
808 bit->time;
809 bit_alloc_cache--;
810 return bit;
811 }
812
813 return malloc(sizeof(*bit));
814}
815
816static inline void __put_trace_last(struct per_dev_info *pdi, struct trace *t)
817{
818 struct per_cpu_info *pci = get_cpu_info(pdi, t->bit->cpu);
819
820 rb_erase(&t->rb_node, &pci->rb_last);
821 pci->rb_last_entries--;
822
823 bit_free(t->bit);
824 t_free(t);
825}
826
827static void put_trace(struct per_dev_info *pdi, struct trace *t)
828{
829 rb_erase(&t->rb_node, &rb_sort_root);
830 rb_sort_entries--;
831
832 trace_rb_insert_last(pdi, t);
833}
834
835static inline int trace_rb_insert(struct trace *t, struct rb_root *root)
836{
837 struct rb_node **p = &root->rb_node;
838 struct rb_node *parent = NULL;
839 struct trace *__t;
840
841 while (*p) {
842 parent = *p;
843
844 __t = rb_entry(parent, struct trace, rb_node);
845
846 if (t->bit->time < __t->bit->time)
847 p = &(*p)->rb_left;
848 else if (t->bit->time > __t->bit->time)
849 p = &(*p)->rb_right;
850 else if (t->bit->device < __t->bit->device)
851 p = &(*p)->rb_left;
852 else if (t->bit->device > __t->bit->device)
853 p = &(*p)->rb_right;
854 else if (t->bit->sequence < __t->bit->sequence)
855 p = &(*p)->rb_left;
856 else /* >= sequence */
857 p = &(*p)->rb_right;
858 }
859
860 rb_link_node(&t->rb_node, parent, p);
861 rb_insert_color(&t->rb_node, root);
862 return 0;
863}
864
865static inline int trace_rb_insert_sort(struct trace *t)
866{
867 if (!trace_rb_insert(t, &rb_sort_root)) {
868 rb_sort_entries++;
869 return 0;
870 }
871
872 return 1;
873}
874
875static int trace_rb_insert_last(struct per_dev_info *pdi, struct trace *t)
876{
877 struct per_cpu_info *pci = get_cpu_info(pdi, t->bit->cpu);
878
879 if (trace_rb_insert(t, &pci->rb_last))
880 return 1;
881
882 pci->rb_last_entries++;
883
884 if (pci->rb_last_entries > rb_batch * pdi->nfiles) {
885 struct rb_node *n = rb_first(&pci->rb_last);
886
887 t = rb_entry(n, struct trace, rb_node);
888 __put_trace_last(pdi, t);
889 }
890
891 return 0;
892}
893
894static struct trace *trace_rb_find(dev_t device, unsigned long sequence,
895 struct rb_root *root, int order)
896{
897 struct rb_node *n = root->rb_node;
898 struct rb_node *prev = NULL;
899 struct trace *__t;
900
901 while (n) {
902 __t = rb_entry(n, struct trace, rb_node);
903 prev = n;
904
905 if (device < __t->bit->device)
906 n = n->rb_left;
907 else if (device > __t->bit->device)
908 n = n->rb_right;
909 else if (sequence < __t->bit->sequence)
910 n = n->rb_left;
911 else if (sequence > __t->bit->sequence)
912 n = n->rb_right;
913 else
914 return __t;
915 }
916
917 /*
918 * hack - the list may not be sequence ordered because some
919 * events don't have sequence and time matched. so we end up
920 * being a little off in the rb lookup here, because we don't
921 * know the time we are looking for. compensate by browsing
922 * a little ahead from the last entry to find the match
923 */
924 if (order && prev) {
925 int max = 5;
926
927 while (((n = rb_next(prev)) != NULL) && max--) {
928 __t = rb_entry(n, struct trace, rb_node);
929
930 if (__t->bit->device == device &&
931 __t->bit->sequence == sequence)
932 return __t;
933
934 prev = n;
935 }
936 }
937
938 return NULL;
939}
940
941static inline struct trace *trace_rb_find_last(struct per_dev_info *pdi,
942 struct per_cpu_info *pci,
943 unsigned long seq)
944{
945 return trace_rb_find(pdi->dev, seq, &pci->rb_last, 0);
946}
947
948static inline int track_rb_insert(struct per_dev_info *pdi,struct io_track *iot)
949{
950 struct rb_node **p = &pdi->rb_track.rb_node;
951 struct rb_node *parent = NULL;
952 struct io_track *__iot;
953
954 while (*p) {
955 parent = *p;
956 __iot = rb_entry(parent, struct io_track, rb_node);
957
958 if (iot->sector < __iot->sector)
959 p = &(*p)->rb_left;
960 else if (iot->sector > __iot->sector)
961 p = &(*p)->rb_right;
962 else {
963 fprintf(stderr,
964 "sector alias (%Lu) on device %d,%d!\n",
965 (unsigned long long) iot->sector,
966 MAJOR(pdi->dev), MINOR(pdi->dev));
967 return 1;
968 }
969 }
970
971 rb_link_node(&iot->rb_node, parent, p);
972 rb_insert_color(&iot->rb_node, &pdi->rb_track);
973 return 0;
974}
975
976static struct io_track *__find_track(struct per_dev_info *pdi, __u64 sector)
977{
978 struct rb_node *n = pdi->rb_track.rb_node;
979 struct io_track *__iot;
980
981 while (n) {
982 __iot = rb_entry(n, struct io_track, rb_node);
983
984 if (sector < __iot->sector)
985 n = n->rb_left;
986 else if (sector > __iot->sector)
987 n = n->rb_right;
988 else
989 return __iot;
990 }
991
992 return NULL;
993}
994
995static struct io_track *find_track(struct per_dev_info *pdi, pid_t pid,
996 __u64 sector)
997{
998 struct io_track *iot;
999
1000 iot = __find_track(pdi, sector);
1001 if (!iot) {
1002 iot = malloc(sizeof(*iot));
1003 iot->ppm = find_ppm(pid);
1004 if (!iot->ppm)
1005 iot->ppm = add_ppm_hash(pid, "unknown");
1006 iot->sector = sector;
1007 iot->allocation_time = -1ULL;
1008 iot->queue_time = -1ULL;
1009 iot->dispatch_time = -1ULL;
1010 iot->completion_time = -1ULL;
1011 track_rb_insert(pdi, iot);
1012 }
1013
1014 return iot;
1015}
1016
1017static void log_track_frontmerge(struct per_dev_info *pdi,
1018 struct blk_io_trace *t)
1019{
1020 struct io_track *iot;
1021
1022 if (!track_ios)
1023 return;
1024
1025 iot = __find_track(pdi, t->sector + t_sec(t));
1026 if (!iot) {
1027 if (verbose)
1028 fprintf(stderr, "merge not found for (%d,%d): %llu\n",
1029 MAJOR(t->device), MINOR(t->device),
1030 (unsigned long long) t->sector + t_sec(t));
1031 return;
1032 }
1033
1034 rb_erase(&iot->rb_node, &pdi->rb_track);
1035 iot->sector -= t_sec(t);
1036 track_rb_insert(pdi, iot);
1037}
1038
1039static void log_track_getrq(struct per_dev_info *pdi, struct blk_io_trace *t)
1040{
1041 struct io_track *iot;
1042
1043 if (!track_ios)
1044 return;
1045
1046 iot = find_track(pdi, t->pid, t->sector);
1047 io_warn_unless(t, iot->allocation_time == -1ULL,
1048 "confused about %s time", "allocation");
1049 iot->allocation_time = t->time;
1050}
1051
1052/*
1053 * for md/dm setups, the interesting cycle is Q -> C. So track queueing
1054 * time here, as dispatch time
1055 */
1056static void log_track_queue(struct per_dev_info *pdi, struct blk_io_trace *t)
1057{
1058 struct io_track *iot;
1059
1060 if (!track_ios)
1061 return;
1062
1063 iot = find_track(pdi, t->pid, t->sector);
1064 io_warn_unless(t, iot->dispatch_time == -1ULL,
1065 "confused about %s time", "dispatch");
1066 iot->dispatch_time = t->time;
1067}
1068
1069/*
1070 * return time between rq allocation and insertion
1071 */
1072static unsigned long long log_track_insert(struct per_dev_info *pdi,
1073 struct blk_io_trace *t)
1074{
1075 unsigned long long elapsed;
1076 struct io_track *iot;
1077
1078 if (!track_ios)
1079 return -1;
1080
1081 iot = find_track(pdi, t->pid, t->sector);
1082 io_warn_unless(t, iot->queue_time == -1ULL,
1083 "confused about %s time", "queue");
1084 iot->queue_time = t->time;
1085
1086 if (iot->allocation_time == -1ULL)
1087 return -1;
1088
1089 elapsed = iot->queue_time - iot->allocation_time;
1090
1091 if (per_process_stats) {
1092 struct per_process_info *ppi = find_ppi(iot->ppm->pid);
1093 int w = (t->action & BLK_TC_ACT(BLK_TC_WRITE)) != 0;
1094
1095 if (ppi && elapsed > ppi->longest_allocation_wait[w])
1096 ppi->longest_allocation_wait[w] = elapsed;
1097 }
1098
1099 return elapsed;
1100}
1101
1102/*
1103 * return time between queue and issue
1104 */
1105static unsigned long long log_track_issue(struct per_dev_info *pdi,
1106 struct blk_io_trace *t)
1107{
1108 unsigned long long elapsed = -1ULL;
1109 struct io_track *iot;
1110
1111 if (!track_ios)
1112 return -1;
1113 if ((t->action & BLK_TC_ACT(BLK_TC_FS)) == 0)
1114 return -1;
1115
1116 iot = __find_track(pdi, t->sector);
1117 if (!iot) {
1118 if (verbose)
1119 fprintf(stderr, "issue not found for (%d,%d): %llu\n",
1120 MAJOR(t->device), MINOR(t->device),
1121 (unsigned long long) t->sector);
1122 return -1;
1123 }
1124
1125 io_warn_unless(t, iot->dispatch_time == -1ULL,
1126 "confused about %s time", "dispatch");
1127 iot->dispatch_time = t->time;
1128 if (iot->queue_time != -1ULL)
1129 elapsed = iot->dispatch_time - iot->queue_time;
1130
1131 if (elapsed != -1ULL && per_process_stats) {
1132 struct per_process_info *ppi = find_ppi(iot->ppm->pid);
1133 int w = (t->action & BLK_TC_ACT(BLK_TC_WRITE)) != 0;
1134
1135 if (ppi && elapsed > ppi->longest_dispatch_wait[w])
1136 ppi->longest_dispatch_wait[w] = elapsed;
1137 }
1138
1139 return elapsed;
1140}
1141
1142/*
1143 * return time between dispatch and complete
1144 */
1145static unsigned long long log_track_complete(struct per_dev_info *pdi,
1146 struct blk_io_trace *t)
1147{
1148 unsigned long long elapsed = -1ULL;
1149 struct io_track *iot;
1150
1151 if (!track_ios)
1152 return -1;
1153
1154 iot = __find_track(pdi, t->sector);
1155 if (!iot) {
1156 if (verbose)
1157 fprintf(stderr,"complete not found for (%d,%d): %llu\n",
1158 MAJOR(t->device), MINOR(t->device),
1159 (unsigned long long) t->sector);
1160 return -1;
1161 }
1162
1163 io_warn_unless(t, iot->completion_time == -1ULL,
1164 "confused about %s time", "completion");
1165 iot->completion_time = t->time;
1166 if (iot->dispatch_time != -1ULL)
1167 elapsed = iot->completion_time - iot->dispatch_time;
1168
1169 if (elapsed != -1ULL && per_process_stats) {
1170 struct per_process_info *ppi = find_ppi(iot->ppm->pid);
1171 int w = (t->action & BLK_TC_ACT(BLK_TC_WRITE)) != 0;
1172
1173 if (ppi && elapsed > ppi->longest_completion_wait[w])
1174 ppi->longest_completion_wait[w] = elapsed;
1175 }
1176
1177 /*
1178 * kill the trace, we don't need it after completion
1179 */
1180 rb_erase(&iot->rb_node, &pdi->rb_track);
1181 free(iot);
1182
1183 return elapsed;
1184}
1185
1186
1187static struct io_stats *find_process_io_stats(pid_t pid)
1188{
1189 struct per_process_info *ppi = find_ppi(pid);
1190
1191 if (!ppi) {
1192 ppi = malloc(sizeof(*ppi));
1193 memset(ppi, 0, sizeof(*ppi));
1194 ppi->ppm = find_ppm(pid);
1195 if (!ppi->ppm)
1196 ppi->ppm = add_ppm_hash(pid, "unknown");
1197 add_ppi_to_hash(ppi);
1198 add_ppi_to_list(ppi);
1199 }
1200
1201 return &ppi->io_stats;
1202}
1203
1204static char *get_dev_name(struct per_dev_info *pdi, char *buffer, int size)
1205{
1206 if (pdi->name)
1207 snprintf(buffer, size, "%s", pdi->name);
1208 else
1209 snprintf(buffer, size, "%d,%d",MAJOR(pdi->dev),MINOR(pdi->dev));
1210 return buffer;
1211}
1212
1213static void check_time(struct per_dev_info *pdi, struct blk_io_trace *bit)
1214{
1215 unsigned long long this = bit->time;
1216 unsigned long long last = pdi->last_reported_time;
1217
1218 pdi->backwards = (this < last) ? 'B' : ' ';
1219 pdi->last_reported_time = this;
1220}
1221
1222static inline void __account_m(struct io_stats *ios, struct blk_io_trace *t,
1223 int rw)
1224{
1225 if (rw) {
1226 ios->mwrites++;
1227 ios->mwrite_kb += t_kb(t);
1228 ios->mwrite_b += t_b(t);
1229 } else {
1230 ios->mreads++;
1231 ios->mread_kb += t_kb(t);
1232 ios->mread_b += t_b(t);
1233 }
1234}
1235
1236static inline void account_m(struct blk_io_trace *t, struct per_cpu_info *pci,
1237 int rw)
1238{
1239 __account_m(&pci->io_stats, t, rw);
1240
1241 if (per_process_stats) {
1242 struct io_stats *ios = find_process_io_stats(t->pid);
1243
1244 __account_m(ios, t, rw);
1245 }
1246}
1247
1248static inline void __account_pc_queue(struct io_stats *ios,
1249 struct blk_io_trace *t, int rw)
1250{
1251 if (rw) {
1252 ios->qwrites_pc++;
1253 ios->qwrite_kb_pc += t_kb(t);
1254 ios->qwrite_b_pc += t_b(t);
1255 } else {
1256 ios->qreads_pc++;
1257 ios->qread_kb += t_kb(t);
1258 ios->qread_b_pc += t_b(t);
1259 }
1260}
1261
1262static inline void account_pc_queue(struct blk_io_trace *t,
1263 struct per_cpu_info *pci, int rw)
1264{
1265 __account_pc_queue(&pci->io_stats, t, rw);
1266
1267 if (per_process_stats) {
1268 struct io_stats *ios = find_process_io_stats(t->pid);
1269
1270 __account_pc_queue(ios, t, rw);
1271 }
1272}
1273
1274static inline void __account_pc_issue(struct io_stats *ios, int rw,
1275 unsigned int bytes)
1276{
1277 if (rw) {
1278 ios->iwrites_pc++;
1279 ios->iwrite_kb_pc += bytes >> 10;
1280 ios->iwrite_b_pc += bytes & 1023;
1281 } else {
1282 ios->ireads_pc++;
1283 ios->iread_kb_pc += bytes >> 10;
1284 ios->iread_b_pc += bytes & 1023;
1285 }
1286}
1287
1288static inline void account_pc_issue(struct blk_io_trace *t,
1289 struct per_cpu_info *pci, int rw)
1290{
1291 __account_pc_issue(&pci->io_stats, rw, t->bytes);
1292
1293 if (per_process_stats) {
1294 struct io_stats *ios = find_process_io_stats(t->pid);
1295
1296 __account_pc_issue(ios, rw, t->bytes);
1297 }
1298}
1299
1300static inline void __account_pc_requeue(struct io_stats *ios,
1301 struct blk_io_trace *t, int rw)
1302{
1303 if (rw) {
1304 ios->wrqueue_pc++;
1305 ios->iwrite_kb_pc -= t_kb(t);
1306 ios->iwrite_b_pc -= t_b(t);
1307 } else {
1308 ios->rrqueue_pc++;
1309 ios->iread_kb_pc -= t_kb(t);
1310 ios->iread_b_pc -= t_b(t);
1311 }
1312}
1313
1314static inline void account_pc_requeue(struct blk_io_trace *t,
1315 struct per_cpu_info *pci, int rw)
1316{
1317 __account_pc_requeue(&pci->io_stats, t, rw);
1318
1319 if (per_process_stats) {
1320 struct io_stats *ios = find_process_io_stats(t->pid);
1321
1322 __account_pc_requeue(ios, t, rw);
1323 }
1324}
1325
1326static inline void __account_pc_c(struct io_stats *ios, int rw)
1327{
1328 if (rw)
1329 ios->cwrites_pc++;
1330 else
1331 ios->creads_pc++;
1332}
1333
1334static inline void account_pc_c(struct blk_io_trace *t,
1335 struct per_cpu_info *pci, int rw)
1336{
1337 __account_pc_c(&pci->io_stats, rw);
1338
1339 if (per_process_stats) {
1340 struct io_stats *ios = find_process_io_stats(t->pid);
1341
1342 __account_pc_c(ios, rw);
1343 }
1344}
1345
1346static inline void __account_queue(struct io_stats *ios, struct blk_io_trace *t,
1347 int rw)
1348{
1349 if (rw) {
1350 ios->qwrites++;
1351 ios->qwrite_kb += t_kb(t);
1352 ios->qwrite_b += t_b(t);
1353 } else {
1354 ios->qreads++;
1355 ios->qread_kb += t_kb(t);
1356 ios->qread_b += t_b(t);
1357 }
1358}
1359
1360static inline void account_queue(struct blk_io_trace *t,
1361 struct per_cpu_info *pci, int rw)
1362{
1363 __account_queue(&pci->io_stats, t, rw);
1364
1365 if (per_process_stats) {
1366 struct io_stats *ios = find_process_io_stats(t->pid);
1367
1368 __account_queue(ios, t, rw);
1369 }
1370}
1371
1372static inline void __account_c(struct io_stats *ios, int rw, int bytes)
1373{
1374 if (rw) {
1375 ios->cwrites++;
1376 ios->cwrite_kb += bytes >> 10;
1377 ios->cwrite_b += bytes & 1023;
1378 } else {
1379 ios->creads++;
1380 ios->cread_kb += bytes >> 10;
1381 ios->cread_b += bytes & 1023;
1382 }
1383}
1384
1385static inline void account_c(struct blk_io_trace *t, struct per_cpu_info *pci,
1386 int rw, int bytes)
1387{
1388 __account_c(&pci->io_stats, rw, bytes);
1389
1390 if (per_process_stats) {
1391 struct io_stats *ios = find_process_io_stats(t->pid);
1392
1393 __account_c(ios, rw, bytes);
1394 }
1395}
1396
1397static inline void __account_issue(struct io_stats *ios, int rw,
1398 unsigned int bytes)
1399{
1400 if (rw) {
1401 ios->iwrites++;
1402 ios->iwrite_kb += bytes >> 10;
1403 ios->iwrite_b += bytes & 1023;
1404 } else {
1405 ios->ireads++;
1406 ios->iread_kb += bytes >> 10;
1407 ios->iread_b += bytes & 1023;
1408 }
1409}
1410
1411static inline void account_issue(struct blk_io_trace *t,
1412 struct per_cpu_info *pci, int rw)
1413{
1414 __account_issue(&pci->io_stats, rw, t->bytes);
1415
1416 if (per_process_stats) {
1417 struct io_stats *ios = find_process_io_stats(t->pid);
1418
1419 __account_issue(ios, rw, t->bytes);
1420 }
1421}
1422
1423static inline void __account_unplug(struct io_stats *ios, int timer)
1424{
1425 if (timer)
1426 ios->timer_unplugs++;
1427 else
1428 ios->io_unplugs++;
1429}
1430
1431static inline void account_unplug(struct blk_io_trace *t,
1432 struct per_cpu_info *pci, int timer)
1433{
1434 __account_unplug(&pci->io_stats, timer);
1435
1436 if (per_process_stats) {
1437 struct io_stats *ios = find_process_io_stats(t->pid);
1438
1439 __account_unplug(ios, timer);
1440 }
1441}
1442
1443static inline void __account_requeue(struct io_stats *ios,
1444 struct blk_io_trace *t, int rw)
1445{
1446 if (rw) {
1447 ios->wrqueue++;
1448 ios->iwrite_kb -= t_kb(t);
1449 ios->iwrite_b -= t_b(t);
1450 } else {
1451 ios->rrqueue++;
1452 ios->iread_kb -= t_kb(t);
1453 ios->iread_b -= t_b(t);
1454 }
1455}
1456
1457static inline void account_requeue(struct blk_io_trace *t,
1458 struct per_cpu_info *pci, int rw)
1459{
1460 __account_requeue(&pci->io_stats, t, rw);
1461
1462 if (per_process_stats) {
1463 struct io_stats *ios = find_process_io_stats(t->pid);
1464
1465 __account_requeue(ios, t, rw);
1466 }
1467}
1468
1469static void log_complete(struct per_dev_info *pdi, struct per_cpu_info *pci,
1470 struct blk_io_trace *t, char *act)
1471{
1472 process_fmt(act, pci, t, log_track_complete(pdi, t), 0, NULL);
1473}
1474
1475static void log_insert(struct per_dev_info *pdi, struct per_cpu_info *pci,
1476 struct blk_io_trace *t, char *act)
1477{
1478 process_fmt(act, pci, t, log_track_insert(pdi, t), 0, NULL);
1479}
1480
1481static void log_queue(struct per_cpu_info *pci, struct blk_io_trace *t,
1482 char *act)
1483{
1484 process_fmt(act, pci, t, -1, 0, NULL);
1485}
1486
1487static void log_issue(struct per_dev_info *pdi, struct per_cpu_info *pci,
1488 struct blk_io_trace *t, char *act)
1489{
1490 process_fmt(act, pci, t, log_track_issue(pdi, t), 0, NULL);
1491}
1492
1493static void log_merge(struct per_dev_info *pdi, struct per_cpu_info *pci,
1494 struct blk_io_trace *t, char *act)
1495{
1496 if (act[0] == 'F')
1497 log_track_frontmerge(pdi, t);
1498
1499 process_fmt(act, pci, t, -1ULL, 0, NULL);
1500}
1501
1502static void log_action(struct per_cpu_info *pci, struct blk_io_trace *t,
1503 char *act)
1504{
1505 process_fmt(act, pci, t, -1ULL, 0, NULL);
1506}
1507
1508static void log_generic(struct per_cpu_info *pci, struct blk_io_trace *t,
1509 char *act)
1510{
1511 process_fmt(act, pci, t, -1ULL, 0, NULL);
1512}
1513
1514static void log_unplug(struct per_cpu_info *pci, struct blk_io_trace *t,
1515 char *act)
1516{
1517 process_fmt(act, pci, t, -1ULL, 0, NULL);
1518}
1519
1520static void log_split(struct per_cpu_info *pci, struct blk_io_trace *t,
1521 char *act)
1522{
1523 process_fmt(act, pci, t, -1ULL, 0, NULL);
1524}
1525
1526static void log_pc(struct per_cpu_info *pci, struct blk_io_trace *t, char *act)
1527{
1528 unsigned char *buf = (unsigned char *) t + sizeof(*t);
1529
1530 process_fmt(act, pci, t, -1ULL, t->pdu_len, buf);
1531}
1532
1533static void dump_trace_pc(struct blk_io_trace *t, struct per_dev_info *pdi,
1534 struct per_cpu_info *pci)
1535{
1536 int w = (t->action & BLK_TC_ACT(BLK_TC_WRITE)) != 0;
1537 int act = t->action & 0xffff;
1538
1539 switch (act) {
1540 case __BLK_TA_QUEUE:
1541 log_generic(pci, t, "Q");
1542 account_pc_queue(t, pci, w);
1543 break;
1544 case __BLK_TA_GETRQ:
1545 log_generic(pci, t, "G");
1546 break;
1547 case __BLK_TA_SLEEPRQ:
1548 log_generic(pci, t, "S");
1549 break;
1550 case __BLK_TA_REQUEUE:
1551 /*
1552 * can happen if we miss traces, don't let it go
1553 * below zero
1554 */
1555 if (pdi->cur_depth[w])
1556 pdi->cur_depth[w]--;
1557 account_pc_requeue(t, pci, w);
1558 log_generic(pci, t, "R");
1559 break;
1560 case __BLK_TA_ISSUE:
1561 account_pc_issue(t, pci, w);
1562 pdi->cur_depth[w]++;
1563 if (pdi->cur_depth[w] > pdi->max_depth[w])
1564 pdi->max_depth[w] = pdi->cur_depth[w];
1565 log_pc(pci, t, "D");
1566 break;
1567 case __BLK_TA_COMPLETE:
1568 if (pdi->cur_depth[w])
1569 pdi->cur_depth[w]--;
1570 log_pc(pci, t, "C");
1571 account_pc_c(t, pci, w);
1572 break;
1573 case __BLK_TA_INSERT:
1574 log_pc(pci, t, "I");
1575 break;
1576 default:
1577 fprintf(stderr, "Bad pc action %x\n", act);
1578 break;
1579 }
1580}
1581
1582static void dump_trace_fs(struct blk_io_trace *t, struct per_dev_info *pdi,
1583 struct per_cpu_info *pci)
1584{
1585 int w = (t->action & BLK_TC_ACT(BLK_TC_WRITE)) != 0;
1586 int act = t->action & 0xffff;
1587
1588 switch (act) {
1589 case __BLK_TA_QUEUE:
1590 log_track_queue(pdi, t);
1591 account_queue(t, pci, w);
1592 log_queue(pci, t, "Q");
1593 break;
1594 case __BLK_TA_INSERT:
1595 log_insert(pdi, pci, t, "I");
1596 break;
1597 case __BLK_TA_BACKMERGE:
1598 account_m(t, pci, w);
1599 log_merge(pdi, pci, t, "M");
1600 break;
1601 case __BLK_TA_FRONTMERGE:
1602 account_m(t, pci, w);
1603 log_merge(pdi, pci, t, "F");
1604 break;
1605 case __BLK_TA_GETRQ:
1606 log_track_getrq(pdi, t);
1607 log_generic(pci, t, "G");
1608 break;
1609 case __BLK_TA_SLEEPRQ:
1610 log_generic(pci, t, "S");
1611 break;
1612 case __BLK_TA_REQUEUE:
1613 /*
1614 * can happen if we miss traces, don't let it go
1615 * below zero
1616 */
1617 if (pdi->cur_depth[w])
1618 pdi->cur_depth[w]--;
1619 account_requeue(t, pci, w);
1620 log_queue(pci, t, "R");
1621 break;
1622 case __BLK_TA_ISSUE:
1623 account_issue(t, pci, w);
1624 pdi->cur_depth[w]++;
1625 if (pdi->cur_depth[w] > pdi->max_depth[w])
1626 pdi->max_depth[w] = pdi->cur_depth[w];
1627 log_issue(pdi, pci, t, "D");
1628 break;
1629 case __BLK_TA_COMPLETE:
1630 if (pdi->cur_depth[w])
1631 pdi->cur_depth[w]--;
1632 account_c(t, pci, w, t->bytes);
1633 log_complete(pdi, pci, t, "C");
1634 break;
1635 case __BLK_TA_PLUG:
1636 log_action(pci, t, "P");
1637 break;
1638 case __BLK_TA_UNPLUG_IO:
1639 account_unplug(t, pci, 0);
1640 log_unplug(pci, t, "U");
1641 break;
1642 case __BLK_TA_UNPLUG_TIMER:
1643 account_unplug(t, pci, 1);
1644 log_unplug(pci, t, "UT");
1645 break;
1646 case __BLK_TA_SPLIT:
1647 log_split(pci, t, "X");
1648 break;
1649 case __BLK_TA_BOUNCE:
1650 log_generic(pci, t, "B");
1651 break;
1652 case __BLK_TA_REMAP:
1653 log_generic(pci, t, "A");
1654 break;
1655 case __BLK_TA_DRV_DATA:
1656 have_drv_data = 1;
1657 /* dump to binary file only */
1658 break;
1659 default:
1660 fprintf(stderr, "Bad fs action %x\n", t->action);
1661 break;
1662 }
1663}
1664
1665static void dump_trace(struct blk_io_trace *t, struct per_cpu_info *pci,
1666 struct per_dev_info *pdi)
1667{
1668 if (text_output) {
1669 if (t->action == BLK_TN_MESSAGE)
1670 handle_notify(t);
1671 else if (t->action & BLK_TC_ACT(BLK_TC_PC))
1672 dump_trace_pc(t, pdi, pci);
1673 else
1674 dump_trace_fs(t, pdi, pci);
1675 }
1676
1677 if (!pdi->events)
1678 pdi->first_reported_time = t->time;
1679
1680 pdi->events++;
1681
1682 if (bin_output_msgs ||
1683 !(t->action & BLK_TC_ACT(BLK_TC_NOTIFY) &&
1684 t->action == BLK_TN_MESSAGE))
1685 output_binary(t, sizeof(*t) + t->pdu_len);
1686}
1687
1688/*
1689 * print in a proper way, not too small and not too big. if more than
1690 * 1000,000K, turn into M and so on
1691 */
1692static char *size_cnv(char *dst, unsigned long long num, int in_kb)
1693{
1694 char suff[] = { '\0', 'K', 'M', 'G', 'P' };
1695 unsigned int i = 0;
1696
1697 if (in_kb)
1698 i++;
1699
1700 while (num > 1000 * 1000ULL && (i < sizeof(suff) - 1)) {
1701 i++;
1702 num /= 1000;
1703 }
1704
1705 sprintf(dst, "%'8Lu%c", num, suff[i]);
1706 return dst;
1707}
1708
1709static void dump_io_stats(struct per_dev_info *pdi, struct io_stats *ios,
1710 char *msg)
1711{
1712 static char x[256], y[256];
1713
1714 fprintf(ofp, "%s\n", msg);
1715
1716 fprintf(ofp, " Reads Queued: %s, %siB\t",
1717 size_cnv(x, ios->qreads, 0),
1718 size_cnv(y, ios->qread_kb + (ios->qread_b>>10), 1));
1719 fprintf(ofp, " Writes Queued: %s, %siB\n",
1720 size_cnv(x, ios->qwrites, 0),
1721 size_cnv(y, ios->qwrite_kb + (ios->qwrite_b>>10), 1));
1722 fprintf(ofp, " Read Dispatches: %s, %siB\t",
1723 size_cnv(x, ios->ireads, 0),
1724 size_cnv(y, ios->iread_kb + (ios->iread_b>>10), 1));
1725 fprintf(ofp, " Write Dispatches: %s, %siB\n",
1726 size_cnv(x, ios->iwrites, 0),
1727 size_cnv(y, ios->iwrite_kb + (ios->iwrite_b>>10), 1));
1728 fprintf(ofp, " Reads Requeued: %s\t\t", size_cnv(x, ios->rrqueue, 0));
1729 fprintf(ofp, " Writes Requeued: %s\n", size_cnv(x, ios->wrqueue, 0));
1730 fprintf(ofp, " Reads Completed: %s, %siB\t",
1731 size_cnv(x, ios->creads, 0),
1732 size_cnv(y, ios->cread_kb + (ios->cread_b>>10), 1));
1733 fprintf(ofp, " Writes Completed: %s, %siB\n",
1734 size_cnv(x, ios->cwrites, 0),
1735 size_cnv(y, ios->cwrite_kb + (ios->cwrite_b>>10), 1));
1736 fprintf(ofp, " Read Merges: %s, %siB\t",
1737 size_cnv(x, ios->mreads, 0),
1738 size_cnv(y, ios->mread_kb + (ios->mread_b>>10), 1));
1739 fprintf(ofp, " Write Merges: %s, %siB\n",
1740 size_cnv(x, ios->mwrites, 0),
1741 size_cnv(y, ios->mwrite_kb + (ios->mwrite_b>>10), 1));
1742 if (pdi) {
1743 fprintf(ofp, " Read depth: %'8u%8c\t", pdi->max_depth[0], ' ');
1744 fprintf(ofp, " Write depth: %'8u\n", pdi->max_depth[1]);
1745 }
1746 if (ios->qreads_pc || ios->qwrites_pc || ios->ireads_pc || ios->iwrites_pc ||
1747 ios->rrqueue_pc || ios->wrqueue_pc || ios->creads_pc || ios->cwrites_pc) {
1748 fprintf(ofp, " PC Reads Queued: %s, %siB\t",
1749 size_cnv(x, ios->qreads_pc, 0),
1750 size_cnv(y,
1751 ios->qread_kb_pc + (ios->qread_b_pc>>10), 1));
1752 fprintf(ofp, " PC Writes Queued: %s, %siB\n",
1753 size_cnv(x, ios->qwrites_pc, 0),
1754 size_cnv(y,
1755 ios->qwrite_kb_pc + (ios->qwrite_b_pc>>10), 1));
1756 fprintf(ofp, " PC Read Disp.: %s, %siB\t",
1757 size_cnv(x, ios->ireads_pc, 0),
1758 size_cnv(y,
1759 ios->iread_kb_pc + (ios->iread_b_pc>>10), 1));
1760 fprintf(ofp, " PC Write Disp.: %s, %siB\n",
1761 size_cnv(x, ios->iwrites_pc, 0),
1762 size_cnv(y,
1763 ios->iwrite_kb_pc + (ios->iwrite_b_pc>>10),
1764 1));
1765 fprintf(ofp, " PC Reads Req.: %s\t\t", size_cnv(x, ios->rrqueue_pc, 0));
1766 fprintf(ofp, " PC Writes Req.: %s\n", size_cnv(x, ios->wrqueue_pc, 0));
1767 fprintf(ofp, " PC Reads Compl.: %s\t\t", size_cnv(x, ios->creads_pc, 0));
1768 fprintf(ofp, " PC Writes Compl.: %s\n", size_cnv(x, ios->cwrites_pc, 0));
1769 }
1770 fprintf(ofp, " IO unplugs: %'8lu%8c\t", ios->io_unplugs, ' ');
1771 fprintf(ofp, " Timer unplugs: %'8lu\n", ios->timer_unplugs);
1772}
1773
1774static void dump_wait_stats(struct per_process_info *ppi)
1775{
1776 unsigned long rawait = ppi->longest_allocation_wait[0] / 1000;
1777 unsigned long rdwait = ppi->longest_dispatch_wait[0] / 1000;
1778 unsigned long rcwait = ppi->longest_completion_wait[0] / 1000;
1779 unsigned long wawait = ppi->longest_allocation_wait[1] / 1000;
1780 unsigned long wdwait = ppi->longest_dispatch_wait[1] / 1000;
1781 unsigned long wcwait = ppi->longest_completion_wait[1] / 1000;
1782
1783 fprintf(ofp, " Allocation wait: %'8lu%8c\t", rawait, ' ');
1784 fprintf(ofp, " Allocation wait: %'8lu\n", wawait);
1785 fprintf(ofp, " Dispatch wait: %'8lu%8c\t", rdwait, ' ');
1786 fprintf(ofp, " Dispatch wait: %'8lu\n", wdwait);
1787 fprintf(ofp, " Completion wait: %'8lu%8c\t", rcwait, ' ');
1788 fprintf(ofp, " Completion wait: %'8lu\n", wcwait);
1789}
1790
1791static int ppi_name_compare(const void *p1, const void *p2)
1792{
1793 struct per_process_info *ppi1 = *((struct per_process_info **) p1);
1794 struct per_process_info *ppi2 = *((struct per_process_info **) p2);
1795 int res;
1796
1797 res = strverscmp(ppi1->ppm->comm, ppi2->ppm->comm);
1798 if (!res)
1799 res = ppi1->ppm->pid > ppi2->ppm->pid;
1800
1801 return res;
1802}
1803
1804static int ppi_event_compare(const void *p1, const void *p2)
1805{
1806 struct per_process_info *ppi1 = *((struct per_process_info **) p1);
1807 struct per_process_info *ppi2 = *((struct per_process_info **) p2);
1808 struct io_stats *ios1 = &ppi1->io_stats;
1809 struct io_stats *ios2 = &ppi2->io_stats;
1810 unsigned long io1, io2;
1811 unsigned long long kb1,kb2;
1812 int sort_by_kb = 1;
1813
1814 io1 = io2 = 0;
1815 kb1 = kb2 = 0;
1816
1817 switch (per_process_stats_event) {
1818 case SORT_PROG_EVENT_QKB: /* KB: Queued read and write */
1819 kb1 = ios1->qwrite_kb + (ios1->qwrite_b>>10) +
1820 ios1->qread_kb + (ios1->qread_b>>10);
1821 kb2 = ios2->qwrite_kb + (ios2->qwrite_b>>10) +
1822 ios2->qread_kb + (ios2->qread_b>>10);
1823 break;
1824 case SORT_PROG_EVENT_RKB: /* KB: Queued Read */
1825 kb1 = ios1->qread_kb + (ios1->qread_b>>10);
1826 kb2 = ios2->qread_kb + (ios2->qread_b>>10);
1827 break;
1828 case SORT_PROG_EVENT_WKB: /* KB: Queued Write */
1829 kb1 = ios1->qwrite_kb + (ios1->qwrite_b>>10);
1830 kb2 = ios2->qwrite_kb + (ios2->qwrite_b>>10);
1831 break;
1832 case SORT_PROG_EVENT_CKB: /* KB: Complete */
1833 kb1 = ios1->cwrite_kb + (ios1->cwrite_b>>10) +
1834 ios1->cread_kb + (ios1->cread_b>>10);
1835 kb2 = ios2->cwrite_kb + (ios2->cwrite_b>>10) +
1836 ios2->cread_kb + (ios2->cread_b>>10);
1837 break;
1838 case SORT_PROG_EVENT_QIO: /* IO: Queued read and write */
1839 sort_by_kb = 0;
1840 io1 = ios1->qreads + ios1->qwrites;
1841 io2 = ios2->qreads + ios2->qwrites;
1842 break;
1843 case SORT_PROG_EVENT_RIO: /* IO: Queued Read */
1844 sort_by_kb = 0;
1845 io1 = ios1->qreads;
1846 io2 = ios2->qreads;
1847 break;
1848 case SORT_PROG_EVENT_WIO: /* IO: Queued Write */
1849 sort_by_kb = 0;
1850 io1 = ios1->qwrites;
1851 io2 = ios2->qwrites;
1852 break;
1853 case SORT_PROG_EVENT_CIO: /* IO: Complete */
1854 sort_by_kb = 0;
1855 io1 = ios1->creads + ios1->cwrites;
1856 io2 = ios2->creads + ios2->cwrites;
1857 break;
1858 }
1859
1860
1861 /* compare kb */
1862 if (sort_by_kb) {
1863 if (kb1 > kb2)
1864 return 1;
1865 else if (kb1 == kb2)
1866 return 0;
1867 return -1;
1868 }
1869
1870 /* compare io */
1871 if (io1 > io2)
1872 return 1;
1873 else if (io1 == io2)
1874 return 0;
1875 return -1;
1876}
1877
1878static int ppi_compare(const void *p1, const void *p2)
1879{
1880 if (per_process_stats_event == SORT_PROG_EVENT_N)
1881 return ppi_name_compare(p1, p2);
1882
1883 return ppi_event_compare(p1, p2);
1884}
1885
1886static void sort_process_list(void)
1887{
1888 struct per_process_info **ppis;
1889 struct per_process_info *ppi;
1890 int i = 0;
1891
1892 ppis = malloc(ppi_list_entries * sizeof(struct per_process_info *));
1893
1894 ppi = ppi_list;
1895 while (ppi) {
1896 ppis[i++] = ppi;
1897 ppi = ppi->list_next;
1898 }
1899
1900 qsort(ppis, ppi_list_entries, sizeof(ppi), ppi_compare);
1901
1902 i = ppi_list_entries - 1;
1903 ppi_list = NULL;
1904 while (i >= 0) {
1905 ppi = ppis[i];
1906
1907 ppi->list_next = ppi_list;
1908 ppi_list = ppi;
1909 i--;
1910 }
1911
1912 free(ppis);
1913}
1914
1915static void show_process_stats(void)
1916{
1917 struct per_process_info *ppi;
1918
1919 sort_process_list();
1920
1921 ppi = ppi_list;
1922 while (ppi) {
1923 struct process_pid_map *ppm = ppi->ppm;
1924 char name[64];
1925
1926 if (ppi->more_than_one)
1927 sprintf(name, "%s (%u, ...)", ppm->comm, ppm->pid);
1928 else
1929 sprintf(name, "%s (%u)", ppm->comm, ppm->pid);
1930
1931 dump_io_stats(NULL, &ppi->io_stats, name);
1932 dump_wait_stats(ppi);
1933 ppi = ppi->list_next;
1934 }
1935
1936 fprintf(ofp, "\n");
1937}
1938
1939static void show_device_and_cpu_stats(void)
1940{
1941 struct per_dev_info *pdi;
1942 struct per_cpu_info *pci;
1943 struct io_stats total, *ios;
1944 unsigned long long rrate, wrate, msec;
1945 int i, j, pci_events;
1946 char line[3 + 8/*cpu*/ + 2 + 32/*dev*/ + 3];
1947 char name[32];
1948 double ratio;
1949
1950 for (pdi = devices, i = 0; i < ndevices; i++, pdi++) {
1951
1952 memset(&total, 0, sizeof(total));
1953 pci_events = 0;
1954
1955 if (i > 0)
1956 fprintf(ofp, "\n");
1957
1958 for (pci = pdi->cpus, j = 0; j < pdi->ncpus; j++, pci++) {
1959 if (!pci->nelems)
1960 continue;
1961
1962 ios = &pci->io_stats;
1963 total.qreads += ios->qreads;
1964 total.qwrites += ios->qwrites;
1965 total.creads += ios->creads;
1966 total.cwrites += ios->cwrites;
1967 total.mreads += ios->mreads;
1968 total.mwrites += ios->mwrites;
1969 total.ireads += ios->ireads;
1970 total.iwrites += ios->iwrites;
1971 total.rrqueue += ios->rrqueue;
1972 total.wrqueue += ios->wrqueue;
1973 total.qread_kb += ios->qread_kb;
1974 total.qwrite_kb += ios->qwrite_kb;
1975 total.cread_kb += ios->cread_kb;
1976 total.cwrite_kb += ios->cwrite_kb;
1977 total.iread_kb += ios->iread_kb;
1978 total.iwrite_kb += ios->iwrite_kb;
1979 total.mread_kb += ios->mread_kb;
1980 total.mwrite_kb += ios->mwrite_kb;
1981 total.qread_b += ios->qread_b;
1982 total.qwrite_b += ios->qwrite_b;
1983 total.cread_b += ios->cread_b;
1984 total.cwrite_b += ios->cwrite_b;
1985 total.iread_b += ios->iread_b;
1986 total.iwrite_b += ios->iwrite_b;
1987 total.mread_b += ios->mread_b;
1988 total.mwrite_b += ios->mwrite_b;
1989
1990 total.qreads_pc += ios->qreads_pc;
1991 total.qwrites_pc += ios->qwrites_pc;
1992 total.creads_pc += ios->creads_pc;
1993 total.cwrites_pc += ios->cwrites_pc;
1994 total.ireads_pc += ios->ireads_pc;
1995 total.iwrites_pc += ios->iwrites_pc;
1996 total.rrqueue_pc += ios->rrqueue_pc;
1997 total.wrqueue_pc += ios->wrqueue_pc;
1998 total.qread_kb_pc += ios->qread_kb_pc;
1999 total.qwrite_kb_pc += ios->qwrite_kb_pc;
2000 total.iread_kb_pc += ios->iread_kb_pc;
2001 total.iwrite_kb_pc += ios->iwrite_kb_pc;
2002 total.qread_b_pc += ios->qread_b_pc;
2003 total.qwrite_b_pc += ios->qwrite_b_pc;
2004 total.iread_b_pc += ios->iread_b_pc;
2005 total.iwrite_b_pc += ios->iwrite_b_pc;
2006
2007 total.timer_unplugs += ios->timer_unplugs;
2008 total.io_unplugs += ios->io_unplugs;
2009
2010 snprintf(line, sizeof(line) - 1, "CPU%d (%s):",
2011 j, get_dev_name(pdi, name, sizeof(name)));
2012 dump_io_stats(pdi, ios, line);
2013 pci_events++;
2014 }
2015
2016 if (pci_events > 1) {
2017 fprintf(ofp, "\n");
2018 snprintf(line, sizeof(line) - 1, "Total (%s):",
2019 get_dev_name(pdi, name, sizeof(name)));
2020 dump_io_stats(NULL, &total, line);
2021 }
2022
2023 wrate = rrate = 0;
2024 msec = (pdi->last_reported_time - pdi->first_reported_time) / 1000000;
2025 if (msec) {
2026 rrate = ((1000 * total.cread_kb) + total.cread_b) /
2027 msec;
2028 wrate = ((1000 * total.cwrite_kb) + total.cwrite_b) /
2029 msec;
2030 }
2031
2032 fprintf(ofp, "\nThroughput (R/W): %'LuKiB/s / %'LuKiB/s\n",
2033 rrate, wrate);
2034 fprintf(ofp, "Events (%s): %'Lu entries\n",
2035 get_dev_name(pdi, line, sizeof(line)), pdi->events);
2036
2037 collect_pdi_skips(pdi);
2038 if (!pdi->skips && !pdi->events)
2039 ratio = 0.0;
2040 else
2041 ratio = 100.0 * ((double)pdi->seq_skips /
2042 (double)(pdi->events + pdi->seq_skips));
2043 fprintf(ofp, "Skips: %'lu forward (%'llu - %5.1lf%%)\n",
2044 pdi->skips, pdi->seq_skips, ratio);
2045 }
2046}
2047
2048static void correct_abs_start_time(void)
2049{
2050 long delta = genesis_time - start_timestamp;
2051
2052 abs_start_time.tv_sec += SECONDS(delta);
2053 abs_start_time.tv_nsec += NANO_SECONDS(delta);
2054 if (abs_start_time.tv_nsec < 0) {
2055 abs_start_time.tv_nsec += 1000000000;
2056 abs_start_time.tv_sec -= 1;
2057 } else
2058 if (abs_start_time.tv_nsec > 1000000000) {
2059 abs_start_time.tv_nsec -= 1000000000;
2060 abs_start_time.tv_sec += 1;
2061 }
2062}
2063
2064static void find_genesis(void)
2065{
2066 struct trace *t = trace_list;
2067
2068 genesis_time = -1ULL;
2069 while (t != NULL) {
2070 if (t->bit->time < genesis_time)
2071 genesis_time = t->bit->time;
2072
2073 t = t->next;
2074 }
2075
2076 /* The time stamp record will usually be the first
2077 * record in the trace, but not always.
2078 */
2079 if (start_timestamp
2080 && start_timestamp != genesis_time) {
2081 correct_abs_start_time();
2082 }
2083}
2084
2085static inline int check_stopwatch(struct blk_io_trace *bit)
2086{
2087 if (bit->time < stopwatch_end &&
2088 bit->time >= stopwatch_start)
2089 return 0;
2090
2091 return 1;
2092}
2093
2094/*
2095 * return youngest entry read
2096 */
2097static int sort_entries(unsigned long long *youngest)
2098{
2099 struct per_dev_info *pdi = NULL;
2100 struct per_cpu_info *pci = NULL;
2101 struct trace *t;
2102
2103 if (!genesis_time)
2104 find_genesis();
2105
2106 *youngest = 0;
2107 while ((t = trace_list) != NULL) {
2108 struct blk_io_trace *bit = t->bit;
2109
2110 trace_list = t->next;
2111
2112 bit->time -= genesis_time;
2113
2114 if (bit->time < *youngest || !*youngest)
2115 *youngest = bit->time;
2116
2117 if (!pdi || pdi->dev != bit->device) {
2118 pdi = get_dev_info(bit->device);
2119 pci = NULL;
2120 }
2121
2122 if (!pci || pci->cpu != bit->cpu)
2123 pci = get_cpu_info(pdi, bit->cpu);
2124
2125 if (bit->sequence < pci->smallest_seq_read)
2126 pci->smallest_seq_read = bit->sequence;
2127
2128 if (check_stopwatch(bit)) {
2129 bit_free(bit);
2130 t_free(t);
2131 continue;
2132 }
2133
2134 if (trace_rb_insert_sort(t))
2135 return -1;
2136 }
2137
2138 return 0;
2139}
2140
2141/*
2142 * to continue, we must have traces from all online cpus in the tree
2143 */
2144static int check_cpu_map(struct per_dev_info *pdi)
2145{
2146 unsigned long *cpu_map;
2147 struct rb_node *n;
2148 struct trace *__t;
2149 unsigned int i;
2150 int ret, cpu;
2151
2152 /*
2153 * create a map of the cpus we have traces for
2154 */
2155 cpu_map = malloc(pdi->cpu_map_max / sizeof(long));
2156 memset(cpu_map, 0, sizeof(*cpu_map));
2157 n = rb_first(&rb_sort_root);
2158 while (n) {
2159 __t = rb_entry(n, struct trace, rb_node);
2160 cpu = __t->bit->cpu;
2161
2162 cpu_map[CPU_IDX(cpu)] |= (1UL << CPU_BIT(cpu));
2163 n = rb_next(n);
2164 }
2165
2166 /*
2167 * we can't continue if pdi->cpu_map has entries set that we don't
2168 * have in the sort rbtree. the opposite is not a problem, though
2169 */
2170 ret = 0;
2171 for (i = 0; i < pdi->cpu_map_max / CPUS_PER_LONG; i++) {
2172 if (pdi->cpu_map[i] & ~(cpu_map[i])) {
2173 ret = 1;
2174 break;
2175 }
2176 }
2177
2178 free(cpu_map);
2179 return ret;
2180}
2181
2182static int check_sequence(struct per_dev_info *pdi, struct trace *t, int force)
2183{
2184 struct blk_io_trace *bit = t->bit;
2185 unsigned long expected_sequence;
2186 struct per_cpu_info *pci;
2187 struct trace *__t;
2188
2189 pci = get_cpu_info(pdi, bit->cpu);
2190 expected_sequence = pci->last_sequence + 1;
2191
2192 if (!expected_sequence) {
2193 /*
2194 * 1 should be the first entry, just allow it
2195 */
2196 if (bit->sequence == 1)
2197 return 0;
2198 if (bit->sequence == pci->smallest_seq_read)
2199 return 0;
2200
2201 return check_cpu_map(pdi);
2202 }
2203
2204 if (bit->sequence == expected_sequence)
2205 return 0;
2206
2207 /*
2208 * we may not have seen that sequence yet. if we are not doing
2209 * the final run, break and wait for more entries.
2210 */
2211 if (expected_sequence < pci->smallest_seq_read) {
2212 __t = trace_rb_find_last(pdi, pci, expected_sequence);
2213 if (!__t)
2214 goto skip;
2215
2216 __put_trace_last(pdi, __t);
2217 return 0;
2218 } else if (!force) {
2219 return 1;
2220 } else {
2221skip:
2222 if (check_current_skips(pci, bit->sequence))
2223 return 0;
2224
2225 if (expected_sequence < bit->sequence)
2226 insert_skip(pci, expected_sequence, bit->sequence - 1);
2227 return 0;
2228 }
2229}
2230
2231static void show_entries_rb(int force)
2232{
2233 struct per_dev_info *pdi = NULL;
2234 struct per_cpu_info *pci = NULL;
2235 struct blk_io_trace *bit;
2236 struct rb_node *n;
2237 struct trace *t;
2238
2239 while ((n = rb_first(&rb_sort_root)) != NULL) {
2240 if (is_done() && !force && !pipeline)
2241 break;
2242
2243 t = rb_entry(n, struct trace, rb_node);
2244 bit = t->bit;
2245
2246 if (read_sequence - t->read_sequence < 1 && !force)
2247 break;
2248
2249 if (!pdi || pdi->dev != bit->device) {
2250 pdi = get_dev_info(bit->device);
2251 pci = NULL;
2252 }
2253
2254 if (!pdi) {
2255 fprintf(stderr, "Unknown device ID? (%d,%d)\n",
2256 MAJOR(bit->device), MINOR(bit->device));
2257 break;
2258 }
2259
2260 if (!(bit->action == BLK_TN_MESSAGE) &&
2261 check_sequence(pdi, t, force))
2262 break;
2263
2264 if (!force && bit->time > last_allowed_time)
2265 break;
2266
2267 check_time(pdi, bit);
2268
2269 if (!pci || pci->cpu != bit->cpu)
2270 pci = get_cpu_info(pdi, bit->cpu);
2271
2272 if (!(bit->action == BLK_TN_MESSAGE))
2273 pci->last_sequence = bit->sequence;
2274
2275 pci->nelems++;
2276
2277 if (bit->action & (act_mask << BLK_TC_SHIFT))
2278 dump_trace(bit, pci, pdi);
2279
2280 put_trace(pdi, t);
2281 }
2282}
2283
2284static int read_data(int fd, void *buffer, int bytes, int block, int *fdblock)
2285{
2286 int ret, bytes_left, fl;
2287 void *p;
2288
2289 if (block != *fdblock) {
2290 fl = fcntl(fd, F_GETFL);
2291
2292 if (!block) {
2293 *fdblock = 0;
2294 fcntl(fd, F_SETFL, fl | O_NONBLOCK);
2295 } else {
2296 *fdblock = 1;
2297 fcntl(fd, F_SETFL, fl & ~O_NONBLOCK);
2298 }
2299 }
2300
2301 bytes_left = bytes;
2302 p = buffer;
2303 while (bytes_left > 0) {
2304 ret = read(fd, p, bytes_left);
2305 if (!ret)
2306 return 1;
2307 else if (ret < 0) {
2308 if (errno != EAGAIN) {
2309 perror("read");
2310 return -1;
2311 }
2312
2313 /*
2314 * never do partial reads. we can return if we
2315 * didn't read anything and we should not block,
2316 * otherwise wait for data
2317 */
2318 if ((bytes_left == bytes) && !block)
2319 return 1;
2320
2321 usleep(10);
2322 continue;
2323 } else {
2324 p += ret;
2325 bytes_left -= ret;
2326 }
2327 }
2328
2329 return 0;
2330}
2331
2332static inline __u16 get_pdulen(struct blk_io_trace *bit)
2333{
2334 if (data_is_native)
2335 return bit->pdu_len;
2336
2337 return __bswap_16(bit->pdu_len);
2338}
2339
2340static inline __u32 get_magic(struct blk_io_trace *bit)
2341{
2342 if (data_is_native)
2343 return bit->magic;
2344
2345 return __bswap_32(bit->magic);
2346}
2347
2348static int read_events(int fd, int always_block, int *fdblock)
2349{
2350 struct per_dev_info *pdi = NULL;
2351 unsigned int events = 0;
2352
2353 while (!is_done() && events < rb_batch) {
2354 struct blk_io_trace *bit;
2355 struct trace *t;
2356 int pdu_len, should_block, ret;
2357 __u32 magic;
2358
2359 bit = bit_alloc();
2360
2361 should_block = !events || always_block;
2362
2363 ret = read_data(fd, bit, sizeof(*bit), should_block, fdblock);
2364 if (ret) {
2365 bit_free(bit);
2366 if (!events && ret < 0)
2367 events = ret;
2368 break;
2369 }
2370
2371 /*
2372 * look at first trace to check whether we need to convert
2373 * data in the future
2374 */
2375 if (data_is_native == -1 && check_data_endianness(bit->magic))
2376 break;
2377
2378 magic = get_magic(bit);
2379 if ((magic & 0xffffff00) != BLK_IO_TRACE_MAGIC) {
2380 fprintf(stderr, "Bad magic %x\n", magic);
2381 break;
2382 }
2383
2384 pdu_len = get_pdulen(bit);
2385 if (pdu_len) {
2386 void *ptr = realloc(bit, sizeof(*bit) + pdu_len);
2387
2388 if (read_data(fd, ptr + sizeof(*bit), pdu_len, 1, fdblock)) {
2389 bit_free(ptr);
2390 break;
2391 }
2392
2393 bit = ptr;
2394 }
2395
2396 trace_to_cpu(bit);
2397
2398 if (verify_trace(bit)) {
2399 bit_free(bit);
2400 continue;
2401 }
2402
2403 /*
2404 * not a real trace, so grab and handle it here
2405 */
2406 if (bit->action & BLK_TC_ACT(BLK_TC_NOTIFY) && bit->action != BLK_TN_MESSAGE) {
2407 handle_notify(bit);
2408 output_binary(bit, sizeof(*bit) + bit->pdu_len);
2409 continue;
2410 }
2411
2412 t = t_alloc();
2413 memset(t, 0, sizeof(*t));
2414 t->bit = bit;
2415 t->read_sequence = read_sequence;
2416
2417 t->next = trace_list;
2418 trace_list = t;
2419
2420 if (!pdi || pdi->dev != bit->device)
2421 pdi = get_dev_info(bit->device);
2422
2423 if (bit->time > pdi->last_read_time)
2424 pdi->last_read_time = bit->time;
2425
2426 events++;
2427 }
2428
2429 return events;
2430}
2431
2432/*
2433 * Managing input streams
2434 */
2435
2436struct ms_stream {
2437 struct ms_stream *next;
2438 struct trace *first, *last;
2439 struct per_dev_info *pdi;
2440 unsigned int cpu;
2441};
2442
2443#define MS_HASH(d, c) ((MAJOR(d) & 0xff) ^ (MINOR(d) & 0xff) ^ (cpu & 0xff))
2444
2445struct ms_stream *ms_head;
2446struct ms_stream *ms_hash[256];
2447
2448static void ms_sort(struct ms_stream *msp);
2449static int ms_prime(struct ms_stream *msp);
2450
2451static inline struct trace *ms_peek(struct ms_stream *msp)
2452{
2453 return (msp == NULL) ? NULL : msp->first;
2454}
2455
2456static inline __u64 ms_peek_time(struct ms_stream *msp)
2457{
2458 return ms_peek(msp)->bit->time;
2459}
2460
2461static inline void ms_resort(struct ms_stream *msp)
2462{
2463 if (msp->next && ms_peek_time(msp) > ms_peek_time(msp->next)) {
2464 ms_head = msp->next;
2465 msp->next = NULL;
2466 ms_sort(msp);
2467 }
2468}
2469
2470static inline void ms_deq(struct ms_stream *msp)
2471{
2472 msp->first = msp->first->next;
2473 if (!msp->first) {
2474 msp->last = NULL;
2475 if (!ms_prime(msp)) {
2476 ms_head = msp->next;
2477 msp->next = NULL;
2478 return;
2479 }
2480 }
2481
2482 ms_resort(msp);
2483}
2484
2485static void ms_sort(struct ms_stream *msp)
2486{
2487 __u64 msp_t = ms_peek_time(msp);
2488 struct ms_stream *this_msp = ms_head;
2489
2490 if (this_msp == NULL)
2491 ms_head = msp;
2492 else if (msp_t < ms_peek_time(this_msp)) {
2493 msp->next = this_msp;
2494 ms_head = msp;
2495 }
2496 else {
2497 while (this_msp->next && ms_peek_time(this_msp->next) < msp_t)
2498 this_msp = this_msp->next;
2499
2500 msp->next = this_msp->next;
2501 this_msp->next = msp;
2502 }
2503}
2504
2505static int ms_prime(struct ms_stream *msp)
2506{
2507 __u32 magic;
2508 unsigned int i;
2509 struct trace *t;
2510 struct per_dev_info *pdi = msp->pdi;
2511 struct per_cpu_info *pci = get_cpu_info(pdi, msp->cpu);
2512 struct blk_io_trace *bit = NULL;
2513 int ret, pdu_len, ndone = 0;
2514
2515 for (i = 0; !is_done() && pci->fd >= 0 && i < rb_batch; i++) {
2516 bit = bit_alloc();
2517 ret = read_data(pci->fd, bit, sizeof(*bit), 1, &pci->fdblock);
2518 if (ret)
2519 goto err;
2520
2521 if (data_is_native == -1 && check_data_endianness(bit->magic))
2522 goto err;
2523
2524 magic = get_magic(bit);
2525 if ((magic & 0xffffff00) != BLK_IO_TRACE_MAGIC) {
2526 fprintf(stderr, "Bad magic %x\n", magic);
2527 goto err;
2528
2529 }
2530
2531 pdu_len = get_pdulen(bit);
2532 if (pdu_len) {
2533 void *ptr = realloc(bit, sizeof(*bit) + pdu_len);
2534 ret = read_data(pci->fd, ptr + sizeof(*bit), pdu_len,
2535 1, &pci->fdblock);
2536 if (ret) {
2537 free(ptr);
2538 bit = NULL;
2539 goto err;
2540 }
2541
2542 bit = ptr;
2543 }
2544
2545 trace_to_cpu(bit);
2546 if (verify_trace(bit))
2547 goto err;
2548
2549 if (bit->cpu != pci->cpu) {
2550 fprintf(stderr, "cpu %d trace info has error cpu %d\n",
2551 pci->cpu, bit->cpu);
2552 continue;
2553 }
2554
2555 if (bit->action & BLK_TC_ACT(BLK_TC_NOTIFY) && bit->action != BLK_TN_MESSAGE) {
2556 handle_notify(bit);
2557 output_binary(bit, sizeof(*bit) + bit->pdu_len);
2558 bit_free(bit);
2559
2560 i -= 1;
2561 continue;
2562 }
2563
2564 if (bit->time > pdi->last_read_time)
2565 pdi->last_read_time = bit->time;
2566
2567 t = t_alloc();
2568 memset(t, 0, sizeof(*t));
2569 t->bit = bit;
2570
2571 if (msp->first == NULL)
2572 msp->first = msp->last = t;
2573 else {
2574 msp->last->next = t;
2575 msp->last = t;
2576 }
2577
2578 ndone++;
2579 }
2580
2581 return ndone;
2582
2583err:
2584 if (bit) bit_free(bit);
2585
2586 cpu_mark_offline(pdi, pci->cpu);
2587 close(pci->fd);
2588 pci->fd = -1;
2589
2590 return ndone;
2591}
2592
2593static struct ms_stream *ms_alloc(struct per_dev_info *pdi, int cpu)
2594{
2595 struct ms_stream *msp = malloc(sizeof(*msp));
2596
2597 msp->next = NULL;
2598 msp->first = msp->last = NULL;
2599 msp->pdi = pdi;
2600 msp->cpu = cpu;
2601
2602 if (ms_prime(msp))
2603 ms_sort(msp);
2604
2605 return msp;
2606}
2607
2608static int setup_file(struct per_dev_info *pdi, int cpu)
2609{
2610 int len = 0;
2611 struct stat st;
2612 char *p, *dname;
2613 struct per_cpu_info *pci = get_cpu_info(pdi, cpu);
2614
2615 pci->cpu = cpu;
2616 pci->fdblock = -1;
2617
2618 p = strdup(pdi->name);
2619 dname = dirname(p);
2620 if (strcmp(dname, ".")) {
2621 input_dir = dname;
2622 p = strdup(pdi->name);
2623 strcpy(pdi->name, basename(p));
2624 }
2625 free(p);
2626
2627 if (input_dir)
2628 len = sprintf(pci->fname, "%s/", input_dir);
2629
2630 snprintf(pci->fname + len, sizeof(pci->fname)-1-len,
2631 "%s.blktrace.%d", pdi->name, pci->cpu);
2632 if (stat(pci->fname, &st) < 0)
2633 return 0;
2634 if (!st.st_size)
2635 return 1;
2636
2637 pci->fd = open(pci->fname, O_RDONLY);
2638 if (pci->fd < 0) {
2639 perror(pci->fname);
2640 return 0;
2641 }
2642
2643 printf("Input file %s added\n", pci->fname);
2644 cpu_mark_online(pdi, pci->cpu);
2645
2646 pdi->nfiles++;
2647 ms_alloc(pdi, pci->cpu);
2648
2649 return 1;
2650}
2651
2652static int handle(struct ms_stream *msp)
2653{
2654 struct trace *t;
2655 struct per_dev_info *pdi;
2656 struct per_cpu_info *pci;
2657 struct blk_io_trace *bit;
2658
2659 t = ms_peek(msp);
2660
2661 bit = t->bit;
2662 pdi = msp->pdi;
2663 pci = get_cpu_info(pdi, msp->cpu);
2664 pci->nelems++;
2665 bit->time -= genesis_time;
2666
2667 if (t->bit->time > stopwatch_end)
2668 return 0;
2669
2670 pdi->last_reported_time = bit->time;
2671 if ((bit->action & (act_mask << BLK_TC_SHIFT))&&
2672 t->bit->time >= stopwatch_start)
2673 dump_trace(bit, pci, pdi);
2674
2675 ms_deq(msp);
2676
2677 if (text_output)
2678 trace_rb_insert_last(pdi, t);
2679 else {
2680 bit_free(t->bit);
2681 t_free(t);
2682 }
2683
2684 return 1;
2685}
2686
2687/*
2688 * Check if we need to sanitize the name. We allow 'foo', or if foo.blktrace.X
2689 * is given, then strip back down to 'foo' to avoid missing files.
2690 */
2691static int name_fixup(char *name)
2692{
2693 char *b;
2694
2695 if (!name)
2696 return 1;
2697
2698 b = strstr(name, ".blktrace.");
2699 if (b)
2700 *b = '\0';
2701
2702 return 0;
2703}
2704
2705static int do_file(void)
2706{
2707 int i, cpu, ret;
2708 struct per_dev_info *pdi;
2709
2710 /*
2711 * first prepare all files for reading
2712 */
2713 for (i = 0; i < ndevices; i++) {
2714 pdi = &devices[i];
2715 ret = name_fixup(pdi->name);
2716 if (ret)
2717 return ret;
2718
2719 for (cpu = 0; setup_file(pdi, cpu); cpu++)
2720 ;
2721
2722 if (!cpu) {
2723 fprintf(stderr,"No input files found for %s\n",
2724 pdi->name);
2725 return 1;
2726 }
2727 }
2728
2729 /*
2730 * Get the initial time stamp
2731 */
2732 if (ms_head)
2733 genesis_time = ms_peek_time(ms_head);
2734
2735 /*
2736 * Correct abs_start_time if necessary
2737 */
2738 if (start_timestamp
2739 && start_timestamp != genesis_time) {
2740 correct_abs_start_time();
2741 }
2742
2743 /*
2744 * Keep processing traces while any are left
2745 */
2746 while (!is_done() && ms_head && handle(ms_head))
2747 ;
2748
2749 return 0;
2750}
2751
2752static void do_pipe(int fd)
2753{
2754 unsigned long long youngest;
2755 int events, fdblock;
2756
2757 last_allowed_time = -1ULL;
2758 fdblock = -1;
2759 while ((events = read_events(fd, 0, &fdblock)) > 0) {
2760 read_sequence++;
2761
2762#if 0
2763 smallest_seq_read = -1U;
2764#endif
2765
2766 if (sort_entries(&youngest))
2767 break;
2768
2769 if (youngest > stopwatch_end)
2770 break;
2771
2772 show_entries_rb(0);
2773 }
2774
2775 if (rb_sort_entries)
2776 show_entries_rb(1);
2777}
2778
2779static int do_fifo(void)
2780{
2781 int fd;
2782
2783 if (!strcmp(pipename, "-"))
2784 fd = dup(STDIN_FILENO);
2785 else
2786 fd = open(pipename, O_RDONLY);
2787
2788 if (fd == -1) {
2789 perror("dup stdin");
2790 return -1;
2791 }
2792
2793 do_pipe(fd);
2794 close(fd);
2795 return 0;
2796}
2797
2798static void show_stats(void)
2799{
2800 if (!ofp)
2801 return;
2802 if (stats_printed)
2803 return;
2804
2805 stats_printed = 1;
2806
2807 if (per_process_stats)
2808 show_process_stats();
2809
2810 if (per_device_and_cpu_stats)
2811 show_device_and_cpu_stats();
2812
2813 fflush(ofp);
2814}
2815
2816static void handle_sigint(__attribute__((__unused__)) int sig)
2817{
2818 done = 1;
2819}
2820
2821/*
2822 * Extract start and duration times from a string, allowing
2823 * us to specify a time interval of interest within a trace.
2824 * Format: "duration" (start is zero) or "start:duration".
2825 */
2826static int find_stopwatch_interval(char *string)
2827{
2828 double value;
2829 char *sp;
2830
2831 value = strtod(string, &sp);
2832 if (sp == string) {
2833 fprintf(stderr,"Invalid stopwatch timer: %s\n", string);
2834 return 1;
2835 }
2836 if (*sp == ':') {
2837 stopwatch_start = DOUBLE_TO_NANO_ULL(value);
2838 string = sp + 1;
2839 value = strtod(string, &sp);
2840 if (sp == string || *sp != '\0') {
2841 fprintf(stderr,"Invalid stopwatch duration time: %s\n",
2842 string);
2843 return 1;
2844 }
2845 } else if (*sp != '\0') {
2846 fprintf(stderr,"Invalid stopwatch start timer: %s\n", string);
2847 return 1;
2848 }
2849 stopwatch_end = DOUBLE_TO_NANO_ULL(value);
2850 if (stopwatch_end <= stopwatch_start) {
2851 fprintf(stderr, "Invalid stopwatch interval: %Lu -> %Lu\n",
2852 stopwatch_start, stopwatch_end);
2853 return 1;
2854 }
2855
2856 return 0;
2857}
2858
2859static int is_pipe(const char *str)
2860{
2861 struct stat st;
2862
2863 if (!strcmp(str, "-"))
2864 return 1;
2865 if (!stat(str, &st) && S_ISFIFO(st.st_mode))
2866 return 1;
2867
2868 return 0;
2869}
2870
2871static int get_program_sort_event(const char *str)
2872{
2873 char evt = str[0];
2874
2875 switch (evt) {
2876 case 'N':
2877 per_process_stats_event = SORT_PROG_EVENT_N;
2878 break;
2879 case 'Q':
2880 per_process_stats_event = SORT_PROG_EVENT_QKB;
2881 break;
2882 case 'q':
2883 per_process_stats_event = SORT_PROG_EVENT_QIO;
2884 break;
2885 case 'R':
2886 per_process_stats_event = SORT_PROG_EVENT_RKB;
2887 break;
2888 case 'r':
2889 per_process_stats_event = SORT_PROG_EVENT_RIO;
2890 break;
2891 case 'W':
2892 per_process_stats_event = SORT_PROG_EVENT_WKB;
2893 break;
2894 case 'w':
2895 per_process_stats_event = SORT_PROG_EVENT_WIO;
2896 break;
2897 case 'C':
2898 per_process_stats_event = SORT_PROG_EVENT_CKB;
2899 break;
2900 case 'c':
2901 per_process_stats_event = SORT_PROG_EVENT_CIO;
2902 break;
2903 default:
2904 return 1;
2905 }
2906
2907 return 0;
2908}
2909
2910#define S_OPTS "a:A:b:D:d:f:F:hi:o:OqsS:tw:vVM"
2911static char usage_str[] = "\n\n" \
2912 "-i <file> | --input=<file>\n" \
2913 "[ -a <action field> | --act-mask=<action field> ]\n" \
2914 "[ -A <action mask> | --set-mask=<action mask> ]\n" \
2915 "[ -b <traces> | --batch=<traces> ]\n" \
2916 "[ -d <file> | --dump-binary=<file> ]\n" \
2917 "[ -D <dir> | --input-directory=<dir> ]\n" \
2918 "[ -f <format> | --format=<format> ]\n" \
2919 "[ -F <spec> | --format-spec=<spec> ]\n" \
2920 "[ -h | --hash-by-name ]\n" \
2921 "[ -o <file> | --output=<file> ]\n" \
2922 "[ -O | --no-text-output ]\n" \
2923 "[ -q | --quiet ]\n" \
2924 "[ -s | --per-program-stats ]\n" \
2925 "[ -S <event> | --sort-program-stats=<event> ]\n" \
2926 "[ -t | --track-ios ]\n" \
2927 "[ -w <time> | --stopwatch=<time> ]\n" \
2928 "[ -M | --no-msgs\n" \
2929 "[ -v | --verbose ]\n" \
2930 "[ -V | --version ]\n\n" \
2931 "\t-a Only trace specified actions. See documentation\n" \
2932 "\t-A Give trace mask as a single value. See documentation\n" \
2933 "\t-b stdin read batching\n" \
2934 "\t-d Output file. If specified, binary data is written to file\n" \
2935 "\t-D Directory to prepend to input file names\n" \
2936 "\t-f Output format. Customize the output format. The format field\n" \
2937 "\t identifies can be found in the documentation\n" \
2938 "\t-F Format specification. Can be found in the documentation\n" \
2939 "\t-h Hash processes by name, not pid\n" \
2940 "\t-i Input file containing trace data, or '-' for stdin\n" \
2941 "\t-o Output file. If not given, output is stdout\n" \
2942 "\t-O Do NOT output text data\n" \
2943 "\t-q Quiet. Don't display any stats at the end of the trace\n" \
2944 "\t-s Show per-program io statistics\n" \
2945 "\t-S Show per-program io statistics sorted by N/Q/q/R/r/W/w/C/c\n" \
2946 "\t N:Name, Q/q:Queued(read & write), R/r:Queued Read, W/w:Queued Write, C/c:Complete.\n" \
2947 "\t Sort programs by how much data(KB): Q,R,W,C.\n" \
2948 "\t Sort programs by how many IO operations: q,r,w,c.\n" \
2949 "\t if -S was used, the -s parameter will be ignored.\n" \
2950 "\t-t Track individual ios. Will tell you the time a request took\n" \
2951 "\t to get queued, to get dispatched, and to get completed\n" \
2952 "\t-w Only parse data between the given time interval in seconds.\n" \
2953 "\t If 'start' isn't given, blkparse defaults the start time to 0\n" \
2954 "\t-M Do not output messages to binary file\n" \
2955 "\t-v More verbose for marginal errors\n" \
2956 "\t-V Print program version info\n\n";
2957
2958static void usage(char *prog)
2959{
2960 fprintf(stderr, "Usage: %s %s", prog, usage_str);
2961}
2962
2963int main(int argc, char *argv[])
2964{
2965 int i, c, ret, mode;
2966 int act_mask_tmp = 0;
2967 char *ofp_buffer = NULL;
2968 char *bin_ofp_buffer = NULL;
2969
2970 while ((c = getopt_long(argc, argv, S_OPTS, l_opts, NULL)) != -1) {
2971 switch (c) {
2972 case 'a':
2973 i = find_mask_map(optarg);
2974 if (i < 0) {
2975 fprintf(stderr,"Invalid action mask %s\n",
2976 optarg);
2977 return 1;
2978 }
2979 act_mask_tmp |= i;
2980 break;
2981
2982 case 'A':
2983 if ((sscanf(optarg, "%x", &i) != 1) ||
2984 !valid_act_opt(i)) {
2985 fprintf(stderr,
2986 "Invalid set action mask %s/0x%x\n",
2987 optarg, i);
2988 return 1;
2989 }
2990 act_mask_tmp = i;
2991 break;
2992 case 'i':
2993 if (is_pipe(optarg) && !pipeline) {
2994 pipeline = 1;
2995 pipename = strdup(optarg);
2996 } else if (resize_devices(optarg) != 0)
2997 return 1;
2998 break;
2999 case 'D':
3000 input_dir = optarg;
3001 break;
3002 case 'o':
3003 output_name = optarg;
3004 break;
3005 case 'O':
3006 text_output = 0;
3007 break;
3008 case 'b':
3009 rb_batch = atoi(optarg);
3010 if (rb_batch <= 0)
3011 rb_batch = RB_BATCH_DEFAULT;
3012 break;
3013 case 's':
3014 per_process_stats = 1;
3015 break;
3016 case 'S':
3017 per_process_stats = 1;
3018 if (get_program_sort_event(optarg))
3019 return 1;
3020 break;
3021 case 't':
3022 track_ios = 1;
3023 break;
3024 case 'q':
3025 per_device_and_cpu_stats = 0;
3026 break;
3027 case 'w':
3028 if (find_stopwatch_interval(optarg) != 0)
3029 return 1;
3030 break;
3031 case 'f':
3032 set_all_format_specs(optarg);
3033 break;
3034 case 'F':
3035 if (add_format_spec(optarg) != 0)
3036 return 1;
3037 break;
3038 case 'h':
3039 ppi_hash_by_pid = 0;
3040 break;
3041 case 'v':
3042 verbose++;
3043 break;
3044 case 'V':
3045 printf("%s version %s\n", argv[0], blkparse_version);
3046 return 0;
3047 case 'd':
3048 dump_binary = optarg;
3049 break;
3050 case 'M':
3051 bin_output_msgs = 0;
3052 break;
3053 default:
3054 usage(argv[0]);
3055 return 1;
3056 }
3057 }
3058
3059 while (optind < argc) {
3060 if (is_pipe(argv[optind]) && !pipeline) {
3061 pipeline = 1;
3062 pipename = strdup(argv[optind]);
3063 } else if (resize_devices(argv[optind]) != 0)
3064 return 1;
3065 optind++;
3066 }
3067
3068 if (!pipeline && !ndevices) {
3069 usage(argv[0]);
3070 return 1;
3071 }
3072
3073 if (act_mask_tmp != 0)
3074 act_mask = act_mask_tmp;
3075
3076 memset(&rb_sort_root, 0, sizeof(rb_sort_root));
3077
3078 signal(SIGINT, handle_sigint);
3079 signal(SIGHUP, handle_sigint);
3080 signal(SIGTERM, handle_sigint);
3081
3082 setlocale(LC_NUMERIC, "en_US");
3083
3084 if (text_output) {
3085 if (!output_name) {
3086 ofp = fdopen(STDOUT_FILENO, "w");
3087 mode = _IOLBF;
3088 } else {
3089 char ofname[PATH_MAX];
3090
3091 snprintf(ofname, sizeof(ofname) - 1, "%s", output_name);
3092 ofp = fopen(ofname, "w");
3093 mode = _IOFBF;
3094 }
3095
3096 if (!ofp) {
3097 perror("fopen");
3098 return 1;
3099 }
3100
3101 ofp_buffer = malloc(4096);
3102 if (setvbuf(ofp, ofp_buffer, mode, 4096)) {
3103 perror("setvbuf");
3104 return 1;
3105 }
3106 }
3107
3108 if (dump_binary) {
3109 if (!strcmp(dump_binary, "-"))
3110 dump_fp = stdout;
3111 else {
3112 dump_fp = fopen(dump_binary, "w");
3113 if (!dump_fp) {
3114 perror(dump_binary);
3115 dump_binary = NULL;
3116 return 1;
3117 }
3118 }
3119 bin_ofp_buffer = malloc(128 * 1024);
3120 if (setvbuf(dump_fp, bin_ofp_buffer, _IOFBF, 128 * 1024)) {
3121 perror("setvbuf binary");
3122 return 1;
3123 }
3124 }
3125
3126 if (pipeline)
3127 ret = do_fifo();
3128 else
3129 ret = do_file();
3130
3131 if (!ret)
3132 show_stats();
3133
3134 if (have_drv_data && !dump_binary)
3135 printf("\ndiscarded traces containing low-level device driver "
3136 "specific data (only available in binary output)\n");
3137
3138 if (ofp_buffer) {
3139 fflush(ofp);
3140 free(ofp_buffer);
3141 }
3142 if (bin_ofp_buffer) {
3143 fflush(dump_fp);
3144 free(bin_ofp_buffer);
3145 }
3146 return ret;
3147}
Block queue io tracer