forked from torvalds/linux
-
Notifications
You must be signed in to change notification settings - Fork 0
/
ring_buffer.c
3880 lines (3244 loc) · 96.2 KB
/
ring_buffer.c
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/*
* Generic ring buffer
*
* Copyright (C) 2008 Steven Rostedt <[email protected]>
*/
#include <linux/ring_buffer.h>
#include <linux/trace_clock.h>
#include <linux/ftrace_irq.h>
#include <linux/spinlock.h>
#include <linux/debugfs.h>
#include <linux/uaccess.h>
#include <linux/hardirq.h>
#include <linux/kmemcheck.h>
#include <linux/module.h>
#include <linux/percpu.h>
#include <linux/mutex.h>
#include <linux/init.h>
#include <linux/hash.h>
#include <linux/list.h>
#include <linux/cpu.h>
#include <linux/fs.h>
#include "trace.h"
/*
* The ring buffer header is special. We must manually up keep it.
*/
int ring_buffer_print_entry_header(struct trace_seq *s)
{
int ret;
ret = trace_seq_printf(s, "# compressed entry header\n");
ret = trace_seq_printf(s, "\ttype_len : 5 bits\n");
ret = trace_seq_printf(s, "\ttime_delta : 27 bits\n");
ret = trace_seq_printf(s, "\tarray : 32 bits\n");
ret = trace_seq_printf(s, "\n");
ret = trace_seq_printf(s, "\tpadding : type == %d\n",
RINGBUF_TYPE_PADDING);
ret = trace_seq_printf(s, "\ttime_extend : type == %d\n",
RINGBUF_TYPE_TIME_EXTEND);
ret = trace_seq_printf(s, "\tdata max type_len == %d\n",
RINGBUF_TYPE_DATA_TYPE_LEN_MAX);
return ret;
}
/*
* The ring buffer is made up of a list of pages. A separate list of pages is
* allocated for each CPU. A writer may only write to a buffer that is
* associated with the CPU it is currently executing on. A reader may read
* from any per cpu buffer.
*
* The reader is special. For each per cpu buffer, the reader has its own
* reader page. When a reader has read the entire reader page, this reader
* page is swapped with another page in the ring buffer.
*
* Now, as long as the writer is off the reader page, the reader can do what
* ever it wants with that page. The writer will never write to that page
* again (as long as it is out of the ring buffer).
*
* Here's some silly ASCII art.
*
* +------+
* |reader| RING BUFFER
* |page |
* +------+ +---+ +---+ +---+
* | |-->| |-->| |
* +---+ +---+ +---+
* ^ |
* | |
* +---------------+
*
*
* +------+
* |reader| RING BUFFER
* |page |------------------v
* +------+ +---+ +---+ +---+
* | |-->| |-->| |
* +---+ +---+ +---+
* ^ |
* | |
* +---------------+
*
*
* +------+
* |reader| RING BUFFER
* |page |------------------v
* +------+ +---+ +---+ +---+
* ^ | |-->| |-->| |
* | +---+ +---+ +---+
* | |
* | |
* +------------------------------+
*
*
* +------+
* |buffer| RING BUFFER
* |page |------------------v
* +------+ +---+ +---+ +---+
* ^ | | | |-->| |
* | New +---+ +---+ +---+
* | Reader------^ |
* | page |
* +------------------------------+
*
*
* After we make this swap, the reader can hand this page off to the splice
* code and be done with it. It can even allocate a new page if it needs to
* and swap that into the ring buffer.
*
* We will be using cmpxchg soon to make all this lockless.
*
*/
/*
* A fast way to enable or disable all ring buffers is to
* call tracing_on or tracing_off. Turning off the ring buffers
* prevents all ring buffers from being recorded to.
* Turning this switch on, makes it OK to write to the
* ring buffer, if the ring buffer is enabled itself.
*
* There's three layers that must be on in order to write
* to the ring buffer.
*
* 1) This global flag must be set.
* 2) The ring buffer must be enabled for recording.
* 3) The per cpu buffer must be enabled for recording.
*
* In case of an anomaly, this global flag has a bit set that
* will permantly disable all ring buffers.
*/
/*
* Global flag to disable all recording to ring buffers
* This has two bits: ON, DISABLED
*
* ON DISABLED
* ---- ----------
* 0 0 : ring buffers are off
* 1 0 : ring buffers are on
* X 1 : ring buffers are permanently disabled
*/
enum {
RB_BUFFERS_ON_BIT = 0,
RB_BUFFERS_DISABLED_BIT = 1,
};
enum {
RB_BUFFERS_ON = 1 << RB_BUFFERS_ON_BIT,
RB_BUFFERS_DISABLED = 1 << RB_BUFFERS_DISABLED_BIT,
};
static unsigned long ring_buffer_flags __read_mostly = RB_BUFFERS_ON;
#define BUF_PAGE_HDR_SIZE offsetof(struct buffer_data_page, data)
/**
* tracing_on - enable all tracing buffers
*
* This function enables all tracing buffers that may have been
* disabled with tracing_off.
*/
void tracing_on(void)
{
set_bit(RB_BUFFERS_ON_BIT, &ring_buffer_flags);
}
EXPORT_SYMBOL_GPL(tracing_on);
/**
* tracing_off - turn off all tracing buffers
*
* This function stops all tracing buffers from recording data.
* It does not disable any overhead the tracers themselves may
* be causing. This function simply causes all recording to
* the ring buffers to fail.
*/
void tracing_off(void)
{
clear_bit(RB_BUFFERS_ON_BIT, &ring_buffer_flags);
}
EXPORT_SYMBOL_GPL(tracing_off);
/**
* tracing_off_permanent - permanently disable ring buffers
*
* This function, once called, will disable all ring buffers
* permanently.
*/
void tracing_off_permanent(void)
{
set_bit(RB_BUFFERS_DISABLED_BIT, &ring_buffer_flags);
}
/**
* tracing_is_on - show state of ring buffers enabled
*/
int tracing_is_on(void)
{
return ring_buffer_flags == RB_BUFFERS_ON;
}
EXPORT_SYMBOL_GPL(tracing_is_on);
#include "trace.h"
#define RB_EVNT_HDR_SIZE (offsetof(struct ring_buffer_event, array))
#define RB_ALIGNMENT 4U
#define RB_MAX_SMALL_DATA (RB_ALIGNMENT * RINGBUF_TYPE_DATA_TYPE_LEN_MAX)
#define RB_EVNT_MIN_SIZE 8U /* two 32bit words */
/* define RINGBUF_TYPE_DATA for 'case RINGBUF_TYPE_DATA:' */
#define RINGBUF_TYPE_DATA 0 ... RINGBUF_TYPE_DATA_TYPE_LEN_MAX
enum {
RB_LEN_TIME_EXTEND = 8,
RB_LEN_TIME_STAMP = 16,
};
static inline int rb_null_event(struct ring_buffer_event *event)
{
return event->type_len == RINGBUF_TYPE_PADDING && !event->time_delta;
}
static void rb_event_set_padding(struct ring_buffer_event *event)
{
/* padding has a NULL time_delta */
event->type_len = RINGBUF_TYPE_PADDING;
event->time_delta = 0;
}
static unsigned
rb_event_data_length(struct ring_buffer_event *event)
{
unsigned length;
if (event->type_len)
length = event->type_len * RB_ALIGNMENT;
else
length = event->array[0];
return length + RB_EVNT_HDR_SIZE;
}
/* inline for ring buffer fast paths */
static unsigned
rb_event_length(struct ring_buffer_event *event)
{
switch (event->type_len) {
case RINGBUF_TYPE_PADDING:
if (rb_null_event(event))
/* undefined */
return -1;
return event->array[0] + RB_EVNT_HDR_SIZE;
case RINGBUF_TYPE_TIME_EXTEND:
return RB_LEN_TIME_EXTEND;
case RINGBUF_TYPE_TIME_STAMP:
return RB_LEN_TIME_STAMP;
case RINGBUF_TYPE_DATA:
return rb_event_data_length(event);
default:
BUG();
}
/* not hit */
return 0;
}
/**
* ring_buffer_event_length - return the length of the event
* @event: the event to get the length of
*/
unsigned ring_buffer_event_length(struct ring_buffer_event *event)
{
unsigned length = rb_event_length(event);
if (event->type_len > RINGBUF_TYPE_DATA_TYPE_LEN_MAX)
return length;
length -= RB_EVNT_HDR_SIZE;
if (length > RB_MAX_SMALL_DATA + sizeof(event->array[0]))
length -= sizeof(event->array[0]);
return length;
}
EXPORT_SYMBOL_GPL(ring_buffer_event_length);
/* inline for ring buffer fast paths */
static void *
rb_event_data(struct ring_buffer_event *event)
{
BUG_ON(event->type_len > RINGBUF_TYPE_DATA_TYPE_LEN_MAX);
/* If length is in len field, then array[0] has the data */
if (event->type_len)
return (void *)&event->array[0];
/* Otherwise length is in array[0] and array[1] has the data */
return (void *)&event->array[1];
}
/**
* ring_buffer_event_data - return the data of the event
* @event: the event to get the data from
*/
void *ring_buffer_event_data(struct ring_buffer_event *event)
{
return rb_event_data(event);
}
EXPORT_SYMBOL_GPL(ring_buffer_event_data);
#define for_each_buffer_cpu(buffer, cpu) \
for_each_cpu(cpu, buffer->cpumask)
#define TS_SHIFT 27
#define TS_MASK ((1ULL << TS_SHIFT) - 1)
#define TS_DELTA_TEST (~TS_MASK)
struct buffer_data_page {
u64 time_stamp; /* page time stamp */
local_t commit; /* write committed index */
unsigned char data[]; /* data of buffer page */
};
/*
* Note, the buffer_page list must be first. The buffer pages
* are allocated in cache lines, which means that each buffer
* page will be at the beginning of a cache line, and thus
* the least significant bits will be zero. We use this to
* add flags in the list struct pointers, to make the ring buffer
* lockless.
*/
struct buffer_page {
struct list_head list; /* list of buffer pages */
local_t write; /* index for next write */
unsigned read; /* index for next read */
local_t entries; /* entries on this page */
struct buffer_data_page *page; /* Actual data page */
};
/*
* The buffer page counters, write and entries, must be reset
* atomically when crossing page boundaries. To synchronize this
* update, two counters are inserted into the number. One is
* the actual counter for the write position or count on the page.
*
* The other is a counter of updaters. Before an update happens
* the update partition of the counter is incremented. This will
* allow the updater to update the counter atomically.
*
* The counter is 20 bits, and the state data is 12.
*/
#define RB_WRITE_MASK 0xfffff
#define RB_WRITE_INTCNT (1 << 20)
static void rb_init_page(struct buffer_data_page *bpage)
{
local_set(&bpage->commit, 0);
}
/**
* ring_buffer_page_len - the size of data on the page.
* @page: The page to read
*
* Returns the amount of data on the page, including buffer page header.
*/
size_t ring_buffer_page_len(void *page)
{
return local_read(&((struct buffer_data_page *)page)->commit)
+ BUF_PAGE_HDR_SIZE;
}
/*
* Also stolen from mm/slob.c. Thanks to Mathieu Desnoyers for pointing
* this issue out.
*/
static void free_buffer_page(struct buffer_page *bpage)
{
free_page((unsigned long)bpage->page);
kfree(bpage);
}
/*
* We need to fit the time_stamp delta into 27 bits.
*/
static inline int test_time_stamp(u64 delta)
{
if (delta & TS_DELTA_TEST)
return 1;
return 0;
}
#define BUF_PAGE_SIZE (PAGE_SIZE - BUF_PAGE_HDR_SIZE)
/* Max payload is BUF_PAGE_SIZE - header (8bytes) */
#define BUF_MAX_DATA_SIZE (BUF_PAGE_SIZE - (sizeof(u32) * 2))
/* Max number of timestamps that can fit on a page */
#define RB_TIMESTAMPS_PER_PAGE (BUF_PAGE_SIZE / RB_LEN_TIME_STAMP)
int ring_buffer_print_page_header(struct trace_seq *s)
{
struct buffer_data_page field;
int ret;
ret = trace_seq_printf(s, "\tfield: u64 timestamp;\t"
"offset:0;\tsize:%u;\n",
(unsigned int)sizeof(field.time_stamp));
ret = trace_seq_printf(s, "\tfield: local_t commit;\t"
"offset:%u;\tsize:%u;\n",
(unsigned int)offsetof(typeof(field), commit),
(unsigned int)sizeof(field.commit));
ret = trace_seq_printf(s, "\tfield: char data;\t"
"offset:%u;\tsize:%u;\n",
(unsigned int)offsetof(typeof(field), data),
(unsigned int)BUF_PAGE_SIZE);
return ret;
}
/*
* head_page == tail_page && head == tail then buffer is empty.
*/
struct ring_buffer_per_cpu {
int cpu;
struct ring_buffer *buffer;
spinlock_t reader_lock; /* serialize readers */
raw_spinlock_t lock;
struct lock_class_key lock_key;
struct list_head *pages;
struct buffer_page *head_page; /* read from head */
struct buffer_page *tail_page; /* write to tail */
struct buffer_page *commit_page; /* committed pages */
struct buffer_page *reader_page;
local_t commit_overrun;
local_t overrun;
local_t entries;
local_t committing;
local_t commits;
unsigned long read;
u64 write_stamp;
u64 read_stamp;
atomic_t record_disabled;
};
struct ring_buffer {
unsigned pages;
unsigned flags;
int cpus;
atomic_t record_disabled;
cpumask_var_t cpumask;
struct lock_class_key *reader_lock_key;
struct mutex mutex;
struct ring_buffer_per_cpu **buffers;
#ifdef CONFIG_HOTPLUG_CPU
struct notifier_block cpu_notify;
#endif
u64 (*clock)(void);
};
struct ring_buffer_iter {
struct ring_buffer_per_cpu *cpu_buffer;
unsigned long head;
struct buffer_page *head_page;
u64 read_stamp;
};
/* buffer may be either ring_buffer or ring_buffer_per_cpu */
#define RB_WARN_ON(b, cond) \
({ \
int _____ret = unlikely(cond); \
if (_____ret) { \
if (__same_type(*(b), struct ring_buffer_per_cpu)) { \
struct ring_buffer_per_cpu *__b = \
(void *)b; \
atomic_inc(&__b->buffer->record_disabled); \
} else \
atomic_inc(&b->record_disabled); \
WARN_ON(1); \
} \
_____ret; \
})
/* Up this if you want to test the TIME_EXTENTS and normalization */
#define DEBUG_SHIFT 0
static inline u64 rb_time_stamp(struct ring_buffer *buffer, int cpu)
{
/* shift to debug/test normalization and TIME_EXTENTS */
return buffer->clock() << DEBUG_SHIFT;
}
u64 ring_buffer_time_stamp(struct ring_buffer *buffer, int cpu)
{
u64 time;
preempt_disable_notrace();
time = rb_time_stamp(buffer, cpu);
preempt_enable_no_resched_notrace();
return time;
}
EXPORT_SYMBOL_GPL(ring_buffer_time_stamp);
void ring_buffer_normalize_time_stamp(struct ring_buffer *buffer,
int cpu, u64 *ts)
{
/* Just stupid testing the normalize function and deltas */
*ts >>= DEBUG_SHIFT;
}
EXPORT_SYMBOL_GPL(ring_buffer_normalize_time_stamp);
/*
* Making the ring buffer lockless makes things tricky.
* Although writes only happen on the CPU that they are on,
* and they only need to worry about interrupts. Reads can
* happen on any CPU.
*
* The reader page is always off the ring buffer, but when the
* reader finishes with a page, it needs to swap its page with
* a new one from the buffer. The reader needs to take from
* the head (writes go to the tail). But if a writer is in overwrite
* mode and wraps, it must push the head page forward.
*
* Here lies the problem.
*
* The reader must be careful to replace only the head page, and
* not another one. As described at the top of the file in the
* ASCII art, the reader sets its old page to point to the next
* page after head. It then sets the page after head to point to
* the old reader page. But if the writer moves the head page
* during this operation, the reader could end up with the tail.
*
* We use cmpxchg to help prevent this race. We also do something
* special with the page before head. We set the LSB to 1.
*
* When the writer must push the page forward, it will clear the
* bit that points to the head page, move the head, and then set
* the bit that points to the new head page.
*
* We also don't want an interrupt coming in and moving the head
* page on another writer. Thus we use the second LSB to catch
* that too. Thus:
*
* head->list->prev->next bit 1 bit 0
* ------- -------
* Normal page 0 0
* Points to head page 0 1
* New head page 1 0
*
* Note we can not trust the prev pointer of the head page, because:
*
* +----+ +-----+ +-----+
* | |------>| T |---X--->| N |
* | |<------| | | |
* +----+ +-----+ +-----+
* ^ ^ |
* | +-----+ | |
* +----------| R |----------+ |
* | |<-----------+
* +-----+
*
* Key: ---X--> HEAD flag set in pointer
* T Tail page
* R Reader page
* N Next page
*
* (see __rb_reserve_next() to see where this happens)
*
* What the above shows is that the reader just swapped out
* the reader page with a page in the buffer, but before it
* could make the new header point back to the new page added
* it was preempted by a writer. The writer moved forward onto
* the new page added by the reader and is about to move forward
* again.
*
* You can see, it is legitimate for the previous pointer of
* the head (or any page) not to point back to itself. But only
* temporarially.
*/
#define RB_PAGE_NORMAL 0UL
#define RB_PAGE_HEAD 1UL
#define RB_PAGE_UPDATE 2UL
#define RB_FLAG_MASK 3UL
/* PAGE_MOVED is not part of the mask */
#define RB_PAGE_MOVED 4UL
/*
* rb_list_head - remove any bit
*/
static struct list_head *rb_list_head(struct list_head *list)
{
unsigned long val = (unsigned long)list;
return (struct list_head *)(val & ~RB_FLAG_MASK);
}
/*
* rb_is_head_page - test if the give page is the head page
*
* Because the reader may move the head_page pointer, we can
* not trust what the head page is (it may be pointing to
* the reader page). But if the next page is a header page,
* its flags will be non zero.
*/
static int inline
rb_is_head_page(struct ring_buffer_per_cpu *cpu_buffer,
struct buffer_page *page, struct list_head *list)
{
unsigned long val;
val = (unsigned long)list->next;
if ((val & ~RB_FLAG_MASK) != (unsigned long)&page->list)
return RB_PAGE_MOVED;
return val & RB_FLAG_MASK;
}
/*
* rb_is_reader_page
*
* The unique thing about the reader page, is that, if the
* writer is ever on it, the previous pointer never points
* back to the reader page.
*/
static int rb_is_reader_page(struct buffer_page *page)
{
struct list_head *list = page->list.prev;
return rb_list_head(list->next) != &page->list;
}
/*
* rb_set_list_to_head - set a list_head to be pointing to head.
*/
static void rb_set_list_to_head(struct ring_buffer_per_cpu *cpu_buffer,
struct list_head *list)
{
unsigned long *ptr;
ptr = (unsigned long *)&list->next;
*ptr |= RB_PAGE_HEAD;
*ptr &= ~RB_PAGE_UPDATE;
}
/*
* rb_head_page_activate - sets up head page
*/
static void rb_head_page_activate(struct ring_buffer_per_cpu *cpu_buffer)
{
struct buffer_page *head;
head = cpu_buffer->head_page;
if (!head)
return;
/*
* Set the previous list pointer to have the HEAD flag.
*/
rb_set_list_to_head(cpu_buffer, head->list.prev);
}
static void rb_list_head_clear(struct list_head *list)
{
unsigned long *ptr = (unsigned long *)&list->next;
*ptr &= ~RB_FLAG_MASK;
}
/*
* rb_head_page_dactivate - clears head page ptr (for free list)
*/
static void
rb_head_page_deactivate(struct ring_buffer_per_cpu *cpu_buffer)
{
struct list_head *hd;
/* Go through the whole list and clear any pointers found. */
rb_list_head_clear(cpu_buffer->pages);
list_for_each(hd, cpu_buffer->pages)
rb_list_head_clear(hd);
}
static int rb_head_page_set(struct ring_buffer_per_cpu *cpu_buffer,
struct buffer_page *head,
struct buffer_page *prev,
int old_flag, int new_flag)
{
struct list_head *list;
unsigned long val = (unsigned long)&head->list;
unsigned long ret;
list = &prev->list;
val &= ~RB_FLAG_MASK;
ret = (unsigned long)cmpxchg(&list->next,
val | old_flag, val | new_flag);
/* check if the reader took the page */
if ((ret & ~RB_FLAG_MASK) != val)
return RB_PAGE_MOVED;
return ret & RB_FLAG_MASK;
}
static int rb_head_page_set_update(struct ring_buffer_per_cpu *cpu_buffer,
struct buffer_page *head,
struct buffer_page *prev,
int old_flag)
{
return rb_head_page_set(cpu_buffer, head, prev,
old_flag, RB_PAGE_UPDATE);
}
static int rb_head_page_set_head(struct ring_buffer_per_cpu *cpu_buffer,
struct buffer_page *head,
struct buffer_page *prev,
int old_flag)
{
return rb_head_page_set(cpu_buffer, head, prev,
old_flag, RB_PAGE_HEAD);
}
static int rb_head_page_set_normal(struct ring_buffer_per_cpu *cpu_buffer,
struct buffer_page *head,
struct buffer_page *prev,
int old_flag)
{
return rb_head_page_set(cpu_buffer, head, prev,
old_flag, RB_PAGE_NORMAL);
}
static inline void rb_inc_page(struct ring_buffer_per_cpu *cpu_buffer,
struct buffer_page **bpage)
{
struct list_head *p = rb_list_head((*bpage)->list.next);
*bpage = list_entry(p, struct buffer_page, list);
}
static struct buffer_page *
rb_set_head_page(struct ring_buffer_per_cpu *cpu_buffer)
{
struct buffer_page *head;
struct buffer_page *page;
struct list_head *list;
int i;
if (RB_WARN_ON(cpu_buffer, !cpu_buffer->head_page))
return NULL;
/* sanity check */
list = cpu_buffer->pages;
if (RB_WARN_ON(cpu_buffer, rb_list_head(list->prev->next) != list))
return NULL;
page = head = cpu_buffer->head_page;
/*
* It is possible that the writer moves the header behind
* where we started, and we miss in one loop.
* A second loop should grab the header, but we'll do
* three loops just because I'm paranoid.
*/
for (i = 0; i < 3; i++) {
do {
if (rb_is_head_page(cpu_buffer, page, page->list.prev)) {
cpu_buffer->head_page = page;
return page;
}
rb_inc_page(cpu_buffer, &page);
} while (page != head);
}
RB_WARN_ON(cpu_buffer, 1);
return NULL;
}
static int rb_head_page_replace(struct buffer_page *old,
struct buffer_page *new)
{
unsigned long *ptr = (unsigned long *)&old->list.prev->next;
unsigned long val;
unsigned long ret;
val = *ptr & ~RB_FLAG_MASK;
val |= RB_PAGE_HEAD;
ret = cmpxchg(ptr, val, &new->list);
return ret == val;
}
/*
* rb_tail_page_update - move the tail page forward
*
* Returns 1 if moved tail page, 0 if someone else did.
*/
static int rb_tail_page_update(struct ring_buffer_per_cpu *cpu_buffer,
struct buffer_page *tail_page,
struct buffer_page *next_page)
{
struct buffer_page *old_tail;
unsigned long old_entries;
unsigned long old_write;
int ret = 0;
/*
* The tail page now needs to be moved forward.
*
* We need to reset the tail page, but without messing
* with possible erasing of data brought in by interrupts
* that have moved the tail page and are currently on it.
*
* We add a counter to the write field to denote this.
*/
old_write = local_add_return(RB_WRITE_INTCNT, &next_page->write);
old_entries = local_add_return(RB_WRITE_INTCNT, &next_page->entries);
/*
* Just make sure we have seen our old_write and synchronize
* with any interrupts that come in.
*/
barrier();
/*
* If the tail page is still the same as what we think
* it is, then it is up to us to update the tail
* pointer.
*/
if (tail_page == cpu_buffer->tail_page) {
/* Zero the write counter */
unsigned long val = old_write & ~RB_WRITE_MASK;
unsigned long eval = old_entries & ~RB_WRITE_MASK;
/*
* This will only succeed if an interrupt did
* not come in and change it. In which case, we
* do not want to modify it.
*
* We add (void) to let the compiler know that we do not care
* about the return value of these functions. We use the
* cmpxchg to only update if an interrupt did not already
* do it for us. If the cmpxchg fails, we don't care.
*/
(void)local_cmpxchg(&next_page->write, old_write, val);
(void)local_cmpxchg(&next_page->entries, old_entries, eval);
/*
* No need to worry about races with clearing out the commit.
* it only can increment when a commit takes place. But that
* only happens in the outer most nested commit.
*/
local_set(&next_page->page->commit, 0);
old_tail = cmpxchg(&cpu_buffer->tail_page,
tail_page, next_page);
if (old_tail == tail_page)
ret = 1;
}
return ret;
}
static int rb_check_bpage(struct ring_buffer_per_cpu *cpu_buffer,
struct buffer_page *bpage)
{
unsigned long val = (unsigned long)bpage;
if (RB_WARN_ON(cpu_buffer, val & RB_FLAG_MASK))
return 1;
return 0;
}
/**
* rb_check_list - make sure a pointer to a list has the last bits zero
*/
static int rb_check_list(struct ring_buffer_per_cpu *cpu_buffer,
struct list_head *list)
{
if (RB_WARN_ON(cpu_buffer, rb_list_head(list->prev) != list->prev))
return 1;
if (RB_WARN_ON(cpu_buffer, rb_list_head(list->next) != list->next))
return 1;
return 0;
}
/**
* check_pages - integrity check of buffer pages
* @cpu_buffer: CPU buffer with pages to test
*
* As a safety measure we check to make sure the data pages have not
* been corrupted.
*/
static int rb_check_pages(struct ring_buffer_per_cpu *cpu_buffer)
{
struct list_head *head = cpu_buffer->pages;
struct buffer_page *bpage, *tmp;
rb_head_page_deactivate(cpu_buffer);
if (RB_WARN_ON(cpu_buffer, head->next->prev != head))
return -1;
if (RB_WARN_ON(cpu_buffer, head->prev->next != head))
return -1;
if (rb_check_list(cpu_buffer, head))
return -1;
list_for_each_entry_safe(bpage, tmp, head, list) {
if (RB_WARN_ON(cpu_buffer,
bpage->list.next->prev != &bpage->list))
return -1;
if (RB_WARN_ON(cpu_buffer,
bpage->list.prev->next != &bpage->list))
return -1;
if (rb_check_list(cpu_buffer, &bpage->list))
return -1;
}
rb_head_page_activate(cpu_buffer);
return 0;
}
static int rb_allocate_pages(struct ring_buffer_per_cpu *cpu_buffer,
unsigned nr_pages)
{
struct buffer_page *bpage, *tmp;
unsigned long addr;
LIST_HEAD(pages);
unsigned i;
WARN_ON(!nr_pages);
for (i = 0; i < nr_pages; i++) {
bpage = kzalloc_node(ALIGN(sizeof(*bpage), cache_line_size()),
GFP_KERNEL, cpu_to_node(cpu_buffer->cpu));
if (!bpage)
goto free_pages;
rb_check_bpage(cpu_buffer, bpage);
list_add(&bpage->list, &pages);
addr = __get_free_page(GFP_KERNEL);
if (!addr)
goto free_pages;
bpage->page = (void *)addr;
rb_init_page(bpage->page);
}
/*
* The ring buffer page list is a circular list that does not
* start and end with a list head. All page list items point to
* other pages.
*/
cpu_buffer->pages = pages.next;
list_del(&pages);
rb_check_pages(cpu_buffer);
return 0;
free_pages:
list_for_each_entry_safe(bpage, tmp, &pages, list) {
list_del_init(&bpage->list);
free_buffer_page(bpage);
}
return -ENOMEM;
}
static struct ring_buffer_per_cpu *
rb_allocate_cpu_buffer(struct ring_buffer *buffer, int cpu)
{
struct ring_buffer_per_cpu *cpu_buffer;
struct buffer_page *bpage;
unsigned long addr;
int ret;
cpu_buffer = kzalloc_node(ALIGN(sizeof(*cpu_buffer), cache_line_size()),
GFP_KERNEL, cpu_to_node(cpu));
if (!cpu_buffer)
return NULL;
cpu_buffer->cpu = cpu;
cpu_buffer->buffer = buffer;
spin_lock_init(&cpu_buffer->reader_lock);
lockdep_set_class(&cpu_buffer->reader_lock, buffer->reader_lock_key);
cpu_buffer->lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;