forked from efficient/cicada-exp-sigmod2017-silo
-
Notifications
You must be signed in to change notification settings - Fork 0
/
txn.h
865 lines (752 loc) · 24.1 KB
/
txn.h
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
#ifndef _NDB_TXN_H_
#define _NDB_TXN_H_
#include <malloc.h>
#include <stdint.h>
#include <sys/types.h>
#include <pthread.h>
#include <map>
#include <iostream>
#include <vector>
#include <string>
#include <utility>
#include <stdexcept>
#include <limits>
#include <type_traits>
#include <tuple>
#include <unordered_map>
#include "amd64.h"
#include "btree_choice.h"
#include "core.h"
#include "counter.h"
#include "macros.h"
#include "varkey.h"
#include "util.h"
#include "rcu.h"
#include "thread.h"
#include "spinlock.h"
#include "small_unordered_map.h"
#include "static_unordered_map.h"
#include "static_vector.h"
#include "prefetch.h"
#include "tuple.h"
#include "scopedperf.hh"
#include "marked_ptr.h"
#include "ndb_type_traits.h"
// forward decl
template <template <typename> class Transaction, typename P>
class base_txn_btree;
class transaction_unusable_exception {};
class transaction_read_only_exception {};
// XXX: hacky
extern std::string (*g_proto_version_str)(uint64_t v);
// base class with very simple definitions- nothing too exciting yet
class transaction_base {
template <template <typename> class T, typename P>
friend class base_txn_btree;
public:
typedef dbtuple::tid_t tid_t;
typedef dbtuple::size_type size_type;
typedef dbtuple::string_type string_type;
// TXN_EMBRYO - the transaction object has been allocated but has not
// done any operations yet
enum txn_state { TXN_EMBRYO, TXN_ACTIVE, TXN_COMMITED, TXN_ABRT, };
enum {
// use the low-level scan protocol for checking scan consistency,
// instead of keeping track of absent ranges
TXN_FLAG_LOW_LEVEL_SCAN = 0x1,
// true to mark a read-only transaction- if a txn marked read-only
// does a write, a transaction_read_only_exception is thrown and the
// txn is aborted
TXN_FLAG_READ_ONLY = 0x2,
// XXX: more flags in the future, things like consistency levels
};
#define ABORT_REASONS(x) \
x(ABORT_REASON_NONE) \
x(ABORT_REASON_USER) \
x(ABORT_REASON_UNSTABLE_READ) \
x(ABORT_REASON_FUTURE_TID_READ) \
x(ABORT_REASON_NODE_SCAN_WRITE_VERSION_CHANGED) \
x(ABORT_REASON_NODE_SCAN_READ_VERSION_CHANGED) \
x(ABORT_REASON_WRITE_NODE_INTERFERENCE) \
x(ABORT_REASON_INSERT_NODE_INTERFERENCE) \
x(ABORT_REASON_READ_NODE_INTEREFERENCE) \
x(ABORT_REASON_READ_ABSENCE_INTEREFERENCE)
enum abort_reason {
#define ENUM_X(x) x,
ABORT_REASONS(ENUM_X)
#undef ENUM_X
};
static const char *
AbortReasonStr(abort_reason reason)
{
switch (reason) {
#define CASE_X(x) case x: return #x;
ABORT_REASONS(CASE_X)
#undef CASE_X
default:
break;
}
ALWAYS_ASSERT(false);
return 0;
}
transaction_base(uint64_t flags)
: state(TXN_EMBRYO),
reason(ABORT_REASON_NONE),
flags(flags) {}
transaction_base(const transaction_base &) = delete;
transaction_base(transaction_base &&) = delete;
transaction_base &operator=(const transaction_base &) = delete;
protected:
#define EVENT_COUNTER_DEF_X(x) \
static event_counter g_ ## x ## _ctr;
ABORT_REASONS(EVENT_COUNTER_DEF_X)
#undef EVENT_COUNTER_DEF_X
static event_counter *
AbortReasonCounter(abort_reason reason)
{
switch (reason) {
#define EVENT_COUNTER_CASE_X(x) case x: return &g_ ## x ## _ctr;
ABORT_REASONS(EVENT_COUNTER_CASE_X)
#undef EVENT_COUNTER_CASE_X
default:
break;
}
ALWAYS_ASSERT(false);
return 0;
}
public:
// only fires during invariant checking
inline void
ensure_active()
{
if (state == TXN_EMBRYO)
state = TXN_ACTIVE;
INVARIANT(state == TXN_ACTIVE);
}
inline uint64_t
get_flags() const
{
return flags;
}
protected:
// the read set is a mapping from (tuple -> tid_read).
// "write_set" is used to indicate if this read tuple
// also belongs in the write set.
struct read_record_t {
constexpr read_record_t() : tuple(), t() {}
constexpr read_record_t(const dbtuple *tuple, tid_t t)
: tuple(tuple), t(t) {}
inline const dbtuple *
get_tuple() const
{
return tuple;
}
inline tid_t
get_tid() const
{
return t;
}
private:
const dbtuple *tuple;
tid_t t;
};
friend std::ostream &
operator<<(std::ostream &o, const read_record_t &r);
// the write set is logically a mapping from (tuple -> value_to_write).
struct write_record_t {
enum {
FLAGS_INSERT = 0x1,
FLAGS_DOWRITE = 0x1 << 1,
};
constexpr inline write_record_t()
: tuple(), k(), r(), w(), btr()
{}
// all inputs are assumed to be stable
inline write_record_t(dbtuple *tuple,
const string_type *k,
const void *r,
dbtuple::tuple_writer_t w,
concurrent_btree *btr,
bool insert)
: tuple(tuple),
k(k),
r(r),
w(w),
btr(btr)
{
this->btr.set_flags(insert ? FLAGS_INSERT : 0);
}
inline dbtuple *
get_tuple()
{
return tuple;
}
inline const dbtuple *
get_tuple() const
{
return tuple;
}
inline bool
is_insert() const
{
return btr.get_flags() & FLAGS_INSERT;
}
inline bool
do_write() const
{
return btr.get_flags() & FLAGS_DOWRITE;
}
inline void
set_do_write()
{
INVARIANT(!do_write());
btr.or_flags(FLAGS_DOWRITE);
}
inline concurrent_btree *
get_btree() const
{
return btr.get();
}
inline const string_type &
get_key() const
{
return *k;
}
inline const void *
get_value() const
{
return r;
}
inline dbtuple::tuple_writer_t
get_writer() const
{
return w;
}
private:
dbtuple *tuple;
const string_type *k;
const void *r;
dbtuple::tuple_writer_t w;
marked_ptr<concurrent_btree> btr; // first bit for inserted, 2nd for dowrite
};
friend std::ostream &
operator<<(std::ostream &o, const write_record_t &r);
// the absent set is a mapping from (btree_node -> version_number).
struct absent_record_t { uint64_t version; };
friend std::ostream &
operator<<(std::ostream &o, const absent_record_t &r);
struct dbtuple_write_info {
enum {
FLAGS_LOCKED = 0x1,
FLAGS_INSERT = 0x1 << 1,
};
dbtuple_write_info() : tuple(), entry(nullptr), pos() {}
dbtuple_write_info(dbtuple *tuple, write_record_t *entry,
bool is_insert, size_t pos)
: tuple(tuple), entry(entry), pos(pos)
{
if (is_insert)
this->tuple.set_flags(FLAGS_LOCKED | FLAGS_INSERT);
}
// XXX: for searching only
explicit dbtuple_write_info(const dbtuple *tuple)
: tuple(const_cast<dbtuple *>(tuple)), entry(), pos() {}
inline dbtuple *
get_tuple()
{
return tuple.get();
}
inline const dbtuple *
get_tuple() const
{
return tuple.get();
}
inline ALWAYS_INLINE void
mark_locked()
{
INVARIANT(!is_locked());
tuple.or_flags(FLAGS_LOCKED);
INVARIANT(is_locked());
}
inline ALWAYS_INLINE bool
is_locked() const
{
return tuple.get_flags() & FLAGS_LOCKED;
}
inline ALWAYS_INLINE bool
is_insert() const
{
return tuple.get_flags() & FLAGS_INSERT;
}
inline ALWAYS_INLINE
bool operator<(const dbtuple_write_info &o) const
{
// the unique key is [tuple, !is_insert, pos]
return tuple < o.tuple ||
(tuple == o.tuple && !is_insert() < !o.is_insert()) ||
(tuple == o.tuple && !is_insert() == !o.is_insert() && pos < o.pos);
}
marked_ptr<dbtuple> tuple;
write_record_t *entry;
size_t pos;
};
static event_counter g_evt_read_logical_deleted_node_search;
static event_counter g_evt_read_logical_deleted_node_scan;
static event_counter g_evt_dbtuple_write_search_failed;
static event_counter g_evt_dbtuple_write_insert_failed;
static event_counter evt_local_search_lookups;
static event_counter evt_local_search_write_set_hits;
static event_counter evt_dbtuple_latest_replacement;
CLASS_STATIC_COUNTER_DECL(scopedperf::tsc_ctr, g_txn_commit_probe0, g_txn_commit_probe0_cg);
CLASS_STATIC_COUNTER_DECL(scopedperf::tsc_ctr, g_txn_commit_probe1, g_txn_commit_probe1_cg);
CLASS_STATIC_COUNTER_DECL(scopedperf::tsc_ctr, g_txn_commit_probe2, g_txn_commit_probe2_cg);
CLASS_STATIC_COUNTER_DECL(scopedperf::tsc_ctr, g_txn_commit_probe3, g_txn_commit_probe3_cg);
CLASS_STATIC_COUNTER_DECL(scopedperf::tsc_ctr, g_txn_commit_probe4, g_txn_commit_probe4_cg);
CLASS_STATIC_COUNTER_DECL(scopedperf::tsc_ctr, g_txn_commit_probe5, g_txn_commit_probe5_cg);
CLASS_STATIC_COUNTER_DECL(scopedperf::tsc_ctr, g_txn_commit_probe6, g_txn_commit_probe6_cg);
txn_state state;
abort_reason reason;
const uint64_t flags;
};
// type specializations
namespace private_ {
template <>
struct is_trivially_destructible<transaction_base::read_record_t> {
static const bool value = true;
};
template <>
struct is_trivially_destructible<transaction_base::write_record_t> {
static const bool value = true;
};
template <>
struct is_trivially_destructible<transaction_base::absent_record_t> {
static const bool value = true;
};
template <>
struct is_trivially_destructible<transaction_base::dbtuple_write_info> {
static const bool value = true;
};
}
inline ALWAYS_INLINE std::ostream &
operator<<(std::ostream &o, const transaction_base::read_record_t &r)
{
//o << "[tuple=" << util::hexify(r.get_tuple())
o << "[tuple=" << *r.get_tuple()
<< ", tid_read=" << g_proto_version_str(r.get_tid())
<< "]";
return o;
}
inline ALWAYS_INLINE std::ostream &
operator<<(
std::ostream &o,
const transaction_base::write_record_t &r)
{
o << "[tuple=" << r.get_tuple()
<< ", key=" << util::hexify(r.get_key())
<< ", value=" << util::hexify(r.get_value())
<< ", insert=" << r.is_insert()
<< ", do_write=" << r.do_write()
<< ", btree=" << r.get_btree()
<< "]";
return o;
}
inline ALWAYS_INLINE std::ostream &
operator<<(std::ostream &o, const transaction_base::absent_record_t &r)
{
o << "[v=" << r.version << "]";
return o;
}
struct default_transaction_traits {
static const size_t read_set_expected_size = SMALL_SIZE_MAP;
static const size_t absent_set_expected_size = EXTRA_SMALL_SIZE_MAP;
static const size_t write_set_expected_size = SMALL_SIZE_MAP;
static const bool stable_input_memory = false;
static const bool hard_expected_sizes = false; // true if the expected sizes are hard maximums
static const bool read_own_writes = true; // if we read a key which we previous put(), are we guaranteed
// to read our latest (uncommited) values? this comes at a
// performance penality [you should not need this behavior to
// write txns, since you *know* the values you inserted]
typedef util::default_string_allocator StringAllocator;
};
struct default_stable_transaction_traits : public default_transaction_traits {
static const bool stable_input_memory = true;
};
template <template <typename> class Protocol, typename Traits>
class transaction : public transaction_base {
// XXX: weaker than necessary
template <template <typename> class, typename>
friend class base_txn_btree;
friend Protocol<Traits>;
public:
// KeyWriter is expected to implement:
// [1-arg constructor]
// KeyWriter(const Key *)
// [fully materialize]
// template <typename StringAllocator>
// const std::string * fully_materialize(bool, StringAllocator &)
// ValueWriter is expected to implement:
// [1-arg constructor]
// ValueWriter(const Value *, ValueInfo)
// [compute new size from old value]
// size_t compute_needed(const uint8_t *, size_t)
// [fully materialize]
// template <typename StringAllocator>
// const std::string * fully_materialize(bool, StringAllocator &)
// [perform write]
// void operator()(uint8_t *, size_t)
//
// ValueWriter does not have to be move/copy constructable. The value passed
// into the ValueWriter constructor is guaranteed to be valid throughout the
// lifetime of a ValueWriter instance.
// KeyReader Interface
//
// KeyReader is a simple transformation from (const std::string &) => const Key &.
// The input is guaranteed to be stable, so it has a simple interface:
//
// const Key &operator()(const std::string &)
//
// The KeyReader is expect to preserve the following property: After a call
// to operator(), but before the next, the returned value is guaranteed to be
// valid and remain stable.
// ValueReader Interface
//
// ValueReader is a more complex transformation from (const uint8_t *, size_t) => Value &.
// The input is not guaranteed to be stable, so it has a more complex interface:
//
// template <typename StringAllocator>
// bool operator()(const uint8_t *, size_t, StringAllocator &)
//
// This interface returns false if there was not enough buffer space to
// finish the read, true otherwise. Note that this interface returning true
// does NOT mean that a read was stable, but it just means there were enough
// bytes in the buffer to perform the tentative read.
//
// Note that ValueReader also exposes a dup interface
//
// template <typename StringAllocator>
// void dup(const Value &, StringAllocator &)
//
// ValueReader also exposes a means to fetch results:
//
// Value &results()
//
// The ValueReader is expected to preserve the following property: After a
// call to operator(), if it returns true, then the value returned from
// results() should remain valid and stable until the next call to
// operator().
//typedef typename P::Key key_type;
//typedef typename P::Value value_type;
//typedef typename P::ValueInfo value_info_type;
//typedef typename P::KeyWriter key_writer_type;
//typedef typename P::ValueWriter value_writer_type;
//typedef typename P::KeyReader key_reader_type;
//typedef typename P::SingleValueReader single_value_reader_type;
//typedef typename P::ValueReader value_reader_type;
typedef Traits traits_type;
typedef typename Traits::StringAllocator string_allocator_type;
protected:
// data structures
inline ALWAYS_INLINE Protocol<Traits> *
cast()
{
return static_cast<Protocol<Traits> *>(this);
}
inline ALWAYS_INLINE const Protocol<Traits> *
cast() const
{
return static_cast<const Protocol<Traits> *>(this);
}
// XXX: we have baked in b-tree into the protocol- other indexes are possible
// but we would need to abstract it away. we don't bother for now.
#ifdef USE_SMALL_CONTAINER_OPT
// XXX: use parameterized typedef to avoid duplication
// small types
typedef small_vector<
read_record_t,
traits_type::read_set_expected_size> read_set_map_small;
typedef small_vector<
write_record_t,
traits_type::write_set_expected_size> write_set_map_small;
typedef small_unordered_map<
const typename concurrent_btree::node_opaque_t *, absent_record_t,
traits_type::absent_set_expected_size> absent_set_map_small;
// static types
typedef static_vector<
read_record_t,
traits_type::read_set_expected_size> read_set_map_static;
typedef static_vector<
write_record_t,
traits_type::write_set_expected_size> write_set_map_static;
typedef static_unordered_map<
const typename concurrent_btree::node_opaque_t *, absent_record_t,
traits_type::absent_set_expected_size> absent_set_map_static;
// helper types for log writing
typedef small_vector<
uint32_t,
traits_type::write_set_expected_size> write_set_u32_vec_small;
typedef static_vector<
uint32_t,
traits_type::write_set_expected_size> write_set_u32_vec_static;
// use static types if the expected sizes are guarantees
typedef
typename std::conditional<
traits_type::hard_expected_sizes,
read_set_map_static, read_set_map_small>::type read_set_map;
typedef
typename std::conditional<
traits_type::hard_expected_sizes,
write_set_map_static, write_set_map_small>::type write_set_map;
typedef
typename std::conditional<
traits_type::hard_expected_sizes,
absent_set_map_static, absent_set_map_small>::type absent_set_map;
typedef
typename std::conditional<
traits_type::hard_expected_sizes,
write_set_u32_vec_static, write_set_u32_vec_small>::type write_set_u32_vec;
#else
typedef std::vector<read_record_t> read_set_map;
typedef std::vector<write_record_t> write_set_map;
typedef std::vector<absent_record_t> absent_set_map;
typedef std::vector<uint32_t> write_set_u32_vec;
#endif
template <typename T>
using write_set_sized_vec =
typename std::conditional<
traits_type::hard_expected_sizes,
static_vector<T, traits_type::write_set_expected_size>,
typename util::vec<T, traits_type::write_set_expected_size>::type
>::type;
// small type
typedef
typename util::vec<
dbtuple_write_info, traits_type::write_set_expected_size>::type
dbtuple_write_info_vec_small;
// static type
typedef
static_vector<
dbtuple_write_info, traits_type::write_set_expected_size>
dbtuple_write_info_vec_static;
// chosen type
typedef
typename std::conditional<
traits_type::hard_expected_sizes,
dbtuple_write_info_vec_static, dbtuple_write_info_vec_small>::type
dbtuple_write_info_vec;
static inline bool
sorted_dbtuples_contains(
const dbtuple_write_info_vec &dbtuples,
const dbtuple *tuple)
{
// XXX: skip binary search for small-sized dbtuples?
return std::binary_search(
dbtuples.begin(), dbtuples.end(),
dbtuple_write_info(tuple),
[](const dbtuple_write_info &lhs, const dbtuple_write_info &rhs)
{ return lhs.get_tuple() < rhs.get_tuple(); });
}
public:
inline transaction(uint64_t flags, string_allocator_type &sa);
inline ~transaction();
// returns TRUE on successful commit, FALSE on abort
// if doThrow, signals success by returning true, and
// failure by throwing an abort exception
bool commit(bool doThrow = false);
// abort() always succeeds
inline void
abort()
{
abort_impl(ABORT_REASON_USER);
}
void dump_debug_info() const;
#ifdef DIE_ON_ABORT
void
abort_trap(abort_reason reason)
{
AbortReasonCounter(reason)->inc();
this->reason = reason; // for dump_debug_info() to see
dump_debug_info();
::abort();
}
#else
inline ALWAYS_INLINE void
abort_trap(abort_reason reason)
{
AbortReasonCounter(reason)->inc();
}
#endif
std::map<std::string, uint64_t> get_txn_counters() const;
inline ALWAYS_INLINE bool
is_snapshot() const
{
return get_flags() & TXN_FLAG_READ_ONLY;
}
// for debugging purposes only
inline const read_set_map &
get_read_set() const
{
return read_set;
}
inline const write_set_map &
get_write_set() const
{
return write_set;
}
inline const absent_set_map &
get_absent_set() const
{
return absent_set;
}
protected:
inline void abort_impl(abort_reason r);
// assumes lock on marker is held on marker by caller, and marker is the
// latest: removes marker from tree, and clears latest
void cleanup_inserted_tuple_marker(
dbtuple *marker, const std::string &key,
concurrent_btree *btr);
// low-level API for txn_btree
// try to insert a new "tentative" tuple into the underlying
// btree associated with the given context.
//
// if return.first is not null, then this function will
// 1) mutate the transaction such that the absent_set is aware of any
// mutating changes made to the underlying btree.
// 2) add the new tuple to the write_set
//
// if return.second is true, then this txn should abort, because a conflict
// was detected w/ the absent_set.
//
// if return.first is not null, the returned tuple is locked()!
//
// if the return.first is null, then this function has no side effects.
//
// NOTE: !ret.first => !ret.second
// NOTE: assumes key/value are stable
std::pair< dbtuple *, bool >
try_insert_new_tuple(
concurrent_btree &btr,
const std::string *key,
const void *value,
dbtuple::tuple_writer_t writer);
// reads the contents of tuple into v
// within this transaction context
template <typename ValueReader>
bool
do_tuple_read(const dbtuple *tuple, ValueReader &value_reader);
void
do_node_read(const typename concurrent_btree::node_opaque_t *n, uint64_t version);
public:
// expected public overrides
/**
* Can we overwrite prev with cur?
*/
bool can_overwrite_record_tid(tid_t prev, tid_t cur) const;
inline string_allocator_type &
string_allocator()
{
return *sa;
}
protected:
// expected protected overrides
/**
* create a new, unique TID for a txn. at the point which gen_commit_tid(),
* it still has not been decided whether or not this txn will commit
* successfully
*/
tid_t gen_commit_tid(const dbtuple_write_info_vec &write_tuples);
bool can_read_tid(tid_t t) const;
// For GC handlers- note that on_dbtuple_spill() is called
// with the lock on ln held, to simplify GC code
//
// Is also called within an RCU read region
void on_dbtuple_spill(dbtuple *tuple_ahead, dbtuple *tuple);
// Called when the latest value written to ln is an empty
// (delete) marker. The protocol can then decide how to schedule
// the logical node for actual deletion
void on_logical_delete(dbtuple *tuple, const std::string &key, concurrent_btree *btr);
// if gen_commit_tid() is called, then on_tid_finish() will be called
// with the commit tid. before on_tid_finish() is called, state is updated
// with the resolution (commited, aborted) of this txn
void on_tid_finish(tid_t commit_tid);
void on_post_rcu_region_completion();
protected:
inline void clear();
// SLOW accessor methods- used for invariant checking
typename read_set_map::iterator
find_read_set(const dbtuple *tuple)
{
// linear scan- returns the *first* entry found
// (a tuple can exist in the read_set more than once)
typename read_set_map::iterator it = read_set.begin();
typename read_set_map::iterator it_end = read_set.end();
for (; it != it_end; ++it)
if (it->get_tuple() == tuple)
break;
return it;
}
inline typename read_set_map::const_iterator
find_read_set(const dbtuple *tuple) const
{
return const_cast<transaction *>(this)->find_read_set(tuple);
}
typename write_set_map::iterator
find_write_set(dbtuple *tuple)
{
// linear scan- returns the *first* entry found
// (a tuple can exist in the write_set more than once)
typename write_set_map::iterator it = write_set.begin();
typename write_set_map::iterator it_end = write_set.end();
for (; it != it_end; ++it)
if (it->get_tuple() == tuple)
break;
return it;
}
inline typename write_set_map::const_iterator
find_write_set(const dbtuple *tuple) const
{
return const_cast<transaction *>(this)->find_write_set(tuple);
}
inline bool
handle_last_tuple_in_group(
dbtuple_write_info &info, bool did_group_insert);
read_set_map read_set;
write_set_map write_set;
absent_set_map absent_set;
string_allocator_type *sa;
unmanaged<scoped_rcu_region> rcu_guard_;
};
class transaction_abort_exception : public std::exception {
public:
transaction_abort_exception(transaction_base::abort_reason r)
: r(r) {}
inline transaction_base::abort_reason
get_reason() const
{
return r;
}
virtual const char *
what() const throw()
{
return transaction_base::AbortReasonStr(r);
}
private:
transaction_base::abort_reason r;
};
// XXX(stephentu): stupid hacks
// XXX(stephentu): txn_epoch_sync is a misnomer
template <template <typename> class Transaction>
struct txn_epoch_sync {
// block until the next epoch
static inline void sync() {}
// finish any async jobs
static inline void finish() {}
// run this code when a benchmark worker finishes
static inline void thread_end() {}
// how many txns have we persisted in total, from
// the last reset invocation?
static inline std::pair<uint64_t, double>
compute_ntxn_persisted() { return {0, 0.0}; }
// reset the persisted counters
static inline void reset_ntxn_persisted() {}
};
#endif /* _NDB_TXN_H_ */