summaryrefslogtreecommitdiffstats
path: root/xlators/cluster/afr/src/afr.h
blob: 855d3a3680e0b35cd2b2c33a0fb74f79c50c84cc (plain)
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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
/*
  Copyright (c) 2008-2012 Red Hat, Inc. <http://www.redhat.com>
  This file is part of GlusterFS.

  This file is licensed to you under your choice of the GNU Lesser
  General Public License, version 3 or any later version (LGPLv3 or
  later), or the GNU General Public License, version 2 (GPLv2), in all
  cases as published by the Free Software Foundation.
*/


#ifndef __AFR_H__
#define __AFR_H__

#ifndef _CONFIG_H
#define _CONFIG_H
#include "config.h"
#endif

#include "call-stub.h"
#include "compat-errno.h"
#include "afr-mem-types.h"

#include "libxlator.h"
#include "timer.h"
#include "syncop.h"

#include "afr-self-heald.h"
#include "afr-messages.h"

#define AFR_XATTR_PREFIX "trusted.afr"
#define AFR_PATHINFO_HEADER "REPLICATE:"
#define AFR_SH_READDIR_SIZE_KEY "self-heal-readdir-size"
#define AFR_SH_DATA_DOMAIN_FMT "%s:self-heal"
#define AFR_DIRTY_DEFAULT AFR_XATTR_PREFIX ".dirty"
#define AFR_DIRTY (((afr_private_t *) (THIS->private))->afr_dirty)

#define AFR_LOCKEE_COUNT_MAX    3
#define AFR_DOM_COUNT_MAX    3
#define AFR_NUM_CHANGE_LOGS            3 /*data + metadata + entry*/
#define AFR_DEFAULT_SPB_CHOICE_TIMEOUT 300 /*in seconds*/

#define ARBITER_BRICK_INDEX 2

typedef int (*afr_lock_cbk_t) (call_frame_t *frame, xlator_t *this);

typedef int (*afr_read_txn_wind_t) (call_frame_t *frame, xlator_t *this, int subvol);

typedef int (*afr_inode_refresh_cbk_t) (call_frame_t *frame, xlator_t *this, int err);

typedef int (*afr_changelog_resume_t) (call_frame_t *frame, xlator_t *this);

#define alloca0(size) ({void *__ptr; __ptr = alloca(size); memset(__ptr, 0, size); __ptr;})
#define AFR_COUNT(array,max) ({int __i; int __res = 0; for (__i = 0; __i < max; __i++) if (array[__i]) __res++; __res;})
#define AFR_INTERSECT(dst,src1,src2,max) ({int __i; for (__i = 0; __i < max; __i++) dst[__i] = src1[__i] && src2[__i];})
#define AFR_CMP(a1,a2,len) ({int __cmp = 0; int __i; for (__i = 0; __i < len; __i++) if (a1[__i] != a2[__i]) { __cmp = 1; break;} __cmp;})

typedef struct _afr_private {
        gf_lock_t lock;               /* to guard access to child_count, etc */
        unsigned int child_count;     /* total number of children   */
        unsigned int arbiter_count;   /*subset of child_count.
                                        Has to be 0 or 1.*/

        xlator_t **children;

        inode_t *root_inode;

        unsigned char *child_up;

        char **pending_key;

        char         *data_self_heal;              /* on/off/open */
        char *       data_self_heal_algorithm;    /* name of algorithm */
        unsigned int data_self_heal_window_size;  /* max number of pipelined
                                                     read/writes */

        unsigned int background_self_heal_count;
        unsigned int background_self_heals_started;
        gf_boolean_t metadata_self_heal;   /* on/off */
        gf_boolean_t entry_self_heal;      /* on/off */

        gf_boolean_t data_change_log;       /* on/off */
        gf_boolean_t metadata_change_log;   /* on/off */
        gf_boolean_t entry_change_log;      /* on/off */

	gf_boolean_t metadata_splitbrain_forced_heal; /* on/off */
        int read_child;               /* read-subvolume */
        unsigned int hash_mode;       /* for when read_child is not set */
        int favorite_child;  /* subvolume to be preferred in resolving
                                         split-brain cases */

        gf_boolean_t inodelk_trace;
        gf_boolean_t entrylk_trace;

        unsigned int wait_count;      /* # of servers to wait for success */

        uint64_t up_count;      /* number of CHILD_UPs we have seen */
        uint64_t down_count;    /* number of CHILD_DOWNs we have seen */

        gf_boolean_t      optimistic_change_log;
        gf_boolean_t      eager_lock;
        gf_boolean_t      pre_op_compat;      /* on/off */
	uint32_t          post_op_delay_secs;
        unsigned int      quorum_count;
        gf_boolean_t      quorum_reads;

        char                   vol_uuid[UUID_SIZE + 1];
        int32_t                *last_event;

	/* @event_generation: Keeps count of number of events received which can
	   potentially impact consistency decisions. The events are CHILD_UP
	   and CHILD_DOWN, when we have to recalculate the freshness/staleness
	   of copies to detect if changes had happened while the other server
	   was down. CHILD_DOWN and CHILD_UP can also be received on network
	   disconnect/reconnects and not necessarily server going down/up.
	   Recalculating freshness/staleness on network events is equally
	   important as we might have had a network split brain.
	*/
	uint32_t               event_generation;

        gf_boolean_t           choose_local;
        gf_boolean_t           did_discovery;
        uint64_t               sh_readdir_size;
        gf_boolean_t           ensure_durability;
        char                   *sh_domain;
	char                   *afr_dirty;

	afr_self_heald_t       shd;

	/* pump dependencies */
	void                   *pump_private;
	gf_boolean_t           use_afr_in_pump;
        gf_boolean_t           consistent_metadata;
        uint64_t               spb_choice_timeout;
} afr_private_t;


typedef enum {
        AFR_DATA_TRANSACTION,          /* truncate, write, ... */
        AFR_METADATA_TRANSACTION,      /* chmod, chown, ... */
        AFR_ENTRY_TRANSACTION,         /* create, rmdir, ... */
        AFR_ENTRY_RENAME_TRANSACTION,  /* rename */
} afr_transaction_type;

typedef enum {
        AFR_TRANSACTION_LK,
        AFR_SELFHEAL_LK,
} transaction_lk_type_t;

typedef enum {
        AFR_LOCK_OP,
        AFR_UNLOCK_OP,
} afr_lock_op_type_t;

typedef enum {
        AFR_DATA_SELF_HEAL_LK,
        AFR_METADATA_SELF_HEAL_LK,
        AFR_ENTRY_SELF_HEAL_LK,
}selfheal_lk_type_t;

typedef enum {
        AFR_INODELK_TRANSACTION,
        AFR_INODELK_NB_TRANSACTION,
        AFR_ENTRYLK_TRANSACTION,
        AFR_ENTRYLK_NB_TRANSACTION,
        AFR_INODELK_SELFHEAL,
        AFR_INODELK_NB_SELFHEAL,
        AFR_ENTRYLK_SELFHEAL,
        AFR_ENTRYLK_NB_SELFHEAL,
} afr_lock_call_type_t;

/*
  xattr format: trusted.afr.volume = [x y z]
  x - data pending
  y - metadata pending
  z - entry pending
*/

static inline int
afr_index_for_transaction_type (afr_transaction_type type)
{
        switch (type) {

        case AFR_DATA_TRANSACTION:
                return 0;

        case AFR_METADATA_TRANSACTION:
                return 1;

        case AFR_ENTRY_TRANSACTION:
        case AFR_ENTRY_RENAME_TRANSACTION:
                return 2;
        }

        return -1;  /* make gcc happy */
}

static inline int
afr_index_from_ia_type (ia_type_t type)
{
        switch (type) {
        case IA_IFDIR:
                return afr_index_for_transaction_type (AFR_ENTRY_TRANSACTION);
        case IA_IFREG:
                return afr_index_for_transaction_type (AFR_DATA_TRANSACTION);
        default: return -1;
        }
}

typedef struct {
        loc_t                   loc;
        char                    *basename;
        unsigned char           *locked_nodes;
        int                     locked_count;

} afr_entry_lockee_t;

int
afr_entry_lockee_cmp (const void *l1, const void *l2);

typedef struct {
        char    *domain; /* Domain on which inodelk is taken */
        struct gf_flock flock;
        unsigned char *locked_nodes;
        int32_t lock_count;
} afr_inodelk_t;

typedef struct {
        loc_t *lk_loc;

        int                     lockee_count;
        afr_entry_lockee_t      lockee[AFR_LOCKEE_COUNT_MAX];

        afr_inodelk_t       inodelk[AFR_DOM_COUNT_MAX];
        const char *lk_basename;
        const char *lower_basename;
        const char *higher_basename;
        char lower_locked;
        char higher_locked;

        unsigned char *locked_nodes;
        unsigned char *lower_locked_nodes;

        selfheal_lk_type_t selfheal_lk_type;
        transaction_lk_type_t transaction_lk_type;

        int32_t lock_count;
        int32_t entrylk_lock_count;

        uint64_t lock_number;
        int32_t lk_call_count;
        int32_t lk_expected_count;
        int32_t lk_attempted_count;

        int32_t lock_op_ret;
        int32_t lock_op_errno;
        afr_lock_cbk_t lock_cbk;
        char *domain; /* Domain on which inode/entry lock/unlock in progress.*/
} afr_internal_lock_t;

struct afr_reply {
	int	valid;
	int32_t	op_ret;
	int32_t	op_errno;
	dict_t *xdata;
	struct iatt poststat;
	struct iatt postparent;
	struct iatt prestat;
	struct iatt preparent;
	struct iatt preparent2;
	struct iatt postparent2;
	uint8_t checksum[MD5_DIGEST_LENGTH];
};

typedef enum {
        AFR_FD_NOT_OPENED,
        AFR_FD_OPENED,
        AFR_FD_OPENING
} afr_fd_open_status_t;

typedef struct {
        unsigned int *pre_op_done[AFR_NUM_CHANGE_LOGS];
	int inherited[AFR_NUM_CHANGE_LOGS];
	int on_disk[AFR_NUM_CHANGE_LOGS];
        afr_fd_open_status_t *opened_on; /* which subvolumes the fd is open on */

        unsigned int *lock_piggyback;
        unsigned int *lock_acquired;

        int flags;

	/* used for delayed-post-op optimization */
	pthread_mutex_t    delay_lock;
	gf_timer_t        *delay_timer;
	call_frame_t      *delay_frame;

	/* set if any write on this fd was a non stable write
	   (i.e, without O_SYNC or O_DSYNC)
	*/
	gf_boolean_t      witnessed_unstable_write;

	/* @open_fd_count:
	   Number of open FDs queried from the server, as queried through
	   xdata in FOPs. Currently, used to decide if eager-locking must be
	   temporarily disabled.
	*/
        uint32_t        open_fd_count;


	/* list of frames currently in progress */
	struct list_head  eager_locked;

	/* the subvolume on which the latest sequence of readdirs (starting
	   at offset 0) has begun. Till the next readdir request with 0 offset
	   arrives, we continue to read off this subvol.
	*/
	int readdir_subvol;
} afr_fd_ctx_t;


typedef struct _afr_local {
	glusterfs_fop_t  op;
        unsigned int call_count;

	/* @event_generation: copy of priv->event_generation taken at the
	   time of starting the transaction. The copy is made so that we
	   have a stable value through the various phases of the transaction.
	*/
	unsigned int event_generation;

        uint32_t     open_fd_count;
        gf_boolean_t update_open_fd_count;

	gf_lkowner_t  saved_lk_owner;

        int32_t op_ret;
        int32_t op_errno;

        int32_t **pending;

	int dirty[AFR_NUM_CHANGE_LOGS];

        loc_t loc;
        loc_t newloc;

        fd_t *fd;
	afr_fd_ctx_t *fd_ctx;

	/* @child_up: copy of priv->child_up taken at the time of transaction
	   start. The copy is taken so that we have a stable child_up array
	   through the phases of the transaction as priv->child_up[i] can keep
	   changing through time.
	*/
        unsigned char *child_up;

	/* @read_attempted:
	   array of flags representing subvolumes where read operations of
	   the read transaction have already been attempted. The array is
	   first pre-filled with down subvolumes, and as reads are performed
	   on other subvolumes, those are set as well. This way if the read
	   operation fails we do not retry on that subvolume again.
	*/
	unsigned char *read_attempted;

	/* @readfn:

	   pointer to function which will perform the read operation on a given
	   subvolume. Used in read transactions.
	*/

	afr_read_txn_wind_t readfn;

	/* @refreshed:

	   the inode was "refreshed" (i.e, pending xattrs from all subvols
	   freshly inspected and inode ctx updated accordingly) as part of
	   this transaction already.
	*/
	gf_boolean_t refreshed;

	/* @inode:

	   the inode on which the read txn is performed on. ref'ed and copied
	   from either fd->inode or loc.inode
	*/

	inode_t *inode;

	/* @parent[2]:

	   parent inode[s] on which directory transactions are performed.
	*/

	inode_t *parent;
	inode_t *parent2;

	/* @readable:

	   array of flags representing servers from which a read can be
	   performed. This is the output of afr_inode_refresh()
	*/
	unsigned char *readable;

	afr_inode_refresh_cbk_t refreshfn;

	/* @refreshinode:

	   Inode currently getting refreshed.
	*/
	inode_t *refreshinode;

	/*
	  @pre_op_compat:

	  compatibility mode of pre-op. send a separate pre-op and
	  op operations as part of transaction, rather than combining
	*/

	gf_boolean_t pre_op_compat;

        dict_t  *xattr_req;

        afr_internal_lock_t internal_lock;

        dict_t  *dict;

        int      optimistic_change_log;
	gf_boolean_t      delayed_post_op;

	/* Is the current writev() going to perform a stable write?
	   i.e, is fd->flags or @flags writev param have O_SYNC or
	   O_DSYNC?
	*/
	gf_boolean_t      stable_write;

	/* This write appended to the file. Nnot necessarily O_APPEND,
	   just means the offset of write was at the end of file.
	*/
	gf_boolean_t      append_write;

        /*
          This struct contains the arguments for the "continuation"
          (scheme-like) of fops
        */

        struct {
                struct {
                        gf_boolean_t needs_fresh_lookup;
                        uuid_t gfid_req;
                } lookup;

                struct {
                        unsigned char buf_set;
                        struct statvfs buf;
                } statfs;

                struct {
                        int32_t flags;
                } open;

                struct {
                        int32_t cmd;
                        struct gf_flock user_flock;
                        struct gf_flock ret_flock;
                        unsigned char *locked_nodes;
                } lk;

                /* inode read */

                struct {
                        int32_t mask;
                        int last_index;  /* index of the child we tried previously */
                } access;

                struct {
                        int last_index;
                } stat;

                struct {
                        int last_index;
                } fstat;

                struct {
                        size_t size;
                        int last_index;
                } readlink;

                struct {
                        char *name;
                        int last_index;
                        long xattr_len;
                } getxattr;

                struct {
                        size_t size;
                        off_t offset;
                        int last_index;
                        uint32_t flags;
                } readv;

                /* dir read */

                struct {
                        int success_count;
                        int32_t op_ret;
                        int32_t op_errno;

                        uint32_t *checksum;
                } opendir;

                struct {
                        int32_t op_ret;
                        int32_t op_errno;
                        size_t size;
                        off_t offset;
                        dict_t *dict;
                        gf_boolean_t failed;
                        int last_index;
                } readdir;
                /* inode write */

                struct {
                        struct iatt prebuf;
                        struct iatt postbuf;
                } inode_wfop; //common structure for all inode-write-fops

                struct {
                        int32_t op_ret;

                        struct iovec *vector;
                        struct iobref *iobref;
                        int32_t count;
                        off_t offset;
                        uint32_t flags;
                } writev;

                struct {
                        off_t offset;
                } truncate;

                struct {
                        off_t offset;
                } ftruncate;

                struct {
                        struct iatt in_buf;
                        int32_t valid;
                } setattr;

                struct {
                        struct iatt in_buf;
                        int32_t valid;
                } fsetattr;

                struct {
                        dict_t *dict;
                        int32_t flags;
                } setxattr;

                struct {
                        dict_t *dict;
                        int32_t flags;
                } fsetxattr;

                struct {
                        char *name;
                } removexattr;

                struct {
                        dict_t *xattr;
                } xattrop;

                struct {
                        dict_t *xattr;
                } fxattrop;

                /* dir write */

                struct {
                        inode_t *inode;
                        struct iatt buf;
                        struct iatt preparent;
                        struct iatt postparent;
                        struct iatt prenewparent;
                        struct iatt postnewparent;
                } dir_fop; //common structure for all dir fops

                struct {
                        fd_t *fd;
                        dict_t *params;
                        int32_t flags;
                        mode_t mode;
                } create;

                struct {
                        dev_t dev;
                        mode_t mode;
                        dict_t *params;
                } mknod;

                struct {
                        int32_t mode;
                        dict_t *params;
                } mkdir;

                struct {
                        int flags;
                } rmdir;

                struct {
                        dict_t *params;
                        char *linkpath;
                } symlink;

		struct {
			int32_t mode;
			off_t offset;
			size_t len;
		} fallocate;

		struct {
			off_t offset;
			size_t len;
		} discard;

                struct {
                        off_t offset;
                        off_t len;
                        struct iatt prebuf;
                        struct iatt postbuf;
                } zerofill;

                struct {
                        char *volume;
                        int32_t cmd;
                        struct gf_flock flock;
                } inodelk;

        } cont;

        struct {
                off_t start, len;

                gf_boolean_t    eager_lock_on;
                int *eager_lock;

                char *basename;
                char *new_basename;

                loc_t parent_loc;
                loc_t new_parent_loc;

                afr_transaction_type type;

		/* stub to resume on destruction
		   of the transaction frame */
		call_stub_t      *resume_stub;

		struct list_head  eager_locked;

                unsigned char   *pre_op;

                /* For arbiter configuration only. */
                dict_t **pre_op_xdata;
                unsigned char *pre_op_sources;

		/* @fop_subvols: subvolumes on which FOP will be attempted */
                unsigned char   *fop_subvols;

		/* @failed_subvols: subvolumes on which FOP failed. Always
		   a subset of @fop_subvols */
                unsigned char   *failed_subvols;

		/* @dirtied: flag which indicates whether we set dirty flag
		   in the OP. Typically true when we are performing operation
		   on more than one subvol and optimistic changelog is disabled

		   A 'true' value set in @dirtied flag means an 'undirtying'
		   has to be done in POST-OP phase.
		*/
		gf_boolean_t  dirtied;

		/* @inherited: flag which indicates that the dirty flags
		   of the previous transaction were inherited
		*/
		gf_boolean_t  inherited;

		/*
		  @no_uninherit: flag which indicates that a pre_op_uninherit()
		  must _not_ be attempted (and returned as failure) always. This
		  flag is set when a hard pre-op is performed, but not accounted
		  for it in fd_ctx->on_disk[]. Such transactions are "isolated"
		  from the pre-op piggybacking entirely and therefore uninherit
		  must not be attempted.
		*/
		gf_boolean_t no_uninherit;

		/* @uninherit_done:
		   @uninherit_value:

		   The above pair variables make pre_op_uninherit() idempotent.
		   Both are FALSE initially. The first call to pre_op_uninherit
		   sets @uninherit_done to TRUE and the return value to
		   @uninherit_value. Further calls will check for @uninherit_done
		   to be TRUE and if so will simply return @uninherit_value.
		*/
		gf_boolean_t uninherit_done;
		gf_boolean_t uninherit_value;

		/* @changelog_resume: function to be called after changlogging
		   (either pre-op or post-op) is done
		*/

		afr_changelog_resume_t changelog_resume;

                call_frame_t *main_frame;

                int (*wind) (call_frame_t *frame, xlator_t *this, int subvol);

                int (*fop) (call_frame_t *frame, xlator_t *this);

                int (*done) (call_frame_t *frame, xlator_t *this);

                int (*resume) (call_frame_t *frame, xlator_t *this);

                int (*unwind) (call_frame_t *frame, xlator_t *this);

                /* post-op hook */
        } transaction;

	syncbarrier_t barrier;

        /* extra data for fops */
        dict_t         *xdata_req;
        dict_t         *xdata_rsp;

        mode_t          umask;
        int             xflag;
        gf_boolean_t    do_discovery;
	struct afr_reply *replies;
} afr_local_t;


typedef struct _afr_inode_ctx {
        uint64_t        read_subvol;
        int             spb_choice;
        gf_timer_t      *timer;
} afr_inode_ctx_t;

typedef struct afr_spbc_timeout {
        call_frame_t *frame;
        gf_boolean_t d_spb;
        gf_boolean_t m_spb;
        loc_t        *loc;
        int          spb_child_index;
} afr_spbc_timeout_t;

/* did a call fail due to a child failing? */
#define child_went_down(op_ret, op_errno) (((op_ret) < 0) &&            \
                                           ((op_errno == ENOTCONN) ||   \
                                            (op_errno == EBADFD)))

int
afr_inode_read_subvol_get (inode_t *inode, xlator_t *this,
			   unsigned char *data_subvols,
			   unsigned char *metadata_subvols,
			   int *event_generation);
int
__afr_inode_read_subvol_get (inode_t *inode, xlator_t *this,
			     unsigned char *data_subvols,
			     unsigned char *metadata_subvols,
			     int *event_generation);

int
__afr_inode_read_subvol_set (inode_t *inode, xlator_t *this,
			     unsigned char *data_subvols,
			     unsigned char *metadata_subvol,
			     int event_generation);
int
afr_inode_read_subvol_set (inode_t *inode, xlator_t *this,
			   unsigned char *data_subvols,
			   unsigned char *metadata_subvols,
			   int event_generation);

int
afr_inode_read_subvol_reset (inode_t *inode, xlator_t *this);

int
afr_read_subvol_select_by_policy (inode_t *inode, xlator_t *this,
				  unsigned char *readable);

int
afr_inode_read_subvol_type_get (inode_t *inode, xlator_t *this,
				unsigned char *readable, int *event_p,
				int type);
int
afr_read_subvol_get (inode_t *inode, xlator_t *this, int *subvol_p,
		     int *event_p, afr_transaction_type type);

#define afr_data_subvol_get(i, t, s, e) \
	afr_read_subvol_get(i, t, s, e, AFR_DATA_TRANSACTION)

#define afr_metadata_subvol_get(i, t, s, e) \
	afr_read_subvol_get(i, t, s, e, AFR_METADATA_TRANSACTION)

int
afr_inode_refresh (call_frame_t *frame, xlator_t *this, inode_t *inode,
		   afr_inode_refresh_cbk_t cbk);

int32_t
afr_notify (xlator_t *this, int32_t event, void *data, void *data2);

int
xattr_is_equal (dict_t *this, char *key1, data_t *value1, void *data);

int
afr_init_entry_lockee (afr_entry_lockee_t *lockee, afr_local_t *local,
                       loc_t *loc, char *basename, int child_count);

void
afr_entry_lockee_cleanup (afr_internal_lock_t *int_lock);

int
afr_attempt_lock_recovery (xlator_t *this, int32_t child_index);

int
afr_mark_locked_nodes (xlator_t *this, fd_t *fd,
                       unsigned char *locked_nodes);

void
afr_set_lk_owner (call_frame_t *frame, xlator_t *this, void *lk_owner);

int
afr_set_lock_number (call_frame_t *frame, xlator_t *this);

int32_t
afr_unlock (call_frame_t *frame, xlator_t *this);

int
afr_nonblocking_entrylk (call_frame_t *frame, xlator_t *this);

int
afr_nonblocking_inodelk (call_frame_t *frame, xlator_t *this);

int
afr_blocking_lock (call_frame_t *frame, xlator_t *this);

int
afr_internal_lock_finish (call_frame_t *frame, xlator_t *this);

int
afr_lk_transfer_datalock (call_frame_t *dst, call_frame_t *src, char *dom,
                          unsigned int child_count);

int
__afr_fd_ctx_set (xlator_t *this, fd_t *fd);

int
afr_fd_ctx_set (xlator_t *this, fd_t *fd);

afr_fd_ctx_t *
afr_fd_ctx_get (fd_t *fd, xlator_t *this);

int
afr_build_parent_loc (loc_t *parent, loc_t *child, int32_t *op_errno);

int
afr_locked_nodes_count (unsigned char *locked_nodes, int child_count);

int
afr_replies_interpret (call_frame_t *frame, xlator_t *this, inode_t *inode);

void
afr_local_replies_wipe (afr_local_t *local, afr_private_t *priv);

void
afr_local_cleanup (afr_local_t *local, xlator_t *this);

int
afr_frame_return (call_frame_t *frame);

int
afr_open (call_frame_t *frame, xlator_t *this, loc_t *loc, int32_t flags,
          fd_t *fd, dict_t *xdata);

void
afr_local_transaction_cleanup (afr_local_t *local, xlator_t *this);

int
afr_cleanup_fd_ctx (xlator_t *this, fd_t *fd);

#define AFR_STACK_UNWIND(fop, frame, params ...)                \
        do {                                                    \
                afr_local_t *__local = NULL;                    \
                xlator_t    *__this = NULL;                     \
                if (frame) {                                    \
                        __local = frame->local;                 \
                        __this = frame->this;                   \
                        frame->local = NULL;                    \
                }                                               \
                STACK_UNWIND_STRICT (fop, frame, params);       \
                if (__local) {                                  \
                        afr_local_cleanup (__local, __this);    \
                        mem_put (__local);                      \
                }                                               \
        } while (0)

#define AFR_STACK_DESTROY(frame)                                \
        do {                                                    \
                afr_local_t *__local = NULL;                    \
                xlator_t    *__this = NULL;                     \
                __local = frame->local;                         \
                __this = frame->this;                           \
                frame->local = NULL;                            \
                STACK_DESTROY (frame->root);                    \
                if (__local) {                                  \
                        afr_local_cleanup (__local, __this);    \
                        mem_put (__local);                      \
                }                                               \
        } while (0);

#define AFR_FRAME_INIT(frame, op_errno)				       \
	({frame->local = mem_get0 (THIS->local_pool);		       \
	if (afr_local_init (frame->local, THIS->private, &op_errno)) { \
		afr_local_cleanup (frame->local, THIS);		       \
		mem_put (frame->local);				       \
		frame->local = NULL; };				       \
	frame->local;})

#define AFR_STACK_RESET(frame)                                         \
        do {                                                           \
                afr_local_t *__local = NULL;                           \
                xlator_t    *__this = NULL;                            \
                __local = frame->local;                                \
                __this = frame->this;                                  \
                frame->local = NULL;                                   \
                int __opr;                                             \
                STACK_RESET (frame->root);                             \
                if (__local) {                                         \
                        afr_local_cleanup (__local, __this);           \
                        mem_put (__local);                             \
                }                                                      \
                AFR_FRAME_INIT (frame, __opr);                         \
        } while (0)

/* allocate and return a string that is the basename of argument */
static inline char *
AFR_BASENAME (const char *str)
{
        char *__tmp_str = NULL;
        char *__basename_str = NULL;
        __tmp_str = gf_strdup (str);
        __basename_str = gf_strdup (basename (__tmp_str));
        GF_FREE (__tmp_str);
        return __basename_str;
}

call_frame_t *
afr_copy_frame (call_frame_t *base);

int
afr_transaction_local_init (afr_local_t *local, xlator_t *this);

int32_t
afr_marker_getxattr (call_frame_t *frame, xlator_t *this,
                     loc_t *loc, const char *name,afr_local_t *local, afr_private_t *priv );

int
afr_local_init (afr_local_t *local, afr_private_t *priv, int32_t *op_errno);

int
afr_internal_lock_init (afr_internal_lock_t *lk, size_t child_count,
                        transaction_lk_type_t lk_type);

int
afr_higher_errno (int32_t old_errno, int32_t new_errno);

int
afr_final_errno (afr_local_t *local, afr_private_t *priv);

int
afr_xattr_req_prepare (xlator_t *this, dict_t *xattr_req);

void
afr_fix_open (fd_t *fd, xlator_t *this);

afr_fd_ctx_t *
afr_fd_ctx_get (fd_t *fd, xlator_t *this);

void
afr_set_low_priority (call_frame_t *frame);
int
afr_child_fd_ctx_set (xlator_t *this, fd_t *fd, int32_t child,
                      int flags);

void
afr_matrix_cleanup (int32_t **pending, unsigned int m);

int32_t**
afr_matrix_create (unsigned int m, unsigned int n);

int**
afr_mark_pending_changelog (afr_private_t *priv, unsigned char *pending,
                            dict_t *xattr, ia_type_t iat);

void
afr_filter_xattrs (dict_t *xattr);

/*
 * Special value indicating we should use the "auto" quorum method instead of
 * a fixed value (including zero to turn off quorum enforcement).
 */
#define AFR_QUORUM_AUTO INT_MAX

int
afr_fd_report_unstable_write (xlator_t *this, fd_t *fd);

gf_boolean_t
afr_fd_has_witnessed_unstable_write (xlator_t *this, fd_t *fd);

void
afr_delayed_changelog_wake_resume (xlator_t *this, fd_t *fd, call_stub_t *stub);

int
afr_inodelk_init (afr_inodelk_t *lk, char *dom, size_t child_count);

void
afr_handle_open_fd_count (call_frame_t *frame, xlator_t *this);

void
afr_remove_eager_lock_stub (afr_local_t *local);

void
afr_replies_wipe (struct afr_reply *replies, int count);

gf_boolean_t
afr_xattrs_are_equal (dict_t *dict1, dict_t *dict2);

gf_boolean_t
afr_is_xattr_ignorable (char *key);

int
afr_get_heal_info (call_frame_t *frame, xlator_t *this, loc_t *loc,
                   dict_t *xdata);

int
afr_heal_splitbrain_file(call_frame_t *frame, xlator_t *this, loc_t *loc);

int
afr_get_split_brain_status (call_frame_t *frame, xlator_t *this, loc_t *loc);

int
afr_inode_split_brain_choice_set (inode_t *inode, xlator_t *this,
                                  int spb_choice);
int
afr_inode_split_brain_choice_get (inode_t *inode, xlator_t *this,
                                  int *spb_choice);
int
afr_get_child_index_from_name (xlator_t *this, char *name);

int
afr_is_split_brain (call_frame_t *frame, xlator_t *this, inode_t *inode,
                    uuid_t gfid, gf_boolean_t *d_spb, gf_boolean_t *m_spb);
int
afr_spb_choice_timeout_cancel (xlator_t *this, inode_t *inode);

int
afr_set_split_brain_choice (int ret, call_frame_t *frame, void *opaque);
#endif /* __AFR_H__ */