-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathClient.h
1956 lines (1641 loc) · 67.9 KB
/
Client.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
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
// -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*-
// vim: ts=8 sw=2 smarttab
/*
* Ceph - scalable distributed file system
*
* Copyright (C) 2004-2006 Sage Weil <[email protected]>
*
* This is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License version 2.1, as published by the Free Software
* Foundation. See file COPYING.
*
*/
#ifndef CEPH_CLIENT_H
#define CEPH_CLIENT_H
#include "common/CommandTable.h"
#include "common/Finisher.h"
#include "common/Timer.h"
#include "common/ceph_mutex.h"
#include "common/cmdparse.h"
#include "common/compiler_extensions.h"
#include "include/common_fwd.h"
#include "include/cephfs/ceph_ll_client.h"
#include "include/filepath.h"
#include "include/interval_set.h"
#include "include/lru.h"
#include "include/types.h"
#include "include/unordered_map.h"
#include "include/unordered_set.h"
#include "include/cephfs/metrics/Types.h"
#include "mds/mdstypes.h"
#include "mds/MDSAuthCaps.h"
#include "include/cephfs/types.h"
#include "msg/Dispatcher.h"
#include "msg/MessageRef.h"
#include "msg/Messenger.h"
#include "osdc/ObjectCacher.h"
#include "RWRef.h"
#include "InodeRef.h"
#include "MetaSession.h"
#include "UserPerm.h"
#include <fstream>
#include <map>
#include <memory>
#include <set>
#include <string>
#include <thread>
using std::set;
using std::map;
using std::fstream;
class FSMap;
class FSMapUser;
class MonClient;
struct DirStat;
struct LeaseStat;
struct InodeStat;
class Filer;
class Objecter;
class WritebackHandler;
class MDSMap;
class Message;
class destructive_lock_ref_t;
enum {
l_c_first = 20000,
l_c_reply,
l_c_lat,
l_c_wrlat,
l_c_read,
l_c_fsync,
l_c_md_avg,
l_c_md_sqsum,
l_c_md_ops,
l_c_rd_avg,
l_c_rd_sqsum,
l_c_rd_ops,
l_c_wr_avg,
l_c_wr_sqsum,
l_c_wr_ops,
l_c_last,
};
class MDSCommandOp : public CommandOp
{
public:
mds_gid_t mds_gid;
bool one_shot = false;
explicit MDSCommandOp(ceph_tid_t t) : CommandOp(t) {}
explicit MDSCommandOp(ceph_tid_t t, ceph_tid_t multi_id) : CommandOp(t, multi_id) {}
};
/* error code for ceph_fuse */
#define CEPH_FUSE_NO_MDS_UP -((1<<16)+0) /* no mds up deteced in ceph_fuse */
#define CEPH_FUSE_LAST -((1<<16)+1) /* (unused) */
// ============================================
// types for my local metadata cache
/* basic structure:
- Dentries live in an LRU loop. they get expired based on last access.
see include/lru.h. items can be bumped to "mid" or "top" of list, etc.
- Inode has ref count for each Fh, Dir, or Dentry that points to it.
- when Inode ref goes to 0, it's expired.
- when Dir is empty, it's removed (and it's Inode ref--)
*/
/* getdir result */
struct DirEntry {
explicit DirEntry(const std::string &s) : d_name(s), stmask(0) {}
DirEntry(const std::string &n, struct stat& s, int stm)
: d_name(n), st(s), stmask(stm) {}
std::string d_name;
struct stat st;
int stmask;
};
struct Cap;
class Dir;
class Dentry;
struct SnapRealm;
struct Fh;
struct CapSnap;
struct MetaRequest;
class ceph_lock_state_t;
// ========================================================
// client interface
struct dir_result_t {
static const int SHIFT = 28;
static const int64_t MASK = (1 << SHIFT) - 1;
static const int64_t HASH = 0xFFULL << (SHIFT + 24); // impossible frag bits
static const loff_t END = 1ULL << (SHIFT + 32);
struct dentry {
int64_t offset;
std::string name;
std::string alternate_name;
InodeRef inode;
explicit dentry(int64_t o) : offset(o) {}
dentry(int64_t o, std::string n, std::string an, InodeRef in) :
offset(o), name(std::move(n)), alternate_name(std::move(an)), inode(std::move(in)) {}
};
struct dentry_off_lt {
bool operator()(const dentry& d, int64_t off) const {
return dir_result_t::fpos_cmp(d.offset, off) < 0;
}
};
explicit dir_result_t(Inode *in, const UserPerm& perms, int fd);
static uint64_t make_fpos(unsigned h, unsigned l, bool hash) {
uint64_t v = ((uint64_t)h<< SHIFT) | (uint64_t)l;
if (hash)
v |= HASH;
else
ceph_assert((v & HASH) != HASH);
return v;
}
static unsigned fpos_high(uint64_t p) {
unsigned v = (p & (END-1)) >> SHIFT;
if ((p & HASH) == HASH)
return ceph_frag_value(v);
return v;
}
static unsigned fpos_low(uint64_t p) {
return p & MASK;
}
static int fpos_cmp(uint64_t l, uint64_t r) {
int c = ceph_frag_compare(fpos_high(l), fpos_high(r));
if (c)
return c;
if (fpos_low(l) == fpos_low(r))
return 0;
return fpos_low(l) < fpos_low(r) ? -1 : 1;
}
unsigned offset_high() { return fpos_high(offset); }
unsigned offset_low() { return fpos_low(offset); }
void set_end() { offset |= END; }
bool at_end() { return (offset & END); }
void set_hash_order() { offset |= HASH; }
bool hash_order() { return (offset & HASH) == HASH; }
bool is_cached() {
if (buffer.empty())
return false;
if (hash_order()) {
return buffer_frag.contains(offset_high());
} else {
return buffer_frag == frag_t(offset_high());
}
}
void reset() {
last_name.clear();
next_offset = 2;
offset = 0;
ordered_count = 0;
cache_index = 0;
buffer.clear();
}
InodeRef inode;
int64_t offset; // hash order:
// (0xff << 52) | ((24 bits hash) << 28) |
// (the nth entry has hash collision);
// frag+name order;
// ((frag value) << 28) | (the nth entry in frag);
unsigned next_offset; // offset of next chunk (last_name's + 1)
std::string last_name; // last entry in previous chunk
uint64_t release_count;
uint64_t ordered_count;
unsigned cache_index;
int start_shared_gen; // dir shared_gen at start of readdir
UserPerm perms;
frag_t buffer_frag;
std::vector<dentry> buffer;
struct dirent de;
int fd; // fd attached using fdopendir (-1 if none)
};
class Client : public Dispatcher, public md_config_obs_t {
public:
friend class C_Block_Sync; // Calls block map and protected helpers
friend class C_Client_CacheInvalidate; // calls ino_invalidate_cb
friend class C_Client_DentryInvalidate; // calls dentry_invalidate_cb
friend class C_Client_FlushComplete; // calls put_inode()
friend class C_Client_Remount;
friend class C_Client_RequestInterrupt;
friend class C_Deleg_Timeout; // Asserts on client_lock, called when a delegation is unreturned
friend class C_Client_CacheRelease; // Asserts on client_lock
friend class SyntheticClient;
friend void intrusive_ptr_release(Inode *in);
template <typename T> friend struct RWRefState;
template <typename T> friend class RWRef;
using Dispatcher::cct;
using clock = ceph::coarse_mono_clock;
typedef int (*add_dirent_cb_t)(void *p, struct dirent *de, struct ceph_statx *stx, off_t off, Inode *in);
struct walk_dentry_result {
InodeRef in;
std::string alternate_name;
};
class CommandHook : public AdminSocketHook {
public:
explicit CommandHook(Client *client);
int call(std::string_view command, const cmdmap_t& cmdmap,
const bufferlist&,
Formatter *f,
std::ostream& errss,
bufferlist& out) override;
private:
Client *m_client;
};
// snapshot info returned via get_snap_info(). nothing to do
// with SnapInfo on the MDS.
struct SnapInfo {
snapid_t id;
std::map<std::string, std::string> metadata;
};
Client(Messenger *m, MonClient *mc, Objecter *objecter_);
Client(const Client&) = delete;
Client(const Client&&) = delete;
virtual ~Client() override;
static UserPerm pick_my_perms(CephContext *c) {
uid_t uid = c->_conf->client_mount_uid >= 0 ? c->_conf->client_mount_uid : -1;
gid_t gid = c->_conf->client_mount_gid >= 0 ? c->_conf->client_mount_gid : -1;
return UserPerm(uid, gid);
}
UserPerm pick_my_perms() {
uid_t uid = user_id >= 0 ? user_id : -1;
gid_t gid = group_id >= 0 ? group_id : -1;
return UserPerm(uid, gid);
}
int mount(const std::string &mount_root, const UserPerm& perms,
bool require_mds=false, const std::string &fs_name="");
void unmount();
bool is_unmounting() const {
return mount_state.check_current_state(CLIENT_UNMOUNTING);
}
bool is_mounted() const {
return mount_state.check_current_state(CLIENT_MOUNTED);
}
bool is_mounting() const {
return mount_state.check_current_state(CLIENT_MOUNTING);
}
bool is_initialized() const {
return initialize_state.check_current_state(CLIENT_INITIALIZED);
}
void abort_conn();
void set_uuid(const std::string& uuid);
void set_session_timeout(unsigned timeout);
int start_reclaim(const std::string& uuid, unsigned flags,
const std::string& fs_name);
void finish_reclaim();
fs_cluster_id_t get_fs_cid() {
return fscid;
}
int mds_command(
const std::string &mds_spec,
const std::vector<std::string>& cmd,
const bufferlist& inbl,
bufferlist *poutbl, std::string *prs, Context *onfinish, bool one_shot = false);
// these should (more or less) mirror the actual system calls.
int statfs(const char *path, struct statvfs *stbuf, const UserPerm& perms);
// crap
int chdir(const char *s, std::string &new_cwd, const UserPerm& perms);
void _getcwd(std::string& cwd, const UserPerm& perms);
void getcwd(std::string& cwd, const UserPerm& perms);
// namespace ops
int opendir(const char *name, dir_result_t **dirpp, const UserPerm& perms);
int fdopendir(int dirfd, dir_result_t **dirpp, const UserPerm& perms);
int closedir(dir_result_t *dirp);
/**
* Fill a directory listing from dirp, invoking cb for each entry
* with the given pointer, the dirent, the struct stat, the stmask,
* and the offset.
*
* Returns 0 if it reached the end of the directory.
* If @a cb returns a negative error code, stop and return that.
*/
int readdir_r_cb(dir_result_t *dirp, add_dirent_cb_t cb, void *p,
unsigned want=0, unsigned flags=AT_STATX_DONT_SYNC,
bool getref=false);
struct dirent * readdir(dir_result_t *d);
int readdir_r(dir_result_t *dirp, struct dirent *de);
int readdirplus_r(dir_result_t *dirp, struct dirent *de, struct ceph_statx *stx, unsigned want, unsigned flags, Inode **out);
/*
* Get the next snapshot delta entry.
*
*/
int readdir_snapdiff(dir_result_t* dir1, snapid_t snap2,
struct dirent* out_de, snapid_t* out_snap);
int getdir(const char *relpath, std::list<std::string>& names,
const UserPerm& perms); // get the whole dir at once.
/**
* Returns the length of the buffer that got filled in, or -errno.
* If it returns -CEPHFS_ERANGE you just need to increase the size of the
* buffer and try again.
*/
int _getdents(dir_result_t *dirp, char *buf, int buflen, bool ful); // get a bunch of dentries at once
int getdents(dir_result_t *dirp, char *buf, int buflen) {
return _getdents(dirp, buf, buflen, true);
}
int getdnames(dir_result_t *dirp, char *buf, int buflen) {
return _getdents(dirp, buf, buflen, false);
}
void rewinddir(dir_result_t *dirp);
loff_t telldir(dir_result_t *dirp);
void seekdir(dir_result_t *dirp, loff_t offset);
int may_delete(const char *relpath, const UserPerm& perms);
int link(const char *existing, const char *newname, const UserPerm& perm, std::string alternate_name="");
int unlink(const char *path, const UserPerm& perm);
int unlinkat(int dirfd, const char *relpath, int flags, const UserPerm& perm);
int rename(const char *from, const char *to, const UserPerm& perm, std::string alternate_name="");
// dirs
int mkdir(const char *path, mode_t mode, const UserPerm& perm, std::string alternate_name="");
int mkdirat(int dirfd, const char *relpath, mode_t mode, const UserPerm& perm,
std::string alternate_name="");
int mkdirs(const char *path, mode_t mode, const UserPerm& perms);
int rmdir(const char *path, const UserPerm& perms);
// symlinks
int readlink(const char *path, char *buf, loff_t size, const UserPerm& perms);
int readlinkat(int dirfd, const char *relpath, char *buf, loff_t size, const UserPerm& perms);
int symlink(const char *existing, const char *newname, const UserPerm& perms, std::string alternate_name="");
int symlinkat(const char *target, int dirfd, const char *relpath, const UserPerm& perms,
std::string alternate_name="");
// path traversal for high-level interface
int walk(std::string_view path, struct walk_dentry_result* result, const UserPerm& perms, bool followsym=true);
// inode stuff
unsigned statx_to_mask(unsigned int flags, unsigned int want);
int stat(const char *path, struct stat *stbuf, const UserPerm& perms,
frag_info_t *dirstat=0, int mask=CEPH_STAT_CAP_INODE_ALL);
int statx(const char *path, struct ceph_statx *stx,
const UserPerm& perms,
unsigned int want, unsigned int flags);
int lstat(const char *path, struct stat *stbuf, const UserPerm& perms,
frag_info_t *dirstat=0, int mask=CEPH_STAT_CAP_INODE_ALL);
int setattr(const char *relpath, struct stat *attr, int mask,
const UserPerm& perms);
int setattrx(const char *relpath, struct ceph_statx *stx, int mask,
const UserPerm& perms, int flags=0);
int fsetattr(int fd, struct stat *attr, int mask, const UserPerm& perms);
int fsetattrx(int fd, struct ceph_statx *stx, int mask, const UserPerm& perms);
int chmod(const char *path, mode_t mode, const UserPerm& perms);
int fchmod(int fd, mode_t mode, const UserPerm& perms);
int chmodat(int dirfd, const char *relpath, mode_t mode, int flags, const UserPerm& perms);
int lchmod(const char *path, mode_t mode, const UserPerm& perms);
int chown(const char *path, uid_t new_uid, gid_t new_gid,
const UserPerm& perms);
int fchown(int fd, uid_t new_uid, gid_t new_gid, const UserPerm& perms);
int lchown(const char *path, uid_t new_uid, gid_t new_gid,
const UserPerm& perms);
int chownat(int dirfd, const char *relpath, uid_t new_uid, gid_t new_gid,
int flags, const UserPerm& perms);
int utime(const char *path, struct utimbuf *buf, const UserPerm& perms);
int lutime(const char *path, struct utimbuf *buf, const UserPerm& perms);
int futime(int fd, struct utimbuf *buf, const UserPerm& perms);
int utimes(const char *relpath, struct timeval times[2], const UserPerm& perms);
int lutimes(const char *relpath, struct timeval times[2], const UserPerm& perms);
int futimes(int fd, struct timeval times[2], const UserPerm& perms);
int futimens(int fd, struct timespec times[2], const UserPerm& perms);
int utimensat(int dirfd, const char *relpath, struct timespec times[2], int flags,
const UserPerm& perms);
int flock(int fd, int operation, uint64_t owner);
int truncate(const char *path, loff_t size, const UserPerm& perms);
// file ops
int mknod(const char *path, mode_t mode, const UserPerm& perms, dev_t rdev=0);
int create_and_open(int dirfd, const char *relpath, int flags, const UserPerm& perms,
mode_t mode, int stripe_unit, int stripe_count, int object_size,
const char *data_pool, std::string alternate_name);
int open(const char *path, int flags, const UserPerm& perms, mode_t mode=0, std::string alternate_name="") {
return open(path, flags, perms, mode, 0, 0, 0, NULL, alternate_name);
}
int open(const char *path, int flags, const UserPerm& perms,
mode_t mode, int stripe_unit, int stripe_count, int object_size,
const char *data_pool, std::string alternate_name="");
int openat(int dirfd, const char *relpath, int flags, const UserPerm& perms,
mode_t mode, int stripe_unit, int stripe_count,
int object_size, const char *data_pool, std::string alternate_name);
int openat(int dirfd, const char *path, int flags, const UserPerm& perms, mode_t mode=0,
std::string alternate_name="") {
return openat(dirfd, path, flags, perms, mode, 0, 0, 0, NULL, alternate_name);
}
int lookup_hash(inodeno_t ino, inodeno_t dirino, const char *name,
const UserPerm& perms);
int lookup_ino(inodeno_t ino, const UserPerm& perms, Inode **inode=NULL);
int lookup_name(Inode *in, Inode *parent, const UserPerm& perms);
int _close(int fd);
int close(int fd);
loff_t lseek(int fd, loff_t offset, int whence);
int read(int fd, char *buf, loff_t size, loff_t offset=-1);
int preadv(int fd, const struct iovec *iov, int iovcnt, loff_t offset=-1);
int write(int fd, const char *buf, loff_t size, loff_t offset=-1);
int pwritev(int fd, const struct iovec *iov, int iovcnt, loff_t offset=-1);
int ftruncate(int fd, loff_t size, const UserPerm& perms);
int fsync(int fd, bool syncdataonly);
int fstat(int fd, struct stat *stbuf, const UserPerm& perms,
int mask=CEPH_STAT_CAP_INODE_ALL);
int fstatx(int fd, struct ceph_statx *stx, const UserPerm& perms,
unsigned int want, unsigned int flags);
int statxat(int dirfd, const char *relpath,
struct ceph_statx *stx, const UserPerm& perms,
unsigned int want, unsigned int flags);
int fallocate(int fd, int mode, loff_t offset, loff_t length);
// full path xattr ops
int getxattr(const char *path, const char *name, void *value, size_t size,
const UserPerm& perms);
int lgetxattr(const char *path, const char *name, void *value, size_t size,
const UserPerm& perms);
int fgetxattr(int fd, const char *name, void *value, size_t size,
const UserPerm& perms);
int listxattr(const char *path, char *list, size_t size, const UserPerm& perms);
int llistxattr(const char *path, char *list, size_t size, const UserPerm& perms);
int flistxattr(int fd, char *list, size_t size, const UserPerm& perms);
int removexattr(const char *path, const char *name, const UserPerm& perms);
int lremovexattr(const char *path, const char *name, const UserPerm& perms);
int fremovexattr(int fd, const char *name, const UserPerm& perms);
int setxattr(const char *path, const char *name, const void *value,
size_t size, int flags, const UserPerm& perms);
int lsetxattr(const char *path, const char *name, const void *value,
size_t size, int flags, const UserPerm& perms);
int fsetxattr(int fd, const char *name, const void *value, size_t size,
int flags, const UserPerm& perms);
int sync_fs();
int64_t drop_caches();
int get_snap_info(const char *path, const UserPerm &perms, SnapInfo *snap_info);
// hpc lazyio
int lazyio(int fd, int enable);
int lazyio_propagate(int fd, loff_t offset, size_t count);
int lazyio_synchronize(int fd, loff_t offset, size_t count);
// expose file layout
int describe_layout(const char *path, file_layout_t* layout,
const UserPerm& perms);
int fdescribe_layout(int fd, file_layout_t* layout);
int get_file_stripe_address(int fd, loff_t offset, std::vector<entity_addr_t>& address);
int get_file_extent_osds(int fd, loff_t off, loff_t *len, std::vector<int>& osds);
int get_osd_addr(int osd, entity_addr_t& addr);
// expose mdsmap
int64_t get_default_pool_id();
// expose osdmap
int get_local_osd();
int get_pool_replication(int64_t pool);
int64_t get_pool_id(const char *pool_name);
std::string get_pool_name(int64_t pool);
int get_osd_crush_location(int id, std::vector<std::pair<std::string, std::string> >& path);
int enumerate_layout(int fd, std::vector<ObjectExtent>& result,
loff_t length, loff_t offset);
int mksnap(const char *path, const char *name, const UserPerm& perm,
mode_t mode=0, const std::map<std::string, std::string> &metadata={});
int rmsnap(const char *path, const char *name, const UserPerm& perm, bool check_perms=false);
// cephx mds auth caps checking
int mds_check_access(std::string& path, const UserPerm& perms, int mask);
// Inode permission checking
int inode_permission(Inode *in, const UserPerm& perms, unsigned want);
// expose caps
int get_caps_issued(int fd);
int get_caps_issued(const char *path, const UserPerm& perms);
snapid_t ll_get_snapid(Inode *in);
vinodeno_t ll_get_vino(Inode *in) {
std::lock_guard lock(client_lock);
return _get_vino(in);
}
// get inode from faked ino
Inode *ll_get_inode(ino_t ino);
Inode *ll_get_inode(vinodeno_t vino);
int ll_lookup(Inode *parent, const char *name, struct stat *attr,
Inode **out, const UserPerm& perms);
int ll_lookup_inode(struct inodeno_t ino, const UserPerm& perms, Inode **inode);
int ll_lookup_vino(vinodeno_t vino, const UserPerm& perms, Inode **inode);
int ll_lookupx(Inode *parent, const char *name, Inode **out,
struct ceph_statx *stx, unsigned want, unsigned flags,
const UserPerm& perms);
bool ll_forget(Inode *in, uint64_t count);
bool ll_put(Inode *in);
int ll_get_snap_ref(snapid_t snap);
int ll_getattr(Inode *in, struct stat *st, const UserPerm& perms);
int ll_getattrx(Inode *in, struct ceph_statx *stx, unsigned int want,
unsigned int flags, const UserPerm& perms);
int ll_setattrx(Inode *in, struct ceph_statx *stx, int mask,
const UserPerm& perms);
int ll_setattr(Inode *in, struct stat *st, int mask,
const UserPerm& perms);
int ll_getxattr(Inode *in, const char *name, void *value, size_t size,
const UserPerm& perms);
int ll_setxattr(Inode *in, const char *name, const void *value, size_t size,
int flags, const UserPerm& perms);
int ll_removexattr(Inode *in, const char *name, const UserPerm& perms);
int ll_listxattr(Inode *in, char *list, size_t size, const UserPerm& perms);
int ll_opendir(Inode *in, int flags, dir_result_t **dirpp,
const UserPerm& perms);
int ll_releasedir(dir_result_t* dirp);
int ll_fsyncdir(dir_result_t* dirp);
int ll_readlink(Inode *in, char *buf, size_t bufsize, const UserPerm& perms);
int ll_mknod(Inode *in, const char *name, mode_t mode, dev_t rdev,
struct stat *attr, Inode **out, const UserPerm& perms);
int ll_mknodx(Inode *parent, const char *name, mode_t mode, dev_t rdev,
Inode **out, struct ceph_statx *stx, unsigned want,
unsigned flags, const UserPerm& perms);
int ll_mkdir(Inode *in, const char *name, mode_t mode, struct stat *attr,
Inode **out, const UserPerm& perm);
int ll_mkdirx(Inode *parent, const char *name, mode_t mode, Inode **out,
struct ceph_statx *stx, unsigned want, unsigned flags,
const UserPerm& perms);
int ll_symlink(Inode *in, const char *name, const char *value,
struct stat *attr, Inode **out, const UserPerm& perms);
int ll_symlinkx(Inode *parent, const char *name, const char *value,
Inode **out, struct ceph_statx *stx, unsigned want,
unsigned flags, const UserPerm& perms);
int ll_unlink(Inode *in, const char *name, const UserPerm& perm);
int ll_rmdir(Inode *in, const char *name, const UserPerm& perms);
int ll_rename(Inode *parent, const char *name, Inode *newparent,
const char *newname, const UserPerm& perm);
int ll_link(Inode *in, Inode *newparent, const char *newname,
const UserPerm& perm);
int ll_open(Inode *in, int flags, Fh **fh, const UserPerm& perms);
int _ll_create(Inode *parent, const char *name, mode_t mode,
int flags, InodeRef *in, int caps, Fh **fhp,
const UserPerm& perms);
int ll_create(Inode *parent, const char *name, mode_t mode, int flags,
struct stat *attr, Inode **out, Fh **fhp,
const UserPerm& perms);
int ll_createx(Inode *parent, const char *name, mode_t mode,
int oflags, Inode **outp, Fh **fhp,
struct ceph_statx *stx, unsigned want, unsigned lflags,
const UserPerm& perms);
int ll_read_block(Inode *in, uint64_t blockid, char *buf, uint64_t offset,
uint64_t length, file_layout_t* layout);
int ll_write_block(Inode *in, uint64_t blockid,
char* buf, uint64_t offset,
uint64_t length, file_layout_t* layout,
uint64_t snapseq, uint32_t sync);
int ll_commit_blocks(Inode *in, uint64_t offset, uint64_t length);
int ll_statfs(Inode *in, struct statvfs *stbuf, const UserPerm& perms);
int ll_walk(const char* name, Inode **i, struct ceph_statx *stx,
unsigned int want, unsigned int flags, const UserPerm& perms);
uint32_t ll_stripe_unit(Inode *in);
int ll_file_layout(Inode *in, file_layout_t *layout);
uint64_t ll_snap_seq(Inode *in);
int ll_read(Fh *fh, loff_t off, loff_t len, bufferlist *bl);
int ll_write(Fh *fh, loff_t off, loff_t len, const char *data);
int64_t ll_readv(struct Fh *fh, const struct iovec *iov, int iovcnt, int64_t off);
int64_t ll_writev(struct Fh *fh, const struct iovec *iov, int iovcnt, int64_t off);
int64_t ll_preadv_pwritev(struct Fh *fh, const struct iovec *iov, int iovcnt,
int64_t offset, bool write,
Context *onfinish = nullptr,
bufferlist *blp = nullptr,
bool do_fsync = false, bool syncdataonly = false);
loff_t ll_lseek(Fh *fh, loff_t offset, int whence);
int ll_flush(Fh *fh);
int ll_fsync(Fh *fh, bool syncdataonly);
int ll_sync_inode(Inode *in, bool syncdataonly);
int ll_fallocate(Fh *fh, int mode, int64_t offset, int64_t length);
int ll_release(Fh *fh);
int ll_getlk(Fh *fh, struct flock *fl, uint64_t owner);
int ll_setlk(Fh *fh, struct flock *fl, uint64_t owner, int sleep);
int ll_flock(Fh *fh, int cmd, uint64_t owner);
int ll_lazyio(Fh *fh, int enable);
int ll_file_layout(Fh *fh, file_layout_t *layout);
void ll_interrupt(void *d);
bool ll_handle_umask() {
return acl_type != NO_ACL;
}
int ll_get_stripe_osd(struct Inode *in, uint64_t blockno,
file_layout_t* layout);
uint64_t ll_get_internal_offset(struct Inode *in, uint64_t blockno);
int ll_num_osds(void);
int ll_osdaddr(int osd, uint32_t *addr);
int ll_osdaddr(int osd, char* buf, size_t size);
void _ll_register_callbacks(struct ceph_client_callback_args *args);
void ll_register_callbacks(struct ceph_client_callback_args *args); // deprecated
int ll_register_callbacks2(struct ceph_client_callback_args *args);
std::pair<int, bool> test_dentry_handling(bool can_invalidate);
const char** get_tracked_conf_keys() const override;
void handle_conf_change(const ConfigProxy& conf,
const std::set <std::string> &changed) override;
uint32_t get_deleg_timeout() { return deleg_timeout; }
int set_deleg_timeout(uint32_t timeout);
int ll_delegation(Fh *fh, unsigned cmd, ceph_deleg_cb_t cb, void *priv);
entity_name_t get_myname() { return messenger->get_myname(); }
void wait_on_list(std::list<ceph::condition_variable*>& ls);
void signal_cond_list(std::list<ceph::condition_variable*>& ls);
void set_filer_flags(int flags);
void clear_filer_flags(int flags);
void tear_down_cache();
void update_metadata(std::string const &k, std::string const &v);
client_t get_nodeid() { return whoami; }
inodeno_t _get_root_ino(bool fake=true);
inodeno_t get_root_ino();
Inode *get_root();
virtual int init();
virtual void shutdown();
// messaging
int cancel_commands_if(std::regular_invocable<MDSCommandOp const&> auto && error_for_op)
{
std::vector<ceph_tid_t> cancel_ops;
std::scoped_lock cmd_lock(command_lock);
auto& commands = command_table.get_commands();
for (const auto &[tid, op]: commands) {
int rc = static_cast<int>(error_for_op(op));
if (rc) {
cancel_ops.push_back(tid);
if (op.on_finish)
op.on_finish->complete(rc);
}
}
for (const auto& tid : cancel_ops)
command_table.erase(tid);
return cancel_ops.size();
}
void cancel_commands(const MDSMap& newmap);
void handle_mds_map(const MConstRef<MMDSMap>& m);
void handle_fs_map(const MConstRef<MFSMap>& m);
void handle_fs_map_user(const MConstRef<MFSMapUser>& m);
void handle_osd_map(const MConstRef<MOSDMap>& m);
void handle_lease(const MConstRef<MClientLease>& m);
// inline data
int uninline_data(Inode *in, Context *onfinish);
// file caps
void check_cap_issue(Inode *in, unsigned issued);
void add_update_cap(Inode *in, MetaSession *session, uint64_t cap_id,
unsigned issued, unsigned wanted, unsigned seq, unsigned mseq,
inodeno_t realm, int flags, const UserPerm& perms);
void remove_cap(Cap *cap, bool queue_release);
void remove_all_caps(Inode *in);
void remove_session_caps(MetaSession *session, int err);
int mark_caps_flushing(Inode *in, ceph_tid_t *ptid);
void adjust_session_flushing_caps(Inode *in, MetaSession *old_s, MetaSession *new_s);
void flush_caps_sync();
void kick_flushing_caps(Inode *in, MetaSession *session);
void kick_flushing_caps(MetaSession *session);
void early_kick_flushing_caps(MetaSession *session);
int get_caps(Fh *fh, int need, int want, int *have, loff_t endoff);
int get_caps_used(Inode *in);
void maybe_update_snaprealm(SnapRealm *realm, snapid_t snap_created, snapid_t snap_highwater,
std::vector<snapid_t>& snaps);
void handle_quota(const MConstRef<MClientQuota>& m);
void handle_snap(const MConstRef<MClientSnap>& m);
void handle_caps(const MConstRef<MClientCaps>& m);
void handle_cap_import(MetaSession *session, Inode *in, const MConstRef<MClientCaps>& m);
void handle_cap_export(MetaSession *session, Inode *in, const MConstRef<MClientCaps>& m);
void handle_cap_trunc(MetaSession *session, Inode *in, const MConstRef<MClientCaps>& m);
void handle_cap_flush_ack(MetaSession *session, Inode *in, Cap *cap, const MConstRef<MClientCaps>& m);
void handle_cap_flushsnap_ack(MetaSession *session, Inode *in, const MConstRef<MClientCaps>& m);
void handle_cap_grant(MetaSession *session, Inode *in, Cap *cap, const MConstRef<MClientCaps>& m);
void cap_delay_requeue(Inode *in);
void send_cap(Inode *in, MetaSession *session, Cap *cap, int flags,
int used, int want, int retain, int flush,
ceph_tid_t flush_tid);
void send_flush_snap(Inode *in, MetaSession *session, snapid_t follows, CapSnap& capsnap);
void flush_snaps(Inode *in);
void get_cap_ref(Inode *in, int cap);
void put_cap_ref(Inode *in, int cap);
void submit_sync_caps(Inode *in, ceph_tid_t want, Context *onfinish);
void wait_sync_caps(Inode *in, ceph_tid_t want);
void wait_sync_caps(ceph_tid_t want);
void queue_cap_snap(Inode *in, SnapContext &old_snapc);
void finish_cap_snap(Inode *in, CapSnap &capsnap, int used);
void _schedule_invalidate_dentry_callback(Dentry *dn, bool del);
void _async_dentry_invalidate(vinodeno_t dirino, vinodeno_t ino, std::string& name);
void _try_to_trim_inode(Inode *in, bool sched_inval);
void _schedule_invalidate_callback(Inode *in, int64_t off, int64_t len);
void _invalidate_inode_cache(Inode *in);
void _invalidate_inode_cache(Inode *in, int64_t off, int64_t len);
void _async_invalidate(vinodeno_t ino, int64_t off, int64_t len);
void _schedule_ino_release_callback(Inode *in);
void _async_inode_release(vinodeno_t ino);
bool _release(Inode *in);
/**
* Initiate a flush of the data associated with the given inode.
* If you specify a Context, you are responsible for holding an inode
* reference for the duration of the flush. If not, _flush() will
* take the reference for you.
* @param in The Inode whose data you wish to flush.
* @param c The Context you wish us to complete once the data is
* flushed. If already flushed, this will be called in-line.
*
* @returns true if the data was already flushed, false otherwise.
*/
bool _flush(Inode *in, Context *c);
void _flush_range(Inode *in, int64_t off, uint64_t size);
void _flushed(Inode *in);
void flush_set_callback(ObjectCacher::ObjectSet *oset);
void close_release(Inode *in);
void close_safe(Inode *in);
void lock_fh_pos(Fh *f);
void unlock_fh_pos(Fh *f);
// metadata cache
void update_dir_dist(Inode *in, DirStat *st, mds_rank_t from);
void clear_dir_complete_and_ordered(Inode *diri, bool complete);
void insert_readdir_results(MetaRequest *request, MetaSession *session,
Inode *diri, Inode *diri_other);
Inode* insert_trace(MetaRequest *request, MetaSession *session);
void update_inode_file_size(Inode *in, int issued, uint64_t size,
uint64_t truncate_seq, uint64_t truncate_size);
void update_inode_file_time(Inode *in, int issued, uint64_t time_warp_seq,
utime_t ctime, utime_t mtime, utime_t atime);
Inode *add_update_inode(InodeStat *st, utime_t ttl, MetaSession *session,
const UserPerm& request_perms);
Dentry *insert_dentry_inode(Dir *dir, const std::string& dname, LeaseStat *dlease,
Inode *in, utime_t from, MetaSession *session,
Dentry *old_dentry = NULL);
void update_dentry_lease(Dentry *dn, LeaseStat *dlease, utime_t from, MetaSession *session);
bool use_faked_inos() { return _use_faked_inos; }
vinodeno_t map_faked_ino(ino_t ino);
//notify the mds to flush the mdlog
void flush_mdlog_sync(Inode *in);
void flush_mdlog_sync();
void flush_mdlog(MetaSession *session);
void renew_caps();
void renew_caps(MetaSession *session);
void flush_cap_releases();
void renew_and_flush_cap_releases();
void tick();
void start_tick_thread();
void update_read_io_size(size_t size) {
total_read_ops++;
total_read_size += size;
}
void update_write_io_size(size_t size) {
total_write_ops++;
total_write_size += size;
}
void inc_dentry_nr() {
++dentry_nr;
}
void dec_dentry_nr() {
--dentry_nr;
}
void dlease_hit() {
++dlease_hits;
}
void dlease_miss() {
++dlease_misses;
}
std::tuple<uint64_t, uint64_t, uint64_t> get_dlease_hit_rates() {
return std::make_tuple(dlease_hits, dlease_misses, dentry_nr);
}
void cap_hit() {
++cap_hits;
}
void cap_miss() {
++cap_misses;
}
std::pair<uint64_t, uint64_t> get_cap_hit_rates() {
return std::make_pair(cap_hits, cap_misses);
}
void inc_opened_files() {
++opened_files;
}
void dec_opened_files() {
--opened_files;
}
std::pair<uint64_t, uint64_t> get_opened_files_rates() {
return std::make_pair(opened_files, inode_map.size());
}
void inc_pinned_icaps() {
++pinned_icaps;
}
void dec_pinned_icaps(uint64_t nr=1) {
pinned_icaps -= nr;
}
std::pair<uint64_t, uint64_t> get_pinned_icaps_rates() {
return std::make_pair(pinned_icaps, inode_map.size());
}
void inc_opened_inodes() {
++opened_inodes;
}
void dec_opened_inodes() {
--opened_inodes;
}
std::pair<uint64_t, uint64_t> get_opened_inodes_rates() {
return std::make_pair(opened_inodes, inode_map.size());
}
/* timer_lock for 'timer' */
ceph::mutex timer_lock = ceph::make_mutex("Client::timer_lock");
SafeTimer timer;
/* tick thread */
std::thread upkeeper;
ceph::condition_variable upkeep_cond;
bool tick_thread_stopped = false;
std::unique_ptr<PerfCounters> logger;
std::unique_ptr<MDSMap> mdsmap;
bool fuse_default_permissions;
bool _collect_and_send_global_metrics;
protected:
std::list<ceph::condition_variable*> waiting_for_reclaim;
/* Flags for check_caps() */
static const unsigned CHECK_CAPS_NODELAY = 0x1;
static const unsigned CHECK_CAPS_SYNCHRONOUS = 0x2;
void check_caps(Inode *in, unsigned flags);
void set_cap_epoch_barrier(epoch_t e);
void handle_command_reply(const MConstRef<MCommandReply>& m);
int fetch_fsmap(bool user);
int resolve_mds(
const std::string &mds_spec,
std::vector<mds_gid_t> *targets);
void get_session_metadata(std::map<std::string, std::string> *meta) const;
bool have_open_session(mds_rank_t mds);
void got_mds_push(MetaSession *s);
MetaSessionRef _get_mds_session(mds_rank_t mds, Connection *con); ///< return session for mds *and* con; null otherwise
MetaSessionRef _get_or_open_mds_session(mds_rank_t mds);
MetaSessionRef _open_mds_session(mds_rank_t mds);
void _close_mds_session(MetaSession *s);
void _closed_mds_session(MetaSession *s, int err=0, bool rejected=false);
bool _any_stale_sessions() const;
void _kick_stale_sessions();
void handle_client_session(const MConstRef<MClientSession>& m);
void send_reconnect(MetaSession *s);
void resend_unsafe_requests(MetaSession *s);
void wait_unsafe_requests();
void dump_mds_requests(Formatter *f);
void dump_mds_sessions(Formatter *f, bool cap_dump=false);
int make_request(MetaRequest *req, const UserPerm& perms,
InodeRef *ptarget = 0, bool *pcreated = 0,
mds_rank_t use_mds=-1, bufferlist *pdirbl=0,
size_t feature_needed=ULONG_MAX);
void put_request(MetaRequest *request);
void unregister_request(MetaRequest *request);
int verify_reply_trace(int r, MetaSession *session, MetaRequest *request,
const MConstRef<MClientReply>& reply,
InodeRef *ptarget, bool *pcreated,
const UserPerm& perms);
void encode_cap_releases(MetaRequest *request, mds_rank_t mds);
int encode_inode_release(Inode *in, MetaRequest *req,
mds_rank_t mds, int drop,
int unless,int force=0);
void encode_dentry_release(Dentry *dn, MetaRequest *req,
mds_rank_t mds, int drop, int unless);
mds_rank_t choose_target_mds(MetaRequest *req, Inode** phash_diri=NULL);
void connect_mds_targets(mds_rank_t mds);
void send_request(MetaRequest *request, MetaSession *session,