1/* fs.h : interface to Subversion filesystem, private to libsvn_fs
2 *
3 * ====================================================================
4 *    Licensed to the Apache Software Foundation (ASF) under one
5 *    or more contributor license agreements.  See the NOTICE file
6 *    distributed with this work for additional information
7 *    regarding copyright ownership.  The ASF licenses this file
8 *    to you under the Apache License, Version 2.0 (the
9 *    "License"); you may not use this file except in compliance
10 *    with the License.  You may obtain a copy of the License at
11 *
12 *      http://www.apache.org/licenses/LICENSE-2.0
13 *
14 *    Unless required by applicable law or agreed to in writing,
15 *    software distributed under the License is distributed on an
16 *    "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
17 *    KIND, either express or implied.  See the License for the
18 *    specific language governing permissions and limitations
19 *    under the License.
20 * ====================================================================
21 */
22
23#ifndef SVN_LIBSVN_FS_X_H
24#define SVN_LIBSVN_FS_X_H
25
26#include <apr_pools.h>
27#include <apr_hash.h>
28#include <apr_network_io.h>
29#include <apr_md5.h>
30#include <apr_sha1.h>
31
32#include "svn_fs.h"
33#include "svn_config.h"
34#include "private/svn_atomic.h"
35#include "private/svn_cache.h"
36#include "private/svn_fs_private.h"
37#include "private/svn_sqlite.h"
38#include "private/svn_mutex.h"
39
40#include "id.h"
41
42#ifdef __cplusplus
43extern "C" {
44#endif /* __cplusplus */
45
46
47/*** The filesystem structure.  ***/
48
49/* Following are defines that specify the textual elements of the
50   native filesystem directories and revision files. */
51
52/* Names of special files in the fs_x filesystem. */
53#define PATH_FORMAT           "format"           /* Contains format number */
54#define PATH_UUID             "uuid"             /* Contains UUID */
55#define PATH_CURRENT          "current"          /* Youngest revision */
56#define PATH_LOCK_FILE        "write-lock"       /* Revision lock file */
57#define PATH_PACK_LOCK_FILE   "pack-lock"        /* Pack lock file */
58#define PATH_REVS_DIR         "revs"             /* Directory of revisions */
59#define PATH_REVPROPS_DIR     "revprops"         /* Directory of revprops */
60#define PATH_TXNS_DIR         "transactions"     /* Directory of transactions */
61#define PATH_NODE_ORIGINS_DIR "node-origins"     /* Lazy node-origin cache */
62#define PATH_TXN_PROTOS_DIR   "txn-protorevs"    /* Directory of proto-revs */
63#define PATH_TXN_CURRENT      "txn-current"      /* File with next txn key */
64#define PATH_TXN_CURRENT_LOCK "txn-current-lock" /* Lock for txn-current */
65#define PATH_LOCKS_DIR        "locks"            /* Directory of locks */
66#define PATH_MIN_UNPACKED_REV "min-unpacked-rev" /* Oldest revision which
67                                                    has not been packed. */
68#define PATH_REVPROP_GENERATION "revprop-generation"
69                                                 /* Current revprop generation*/
70#define PATH_MANIFEST         "manifest"         /* Manifest file name */
71#define PATH_PACKED           "pack"             /* Packed revision data file */
72#define PATH_EXT_PACKED_SHARD ".pack"            /* Extension for packed
73                                                    shards */
74#define PATH_EXT_L2P_INDEX    ".l2p"             /* extension of the log-
75                                                    to-phys index */
76#define PATH_EXT_P2L_INDEX    ".p2l"             /* extension of the phys-
77                                                    to-log index */
78/* If you change this, look at tests/svn_test_fs.c(maybe_install_fsx_conf) */
79#define PATH_CONFIG           "fsx.conf"         /* Configuration */
80
81/* Names of special files and file extensions for transactions */
82#define PATH_CHANGES       "changes"       /* Records changes made so far */
83#define PATH_TXN_PROPS     "props"         /* Transaction properties */
84#define PATH_TXN_PROPS_FINAL "props-final" /* Final transaction properties
85                                              before moving to revprops */
86#define PATH_NEXT_IDS      "next-ids"      /* Next temporary ID assignments */
87#define PATH_PREFIX_NODE   "node."         /* Prefix for node filename */
88#define PATH_EXT_TXN       ".txn"          /* Extension of txn dir */
89#define PATH_EXT_CHILDREN  ".children"     /* Extension for dir contents */
90#define PATH_EXT_PROPS     ".props"        /* Extension for node props */
91#define PATH_EXT_REV       ".rev"          /* Extension of protorev file */
92#define PATH_EXT_REV_LOCK  ".rev-lock"     /* Extension of protorev lock file */
93#define PATH_TXN_ITEM_INDEX "itemidx"      /* File containing the current item
94                                             index number */
95#define PATH_INDEX          "index"        /* name of index files w/o ext */
96
97/* Names of files in legacy FS formats */
98#define PATH_REV           "rev"           /* Proto rev file */
99#define PATH_REV_LOCK      "rev-lock"      /* Proto rev (write) lock file */
100
101/* Names of sections and options in fsx.conf. */
102#define CONFIG_SECTION_CACHES            "caches"
103#define CONFIG_OPTION_FAIL_STOP          "fail-stop"
104#define CONFIG_SECTION_REP_SHARING       "rep-sharing"
105#define CONFIG_OPTION_ENABLE_REP_SHARING "enable-rep-sharing"
106#define CONFIG_SECTION_DELTIFICATION     "deltification"
107#define CONFIG_OPTION_MAX_DELTIFICATION_WALK     "max-deltification-walk"
108#define CONFIG_OPTION_MAX_LINEAR_DELTIFICATION   "max-linear-deltification"
109#define CONFIG_OPTION_COMPRESSION_LEVEL  "compression-level"
110#define CONFIG_SECTION_PACKED_REVPROPS   "packed-revprops"
111#define CONFIG_OPTION_REVPROP_PACK_SIZE  "revprop-pack-size"
112#define CONFIG_OPTION_COMPRESS_PACKED_REVPROPS  "compress-packed-revprops"
113#define CONFIG_SECTION_IO                "io"
114#define CONFIG_OPTION_BLOCK_SIZE         "block-size"
115#define CONFIG_OPTION_L2P_PAGE_SIZE      "l2p-page-size"
116#define CONFIG_OPTION_P2L_PAGE_SIZE      "p2l-page-size"
117#define CONFIG_SECTION_DEBUG             "debug"
118#define CONFIG_OPTION_PACK_AFTER_COMMIT  "pack-after-commit"
119
120/* The format number of this filesystem.
121   This is independent of the repository format number, and
122   independent of any other FS back ends.
123
124   Note: If you bump this, please update the switch statement in
125         svn_fs_x__create() as well.
126 */
127#define SVN_FS_X__FORMAT_NUMBER   1
128
129/* On most operating systems apr implements file locks per process, not
130   per file.  On Windows apr implements the locking as per file handle
131   locks, so we don't have to add our own mutex for just in-process
132   synchronization. */
133#if APR_HAS_THREADS && !defined(WIN32)
134#define SVN_FS_X__USE_LOCK_MUTEX 1
135#else
136#define SVN_FS_X__USE_LOCK_MUTEX 0
137#endif
138
139/* Private FSX-specific data shared between all svn_txn_t objects that
140   relate to a particular transaction in a filesystem (as identified
141   by transaction id and filesystem UUID).  Objects of this type are
142   allocated in their own subpool of the common pool. */
143typedef struct svn_fs_x__shared_txn_data_t
144{
145  /* The next transaction in the list, or NULL if there is no following
146     transaction. */
147  struct svn_fs_x__shared_txn_data_t *next;
148
149  /* ID of this transaction. */
150  svn_fs_x__txn_id_t txn_id;
151
152  /* Whether the transaction's prototype revision file is locked for
153     writing by any thread in this process (including the current
154     thread; recursive locks are not permitted).  This is effectively
155     a non-recursive mutex. */
156  svn_boolean_t being_written;
157
158  /* The pool in which this object has been allocated; a subpool of the
159     common pool. */
160  apr_pool_t *pool;
161} svn_fs_x__shared_txn_data_t;
162
163/* Private FSX-specific data shared between all svn_fs_t objects that
164   relate to a particular filesystem, as identified by filesystem UUID.
165   Objects of this type are allocated in the common pool. */
166typedef struct svn_fs_x__shared_data_t
167{
168  /* A list of shared transaction objects for each transaction that is
169     currently active, or NULL if none are.  All access to this list,
170     including the contents of the objects stored in it, is synchronised
171     under TXN_LIST_LOCK. */
172  svn_fs_x__shared_txn_data_t *txns;
173
174  /* A free transaction object, or NULL if there is no free object.
175     Access to this object is synchronised under TXN_LIST_LOCK. */
176  svn_fs_x__shared_txn_data_t *free_txn;
177
178  /* The following lock must be taken out in reverse order of their
179     declaration here.  Any subset may be acquired and held at any given
180     time but their relative acquisition order must not change.
181
182     (lock 'txn-current' before 'pack' before 'write' before 'txn-list') */
183
184  /* A lock for intra-process synchronization when accessing the TXNS list. */
185  svn_mutex__t *txn_list_lock;
186
187  /* A lock for intra-process synchronization when grabbing the
188     repository write lock. */
189  svn_mutex__t *fs_write_lock;
190
191  /* A lock for intra-process synchronization when grabbing the
192     repository pack operation lock. */
193  svn_mutex__t *fs_pack_lock;
194
195  /* A lock for intra-process synchronization when locking the
196     txn-current file. */
197  svn_mutex__t *txn_current_lock;
198
199  /* The common pool, under which this object is allocated, subpools
200     of which are used to allocate the transaction objects. */
201  apr_pool_t *common_pool;
202} svn_fs_x__shared_data_t;
203
204/* Data structure for the 1st level DAG node cache. */
205typedef struct svn_fs_x__dag_cache_t svn_fs_x__dag_cache_t;
206
207/* Key type for all caches that use revision + offset / counter as key.
208
209   Note: Cache keys should be 16 bytes for best performance and there
210         should be no padding. */
211typedef struct svn_fs_x__pair_cache_key_t
212{
213  /* The object's revision.  Use the 64 data type to prevent padding. */
214  apr_int64_t revision;
215
216  /* Sub-address: item index, revprop generation, packed flag, etc. */
217  apr_int64_t second;
218} svn_fs_x__pair_cache_key_t;
219
220/* Key type that identifies a representation / rep header.
221
222   Note: Cache keys should require no padding. */
223typedef struct svn_fs_x__representation_cache_key_t
224{
225  /* Revision that contains the representation */
226  apr_int64_t revision;
227
228  /* Packed or non-packed representation (boolean)? */
229  apr_int64_t is_packed;
230
231  /* Item index of the representation */
232  apr_uint64_t item_index;
233} svn_fs_x__representation_cache_key_t;
234
235/* Key type that identifies a txdelta window.
236
237   Note: Cache keys should require no padding. */
238typedef struct svn_fs_x__window_cache_key_t
239{
240  /* The object's revision.  Use the 64 data type to prevent padding. */
241  apr_int64_t revision;
242
243  /* Window number within that representation. */
244  apr_int64_t chunk_index;
245
246  /* Item index of the representation */
247  apr_uint64_t item_index;
248} svn_fs_x__window_cache_key_t;
249
250/* Private (non-shared) FSX-specific data for each svn_fs_t object.
251   Any caches in here may be NULL. */
252typedef struct svn_fs_x__data_t
253{
254  /* The format number of this FS. */
255  int format;
256
257  /* The maximum number of files to store per directory. */
258  int max_files_per_dir;
259
260  /* Rev / pack file read granularity in bytes. */
261  apr_int64_t block_size;
262
263  /* Rev / pack file granularity (in bytes) covered by a single phys-to-log
264   * index page. */
265  /* Capacity in entries of log-to-phys index pages */
266  apr_int64_t l2p_page_size;
267
268  /* Rev / pack file granularity covered by phys-to-log index pages */
269  apr_int64_t p2l_page_size;
270
271  /* The revision that was youngest, last time we checked. */
272  svn_revnum_t youngest_rev_cache;
273
274  /* Caches of immutable data.  (Note that these may be shared between
275     multiple svn_fs_t's for the same filesystem.) */
276
277  /* Access to the configured memcached instances.  May be NULL. */
278  svn_memcache_t *memcache;
279
280  /* If TRUE, don't ignore any cache-related errors.  If FALSE, errors from
281     e.g. memcached may be ignored as caching is an optional feature. */
282  svn_boolean_t fail_stop;
283
284  /* Caches native dag_node_t* instances and acts as a 1st level cache */
285  svn_fs_x__dag_cache_t *dag_node_cache;
286
287  /* DAG node cache for immutable nodes.  Maps (revision, fspath)
288     to (dag_node_t *). This is the 2nd level cache for DAG nodes. */
289  svn_cache__t *rev_node_cache;
290
291  /* A cache of the contents of immutable directories; maps from
292     unparsed FS ID to a apr_hash_t * mapping (const char *) dirent
293     names to (svn_fs_x__dirent_t *). */
294  svn_cache__t *dir_cache;
295
296  /* Fulltext cache; currently only used with memcached.  Maps from
297     rep key (revision/offset) to svn_stringbuf_t. */
298  svn_cache__t *fulltext_cache;
299
300  /* Access object to the revprop "generation". Will be NULL until
301     the first access.  May be also get closed and set to NULL again. */
302  apr_file_t *revprop_generation_file;
303
304  /* Revision property cache.  Maps from (rev,generation) to apr_hash_t. */
305  svn_cache__t *revprop_cache;
306
307  /* Node properties cache.  Maps from rep key to apr_hash_t. */
308  svn_cache__t *properties_cache;
309
310  /* Pack manifest cache; a cache mapping (svn_revnum_t) shard number to
311     a manifest; and a manifest is a mapping from (svn_revnum_t) revision
312     number offset within a shard to (apr_off_t) byte-offset in the
313     respective pack file. */
314  svn_cache__t *packed_offset_cache;
315
316  /* Cache for txdelta_window_t objects;
317   * the key is svn_fs_x__window_cache_key_t */
318  svn_cache__t *txdelta_window_cache;
319
320  /* Cache for combined windows as svn_stringbuf_t objects;
321     the key is svn_fs_x__window_cache_key_t */
322  svn_cache__t *combined_window_cache;
323
324  /* Cache for svn_fs_x__rep_header_t objects;
325   * the key is (revision, item index) */
326  svn_cache__t *node_revision_cache;
327
328  /* Cache for noderevs_t containers;
329     the key is a (pack file revision, file offset) pair */
330  svn_cache__t *noderevs_container_cache;
331
332  /* Cache for change lists as APR arrays of svn_fs_x__change_t * objects;
333     the key is the revision */
334  svn_cache__t *changes_cache;
335
336  /* Cache for change_list_t containers;
337     the key is a (pack file revision, file offset) pair */
338  svn_cache__t *changes_container_cache;
339
340  /* Cache for star-delta / representation containers;
341     the key is a (pack file revision, file offset) pair */
342  svn_cache__t *reps_container_cache;
343
344  /* Cache for svn_fs_x__rep_header_t objects; the key is a
345     (revision, item index) pair */
346  svn_cache__t *rep_header_cache;
347
348  /* Cache for svn_mergeinfo_t objects; the key is a combination of
349     revision, inheritance flags and path. */
350  svn_cache__t *mergeinfo_cache;
351
352  /* Cache for presence of svn_mergeinfo_t on a noderev; the key is a
353     combination of revision, inheritance flags and path; value is "1"
354     if the node has mergeinfo, "0" if it doesn't. */
355  svn_cache__t *mergeinfo_existence_cache;
356
357  /* Cache for l2p_header_t objects; the key is (revision, is-packed).
358     Will be NULL for pre-format7 repos */
359  svn_cache__t *l2p_header_cache;
360
361  /* Cache for l2p_page_t objects; the key is svn_fs_x__page_cache_key_t.
362     Will be NULL for pre-format7 repos */
363  svn_cache__t *l2p_page_cache;
364
365  /* Cache for p2l_header_t objects; the key is (revision, is-packed).
366     Will be NULL for pre-format7 repos */
367  svn_cache__t *p2l_header_cache;
368
369  /* Cache for apr_array_header_t objects containing svn_fs_x__p2l_entry_t
370     elements; the key is svn_fs_x__page_cache_key_t.
371     Will be NULL for pre-format7 repos */
372  svn_cache__t *p2l_page_cache;
373
374  /* TRUE while the we hold a lock on the write lock file. */
375  svn_boolean_t has_write_lock;
376
377  /* Data shared between all svn_fs_t objects for a given filesystem. */
378  svn_fs_x__shared_data_t *shared;
379
380  /* The sqlite database used for rep caching. */
381  svn_sqlite__db_t *rep_cache_db;
382
383  /* Thread-safe boolean */
384  svn_atomic_t rep_cache_db_opened;
385
386  /* The oldest revision not in a pack file.  It also applies to revprops
387   * if revprop packing has been enabled by the FSX format version. */
388  svn_revnum_t min_unpacked_rev;
389
390  /* Whether rep-sharing is supported by the filesystem
391   * and allowed by the configuration. */
392  svn_boolean_t rep_sharing_allowed;
393
394  /* File size limit in bytes up to which multiple revprops shall be packed
395   * into a single file. */
396  apr_int64_t revprop_pack_size;
397
398  /* Whether packed revprop files shall be compressed. */
399  svn_boolean_t compress_packed_revprops;
400
401  /* Restart deltification histories after each multiple of this value */
402  apr_int64_t max_deltification_walk;
403
404  /* Maximum number of length of the linear part at the top of the
405   * deltification history after which skip deltas will be used. */
406  apr_int64_t max_linear_deltification;
407
408  /* Compression level to use with txdelta storage format in new revs. */
409  int delta_compression_level;
410
411  /* Pack after every commit. */
412  svn_boolean_t pack_after_commit;
413
414  /* Per-instance filesystem ID, which provides an additional level of
415     uniqueness for filesystems that share the same UUID, but should
416     still be distinguishable (e.g. backups produced by svn_fs_hotcopy()
417     or dump / load cycles). */
418  const char *instance_id;
419
420  /* Pointer to svn_fs_open. */
421  svn_error_t *(*svn_fs_open_)(svn_fs_t **, const char *, apr_hash_t *,
422                               apr_pool_t *, apr_pool_t *);
423} svn_fs_x__data_t;
424
425
426/*** Filesystem Transaction ***/
427typedef struct svn_fs_x__transaction_t
428{
429  /* property list (const char * name, svn_string_t * value).
430     may be NULL if there are no properties.  */
431  apr_hash_t *proplist;
432
433  /* revision upon which this txn is base.  (unfinished only) */
434  svn_revnum_t base_rev;
435
436  /* copies list (const char * copy_ids), or NULL if there have been
437     no copies in this transaction.  */
438  apr_array_header_t *copies;
439
440} svn_fs_x__transaction_t;
441
442
443/*** Representation ***/
444/* If you add fields to this, check to see if you need to change
445 * svn_fs_x__rep_copy. */
446typedef struct svn_fs_x__representation_t
447{
448  /* Checksums digests for the contents produced by this representation.
449     This checksum is for the contents the rep shows to consumers,
450     regardless of how the rep stores the data under the hood.  It is
451     independent of the storage (fulltext, delta, whatever).
452
453     If has_sha1 is FALSE, then for compatibility behave as though this
454     checksum matches the expected checksum.
455
456     The md5 checksum is always filled, unless this is rep which was
457     retrieved from the rep-cache.  The sha1 checksum is only computed on
458     a write, for use with rep-sharing. */
459  svn_boolean_t has_sha1;
460  unsigned char sha1_digest[APR_SHA1_DIGESTSIZE];
461  unsigned char md5_digest[APR_MD5_DIGESTSIZE];
462
463  /* Change set and item number where this representation is located. */
464  svn_fs_x__id_t id;
465
466  /* The size of the representation in bytes as seen in the revision
467     file. */
468  svn_filesize_t size;
469
470  /* The size of the fulltext of the representation. */
471  svn_filesize_t expanded_size;
472
473} svn_fs_x__representation_t;
474
475
476/*** Node-Revision ***/
477/* If you add fields to this, check to see if you need to change
478 * copy_node_revision in dag.c. */
479typedef struct svn_fs_x__noderev_t
480{
481  /* Predecessor node revision id.  Will be "unused" if there is no
482     predecessor for this node revision. */
483  svn_fs_x__id_t predecessor_id;
484
485  /* The ID of this noderev */
486  svn_fs_x__id_t noderev_id;
487
488  /* Identifier of the node that this noderev belongs to. */
489  svn_fs_x__id_t node_id;
490
491  /* Copy identifier of this line of history. */
492  svn_fs_x__id_t copy_id;
493
494  /* If this node-rev is a copy, where was it copied from? */
495  const char *copyfrom_path;
496  svn_revnum_t copyfrom_rev;
497
498  /* Helper for history tracing, root of the parent tree from whence
499     this node-rev was copied. */
500  svn_revnum_t copyroot_rev;
501  const char *copyroot_path;
502
503  /* node kind */
504  svn_node_kind_t kind;
505
506  /* number of predecessors this node revision has (recursively). */
507  int predecessor_count;
508
509  /* representation key for this node's properties.  may be NULL if
510     there are no properties.  */
511  svn_fs_x__representation_t *prop_rep;
512
513  /* representation for this node's data.  may be NULL if there is
514     no data. */
515  svn_fs_x__representation_t *data_rep;
516
517  /* path at which this node first came into existence.  */
518  const char *created_path;
519
520  /* Does this node itself have svn:mergeinfo? */
521  svn_boolean_t has_mergeinfo;
522
523  /* Number of nodes with svn:mergeinfo properties that are
524     descendants of this node (including it itself) */
525  apr_int64_t mergeinfo_count;
526
527} svn_fs_x__noderev_t;
528
529
530/** The type of a directory entry.  */
531typedef struct svn_fs_x__dirent_t
532{
533
534  /** The name of this directory entry.  */
535  const char *name;
536
537  /** The node revision ID it names.  */
538  svn_fs_x__id_t id;
539
540  /** The node kind. */
541  svn_node_kind_t kind;
542} svn_fs_x__dirent_t;
543
544
545/*** Change ***/
546typedef struct svn_fs_x__change_t
547{
548  /* Path of the change. */
549  svn_string_t path;
550
551  /* node revision id of changed path */
552  svn_fs_x__id_t noderev_id;
553
554  /* See svn_fs_path_change2_t for a description for the remaining elements.
555   */
556  svn_fs_path_change_kind_t change_kind;
557
558  svn_boolean_t text_mod;
559  svn_boolean_t prop_mod;
560  svn_node_kind_t node_kind;
561
562  svn_boolean_t copyfrom_known;
563  svn_revnum_t copyfrom_rev;
564  const char *copyfrom_path;
565
566  svn_tristate_t mergeinfo_mod;
567} svn_fs_x__change_t;
568
569
570#ifdef __cplusplus
571}
572#endif /* __cplusplus */
573
574#endif /* SVN_LIBSVN_FS_X_H */
575