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
|
/* -*- mode: c; c-basic-offset: 8; -*-
* vim: noexpandtab sw=8 ts=8 sts=0:
*
* journal.h
*
* Defines journalling api and structures.
*
* Copyright (C) 2003, 2005 Oracle. All rights reserved.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public
* License along with this program; if not, write to the
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
* Boston, MA 021110-1307, USA.
*/
#ifndef OCFS2_JOURNAL_H
#define OCFS2_JOURNAL_H
#include <linux/fs.h>
#ifndef CONFIG_OCFS2_COMPAT_JBD
# include <linux/jbd2.h>
#else
# include <linux/jbd.h>
# include "ocfs2_jbd_compat.h"
#endif
enum ocfs2_journal_state {
OCFS2_JOURNAL_FREE = 0,
OCFS2_JOURNAL_LOADED,
OCFS2_JOURNAL_IN_SHUTDOWN,
};
struct ocfs2_super;
struct ocfs2_dinode;
struct ocfs2_journal {
enum ocfs2_journal_state j_state; /* Journals current state */
journal_t *j_journal; /* The kernels journal type */
struct inode *j_inode; /* Kernel inode pointing to
* this journal */
struct ocfs2_super *j_osb; /* pointer to the super
* block for the node
* we're currently
* running on -- not
* necessarily the super
* block from the node
* which we usually run
* from (recovery,
* etc) */
struct buffer_head *j_bh; /* Journal disk inode block */
atomic_t j_num_trans; /* Number of transactions
* currently in the system. */
unsigned long j_trans_id;
struct rw_semaphore j_trans_barrier;
wait_queue_head_t j_checkpointed;
spinlock_t j_lock;
struct list_head j_la_cleanups;
struct work_struct j_recovery_work;
};
extern spinlock_t trans_inc_lock;
/* wrap j_trans_id so we never have it equal to zero. */
static inline unsigned long ocfs2_inc_trans_id(struct ocfs2_journal *j)
{
unsigned long old_id;
spin_lock(&trans_inc_lock);
old_id = j->j_trans_id++;
if (unlikely(!j->j_trans_id))
j->j_trans_id = 1;
spin_unlock(&trans_inc_lock);
return old_id;
}
static inline void ocfs2_set_inode_lock_trans(struct ocfs2_journal *journal,
struct inode *inode)
{
spin_lock(&trans_inc_lock);
OCFS2_I(inode)->ip_last_trans = journal->j_trans_id;
spin_unlock(&trans_inc_lock);
}
/* Used to figure out whether it's safe to drop a metadata lock on an
* inode. Returns true if all the inodes changes have been
* checkpointed to disk. You should be holding the spinlock on the
* metadata lock while calling this to be sure that nobody can take
* the lock and put it on another transaction. */
static inline int ocfs2_inode_fully_checkpointed(struct inode *inode)
{
int ret;
struct ocfs2_journal *journal = OCFS2_SB(inode->i_sb)->journal;
spin_lock(&trans_inc_lock);
ret = time_after(journal->j_trans_id, OCFS2_I(inode)->ip_last_trans);
spin_unlock(&trans_inc_lock);
return ret;
}
/* convenience function to check if an inode is still new (has never
* hit disk) Will do you a favor and set created_trans = 0 when you've
* been checkpointed. returns '1' if the inode is still new. */
static inline int ocfs2_inode_is_new(struct inode *inode)
{
int ret;
/* System files are never "new" as they're written out by
* mkfs. This helps us early during mount, before we have the
* journal open and j_trans_id could be junk. */
if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_SYSTEM_FILE)
return 0;
spin_lock(&trans_inc_lock);
ret = !(time_after(OCFS2_SB(inode->i_sb)->journal->j_trans_id,
OCFS2_I(inode)->ip_created_trans));
if (!ret)
OCFS2_I(inode)->ip_created_trans = 0;
spin_unlock(&trans_inc_lock);
return ret;
}
static inline void ocfs2_inode_set_new(struct ocfs2_super *osb,
struct inode *inode)
{
spin_lock(&trans_inc_lock);
OCFS2_I(inode)->ip_created_trans = osb->journal->j_trans_id;
spin_unlock(&trans_inc_lock);
}
/* Exported only for the journal struct init code in super.c. Do not call. */
void ocfs2_complete_recovery(struct work_struct *work);
void ocfs2_wait_for_recovery(struct ocfs2_super *osb);
int ocfs2_recovery_init(struct ocfs2_super *osb);
void ocfs2_recovery_exit(struct ocfs2_super *osb);
/*
* Journal Control:
* Initialize, Load, Shutdown, Wipe a journal.
*
* ocfs2_journal_init - Initialize journal structures in the OSB.
* ocfs2_journal_load - Load the given journal off disk. Replay it if
* there's transactions still in there.
* ocfs2_journal_shutdown - Shutdown a journal, this will flush all
* uncommitted, uncheckpointed transactions.
* ocfs2_journal_wipe - Wipe transactions from a journal. Optionally
* zero out each block.
* ocfs2_recovery_thread - Perform recovery on a node. osb is our own osb.
* ocfs2_mark_dead_nodes - Start recovery on nodes we won't get a heartbeat
* event on.
* ocfs2_start_checkpoint - Kick the commit thread to do a checkpoint.
*/
void ocfs2_set_journal_params(struct ocfs2_super *osb);
int ocfs2_journal_init(struct ocfs2_journal *journal,
int *dirty);
void ocfs2_journal_shutdown(struct ocfs2_super *osb);
int ocfs2_journal_wipe(struct ocfs2_journal *journal,
int full);
int ocfs2_journal_load(struct ocfs2_journal *journal, int local,
int replayed);
int ocfs2_check_journals_nolocks(struct ocfs2_super *osb);
void ocfs2_recovery_thread(struct ocfs2_super *osb,
int node_num);
int ocfs2_mark_dead_nodes(struct ocfs2_super *osb);
void ocfs2_complete_mount_recovery(struct ocfs2_super *osb);
static inline void ocfs2_start_checkpoint(struct ocfs2_super *osb)
{
atomic_set(&osb->needs_checkpoint, 1);
wake_up(&osb->checkpoint_event);
}
static inline void ocfs2_checkpoint_inode(struct inode *inode)
{
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
if (ocfs2_mount_local(osb))
return;
if (!ocfs2_inode_fully_checkpointed(inode)) {
/* WARNING: This only kicks off a single
* checkpoint. If someone races you and adds more
* metadata to the journal, you won't know, and will
* wind up waiting *alot* longer than necessary. Right
* now we only use this in clear_inode so that's
* OK. */
ocfs2_start_checkpoint(osb);
wait_event(osb->journal->j_checkpointed,
ocfs2_inode_fully_checkpointed(inode));
}
}
/*
* Transaction Handling:
* Manage the lifetime of a transaction handle.
*
* ocfs2_start_trans - Begin a transaction. Give it an upper estimate of
* the number of blocks that will be changed during
* this handle.
* ocfs2_commit_trans - Complete a handle. It might return -EIO if
* the journal was aborted. The majority of paths don't
* check the return value as an error there comes too
* late to do anything (and will be picked up in a
* later transaction).
* ocfs2_extend_trans - Extend a handle by nblocks credits. This may
* commit the handle to disk in the process, but will
* not release any locks taken during the transaction.
* ocfs2_journal_access - Notify the handle that we want to journal this
* buffer. Will have to call ocfs2_journal_dirty once
* we've actually dirtied it. Type is one of . or .
* ocfs2_journal_dirty - Mark a journalled buffer as having dirty data.
* ocfs2_jbd2_file_inode - Mark an inode so that its data goes out before
* the current handle commits.
*/
/* You must always start_trans with a number of buffs > 0, but it's
* perfectly legal to go through an entire transaction without having
* dirtied any buffers. */
handle_t *ocfs2_start_trans(struct ocfs2_super *osb,
int max_buffs);
int ocfs2_commit_trans(struct ocfs2_super *osb,
handle_t *handle);
int ocfs2_extend_trans(handle_t *handle, int nblocks);
/*
* Create access is for when we get a newly created buffer and we're
* not gonna read it off disk, but rather fill it ourselves. Right
* now, we don't do anything special with this (it turns into a write
* request), but this is a good placeholder in case we do...
*
* Write access is for when we read a block off disk and are going to
* modify it. This way the journalling layer knows it may need to make
* a copy of that block (if it's part of another, uncommitted
* transaction) before we do so.
*/
#define OCFS2_JOURNAL_ACCESS_CREATE 0
#define OCFS2_JOURNAL_ACCESS_WRITE 1
#define OCFS2_JOURNAL_ACCESS_UNDO 2
int ocfs2_journal_access(handle_t *handle,
struct inode *inode,
struct buffer_head *bh,
int type);
/*
* A word about the journal_access/journal_dirty "dance". It is
* entirely legal to journal_access a buffer more than once (as long
* as the access type is the same -- I'm not sure what will happen if
* access type is different but this should never happen anyway) It is
* also legal to journal_dirty a buffer more than once. In fact, you
* can even journal_access a buffer after you've done a
* journal_access/journal_dirty pair. The only thing you cannot do
* however, is journal_dirty a buffer which you haven't yet passed to
* journal_access at least once.
*
* That said, 99% of the time this doesn't matter and this is what the
* path looks like:
*
* <read a bh>
* ocfs2_journal_access(handle, bh, OCFS2_JOURNAL_ACCESS_WRITE);
* <modify the bh>
* ocfs2_journal_dirty(handle, bh);
*/
int ocfs2_journal_dirty(handle_t *handle,
struct buffer_head *bh);
#ifdef CONFIG_OCFS2_COMPAT_JBD
int ocfs2_journal_dirty_data(handle_t *handle,
struct buffer_head *bh);
#endif
/*
* Credit Macros:
* Convenience macros to calculate number of credits needed.
*
* For convenience sake, I have a set of macros here which calculate
* the *maximum* number of sectors which will be changed for various
* metadata updates.
*/
/* simple file updates like chmod, etc. */
#define OCFS2_INODE_UPDATE_CREDITS 1
/* extended attribute block update */
#define OCFS2_XATTR_BLOCK_UPDATE_CREDITS 1
/* group extend. inode update and last group update. */
#define OCFS2_GROUP_EXTEND_CREDITS (OCFS2_INODE_UPDATE_CREDITS + 1)
/* group add. inode update and the new group update. */
#define OCFS2_GROUP_ADD_CREDITS (OCFS2_INODE_UPDATE_CREDITS + 1)
/* get one bit out of a suballocator: dinode + group descriptor +
* prev. group desc. if we relink. */
#define OCFS2_SUBALLOC_ALLOC (3)
#define OCFS2_INLINE_TO_EXTENTS_CREDITS (OCFS2_SUBALLOC_ALLOC \
+ OCFS2_INODE_UPDATE_CREDITS)
/* dinode + group descriptor update. We don't relink on free yet. */
#define OCFS2_SUBALLOC_FREE (2)
#define OCFS2_TRUNCATE_LOG_UPDATE OCFS2_INODE_UPDATE_CREDITS
#define OCFS2_TRUNCATE_LOG_FLUSH_ONE_REC (OCFS2_SUBALLOC_FREE \
+ OCFS2_TRUNCATE_LOG_UPDATE)
#define OCFS2_REMOVE_EXTENT_CREDITS (OCFS2_TRUNCATE_LOG_UPDATE + OCFS2_INODE_UPDATE_CREDITS)
/* data block for new dir/symlink, 2 for bitmap updates (bitmap fe +
* bitmap block for the new bit) */
#define OCFS2_DIR_LINK_ADDITIONAL_CREDITS (1 + 2)
/* parent fe, parent block, new file entry, inode alloc fe, inode alloc
* group descriptor + mkdir/symlink blocks */
#define OCFS2_MKNOD_CREDITS (3 + OCFS2_SUBALLOC_ALLOC \
+ OCFS2_DIR_LINK_ADDITIONAL_CREDITS)
/* local alloc metadata change + main bitmap updates */
#define OCFS2_WINDOW_MOVE_CREDITS (OCFS2_INODE_UPDATE_CREDITS \
+ OCFS2_SUBALLOC_ALLOC + OCFS2_SUBALLOC_FREE)
/* used when we don't need an allocation change for a dir extend. One
* for the dinode, one for the new block. */
#define OCFS2_SIMPLE_DIR_EXTEND_CREDITS (2)
/* file update (nlink, etc) + directory mtime/ctime + dir entry block */
#define OCFS2_LINK_CREDITS (2*OCFS2_INODE_UPDATE_CREDITS + 1)
/* inode + dir inode (if we unlink a dir), + dir entry block + orphan
* dir inode link */
#define OCFS2_UNLINK_CREDITS (2 * OCFS2_INODE_UPDATE_CREDITS + 1 \
+ OCFS2_LINK_CREDITS)
/* dinode + orphan dir dinode + inode alloc dinode + orphan dir entry +
* inode alloc group descriptor */
#define OCFS2_DELETE_INODE_CREDITS (3 * OCFS2_INODE_UPDATE_CREDITS + 1 + 1)
/* dinode update, old dir dinode update, new dir dinode update, old
* dir dir entry, new dir dir entry, dir entry update for renaming
* directory + target unlink */
#define OCFS2_RENAME_CREDITS (3 * OCFS2_INODE_UPDATE_CREDITS + 3 \
+ OCFS2_UNLINK_CREDITS)
/* global bitmap dinode, group desc., relinked group,
* suballocator dinode, group desc., relinked group,
* dinode, xattr block */
#define OCFS2_XATTR_BLOCK_CREATE_CREDITS (OCFS2_SUBALLOC_ALLOC * 2 + \
+ OCFS2_INODE_UPDATE_CREDITS \
+ OCFS2_XATTR_BLOCK_UPDATE_CREDITS)
/*
* Please note that the caller must make sure that root_el is the root
* of extent tree. So for an inode, it should be &fe->id2.i_list. Otherwise
* the result may be wrong.
*/
static inline int ocfs2_calc_extend_credits(struct super_block *sb,
struct ocfs2_extent_list *root_el,
u32 bits_wanted)
{
int bitmap_blocks, sysfile_bitmap_blocks, extent_blocks;
/* bitmap dinode, group desc. + relinked group. */
bitmap_blocks = OCFS2_SUBALLOC_ALLOC;
/* we might need to shift tree depth so lets assume an
* absolute worst case of complete fragmentation. Even with
* that, we only need one update for the dinode, and then
* however many metadata chunks needed * a remaining suballoc
* alloc. */
sysfile_bitmap_blocks = 1 +
|