blob: 47878cf164184cf910bcbbebc10da133aa2e0c64 [file] [log] [blame]
Mark Fashehccd979b2005-12-15 14:31:24 -08001/* -*- mode: c; c-basic-offset: 8; -*-
2 * vim: noexpandtab sw=8 ts=8 sts=0:
3 *
4 * journal.c
5 *
6 * Defines functions of journalling api
7 *
8 * Copyright (C) 2003, 2004 Oracle. All rights reserved.
9 *
10 * This program is free software; you can redistribute it and/or
11 * modify it under the terms of the GNU General Public
12 * License as published by the Free Software Foundation; either
13 * version 2 of the License, or (at your option) any later version.
14 *
15 * This program is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
18 * General Public License for more details.
19 *
20 * You should have received a copy of the GNU General Public
21 * License along with this program; if not, write to the
22 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
23 * Boston, MA 021110-1307, USA.
24 */
25
26#include <linux/fs.h>
27#include <linux/types.h>
28#include <linux/slab.h>
29#include <linux/highmem.h>
30#include <linux/kthread.h>
Srinivas Eeda83273932009-06-03 17:02:55 -070031#include <linux/time.h>
32#include <linux/random.h>
Mark Fashehccd979b2005-12-15 14:31:24 -080033
34#define MLOG_MASK_PREFIX ML_JOURNAL
35#include <cluster/masklog.h>
36
37#include "ocfs2.h"
38
39#include "alloc.h"
Joel Becker50655ae2008-09-11 15:53:07 -070040#include "blockcheck.h"
Mark Fasheh316f4b92007-09-07 18:21:26 -070041#include "dir.h"
Mark Fashehccd979b2005-12-15 14:31:24 -080042#include "dlmglue.h"
43#include "extent_map.h"
44#include "heartbeat.h"
45#include "inode.h"
46#include "journal.h"
47#include "localalloc.h"
Mark Fashehccd979b2005-12-15 14:31:24 -080048#include "slot_map.h"
49#include "super.h"
Mark Fashehccd979b2005-12-15 14:31:24 -080050#include "sysfile.h"
Joel Becker0cf2f762009-02-12 16:41:25 -080051#include "uptodate.h"
Jan Kara22053632008-10-20 23:50:38 +020052#include "quota.h"
Mark Fashehccd979b2005-12-15 14:31:24 -080053
54#include "buffer_head_io.h"
55
Ingo Molnar34af9462006-06-27 02:53:55 -070056DEFINE_SPINLOCK(trans_inc_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -080057
Srinivas Eeda83273932009-06-03 17:02:55 -070058#define ORPHAN_SCAN_SCHEDULE_TIMEOUT 300000
59
Mark Fashehccd979b2005-12-15 14:31:24 -080060static int ocfs2_force_read_journal(struct inode *inode);
61static int ocfs2_recover_node(struct ocfs2_super *osb,
Jan Kara22053632008-10-20 23:50:38 +020062 int node_num, int slot_num);
Mark Fashehccd979b2005-12-15 14:31:24 -080063static int __ocfs2_recovery_thread(void *arg);
64static int ocfs2_commit_cache(struct ocfs2_super *osb);
Jan Kara19ece542008-08-21 20:13:17 +020065static int __ocfs2_wait_on_mount(struct ocfs2_super *osb, int quota);
Mark Fashehccd979b2005-12-15 14:31:24 -080066static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
Sunil Mushran539d8262008-07-14 17:31:10 -070067 int dirty, int replayed);
Mark Fashehccd979b2005-12-15 14:31:24 -080068static int ocfs2_trylock_journal(struct ocfs2_super *osb,
69 int slot_num);
70static int ocfs2_recover_orphans(struct ocfs2_super *osb,
71 int slot);
72static int ocfs2_commit_thread(void *arg);
Srinivas Eeda9140db02009-03-06 14:21:46 -080073static void ocfs2_queue_recovery_completion(struct ocfs2_journal *journal,
74 int slot_num,
75 struct ocfs2_dinode *la_dinode,
76 struct ocfs2_dinode *tl_dinode,
77 struct ocfs2_quota_recovery *qrec);
Mark Fashehccd979b2005-12-15 14:31:24 -080078
Jan Kara19ece542008-08-21 20:13:17 +020079static inline int ocfs2_wait_on_mount(struct ocfs2_super *osb)
80{
81 return __ocfs2_wait_on_mount(osb, 0);
82}
83
84static inline int ocfs2_wait_on_quotas(struct ocfs2_super *osb)
85{
86 return __ocfs2_wait_on_mount(osb, 1);
87}
88
Srinivas Eeda9140db02009-03-06 14:21:46 -080089/*
90 * This replay_map is to track online/offline slots, so we could recover
91 * offline slots during recovery and mount
92 */
93
94enum ocfs2_replay_state {
95 REPLAY_UNNEEDED = 0, /* Replay is not needed, so ignore this map */
96 REPLAY_NEEDED, /* Replay slots marked in rm_replay_slots */
97 REPLAY_DONE /* Replay was already queued */
98};
99
100struct ocfs2_replay_map {
101 unsigned int rm_slots;
102 enum ocfs2_replay_state rm_state;
103 unsigned char rm_replay_slots[0];
104};
105
106void ocfs2_replay_map_set_state(struct ocfs2_super *osb, int state)
107{
108 if (!osb->replay_map)
109 return;
110
111 /* If we've already queued the replay, we don't have any more to do */
112 if (osb->replay_map->rm_state == REPLAY_DONE)
113 return;
114
115 osb->replay_map->rm_state = state;
116}
117
118int ocfs2_compute_replay_slots(struct ocfs2_super *osb)
119{
120 struct ocfs2_replay_map *replay_map;
121 int i, node_num;
122
123 /* If replay map is already set, we don't do it again */
124 if (osb->replay_map)
125 return 0;
126
127 replay_map = kzalloc(sizeof(struct ocfs2_replay_map) +
128 (osb->max_slots * sizeof(char)), GFP_KERNEL);
129
130 if (!replay_map) {
131 mlog_errno(-ENOMEM);
132 return -ENOMEM;
133 }
134
135 spin_lock(&osb->osb_lock);
136
137 replay_map->rm_slots = osb->max_slots;
138 replay_map->rm_state = REPLAY_UNNEEDED;
139
140 /* set rm_replay_slots for offline slot(s) */
141 for (i = 0; i < replay_map->rm_slots; i++) {
142 if (ocfs2_slot_to_node_num_locked(osb, i, &node_num) == -ENOENT)
143 replay_map->rm_replay_slots[i] = 1;
144 }
145
146 osb->replay_map = replay_map;
147 spin_unlock(&osb->osb_lock);
148 return 0;
149}
150
151void ocfs2_queue_replay_slots(struct ocfs2_super *osb)
152{
153 struct ocfs2_replay_map *replay_map = osb->replay_map;
154 int i;
155
156 if (!replay_map)
157 return;
158
159 if (replay_map->rm_state != REPLAY_NEEDED)
160 return;
161
162 for (i = 0; i < replay_map->rm_slots; i++)
163 if (replay_map->rm_replay_slots[i])
164 ocfs2_queue_recovery_completion(osb->journal, i, NULL,
165 NULL, NULL);
166 replay_map->rm_state = REPLAY_DONE;
167}
168
169void ocfs2_free_replay_slots(struct ocfs2_super *osb)
170{
171 struct ocfs2_replay_map *replay_map = osb->replay_map;
172
173 if (!osb->replay_map)
174 return;
175
176 kfree(replay_map);
177 osb->replay_map = NULL;
178}
179
Joel Becker553abd02008-02-01 12:03:57 -0800180int ocfs2_recovery_init(struct ocfs2_super *osb)
181{
182 struct ocfs2_recovery_map *rm;
183
184 mutex_init(&osb->recovery_lock);
185 osb->disable_recovery = 0;
186 osb->recovery_thread_task = NULL;
187 init_waitqueue_head(&osb->recovery_event);
188
189 rm = kzalloc(sizeof(struct ocfs2_recovery_map) +
190 osb->max_slots * sizeof(unsigned int),
191 GFP_KERNEL);
192 if (!rm) {
193 mlog_errno(-ENOMEM);
194 return -ENOMEM;
195 }
196
197 rm->rm_entries = (unsigned int *)((char *)rm +
198 sizeof(struct ocfs2_recovery_map));
199 osb->recovery_map = rm;
200
201 return 0;
202}
203
204/* we can't grab the goofy sem lock from inside wait_event, so we use
205 * memory barriers to make sure that we'll see the null task before
206 * being woken up */
207static int ocfs2_recovery_thread_running(struct ocfs2_super *osb)
208{
209 mb();
210 return osb->recovery_thread_task != NULL;
211}
212
213void ocfs2_recovery_exit(struct ocfs2_super *osb)
214{
215 struct ocfs2_recovery_map *rm;
216
217 /* disable any new recovery threads and wait for any currently
218 * running ones to exit. Do this before setting the vol_state. */
219 mutex_lock(&osb->recovery_lock);
220 osb->disable_recovery = 1;
221 mutex_unlock(&osb->recovery_lock);
222 wait_event(osb->recovery_event, !ocfs2_recovery_thread_running(osb));
223
224 /* At this point, we know that no more recovery threads can be
225 * launched, so wait for any recovery completion work to
226 * complete. */
227 flush_workqueue(ocfs2_wq);
228
229 /*
230 * Now that recovery is shut down, and the osb is about to be
231 * freed, the osb_lock is not taken here.
232 */
233 rm = osb->recovery_map;
234 /* XXX: Should we bug if there are dirty entries? */
235
236 kfree(rm);
237}
238
239static int __ocfs2_recovery_map_test(struct ocfs2_super *osb,
240 unsigned int node_num)
241{
242 int i;
243 struct ocfs2_recovery_map *rm = osb->recovery_map;
244
245 assert_spin_locked(&osb->osb_lock);
246
247 for (i = 0; i < rm->rm_used; i++) {
248 if (rm->rm_entries[i] == node_num)
249 return 1;
250 }
251
252 return 0;
253}
254
255/* Behaves like test-and-set. Returns the previous value */
256static int ocfs2_recovery_map_set(struct ocfs2_super *osb,
257 unsigned int node_num)
258{
259 struct ocfs2_recovery_map *rm = osb->recovery_map;
260
261 spin_lock(&osb->osb_lock);
262 if (__ocfs2_recovery_map_test(osb, node_num)) {
263 spin_unlock(&osb->osb_lock);
264 return 1;
265 }
266
267 /* XXX: Can this be exploited? Not from o2dlm... */
268 BUG_ON(rm->rm_used >= osb->max_slots);
269
270 rm->rm_entries[rm->rm_used] = node_num;
271 rm->rm_used++;
272 spin_unlock(&osb->osb_lock);
273
274 return 0;
275}
276
277static void ocfs2_recovery_map_clear(struct ocfs2_super *osb,
278 unsigned int node_num)
279{
280 int i;
281 struct ocfs2_recovery_map *rm = osb->recovery_map;
282
283 spin_lock(&osb->osb_lock);
284
285 for (i = 0; i < rm->rm_used; i++) {
286 if (rm->rm_entries[i] == node_num)
287 break;
288 }
289
290 if (i < rm->rm_used) {
291 /* XXX: be careful with the pointer math */
292 memmove(&(rm->rm_entries[i]), &(rm->rm_entries[i + 1]),
293 (rm->rm_used - i - 1) * sizeof(unsigned int));
294 rm->rm_used--;
295 }
296
297 spin_unlock(&osb->osb_lock);
298}
299
Mark Fashehccd979b2005-12-15 14:31:24 -0800300static int ocfs2_commit_cache(struct ocfs2_super *osb)
301{
302 int status = 0;
303 unsigned int flushed;
304 unsigned long old_id;
305 struct ocfs2_journal *journal = NULL;
306
307 mlog_entry_void();
308
309 journal = osb->journal;
310
311 /* Flush all pending commits and checkpoint the journal. */
312 down_write(&journal->j_trans_barrier);
313
314 if (atomic_read(&journal->j_num_trans) == 0) {
315 up_write(&journal->j_trans_barrier);
316 mlog(0, "No transactions for me to flush!\n");
317 goto finally;
318 }
319
Joel Becker2b4e30f2008-09-03 20:03:41 -0700320 jbd2_journal_lock_updates(journal->j_journal);
321 status = jbd2_journal_flush(journal->j_journal);
322 jbd2_journal_unlock_updates(journal->j_journal);
Mark Fashehccd979b2005-12-15 14:31:24 -0800323 if (status < 0) {
324 up_write(&journal->j_trans_barrier);
325 mlog_errno(status);
326 goto finally;
327 }
328
329 old_id = ocfs2_inc_trans_id(journal);
330
331 flushed = atomic_read(&journal->j_num_trans);
332 atomic_set(&journal->j_num_trans, 0);
333 up_write(&journal->j_trans_barrier);
334
335 mlog(0, "commit_thread: flushed transaction %lu (%u handles)\n",
336 journal->j_trans_id, flushed);
337
Mark Fasheh34d024f2007-09-24 15:56:19 -0700338 ocfs2_wake_downconvert_thread(osb);
Mark Fashehccd979b2005-12-15 14:31:24 -0800339 wake_up(&journal->j_checkpointed);
340finally:
341 mlog_exit(status);
342 return status;
343}
344
Mark Fashehccd979b2005-12-15 14:31:24 -0800345/* pass it NULL and it will allocate a new handle object for you. If
346 * you pass it a handle however, it may still return error, in which
347 * case it has free'd the passed handle for you. */
Mark Fasheh1fabe142006-10-09 18:11:45 -0700348handle_t *ocfs2_start_trans(struct ocfs2_super *osb, int max_buffs)
Mark Fashehccd979b2005-12-15 14:31:24 -0800349{
Mark Fashehccd979b2005-12-15 14:31:24 -0800350 journal_t *journal = osb->journal->j_journal;
Mark Fasheh1fabe142006-10-09 18:11:45 -0700351 handle_t *handle;
Mark Fashehccd979b2005-12-15 14:31:24 -0800352
Eric Sesterhenn / snakebyteebdec832006-01-27 10:32:52 +0100353 BUG_ON(!osb || !osb->journal->j_journal);
Mark Fashehccd979b2005-12-15 14:31:24 -0800354
Mark Fasheh65eff9c2006-10-09 17:26:22 -0700355 if (ocfs2_is_hard_readonly(osb))
356 return ERR_PTR(-EROFS);
Mark Fashehccd979b2005-12-15 14:31:24 -0800357
358 BUG_ON(osb->journal->j_state == OCFS2_JOURNAL_FREE);
359 BUG_ON(max_buffs <= 0);
360
Jan Kara90e86a62008-08-27 22:30:28 +0200361 /* Nested transaction? Just return the handle... */
362 if (journal_current_handle())
363 return jbd2_journal_start(journal, max_buffs);
Mark Fashehccd979b2005-12-15 14:31:24 -0800364
Mark Fashehccd979b2005-12-15 14:31:24 -0800365 down_read(&osb->journal->j_trans_barrier);
366
Joel Becker2b4e30f2008-09-03 20:03:41 -0700367 handle = jbd2_journal_start(journal, max_buffs);
Mark Fasheh1fabe142006-10-09 18:11:45 -0700368 if (IS_ERR(handle)) {
Mark Fashehccd979b2005-12-15 14:31:24 -0800369 up_read(&osb->journal->j_trans_barrier);
370
Mark Fasheh1fabe142006-10-09 18:11:45 -0700371 mlog_errno(PTR_ERR(handle));
Mark Fashehccd979b2005-12-15 14:31:24 -0800372
373 if (is_journal_aborted(journal)) {
374 ocfs2_abort(osb->sb, "Detected aborted journal");
Mark Fasheh1fabe142006-10-09 18:11:45 -0700375 handle = ERR_PTR(-EROFS);
Mark Fashehccd979b2005-12-15 14:31:24 -0800376 }
Sunil Mushranc271c5c2006-12-05 17:56:35 -0800377 } else {
378 if (!ocfs2_mount_local(osb))
379 atomic_inc(&(osb->journal->j_num_trans));
380 }
Mark Fashehccd979b2005-12-15 14:31:24 -0800381
Mark Fashehccd979b2005-12-15 14:31:24 -0800382 return handle;
Mark Fashehccd979b2005-12-15 14:31:24 -0800383}
384
Mark Fasheh1fabe142006-10-09 18:11:45 -0700385int ocfs2_commit_trans(struct ocfs2_super *osb,
386 handle_t *handle)
Mark Fashehccd979b2005-12-15 14:31:24 -0800387{
Jan Kara90e86a62008-08-27 22:30:28 +0200388 int ret, nested;
Mark Fasheh02dc1af2006-10-09 16:48:10 -0700389 struct ocfs2_journal *journal = osb->journal;
Mark Fashehccd979b2005-12-15 14:31:24 -0800390
391 BUG_ON(!handle);
392
Jan Kara90e86a62008-08-27 22:30:28 +0200393 nested = handle->h_ref > 1;
Joel Becker2b4e30f2008-09-03 20:03:41 -0700394 ret = jbd2_journal_stop(handle);
Mark Fasheh1fabe142006-10-09 18:11:45 -0700395 if (ret < 0)
396 mlog_errno(ret);
Mark Fashehccd979b2005-12-15 14:31:24 -0800397
Jan Kara90e86a62008-08-27 22:30:28 +0200398 if (!nested)
399 up_read(&journal->j_trans_barrier);
Mark Fashehccd979b2005-12-15 14:31:24 -0800400
Mark Fasheh1fabe142006-10-09 18:11:45 -0700401 return ret;
Mark Fashehccd979b2005-12-15 14:31:24 -0800402}
403
404/*
Tao Mac901fb02010-04-26 14:34:57 +0800405 * 'nblocks' is what you want to add to the current transaction.
Mark Fashehccd979b2005-12-15 14:31:24 -0800406 *
Joel Becker2b4e30f2008-09-03 20:03:41 -0700407 * This might call jbd2_journal_restart() which will commit dirty buffers
Mark Fashehe8aed342007-12-03 16:43:01 -0800408 * and then restart the transaction. Before calling
409 * ocfs2_extend_trans(), any changed blocks should have been
410 * dirtied. After calling it, all blocks which need to be changed must
411 * go through another set of journal_access/journal_dirty calls.
412 *
Mark Fashehccd979b2005-12-15 14:31:24 -0800413 * WARNING: This will not release any semaphores or disk locks taken
414 * during the transaction, so make sure they were taken *before*
415 * start_trans or we'll have ordering deadlocks.
416 *
417 * WARNING2: Note that we do *not* drop j_trans_barrier here. This is
418 * good because transaction ids haven't yet been recorded on the
419 * cluster locks associated with this handle.
420 */
Mark Fasheh1fc58142006-10-05 14:15:36 -0700421int ocfs2_extend_trans(handle_t *handle, int nblocks)
Mark Fashehccd979b2005-12-15 14:31:24 -0800422{
Tao Mac901fb02010-04-26 14:34:57 +0800423 int status, old_nblocks;
Mark Fashehccd979b2005-12-15 14:31:24 -0800424
425 BUG_ON(!handle);
Tao Mac901fb02010-04-26 14:34:57 +0800426 BUG_ON(nblocks < 0);
Mark Fashehccd979b2005-12-15 14:31:24 -0800427
Tao Mac901fb02010-04-26 14:34:57 +0800428 if (!nblocks)
429 return 0;
430
431 old_nblocks = handle->h_buffer_credits;
Mark Fashehccd979b2005-12-15 14:31:24 -0800432 mlog_entry_void();
433
434 mlog(0, "Trying to extend transaction by %d blocks\n", nblocks);
435
Joel Beckere407e392008-06-12 22:35:39 -0700436#ifdef CONFIG_OCFS2_DEBUG_FS
Mark Fasheh0879c582007-12-03 16:42:19 -0800437 status = 1;
438#else
Joel Becker2b4e30f2008-09-03 20:03:41 -0700439 status = jbd2_journal_extend(handle, nblocks);
Mark Fashehccd979b2005-12-15 14:31:24 -0800440 if (status < 0) {
441 mlog_errno(status);
442 goto bail;
443 }
Mark Fasheh0879c582007-12-03 16:42:19 -0800444#endif
Mark Fashehccd979b2005-12-15 14:31:24 -0800445
446 if (status > 0) {
Joel Becker2b4e30f2008-09-03 20:03:41 -0700447 mlog(0,
448 "jbd2_journal_extend failed, trying "
449 "jbd2_journal_restart\n");
Tao Mac901fb02010-04-26 14:34:57 +0800450 status = jbd2_journal_restart(handle,
451 old_nblocks + nblocks);
Mark Fashehccd979b2005-12-15 14:31:24 -0800452 if (status < 0) {
Mark Fashehccd979b2005-12-15 14:31:24 -0800453 mlog_errno(status);
454 goto bail;
455 }
Mark Fasheh01ddf1e2006-10-05 13:54:39 -0700456 }
Mark Fashehccd979b2005-12-15 14:31:24 -0800457
458 status = 0;
459bail:
460
461 mlog_exit(status);
462 return status;
463}
464
Joel Becker50655ae2008-09-11 15:53:07 -0700465struct ocfs2_triggers {
466 struct jbd2_buffer_trigger_type ot_triggers;
467 int ot_offset;
468};
469
470static inline struct ocfs2_triggers *to_ocfs2_trigger(struct jbd2_buffer_trigger_type *triggers)
471{
472 return container_of(triggers, struct ocfs2_triggers, ot_triggers);
473}
474
475static void ocfs2_commit_trigger(struct jbd2_buffer_trigger_type *triggers,
476 struct buffer_head *bh,
477 void *data, size_t size)
478{
479 struct ocfs2_triggers *ot = to_ocfs2_trigger(triggers);
480
481 /*
482 * We aren't guaranteed to have the superblock here, so we
483 * must unconditionally compute the ecc data.
484 * __ocfs2_journal_access() will only set the triggers if
485 * metaecc is enabled.
486 */
487 ocfs2_block_check_compute(data, size, data + ot->ot_offset);
488}
489
490/*
491 * Quota blocks have their own trigger because the struct ocfs2_block_check
492 * offset depends on the blocksize.
493 */
494static void ocfs2_dq_commit_trigger(struct jbd2_buffer_trigger_type *triggers,
495 struct buffer_head *bh,
496 void *data, size_t size)
497{
498 struct ocfs2_disk_dqtrailer *dqt =
499 ocfs2_block_dqtrailer(size, data);
500
501 /*
502 * We aren't guaranteed to have the superblock here, so we
503 * must unconditionally compute the ecc data.
504 * __ocfs2_journal_access() will only set the triggers if
505 * metaecc is enabled.
506 */
507 ocfs2_block_check_compute(data, size, &dqt->dq_check);
508}
509
Joel Beckerc175a512008-12-10 17:58:22 -0800510/*
511 * Directory blocks also have their own trigger because the
512 * struct ocfs2_block_check offset depends on the blocksize.
513 */
514static void ocfs2_db_commit_trigger(struct jbd2_buffer_trigger_type *triggers,
515 struct buffer_head *bh,
516 void *data, size_t size)
517{
518 struct ocfs2_dir_block_trailer *trailer =
519 ocfs2_dir_trailer_from_size(size, data);
520
521 /*
522 * We aren't guaranteed to have the superblock here, so we
523 * must unconditionally compute the ecc data.
524 * __ocfs2_journal_access() will only set the triggers if
525 * metaecc is enabled.
526 */
527 ocfs2_block_check_compute(data, size, &trailer->db_check);
528}
529
Joel Becker50655ae2008-09-11 15:53:07 -0700530static void ocfs2_abort_trigger(struct jbd2_buffer_trigger_type *triggers,
531 struct buffer_head *bh)
532{
533 mlog(ML_ERROR,
534 "ocfs2_abort_trigger called by JBD2. bh = 0x%lx, "
535 "bh->b_blocknr = %llu\n",
536 (unsigned long)bh,
537 (unsigned long long)bh->b_blocknr);
538
539 /* We aren't guaranteed to have the superblock here - but if we
540 * don't, it'll just crash. */
541 ocfs2_error(bh->b_assoc_map->host->i_sb,
542 "JBD2 has aborted our journal, ocfs2 cannot continue\n");
543}
544
545static struct ocfs2_triggers di_triggers = {
546 .ot_triggers = {
547 .t_commit = ocfs2_commit_trigger,
548 .t_abort = ocfs2_abort_trigger,
549 },
550 .ot_offset = offsetof(struct ocfs2_dinode, i_check),
551};
552
553static struct ocfs2_triggers eb_triggers = {
554 .ot_triggers = {
555 .t_commit = ocfs2_commit_trigger,
556 .t_abort = ocfs2_abort_trigger,
557 },
558 .ot_offset = offsetof(struct ocfs2_extent_block, h_check),
559};
560
Tao Ma93c97082009-08-18 11:19:20 +0800561static struct ocfs2_triggers rb_triggers = {
562 .ot_triggers = {
563 .t_commit = ocfs2_commit_trigger,
564 .t_abort = ocfs2_abort_trigger,
565 },
566 .ot_offset = offsetof(struct ocfs2_refcount_block, rf_check),
567};
568
Joel Becker50655ae2008-09-11 15:53:07 -0700569static struct ocfs2_triggers gd_triggers = {
570 .ot_triggers = {
571 .t_commit = ocfs2_commit_trigger,
572 .t_abort = ocfs2_abort_trigger,
573 },
574 .ot_offset = offsetof(struct ocfs2_group_desc, bg_check),
575};
576
Joel Beckerc175a512008-12-10 17:58:22 -0800577static struct ocfs2_triggers db_triggers = {
578 .ot_triggers = {
579 .t_commit = ocfs2_db_commit_trigger,
580 .t_abort = ocfs2_abort_trigger,
581 },
582};
583
Joel Becker50655ae2008-09-11 15:53:07 -0700584static struct ocfs2_triggers xb_triggers = {
585 .ot_triggers = {
586 .t_commit = ocfs2_commit_trigger,
587 .t_abort = ocfs2_abort_trigger,
588 },
589 .ot_offset = offsetof(struct ocfs2_xattr_block, xb_check),
590};
591
592static struct ocfs2_triggers dq_triggers = {
593 .ot_triggers = {
594 .t_commit = ocfs2_dq_commit_trigger,
595 .t_abort = ocfs2_abort_trigger,
596 },
597};
598
Mark Fasheh9b7895e2008-11-12 16:27:44 -0800599static struct ocfs2_triggers dr_triggers = {
600 .ot_triggers = {
601 .t_commit = ocfs2_commit_trigger,
602 .t_abort = ocfs2_abort_trigger,
603 },
604 .ot_offset = offsetof(struct ocfs2_dx_root_block, dr_check),
605};
606
607static struct ocfs2_triggers dl_triggers = {
608 .ot_triggers = {
609 .t_commit = ocfs2_commit_trigger,
610 .t_abort = ocfs2_abort_trigger,
611 },
612 .ot_offset = offsetof(struct ocfs2_dx_leaf, dl_check),
613};
614
Joel Becker50655ae2008-09-11 15:53:07 -0700615static int __ocfs2_journal_access(handle_t *handle,
Joel Becker0cf2f762009-02-12 16:41:25 -0800616 struct ocfs2_caching_info *ci,
Joel Becker50655ae2008-09-11 15:53:07 -0700617 struct buffer_head *bh,
618 struct ocfs2_triggers *triggers,
619 int type)
Mark Fashehccd979b2005-12-15 14:31:24 -0800620{
621 int status;
Joel Becker0cf2f762009-02-12 16:41:25 -0800622 struct ocfs2_super *osb =
623 OCFS2_SB(ocfs2_metadata_cache_get_super(ci));
Mark Fashehccd979b2005-12-15 14:31:24 -0800624
Joel Becker0cf2f762009-02-12 16:41:25 -0800625 BUG_ON(!ci || !ci->ci_ops);
Mark Fashehccd979b2005-12-15 14:31:24 -0800626 BUG_ON(!handle);
627 BUG_ON(!bh);
Mark Fashehccd979b2005-12-15 14:31:24 -0800628
Badari Pulavarty205f87f2006-03-26 01:38:00 -0800629 mlog_entry("bh->b_blocknr=%llu, type=%d (\"%s\"), bh->b_size = %zu\n",
Mark Fashehccd979b2005-12-15 14:31:24 -0800630 (unsigned long long)bh->b_blocknr, type,
631 (type == OCFS2_JOURNAL_ACCESS_CREATE) ?
632 "OCFS2_JOURNAL_ACCESS_CREATE" :
633 "OCFS2_JOURNAL_ACCESS_WRITE",
634 bh->b_size);
635
636 /* we can safely remove this assertion after testing. */
637 if (!buffer_uptodate(bh)) {
638 mlog(ML_ERROR, "giving me a buffer that's not uptodate!\n");
639 mlog(ML_ERROR, "b_blocknr=%llu\n",
640 (unsigned long long)bh->b_blocknr);
641 BUG();
642 }
643
Joel Becker0cf2f762009-02-12 16:41:25 -0800644 /* Set the current transaction information on the ci so
Mark Fashehccd979b2005-12-15 14:31:24 -0800645 * that the locking code knows whether it can drop it's locks
Joel Becker0cf2f762009-02-12 16:41:25 -0800646 * on this ci or not. We're protected from the commit
Mark Fashehccd979b2005-12-15 14:31:24 -0800647 * thread updating the current transaction id until
648 * ocfs2_commit_trans() because ocfs2_start_trans() took
649 * j_trans_barrier for us. */
Joel Becker0cf2f762009-02-12 16:41:25 -0800650 ocfs2_set_ci_lock_trans(osb->journal, ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800651
Joel Becker0cf2f762009-02-12 16:41:25 -0800652 ocfs2_metadata_cache_io_lock(ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800653 switch (type) {
654 case OCFS2_JOURNAL_ACCESS_CREATE:
655 case OCFS2_JOURNAL_ACCESS_WRITE:
Joel Becker2b4e30f2008-09-03 20:03:41 -0700656 status = jbd2_journal_get_write_access(handle, bh);
Mark Fashehccd979b2005-12-15 14:31:24 -0800657 break;
658
659 case OCFS2_JOURNAL_ACCESS_UNDO:
Joel Becker2b4e30f2008-09-03 20:03:41 -0700660 status = jbd2_journal_get_undo_access(handle, bh);
Mark Fashehccd979b2005-12-15 14:31:24 -0800661 break;
662
663 default:
664 status = -EINVAL;
André Goddard Rosaaf901ca2009-11-14 13:09:05 -0200665 mlog(ML_ERROR, "Unknown access type!\n");
Mark Fashehccd979b2005-12-15 14:31:24 -0800666 }
Joel Becker0cf2f762009-02-12 16:41:25 -0800667 if (!status && ocfs2_meta_ecc(osb) && triggers)
Joel Becker50655ae2008-09-11 15:53:07 -0700668 jbd2_journal_set_triggers(bh, &triggers->ot_triggers);
Joel Becker0cf2f762009-02-12 16:41:25 -0800669 ocfs2_metadata_cache_io_unlock(ci);
Mark Fashehccd979b2005-12-15 14:31:24 -0800670
671 if (status < 0)
672 mlog(ML_ERROR, "Error %d getting %d access to buffer!\n",
673 status, type);
674
675 mlog_exit(status);
676 return status;
677}
678
Joel Becker0cf2f762009-02-12 16:41:25 -0800679int ocfs2_journal_access_di(handle_t *handle, struct ocfs2_caching_info *ci,
Joel Becker50655ae2008-09-11 15:53:07 -0700680 struct buffer_head *bh, int type)
681{
Joel Becker0cf2f762009-02-12 16:41:25 -0800682 return __ocfs2_journal_access(handle, ci, bh, &di_triggers, type);
Joel Becker50655ae2008-09-11 15:53:07 -0700683}
684
Joel Becker0cf2f762009-02-12 16:41:25 -0800685int ocfs2_journal_access_eb(handle_t *handle, struct ocfs2_caching_info *ci,
Joel Becker50655ae2008-09-11 15:53:07 -0700686 struct buffer_head *bh, int type)
687{
Joel Becker0cf2f762009-02-12 16:41:25 -0800688 return __ocfs2_journal_access(handle, ci, bh, &eb_triggers, type);
Joel Becker50655ae2008-09-11 15:53:07 -0700689}
690
Tao Ma93c97082009-08-18 11:19:20 +0800691int ocfs2_journal_access_rb(handle_t *handle, struct ocfs2_caching_info *ci,
692 struct buffer_head *bh, int type)
693{
694 return __ocfs2_journal_access(handle, ci, bh, &rb_triggers,
695 type);
696}
697
Joel Becker0cf2f762009-02-12 16:41:25 -0800698int ocfs2_journal_access_gd(handle_t *handle, struct ocfs2_caching_info *ci,
Joel Becker50655ae2008-09-11 15:53:07 -0700699 struct buffer_head *bh, int type)
700{
Joel Becker0cf2f762009-02-12 16:41:25 -0800701 return __ocfs2_journal_access(handle, ci, bh, &gd_triggers, type);
Joel Becker50655ae2008-09-11 15:53:07 -0700702}
703
Joel Becker0cf2f762009-02-12 16:41:25 -0800704int ocfs2_journal_access_db(handle_t *handle, struct ocfs2_caching_info *ci,
Joel Becker50655ae2008-09-11 15:53:07 -0700705 struct buffer_head *bh, int type)
706{
Joel Becker0cf2f762009-02-12 16:41:25 -0800707 return __ocfs2_journal_access(handle, ci, bh, &db_triggers, type);
Joel Becker50655ae2008-09-11 15:53:07 -0700708}
709
Joel Becker0cf2f762009-02-12 16:41:25 -0800710int ocfs2_journal_access_xb(handle_t *handle, struct ocfs2_caching_info *ci,
Joel Becker50655ae2008-09-11 15:53:07 -0700711 struct buffer_head *bh, int type)
712{
Joel Becker0cf2f762009-02-12 16:41:25 -0800713 return __ocfs2_journal_access(handle, ci, bh, &xb_triggers, type);
Joel Becker50655ae2008-09-11 15:53:07 -0700714}
715
Joel Becker0cf2f762009-02-12 16:41:25 -0800716int ocfs2_journal_access_dq(handle_t *handle, struct ocfs2_caching_info *ci,
Mark Fasheh9b7895e2008-11-12 16:27:44 -0800717 struct buffer_head *bh, int type)
718{
Joel Becker0cf2f762009-02-12 16:41:25 -0800719 return __ocfs2_journal_access(handle, ci, bh, &dq_triggers, type);
Mark Fasheh9b7895e2008-11-12 16:27:44 -0800720}
721
Joel Becker0cf2f762009-02-12 16:41:25 -0800722int ocfs2_journal_access_dr(handle_t *handle, struct ocfs2_caching_info *ci,
Mark Fasheh9b7895e2008-11-12 16:27:44 -0800723 struct buffer_head *bh, int type)
724{
Joel Becker0cf2f762009-02-12 16:41:25 -0800725 return __ocfs2_journal_access(handle, ci, bh, &dr_triggers, type);
Mark Fasheh9b7895e2008-11-12 16:27:44 -0800726}
727
Joel Becker0cf2f762009-02-12 16:41:25 -0800728int ocfs2_journal_access_dl(handle_t *handle, struct ocfs2_caching_info *ci,
729 struct buffer_head *bh, int type)
730{
731 return __ocfs2_journal_access(handle, ci, bh, &dl_triggers, type);
732}
733
734int ocfs2_journal_access(handle_t *handle, struct ocfs2_caching_info *ci,
Joel Becker50655ae2008-09-11 15:53:07 -0700735 struct buffer_head *bh, int type)
736{
Joel Becker0cf2f762009-02-12 16:41:25 -0800737 return __ocfs2_journal_access(handle, ci, bh, NULL, type);
Joel Becker50655ae2008-09-11 15:53:07 -0700738}
739
Joel Beckerec20cec2010-03-19 14:13:52 -0700740void ocfs2_journal_dirty(handle_t *handle, struct buffer_head *bh)
Mark Fashehccd979b2005-12-15 14:31:24 -0800741{
742 int status;
743
Mark Fashehccd979b2005-12-15 14:31:24 -0800744 mlog_entry("(bh->b_blocknr=%llu)\n",
745 (unsigned long long)bh->b_blocknr);
746
Joel Becker2b4e30f2008-09-03 20:03:41 -0700747 status = jbd2_journal_dirty_metadata(handle, bh);
Joel Beckerec20cec2010-03-19 14:13:52 -0700748 BUG_ON(status);
Mark Fashehccd979b2005-12-15 14:31:24 -0800749
Joel Beckerec20cec2010-03-19 14:13:52 -0700750 mlog_exit_void();
Mark Fashehccd979b2005-12-15 14:31:24 -0800751}
752
Joel Becker2b4e30f2008-09-03 20:03:41 -0700753#define OCFS2_DEFAULT_COMMIT_INTERVAL (HZ * JBD2_DEFAULT_MAX_COMMIT_AGE)
Mark Fashehccd979b2005-12-15 14:31:24 -0800754
755void ocfs2_set_journal_params(struct ocfs2_super *osb)
756{
757 journal_t *journal = osb->journal->j_journal;
Mark Fashehd147b3d2007-11-07 14:40:36 -0800758 unsigned long commit_interval = OCFS2_DEFAULT_COMMIT_INTERVAL;
759
760 if (osb->osb_commit_interval)
761 commit_interval = osb->osb_commit_interval;
Mark Fashehccd979b2005-12-15 14:31:24 -0800762
763 spin_lock(&journal->j_state_lock);
Mark Fashehd147b3d2007-11-07 14:40:36 -0800764 journal->j_commit_interval = commit_interval;
Mark Fashehccd979b2005-12-15 14:31:24 -0800765 if (osb->s_mount_opt & OCFS2_MOUNT_BARRIER)
Joel Becker2b4e30f2008-09-03 20:03:41 -0700766 journal->j_flags |= JBD2_BARRIER;
Mark Fashehccd979b2005-12-15 14:31:24 -0800767 else
Joel Becker2b4e30f2008-09-03 20:03:41 -0700768 journal->j_flags &= ~JBD2_BARRIER;
Mark Fashehccd979b2005-12-15 14:31:24 -0800769 spin_unlock(&journal->j_state_lock);
770}
771
772int ocfs2_journal_init(struct ocfs2_journal *journal, int *dirty)
773{
774 int status = -1;
775 struct inode *inode = NULL; /* the journal inode */
776 journal_t *j_journal = NULL;
777 struct ocfs2_dinode *di = NULL;
778 struct buffer_head *bh = NULL;
779 struct ocfs2_super *osb;
Mark Fashehe63aecb62007-10-18 15:30:42 -0700780 int inode_lock = 0;
Mark Fashehccd979b2005-12-15 14:31:24 -0800781
782 mlog_entry_void();
783
784 BUG_ON(!journal);
785
786 osb = journal->j_osb;
787
788 /* already have the inode for our journal */
789 inode = ocfs2_get_system_file_inode(osb, JOURNAL_SYSTEM_INODE,
790 osb->slot_num);
791 if (inode == NULL) {
792 status = -EACCES;
793 mlog_errno(status);
794 goto done;
795 }
796 if (is_bad_inode(inode)) {
797 mlog(ML_ERROR, "access error (bad inode)\n");
798 iput(inode);
799 inode = NULL;
800 status = -EACCES;
801 goto done;
802 }
803
804 SET_INODE_JOURNAL(inode);
805 OCFS2_I(inode)->ip_open_count++;
806
Mark Fasheh6eff5792006-01-18 10:31:47 -0800807 /* Skip recovery waits here - journal inode metadata never
808 * changes in a live cluster so it can be considered an
809 * exception to the rule. */
Mark Fashehe63aecb62007-10-18 15:30:42 -0700810 status = ocfs2_inode_lock_full(inode, &bh, 1, OCFS2_META_LOCK_RECOVERY);
Mark Fashehccd979b2005-12-15 14:31:24 -0800811 if (status < 0) {
812 if (status != -ERESTARTSYS)
813 mlog(ML_ERROR, "Could not get lock on journal!\n");
814 goto done;
815 }
816
Mark Fashehe63aecb62007-10-18 15:30:42 -0700817 inode_lock = 1;
Mark Fashehccd979b2005-12-15 14:31:24 -0800818 di = (struct ocfs2_dinode *)bh->b_data;
819
820 if (inode->i_size < OCFS2_MIN_JOURNAL_SIZE) {
821 mlog(ML_ERROR, "Journal file size (%lld) is too small!\n",
822 inode->i_size);
823 status = -EINVAL;
824 goto done;
825 }
826
827 mlog(0, "inode->i_size = %lld\n", inode->i_size);
Andrew Morton5515eff2006-03-26 01:37:53 -0800828 mlog(0, "inode->i_blocks = %llu\n",
829 (unsigned long long)inode->i_blocks);
Mark Fashehccd979b2005-12-15 14:31:24 -0800830 mlog(0, "inode->ip_clusters = %u\n", OCFS2_I(inode)->ip_clusters);
831
832 /* call the kernels journal init function now */
Joel Becker2b4e30f2008-09-03 20:03:41 -0700833 j_journal = jbd2_journal_init_inode(inode);
Mark Fashehccd979b2005-12-15 14:31:24 -0800834 if (j_journal == NULL) {
835 mlog(ML_ERROR, "Linux journal layer error\n");
836 status = -EINVAL;
837 goto done;
838 }
839
Joel Becker2b4e30f2008-09-03 20:03:41 -0700840 mlog(0, "Returned from jbd2_journal_init_inode\n");
Mark Fashehccd979b2005-12-15 14:31:24 -0800841 mlog(0, "j_journal->j_maxlen = %u\n", j_journal->j_maxlen);
842
843 *dirty = (le32_to_cpu(di->id1.journal1.ij_flags) &
844 OCFS2_JOURNAL_DIRTY_FL);
845
846 journal->j_journal = j_journal;
847 journal->j_inode = inode;
848 journal->j_bh = bh;
849
850 ocfs2_set_journal_params(osb);
851
852 journal->j_state = OCFS2_JOURNAL_LOADED;
853
854 status = 0;
855done:
856 if (status < 0) {
Mark Fashehe63aecb62007-10-18 15:30:42 -0700857 if (inode_lock)
858 ocfs2_inode_unlock(inode, 1);
Mark Fasheha81cb882008-10-07 14:25:16 -0700859 brelse(bh);
Mark Fashehccd979b2005-12-15 14:31:24 -0800860 if (inode) {
861 OCFS2_I(inode)->ip_open_count--;
862 iput(inode);
863 }
864 }
865
866 mlog_exit(status);
867 return status;
868}
869
Sunil Mushran539d8262008-07-14 17:31:10 -0700870static void ocfs2_bump_recovery_generation(struct ocfs2_dinode *di)
871{
872 le32_add_cpu(&(di->id1.journal1.ij_recovery_generation), 1);
873}
874
875static u32 ocfs2_get_recovery_generation(struct ocfs2_dinode *di)
876{
877 return le32_to_cpu(di->id1.journal1.ij_recovery_generation);
878}
879
Mark Fashehccd979b2005-12-15 14:31:24 -0800880static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
Sunil Mushran539d8262008-07-14 17:31:10 -0700881 int dirty, int replayed)
Mark Fashehccd979b2005-12-15 14:31:24 -0800882{
883 int status;
884 unsigned int flags;
885 struct ocfs2_journal *journal = osb->journal;
886 struct buffer_head *bh = journal->j_bh;
887 struct ocfs2_dinode *fe;
888
889 mlog_entry_void();
890
891 fe = (struct ocfs2_dinode *)bh->b_data;
Joel Becker10995aa2008-11-13 14:49:12 -0800892
893 /* The journal bh on the osb always comes from ocfs2_journal_init()
894 * and was validated there inside ocfs2_inode_lock_full(). It's a
895 * code bug if we mess it up. */
896 BUG_ON(!OCFS2_IS_VALID_DINODE(fe));
Mark Fashehccd979b2005-12-15 14:31:24 -0800897
898 flags = le32_to_cpu(fe->id1.journal1.ij_flags);
899 if (dirty)
900 flags |= OCFS2_JOURNAL_DIRTY_FL;
901 else
902 flags &= ~OCFS2_JOURNAL_DIRTY_FL;
903 fe->id1.journal1.ij_flags = cpu_to_le32(flags);
904
Sunil Mushran539d8262008-07-14 17:31:10 -0700905 if (replayed)
906 ocfs2_bump_recovery_generation(fe);
907
Joel Becker13723d02008-10-17 19:25:01 -0700908 ocfs2_compute_meta_ecc(osb->sb, bh->b_data, &fe->i_check);
Joel Becker8cb471e2009-02-10 20:00:41 -0800909 status = ocfs2_write_block(osb, bh, INODE_CACHE(journal->j_inode));
Mark Fashehccd979b2005-12-15 14:31:24 -0800910 if (status < 0)
911 mlog_errno(status);
912
Mark Fashehccd979b2005-12-15 14:31:24 -0800913 mlog_exit(status);
914 return status;
915}
916
917/*
918 * If the journal has been kmalloc'd it needs to be freed after this
919 * call.
920 */
921void ocfs2_journal_shutdown(struct ocfs2_super *osb)
922{
923 struct ocfs2_journal *journal = NULL;
924 int status = 0;
925 struct inode *inode = NULL;
926 int num_running_trans = 0;
927
928 mlog_entry_void();
929
Eric Sesterhenn / snakebyteebdec832006-01-27 10:32:52 +0100930 BUG_ON(!osb);
Mark Fashehccd979b2005-12-15 14:31:24 -0800931
932 journal = osb->journal;
933 if (!journal)
934 goto done;
935
936 inode = journal->j_inode;
937
938 if (journal->j_state != OCFS2_JOURNAL_LOADED)
939 goto done;
940
Joel Becker2b4e30f2008-09-03 20:03:41 -0700941 /* need to inc inode use count - jbd2_journal_destroy will iput. */
Mark Fashehccd979b2005-12-15 14:31:24 -0800942 if (!igrab(inode))
943 BUG();
944
945 num_running_trans = atomic_read(&(osb->journal->j_num_trans));
946 if (num_running_trans > 0)
947 mlog(0, "Shutting down journal: must wait on %d "
948 "running transactions!\n",
949 num_running_trans);
950
951 /* Do a commit_cache here. It will flush our journal, *and*
952 * release any locks that are still held.
953 * set the SHUTDOWN flag and release the trans lock.
954 * the commit thread will take the trans lock for us below. */
955 journal->j_state = OCFS2_JOURNAL_IN_SHUTDOWN;
956
957 /* The OCFS2_JOURNAL_IN_SHUTDOWN will signal to commit_cache to not
958 * drop the trans_lock (which we want to hold until we
959 * completely destroy the journal. */
960 if (osb->commit_task) {
961 /* Wait for the commit thread */
962 mlog(0, "Waiting for ocfs2commit to exit....\n");
963 kthread_stop(osb->commit_task);
964 osb->commit_task = NULL;
965 }
966
967 BUG_ON(atomic_read(&(osb->journal->j_num_trans)) != 0);
968
Sunil Mushranc271c5c2006-12-05 17:56:35 -0800969 if (ocfs2_mount_local(osb)) {
Joel Becker2b4e30f2008-09-03 20:03:41 -0700970 jbd2_journal_lock_updates(journal->j_journal);
971 status = jbd2_journal_flush(journal->j_journal);
972 jbd2_journal_unlock_updates(journal->j_journal);
Sunil Mushranc271c5c2006-12-05 17:56:35 -0800973 if (status < 0)
974 mlog_errno(status);
975 }
976
977 if (status == 0) {
978 /*
979 * Do not toggle if flush was unsuccessful otherwise
980 * will leave dirty metadata in a "clean" journal
981 */
Sunil Mushran539d8262008-07-14 17:31:10 -0700982 status = ocfs2_journal_toggle_dirty(osb, 0, 0);
Sunil Mushranc271c5c2006-12-05 17:56:35 -0800983 if (status < 0)
984 mlog_errno(status);
985 }
Mark Fashehccd979b2005-12-15 14:31:24 -0800986
987 /* Shutdown the kernel journal system */
Joel Becker2b4e30f2008-09-03 20:03:41 -0700988 jbd2_journal_destroy(journal->j_journal);
Sunil Mushranae0dff62008-10-22 13:24:29 -0700989 journal->j_journal = NULL;
Mark Fashehccd979b2005-12-15 14:31:24 -0800990
991 OCFS2_I(inode)->ip_open_count--;
992
993 /* unlock our journal */
Mark Fashehe63aecb62007-10-18 15:30:42 -0700994 ocfs2_inode_unlock(inode, 1);
Mark Fashehccd979b2005-12-15 14:31:24 -0800995
996 brelse(journal->j_bh);
997 journal->j_bh = NULL;
998
999 journal->j_state = OCFS2_JOURNAL_FREE;
1000
1001// up_write(&journal->j_trans_barrier);
1002done:
1003 if (inode)
1004 iput(inode);
1005 mlog_exit_void();
1006}
1007
1008static void ocfs2_clear_journal_error(struct super_block *sb,
1009 journal_t *journal,
1010 int slot)
1011{
1012 int olderr;
1013
Joel Becker2b4e30f2008-09-03 20:03:41 -07001014 olderr = jbd2_journal_errno(journal);
Mark Fashehccd979b2005-12-15 14:31:24 -08001015 if (olderr) {
1016 mlog(ML_ERROR, "File system error %d recorded in "
1017 "journal %u.\n", olderr, slot);
1018 mlog(ML_ERROR, "File system on device %s needs checking.\n",
1019 sb->s_id);
1020
Joel Becker2b4e30f2008-09-03 20:03:41 -07001021 jbd2_journal_ack_err(journal);
1022 jbd2_journal_clear_err(journal);
Mark Fashehccd979b2005-12-15 14:31:24 -08001023 }
1024}
1025
Sunil Mushran539d8262008-07-14 17:31:10 -07001026int ocfs2_journal_load(struct ocfs2_journal *journal, int local, int replayed)
Mark Fashehccd979b2005-12-15 14:31:24 -08001027{
1028 int status = 0;
1029 struct ocfs2_super *osb;
1030
1031 mlog_entry_void();
1032
Julia Lawallb1f35502008-03-04 15:21:05 -08001033 BUG_ON(!journal);
Mark Fashehccd979b2005-12-15 14:31:24 -08001034
1035 osb = journal->j_osb;
1036
Joel Becker2b4e30f2008-09-03 20:03:41 -07001037 status = jbd2_journal_load(journal->j_journal);
Mark Fashehccd979b2005-12-15 14:31:24 -08001038 if (status < 0) {
1039 mlog(ML_ERROR, "Failed to load journal!\n");
1040 goto done;
1041 }
1042
1043 ocfs2_clear_journal_error(osb->sb, journal->j_journal, osb->slot_num);
1044
Sunil Mushran539d8262008-07-14 17:31:10 -07001045 status = ocfs2_journal_toggle_dirty(osb, 1, replayed);
Mark Fashehccd979b2005-12-15 14:31:24 -08001046 if (status < 0) {
1047 mlog_errno(status);
1048 goto done;
1049 }
1050
1051 /* Launch the commit thread */
Sunil Mushranc271c5c2006-12-05 17:56:35 -08001052 if (!local) {
1053 osb->commit_task = kthread_run(ocfs2_commit_thread, osb,
1054 "ocfs2cmt");
1055 if (IS_ERR(osb->commit_task)) {
1056 status = PTR_ERR(osb->commit_task);
1057 osb->commit_task = NULL;
1058 mlog(ML_ERROR, "unable to launch ocfs2commit thread, "
1059 "error=%d", status);
1060 goto done;
1061 }
1062 } else
Mark Fashehccd979b2005-12-15 14:31:24 -08001063 osb->commit_task = NULL;
Mark Fashehccd979b2005-12-15 14:31:24 -08001064
1065done:
1066 mlog_exit(status);
1067 return status;
1068}
1069
1070
1071/* 'full' flag tells us whether we clear out all blocks or if we just
1072 * mark the journal clean */
1073int ocfs2_journal_wipe(struct ocfs2_journal *journal, int full)
1074{
1075 int status;
1076
1077 mlog_entry_void();
1078
Eric Sesterhenn / snakebyteebdec832006-01-27 10:32:52 +01001079 BUG_ON(!journal);
Mark Fashehccd979b2005-12-15 14:31:24 -08001080
Joel Becker2b4e30f2008-09-03 20:03:41 -07001081 status = jbd2_journal_wipe(journal->j_journal, full);
Mark Fashehccd979b2005-12-15 14:31:24 -08001082 if (status < 0) {
1083 mlog_errno(status);
1084 goto bail;
1085 }
1086
Sunil Mushran539d8262008-07-14 17:31:10 -07001087 status = ocfs2_journal_toggle_dirty(journal->j_osb, 0, 0);
Mark Fashehccd979b2005-12-15 14:31:24 -08001088 if (status < 0)
1089 mlog_errno(status);
1090
1091bail:
1092 mlog_exit(status);
1093 return status;
1094}
1095
Joel Becker553abd02008-02-01 12:03:57 -08001096static int ocfs2_recovery_completed(struct ocfs2_super *osb)
1097{
1098 int empty;
1099 struct ocfs2_recovery_map *rm = osb->recovery_map;
1100
1101 spin_lock(&osb->osb_lock);
1102 empty = (rm->rm_used == 0);
1103 spin_unlock(&osb->osb_lock);
1104
1105 return empty;
1106}
1107
1108void ocfs2_wait_for_recovery(struct ocfs2_super *osb)
1109{
1110 wait_event(osb->recovery_event, ocfs2_recovery_completed(osb));
1111}
1112
Mark Fashehccd979b2005-12-15 14:31:24 -08001113/*
1114 * JBD Might read a cached version of another nodes journal file. We
1115 * don't want this as this file changes often and we get no
1116 * notification on those changes. The only way to be sure that we've
1117 * got the most up to date version of those blocks then is to force
1118 * read them off disk. Just searching through the buffer cache won't
1119 * work as there may be pages backing this file which are still marked
1120 * up to date. We know things can't change on this file underneath us
1121 * as we have the lock by now :)
1122 */
1123static int ocfs2_force_read_journal(struct inode *inode)
1124{
1125 int status = 0;
Mark Fasheh4f902c32007-03-09 16:26:50 -08001126 int i;
Mark Fasheh8110b072007-03-22 16:53:23 -07001127 u64 v_blkno, p_blkno, p_blocks, num_blocks;
Mark Fasheh4f902c32007-03-09 16:26:50 -08001128#define CONCURRENT_JOURNAL_FILL 32ULL
Mark Fashehccd979b2005-12-15 14:31:24 -08001129 struct buffer_head *bhs[CONCURRENT_JOURNAL_FILL];
1130
1131 mlog_entry_void();
1132
Mark Fashehccd979b2005-12-15 14:31:24 -08001133 memset(bhs, 0, sizeof(struct buffer_head *) * CONCURRENT_JOURNAL_FILL);
1134
Mark Fasheh8110b072007-03-22 16:53:23 -07001135 num_blocks = ocfs2_blocks_for_bytes(inode->i_sb, inode->i_size);
Mark Fashehccd979b2005-12-15 14:31:24 -08001136 v_blkno = 0;
Mark Fasheh8110b072007-03-22 16:53:23 -07001137 while (v_blkno < num_blocks) {
Mark Fashehccd979b2005-12-15 14:31:24 -08001138 status = ocfs2_extent_map_get_blocks(inode, v_blkno,
Mark Fasheh49cb8d22007-03-09 16:21:46 -08001139 &p_blkno, &p_blocks, NULL);
Mark Fashehccd979b2005-12-15 14:31:24 -08001140 if (status < 0) {
1141 mlog_errno(status);
1142 goto bail;
1143 }
1144
1145 if (p_blocks > CONCURRENT_JOURNAL_FILL)
1146 p_blocks = CONCURRENT_JOURNAL_FILL;
1147
Mark Fashehdd4a2c22006-04-12 14:24:05 -07001148 /* We are reading journal data which should not
1149 * be put in the uptodate cache */
Joel Beckerda1e9092008-10-09 17:20:29 -07001150 status = ocfs2_read_blocks_sync(OCFS2_SB(inode->i_sb),
1151 p_blkno, p_blocks, bhs);
Mark Fashehccd979b2005-12-15 14:31:24 -08001152 if (status < 0) {
1153 mlog_errno(status);
1154 goto bail;
1155 }
1156
1157 for(i = 0; i < p_blocks; i++) {
1158 brelse(bhs[i]);
1159 bhs[i] = NULL;
1160 }
1161
1162 v_blkno += p_blocks;
1163 }
1164
1165bail:
1166 for(i = 0; i < CONCURRENT_JOURNAL_FILL; i++)
Mark Fasheha81cb882008-10-07 14:25:16 -07001167 brelse(bhs[i]);
Mark Fashehccd979b2005-12-15 14:31:24 -08001168 mlog_exit(status);
1169 return status;
1170}
1171
1172struct ocfs2_la_recovery_item {
1173 struct list_head lri_list;
1174 int lri_slot;
1175 struct ocfs2_dinode *lri_la_dinode;
1176 struct ocfs2_dinode *lri_tl_dinode;
Jan Kara22053632008-10-20 23:50:38 +02001177 struct ocfs2_quota_recovery *lri_qrec;
Mark Fashehccd979b2005-12-15 14:31:24 -08001178};
1179
1180/* Does the second half of the recovery process. By this point, the
1181 * node is marked clean and can actually be considered recovered,
1182 * hence it's no longer in the recovery map, but there's still some
1183 * cleanup we can do which shouldn't happen within the recovery thread
1184 * as locking in that context becomes very difficult if we are to take
1185 * recovering nodes into account.
1186 *
1187 * NOTE: This function can and will sleep on recovery of other nodes
1188 * during cluster locking, just like any other ocfs2 process.
1189 */
David Howellsc4028952006-11-22 14:57:56 +00001190void ocfs2_complete_recovery(struct work_struct *work)
Mark Fashehccd979b2005-12-15 14:31:24 -08001191{
1192 int ret;
David Howellsc4028952006-11-22 14:57:56 +00001193 struct ocfs2_journal *journal =
1194 container_of(work, struct ocfs2_journal, j_recovery_work);
1195 struct ocfs2_super *osb = journal->j_osb;
Mark Fashehccd979b2005-12-15 14:31:24 -08001196 struct ocfs2_dinode *la_dinode, *tl_dinode;
Christoph Hellwig800deef2007-05-17 16:03:13 +02001197 struct ocfs2_la_recovery_item *item, *n;
Jan Kara22053632008-10-20 23:50:38 +02001198 struct ocfs2_quota_recovery *qrec;
Mark Fashehccd979b2005-12-15 14:31:24 -08001199 LIST_HEAD(tmp_la_list);
1200
1201 mlog_entry_void();
1202
1203 mlog(0, "completing recovery from keventd\n");
1204
1205 spin_lock(&journal->j_lock);
1206 list_splice_init(&journal->j_la_cleanups, &tmp_la_list);
1207 spin_unlock(&journal->j_lock);
1208
Christoph Hellwig800deef2007-05-17 16:03:13 +02001209 list_for_each_entry_safe(item, n, &tmp_la_list, lri_list) {
Mark Fashehccd979b2005-12-15 14:31:24 -08001210 list_del_init(&item->lri_list);
1211
1212 mlog(0, "Complete recovery for slot %d\n", item->lri_slot);
1213
Jan Kara19ece542008-08-21 20:13:17 +02001214 ocfs2_wait_on_quotas(osb);
1215
Mark Fashehccd979b2005-12-15 14:31:24 -08001216 la_dinode = item->lri_la_dinode;
1217 if (la_dinode) {
Mark Fashehb0697052006-03-03 10:24:33 -08001218 mlog(0, "Clean up local alloc %llu\n",
Mark Fasheh1ca1a112007-04-27 16:01:25 -07001219 (unsigned long long)le64_to_cpu(la_dinode->i_blkno));
Mark Fashehccd979b2005-12-15 14:31:24 -08001220
1221 ret = ocfs2_complete_local_alloc_recovery(osb,
1222 la_dinode);
1223 if (ret < 0)
1224 mlog_errno(ret);
1225
1226 kfree(la_dinode);
1227 }
1228
1229 tl_dinode = item->lri_tl_dinode;
1230 if (tl_dinode) {
Mark Fashehb0697052006-03-03 10:24:33 -08001231 mlog(0, "Clean up truncate log %llu\n",
Mark Fasheh1ca1a112007-04-27 16:01:25 -07001232 (unsigned long long)le64_to_cpu(tl_dinode->i_blkno));
Mark Fashehccd979b2005-12-15 14:31:24 -08001233
1234 ret = ocfs2_complete_truncate_log_recovery(osb,
1235 tl_dinode);
1236 if (ret < 0)
1237 mlog_errno(ret);
1238
1239 kfree(tl_dinode);
1240 }
1241
1242 ret = ocfs2_recover_orphans(osb, item->lri_slot);
1243 if (ret < 0)
1244 mlog_errno(ret);
1245
Jan Kara22053632008-10-20 23:50:38 +02001246 qrec = item->lri_qrec;
1247 if (qrec) {
1248 mlog(0, "Recovering quota files");
1249 ret = ocfs2_finish_quota_recovery(osb, qrec,
1250 item->lri_slot);
1251 if (ret < 0)
1252 mlog_errno(ret);
1253 /* Recovery info is already freed now */
1254 }
1255
Mark Fashehccd979b2005-12-15 14:31:24 -08001256 kfree(item);
1257 }
1258
1259 mlog(0, "Recovery completion\n");
1260 mlog_exit_void();
1261}
1262
1263/* NOTE: This function always eats your references to la_dinode and
1264 * tl_dinode, either manually on error, or by passing them to
1265 * ocfs2_complete_recovery */
1266static void ocfs2_queue_recovery_completion(struct ocfs2_journal *journal,
1267 int slot_num,
1268 struct ocfs2_dinode *la_dinode,
Jan Kara22053632008-10-20 23:50:38 +02001269 struct ocfs2_dinode *tl_dinode,
1270 struct ocfs2_quota_recovery *qrec)
Mark Fashehccd979b2005-12-15 14:31:24 -08001271{
1272 struct ocfs2_la_recovery_item *item;
1273
Sunil Mushranafae00ab2006-04-12 14:37:00 -07001274 item = kmalloc(sizeof(struct ocfs2_la_recovery_item), GFP_NOFS);
Mark Fashehccd979b2005-12-15 14:31:24 -08001275 if (!item) {
1276 /* Though we wish to avoid it, we are in fact safe in
1277 * skipping local alloc cleanup as fsck.ocfs2 is more
1278 * than capable of reclaiming unused space. */
1279 if (la_dinode)
1280 kfree(la_dinode);
1281
1282 if (tl_dinode)
1283 kfree(tl_dinode);
1284
Jan Kara22053632008-10-20 23:50:38 +02001285 if (qrec)
1286 ocfs2_free_quota_recovery(qrec);
1287
Mark Fashehccd979b2005-12-15 14:31:24 -08001288 mlog_errno(-ENOMEM);
1289 return;
1290 }
1291
1292 INIT_LIST_HEAD(&item->lri_list);
1293 item->lri_la_dinode = la_dinode;
1294 item->lri_slot = slot_num;
1295 item->lri_tl_dinode = tl_dinode;
Jan Kara22053632008-10-20 23:50:38 +02001296 item->lri_qrec = qrec;
Mark Fashehccd979b2005-12-15 14:31:24 -08001297
1298 spin_lock(&journal->j_lock);
1299 list_add_tail(&item->lri_list, &journal->j_la_cleanups);
1300 queue_work(ocfs2_wq, &journal->j_recovery_work);
1301 spin_unlock(&journal->j_lock);
1302}
1303
1304/* Called by the mount code to queue recovery the last part of
Srinivas Eeda9140db02009-03-06 14:21:46 -08001305 * recovery for it's own and offline slot(s). */
Mark Fashehccd979b2005-12-15 14:31:24 -08001306void ocfs2_complete_mount_recovery(struct ocfs2_super *osb)
1307{
1308 struct ocfs2_journal *journal = osb->journal;
1309
Srinivas Eeda9140db02009-03-06 14:21:46 -08001310 /* No need to queue up our truncate_log as regular cleanup will catch
1311 * that */
1312 ocfs2_queue_recovery_completion(journal, osb->slot_num,
1313 osb->local_alloc_copy, NULL, NULL);
1314 ocfs2_schedule_truncate_log_flush(osb, 0);
Mark Fashehccd979b2005-12-15 14:31:24 -08001315
Srinivas Eeda9140db02009-03-06 14:21:46 -08001316 osb->local_alloc_copy = NULL;
1317 osb->dirty = 0;
1318
1319 /* queue to recover orphan slots for all offline slots */
1320 ocfs2_replay_map_set_state(osb, REPLAY_NEEDED);
1321 ocfs2_queue_replay_slots(osb);
1322 ocfs2_free_replay_slots(osb);
Mark Fashehccd979b2005-12-15 14:31:24 -08001323}
1324
Jan Kara22053632008-10-20 23:50:38 +02001325void ocfs2_complete_quota_recovery(struct ocfs2_super *osb)
1326{
1327 if (osb->quota_rec) {
1328 ocfs2_queue_recovery_completion(osb->journal,
1329 osb->slot_num,
1330 NULL,
1331 NULL,
1332 osb->quota_rec);
1333 osb->quota_rec = NULL;
1334 }
1335}
1336
Mark Fashehccd979b2005-12-15 14:31:24 -08001337static int __ocfs2_recovery_thread(void *arg)
1338{
Jan Kara22053632008-10-20 23:50:38 +02001339 int status, node_num, slot_num;
Mark Fashehccd979b2005-12-15 14:31:24 -08001340 struct ocfs2_super *osb = arg;
Joel Becker553abd02008-02-01 12:03:57 -08001341 struct ocfs2_recovery_map *rm = osb->recovery_map;
Jan Kara22053632008-10-20 23:50:38 +02001342 int *rm_quota = NULL;
1343 int rm_quota_used = 0, i;
1344 struct ocfs2_quota_recovery *qrec;
Mark Fashehccd979b2005-12-15 14:31:24 -08001345
1346 mlog_entry_void();
1347
1348 status = ocfs2_wait_on_mount(osb);
1349 if (status < 0) {
1350 goto bail;
1351 }
1352
Jan Kara22053632008-10-20 23:50:38 +02001353 rm_quota = kzalloc(osb->max_slots * sizeof(int), GFP_NOFS);
1354 if (!rm_quota) {
1355 status = -ENOMEM;
1356 goto bail;
1357 }
Mark Fashehccd979b2005-12-15 14:31:24 -08001358restart:
1359 status = ocfs2_super_lock(osb, 1);
1360 if (status < 0) {
1361 mlog_errno(status);
1362 goto bail;
1363 }
1364
Srinivas Eeda9140db02009-03-06 14:21:46 -08001365 status = ocfs2_compute_replay_slots(osb);
1366 if (status < 0)
1367 mlog_errno(status);
1368
1369 /* queue recovery for our own slot */
1370 ocfs2_queue_recovery_completion(osb->journal, osb->slot_num, NULL,
1371 NULL, NULL);
1372
Joel Becker553abd02008-02-01 12:03:57 -08001373 spin_lock(&osb->osb_lock);
1374 while (rm->rm_used) {
1375 /* It's always safe to remove entry zero, as we won't
1376 * clear it until ocfs2_recover_node() has succeeded. */
1377 node_num = rm->rm_entries[0];
1378 spin_unlock(&osb->osb_lock);
Jan Kara22053632008-10-20 23:50:38 +02001379 mlog(0, "checking node %d\n", node_num);
1380 slot_num = ocfs2_node_num_to_slot(osb, node_num);
1381 if (slot_num == -ENOENT) {
1382 status = 0;
1383 mlog(0, "no slot for this node, so no recovery"
1384 "required.\n");
1385 goto skip_recovery;
1386 }
1387 mlog(0, "node %d was using slot %d\n", node_num, slot_num);
Mark Fashehccd979b2005-12-15 14:31:24 -08001388
Jan Kara22053632008-10-20 23:50:38 +02001389 /* It is a bit subtle with quota recovery. We cannot do it
1390 * immediately because we have to obtain cluster locks from
1391 * quota files and we also don't want to just skip it because
1392 * then quota usage would be out of sync until some node takes
1393 * the slot. So we remember which nodes need quota recovery
1394 * and when everything else is done, we recover quotas. */
1395 for (i = 0; i < rm_quota_used && rm_quota[i] != slot_num; i++);
1396 if (i == rm_quota_used)
1397 rm_quota[rm_quota_used++] = slot_num;
1398
1399 status = ocfs2_recover_node(osb, node_num, slot_num);
1400skip_recovery:
Joel Becker553abd02008-02-01 12:03:57 -08001401 if (!status) {
1402 ocfs2_recovery_map_clear(osb, node_num);
1403 } else {
Mark Fashehccd979b2005-12-15 14:31:24 -08001404 mlog(ML_ERROR,
1405 "Error %d recovering node %d on device (%u,%u)!\n",
1406 status, node_num,
1407 MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev));
1408 mlog(ML_ERROR, "Volume requires unmount.\n");
Mark Fashehccd979b2005-12-15 14:31:24 -08001409 }
1410
Joel Becker553abd02008-02-01 12:03:57 -08001411 spin_lock(&osb->osb_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -08001412 }
Joel Becker553abd02008-02-01 12:03:57 -08001413 spin_unlock(&osb->osb_lock);
1414 mlog(0, "All nodes recovered\n");
1415
Sunil Mushran539d8262008-07-14 17:31:10 -07001416 /* Refresh all journal recovery generations from disk */
1417 status = ocfs2_check_journals_nolocks(osb);
1418 status = (status == -EROFS) ? 0 : status;
1419 if (status < 0)
1420 mlog_errno(status);
1421
Jan Kara22053632008-10-20 23:50:38 +02001422 /* Now it is right time to recover quotas... We have to do this under
1423 * superblock lock so that noone can start using the slot (and crash)
1424 * before we recover it */
1425 for (i = 0; i < rm_quota_used; i++) {
1426 qrec = ocfs2_begin_quota_recovery(osb, rm_quota[i]);
1427 if (IS_ERR(qrec)) {
1428 status = PTR_ERR(qrec);
1429 mlog_errno(status);
1430 continue;
1431 }
1432 ocfs2_queue_recovery_completion(osb->journal, rm_quota[i],
1433 NULL, NULL, qrec);
1434 }
1435
Mark Fashehccd979b2005-12-15 14:31:24 -08001436 ocfs2_super_unlock(osb, 1);
1437
Srinivas Eeda9140db02009-03-06 14:21:46 -08001438 /* queue recovery for offline slots */
1439 ocfs2_queue_replay_slots(osb);
Mark Fashehccd979b2005-12-15 14:31:24 -08001440
1441bail:
Arjan van de Venc74ec2f2006-01-13 21:54:23 -08001442 mutex_lock(&osb->recovery_lock);
Joel Becker553abd02008-02-01 12:03:57 -08001443 if (!status && !ocfs2_recovery_completed(osb)) {
Arjan van de Venc74ec2f2006-01-13 21:54:23 -08001444 mutex_unlock(&osb->recovery_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -08001445 goto restart;
1446 }
1447
Srinivas Eeda9140db02009-03-06 14:21:46 -08001448 ocfs2_free_replay_slots(osb);
Mark Fashehccd979b2005-12-15 14:31:24 -08001449 osb->recovery_thread_task = NULL;
1450 mb(); /* sync with ocfs2_recovery_thread_running */
1451 wake_up(&osb->recovery_event);
1452
Arjan van de Venc74ec2f2006-01-13 21:54:23 -08001453 mutex_unlock(&osb->recovery_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -08001454
Jan Kara22053632008-10-20 23:50:38 +02001455 if (rm_quota)
1456 kfree(rm_quota);
1457
Mark Fashehccd979b2005-12-15 14:31:24 -08001458 mlog_exit(status);
1459 /* no one is callint kthread_stop() for us so the kthread() api
1460 * requires that we call do_exit(). And it isn't exported, but
1461 * complete_and_exit() seems to be a minimal wrapper around it. */
1462 complete_and_exit(NULL, status);
1463 return status;
1464}
1465
1466void ocfs2_recovery_thread(struct ocfs2_super *osb, int node_num)
1467{
1468 mlog_entry("(node_num=%d, osb->node_num = %d)\n",
1469 node_num, osb->node_num);
1470
Arjan van de Venc74ec2f2006-01-13 21:54:23 -08001471 mutex_lock(&osb->recovery_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -08001472 if (osb->disable_recovery)
1473 goto out;
1474
1475 /* People waiting on recovery will wait on
1476 * the recovery map to empty. */
Joel Becker553abd02008-02-01 12:03:57 -08001477 if (ocfs2_recovery_map_set(osb, node_num))
1478 mlog(0, "node %d already in recovery map.\n", node_num);
Mark Fashehccd979b2005-12-15 14:31:24 -08001479
1480 mlog(0, "starting recovery thread...\n");
1481
1482 if (osb->recovery_thread_task)
1483 goto out;
1484
1485 osb->recovery_thread_task = kthread_run(__ocfs2_recovery_thread, osb,
Mark Fasheh78427042006-05-04 12:03:26 -07001486 "ocfs2rec");
Mark Fashehccd979b2005-12-15 14:31:24 -08001487 if (IS_ERR(osb->recovery_thread_task)) {
1488 mlog_errno((int)PTR_ERR(osb->recovery_thread_task));
1489 osb->recovery_thread_task = NULL;
1490 }
1491
1492out:
Arjan van de Venc74ec2f2006-01-13 21:54:23 -08001493 mutex_unlock(&osb->recovery_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -08001494 wake_up(&osb->recovery_event);
1495
1496 mlog_exit_void();
1497}
1498
Sunil Mushran539d8262008-07-14 17:31:10 -07001499static int ocfs2_read_journal_inode(struct ocfs2_super *osb,
1500 int slot_num,
1501 struct buffer_head **bh,
1502 struct inode **ret_inode)
1503{
1504 int status = -EACCES;
1505 struct inode *inode = NULL;
1506
1507 BUG_ON(slot_num >= osb->max_slots);
1508
1509 inode = ocfs2_get_system_file_inode(osb, JOURNAL_SYSTEM_INODE,
1510 slot_num);
1511 if (!inode || is_bad_inode(inode)) {
1512 mlog_errno(status);
1513 goto bail;
1514 }
1515 SET_INODE_JOURNAL(inode);
1516
Joel Beckerb657c952008-11-13 14:49:11 -08001517 status = ocfs2_read_inode_block_full(inode, bh, OCFS2_BH_IGNORE_CACHE);
Sunil Mushran539d8262008-07-14 17:31:10 -07001518 if (status < 0) {
1519 mlog_errno(status);
1520 goto bail;
1521 }
1522
1523 status = 0;
1524
1525bail:
1526 if (inode) {
1527 if (status || !ret_inode)
1528 iput(inode);
1529 else
1530 *ret_inode = inode;
1531 }
1532 return status;
1533}
1534
Mark Fashehccd979b2005-12-15 14:31:24 -08001535/* Does the actual journal replay and marks the journal inode as
1536 * clean. Will only replay if the journal inode is marked dirty. */
1537static int ocfs2_replay_journal(struct ocfs2_super *osb,
1538 int node_num,
1539 int slot_num)
1540{
1541 int status;
1542 int got_lock = 0;
1543 unsigned int flags;
1544 struct inode *inode = NULL;
1545 struct ocfs2_dinode *fe;
1546 journal_t *journal = NULL;
1547 struct buffer_head *bh = NULL;
Sunil Mushran539d8262008-07-14 17:31:10 -07001548 u32 slot_reco_gen;
Mark Fashehccd979b2005-12-15 14:31:24 -08001549
Sunil Mushran539d8262008-07-14 17:31:10 -07001550 status = ocfs2_read_journal_inode(osb, slot_num, &bh, &inode);
1551 if (status) {
Mark Fashehccd979b2005-12-15 14:31:24 -08001552 mlog_errno(status);
1553 goto done;
1554 }
Sunil Mushran539d8262008-07-14 17:31:10 -07001555
1556 fe = (struct ocfs2_dinode *)bh->b_data;
1557 slot_reco_gen = ocfs2_get_recovery_generation(fe);
1558 brelse(bh);
1559 bh = NULL;
1560
1561 /*
1562 * As the fs recovery is asynchronous, there is a small chance that
1563 * another node mounted (and recovered) the slot before the recovery
1564 * thread could get the lock. To handle that, we dirty read the journal
1565 * inode for that slot to get the recovery generation. If it is
1566 * different than what we expected, the slot has been recovered.
1567 * If not, it needs recovery.
1568 */
1569 if (osb->slot_recovery_generations[slot_num] != slot_reco_gen) {
1570 mlog(0, "Slot %u already recovered (old/new=%u/%u)\n", slot_num,
1571 osb->slot_recovery_generations[slot_num], slot_reco_gen);
1572 osb->slot_recovery_generations[slot_num] = slot_reco_gen;
1573 status = -EBUSY;
Mark Fashehccd979b2005-12-15 14:31:24 -08001574 goto done;
1575 }
Sunil Mushran539d8262008-07-14 17:31:10 -07001576
1577 /* Continue with recovery as the journal has not yet been recovered */
Mark Fashehccd979b2005-12-15 14:31:24 -08001578
Mark Fashehe63aecb62007-10-18 15:30:42 -07001579 status = ocfs2_inode_lock_full(inode, &bh, 1, OCFS2_META_LOCK_RECOVERY);
Mark Fashehccd979b2005-12-15 14:31:24 -08001580 if (status < 0) {
Mark Fashehe63aecb62007-10-18 15:30:42 -07001581 mlog(0, "status returned from ocfs2_inode_lock=%d\n", status);
Mark Fashehccd979b2005-12-15 14:31:24 -08001582 if (status != -ERESTARTSYS)
1583 mlog(ML_ERROR, "Could not lock journal!\n");
1584 goto done;
1585 }
1586 got_lock = 1;
1587
1588 fe = (struct ocfs2_dinode *) bh->b_data;
1589
1590 flags = le32_to_cpu(fe->id1.journal1.ij_flags);
Sunil Mushran539d8262008-07-14 17:31:10 -07001591 slot_reco_gen = ocfs2_get_recovery_generation(fe);
Mark Fashehccd979b2005-12-15 14:31:24 -08001592
1593 if (!(flags & OCFS2_JOURNAL_DIRTY_FL)) {
1594 mlog(0, "No recovery required for node %d\n", node_num);
Sunil Mushran539d8262008-07-14 17:31:10 -07001595 /* Refresh recovery generation for the slot */
1596 osb->slot_recovery_generations[slot_num] = slot_reco_gen;
Mark Fashehccd979b2005-12-15 14:31:24 -08001597 goto done;
1598 }
1599
Srinivas Eeda9140db02009-03-06 14:21:46 -08001600 /* we need to run complete recovery for offline orphan slots */
1601 ocfs2_replay_map_set_state(osb, REPLAY_NEEDED);
1602
Mark Fashehccd979b2005-12-15 14:31:24 -08001603 mlog(ML_NOTICE, "Recovering node %d from slot %d on device (%u,%u)\n",
1604 node_num, slot_num,
1605 MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev));
1606
1607 OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
1608
1609 status = ocfs2_force_read_journal(inode);
1610 if (status < 0) {
1611 mlog_errno(status);
1612 goto done;
1613 }
1614
1615 mlog(0, "calling journal_init_inode\n");
Joel Becker2b4e30f2008-09-03 20:03:41 -07001616 journal = jbd2_journal_init_inode(inode);
Mark Fashehccd979b2005-12-15 14:31:24 -08001617 if (journal == NULL) {
1618 mlog(ML_ERROR, "Linux journal layer error\n");
1619 status = -EIO;
1620 goto done;
1621 }
1622
Joel Becker2b4e30f2008-09-03 20:03:41 -07001623 status = jbd2_journal_load(journal);
Mark Fashehccd979b2005-12-15 14:31:24 -08001624 if (status < 0) {
1625 mlog_errno(status);
1626 if (!igrab(inode))
1627 BUG();
Joel Becker2b4e30f2008-09-03 20:03:41 -07001628 jbd2_journal_destroy(journal);
Mark Fashehccd979b2005-12-15 14:31:24 -08001629 goto done;
1630 }
1631
1632 ocfs2_clear_journal_error(osb->sb, journal, slot_num);
1633
1634 /* wipe the journal */
1635 mlog(0, "flushing the journal.\n");
Joel Becker2b4e30f2008-09-03 20:03:41 -07001636 jbd2_journal_lock_updates(journal);
1637 status = jbd2_journal_flush(journal);
1638 jbd2_journal_unlock_updates(journal);
Mark Fashehccd979b2005-12-15 14:31:24 -08001639 if (status < 0)
1640 mlog_errno(status);
1641
1642 /* This will mark the node clean */
1643 flags = le32_to_cpu(fe->id1.journal1.ij_flags);
1644 flags &= ~OCFS2_JOURNAL_DIRTY_FL;
1645 fe->id1.journal1.ij_flags = cpu_to_le32(flags);
1646
Sunil Mushran539d8262008-07-14 17:31:10 -07001647 /* Increment recovery generation to indicate successful recovery */
1648 ocfs2_bump_recovery_generation(fe);
1649 osb->slot_recovery_generations[slot_num] =
1650 ocfs2_get_recovery_generation(fe);
1651
Joel Becker13723d02008-10-17 19:25:01 -07001652 ocfs2_compute_meta_ecc(osb->sb, bh->b_data, &fe->i_check);
Joel Becker8cb471e2009-02-10 20:00:41 -08001653 status = ocfs2_write_block(osb, bh, INODE_CACHE(inode));
Mark Fashehccd979b2005-12-15 14:31:24 -08001654 if (status < 0)
1655 mlog_errno(status);
1656
1657 if (!igrab(inode))
1658 BUG();
1659
Joel Becker2b4e30f2008-09-03 20:03:41 -07001660 jbd2_journal_destroy(journal);
Mark Fashehccd979b2005-12-15 14:31:24 -08001661
1662done:
1663 /* drop the lock on this nodes journal */
1664 if (got_lock)
Mark Fashehe63aecb62007-10-18 15:30:42 -07001665 ocfs2_inode_unlock(inode, 1);
Mark Fashehccd979b2005-12-15 14:31:24 -08001666
1667 if (inode)
1668 iput(inode);
1669
Mark Fasheha81cb882008-10-07 14:25:16 -07001670 brelse(bh);
Mark Fashehccd979b2005-12-15 14:31:24 -08001671
1672 mlog_exit(status);
1673 return status;
1674}
1675
1676/*
1677 * Do the most important parts of node recovery:
1678 * - Replay it's journal
1679 * - Stamp a clean local allocator file
1680 * - Stamp a clean truncate log
1681 * - Mark the node clean
1682 *
1683 * If this function completes without error, a node in OCFS2 can be
1684 * said to have been safely recovered. As a result, failure during the
1685 * second part of a nodes recovery process (local alloc recovery) is
1686 * far less concerning.
1687 */
1688static int ocfs2_recover_node(struct ocfs2_super *osb,
Jan Kara22053632008-10-20 23:50:38 +02001689 int node_num, int slot_num)
Mark Fashehccd979b2005-12-15 14:31:24 -08001690{
1691 int status = 0;
Mark Fashehccd979b2005-12-15 14:31:24 -08001692 struct ocfs2_dinode *la_copy = NULL;
1693 struct ocfs2_dinode *tl_copy = NULL;
1694
Jan Kara22053632008-10-20 23:50:38 +02001695 mlog_entry("(node_num=%d, slot_num=%d, osb->node_num = %d)\n",
1696 node_num, slot_num, osb->node_num);
Mark Fashehccd979b2005-12-15 14:31:24 -08001697
1698 /* Should not ever be called to recover ourselves -- in that
1699 * case we should've called ocfs2_journal_load instead. */
Eric Sesterhenn / snakebyteebdec832006-01-27 10:32:52 +01001700 BUG_ON(osb->node_num == node_num);
Mark Fashehccd979b2005-12-15 14:31:24 -08001701
Mark Fashehccd979b2005-12-15 14:31:24 -08001702 status = ocfs2_replay_journal(osb, node_num, slot_num);
1703 if (status < 0) {
Sunil Mushran539d8262008-07-14 17:31:10 -07001704 if (status == -EBUSY) {
1705 mlog(0, "Skipping recovery for slot %u (node %u) "
1706 "as another node has recovered it\n", slot_num,
1707 node_num);
1708 status = 0;
1709 goto done;
1710 }
Mark Fashehccd979b2005-12-15 14:31:24 -08001711 mlog_errno(status);
1712 goto done;
1713 }
1714
1715 /* Stamp a clean local alloc file AFTER recovering the journal... */
1716 status = ocfs2_begin_local_alloc_recovery(osb, slot_num, &la_copy);
1717 if (status < 0) {
1718 mlog_errno(status);
1719 goto done;
1720 }
1721
1722 /* An error from begin_truncate_log_recovery is not
1723 * serious enough to warrant halting the rest of
1724 * recovery. */
1725 status = ocfs2_begin_truncate_log_recovery(osb, slot_num, &tl_copy);
1726 if (status < 0)
1727 mlog_errno(status);
1728
1729 /* Likewise, this would be a strange but ultimately not so
1730 * harmful place to get an error... */
Mark Fasheh8e8a4602008-02-01 11:59:09 -08001731 status = ocfs2_clear_slot(osb, slot_num);
Mark Fashehccd979b2005-12-15 14:31:24 -08001732 if (status < 0)
1733 mlog_errno(status);
1734
1735 /* This will kfree the memory pointed to by la_copy and tl_copy */
1736 ocfs2_queue_recovery_completion(osb->journal, slot_num, la_copy,
Jan Kara22053632008-10-20 23:50:38 +02001737 tl_copy, NULL);
Mark Fashehccd979b2005-12-15 14:31:24 -08001738
1739 status = 0;
1740done:
1741
1742 mlog_exit(status);
1743 return status;
1744}
1745
1746/* Test node liveness by trylocking his journal. If we get the lock,
1747 * we drop it here. Return 0 if we got the lock, -EAGAIN if node is
1748 * still alive (we couldn't get the lock) and < 0 on error. */
1749static int ocfs2_trylock_journal(struct ocfs2_super *osb,
1750 int slot_num)
1751{
1752 int status, flags;
1753 struct inode *inode = NULL;
1754
1755 inode = ocfs2_get_system_file_inode(osb, JOURNAL_SYSTEM_INODE,
1756 slot_num);
1757 if (inode == NULL) {
1758 mlog(ML_ERROR, "access error\n");
1759 status = -EACCES;
1760 goto bail;
1761 }
1762 if (is_bad_inode(inode)) {
1763 mlog(ML_ERROR, "access error (bad inode)\n");
1764 iput(inode);
1765 inode = NULL;
1766 status = -EACCES;
1767 goto bail;
1768 }
1769 SET_INODE_JOURNAL(inode);
1770
1771 flags = OCFS2_META_LOCK_RECOVERY | OCFS2_META_LOCK_NOQUEUE;
Mark Fashehe63aecb62007-10-18 15:30:42 -07001772 status = ocfs2_inode_lock_full(inode, NULL, 1, flags);
Mark Fashehccd979b2005-12-15 14:31:24 -08001773 if (status < 0) {
1774 if (status != -EAGAIN)
1775 mlog_errno(status);
1776 goto bail;
1777 }
1778
Mark Fashehe63aecb62007-10-18 15:30:42 -07001779 ocfs2_inode_unlock(inode, 1);
Mark Fashehccd979b2005-12-15 14:31:24 -08001780bail:
1781 if (inode)
1782 iput(inode);
1783
1784 return status;
1785}
1786
1787/* Call this underneath ocfs2_super_lock. It also assumes that the
1788 * slot info struct has been updated from disk. */
1789int ocfs2_mark_dead_nodes(struct ocfs2_super *osb)
1790{
Joel Beckerd85b20e2008-02-01 12:01:05 -08001791 unsigned int node_num;
1792 int status, i;
Mark Fasheha1af7d12008-08-19 17:20:28 -07001793 u32 gen;
Sunil Mushran539d8262008-07-14 17:31:10 -07001794 struct buffer_head *bh = NULL;
1795 struct ocfs2_dinode *di;
Mark Fashehccd979b2005-12-15 14:31:24 -08001796
1797 /* This is called with the super block cluster lock, so we
1798 * know that the slot map can't change underneath us. */
1799
Joel Beckerd85b20e2008-02-01 12:01:05 -08001800 for (i = 0; i < osb->max_slots; i++) {
Sunil Mushran539d8262008-07-14 17:31:10 -07001801 /* Read journal inode to get the recovery generation */
1802 status = ocfs2_read_journal_inode(osb, i, &bh, NULL);
1803 if (status) {
1804 mlog_errno(status);
1805 goto bail;
1806 }
1807 di = (struct ocfs2_dinode *)bh->b_data;
Mark Fasheha1af7d12008-08-19 17:20:28 -07001808 gen = ocfs2_get_recovery_generation(di);
Sunil Mushran539d8262008-07-14 17:31:10 -07001809 brelse(bh);
1810 bh = NULL;
1811
Mark Fasheha1af7d12008-08-19 17:20:28 -07001812 spin_lock(&osb->osb_lock);
1813 osb->slot_recovery_generations[i] = gen;
1814
Sunil Mushran539d8262008-07-14 17:31:10 -07001815 mlog(0, "Slot %u recovery generation is %u\n", i,
1816 osb->slot_recovery_generations[i]);
1817
Mark Fasheha1af7d12008-08-19 17:20:28 -07001818 if (i == osb->slot_num) {
1819 spin_unlock(&osb->osb_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -08001820 continue;
Mark Fasheha1af7d12008-08-19 17:20:28 -07001821 }
Joel Beckerd85b20e2008-02-01 12:01:05 -08001822
1823 status = ocfs2_slot_to_node_num_locked(osb, i, &node_num);
Mark Fasheha1af7d12008-08-19 17:20:28 -07001824 if (status == -ENOENT) {
1825 spin_unlock(&osb->osb_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -08001826 continue;
Mark Fasheha1af7d12008-08-19 17:20:28 -07001827 }
Mark Fashehccd979b2005-12-15 14:31:24 -08001828
Mark Fasheha1af7d12008-08-19 17:20:28 -07001829 if (__ocfs2_recovery_map_test(osb, node_num)) {
1830 spin_unlock(&osb->osb_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -08001831 continue;
Mark Fasheha1af7d12008-08-19 17:20:28 -07001832 }
Joel Beckerd85b20e2008-02-01 12:01:05 -08001833 spin_unlock(&osb->osb_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -08001834
1835 /* Ok, we have a slot occupied by another node which
1836 * is not in the recovery map. We trylock his journal
1837 * file here to test if he's alive. */
1838 status = ocfs2_trylock_journal(osb, i);
1839 if (!status) {
1840 /* Since we're called from mount, we know that
1841 * the recovery thread can't race us on
1842 * setting / checking the recovery bits. */
1843 ocfs2_recovery_thread(osb, node_num);
1844 } else if ((status < 0) && (status != -EAGAIN)) {
1845 mlog_errno(status);
1846 goto bail;
1847 }
Mark Fashehccd979b2005-12-15 14:31:24 -08001848 }
Mark Fashehccd979b2005-12-15 14:31:24 -08001849
1850 status = 0;
1851bail:
1852 mlog_exit(status);
1853 return status;
1854}
1855
Srinivas Eeda83273932009-06-03 17:02:55 -07001856/*
1857 * Scan timer should get fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT. Add some
1858 * randomness to the timeout to minimize multple nodes firing the timer at the
1859 * same time.
1860 */
1861static inline unsigned long ocfs2_orphan_scan_timeout(void)
1862{
1863 unsigned long time;
1864
1865 get_random_bytes(&time, sizeof(time));
1866 time = ORPHAN_SCAN_SCHEDULE_TIMEOUT + (time % 5000);
1867 return msecs_to_jiffies(time);
1868}
1869
1870/*
1871 * ocfs2_queue_orphan_scan calls ocfs2_queue_recovery_completion for
1872 * every slot, queuing a recovery of the slot on the ocfs2_wq thread. This
1873 * is done to catch any orphans that are left over in orphan directories.
1874 *
1875 * ocfs2_queue_orphan_scan gets called every ORPHAN_SCAN_SCHEDULE_TIMEOUT
1876 * seconds. It gets an EX lock on os_lockres and checks sequence number
1877 * stored in LVB. If the sequence number has changed, it means some other
1878 * node has done the scan. This node skips the scan and tracks the
1879 * sequence number. If the sequence number didn't change, it means a scan
1880 * hasn't happened. The node queues a scan and increments the
1881 * sequence number in the LVB.
1882 */
1883void ocfs2_queue_orphan_scan(struct ocfs2_super *osb)
1884{
1885 struct ocfs2_orphan_scan *os;
1886 int status, i;
1887 u32 seqno = 0;
1888
1889 os = &osb->osb_orphan_scan;
1890
Sunil Mushran692684e2009-06-19 16:53:17 -07001891 if (atomic_read(&os->os_state) == ORPHAN_SCAN_INACTIVE)
1892 goto out;
1893
Sunil Mushrandf152c22009-06-22 11:40:07 -07001894 status = ocfs2_orphan_scan_lock(osb, &seqno);
Srinivas Eeda83273932009-06-03 17:02:55 -07001895 if (status < 0) {
1896 if (status != -EAGAIN)
1897 mlog_errno(status);
1898 goto out;
1899 }
1900
Sunil Mushran692684e2009-06-19 16:53:17 -07001901 /* Do no queue the tasks if the volume is being umounted */
1902 if (atomic_read(&os->os_state) == ORPHAN_SCAN_INACTIVE)
1903 goto unlock;
1904
Srinivas Eeda83273932009-06-03 17:02:55 -07001905 if (os->os_seqno != seqno) {
1906 os->os_seqno = seqno;
1907 goto unlock;
1908 }
1909
1910 for (i = 0; i < osb->max_slots; i++)
1911 ocfs2_queue_recovery_completion(osb->journal, i, NULL, NULL,
1912 NULL);
1913 /*
1914 * We queued a recovery on orphan slots, increment the sequence
1915 * number and update LVB so other node will skip the scan for a while
1916 */
1917 seqno++;
Srinivas Eeda15633a22009-06-03 17:02:56 -07001918 os->os_count++;
1919 os->os_scantime = CURRENT_TIME;
Srinivas Eeda83273932009-06-03 17:02:55 -07001920unlock:
Sunil Mushrandf152c22009-06-22 11:40:07 -07001921 ocfs2_orphan_scan_unlock(osb, seqno);
Srinivas Eeda83273932009-06-03 17:02:55 -07001922out:
1923 return;
1924}
1925
1926/* Worker task that gets fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT millsec */
1927void ocfs2_orphan_scan_work(struct work_struct *work)
1928{
1929 struct ocfs2_orphan_scan *os;
1930 struct ocfs2_super *osb;
1931
1932 os = container_of(work, struct ocfs2_orphan_scan,
1933 os_orphan_scan_work.work);
1934 osb = os->os_osb;
1935
1936 mutex_lock(&os->os_lock);
1937 ocfs2_queue_orphan_scan(osb);
Sunil Mushran692684e2009-06-19 16:53:17 -07001938 if (atomic_read(&os->os_state) == ORPHAN_SCAN_ACTIVE)
1939 schedule_delayed_work(&os->os_orphan_scan_work,
1940 ocfs2_orphan_scan_timeout());
Srinivas Eeda83273932009-06-03 17:02:55 -07001941 mutex_unlock(&os->os_lock);
1942}
1943
1944void ocfs2_orphan_scan_stop(struct ocfs2_super *osb)
1945{
1946 struct ocfs2_orphan_scan *os;
1947
1948 os = &osb->osb_orphan_scan;
Sunil Mushrandf152c22009-06-22 11:40:07 -07001949 if (atomic_read(&os->os_state) == ORPHAN_SCAN_ACTIVE) {
1950 atomic_set(&os->os_state, ORPHAN_SCAN_INACTIVE);
1951 mutex_lock(&os->os_lock);
1952 cancel_delayed_work(&os->os_orphan_scan_work);
1953 mutex_unlock(&os->os_lock);
1954 }
Srinivas Eeda83273932009-06-03 17:02:55 -07001955}
1956
Sunil Mushrandf152c22009-06-22 11:40:07 -07001957void ocfs2_orphan_scan_init(struct ocfs2_super *osb)
Srinivas Eeda83273932009-06-03 17:02:55 -07001958{
1959 struct ocfs2_orphan_scan *os;
1960
1961 os = &osb->osb_orphan_scan;
1962 os->os_osb = osb;
Srinivas Eeda15633a22009-06-03 17:02:56 -07001963 os->os_count = 0;
Sunil Mushran32119492009-06-19 16:53:18 -07001964 os->os_seqno = 0;
Srinivas Eeda83273932009-06-03 17:02:55 -07001965 mutex_init(&os->os_lock);
Sunil Mushrandf152c22009-06-22 11:40:07 -07001966 INIT_DELAYED_WORK(&os->os_orphan_scan_work, ocfs2_orphan_scan_work);
Jeff Mahoney8b712cd2009-07-07 17:22:12 -04001967}
Srinivas Eeda83273932009-06-03 17:02:55 -07001968
Jeff Mahoney8b712cd2009-07-07 17:22:12 -04001969void ocfs2_orphan_scan_start(struct ocfs2_super *osb)
1970{
1971 struct ocfs2_orphan_scan *os;
1972
1973 os = &osb->osb_orphan_scan;
1974 os->os_scantime = CURRENT_TIME;
Sunil Mushrandf152c22009-06-22 11:40:07 -07001975 if (ocfs2_is_hard_readonly(osb) || ocfs2_mount_local(osb))
1976 atomic_set(&os->os_state, ORPHAN_SCAN_INACTIVE);
1977 else {
1978 atomic_set(&os->os_state, ORPHAN_SCAN_ACTIVE);
1979 schedule_delayed_work(&os->os_orphan_scan_work,
1980 ocfs2_orphan_scan_timeout());
1981 }
Srinivas Eeda83273932009-06-03 17:02:55 -07001982}
1983
Mark Fasheh5eae5b92007-09-10 17:50:51 -07001984struct ocfs2_orphan_filldir_priv {
1985 struct inode *head;
1986 struct ocfs2_super *osb;
1987};
1988
1989static int ocfs2_orphan_filldir(void *priv, const char *name, int name_len,
1990 loff_t pos, u64 ino, unsigned type)
1991{
1992 struct ocfs2_orphan_filldir_priv *p = priv;
1993 struct inode *iter;
1994
1995 if (name_len == 1 && !strncmp(".", name, 1))
1996 return 0;
1997 if (name_len == 2 && !strncmp("..", name, 2))
1998 return 0;
1999
2000 /* Skip bad inodes so that recovery can continue */
2001 iter = ocfs2_iget(p->osb, ino,
Jan Kara5fa06132008-01-11 00:11:45 +01002002 OCFS2_FI_FLAG_ORPHAN_RECOVERY, 0);
Mark Fasheh5eae5b92007-09-10 17:50:51 -07002003 if (IS_ERR(iter))
2004 return 0;
2005
2006 mlog(0, "queue orphan %llu\n",
2007 (unsigned long long)OCFS2_I(iter)->ip_blkno);
2008 /* No locking is required for the next_orphan queue as there
2009 * is only ever a single process doing orphan recovery. */
2010 OCFS2_I(iter)->ip_next_orphan = p->head;
2011 p->head = iter;
2012
2013 return 0;
2014}
2015
Mark Fashehb4df6ed2006-02-22 17:35:08 -08002016static int ocfs2_queue_orphans(struct ocfs2_super *osb,
2017 int slot,
2018 struct inode **head)
Mark Fashehccd979b2005-12-15 14:31:24 -08002019{
Mark Fashehb4df6ed2006-02-22 17:35:08 -08002020 int status;
Mark Fashehccd979b2005-12-15 14:31:24 -08002021 struct inode *orphan_dir_inode = NULL;
Mark Fasheh5eae5b92007-09-10 17:50:51 -07002022 struct ocfs2_orphan_filldir_priv priv;
2023 loff_t pos = 0;
2024
2025 priv.osb = osb;
2026 priv.head = *head;
Mark Fashehccd979b2005-12-15 14:31:24 -08002027
2028 orphan_dir_inode = ocfs2_get_system_file_inode(osb,
2029 ORPHAN_DIR_SYSTEM_INODE,
2030 slot);
2031 if (!orphan_dir_inode) {
2032 status = -ENOENT;
2033 mlog_errno(status);
Mark Fashehb4df6ed2006-02-22 17:35:08 -08002034 return status;
Sunil Mushran2bd63212010-01-25 16:57:38 -08002035 }
Mark Fashehccd979b2005-12-15 14:31:24 -08002036
Jes Sorensen1b1dcc12006-01-09 15:59:24 -08002037 mutex_lock(&orphan_dir_inode->i_mutex);
Mark Fashehe63aecb62007-10-18 15:30:42 -07002038 status = ocfs2_inode_lock(orphan_dir_inode, NULL, 0);
Mark Fashehccd979b2005-12-15 14:31:24 -08002039 if (status < 0) {
Mark Fashehccd979b2005-12-15 14:31:24 -08002040 mlog_errno(status);
2041 goto out;
2042 }
Mark Fashehccd979b2005-12-15 14:31:24 -08002043
Mark Fasheh5eae5b92007-09-10 17:50:51 -07002044 status = ocfs2_dir_foreach(orphan_dir_inode, &pos, &priv,
2045 ocfs2_orphan_filldir);
2046 if (status) {
2047 mlog_errno(status);
Mark Fasheha86370f2007-12-03 14:06:23 -08002048 goto out_cluster;
Mark Fashehccd979b2005-12-15 14:31:24 -08002049 }
Mark Fashehccd979b2005-12-15 14:31:24 -08002050
Mark Fasheh5eae5b92007-09-10 17:50:51 -07002051 *head = priv.head;
2052
Mark Fasheha86370f2007-12-03 14:06:23 -08002053out_cluster:
Mark Fashehe63aecb62007-10-18 15:30:42 -07002054 ocfs2_inode_unlock(orphan_dir_inode, 0);
Mark Fashehb4df6ed2006-02-22 17:35:08 -08002055out:
2056 mutex_unlock(&orphan_dir_inode->i_mutex);
Mark Fashehccd979b2005-12-15 14:31:24 -08002057 iput(orphan_dir_inode);
Mark Fashehb4df6ed2006-02-22 17:35:08 -08002058 return status;
2059}
2060
2061static int ocfs2_orphan_recovery_can_continue(struct ocfs2_super *osb,
2062 int slot)
2063{
2064 int ret;
2065
2066 spin_lock(&osb->osb_lock);
2067 ret = !osb->osb_orphan_wipes[slot];
2068 spin_unlock(&osb->osb_lock);
2069 return ret;
2070}
2071
2072static void ocfs2_mark_recovering_orphan_dir(struct ocfs2_super *osb,
2073 int slot)
2074{
2075 spin_lock(&osb->osb_lock);
2076 /* Mark ourselves such that new processes in delete_inode()
2077 * know to quit early. */
2078 ocfs2_node_map_set_bit(osb, &osb->osb_recovering_orphan_dirs, slot);
2079 while (osb->osb_orphan_wipes[slot]) {
2080 /* If any processes are already in the middle of an
2081 * orphan wipe on this dir, then we need to wait for
2082 * them. */
2083 spin_unlock(&osb->osb_lock);
2084 wait_event_interruptible(osb->osb_wipe_event,
2085 ocfs2_orphan_recovery_can_continue(osb, slot));
2086 spin_lock(&osb->osb_lock);
2087 }
2088 spin_unlock(&osb->osb_lock);
2089}
2090
2091static void ocfs2_clear_recovering_orphan_dir(struct ocfs2_super *osb,
2092 int slot)
2093{
2094 ocfs2_node_map_clear_bit(osb, &osb->osb_recovering_orphan_dirs, slot);
2095}
2096
2097/*
2098 * Orphan recovery. Each mounted node has it's own orphan dir which we
2099 * must run during recovery. Our strategy here is to build a list of
2100 * the inodes in the orphan dir and iget/iput them. The VFS does
2101 * (most) of the rest of the work.
2102 *
2103 * Orphan recovery can happen at any time, not just mount so we have a
2104 * couple of extra considerations.
2105 *
2106 * - We grab as many inodes as we can under the orphan dir lock -
2107 * doing iget() outside the orphan dir risks getting a reference on
2108 * an invalid inode.
2109 * - We must be sure not to deadlock with other processes on the
2110 * system wanting to run delete_inode(). This can happen when they go
2111 * to lock the orphan dir and the orphan recovery process attempts to
2112 * iget() inside the orphan dir lock. This can be avoided by
2113 * advertising our state to ocfs2_delete_inode().
2114 */
2115static int ocfs2_recover_orphans(struct ocfs2_super *osb,
2116 int slot)
2117{
2118 int ret = 0;
2119 struct inode *inode = NULL;
2120 struct inode *iter;
2121 struct ocfs2_inode_info *oi;
2122
2123 mlog(0, "Recover inodes from orphan dir in slot %d\n", slot);
2124
2125 ocfs2_mark_recovering_orphan_dir(osb, slot);
2126 ret = ocfs2_queue_orphans(osb, slot, &inode);
2127 ocfs2_clear_recovering_orphan_dir(osb, slot);
2128
2129 /* Error here should be noted, but we want to continue with as
2130 * many queued inodes as we've got. */
2131 if (ret)
2132 mlog_errno(ret);
Mark Fashehccd979b2005-12-15 14:31:24 -08002133
2134 while (inode) {
2135 oi = OCFS2_I(inode);
Mark Fashehb0697052006-03-03 10:24:33 -08002136 mlog(0, "iput orphan %llu\n", (unsigned long long)oi->ip_blkno);
Mark Fashehccd979b2005-12-15 14:31:24 -08002137
2138 iter = oi->ip_next_orphan;
2139
2140 spin_lock(&oi->ip_lock);
Mark Fasheh34d024f2007-09-24 15:56:19 -07002141 /* The remote delete code may have set these on the
2142 * assumption that the other node would wipe them
2143 * successfully. If they are still in the node's
2144 * orphan dir, we need to reset that state. */
Mark Fashehccd979b2005-12-15 14:31:24 -08002145 oi->ip_flags &= ~(OCFS2_INODE_DELETED|OCFS2_INODE_SKIP_DELETE);
2146
2147 /* Set the proper information to get us going into
2148 * ocfs2_delete_inode. */
2149 oi->ip_flags |= OCFS2_INODE_MAYBE_ORPHANED;
Mark Fashehccd979b2005-12-15 14:31:24 -08002150 spin_unlock(&oi->ip_lock);
2151
2152 iput(inode);
2153
2154 inode = iter;
2155 }
2156
Mark Fashehb4df6ed2006-02-22 17:35:08 -08002157 return ret;
Mark Fashehccd979b2005-12-15 14:31:24 -08002158}
2159
Jan Kara19ece542008-08-21 20:13:17 +02002160static int __ocfs2_wait_on_mount(struct ocfs2_super *osb, int quota)
Mark Fashehccd979b2005-12-15 14:31:24 -08002161{
2162 /* This check is good because ocfs2 will wait on our recovery
2163 * thread before changing it to something other than MOUNTED
2164 * or DISABLED. */
2165 wait_event(osb->osb_mount_event,
Jan Kara19ece542008-08-21 20:13:17 +02002166 (!quota && atomic_read(&osb->vol_state) == VOLUME_MOUNTED) ||
2167 atomic_read(&osb->vol_state) == VOLUME_MOUNTED_QUOTAS ||
Mark Fashehccd979b2005-12-15 14:31:24 -08002168 atomic_read(&osb->vol_state) == VOLUME_DISABLED);
2169
2170 /* If there's an error on mount, then we may never get to the
2171 * MOUNTED flag, but this is set right before
2172 * dismount_volume() so we can trust it. */
2173 if (atomic_read(&osb->vol_state) == VOLUME_DISABLED) {
2174 mlog(0, "mount error, exiting!\n");
2175 return -EBUSY;
2176 }
2177
2178 return 0;
2179}
2180
2181static int ocfs2_commit_thread(void *arg)
2182{
2183 int status;
2184 struct ocfs2_super *osb = arg;
2185 struct ocfs2_journal *journal = osb->journal;
2186
2187 /* we can trust j_num_trans here because _should_stop() is only set in
2188 * shutdown and nobody other than ourselves should be able to start
2189 * transactions. committing on shutdown might take a few iterations
2190 * as final transactions put deleted inodes on the list */
2191 while (!(kthread_should_stop() &&
2192 atomic_read(&journal->j_num_trans) == 0)) {
2193
Mark Fasheh745ae8ba2006-02-09 13:23:39 -08002194 wait_event_interruptible(osb->checkpoint_event,
2195 atomic_read(&journal->j_num_trans)
2196 || kthread_should_stop());
Mark Fashehccd979b2005-12-15 14:31:24 -08002197
2198 status = ocfs2_commit_cache(osb);
2199 if (status < 0)
2200 mlog_errno(status);
2201
2202 if (kthread_should_stop() && atomic_read(&journal->j_num_trans)){
2203 mlog(ML_KTHREAD,
2204 "commit_thread: %u transactions pending on "
2205 "shutdown\n",
2206 atomic_read(&journal->j_num_trans));
2207 }
2208 }
2209
2210 return 0;
2211}
2212
Sunil Mushran539d8262008-07-14 17:31:10 -07002213/* Reads all the journal inodes without taking any cluster locks. Used
2214 * for hard readonly access to determine whether any journal requires
2215 * recovery. Also used to refresh the recovery generation numbers after
2216 * a journal has been recovered by another node.
2217 */
Mark Fashehccd979b2005-12-15 14:31:24 -08002218int ocfs2_check_journals_nolocks(struct ocfs2_super *osb)
2219{
2220 int ret = 0;
2221 unsigned int slot;
Sunil Mushran539d8262008-07-14 17:31:10 -07002222 struct buffer_head *di_bh = NULL;
Mark Fashehccd979b2005-12-15 14:31:24 -08002223 struct ocfs2_dinode *di;
Sunil Mushran539d8262008-07-14 17:31:10 -07002224 int journal_dirty = 0;
Mark Fashehccd979b2005-12-15 14:31:24 -08002225
2226 for(slot = 0; slot < osb->max_slots; slot++) {
Sunil Mushran539d8262008-07-14 17:31:10 -07002227 ret = ocfs2_read_journal_inode(osb, slot, &di_bh, NULL);
2228 if (ret) {
Mark Fashehccd979b2005-12-15 14:31:24 -08002229 mlog_errno(ret);
2230 goto out;
2231 }
2232
2233 di = (struct ocfs2_dinode *) di_bh->b_data;
2234
Sunil Mushran539d8262008-07-14 17:31:10 -07002235 osb->slot_recovery_generations[slot] =
2236 ocfs2_get_recovery_generation(di);
2237
Mark Fashehccd979b2005-12-15 14:31:24 -08002238 if (le32_to_cpu(di->id1.journal1.ij_flags) &
2239 OCFS2_JOURNAL_DIRTY_FL)
Sunil Mushran539d8262008-07-14 17:31:10 -07002240 journal_dirty = 1;
Mark Fashehccd979b2005-12-15 14:31:24 -08002241
2242 brelse(di_bh);
Sunil Mushran539d8262008-07-14 17:31:10 -07002243 di_bh = NULL;
Mark Fashehccd979b2005-12-15 14:31:24 -08002244 }
2245
2246out:
Sunil Mushran539d8262008-07-14 17:31:10 -07002247 if (journal_dirty)
2248 ret = -EROFS;
Mark Fashehccd979b2005-12-15 14:31:24 -08002249 return ret;
2250}