projects / GitHub / mt8127 / android_kernel_alcatel_ttab.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
libfc: use ARRAY_SIZE
[GitHub/mt8127/android_kernel_alcatel_ttab.git] / fs / ocfs2 / dlmglue.c
CommitLineData
ccd979bd
MF		 1/* -*- mode: c; c-basic-offset: 8; -*-
2 * vim: noexpandtab sw=8 ts=8 sts=0:
3 *
4 * dlmglue.c
5 *
6 * Code which implements an OCFS2 specific interface to our DLM.
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/types.h>
27#include <linux/slab.h>
28#include <linux/highmem.h>
29#include <linux/mm.h>
ccd979bd
MF		 30#include <linux/kthread.h>
31#include <linux/pagemap.h>
32#include <linux/debugfs.h>
33#include <linux/seq_file.h>
8ddb7b00 34#include <linux/time.h>
9e33d69f 35#include <linux/quotaops.h>
ccd979bd 36
ccd979bd
MF		 37#define MLOG_MASK_PREFIX ML_DLM_GLUE
38#include <cluster/masklog.h>
39
40#include "ocfs2.h"
d24fbcda 41#include "ocfs2_lockingver.h"
ccd979bd
MF		 42
43#include "alloc.h"
d680efe9 44#include "dcache.h"
ccd979bd
MF		 45#include "dlmglue.h"
46#include "extent_map.h"
7f1a37e3 47#include "file.h"
ccd979bd
MF		 48#include "heartbeat.h"
49#include "inode.h"
50#include "journal.h"
24ef1815 51#include "stackglue.h"
ccd979bd
MF		 52#include "slot_map.h"
53#include "super.h"
54#include "uptodate.h"
9e33d69f 55#include "quota.h"
8dec98ed 56#include "refcounttree.h"
ccd979bd
MF		 57
58#include "buffer_head_io.h"
59
60struct ocfs2_mask_waiter {
61 struct list_head mw_item;
62 int mw_status;
63 struct completion mw_complete;
64 unsigned long mw_mask;
65 unsigned long mw_goal;
8ddb7b00
SM		 66#ifdef CONFIG_OCFS2_FS_STATS
67 unsigned long long mw_lock_start;
68#endif
ccd979bd
MF		 69};
70
54a7e755
MF		 71static struct ocfs2_super *ocfs2_get_dentry_osb(struct ocfs2_lock_res *lockres);
72static struct ocfs2_super *ocfs2_get_inode_osb(struct ocfs2_lock_res *lockres);
cf8e06f1 73static struct ocfs2_super *ocfs2_get_file_osb(struct ocfs2_lock_res *lockres);
9e33d69f 74static struct ocfs2_super *ocfs2_get_qinfo_osb(struct ocfs2_lock_res *lockres);
ccd979bd 75
d680efe9 76/*
cc567d89 77 * Return value from ->downconvert_worker functions.
d680efe9 78 *
b5e500e2 79 * These control the precise actions of ocfs2_unblock_lock()
d680efe9
MF		 80 * and ocfs2_process_blocked_lock()
81 *
82 */
83enum ocfs2_unblock_action {
84 UNBLOCK_CONTINUE = 0, /* Continue downconvert */
85 UNBLOCK_CONTINUE_POST = 1, /* Continue downconvert, fire
86 * ->post_unlock callback */
87 UNBLOCK_STOP_POST = 2, /* Do not downconvert, fire
88 * ->post_unlock() callback. */
89};
90
91struct ocfs2_unblock_ctl {
92 int requeue;
93 enum ocfs2_unblock_action unblock_action;
94};
95
cb25797d
JK		 96/* Lockdep class keys */
97struct lock_class_key lockdep_keys[OCFS2_NUM_LOCK_TYPES];
98
810d5aeb
MF		 99static int ocfs2_check_meta_downconvert(struct ocfs2_lock_res *lockres,
100 int new_level);
101static void ocfs2_set_meta_lvb(struct ocfs2_lock_res *lockres);
102
cc567d89
MF		 103static int ocfs2_data_convert_worker(struct ocfs2_lock_res *lockres,
104 int blocking);
105
cc567d89
MF		 106static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res *lockres,
107 int blocking);
d680efe9
MF		 108
109static void ocfs2_dentry_post_unlock(struct ocfs2_super *osb,
110 struct ocfs2_lock_res *lockres);
ccd979bd 111
9e33d69f 112static void ocfs2_set_qinfo_lvb(struct ocfs2_lock_res *lockres);
6cb129f5 113
8dec98ed
TM		 114static int ocfs2_check_refcount_downconvert(struct ocfs2_lock_res *lockres,
115 int new_level);
116static int ocfs2_refcount_convert_worker(struct ocfs2_lock_res *lockres,
117 int blocking);
118
6cb129f5
AB		 119#define mlog_meta_lvb(__level, __lockres) ocfs2_dump_meta_lvb_info(__level, __PRETTY_FUNCTION__, __LINE__, __lockres)
120
121/* This aids in debugging situations where a bad LVB might be involved. */
122static void ocfs2_dump_meta_lvb_info(u64 level,
123 const char *function,
124 unsigned int line,
125 struct ocfs2_lock_res *lockres)
126{
a641dc2a 127 struct ocfs2_meta_lvb *lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
6cb129f5
AB		 128
129 mlog(level, "LVB information for %s (called from %s:%u):\n",
130 lockres->l_name, function, line);
131 mlog(level, "version: %u, clusters: %u, generation: 0x%x\n",
132 lvb->lvb_version, be32_to_cpu(lvb->lvb_iclusters),
133 be32_to_cpu(lvb->lvb_igeneration));
134 mlog(level, "size: %llu, uid %u, gid %u, mode 0x%x\n",
135 (unsigned long long)be64_to_cpu(lvb->lvb_isize),
136 be32_to_cpu(lvb->lvb_iuid), be32_to_cpu(lvb->lvb_igid),
137 be16_to_cpu(lvb->lvb_imode));
138 mlog(level, "nlink %u, atime_packed 0x%llx, ctime_packed 0x%llx, "
139 "mtime_packed 0x%llx iattr 0x%x\n", be16_to_cpu(lvb->lvb_inlink),
140 (long long)be64_to_cpu(lvb->lvb_iatime_packed),
141 (long long)be64_to_cpu(lvb->lvb_ictime_packed),
142 (long long)be64_to_cpu(lvb->lvb_imtime_packed),
143 be32_to_cpu(lvb->lvb_iattr));
144}
145
146
f625c979
MF		 147/*
148 * OCFS2 Lock Resource Operations
149 *
150 * These fine tune the behavior of the generic dlmglue locking infrastructure.
0d5dc6c2
MF		 151 *
152 * The most basic of lock types can point ->l_priv to their respective
153 * struct ocfs2_super and allow the default actions to manage things.
154 *
155 * Right now, each lock type also needs to implement an init function,
156 * and trivial lock/unlock wrappers. ocfs2_simple_drop_lockres()
157 * should be called when the lock is no longer needed (i.e., object
158 * destruction time).
f625c979 159 */
ccd979bd 160struct ocfs2_lock_res_ops {
54a7e755
MF		 161 /*
162 * Translate an ocfs2_lock_res * into an ocfs2_super *. Define
163 * this callback if ->l_priv is not an ocfs2_super pointer
164 */
165 struct ocfs2_super * (*get_osb)(struct ocfs2_lock_res *);
b5e500e2 166
0d5dc6c2 167 /*
34d024f8
MF		 168 * Optionally called in the downconvert thread after a
169 * successful downconvert. The lockres will not be referenced
170 * after this callback is called, so it is safe to free
171 * memory, etc.
0d5dc6c2
MF		 172 *
173 * The exact semantics of when this is called are controlled
174 * by ->downconvert_worker()
175 */
d680efe9 176 void (*post_unlock)(struct ocfs2_super *, struct ocfs2_lock_res *);
f625c979 177
16d5b956
MF		 178 /*
179 * Allow a lock type to add checks to determine whether it is
180 * safe to downconvert a lock. Return 0 to re-queue the
181 * downconvert at a later time, nonzero to continue.
182 *
183 * For most locks, the default checks that there are no
184 * incompatible holders are sufficient.
185 *
186 * Called with the lockres spinlock held.
187 */
188 int (*check_downconvert)(struct ocfs2_lock_res *, int);
189
5ef0d4ea
MF		 190 /*
191 * Allows a lock type to populate the lock value block. This
192 * is called on downconvert, and when we drop a lock.
193 *
194 * Locks that want to use this should set LOCK_TYPE_USES_LVB
195 * in the flags field.
196 *
197 * Called with the lockres spinlock held.
198 */
199 void (*set_lvb)(struct ocfs2_lock_res *);
200
cc567d89
MF		 201 /*
202 * Called from the downconvert thread when it is determined
203 * that a lock will be downconverted. This is called without
204 * any locks held so the function can do work that might
205 * schedule (syncing out data, etc).
206 *
207 * This should return any one of the ocfs2_unblock_action
208 * values, depending on what it wants the thread to do.
209 */
210 int (*downconvert_worker)(struct ocfs2_lock_res *, int);
211
f625c979
MF		 212 /*
213 * LOCK_TYPE_* flags which describe the specific requirements
214 * of a lock type. Descriptions of each individual flag follow.
215 */
216 int flags;
ccd979bd
MF		 217};
218
f625c979
MF		 219/*
220 * Some locks want to "refresh" potentially stale data when a
221 * meaningful (PRMODE or EXMODE) lock level is first obtained. If this
222 * flag is set, the OCFS2_LOCK_NEEDS_REFRESH flag will be set on the
223 * individual lockres l_flags member from the ast function. It is
224 * expected that the locking wrapper will clear the
225 * OCFS2_LOCK_NEEDS_REFRESH flag when done.
226 */
227#define LOCK_TYPE_REQUIRES_REFRESH 0x1
228
b80fc012 229/*
5ef0d4ea
MF		 230 * Indicate that a lock type makes use of the lock value block. The
231 * ->set_lvb lock type callback must be defined.
b80fc012
MF		 232 */
233#define LOCK_TYPE_USES_LVB 0x2
234
ccd979bd 235static struct ocfs2_lock_res_ops ocfs2_inode_rw_lops = {
54a7e755 236 .get_osb = ocfs2_get_inode_osb,
f625c979 237 .flags = 0,
ccd979bd
MF		 238};
239
e63aecb6 240static struct ocfs2_lock_res_ops ocfs2_inode_inode_lops = {
54a7e755 241 .get_osb = ocfs2_get_inode_osb,
810d5aeb
MF		 242 .check_downconvert = ocfs2_check_meta_downconvert,
243 .set_lvb = ocfs2_set_meta_lvb,
f1f54068 244 .downconvert_worker = ocfs2_data_convert_worker,
b80fc012 245 .flags = LOCK_TYPE_REQUIRES_REFRESH|LOCK_TYPE_USES_LVB,
ccd979bd
MF		 246};
247
ccd979bd 248static struct ocfs2_lock_res_ops ocfs2_super_lops = {
f625c979 249 .flags = LOCK_TYPE_REQUIRES_REFRESH,
ccd979bd
MF		 250};
251
252static struct ocfs2_lock_res_ops ocfs2_rename_lops = {
f625c979 253 .flags = 0,
ccd979bd
MF		 254};
255
6ca497a8 256static struct ocfs2_lock_res_ops ocfs2_nfs_sync_lops = {
257 .flags = 0,
258};
259
83273932
SE		 260static struct ocfs2_lock_res_ops ocfs2_orphan_scan_lops = {
261 .flags = LOCK_TYPE_REQUIRES_REFRESH|LOCK_TYPE_USES_LVB,
262};
263
d680efe9 264static struct ocfs2_lock_res_ops ocfs2_dentry_lops = {
54a7e755 265 .get_osb = ocfs2_get_dentry_osb,
d680efe9 266 .post_unlock = ocfs2_dentry_post_unlock,
cc567d89 267 .downconvert_worker = ocfs2_dentry_convert_worker,
f625c979 268 .flags = 0,
d680efe9
MF		 269};
270
50008630
TY		 271static struct ocfs2_lock_res_ops ocfs2_inode_open_lops = {
272 .get_osb = ocfs2_get_inode_osb,
273 .flags = 0,
274};
275
cf8e06f1
MF		 276static struct ocfs2_lock_res_ops ocfs2_flock_lops = {
277 .get_osb = ocfs2_get_file_osb,
278 .flags = 0,
279};
280
9e33d69f
JK		 281static struct ocfs2_lock_res_ops ocfs2_qinfo_lops = {
282 .set_lvb = ocfs2_set_qinfo_lvb,
283 .get_osb = ocfs2_get_qinfo_osb,
284 .flags = LOCK_TYPE_REQUIRES_REFRESH | LOCK_TYPE_USES_LVB,
285};
286
8dec98ed
TM		 287static struct ocfs2_lock_res_ops ocfs2_refcount_block_lops = {
288 .check_downconvert = ocfs2_check_refcount_downconvert,
289 .downconvert_worker = ocfs2_refcount_convert_worker,
290 .flags = 0,
291};
292
ccd979bd
MF		 293static inline int ocfs2_is_inode_lock(struct ocfs2_lock_res *lockres)
294{
295 return lockres->l_type == OCFS2_LOCK_TYPE_META ||
50008630
TY		 296 lockres->l_type == OCFS2_LOCK_TYPE_RW ||
297 lockres->l_type == OCFS2_LOCK_TYPE_OPEN;
ccd979bd
MF		 298}
299
c0e41338 300static inline struct ocfs2_lock_res *ocfs2_lksb_to_lock_res(struct ocfs2_dlm_lksb *lksb)
a796d286
JB		 301{
302 return container_of(lksb, struct ocfs2_lock_res, l_lksb);
303}
304
ccd979bd
MF		 305static inline struct inode *ocfs2_lock_res_inode(struct ocfs2_lock_res *lockres)
306{
307 BUG_ON(!ocfs2_is_inode_lock(lockres));
308
309 return (struct inode *) lockres->l_priv;
310}
311
d680efe9
MF		 312static inline struct ocfs2_dentry_lock *ocfs2_lock_res_dl(struct ocfs2_lock_res *lockres)
313{
314 BUG_ON(lockres->l_type != OCFS2_LOCK_TYPE_DENTRY);
315
316 return (struct ocfs2_dentry_lock *)lockres->l_priv;
317}
318
9e33d69f
JK		 319static inline struct ocfs2_mem_dqinfo *ocfs2_lock_res_qinfo(struct ocfs2_lock_res *lockres)
320{
321 BUG_ON(lockres->l_type != OCFS2_LOCK_TYPE_QINFO);
322
323 return (struct ocfs2_mem_dqinfo *)lockres->l_priv;
324}
325
8dec98ed
TM		 326static inline struct ocfs2_refcount_tree *
327ocfs2_lock_res_refcount_tree(struct ocfs2_lock_res *res)
328{
329 return container_of(res, struct ocfs2_refcount_tree, rf_lockres);
330}
331
54a7e755
MF		 332static inline struct ocfs2_super *ocfs2_get_lockres_osb(struct ocfs2_lock_res *lockres)
333{
334 if (lockres->l_ops->get_osb)
335 return lockres->l_ops->get_osb(lockres);
336
337 return (struct ocfs2_super *)lockres->l_priv;
338}
339
ccd979bd
MF		 340static int ocfs2_lock_create(struct ocfs2_super *osb,
341 struct ocfs2_lock_res *lockres,
342 int level,
bd3e7610 343 u32 dlm_flags);
ccd979bd
MF		 344static inline int ocfs2_may_continue_on_blocked_lock(struct ocfs2_lock_res *lockres,
345 int wanted);
cb25797d
JK		 346static void __ocfs2_cluster_unlock(struct ocfs2_super *osb,
347 struct ocfs2_lock_res *lockres,
348 int level, unsigned long caller_ip);
349static inline void ocfs2_cluster_unlock(struct ocfs2_super *osb,
350 struct ocfs2_lock_res *lockres,
351 int level)
352{
353 __ocfs2_cluster_unlock(osb, lockres, level, _RET_IP_);
354}
355
ccd979bd
MF		 356static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res *lockres);
357static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res *lockres);
358static inline void ocfs2_generic_handle_attach_action(struct ocfs2_lock_res *lockres);
359static int ocfs2_generic_handle_bast(struct ocfs2_lock_res *lockres, int level);
360static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb,
361 struct ocfs2_lock_res *lockres);
362static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres,
363 int convert);
c74ff8bb
SM		 364#define ocfs2_log_dlm_error(_func, _err, _lockres) do { \
365 if ((_lockres)->l_type != OCFS2_LOCK_TYPE_DENTRY) \
366 mlog(ML_ERROR, "DLM error %d while calling %s on resource %s\n", \
367 _err, _func, _lockres->l_name); \
368 else \
369 mlog(ML_ERROR, "DLM error %d while calling %s on resource %.*s%08x\n", \
370 _err, _func, OCFS2_DENTRY_LOCK_INO_START - 1, (_lockres)->l_name, \
371 (unsigned int)ocfs2_get_dentry_lock_ino(_lockres)); \
ccd979bd 372} while (0)
34d024f8
MF		 373static int ocfs2_downconvert_thread(void *arg);
374static void ocfs2_downconvert_on_unlock(struct ocfs2_super *osb,
375 struct ocfs2_lock_res *lockres);
e63aecb6 376static int ocfs2_inode_lock_update(struct inode *inode,
ccd979bd
MF		 377 struct buffer_head **bh);
378static void ocfs2_drop_osb_locks(struct ocfs2_super *osb);
379static inline int ocfs2_highest_compat_lock_level(int level);
de551246
JB		 380static unsigned int ocfs2_prepare_downconvert(struct ocfs2_lock_res *lockres,
381 int new_level);
cf8e06f1
MF		 382static int ocfs2_downconvert_lock(struct ocfs2_super *osb,
383 struct ocfs2_lock_res *lockres,
384 int new_level,
de551246
JB		 385 int lvb,
386 unsigned int generation);
cf8e06f1
MF		 387static int ocfs2_prepare_cancel_convert(struct ocfs2_super *osb,
388 struct ocfs2_lock_res *lockres);
389static int ocfs2_cancel_convert(struct ocfs2_super *osb,
390 struct ocfs2_lock_res *lockres);
391
ccd979bd 392
ccd979bd
MF		 393static void ocfs2_build_lock_name(enum ocfs2_lock_type type,
394 u64 blkno,
395 u32 generation,
396 char *name)
397{
398 int len;
399
400 mlog_entry_void();
401
402 BUG_ON(type >= OCFS2_NUM_LOCK_TYPES);
403
b0697053
MF		 404 len = snprintf(name, OCFS2_LOCK_ID_MAX_LEN, "%c%s%016llx%08x",
405 ocfs2_lock_type_char(type), OCFS2_LOCK_ID_PAD,
406 (long long)blkno, generation);
ccd979bd
MF		 407
408 BUG_ON(len != (OCFS2_LOCK_ID_MAX_LEN - 1));
409
410 mlog(0, "built lock resource with name: %s\n", name);
411
412 mlog_exit_void();
413}
414
34af946a 415static DEFINE_SPINLOCK(ocfs2_dlm_tracking_lock);
ccd979bd
MF		 416
417static void ocfs2_add_lockres_tracking(struct ocfs2_lock_res *res,
418 struct ocfs2_dlm_debug *dlm_debug)
419{
420 mlog(0, "Add tracking for lockres %s\n", res->l_name);
421
422 spin_lock(&ocfs2_dlm_tracking_lock);
423 list_add(&res->l_debug_list, &dlm_debug->d_lockres_tracking);
424 spin_unlock(&ocfs2_dlm_tracking_lock);
425}
426
427static void ocfs2_remove_lockres_tracking(struct ocfs2_lock_res *res)
428{
429 spin_lock(&ocfs2_dlm_tracking_lock);
430 if (!list_empty(&res->l_debug_list))
431 list_del_init(&res->l_debug_list);
432 spin_unlock(&ocfs2_dlm_tracking_lock);
433}
434
8ddb7b00
SM		 435#ifdef CONFIG_OCFS2_FS_STATS
436static void ocfs2_init_lock_stats(struct ocfs2_lock_res *res)
437{
438 res->l_lock_num_prmode = 0;
439 res->l_lock_num_prmode_failed = 0;
440 res->l_lock_total_prmode = 0;
441 res->l_lock_max_prmode = 0;
442 res->l_lock_num_exmode = 0;
443 res->l_lock_num_exmode_failed = 0;
444 res->l_lock_total_exmode = 0;
445 res->l_lock_max_exmode = 0;
446 res->l_lock_refresh = 0;
447}
448
449static void ocfs2_update_lock_stats(struct ocfs2_lock_res *res, int level,
450 struct ocfs2_mask_waiter *mw, int ret)
451{
452 unsigned long long *num, *sum;
453 unsigned int *max, *failed;
454 struct timespec ts = current_kernel_time();
455 unsigned long long time = timespec_to_ns(&ts) - mw->mw_lock_start;
456
457 if (level == LKM_PRMODE) {
458 num = &res->l_lock_num_prmode;
459 sum = &res->l_lock_total_prmode;
460 max = &res->l_lock_max_prmode;
461 failed = &res->l_lock_num_prmode_failed;
462 } else if (level == LKM_EXMODE) {
463 num = &res->l_lock_num_exmode;
464 sum = &res->l_lock_total_exmode;
465 max = &res->l_lock_max_exmode;
466 failed = &res->l_lock_num_exmode_failed;
467 } else
468 return;
469
470 (*num)++;
471 (*sum) += time;
472 if (time > *max)
473 *max = time;
474 if (ret)
475 (*failed)++;
476}
477
478static inline void ocfs2_track_lock_refresh(struct ocfs2_lock_res *lockres)
479{
480 lockres->l_lock_refresh++;
481}
482
483static inline void ocfs2_init_start_time(struct ocfs2_mask_waiter *mw)
484{
485 struct timespec ts = current_kernel_time();
486 mw->mw_lock_start = timespec_to_ns(&ts);
487}
488#else
489static inline void ocfs2_init_lock_stats(struct ocfs2_lock_res *res)
490{
491}
492static inline void ocfs2_update_lock_stats(struct ocfs2_lock_res *res,
493 int level, struct ocfs2_mask_waiter *mw, int ret)
494{
495}
496static inline void ocfs2_track_lock_refresh(struct ocfs2_lock_res *lockres)
497{
498}
499static inline void ocfs2_init_start_time(struct ocfs2_mask_waiter *mw)
500{
501}
502#endif
503
ccd979bd
MF		 504static void ocfs2_lock_res_init_common(struct ocfs2_super *osb,
505 struct ocfs2_lock_res *res,
506 enum ocfs2_lock_type type,
ccd979bd
MF		 507 struct ocfs2_lock_res_ops *ops,
508 void *priv)
509{
ccd979bd
MF		 510 res->l_type = type;
511 res->l_ops = ops;
512 res->l_priv = priv;
513
bd3e7610
JB		 514 res->l_level = DLM_LOCK_IV;
515 res->l_requested = DLM_LOCK_IV;
516 res->l_blocking = DLM_LOCK_IV;
ccd979bd
MF		 517 res->l_action = OCFS2_AST_INVALID;
518 res->l_unlock_action = OCFS2_UNLOCK_INVALID;
519
520 res->l_flags = OCFS2_LOCK_INITIALIZED;
521
522 ocfs2_add_lockres_tracking(res, osb->osb_dlm_debug);
8ddb7b00
SM		 523
524 ocfs2_init_lock_stats(res);
cb25797d
JK		 525#ifdef CONFIG_DEBUG_LOCK_ALLOC
526 if (type != OCFS2_LOCK_TYPE_OPEN)
527 lockdep_init_map(&res->l_lockdep_map, ocfs2_lock_type_strings[type],
528 &lockdep_keys[type], 0);
529 else
530 res->l_lockdep_map.key = NULL;
531#endif
ccd979bd
MF		 532}
533
534void ocfs2_lock_res_init_once(struct ocfs2_lock_res *res)
535{
536 /* This also clears out the lock status block */
537 memset(res, 0, sizeof(struct ocfs2_lock_res));
538 spin_lock_init(&res->l_lock);
539 init_waitqueue_head(&res->l_event);
540 INIT_LIST_HEAD(&res->l_blocked_list);
541 INIT_LIST_HEAD(&res->l_mask_waiters);
542}
543
544void ocfs2_inode_lock_res_init(struct ocfs2_lock_res *res,
545 enum ocfs2_lock_type type,
24c19ef4 546 unsigned int generation,
ccd979bd
MF		 547 struct inode *inode)
548{
549 struct ocfs2_lock_res_ops *ops;
550
551 switch(type) {
552 case OCFS2_LOCK_TYPE_RW:
553 ops = &ocfs2_inode_rw_lops;
554 break;
555 case OCFS2_LOCK_TYPE_META:
e63aecb6 556 ops = &ocfs2_inode_inode_lops;
ccd979bd 557 break;
50008630
TY		 558 case OCFS2_LOCK_TYPE_OPEN:
559 ops = &ocfs2_inode_open_lops;
560 break;
ccd979bd
MF		 561 default:
562 mlog_bug_on_msg(1, "type: %d\n", type);
563 ops = NULL; /* thanks, gcc */
564 break;
565 };
566
d680efe9 567 ocfs2_build_lock_name(type, OCFS2_I(inode)->ip_blkno,
24c19ef4 568 generation, res->l_name);
d680efe9
MF		 569 ocfs2_lock_res_init_common(OCFS2_SB(inode->i_sb), res, type, ops, inode);
570}
571
54a7e755
MF		 572static struct ocfs2_super *ocfs2_get_inode_osb(struct ocfs2_lock_res *lockres)
573{
574 struct inode *inode = ocfs2_lock_res_inode(lockres);
575
576 return OCFS2_SB(inode->i_sb);
577}
578
9e33d69f
JK		 579static struct ocfs2_super *ocfs2_get_qinfo_osb(struct ocfs2_lock_res *lockres)
580{
581 struct ocfs2_mem_dqinfo *info = lockres->l_priv;
582
583 return OCFS2_SB(info->dqi_gi.dqi_sb);
584}
585
cf8e06f1
MF		 586static struct ocfs2_super *ocfs2_get_file_osb(struct ocfs2_lock_res *lockres)
587{
588 struct ocfs2_file_private *fp = lockres->l_priv;
589
590 return OCFS2_SB(fp->fp_file->f_mapping->host->i_sb);
591}
592
d680efe9
MF		 593static __u64 ocfs2_get_dentry_lock_ino(struct ocfs2_lock_res *lockres)
594{
595 __be64 inode_blkno_be;
596
597 memcpy(&inode_blkno_be, &lockres->l_name[OCFS2_DENTRY_LOCK_INO_START],
598 sizeof(__be64));
599
600 return be64_to_cpu(inode_blkno_be);
601}
602
54a7e755
MF		 603static struct ocfs2_super *ocfs2_get_dentry_osb(struct ocfs2_lock_res *lockres)
604{
605 struct ocfs2_dentry_lock *dl = lockres->l_priv;
606
607 return OCFS2_SB(dl->dl_inode->i_sb);
608}
609
d680efe9
MF		 610void ocfs2_dentry_lock_res_init(struct ocfs2_dentry_lock *dl,
611 u64 parent, struct inode *inode)
612{
613 int len;
614 u64 inode_blkno = OCFS2_I(inode)->ip_blkno;
615 __be64 inode_blkno_be = cpu_to_be64(inode_blkno);
616 struct ocfs2_lock_res *lockres = &dl->dl_lockres;
617
618 ocfs2_lock_res_init_once(lockres);
619
620 /*
621 * Unfortunately, the standard lock naming scheme won't work
622 * here because we have two 16 byte values to use. Instead,
623 * we'll stuff the inode number as a binary value. We still
624 * want error prints to show something without garbling the
625 * display, so drop a null byte in there before the inode
626 * number. A future version of OCFS2 will likely use all
627 * binary lock names. The stringified names have been a
628 * tremendous aid in debugging, but now that the debugfs
629 * interface exists, we can mangle things there if need be.
630 *
631 * NOTE: We also drop the standard "pad" value (the total lock
632 * name size stays the same though - the last part is all
633 * zeros due to the memset in ocfs2_lock_res_init_once()
634 */
635 len = snprintf(lockres->l_name, OCFS2_DENTRY_LOCK_INO_START,
636 "%c%016llx",
637 ocfs2_lock_type_char(OCFS2_LOCK_TYPE_DENTRY),
638 (long long)parent);
639
640 BUG_ON(len != (OCFS2_DENTRY_LOCK_INO_START - 1));
641
642 memcpy(&lockres->l_name[OCFS2_DENTRY_LOCK_INO_START], &inode_blkno_be,
643 sizeof(__be64));
644
645 ocfs2_lock_res_init_common(OCFS2_SB(inode->i_sb), lockres,
646 OCFS2_LOCK_TYPE_DENTRY, &ocfs2_dentry_lops,
647 dl);
ccd979bd
MF		 648}
649
650static void ocfs2_super_lock_res_init(struct ocfs2_lock_res *res,
651 struct ocfs2_super *osb)
652{
653 /* Superblock lockres doesn't come from a slab so we call init
654 * once on it manually. */
655 ocfs2_lock_res_init_once(res);
d680efe9
MF		 656 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_SUPER, OCFS2_SUPER_BLOCK_BLKNO,
657 0, res->l_name);
ccd979bd 658 ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_SUPER,
ccd979bd
MF		 659 &ocfs2_super_lops, osb);
660}
661
662static void ocfs2_rename_lock_res_init(struct ocfs2_lock_res *res,
663 struct ocfs2_super *osb)
664{
665 /* Rename lockres doesn't come from a slab so we call init
666 * once on it manually. */
667 ocfs2_lock_res_init_once(res);
d680efe9
MF		 668 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_RENAME, 0, 0, res->l_name);
669 ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_RENAME,
ccd979bd
MF		 670 &ocfs2_rename_lops, osb);
671}
672
6ca497a8 673static void ocfs2_nfs_sync_lock_res_init(struct ocfs2_lock_res *res,
674 struct ocfs2_super *osb)
675{
676 /* nfs_sync lockres doesn't come from a slab so we call init
677 * once on it manually. */
678 ocfs2_lock_res_init_once(res);
679 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_NFS_SYNC, 0, 0, res->l_name);
680 ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_NFS_SYNC,
681 &ocfs2_nfs_sync_lops, osb);
682}
683
83273932
SE		 684static void ocfs2_orphan_scan_lock_res_init(struct ocfs2_lock_res *res,
685 struct ocfs2_super *osb)
686{
83273932
SE		 687 ocfs2_lock_res_init_once(res);
688 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_ORPHAN_SCAN, 0, 0, res->l_name);
689 ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_ORPHAN_SCAN,
690 &ocfs2_orphan_scan_lops, osb);
83273932
SE		 691}
692
cf8e06f1
MF		 693void ocfs2_file_lock_res_init(struct ocfs2_lock_res *lockres,
694 struct ocfs2_file_private *fp)
695{
696 struct inode *inode = fp->fp_file->f_mapping->host;
697 struct ocfs2_inode_info *oi = OCFS2_I(inode);
698
699 ocfs2_lock_res_init_once(lockres);
700 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_FLOCK, oi->ip_blkno,
701 inode->i_generation, lockres->l_name);
702 ocfs2_lock_res_init_common(OCFS2_SB(inode->i_sb), lockres,
703 OCFS2_LOCK_TYPE_FLOCK, &ocfs2_flock_lops,
704 fp);
705 lockres->l_flags |= OCFS2_LOCK_NOCACHE;
706}
707
9e33d69f
JK		 708void ocfs2_qinfo_lock_res_init(struct ocfs2_lock_res *lockres,
709 struct ocfs2_mem_dqinfo *info)
710{
711 ocfs2_lock_res_init_once(lockres);
712 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_QINFO, info->dqi_gi.dqi_type,
713 0, lockres->l_name);
714 ocfs2_lock_res_init_common(OCFS2_SB(info->dqi_gi.dqi_sb), lockres,
715 OCFS2_LOCK_TYPE_QINFO, &ocfs2_qinfo_lops,
716 info);
717}
718
8dec98ed
TM		 719void ocfs2_refcount_lock_res_init(struct ocfs2_lock_res *lockres,
720 struct ocfs2_super *osb, u64 ref_blkno,
721 unsigned int generation)
722{
723 ocfs2_lock_res_init_once(lockres);
724 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_REFCOUNT, ref_blkno,
725 generation, lockres->l_name);
726 ocfs2_lock_res_init_common(osb, lockres, OCFS2_LOCK_TYPE_REFCOUNT,
727 &ocfs2_refcount_block_lops, osb);
728}
729
ccd979bd
MF		 730void ocfs2_lock_res_free(struct ocfs2_lock_res *res)
731{
732 mlog_entry_void();
733
734 if (!(res->l_flags & OCFS2_LOCK_INITIALIZED))
735 return;
736
737 ocfs2_remove_lockres_tracking(res);
738
739 mlog_bug_on_msg(!list_empty(&res->l_blocked_list),
740 "Lockres %s is on the blocked list\n",
741 res->l_name);
742 mlog_bug_on_msg(!list_empty(&res->l_mask_waiters),
743 "Lockres %s has mask waiters pending\n",
744 res->l_name);
745 mlog_bug_on_msg(spin_is_locked(&res->l_lock),
746 "Lockres %s is locked\n",
747 res->l_name);
748 mlog_bug_on_msg(res->l_ro_holders,
749 "Lockres %s has %u ro holders\n",
750 res->l_name, res->l_ro_holders);
751 mlog_bug_on_msg(res->l_ex_holders,
752 "Lockres %s has %u ex holders\n",
753 res->l_name, res->l_ex_holders);
754
755 /* Need to clear out the lock status block for the dlm */
756 memset(&res->l_lksb, 0, sizeof(res->l_lksb));
757
758 res->l_flags = 0UL;
759 mlog_exit_void();
760}
761
762static inline void ocfs2_inc_holders(struct ocfs2_lock_res *lockres,
763 int level)
764{
765 mlog_entry_void();
766
767 BUG_ON(!lockres);
768
769 switch(level) {
bd3e7610 770 case DLM_LOCK_EX:
ccd979bd
MF		 771 lockres->l_ex_holders++;
772 break;
bd3e7610 773 case DLM_LOCK_PR:
ccd979bd
MF		 774 lockres->l_ro_holders++;
775 break;
776 default:
777 BUG();
778 }
779
780 mlog_exit_void();
781}
782
783static inline void ocfs2_dec_holders(struct ocfs2_lock_res *lockres,
784 int level)
785{
786 mlog_entry_void();
787
788 BUG_ON(!lockres);
789
790 switch(level) {
bd3e7610 791 case DLM_LOCK_EX:
ccd979bd
MF		 792 BUG_ON(!lockres->l_ex_holders);
793 lockres->l_ex_holders--;
794 break;
bd3e7610 795 case DLM_LOCK_PR:
ccd979bd
MF		 796 BUG_ON(!lockres->l_ro_holders);
797 lockres->l_ro_holders--;
798 break;
799 default:
800 BUG();
801 }
802 mlog_exit_void();
803}
804
805/* WARNING: This function lives in a world where the only three lock
806 * levels are EX, PR, and NL. It *will* have to be adjusted when more
807 * lock types are added. */
808static inline int ocfs2_highest_compat_lock_level(int level)
809{
bd3e7610 810 int new_level = DLM_LOCK_EX;
ccd979bd 811
bd3e7610
JB		 812 if (level == DLM_LOCK_EX)
813 new_level = DLM_LOCK_NL;
814 else if (level == DLM_LOCK_PR)
815 new_level = DLM_LOCK_PR;
ccd979bd
MF		 816 return new_level;
817}
818
819static void lockres_set_flags(struct ocfs2_lock_res *lockres,
820 unsigned long newflags)
821{
800deef3 822 struct ocfs2_mask_waiter *mw, *tmp;
ccd979bd
MF		 823
824 assert_spin_locked(&lockres->l_lock);
825
826 lockres->l_flags = newflags;
827
800deef3 828 list_for_each_entry_safe(mw, tmp, &lockres->l_mask_waiters, mw_item) {
ccd979bd
MF		 829 if ((lockres->l_flags & mw->mw_mask) != mw->mw_goal)
830 continue;
831
832 list_del_init(&mw->mw_item);
833 mw->mw_status = 0;
834 complete(&mw->mw_complete);
835 }
836}
837static void lockres_or_flags(struct ocfs2_lock_res *lockres, unsigned long or)
838{
839 lockres_set_flags(lockres, lockres->l_flags | or);
840}
841static void lockres_clear_flags(struct ocfs2_lock_res *lockres,
842 unsigned long clear)
843{
844 lockres_set_flags(lockres, lockres->l_flags & ~clear);
845}
846
847static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res *lockres)
848{
849 mlog_entry_void();
850
851 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY));
852 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_ATTACHED));
853 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BLOCKED));
bd3e7610 854 BUG_ON(lockres->l_blocking <= DLM_LOCK_NL);
ccd979bd
MF		 855
856 lockres->l_level = lockres->l_requested;
857 if (lockres->l_level <=
858 ocfs2_highest_compat_lock_level(lockres->l_blocking)) {
bd3e7610 859 lockres->l_blocking = DLM_LOCK_NL;
ccd979bd
MF		 860 lockres_clear_flags(lockres, OCFS2_LOCK_BLOCKED);
861 }
862 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
863
864 mlog_exit_void();
865}
866
867static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res *lockres)
868{
869 mlog_entry_void();
870
871 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY));
872 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_ATTACHED));
873
874 /* Convert from RO to EX doesn't really need anything as our
875 * information is already up to data. Convert from NL to
876 * *anything* however should mark ourselves as needing an
877 * update */
bd3e7610 878 if (lockres->l_level == DLM_LOCK_NL &&
f625c979 879 lockres->l_ops->flags & LOCK_TYPE_REQUIRES_REFRESH)
ccd979bd
MF		 880 lockres_or_flags(lockres, OCFS2_LOCK_NEEDS_REFRESH);
881
882 lockres->l_level = lockres->l_requested;
a1912826
SM		 883
884 /*
885 * We set the OCFS2_LOCK_UPCONVERT_FINISHING flag before clearing
886 * the OCFS2_LOCK_BUSY flag to prevent the dc thread from
887 * downconverting the lock before the upconvert has fully completed.
888 */
889 lockres_or_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
890
ccd979bd
MF		 891 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
892
893 mlog_exit_void();
894}
895
896static inline void ocfs2_generic_handle_attach_action(struct ocfs2_lock_res *lockres)
897{
898 mlog_entry_void();
899
3cf0c507 900 BUG_ON((!(lockres->l_flags & OCFS2_LOCK_BUSY)));
ccd979bd
MF		 901 BUG_ON(lockres->l_flags & OCFS2_LOCK_ATTACHED);
902
bd3e7610 903 if (lockres->l_requested > DLM_LOCK_NL &&
f625c979
MF		 904 !(lockres->l_flags & OCFS2_LOCK_LOCAL) &&
905 lockres->l_ops->flags & LOCK_TYPE_REQUIRES_REFRESH)
ccd979bd
MF		 906 lockres_or_flags(lockres, OCFS2_LOCK_NEEDS_REFRESH);
907
908 lockres->l_level = lockres->l_requested;
909 lockres_or_flags(lockres, OCFS2_LOCK_ATTACHED);
910 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
911
912 mlog_exit_void();
913}
914
ccd979bd
MF		 915static int ocfs2_generic_handle_bast(struct ocfs2_lock_res *lockres,
916 int level)
917{
918 int needs_downconvert = 0;
919 mlog_entry_void();
920
921 assert_spin_locked(&lockres->l_lock);
922
ccd979bd
MF		 923 if (level > lockres->l_blocking) {
924 /* only schedule a downconvert if we haven't already scheduled
925 * one that goes low enough to satisfy the level we're
926 * blocking. this also catches the case where we get
927 * duplicate BASTs */
928 if (ocfs2_highest_compat_lock_level(level) <
929 ocfs2_highest_compat_lock_level(lockres->l_blocking))
930 needs_downconvert = 1;
931
932 lockres->l_blocking = level;
933 }
934
9b915181
SM		 935 mlog(ML_BASTS, "lockres %s, block %d, level %d, l_block %d, dwn %d\n",
936 lockres->l_name, level, lockres->l_level, lockres->l_blocking,
937 needs_downconvert);
938
0b94a909
WW		 939 if (needs_downconvert)
940 lockres_or_flags(lockres, OCFS2_LOCK_BLOCKED);
941
ccd979bd
MF		 942 mlog_exit(needs_downconvert);
943 return needs_downconvert;
944}
945
de551246
JB		 946/*
947 * OCFS2_LOCK_PENDING and l_pending_gen.
948 *
949 * Why does OCFS2_LOCK_PENDING exist? To close a race between setting
950 * OCFS2_LOCK_BUSY and calling ocfs2_dlm_lock(). See ocfs2_unblock_lock()
951 * for more details on the race.
952 *
953 * OCFS2_LOCK_PENDING closes the race quite nicely. However, it introduces
954 * a race on itself. In o2dlm, we can get the ast before ocfs2_dlm_lock()
955 * returns. The ast clears OCFS2_LOCK_BUSY, and must therefore clear
956 * OCFS2_LOCK_PENDING at the same time. When ocfs2_dlm_lock() returns,
957 * the caller is going to try to clear PENDING again. If nothing else is
958 * happening, __lockres_clear_pending() sees PENDING is unset and does
959 * nothing.
960 *
961 * But what if another path (eg downconvert thread) has just started a
962 * new locking action? The other path has re-set PENDING. Our path
963 * cannot clear PENDING, because that will re-open the original race
964 * window.
965 *
966 * [Example]
967 *
968 * ocfs2_meta_lock()
969 * ocfs2_cluster_lock()
970 * set BUSY
971 * set PENDING
972 * drop l_lock
973 * ocfs2_dlm_lock()
974 * ocfs2_locking_ast() ocfs2_downconvert_thread()
975 * clear PENDING ocfs2_unblock_lock()
976 * take_l_lock
977 * !BUSY
978 * ocfs2_prepare_downconvert()
979 * set BUSY
980 * set PENDING
981 * drop l_lock
982 * take l_lock
983 * clear PENDING
984 * drop l_lock
985 * <window>
986 * ocfs2_dlm_lock()
987 *
988 * So as you can see, we now have a window where l_lock is not held,
989 * PENDING is not set, and ocfs2_dlm_lock() has not been called.
990 *
991 * The core problem is that ocfs2_cluster_lock() has cleared the PENDING
992 * set by ocfs2_prepare_downconvert(). That wasn't nice.
993 *
994 * To solve this we introduce l_pending_gen. A call to
995 * lockres_clear_pending() will only do so when it is passed a generation
996 * number that matches the lockres. lockres_set_pending() will return the
997 * current generation number. When ocfs2_cluster_lock() goes to clear
998 * PENDING, it passes the generation it got from set_pending(). In our
999 * example above, the generation numbers will *not* match. Thus,
1000 * ocfs2_cluster_lock() will not clear the PENDING set by
1001 * ocfs2_prepare_downconvert().
1002 */
1003
1004/* Unlocked version for ocfs2_locking_ast() */
1005static void __lockres_clear_pending(struct ocfs2_lock_res *lockres,
1006 unsigned int generation,
1007 struct ocfs2_super *osb)
1008{
1009 assert_spin_locked(&lockres->l_lock);
1010
1011 /*
1012 * The ast and locking functions can race us here. The winner
1013 * will clear pending, the loser will not.
1014 */
1015 if (!(lockres->l_flags & OCFS2_LOCK_PENDING) ||
1016 (lockres->l_pending_gen != generation))
1017 return;
1018
1019 lockres_clear_flags(lockres, OCFS2_LOCK_PENDING);
1020 lockres->l_pending_gen++;
1021
1022 /*
1023 * The downconvert thread may have skipped us because we
1024 * were PENDING. Wake it up.
1025 */
1026 if (lockres->l_flags & OCFS2_LOCK_BLOCKED)
1027 ocfs2_wake_downconvert_thread(osb);
1028}
1029
1030/* Locked version for callers of ocfs2_dlm_lock() */
1031static void lockres_clear_pending(struct ocfs2_lock_res *lockres,
1032 unsigned int generation,
1033 struct ocfs2_super *osb)
1034{
1035 unsigned long flags;
1036
1037 spin_lock_irqsave(&lockres->l_lock, flags);
1038 __lockres_clear_pending(lockres, generation, osb);
1039 spin_unlock_irqrestore(&lockres->l_lock, flags);
1040}
1041
1042static unsigned int lockres_set_pending(struct ocfs2_lock_res *lockres)
1043{
1044 assert_spin_locked(&lockres->l_lock);
1045 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY));
1046
1047 lockres_or_flags(lockres, OCFS2_LOCK_PENDING);
1048
1049 return lockres->l_pending_gen;
1050}
1051
c0e41338 1052static void ocfs2_blocking_ast(struct ocfs2_dlm_lksb *lksb, int level)
ccd979bd 1053{
a796d286 1054 struct ocfs2_lock_res *lockres = ocfs2_lksb_to_lock_res(lksb);
aa2623ad 1055 struct ocfs2_super *osb = ocfs2_get_lockres_osb(lockres);
ccd979bd
MF		 1056 int needs_downconvert;
1057 unsigned long flags;
1058
bd3e7610 1059 BUG_ON(level <= DLM_LOCK_NL);
ccd979bd 1060
9b915181
SM		 1061 mlog(ML_BASTS, "BAST fired for lockres %s, blocking %d, level %d, "
1062 "type %s\n", lockres->l_name, level, lockres->l_level,
aa2623ad
MF		 1063 ocfs2_lock_type_string(lockres->l_type));
1064
cf8e06f1
MF		 1065 /*
1066 * We can skip the bast for locks which don't enable caching -
1067 * they'll be dropped at the earliest possible time anyway.
1068 */
1069 if (lockres->l_flags & OCFS2_LOCK_NOCACHE)
1070 return;
1071
ccd979bd
MF		 1072 spin_lock_irqsave(&lockres->l_lock, flags);
1073 needs_downconvert = ocfs2_generic_handle_bast(lockres, level);
1074 if (needs_downconvert)
1075 ocfs2_schedule_blocked_lock(osb, lockres);
1076 spin_unlock_irqrestore(&lockres->l_lock, flags);
1077
d680efe9
MF		 1078 wake_up(&lockres->l_event);
1079
34d024f8 1080 ocfs2_wake_downconvert_thread(osb);
ccd979bd
MF		 1081}
1082
c0e41338 1083static void ocfs2_locking_ast(struct ocfs2_dlm_lksb *lksb)
ccd979bd 1084{
a796d286 1085 struct ocfs2_lock_res *lockres = ocfs2_lksb_to_lock_res(lksb);
de551246 1086 struct ocfs2_super *osb = ocfs2_get_lockres_osb(lockres);
ccd979bd 1087 unsigned long flags;
1693a5c0 1088 int status;
ccd979bd
MF		 1089
1090 spin_lock_irqsave(&lockres->l_lock, flags);
1091
1693a5c0
DT		 1092 status = ocfs2_dlm_lock_status(&lockres->l_lksb);
1093
1094 if (status == -EAGAIN) {
1095 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
1096 goto out;
1097 }
1098
1099 if (status) {
8f2c9c1b 1100 mlog(ML_ERROR, "lockres %s: lksb status value of %d!\n",
1693a5c0 1101 lockres->l_name, status);
ccd979bd
MF		 1102 spin_unlock_irqrestore(&lockres->l_lock, flags);
1103 return;
1104 }
1105
9b915181
SM		 1106 mlog(ML_BASTS, "AST fired for lockres %s, action %d, unlock %d, "
1107 "level %d => %d\n", lockres->l_name, lockres->l_action,
1108 lockres->l_unlock_action, lockres->l_level, lockres->l_requested);
1109
ccd979bd
MF		 1110 switch(lockres->l_action) {
1111 case OCFS2_AST_ATTACH:
1112 ocfs2_generic_handle_attach_action(lockres);
e92d57df 1113 lockres_clear_flags(lockres, OCFS2_LOCK_LOCAL);
ccd979bd
MF		 1114 break;
1115 case OCFS2_AST_CONVERT:
1116 ocfs2_generic_handle_convert_action(lockres);
1117 break;
1118 case OCFS2_AST_DOWNCONVERT:
1119 ocfs2_generic_handle_downconvert_action(lockres);
1120 break;
1121 default:
9b915181
SM		 1122 mlog(ML_ERROR, "lockres %s: AST fired with invalid action: %u, "
1123 "flags 0x%lx, unlock: %u\n",
e92d57df
MF		 1124 lockres->l_name, lockres->l_action, lockres->l_flags,
1125 lockres->l_unlock_action);
ccd979bd
MF		 1126 BUG();
1127 }
1693a5c0 1128out:
ccd979bd
MF		 1129 /* set it to something invalid so if we get called again we
1130 * can catch it. */
1131 lockres->l_action = OCFS2_AST_INVALID;
ccd979bd 1132
de551246
JB		 1133 /* Did we try to cancel this lock? Clear that state */
1134 if (lockres->l_unlock_action == OCFS2_UNLOCK_CANCEL_CONVERT)
1135 lockres->l_unlock_action = OCFS2_UNLOCK_INVALID;
1136
1137 /*
1138 * We may have beaten the locking functions here. We certainly
1139 * know that dlm_lock() has been called :-)
1140 * Because we can't have two lock calls in flight at once, we
1141 * can use lockres->l_pending_gen.
1142 */
1143 __lockres_clear_pending(lockres, lockres->l_pending_gen, osb);
1144
ccd979bd 1145 wake_up(&lockres->l_event);
d680efe9 1146 spin_unlock_irqrestore(&lockres->l_lock, flags);
ccd979bd
MF		 1147}
1148
553b5eb9
JB		 1149static void ocfs2_unlock_ast(struct ocfs2_dlm_lksb *lksb, int error)
1150{
1151 struct ocfs2_lock_res *lockres = ocfs2_lksb_to_lock_res(lksb);
1152 unsigned long flags;
1153
1154 mlog_entry_void();
1155
9b915181
SM		 1156 mlog(ML_BASTS, "UNLOCK AST fired for lockres %s, action = %d\n",
1157 lockres->l_name, lockres->l_unlock_action);
553b5eb9
JB		 1158
1159 spin_lock_irqsave(&lockres->l_lock, flags);
1160 if (error) {
1161 mlog(ML_ERROR, "Dlm passes error %d for lock %s, "
1162 "unlock_action %d\n", error, lockres->l_name,
1163 lockres->l_unlock_action);
1164 spin_unlock_irqrestore(&lockres->l_lock, flags);
1165 mlog_exit_void();
1166 return;
1167 }
1168
1169 switch(lockres->l_unlock_action) {
1170 case OCFS2_UNLOCK_CANCEL_CONVERT:
1171 mlog(0, "Cancel convert success for %s\n", lockres->l_name);
1172 lockres->l_action = OCFS2_AST_INVALID;
1173 /* Downconvert thread may have requeued this lock, we
1174 * need to wake it. */
1175 if (lockres->l_flags & OCFS2_LOCK_BLOCKED)
1176 ocfs2_wake_downconvert_thread(ocfs2_get_lockres_osb(lockres));
1177 break;
1178 case OCFS2_UNLOCK_DROP_LOCK:
1179 lockres->l_level = DLM_LOCK_IV;
1180 break;
1181 default:
1182 BUG();
1183 }
1184
1185 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
1186 lockres->l_unlock_action = OCFS2_UNLOCK_INVALID;
1187 wake_up(&lockres->l_event);
1188 spin_unlock_irqrestore(&lockres->l_lock, flags);
1189
1190 mlog_exit_void();
1191}
1192
1193/*
1194 * This is the filesystem locking protocol. It provides the lock handling
1195 * hooks for the underlying DLM. It has a maximum version number.
1196 * The version number allows interoperability with systems running at
1197 * the same major number and an equal or smaller minor number.
1198 *
1199 * Whenever the filesystem does new things with locks (adds or removes a
1200 * lock, orders them differently, does different things underneath a lock),
1201 * the version must be changed. The protocol is negotiated when joining
1202 * the dlm domain. A node may join the domain if its major version is
1203 * identical to all other nodes and its minor version is greater than
1204 * or equal to all other nodes. When its minor version is greater than
1205 * the other nodes, it will run at the minor version specified by the
1206 * other nodes.
1207 *
1208 * If a locking change is made that will not be compatible with older
1209 * versions, the major number must be increased and the minor version set
1210 * to zero. If a change merely adds a behavior that can be disabled when
1211 * speaking to older versions, the minor version must be increased. If a
1212 * change adds a fully backwards compatible change (eg, LVB changes that
1213 * are just ignored by older versions), the version does not need to be
1214 * updated.
1215 */
1216static struct ocfs2_locking_protocol lproto = {
1217 .lp_max_version = {
1218 .pv_major = OCFS2_LOCKING_PROTOCOL_MAJOR,
1219 .pv_minor = OCFS2_LOCKING_PROTOCOL_MINOR,
1220 },
1221 .lp_lock_ast = ocfs2_locking_ast,
1222 .lp_blocking_ast = ocfs2_blocking_ast,
1223 .lp_unlock_ast = ocfs2_unlock_ast,
1224};
1225
1226void ocfs2_set_locking_protocol(void)
1227{
1228 ocfs2_stack_glue_set_max_proto_version(&lproto.lp_max_version);
1229}
1230
ccd979bd
MF		 1231static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres,
1232 int convert)
1233{
1234 unsigned long flags;
1235
1236 mlog_entry_void();
1237 spin_lock_irqsave(&lockres->l_lock, flags);
1238 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
a1912826 1239 lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
ccd979bd
MF		 1240 if (convert)
1241 lockres->l_action = OCFS2_AST_INVALID;
1242 else
1243 lockres->l_unlock_action = OCFS2_UNLOCK_INVALID;
1244 spin_unlock_irqrestore(&lockres->l_lock, flags);
1245
1246 wake_up(&lockres->l_event);
1247 mlog_exit_void();
1248}
1249
1250/* Note: If we detect another process working on the lock (i.e.,
1251 * OCFS2_LOCK_BUSY), we'll bail out returning 0. It's up to the caller
1252 * to do the right thing in that case.
1253 */
1254static int ocfs2_lock_create(struct ocfs2_super *osb,
1255 struct ocfs2_lock_res *lockres,
1256 int level,
bd3e7610 1257 u32 dlm_flags)
ccd979bd
MF		 1258{
1259 int ret = 0;
ccd979bd 1260 unsigned long flags;
de551246 1261 unsigned int gen;
ccd979bd
MF		 1262
1263 mlog_entry_void();
1264
bd3e7610 1265 mlog(0, "lock %s, level = %d, flags = %u\n", lockres->l_name, level,
ccd979bd
MF		 1266 dlm_flags);
1267
1268 spin_lock_irqsave(&lockres->l_lock, flags);
1269 if ((lockres->l_flags & OCFS2_LOCK_ATTACHED) ||
1270 (lockres->l_flags & OCFS2_LOCK_BUSY)) {
1271 spin_unlock_irqrestore(&lockres->l_lock, flags);
1272 goto bail;
1273 }
1274
1275 lockres->l_action = OCFS2_AST_ATTACH;
1276 lockres->l_requested = level;
1277 lockres_or_flags(lockres, OCFS2_LOCK_BUSY);
de551246 1278 gen = lockres_set_pending(lockres);
ccd979bd
MF		 1279 spin_unlock_irqrestore(&lockres->l_lock, flags);
1280
4670c46d 1281 ret = ocfs2_dlm_lock(osb->cconn,
7431cd7e
JB		 1282 level,
1283 &lockres->l_lksb,
1284 dlm_flags,
1285 lockres->l_name,
a796d286 1286 OCFS2_LOCK_ID_MAX_LEN - 1);
de551246 1287 lockres_clear_pending(lockres, gen, osb);
7431cd7e
JB		 1288 if (ret) {
1289 ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres);
ccd979bd
MF		 1290 ocfs2_recover_from_dlm_error(lockres, 1);
1291 }
1292
7431cd7e 1293 mlog(0, "lock %s, return from ocfs2_dlm_lock\n", lockres->l_name);
ccd979bd
MF		 1294
1295bail:
1296 mlog_exit(ret);
1297 return ret;
1298}
1299
1300static inline int ocfs2_check_wait_flag(struct ocfs2_lock_res *lockres,
1301 int flag)
1302{
1303 unsigned long flags;
1304 int ret;
1305
1306 spin_lock_irqsave(&lockres->l_lock, flags);
1307 ret = lockres->l_flags & flag;
1308 spin_unlock_irqrestore(&lockres->l_lock, flags);
1309
1310 return ret;
1311}
1312
1313static inline void ocfs2_wait_on_busy_lock(struct ocfs2_lock_res *lockres)
1314
1315{
1316 wait_event(lockres->l_event,
1317 !ocfs2_check_wait_flag(lockres, OCFS2_LOCK_BUSY));
1318}
1319
1320static inline void ocfs2_wait_on_refreshing_lock(struct ocfs2_lock_res *lockres)
1321
1322{
1323 wait_event(lockres->l_event,
1324 !ocfs2_check_wait_flag(lockres, OCFS2_LOCK_REFRESHING));
1325}
1326
1327/* predict what lock level we'll be dropping down to on behalf
1328 * of another node, and return true if the currently wanted
1329 * level will be compatible with it. */
1330static inline int ocfs2_may_continue_on_blocked_lock(struct ocfs2_lock_res *lockres,
1331 int wanted)
1332{
1333 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BLOCKED));
1334
1335 return wanted <= ocfs2_highest_compat_lock_level(lockres->l_blocking);
1336}
1337
1338static void ocfs2_init_mask_waiter(struct ocfs2_mask_waiter *mw)
1339{
1340 INIT_LIST_HEAD(&mw->mw_item);
1341 init_completion(&mw->mw_complete);
8ddb7b00 1342 ocfs2_init_start_time(mw);
ccd979bd
MF		 1343}
1344
1345static int ocfs2_wait_for_mask(struct ocfs2_mask_waiter *mw)
1346{
1347 wait_for_completion(&mw->mw_complete);
1348 /* Re-arm the completion in case we want to wait on it again */
1349 INIT_COMPLETION(mw->mw_complete);
1350 return mw->mw_status;
1351}
1352
1353static void lockres_add_mask_waiter(struct ocfs2_lock_res *lockres,
1354 struct ocfs2_mask_waiter *mw,
1355 unsigned long mask,
1356 unsigned long goal)
1357{
1358 BUG_ON(!list_empty(&mw->mw_item));
1359
1360 assert_spin_locked(&lockres->l_lock);
1361
1362 list_add_tail(&mw->mw_item, &lockres->l_mask_waiters);
1363 mw->mw_mask = mask;
1364 mw->mw_goal = goal;
1365}
1366
1367/* returns 0 if the mw that was removed was already satisfied, -EBUSY
1368 * if the mask still hadn't reached its goal */
1369static int lockres_remove_mask_waiter(struct ocfs2_lock_res *lockres,
1370 struct ocfs2_mask_waiter *mw)
1371{
1372 unsigned long flags;
1373 int ret = 0;
1374
1375 spin_lock_irqsave(&lockres->l_lock, flags);
1376 if (!list_empty(&mw->mw_item)) {
1377 if ((lockres->l_flags & mw->mw_mask) != mw->mw_goal)
1378 ret = -EBUSY;
1379
1380 list_del_init(&mw->mw_item);
1381 init_completion(&mw->mw_complete);
1382 }
1383 spin_unlock_irqrestore(&lockres->l_lock, flags);
1384
1385 return ret;
1386
1387}
1388
cf8e06f1
MF		 1389static int ocfs2_wait_for_mask_interruptible(struct ocfs2_mask_waiter *mw,
1390 struct ocfs2_lock_res *lockres)
1391{
1392 int ret;
1393
1394 ret = wait_for_completion_interruptible(&mw->mw_complete);
1395 if (ret)
1396 lockres_remove_mask_waiter(lockres, mw);
1397 else
1398 ret = mw->mw_status;
1399 /* Re-arm the completion in case we want to wait on it again */
1400 INIT_COMPLETION(mw->mw_complete);
1401 return ret;
1402}
1403
cb25797d
JK		 1404static int __ocfs2_cluster_lock(struct ocfs2_super *osb,
1405 struct ocfs2_lock_res *lockres,
1406 int level,
1407 u32 lkm_flags,
1408 int arg_flags,
1409 int l_subclass,
1410 unsigned long caller_ip)
ccd979bd
MF		 1411{
1412 struct ocfs2_mask_waiter mw;
ccd979bd
MF		 1413 int wait, catch_signals = !(osb->s_mount_opt & OCFS2_MOUNT_NOINTR);
1414 int ret = 0; /* gcc doesn't realize wait = 1 guarantees ret is set */
1415 unsigned long flags;
de551246 1416 unsigned int gen;
1693a5c0 1417 int noqueue_attempted = 0;
ccd979bd
MF		 1418
1419 mlog_entry_void();
1420
1421 ocfs2_init_mask_waiter(&mw);
1422
b80fc012 1423 if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB)
bd3e7610 1424 lkm_flags |= DLM_LKF_VALBLK;
b80fc012 1425
ccd979bd
MF		 1426again:
1427 wait = 0;
1428
a1912826
SM		 1429 spin_lock_irqsave(&lockres->l_lock, flags);
1430
ccd979bd
MF		 1431 if (catch_signals && signal_pending(current)) {
1432 ret = -ERESTARTSYS;
a1912826 1433 goto unlock;
ccd979bd
MF		 1434 }
1435
ccd979bd
MF		 1436 mlog_bug_on_msg(lockres->l_flags & OCFS2_LOCK_FREEING,
1437 "Cluster lock called on freeing lockres %s! flags "
1438 "0x%lx\n", lockres->l_name, lockres->l_flags);
1439
1440 /* We only compare against the currently granted level
1441 * here. If the lock is blocked waiting on a downconvert,
1442 * we'll get caught below. */
1443 if (lockres->l_flags & OCFS2_LOCK_BUSY &&
1444 level > lockres->l_level) {
1445 /* is someone sitting in dlm_lock? If so, wait on
1446 * them. */
1447 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0);
1448 wait = 1;
1449 goto unlock;
1450 }
1451
a1912826
SM		 1452 if (lockres->l_flags & OCFS2_LOCK_UPCONVERT_FINISHING) {
1453 /*
1454 * We've upconverted. If the lock now has a level we can
1455 * work with, we take it. If, however, the lock is not at the
1456 * required level, we go thru the full cycle. One way this could
1457 * happen is if a process requesting an upconvert to PR is
1458 * closely followed by another requesting upconvert to an EX.
1459 * If the process requesting EX lands here, we want it to
1460 * continue attempting to upconvert and let the process
1461 * requesting PR take the lock.
1462 * If multiple processes request upconvert to PR, the first one
1463 * here will take the lock. The others will have to go thru the
1464 * OCFS2_LOCK_BLOCKED check to ensure that there is no pending
1465 * downconvert request.
1466 */
1467 if (level <= lockres->l_level)
1468 goto update_holders;
1469 }
1470
ccd979bd
MF		 1471 if (lockres->l_flags & OCFS2_LOCK_BLOCKED &&
1472 !ocfs2_may_continue_on_blocked_lock(lockres, level)) {
1473 /* is the lock is currently blocked on behalf of
1474 * another node */
1475 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BLOCKED, 0);
1476 wait = 1;
1477 goto unlock;
1478 }
1479
1480 if (level > lockres->l_level) {
1693a5c0
DT		 1481 if (noqueue_attempted > 0) {
1482 ret = -EAGAIN;
1483 goto unlock;
1484 }
1485 if (lkm_flags & DLM_LKF_NOQUEUE)
1486 noqueue_attempted = 1;
1487
ccd979bd
MF		 1488 if (lockres->l_action != OCFS2_AST_INVALID)
1489 mlog(ML_ERROR, "lockres %s has action %u pending\n",
1490 lockres->l_name, lockres->l_action);
1491
019d1b22
MF		 1492 if (!(lockres->l_flags & OCFS2_LOCK_ATTACHED)) {
1493 lockres->l_action = OCFS2_AST_ATTACH;
bd3e7610 1494 lkm_flags &= ~DLM_LKF_CONVERT;
019d1b22
MF		 1495 } else {
1496 lockres->l_action = OCFS2_AST_CONVERT;
bd3e7610 1497 lkm_flags |= DLM_LKF_CONVERT;
019d1b22
MF		 1498 }
1499
ccd979bd
MF		 1500 lockres->l_requested = level;
1501 lockres_or_flags(lockres, OCFS2_LOCK_BUSY);
de551246 1502 gen = lockres_set_pending(lockres);
ccd979bd
MF		 1503 spin_unlock_irqrestore(&lockres->l_lock, flags);
1504
bd3e7610
JB		 1505 BUG_ON(level == DLM_LOCK_IV);
1506 BUG_ON(level == DLM_LOCK_NL);
ccd979bd 1507
9b915181 1508 mlog(ML_BASTS, "lockres %s, convert from %d to %d\n",
ccd979bd
MF		 1509 lockres->l_name, lockres->l_level, level);
1510
1511 /* call dlm_lock to upgrade lock now */
4670c46d 1512 ret = ocfs2_dlm_lock(osb->cconn,
7431cd7e
JB		 1513 level,
1514 &lockres->l_lksb,
1515 lkm_flags,
1516 lockres->l_name,
a796d286 1517 OCFS2_LOCK_ID_MAX_LEN - 1);
de551246 1518 lockres_clear_pending(lockres, gen, osb);
7431cd7e
JB		 1519 if (ret) {
1520 if (!(lkm_flags & DLM_LKF_NOQUEUE) ||
1521 (ret != -EAGAIN)) {
24ef1815 1522 ocfs2_log_dlm_error("ocfs2_dlm_lock",
7431cd7e 1523 ret, lockres);
ccd979bd
MF		 1524 }
1525 ocfs2_recover_from_dlm_error(lockres, 1);
1526 goto out;
1527 }
1528
73ac36ea 1529 mlog(0, "lock %s, successful return from ocfs2_dlm_lock\n",
ccd979bd
MF		 1530 lockres->l_name);
1531
1532 /* At this point we've gone inside the dlm and need to
1533 * complete our work regardless. */
1534 catch_signals = 0;
1535
1536 /* wait for busy to clear and carry on */
1537 goto again;
1538 }
1539
a1912826 1540update_holders:
ccd979bd
MF		 1541 /* Ok, if we get here then we're good to go. */
1542 ocfs2_inc_holders(lockres, level);
1543
1544 ret = 0;
1545unlock:
a1912826
SM		 1546 lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
1547
ccd979bd
MF		 1548 spin_unlock_irqrestore(&lockres->l_lock, flags);
1549out:
1550 /*
1551 * This is helping work around a lock inversion between the page lock
1552 * and dlm locks. One path holds the page lock while calling aops
1553 * which block acquiring dlm locks. The voting thread holds dlm
1554 * locks while acquiring page locks while down converting data locks.
1555 * This block is helping an aop path notice the inversion and back
1556 * off to unlock its page lock before trying the dlm lock again.
1557 */
1558 if (wait && arg_flags & OCFS2_LOCK_NONBLOCK &&
1559 mw.mw_mask & (OCFS2_LOCK_BUSY|OCFS2_LOCK_BLOCKED)) {
1560 wait = 0;
1561 if (lockres_remove_mask_waiter(lockres, &mw))
1562 ret = -EAGAIN;
1563 else
1564 goto again;
1565 }
1566 if (wait) {
1567 ret = ocfs2_wait_for_mask(&mw);
1568 if (ret == 0)
1569 goto again;
1570 mlog_errno(ret);
1571 }
8ddb7b00 1572 ocfs2_update_lock_stats(lockres, level, &mw, ret);
ccd979bd 1573
cb25797d
JK		 1574#ifdef CONFIG_DEBUG_LOCK_ALLOC
1575 if (!ret && lockres->l_lockdep_map.key != NULL) {
1576 if (level == DLM_LOCK_PR)
1577 rwsem_acquire_read(&lockres->l_lockdep_map, l_subclass,
1578 !!(arg_flags & OCFS2_META_LOCK_NOQUEUE),
1579 caller_ip);
1580 else
1581 rwsem_acquire(&lockres->l_lockdep_map, l_subclass,
1582 !!(arg_flags & OCFS2_META_LOCK_NOQUEUE),
1583 caller_ip);
1584 }
1585#endif
ccd979bd
MF		 1586 mlog_exit(ret);
1587 return ret;
1588}
1589
cb25797d
JK		 1590static inline int ocfs2_cluster_lock(struct ocfs2_super *osb,
1591 struct ocfs2_lock_res *lockres,
1592 int level,
1593 u32 lkm_flags,
1594 int arg_flags)
1595{
1596 return __ocfs2_cluster_lock(osb, lockres, level, lkm_flags, arg_flags,
1597 0, _RET_IP_);
1598}
1599
1600
1601static void __ocfs2_cluster_unlock(struct ocfs2_super *osb,
1602 struct ocfs2_lock_res *lockres,
1603 int level,
1604 unsigned long caller_ip)
ccd979bd
MF		 1605{
1606 unsigned long flags;
1607
1608 mlog_entry_void();
1609 spin_lock_irqsave(&lockres->l_lock, flags);
1610 ocfs2_dec_holders(lockres, level);
34d024f8 1611 ocfs2_downconvert_on_unlock(osb, lockres);
ccd979bd 1612 spin_unlock_irqrestore(&lockres->l_lock, flags);
cb25797d
JK		 1613#ifdef CONFIG_DEBUG_LOCK_ALLOC
1614 if (lockres->l_lockdep_map.key != NULL)
1615 rwsem_release(&lockres->l_lockdep_map, 1, caller_ip);
1616#endif
ccd979bd
MF		 1617 mlog_exit_void();
1618}
1619
da66116e
AB		 1620static int ocfs2_create_new_lock(struct ocfs2_super *osb,
1621 struct ocfs2_lock_res *lockres,
1622 int ex,
1623 int local)
ccd979bd 1624{
bd3e7610 1625 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
ccd979bd 1626 unsigned long flags;
bd3e7610 1627 u32 lkm_flags = local ? DLM_LKF_LOCAL : 0;
ccd979bd
MF		 1628
1629 spin_lock_irqsave(&lockres->l_lock, flags);
1630 BUG_ON(lockres->l_flags & OCFS2_LOCK_ATTACHED);
1631 lockres_or_flags(lockres, OCFS2_LOCK_LOCAL);
1632 spin_unlock_irqrestore(&lockres->l_lock, flags);
1633
24c19ef4 1634 return ocfs2_lock_create(osb, lockres, level, lkm_flags);
ccd979bd
MF		 1635}
1636
1637/* Grants us an EX lock on the data and metadata resources, skipping
1638 * the normal cluster directory lookup. Use this ONLY on newly created
1639 * inodes which other nodes can't possibly see, and which haven't been
1640 * hashed in the inode hash yet. This can give us a good performance
1641 * increase as it'll skip the network broadcast normally associated
1642 * with creating a new lock resource. */
1643int ocfs2_create_new_inode_locks(struct inode *inode)
1644{
1645 int ret;
d680efe9 1646 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
ccd979bd
MF		 1647
1648 BUG_ON(!inode);
1649 BUG_ON(!ocfs2_inode_is_new(inode));
1650
1651 mlog_entry_void();
1652
b0697053 1653 mlog(0, "Inode %llu\n", (unsigned long long)OCFS2_I(inode)->ip_blkno);
ccd979bd
MF		 1654
1655 /* NOTE: That we don't increment any of the holder counts, nor
1656 * do we add anything to a journal handle. Since this is
1657 * supposed to be a new inode which the cluster doesn't know
1658 * about yet, there is no need to. As far as the LVB handling
1659 * is concerned, this is basically like acquiring an EX lock
1660 * on a resource which has an invalid one -- we'll set it
1661 * valid when we release the EX. */
1662
24c19ef4 1663 ret = ocfs2_create_new_lock(osb, &OCFS2_I(inode)->ip_rw_lockres, 1, 1);
ccd979bd
MF		 1664 if (ret) {
1665 mlog_errno(ret);
1666 goto bail;
1667 }
1668
24c19ef4 1669 /*
bd3e7610 1670 * We don't want to use DLM_LKF_LOCAL on a meta data lock as they
24c19ef4
MF		 1671 * don't use a generation in their lock names.
1672 */
e63aecb6 1673 ret = ocfs2_create_new_lock(osb, &OCFS2_I(inode)->ip_inode_lockres, 1, 0);
ccd979bd
MF		 1674 if (ret) {
1675 mlog_errno(ret);
1676 goto bail;
1677 }
1678
50008630
TY		 1679 ret = ocfs2_create_new_lock(osb, &OCFS2_I(inode)->ip_open_lockres, 0, 0);
1680 if (ret) {
1681 mlog_errno(ret);
1682 goto bail;
1683 }
1684
ccd979bd
MF		 1685bail:
1686 mlog_exit(ret);
1687 return ret;
1688}
1689
1690int ocfs2_rw_lock(struct inode *inode, int write)
1691{
1692 int status, level;
1693 struct ocfs2_lock_res *lockres;
c271c5c2 1694 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
ccd979bd
MF		 1695
1696 BUG_ON(!inode);
1697
1698 mlog_entry_void();
1699
b0697053
MF		 1700 mlog(0, "inode %llu take %s RW lock\n",
1701 (unsigned long long)OCFS2_I(inode)->ip_blkno,
ccd979bd
MF		 1702 write ? "EXMODE" : "PRMODE");
1703
d92bc512
CL		 1704 if (ocfs2_mount_local(osb)) {
1705 mlog_exit(0);
c271c5c2 1706 return 0;
d92bc512 1707 }
c271c5c2 1708
ccd979bd
MF		 1709 lockres = &OCFS2_I(inode)->ip_rw_lockres;
1710
bd3e7610 1711 level = write ? DLM_LOCK_EX : DLM_LOCK_PR;
ccd979bd
MF		 1712
1713 status = ocfs2_cluster_lock(OCFS2_SB(inode->i_sb), lockres, level, 0,
1714 0);
1715 if (status < 0)
1716 mlog_errno(status);
1717
1718 mlog_exit(status);
1719 return status;
1720}
1721
1722void ocfs2_rw_unlock(struct inode *inode, int write)
1723{
bd3e7610 1724 int level = write ? DLM_LOCK_EX : DLM_LOCK_PR;
ccd979bd 1725 struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_rw_lockres;
c271c5c2 1726 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
ccd979bd
MF		 1727
1728 mlog_entry_void();
1729
b0697053
MF		 1730 mlog(0, "inode %llu drop %s RW lock\n",
1731 (unsigned long long)OCFS2_I(inode)->ip_blkno,
ccd979bd
MF		 1732 write ? "EXMODE" : "PRMODE");
1733
c271c5c2
SM		 1734 if (!ocfs2_mount_local(osb))
1735 ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, level);
ccd979bd
MF		 1736
1737 mlog_exit_void();
1738}
1739
50008630
TY		 1740/*
1741 * ocfs2_open_lock always get PR mode lock.
1742 */
1743int ocfs2_open_lock(struct inode *inode)
1744{
1745 int status = 0;
1746 struct ocfs2_lock_res *lockres;
1747 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1748
1749 BUG_ON(!inode);
1750
1751 mlog_entry_void();
1752
1753 mlog(0, "inode %llu take PRMODE open lock\n",
1754 (unsigned long long)OCFS2_I(inode)->ip_blkno);
1755
1756 if (ocfs2_mount_local(osb))
1757 goto out;
1758
1759 lockres = &OCFS2_I(inode)->ip_open_lockres;
1760
1761 status = ocfs2_cluster_lock(OCFS2_SB(inode->i_sb), lockres,
bd3e7610 1762 DLM_LOCK_PR, 0, 0);
50008630
TY		 1763 if (status < 0)
1764 mlog_errno(status);
1765
1766out:
1767 mlog_exit(status);
1768 return status;
1769}
1770
1771int ocfs2_try_open_lock(struct inode *inode, int write)
1772{
1773 int status = 0, level;
1774 struct ocfs2_lock_res *lockres;
1775 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1776
1777 BUG_ON(!inode);
1778
1779 mlog_entry_void();
1780
1781 mlog(0, "inode %llu try to take %s open lock\n",
1782 (unsigned long long)OCFS2_I(inode)->ip_blkno,
1783 write ? "EXMODE" : "PRMODE");
1784
1785 if (ocfs2_mount_local(osb))
1786 goto out;
1787
1788 lockres = &OCFS2_I(inode)->ip_open_lockres;
1789
bd3e7610 1790 level = write ? DLM_LOCK_EX : DLM_LOCK_PR;
50008630
TY		 1791
1792 /*
1793 * The file system may already holding a PRMODE/EXMODE open lock.
bd3e7610 1794 * Since we pass DLM_LKF_NOQUEUE, the request won't block waiting on
50008630
TY		 1795 * other nodes and the -EAGAIN will indicate to the caller that
1796 * this inode is still in use.
1797 */
1798 status = ocfs2_cluster_lock(OCFS2_SB(inode->i_sb), lockres,
bd3e7610 1799 level, DLM_LKF_NOQUEUE, 0);
50008630
TY		 1800
1801out:
1802 mlog_exit(status);
1803 return status;
1804}
1805
1806/*
1807 * ocfs2_open_unlock unlock PR and EX mode open locks.
1808 */
1809void ocfs2_open_unlock(struct inode *inode)
1810{
1811 struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_open_lockres;
1812 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1813
1814 mlog_entry_void();
1815
1816 mlog(0, "inode %llu drop open lock\n",
1817 (unsigned long long)OCFS2_I(inode)->ip_blkno);
1818
1819 if (ocfs2_mount_local(osb))
1820 goto out;
1821
1822 if(lockres->l_ro_holders)
1823 ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres,
bd3e7610 1824 DLM_LOCK_PR);
50008630
TY		 1825 if(lockres->l_ex_holders)
1826 ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres,
bd3e7610 1827 DLM_LOCK_EX);
50008630
TY		 1828
1829out:
1830 mlog_exit_void();
1831}
1832
cf8e06f1
MF		 1833static int ocfs2_flock_handle_signal(struct ocfs2_lock_res *lockres,
1834 int level)
1835{
1836 int ret;
1837 struct ocfs2_super *osb = ocfs2_get_lockres_osb(lockres);
1838 unsigned long flags;
1839 struct ocfs2_mask_waiter mw;
1840
1841 ocfs2_init_mask_waiter(&mw);
1842
1843retry_cancel:
1844 spin_lock_irqsave(&lockres->l_lock, flags);
1845 if (lockres->l_flags & OCFS2_LOCK_BUSY) {
1846 ret = ocfs2_prepare_cancel_convert(osb, lockres);
1847 if (ret) {
1848 spin_unlock_irqrestore(&lockres->l_lock, flags);
1849 ret = ocfs2_cancel_convert(osb, lockres);
1850 if (ret < 0) {
1851 mlog_errno(ret);
1852 goto out;
1853 }
1854 goto retry_cancel;
1855 }
1856 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0);
1857 spin_unlock_irqrestore(&lockres->l_lock, flags);
1858
1859 ocfs2_wait_for_mask(&mw);
1860 goto retry_cancel;
1861 }
1862
1863 ret = -ERESTARTSYS;
1864 /*
1865 * We may still have gotten the lock, in which case there's no
1866 * point to restarting the syscall.
1867 */
1868 if (lockres->l_level == level)
1869 ret = 0;
1870
1871 mlog(0, "Cancel returning %d. flags: 0x%lx, level: %d, act: %d\n", ret,
1872 lockres->l_flags, lockres->l_level, lockres->l_action);
1873
1874 spin_unlock_irqrestore(&lockres->l_lock, flags);
1875
1876out:
1877 return ret;
1878}
1879
1880/*
1881 * ocfs2_file_lock() and ocfs2_file_unlock() map to a single pair of
1882 * flock() calls. The locking approach this requires is sufficiently
1883 * different from all other cluster lock types that we implement a
3ad2f3fb 1884 * separate path to the "low-level" dlm calls. In particular:
cf8e06f1
MF		 1885 *
1886 * - No optimization of lock levels is done - we take at exactly
1887 * what's been requested.
1888 *
1889 * - No lock caching is employed. We immediately downconvert to
1890 * no-lock at unlock time. This also means flock locks never go on
1891 * the blocking list).
1892 *
1893 * - Since userspace can trivially deadlock itself with flock, we make
1894 * sure to allow cancellation of a misbehaving applications flock()
1895 * request.
1896 *
1897 * - Access to any flock lockres doesn't require concurrency, so we
1898 * can simplify the code by requiring the caller to guarantee
1899 * serialization of dlmglue flock calls.
1900 */
1901int ocfs2_file_lock(struct file *file, int ex, int trylock)
1902{
e988cf1c
MF		 1903 int ret, level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
1904 unsigned int lkm_flags = trylock ? DLM_LKF_NOQUEUE : 0;
cf8e06f1
MF		 1905 unsigned long flags;
1906 struct ocfs2_file_private *fp = file->private_data;
1907 struct ocfs2_lock_res *lockres = &fp->fp_flock;
1908 struct ocfs2_super *osb = OCFS2_SB(file->f_mapping->host->i_sb);
1909 struct ocfs2_mask_waiter mw;
1910
1911 ocfs2_init_mask_waiter(&mw);
1912
1913 if ((lockres->l_flags & OCFS2_LOCK_BUSY) ||
bd3e7610 1914 (lockres->l_level > DLM_LOCK_NL)) {
cf8e06f1
MF		 1915 mlog(ML_ERROR,
1916 "File lock \"%s\" has busy or locked state: flags: 0x%lx, "
1917 "level: %u\n", lockres->l_name, lockres->l_flags,
1918 lockres->l_level);
1919 return -EINVAL;
1920 }
1921
1922 spin_lock_irqsave(&lockres->l_lock, flags);
1923 if (!(lockres->l_flags & OCFS2_LOCK_ATTACHED)) {
1924 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0);
1925 spin_unlock_irqrestore(&lockres->l_lock, flags);
1926
1927 /*
1928 * Get the lock at NLMODE to start - that way we
1929 * can cancel the upconvert request if need be.
1930 */
e988cf1c 1931 ret = ocfs2_lock_create(osb, lockres, DLM_LOCK_NL, 0);
cf8e06f1
MF		 1932 if (ret < 0) {
1933 mlog_errno(ret);
1934 goto out;
1935 }
1936
1937 ret = ocfs2_wait_for_mask(&mw);
1938 if (ret) {
1939 mlog_errno(ret);
1940 goto out;
1941 }
1942 spin_lock_irqsave(&lockres->l_lock, flags);
1943 }
1944
1945 lockres->l_action = OCFS2_AST_CONVERT;
e988cf1c 1946 lkm_flags |= DLM_LKF_CONVERT;
cf8e06f1
MF		 1947 lockres->l_requested = level;
1948 lockres_or_flags(lockres, OCFS2_LOCK_BUSY);
1949
1950 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0);
1951 spin_unlock_irqrestore(&lockres->l_lock, flags);
1952
4670c46d 1953 ret = ocfs2_dlm_lock(osb->cconn, level, &lockres->l_lksb, lkm_flags,
a796d286 1954 lockres->l_name, OCFS2_LOCK_ID_MAX_LEN - 1);
7431cd7e
JB		 1955 if (ret) {
1956 if (!trylock || (ret != -EAGAIN)) {
24ef1815 1957 ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres);
cf8e06f1
MF		 1958 ret = -EINVAL;
1959 }
1960
1961 ocfs2_recover_from_dlm_error(lockres, 1);
1962 lockres_remove_mask_waiter(lockres, &mw);
1963 goto out;
1964 }
1965
1966 ret = ocfs2_wait_for_mask_interruptible(&mw, lockres);
1967 if (ret == -ERESTARTSYS) {
1968 /*
1969 * Userspace can cause deadlock itself with
1970 * flock(). Current behavior locally is to allow the
1971 * deadlock, but abort the system call if a signal is
1972 * received. We follow this example, otherwise a
1973 * poorly written program could sit in kernel until
1974 * reboot.
1975 *
1976 * Handling this is a bit more complicated for Ocfs2
1977 * though. We can't exit this function with an
1978 * outstanding lock request, so a cancel convert is
1979 * required. We intentionally overwrite 'ret' - if the
1980 * cancel fails and the lock was granted, it's easier
af901ca1 1981 * to just bubble success back up to the user.
cf8e06f1
MF		 1982 */
1983 ret = ocfs2_flock_handle_signal(lockres, level);
1693a5c0
DT		 1984 } else if (!ret && (level > lockres->l_level)) {
1985 /* Trylock failed asynchronously */
1986 BUG_ON(!trylock);
1987 ret = -EAGAIN;
cf8e06f1
MF		 1988 }
1989
1990out:
1991
1992 mlog(0, "Lock: \"%s\" ex: %d, trylock: %d, returns: %d\n",
1993 lockres->l_name, ex, trylock, ret);
1994 return ret;
1995}
1996
1997void ocfs2_file_unlock(struct file *file)
1998{
1999 int ret;
de551246 2000 unsigned int gen;
cf8e06f1
MF		 2001 unsigned long flags;
2002 struct ocfs2_file_private *fp = file->private_data;
2003 struct ocfs2_lock_res *lockres = &fp->fp_flock;
2004 struct ocfs2_super *osb = OCFS2_SB(file->f_mapping->host->i_sb);
2005 struct ocfs2_mask_waiter mw;
2006
2007 ocfs2_init_mask_waiter(&mw);
2008
2009 if (!(lockres->l_flags & OCFS2_LOCK_ATTACHED))
2010 return;
2011
e988cf1c 2012 if (lockres->l_level == DLM_LOCK_NL)
cf8e06f1
MF		 2013 return;
2014
2015 mlog(0, "Unlock: \"%s\" flags: 0x%lx, level: %d, act: %d\n",
2016 lockres->l_name, lockres->l_flags, lockres->l_level,
2017 lockres->l_action);
2018
2019 spin_lock_irqsave(&lockres->l_lock, flags);
2020 /*
2021 * Fake a blocking ast for the downconvert code.
2022 */
2023 lockres_or_flags(lockres, OCFS2_LOCK_BLOCKED);
bd3e7610 2024 lockres->l_blocking = DLM_LOCK_EX;
cf8e06f1 2025
e988cf1c 2026 gen = ocfs2_prepare_downconvert(lockres, DLM_LOCK_NL);
cf8e06f1
MF		 2027 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0);
2028 spin_unlock_irqrestore(&lockres->l_lock, flags);
2029
e988cf1c 2030 ret = ocfs2_downconvert_lock(osb, lockres, DLM_LOCK_NL, 0, gen);
cf8e06f1
MF		 2031 if (ret) {
2032 mlog_errno(ret);
2033 return;
2034 }
2035
2036 ret = ocfs2_wait_for_mask(&mw);
2037 if (ret)
2038 mlog_errno(ret);
2039}
2040
34d024f8
MF		 2041static void ocfs2_downconvert_on_unlock(struct ocfs2_super *osb,
2042 struct ocfs2_lock_res *lockres)
ccd979bd
MF		 2043{
2044 int kick = 0;
2045
2046 mlog_entry_void();
2047
2048 /* If we know that another node is waiting on our lock, kick
34d024f8 2049 * the downconvert thread * pre-emptively when we reach a release
ccd979bd
MF		 2050 * condition. */
2051 if (lockres->l_flags & OCFS2_LOCK_BLOCKED) {
2052 switch(lockres->l_blocking) {
bd3e7610 2053 case DLM_LOCK_EX:
ccd979bd
MF		 2054 if (!lockres->l_ex_holders && !lockres->l_ro_holders)
2055 kick = 1;
2056 break;
bd3e7610 2057 case DLM_LOCK_PR:
ccd979bd
MF		 2058 if (!lockres->l_ex_holders)
2059 kick = 1;
2060 break;
2061 default:
2062 BUG();
2063 }
2064 }
2065
2066 if (kick)
34d024f8 2067 ocfs2_wake_downconvert_thread(osb);
ccd979bd
MF		 2068
2069 mlog_exit_void();
2070}
2071
ccd979bd
MF		 2072#define OCFS2_SEC_BITS 34
2073#define OCFS2_SEC_SHIFT (64 - 34)
2074#define OCFS2_NSEC_MASK ((1ULL << OCFS2_SEC_SHIFT) - 1)
2075
2076/* LVB only has room for 64 bits of time here so we pack it for
2077 * now. */
2078static u64 ocfs2_pack_timespec(struct timespec *spec)
2079{
2080 u64 res;
2081 u64 sec = spec->tv_sec;
2082 u32 nsec = spec->tv_nsec;
2083
2084 res = (sec << OCFS2_SEC_SHIFT) | (nsec & OCFS2_NSEC_MASK);
2085
2086 return res;
2087}
2088
2089/* Call this with the lockres locked. I am reasonably sure we don't
2090 * need ip_lock in this function as anyone who would be changing those
e63aecb6 2091 * values is supposed to be blocked in ocfs2_inode_lock right now. */
ccd979bd
MF		 2092static void __ocfs2_stuff_meta_lvb(struct inode *inode)
2093{
2094 struct ocfs2_inode_info *oi = OCFS2_I(inode);
e63aecb6 2095 struct ocfs2_lock_res *lockres = &oi->ip_inode_lockres;
ccd979bd
MF		 2096 struct ocfs2_meta_lvb *lvb;
2097
2098 mlog_entry_void();
2099
a641dc2a 2100 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
ccd979bd 2101
24c19ef4
MF		 2102 /*
2103 * Invalidate the LVB of a deleted inode - this way other
2104 * nodes are forced to go to disk and discover the new inode
2105 * status.
2106 */
2107 if (oi->ip_flags & OCFS2_INODE_DELETED) {
2108 lvb->lvb_version = 0;
2109 goto out;
2110 }
2111
4d3b83f7 2112 lvb->lvb_version = OCFS2_LVB_VERSION;
ccd979bd
MF		 2113 lvb->lvb_isize = cpu_to_be64(i_size_read(inode));
2114 lvb->lvb_iclusters = cpu_to_be32(oi->ip_clusters);
2115 lvb->lvb_iuid = cpu_to_be32(inode->i_uid);
2116 lvb->lvb_igid = cpu_to_be32(inode->i_gid);
2117 lvb->lvb_imode = cpu_to_be16(inode->i_mode);
2118 lvb->lvb_inlink = cpu_to_be16(inode->i_nlink);
2119 lvb->lvb_iatime_packed =
2120 cpu_to_be64(ocfs2_pack_timespec(&inode->i_atime));
2121 lvb->lvb_ictime_packed =
2122 cpu_to_be64(ocfs2_pack_timespec(&inode->i_ctime));
2123 lvb->lvb_imtime_packed =
2124 cpu_to_be64(ocfs2_pack_timespec(&inode->i_mtime));
ca4d147e 2125 lvb->lvb_iattr = cpu_to_be32(oi->ip_attr);
15b1e36b 2126 lvb->lvb_idynfeatures = cpu_to_be16(oi->ip_dyn_features);
f9e2d82e 2127 lvb->lvb_igeneration = cpu_to_be32(inode->i_generation);
ccd979bd 2128
24c19ef4 2129out:
ccd979bd
MF		 2130 mlog_meta_lvb(0, lockres);
2131
2132 mlog_exit_void();
2133}
2134
2135static void ocfs2_unpack_timespec(struct timespec *spec,
2136 u64 packed_time)
2137{
2138 spec->tv_sec = packed_time >> OCFS2_SEC_SHIFT;
2139 spec->tv_nsec = packed_time & OCFS2_NSEC_MASK;
2140}
2141
2142static void ocfs2_refresh_inode_from_lvb(struct inode *inode)
2143{
2144 struct ocfs2_inode_info *oi = OCFS2_I(inode);
e63aecb6 2145 struct ocfs2_lock_res *lockres = &oi->ip_inode_lockres;
ccd979bd
MF		 2146 struct ocfs2_meta_lvb *lvb;
2147
2148 mlog_entry_void();
2149
2150 mlog_meta_lvb(0, lockres);
2151
a641dc2a 2152 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
ccd979bd
MF		 2153
2154 /* We're safe here without the lockres lock... */
2155 spin_lock(&oi->ip_lock);
2156 oi->ip_clusters = be32_to_cpu(lvb->lvb_iclusters);
2157 i_size_write(inode, be64_to_cpu(lvb->lvb_isize));
2158
ca4d147e 2159 oi->ip_attr = be32_to_cpu(lvb->lvb_iattr);
15b1e36b 2160 oi->ip_dyn_features = be16_to_cpu(lvb->lvb_idynfeatures);
ca4d147e
HP		 2161 ocfs2_set_inode_flags(inode);
2162
ccd979bd
MF		 2163 /* fast-symlinks are a special case */
2164 if (S_ISLNK(inode->i_mode) && !oi->ip_clusters)
2165 inode->i_blocks = 0;
2166 else
8110b073 2167 inode->i_blocks = ocfs2_inode_sector_count(inode);
ccd979bd
MF		 2168
2169 inode->i_uid = be32_to_cpu(lvb->lvb_iuid);
2170 inode->i_gid = be32_to_cpu(lvb->lvb_igid);
2171 inode->i_mode = be16_to_cpu(lvb->lvb_imode);
2172 inode->i_nlink = be16_to_cpu(lvb->lvb_inlink);
2173 ocfs2_unpack_timespec(&inode->i_atime,
2174 be64_to_cpu(lvb->lvb_iatime_packed));
2175 ocfs2_unpack_timespec(&inode->i_mtime,
2176 be64_to_cpu(lvb->lvb_imtime_packed));
2177 ocfs2_unpack_timespec(&inode->i_ctime,
2178 be64_to_cpu(lvb->lvb_ictime_packed));
2179 spin_unlock(&oi->ip_lock);
2180
2181 mlog_exit_void();
2182}
2183
f9e2d82e
MF		 2184static inline int ocfs2_meta_lvb_is_trustable(struct inode *inode,
2185 struct ocfs2_lock_res *lockres)
ccd979bd 2186{
a641dc2a 2187 struct ocfs2_meta_lvb *lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
ccd979bd 2188
1c520dfb
JB		 2189 if (ocfs2_dlm_lvb_valid(&lockres->l_lksb)
2190 && lvb->lvb_version == OCFS2_LVB_VERSION
f9e2d82e 2191 && be32_to_cpu(lvb->lvb_igeneration) == inode->i_generation)
ccd979bd
MF		 2192 return 1;
2193 return 0;
2194}
2195
2196/* Determine whether a lock resource needs to be refreshed, and
2197 * arbitrate who gets to refresh it.
2198 *
2199 * 0 means no refresh needed.
2200 *
2201 * > 0 means you need to refresh this and you MUST call
2202 * ocfs2_complete_lock_res_refresh afterwards. */
2203static int ocfs2_should_refresh_lock_res(struct ocfs2_lock_res *lockres)
2204{
2205 unsigned long flags;
2206 int status = 0;
2207
2208 mlog_entry_void();
2209
2210refresh_check:
2211 spin_lock_irqsave(&lockres->l_lock, flags);
2212 if (!(lockres->l_flags & OCFS2_LOCK_NEEDS_REFRESH)) {
2213 spin_unlock_irqrestore(&lockres->l_lock, flags);
2214 goto bail;
2215 }
2216
2217 if (lockres->l_flags & OCFS2_LOCK_REFRESHING) {
2218 spin_unlock_irqrestore(&lockres->l_lock, flags);
2219
2220 ocfs2_wait_on_refreshing_lock(lockres);
2221 goto refresh_check;
2222 }
2223
2224 /* Ok, I'll be the one to refresh this lock. */
2225 lockres_or_flags(lockres, OCFS2_LOCK_REFRESHING);
2226 spin_unlock_irqrestore(&lockres->l_lock, flags);
2227
2228 status = 1;
2229bail:
2230 mlog_exit(status);
2231 return status;
2232}
2233
2234/* If status is non zero, I'll mark it as not being in refresh
2235 * anymroe, but i won't clear the needs refresh flag. */
2236static inline void ocfs2_complete_lock_res_refresh(struct ocfs2_lock_res *lockres,
2237 int status)
2238{
2239 unsigned long flags;
2240 mlog_entry_void();
2241
2242 spin_lock_irqsave(&lockres->l_lock, flags);
2243 lockres_clear_flags(lockres, OCFS2_LOCK_REFRESHING);
2244 if (!status)
2245 lockres_clear_flags(lockres, OCFS2_LOCK_NEEDS_REFRESH);
2246 spin_unlock_irqrestore(&lockres->l_lock, flags);
2247
2248 wake_up(&lockres->l_event);
2249
2250 mlog_exit_void();
2251}
2252
2253/* may or may not return a bh if it went to disk. */
e63aecb6 2254static int ocfs2_inode_lock_update(struct inode *inode,
ccd979bd
MF		 2255 struct buffer_head **bh)
2256{
2257 int status = 0;
2258 struct ocfs2_inode_info *oi = OCFS2_I(inode);
e63aecb6 2259 struct ocfs2_lock_res *lockres = &oi->ip_inode_lockres;
ccd979bd 2260 struct ocfs2_dinode *fe;
c271c5c2 2261 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
ccd979bd
MF		 2262
2263 mlog_entry_void();
2264
be9e986b
MF		 2265 if (ocfs2_mount_local(osb))
2266 goto bail;
2267
ccd979bd
MF		 2268 spin_lock(&oi->ip_lock);
2269 if (oi->ip_flags & OCFS2_INODE_DELETED) {
b0697053 2270 mlog(0, "Orphaned inode %llu was deleted while we "
ccd979bd 2271 "were waiting on a lock. ip_flags = 0x%x\n",
b0697053 2272 (unsigned long long)oi->ip_blkno, oi->ip_flags);
ccd979bd
MF		 2273 spin_unlock(&oi->ip_lock);
2274 status = -ENOENT;
2275 goto bail;
2276 }
2277 spin_unlock(&oi->ip_lock);
2278
be9e986b
MF		 2279 if (!ocfs2_should_refresh_lock_res(lockres))
2280 goto bail;
ccd979bd
MF		 2281
2282 /* This will discard any caching information we might have had
2283 * for the inode metadata. */
8cb471e8 2284 ocfs2_metadata_cache_purge(INODE_CACHE(inode));
ccd979bd 2285
83418978
MF		 2286 ocfs2_extent_map_trunc(inode, 0);
2287
be9e986b 2288 if (ocfs2_meta_lvb_is_trustable(inode, lockres)) {
b0697053
MF		 2289 mlog(0, "Trusting LVB on inode %llu\n",
2290 (unsigned long long)oi->ip_blkno);
ccd979bd
MF		 2291 ocfs2_refresh_inode_from_lvb(inode);
2292 } else {
2293 /* Boo, we have to go to disk. */
2294 /* read bh, cast, ocfs2_refresh_inode */
b657c95c 2295 status = ocfs2_read_inode_block(inode, bh);
ccd979bd
MF		 2296 if (status < 0) {
2297 mlog_errno(status);
2298 goto bail_refresh;
2299 }
2300 fe = (struct ocfs2_dinode *) (*bh)->b_data;
2301
2302 /* This is a good chance to make sure we're not
b657c95c
JB		 2303 * locking an invalid object. ocfs2_read_inode_block()
2304 * already checked that the inode block is sane.
ccd979bd
MF		 2305 *
2306 * We bug on a stale inode here because we checked
2307 * above whether it was wiped from disk. The wiping
2308 * node provides a guarantee that we receive that
2309 * message and can mark the inode before dropping any
2310 * locks associated with it. */
ccd979bd
MF		 2311 mlog_bug_on_msg(inode->i_generation !=
2312 le32_to_cpu(fe->i_generation),
b0697053 2313 "Invalid dinode %llu disk generation: %u "
ccd979bd 2314 "inode->i_generation: %u\n",
b0697053
MF		 2315 (unsigned long long)oi->ip_blkno,
2316 le32_to_cpu(fe->i_generation),
ccd979bd
MF		 2317 inode->i_generation);
2318 mlog_bug_on_msg(le64_to_cpu(fe->i_dtime) ||
2319 !(fe->i_flags & cpu_to_le32(OCFS2_VALID_FL)),
b0697053
MF		 2320 "Stale dinode %llu dtime: %llu flags: 0x%x\n",
2321 (unsigned long long)oi->ip_blkno,
2322 (unsigned long long)le64_to_cpu(fe->i_dtime),
ccd979bd
MF		 2323 le32_to_cpu(fe->i_flags));
2324
2325 ocfs2_refresh_inode(inode, fe);
8ddb7b00 2326 ocfs2_track_lock_refresh(lockres);
ccd979bd
MF		 2327 }
2328
2329 status = 0;
2330bail_refresh:
be9e986b 2331 ocfs2_complete_lock_res_refresh(lockres, status);
ccd979bd
MF		 2332bail:
2333 mlog_exit(status);
2334 return status;
2335}
2336
2337static int ocfs2_assign_bh(struct inode *inode,
2338 struct buffer_head **ret_bh,
2339 struct buffer_head *passed_bh)
2340{
2341 int status;
2342
2343 if (passed_bh) {
2344 /* Ok, the update went to disk for us, use the
2345 * returned bh. */
2346 *ret_bh = passed_bh;
2347 get_bh(*ret_bh);
2348
2349 return 0;
2350 }
2351
b657c95c 2352 status = ocfs2_read_inode_block(inode, ret_bh);
ccd979bd
MF		 2353 if (status < 0)
2354 mlog_errno(status);
2355
2356 return status;
2357}
2358
2359/*
2360 * returns < 0 error if the callback will never be called, otherwise
2361 * the result of the lock will be communicated via the callback.
2362 */
cb25797d
JK		 2363int ocfs2_inode_lock_full_nested(struct inode *inode,
2364 struct buffer_head **ret_bh,
2365 int ex,
2366 int arg_flags,
2367 int subclass)
ccd979bd 2368{
bd3e7610
JB		 2369 int status, level, acquired;
2370 u32 dlm_flags;
c271c5c2 2371 struct ocfs2_lock_res *lockres = NULL;
ccd979bd
MF		 2372 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
2373 struct buffer_head *local_bh = NULL;
2374
2375 BUG_ON(!inode);
2376
2377 mlog_entry_void();
2378
b0697053
MF		 2379 mlog(0, "inode %llu, take %s META lock\n",
2380 (unsigned long long)OCFS2_I(inode)->ip_blkno,
ccd979bd
MF		 2381 ex ? "EXMODE" : "PRMODE");
2382
2383 status = 0;
2384 acquired = 0;
2385 /* We'll allow faking a readonly metadata lock for
2386 * rodevices. */
2387 if (ocfs2_is_hard_readonly(osb)) {
2388 if (ex)
2389 status = -EROFS;
2390 goto bail;
2391 }
2392
c271c5c2
SM		 2393 if (ocfs2_mount_local(osb))
2394 goto local;
2395
ccd979bd 2396 if (!(arg_flags & OCFS2_META_LOCK_RECOVERY))
553abd04 2397 ocfs2_wait_for_recovery(osb);
ccd979bd 2398
e63aecb6 2399 lockres = &OCFS2_I(inode)->ip_inode_lockres;
bd3e7610 2400 level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
ccd979bd
MF		 2401 dlm_flags = 0;
2402 if (arg_flags & OCFS2_META_LOCK_NOQUEUE)
bd3e7610 2403 dlm_flags |= DLM_LKF_NOQUEUE;
ccd979bd 2404
cb25797d
JK		 2405 status = __ocfs2_cluster_lock(osb, lockres, level, dlm_flags,
2406 arg_flags, subclass, _RET_IP_);
ccd979bd
MF		 2407 if (status < 0) {
2408 if (status != -EAGAIN && status != -EIOCBRETRY)
2409 mlog_errno(status);
2410 goto bail;
2411 }
2412
2413 /* Notify the error cleanup path to drop the cluster lock. */
2414 acquired = 1;
2415
2416 /* We wait twice because a node may have died while we were in
2417 * the lower dlm layers. The second time though, we've
2418 * committed to owning this lock so we don't allow signals to
2419 * abort the operation. */
2420 if (!(arg_flags & OCFS2_META_LOCK_RECOVERY))
553abd04 2421 ocfs2_wait_for_recovery(osb);
ccd979bd 2422
c271c5c2 2423local:
24c19ef4
MF		 2424 /*
2425 * We only see this flag if we're being called from
2426 * ocfs2_read_locked_inode(). It means we're locking an inode
2427 * which hasn't been populated yet, so clear the refresh flag
2428 * and let the caller handle it.
2429 */
2430 if (inode->i_state & I_NEW) {
2431 status = 0;
c271c5c2
SM		 2432 if (lockres)
2433 ocfs2_complete_lock_res_refresh(lockres, 0);
24c19ef4
MF		 2434 goto bail;
2435 }
2436
ccd979bd 2437 /* This is fun. The caller may want a bh back, or it may
e63aecb6 2438 * not. ocfs2_inode_lock_update definitely wants one in, but
ccd979bd
MF		 2439 * may or may not read one, depending on what's in the
2440 * LVB. The result of all of this is that we've *only* gone to
2441 * disk if we have to, so the complexity is worthwhile. */
e63aecb6 2442 status = ocfs2_inode_lock_update(inode, &local_bh);
ccd979bd
MF		 2443 if (status < 0) {
2444 if (status != -ENOENT)
2445 mlog_errno(status);
2446 goto bail;
2447 }
2448
2449 if (ret_bh) {
2450 status = ocfs2_assign_bh(inode, ret_bh, local_bh);
2451 if (status < 0) {
2452 mlog_errno(status);
2453 goto bail;
2454 }
2455 }
2456
ccd979bd
MF		 2457bail:
2458 if (status < 0) {
2459 if (ret_bh && (*ret_bh)) {
2460 brelse(*ret_bh);
2461 *ret_bh = NULL;
2462 }
2463 if (acquired)
e63aecb6 2464 ocfs2_inode_unlock(inode, ex);
ccd979bd
MF		 2465 }
2466
2467 if (local_bh)
2468 brelse(local_bh);
2469
2470 mlog_exit(status);
2471 return status;
2472}
2473
2474/*
34d024f8
MF		 2475 * This is working around a lock inversion between tasks acquiring DLM
2476 * locks while holding a page lock and the downconvert thread which
2477 * blocks dlm lock acquiry while acquiring page locks.
ccd979bd
MF		 2478 *
2479 * ** These _with_page variantes are only intended to be called from aop
2480 * methods that hold page locks and return a very specific *positive* error
2481 * code that aop methods pass up to the VFS -- test for errors with != 0. **
2482 *
34d024f8
MF		 2483 * The DLM is called such that it returns -EAGAIN if it would have
2484 * blocked waiting for the downconvert thread. In that case we unlock
2485 * our page so the downconvert thread can make progress. Once we've
2486 * done this we have to return AOP_TRUNCATED_PAGE so the aop method
2487 * that called us can bubble that back up into the VFS who will then
2488 * immediately retry the aop call.
ccd979bd
MF		 2489 *
2490 * We do a blocking lock and immediate unlock before returning, though, so that
2491 * the lock has a great chance of being cached on this node by the time the VFS
2492 * calls back to retry the aop. This has a potential to livelock as nodes
2493 * ping locks back and forth, but that's a risk we're willing to take to avoid
2494 * the lock inversion simply.
2495 */
e63aecb6 2496int ocfs2_inode_lock_with_page(struct inode *inode,
ccd979bd
MF		 2497 struct buffer_head **ret_bh,
2498 int ex,
2499 struct page *page)
2500{
2501 int ret;
2502
e63aecb6 2503 ret = ocfs2_inode_lock_full(inode, ret_bh, ex, OCFS2_LOCK_NONBLOCK);
ccd979bd
MF		 2504 if (ret == -EAGAIN) {
2505 unlock_page(page);
e63aecb6
MF		 2506 if (ocfs2_inode_lock(inode, ret_bh, ex) == 0)
2507 ocfs2_inode_unlock(inode, ex);
ccd979bd
MF		 2508 ret = AOP_TRUNCATED_PAGE;
2509 }
2510
2511 return ret;
2512}
2513
e63aecb6 2514int ocfs2_inode_lock_atime(struct inode *inode,
7f1a37e3
TY		 2515 struct vfsmount *vfsmnt,
2516 int *level)
2517{
2518 int ret;
2519
2520 mlog_entry_void();
e63aecb6 2521 ret = ocfs2_inode_lock(inode, NULL, 0);
7f1a37e3
TY		 2522 if (ret < 0) {
2523 mlog_errno(ret);
2524 return ret;
2525 }
2526
2527 /*
2528 * If we should update atime, we will get EX lock,
2529 * otherwise we just get PR lock.
2530 */
2531 if (ocfs2_should_update_atime(inode, vfsmnt)) {
2532 struct buffer_head *bh = NULL;
2533
e63aecb6
MF		 2534 ocfs2_inode_unlock(inode, 0);
2535 ret = ocfs2_inode_lock(inode, &bh, 1);
7f1a37e3
TY		 2536 if (ret < 0) {
2537 mlog_errno(ret);
2538 return ret;
2539 }
2540 *level = 1;
2541 if (ocfs2_should_update_atime(inode, vfsmnt))
2542 ocfs2_update_inode_atime(inode, bh);
2543 if (bh)
2544 brelse(bh);
2545 } else
2546 *level = 0;
2547
2548 mlog_exit(ret);
2549 return ret;
2550}
2551
e63aecb6 2552void ocfs2_inode_unlock(struct inode *inode,
ccd979bd
MF		 2553 int ex)
2554{
bd3e7610 2555 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
e63aecb6 2556 struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_inode_lockres;
c271c5c2 2557 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
ccd979bd
MF		 2558
2559 mlog_entry_void();
2560
b0697053
MF		 2561 mlog(0, "inode %llu drop %s META lock\n",
2562 (unsigned long long)OCFS2_I(inode)->ip_blkno,
ccd979bd
MF		 2563 ex ? "EXMODE" : "PRMODE");
2564
c271c5c2
SM		 2565 if (!ocfs2_is_hard_readonly(OCFS2_SB(inode->i_sb)) &&
2566 !ocfs2_mount_local(osb))
ccd979bd
MF		 2567 ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, level);
2568
2569 mlog_exit_void();
2570}
2571
df152c24 2572int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno)
83273932
SE		 2573{
2574 struct ocfs2_lock_res *lockres;
2575 struct ocfs2_orphan_scan_lvb *lvb;
83273932
SE		 2576 int status = 0;
2577
df152c24
SM		 2578 if (ocfs2_is_hard_readonly(osb))
2579 return -EROFS;
2580
2581 if (ocfs2_mount_local(osb))
2582 return 0;
2583
83273932 2584 lockres = &osb->osb_orphan_scan.os_lockres;
df152c24 2585 status = ocfs2_cluster_lock(osb, lockres, DLM_LOCK_EX, 0, 0);
83273932
SE		 2586 if (status < 0)
2587 return status;
2588
2589 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
1c520dfb
JB		 2590 if (ocfs2_dlm_lvb_valid(&lockres->l_lksb) &&
2591 lvb->lvb_version == OCFS2_ORPHAN_LVB_VERSION)
83273932 2592 *seqno = be32_to_cpu(lvb->lvb_os_seqno);
3211949f
SM		 2593 else
2594 *seqno = osb->osb_orphan_scan.os_seqno + 1;
2595
83273932
SE		 2596 return status;
2597}
2598
df152c24 2599void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno)
83273932
SE		 2600{
2601 struct ocfs2_lock_res *lockres;
2602 struct ocfs2_orphan_scan_lvb *lvb;
83273932 2603
df152c24
SM		 2604 if (!ocfs2_is_hard_readonly(osb) && !ocfs2_mount_local(osb)) {
2605 lockres = &osb->osb_orphan_scan.os_lockres;
2606 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
2607 lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
2608 lvb->lvb_os_seqno = cpu_to_be32(seqno);
2609 ocfs2_cluster_unlock(osb, lockres, DLM_LOCK_EX);
2610 }
83273932
SE		 2611}
2612
ccd979bd
MF		 2613int ocfs2_super_lock(struct ocfs2_super *osb,
2614 int ex)
2615{
c271c5c2 2616 int status = 0;
bd3e7610 2617 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
ccd979bd 2618 struct ocfs2_lock_res *lockres = &osb->osb_super_lockres;
ccd979bd
MF		 2619
2620 mlog_entry_void();
2621
2622 if (ocfs2_is_hard_readonly(osb))
2623 return -EROFS;
2624
c271c5c2
SM		 2625 if (ocfs2_mount_local(osb))
2626 goto bail;
2627
ccd979bd
MF		 2628 status = ocfs2_cluster_lock(osb, lockres, level, 0, 0);
2629 if (status < 0) {
2630 mlog_errno(status);
2631 goto bail;
2632 }
2633
2634 /* The super block lock path is really in the best position to
2635 * know when resources covered by the lock need to be
2636 * refreshed, so we do it here. Of course, making sense of
2637 * everything is up to the caller :) */
2638 status = ocfs2_should_refresh_lock_res(lockres);
2639 if (status < 0) {
2640 mlog_errno(status);
2641 goto bail;
2642 }
2643 if (status) {
8e8a4603 2644 status = ocfs2_refresh_slot_info(osb);
ccd979bd
MF		 2645
2646 ocfs2_complete_lock_res_refresh(lockres, status);
2647
2648 if (status < 0)
2649 mlog_errno(status);
8ddb7b00 2650 ocfs2_track_lock_refresh(lockres);
ccd979bd
MF		 2651 }
2652bail:
2653 mlog_exit(status);
2654 return status;
2655}
2656
2657void ocfs2_super_unlock(struct ocfs2_super *osb,
2658 int ex)
2659{
bd3e7610 2660 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
ccd979bd
MF		 2661 struct ocfs2_lock_res *lockres = &osb->osb_super_lockres;
2662
c271c5c2
SM		 2663 if (!ocfs2_mount_local(osb))
2664 ocfs2_cluster_unlock(osb, lockres, level);
ccd979bd
MF		 2665}
2666
2667int ocfs2_rename_lock(struct ocfs2_super *osb)
2668{
2669 int status;
2670 struct ocfs2_lock_res *lockres = &osb->osb_rename_lockres;
2671
2672 if (ocfs2_is_hard_readonly(osb))
2673 return -EROFS;
2674
c271c5c2
SM		 2675 if (ocfs2_mount_local(osb))
2676 return 0;
2677
bd3e7610 2678 status = ocfs2_cluster_lock(osb, lockres, DLM_LOCK_EX, 0, 0);
ccd979bd
MF		 2679 if (status < 0)
2680 mlog_errno(status);
2681
2682 return status;
2683}
2684
2685void ocfs2_rename_unlock(struct ocfs2_super *osb)
2686{
2687 struct ocfs2_lock_res *lockres = &osb->osb_rename_lockres;
2688
c271c5c2 2689 if (!ocfs2_mount_local(osb))
bd3e7610 2690 ocfs2_cluster_unlock(osb, lockres, DLM_LOCK_EX);
ccd979bd
MF		 2691}
2692
6ca497a8 2693int ocfs2_nfs_sync_lock(struct ocfs2_super *osb, int ex)
2694{
2695 int status;
2696 struct ocfs2_lock_res *lockres = &osb->osb_nfs_sync_lockres;
2697
2698 if (ocfs2_is_hard_readonly(osb))
2699 return -EROFS;
2700
2701 if (ocfs2_mount_local(osb))
2702 return 0;
2703
2704 status = ocfs2_cluster_lock(osb, lockres, ex ? LKM_EXMODE : LKM_PRMODE,
2705 0, 0);
2706 if (status < 0)
2707 mlog(ML_ERROR, "lock on nfs sync lock failed %d\n", status);
2708
2709 return status;
2710}
2711
2712void ocfs2_nfs_sync_unlock(struct ocfs2_super *osb, int ex)
2713{
2714 struct ocfs2_lock_res *lockres = &osb->osb_nfs_sync_lockres;
2715
2716 if (!ocfs2_mount_local(osb))
2717 ocfs2_cluster_unlock(osb, lockres,
2718 ex ? LKM_EXMODE : LKM_PRMODE);
2719}
2720
d680efe9
MF		 2721int ocfs2_dentry_lock(struct dentry *dentry, int ex)
2722{
2723 int ret;
bd3e7610 2724 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
d680efe9
MF		 2725 struct ocfs2_dentry_lock *dl = dentry->d_fsdata;
2726 struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
2727
2728 BUG_ON(!dl);
2729
2730 if (ocfs2_is_hard_readonly(osb))
2731 return -EROFS;
2732
c271c5c2
SM		 2733 if (ocfs2_mount_local(osb))
2734 return 0;
2735
d680efe9
MF		 2736 ret = ocfs2_cluster_lock(osb, &dl->dl_lockres, level, 0, 0);
2737 if (ret < 0)
2738 mlog_errno(ret);
2739
2740 return ret;
2741}
2742
2743void ocfs2_dentry_unlock(struct dentry *dentry, int ex)
2744{
bd3e7610 2745 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
d680efe9
MF		 2746 struct ocfs2_dentry_lock *dl = dentry->d_fsdata;
2747 struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
2748
c271c5c2
SM		 2749 if (!ocfs2_mount_local(osb))
2750 ocfs2_cluster_unlock(osb, &dl->dl_lockres, level);
d680efe9
MF		 2751}
2752
ccd979bd
MF		 2753/* Reference counting of the dlm debug structure. We want this because
2754 * open references on the debug inodes can live on after a mount, so
2755 * we can't rely on the ocfs2_super to always exist. */
2756static void ocfs2_dlm_debug_free(struct kref *kref)
2757{
2758 struct ocfs2_dlm_debug *dlm_debug;
2759
2760 dlm_debug = container_of(kref, struct ocfs2_dlm_debug, d_refcnt);
2761
2762 kfree(dlm_debug);
2763}
2764
2765void ocfs2_put_dlm_debug(struct ocfs2_dlm_debug *dlm_debug)
2766{
2767 if (dlm_debug)
2768 kref_put(&dlm_debug->d_refcnt, ocfs2_dlm_debug_free);
2769}
2770
2771static void ocfs2_get_dlm_debug(struct ocfs2_dlm_debug *debug)
2772{
2773 kref_get(&debug->d_refcnt);
2774}
2775
2776struct ocfs2_dlm_debug *ocfs2_new_dlm_debug(void)
2777{
2778 struct ocfs2_dlm_debug *dlm_debug;
2779
2780 dlm_debug = kmalloc(sizeof(struct ocfs2_dlm_debug), GFP_KERNEL);
2781 if (!dlm_debug) {
2782 mlog_errno(-ENOMEM);
2783 goto out;
2784 }
2785
2786 kref_init(&dlm_debug->d_refcnt);
2787 INIT_LIST_HEAD(&dlm_debug->d_lockres_tracking);
2788 dlm_debug->d_locking_state = NULL;
2789out:
2790 return dlm_debug;
2791}
2792
2793/* Access to this is arbitrated for us via seq_file->sem. */
2794struct ocfs2_dlm_seq_priv {
2795 struct ocfs2_dlm_debug *p_dlm_debug;
2796 struct ocfs2_lock_res p_iter_res;
2797 struct ocfs2_lock_res p_tmp_res;
2798};
2799
2800static struct ocfs2_lock_res *ocfs2_dlm_next_res(struct ocfs2_lock_res *start,
2801 struct ocfs2_dlm_seq_priv *priv)
2802{
2803 struct ocfs2_lock_res *iter, *ret = NULL;
2804 struct ocfs2_dlm_debug *dlm_debug = priv->p_dlm_debug;
2805
2806 assert_spin_locked(&ocfs2_dlm_tracking_lock);
2807
2808 list_for_each_entry(iter, &start->l_debug_list, l_debug_list) {
2809 /* discover the head of the list */
2810 if (&iter->l_debug_list == &dlm_debug->d_lockres_tracking) {
2811 mlog(0, "End of list found, %p\n", ret);
2812 break;
2813 }
2814
2815 /* We track our "dummy" iteration lockres' by a NULL
2816 * l_ops field. */
2817 if (iter->l_ops != NULL) {
2818 ret = iter;
2819 break;
2820 }
2821 }
2822
2823 return ret;
2824}
2825
2826static void *ocfs2_dlm_seq_start(struct seq_file *m, loff_t *pos)
2827{
2828 struct ocfs2_dlm_seq_priv *priv = m->private;
2829 struct ocfs2_lock_res *iter;
2830
2831 spin_lock(&ocfs2_dlm_tracking_lock);
2832 iter = ocfs2_dlm_next_res(&priv->p_iter_res, priv);
2833 if (iter) {
2834 /* Since lockres' have the lifetime of their container
2835 * (which can be inodes, ocfs2_supers, etc) we want to
2836 * copy this out to a temporary lockres while still
2837 * under the spinlock. Obviously after this we can't
2838 * trust any pointers on the copy returned, but that's
2839 * ok as the information we want isn't typically held
2840 * in them. */
2841 priv->p_tmp_res = *iter;
2842 iter = &priv->p_tmp_res;
2843 }
2844 spin_unlock(&ocfs2_dlm_tracking_lock);
2845
2846 return iter;
2847}
2848
2849static void ocfs2_dlm_seq_stop(struct seq_file *m, void *v)
2850{
2851}
2852
2853static void *ocfs2_dlm_seq_next(struct seq_file *m, void *v, loff_t *pos)
2854{
2855 struct ocfs2_dlm_seq_priv *priv = m->private;
2856 struct ocfs2_lock_res *iter = v;
2857 struct ocfs2_lock_res *dummy = &priv->p_iter_res;
2858
2859 spin_lock(&ocfs2_dlm_tracking_lock);
2860 iter = ocfs2_dlm_next_res(iter, priv);
2861 list_del_init(&dummy->l_debug_list);
2862 if (iter) {
2863 list_add(&dummy->l_debug_list, &iter->l_debug_list);
2864 priv->p_tmp_res = *iter;
2865 iter = &priv->p_tmp_res;
2866 }
2867 spin_unlock(&ocfs2_dlm_tracking_lock);
2868
2869 return iter;
2870}
2871
2872/* So that debugfs.ocfs2 can determine which format is being used */
8ddb7b00 2873#define OCFS2_DLM_DEBUG_STR_VERSION 2
ccd979bd
MF		 2874static int ocfs2_dlm_seq_show(struct seq_file *m, void *v)
2875{
2876 int i;
2877 char *lvb;
2878 struct ocfs2_lock_res *lockres = v;
2879
2880 if (!lockres)
2881 return -EINVAL;
2882
d680efe9
MF		 2883 seq_printf(m, "0x%x\t", OCFS2_DLM_DEBUG_STR_VERSION);
2884
2885 if (lockres->l_type == OCFS2_LOCK_TYPE_DENTRY)
2886 seq_printf(m, "%.*s%08x\t", OCFS2_DENTRY_LOCK_INO_START - 1,
2887 lockres->l_name,
2888 (unsigned int)ocfs2_get_dentry_lock_ino(lockres));
2889 else
2890 seq_printf(m, "%.*s\t", OCFS2_LOCK_ID_MAX_LEN, lockres->l_name);
2891
2892 seq_printf(m, "%d\t"
ccd979bd
MF		 2893 "0x%lx\t"
2894 "0x%x\t"
2895 "0x%x\t"
2896 "%u\t"
2897 "%u\t"
2898 "%d\t"
2899 "%d\t",
ccd979bd
MF		 2900 lockres->l_level,
2901 lockres->l_flags,
2902 lockres->l_action,
2903 lockres->l_unlock_action,
2904 lockres->l_ro_holders,
2905 lockres->l_ex_holders,
2906 lockres->l_requested,
2907 lockres->l_blocking);
2908
2909 /* Dump the raw LVB */
8f2c9c1b 2910 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
ccd979bd
MF		 2911 for(i = 0; i < DLM_LVB_LEN; i++)
2912 seq_printf(m, "0x%x\t", lvb[i]);
2913
8ddb7b00
SM		 2914#ifdef CONFIG_OCFS2_FS_STATS
2915# define lock_num_prmode(_l) (_l)->l_lock_num_prmode
2916# define lock_num_exmode(_l) (_l)->l_lock_num_exmode
2917# define lock_num_prmode_failed(_l) (_l)->l_lock_num_prmode_failed
2918# define lock_num_exmode_failed(_l) (_l)->l_lock_num_exmode_failed
2919# define lock_total_prmode(_l) (_l)->l_lock_total_prmode
2920# define lock_total_exmode(_l) (_l)->l_lock_total_exmode
2921# define lock_max_prmode(_l) (_l)->l_lock_max_prmode
2922# define lock_max_exmode(_l) (_l)->l_lock_max_exmode
2923# define lock_refresh(_l) (_l)->l_lock_refresh
2924#else
dd25e55e
RD		 2925# define lock_num_prmode(_l) (0ULL)
2926# define lock_num_exmode(_l) (0ULL)
8ddb7b00
SM		 2927# define lock_num_prmode_failed(_l) (0)
2928# define lock_num_exmode_failed(_l) (0)
dd25e55e
RD		 2929# define lock_total_prmode(_l) (0ULL)
2930# define lock_total_exmode(_l) (0ULL)
8ddb7b00
SM		 2931# define lock_max_prmode(_l) (0)
2932# define lock_max_exmode(_l) (0)
2933# define lock_refresh(_l) (0)
2934#endif
2935 /* The following seq_print was added in version 2 of this output */
2936 seq_printf(m, "%llu\t"
2937 "%llu\t"
2938 "%u\t"
2939 "%u\t"
2940 "%llu\t"
2941 "%llu\t"
2942 "%u\t"
2943 "%u\t"
2944 "%u\t",
2945 lock_num_prmode(lockres),
2946 lock_num_exmode(lockres),
2947 lock_num_prmode_failed(lockres),
2948 lock_num_exmode_failed(lockres),
2949 lock_total_prmode(lockres),
2950 lock_total_exmode(lockres),
2951 lock_max_prmode(lockres),
2952 lock_max_exmode(lockres),
2953 lock_refresh(lockres));
2954
ccd979bd
MF		 2955 /* End the line */
2956 seq_printf(m, "\n");
2957 return 0;
2958}
2959
90d99779 2960static const struct seq_operations ocfs2_dlm_seq_ops = {
ccd979bd
MF		 2961 .start = ocfs2_dlm_seq_start,
2962 .stop = ocfs2_dlm_seq_stop,
2963 .next = ocfs2_dlm_seq_next,
2964 .show = ocfs2_dlm_seq_show,
2965};
2966
2967static int ocfs2_dlm_debug_release(struct inode *inode, struct file *file)
2968{
2969 struct seq_file *seq = (struct seq_file *) file->private_data;
2970 struct ocfs2_dlm_seq_priv *priv = seq->private;
2971 struct ocfs2_lock_res *res = &priv->p_iter_res;
2972
2973 ocfs2_remove_lockres_tracking(res);
2974 ocfs2_put_dlm_debug(priv->p_dlm_debug);
2975 return seq_release_private(inode, file);
2976}
2977
2978static int ocfs2_dlm_debug_open(struct inode *inode, struct file *file)
2979{
2980 int ret;
2981 struct ocfs2_dlm_seq_priv *priv;
2982 struct seq_file *seq;
2983 struct ocfs2_super *osb;
2984
2985 priv = kzalloc(sizeof(struct ocfs2_dlm_seq_priv), GFP_KERNEL);
2986 if (!priv) {
2987 ret = -ENOMEM;
2988 mlog_errno(ret);
2989 goto out;
2990 }
8e18e294 2991 osb = inode->i_private;
ccd979bd
MF		 2992 ocfs2_get_dlm_debug(osb->osb_dlm_debug);
2993 priv->p_dlm_debug = osb->osb_dlm_debug;
2994 INIT_LIST_HEAD(&priv->p_iter_res.l_debug_list);
2995
2996 ret = seq_open(file, &ocfs2_dlm_seq_ops);
2997 if (ret) {
2998 kfree(priv);
2999 mlog_errno(ret);
3000 goto out;
3001 }
3002
3003 seq = (struct seq_file *) file->private_data;
3004 seq->private = priv;
3005
3006 ocfs2_add_lockres_tracking(&priv->p_iter_res,
3007 priv->p_dlm_debug);
3008
3009out:
3010 return ret;
3011}
3012
4b6f5d20 3013static const struct file_operations ocfs2_dlm_debug_fops = {
ccd979bd
MF		 3014 .open = ocfs2_dlm_debug_open,
3015 .release = ocfs2_dlm_debug_release,
3016 .read = seq_read,
3017 .llseek = seq_lseek,
3018};
3019
3020static int ocfs2_dlm_init_debug(struct ocfs2_super *osb)
3021{
3022 int ret = 0;
3023 struct ocfs2_dlm_debug *dlm_debug = osb->osb_dlm_debug;
3024
3025 dlm_debug->d_locking_state = debugfs_create_file("locking_state",
3026 S_IFREG|S_IRUSR,
3027 osb->osb_debug_root,
3028 osb,
3029 &ocfs2_dlm_debug_fops);
3030 if (!dlm_debug->d_locking_state) {
3031 ret = -EINVAL;
3032 mlog(ML_ERROR,
3033 "Unable to create locking state debugfs file.\n");
3034 goto out;
3035 }
3036
3037 ocfs2_get_dlm_debug(dlm_debug);
3038out:
3039 return ret;
3040}
3041
3042static void ocfs2_dlm_shutdown_debug(struct ocfs2_super *osb)
3043{
3044 struct ocfs2_dlm_debug *dlm_debug = osb->osb_dlm_debug;
3045
3046 if (dlm_debug) {
3047 debugfs_remove(dlm_debug->d_locking_state);
3048 ocfs2_put_dlm_debug(dlm_debug);
3049 }
3050}
3051
3052int ocfs2_dlm_init(struct ocfs2_super *osb)
3053{
c271c5c2 3054 int status = 0;
4670c46d 3055 struct ocfs2_cluster_connection *conn = NULL;
ccd979bd
MF		 3056
3057 mlog_entry_void();
3058
0abd6d18
MF		 3059 if (ocfs2_mount_local(osb)) {
3060 osb->node_num = 0;
c271c5c2 3061 goto local;
0abd6d18 3062 }
c271c5c2 3063
ccd979bd
MF		 3064 status = ocfs2_dlm_init_debug(osb);
3065 if (status < 0) {
3066 mlog_errno(status);
3067 goto bail;
3068 }
3069
34d024f8
MF		 3070 /* launch downconvert thread */
3071 osb->dc_task = kthread_run(ocfs2_downconvert_thread, osb, "ocfs2dc");
3072 if (IS_ERR(osb->dc_task)) {
3073 status = PTR_ERR(osb->dc_task);
3074 osb->dc_task = NULL;
ccd979bd
MF		 3075 mlog_errno(status);
3076 goto bail;
3077 }
3078
ccd979bd 3079 /* for now, uuid == domain */
9c6c877c
JB		 3080 status = ocfs2_cluster_connect(osb->osb_cluster_stack,
3081 osb->uuid_str,
4670c46d 3082 strlen(osb->uuid_str),
553b5eb9 3083 &lproto, ocfs2_do_node_down, osb,
4670c46d
JB		 3084 &conn);
3085 if (status) {
ccd979bd
MF		 3086 mlog_errno(status);
3087 goto bail;
3088 }
3089
0abd6d18
MF		 3090 status = ocfs2_cluster_this_node(&osb->node_num);
3091 if (status < 0) {
3092 mlog_errno(status);
3093 mlog(ML_ERROR,
3094 "could not find this host's node number\n");
286eaa95 3095 ocfs2_cluster_disconnect(conn, 0);
0abd6d18
MF		 3096 goto bail;
3097 }
3098
c271c5c2 3099local:
ccd979bd
MF		 3100 ocfs2_super_lock_res_init(&osb->osb_super_lockres, osb);
3101 ocfs2_rename_lock_res_init(&osb->osb_rename_lockres, osb);
6ca497a8 3102 ocfs2_nfs_sync_lock_res_init(&osb->osb_nfs_sync_lockres, osb);
83273932 3103 ocfs2_orphan_scan_lock_res_init(&osb->osb_orphan_scan.os_lockres, osb);
ccd979bd 3104
4670c46d 3105 osb->cconn = conn;
ccd979bd
MF		 3106
3107 status = 0;
3108bail:
3109 if (status < 0) {
3110 ocfs2_dlm_shutdown_debug(osb);
34d024f8
MF		 3111 if (osb->dc_task)
3112 kthread_stop(osb->dc_task);
ccd979bd
MF		 3113 }
3114
3115 mlog_exit(status);
3116 return status;
3117}
3118
286eaa95
JB		 3119void ocfs2_dlm_shutdown(struct ocfs2_super *osb,
3120 int hangup_pending)
ccd979bd
MF		 3121{
3122 mlog_entry_void();
3123
ccd979bd
MF		 3124 ocfs2_drop_osb_locks(osb);
3125
4670c46d
JB		 3126 /*
3127 * Now that we have dropped all locks and ocfs2_dismount_volume()
3128 * has disabled recovery, the DLM won't be talking to us. It's
3129 * safe to tear things down before disconnecting the cluster.
3130 */
3131
34d024f8
MF		 3132 if (osb->dc_task) {
3133 kthread_stop(osb->dc_task);
3134 osb->dc_task = NULL;
ccd979bd
MF		 3135 }
3136
3137 ocfs2_lock_res_free(&osb->osb_super_lockres);
3138 ocfs2_lock_res_free(&osb->osb_rename_lockres);
6ca497a8 3139 ocfs2_lock_res_free(&osb->osb_nfs_sync_lockres);
83273932 3140 ocfs2_lock_res_free(&osb->osb_orphan_scan.os_lockres);
ccd979bd 3141
286eaa95 3142 ocfs2_cluster_disconnect(osb->cconn, hangup_pending);
4670c46d 3143 osb->cconn = NULL;
ccd979bd
MF		 3144
3145 ocfs2_dlm_shutdown_debug(osb);
3146
3147 mlog_exit_void();
3148}
3149
ccd979bd 3150static int ocfs2_drop_lock(struct ocfs2_super *osb,
0d5dc6c2 3151 struct ocfs2_lock_res *lockres)
ccd979bd 3152{
7431cd7e 3153 int ret;
ccd979bd 3154 unsigned long flags;
bd3e7610 3155 u32 lkm_flags = 0;
ccd979bd
MF		 3156
3157 /* We didn't get anywhere near actually using this lockres. */
3158 if (!(lockres->l_flags & OCFS2_LOCK_INITIALIZED))
3159 goto out;
3160
b80fc012 3161 if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB)
bd3e7610 3162 lkm_flags |= DLM_LKF_VALBLK;
b80fc012 3163
ccd979bd
MF		 3164 spin_lock_irqsave(&lockres->l_lock, flags);
3165
3166 mlog_bug_on_msg(!(lockres->l_flags & OCFS2_LOCK_FREEING),
3167 "lockres %s, flags 0x%lx\n",
3168 lockres->l_name, lockres->l_flags);
3169
3170 while (lockres->l_flags & OCFS2_LOCK_BUSY) {
3171 mlog(0, "waiting on busy lock \"%s\": flags = %lx, action = "
3172 "%u, unlock_action = %u\n",
3173 lockres->l_name, lockres->l_flags, lockres->l_action,
3174 lockres->l_unlock_action);
3175
3176 spin_unlock_irqrestore(&lockres->l_lock, flags);
3177
3178 /* XXX: Today we just wait on any busy
3179 * locks... Perhaps we need to cancel converts in the
3180 * future? */
3181 ocfs2_wait_on_busy_lock(lockres);
3182
3183 spin_lock_irqsave(&lockres->l_lock, flags);
3184 }
3185
0d5dc6c2
MF		 3186 if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB) {
3187 if (lockres->l_flags & OCFS2_LOCK_ATTACHED &&
bd3e7610 3188 lockres->l_level == DLM_LOCK_EX &&
0d5dc6c2
MF		 3189 !(lockres->l_flags & OCFS2_LOCK_NEEDS_REFRESH))
3190 lockres->l_ops->set_lvb(lockres);
3191 }
ccd979bd
MF		 3192
3193 if (lockres->l_flags & OCFS2_LOCK_BUSY)
3194 mlog(ML_ERROR, "destroying busy lock: \"%s\"\n",
3195 lockres->l_name);
3196 if (lockres->l_flags & OCFS2_LOCK_BLOCKED)
3197 mlog(0, "destroying blocked lock: \"%s\"\n", lockres->l_name);
3198
3199 if (!(lockres->l_flags & OCFS2_LOCK_ATTACHED)) {
3200 spin_unlock_irqrestore(&lockres->l_lock, flags);
3201 goto out;
3202 }
3203
3204 lockres_clear_flags(lockres, OCFS2_LOCK_ATTACHED);
3205
3206 /* make sure we never get here while waiting for an ast to
3207 * fire. */
3208 BUG_ON(lockres->l_action != OCFS2_AST_INVALID);
3209
3210 /* is this necessary? */
3211 lockres_or_flags(lockres, OCFS2_LOCK_BUSY);
3212 lockres->l_unlock_action = OCFS2_UNLOCK_DROP_LOCK;
3213 spin_unlock_irqrestore(&lockres->l_lock, flags);
3214
3215 mlog(0, "lock %s\n", lockres->l_name);
3216
a796d286 3217 ret = ocfs2_dlm_unlock(osb->cconn, &lockres->l_lksb, lkm_flags);
7431cd7e
JB		 3218 if (ret) {
3219 ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres);
ccd979bd 3220 mlog(ML_ERROR, "lockres flags: %lu\n", lockres->l_flags);
cf0acdcd 3221 ocfs2_dlm_dump_lksb(&lockres->l_lksb);
ccd979bd
MF		 3222 BUG();
3223 }
73ac36ea 3224 mlog(0, "lock %s, successful return from ocfs2_dlm_unlock\n",
ccd979bd
MF		 3225 lockres->l_name);
3226
3227 ocfs2_wait_on_busy_lock(lockres);
3228out:
3229 mlog_exit(0);
3230 return 0;
3231}
3232
3233/* Mark the lockres as being dropped. It will no longer be
3234 * queued if blocking, but we still may have to wait on it
34d024f8 3235 * being dequeued from the downconvert thread before we can consider
2bd63216 3236 * it safe to drop.
ccd979bd
MF		 3237 *
3238 * You can *not* attempt to call cluster_lock on this lockres anymore. */
3239void ocfs2_mark_lockres_freeing(struct ocfs2_lock_res *lockres)
3240{
3241 int status;
3242 struct ocfs2_mask_waiter mw;
3243 unsigned long flags;
3244
3245 ocfs2_init_mask_waiter(&mw);
3246
3247 spin_lock_irqsave(&lockres->l_lock, flags);
3248 lockres->l_flags |= OCFS2_LOCK_FREEING;
3249 while (lockres->l_flags & OCFS2_LOCK_QUEUED) {
3250 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_QUEUED, 0);
3251 spin_unlock_irqrestore(&lockres->l_lock, flags);
3252
3253 mlog(0, "Waiting on lockres %s\n", lockres->l_name);
3254
3255 status = ocfs2_wait_for_mask(&mw);
3256 if (status)
3257 mlog_errno(status);
3258
3259 spin_lock_irqsave(&lockres->l_lock, flags);
3260 }
3261 spin_unlock_irqrestore(&lockres->l_lock, flags);
3262}
3263
d680efe9
MF		 3264void ocfs2_simple_drop_lockres(struct ocfs2_super *osb,
3265 struct ocfs2_lock_res *lockres)
ccd979bd 3266{
d680efe9 3267 int ret;
ccd979bd 3268
d680efe9 3269 ocfs2_mark_lockres_freeing(lockres);
0d5dc6c2 3270 ret = ocfs2_drop_lock(osb, lockres);
d680efe9
MF		 3271 if (ret)
3272 mlog_errno(ret);
3273}
ccd979bd 3274
d680efe9
MF		 3275static void ocfs2_drop_osb_locks(struct ocfs2_super *osb)
3276{
3277 ocfs2_simple_drop_lockres(osb, &osb->osb_super_lockres);
3278 ocfs2_simple_drop_lockres(osb, &osb->osb_rename_lockres);
6ca497a8 3279 ocfs2_simple_drop_lockres(osb, &osb->osb_nfs_sync_lockres);
83273932 3280 ocfs2_simple_drop_lockres(osb, &osb->osb_orphan_scan.os_lockres);
ccd979bd
MF		 3281}
3282
ccd979bd
MF		 3283int ocfs2_drop_inode_locks(struct inode *inode)
3284{
3285 int status, err;
ccd979bd
MF		 3286
3287 mlog_entry_void();
3288
3289 /* No need to call ocfs2_mark_lockres_freeing here -
3290 * ocfs2_clear_inode has done it for us. */
3291
3292 err = ocfs2_drop_lock(OCFS2_SB(inode->i_sb),
50008630 3293 &OCFS2_I(inode)->ip_open_lockres);
ccd979bd
MF		 3294 if (err < 0)
3295 mlog_errno(err);
3296
3297 status = err;
3298
3299 err = ocfs2_drop_lock(OCFS2_SB(inode->i_sb),
e63aecb6 3300 &OCFS2_I(inode)->ip_inode_lockres);
ccd979bd
MF		 3301 if (err < 0)
3302 mlog_errno(err);
3303 if (err < 0 && !status)
3304 status = err;
3305
3306 err = ocfs2_drop_lock(OCFS2_SB(inode->i_sb),
0d5dc6c2 3307 &OCFS2_I(inode)->ip_rw_lockres);
ccd979bd
MF		 3308 if (err < 0)
3309 mlog_errno(err);
3310 if (err < 0 && !status)
3311 status = err;
3312
3313 mlog_exit(status);
3314 return status;
3315}
3316
de551246
JB		 3317static unsigned int ocfs2_prepare_downconvert(struct ocfs2_lock_res *lockres,
3318 int new_level)
ccd979bd
MF		 3319{
3320 assert_spin_locked(&lockres->l_lock);
3321
bd3e7610 3322 BUG_ON(lockres->l_blocking <= DLM_LOCK_NL);
ccd979bd
MF		 3323
3324 if (lockres->l_level <= new_level) {
9b915181
SM		 3325 mlog(ML_ERROR, "lockres %s, lvl %d <= %d, blcklst %d, mask %d, "
3326 "type %d, flags 0x%lx, hold %d %d, act %d %d, req %d, "
3327 "block %d, pgen %d\n", lockres->l_name, lockres->l_level,
3328 new_level, list_empty(&lockres->l_blocked_list),
3329 list_empty(&lockres->l_mask_waiters), lockres->l_type,
3330 lockres->l_flags, lockres->l_ro_holders,
3331 lockres->l_ex_holders, lockres->l_action,
3332 lockres->l_unlock_action, lockres->l_requested,
3333 lockres->l_blocking, lockres->l_pending_gen);
ccd979bd
MF		 3334 BUG();
3335 }
3336
9b915181
SM		 3337 mlog(ML_BASTS, "lockres %s, level %d => %d, blocking %d\n",
3338 lockres->l_name, lockres->l_level, new_level, lockres->l_blocking);
ccd979bd
MF		 3339
3340 lockres->l_action = OCFS2_AST_DOWNCONVERT;
3341 lockres->l_requested = new_level;
3342 lockres_or_flags(lockres, OCFS2_LOCK_BUSY);
de551246 3343 return lockres_set_pending(lockres);
ccd979bd
MF		 3344}
3345
3346static int ocfs2_downconvert_lock(struct ocfs2_super *osb,
3347 struct ocfs2_lock_res *lockres,
3348 int new_level,
de551246
JB		 3349 int lvb,
3350 unsigned int generation)
ccd979bd 3351{
bd3e7610
JB		 3352 int ret;
3353 u32 dlm_flags = DLM_LKF_CONVERT;
ccd979bd
MF		 3354
3355 mlog_entry_void();
3356
9b915181
SM		 3357 mlog(ML_BASTS, "lockres %s, level %d => %d\n", lockres->l_name,
3358 lockres->l_level, new_level);
3359
ccd979bd 3360 if (lvb)
bd3e7610 3361 dlm_flags |= DLM_LKF_VALBLK;
ccd979bd 3362
4670c46d 3363 ret = ocfs2_dlm_lock(osb->cconn,
7431cd7e
JB		 3364 new_level,
3365 &lockres->l_lksb,
3366 dlm_flags,
3367 lockres->l_name,
a796d286 3368 OCFS2_LOCK_ID_MAX_LEN - 1);
de551246 3369 lockres_clear_pending(lockres, generation, osb);
7431cd7e
JB		 3370 if (ret) {
3371 ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres);
ccd979bd
MF		 3372 ocfs2_recover_from_dlm_error(lockres, 1);
3373 goto bail;
3374 }
3375
3376 ret = 0;
3377bail:
3378 mlog_exit(ret);
3379 return ret;
3380}
3381
24ef1815 3382/* returns 1 when the caller should unlock and call ocfs2_dlm_unlock */
ccd979bd
MF		 3383static int ocfs2_prepare_cancel_convert(struct ocfs2_super *osb,
3384 struct ocfs2_lock_res *lockres)
3385{
3386 assert_spin_locked(&lockres->l_lock);
3387
3388 mlog_entry_void();
ccd979bd
MF		 3389
3390 if (lockres->l_unlock_action == OCFS2_UNLOCK_CANCEL_CONVERT) {
3391 /* If we're already trying to cancel a lock conversion
3392 * then just drop the spinlock and allow the caller to
3393 * requeue this lock. */
9b915181 3394 mlog(ML_BASTS, "lockres %s, skip convert\n", lockres->l_name);
ccd979bd
MF		 3395 return 0;
3396 }
3397
3398 /* were we in a convert when we got the bast fire? */
3399 BUG_ON(lockres->l_action != OCFS2_AST_CONVERT &&
3400 lockres->l_action != OCFS2_AST_DOWNCONVERT);
3401 /* set things up for the unlockast to know to just
3402 * clear out the ast_action and unset busy, etc. */
3403 lockres->l_unlock_action = OCFS2_UNLOCK_CANCEL_CONVERT;
3404
3405 mlog_bug_on_msg(!(lockres->l_flags & OCFS2_LOCK_BUSY),
3406 "lock %s, invalid flags: 0x%lx\n",
3407 lockres->l_name, lockres->l_flags);
3408
9b915181
SM		 3409 mlog(ML_BASTS, "lockres %s\n", lockres->l_name);
3410
ccd979bd
MF		 3411 return 1;
3412}
3413
3414static int ocfs2_cancel_convert(struct ocfs2_super *osb,
3415 struct ocfs2_lock_res *lockres)
3416{
3417 int ret;
ccd979bd
MF		 3418
3419 mlog_entry_void();
ccd979bd 3420
4670c46d 3421 ret = ocfs2_dlm_unlock(osb->cconn, &lockres->l_lksb,
a796d286 3422 DLM_LKF_CANCEL);
7431cd7e
JB		 3423 if (ret) {
3424 ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres);
ccd979bd
MF		 3425 ocfs2_recover_from_dlm_error(lockres, 0);
3426 }
3427
9b915181 3428 mlog(ML_BASTS, "lockres %s\n", lockres->l_name);
ccd979bd 3429
ccd979bd
MF		 3430 mlog_exit(ret);
3431 return ret;
3432}
3433
b5e500e2
MF		 3434static int ocfs2_unblock_lock(struct ocfs2_super *osb,
3435 struct ocfs2_lock_res *lockres,
3436 struct ocfs2_unblock_ctl *ctl)
ccd979bd
MF		 3437{
3438 unsigned long flags;
3439 int blocking;
3440 int new_level;
079b8057 3441 int level;
ccd979bd 3442 int ret = 0;
5ef0d4ea 3443 int set_lvb = 0;
de551246 3444 unsigned int gen;
ccd979bd
MF		 3445
3446 mlog_entry_void();
3447
3448 spin_lock_irqsave(&lockres->l_lock, flags);
3449
ccd979bd 3450recheck:
db0f6ce6
SM		 3451 /*
3452 * Is it still blocking? If not, we have no more work to do.
3453 */
3454 if (!(lockres->l_flags & OCFS2_LOCK_BLOCKED)) {
3455 BUG_ON(lockres->l_blocking != DLM_LOCK_NL);
3456 spin_unlock_irqrestore(&lockres->l_lock, flags);
3457 ret = 0;
3458 goto leave;
3459 }
3460
ccd979bd 3461 if (lockres->l_flags & OCFS2_LOCK_BUSY) {
de551246
JB		 3462 /* XXX
3463 * This is a *big* race. The OCFS2_LOCK_PENDING flag
3464 * exists entirely for one reason - another thread has set
3465 * OCFS2_LOCK_BUSY, but has *NOT* yet called dlm_lock().
3466 *
3467 * If we do ocfs2_cancel_convert() before the other thread
3468 * calls dlm_lock(), our cancel will do nothing. We will
3469 * get no ast, and we will have no way of knowing the
3470 * cancel failed. Meanwhile, the other thread will call
3471 * into dlm_lock() and wait...forever.
3472 *
3473 * Why forever? Because another node has asked for the
3474 * lock first; that's why we're here in unblock_lock().
3475 *
3476 * The solution is OCFS2_LOCK_PENDING. When PENDING is
3477 * set, we just requeue the unblock. Only when the other
3478 * thread has called dlm_lock() and cleared PENDING will
3479 * we then cancel their request.
3480 *
3481 * All callers of dlm_lock() must set OCFS2_DLM_PENDING
3482 * at the same time they set OCFS2_DLM_BUSY. They must
3483 * clear OCFS2_DLM_PENDING after dlm_lock() returns.
3484 */
9b915181
SM		 3485 if (lockres->l_flags & OCFS2_LOCK_PENDING) {
3486 mlog(ML_BASTS, "lockres %s, ReQ: Pending\n",
3487 lockres->l_name);
de551246 3488 goto leave_requeue;
9b915181 3489 }
de551246 3490
d680efe9 3491 ctl->requeue = 1;
ccd979bd
MF		 3492 ret = ocfs2_prepare_cancel_convert(osb, lockres);
3493 spin_unlock_irqrestore(&lockres->l_lock, flags);
3494 if (ret) {
3495 ret = ocfs2_cancel_convert(osb, lockres);
3496 if (ret < 0)
3497 mlog_errno(ret);
3498 }
3499 goto leave;
3500 }
3501
a1912826
SM		 3502 /*
3503 * This prevents livelocks. OCFS2_LOCK_UPCONVERT_FINISHING flag is
3504 * set when the ast is received for an upconvert just before the
3505 * OCFS2_LOCK_BUSY flag is cleared. Now if the fs received a bast
3506 * on the heels of the ast, we want to delay the downconvert just
3507 * enough to allow the up requestor to do its task. Because this
3508 * lock is in the blocked queue, the lock will be downconverted
3509 * as soon as the requestor is done with the lock.
3510 */
3511 if (lockres->l_flags & OCFS2_LOCK_UPCONVERT_FINISHING)
3512 goto leave_requeue;
3513
0d74125a
SM		 3514 /*
3515 * How can we block and yet be at NL? We were trying to upconvert
3516 * from NL and got canceled. The code comes back here, and now
3517 * we notice and clear BLOCKING.
3518 */
3519 if (lockres->l_level == DLM_LOCK_NL) {
3520 BUG_ON(lockres->l_ex_holders || lockres->l_ro_holders);
9b915181 3521 mlog(ML_BASTS, "lockres %s, Aborting dc\n", lockres->l_name);
0d74125a
SM		 3522 lockres->l_blocking = DLM_LOCK_NL;
3523 lockres_clear_flags(lockres, OCFS2_LOCK_BLOCKED);
3524 spin_unlock_irqrestore(&lockres->l_lock, flags);
3525 goto leave;
3526 }
3527
ccd979bd
MF		 3528 /* if we're blocking an exclusive and we have *any* holders,
3529 * then requeue. */
bd3e7610 3530 if ((lockres->l_blocking == DLM_LOCK_EX)
9b915181
SM		 3531 && (lockres->l_ex_holders || lockres->l_ro_holders)) {
3532 mlog(ML_BASTS, "lockres %s, ReQ: EX/PR Holders %u,%u\n",
3533 lockres->l_name, lockres->l_ex_holders,
3534 lockres->l_ro_holders);
f7fbfdd1 3535 goto leave_requeue;
9b915181 3536 }
ccd979bd
MF		 3537
3538 /* If it's a PR we're blocking, then only
3539 * requeue if we've got any EX holders */
bd3e7610 3540 if (lockres->l_blocking == DLM_LOCK_PR &&
9b915181
SM		 3541 lockres->l_ex_holders) {
3542 mlog(ML_BASTS, "lockres %s, ReQ: EX Holders %u\n",
3543 lockres->l_name, lockres->l_ex_holders);
f7fbfdd1 3544 goto leave_requeue;
9b915181 3545 }
f7fbfdd1
MF		 3546
3547 /*
3548 * Can we get a lock in this state if the holder counts are
3549 * zero? The meta data unblock code used to check this.
3550 */
3551 if ((lockres->l_ops->flags & LOCK_TYPE_REQUIRES_REFRESH)
9b915181
SM		 3552 && (lockres->l_flags & OCFS2_LOCK_REFRESHING)) {
3553 mlog(ML_BASTS, "lockres %s, ReQ: Lock Refreshing\n",
3554 lockres->l_name);
f7fbfdd1 3555 goto leave_requeue;
9b915181 3556 }
ccd979bd 3557
16d5b956
MF		 3558 new_level = ocfs2_highest_compat_lock_level(lockres->l_blocking);
3559
3560 if (lockres->l_ops->check_downconvert
9b915181
SM		 3561 && !lockres->l_ops->check_downconvert(lockres, new_level)) {
3562 mlog(ML_BASTS, "lockres %s, ReQ: Checkpointing\n",
3563 lockres->l_name);
16d5b956 3564 goto leave_requeue;
9b915181 3565 }
16d5b956 3566
ccd979bd
MF		 3567 /* If we get here, then we know that there are no more
3568 * incompatible holders (and anyone asking for an incompatible
3569 * lock is blocked). We can now downconvert the lock */
cc567d89 3570 if (!lockres->l_ops->downconvert_worker)
ccd979bd
MF		 3571 goto downconvert;
3572
3573 /* Some lockres types want to do a bit of work before
3574 * downconverting a lock. Allow that here. The worker function
3575 * may sleep, so we save off a copy of what we're blocking as
3576 * it may change while we're not holding the spin lock. */
3577 blocking = lockres->l_blocking;
079b8057 3578 level = lockres->l_level;
ccd979bd
MF		 3579 spin_unlock_irqrestore(&lockres->l_lock, flags);
3580
cc567d89 3581 ctl->unblock_action = lockres->l_ops->downconvert_worker(lockres, blocking);
d680efe9 3582
9b915181
SM		 3583 if (ctl->unblock_action == UNBLOCK_STOP_POST) {
3584 mlog(ML_BASTS, "lockres %s, UNBLOCK_STOP_POST\n",
3585 lockres->l_name);
d680efe9 3586 goto leave;
9b915181 3587 }
ccd979bd
MF		 3588
3589 spin_lock_irqsave(&lockres->l_lock, flags);
079b8057 3590 if ((blocking != lockres->l_blocking) || (level != lockres->l_level)) {
ccd979bd
MF		 3591 /* If this changed underneath us, then we can't drop
3592 * it just yet. */
9b915181
SM		 3593 mlog(ML_BASTS, "lockres %s, block=%d:%d, level=%d:%d, "
3594 "Recheck\n", lockres->l_name, blocking,
3595 lockres->l_blocking, level, lockres->l_level);
ccd979bd
MF		 3596 goto recheck;
3597 }
3598
3599downconvert:
d680efe9 3600 ctl->requeue = 0;
ccd979bd 3601
5ef0d4ea 3602 if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB) {
bd3e7610 3603 if (lockres->l_level == DLM_LOCK_EX)
5ef0d4ea
MF		 3604 set_lvb = 1;
3605
3606 /*
3607 * We only set the lvb if the lock has been fully
3608 * refreshed - otherwise we risk setting stale
3609 * data. Otherwise, there's no need to actually clear
3610 * out the lvb here as it's value is still valid.
3611 */
3612 if (set_lvb && !(lockres->l_flags & OCFS2_LOCK_NEEDS_REFRESH))
3613 lockres->l_ops->set_lvb(lockres);
3614 }
3615
de551246 3616 gen = ocfs2_prepare_downconvert(lockres, new_level);
ccd979bd 3617 spin_unlock_irqrestore(&lockres->l_lock, flags);
de551246
JB		 3618 ret = ocfs2_downconvert_lock(osb, lockres, new_level, set_lvb,
3619 gen);
3620
ccd979bd
MF		 3621leave:
3622 mlog_exit(ret);
3623 return ret;
f7fbfdd1
MF		 3624
3625leave_requeue:
3626 spin_unlock_irqrestore(&lockres->l_lock, flags);
3627 ctl->requeue = 1;
3628
3629 mlog_exit(0);
3630 return 0;
ccd979bd
MF		 3631}
3632
d680efe9
MF		 3633static int ocfs2_data_convert_worker(struct ocfs2_lock_res *lockres,
3634 int blocking)
ccd979bd
MF		 3635{
3636 struct inode *inode;
3637 struct address_space *mapping;
3638
ccd979bd
MF		 3639 inode = ocfs2_lock_res_inode(lockres);
3640 mapping = inode->i_mapping;
3641
1044e401 3642 if (!S_ISREG(inode->i_mode))
f1f54068
MF		 3643 goto out;
3644
7f4a2a97
MF		 3645 /*
3646 * We need this before the filemap_fdatawrite() so that it can
3647 * transfer the dirty bit from the PTE to the
3648 * page. Unfortunately this means that even for EX->PR
3649 * downconverts, we'll lose our mappings and have to build
3650 * them up again.
3651 */
3652 unmap_mapping_range(mapping, 0, 0, 0);
3653
ccd979bd 3654 if (filemap_fdatawrite(mapping)) {
b0697053
MF		 3655 mlog(ML_ERROR, "Could not sync inode %llu for downconvert!",
3656 (unsigned long long)OCFS2_I(inode)->ip_blkno);
ccd979bd
MF		 3657 }
3658 sync_mapping_buffers(mapping);
bd3e7610 3659 if (blocking == DLM_LOCK_EX) {
ccd979bd 3660 truncate_inode_pages(mapping, 0);
ccd979bd
MF		 3661 } else {
3662 /* We only need to wait on the I/O if we're not also
3663 * truncating pages because truncate_inode_pages waits
3664 * for us above. We don't truncate pages if we're
3665 * blocking anything < EXMODE because we want to keep
3666 * them around in that case. */
3667 filemap_fdatawait(mapping);
3668 }
3669
f1f54068 3670out:
d680efe9 3671 return UNBLOCK_CONTINUE;
ccd979bd
MF		 3672}
3673
a4338481
TM		 3674static int ocfs2_ci_checkpointed(struct ocfs2_caching_info *ci,
3675 struct ocfs2_lock_res *lockres,
3676 int new_level)
810d5aeb 3677{
a4338481 3678 int checkpointed = ocfs2_ci_fully_checkpointed(ci);
810d5aeb 3679
bd3e7610
JB		 3680 BUG_ON(new_level != DLM_LOCK_NL && new_level != DLM_LOCK_PR);
3681 BUG_ON(lockres->l_level != DLM_LOCK_EX && !checkpointed);
810d5aeb
MF		 3682
3683 if (checkpointed)
3684 return 1;
3685
a4338481 3686 ocfs2_start_checkpoint(OCFS2_SB(ocfs2_metadata_cache_get_super(ci)));
810d5aeb
MF		 3687 return 0;
3688}
3689
a4338481
TM		 3690static int ocfs2_check_meta_downconvert(struct ocfs2_lock_res *lockres,
3691 int new_level)
3692{
3693 struct inode *inode = ocfs2_lock_res_inode(lockres);
3694
3695 return ocfs2_ci_checkpointed(INODE_CACHE(inode), lockres, new_level);
3696}
3697
810d5aeb
MF		 3698static void ocfs2_set_meta_lvb(struct ocfs2_lock_res *lockres)
3699{
3700 struct inode *inode = ocfs2_lock_res_inode(lockres);
3701
3702 __ocfs2_stuff_meta_lvb(inode);
3703}
3704
d680efe9
MF		 3705/*
3706 * Does the final reference drop on our dentry lock. Right now this
34d024f8 3707 * happens in the downconvert thread, but we could choose to simplify the
d680efe9
MF		 3708 * dlmglue API and push these off to the ocfs2_wq in the future.
3709 */
3710static void ocfs2_dentry_post_unlock(struct ocfs2_super *osb,
3711 struct ocfs2_lock_res *lockres)
3712{
3713 struct ocfs2_dentry_lock *dl = ocfs2_lock_res_dl(lockres);
3714 ocfs2_dentry_lock_put(osb, dl);
3715}
3716
3717/*
3718 * d_delete() matching dentries before the lock downconvert.
3719 *
3720 * At this point, any process waiting to destroy the
3721 * dentry_lock due to last ref count is stopped by the
3722 * OCFS2_LOCK_QUEUED flag.
3723 *
3724 * We have two potential problems
3725 *
3726 * 1) If we do the last reference drop on our dentry_lock (via dput)
3727 * we'll wind up in ocfs2_release_dentry_lock(), waiting on
3728 * the downconvert to finish. Instead we take an elevated
3729 * reference and push the drop until after we've completed our
3730 * unblock processing.
3731 *
3732 * 2) There might be another process with a final reference,
3733 * waiting on us to finish processing. If this is the case, we
3734 * detect it and exit out - there's no more dentries anyway.
3735 */
3736static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res *lockres,
3737 int blocking)
3738{
3739 struct ocfs2_dentry_lock *dl = ocfs2_lock_res_dl(lockres);
3740 struct ocfs2_inode_info *oi = OCFS2_I(dl->dl_inode);
3741 struct dentry *dentry;
3742 unsigned long flags;
3743 int extra_ref = 0;
3744
3745 /*
3746 * This node is blocking another node from getting a read
3747 * lock. This happens when we've renamed within a
3748 * directory. We've forced the other nodes to d_delete(), but
3749 * we never actually dropped our lock because it's still
3750 * valid. The downconvert code will retain a PR for this node,
3751 * so there's no further work to do.
3752 */
bd3e7610 3753 if (blocking == DLM_LOCK_PR)
d680efe9
MF		 3754 return UNBLOCK_CONTINUE;
3755
3756 /*
3757 * Mark this inode as potentially orphaned. The code in
3758 * ocfs2_delete_inode() will figure out whether it actually
3759 * needs to be freed or not.
3760 */
3761 spin_lock(&oi->ip_lock);
3762 oi->ip_flags |= OCFS2_INODE_MAYBE_ORPHANED;
3763 spin_unlock(&oi->ip_lock);
3764
3765 /*
3766 * Yuck. We need to make sure however that the check of
3767 * OCFS2_LOCK_FREEING and the extra reference are atomic with
3768 * respect to a reference decrement or the setting of that
3769 * flag.
3770 */
3771 spin_lock_irqsave(&lockres->l_lock, flags);
3772 spin_lock(&dentry_attach_lock);
3773 if (!(lockres->l_flags & OCFS2_LOCK_FREEING)
3774 && dl->dl_count) {
3775 dl->dl_count++;
3776 extra_ref = 1;
3777 }
3778 spin_unlock(&dentry_attach_lock);
3779 spin_unlock_irqrestore(&lockres->l_lock, flags);
3780
3781 mlog(0, "extra_ref = %d\n", extra_ref);
3782
3783 /*
3784 * We have a process waiting on us in ocfs2_dentry_iput(),
3785 * which means we can't have any more outstanding
3786 * aliases. There's no need to do any more work.
3787 */
3788 if (!extra_ref)
3789 return UNBLOCK_CONTINUE;
3790
3791 spin_lock(&dentry_attach_lock);
3792 while (1) {
3793 dentry = ocfs2_find_local_alias(dl->dl_inode,
3794 dl->dl_parent_blkno, 1);
3795 if (!dentry)
3796 break;
3797 spin_unlock(&dentry_attach_lock);
3798
3799 mlog(0, "d_delete(%.*s);\n", dentry->d_name.len,
3800 dentry->d_name.name);
3801
3802 /*
3803 * The following dcache calls may do an
3804 * iput(). Normally we don't want that from the
3805 * downconverting thread, but in this case it's ok
3806 * because the requesting node already has an
3807 * exclusive lock on the inode, so it can't be queued
3808 * for a downconvert.
3809 */
3810 d_delete(dentry);
3811 dput(dentry);
3812
3813 spin_lock(&dentry_attach_lock);
3814 }
3815 spin_unlock(&dentry_attach_lock);
3816
3817 /*
3818 * If we are the last holder of this dentry lock, there is no
3819 * reason to downconvert so skip straight to the unlock.
3820 */
3821 if (dl->dl_count == 1)
3822 return UNBLOCK_STOP_POST;
3823
3824 return UNBLOCK_CONTINUE_POST;
3825}
3826
8dec98ed
TM		 3827static int ocfs2_check_refcount_downconvert(struct ocfs2_lock_res *lockres,
3828 int new_level)
3829{
3830 struct ocfs2_refcount_tree *tree =
3831 ocfs2_lock_res_refcount_tree(lockres);
3832
3833 return ocfs2_ci_checkpointed(&tree->rf_ci, lockres, new_level);
3834}
3835
3836static int ocfs2_refcount_convert_worker(struct ocfs2_lock_res *lockres,
3837 int blocking)
3838{
3839 struct ocfs2_refcount_tree *tree =
3840 ocfs2_lock_res_refcount_tree(lockres);
3841
3842 ocfs2_metadata_cache_purge(&tree->rf_ci);
3843
3844 return UNBLOCK_CONTINUE;
3845}
3846
9e33d69f
JK		 3847static void ocfs2_set_qinfo_lvb(struct ocfs2_lock_res *lockres)
3848{
3849 struct ocfs2_qinfo_lvb *lvb;
3850 struct ocfs2_mem_dqinfo *oinfo = ocfs2_lock_res_qinfo(lockres);
3851 struct mem_dqinfo *info = sb_dqinfo(oinfo->dqi_gi.dqi_sb,
3852 oinfo->dqi_gi.dqi_type);
3853
3854 mlog_entry_void();
3855
a641dc2a 3856 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
9e33d69f
JK		 3857 lvb->lvb_version = OCFS2_QINFO_LVB_VERSION;
3858 lvb->lvb_bgrace = cpu_to_be32(info->dqi_bgrace);
3859 lvb->lvb_igrace = cpu_to_be32(info->dqi_igrace);
3860 lvb->lvb_syncms = cpu_to_be32(oinfo->dqi_syncms);
3861 lvb->lvb_blocks = cpu_to_be32(oinfo->dqi_gi.dqi_blocks);
3862 lvb->lvb_free_blk = cpu_to_be32(oinfo->dqi_gi.dqi_free_blk);
3863 lvb->lvb_free_entry = cpu_to_be32(oinfo->dqi_gi.dqi_free_entry);
3864
3865 mlog_exit_void();
3866}
3867
3868void ocfs2_qinfo_unlock(struct ocfs2_mem_dqinfo *oinfo, int ex)
3869{
3870 struct ocfs2_lock_res *lockres = &oinfo->dqi_gqlock;
3871 struct ocfs2_super *osb = OCFS2_SB(oinfo->dqi_gi.dqi_sb);
3872 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
3873
3874 mlog_entry_void();
3875 if (!ocfs2_is_hard_readonly(osb) && !ocfs2_mount_local(osb))
3876 ocfs2_cluster_unlock(osb, lockres, level);
3877 mlog_exit_void();
3878}
3879
3880static int ocfs2_refresh_qinfo(struct ocfs2_mem_dqinfo *oinfo)
3881{
3882 struct mem_dqinfo *info = sb_dqinfo(oinfo->dqi_gi.dqi_sb,
3883 oinfo->dqi_gi.dqi_type);
3884 struct ocfs2_lock_res *lockres = &oinfo->dqi_gqlock;
3885 struct ocfs2_qinfo_lvb *lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
85eb8b73 3886 struct buffer_head *bh = NULL;
9e33d69f
JK		 3887 struct ocfs2_global_disk_dqinfo *gdinfo;
3888 int status = 0;
3889
1c520dfb
JB		 3890 if (ocfs2_dlm_lvb_valid(&lockres->l_lksb) &&
3891 lvb->lvb_version == OCFS2_QINFO_LVB_VERSION) {
9e33d69f
JK		 3892 info->dqi_bgrace = be32_to_cpu(lvb->lvb_bgrace);
3893 info->dqi_igrace = be32_to_cpu(lvb->lvb_igrace);
3894 oinfo->dqi_syncms = be32_to_cpu(lvb->lvb_syncms);
3895 oinfo->dqi_gi.dqi_blocks = be32_to_cpu(lvb->lvb_blocks);
3896 oinfo->dqi_gi.dqi_free_blk = be32_to_cpu(lvb->lvb_free_blk);
3897 oinfo->dqi_gi.dqi_free_entry =
3898 be32_to_cpu(lvb->lvb_free_entry);
3899 } else {
ae4f6ef1
JK		 3900 status = ocfs2_read_quota_phys_block(oinfo->dqi_gqinode,
3901 oinfo->dqi_giblk, &bh);
85eb8b73 3902 if (status) {
9e33d69f
JK		 3903 mlog_errno(status);
3904 goto bail;
3905 }
3906 gdinfo = (struct ocfs2_global_disk_dqinfo *)
3907 (bh->b_data + OCFS2_GLOBAL_INFO_OFF);
3908 info->dqi_bgrace = le32_to_cpu(gdinfo->dqi_bgrace);
3909 info->dqi_igrace = le32_to_cpu(gdinfo->dqi_igrace);
3910 oinfo->dqi_syncms = le32_to_cpu(gdinfo->dqi_syncms);
3911 oinfo->dqi_gi.dqi_blocks = le32_to_cpu(gdinfo->dqi_blocks);
3912 oinfo->dqi_gi.dqi_free_blk = le32_to_cpu(gdinfo->dqi_free_blk);
3913 oinfo->dqi_gi.dqi_free_entry =
3914 le32_to_cpu(gdinfo->dqi_free_entry);
3915 brelse(bh);
3916 ocfs2_track_lock_refresh(lockres);
3917 }
3918
3919bail:
3920 return status;
3921}
3922
3923/* Lock quota info, this function expects at least shared lock on the quota file
3924 * so that we can safely refresh quota info from disk. */
3925int ocfs2_qinfo_lock(struct ocfs2_mem_dqinfo *oinfo, int ex)
3926{
3927 struct ocfs2_lock_res *lockres = &oinfo->dqi_gqlock;
3928 struct ocfs2_super *osb = OCFS2_SB(oinfo->dqi_gi.dqi_sb);
3929 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
3930 int status = 0;
3931
3932 mlog_entry_void();
3933
3934 /* On RO devices, locking really isn't needed... */
3935 if (ocfs2_is_hard_readonly(osb)) {
3936 if (ex)
3937 status = -EROFS;
3938 goto bail;
3939 }
3940 if (ocfs2_mount_local(osb))
3941 goto bail;
3942
3943 status = ocfs2_cluster_lock(osb, lockres, level, 0, 0);
3944 if (status < 0) {
3945 mlog_errno(status);
3946 goto bail;
3947 }
3948 if (!ocfs2_should_refresh_lock_res(lockres))
3949 goto bail;
3950 /* OK, we have the lock but we need to refresh the quota info */
3951 status = ocfs2_refresh_qinfo(oinfo);
3952 if (status)
3953 ocfs2_qinfo_unlock(oinfo, ex);
3954 ocfs2_complete_lock_res_refresh(lockres, status);
3955bail:
3956 mlog_exit(status);
3957 return status;
3958}
3959
8dec98ed
TM		 3960int ocfs2_refcount_lock(struct ocfs2_refcount_tree *ref_tree, int ex)
3961{
3962 int status;
3963 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
3964 struct ocfs2_lock_res *lockres = &ref_tree->rf_lockres;
3965 struct ocfs2_super *osb = lockres->l_priv;
3966
3967
3968 if (ocfs2_is_hard_readonly(osb))
3969 return -EROFS;
3970
3971 if (ocfs2_mount_local(osb))
3972 return 0;
3973
3974 status = ocfs2_cluster_lock(osb, lockres, level, 0, 0);
3975 if (status < 0)
3976 mlog_errno(status);
3977
3978 return status;
3979}
3980
3981void ocfs2_refcount_unlock(struct ocfs2_refcount_tree *ref_tree, int ex)
3982{
3983 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
3984 struct ocfs2_lock_res *lockres = &ref_tree->rf_lockres;
3985 struct ocfs2_super *osb = lockres->l_priv;
3986
3987 if (!ocfs2_mount_local(osb))
3988 ocfs2_cluster_unlock(osb, lockres, level);
3989}
3990
00600056
AB		 3991static void ocfs2_process_blocked_lock(struct ocfs2_super *osb,
3992 struct ocfs2_lock_res *lockres)
ccd979bd
MF		 3993{
3994 int status;
d680efe9 3995 struct ocfs2_unblock_ctl ctl = {0, 0,};
ccd979bd
MF		 3996 unsigned long flags;
3997
3998 /* Our reference to the lockres in this function can be
3999 * considered valid until we remove the OCFS2_LOCK_QUEUED
4000 * flag. */
4001
4002 mlog_entry_void();
4003
4004 BUG_ON(!lockres);
4005 BUG_ON(!lockres->l_ops);
ccd979bd 4006
9b915181 4007 mlog(ML_BASTS, "lockres %s blocked\n", lockres->l_name);
ccd979bd
MF		 4008
4009 /* Detect whether a lock has been marked as going away while
34d024f8 4010 * the downconvert thread was processing other things. A lock can
ccd979bd
MF		 4011 * still be marked with OCFS2_LOCK_FREEING after this check,
4012 * but short circuiting here will still save us some
4013 * performance. */
4014 spin_lock_irqsave(&lockres->l_lock, flags);
4015 if (lockres->l_flags & OCFS2_LOCK_FREEING)
4016 goto unqueue;
4017 spin_unlock_irqrestore(&lockres->l_lock, flags);
4018
b5e500e2 4019 status = ocfs2_unblock_lock(osb, lockres, &ctl);
ccd979bd
MF		 4020 if (status < 0)
4021 mlog_errno(status);
4022
4023 spin_lock_irqsave(&lockres->l_lock, flags);
4024unqueue:
d680efe9 4025 if (lockres->l_flags & OCFS2_LOCK_FREEING || !ctl.requeue) {
ccd979bd
MF		 4026 lockres_clear_flags(lockres, OCFS2_LOCK_QUEUED);
4027 } else
4028 ocfs2_schedule_blocked_lock(osb, lockres);
4029
9b915181 4030 mlog(ML_BASTS, "lockres %s, requeue = %s.\n", lockres->l_name,
d680efe9 4031 ctl.requeue ? "yes" : "no");
ccd979bd
MF		 4032 spin_unlock_irqrestore(&lockres->l_lock, flags);
4033
d680efe9
MF		 4034 if (ctl.unblock_action != UNBLOCK_CONTINUE
4035 && lockres->l_ops->post_unlock)
4036 lockres->l_ops->post_unlock(osb, lockres);
4037
ccd979bd
MF		 4038 mlog_exit_void();
4039}
4040
4041static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb,
4042 struct ocfs2_lock_res *lockres)
4043{
4044 mlog_entry_void();
4045
4046 assert_spin_locked(&lockres->l_lock);
4047
4048 if (lockres->l_flags & OCFS2_LOCK_FREEING) {
4049 /* Do not schedule a lock for downconvert when it's on
4050 * the way to destruction - any nodes wanting access
4051 * to the resource will get it soon. */
9b915181 4052 mlog(ML_BASTS, "lockres %s won't be scheduled: flags 0x%lx\n",
ccd979bd
MF		 4053 lockres->l_name, lockres->l_flags);
4054 return;
4055 }
4056
4057 lockres_or_flags(lockres, OCFS2_LOCK_QUEUED);
4058
34d024f8 4059 spin_lock(&osb->dc_task_lock);
ccd979bd
MF		 4060 if (list_empty(&lockres->l_blocked_list)) {
4061 list_add_tail(&lockres->l_blocked_list,
4062 &osb->blocked_lock_list);
4063 osb->blocked_lock_count++;
4064 }
34d024f8 4065 spin_unlock(&osb->dc_task_lock);
ccd979bd
MF		 4066
4067 mlog_exit_void();
4068}
34d024f8
MF		 4069
4070static void ocfs2_downconvert_thread_do_work(struct ocfs2_super *osb)
4071{
4072 unsigned long processed;
4073 struct ocfs2_lock_res *lockres;
4074
4075 mlog_entry_void();
4076
4077 spin_lock(&osb->dc_task_lock);
4078 /* grab this early so we know to try again if a state change and
4079 * wake happens part-way through our work */
4080 osb->dc_work_sequence = osb->dc_wake_sequence;
4081
4082 processed = osb->blocked_lock_count;
4083 while (processed) {
4084 BUG_ON(list_empty(&osb->blocked_lock_list));
4085
4086 lockres = list_entry(osb->blocked_lock_list.next,
4087 struct ocfs2_lock_res, l_blocked_list);
4088 list_del_init(&lockres->l_blocked_list);
4089 osb->blocked_lock_count--;
4090 spin_unlock(&osb->dc_task_lock);
4091
4092 BUG_ON(!processed);
4093 processed--;
4094
4095 ocfs2_process_blocked_lock(osb, lockres);
4096
4097 spin_lock(&osb->dc_task_lock);
4098 }
4099 spin_unlock(&osb->dc_task_lock);
4100
4101 mlog_exit_void();
4102}
4103
4104static int ocfs2_downconvert_thread_lists_empty(struct ocfs2_super *osb)
4105{
4106 int empty = 0;
4107
4108 spin_lock(&osb->dc_task_lock);
4109 if (list_empty(&osb->blocked_lock_list))
4110 empty = 1;
4111
4112 spin_unlock(&osb->dc_task_lock);
4113 return empty;
4114}
4115
4116static int ocfs2_downconvert_thread_should_wake(struct ocfs2_super *osb)
4117{
4118 int should_wake = 0;
4119
4120 spin_lock(&osb->dc_task_lock);
4121 if (osb->dc_work_sequence != osb->dc_wake_sequence)
4122 should_wake = 1;
4123 spin_unlock(&osb->dc_task_lock);
4124
4125 return should_wake;
4126}
4127
200bfae3 4128static int ocfs2_downconvert_thread(void *arg)
34d024f8
MF		 4129{
4130 int status = 0;
4131 struct ocfs2_super *osb = arg;
4132
4133 /* only quit once we've been asked to stop and there is no more
4134 * work available */
4135 while (!(kthread_should_stop() &&
4136 ocfs2_downconvert_thread_lists_empty(osb))) {
4137
4138 wait_event_interruptible(osb->dc_event,
4139 ocfs2_downconvert_thread_should_wake(osb) ||
4140 kthread_should_stop());
4141
4142 mlog(0, "downconvert_thread: awoken\n");
4143
4144 ocfs2_downconvert_thread_do_work(osb);
4145 }
4146
4147 osb->dc_task = NULL;
4148 return status;
4149}
4150
4151void ocfs2_wake_downconvert_thread(struct ocfs2_super *osb)
4152{
4153 spin_lock(&osb->dc_task_lock);
4154 /* make sure the voting thread gets a swipe at whatever changes
4155 * the caller may have made to the voting state */
4156 osb->dc_wake_sequence++;
4157 spin_unlock(&osb->dc_task_lock);
4158 wake_up(&osb->dc_event);
4159}
kernel source for telekom puls (alcatel/mediatek)