1 /* -*- mode: c; c-basic-offset: 8; -*-
2 * vim: noexpandtab sw=8 ts=8 sts=0:
6 * Code which implements an OCFS2 specific interface to our DLM.
8 * Copyright (C) 2003, 2004 Oracle. All rights reserved.
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.
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.
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.
26 #include <linux/types.h>
27 #include <linux/slab.h>
28 #include <linux/highmem.h>
30 #include <linux/crc32.h>
31 #include <linux/kthread.h>
32 #include <linux/pagemap.h>
33 #include <linux/debugfs.h>
34 #include <linux/seq_file.h>
36 #include <cluster/heartbeat.h>
37 #include <cluster/nodemanager.h>
38 #include <cluster/tcp.h>
40 #define MLOG_MASK_PREFIX ML_DLM_GLUE
41 #include <cluster/masklog.h>
44 #include "ocfs2_lockingver.h"
49 #include "extent_map.h"
51 #include "heartbeat.h"
54 #include "stackglue.h"
59 #include "buffer_head_io.h"
61 struct ocfs2_mask_waiter
{
62 struct list_head mw_item
;
64 struct completion mw_complete
;
65 unsigned long mw_mask
;
66 unsigned long mw_goal
;
69 static struct ocfs2_super
*ocfs2_get_dentry_osb(struct ocfs2_lock_res
*lockres
);
70 static struct ocfs2_super
*ocfs2_get_inode_osb(struct ocfs2_lock_res
*lockres
);
71 static struct ocfs2_super
*ocfs2_get_file_osb(struct ocfs2_lock_res
*lockres
);
74 * Return value from ->downconvert_worker functions.
76 * These control the precise actions of ocfs2_unblock_lock()
77 * and ocfs2_process_blocked_lock()
80 enum ocfs2_unblock_action
{
81 UNBLOCK_CONTINUE
= 0, /* Continue downconvert */
82 UNBLOCK_CONTINUE_POST
= 1, /* Continue downconvert, fire
83 * ->post_unlock callback */
84 UNBLOCK_STOP_POST
= 2, /* Do not downconvert, fire
85 * ->post_unlock() callback. */
88 struct ocfs2_unblock_ctl
{
90 enum ocfs2_unblock_action unblock_action
;
93 static int ocfs2_check_meta_downconvert(struct ocfs2_lock_res
*lockres
,
95 static void ocfs2_set_meta_lvb(struct ocfs2_lock_res
*lockres
);
97 static int ocfs2_data_convert_worker(struct ocfs2_lock_res
*lockres
,
100 static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res
*lockres
,
103 static void ocfs2_dentry_post_unlock(struct ocfs2_super
*osb
,
104 struct ocfs2_lock_res
*lockres
);
107 #define mlog_meta_lvb(__level, __lockres) ocfs2_dump_meta_lvb_info(__level, __PRETTY_FUNCTION__, __LINE__, __lockres)
109 /* This aids in debugging situations where a bad LVB might be involved. */
110 static void ocfs2_dump_meta_lvb_info(u64 level
,
111 const char *function
,
113 struct ocfs2_lock_res
*lockres
)
115 struct ocfs2_meta_lvb
*lvb
=
116 (struct ocfs2_meta_lvb
*)ocfs2_dlm_lvb(&lockres
->l_lksb
);
118 mlog(level
, "LVB information for %s (called from %s:%u):\n",
119 lockres
->l_name
, function
, line
);
120 mlog(level
, "version: %u, clusters: %u, generation: 0x%x\n",
121 lvb
->lvb_version
, be32_to_cpu(lvb
->lvb_iclusters
),
122 be32_to_cpu(lvb
->lvb_igeneration
));
123 mlog(level
, "size: %llu, uid %u, gid %u, mode 0x%x\n",
124 (unsigned long long)be64_to_cpu(lvb
->lvb_isize
),
125 be32_to_cpu(lvb
->lvb_iuid
), be32_to_cpu(lvb
->lvb_igid
),
126 be16_to_cpu(lvb
->lvb_imode
));
127 mlog(level
, "nlink %u, atime_packed 0x%llx, ctime_packed 0x%llx, "
128 "mtime_packed 0x%llx iattr 0x%x\n", be16_to_cpu(lvb
->lvb_inlink
),
129 (long long)be64_to_cpu(lvb
->lvb_iatime_packed
),
130 (long long)be64_to_cpu(lvb
->lvb_ictime_packed
),
131 (long long)be64_to_cpu(lvb
->lvb_imtime_packed
),
132 be32_to_cpu(lvb
->lvb_iattr
));
137 * OCFS2 Lock Resource Operations
139 * These fine tune the behavior of the generic dlmglue locking infrastructure.
141 * The most basic of lock types can point ->l_priv to their respective
142 * struct ocfs2_super and allow the default actions to manage things.
144 * Right now, each lock type also needs to implement an init function,
145 * and trivial lock/unlock wrappers. ocfs2_simple_drop_lockres()
146 * should be called when the lock is no longer needed (i.e., object
149 struct ocfs2_lock_res_ops
{
151 * Translate an ocfs2_lock_res * into an ocfs2_super *. Define
152 * this callback if ->l_priv is not an ocfs2_super pointer
154 struct ocfs2_super
* (*get_osb
)(struct ocfs2_lock_res
*);
157 * Optionally called in the downconvert thread after a
158 * successful downconvert. The lockres will not be referenced
159 * after this callback is called, so it is safe to free
162 * The exact semantics of when this is called are controlled
163 * by ->downconvert_worker()
165 void (*post_unlock
)(struct ocfs2_super
*, struct ocfs2_lock_res
*);
168 * Allow a lock type to add checks to determine whether it is
169 * safe to downconvert a lock. Return 0 to re-queue the
170 * downconvert at a later time, nonzero to continue.
172 * For most locks, the default checks that there are no
173 * incompatible holders are sufficient.
175 * Called with the lockres spinlock held.
177 int (*check_downconvert
)(struct ocfs2_lock_res
*, int);
180 * Allows a lock type to populate the lock value block. This
181 * is called on downconvert, and when we drop a lock.
183 * Locks that want to use this should set LOCK_TYPE_USES_LVB
184 * in the flags field.
186 * Called with the lockres spinlock held.
188 void (*set_lvb
)(struct ocfs2_lock_res
*);
191 * Called from the downconvert thread when it is determined
192 * that a lock will be downconverted. This is called without
193 * any locks held so the function can do work that might
194 * schedule (syncing out data, etc).
196 * This should return any one of the ocfs2_unblock_action
197 * values, depending on what it wants the thread to do.
199 int (*downconvert_worker
)(struct ocfs2_lock_res
*, int);
202 * LOCK_TYPE_* flags which describe the specific requirements
203 * of a lock type. Descriptions of each individual flag follow.
209 * Some locks want to "refresh" potentially stale data when a
210 * meaningful (PRMODE or EXMODE) lock level is first obtained. If this
211 * flag is set, the OCFS2_LOCK_NEEDS_REFRESH flag will be set on the
212 * individual lockres l_flags member from the ast function. It is
213 * expected that the locking wrapper will clear the
214 * OCFS2_LOCK_NEEDS_REFRESH flag when done.
216 #define LOCK_TYPE_REQUIRES_REFRESH 0x1
219 * Indicate that a lock type makes use of the lock value block. The
220 * ->set_lvb lock type callback must be defined.
222 #define LOCK_TYPE_USES_LVB 0x2
224 static struct ocfs2_lock_res_ops ocfs2_inode_rw_lops
= {
225 .get_osb
= ocfs2_get_inode_osb
,
229 static struct ocfs2_lock_res_ops ocfs2_inode_inode_lops
= {
230 .get_osb
= ocfs2_get_inode_osb
,
231 .check_downconvert
= ocfs2_check_meta_downconvert
,
232 .set_lvb
= ocfs2_set_meta_lvb
,
233 .downconvert_worker
= ocfs2_data_convert_worker
,
234 .flags
= LOCK_TYPE_REQUIRES_REFRESH
|LOCK_TYPE_USES_LVB
,
237 static struct ocfs2_lock_res_ops ocfs2_super_lops
= {
238 .flags
= LOCK_TYPE_REQUIRES_REFRESH
,
241 static struct ocfs2_lock_res_ops ocfs2_rename_lops
= {
245 static struct ocfs2_lock_res_ops ocfs2_dentry_lops
= {
246 .get_osb
= ocfs2_get_dentry_osb
,
247 .post_unlock
= ocfs2_dentry_post_unlock
,
248 .downconvert_worker
= ocfs2_dentry_convert_worker
,
252 static struct ocfs2_lock_res_ops ocfs2_inode_open_lops
= {
253 .get_osb
= ocfs2_get_inode_osb
,
257 static struct ocfs2_lock_res_ops ocfs2_flock_lops
= {
258 .get_osb
= ocfs2_get_file_osb
,
263 * This is the filesystem locking protocol version.
265 * Whenever the filesystem does new things with locks (adds or removes a
266 * lock, orders them differently, does different things underneath a lock),
267 * the version must be changed. The protocol is negotiated when joining
268 * the dlm domain. A node may join the domain if its major version is
269 * identical to all other nodes and its minor version is greater than
270 * or equal to all other nodes. When its minor version is greater than
271 * the other nodes, it will run at the minor version specified by the
274 * If a locking change is made that will not be compatible with older
275 * versions, the major number must be increased and the minor version set
276 * to zero. If a change merely adds a behavior that can be disabled when
277 * speaking to older versions, the minor version must be increased. If a
278 * change adds a fully backwards compatible change (eg, LVB changes that
279 * are just ignored by older versions), the version does not need to be
282 const struct dlm_protocol_version ocfs2_locking_protocol
= {
283 .pv_major
= OCFS2_LOCKING_PROTOCOL_MAJOR
,
284 .pv_minor
= OCFS2_LOCKING_PROTOCOL_MINOR
,
287 static inline int ocfs2_is_inode_lock(struct ocfs2_lock_res
*lockres
)
289 return lockres
->l_type
== OCFS2_LOCK_TYPE_META
||
290 lockres
->l_type
== OCFS2_LOCK_TYPE_RW
||
291 lockres
->l_type
== OCFS2_LOCK_TYPE_OPEN
;
294 static inline struct inode
*ocfs2_lock_res_inode(struct ocfs2_lock_res
*lockres
)
296 BUG_ON(!ocfs2_is_inode_lock(lockres
));
298 return (struct inode
*) lockres
->l_priv
;
301 static inline struct ocfs2_dentry_lock
*ocfs2_lock_res_dl(struct ocfs2_lock_res
*lockres
)
303 BUG_ON(lockres
->l_type
!= OCFS2_LOCK_TYPE_DENTRY
);
305 return (struct ocfs2_dentry_lock
*)lockres
->l_priv
;
308 static inline struct ocfs2_super
*ocfs2_get_lockres_osb(struct ocfs2_lock_res
*lockres
)
310 if (lockres
->l_ops
->get_osb
)
311 return lockres
->l_ops
->get_osb(lockres
);
313 return (struct ocfs2_super
*)lockres
->l_priv
;
316 static int ocfs2_lock_create(struct ocfs2_super
*osb
,
317 struct ocfs2_lock_res
*lockres
,
320 static inline int ocfs2_may_continue_on_blocked_lock(struct ocfs2_lock_res
*lockres
,
322 static void ocfs2_cluster_unlock(struct ocfs2_super
*osb
,
323 struct ocfs2_lock_res
*lockres
,
325 static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res
*lockres
);
326 static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res
*lockres
);
327 static inline void ocfs2_generic_handle_attach_action(struct ocfs2_lock_res
*lockres
);
328 static int ocfs2_generic_handle_bast(struct ocfs2_lock_res
*lockres
, int level
);
329 static void ocfs2_schedule_blocked_lock(struct ocfs2_super
*osb
,
330 struct ocfs2_lock_res
*lockres
);
331 static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res
*lockres
,
333 #define ocfs2_log_dlm_error(_func, _err, _lockres) do { \
334 mlog(ML_ERROR, "DLM error %d while calling %s on resource %s\n", \
335 _err, _func, _lockres->l_name); \
337 static int ocfs2_downconvert_thread(void *arg
);
338 static void ocfs2_downconvert_on_unlock(struct ocfs2_super
*osb
,
339 struct ocfs2_lock_res
*lockres
);
340 static int ocfs2_inode_lock_update(struct inode
*inode
,
341 struct buffer_head
**bh
);
342 static void ocfs2_drop_osb_locks(struct ocfs2_super
*osb
);
343 static inline int ocfs2_highest_compat_lock_level(int level
);
344 static void ocfs2_prepare_downconvert(struct ocfs2_lock_res
*lockres
,
346 static int ocfs2_downconvert_lock(struct ocfs2_super
*osb
,
347 struct ocfs2_lock_res
*lockres
,
350 static int ocfs2_prepare_cancel_convert(struct ocfs2_super
*osb
,
351 struct ocfs2_lock_res
*lockres
);
352 static int ocfs2_cancel_convert(struct ocfs2_super
*osb
,
353 struct ocfs2_lock_res
*lockres
);
356 static void ocfs2_build_lock_name(enum ocfs2_lock_type type
,
365 BUG_ON(type
>= OCFS2_NUM_LOCK_TYPES
);
367 len
= snprintf(name
, OCFS2_LOCK_ID_MAX_LEN
, "%c%s%016llx%08x",
368 ocfs2_lock_type_char(type
), OCFS2_LOCK_ID_PAD
,
369 (long long)blkno
, generation
);
371 BUG_ON(len
!= (OCFS2_LOCK_ID_MAX_LEN
- 1));
373 mlog(0, "built lock resource with name: %s\n", name
);
378 static DEFINE_SPINLOCK(ocfs2_dlm_tracking_lock
);
380 static void ocfs2_add_lockres_tracking(struct ocfs2_lock_res
*res
,
381 struct ocfs2_dlm_debug
*dlm_debug
)
383 mlog(0, "Add tracking for lockres %s\n", res
->l_name
);
385 spin_lock(&ocfs2_dlm_tracking_lock
);
386 list_add(&res
->l_debug_list
, &dlm_debug
->d_lockres_tracking
);
387 spin_unlock(&ocfs2_dlm_tracking_lock
);
390 static void ocfs2_remove_lockres_tracking(struct ocfs2_lock_res
*res
)
392 spin_lock(&ocfs2_dlm_tracking_lock
);
393 if (!list_empty(&res
->l_debug_list
))
394 list_del_init(&res
->l_debug_list
);
395 spin_unlock(&ocfs2_dlm_tracking_lock
);
398 static void ocfs2_lock_res_init_common(struct ocfs2_super
*osb
,
399 struct ocfs2_lock_res
*res
,
400 enum ocfs2_lock_type type
,
401 struct ocfs2_lock_res_ops
*ops
,
408 res
->l_level
= DLM_LOCK_IV
;
409 res
->l_requested
= DLM_LOCK_IV
;
410 res
->l_blocking
= DLM_LOCK_IV
;
411 res
->l_action
= OCFS2_AST_INVALID
;
412 res
->l_unlock_action
= OCFS2_UNLOCK_INVALID
;
414 res
->l_flags
= OCFS2_LOCK_INITIALIZED
;
416 ocfs2_add_lockres_tracking(res
, osb
->osb_dlm_debug
);
419 void ocfs2_lock_res_init_once(struct ocfs2_lock_res
*res
)
421 /* This also clears out the lock status block */
422 memset(res
, 0, sizeof(struct ocfs2_lock_res
));
423 spin_lock_init(&res
->l_lock
);
424 init_waitqueue_head(&res
->l_event
);
425 INIT_LIST_HEAD(&res
->l_blocked_list
);
426 INIT_LIST_HEAD(&res
->l_mask_waiters
);
429 void ocfs2_inode_lock_res_init(struct ocfs2_lock_res
*res
,
430 enum ocfs2_lock_type type
,
431 unsigned int generation
,
434 struct ocfs2_lock_res_ops
*ops
;
437 case OCFS2_LOCK_TYPE_RW
:
438 ops
= &ocfs2_inode_rw_lops
;
440 case OCFS2_LOCK_TYPE_META
:
441 ops
= &ocfs2_inode_inode_lops
;
443 case OCFS2_LOCK_TYPE_OPEN
:
444 ops
= &ocfs2_inode_open_lops
;
447 mlog_bug_on_msg(1, "type: %d\n", type
);
448 ops
= NULL
; /* thanks, gcc */
452 ocfs2_build_lock_name(type
, OCFS2_I(inode
)->ip_blkno
,
453 generation
, res
->l_name
);
454 ocfs2_lock_res_init_common(OCFS2_SB(inode
->i_sb
), res
, type
, ops
, inode
);
457 static struct ocfs2_super
*ocfs2_get_inode_osb(struct ocfs2_lock_res
*lockres
)
459 struct inode
*inode
= ocfs2_lock_res_inode(lockres
);
461 return OCFS2_SB(inode
->i_sb
);
464 static struct ocfs2_super
*ocfs2_get_file_osb(struct ocfs2_lock_res
*lockres
)
466 struct ocfs2_file_private
*fp
= lockres
->l_priv
;
468 return OCFS2_SB(fp
->fp_file
->f_mapping
->host
->i_sb
);
471 static __u64
ocfs2_get_dentry_lock_ino(struct ocfs2_lock_res
*lockres
)
473 __be64 inode_blkno_be
;
475 memcpy(&inode_blkno_be
, &lockres
->l_name
[OCFS2_DENTRY_LOCK_INO_START
],
478 return be64_to_cpu(inode_blkno_be
);
481 static struct ocfs2_super
*ocfs2_get_dentry_osb(struct ocfs2_lock_res
*lockres
)
483 struct ocfs2_dentry_lock
*dl
= lockres
->l_priv
;
485 return OCFS2_SB(dl
->dl_inode
->i_sb
);
488 void ocfs2_dentry_lock_res_init(struct ocfs2_dentry_lock
*dl
,
489 u64 parent
, struct inode
*inode
)
492 u64 inode_blkno
= OCFS2_I(inode
)->ip_blkno
;
493 __be64 inode_blkno_be
= cpu_to_be64(inode_blkno
);
494 struct ocfs2_lock_res
*lockres
= &dl
->dl_lockres
;
496 ocfs2_lock_res_init_once(lockres
);
499 * Unfortunately, the standard lock naming scheme won't work
500 * here because we have two 16 byte values to use. Instead,
501 * we'll stuff the inode number as a binary value. We still
502 * want error prints to show something without garbling the
503 * display, so drop a null byte in there before the inode
504 * number. A future version of OCFS2 will likely use all
505 * binary lock names. The stringified names have been a
506 * tremendous aid in debugging, but now that the debugfs
507 * interface exists, we can mangle things there if need be.
509 * NOTE: We also drop the standard "pad" value (the total lock
510 * name size stays the same though - the last part is all
511 * zeros due to the memset in ocfs2_lock_res_init_once()
513 len
= snprintf(lockres
->l_name
, OCFS2_DENTRY_LOCK_INO_START
,
515 ocfs2_lock_type_char(OCFS2_LOCK_TYPE_DENTRY
),
518 BUG_ON(len
!= (OCFS2_DENTRY_LOCK_INO_START
- 1));
520 memcpy(&lockres
->l_name
[OCFS2_DENTRY_LOCK_INO_START
], &inode_blkno_be
,
523 ocfs2_lock_res_init_common(OCFS2_SB(inode
->i_sb
), lockres
,
524 OCFS2_LOCK_TYPE_DENTRY
, &ocfs2_dentry_lops
,
528 static void ocfs2_super_lock_res_init(struct ocfs2_lock_res
*res
,
529 struct ocfs2_super
*osb
)
531 /* Superblock lockres doesn't come from a slab so we call init
532 * once on it manually. */
533 ocfs2_lock_res_init_once(res
);
534 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_SUPER
, OCFS2_SUPER_BLOCK_BLKNO
,
536 ocfs2_lock_res_init_common(osb
, res
, OCFS2_LOCK_TYPE_SUPER
,
537 &ocfs2_super_lops
, osb
);
540 static void ocfs2_rename_lock_res_init(struct ocfs2_lock_res
*res
,
541 struct ocfs2_super
*osb
)
543 /* Rename lockres doesn't come from a slab so we call init
544 * once on it manually. */
545 ocfs2_lock_res_init_once(res
);
546 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_RENAME
, 0, 0, res
->l_name
);
547 ocfs2_lock_res_init_common(osb
, res
, OCFS2_LOCK_TYPE_RENAME
,
548 &ocfs2_rename_lops
, osb
);
551 void ocfs2_file_lock_res_init(struct ocfs2_lock_res
*lockres
,
552 struct ocfs2_file_private
*fp
)
554 struct inode
*inode
= fp
->fp_file
->f_mapping
->host
;
555 struct ocfs2_inode_info
*oi
= OCFS2_I(inode
);
557 ocfs2_lock_res_init_once(lockres
);
558 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_FLOCK
, oi
->ip_blkno
,
559 inode
->i_generation
, lockres
->l_name
);
560 ocfs2_lock_res_init_common(OCFS2_SB(inode
->i_sb
), lockres
,
561 OCFS2_LOCK_TYPE_FLOCK
, &ocfs2_flock_lops
,
563 lockres
->l_flags
|= OCFS2_LOCK_NOCACHE
;
566 void ocfs2_lock_res_free(struct ocfs2_lock_res
*res
)
570 if (!(res
->l_flags
& OCFS2_LOCK_INITIALIZED
))
573 ocfs2_remove_lockres_tracking(res
);
575 mlog_bug_on_msg(!list_empty(&res
->l_blocked_list
),
576 "Lockres %s is on the blocked list\n",
578 mlog_bug_on_msg(!list_empty(&res
->l_mask_waiters
),
579 "Lockres %s has mask waiters pending\n",
581 mlog_bug_on_msg(spin_is_locked(&res
->l_lock
),
582 "Lockres %s is locked\n",
584 mlog_bug_on_msg(res
->l_ro_holders
,
585 "Lockres %s has %u ro holders\n",
586 res
->l_name
, res
->l_ro_holders
);
587 mlog_bug_on_msg(res
->l_ex_holders
,
588 "Lockres %s has %u ex holders\n",
589 res
->l_name
, res
->l_ex_holders
);
591 /* Need to clear out the lock status block for the dlm */
592 memset(&res
->l_lksb
, 0, sizeof(res
->l_lksb
));
598 static inline void ocfs2_inc_holders(struct ocfs2_lock_res
*lockres
,
607 lockres
->l_ex_holders
++;
610 lockres
->l_ro_holders
++;
619 static inline void ocfs2_dec_holders(struct ocfs2_lock_res
*lockres
,
628 BUG_ON(!lockres
->l_ex_holders
);
629 lockres
->l_ex_holders
--;
632 BUG_ON(!lockres
->l_ro_holders
);
633 lockres
->l_ro_holders
--;
641 /* WARNING: This function lives in a world where the only three lock
642 * levels are EX, PR, and NL. It *will* have to be adjusted when more
643 * lock types are added. */
644 static inline int ocfs2_highest_compat_lock_level(int level
)
646 int new_level
= DLM_LOCK_EX
;
648 if (level
== DLM_LOCK_EX
)
649 new_level
= DLM_LOCK_NL
;
650 else if (level
== DLM_LOCK_PR
)
651 new_level
= DLM_LOCK_PR
;
655 static void lockres_set_flags(struct ocfs2_lock_res
*lockres
,
656 unsigned long newflags
)
658 struct ocfs2_mask_waiter
*mw
, *tmp
;
660 assert_spin_locked(&lockres
->l_lock
);
662 lockres
->l_flags
= newflags
;
664 list_for_each_entry_safe(mw
, tmp
, &lockres
->l_mask_waiters
, mw_item
) {
665 if ((lockres
->l_flags
& mw
->mw_mask
) != mw
->mw_goal
)
668 list_del_init(&mw
->mw_item
);
670 complete(&mw
->mw_complete
);
673 static void lockres_or_flags(struct ocfs2_lock_res
*lockres
, unsigned long or)
675 lockres_set_flags(lockres
, lockres
->l_flags
| or);
677 static void lockres_clear_flags(struct ocfs2_lock_res
*lockres
,
680 lockres_set_flags(lockres
, lockres
->l_flags
& ~clear
);
683 static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res
*lockres
)
687 BUG_ON(!(lockres
->l_flags
& OCFS2_LOCK_BUSY
));
688 BUG_ON(!(lockres
->l_flags
& OCFS2_LOCK_ATTACHED
));
689 BUG_ON(!(lockres
->l_flags
& OCFS2_LOCK_BLOCKED
));
690 BUG_ON(lockres
->l_blocking
<= DLM_LOCK_NL
);
692 lockres
->l_level
= lockres
->l_requested
;
693 if (lockres
->l_level
<=
694 ocfs2_highest_compat_lock_level(lockres
->l_blocking
)) {
695 lockres
->l_blocking
= DLM_LOCK_NL
;
696 lockres_clear_flags(lockres
, OCFS2_LOCK_BLOCKED
);
698 lockres_clear_flags(lockres
, OCFS2_LOCK_BUSY
);
703 static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res
*lockres
)
707 BUG_ON(!(lockres
->l_flags
& OCFS2_LOCK_BUSY
));
708 BUG_ON(!(lockres
->l_flags
& OCFS2_LOCK_ATTACHED
));
710 /* Convert from RO to EX doesn't really need anything as our
711 * information is already up to data. Convert from NL to
712 * *anything* however should mark ourselves as needing an
714 if (lockres
->l_level
== DLM_LOCK_NL
&&
715 lockres
->l_ops
->flags
& LOCK_TYPE_REQUIRES_REFRESH
)
716 lockres_or_flags(lockres
, OCFS2_LOCK_NEEDS_REFRESH
);
718 lockres
->l_level
= lockres
->l_requested
;
719 lockres_clear_flags(lockres
, OCFS2_LOCK_BUSY
);
724 static inline void ocfs2_generic_handle_attach_action(struct ocfs2_lock_res
*lockres
)
728 BUG_ON((!(lockres
->l_flags
& OCFS2_LOCK_BUSY
)));
729 BUG_ON(lockres
->l_flags
& OCFS2_LOCK_ATTACHED
);
731 if (lockres
->l_requested
> DLM_LOCK_NL
&&
732 !(lockres
->l_flags
& OCFS2_LOCK_LOCAL
) &&
733 lockres
->l_ops
->flags
& LOCK_TYPE_REQUIRES_REFRESH
)
734 lockres_or_flags(lockres
, OCFS2_LOCK_NEEDS_REFRESH
);
736 lockres
->l_level
= lockres
->l_requested
;
737 lockres_or_flags(lockres
, OCFS2_LOCK_ATTACHED
);
738 lockres_clear_flags(lockres
, OCFS2_LOCK_BUSY
);
743 static int ocfs2_generic_handle_bast(struct ocfs2_lock_res
*lockres
,
746 int needs_downconvert
= 0;
749 assert_spin_locked(&lockres
->l_lock
);
751 lockres_or_flags(lockres
, OCFS2_LOCK_BLOCKED
);
753 if (level
> lockres
->l_blocking
) {
754 /* only schedule a downconvert if we haven't already scheduled
755 * one that goes low enough to satisfy the level we're
756 * blocking. this also catches the case where we get
758 if (ocfs2_highest_compat_lock_level(level
) <
759 ocfs2_highest_compat_lock_level(lockres
->l_blocking
))
760 needs_downconvert
= 1;
762 lockres
->l_blocking
= level
;
765 mlog_exit(needs_downconvert
);
766 return needs_downconvert
;
769 static void ocfs2_blocking_ast(void *opaque
, int level
)
771 struct ocfs2_lock_res
*lockres
= opaque
;
772 struct ocfs2_super
*osb
= ocfs2_get_lockres_osb(lockres
);
773 int needs_downconvert
;
776 BUG_ON(level
<= DLM_LOCK_NL
);
778 mlog(0, "BAST fired for lockres %s, blocking %d, level %d type %s\n",
779 lockres
->l_name
, level
, lockres
->l_level
,
780 ocfs2_lock_type_string(lockres
->l_type
));
783 * We can skip the bast for locks which don't enable caching -
784 * they'll be dropped at the earliest possible time anyway.
786 if (lockres
->l_flags
& OCFS2_LOCK_NOCACHE
)
789 spin_lock_irqsave(&lockres
->l_lock
, flags
);
790 needs_downconvert
= ocfs2_generic_handle_bast(lockres
, level
);
791 if (needs_downconvert
)
792 ocfs2_schedule_blocked_lock(osb
, lockres
);
793 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
795 wake_up(&lockres
->l_event
);
797 ocfs2_wake_downconvert_thread(osb
);
800 static void ocfs2_locking_ast(void *opaque
)
802 struct ocfs2_lock_res
*lockres
= opaque
;
805 spin_lock_irqsave(&lockres
->l_lock
, flags
);
807 if (ocfs2_dlm_lock_status(&lockres
->l_lksb
)) {
808 mlog(ML_ERROR
, "lockres %s: lksb status value of %d!\n",
810 ocfs2_dlm_lock_status(&lockres
->l_lksb
));
811 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
815 switch(lockres
->l_action
) {
816 case OCFS2_AST_ATTACH
:
817 ocfs2_generic_handle_attach_action(lockres
);
818 lockres_clear_flags(lockres
, OCFS2_LOCK_LOCAL
);
820 case OCFS2_AST_CONVERT
:
821 ocfs2_generic_handle_convert_action(lockres
);
823 case OCFS2_AST_DOWNCONVERT
:
824 ocfs2_generic_handle_downconvert_action(lockres
);
827 mlog(ML_ERROR
, "lockres %s: ast fired with invalid action: %u "
828 "lockres flags = 0x%lx, unlock action: %u\n",
829 lockres
->l_name
, lockres
->l_action
, lockres
->l_flags
,
830 lockres
->l_unlock_action
);
834 /* set it to something invalid so if we get called again we
836 lockres
->l_action
= OCFS2_AST_INVALID
;
838 wake_up(&lockres
->l_event
);
839 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
842 static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res
*lockres
,
848 spin_lock_irqsave(&lockres
->l_lock
, flags
);
849 lockres_clear_flags(lockres
, OCFS2_LOCK_BUSY
);
851 lockres
->l_action
= OCFS2_AST_INVALID
;
853 lockres
->l_unlock_action
= OCFS2_UNLOCK_INVALID
;
854 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
856 wake_up(&lockres
->l_event
);
860 /* Note: If we detect another process working on the lock (i.e.,
861 * OCFS2_LOCK_BUSY), we'll bail out returning 0. It's up to the caller
862 * to do the right thing in that case.
864 static int ocfs2_lock_create(struct ocfs2_super
*osb
,
865 struct ocfs2_lock_res
*lockres
,
874 mlog(0, "lock %s, level = %d, flags = %u\n", lockres
->l_name
, level
,
877 spin_lock_irqsave(&lockres
->l_lock
, flags
);
878 if ((lockres
->l_flags
& OCFS2_LOCK_ATTACHED
) ||
879 (lockres
->l_flags
& OCFS2_LOCK_BUSY
)) {
880 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
884 lockres
->l_action
= OCFS2_AST_ATTACH
;
885 lockres
->l_requested
= level
;
886 lockres_or_flags(lockres
, OCFS2_LOCK_BUSY
);
887 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
889 ret
= ocfs2_dlm_lock(osb
->dlm
,
894 OCFS2_LOCK_ID_MAX_LEN
- 1,
897 ocfs2_log_dlm_error("ocfs2_dlm_lock", ret
, lockres
);
898 ocfs2_recover_from_dlm_error(lockres
, 1);
901 mlog(0, "lock %s, return from ocfs2_dlm_lock\n", lockres
->l_name
);
908 static inline int ocfs2_check_wait_flag(struct ocfs2_lock_res
*lockres
,
914 spin_lock_irqsave(&lockres
->l_lock
, flags
);
915 ret
= lockres
->l_flags
& flag
;
916 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
921 static inline void ocfs2_wait_on_busy_lock(struct ocfs2_lock_res
*lockres
)
924 wait_event(lockres
->l_event
,
925 !ocfs2_check_wait_flag(lockres
, OCFS2_LOCK_BUSY
));
928 static inline void ocfs2_wait_on_refreshing_lock(struct ocfs2_lock_res
*lockres
)
931 wait_event(lockres
->l_event
,
932 !ocfs2_check_wait_flag(lockres
, OCFS2_LOCK_REFRESHING
));
935 /* predict what lock level we'll be dropping down to on behalf
936 * of another node, and return true if the currently wanted
937 * level will be compatible with it. */
938 static inline int ocfs2_may_continue_on_blocked_lock(struct ocfs2_lock_res
*lockres
,
941 BUG_ON(!(lockres
->l_flags
& OCFS2_LOCK_BLOCKED
));
943 return wanted
<= ocfs2_highest_compat_lock_level(lockres
->l_blocking
);
946 static void ocfs2_init_mask_waiter(struct ocfs2_mask_waiter
*mw
)
948 INIT_LIST_HEAD(&mw
->mw_item
);
949 init_completion(&mw
->mw_complete
);
952 static int ocfs2_wait_for_mask(struct ocfs2_mask_waiter
*mw
)
954 wait_for_completion(&mw
->mw_complete
);
955 /* Re-arm the completion in case we want to wait on it again */
956 INIT_COMPLETION(mw
->mw_complete
);
957 return mw
->mw_status
;
960 static void lockres_add_mask_waiter(struct ocfs2_lock_res
*lockres
,
961 struct ocfs2_mask_waiter
*mw
,
965 BUG_ON(!list_empty(&mw
->mw_item
));
967 assert_spin_locked(&lockres
->l_lock
);
969 list_add_tail(&mw
->mw_item
, &lockres
->l_mask_waiters
);
974 /* returns 0 if the mw that was removed was already satisfied, -EBUSY
975 * if the mask still hadn't reached its goal */
976 static int lockres_remove_mask_waiter(struct ocfs2_lock_res
*lockres
,
977 struct ocfs2_mask_waiter
*mw
)
982 spin_lock_irqsave(&lockres
->l_lock
, flags
);
983 if (!list_empty(&mw
->mw_item
)) {
984 if ((lockres
->l_flags
& mw
->mw_mask
) != mw
->mw_goal
)
987 list_del_init(&mw
->mw_item
);
988 init_completion(&mw
->mw_complete
);
990 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
996 static int ocfs2_wait_for_mask_interruptible(struct ocfs2_mask_waiter
*mw
,
997 struct ocfs2_lock_res
*lockres
)
1001 ret
= wait_for_completion_interruptible(&mw
->mw_complete
);
1003 lockres_remove_mask_waiter(lockres
, mw
);
1005 ret
= mw
->mw_status
;
1006 /* Re-arm the completion in case we want to wait on it again */
1007 INIT_COMPLETION(mw
->mw_complete
);
1011 static int ocfs2_cluster_lock(struct ocfs2_super
*osb
,
1012 struct ocfs2_lock_res
*lockres
,
1017 struct ocfs2_mask_waiter mw
;
1018 int wait
, catch_signals
= !(osb
->s_mount_opt
& OCFS2_MOUNT_NOINTR
);
1019 int ret
= 0; /* gcc doesn't realize wait = 1 guarantees ret is set */
1020 unsigned long flags
;
1024 ocfs2_init_mask_waiter(&mw
);
1026 if (lockres
->l_ops
->flags
& LOCK_TYPE_USES_LVB
)
1027 lkm_flags
|= DLM_LKF_VALBLK
;
1032 if (catch_signals
&& signal_pending(current
)) {
1037 spin_lock_irqsave(&lockres
->l_lock
, flags
);
1039 mlog_bug_on_msg(lockres
->l_flags
& OCFS2_LOCK_FREEING
,
1040 "Cluster lock called on freeing lockres %s! flags "
1041 "0x%lx\n", lockres
->l_name
, lockres
->l_flags
);
1043 /* We only compare against the currently granted level
1044 * here. If the lock is blocked waiting on a downconvert,
1045 * we'll get caught below. */
1046 if (lockres
->l_flags
& OCFS2_LOCK_BUSY
&&
1047 level
> lockres
->l_level
) {
1048 /* is someone sitting in dlm_lock? If so, wait on
1050 lockres_add_mask_waiter(lockres
, &mw
, OCFS2_LOCK_BUSY
, 0);
1055 if (lockres
->l_flags
& OCFS2_LOCK_BLOCKED
&&
1056 !ocfs2_may_continue_on_blocked_lock(lockres
, level
)) {
1057 /* is the lock is currently blocked on behalf of
1059 lockres_add_mask_waiter(lockres
, &mw
, OCFS2_LOCK_BLOCKED
, 0);
1064 if (level
> lockres
->l_level
) {
1065 if (lockres
->l_action
!= OCFS2_AST_INVALID
)
1066 mlog(ML_ERROR
, "lockres %s has action %u pending\n",
1067 lockres
->l_name
, lockres
->l_action
);
1069 if (!(lockres
->l_flags
& OCFS2_LOCK_ATTACHED
)) {
1070 lockres
->l_action
= OCFS2_AST_ATTACH
;
1071 lkm_flags
&= ~DLM_LKF_CONVERT
;
1073 lockres
->l_action
= OCFS2_AST_CONVERT
;
1074 lkm_flags
|= DLM_LKF_CONVERT
;
1077 lockres
->l_requested
= level
;
1078 lockres_or_flags(lockres
, OCFS2_LOCK_BUSY
);
1079 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1081 BUG_ON(level
== DLM_LOCK_IV
);
1082 BUG_ON(level
== DLM_LOCK_NL
);
1084 mlog(0, "lock %s, convert from %d to level = %d\n",
1085 lockres
->l_name
, lockres
->l_level
, level
);
1087 /* call dlm_lock to upgrade lock now */
1088 ret
= ocfs2_dlm_lock(osb
->dlm
,
1093 OCFS2_LOCK_ID_MAX_LEN
- 1,
1096 if (!(lkm_flags
& DLM_LKF_NOQUEUE
) ||
1098 ocfs2_log_dlm_error("ocfs2_dlm_lock",
1101 ocfs2_recover_from_dlm_error(lockres
, 1);
1105 mlog(0, "lock %s, successfull return from ocfs2_dlm_lock\n",
1108 /* At this point we've gone inside the dlm and need to
1109 * complete our work regardless. */
1112 /* wait for busy to clear and carry on */
1116 /* Ok, if we get here then we're good to go. */
1117 ocfs2_inc_holders(lockres
, level
);
1121 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1124 * This is helping work around a lock inversion between the page lock
1125 * and dlm locks. One path holds the page lock while calling aops
1126 * which block acquiring dlm locks. The voting thread holds dlm
1127 * locks while acquiring page locks while down converting data locks.
1128 * This block is helping an aop path notice the inversion and back
1129 * off to unlock its page lock before trying the dlm lock again.
1131 if (wait
&& arg_flags
& OCFS2_LOCK_NONBLOCK
&&
1132 mw
.mw_mask
& (OCFS2_LOCK_BUSY
|OCFS2_LOCK_BLOCKED
)) {
1134 if (lockres_remove_mask_waiter(lockres
, &mw
))
1140 ret
= ocfs2_wait_for_mask(&mw
);
1150 static void ocfs2_cluster_unlock(struct ocfs2_super
*osb
,
1151 struct ocfs2_lock_res
*lockres
,
1154 unsigned long flags
;
1157 spin_lock_irqsave(&lockres
->l_lock
, flags
);
1158 ocfs2_dec_holders(lockres
, level
);
1159 ocfs2_downconvert_on_unlock(osb
, lockres
);
1160 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1164 static int ocfs2_create_new_lock(struct ocfs2_super
*osb
,
1165 struct ocfs2_lock_res
*lockres
,
1169 int level
= ex
? DLM_LOCK_EX
: DLM_LOCK_PR
;
1170 unsigned long flags
;
1171 u32 lkm_flags
= local
? DLM_LKF_LOCAL
: 0;
1173 spin_lock_irqsave(&lockres
->l_lock
, flags
);
1174 BUG_ON(lockres
->l_flags
& OCFS2_LOCK_ATTACHED
);
1175 lockres_or_flags(lockres
, OCFS2_LOCK_LOCAL
);
1176 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1178 return ocfs2_lock_create(osb
, lockres
, level
, lkm_flags
);
1181 /* Grants us an EX lock on the data and metadata resources, skipping
1182 * the normal cluster directory lookup. Use this ONLY on newly created
1183 * inodes which other nodes can't possibly see, and which haven't been
1184 * hashed in the inode hash yet. This can give us a good performance
1185 * increase as it'll skip the network broadcast normally associated
1186 * with creating a new lock resource. */
1187 int ocfs2_create_new_inode_locks(struct inode
*inode
)
1190 struct ocfs2_super
*osb
= OCFS2_SB(inode
->i_sb
);
1193 BUG_ON(!ocfs2_inode_is_new(inode
));
1197 mlog(0, "Inode %llu\n", (unsigned long long)OCFS2_I(inode
)->ip_blkno
);
1199 /* NOTE: That we don't increment any of the holder counts, nor
1200 * do we add anything to a journal handle. Since this is
1201 * supposed to be a new inode which the cluster doesn't know
1202 * about yet, there is no need to. As far as the LVB handling
1203 * is concerned, this is basically like acquiring an EX lock
1204 * on a resource which has an invalid one -- we'll set it
1205 * valid when we release the EX. */
1207 ret
= ocfs2_create_new_lock(osb
, &OCFS2_I(inode
)->ip_rw_lockres
, 1, 1);
1214 * We don't want to use DLM_LKF_LOCAL on a meta data lock as they
1215 * don't use a generation in their lock names.
1217 ret
= ocfs2_create_new_lock(osb
, &OCFS2_I(inode
)->ip_inode_lockres
, 1, 0);
1223 ret
= ocfs2_create_new_lock(osb
, &OCFS2_I(inode
)->ip_open_lockres
, 0, 0);
1234 int ocfs2_rw_lock(struct inode
*inode
, int write
)
1237 struct ocfs2_lock_res
*lockres
;
1238 struct ocfs2_super
*osb
= OCFS2_SB(inode
->i_sb
);
1244 mlog(0, "inode %llu take %s RW lock\n",
1245 (unsigned long long)OCFS2_I(inode
)->ip_blkno
,
1246 write
? "EXMODE" : "PRMODE");
1248 if (ocfs2_mount_local(osb
))
1251 lockres
= &OCFS2_I(inode
)->ip_rw_lockres
;
1253 level
= write
? DLM_LOCK_EX
: DLM_LOCK_PR
;
1255 status
= ocfs2_cluster_lock(OCFS2_SB(inode
->i_sb
), lockres
, level
, 0,
1264 void ocfs2_rw_unlock(struct inode
*inode
, int write
)
1266 int level
= write
? DLM_LOCK_EX
: DLM_LOCK_PR
;
1267 struct ocfs2_lock_res
*lockres
= &OCFS2_I(inode
)->ip_rw_lockres
;
1268 struct ocfs2_super
*osb
= OCFS2_SB(inode
->i_sb
);
1272 mlog(0, "inode %llu drop %s RW lock\n",
1273 (unsigned long long)OCFS2_I(inode
)->ip_blkno
,
1274 write
? "EXMODE" : "PRMODE");
1276 if (!ocfs2_mount_local(osb
))
1277 ocfs2_cluster_unlock(OCFS2_SB(inode
->i_sb
), lockres
, level
);
1283 * ocfs2_open_lock always get PR mode lock.
1285 int ocfs2_open_lock(struct inode
*inode
)
1288 struct ocfs2_lock_res
*lockres
;
1289 struct ocfs2_super
*osb
= OCFS2_SB(inode
->i_sb
);
1295 mlog(0, "inode %llu take PRMODE open lock\n",
1296 (unsigned long long)OCFS2_I(inode
)->ip_blkno
);
1298 if (ocfs2_mount_local(osb
))
1301 lockres
= &OCFS2_I(inode
)->ip_open_lockres
;
1303 status
= ocfs2_cluster_lock(OCFS2_SB(inode
->i_sb
), lockres
,
1313 int ocfs2_try_open_lock(struct inode
*inode
, int write
)
1315 int status
= 0, level
;
1316 struct ocfs2_lock_res
*lockres
;
1317 struct ocfs2_super
*osb
= OCFS2_SB(inode
->i_sb
);
1323 mlog(0, "inode %llu try to take %s open lock\n",
1324 (unsigned long long)OCFS2_I(inode
)->ip_blkno
,
1325 write
? "EXMODE" : "PRMODE");
1327 if (ocfs2_mount_local(osb
))
1330 lockres
= &OCFS2_I(inode
)->ip_open_lockres
;
1332 level
= write
? DLM_LOCK_EX
: DLM_LOCK_PR
;
1335 * The file system may already holding a PRMODE/EXMODE open lock.
1336 * Since we pass DLM_LKF_NOQUEUE, the request won't block waiting on
1337 * other nodes and the -EAGAIN will indicate to the caller that
1338 * this inode is still in use.
1340 status
= ocfs2_cluster_lock(OCFS2_SB(inode
->i_sb
), lockres
,
1341 level
, DLM_LKF_NOQUEUE
, 0);
1349 * ocfs2_open_unlock unlock PR and EX mode open locks.
1351 void ocfs2_open_unlock(struct inode
*inode
)
1353 struct ocfs2_lock_res
*lockres
= &OCFS2_I(inode
)->ip_open_lockres
;
1354 struct ocfs2_super
*osb
= OCFS2_SB(inode
->i_sb
);
1358 mlog(0, "inode %llu drop open lock\n",
1359 (unsigned long long)OCFS2_I(inode
)->ip_blkno
);
1361 if (ocfs2_mount_local(osb
))
1364 if(lockres
->l_ro_holders
)
1365 ocfs2_cluster_unlock(OCFS2_SB(inode
->i_sb
), lockres
,
1367 if(lockres
->l_ex_holders
)
1368 ocfs2_cluster_unlock(OCFS2_SB(inode
->i_sb
), lockres
,
1375 static int ocfs2_flock_handle_signal(struct ocfs2_lock_res
*lockres
,
1379 struct ocfs2_super
*osb
= ocfs2_get_lockres_osb(lockres
);
1380 unsigned long flags
;
1381 struct ocfs2_mask_waiter mw
;
1383 ocfs2_init_mask_waiter(&mw
);
1386 spin_lock_irqsave(&lockres
->l_lock
, flags
);
1387 if (lockres
->l_flags
& OCFS2_LOCK_BUSY
) {
1388 ret
= ocfs2_prepare_cancel_convert(osb
, lockres
);
1390 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1391 ret
= ocfs2_cancel_convert(osb
, lockres
);
1398 lockres_add_mask_waiter(lockres
, &mw
, OCFS2_LOCK_BUSY
, 0);
1399 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1401 ocfs2_wait_for_mask(&mw
);
1407 * We may still have gotten the lock, in which case there's no
1408 * point to restarting the syscall.
1410 if (lockres
->l_level
== level
)
1413 mlog(0, "Cancel returning %d. flags: 0x%lx, level: %d, act: %d\n", ret
,
1414 lockres
->l_flags
, lockres
->l_level
, lockres
->l_action
);
1416 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1423 * ocfs2_file_lock() and ocfs2_file_unlock() map to a single pair of
1424 * flock() calls. The locking approach this requires is sufficiently
1425 * different from all other cluster lock types that we implement a
1426 * seperate path to the "low-level" dlm calls. In particular:
1428 * - No optimization of lock levels is done - we take at exactly
1429 * what's been requested.
1431 * - No lock caching is employed. We immediately downconvert to
1432 * no-lock at unlock time. This also means flock locks never go on
1433 * the blocking list).
1435 * - Since userspace can trivially deadlock itself with flock, we make
1436 * sure to allow cancellation of a misbehaving applications flock()
1439 * - Access to any flock lockres doesn't require concurrency, so we
1440 * can simplify the code by requiring the caller to guarantee
1441 * serialization of dlmglue flock calls.
1443 int ocfs2_file_lock(struct file
*file
, int ex
, int trylock
)
1445 int ret
, level
= ex
? LKM_EXMODE
: LKM_PRMODE
;
1446 unsigned int lkm_flags
= trylock
? LKM_NOQUEUE
: 0;
1447 unsigned long flags
;
1448 struct ocfs2_file_private
*fp
= file
->private_data
;
1449 struct ocfs2_lock_res
*lockres
= &fp
->fp_flock
;
1450 struct ocfs2_super
*osb
= OCFS2_SB(file
->f_mapping
->host
->i_sb
);
1451 struct ocfs2_mask_waiter mw
;
1453 ocfs2_init_mask_waiter(&mw
);
1455 if ((lockres
->l_flags
& OCFS2_LOCK_BUSY
) ||
1456 (lockres
->l_level
> DLM_LOCK_NL
)) {
1458 "File lock \"%s\" has busy or locked state: flags: 0x%lx, "
1459 "level: %u\n", lockres
->l_name
, lockres
->l_flags
,
1464 spin_lock_irqsave(&lockres
->l_lock
, flags
);
1465 if (!(lockres
->l_flags
& OCFS2_LOCK_ATTACHED
)) {
1466 lockres_add_mask_waiter(lockres
, &mw
, OCFS2_LOCK_BUSY
, 0);
1467 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1470 * Get the lock at NLMODE to start - that way we
1471 * can cancel the upconvert request if need be.
1473 ret
= ocfs2_lock_create(osb
, lockres
, LKM_NLMODE
, 0);
1479 ret
= ocfs2_wait_for_mask(&mw
);
1484 spin_lock_irqsave(&lockres
->l_lock
, flags
);
1487 lockres
->l_action
= OCFS2_AST_CONVERT
;
1488 lkm_flags
|= LKM_CONVERT
;
1489 lockres
->l_requested
= level
;
1490 lockres_or_flags(lockres
, OCFS2_LOCK_BUSY
);
1492 lockres_add_mask_waiter(lockres
, &mw
, OCFS2_LOCK_BUSY
, 0);
1493 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1495 ret
= ocfs2_dlm_lock(osb
->dlm
, level
, &lockres
->l_lksb
, lkm_flags
,
1496 lockres
->l_name
, OCFS2_LOCK_ID_MAX_LEN
- 1,
1499 if (!trylock
|| (ret
!= -EAGAIN
)) {
1500 ocfs2_log_dlm_error("ocfs2_dlm_lock", ret
, lockres
);
1504 ocfs2_recover_from_dlm_error(lockres
, 1);
1505 lockres_remove_mask_waiter(lockres
, &mw
);
1509 ret
= ocfs2_wait_for_mask_interruptible(&mw
, lockres
);
1510 if (ret
== -ERESTARTSYS
) {
1512 * Userspace can cause deadlock itself with
1513 * flock(). Current behavior locally is to allow the
1514 * deadlock, but abort the system call if a signal is
1515 * received. We follow this example, otherwise a
1516 * poorly written program could sit in kernel until
1519 * Handling this is a bit more complicated for Ocfs2
1520 * though. We can't exit this function with an
1521 * outstanding lock request, so a cancel convert is
1522 * required. We intentionally overwrite 'ret' - if the
1523 * cancel fails and the lock was granted, it's easier
1524 * to just bubble sucess back up to the user.
1526 ret
= ocfs2_flock_handle_signal(lockres
, level
);
1531 mlog(0, "Lock: \"%s\" ex: %d, trylock: %d, returns: %d\n",
1532 lockres
->l_name
, ex
, trylock
, ret
);
1536 void ocfs2_file_unlock(struct file
*file
)
1539 unsigned long flags
;
1540 struct ocfs2_file_private
*fp
= file
->private_data
;
1541 struct ocfs2_lock_res
*lockres
= &fp
->fp_flock
;
1542 struct ocfs2_super
*osb
= OCFS2_SB(file
->f_mapping
->host
->i_sb
);
1543 struct ocfs2_mask_waiter mw
;
1545 ocfs2_init_mask_waiter(&mw
);
1547 if (!(lockres
->l_flags
& OCFS2_LOCK_ATTACHED
))
1550 if (lockres
->l_level
== LKM_NLMODE
)
1553 mlog(0, "Unlock: \"%s\" flags: 0x%lx, level: %d, act: %d\n",
1554 lockres
->l_name
, lockres
->l_flags
, lockres
->l_level
,
1557 spin_lock_irqsave(&lockres
->l_lock
, flags
);
1559 * Fake a blocking ast for the downconvert code.
1561 lockres_or_flags(lockres
, OCFS2_LOCK_BLOCKED
);
1562 lockres
->l_blocking
= DLM_LOCK_EX
;
1564 ocfs2_prepare_downconvert(lockres
, LKM_NLMODE
);
1565 lockres_add_mask_waiter(lockres
, &mw
, OCFS2_LOCK_BUSY
, 0);
1566 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1568 ret
= ocfs2_downconvert_lock(osb
, lockres
, LKM_NLMODE
, 0);
1574 ret
= ocfs2_wait_for_mask(&mw
);
1579 static void ocfs2_downconvert_on_unlock(struct ocfs2_super
*osb
,
1580 struct ocfs2_lock_res
*lockres
)
1586 /* If we know that another node is waiting on our lock, kick
1587 * the downconvert thread * pre-emptively when we reach a release
1589 if (lockres
->l_flags
& OCFS2_LOCK_BLOCKED
) {
1590 switch(lockres
->l_blocking
) {
1592 if (!lockres
->l_ex_holders
&& !lockres
->l_ro_holders
)
1596 if (!lockres
->l_ex_holders
)
1605 ocfs2_wake_downconvert_thread(osb
);
1610 #define OCFS2_SEC_BITS 34
1611 #define OCFS2_SEC_SHIFT (64 - 34)
1612 #define OCFS2_NSEC_MASK ((1ULL << OCFS2_SEC_SHIFT) - 1)
1614 /* LVB only has room for 64 bits of time here so we pack it for
1616 static u64
ocfs2_pack_timespec(struct timespec
*spec
)
1619 u64 sec
= spec
->tv_sec
;
1620 u32 nsec
= spec
->tv_nsec
;
1622 res
= (sec
<< OCFS2_SEC_SHIFT
) | (nsec
& OCFS2_NSEC_MASK
);
1627 /* Call this with the lockres locked. I am reasonably sure we don't
1628 * need ip_lock in this function as anyone who would be changing those
1629 * values is supposed to be blocked in ocfs2_inode_lock right now. */
1630 static void __ocfs2_stuff_meta_lvb(struct inode
*inode
)
1632 struct ocfs2_inode_info
*oi
= OCFS2_I(inode
);
1633 struct ocfs2_lock_res
*lockres
= &oi
->ip_inode_lockres
;
1634 struct ocfs2_meta_lvb
*lvb
;
1638 lvb
= (struct ocfs2_meta_lvb
*)ocfs2_dlm_lvb(&lockres
->l_lksb
);
1641 * Invalidate the LVB of a deleted inode - this way other
1642 * nodes are forced to go to disk and discover the new inode
1645 if (oi
->ip_flags
& OCFS2_INODE_DELETED
) {
1646 lvb
->lvb_version
= 0;
1650 lvb
->lvb_version
= OCFS2_LVB_VERSION
;
1651 lvb
->lvb_isize
= cpu_to_be64(i_size_read(inode
));
1652 lvb
->lvb_iclusters
= cpu_to_be32(oi
->ip_clusters
);
1653 lvb
->lvb_iuid
= cpu_to_be32(inode
->i_uid
);
1654 lvb
->lvb_igid
= cpu_to_be32(inode
->i_gid
);
1655 lvb
->lvb_imode
= cpu_to_be16(inode
->i_mode
);
1656 lvb
->lvb_inlink
= cpu_to_be16(inode
->i_nlink
);
1657 lvb
->lvb_iatime_packed
=
1658 cpu_to_be64(ocfs2_pack_timespec(&inode
->i_atime
));
1659 lvb
->lvb_ictime_packed
=
1660 cpu_to_be64(ocfs2_pack_timespec(&inode
->i_ctime
));
1661 lvb
->lvb_imtime_packed
=
1662 cpu_to_be64(ocfs2_pack_timespec(&inode
->i_mtime
));
1663 lvb
->lvb_iattr
= cpu_to_be32(oi
->ip_attr
);
1664 lvb
->lvb_idynfeatures
= cpu_to_be16(oi
->ip_dyn_features
);
1665 lvb
->lvb_igeneration
= cpu_to_be32(inode
->i_generation
);
1668 mlog_meta_lvb(0, lockres
);
1673 static void ocfs2_unpack_timespec(struct timespec
*spec
,
1676 spec
->tv_sec
= packed_time
>> OCFS2_SEC_SHIFT
;
1677 spec
->tv_nsec
= packed_time
& OCFS2_NSEC_MASK
;
1680 static void ocfs2_refresh_inode_from_lvb(struct inode
*inode
)
1682 struct ocfs2_inode_info
*oi
= OCFS2_I(inode
);
1683 struct ocfs2_lock_res
*lockres
= &oi
->ip_inode_lockres
;
1684 struct ocfs2_meta_lvb
*lvb
;
1688 mlog_meta_lvb(0, lockres
);
1690 lvb
= (struct ocfs2_meta_lvb
*)ocfs2_dlm_lvb(&lockres
->l_lksb
);
1692 /* We're safe here without the lockres lock... */
1693 spin_lock(&oi
->ip_lock
);
1694 oi
->ip_clusters
= be32_to_cpu(lvb
->lvb_iclusters
);
1695 i_size_write(inode
, be64_to_cpu(lvb
->lvb_isize
));
1697 oi
->ip_attr
= be32_to_cpu(lvb
->lvb_iattr
);
1698 oi
->ip_dyn_features
= be16_to_cpu(lvb
->lvb_idynfeatures
);
1699 ocfs2_set_inode_flags(inode
);
1701 /* fast-symlinks are a special case */
1702 if (S_ISLNK(inode
->i_mode
) && !oi
->ip_clusters
)
1703 inode
->i_blocks
= 0;
1705 inode
->i_blocks
= ocfs2_inode_sector_count(inode
);
1707 inode
->i_uid
= be32_to_cpu(lvb
->lvb_iuid
);
1708 inode
->i_gid
= be32_to_cpu(lvb
->lvb_igid
);
1709 inode
->i_mode
= be16_to_cpu(lvb
->lvb_imode
);
1710 inode
->i_nlink
= be16_to_cpu(lvb
->lvb_inlink
);
1711 ocfs2_unpack_timespec(&inode
->i_atime
,
1712 be64_to_cpu(lvb
->lvb_iatime_packed
));
1713 ocfs2_unpack_timespec(&inode
->i_mtime
,
1714 be64_to_cpu(lvb
->lvb_imtime_packed
));
1715 ocfs2_unpack_timespec(&inode
->i_ctime
,
1716 be64_to_cpu(lvb
->lvb_ictime_packed
));
1717 spin_unlock(&oi
->ip_lock
);
1722 static inline int ocfs2_meta_lvb_is_trustable(struct inode
*inode
,
1723 struct ocfs2_lock_res
*lockres
)
1725 struct ocfs2_meta_lvb
*lvb
=
1726 (struct ocfs2_meta_lvb
*)ocfs2_dlm_lvb(&lockres
->l_lksb
);
1728 if (lvb
->lvb_version
== OCFS2_LVB_VERSION
1729 && be32_to_cpu(lvb
->lvb_igeneration
) == inode
->i_generation
)
1734 /* Determine whether a lock resource needs to be refreshed, and
1735 * arbitrate who gets to refresh it.
1737 * 0 means no refresh needed.
1739 * > 0 means you need to refresh this and you MUST call
1740 * ocfs2_complete_lock_res_refresh afterwards. */
1741 static int ocfs2_should_refresh_lock_res(struct ocfs2_lock_res
*lockres
)
1743 unsigned long flags
;
1749 spin_lock_irqsave(&lockres
->l_lock
, flags
);
1750 if (!(lockres
->l_flags
& OCFS2_LOCK_NEEDS_REFRESH
)) {
1751 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1755 if (lockres
->l_flags
& OCFS2_LOCK_REFRESHING
) {
1756 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1758 ocfs2_wait_on_refreshing_lock(lockres
);
1762 /* Ok, I'll be the one to refresh this lock. */
1763 lockres_or_flags(lockres
, OCFS2_LOCK_REFRESHING
);
1764 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1772 /* If status is non zero, I'll mark it as not being in refresh
1773 * anymroe, but i won't clear the needs refresh flag. */
1774 static inline void ocfs2_complete_lock_res_refresh(struct ocfs2_lock_res
*lockres
,
1777 unsigned long flags
;
1780 spin_lock_irqsave(&lockres
->l_lock
, flags
);
1781 lockres_clear_flags(lockres
, OCFS2_LOCK_REFRESHING
);
1783 lockres_clear_flags(lockres
, OCFS2_LOCK_NEEDS_REFRESH
);
1784 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
1786 wake_up(&lockres
->l_event
);
1791 /* may or may not return a bh if it went to disk. */
1792 static int ocfs2_inode_lock_update(struct inode
*inode
,
1793 struct buffer_head
**bh
)
1796 struct ocfs2_inode_info
*oi
= OCFS2_I(inode
);
1797 struct ocfs2_lock_res
*lockres
= &oi
->ip_inode_lockres
;
1798 struct ocfs2_dinode
*fe
;
1799 struct ocfs2_super
*osb
= OCFS2_SB(inode
->i_sb
);
1803 if (ocfs2_mount_local(osb
))
1806 spin_lock(&oi
->ip_lock
);
1807 if (oi
->ip_flags
& OCFS2_INODE_DELETED
) {
1808 mlog(0, "Orphaned inode %llu was deleted while we "
1809 "were waiting on a lock. ip_flags = 0x%x\n",
1810 (unsigned long long)oi
->ip_blkno
, oi
->ip_flags
);
1811 spin_unlock(&oi
->ip_lock
);
1815 spin_unlock(&oi
->ip_lock
);
1817 if (!ocfs2_should_refresh_lock_res(lockres
))
1820 /* This will discard any caching information we might have had
1821 * for the inode metadata. */
1822 ocfs2_metadata_cache_purge(inode
);
1824 ocfs2_extent_map_trunc(inode
, 0);
1826 if (ocfs2_meta_lvb_is_trustable(inode
, lockres
)) {
1827 mlog(0, "Trusting LVB on inode %llu\n",
1828 (unsigned long long)oi
->ip_blkno
);
1829 ocfs2_refresh_inode_from_lvb(inode
);
1831 /* Boo, we have to go to disk. */
1832 /* read bh, cast, ocfs2_refresh_inode */
1833 status
= ocfs2_read_block(OCFS2_SB(inode
->i_sb
), oi
->ip_blkno
,
1834 bh
, OCFS2_BH_CACHED
, inode
);
1839 fe
= (struct ocfs2_dinode
*) (*bh
)->b_data
;
1841 /* This is a good chance to make sure we're not
1842 * locking an invalid object.
1844 * We bug on a stale inode here because we checked
1845 * above whether it was wiped from disk. The wiping
1846 * node provides a guarantee that we receive that
1847 * message and can mark the inode before dropping any
1848 * locks associated with it. */
1849 if (!OCFS2_IS_VALID_DINODE(fe
)) {
1850 OCFS2_RO_ON_INVALID_DINODE(inode
->i_sb
, fe
);
1854 mlog_bug_on_msg(inode
->i_generation
!=
1855 le32_to_cpu(fe
->i_generation
),
1856 "Invalid dinode %llu disk generation: %u "
1857 "inode->i_generation: %u\n",
1858 (unsigned long long)oi
->ip_blkno
,
1859 le32_to_cpu(fe
->i_generation
),
1860 inode
->i_generation
);
1861 mlog_bug_on_msg(le64_to_cpu(fe
->i_dtime
) ||
1862 !(fe
->i_flags
& cpu_to_le32(OCFS2_VALID_FL
)),
1863 "Stale dinode %llu dtime: %llu flags: 0x%x\n",
1864 (unsigned long long)oi
->ip_blkno
,
1865 (unsigned long long)le64_to_cpu(fe
->i_dtime
),
1866 le32_to_cpu(fe
->i_flags
));
1868 ocfs2_refresh_inode(inode
, fe
);
1873 ocfs2_complete_lock_res_refresh(lockres
, status
);
1879 static int ocfs2_assign_bh(struct inode
*inode
,
1880 struct buffer_head
**ret_bh
,
1881 struct buffer_head
*passed_bh
)
1886 /* Ok, the update went to disk for us, use the
1888 *ret_bh
= passed_bh
;
1894 status
= ocfs2_read_block(OCFS2_SB(inode
->i_sb
),
1895 OCFS2_I(inode
)->ip_blkno
,
1906 * returns < 0 error if the callback will never be called, otherwise
1907 * the result of the lock will be communicated via the callback.
1909 int ocfs2_inode_lock_full(struct inode
*inode
,
1910 struct buffer_head
**ret_bh
,
1914 int status
, level
, acquired
;
1916 struct ocfs2_lock_res
*lockres
= NULL
;
1917 struct ocfs2_super
*osb
= OCFS2_SB(inode
->i_sb
);
1918 struct buffer_head
*local_bh
= NULL
;
1924 mlog(0, "inode %llu, take %s META lock\n",
1925 (unsigned long long)OCFS2_I(inode
)->ip_blkno
,
1926 ex
? "EXMODE" : "PRMODE");
1930 /* We'll allow faking a readonly metadata lock for
1932 if (ocfs2_is_hard_readonly(osb
)) {
1938 if (ocfs2_mount_local(osb
))
1941 if (!(arg_flags
& OCFS2_META_LOCK_RECOVERY
))
1942 ocfs2_wait_for_recovery(osb
);
1944 lockres
= &OCFS2_I(inode
)->ip_inode_lockres
;
1945 level
= ex
? DLM_LOCK_EX
: DLM_LOCK_PR
;
1947 if (arg_flags
& OCFS2_META_LOCK_NOQUEUE
)
1948 dlm_flags
|= DLM_LKF_NOQUEUE
;
1950 status
= ocfs2_cluster_lock(osb
, lockres
, level
, dlm_flags
, arg_flags
);
1952 if (status
!= -EAGAIN
&& status
!= -EIOCBRETRY
)
1957 /* Notify the error cleanup path to drop the cluster lock. */
1960 /* We wait twice because a node may have died while we were in
1961 * the lower dlm layers. The second time though, we've
1962 * committed to owning this lock so we don't allow signals to
1963 * abort the operation. */
1964 if (!(arg_flags
& OCFS2_META_LOCK_RECOVERY
))
1965 ocfs2_wait_for_recovery(osb
);
1969 * We only see this flag if we're being called from
1970 * ocfs2_read_locked_inode(). It means we're locking an inode
1971 * which hasn't been populated yet, so clear the refresh flag
1972 * and let the caller handle it.
1974 if (inode
->i_state
& I_NEW
) {
1977 ocfs2_complete_lock_res_refresh(lockres
, 0);
1981 /* This is fun. The caller may want a bh back, or it may
1982 * not. ocfs2_inode_lock_update definitely wants one in, but
1983 * may or may not read one, depending on what's in the
1984 * LVB. The result of all of this is that we've *only* gone to
1985 * disk if we have to, so the complexity is worthwhile. */
1986 status
= ocfs2_inode_lock_update(inode
, &local_bh
);
1988 if (status
!= -ENOENT
)
1994 status
= ocfs2_assign_bh(inode
, ret_bh
, local_bh
);
2003 if (ret_bh
&& (*ret_bh
)) {
2008 ocfs2_inode_unlock(inode
, ex
);
2019 * This is working around a lock inversion between tasks acquiring DLM
2020 * locks while holding a page lock and the downconvert thread which
2021 * blocks dlm lock acquiry while acquiring page locks.
2023 * ** These _with_page variantes are only intended to be called from aop
2024 * methods that hold page locks and return a very specific *positive* error
2025 * code that aop methods pass up to the VFS -- test for errors with != 0. **
2027 * The DLM is called such that it returns -EAGAIN if it would have
2028 * blocked waiting for the downconvert thread. In that case we unlock
2029 * our page so the downconvert thread can make progress. Once we've
2030 * done this we have to return AOP_TRUNCATED_PAGE so the aop method
2031 * that called us can bubble that back up into the VFS who will then
2032 * immediately retry the aop call.
2034 * We do a blocking lock and immediate unlock before returning, though, so that
2035 * the lock has a great chance of being cached on this node by the time the VFS
2036 * calls back to retry the aop. This has a potential to livelock as nodes
2037 * ping locks back and forth, but that's a risk we're willing to take to avoid
2038 * the lock inversion simply.
2040 int ocfs2_inode_lock_with_page(struct inode
*inode
,
2041 struct buffer_head
**ret_bh
,
2047 ret
= ocfs2_inode_lock_full(inode
, ret_bh
, ex
, OCFS2_LOCK_NONBLOCK
);
2048 if (ret
== -EAGAIN
) {
2050 if (ocfs2_inode_lock(inode
, ret_bh
, ex
) == 0)
2051 ocfs2_inode_unlock(inode
, ex
);
2052 ret
= AOP_TRUNCATED_PAGE
;
2058 int ocfs2_inode_lock_atime(struct inode
*inode
,
2059 struct vfsmount
*vfsmnt
,
2065 ret
= ocfs2_inode_lock(inode
, NULL
, 0);
2072 * If we should update atime, we will get EX lock,
2073 * otherwise we just get PR lock.
2075 if (ocfs2_should_update_atime(inode
, vfsmnt
)) {
2076 struct buffer_head
*bh
= NULL
;
2078 ocfs2_inode_unlock(inode
, 0);
2079 ret
= ocfs2_inode_lock(inode
, &bh
, 1);
2085 if (ocfs2_should_update_atime(inode
, vfsmnt
))
2086 ocfs2_update_inode_atime(inode
, bh
);
2096 void ocfs2_inode_unlock(struct inode
*inode
,
2099 int level
= ex
? DLM_LOCK_EX
: DLM_LOCK_PR
;
2100 struct ocfs2_lock_res
*lockres
= &OCFS2_I(inode
)->ip_inode_lockres
;
2101 struct ocfs2_super
*osb
= OCFS2_SB(inode
->i_sb
);
2105 mlog(0, "inode %llu drop %s META lock\n",
2106 (unsigned long long)OCFS2_I(inode
)->ip_blkno
,
2107 ex
? "EXMODE" : "PRMODE");
2109 if (!ocfs2_is_hard_readonly(OCFS2_SB(inode
->i_sb
)) &&
2110 !ocfs2_mount_local(osb
))
2111 ocfs2_cluster_unlock(OCFS2_SB(inode
->i_sb
), lockres
, level
);
2116 int ocfs2_super_lock(struct ocfs2_super
*osb
,
2120 int level
= ex
? DLM_LOCK_EX
: DLM_LOCK_PR
;
2121 struct ocfs2_lock_res
*lockres
= &osb
->osb_super_lockres
;
2125 if (ocfs2_is_hard_readonly(osb
))
2128 if (ocfs2_mount_local(osb
))
2131 status
= ocfs2_cluster_lock(osb
, lockres
, level
, 0, 0);
2137 /* The super block lock path is really in the best position to
2138 * know when resources covered by the lock need to be
2139 * refreshed, so we do it here. Of course, making sense of
2140 * everything is up to the caller :) */
2141 status
= ocfs2_should_refresh_lock_res(lockres
);
2147 status
= ocfs2_refresh_slot_info(osb
);
2149 ocfs2_complete_lock_res_refresh(lockres
, status
);
2159 void ocfs2_super_unlock(struct ocfs2_super
*osb
,
2162 int level
= ex
? DLM_LOCK_EX
: DLM_LOCK_PR
;
2163 struct ocfs2_lock_res
*lockres
= &osb
->osb_super_lockres
;
2165 if (!ocfs2_mount_local(osb
))
2166 ocfs2_cluster_unlock(osb
, lockres
, level
);
2169 int ocfs2_rename_lock(struct ocfs2_super
*osb
)
2172 struct ocfs2_lock_res
*lockres
= &osb
->osb_rename_lockres
;
2174 if (ocfs2_is_hard_readonly(osb
))
2177 if (ocfs2_mount_local(osb
))
2180 status
= ocfs2_cluster_lock(osb
, lockres
, DLM_LOCK_EX
, 0, 0);
2187 void ocfs2_rename_unlock(struct ocfs2_super
*osb
)
2189 struct ocfs2_lock_res
*lockres
= &osb
->osb_rename_lockres
;
2191 if (!ocfs2_mount_local(osb
))
2192 ocfs2_cluster_unlock(osb
, lockres
, DLM_LOCK_EX
);
2195 int ocfs2_dentry_lock(struct dentry
*dentry
, int ex
)
2198 int level
= ex
? DLM_LOCK_EX
: DLM_LOCK_PR
;
2199 struct ocfs2_dentry_lock
*dl
= dentry
->d_fsdata
;
2200 struct ocfs2_super
*osb
= OCFS2_SB(dentry
->d_sb
);
2204 if (ocfs2_is_hard_readonly(osb
))
2207 if (ocfs2_mount_local(osb
))
2210 ret
= ocfs2_cluster_lock(osb
, &dl
->dl_lockres
, level
, 0, 0);
2217 void ocfs2_dentry_unlock(struct dentry
*dentry
, int ex
)
2219 int level
= ex
? DLM_LOCK_EX
: DLM_LOCK_PR
;
2220 struct ocfs2_dentry_lock
*dl
= dentry
->d_fsdata
;
2221 struct ocfs2_super
*osb
= OCFS2_SB(dentry
->d_sb
);
2223 if (!ocfs2_mount_local(osb
))
2224 ocfs2_cluster_unlock(osb
, &dl
->dl_lockres
, level
);
2227 /* Reference counting of the dlm debug structure. We want this because
2228 * open references on the debug inodes can live on after a mount, so
2229 * we can't rely on the ocfs2_super to always exist. */
2230 static void ocfs2_dlm_debug_free(struct kref
*kref
)
2232 struct ocfs2_dlm_debug
*dlm_debug
;
2234 dlm_debug
= container_of(kref
, struct ocfs2_dlm_debug
, d_refcnt
);
2239 void ocfs2_put_dlm_debug(struct ocfs2_dlm_debug
*dlm_debug
)
2242 kref_put(&dlm_debug
->d_refcnt
, ocfs2_dlm_debug_free
);
2245 static void ocfs2_get_dlm_debug(struct ocfs2_dlm_debug
*debug
)
2247 kref_get(&debug
->d_refcnt
);
2250 struct ocfs2_dlm_debug
*ocfs2_new_dlm_debug(void)
2252 struct ocfs2_dlm_debug
*dlm_debug
;
2254 dlm_debug
= kmalloc(sizeof(struct ocfs2_dlm_debug
), GFP_KERNEL
);
2256 mlog_errno(-ENOMEM
);
2260 kref_init(&dlm_debug
->d_refcnt
);
2261 INIT_LIST_HEAD(&dlm_debug
->d_lockres_tracking
);
2262 dlm_debug
->d_locking_state
= NULL
;
2267 /* Access to this is arbitrated for us via seq_file->sem. */
2268 struct ocfs2_dlm_seq_priv
{
2269 struct ocfs2_dlm_debug
*p_dlm_debug
;
2270 struct ocfs2_lock_res p_iter_res
;
2271 struct ocfs2_lock_res p_tmp_res
;
2274 static struct ocfs2_lock_res
*ocfs2_dlm_next_res(struct ocfs2_lock_res
*start
,
2275 struct ocfs2_dlm_seq_priv
*priv
)
2277 struct ocfs2_lock_res
*iter
, *ret
= NULL
;
2278 struct ocfs2_dlm_debug
*dlm_debug
= priv
->p_dlm_debug
;
2280 assert_spin_locked(&ocfs2_dlm_tracking_lock
);
2282 list_for_each_entry(iter
, &start
->l_debug_list
, l_debug_list
) {
2283 /* discover the head of the list */
2284 if (&iter
->l_debug_list
== &dlm_debug
->d_lockres_tracking
) {
2285 mlog(0, "End of list found, %p\n", ret
);
2289 /* We track our "dummy" iteration lockres' by a NULL
2291 if (iter
->l_ops
!= NULL
) {
2300 static void *ocfs2_dlm_seq_start(struct seq_file
*m
, loff_t
*pos
)
2302 struct ocfs2_dlm_seq_priv
*priv
= m
->private;
2303 struct ocfs2_lock_res
*iter
;
2305 spin_lock(&ocfs2_dlm_tracking_lock
);
2306 iter
= ocfs2_dlm_next_res(&priv
->p_iter_res
, priv
);
2308 /* Since lockres' have the lifetime of their container
2309 * (which can be inodes, ocfs2_supers, etc) we want to
2310 * copy this out to a temporary lockres while still
2311 * under the spinlock. Obviously after this we can't
2312 * trust any pointers on the copy returned, but that's
2313 * ok as the information we want isn't typically held
2315 priv
->p_tmp_res
= *iter
;
2316 iter
= &priv
->p_tmp_res
;
2318 spin_unlock(&ocfs2_dlm_tracking_lock
);
2323 static void ocfs2_dlm_seq_stop(struct seq_file
*m
, void *v
)
2327 static void *ocfs2_dlm_seq_next(struct seq_file
*m
, void *v
, loff_t
*pos
)
2329 struct ocfs2_dlm_seq_priv
*priv
= m
->private;
2330 struct ocfs2_lock_res
*iter
= v
;
2331 struct ocfs2_lock_res
*dummy
= &priv
->p_iter_res
;
2333 spin_lock(&ocfs2_dlm_tracking_lock
);
2334 iter
= ocfs2_dlm_next_res(iter
, priv
);
2335 list_del_init(&dummy
->l_debug_list
);
2337 list_add(&dummy
->l_debug_list
, &iter
->l_debug_list
);
2338 priv
->p_tmp_res
= *iter
;
2339 iter
= &priv
->p_tmp_res
;
2341 spin_unlock(&ocfs2_dlm_tracking_lock
);
2346 /* So that debugfs.ocfs2 can determine which format is being used */
2347 #define OCFS2_DLM_DEBUG_STR_VERSION 1
2348 static int ocfs2_dlm_seq_show(struct seq_file
*m
, void *v
)
2352 struct ocfs2_lock_res
*lockres
= v
;
2357 seq_printf(m
, "0x%x\t", OCFS2_DLM_DEBUG_STR_VERSION
);
2359 if (lockres
->l_type
== OCFS2_LOCK_TYPE_DENTRY
)
2360 seq_printf(m
, "%.*s%08x\t", OCFS2_DENTRY_LOCK_INO_START
- 1,
2362 (unsigned int)ocfs2_get_dentry_lock_ino(lockres
));
2364 seq_printf(m
, "%.*s\t", OCFS2_LOCK_ID_MAX_LEN
, lockres
->l_name
);
2366 seq_printf(m
, "%d\t"
2377 lockres
->l_unlock_action
,
2378 lockres
->l_ro_holders
,
2379 lockres
->l_ex_holders
,
2380 lockres
->l_requested
,
2381 lockres
->l_blocking
);
2383 /* Dump the raw LVB */
2384 lvb
= ocfs2_dlm_lvb(&lockres
->l_lksb
);
2385 for(i
= 0; i
< DLM_LVB_LEN
; i
++)
2386 seq_printf(m
, "0x%x\t", lvb
[i
]);
2389 seq_printf(m
, "\n");
2393 static const struct seq_operations ocfs2_dlm_seq_ops
= {
2394 .start
= ocfs2_dlm_seq_start
,
2395 .stop
= ocfs2_dlm_seq_stop
,
2396 .next
= ocfs2_dlm_seq_next
,
2397 .show
= ocfs2_dlm_seq_show
,
2400 static int ocfs2_dlm_debug_release(struct inode
*inode
, struct file
*file
)
2402 struct seq_file
*seq
= (struct seq_file
*) file
->private_data
;
2403 struct ocfs2_dlm_seq_priv
*priv
= seq
->private;
2404 struct ocfs2_lock_res
*res
= &priv
->p_iter_res
;
2406 ocfs2_remove_lockres_tracking(res
);
2407 ocfs2_put_dlm_debug(priv
->p_dlm_debug
);
2408 return seq_release_private(inode
, file
);
2411 static int ocfs2_dlm_debug_open(struct inode
*inode
, struct file
*file
)
2414 struct ocfs2_dlm_seq_priv
*priv
;
2415 struct seq_file
*seq
;
2416 struct ocfs2_super
*osb
;
2418 priv
= kzalloc(sizeof(struct ocfs2_dlm_seq_priv
), GFP_KERNEL
);
2424 osb
= inode
->i_private
;
2425 ocfs2_get_dlm_debug(osb
->osb_dlm_debug
);
2426 priv
->p_dlm_debug
= osb
->osb_dlm_debug
;
2427 INIT_LIST_HEAD(&priv
->p_iter_res
.l_debug_list
);
2429 ret
= seq_open(file
, &ocfs2_dlm_seq_ops
);
2436 seq
= (struct seq_file
*) file
->private_data
;
2437 seq
->private = priv
;
2439 ocfs2_add_lockres_tracking(&priv
->p_iter_res
,
2446 static const struct file_operations ocfs2_dlm_debug_fops
= {
2447 .open
= ocfs2_dlm_debug_open
,
2448 .release
= ocfs2_dlm_debug_release
,
2450 .llseek
= seq_lseek
,
2453 static int ocfs2_dlm_init_debug(struct ocfs2_super
*osb
)
2456 struct ocfs2_dlm_debug
*dlm_debug
= osb
->osb_dlm_debug
;
2458 dlm_debug
->d_locking_state
= debugfs_create_file("locking_state",
2460 osb
->osb_debug_root
,
2462 &ocfs2_dlm_debug_fops
);
2463 if (!dlm_debug
->d_locking_state
) {
2466 "Unable to create locking state debugfs file.\n");
2470 ocfs2_get_dlm_debug(dlm_debug
);
2475 static void ocfs2_dlm_shutdown_debug(struct ocfs2_super
*osb
)
2477 struct ocfs2_dlm_debug
*dlm_debug
= osb
->osb_dlm_debug
;
2480 debugfs_remove(dlm_debug
->d_locking_state
);
2481 ocfs2_put_dlm_debug(dlm_debug
);
2485 int ocfs2_dlm_init(struct ocfs2_super
*osb
)
2489 struct dlm_ctxt
*dlm
= NULL
;
2493 if (ocfs2_mount_local(osb
))
2496 status
= ocfs2_dlm_init_debug(osb
);
2502 /* launch downconvert thread */
2503 osb
->dc_task
= kthread_run(ocfs2_downconvert_thread
, osb
, "ocfs2dc");
2504 if (IS_ERR(osb
->dc_task
)) {
2505 status
= PTR_ERR(osb
->dc_task
);
2506 osb
->dc_task
= NULL
;
2511 /* used by the dlm code to make message headers unique, each
2512 * node in this domain must agree on this. */
2513 dlm_key
= crc32_le(0, osb
->uuid_str
, strlen(osb
->uuid_str
));
2515 /* for now, uuid == domain */
2516 dlm
= dlm_register_domain(osb
->uuid_str
, dlm_key
,
2517 &osb
->osb_locking_proto
);
2519 status
= PTR_ERR(dlm
);
2524 dlm_register_eviction_cb(dlm
, &osb
->osb_eviction_cb
);
2527 ocfs2_super_lock_res_init(&osb
->osb_super_lockres
, osb
);
2528 ocfs2_rename_lock_res_init(&osb
->osb_rename_lockres
, osb
);
2535 ocfs2_dlm_shutdown_debug(osb
);
2537 kthread_stop(osb
->dc_task
);
2544 void ocfs2_dlm_shutdown(struct ocfs2_super
*osb
)
2548 dlm_unregister_eviction_cb(&osb
->osb_eviction_cb
);
2550 ocfs2_drop_osb_locks(osb
);
2553 kthread_stop(osb
->dc_task
);
2554 osb
->dc_task
= NULL
;
2557 ocfs2_lock_res_free(&osb
->osb_super_lockres
);
2558 ocfs2_lock_res_free(&osb
->osb_rename_lockres
);
2560 dlm_unregister_domain(osb
->dlm
);
2563 ocfs2_dlm_shutdown_debug(osb
);
2568 static void ocfs2_unlock_ast(void *opaque
, int error
)
2570 struct ocfs2_lock_res
*lockres
= opaque
;
2571 unsigned long flags
;
2575 mlog(0, "UNLOCK AST called on lock %s, action = %d\n", lockres
->l_name
,
2576 lockres
->l_unlock_action
);
2578 spin_lock_irqsave(&lockres
->l_lock
, flags
);
2579 /* We tried to cancel a convert request, but it was already
2580 * granted. All we want to do here is clear our unlock
2581 * state. The wake_up call done at the bottom is redundant
2582 * (ocfs2_prepare_cancel_convert doesn't sleep on this) but doesn't
2583 * hurt anything anyway */
2584 if (error
== -DLM_ECANCEL
&&
2585 lockres
->l_unlock_action
== OCFS2_UNLOCK_CANCEL_CONVERT
) {
2586 mlog(0, "Got cancelgrant for %s\n", lockres
->l_name
);
2588 /* We don't clear the busy flag in this case as it
2589 * should have been cleared by the ast which the dlm
2591 goto complete_unlock
;
2594 /* DLM_EUNLOCK is the success code for unlock */
2595 if (error
!= -DLM_EUNLOCK
) {
2596 mlog(ML_ERROR
, "Dlm passes error %d for lock %s, "
2597 "unlock_action %d\n", error
, lockres
->l_name
,
2598 lockres
->l_unlock_action
);
2599 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
2603 switch(lockres
->l_unlock_action
) {
2604 case OCFS2_UNLOCK_CANCEL_CONVERT
:
2605 mlog(0, "Cancel convert success for %s\n", lockres
->l_name
);
2606 lockres
->l_action
= OCFS2_AST_INVALID
;
2608 case OCFS2_UNLOCK_DROP_LOCK
:
2609 lockres
->l_level
= DLM_LOCK_IV
;
2615 lockres_clear_flags(lockres
, OCFS2_LOCK_BUSY
);
2617 lockres
->l_unlock_action
= OCFS2_UNLOCK_INVALID
;
2618 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
2620 wake_up(&lockres
->l_event
);
2625 static int ocfs2_drop_lock(struct ocfs2_super
*osb
,
2626 struct ocfs2_lock_res
*lockres
)
2629 unsigned long flags
;
2632 /* We didn't get anywhere near actually using this lockres. */
2633 if (!(lockres
->l_flags
& OCFS2_LOCK_INITIALIZED
))
2636 if (lockres
->l_ops
->flags
& LOCK_TYPE_USES_LVB
)
2637 lkm_flags
|= DLM_LKF_VALBLK
;
2639 spin_lock_irqsave(&lockres
->l_lock
, flags
);
2641 mlog_bug_on_msg(!(lockres
->l_flags
& OCFS2_LOCK_FREEING
),
2642 "lockres %s, flags 0x%lx\n",
2643 lockres
->l_name
, lockres
->l_flags
);
2645 while (lockres
->l_flags
& OCFS2_LOCK_BUSY
) {
2646 mlog(0, "waiting on busy lock \"%s\": flags = %lx, action = "
2647 "%u, unlock_action = %u\n",
2648 lockres
->l_name
, lockres
->l_flags
, lockres
->l_action
,
2649 lockres
->l_unlock_action
);
2651 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
2653 /* XXX: Today we just wait on any busy
2654 * locks... Perhaps we need to cancel converts in the
2656 ocfs2_wait_on_busy_lock(lockres
);
2658 spin_lock_irqsave(&lockres
->l_lock
, flags
);
2661 if (lockres
->l_ops
->flags
& LOCK_TYPE_USES_LVB
) {
2662 if (lockres
->l_flags
& OCFS2_LOCK_ATTACHED
&&
2663 lockres
->l_level
== DLM_LOCK_EX
&&
2664 !(lockres
->l_flags
& OCFS2_LOCK_NEEDS_REFRESH
))
2665 lockres
->l_ops
->set_lvb(lockres
);
2668 if (lockres
->l_flags
& OCFS2_LOCK_BUSY
)
2669 mlog(ML_ERROR
, "destroying busy lock: \"%s\"\n",
2671 if (lockres
->l_flags
& OCFS2_LOCK_BLOCKED
)
2672 mlog(0, "destroying blocked lock: \"%s\"\n", lockres
->l_name
);
2674 if (!(lockres
->l_flags
& OCFS2_LOCK_ATTACHED
)) {
2675 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
2679 lockres_clear_flags(lockres
, OCFS2_LOCK_ATTACHED
);
2681 /* make sure we never get here while waiting for an ast to
2683 BUG_ON(lockres
->l_action
!= OCFS2_AST_INVALID
);
2685 /* is this necessary? */
2686 lockres_or_flags(lockres
, OCFS2_LOCK_BUSY
);
2687 lockres
->l_unlock_action
= OCFS2_UNLOCK_DROP_LOCK
;
2688 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
2690 mlog(0, "lock %s\n", lockres
->l_name
);
2692 ret
= ocfs2_dlm_unlock(osb
->dlm
, &lockres
->l_lksb
, lkm_flags
,
2695 ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret
, lockres
);
2696 mlog(ML_ERROR
, "lockres flags: %lu\n", lockres
->l_flags
);
2697 /* XXX Need to abstract this */
2698 dlm_print_one_lock(lockres
->l_lksb
.lksb_o2dlm
.lockid
);
2701 mlog(0, "lock %s, successfull return from ocfs2_dlm_unlock\n",
2704 ocfs2_wait_on_busy_lock(lockres
);
2710 /* Mark the lockres as being dropped. It will no longer be
2711 * queued if blocking, but we still may have to wait on it
2712 * being dequeued from the downconvert thread before we can consider
2715 * You can *not* attempt to call cluster_lock on this lockres anymore. */
2716 void ocfs2_mark_lockres_freeing(struct ocfs2_lock_res
*lockres
)
2719 struct ocfs2_mask_waiter mw
;
2720 unsigned long flags
;
2722 ocfs2_init_mask_waiter(&mw
);
2724 spin_lock_irqsave(&lockres
->l_lock
, flags
);
2725 lockres
->l_flags
|= OCFS2_LOCK_FREEING
;
2726 while (lockres
->l_flags
& OCFS2_LOCK_QUEUED
) {
2727 lockres_add_mask_waiter(lockres
, &mw
, OCFS2_LOCK_QUEUED
, 0);
2728 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
2730 mlog(0, "Waiting on lockres %s\n", lockres
->l_name
);
2732 status
= ocfs2_wait_for_mask(&mw
);
2736 spin_lock_irqsave(&lockres
->l_lock
, flags
);
2738 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
2741 void ocfs2_simple_drop_lockres(struct ocfs2_super
*osb
,
2742 struct ocfs2_lock_res
*lockres
)
2746 ocfs2_mark_lockres_freeing(lockres
);
2747 ret
= ocfs2_drop_lock(osb
, lockres
);
2752 static void ocfs2_drop_osb_locks(struct ocfs2_super
*osb
)
2754 ocfs2_simple_drop_lockres(osb
, &osb
->osb_super_lockres
);
2755 ocfs2_simple_drop_lockres(osb
, &osb
->osb_rename_lockres
);
2758 int ocfs2_drop_inode_locks(struct inode
*inode
)
2764 /* No need to call ocfs2_mark_lockres_freeing here -
2765 * ocfs2_clear_inode has done it for us. */
2767 err
= ocfs2_drop_lock(OCFS2_SB(inode
->i_sb
),
2768 &OCFS2_I(inode
)->ip_open_lockres
);
2774 err
= ocfs2_drop_lock(OCFS2_SB(inode
->i_sb
),
2775 &OCFS2_I(inode
)->ip_inode_lockres
);
2778 if (err
< 0 && !status
)
2781 err
= ocfs2_drop_lock(OCFS2_SB(inode
->i_sb
),
2782 &OCFS2_I(inode
)->ip_rw_lockres
);
2785 if (err
< 0 && !status
)
2792 static void ocfs2_prepare_downconvert(struct ocfs2_lock_res
*lockres
,
2795 assert_spin_locked(&lockres
->l_lock
);
2797 BUG_ON(lockres
->l_blocking
<= DLM_LOCK_NL
);
2799 if (lockres
->l_level
<= new_level
) {
2800 mlog(ML_ERROR
, "lockres->l_level (%d) <= new_level (%d)\n",
2801 lockres
->l_level
, new_level
);
2805 mlog(0, "lock %s, new_level = %d, l_blocking = %d\n",
2806 lockres
->l_name
, new_level
, lockres
->l_blocking
);
2808 lockres
->l_action
= OCFS2_AST_DOWNCONVERT
;
2809 lockres
->l_requested
= new_level
;
2810 lockres_or_flags(lockres
, OCFS2_LOCK_BUSY
);
2813 static int ocfs2_downconvert_lock(struct ocfs2_super
*osb
,
2814 struct ocfs2_lock_res
*lockres
,
2819 u32 dlm_flags
= DLM_LKF_CONVERT
;
2824 dlm_flags
|= DLM_LKF_VALBLK
;
2826 ret
= ocfs2_dlm_lock(osb
->dlm
,
2831 OCFS2_LOCK_ID_MAX_LEN
- 1,
2834 ocfs2_log_dlm_error("ocfs2_dlm_lock", ret
, lockres
);
2835 ocfs2_recover_from_dlm_error(lockres
, 1);
2845 /* returns 1 when the caller should unlock and call ocfs2_dlm_unlock */
2846 static int ocfs2_prepare_cancel_convert(struct ocfs2_super
*osb
,
2847 struct ocfs2_lock_res
*lockres
)
2849 assert_spin_locked(&lockres
->l_lock
);
2852 mlog(0, "lock %s\n", lockres
->l_name
);
2854 if (lockres
->l_unlock_action
== OCFS2_UNLOCK_CANCEL_CONVERT
) {
2855 /* If we're already trying to cancel a lock conversion
2856 * then just drop the spinlock and allow the caller to
2857 * requeue this lock. */
2859 mlog(0, "Lockres %s, skip convert\n", lockres
->l_name
);
2863 /* were we in a convert when we got the bast fire? */
2864 BUG_ON(lockres
->l_action
!= OCFS2_AST_CONVERT
&&
2865 lockres
->l_action
!= OCFS2_AST_DOWNCONVERT
);
2866 /* set things up for the unlockast to know to just
2867 * clear out the ast_action and unset busy, etc. */
2868 lockres
->l_unlock_action
= OCFS2_UNLOCK_CANCEL_CONVERT
;
2870 mlog_bug_on_msg(!(lockres
->l_flags
& OCFS2_LOCK_BUSY
),
2871 "lock %s, invalid flags: 0x%lx\n",
2872 lockres
->l_name
, lockres
->l_flags
);
2877 static int ocfs2_cancel_convert(struct ocfs2_super
*osb
,
2878 struct ocfs2_lock_res
*lockres
)
2883 mlog(0, "lock %s\n", lockres
->l_name
);
2885 ret
= ocfs2_dlm_unlock(osb
->dlm
, &lockres
->l_lksb
,
2886 DLM_LKF_CANCEL
, lockres
);
2888 ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret
, lockres
);
2889 ocfs2_recover_from_dlm_error(lockres
, 0);
2892 mlog(0, "lock %s return from ocfs2_dlm_unlock\n", lockres
->l_name
);
2898 static int ocfs2_unblock_lock(struct ocfs2_super
*osb
,
2899 struct ocfs2_lock_res
*lockres
,
2900 struct ocfs2_unblock_ctl
*ctl
)
2902 unsigned long flags
;
2910 spin_lock_irqsave(&lockres
->l_lock
, flags
);
2912 BUG_ON(!(lockres
->l_flags
& OCFS2_LOCK_BLOCKED
));
2915 if (lockres
->l_flags
& OCFS2_LOCK_BUSY
) {
2917 ret
= ocfs2_prepare_cancel_convert(osb
, lockres
);
2918 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
2920 ret
= ocfs2_cancel_convert(osb
, lockres
);
2927 /* if we're blocking an exclusive and we have *any* holders,
2929 if ((lockres
->l_blocking
== DLM_LOCK_EX
)
2930 && (lockres
->l_ex_holders
|| lockres
->l_ro_holders
))
2933 /* If it's a PR we're blocking, then only
2934 * requeue if we've got any EX holders */
2935 if (lockres
->l_blocking
== DLM_LOCK_PR
&&
2936 lockres
->l_ex_holders
)
2940 * Can we get a lock in this state if the holder counts are
2941 * zero? The meta data unblock code used to check this.
2943 if ((lockres
->l_ops
->flags
& LOCK_TYPE_REQUIRES_REFRESH
)
2944 && (lockres
->l_flags
& OCFS2_LOCK_REFRESHING
))
2947 new_level
= ocfs2_highest_compat_lock_level(lockres
->l_blocking
);
2949 if (lockres
->l_ops
->check_downconvert
2950 && !lockres
->l_ops
->check_downconvert(lockres
, new_level
))
2953 /* If we get here, then we know that there are no more
2954 * incompatible holders (and anyone asking for an incompatible
2955 * lock is blocked). We can now downconvert the lock */
2956 if (!lockres
->l_ops
->downconvert_worker
)
2959 /* Some lockres types want to do a bit of work before
2960 * downconverting a lock. Allow that here. The worker function
2961 * may sleep, so we save off a copy of what we're blocking as
2962 * it may change while we're not holding the spin lock. */
2963 blocking
= lockres
->l_blocking
;
2964 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
2966 ctl
->unblock_action
= lockres
->l_ops
->downconvert_worker(lockres
, blocking
);
2968 if (ctl
->unblock_action
== UNBLOCK_STOP_POST
)
2971 spin_lock_irqsave(&lockres
->l_lock
, flags
);
2972 if (blocking
!= lockres
->l_blocking
) {
2973 /* If this changed underneath us, then we can't drop
2981 if (lockres
->l_ops
->flags
& LOCK_TYPE_USES_LVB
) {
2982 if (lockres
->l_level
== DLM_LOCK_EX
)
2986 * We only set the lvb if the lock has been fully
2987 * refreshed - otherwise we risk setting stale
2988 * data. Otherwise, there's no need to actually clear
2989 * out the lvb here as it's value is still valid.
2991 if (set_lvb
&& !(lockres
->l_flags
& OCFS2_LOCK_NEEDS_REFRESH
))
2992 lockres
->l_ops
->set_lvb(lockres
);
2995 ocfs2_prepare_downconvert(lockres
, new_level
);
2996 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
2997 ret
= ocfs2_downconvert_lock(osb
, lockres
, new_level
, set_lvb
);
3003 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
3010 static int ocfs2_data_convert_worker(struct ocfs2_lock_res
*lockres
,
3013 struct inode
*inode
;
3014 struct address_space
*mapping
;
3016 inode
= ocfs2_lock_res_inode(lockres
);
3017 mapping
= inode
->i_mapping
;
3019 if (!S_ISREG(inode
->i_mode
))
3023 * We need this before the filemap_fdatawrite() so that it can
3024 * transfer the dirty bit from the PTE to the
3025 * page. Unfortunately this means that even for EX->PR
3026 * downconverts, we'll lose our mappings and have to build
3029 unmap_mapping_range(mapping
, 0, 0, 0);
3031 if (filemap_fdatawrite(mapping
)) {
3032 mlog(ML_ERROR
, "Could not sync inode %llu for downconvert!",
3033 (unsigned long long)OCFS2_I(inode
)->ip_blkno
);
3035 sync_mapping_buffers(mapping
);
3036 if (blocking
== DLM_LOCK_EX
) {
3037 truncate_inode_pages(mapping
, 0);
3039 /* We only need to wait on the I/O if we're not also
3040 * truncating pages because truncate_inode_pages waits
3041 * for us above. We don't truncate pages if we're
3042 * blocking anything < EXMODE because we want to keep
3043 * them around in that case. */
3044 filemap_fdatawait(mapping
);
3048 return UNBLOCK_CONTINUE
;
3051 static int ocfs2_check_meta_downconvert(struct ocfs2_lock_res
*lockres
,
3054 struct inode
*inode
= ocfs2_lock_res_inode(lockres
);
3055 int checkpointed
= ocfs2_inode_fully_checkpointed(inode
);
3057 BUG_ON(new_level
!= DLM_LOCK_NL
&& new_level
!= DLM_LOCK_PR
);
3058 BUG_ON(lockres
->l_level
!= DLM_LOCK_EX
&& !checkpointed
);
3063 ocfs2_start_checkpoint(OCFS2_SB(inode
->i_sb
));
3067 static void ocfs2_set_meta_lvb(struct ocfs2_lock_res
*lockres
)
3069 struct inode
*inode
= ocfs2_lock_res_inode(lockres
);
3071 __ocfs2_stuff_meta_lvb(inode
);
3075 * Does the final reference drop on our dentry lock. Right now this
3076 * happens in the downconvert thread, but we could choose to simplify the
3077 * dlmglue API and push these off to the ocfs2_wq in the future.
3079 static void ocfs2_dentry_post_unlock(struct ocfs2_super
*osb
,
3080 struct ocfs2_lock_res
*lockres
)
3082 struct ocfs2_dentry_lock
*dl
= ocfs2_lock_res_dl(lockres
);
3083 ocfs2_dentry_lock_put(osb
, dl
);
3087 * d_delete() matching dentries before the lock downconvert.
3089 * At this point, any process waiting to destroy the
3090 * dentry_lock due to last ref count is stopped by the
3091 * OCFS2_LOCK_QUEUED flag.
3093 * We have two potential problems
3095 * 1) If we do the last reference drop on our dentry_lock (via dput)
3096 * we'll wind up in ocfs2_release_dentry_lock(), waiting on
3097 * the downconvert to finish. Instead we take an elevated
3098 * reference and push the drop until after we've completed our
3099 * unblock processing.
3101 * 2) There might be another process with a final reference,
3102 * waiting on us to finish processing. If this is the case, we
3103 * detect it and exit out - there's no more dentries anyway.
3105 static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res
*lockres
,
3108 struct ocfs2_dentry_lock
*dl
= ocfs2_lock_res_dl(lockres
);
3109 struct ocfs2_inode_info
*oi
= OCFS2_I(dl
->dl_inode
);
3110 struct dentry
*dentry
;
3111 unsigned long flags
;
3115 * This node is blocking another node from getting a read
3116 * lock. This happens when we've renamed within a
3117 * directory. We've forced the other nodes to d_delete(), but
3118 * we never actually dropped our lock because it's still
3119 * valid. The downconvert code will retain a PR for this node,
3120 * so there's no further work to do.
3122 if (blocking
== DLM_LOCK_PR
)
3123 return UNBLOCK_CONTINUE
;
3126 * Mark this inode as potentially orphaned. The code in
3127 * ocfs2_delete_inode() will figure out whether it actually
3128 * needs to be freed or not.
3130 spin_lock(&oi
->ip_lock
);
3131 oi
->ip_flags
|= OCFS2_INODE_MAYBE_ORPHANED
;
3132 spin_unlock(&oi
->ip_lock
);
3135 * Yuck. We need to make sure however that the check of
3136 * OCFS2_LOCK_FREEING and the extra reference are atomic with
3137 * respect to a reference decrement or the setting of that
3140 spin_lock_irqsave(&lockres
->l_lock
, flags
);
3141 spin_lock(&dentry_attach_lock
);
3142 if (!(lockres
->l_flags
& OCFS2_LOCK_FREEING
)
3147 spin_unlock(&dentry_attach_lock
);
3148 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
3150 mlog(0, "extra_ref = %d\n", extra_ref
);
3153 * We have a process waiting on us in ocfs2_dentry_iput(),
3154 * which means we can't have any more outstanding
3155 * aliases. There's no need to do any more work.
3158 return UNBLOCK_CONTINUE
;
3160 spin_lock(&dentry_attach_lock
);
3162 dentry
= ocfs2_find_local_alias(dl
->dl_inode
,
3163 dl
->dl_parent_blkno
, 1);
3166 spin_unlock(&dentry_attach_lock
);
3168 mlog(0, "d_delete(%.*s);\n", dentry
->d_name
.len
,
3169 dentry
->d_name
.name
);
3172 * The following dcache calls may do an
3173 * iput(). Normally we don't want that from the
3174 * downconverting thread, but in this case it's ok
3175 * because the requesting node already has an
3176 * exclusive lock on the inode, so it can't be queued
3177 * for a downconvert.
3182 spin_lock(&dentry_attach_lock
);
3184 spin_unlock(&dentry_attach_lock
);
3187 * If we are the last holder of this dentry lock, there is no
3188 * reason to downconvert so skip straight to the unlock.
3190 if (dl
->dl_count
== 1)
3191 return UNBLOCK_STOP_POST
;
3193 return UNBLOCK_CONTINUE_POST
;
3196 static struct ocfs2_locking_protocol lproto
= {
3197 .lp_lock_ast
= ocfs2_locking_ast
,
3198 .lp_blocking_ast
= ocfs2_blocking_ast
,
3199 .lp_unlock_ast
= ocfs2_unlock_ast
,
3202 /* This interface isn't the final one, hence the less-than-perfect names */
3203 void dlmglue_init_stack(void)
3205 o2cb_get_stack(&lproto
);
3208 void dlmglue_exit_stack(void)
3213 static void ocfs2_process_blocked_lock(struct ocfs2_super
*osb
,
3214 struct ocfs2_lock_res
*lockres
)
3217 struct ocfs2_unblock_ctl ctl
= {0, 0,};
3218 unsigned long flags
;
3220 /* Our reference to the lockres in this function can be
3221 * considered valid until we remove the OCFS2_LOCK_QUEUED
3227 BUG_ON(!lockres
->l_ops
);
3229 mlog(0, "lockres %s blocked.\n", lockres
->l_name
);
3231 /* Detect whether a lock has been marked as going away while
3232 * the downconvert thread was processing other things. A lock can
3233 * still be marked with OCFS2_LOCK_FREEING after this check,
3234 * but short circuiting here will still save us some
3236 spin_lock_irqsave(&lockres
->l_lock
, flags
);
3237 if (lockres
->l_flags
& OCFS2_LOCK_FREEING
)
3239 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
3241 status
= ocfs2_unblock_lock(osb
, lockres
, &ctl
);
3245 spin_lock_irqsave(&lockres
->l_lock
, flags
);
3247 if (lockres
->l_flags
& OCFS2_LOCK_FREEING
|| !ctl
.requeue
) {
3248 lockres_clear_flags(lockres
, OCFS2_LOCK_QUEUED
);
3250 ocfs2_schedule_blocked_lock(osb
, lockres
);
3252 mlog(0, "lockres %s, requeue = %s.\n", lockres
->l_name
,
3253 ctl
.requeue
? "yes" : "no");
3254 spin_unlock_irqrestore(&lockres
->l_lock
, flags
);
3256 if (ctl
.unblock_action
!= UNBLOCK_CONTINUE
3257 && lockres
->l_ops
->post_unlock
)
3258 lockres
->l_ops
->post_unlock(osb
, lockres
);
3263 static void ocfs2_schedule_blocked_lock(struct ocfs2_super
*osb
,
3264 struct ocfs2_lock_res
*lockres
)
3268 assert_spin_locked(&lockres
->l_lock
);
3270 if (lockres
->l_flags
& OCFS2_LOCK_FREEING
) {
3271 /* Do not schedule a lock for downconvert when it's on
3272 * the way to destruction - any nodes wanting access
3273 * to the resource will get it soon. */
3274 mlog(0, "Lockres %s won't be scheduled: flags 0x%lx\n",
3275 lockres
->l_name
, lockres
->l_flags
);
3279 lockres_or_flags(lockres
, OCFS2_LOCK_QUEUED
);
3281 spin_lock(&osb
->dc_task_lock
);
3282 if (list_empty(&lockres
->l_blocked_list
)) {
3283 list_add_tail(&lockres
->l_blocked_list
,
3284 &osb
->blocked_lock_list
);
3285 osb
->blocked_lock_count
++;
3287 spin_unlock(&osb
->dc_task_lock
);
3292 static void ocfs2_downconvert_thread_do_work(struct ocfs2_super
*osb
)
3294 unsigned long processed
;
3295 struct ocfs2_lock_res
*lockres
;
3299 spin_lock(&osb
->dc_task_lock
);
3300 /* grab this early so we know to try again if a state change and
3301 * wake happens part-way through our work */
3302 osb
->dc_work_sequence
= osb
->dc_wake_sequence
;
3304 processed
= osb
->blocked_lock_count
;
3306 BUG_ON(list_empty(&osb
->blocked_lock_list
));
3308 lockres
= list_entry(osb
->blocked_lock_list
.next
,
3309 struct ocfs2_lock_res
, l_blocked_list
);
3310 list_del_init(&lockres
->l_blocked_list
);
3311 osb
->blocked_lock_count
--;
3312 spin_unlock(&osb
->dc_task_lock
);
3317 ocfs2_process_blocked_lock(osb
, lockres
);
3319 spin_lock(&osb
->dc_task_lock
);
3321 spin_unlock(&osb
->dc_task_lock
);
3326 static int ocfs2_downconvert_thread_lists_empty(struct ocfs2_super
*osb
)
3330 spin_lock(&osb
->dc_task_lock
);
3331 if (list_empty(&osb
->blocked_lock_list
))
3334 spin_unlock(&osb
->dc_task_lock
);
3338 static int ocfs2_downconvert_thread_should_wake(struct ocfs2_super
*osb
)
3340 int should_wake
= 0;
3342 spin_lock(&osb
->dc_task_lock
);
3343 if (osb
->dc_work_sequence
!= osb
->dc_wake_sequence
)
3345 spin_unlock(&osb
->dc_task_lock
);
3350 static int ocfs2_downconvert_thread(void *arg
)
3353 struct ocfs2_super
*osb
= arg
;
3355 /* only quit once we've been asked to stop and there is no more
3357 while (!(kthread_should_stop() &&
3358 ocfs2_downconvert_thread_lists_empty(osb
))) {
3360 wait_event_interruptible(osb
->dc_event
,
3361 ocfs2_downconvert_thread_should_wake(osb
) ||
3362 kthread_should_stop());
3364 mlog(0, "downconvert_thread: awoken\n");
3366 ocfs2_downconvert_thread_do_work(osb
);
3369 osb
->dc_task
= NULL
;
3373 void ocfs2_wake_downconvert_thread(struct ocfs2_super
*osb
)
3375 spin_lock(&osb
->dc_task_lock
);
3376 /* make sure the voting thread gets a swipe at whatever changes
3377 * the caller may have made to the voting state */
3378 osb
->dc_wake_sequence
++;
3379 spin_unlock(&osb
->dc_task_lock
);
3380 wake_up(&osb
->dc_event
);