blob: 9479f99c2145c623c9c740eda6ac8c3e21b93a67 [file] [log] [blame]
Mark Fashehccd979b2005-12-15 14:31:24 -08001/* -*- mode: c; c-basic-offset: 8; -*-
2 * vim: noexpandtab sw=8 ts=8 sts=0:
3 *
4 * 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>
Mark Fashehccd979b2005-12-15 14:31:24 -080030#include <linux/kthread.h>
31#include <linux/pagemap.h>
32#include <linux/debugfs.h>
33#include <linux/seq_file.h>
Sunil Mushran8ddb7b02008-05-13 13:45:15 -070034#include <linux/time.h>
Jan Kara9e33d692008-08-25 19:56:50 +020035#include <linux/quotaops.h>
Ingo Molnar174cd4b2017-02-02 19:15:33 +010036#include <linux/sched/signal.h>
Mark Fashehccd979b2005-12-15 14:31:24 -080037
Mark Fashehccd979b2005-12-15 14:31:24 -080038#define MLOG_MASK_PREFIX ML_DLM_GLUE
39#include <cluster/masklog.h>
40
41#include "ocfs2.h"
Joel Beckerd24fbcd2008-01-25 17:02:21 -080042#include "ocfs2_lockingver.h"
Mark Fashehccd979b2005-12-15 14:31:24 -080043
44#include "alloc.h"
Mark Fashehd680efe2006-09-08 14:14:34 -070045#include "dcache.h"
Mark Fashehccd979b2005-12-15 14:31:24 -080046#include "dlmglue.h"
47#include "extent_map.h"
Tiger Yang7f1a37e2006-11-15 15:48:42 +080048#include "file.h"
Mark Fashehccd979b2005-12-15 14:31:24 -080049#include "heartbeat.h"
50#include "inode.h"
51#include "journal.h"
Joel Becker24ef1812008-01-29 17:37:32 -080052#include "stackglue.h"
Mark Fashehccd979b2005-12-15 14:31:24 -080053#include "slot_map.h"
54#include "super.h"
55#include "uptodate.h"
Jan Kara9e33d692008-08-25 19:56:50 +020056#include "quota.h"
Tao Ma8dec98e2009-08-18 11:19:58 +080057#include "refcounttree.h"
Andreas Gruenbacherb8a7a3a2016-03-24 14:38:37 +010058#include "acl.h"
Mark Fashehccd979b2005-12-15 14:31:24 -080059
60#include "buffer_head_io.h"
61
62struct ocfs2_mask_waiter {
63 struct list_head mw_item;
64 int mw_status;
65 struct completion mw_complete;
66 unsigned long mw_mask;
67 unsigned long mw_goal;
Sunil Mushran8ddb7b02008-05-13 13:45:15 -070068#ifdef CONFIG_OCFS2_FS_STATS
Sunil Mushran5bc970e2010-12-28 23:26:03 -080069 ktime_t mw_lock_start;
Sunil Mushran8ddb7b02008-05-13 13:45:15 -070070#endif
Mark Fashehccd979b2005-12-15 14:31:24 -080071};
72
Mark Fasheh54a7e752006-09-12 21:49:13 -070073static struct ocfs2_super *ocfs2_get_dentry_osb(struct ocfs2_lock_res *lockres);
74static struct ocfs2_super *ocfs2_get_inode_osb(struct ocfs2_lock_res *lockres);
Mark Fashehcf8e06f2007-12-20 16:43:10 -080075static struct ocfs2_super *ocfs2_get_file_osb(struct ocfs2_lock_res *lockres);
Jan Kara9e33d692008-08-25 19:56:50 +020076static struct ocfs2_super *ocfs2_get_qinfo_osb(struct ocfs2_lock_res *lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -080077
Mark Fashehd680efe2006-09-08 14:14:34 -070078/*
Mark Fashehcc567d82006-09-13 21:52:21 -070079 * Return value from ->downconvert_worker functions.
Mark Fashehd680efe2006-09-08 14:14:34 -070080 *
Mark Fashehb5e500e2006-09-13 22:01:16 -070081 * These control the precise actions of ocfs2_unblock_lock()
Mark Fashehd680efe2006-09-08 14:14:34 -070082 * and ocfs2_process_blocked_lock()
83 *
84 */
85enum ocfs2_unblock_action {
86 UNBLOCK_CONTINUE = 0, /* Continue downconvert */
87 UNBLOCK_CONTINUE_POST = 1, /* Continue downconvert, fire
88 * ->post_unlock callback */
89 UNBLOCK_STOP_POST = 2, /* Do not downconvert, fire
90 * ->post_unlock() callback. */
91};
92
93struct ocfs2_unblock_ctl {
94 int requeue;
95 enum ocfs2_unblock_action unblock_action;
96};
97
Jan Karacb257972009-06-04 15:26:50 +020098/* Lockdep class keys */
99struct lock_class_key lockdep_keys[OCFS2_NUM_LOCK_TYPES];
100
Mark Fasheh810d5ae2006-09-13 21:39:52 -0700101static int ocfs2_check_meta_downconvert(struct ocfs2_lock_res *lockres,
102 int new_level);
103static void ocfs2_set_meta_lvb(struct ocfs2_lock_res *lockres);
104
Mark Fashehcc567d82006-09-13 21:52:21 -0700105static int ocfs2_data_convert_worker(struct ocfs2_lock_res *lockres,
106 int blocking);
107
Mark Fashehcc567d82006-09-13 21:52:21 -0700108static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res *lockres,
109 int blocking);
Mark Fashehd680efe2006-09-08 14:14:34 -0700110
111static void ocfs2_dentry_post_unlock(struct ocfs2_super *osb,
112 struct ocfs2_lock_res *lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -0800113
Jan Kara9e33d692008-08-25 19:56:50 +0200114static void ocfs2_set_qinfo_lvb(struct ocfs2_lock_res *lockres);
Adrian Bunk6cb129f2007-04-26 00:29:35 -0700115
Tao Ma8dec98e2009-08-18 11:19:58 +0800116static int ocfs2_check_refcount_downconvert(struct ocfs2_lock_res *lockres,
117 int new_level);
118static int ocfs2_refcount_convert_worker(struct ocfs2_lock_res *lockres,
119 int blocking);
120
Adrian Bunk6cb129f2007-04-26 00:29:35 -0700121#define mlog_meta_lvb(__level, __lockres) ocfs2_dump_meta_lvb_info(__level, __PRETTY_FUNCTION__, __LINE__, __lockres)
122
123/* This aids in debugging situations where a bad LVB might be involved. */
124static void ocfs2_dump_meta_lvb_info(u64 level,
125 const char *function,
126 unsigned int line,
127 struct ocfs2_lock_res *lockres)
128{
Mark Fasheha641dc22008-12-24 16:03:48 -0800129 struct ocfs2_meta_lvb *lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
Adrian Bunk6cb129f2007-04-26 00:29:35 -0700130
131 mlog(level, "LVB information for %s (called from %s:%u):\n",
132 lockres->l_name, function, line);
133 mlog(level, "version: %u, clusters: %u, generation: 0x%x\n",
134 lvb->lvb_version, be32_to_cpu(lvb->lvb_iclusters),
135 be32_to_cpu(lvb->lvb_igeneration));
136 mlog(level, "size: %llu, uid %u, gid %u, mode 0x%x\n",
137 (unsigned long long)be64_to_cpu(lvb->lvb_isize),
138 be32_to_cpu(lvb->lvb_iuid), be32_to_cpu(lvb->lvb_igid),
139 be16_to_cpu(lvb->lvb_imode));
140 mlog(level, "nlink %u, atime_packed 0x%llx, ctime_packed 0x%llx, "
141 "mtime_packed 0x%llx iattr 0x%x\n", be16_to_cpu(lvb->lvb_inlink),
142 (long long)be64_to_cpu(lvb->lvb_iatime_packed),
143 (long long)be64_to_cpu(lvb->lvb_ictime_packed),
144 (long long)be64_to_cpu(lvb->lvb_imtime_packed),
145 be32_to_cpu(lvb->lvb_iattr));
146}
147
148
Mark Fashehf625c972006-09-12 21:24:53 -0700149/*
150 * OCFS2 Lock Resource Operations
151 *
152 * These fine tune the behavior of the generic dlmglue locking infrastructure.
Mark Fasheh0d5dc6c2006-09-14 14:44:51 -0700153 *
154 * The most basic of lock types can point ->l_priv to their respective
155 * struct ocfs2_super and allow the default actions to manage things.
156 *
157 * Right now, each lock type also needs to implement an init function,
158 * and trivial lock/unlock wrappers. ocfs2_simple_drop_lockres()
159 * should be called when the lock is no longer needed (i.e., object
160 * destruction time).
Mark Fashehf625c972006-09-12 21:24:53 -0700161 */
Mark Fashehccd979b2005-12-15 14:31:24 -0800162struct ocfs2_lock_res_ops {
Mark Fasheh54a7e752006-09-12 21:49:13 -0700163 /*
164 * Translate an ocfs2_lock_res * into an ocfs2_super *. Define
165 * this callback if ->l_priv is not an ocfs2_super pointer
166 */
167 struct ocfs2_super * (*get_osb)(struct ocfs2_lock_res *);
Mark Fashehb5e500e2006-09-13 22:01:16 -0700168
Mark Fasheh0d5dc6c2006-09-14 14:44:51 -0700169 /*
Mark Fasheh34d024f2007-09-24 15:56:19 -0700170 * Optionally called in the downconvert thread after a
171 * successful downconvert. The lockres will not be referenced
172 * after this callback is called, so it is safe to free
173 * memory, etc.
Mark Fasheh0d5dc6c2006-09-14 14:44:51 -0700174 *
175 * The exact semantics of when this is called are controlled
176 * by ->downconvert_worker()
177 */
Mark Fashehd680efe2006-09-08 14:14:34 -0700178 void (*post_unlock)(struct ocfs2_super *, struct ocfs2_lock_res *);
Mark Fashehf625c972006-09-12 21:24:53 -0700179
180 /*
Mark Fasheh16d5b952006-09-13 21:10:12 -0700181 * Allow a lock type to add checks to determine whether it is
182 * safe to downconvert a lock. Return 0 to re-queue the
183 * downconvert at a later time, nonzero to continue.
184 *
185 * For most locks, the default checks that there are no
186 * incompatible holders are sufficient.
187 *
188 * Called with the lockres spinlock held.
189 */
190 int (*check_downconvert)(struct ocfs2_lock_res *, int);
191
192 /*
Mark Fasheh5ef0d4e2006-09-13 21:21:52 -0700193 * Allows a lock type to populate the lock value block. This
194 * is called on downconvert, and when we drop a lock.
195 *
196 * Locks that want to use this should set LOCK_TYPE_USES_LVB
197 * in the flags field.
198 *
199 * Called with the lockres spinlock held.
200 */
201 void (*set_lvb)(struct ocfs2_lock_res *);
202
203 /*
Mark Fashehcc567d82006-09-13 21:52:21 -0700204 * Called from the downconvert thread when it is determined
205 * that a lock will be downconverted. This is called without
206 * any locks held so the function can do work that might
207 * schedule (syncing out data, etc).
208 *
209 * This should return any one of the ocfs2_unblock_action
210 * values, depending on what it wants the thread to do.
211 */
212 int (*downconvert_worker)(struct ocfs2_lock_res *, int);
213
214 /*
Mark Fashehf625c972006-09-12 21:24:53 -0700215 * LOCK_TYPE_* flags which describe the specific requirements
216 * of a lock type. Descriptions of each individual flag follow.
217 */
218 int flags;
Mark Fashehccd979b2005-12-15 14:31:24 -0800219};
220
Mark Fashehf625c972006-09-12 21:24:53 -0700221/*
222 * Some locks want to "refresh" potentially stale data when a
223 * meaningful (PRMODE or EXMODE) lock level is first obtained. If this
224 * flag is set, the OCFS2_LOCK_NEEDS_REFRESH flag will be set on the
225 * individual lockres l_flags member from the ast function. It is
226 * expected that the locking wrapper will clear the
227 * OCFS2_LOCK_NEEDS_REFRESH flag when done.
228 */
229#define LOCK_TYPE_REQUIRES_REFRESH 0x1
230
Mark Fashehb80fc012006-09-12 22:08:14 -0700231/*
Mark Fasheh5ef0d4e2006-09-13 21:21:52 -0700232 * Indicate that a lock type makes use of the lock value block. The
233 * ->set_lvb lock type callback must be defined.
Mark Fashehb80fc012006-09-12 22:08:14 -0700234 */
235#define LOCK_TYPE_USES_LVB 0x2
236
Mark Fashehccd979b2005-12-15 14:31:24 -0800237static struct ocfs2_lock_res_ops ocfs2_inode_rw_lops = {
Mark Fasheh54a7e752006-09-12 21:49:13 -0700238 .get_osb = ocfs2_get_inode_osb,
Mark Fashehf625c972006-09-12 21:24:53 -0700239 .flags = 0,
Mark Fashehccd979b2005-12-15 14:31:24 -0800240};
241
Mark Fashehe63aecb62007-10-18 15:30:42 -0700242static struct ocfs2_lock_res_ops ocfs2_inode_inode_lops = {
Mark Fasheh54a7e752006-09-12 21:49:13 -0700243 .get_osb = ocfs2_get_inode_osb,
Mark Fasheh810d5ae2006-09-13 21:39:52 -0700244 .check_downconvert = ocfs2_check_meta_downconvert,
245 .set_lvb = ocfs2_set_meta_lvb,
Mark Fashehf1f54062007-10-18 15:13:59 -0700246 .downconvert_worker = ocfs2_data_convert_worker,
Mark Fashehb80fc012006-09-12 22:08:14 -0700247 .flags = LOCK_TYPE_REQUIRES_REFRESH|LOCK_TYPE_USES_LVB,
Mark Fashehccd979b2005-12-15 14:31:24 -0800248};
249
Mark Fashehccd979b2005-12-15 14:31:24 -0800250static struct ocfs2_lock_res_ops ocfs2_super_lops = {
Mark Fashehf625c972006-09-12 21:24:53 -0700251 .flags = LOCK_TYPE_REQUIRES_REFRESH,
Mark Fashehccd979b2005-12-15 14:31:24 -0800252};
253
254static struct ocfs2_lock_res_ops ocfs2_rename_lops = {
Mark Fashehf625c972006-09-12 21:24:53 -0700255 .flags = 0,
Mark Fashehccd979b2005-12-15 14:31:24 -0800256};
257
wengang wang6ca497a2009-03-06 21:29:10 +0800258static struct ocfs2_lock_res_ops ocfs2_nfs_sync_lops = {
259 .flags = 0,
260};
261
Gang He4882abe2018-01-31 16:15:10 -0800262static struct ocfs2_lock_res_ops ocfs2_trim_fs_lops = {
263 .flags = LOCK_TYPE_REQUIRES_REFRESH|LOCK_TYPE_USES_LVB,
264};
265
Srinivas Eeda83273932009-06-03 17:02:55 -0700266static struct ocfs2_lock_res_ops ocfs2_orphan_scan_lops = {
267 .flags = LOCK_TYPE_REQUIRES_REFRESH|LOCK_TYPE_USES_LVB,
268};
269
Mark Fashehd680efe2006-09-08 14:14:34 -0700270static struct ocfs2_lock_res_ops ocfs2_dentry_lops = {
Mark Fasheh54a7e752006-09-12 21:49:13 -0700271 .get_osb = ocfs2_get_dentry_osb,
Mark Fashehd680efe2006-09-08 14:14:34 -0700272 .post_unlock = ocfs2_dentry_post_unlock,
Mark Fashehcc567d82006-09-13 21:52:21 -0700273 .downconvert_worker = ocfs2_dentry_convert_worker,
Mark Fashehf625c972006-09-12 21:24:53 -0700274 .flags = 0,
Mark Fashehd680efe2006-09-08 14:14:34 -0700275};
276
Tiger Yang50008632007-03-20 16:01:38 -0700277static struct ocfs2_lock_res_ops ocfs2_inode_open_lops = {
278 .get_osb = ocfs2_get_inode_osb,
279 .flags = 0,
280};
281
Mark Fashehcf8e06f2007-12-20 16:43:10 -0800282static struct ocfs2_lock_res_ops ocfs2_flock_lops = {
283 .get_osb = ocfs2_get_file_osb,
284 .flags = 0,
285};
286
Jan Kara9e33d692008-08-25 19:56:50 +0200287static struct ocfs2_lock_res_ops ocfs2_qinfo_lops = {
288 .set_lvb = ocfs2_set_qinfo_lvb,
289 .get_osb = ocfs2_get_qinfo_osb,
290 .flags = LOCK_TYPE_REQUIRES_REFRESH | LOCK_TYPE_USES_LVB,
291};
292
Tao Ma8dec98e2009-08-18 11:19:58 +0800293static struct ocfs2_lock_res_ops ocfs2_refcount_block_lops = {
294 .check_downconvert = ocfs2_check_refcount_downconvert,
295 .downconvert_worker = ocfs2_refcount_convert_worker,
296 .flags = 0,
297};
298
Mark Fashehccd979b2005-12-15 14:31:24 -0800299static inline int ocfs2_is_inode_lock(struct ocfs2_lock_res *lockres)
300{
301 return lockres->l_type == OCFS2_LOCK_TYPE_META ||
Tiger Yang50008632007-03-20 16:01:38 -0700302 lockres->l_type == OCFS2_LOCK_TYPE_RW ||
303 lockres->l_type == OCFS2_LOCK_TYPE_OPEN;
Mark Fashehccd979b2005-12-15 14:31:24 -0800304}
305
Joel Beckerc0e413382010-01-29 14:46:44 -0800306static inline struct ocfs2_lock_res *ocfs2_lksb_to_lock_res(struct ocfs2_dlm_lksb *lksb)
Joel Beckera796d282010-01-28 19:22:39 -0800307{
308 return container_of(lksb, struct ocfs2_lock_res, l_lksb);
309}
310
Mark Fashehccd979b2005-12-15 14:31:24 -0800311static inline struct inode *ocfs2_lock_res_inode(struct ocfs2_lock_res *lockres)
312{
313 BUG_ON(!ocfs2_is_inode_lock(lockres));
314
315 return (struct inode *) lockres->l_priv;
316}
317
Mark Fashehd680efe2006-09-08 14:14:34 -0700318static inline struct ocfs2_dentry_lock *ocfs2_lock_res_dl(struct ocfs2_lock_res *lockres)
319{
320 BUG_ON(lockres->l_type != OCFS2_LOCK_TYPE_DENTRY);
321
322 return (struct ocfs2_dentry_lock *)lockres->l_priv;
323}
324
Jan Kara9e33d692008-08-25 19:56:50 +0200325static inline struct ocfs2_mem_dqinfo *ocfs2_lock_res_qinfo(struct ocfs2_lock_res *lockres)
326{
327 BUG_ON(lockres->l_type != OCFS2_LOCK_TYPE_QINFO);
328
329 return (struct ocfs2_mem_dqinfo *)lockres->l_priv;
330}
331
Tao Ma8dec98e2009-08-18 11:19:58 +0800332static inline struct ocfs2_refcount_tree *
333ocfs2_lock_res_refcount_tree(struct ocfs2_lock_res *res)
334{
335 return container_of(res, struct ocfs2_refcount_tree, rf_lockres);
336}
337
Mark Fasheh54a7e752006-09-12 21:49:13 -0700338static inline struct ocfs2_super *ocfs2_get_lockres_osb(struct ocfs2_lock_res *lockres)
339{
340 if (lockres->l_ops->get_osb)
341 return lockres->l_ops->get_osb(lockres);
342
343 return (struct ocfs2_super *)lockres->l_priv;
344}
345
Mark Fashehccd979b2005-12-15 14:31:24 -0800346static int ocfs2_lock_create(struct ocfs2_super *osb,
347 struct ocfs2_lock_res *lockres,
348 int level,
Joel Beckerbd3e7612008-02-01 12:14:57 -0800349 u32 dlm_flags);
Mark Fashehccd979b2005-12-15 14:31:24 -0800350static inline int ocfs2_may_continue_on_blocked_lock(struct ocfs2_lock_res *lockres,
351 int wanted);
Jan Karacb257972009-06-04 15:26:50 +0200352static void __ocfs2_cluster_unlock(struct ocfs2_super *osb,
353 struct ocfs2_lock_res *lockres,
354 int level, unsigned long caller_ip);
355static inline void ocfs2_cluster_unlock(struct ocfs2_super *osb,
356 struct ocfs2_lock_res *lockres,
357 int level)
358{
359 __ocfs2_cluster_unlock(osb, lockres, level, _RET_IP_);
360}
361
Mark Fashehccd979b2005-12-15 14:31:24 -0800362static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res *lockres);
363static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res *lockres);
364static inline void ocfs2_generic_handle_attach_action(struct ocfs2_lock_res *lockres);
365static int ocfs2_generic_handle_bast(struct ocfs2_lock_res *lockres, int level);
366static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb,
367 struct ocfs2_lock_res *lockres);
368static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres,
369 int convert);
Sunil Mushranc74ff8b2009-02-03 12:37:14 -0800370#define ocfs2_log_dlm_error(_func, _err, _lockres) do { \
371 if ((_lockres)->l_type != OCFS2_LOCK_TYPE_DENTRY) \
372 mlog(ML_ERROR, "DLM error %d while calling %s on resource %s\n", \
373 _err, _func, _lockres->l_name); \
374 else \
375 mlog(ML_ERROR, "DLM error %d while calling %s on resource %.*s%08x\n", \
376 _err, _func, OCFS2_DENTRY_LOCK_INO_START - 1, (_lockres)->l_name, \
377 (unsigned int)ocfs2_get_dentry_lock_ino(_lockres)); \
Mark Fashehccd979b2005-12-15 14:31:24 -0800378} while (0)
Mark Fasheh34d024f2007-09-24 15:56:19 -0700379static int ocfs2_downconvert_thread(void *arg);
380static void ocfs2_downconvert_on_unlock(struct ocfs2_super *osb,
381 struct ocfs2_lock_res *lockres);
Mark Fashehe63aecb62007-10-18 15:30:42 -0700382static int ocfs2_inode_lock_update(struct inode *inode,
Mark Fashehccd979b2005-12-15 14:31:24 -0800383 struct buffer_head **bh);
384static void ocfs2_drop_osb_locks(struct ocfs2_super *osb);
385static inline int ocfs2_highest_compat_lock_level(int level);
Joel Beckerde551242008-02-01 14:45:08 -0800386static unsigned int ocfs2_prepare_downconvert(struct ocfs2_lock_res *lockres,
387 int new_level);
Mark Fashehcf8e06f2007-12-20 16:43:10 -0800388static int ocfs2_downconvert_lock(struct ocfs2_super *osb,
389 struct ocfs2_lock_res *lockres,
390 int new_level,
Joel Beckerde551242008-02-01 14:45:08 -0800391 int lvb,
392 unsigned int generation);
Mark Fashehcf8e06f2007-12-20 16:43:10 -0800393static int ocfs2_prepare_cancel_convert(struct ocfs2_super *osb,
394 struct ocfs2_lock_res *lockres);
395static int ocfs2_cancel_convert(struct ocfs2_super *osb,
396 struct ocfs2_lock_res *lockres);
397
Mark Fashehccd979b2005-12-15 14:31:24 -0800398
Mark Fashehccd979b2005-12-15 14:31:24 -0800399static void ocfs2_build_lock_name(enum ocfs2_lock_type type,
400 u64 blkno,
401 u32 generation,
402 char *name)
403{
404 int len;
405
Mark Fashehccd979b2005-12-15 14:31:24 -0800406 BUG_ON(type >= OCFS2_NUM_LOCK_TYPES);
407
Mark Fashehb06970532006-03-03 10:24:33 -0800408 len = snprintf(name, OCFS2_LOCK_ID_MAX_LEN, "%c%s%016llx%08x",
409 ocfs2_lock_type_char(type), OCFS2_LOCK_ID_PAD,
410 (long long)blkno, generation);
Mark Fashehccd979b2005-12-15 14:31:24 -0800411
412 BUG_ON(len != (OCFS2_LOCK_ID_MAX_LEN - 1));
413
414 mlog(0, "built lock resource with name: %s\n", name);
Mark Fashehccd979b2005-12-15 14:31:24 -0800415}
416
Ingo Molnar34af9462006-06-27 02:53:55 -0700417static DEFINE_SPINLOCK(ocfs2_dlm_tracking_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -0800418
419static void ocfs2_add_lockres_tracking(struct ocfs2_lock_res *res,
420 struct ocfs2_dlm_debug *dlm_debug)
421{
422 mlog(0, "Add tracking for lockres %s\n", res->l_name);
423
424 spin_lock(&ocfs2_dlm_tracking_lock);
425 list_add(&res->l_debug_list, &dlm_debug->d_lockres_tracking);
426 spin_unlock(&ocfs2_dlm_tracking_lock);
427}
428
429static void ocfs2_remove_lockres_tracking(struct ocfs2_lock_res *res)
430{
431 spin_lock(&ocfs2_dlm_tracking_lock);
432 if (!list_empty(&res->l_debug_list))
433 list_del_init(&res->l_debug_list);
434 spin_unlock(&ocfs2_dlm_tracking_lock);
435}
436
Sunil Mushran8ddb7b02008-05-13 13:45:15 -0700437#ifdef CONFIG_OCFS2_FS_STATS
438static void ocfs2_init_lock_stats(struct ocfs2_lock_res *res)
439{
Sunil Mushran8ddb7b02008-05-13 13:45:15 -0700440 res->l_lock_refresh = 0;
Sunil Mushran5bc970e2010-12-28 23:26:03 -0800441 memset(&res->l_lock_prmode, 0, sizeof(struct ocfs2_lock_stats));
442 memset(&res->l_lock_exmode, 0, sizeof(struct ocfs2_lock_stats));
Sunil Mushran8ddb7b02008-05-13 13:45:15 -0700443}
444
445static void ocfs2_update_lock_stats(struct ocfs2_lock_res *res, int level,
446 struct ocfs2_mask_waiter *mw, int ret)
447{
Sunil Mushran5bc970e2010-12-28 23:26:03 -0800448 u32 usec;
449 ktime_t kt;
450 struct ocfs2_lock_stats *stats;
Sunil Mushran8ddb7b02008-05-13 13:45:15 -0700451
Sunil Mushran5bc970e2010-12-28 23:26:03 -0800452 if (level == LKM_PRMODE)
453 stats = &res->l_lock_prmode;
454 else if (level == LKM_EXMODE)
455 stats = &res->l_lock_exmode;
456 else
Sunil Mushran8ddb7b02008-05-13 13:45:15 -0700457 return;
458
Sunil Mushran5bc970e2010-12-28 23:26:03 -0800459 kt = ktime_sub(ktime_get(), mw->mw_lock_start);
460 usec = ktime_to_us(kt);
461
462 stats->ls_gets++;
463 stats->ls_total += ktime_to_ns(kt);
464 /* overflow */
roel16865b72011-12-12 23:40:51 +0100465 if (unlikely(stats->ls_gets == 0)) {
Sunil Mushran5bc970e2010-12-28 23:26:03 -0800466 stats->ls_gets++;
467 stats->ls_total = ktime_to_ns(kt);
468 }
469
470 if (stats->ls_max < usec)
471 stats->ls_max = usec;
472
Sunil Mushran8ddb7b02008-05-13 13:45:15 -0700473 if (ret)
Sunil Mushran5bc970e2010-12-28 23:26:03 -0800474 stats->ls_fail++;
Sunil Mushran8ddb7b02008-05-13 13:45:15 -0700475}
476
477static inline void ocfs2_track_lock_refresh(struct ocfs2_lock_res *lockres)
478{
479 lockres->l_lock_refresh++;
480}
481
482static inline void ocfs2_init_start_time(struct ocfs2_mask_waiter *mw)
483{
Sunil Mushran5bc970e2010-12-28 23:26:03 -0800484 mw->mw_lock_start = ktime_get();
Sunil Mushran8ddb7b02008-05-13 13:45:15 -0700485}
486#else
487static inline void ocfs2_init_lock_stats(struct ocfs2_lock_res *res)
488{
489}
490static inline void ocfs2_update_lock_stats(struct ocfs2_lock_res *res,
491 int level, struct ocfs2_mask_waiter *mw, int ret)
492{
493}
494static inline void ocfs2_track_lock_refresh(struct ocfs2_lock_res *lockres)
495{
496}
497static inline void ocfs2_init_start_time(struct ocfs2_mask_waiter *mw)
498{
499}
500#endif
501
Mark Fashehccd979b2005-12-15 14:31:24 -0800502static void ocfs2_lock_res_init_common(struct ocfs2_super *osb,
503 struct ocfs2_lock_res *res,
504 enum ocfs2_lock_type type,
Mark Fashehccd979b2005-12-15 14:31:24 -0800505 struct ocfs2_lock_res_ops *ops,
506 void *priv)
507{
Mark Fashehccd979b2005-12-15 14:31:24 -0800508 res->l_type = type;
509 res->l_ops = ops;
510 res->l_priv = priv;
511
Joel Beckerbd3e7612008-02-01 12:14:57 -0800512 res->l_level = DLM_LOCK_IV;
513 res->l_requested = DLM_LOCK_IV;
514 res->l_blocking = DLM_LOCK_IV;
Mark Fashehccd979b2005-12-15 14:31:24 -0800515 res->l_action = OCFS2_AST_INVALID;
516 res->l_unlock_action = OCFS2_UNLOCK_INVALID;
517
518 res->l_flags = OCFS2_LOCK_INITIALIZED;
519
520 ocfs2_add_lockres_tracking(res, osb->osb_dlm_debug);
Sunil Mushran8ddb7b02008-05-13 13:45:15 -0700521
522 ocfs2_init_lock_stats(res);
Jan Karacb257972009-06-04 15:26:50 +0200523#ifdef CONFIG_DEBUG_LOCK_ALLOC
524 if (type != OCFS2_LOCK_TYPE_OPEN)
525 lockdep_init_map(&res->l_lockdep_map, ocfs2_lock_type_strings[type],
526 &lockdep_keys[type], 0);
527 else
528 res->l_lockdep_map.key = NULL;
529#endif
Mark Fashehccd979b2005-12-15 14:31:24 -0800530}
531
532void ocfs2_lock_res_init_once(struct ocfs2_lock_res *res)
533{
534 /* This also clears out the lock status block */
535 memset(res, 0, sizeof(struct ocfs2_lock_res));
536 spin_lock_init(&res->l_lock);
537 init_waitqueue_head(&res->l_event);
538 INIT_LIST_HEAD(&res->l_blocked_list);
539 INIT_LIST_HEAD(&res->l_mask_waiters);
Eric Ren439a36b2017-02-22 15:40:41 -0800540 INIT_LIST_HEAD(&res->l_holders);
Mark Fashehccd979b2005-12-15 14:31:24 -0800541}
542
543void ocfs2_inode_lock_res_init(struct ocfs2_lock_res *res,
544 enum ocfs2_lock_type type,
Mark Fasheh24c19ef2006-09-22 17:28:19 -0700545 unsigned int generation,
Mark Fashehccd979b2005-12-15 14:31:24 -0800546 struct inode *inode)
547{
548 struct ocfs2_lock_res_ops *ops;
549
550 switch(type) {
551 case OCFS2_LOCK_TYPE_RW:
552 ops = &ocfs2_inode_rw_lops;
553 break;
554 case OCFS2_LOCK_TYPE_META:
Mark Fashehe63aecb62007-10-18 15:30:42 -0700555 ops = &ocfs2_inode_inode_lops;
Mark Fashehccd979b2005-12-15 14:31:24 -0800556 break;
Tiger Yang50008632007-03-20 16:01:38 -0700557 case OCFS2_LOCK_TYPE_OPEN:
558 ops = &ocfs2_inode_open_lops;
559 break;
Mark Fashehccd979b2005-12-15 14:31:24 -0800560 default:
561 mlog_bug_on_msg(1, "type: %d\n", type);
562 ops = NULL; /* thanks, gcc */
563 break;
564 };
565
Mark Fashehd680efe2006-09-08 14:14:34 -0700566 ocfs2_build_lock_name(type, OCFS2_I(inode)->ip_blkno,
Mark Fasheh24c19ef2006-09-22 17:28:19 -0700567 generation, res->l_name);
Mark Fashehd680efe2006-09-08 14:14:34 -0700568 ocfs2_lock_res_init_common(OCFS2_SB(inode->i_sb), res, type, ops, inode);
569}
570
Mark Fasheh54a7e752006-09-12 21:49:13 -0700571static struct ocfs2_super *ocfs2_get_inode_osb(struct ocfs2_lock_res *lockres)
572{
573 struct inode *inode = ocfs2_lock_res_inode(lockres);
574
575 return OCFS2_SB(inode->i_sb);
576}
577
Jan Kara9e33d692008-08-25 19:56:50 +0200578static struct ocfs2_super *ocfs2_get_qinfo_osb(struct ocfs2_lock_res *lockres)
579{
580 struct ocfs2_mem_dqinfo *info = lockres->l_priv;
581
582 return OCFS2_SB(info->dqi_gi.dqi_sb);
583}
584
Mark Fashehcf8e06f2007-12-20 16:43:10 -0800585static struct ocfs2_super *ocfs2_get_file_osb(struct ocfs2_lock_res *lockres)
586{
587 struct ocfs2_file_private *fp = lockres->l_priv;
588
589 return OCFS2_SB(fp->fp_file->f_mapping->host->i_sb);
590}
591
Mark Fashehd680efe2006-09-08 14:14:34 -0700592static __u64 ocfs2_get_dentry_lock_ino(struct ocfs2_lock_res *lockres)
593{
594 __be64 inode_blkno_be;
595
596 memcpy(&inode_blkno_be, &lockres->l_name[OCFS2_DENTRY_LOCK_INO_START],
597 sizeof(__be64));
598
599 return be64_to_cpu(inode_blkno_be);
600}
601
Mark Fasheh54a7e752006-09-12 21:49:13 -0700602static struct ocfs2_super *ocfs2_get_dentry_osb(struct ocfs2_lock_res *lockres)
603{
604 struct ocfs2_dentry_lock *dl = lockres->l_priv;
605
606 return OCFS2_SB(dl->dl_inode->i_sb);
607}
608
Mark Fashehd680efe2006-09-08 14:14:34 -0700609void ocfs2_dentry_lock_res_init(struct ocfs2_dentry_lock *dl,
610 u64 parent, struct inode *inode)
611{
612 int len;
613 u64 inode_blkno = OCFS2_I(inode)->ip_blkno;
614 __be64 inode_blkno_be = cpu_to_be64(inode_blkno);
615 struct ocfs2_lock_res *lockres = &dl->dl_lockres;
616
617 ocfs2_lock_res_init_once(lockres);
618
619 /*
620 * Unfortunately, the standard lock naming scheme won't work
621 * here because we have two 16 byte values to use. Instead,
622 * we'll stuff the inode number as a binary value. We still
623 * want error prints to show something without garbling the
624 * display, so drop a null byte in there before the inode
625 * number. A future version of OCFS2 will likely use all
626 * binary lock names. The stringified names have been a
627 * tremendous aid in debugging, but now that the debugfs
628 * interface exists, we can mangle things there if need be.
629 *
630 * NOTE: We also drop the standard "pad" value (the total lock
631 * name size stays the same though - the last part is all
632 * zeros due to the memset in ocfs2_lock_res_init_once()
633 */
634 len = snprintf(lockres->l_name, OCFS2_DENTRY_LOCK_INO_START,
635 "%c%016llx",
636 ocfs2_lock_type_char(OCFS2_LOCK_TYPE_DENTRY),
637 (long long)parent);
638
639 BUG_ON(len != (OCFS2_DENTRY_LOCK_INO_START - 1));
640
641 memcpy(&lockres->l_name[OCFS2_DENTRY_LOCK_INO_START], &inode_blkno_be,
642 sizeof(__be64));
643
644 ocfs2_lock_res_init_common(OCFS2_SB(inode->i_sb), lockres,
645 OCFS2_LOCK_TYPE_DENTRY, &ocfs2_dentry_lops,
646 dl);
Mark Fashehccd979b2005-12-15 14:31:24 -0800647}
648
649static void ocfs2_super_lock_res_init(struct ocfs2_lock_res *res,
650 struct ocfs2_super *osb)
651{
652 /* Superblock lockres doesn't come from a slab so we call init
653 * once on it manually. */
654 ocfs2_lock_res_init_once(res);
Mark Fashehd680efe2006-09-08 14:14:34 -0700655 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_SUPER, OCFS2_SUPER_BLOCK_BLKNO,
656 0, res->l_name);
Mark Fashehccd979b2005-12-15 14:31:24 -0800657 ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_SUPER,
Mark Fashehccd979b2005-12-15 14:31:24 -0800658 &ocfs2_super_lops, osb);
659}
660
661static void ocfs2_rename_lock_res_init(struct ocfs2_lock_res *res,
662 struct ocfs2_super *osb)
663{
664 /* Rename lockres doesn't come from a slab so we call init
665 * once on it manually. */
666 ocfs2_lock_res_init_once(res);
Mark Fashehd680efe2006-09-08 14:14:34 -0700667 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_RENAME, 0, 0, res->l_name);
668 ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_RENAME,
Mark Fashehccd979b2005-12-15 14:31:24 -0800669 &ocfs2_rename_lops, osb);
670}
671
wengang wang6ca497a2009-03-06 21:29:10 +0800672static void ocfs2_nfs_sync_lock_res_init(struct ocfs2_lock_res *res,
673 struct ocfs2_super *osb)
674{
675 /* nfs_sync lockres doesn't come from a slab so we call init
676 * once on it manually. */
677 ocfs2_lock_res_init_once(res);
678 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_NFS_SYNC, 0, 0, res->l_name);
679 ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_NFS_SYNC,
680 &ocfs2_nfs_sync_lops, osb);
681}
682
Gang He4882abe2018-01-31 16:15:10 -0800683void ocfs2_trim_fs_lock_res_init(struct ocfs2_super *osb)
684{
685 struct ocfs2_lock_res *lockres = &osb->osb_trim_fs_lockres;
686
687 ocfs2_lock_res_init_once(lockres);
688 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_TRIM_FS, 0, 0, lockres->l_name);
689 ocfs2_lock_res_init_common(osb, lockres, OCFS2_LOCK_TYPE_TRIM_FS,
690 &ocfs2_trim_fs_lops, osb);
691}
692
693void ocfs2_trim_fs_lock_res_uninit(struct ocfs2_super *osb)
694{
695 struct ocfs2_lock_res *lockres = &osb->osb_trim_fs_lockres;
696
697 ocfs2_simple_drop_lockres(osb, lockres);
698 ocfs2_lock_res_free(lockres);
699}
700
Srinivas Eeda83273932009-06-03 17:02:55 -0700701static void ocfs2_orphan_scan_lock_res_init(struct ocfs2_lock_res *res,
702 struct ocfs2_super *osb)
703{
Srinivas Eeda83273932009-06-03 17:02:55 -0700704 ocfs2_lock_res_init_once(res);
705 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_ORPHAN_SCAN, 0, 0, res->l_name);
706 ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_ORPHAN_SCAN,
707 &ocfs2_orphan_scan_lops, osb);
Srinivas Eeda83273932009-06-03 17:02:55 -0700708}
709
Mark Fashehcf8e06f2007-12-20 16:43:10 -0800710void ocfs2_file_lock_res_init(struct ocfs2_lock_res *lockres,
711 struct ocfs2_file_private *fp)
712{
713 struct inode *inode = fp->fp_file->f_mapping->host;
714 struct ocfs2_inode_info *oi = OCFS2_I(inode);
715
716 ocfs2_lock_res_init_once(lockres);
717 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_FLOCK, oi->ip_blkno,
718 inode->i_generation, lockres->l_name);
719 ocfs2_lock_res_init_common(OCFS2_SB(inode->i_sb), lockres,
720 OCFS2_LOCK_TYPE_FLOCK, &ocfs2_flock_lops,
721 fp);
722 lockres->l_flags |= OCFS2_LOCK_NOCACHE;
723}
724
Jan Kara9e33d692008-08-25 19:56:50 +0200725void ocfs2_qinfo_lock_res_init(struct ocfs2_lock_res *lockres,
726 struct ocfs2_mem_dqinfo *info)
727{
728 ocfs2_lock_res_init_once(lockres);
729 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_QINFO, info->dqi_gi.dqi_type,
730 0, lockres->l_name);
731 ocfs2_lock_res_init_common(OCFS2_SB(info->dqi_gi.dqi_sb), lockres,
732 OCFS2_LOCK_TYPE_QINFO, &ocfs2_qinfo_lops,
733 info);
734}
735
Tao Ma8dec98e2009-08-18 11:19:58 +0800736void ocfs2_refcount_lock_res_init(struct ocfs2_lock_res *lockres,
737 struct ocfs2_super *osb, u64 ref_blkno,
738 unsigned int generation)
739{
740 ocfs2_lock_res_init_once(lockres);
741 ocfs2_build_lock_name(OCFS2_LOCK_TYPE_REFCOUNT, ref_blkno,
742 generation, lockres->l_name);
743 ocfs2_lock_res_init_common(osb, lockres, OCFS2_LOCK_TYPE_REFCOUNT,
744 &ocfs2_refcount_block_lops, osb);
745}
746
Mark Fashehccd979b2005-12-15 14:31:24 -0800747void ocfs2_lock_res_free(struct ocfs2_lock_res *res)
748{
Mark Fashehccd979b2005-12-15 14:31:24 -0800749 if (!(res->l_flags & OCFS2_LOCK_INITIALIZED))
750 return;
751
752 ocfs2_remove_lockres_tracking(res);
753
754 mlog_bug_on_msg(!list_empty(&res->l_blocked_list),
755 "Lockres %s is on the blocked list\n",
756 res->l_name);
757 mlog_bug_on_msg(!list_empty(&res->l_mask_waiters),
758 "Lockres %s has mask waiters pending\n",
759 res->l_name);
760 mlog_bug_on_msg(spin_is_locked(&res->l_lock),
761 "Lockres %s is locked\n",
762 res->l_name);
763 mlog_bug_on_msg(res->l_ro_holders,
764 "Lockres %s has %u ro holders\n",
765 res->l_name, res->l_ro_holders);
766 mlog_bug_on_msg(res->l_ex_holders,
767 "Lockres %s has %u ex holders\n",
768 res->l_name, res->l_ex_holders);
769
770 /* Need to clear out the lock status block for the dlm */
771 memset(&res->l_lksb, 0, sizeof(res->l_lksb));
772
773 res->l_flags = 0UL;
Mark Fashehccd979b2005-12-15 14:31:24 -0800774}
775
Eric Ren439a36b2017-02-22 15:40:41 -0800776/*
777 * Keep a list of processes who have interest in a lockres.
778 * Note: this is now only uesed for check recursive cluster locking.
779 */
780static inline void ocfs2_add_holder(struct ocfs2_lock_res *lockres,
781 struct ocfs2_lock_holder *oh)
782{
783 INIT_LIST_HEAD(&oh->oh_list);
784 oh->oh_owner_pid = get_pid(task_pid(current));
785
786 spin_lock(&lockres->l_lock);
787 list_add_tail(&oh->oh_list, &lockres->l_holders);
788 spin_unlock(&lockres->l_lock);
789}
790
791static inline void ocfs2_remove_holder(struct ocfs2_lock_res *lockres,
792 struct ocfs2_lock_holder *oh)
793{
794 spin_lock(&lockres->l_lock);
795 list_del(&oh->oh_list);
796 spin_unlock(&lockres->l_lock);
797
798 put_pid(oh->oh_owner_pid);
799}
800
801static inline int ocfs2_is_locked_by_me(struct ocfs2_lock_res *lockres)
802{
803 struct ocfs2_lock_holder *oh;
804 struct pid *pid;
805
806 /* look in the list of holders for one with the current task as owner */
807 spin_lock(&lockres->l_lock);
808 pid = task_pid(current);
809 list_for_each_entry(oh, &lockres->l_holders, oh_list) {
810 if (oh->oh_owner_pid == pid) {
811 spin_unlock(&lockres->l_lock);
812 return 1;
813 }
814 }
815 spin_unlock(&lockres->l_lock);
816
817 return 0;
818}
819
Mark Fashehccd979b2005-12-15 14:31:24 -0800820static inline void ocfs2_inc_holders(struct ocfs2_lock_res *lockres,
821 int level)
822{
Mark Fashehccd979b2005-12-15 14:31:24 -0800823 BUG_ON(!lockres);
824
825 switch(level) {
Joel Beckerbd3e7612008-02-01 12:14:57 -0800826 case DLM_LOCK_EX:
Mark Fashehccd979b2005-12-15 14:31:24 -0800827 lockres->l_ex_holders++;
828 break;
Joel Beckerbd3e7612008-02-01 12:14:57 -0800829 case DLM_LOCK_PR:
Mark Fashehccd979b2005-12-15 14:31:24 -0800830 lockres->l_ro_holders++;
831 break;
832 default:
833 BUG();
834 }
Mark Fashehccd979b2005-12-15 14:31:24 -0800835}
836
837static inline void ocfs2_dec_holders(struct ocfs2_lock_res *lockres,
838 int level)
839{
Mark Fashehccd979b2005-12-15 14:31:24 -0800840 BUG_ON(!lockres);
841
842 switch(level) {
Joel Beckerbd3e7612008-02-01 12:14:57 -0800843 case DLM_LOCK_EX:
Mark Fashehccd979b2005-12-15 14:31:24 -0800844 BUG_ON(!lockres->l_ex_holders);
845 lockres->l_ex_holders--;
846 break;
Joel Beckerbd3e7612008-02-01 12:14:57 -0800847 case DLM_LOCK_PR:
Mark Fashehccd979b2005-12-15 14:31:24 -0800848 BUG_ON(!lockres->l_ro_holders);
849 lockres->l_ro_holders--;
850 break;
851 default:
852 BUG();
853 }
Mark Fashehccd979b2005-12-15 14:31:24 -0800854}
855
856/* WARNING: This function lives in a world where the only three lock
857 * levels are EX, PR, and NL. It *will* have to be adjusted when more
858 * lock types are added. */
859static inline int ocfs2_highest_compat_lock_level(int level)
860{
Joel Beckerbd3e7612008-02-01 12:14:57 -0800861 int new_level = DLM_LOCK_EX;
Mark Fashehccd979b2005-12-15 14:31:24 -0800862
Joel Beckerbd3e7612008-02-01 12:14:57 -0800863 if (level == DLM_LOCK_EX)
864 new_level = DLM_LOCK_NL;
865 else if (level == DLM_LOCK_PR)
866 new_level = DLM_LOCK_PR;
Mark Fashehccd979b2005-12-15 14:31:24 -0800867 return new_level;
868}
869
870static void lockres_set_flags(struct ocfs2_lock_res *lockres,
871 unsigned long newflags)
872{
Christoph Hellwig800deef2007-05-17 16:03:13 +0200873 struct ocfs2_mask_waiter *mw, *tmp;
Mark Fashehccd979b2005-12-15 14:31:24 -0800874
875 assert_spin_locked(&lockres->l_lock);
876
877 lockres->l_flags = newflags;
878
Christoph Hellwig800deef2007-05-17 16:03:13 +0200879 list_for_each_entry_safe(mw, tmp, &lockres->l_mask_waiters, mw_item) {
Mark Fashehccd979b2005-12-15 14:31:24 -0800880 if ((lockres->l_flags & mw->mw_mask) != mw->mw_goal)
881 continue;
882
883 list_del_init(&mw->mw_item);
884 mw->mw_status = 0;
885 complete(&mw->mw_complete);
886 }
887}
888static void lockres_or_flags(struct ocfs2_lock_res *lockres, unsigned long or)
889{
890 lockres_set_flags(lockres, lockres->l_flags | or);
891}
892static void lockres_clear_flags(struct ocfs2_lock_res *lockres,
893 unsigned long clear)
894{
895 lockres_set_flags(lockres, lockres->l_flags & ~clear);
896}
897
898static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res *lockres)
899{
Mark Fashehccd979b2005-12-15 14:31:24 -0800900 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY));
901 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_ATTACHED));
902 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BLOCKED));
Joel Beckerbd3e7612008-02-01 12:14:57 -0800903 BUG_ON(lockres->l_blocking <= DLM_LOCK_NL);
Mark Fashehccd979b2005-12-15 14:31:24 -0800904
905 lockres->l_level = lockres->l_requested;
906 if (lockres->l_level <=
907 ocfs2_highest_compat_lock_level(lockres->l_blocking)) {
Joel Beckerbd3e7612008-02-01 12:14:57 -0800908 lockres->l_blocking = DLM_LOCK_NL;
Mark Fashehccd979b2005-12-15 14:31:24 -0800909 lockres_clear_flags(lockres, OCFS2_LOCK_BLOCKED);
910 }
911 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
Mark Fashehccd979b2005-12-15 14:31:24 -0800912}
913
914static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res *lockres)
915{
Mark Fashehccd979b2005-12-15 14:31:24 -0800916 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY));
917 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_ATTACHED));
918
919 /* Convert from RO to EX doesn't really need anything as our
920 * information is already up to data. Convert from NL to
921 * *anything* however should mark ourselves as needing an
922 * update */
Joel Beckerbd3e7612008-02-01 12:14:57 -0800923 if (lockres->l_level == DLM_LOCK_NL &&
Mark Fashehf625c972006-09-12 21:24:53 -0700924 lockres->l_ops->flags & LOCK_TYPE_REQUIRES_REFRESH)
Mark Fashehccd979b2005-12-15 14:31:24 -0800925 lockres_or_flags(lockres, OCFS2_LOCK_NEEDS_REFRESH);
926
927 lockres->l_level = lockres->l_requested;
Sunil Mushrana1912822010-01-21 10:50:03 -0800928
929 /*
930 * We set the OCFS2_LOCK_UPCONVERT_FINISHING flag before clearing
931 * the OCFS2_LOCK_BUSY flag to prevent the dc thread from
932 * downconverting the lock before the upconvert has fully completed.
Xue jiufeid1e782382014-12-10 15:41:59 -0800933 * Do not prevent the dc thread from downconverting if NONBLOCK lock
934 * had already returned.
Sunil Mushrana1912822010-01-21 10:50:03 -0800935 */
Xue jiufeid1e782382014-12-10 15:41:59 -0800936 if (!(lockres->l_flags & OCFS2_LOCK_NONBLOCK_FINISHED))
937 lockres_or_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
938 else
939 lockres_clear_flags(lockres, OCFS2_LOCK_NONBLOCK_FINISHED);
Sunil Mushrana1912822010-01-21 10:50:03 -0800940
Mark Fashehccd979b2005-12-15 14:31:24 -0800941 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
Mark Fashehccd979b2005-12-15 14:31:24 -0800942}
943
944static inline void ocfs2_generic_handle_attach_action(struct ocfs2_lock_res *lockres)
945{
Roel Kluin3cf0c502007-10-27 00:20:36 +0200946 BUG_ON((!(lockres->l_flags & OCFS2_LOCK_BUSY)));
Mark Fashehccd979b2005-12-15 14:31:24 -0800947 BUG_ON(lockres->l_flags & OCFS2_LOCK_ATTACHED);
948
Joel Beckerbd3e7612008-02-01 12:14:57 -0800949 if (lockres->l_requested > DLM_LOCK_NL &&
Mark Fashehf625c972006-09-12 21:24:53 -0700950 !(lockres->l_flags & OCFS2_LOCK_LOCAL) &&
951 lockres->l_ops->flags & LOCK_TYPE_REQUIRES_REFRESH)
Mark Fashehccd979b2005-12-15 14:31:24 -0800952 lockres_or_flags(lockres, OCFS2_LOCK_NEEDS_REFRESH);
953
954 lockres->l_level = lockres->l_requested;
955 lockres_or_flags(lockres, OCFS2_LOCK_ATTACHED);
956 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
Mark Fashehccd979b2005-12-15 14:31:24 -0800957}
958
Mark Fashehccd979b2005-12-15 14:31:24 -0800959static int ocfs2_generic_handle_bast(struct ocfs2_lock_res *lockres,
960 int level)
961{
962 int needs_downconvert = 0;
Mark Fashehccd979b2005-12-15 14:31:24 -0800963
964 assert_spin_locked(&lockres->l_lock);
965
Mark Fashehccd979b2005-12-15 14:31:24 -0800966 if (level > lockres->l_blocking) {
967 /* only schedule a downconvert if we haven't already scheduled
968 * one that goes low enough to satisfy the level we're
969 * blocking. this also catches the case where we get
970 * duplicate BASTs */
971 if (ocfs2_highest_compat_lock_level(level) <
972 ocfs2_highest_compat_lock_level(lockres->l_blocking))
973 needs_downconvert = 1;
974
975 lockres->l_blocking = level;
976 }
977
Sunil Mushran9b915182010-02-26 19:42:44 -0800978 mlog(ML_BASTS, "lockres %s, block %d, level %d, l_block %d, dwn %d\n",
979 lockres->l_name, level, lockres->l_level, lockres->l_blocking,
980 needs_downconvert);
981
Wengang Wang0b94a902010-01-21 10:50:02 -0800982 if (needs_downconvert)
983 lockres_or_flags(lockres, OCFS2_LOCK_BLOCKED);
Tao Mac1e8d352011-03-07 16:43:21 +0800984 mlog(0, "needs_downconvert = %d\n", needs_downconvert);
Mark Fashehccd979b2005-12-15 14:31:24 -0800985 return needs_downconvert;
986}
987
Joel Beckerde551242008-02-01 14:45:08 -0800988/*
989 * OCFS2_LOCK_PENDING and l_pending_gen.
990 *
991 * Why does OCFS2_LOCK_PENDING exist? To close a race between setting
992 * OCFS2_LOCK_BUSY and calling ocfs2_dlm_lock(). See ocfs2_unblock_lock()
993 * for more details on the race.
994 *
995 * OCFS2_LOCK_PENDING closes the race quite nicely. However, it introduces
996 * a race on itself. In o2dlm, we can get the ast before ocfs2_dlm_lock()
997 * returns. The ast clears OCFS2_LOCK_BUSY, and must therefore clear
998 * OCFS2_LOCK_PENDING at the same time. When ocfs2_dlm_lock() returns,
999 * the caller is going to try to clear PENDING again. If nothing else is
1000 * happening, __lockres_clear_pending() sees PENDING is unset and does
1001 * nothing.
1002 *
1003 * But what if another path (eg downconvert thread) has just started a
1004 * new locking action? The other path has re-set PENDING. Our path
1005 * cannot clear PENDING, because that will re-open the original race
1006 * window.
1007 *
1008 * [Example]
1009 *
1010 * ocfs2_meta_lock()
1011 * ocfs2_cluster_lock()
1012 * set BUSY
1013 * set PENDING
1014 * drop l_lock
1015 * ocfs2_dlm_lock()
1016 * ocfs2_locking_ast() ocfs2_downconvert_thread()
1017 * clear PENDING ocfs2_unblock_lock()
1018 * take_l_lock
1019 * !BUSY
1020 * ocfs2_prepare_downconvert()
1021 * set BUSY
1022 * set PENDING
1023 * drop l_lock
1024 * take l_lock
1025 * clear PENDING
1026 * drop l_lock
1027 * <window>
1028 * ocfs2_dlm_lock()
1029 *
1030 * So as you can see, we now have a window where l_lock is not held,
1031 * PENDING is not set, and ocfs2_dlm_lock() has not been called.
1032 *
1033 * The core problem is that ocfs2_cluster_lock() has cleared the PENDING
1034 * set by ocfs2_prepare_downconvert(). That wasn't nice.
1035 *
1036 * To solve this we introduce l_pending_gen. A call to
1037 * lockres_clear_pending() will only do so when it is passed a generation
1038 * number that matches the lockres. lockres_set_pending() will return the
1039 * current generation number. When ocfs2_cluster_lock() goes to clear
1040 * PENDING, it passes the generation it got from set_pending(). In our
1041 * example above, the generation numbers will *not* match. Thus,
1042 * ocfs2_cluster_lock() will not clear the PENDING set by
1043 * ocfs2_prepare_downconvert().
1044 */
1045
1046/* Unlocked version for ocfs2_locking_ast() */
1047static void __lockres_clear_pending(struct ocfs2_lock_res *lockres,
1048 unsigned int generation,
1049 struct ocfs2_super *osb)
1050{
1051 assert_spin_locked(&lockres->l_lock);
1052
1053 /*
1054 * The ast and locking functions can race us here. The winner
1055 * will clear pending, the loser will not.
1056 */
1057 if (!(lockres->l_flags & OCFS2_LOCK_PENDING) ||
1058 (lockres->l_pending_gen != generation))
1059 return;
1060
1061 lockres_clear_flags(lockres, OCFS2_LOCK_PENDING);
1062 lockres->l_pending_gen++;
1063
1064 /*
1065 * The downconvert thread may have skipped us because we
1066 * were PENDING. Wake it up.
1067 */
1068 if (lockres->l_flags & OCFS2_LOCK_BLOCKED)
1069 ocfs2_wake_downconvert_thread(osb);
1070}
1071
1072/* Locked version for callers of ocfs2_dlm_lock() */
1073static void lockres_clear_pending(struct ocfs2_lock_res *lockres,
1074 unsigned int generation,
1075 struct ocfs2_super *osb)
1076{
1077 unsigned long flags;
1078
1079 spin_lock_irqsave(&lockres->l_lock, flags);
1080 __lockres_clear_pending(lockres, generation, osb);
1081 spin_unlock_irqrestore(&lockres->l_lock, flags);
1082}
1083
1084static unsigned int lockres_set_pending(struct ocfs2_lock_res *lockres)
1085{
1086 assert_spin_locked(&lockres->l_lock);
1087 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY));
1088
1089 lockres_or_flags(lockres, OCFS2_LOCK_PENDING);
1090
1091 return lockres->l_pending_gen;
1092}
1093
Joel Beckerc0e413382010-01-29 14:46:44 -08001094static void ocfs2_blocking_ast(struct ocfs2_dlm_lksb *lksb, int level)
Mark Fashehccd979b2005-12-15 14:31:24 -08001095{
Joel Beckera796d282010-01-28 19:22:39 -08001096 struct ocfs2_lock_res *lockres = ocfs2_lksb_to_lock_res(lksb);
Mark Fashehaa2623a2006-09-12 21:58:23 -07001097 struct ocfs2_super *osb = ocfs2_get_lockres_osb(lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08001098 int needs_downconvert;
1099 unsigned long flags;
1100
Joel Beckerbd3e7612008-02-01 12:14:57 -08001101 BUG_ON(level <= DLM_LOCK_NL);
Mark Fashehccd979b2005-12-15 14:31:24 -08001102
Sunil Mushran9b915182010-02-26 19:42:44 -08001103 mlog(ML_BASTS, "BAST fired for lockres %s, blocking %d, level %d, "
1104 "type %s\n", lockres->l_name, level, lockres->l_level,
Mark Fashehaa2623a2006-09-12 21:58:23 -07001105 ocfs2_lock_type_string(lockres->l_type));
1106
Mark Fashehcf8e06f2007-12-20 16:43:10 -08001107 /*
1108 * We can skip the bast for locks which don't enable caching -
1109 * they'll be dropped at the earliest possible time anyway.
1110 */
1111 if (lockres->l_flags & OCFS2_LOCK_NOCACHE)
1112 return;
1113
Mark Fashehccd979b2005-12-15 14:31:24 -08001114 spin_lock_irqsave(&lockres->l_lock, flags);
1115 needs_downconvert = ocfs2_generic_handle_bast(lockres, level);
1116 if (needs_downconvert)
1117 ocfs2_schedule_blocked_lock(osb, lockres);
1118 spin_unlock_irqrestore(&lockres->l_lock, flags);
1119
Mark Fashehd680efe2006-09-08 14:14:34 -07001120 wake_up(&lockres->l_event);
1121
Mark Fasheh34d024f2007-09-24 15:56:19 -07001122 ocfs2_wake_downconvert_thread(osb);
Mark Fashehccd979b2005-12-15 14:31:24 -08001123}
1124
Joel Beckerc0e413382010-01-29 14:46:44 -08001125static void ocfs2_locking_ast(struct ocfs2_dlm_lksb *lksb)
Mark Fashehccd979b2005-12-15 14:31:24 -08001126{
Joel Beckera796d282010-01-28 19:22:39 -08001127 struct ocfs2_lock_res *lockres = ocfs2_lksb_to_lock_res(lksb);
Joel Beckerde551242008-02-01 14:45:08 -08001128 struct ocfs2_super *osb = ocfs2_get_lockres_osb(lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08001129 unsigned long flags;
David Teigland1693a5c2008-01-30 16:52:53 -08001130 int status;
Mark Fashehccd979b2005-12-15 14:31:24 -08001131
1132 spin_lock_irqsave(&lockres->l_lock, flags);
1133
David Teigland1693a5c2008-01-30 16:52:53 -08001134 status = ocfs2_dlm_lock_status(&lockres->l_lksb);
1135
1136 if (status == -EAGAIN) {
1137 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
1138 goto out;
1139 }
1140
1141 if (status) {
Joel Becker8f2c9c12008-02-01 12:16:57 -08001142 mlog(ML_ERROR, "lockres %s: lksb status value of %d!\n",
David Teigland1693a5c2008-01-30 16:52:53 -08001143 lockres->l_name, status);
Mark Fashehccd979b2005-12-15 14:31:24 -08001144 spin_unlock_irqrestore(&lockres->l_lock, flags);
1145 return;
1146 }
1147
Sunil Mushran9b915182010-02-26 19:42:44 -08001148 mlog(ML_BASTS, "AST fired for lockres %s, action %d, unlock %d, "
1149 "level %d => %d\n", lockres->l_name, lockres->l_action,
1150 lockres->l_unlock_action, lockres->l_level, lockres->l_requested);
1151
Mark Fashehccd979b2005-12-15 14:31:24 -08001152 switch(lockres->l_action) {
1153 case OCFS2_AST_ATTACH:
1154 ocfs2_generic_handle_attach_action(lockres);
Mark Fashehe92d57d2006-09-12 21:34:35 -07001155 lockres_clear_flags(lockres, OCFS2_LOCK_LOCAL);
Mark Fashehccd979b2005-12-15 14:31:24 -08001156 break;
1157 case OCFS2_AST_CONVERT:
1158 ocfs2_generic_handle_convert_action(lockres);
1159 break;
1160 case OCFS2_AST_DOWNCONVERT:
1161 ocfs2_generic_handle_downconvert_action(lockres);
1162 break;
1163 default:
Sunil Mushran9b915182010-02-26 19:42:44 -08001164 mlog(ML_ERROR, "lockres %s: AST fired with invalid action: %u, "
1165 "flags 0x%lx, unlock: %u\n",
Mark Fashehe92d57d2006-09-12 21:34:35 -07001166 lockres->l_name, lockres->l_action, lockres->l_flags,
1167 lockres->l_unlock_action);
Mark Fashehccd979b2005-12-15 14:31:24 -08001168 BUG();
1169 }
David Teigland1693a5c2008-01-30 16:52:53 -08001170out:
Mark Fashehccd979b2005-12-15 14:31:24 -08001171 /* set it to something invalid so if we get called again we
1172 * can catch it. */
1173 lockres->l_action = OCFS2_AST_INVALID;
Mark Fashehccd979b2005-12-15 14:31:24 -08001174
Joel Beckerde551242008-02-01 14:45:08 -08001175 /* Did we try to cancel this lock? Clear that state */
1176 if (lockres->l_unlock_action == OCFS2_UNLOCK_CANCEL_CONVERT)
1177 lockres->l_unlock_action = OCFS2_UNLOCK_INVALID;
1178
1179 /*
1180 * We may have beaten the locking functions here. We certainly
1181 * know that dlm_lock() has been called :-)
1182 * Because we can't have two lock calls in flight at once, we
1183 * can use lockres->l_pending_gen.
1184 */
1185 __lockres_clear_pending(lockres, lockres->l_pending_gen, osb);
1186
Mark Fashehccd979b2005-12-15 14:31:24 -08001187 wake_up(&lockres->l_event);
Mark Fashehd680efe2006-09-08 14:14:34 -07001188 spin_unlock_irqrestore(&lockres->l_lock, flags);
Mark Fashehccd979b2005-12-15 14:31:24 -08001189}
1190
Joel Becker553b5eb2010-01-29 17:19:06 -08001191static void ocfs2_unlock_ast(struct ocfs2_dlm_lksb *lksb, int error)
1192{
1193 struct ocfs2_lock_res *lockres = ocfs2_lksb_to_lock_res(lksb);
1194 unsigned long flags;
1195
Sunil Mushran9b915182010-02-26 19:42:44 -08001196 mlog(ML_BASTS, "UNLOCK AST fired for lockres %s, action = %d\n",
1197 lockres->l_name, lockres->l_unlock_action);
Joel Becker553b5eb2010-01-29 17:19:06 -08001198
1199 spin_lock_irqsave(&lockres->l_lock, flags);
1200 if (error) {
1201 mlog(ML_ERROR, "Dlm passes error %d for lock %s, "
1202 "unlock_action %d\n", error, lockres->l_name,
1203 lockres->l_unlock_action);
1204 spin_unlock_irqrestore(&lockres->l_lock, flags);
Joel Becker553b5eb2010-01-29 17:19:06 -08001205 return;
1206 }
1207
1208 switch(lockres->l_unlock_action) {
1209 case OCFS2_UNLOCK_CANCEL_CONVERT:
1210 mlog(0, "Cancel convert success for %s\n", lockres->l_name);
1211 lockres->l_action = OCFS2_AST_INVALID;
1212 /* Downconvert thread may have requeued this lock, we
1213 * need to wake it. */
1214 if (lockres->l_flags & OCFS2_LOCK_BLOCKED)
1215 ocfs2_wake_downconvert_thread(ocfs2_get_lockres_osb(lockres));
1216 break;
1217 case OCFS2_UNLOCK_DROP_LOCK:
1218 lockres->l_level = DLM_LOCK_IV;
1219 break;
1220 default:
1221 BUG();
1222 }
1223
1224 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
1225 lockres->l_unlock_action = OCFS2_UNLOCK_INVALID;
1226 wake_up(&lockres->l_event);
1227 spin_unlock_irqrestore(&lockres->l_lock, flags);
Joel Becker553b5eb2010-01-29 17:19:06 -08001228}
1229
1230/*
1231 * This is the filesystem locking protocol. It provides the lock handling
1232 * hooks for the underlying DLM. It has a maximum version number.
1233 * The version number allows interoperability with systems running at
1234 * the same major number and an equal or smaller minor number.
1235 *
1236 * Whenever the filesystem does new things with locks (adds or removes a
1237 * lock, orders them differently, does different things underneath a lock),
1238 * the version must be changed. The protocol is negotiated when joining
1239 * the dlm domain. A node may join the domain if its major version is
1240 * identical to all other nodes and its minor version is greater than
1241 * or equal to all other nodes. When its minor version is greater than
1242 * the other nodes, it will run at the minor version specified by the
1243 * other nodes.
1244 *
1245 * If a locking change is made that will not be compatible with older
1246 * versions, the major number must be increased and the minor version set
1247 * to zero. If a change merely adds a behavior that can be disabled when
1248 * speaking to older versions, the minor version must be increased. If a
1249 * change adds a fully backwards compatible change (eg, LVB changes that
1250 * are just ignored by older versions), the version does not need to be
1251 * updated.
1252 */
1253static struct ocfs2_locking_protocol lproto = {
1254 .lp_max_version = {
1255 .pv_major = OCFS2_LOCKING_PROTOCOL_MAJOR,
1256 .pv_minor = OCFS2_LOCKING_PROTOCOL_MINOR,
1257 },
1258 .lp_lock_ast = ocfs2_locking_ast,
1259 .lp_blocking_ast = ocfs2_blocking_ast,
1260 .lp_unlock_ast = ocfs2_unlock_ast,
1261};
1262
1263void ocfs2_set_locking_protocol(void)
1264{
1265 ocfs2_stack_glue_set_max_proto_version(&lproto.lp_max_version);
1266}
1267
Mark Fashehccd979b2005-12-15 14:31:24 -08001268static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres,
1269 int convert)
1270{
1271 unsigned long flags;
1272
Mark Fashehccd979b2005-12-15 14:31:24 -08001273 spin_lock_irqsave(&lockres->l_lock, flags);
1274 lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
Sunil Mushrana1912822010-01-21 10:50:03 -08001275 lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
Mark Fashehccd979b2005-12-15 14:31:24 -08001276 if (convert)
1277 lockres->l_action = OCFS2_AST_INVALID;
1278 else
1279 lockres->l_unlock_action = OCFS2_UNLOCK_INVALID;
1280 spin_unlock_irqrestore(&lockres->l_lock, flags);
1281
1282 wake_up(&lockres->l_event);
Mark Fashehccd979b2005-12-15 14:31:24 -08001283}
1284
1285/* Note: If we detect another process working on the lock (i.e.,
1286 * OCFS2_LOCK_BUSY), we'll bail out returning 0. It's up to the caller
1287 * to do the right thing in that case.
1288 */
1289static int ocfs2_lock_create(struct ocfs2_super *osb,
1290 struct ocfs2_lock_res *lockres,
1291 int level,
Joel Beckerbd3e7612008-02-01 12:14:57 -08001292 u32 dlm_flags)
Mark Fashehccd979b2005-12-15 14:31:24 -08001293{
1294 int ret = 0;
Mark Fashehccd979b2005-12-15 14:31:24 -08001295 unsigned long flags;
Joel Beckerde551242008-02-01 14:45:08 -08001296 unsigned int gen;
Mark Fashehccd979b2005-12-15 14:31:24 -08001297
Joel Beckerbd3e7612008-02-01 12:14:57 -08001298 mlog(0, "lock %s, level = %d, flags = %u\n", lockres->l_name, level,
Mark Fashehccd979b2005-12-15 14:31:24 -08001299 dlm_flags);
1300
1301 spin_lock_irqsave(&lockres->l_lock, flags);
1302 if ((lockres->l_flags & OCFS2_LOCK_ATTACHED) ||
1303 (lockres->l_flags & OCFS2_LOCK_BUSY)) {
1304 spin_unlock_irqrestore(&lockres->l_lock, flags);
1305 goto bail;
1306 }
1307
1308 lockres->l_action = OCFS2_AST_ATTACH;
1309 lockres->l_requested = level;
1310 lockres_or_flags(lockres, OCFS2_LOCK_BUSY);
Joel Beckerde551242008-02-01 14:45:08 -08001311 gen = lockres_set_pending(lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08001312 spin_unlock_irqrestore(&lockres->l_lock, flags);
1313
Joel Becker4670c462008-02-01 14:39:35 -08001314 ret = ocfs2_dlm_lock(osb->cconn,
Joel Becker7431cd72008-02-01 12:15:37 -08001315 level,
1316 &lockres->l_lksb,
1317 dlm_flags,
1318 lockres->l_name,
Joel Beckera796d282010-01-28 19:22:39 -08001319 OCFS2_LOCK_ID_MAX_LEN - 1);
Joel Beckerde551242008-02-01 14:45:08 -08001320 lockres_clear_pending(lockres, gen, osb);
Joel Becker7431cd72008-02-01 12:15:37 -08001321 if (ret) {
1322 ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08001323 ocfs2_recover_from_dlm_error(lockres, 1);
1324 }
1325
Joel Becker7431cd72008-02-01 12:15:37 -08001326 mlog(0, "lock %s, return from ocfs2_dlm_lock\n", lockres->l_name);
Mark Fashehccd979b2005-12-15 14:31:24 -08001327
1328bail:
Mark Fashehccd979b2005-12-15 14:31:24 -08001329 return ret;
1330}
1331
1332static inline int ocfs2_check_wait_flag(struct ocfs2_lock_res *lockres,
1333 int flag)
1334{
1335 unsigned long flags;
1336 int ret;
1337
1338 spin_lock_irqsave(&lockres->l_lock, flags);
1339 ret = lockres->l_flags & flag;
1340 spin_unlock_irqrestore(&lockres->l_lock, flags);
1341
1342 return ret;
1343}
1344
1345static inline void ocfs2_wait_on_busy_lock(struct ocfs2_lock_res *lockres)
1346
1347{
1348 wait_event(lockres->l_event,
1349 !ocfs2_check_wait_flag(lockres, OCFS2_LOCK_BUSY));
1350}
1351
1352static inline void ocfs2_wait_on_refreshing_lock(struct ocfs2_lock_res *lockres)
1353
1354{
1355 wait_event(lockres->l_event,
1356 !ocfs2_check_wait_flag(lockres, OCFS2_LOCK_REFRESHING));
1357}
1358
1359/* predict what lock level we'll be dropping down to on behalf
1360 * of another node, and return true if the currently wanted
1361 * level will be compatible with it. */
1362static inline int ocfs2_may_continue_on_blocked_lock(struct ocfs2_lock_res *lockres,
1363 int wanted)
1364{
1365 BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BLOCKED));
1366
1367 return wanted <= ocfs2_highest_compat_lock_level(lockres->l_blocking);
1368}
1369
1370static void ocfs2_init_mask_waiter(struct ocfs2_mask_waiter *mw)
1371{
1372 INIT_LIST_HEAD(&mw->mw_item);
1373 init_completion(&mw->mw_complete);
Sunil Mushran8ddb7b02008-05-13 13:45:15 -07001374 ocfs2_init_start_time(mw);
Mark Fashehccd979b2005-12-15 14:31:24 -08001375}
1376
1377static int ocfs2_wait_for_mask(struct ocfs2_mask_waiter *mw)
1378{
1379 wait_for_completion(&mw->mw_complete);
1380 /* Re-arm the completion in case we want to wait on it again */
Wolfram Sang16735d02013-11-14 14:32:02 -08001381 reinit_completion(&mw->mw_complete);
Mark Fashehccd979b2005-12-15 14:31:24 -08001382 return mw->mw_status;
1383}
1384
1385static void lockres_add_mask_waiter(struct ocfs2_lock_res *lockres,
1386 struct ocfs2_mask_waiter *mw,
1387 unsigned long mask,
1388 unsigned long goal)
1389{
1390 BUG_ON(!list_empty(&mw->mw_item));
1391
1392 assert_spin_locked(&lockres->l_lock);
1393
1394 list_add_tail(&mw->mw_item, &lockres->l_mask_waiters);
1395 mw->mw_mask = mask;
1396 mw->mw_goal = goal;
1397}
1398
1399/* returns 0 if the mw that was removed was already satisfied, -EBUSY
1400 * if the mask still hadn't reached its goal */
Xue jiufeid1e782382014-12-10 15:41:59 -08001401static int __lockres_remove_mask_waiter(struct ocfs2_lock_res *lockres,
Mark Fashehccd979b2005-12-15 14:31:24 -08001402 struct ocfs2_mask_waiter *mw)
1403{
Mark Fashehccd979b2005-12-15 14:31:24 -08001404 int ret = 0;
1405
Xue jiufeid1e782382014-12-10 15:41:59 -08001406 assert_spin_locked(&lockres->l_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -08001407 if (!list_empty(&mw->mw_item)) {
1408 if ((lockres->l_flags & mw->mw_mask) != mw->mw_goal)
1409 ret = -EBUSY;
1410
1411 list_del_init(&mw->mw_item);
1412 init_completion(&mw->mw_complete);
1413 }
Xue jiufeid1e782382014-12-10 15:41:59 -08001414
1415 return ret;
1416}
1417
1418static int lockres_remove_mask_waiter(struct ocfs2_lock_res *lockres,
1419 struct ocfs2_mask_waiter *mw)
1420{
1421 unsigned long flags;
1422 int ret = 0;
1423
1424 spin_lock_irqsave(&lockres->l_lock, flags);
1425 ret = __lockres_remove_mask_waiter(lockres, mw);
Mark Fashehccd979b2005-12-15 14:31:24 -08001426 spin_unlock_irqrestore(&lockres->l_lock, flags);
1427
1428 return ret;
1429
1430}
1431
Mark Fashehcf8e06f2007-12-20 16:43:10 -08001432static int ocfs2_wait_for_mask_interruptible(struct ocfs2_mask_waiter *mw,
1433 struct ocfs2_lock_res *lockres)
1434{
1435 int ret;
1436
1437 ret = wait_for_completion_interruptible(&mw->mw_complete);
1438 if (ret)
1439 lockres_remove_mask_waiter(lockres, mw);
1440 else
1441 ret = mw->mw_status;
1442 /* Re-arm the completion in case we want to wait on it again */
Wolfram Sang16735d02013-11-14 14:32:02 -08001443 reinit_completion(&mw->mw_complete);
Mark Fashehcf8e06f2007-12-20 16:43:10 -08001444 return ret;
1445}
1446
Jan Karacb257972009-06-04 15:26:50 +02001447static int __ocfs2_cluster_lock(struct ocfs2_super *osb,
1448 struct ocfs2_lock_res *lockres,
1449 int level,
1450 u32 lkm_flags,
1451 int arg_flags,
1452 int l_subclass,
1453 unsigned long caller_ip)
Mark Fashehccd979b2005-12-15 14:31:24 -08001454{
1455 struct ocfs2_mask_waiter mw;
Mark Fashehccd979b2005-12-15 14:31:24 -08001456 int wait, catch_signals = !(osb->s_mount_opt & OCFS2_MOUNT_NOINTR);
1457 int ret = 0; /* gcc doesn't realize wait = 1 guarantees ret is set */
1458 unsigned long flags;
Joel Beckerde551242008-02-01 14:45:08 -08001459 unsigned int gen;
David Teigland1693a5c2008-01-30 16:52:53 -08001460 int noqueue_attempted = 0;
Xue jiufeid1e782382014-12-10 15:41:59 -08001461 int dlm_locked = 0;
Tariq Saeedb1b1e152016-01-21 16:40:39 -08001462 int kick_dc = 0;
Mark Fashehccd979b2005-12-15 14:31:24 -08001463
alex chen2f2eca22015-04-14 15:43:49 -07001464 if (!(lockres->l_flags & OCFS2_LOCK_INITIALIZED)) {
1465 mlog_errno(-EINVAL);
1466 return -EINVAL;
1467 }
1468
Mark Fashehccd979b2005-12-15 14:31:24 -08001469 ocfs2_init_mask_waiter(&mw);
1470
Mark Fashehb80fc012006-09-12 22:08:14 -07001471 if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB)
Joel Beckerbd3e7612008-02-01 12:14:57 -08001472 lkm_flags |= DLM_LKF_VALBLK;
Mark Fashehb80fc012006-09-12 22:08:14 -07001473
Mark Fashehccd979b2005-12-15 14:31:24 -08001474again:
1475 wait = 0;
1476
Sunil Mushrana1912822010-01-21 10:50:03 -08001477 spin_lock_irqsave(&lockres->l_lock, flags);
1478
Mark Fashehccd979b2005-12-15 14:31:24 -08001479 if (catch_signals && signal_pending(current)) {
1480 ret = -ERESTARTSYS;
Sunil Mushrana1912822010-01-21 10:50:03 -08001481 goto unlock;
Mark Fashehccd979b2005-12-15 14:31:24 -08001482 }
1483
Mark Fashehccd979b2005-12-15 14:31:24 -08001484 mlog_bug_on_msg(lockres->l_flags & OCFS2_LOCK_FREEING,
1485 "Cluster lock called on freeing lockres %s! flags "
1486 "0x%lx\n", lockres->l_name, lockres->l_flags);
1487
1488 /* We only compare against the currently granted level
1489 * here. If the lock is blocked waiting on a downconvert,
1490 * we'll get caught below. */
1491 if (lockres->l_flags & OCFS2_LOCK_BUSY &&
1492 level > lockres->l_level) {
1493 /* is someone sitting in dlm_lock? If so, wait on
1494 * them. */
1495 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0);
1496 wait = 1;
1497 goto unlock;
1498 }
1499
Sunil Mushrana1912822010-01-21 10:50:03 -08001500 if (lockres->l_flags & OCFS2_LOCK_UPCONVERT_FINISHING) {
1501 /*
1502 * We've upconverted. If the lock now has a level we can
1503 * work with, we take it. If, however, the lock is not at the
1504 * required level, we go thru the full cycle. One way this could
1505 * happen is if a process requesting an upconvert to PR is
1506 * closely followed by another requesting upconvert to an EX.
1507 * If the process requesting EX lands here, we want it to
1508 * continue attempting to upconvert and let the process
1509 * requesting PR take the lock.
1510 * If multiple processes request upconvert to PR, the first one
1511 * here will take the lock. The others will have to go thru the
1512 * OCFS2_LOCK_BLOCKED check to ensure that there is no pending
1513 * downconvert request.
1514 */
1515 if (level <= lockres->l_level)
1516 goto update_holders;
1517 }
1518
Mark Fashehccd979b2005-12-15 14:31:24 -08001519 if (lockres->l_flags & OCFS2_LOCK_BLOCKED &&
1520 !ocfs2_may_continue_on_blocked_lock(lockres, level)) {
1521 /* is the lock is currently blocked on behalf of
1522 * another node */
1523 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BLOCKED, 0);
1524 wait = 1;
1525 goto unlock;
1526 }
1527
1528 if (level > lockres->l_level) {
David Teigland1693a5c2008-01-30 16:52:53 -08001529 if (noqueue_attempted > 0) {
1530 ret = -EAGAIN;
1531 goto unlock;
1532 }
1533 if (lkm_flags & DLM_LKF_NOQUEUE)
1534 noqueue_attempted = 1;
1535
Mark Fashehccd979b2005-12-15 14:31:24 -08001536 if (lockres->l_action != OCFS2_AST_INVALID)
1537 mlog(ML_ERROR, "lockres %s has action %u pending\n",
1538 lockres->l_name, lockres->l_action);
1539
Mark Fasheh019d1b22007-10-05 12:09:05 -07001540 if (!(lockres->l_flags & OCFS2_LOCK_ATTACHED)) {
1541 lockres->l_action = OCFS2_AST_ATTACH;
Joel Beckerbd3e7612008-02-01 12:14:57 -08001542 lkm_flags &= ~DLM_LKF_CONVERT;
Mark Fasheh019d1b22007-10-05 12:09:05 -07001543 } else {
1544 lockres->l_action = OCFS2_AST_CONVERT;
Joel Beckerbd3e7612008-02-01 12:14:57 -08001545 lkm_flags |= DLM_LKF_CONVERT;
Mark Fasheh019d1b22007-10-05 12:09:05 -07001546 }
1547
Mark Fashehccd979b2005-12-15 14:31:24 -08001548 lockres->l_requested = level;
1549 lockres_or_flags(lockres, OCFS2_LOCK_BUSY);
Joel Beckerde551242008-02-01 14:45:08 -08001550 gen = lockres_set_pending(lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08001551 spin_unlock_irqrestore(&lockres->l_lock, flags);
1552
Joel Beckerbd3e7612008-02-01 12:14:57 -08001553 BUG_ON(level == DLM_LOCK_IV);
1554 BUG_ON(level == DLM_LOCK_NL);
Mark Fashehccd979b2005-12-15 14:31:24 -08001555
Sunil Mushran9b915182010-02-26 19:42:44 -08001556 mlog(ML_BASTS, "lockres %s, convert from %d to %d\n",
Mark Fashehccd979b2005-12-15 14:31:24 -08001557 lockres->l_name, lockres->l_level, level);
1558
1559 /* call dlm_lock to upgrade lock now */
Joel Becker4670c462008-02-01 14:39:35 -08001560 ret = ocfs2_dlm_lock(osb->cconn,
Joel Becker7431cd72008-02-01 12:15:37 -08001561 level,
1562 &lockres->l_lksb,
1563 lkm_flags,
1564 lockres->l_name,
Joel Beckera796d282010-01-28 19:22:39 -08001565 OCFS2_LOCK_ID_MAX_LEN - 1);
Joel Beckerde551242008-02-01 14:45:08 -08001566 lockres_clear_pending(lockres, gen, osb);
Joel Becker7431cd72008-02-01 12:15:37 -08001567 if (ret) {
1568 if (!(lkm_flags & DLM_LKF_NOQUEUE) ||
1569 (ret != -EAGAIN)) {
Joel Becker24ef1812008-01-29 17:37:32 -08001570 ocfs2_log_dlm_error("ocfs2_dlm_lock",
Joel Becker7431cd72008-02-01 12:15:37 -08001571 ret, lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08001572 }
1573 ocfs2_recover_from_dlm_error(lockres, 1);
1574 goto out;
1575 }
Xue jiufeid1e782382014-12-10 15:41:59 -08001576 dlm_locked = 1;
Mark Fashehccd979b2005-12-15 14:31:24 -08001577
Coly Li73ac36e2009-01-07 18:09:16 -08001578 mlog(0, "lock %s, successful return from ocfs2_dlm_lock\n",
Mark Fashehccd979b2005-12-15 14:31:24 -08001579 lockres->l_name);
1580
1581 /* At this point we've gone inside the dlm and need to
1582 * complete our work regardless. */
1583 catch_signals = 0;
1584
1585 /* wait for busy to clear and carry on */
1586 goto again;
1587 }
1588
Sunil Mushrana1912822010-01-21 10:50:03 -08001589update_holders:
Mark Fashehccd979b2005-12-15 14:31:24 -08001590 /* Ok, if we get here then we're good to go. */
1591 ocfs2_inc_holders(lockres, level);
1592
1593 ret = 0;
1594unlock:
Sunil Mushrana1912822010-01-21 10:50:03 -08001595 lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
1596
Tariq Saeedb1b1e152016-01-21 16:40:39 -08001597 /* ocfs2_unblock_lock reques on seeing OCFS2_LOCK_UPCONVERT_FINISHING */
1598 kick_dc = (lockres->l_flags & OCFS2_LOCK_BLOCKED);
1599
Mark Fashehccd979b2005-12-15 14:31:24 -08001600 spin_unlock_irqrestore(&lockres->l_lock, flags);
Tariq Saeedb1b1e152016-01-21 16:40:39 -08001601 if (kick_dc)
1602 ocfs2_wake_downconvert_thread(osb);
Mark Fashehccd979b2005-12-15 14:31:24 -08001603out:
1604 /*
1605 * This is helping work around a lock inversion between the page lock
1606 * and dlm locks. One path holds the page lock while calling aops
1607 * which block acquiring dlm locks. The voting thread holds dlm
1608 * locks while acquiring page locks while down converting data locks.
1609 * This block is helping an aop path notice the inversion and back
1610 * off to unlock its page lock before trying the dlm lock again.
1611 */
1612 if (wait && arg_flags & OCFS2_LOCK_NONBLOCK &&
1613 mw.mw_mask & (OCFS2_LOCK_BUSY|OCFS2_LOCK_BLOCKED)) {
1614 wait = 0;
Xue jiufeid1e782382014-12-10 15:41:59 -08001615 spin_lock_irqsave(&lockres->l_lock, flags);
1616 if (__lockres_remove_mask_waiter(lockres, &mw)) {
1617 if (dlm_locked)
1618 lockres_or_flags(lockres,
1619 OCFS2_LOCK_NONBLOCK_FINISHED);
1620 spin_unlock_irqrestore(&lockres->l_lock, flags);
Mark Fashehccd979b2005-12-15 14:31:24 -08001621 ret = -EAGAIN;
Xue jiufeid1e782382014-12-10 15:41:59 -08001622 } else {
1623 spin_unlock_irqrestore(&lockres->l_lock, flags);
Mark Fashehccd979b2005-12-15 14:31:24 -08001624 goto again;
Xue jiufeid1e782382014-12-10 15:41:59 -08001625 }
Mark Fashehccd979b2005-12-15 14:31:24 -08001626 }
1627 if (wait) {
1628 ret = ocfs2_wait_for_mask(&mw);
1629 if (ret == 0)
1630 goto again;
1631 mlog_errno(ret);
1632 }
Sunil Mushran8ddb7b02008-05-13 13:45:15 -07001633 ocfs2_update_lock_stats(lockres, level, &mw, ret);
Mark Fashehccd979b2005-12-15 14:31:24 -08001634
Jan Karacb257972009-06-04 15:26:50 +02001635#ifdef CONFIG_DEBUG_LOCK_ALLOC
1636 if (!ret && lockres->l_lockdep_map.key != NULL) {
1637 if (level == DLM_LOCK_PR)
1638 rwsem_acquire_read(&lockres->l_lockdep_map, l_subclass,
1639 !!(arg_flags & OCFS2_META_LOCK_NOQUEUE),
1640 caller_ip);
1641 else
1642 rwsem_acquire(&lockres->l_lockdep_map, l_subclass,
1643 !!(arg_flags & OCFS2_META_LOCK_NOQUEUE),
1644 caller_ip);
1645 }
1646#endif
Mark Fashehccd979b2005-12-15 14:31:24 -08001647 return ret;
1648}
1649
Jan Karacb257972009-06-04 15:26:50 +02001650static inline int ocfs2_cluster_lock(struct ocfs2_super *osb,
1651 struct ocfs2_lock_res *lockres,
1652 int level,
1653 u32 lkm_flags,
1654 int arg_flags)
1655{
1656 return __ocfs2_cluster_lock(osb, lockres, level, lkm_flags, arg_flags,
1657 0, _RET_IP_);
1658}
1659
1660
1661static void __ocfs2_cluster_unlock(struct ocfs2_super *osb,
1662 struct ocfs2_lock_res *lockres,
1663 int level,
1664 unsigned long caller_ip)
Mark Fashehccd979b2005-12-15 14:31:24 -08001665{
1666 unsigned long flags;
1667
Mark Fashehccd979b2005-12-15 14:31:24 -08001668 spin_lock_irqsave(&lockres->l_lock, flags);
1669 ocfs2_dec_holders(lockres, level);
Mark Fasheh34d024f2007-09-24 15:56:19 -07001670 ocfs2_downconvert_on_unlock(osb, lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08001671 spin_unlock_irqrestore(&lockres->l_lock, flags);
Jan Karacb257972009-06-04 15:26:50 +02001672#ifdef CONFIG_DEBUG_LOCK_ALLOC
1673 if (lockres->l_lockdep_map.key != NULL)
1674 rwsem_release(&lockres->l_lockdep_map, 1, caller_ip);
1675#endif
Mark Fashehccd979b2005-12-15 14:31:24 -08001676}
1677
Adrian Bunkda661162006-11-20 03:24:28 +01001678static int ocfs2_create_new_lock(struct ocfs2_super *osb,
1679 struct ocfs2_lock_res *lockres,
1680 int ex,
1681 int local)
Mark Fashehccd979b2005-12-15 14:31:24 -08001682{
Joel Beckerbd3e7612008-02-01 12:14:57 -08001683 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
Mark Fashehccd979b2005-12-15 14:31:24 -08001684 unsigned long flags;
Joel Beckerbd3e7612008-02-01 12:14:57 -08001685 u32 lkm_flags = local ? DLM_LKF_LOCAL : 0;
Mark Fashehccd979b2005-12-15 14:31:24 -08001686
1687 spin_lock_irqsave(&lockres->l_lock, flags);
1688 BUG_ON(lockres->l_flags & OCFS2_LOCK_ATTACHED);
1689 lockres_or_flags(lockres, OCFS2_LOCK_LOCAL);
1690 spin_unlock_irqrestore(&lockres->l_lock, flags);
1691
Mark Fasheh24c19ef2006-09-22 17:28:19 -07001692 return ocfs2_lock_create(osb, lockres, level, lkm_flags);
Mark Fashehccd979b2005-12-15 14:31:24 -08001693}
1694
1695/* Grants us an EX lock on the data and metadata resources, skipping
1696 * the normal cluster directory lookup. Use this ONLY on newly created
1697 * inodes which other nodes can't possibly see, and which haven't been
1698 * hashed in the inode hash yet. This can give us a good performance
1699 * increase as it'll skip the network broadcast normally associated
1700 * with creating a new lock resource. */
1701int ocfs2_create_new_inode_locks(struct inode *inode)
1702{
1703 int ret;
Mark Fashehd680efe2006-09-08 14:14:34 -07001704 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
Mark Fashehccd979b2005-12-15 14:31:24 -08001705
Mark Fashehccd979b2005-12-15 14:31:24 -08001706 BUG_ON(!ocfs2_inode_is_new(inode));
1707
Mark Fashehb06970532006-03-03 10:24:33 -08001708 mlog(0, "Inode %llu\n", (unsigned long long)OCFS2_I(inode)->ip_blkno);
Mark Fashehccd979b2005-12-15 14:31:24 -08001709
1710 /* NOTE: That we don't increment any of the holder counts, nor
1711 * do we add anything to a journal handle. Since this is
1712 * supposed to be a new inode which the cluster doesn't know
1713 * about yet, there is no need to. As far as the LVB handling
1714 * is concerned, this is basically like acquiring an EX lock
1715 * on a resource which has an invalid one -- we'll set it
1716 * valid when we release the EX. */
1717
Mark Fasheh24c19ef2006-09-22 17:28:19 -07001718 ret = ocfs2_create_new_lock(osb, &OCFS2_I(inode)->ip_rw_lockres, 1, 1);
Mark Fashehccd979b2005-12-15 14:31:24 -08001719 if (ret) {
1720 mlog_errno(ret);
1721 goto bail;
1722 }
1723
Mark Fasheh24c19ef2006-09-22 17:28:19 -07001724 /*
Joel Beckerbd3e7612008-02-01 12:14:57 -08001725 * We don't want to use DLM_LKF_LOCAL on a meta data lock as they
Mark Fasheh24c19ef2006-09-22 17:28:19 -07001726 * don't use a generation in their lock names.
1727 */
Mark Fashehe63aecb62007-10-18 15:30:42 -07001728 ret = ocfs2_create_new_lock(osb, &OCFS2_I(inode)->ip_inode_lockres, 1, 0);
Mark Fashehccd979b2005-12-15 14:31:24 -08001729 if (ret) {
1730 mlog_errno(ret);
1731 goto bail;
1732 }
1733
Tiger Yang50008632007-03-20 16:01:38 -07001734 ret = ocfs2_create_new_lock(osb, &OCFS2_I(inode)->ip_open_lockres, 0, 0);
Joseph Qia8f24f1b2016-07-26 15:21:35 -07001735 if (ret)
Tiger Yang50008632007-03-20 16:01:38 -07001736 mlog_errno(ret);
Tiger Yang50008632007-03-20 16:01:38 -07001737
Mark Fashehccd979b2005-12-15 14:31:24 -08001738bail:
Mark Fashehccd979b2005-12-15 14:31:24 -08001739 return ret;
1740}
1741
1742int ocfs2_rw_lock(struct inode *inode, int write)
1743{
1744 int status, level;
1745 struct ocfs2_lock_res *lockres;
Sunil Mushranc271c5c2006-12-05 17:56:35 -08001746 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
Mark Fashehccd979b2005-12-15 14:31:24 -08001747
Mark Fashehb06970532006-03-03 10:24:33 -08001748 mlog(0, "inode %llu take %s RW lock\n",
1749 (unsigned long long)OCFS2_I(inode)->ip_blkno,
Mark Fashehccd979b2005-12-15 14:31:24 -08001750 write ? "EXMODE" : "PRMODE");
1751
Tao Mac1e8d352011-03-07 16:43:21 +08001752 if (ocfs2_mount_local(osb))
Sunil Mushranc271c5c2006-12-05 17:56:35 -08001753 return 0;
1754
Mark Fashehccd979b2005-12-15 14:31:24 -08001755 lockres = &OCFS2_I(inode)->ip_rw_lockres;
1756
Joel Beckerbd3e7612008-02-01 12:14:57 -08001757 level = write ? DLM_LOCK_EX : DLM_LOCK_PR;
Mark Fashehccd979b2005-12-15 14:31:24 -08001758
1759 status = ocfs2_cluster_lock(OCFS2_SB(inode->i_sb), lockres, level, 0,
1760 0);
1761 if (status < 0)
1762 mlog_errno(status);
1763
Mark Fashehccd979b2005-12-15 14:31:24 -08001764 return status;
1765}
1766
Gang He06e7f132018-01-31 16:15:17 -08001767int ocfs2_try_rw_lock(struct inode *inode, int write)
1768{
1769 int status, level;
1770 struct ocfs2_lock_res *lockres;
1771 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1772
1773 mlog(0, "inode %llu try to take %s RW lock\n",
1774 (unsigned long long)OCFS2_I(inode)->ip_blkno,
1775 write ? "EXMODE" : "PRMODE");
1776
1777 if (ocfs2_mount_local(osb))
1778 return 0;
1779
1780 lockres = &OCFS2_I(inode)->ip_rw_lockres;
1781
1782 level = write ? DLM_LOCK_EX : DLM_LOCK_PR;
1783
1784 status = ocfs2_cluster_lock(osb, lockres, level, DLM_LKF_NOQUEUE, 0);
1785 return status;
1786}
1787
Mark Fashehccd979b2005-12-15 14:31:24 -08001788void ocfs2_rw_unlock(struct inode *inode, int write)
1789{
Joel Beckerbd3e7612008-02-01 12:14:57 -08001790 int level = write ? DLM_LOCK_EX : DLM_LOCK_PR;
Mark Fashehccd979b2005-12-15 14:31:24 -08001791 struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_rw_lockres;
Sunil Mushranc271c5c2006-12-05 17:56:35 -08001792 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
Mark Fashehccd979b2005-12-15 14:31:24 -08001793
Mark Fashehb06970532006-03-03 10:24:33 -08001794 mlog(0, "inode %llu drop %s RW lock\n",
1795 (unsigned long long)OCFS2_I(inode)->ip_blkno,
Mark Fashehccd979b2005-12-15 14:31:24 -08001796 write ? "EXMODE" : "PRMODE");
1797
Sunil Mushranc271c5c2006-12-05 17:56:35 -08001798 if (!ocfs2_mount_local(osb))
1799 ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, level);
Mark Fashehccd979b2005-12-15 14:31:24 -08001800}
1801
Tiger Yang50008632007-03-20 16:01:38 -07001802/*
1803 * ocfs2_open_lock always get PR mode lock.
1804 */
1805int ocfs2_open_lock(struct inode *inode)
1806{
1807 int status = 0;
1808 struct ocfs2_lock_res *lockres;
1809 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1810
Tiger Yang50008632007-03-20 16:01:38 -07001811 mlog(0, "inode %llu take PRMODE open lock\n",
1812 (unsigned long long)OCFS2_I(inode)->ip_blkno);
1813
Tiger Yang03efed82011-05-28 00:34:19 +08001814 if (ocfs2_is_hard_readonly(osb) || ocfs2_mount_local(osb))
Tiger Yang50008632007-03-20 16:01:38 -07001815 goto out;
1816
1817 lockres = &OCFS2_I(inode)->ip_open_lockres;
1818
1819 status = ocfs2_cluster_lock(OCFS2_SB(inode->i_sb), lockres,
Joel Beckerbd3e7612008-02-01 12:14:57 -08001820 DLM_LOCK_PR, 0, 0);
Tiger Yang50008632007-03-20 16:01:38 -07001821 if (status < 0)
1822 mlog_errno(status);
1823
1824out:
Tiger Yang50008632007-03-20 16:01:38 -07001825 return status;
1826}
1827
1828int ocfs2_try_open_lock(struct inode *inode, int write)
1829{
1830 int status = 0, level;
1831 struct ocfs2_lock_res *lockres;
1832 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1833
Tiger Yang50008632007-03-20 16:01:38 -07001834 mlog(0, "inode %llu try to take %s open lock\n",
1835 (unsigned long long)OCFS2_I(inode)->ip_blkno,
1836 write ? "EXMODE" : "PRMODE");
1837
Tiger Yang03efed82011-05-28 00:34:19 +08001838 if (ocfs2_is_hard_readonly(osb)) {
1839 if (write)
1840 status = -EROFS;
1841 goto out;
1842 }
1843
Tiger Yang50008632007-03-20 16:01:38 -07001844 if (ocfs2_mount_local(osb))
1845 goto out;
1846
1847 lockres = &OCFS2_I(inode)->ip_open_lockres;
1848
Joel Beckerbd3e7612008-02-01 12:14:57 -08001849 level = write ? DLM_LOCK_EX : DLM_LOCK_PR;
Tiger Yang50008632007-03-20 16:01:38 -07001850
1851 /*
1852 * The file system may already holding a PRMODE/EXMODE open lock.
Joel Beckerbd3e7612008-02-01 12:14:57 -08001853 * Since we pass DLM_LKF_NOQUEUE, the request won't block waiting on
Tiger Yang50008632007-03-20 16:01:38 -07001854 * other nodes and the -EAGAIN will indicate to the caller that
1855 * this inode is still in use.
1856 */
1857 status = ocfs2_cluster_lock(OCFS2_SB(inode->i_sb), lockres,
Joel Beckerbd3e7612008-02-01 12:14:57 -08001858 level, DLM_LKF_NOQUEUE, 0);
Tiger Yang50008632007-03-20 16:01:38 -07001859
1860out:
Tiger Yang50008632007-03-20 16:01:38 -07001861 return status;
1862}
1863
1864/*
1865 * ocfs2_open_unlock unlock PR and EX mode open locks.
1866 */
1867void ocfs2_open_unlock(struct inode *inode)
1868{
1869 struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_open_lockres;
1870 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
1871
Tiger Yang50008632007-03-20 16:01:38 -07001872 mlog(0, "inode %llu drop open lock\n",
1873 (unsigned long long)OCFS2_I(inode)->ip_blkno);
1874
1875 if (ocfs2_mount_local(osb))
1876 goto out;
1877
1878 if(lockres->l_ro_holders)
1879 ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres,
Joel Beckerbd3e7612008-02-01 12:14:57 -08001880 DLM_LOCK_PR);
Tiger Yang50008632007-03-20 16:01:38 -07001881 if(lockres->l_ex_holders)
1882 ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres,
Joel Beckerbd3e7612008-02-01 12:14:57 -08001883 DLM_LOCK_EX);
Tiger Yang50008632007-03-20 16:01:38 -07001884
1885out:
Tao Mac1e8d352011-03-07 16:43:21 +08001886 return;
Tiger Yang50008632007-03-20 16:01:38 -07001887}
1888
Mark Fashehcf8e06f2007-12-20 16:43:10 -08001889static int ocfs2_flock_handle_signal(struct ocfs2_lock_res *lockres,
1890 int level)
1891{
1892 int ret;
1893 struct ocfs2_super *osb = ocfs2_get_lockres_osb(lockres);
1894 unsigned long flags;
1895 struct ocfs2_mask_waiter mw;
1896
1897 ocfs2_init_mask_waiter(&mw);
1898
1899retry_cancel:
1900 spin_lock_irqsave(&lockres->l_lock, flags);
1901 if (lockres->l_flags & OCFS2_LOCK_BUSY) {
1902 ret = ocfs2_prepare_cancel_convert(osb, lockres);
1903 if (ret) {
1904 spin_unlock_irqrestore(&lockres->l_lock, flags);
1905 ret = ocfs2_cancel_convert(osb, lockres);
1906 if (ret < 0) {
1907 mlog_errno(ret);
1908 goto out;
1909 }
1910 goto retry_cancel;
1911 }
1912 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0);
1913 spin_unlock_irqrestore(&lockres->l_lock, flags);
1914
1915 ocfs2_wait_for_mask(&mw);
1916 goto retry_cancel;
1917 }
1918
1919 ret = -ERESTARTSYS;
1920 /*
1921 * We may still have gotten the lock, in which case there's no
1922 * point to restarting the syscall.
1923 */
1924 if (lockres->l_level == level)
1925 ret = 0;
1926
1927 mlog(0, "Cancel returning %d. flags: 0x%lx, level: %d, act: %d\n", ret,
1928 lockres->l_flags, lockres->l_level, lockres->l_action);
1929
1930 spin_unlock_irqrestore(&lockres->l_lock, flags);
1931
1932out:
1933 return ret;
1934}
1935
1936/*
1937 * ocfs2_file_lock() and ocfs2_file_unlock() map to a single pair of
1938 * flock() calls. The locking approach this requires is sufficiently
1939 * different from all other cluster lock types that we implement a
Daniel Mack3ad2f3fb2010-02-03 08:01:28 +08001940 * separate path to the "low-level" dlm calls. In particular:
Mark Fashehcf8e06f2007-12-20 16:43:10 -08001941 *
1942 * - No optimization of lock levels is done - we take at exactly
1943 * what's been requested.
1944 *
1945 * - No lock caching is employed. We immediately downconvert to
1946 * no-lock at unlock time. This also means flock locks never go on
1947 * the blocking list).
1948 *
1949 * - Since userspace can trivially deadlock itself with flock, we make
1950 * sure to allow cancellation of a misbehaving applications flock()
1951 * request.
1952 *
1953 * - Access to any flock lockres doesn't require concurrency, so we
1954 * can simplify the code by requiring the caller to guarantee
1955 * serialization of dlmglue flock calls.
1956 */
1957int ocfs2_file_lock(struct file *file, int ex, int trylock)
1958{
Mark Fashehe988cf12008-07-10 09:25:39 -07001959 int ret, level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
1960 unsigned int lkm_flags = trylock ? DLM_LKF_NOQUEUE : 0;
Mark Fashehcf8e06f2007-12-20 16:43:10 -08001961 unsigned long flags;
1962 struct ocfs2_file_private *fp = file->private_data;
1963 struct ocfs2_lock_res *lockres = &fp->fp_flock;
1964 struct ocfs2_super *osb = OCFS2_SB(file->f_mapping->host->i_sb);
1965 struct ocfs2_mask_waiter mw;
1966
1967 ocfs2_init_mask_waiter(&mw);
1968
1969 if ((lockres->l_flags & OCFS2_LOCK_BUSY) ||
Joel Beckerbd3e7612008-02-01 12:14:57 -08001970 (lockres->l_level > DLM_LOCK_NL)) {
Mark Fashehcf8e06f2007-12-20 16:43:10 -08001971 mlog(ML_ERROR,
1972 "File lock \"%s\" has busy or locked state: flags: 0x%lx, "
1973 "level: %u\n", lockres->l_name, lockres->l_flags,
1974 lockres->l_level);
1975 return -EINVAL;
1976 }
1977
1978 spin_lock_irqsave(&lockres->l_lock, flags);
1979 if (!(lockres->l_flags & OCFS2_LOCK_ATTACHED)) {
1980 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0);
1981 spin_unlock_irqrestore(&lockres->l_lock, flags);
1982
1983 /*
1984 * Get the lock at NLMODE to start - that way we
1985 * can cancel the upconvert request if need be.
1986 */
Mark Fashehe988cf12008-07-10 09:25:39 -07001987 ret = ocfs2_lock_create(osb, lockres, DLM_LOCK_NL, 0);
Mark Fashehcf8e06f2007-12-20 16:43:10 -08001988 if (ret < 0) {
1989 mlog_errno(ret);
1990 goto out;
1991 }
1992
1993 ret = ocfs2_wait_for_mask(&mw);
1994 if (ret) {
1995 mlog_errno(ret);
1996 goto out;
1997 }
1998 spin_lock_irqsave(&lockres->l_lock, flags);
1999 }
2000
2001 lockres->l_action = OCFS2_AST_CONVERT;
Mark Fashehe988cf12008-07-10 09:25:39 -07002002 lkm_flags |= DLM_LKF_CONVERT;
Mark Fashehcf8e06f2007-12-20 16:43:10 -08002003 lockres->l_requested = level;
2004 lockres_or_flags(lockres, OCFS2_LOCK_BUSY);
2005
2006 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0);
2007 spin_unlock_irqrestore(&lockres->l_lock, flags);
2008
Joel Becker4670c462008-02-01 14:39:35 -08002009 ret = ocfs2_dlm_lock(osb->cconn, level, &lockres->l_lksb, lkm_flags,
Joel Beckera796d282010-01-28 19:22:39 -08002010 lockres->l_name, OCFS2_LOCK_ID_MAX_LEN - 1);
Joel Becker7431cd72008-02-01 12:15:37 -08002011 if (ret) {
2012 if (!trylock || (ret != -EAGAIN)) {
Joel Becker24ef1812008-01-29 17:37:32 -08002013 ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres);
Mark Fashehcf8e06f2007-12-20 16:43:10 -08002014 ret = -EINVAL;
2015 }
2016
2017 ocfs2_recover_from_dlm_error(lockres, 1);
2018 lockres_remove_mask_waiter(lockres, &mw);
2019 goto out;
2020 }
2021
2022 ret = ocfs2_wait_for_mask_interruptible(&mw, lockres);
2023 if (ret == -ERESTARTSYS) {
2024 /*
2025 * Userspace can cause deadlock itself with
2026 * flock(). Current behavior locally is to allow the
2027 * deadlock, but abort the system call if a signal is
2028 * received. We follow this example, otherwise a
2029 * poorly written program could sit in kernel until
2030 * reboot.
2031 *
2032 * Handling this is a bit more complicated for Ocfs2
2033 * though. We can't exit this function with an
2034 * outstanding lock request, so a cancel convert is
2035 * required. We intentionally overwrite 'ret' - if the
2036 * cancel fails and the lock was granted, it's easier
André Goddard Rosaaf901ca2009-11-14 13:09:05 -02002037 * to just bubble success back up to the user.
Mark Fashehcf8e06f2007-12-20 16:43:10 -08002038 */
2039 ret = ocfs2_flock_handle_signal(lockres, level);
David Teigland1693a5c2008-01-30 16:52:53 -08002040 } else if (!ret && (level > lockres->l_level)) {
2041 /* Trylock failed asynchronously */
2042 BUG_ON(!trylock);
2043 ret = -EAGAIN;
Mark Fashehcf8e06f2007-12-20 16:43:10 -08002044 }
2045
2046out:
2047
2048 mlog(0, "Lock: \"%s\" ex: %d, trylock: %d, returns: %d\n",
2049 lockres->l_name, ex, trylock, ret);
2050 return ret;
2051}
2052
2053void ocfs2_file_unlock(struct file *file)
2054{
2055 int ret;
Joel Beckerde551242008-02-01 14:45:08 -08002056 unsigned int gen;
Mark Fashehcf8e06f2007-12-20 16:43:10 -08002057 unsigned long flags;
2058 struct ocfs2_file_private *fp = file->private_data;
2059 struct ocfs2_lock_res *lockres = &fp->fp_flock;
2060 struct ocfs2_super *osb = OCFS2_SB(file->f_mapping->host->i_sb);
2061 struct ocfs2_mask_waiter mw;
2062
2063 ocfs2_init_mask_waiter(&mw);
2064
2065 if (!(lockres->l_flags & OCFS2_LOCK_ATTACHED))
2066 return;
2067
Mark Fashehe988cf12008-07-10 09:25:39 -07002068 if (lockres->l_level == DLM_LOCK_NL)
Mark Fashehcf8e06f2007-12-20 16:43:10 -08002069 return;
2070
2071 mlog(0, "Unlock: \"%s\" flags: 0x%lx, level: %d, act: %d\n",
2072 lockres->l_name, lockres->l_flags, lockres->l_level,
2073 lockres->l_action);
2074
2075 spin_lock_irqsave(&lockres->l_lock, flags);
2076 /*
2077 * Fake a blocking ast for the downconvert code.
2078 */
2079 lockres_or_flags(lockres, OCFS2_LOCK_BLOCKED);
Joel Beckerbd3e7612008-02-01 12:14:57 -08002080 lockres->l_blocking = DLM_LOCK_EX;
Mark Fashehcf8e06f2007-12-20 16:43:10 -08002081
Mark Fashehe988cf12008-07-10 09:25:39 -07002082 gen = ocfs2_prepare_downconvert(lockres, DLM_LOCK_NL);
Mark Fashehcf8e06f2007-12-20 16:43:10 -08002083 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0);
2084 spin_unlock_irqrestore(&lockres->l_lock, flags);
2085
Mark Fashehe988cf12008-07-10 09:25:39 -07002086 ret = ocfs2_downconvert_lock(osb, lockres, DLM_LOCK_NL, 0, gen);
Mark Fashehcf8e06f2007-12-20 16:43:10 -08002087 if (ret) {
2088 mlog_errno(ret);
2089 return;
2090 }
2091
2092 ret = ocfs2_wait_for_mask(&mw);
2093 if (ret)
2094 mlog_errno(ret);
2095}
2096
Mark Fasheh34d024f2007-09-24 15:56:19 -07002097static void ocfs2_downconvert_on_unlock(struct ocfs2_super *osb,
2098 struct ocfs2_lock_res *lockres)
Mark Fashehccd979b2005-12-15 14:31:24 -08002099{
2100 int kick = 0;
2101
Mark Fashehccd979b2005-12-15 14:31:24 -08002102 /* If we know that another node is waiting on our lock, kick
Mark Fasheh34d024f2007-09-24 15:56:19 -07002103 * the downconvert thread * pre-emptively when we reach a release
Mark Fashehccd979b2005-12-15 14:31:24 -08002104 * condition. */
2105 if (lockres->l_flags & OCFS2_LOCK_BLOCKED) {
2106 switch(lockres->l_blocking) {
Joel Beckerbd3e7612008-02-01 12:14:57 -08002107 case DLM_LOCK_EX:
Mark Fashehccd979b2005-12-15 14:31:24 -08002108 if (!lockres->l_ex_holders && !lockres->l_ro_holders)
2109 kick = 1;
2110 break;
Joel Beckerbd3e7612008-02-01 12:14:57 -08002111 case DLM_LOCK_PR:
Mark Fashehccd979b2005-12-15 14:31:24 -08002112 if (!lockres->l_ex_holders)
2113 kick = 1;
2114 break;
2115 default:
2116 BUG();
2117 }
2118 }
2119
2120 if (kick)
Mark Fasheh34d024f2007-09-24 15:56:19 -07002121 ocfs2_wake_downconvert_thread(osb);
Mark Fashehccd979b2005-12-15 14:31:24 -08002122}
2123
Mark Fashehccd979b2005-12-15 14:31:24 -08002124#define OCFS2_SEC_BITS 34
2125#define OCFS2_SEC_SHIFT (64 - 34)
2126#define OCFS2_NSEC_MASK ((1ULL << OCFS2_SEC_SHIFT) - 1)
2127
2128/* LVB only has room for 64 bits of time here so we pack it for
2129 * now. */
2130static u64 ocfs2_pack_timespec(struct timespec *spec)
2131{
2132 u64 res;
2133 u64 sec = spec->tv_sec;
2134 u32 nsec = spec->tv_nsec;
2135
2136 res = (sec << OCFS2_SEC_SHIFT) | (nsec & OCFS2_NSEC_MASK);
2137
2138 return res;
2139}
2140
2141/* Call this with the lockres locked. I am reasonably sure we don't
2142 * need ip_lock in this function as anyone who would be changing those
Mark Fashehe63aecb62007-10-18 15:30:42 -07002143 * values is supposed to be blocked in ocfs2_inode_lock right now. */
Mark Fashehccd979b2005-12-15 14:31:24 -08002144static void __ocfs2_stuff_meta_lvb(struct inode *inode)
2145{
2146 struct ocfs2_inode_info *oi = OCFS2_I(inode);
Mark Fashehe63aecb62007-10-18 15:30:42 -07002147 struct ocfs2_lock_res *lockres = &oi->ip_inode_lockres;
Mark Fashehccd979b2005-12-15 14:31:24 -08002148 struct ocfs2_meta_lvb *lvb;
2149
Mark Fasheha641dc22008-12-24 16:03:48 -08002150 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
Mark Fashehccd979b2005-12-15 14:31:24 -08002151
Mark Fasheh24c19ef2006-09-22 17:28:19 -07002152 /*
2153 * Invalidate the LVB of a deleted inode - this way other
2154 * nodes are forced to go to disk and discover the new inode
2155 * status.
2156 */
2157 if (oi->ip_flags & OCFS2_INODE_DELETED) {
2158 lvb->lvb_version = 0;
2159 goto out;
2160 }
2161
Mark Fasheh4d3b83f2006-09-12 15:22:18 -07002162 lvb->lvb_version = OCFS2_LVB_VERSION;
Mark Fashehccd979b2005-12-15 14:31:24 -08002163 lvb->lvb_isize = cpu_to_be64(i_size_read(inode));
2164 lvb->lvb_iclusters = cpu_to_be32(oi->ip_clusters);
Eric W. Biederman03ab30f2013-01-31 16:59:23 -08002165 lvb->lvb_iuid = cpu_to_be32(i_uid_read(inode));
2166 lvb->lvb_igid = cpu_to_be32(i_gid_read(inode));
Mark Fashehccd979b2005-12-15 14:31:24 -08002167 lvb->lvb_imode = cpu_to_be16(inode->i_mode);
2168 lvb->lvb_inlink = cpu_to_be16(inode->i_nlink);
2169 lvb->lvb_iatime_packed =
2170 cpu_to_be64(ocfs2_pack_timespec(&inode->i_atime));
2171 lvb->lvb_ictime_packed =
2172 cpu_to_be64(ocfs2_pack_timespec(&inode->i_ctime));
2173 lvb->lvb_imtime_packed =
2174 cpu_to_be64(ocfs2_pack_timespec(&inode->i_mtime));
Herbert Poetzlca4d1472006-07-03 17:27:12 -07002175 lvb->lvb_iattr = cpu_to_be32(oi->ip_attr);
Mark Fasheh15b1e362007-09-07 13:58:15 -07002176 lvb->lvb_idynfeatures = cpu_to_be16(oi->ip_dyn_features);
Mark Fashehf9e2d822006-09-12 15:35:49 -07002177 lvb->lvb_igeneration = cpu_to_be32(inode->i_generation);
Mark Fashehccd979b2005-12-15 14:31:24 -08002178
Mark Fasheh24c19ef2006-09-22 17:28:19 -07002179out:
Mark Fashehccd979b2005-12-15 14:31:24 -08002180 mlog_meta_lvb(0, lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08002181}
2182
2183static void ocfs2_unpack_timespec(struct timespec *spec,
2184 u64 packed_time)
2185{
2186 spec->tv_sec = packed_time >> OCFS2_SEC_SHIFT;
2187 spec->tv_nsec = packed_time & OCFS2_NSEC_MASK;
2188}
2189
2190static void ocfs2_refresh_inode_from_lvb(struct inode *inode)
2191{
2192 struct ocfs2_inode_info *oi = OCFS2_I(inode);
Mark Fashehe63aecb62007-10-18 15:30:42 -07002193 struct ocfs2_lock_res *lockres = &oi->ip_inode_lockres;
Mark Fashehccd979b2005-12-15 14:31:24 -08002194 struct ocfs2_meta_lvb *lvb;
2195
Mark Fashehccd979b2005-12-15 14:31:24 -08002196 mlog_meta_lvb(0, lockres);
2197
Mark Fasheha641dc22008-12-24 16:03:48 -08002198 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
Mark Fashehccd979b2005-12-15 14:31:24 -08002199
2200 /* We're safe here without the lockres lock... */
2201 spin_lock(&oi->ip_lock);
2202 oi->ip_clusters = be32_to_cpu(lvb->lvb_iclusters);
2203 i_size_write(inode, be64_to_cpu(lvb->lvb_isize));
2204
Herbert Poetzlca4d1472006-07-03 17:27:12 -07002205 oi->ip_attr = be32_to_cpu(lvb->lvb_iattr);
Mark Fasheh15b1e362007-09-07 13:58:15 -07002206 oi->ip_dyn_features = be16_to_cpu(lvb->lvb_idynfeatures);
Herbert Poetzlca4d1472006-07-03 17:27:12 -07002207 ocfs2_set_inode_flags(inode);
2208
Mark Fashehccd979b2005-12-15 14:31:24 -08002209 /* fast-symlinks are a special case */
2210 if (S_ISLNK(inode->i_mode) && !oi->ip_clusters)
2211 inode->i_blocks = 0;
2212 else
Mark Fasheh8110b072007-03-22 16:53:23 -07002213 inode->i_blocks = ocfs2_inode_sector_count(inode);
Mark Fashehccd979b2005-12-15 14:31:24 -08002214
Eric W. Biederman03ab30f2013-01-31 16:59:23 -08002215 i_uid_write(inode, be32_to_cpu(lvb->lvb_iuid));
2216 i_gid_write(inode, be32_to_cpu(lvb->lvb_igid));
Mark Fashehccd979b2005-12-15 14:31:24 -08002217 inode->i_mode = be16_to_cpu(lvb->lvb_imode);
Miklos Szeredibfe86842011-10-28 14:13:29 +02002218 set_nlink(inode, be16_to_cpu(lvb->lvb_inlink));
Mark Fashehccd979b2005-12-15 14:31:24 -08002219 ocfs2_unpack_timespec(&inode->i_atime,
2220 be64_to_cpu(lvb->lvb_iatime_packed));
2221 ocfs2_unpack_timespec(&inode->i_mtime,
2222 be64_to_cpu(lvb->lvb_imtime_packed));
2223 ocfs2_unpack_timespec(&inode->i_ctime,
2224 be64_to_cpu(lvb->lvb_ictime_packed));
2225 spin_unlock(&oi->ip_lock);
Mark Fashehccd979b2005-12-15 14:31:24 -08002226}
2227
Mark Fashehf9e2d822006-09-12 15:35:49 -07002228static inline int ocfs2_meta_lvb_is_trustable(struct inode *inode,
2229 struct ocfs2_lock_res *lockres)
Mark Fashehccd979b2005-12-15 14:31:24 -08002230{
Mark Fasheha641dc22008-12-24 16:03:48 -08002231 struct ocfs2_meta_lvb *lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
Mark Fashehccd979b2005-12-15 14:31:24 -08002232
Joel Becker1c520df2009-06-19 15:14:13 -07002233 if (ocfs2_dlm_lvb_valid(&lockres->l_lksb)
2234 && lvb->lvb_version == OCFS2_LVB_VERSION
Mark Fashehf9e2d822006-09-12 15:35:49 -07002235 && be32_to_cpu(lvb->lvb_igeneration) == inode->i_generation)
Mark Fashehccd979b2005-12-15 14:31:24 -08002236 return 1;
2237 return 0;
2238}
2239
2240/* Determine whether a lock resource needs to be refreshed, and
2241 * arbitrate who gets to refresh it.
2242 *
2243 * 0 means no refresh needed.
2244 *
2245 * > 0 means you need to refresh this and you MUST call
2246 * ocfs2_complete_lock_res_refresh afterwards. */
2247static int ocfs2_should_refresh_lock_res(struct ocfs2_lock_res *lockres)
2248{
2249 unsigned long flags;
2250 int status = 0;
2251
Mark Fashehccd979b2005-12-15 14:31:24 -08002252refresh_check:
2253 spin_lock_irqsave(&lockres->l_lock, flags);
2254 if (!(lockres->l_flags & OCFS2_LOCK_NEEDS_REFRESH)) {
2255 spin_unlock_irqrestore(&lockres->l_lock, flags);
2256 goto bail;
2257 }
2258
2259 if (lockres->l_flags & OCFS2_LOCK_REFRESHING) {
2260 spin_unlock_irqrestore(&lockres->l_lock, flags);
2261
2262 ocfs2_wait_on_refreshing_lock(lockres);
2263 goto refresh_check;
2264 }
2265
2266 /* Ok, I'll be the one to refresh this lock. */
2267 lockres_or_flags(lockres, OCFS2_LOCK_REFRESHING);
2268 spin_unlock_irqrestore(&lockres->l_lock, flags);
2269
2270 status = 1;
2271bail:
Tao Mac1e8d352011-03-07 16:43:21 +08002272 mlog(0, "status %d\n", status);
Mark Fashehccd979b2005-12-15 14:31:24 -08002273 return status;
2274}
2275
2276/* If status is non zero, I'll mark it as not being in refresh
2277 * anymroe, but i won't clear the needs refresh flag. */
2278static inline void ocfs2_complete_lock_res_refresh(struct ocfs2_lock_res *lockres,
2279 int status)
2280{
2281 unsigned long flags;
Mark Fashehccd979b2005-12-15 14:31:24 -08002282
2283 spin_lock_irqsave(&lockres->l_lock, flags);
2284 lockres_clear_flags(lockres, OCFS2_LOCK_REFRESHING);
2285 if (!status)
2286 lockres_clear_flags(lockres, OCFS2_LOCK_NEEDS_REFRESH);
2287 spin_unlock_irqrestore(&lockres->l_lock, flags);
2288
2289 wake_up(&lockres->l_event);
Mark Fashehccd979b2005-12-15 14:31:24 -08002290}
2291
2292/* may or may not return a bh if it went to disk. */
Mark Fashehe63aecb62007-10-18 15:30:42 -07002293static int ocfs2_inode_lock_update(struct inode *inode,
Mark Fashehccd979b2005-12-15 14:31:24 -08002294 struct buffer_head **bh)
2295{
2296 int status = 0;
2297 struct ocfs2_inode_info *oi = OCFS2_I(inode);
Mark Fashehe63aecb62007-10-18 15:30:42 -07002298 struct ocfs2_lock_res *lockres = &oi->ip_inode_lockres;
Mark Fashehccd979b2005-12-15 14:31:24 -08002299 struct ocfs2_dinode *fe;
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002300 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
Mark Fashehccd979b2005-12-15 14:31:24 -08002301
Mark Fashehbe9e9862007-04-18 15:22:08 -07002302 if (ocfs2_mount_local(osb))
2303 goto bail;
2304
Mark Fashehccd979b2005-12-15 14:31:24 -08002305 spin_lock(&oi->ip_lock);
2306 if (oi->ip_flags & OCFS2_INODE_DELETED) {
Mark Fashehb06970532006-03-03 10:24:33 -08002307 mlog(0, "Orphaned inode %llu was deleted while we "
Mark Fashehccd979b2005-12-15 14:31:24 -08002308 "were waiting on a lock. ip_flags = 0x%x\n",
Mark Fashehb06970532006-03-03 10:24:33 -08002309 (unsigned long long)oi->ip_blkno, oi->ip_flags);
Mark Fashehccd979b2005-12-15 14:31:24 -08002310 spin_unlock(&oi->ip_lock);
2311 status = -ENOENT;
2312 goto bail;
2313 }
2314 spin_unlock(&oi->ip_lock);
2315
Mark Fashehbe9e9862007-04-18 15:22:08 -07002316 if (!ocfs2_should_refresh_lock_res(lockres))
2317 goto bail;
Mark Fashehccd979b2005-12-15 14:31:24 -08002318
2319 /* This will discard any caching information we might have had
2320 * for the inode metadata. */
Joel Becker8cb471e2009-02-10 20:00:41 -08002321 ocfs2_metadata_cache_purge(INODE_CACHE(inode));
Mark Fashehccd979b2005-12-15 14:31:24 -08002322
Mark Fasheh83418972007-04-23 18:53:12 -07002323 ocfs2_extent_map_trunc(inode, 0);
2324
Mark Fashehbe9e9862007-04-18 15:22:08 -07002325 if (ocfs2_meta_lvb_is_trustable(inode, lockres)) {
Mark Fashehb06970532006-03-03 10:24:33 -08002326 mlog(0, "Trusting LVB on inode %llu\n",
2327 (unsigned long long)oi->ip_blkno);
Mark Fashehccd979b2005-12-15 14:31:24 -08002328 ocfs2_refresh_inode_from_lvb(inode);
2329 } else {
2330 /* Boo, we have to go to disk. */
2331 /* read bh, cast, ocfs2_refresh_inode */
Joel Beckerb657c952008-11-13 14:49:11 -08002332 status = ocfs2_read_inode_block(inode, bh);
Mark Fashehccd979b2005-12-15 14:31:24 -08002333 if (status < 0) {
2334 mlog_errno(status);
2335 goto bail_refresh;
2336 }
2337 fe = (struct ocfs2_dinode *) (*bh)->b_data;
2338
2339 /* This is a good chance to make sure we're not
Joel Beckerb657c952008-11-13 14:49:11 -08002340 * locking an invalid object. ocfs2_read_inode_block()
2341 * already checked that the inode block is sane.
Mark Fashehccd979b2005-12-15 14:31:24 -08002342 *
2343 * We bug on a stale inode here because we checked
2344 * above whether it was wiped from disk. The wiping
2345 * node provides a guarantee that we receive that
2346 * message and can mark the inode before dropping any
2347 * locks associated with it. */
Mark Fashehccd979b2005-12-15 14:31:24 -08002348 mlog_bug_on_msg(inode->i_generation !=
2349 le32_to_cpu(fe->i_generation),
Mark Fashehb06970532006-03-03 10:24:33 -08002350 "Invalid dinode %llu disk generation: %u "
Mark Fashehccd979b2005-12-15 14:31:24 -08002351 "inode->i_generation: %u\n",
Mark Fashehb06970532006-03-03 10:24:33 -08002352 (unsigned long long)oi->ip_blkno,
2353 le32_to_cpu(fe->i_generation),
Mark Fashehccd979b2005-12-15 14:31:24 -08002354 inode->i_generation);
2355 mlog_bug_on_msg(le64_to_cpu(fe->i_dtime) ||
2356 !(fe->i_flags & cpu_to_le32(OCFS2_VALID_FL)),
Mark Fashehb06970532006-03-03 10:24:33 -08002357 "Stale dinode %llu dtime: %llu flags: 0x%x\n",
2358 (unsigned long long)oi->ip_blkno,
2359 (unsigned long long)le64_to_cpu(fe->i_dtime),
Mark Fashehccd979b2005-12-15 14:31:24 -08002360 le32_to_cpu(fe->i_flags));
2361
2362 ocfs2_refresh_inode(inode, fe);
Sunil Mushran8ddb7b02008-05-13 13:45:15 -07002363 ocfs2_track_lock_refresh(lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08002364 }
2365
2366 status = 0;
2367bail_refresh:
Mark Fashehbe9e9862007-04-18 15:22:08 -07002368 ocfs2_complete_lock_res_refresh(lockres, status);
Mark Fashehccd979b2005-12-15 14:31:24 -08002369bail:
Mark Fashehccd979b2005-12-15 14:31:24 -08002370 return status;
2371}
2372
2373static int ocfs2_assign_bh(struct inode *inode,
2374 struct buffer_head **ret_bh,
2375 struct buffer_head *passed_bh)
2376{
2377 int status;
2378
2379 if (passed_bh) {
2380 /* Ok, the update went to disk for us, use the
2381 * returned bh. */
2382 *ret_bh = passed_bh;
2383 get_bh(*ret_bh);
2384
2385 return 0;
2386 }
2387
Joel Beckerb657c952008-11-13 14:49:11 -08002388 status = ocfs2_read_inode_block(inode, ret_bh);
Mark Fashehccd979b2005-12-15 14:31:24 -08002389 if (status < 0)
2390 mlog_errno(status);
2391
2392 return status;
2393}
2394
2395/*
2396 * returns < 0 error if the callback will never be called, otherwise
2397 * the result of the lock will be communicated via the callback.
2398 */
Jan Karacb257972009-06-04 15:26:50 +02002399int ocfs2_inode_lock_full_nested(struct inode *inode,
2400 struct buffer_head **ret_bh,
2401 int ex,
2402 int arg_flags,
2403 int subclass)
Mark Fashehccd979b2005-12-15 14:31:24 -08002404{
Joel Beckerbd3e7612008-02-01 12:14:57 -08002405 int status, level, acquired;
2406 u32 dlm_flags;
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002407 struct ocfs2_lock_res *lockres = NULL;
Mark Fashehccd979b2005-12-15 14:31:24 -08002408 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
2409 struct buffer_head *local_bh = NULL;
2410
Mark Fashehb06970532006-03-03 10:24:33 -08002411 mlog(0, "inode %llu, take %s META lock\n",
2412 (unsigned long long)OCFS2_I(inode)->ip_blkno,
Mark Fashehccd979b2005-12-15 14:31:24 -08002413 ex ? "EXMODE" : "PRMODE");
2414
2415 status = 0;
2416 acquired = 0;
2417 /* We'll allow faking a readonly metadata lock for
2418 * rodevices. */
2419 if (ocfs2_is_hard_readonly(osb)) {
2420 if (ex)
2421 status = -EROFS;
Tiger Yang03efed82011-05-28 00:34:19 +08002422 goto getbh;
Mark Fashehccd979b2005-12-15 14:31:24 -08002423 }
2424
Eric Ren439a36b2017-02-22 15:40:41 -08002425 if ((arg_flags & OCFS2_META_LOCK_GETBH) ||
2426 ocfs2_mount_local(osb))
2427 goto update;
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002428
Mark Fashehccd979b2005-12-15 14:31:24 -08002429 if (!(arg_flags & OCFS2_META_LOCK_RECOVERY))
Joel Becker553abd02008-02-01 12:03:57 -08002430 ocfs2_wait_for_recovery(osb);
Mark Fashehccd979b2005-12-15 14:31:24 -08002431
Mark Fashehe63aecb62007-10-18 15:30:42 -07002432 lockres = &OCFS2_I(inode)->ip_inode_lockres;
Joel Beckerbd3e7612008-02-01 12:14:57 -08002433 level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
Mark Fashehccd979b2005-12-15 14:31:24 -08002434 dlm_flags = 0;
2435 if (arg_flags & OCFS2_META_LOCK_NOQUEUE)
Joel Beckerbd3e7612008-02-01 12:14:57 -08002436 dlm_flags |= DLM_LKF_NOQUEUE;
Mark Fashehccd979b2005-12-15 14:31:24 -08002437
Jan Karacb257972009-06-04 15:26:50 +02002438 status = __ocfs2_cluster_lock(osb, lockres, level, dlm_flags,
2439 arg_flags, subclass, _RET_IP_);
Mark Fashehccd979b2005-12-15 14:31:24 -08002440 if (status < 0) {
Zach Brown41003a72013-05-07 16:18:25 -07002441 if (status != -EAGAIN)
Mark Fashehccd979b2005-12-15 14:31:24 -08002442 mlog_errno(status);
2443 goto bail;
2444 }
2445
2446 /* Notify the error cleanup path to drop the cluster lock. */
2447 acquired = 1;
2448
2449 /* We wait twice because a node may have died while we were in
2450 * the lower dlm layers. The second time though, we've
2451 * committed to owning this lock so we don't allow signals to
2452 * abort the operation. */
2453 if (!(arg_flags & OCFS2_META_LOCK_RECOVERY))
Joel Becker553abd02008-02-01 12:03:57 -08002454 ocfs2_wait_for_recovery(osb);
Mark Fashehccd979b2005-12-15 14:31:24 -08002455
Eric Ren439a36b2017-02-22 15:40:41 -08002456update:
Mark Fasheh24c19ef2006-09-22 17:28:19 -07002457 /*
2458 * We only see this flag if we're being called from
2459 * ocfs2_read_locked_inode(). It means we're locking an inode
2460 * which hasn't been populated yet, so clear the refresh flag
2461 * and let the caller handle it.
2462 */
2463 if (inode->i_state & I_NEW) {
2464 status = 0;
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002465 if (lockres)
2466 ocfs2_complete_lock_res_refresh(lockres, 0);
Mark Fasheh24c19ef2006-09-22 17:28:19 -07002467 goto bail;
2468 }
2469
Mark Fashehccd979b2005-12-15 14:31:24 -08002470 /* This is fun. The caller may want a bh back, or it may
Mark Fashehe63aecb62007-10-18 15:30:42 -07002471 * not. ocfs2_inode_lock_update definitely wants one in, but
Mark Fashehccd979b2005-12-15 14:31:24 -08002472 * may or may not read one, depending on what's in the
2473 * LVB. The result of all of this is that we've *only* gone to
2474 * disk if we have to, so the complexity is worthwhile. */
Mark Fashehe63aecb62007-10-18 15:30:42 -07002475 status = ocfs2_inode_lock_update(inode, &local_bh);
Mark Fashehccd979b2005-12-15 14:31:24 -08002476 if (status < 0) {
2477 if (status != -ENOENT)
2478 mlog_errno(status);
2479 goto bail;
2480 }
Tiger Yang03efed82011-05-28 00:34:19 +08002481getbh:
Mark Fashehccd979b2005-12-15 14:31:24 -08002482 if (ret_bh) {
2483 status = ocfs2_assign_bh(inode, ret_bh, local_bh);
2484 if (status < 0) {
2485 mlog_errno(status);
2486 goto bail;
2487 }
2488 }
2489
Mark Fashehccd979b2005-12-15 14:31:24 -08002490bail:
2491 if (status < 0) {
2492 if (ret_bh && (*ret_bh)) {
2493 brelse(*ret_bh);
2494 *ret_bh = NULL;
2495 }
2496 if (acquired)
Mark Fashehe63aecb62007-10-18 15:30:42 -07002497 ocfs2_inode_unlock(inode, ex);
Mark Fashehccd979b2005-12-15 14:31:24 -08002498 }
2499
2500 if (local_bh)
2501 brelse(local_bh);
2502
Mark Fashehccd979b2005-12-15 14:31:24 -08002503 return status;
2504}
2505
2506/*
Mark Fasheh34d024f2007-09-24 15:56:19 -07002507 * This is working around a lock inversion between tasks acquiring DLM
2508 * locks while holding a page lock and the downconvert thread which
2509 * blocks dlm lock acquiry while acquiring page locks.
Mark Fashehccd979b2005-12-15 14:31:24 -08002510 *
2511 * ** These _with_page variantes are only intended to be called from aop
2512 * methods that hold page locks and return a very specific *positive* error
2513 * code that aop methods pass up to the VFS -- test for errors with != 0. **
2514 *
Mark Fasheh34d024f2007-09-24 15:56:19 -07002515 * The DLM is called such that it returns -EAGAIN if it would have
2516 * blocked waiting for the downconvert thread. In that case we unlock
2517 * our page so the downconvert thread can make progress. Once we've
2518 * done this we have to return AOP_TRUNCATED_PAGE so the aop method
2519 * that called us can bubble that back up into the VFS who will then
2520 * immediately retry the aop call.
Mark Fashehccd979b2005-12-15 14:31:24 -08002521 */
Mark Fashehe63aecb62007-10-18 15:30:42 -07002522int ocfs2_inode_lock_with_page(struct inode *inode,
Mark Fashehccd979b2005-12-15 14:31:24 -08002523 struct buffer_head **ret_bh,
2524 int ex,
2525 struct page *page)
2526{
2527 int ret;
2528
Mark Fashehe63aecb62007-10-18 15:30:42 -07002529 ret = ocfs2_inode_lock_full(inode, ret_bh, ex, OCFS2_LOCK_NONBLOCK);
Mark Fashehccd979b2005-12-15 14:31:24 -08002530 if (ret == -EAGAIN) {
2531 unlock_page(page);
Gang Heff26cc12018-01-31 16:14:48 -08002532 /*
2533 * If we can't get inode lock immediately, we should not return
2534 * directly here, since this will lead to a softlockup problem.
2535 * The method is to get a blocking lock and immediately unlock
2536 * before returning, this can avoid CPU resource waste due to
2537 * lots of retries, and benefits fairness in getting lock.
2538 */
2539 if (ocfs2_inode_lock(inode, ret_bh, ex) == 0)
2540 ocfs2_inode_unlock(inode, ex);
Mark Fashehccd979b2005-12-15 14:31:24 -08002541 ret = AOP_TRUNCATED_PAGE;
2542 }
2543
2544 return ret;
2545}
2546
Mark Fashehe63aecb62007-10-18 15:30:42 -07002547int ocfs2_inode_lock_atime(struct inode *inode,
Tiger Yang7f1a37e2006-11-15 15:48:42 +08002548 struct vfsmount *vfsmnt,
Gang Hec4c24162018-01-31 16:15:25 -08002549 int *level, int wait)
Tiger Yang7f1a37e2006-11-15 15:48:42 +08002550{
2551 int ret;
2552
Gang Hec4c24162018-01-31 16:15:25 -08002553 if (wait)
2554 ret = ocfs2_inode_lock(inode, NULL, 0);
2555 else
2556 ret = ocfs2_try_inode_lock(inode, NULL, 0);
2557
Tiger Yang7f1a37e2006-11-15 15:48:42 +08002558 if (ret < 0) {
Gang Hec4c24162018-01-31 16:15:25 -08002559 if (ret != -EAGAIN)
2560 mlog_errno(ret);
Tiger Yang7f1a37e2006-11-15 15:48:42 +08002561 return ret;
2562 }
2563
2564 /*
2565 * If we should update atime, we will get EX lock,
2566 * otherwise we just get PR lock.
2567 */
2568 if (ocfs2_should_update_atime(inode, vfsmnt)) {
2569 struct buffer_head *bh = NULL;
2570
Mark Fashehe63aecb62007-10-18 15:30:42 -07002571 ocfs2_inode_unlock(inode, 0);
Gang Hec4c24162018-01-31 16:15:25 -08002572 if (wait)
2573 ret = ocfs2_inode_lock(inode, &bh, 1);
2574 else
2575 ret = ocfs2_try_inode_lock(inode, &bh, 1);
2576
Tiger Yang7f1a37e2006-11-15 15:48:42 +08002577 if (ret < 0) {
Gang Hec4c24162018-01-31 16:15:25 -08002578 if (ret != -EAGAIN)
2579 mlog_errno(ret);
Tiger Yang7f1a37e2006-11-15 15:48:42 +08002580 return ret;
2581 }
2582 *level = 1;
2583 if (ocfs2_should_update_atime(inode, vfsmnt))
2584 ocfs2_update_inode_atime(inode, bh);
2585 if (bh)
2586 brelse(bh);
2587 } else
2588 *level = 0;
2589
Tiger Yang7f1a37e2006-11-15 15:48:42 +08002590 return ret;
2591}
2592
Mark Fashehe63aecb62007-10-18 15:30:42 -07002593void ocfs2_inode_unlock(struct inode *inode,
Mark Fashehccd979b2005-12-15 14:31:24 -08002594 int ex)
2595{
Joel Beckerbd3e7612008-02-01 12:14:57 -08002596 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
Mark Fashehe63aecb62007-10-18 15:30:42 -07002597 struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_inode_lockres;
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002598 struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
Mark Fashehccd979b2005-12-15 14:31:24 -08002599
Mark Fashehb06970532006-03-03 10:24:33 -08002600 mlog(0, "inode %llu drop %s META lock\n",
2601 (unsigned long long)OCFS2_I(inode)->ip_blkno,
Mark Fashehccd979b2005-12-15 14:31:24 -08002602 ex ? "EXMODE" : "PRMODE");
2603
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002604 if (!ocfs2_is_hard_readonly(OCFS2_SB(inode->i_sb)) &&
2605 !ocfs2_mount_local(osb))
Mark Fashehccd979b2005-12-15 14:31:24 -08002606 ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, level);
Mark Fashehccd979b2005-12-15 14:31:24 -08002607}
2608
Eric Ren439a36b2017-02-22 15:40:41 -08002609/*
2610 * This _tracker variantes are introduced to deal with the recursive cluster
2611 * locking issue. The idea is to keep track of a lock holder on the stack of
2612 * the current process. If there's a lock holder on the stack, we know the
2613 * task context is already protected by cluster locking. Currently, they're
2614 * used in some VFS entry routines.
2615 *
2616 * return < 0 on error, return == 0 if there's no lock holder on the stack
2617 * before this call, return == 1 if this call would be a recursive locking.
2618 */
2619int ocfs2_inode_lock_tracker(struct inode *inode,
2620 struct buffer_head **ret_bh,
2621 int ex,
2622 struct ocfs2_lock_holder *oh)
2623{
2624 int status;
2625 int arg_flags = 0, has_locked;
2626 struct ocfs2_lock_res *lockres;
2627
2628 lockres = &OCFS2_I(inode)->ip_inode_lockres;
2629 has_locked = ocfs2_is_locked_by_me(lockres);
2630 /* Just get buffer head if the cluster lock has been taken */
2631 if (has_locked)
2632 arg_flags = OCFS2_META_LOCK_GETBH;
2633
2634 if (likely(!has_locked || ret_bh)) {
2635 status = ocfs2_inode_lock_full(inode, ret_bh, ex, arg_flags);
2636 if (status < 0) {
2637 if (status != -ENOENT)
2638 mlog_errno(status);
2639 return status;
2640 }
2641 }
2642 if (!has_locked)
2643 ocfs2_add_holder(lockres, oh);
2644
2645 return has_locked;
2646}
2647
2648void ocfs2_inode_unlock_tracker(struct inode *inode,
2649 int ex,
2650 struct ocfs2_lock_holder *oh,
2651 int had_lock)
2652{
2653 struct ocfs2_lock_res *lockres;
2654
2655 lockres = &OCFS2_I(inode)->ip_inode_lockres;
Eric Ren8818efa2017-06-23 15:08:55 -07002656 /* had_lock means that the currect process already takes the cluster
2657 * lock previously. If had_lock is 1, we have nothing to do here, and
2658 * it will get unlocked where we got the lock.
2659 */
Eric Ren439a36b2017-02-22 15:40:41 -08002660 if (!had_lock) {
2661 ocfs2_remove_holder(lockres, oh);
2662 ocfs2_inode_unlock(inode, ex);
2663 }
2664}
2665
Sunil Mushrandf152c22009-06-22 11:40:07 -07002666int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno)
Srinivas Eeda83273932009-06-03 17:02:55 -07002667{
2668 struct ocfs2_lock_res *lockres;
2669 struct ocfs2_orphan_scan_lvb *lvb;
Srinivas Eeda83273932009-06-03 17:02:55 -07002670 int status = 0;
2671
Sunil Mushrandf152c22009-06-22 11:40:07 -07002672 if (ocfs2_is_hard_readonly(osb))
2673 return -EROFS;
2674
2675 if (ocfs2_mount_local(osb))
2676 return 0;
2677
Srinivas Eeda83273932009-06-03 17:02:55 -07002678 lockres = &osb->osb_orphan_scan.os_lockres;
Sunil Mushrandf152c22009-06-22 11:40:07 -07002679 status = ocfs2_cluster_lock(osb, lockres, DLM_LOCK_EX, 0, 0);
Srinivas Eeda83273932009-06-03 17:02:55 -07002680 if (status < 0)
2681 return status;
2682
2683 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
Joel Becker1c520df2009-06-19 15:14:13 -07002684 if (ocfs2_dlm_lvb_valid(&lockres->l_lksb) &&
2685 lvb->lvb_version == OCFS2_ORPHAN_LVB_VERSION)
Srinivas Eeda83273932009-06-03 17:02:55 -07002686 *seqno = be32_to_cpu(lvb->lvb_os_seqno);
Sunil Mushran32119492009-06-19 16:53:18 -07002687 else
2688 *seqno = osb->osb_orphan_scan.os_seqno + 1;
2689
Srinivas Eeda83273932009-06-03 17:02:55 -07002690 return status;
2691}
2692
Sunil Mushrandf152c22009-06-22 11:40:07 -07002693void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno)
Srinivas Eeda83273932009-06-03 17:02:55 -07002694{
2695 struct ocfs2_lock_res *lockres;
2696 struct ocfs2_orphan_scan_lvb *lvb;
Srinivas Eeda83273932009-06-03 17:02:55 -07002697
Sunil Mushrandf152c22009-06-22 11:40:07 -07002698 if (!ocfs2_is_hard_readonly(osb) && !ocfs2_mount_local(osb)) {
2699 lockres = &osb->osb_orphan_scan.os_lockres;
2700 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
2701 lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
2702 lvb->lvb_os_seqno = cpu_to_be32(seqno);
2703 ocfs2_cluster_unlock(osb, lockres, DLM_LOCK_EX);
2704 }
Srinivas Eeda83273932009-06-03 17:02:55 -07002705}
2706
Mark Fashehccd979b2005-12-15 14:31:24 -08002707int ocfs2_super_lock(struct ocfs2_super *osb,
2708 int ex)
2709{
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002710 int status = 0;
Joel Beckerbd3e7612008-02-01 12:14:57 -08002711 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
Mark Fashehccd979b2005-12-15 14:31:24 -08002712 struct ocfs2_lock_res *lockres = &osb->osb_super_lockres;
Mark Fashehccd979b2005-12-15 14:31:24 -08002713
Mark Fashehccd979b2005-12-15 14:31:24 -08002714 if (ocfs2_is_hard_readonly(osb))
2715 return -EROFS;
2716
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002717 if (ocfs2_mount_local(osb))
2718 goto bail;
2719
Mark Fashehccd979b2005-12-15 14:31:24 -08002720 status = ocfs2_cluster_lock(osb, lockres, level, 0, 0);
2721 if (status < 0) {
2722 mlog_errno(status);
2723 goto bail;
2724 }
2725
2726 /* The super block lock path is really in the best position to
2727 * know when resources covered by the lock need to be
2728 * refreshed, so we do it here. Of course, making sense of
2729 * everything is up to the caller :) */
2730 status = ocfs2_should_refresh_lock_res(lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08002731 if (status) {
Mark Fasheh8e8a4602008-02-01 11:59:09 -08002732 status = ocfs2_refresh_slot_info(osb);
Mark Fashehccd979b2005-12-15 14:31:24 -08002733
2734 ocfs2_complete_lock_res_refresh(lockres, status);
2735
Junxiao Bi3278bb72013-02-21 16:42:45 -08002736 if (status < 0) {
2737 ocfs2_cluster_unlock(osb, lockres, level);
Mark Fashehccd979b2005-12-15 14:31:24 -08002738 mlog_errno(status);
Junxiao Bi3278bb72013-02-21 16:42:45 -08002739 }
Sunil Mushran8ddb7b02008-05-13 13:45:15 -07002740 ocfs2_track_lock_refresh(lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08002741 }
2742bail:
Mark Fashehccd979b2005-12-15 14:31:24 -08002743 return status;
2744}
2745
2746void ocfs2_super_unlock(struct ocfs2_super *osb,
2747 int ex)
2748{
Joel Beckerbd3e7612008-02-01 12:14:57 -08002749 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
Mark Fashehccd979b2005-12-15 14:31:24 -08002750 struct ocfs2_lock_res *lockres = &osb->osb_super_lockres;
2751
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002752 if (!ocfs2_mount_local(osb))
2753 ocfs2_cluster_unlock(osb, lockres, level);
Mark Fashehccd979b2005-12-15 14:31:24 -08002754}
2755
2756int ocfs2_rename_lock(struct ocfs2_super *osb)
2757{
2758 int status;
2759 struct ocfs2_lock_res *lockres = &osb->osb_rename_lockres;
2760
2761 if (ocfs2_is_hard_readonly(osb))
2762 return -EROFS;
2763
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002764 if (ocfs2_mount_local(osb))
2765 return 0;
2766
Joel Beckerbd3e7612008-02-01 12:14:57 -08002767 status = ocfs2_cluster_lock(osb, lockres, DLM_LOCK_EX, 0, 0);
Mark Fashehccd979b2005-12-15 14:31:24 -08002768 if (status < 0)
2769 mlog_errno(status);
2770
2771 return status;
2772}
2773
2774void ocfs2_rename_unlock(struct ocfs2_super *osb)
2775{
2776 struct ocfs2_lock_res *lockres = &osb->osb_rename_lockres;
2777
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002778 if (!ocfs2_mount_local(osb))
Joel Beckerbd3e7612008-02-01 12:14:57 -08002779 ocfs2_cluster_unlock(osb, lockres, DLM_LOCK_EX);
Mark Fashehccd979b2005-12-15 14:31:24 -08002780}
2781
wengang wang6ca497a2009-03-06 21:29:10 +08002782int ocfs2_nfs_sync_lock(struct ocfs2_super *osb, int ex)
2783{
2784 int status;
2785 struct ocfs2_lock_res *lockres = &osb->osb_nfs_sync_lockres;
2786
2787 if (ocfs2_is_hard_readonly(osb))
2788 return -EROFS;
2789
2790 if (ocfs2_mount_local(osb))
2791 return 0;
2792
2793 status = ocfs2_cluster_lock(osb, lockres, ex ? LKM_EXMODE : LKM_PRMODE,
2794 0, 0);
2795 if (status < 0)
2796 mlog(ML_ERROR, "lock on nfs sync lock failed %d\n", status);
2797
2798 return status;
2799}
2800
2801void ocfs2_nfs_sync_unlock(struct ocfs2_super *osb, int ex)
2802{
2803 struct ocfs2_lock_res *lockres = &osb->osb_nfs_sync_lockres;
2804
2805 if (!ocfs2_mount_local(osb))
2806 ocfs2_cluster_unlock(osb, lockres,
2807 ex ? LKM_EXMODE : LKM_PRMODE);
2808}
2809
Gang He4882abe2018-01-31 16:15:10 -08002810int ocfs2_trim_fs_lock(struct ocfs2_super *osb,
2811 struct ocfs2_trim_fs_info *info, int trylock)
2812{
2813 int status;
2814 struct ocfs2_trim_fs_lvb *lvb;
2815 struct ocfs2_lock_res *lockres = &osb->osb_trim_fs_lockres;
2816
2817 if (info)
2818 info->tf_valid = 0;
2819
2820 if (ocfs2_is_hard_readonly(osb))
2821 return -EROFS;
2822
2823 if (ocfs2_mount_local(osb))
2824 return 0;
2825
2826 status = ocfs2_cluster_lock(osb, lockres, DLM_LOCK_EX,
2827 trylock ? DLM_LKF_NOQUEUE : 0, 0);
2828 if (status < 0) {
2829 if (status != -EAGAIN)
2830 mlog_errno(status);
2831 return status;
2832 }
2833
2834 if (info) {
2835 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
2836 if (ocfs2_dlm_lvb_valid(&lockres->l_lksb) &&
2837 lvb->lvb_version == OCFS2_TRIMFS_LVB_VERSION) {
2838 info->tf_valid = 1;
2839 info->tf_success = lvb->lvb_success;
2840 info->tf_nodenum = be32_to_cpu(lvb->lvb_nodenum);
2841 info->tf_start = be64_to_cpu(lvb->lvb_start);
2842 info->tf_len = be64_to_cpu(lvb->lvb_len);
2843 info->tf_minlen = be64_to_cpu(lvb->lvb_minlen);
2844 info->tf_trimlen = be64_to_cpu(lvb->lvb_trimlen);
2845 }
2846 }
2847
2848 return status;
2849}
2850
2851void ocfs2_trim_fs_unlock(struct ocfs2_super *osb,
2852 struct ocfs2_trim_fs_info *info)
2853{
2854 struct ocfs2_trim_fs_lvb *lvb;
2855 struct ocfs2_lock_res *lockres = &osb->osb_trim_fs_lockres;
2856
2857 if (ocfs2_mount_local(osb))
2858 return;
2859
2860 if (info) {
2861 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
2862 lvb->lvb_version = OCFS2_TRIMFS_LVB_VERSION;
2863 lvb->lvb_success = info->tf_success;
2864 lvb->lvb_nodenum = cpu_to_be32(info->tf_nodenum);
2865 lvb->lvb_start = cpu_to_be64(info->tf_start);
2866 lvb->lvb_len = cpu_to_be64(info->tf_len);
2867 lvb->lvb_minlen = cpu_to_be64(info->tf_minlen);
2868 lvb->lvb_trimlen = cpu_to_be64(info->tf_trimlen);
2869 }
2870
2871 ocfs2_cluster_unlock(osb, lockres, DLM_LOCK_EX);
2872}
2873
Mark Fashehd680efe2006-09-08 14:14:34 -07002874int ocfs2_dentry_lock(struct dentry *dentry, int ex)
2875{
2876 int ret;
Joel Beckerbd3e7612008-02-01 12:14:57 -08002877 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
Mark Fashehd680efe2006-09-08 14:14:34 -07002878 struct ocfs2_dentry_lock *dl = dentry->d_fsdata;
2879 struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
2880
2881 BUG_ON(!dl);
2882
Tiger Yang03efed82011-05-28 00:34:19 +08002883 if (ocfs2_is_hard_readonly(osb)) {
2884 if (ex)
2885 return -EROFS;
2886 return 0;
2887 }
Mark Fashehd680efe2006-09-08 14:14:34 -07002888
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002889 if (ocfs2_mount_local(osb))
2890 return 0;
2891
Mark Fashehd680efe2006-09-08 14:14:34 -07002892 ret = ocfs2_cluster_lock(osb, &dl->dl_lockres, level, 0, 0);
2893 if (ret < 0)
2894 mlog_errno(ret);
2895
2896 return ret;
2897}
2898
2899void ocfs2_dentry_unlock(struct dentry *dentry, int ex)
2900{
Joel Beckerbd3e7612008-02-01 12:14:57 -08002901 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
Mark Fashehd680efe2006-09-08 14:14:34 -07002902 struct ocfs2_dentry_lock *dl = dentry->d_fsdata;
2903 struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
2904
Tiger Yang03efed82011-05-28 00:34:19 +08002905 if (!ocfs2_is_hard_readonly(osb) && !ocfs2_mount_local(osb))
Sunil Mushranc271c5c2006-12-05 17:56:35 -08002906 ocfs2_cluster_unlock(osb, &dl->dl_lockres, level);
Mark Fashehd680efe2006-09-08 14:14:34 -07002907}
2908
Mark Fashehccd979b2005-12-15 14:31:24 -08002909/* Reference counting of the dlm debug structure. We want this because
2910 * open references on the debug inodes can live on after a mount, so
2911 * we can't rely on the ocfs2_super to always exist. */
2912static void ocfs2_dlm_debug_free(struct kref *kref)
2913{
2914 struct ocfs2_dlm_debug *dlm_debug;
2915
2916 dlm_debug = container_of(kref, struct ocfs2_dlm_debug, d_refcnt);
2917
2918 kfree(dlm_debug);
2919}
2920
2921void ocfs2_put_dlm_debug(struct ocfs2_dlm_debug *dlm_debug)
2922{
2923 if (dlm_debug)
2924 kref_put(&dlm_debug->d_refcnt, ocfs2_dlm_debug_free);
2925}
2926
2927static void ocfs2_get_dlm_debug(struct ocfs2_dlm_debug *debug)
2928{
2929 kref_get(&debug->d_refcnt);
2930}
2931
2932struct ocfs2_dlm_debug *ocfs2_new_dlm_debug(void)
2933{
2934 struct ocfs2_dlm_debug *dlm_debug;
2935
2936 dlm_debug = kmalloc(sizeof(struct ocfs2_dlm_debug), GFP_KERNEL);
2937 if (!dlm_debug) {
2938 mlog_errno(-ENOMEM);
2939 goto out;
2940 }
2941
2942 kref_init(&dlm_debug->d_refcnt);
2943 INIT_LIST_HEAD(&dlm_debug->d_lockres_tracking);
2944 dlm_debug->d_locking_state = NULL;
2945out:
2946 return dlm_debug;
2947}
2948
2949/* Access to this is arbitrated for us via seq_file->sem. */
2950struct ocfs2_dlm_seq_priv {
2951 struct ocfs2_dlm_debug *p_dlm_debug;
2952 struct ocfs2_lock_res p_iter_res;
2953 struct ocfs2_lock_res p_tmp_res;
2954};
2955
2956static struct ocfs2_lock_res *ocfs2_dlm_next_res(struct ocfs2_lock_res *start,
2957 struct ocfs2_dlm_seq_priv *priv)
2958{
2959 struct ocfs2_lock_res *iter, *ret = NULL;
2960 struct ocfs2_dlm_debug *dlm_debug = priv->p_dlm_debug;
2961
2962 assert_spin_locked(&ocfs2_dlm_tracking_lock);
2963
2964 list_for_each_entry(iter, &start->l_debug_list, l_debug_list) {
2965 /* discover the head of the list */
2966 if (&iter->l_debug_list == &dlm_debug->d_lockres_tracking) {
2967 mlog(0, "End of list found, %p\n", ret);
2968 break;
2969 }
2970
2971 /* We track our "dummy" iteration lockres' by a NULL
2972 * l_ops field. */
2973 if (iter->l_ops != NULL) {
2974 ret = iter;
2975 break;
2976 }
2977 }
2978
2979 return ret;
2980}
2981
2982static void *ocfs2_dlm_seq_start(struct seq_file *m, loff_t *pos)
2983{
2984 struct ocfs2_dlm_seq_priv *priv = m->private;
2985 struct ocfs2_lock_res *iter;
2986
2987 spin_lock(&ocfs2_dlm_tracking_lock);
2988 iter = ocfs2_dlm_next_res(&priv->p_iter_res, priv);
2989 if (iter) {
2990 /* Since lockres' have the lifetime of their container
2991 * (which can be inodes, ocfs2_supers, etc) we want to
2992 * copy this out to a temporary lockres while still
2993 * under the spinlock. Obviously after this we can't
2994 * trust any pointers on the copy returned, but that's
2995 * ok as the information we want isn't typically held
2996 * in them. */
2997 priv->p_tmp_res = *iter;
2998 iter = &priv->p_tmp_res;
2999 }
3000 spin_unlock(&ocfs2_dlm_tracking_lock);
3001
3002 return iter;
3003}
3004
3005static void ocfs2_dlm_seq_stop(struct seq_file *m, void *v)
3006{
3007}
3008
3009static void *ocfs2_dlm_seq_next(struct seq_file *m, void *v, loff_t *pos)
3010{
3011 struct ocfs2_dlm_seq_priv *priv = m->private;
3012 struct ocfs2_lock_res *iter = v;
3013 struct ocfs2_lock_res *dummy = &priv->p_iter_res;
3014
3015 spin_lock(&ocfs2_dlm_tracking_lock);
3016 iter = ocfs2_dlm_next_res(iter, priv);
3017 list_del_init(&dummy->l_debug_list);
3018 if (iter) {
3019 list_add(&dummy->l_debug_list, &iter->l_debug_list);
3020 priv->p_tmp_res = *iter;
3021 iter = &priv->p_tmp_res;
3022 }
3023 spin_unlock(&ocfs2_dlm_tracking_lock);
3024
3025 return iter;
3026}
3027
Sunil Mushran5bc970e2010-12-28 23:26:03 -08003028/*
3029 * Version is used by debugfs.ocfs2 to determine the format being used
3030 *
3031 * New in version 2
3032 * - Lock stats printed
3033 * New in version 3
3034 * - Max time in lock stats is in usecs (instead of nsecs)
3035 */
3036#define OCFS2_DLM_DEBUG_STR_VERSION 3
Mark Fashehccd979b2005-12-15 14:31:24 -08003037static int ocfs2_dlm_seq_show(struct seq_file *m, void *v)
3038{
3039 int i;
3040 char *lvb;
3041 struct ocfs2_lock_res *lockres = v;
3042
3043 if (!lockres)
3044 return -EINVAL;
3045
Mark Fashehd680efe2006-09-08 14:14:34 -07003046 seq_printf(m, "0x%x\t", OCFS2_DLM_DEBUG_STR_VERSION);
3047
3048 if (lockres->l_type == OCFS2_LOCK_TYPE_DENTRY)
3049 seq_printf(m, "%.*s%08x\t", OCFS2_DENTRY_LOCK_INO_START - 1,
3050 lockres->l_name,
3051 (unsigned int)ocfs2_get_dentry_lock_ino(lockres));
3052 else
3053 seq_printf(m, "%.*s\t", OCFS2_LOCK_ID_MAX_LEN, lockres->l_name);
3054
3055 seq_printf(m, "%d\t"
Mark Fashehccd979b2005-12-15 14:31:24 -08003056 "0x%lx\t"
3057 "0x%x\t"
3058 "0x%x\t"
3059 "%u\t"
3060 "%u\t"
3061 "%d\t"
3062 "%d\t",
Mark Fashehccd979b2005-12-15 14:31:24 -08003063 lockres->l_level,
3064 lockres->l_flags,
3065 lockres->l_action,
3066 lockres->l_unlock_action,
3067 lockres->l_ro_holders,
3068 lockres->l_ex_holders,
3069 lockres->l_requested,
3070 lockres->l_blocking);
3071
3072 /* Dump the raw LVB */
Joel Becker8f2c9c12008-02-01 12:16:57 -08003073 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
Mark Fashehccd979b2005-12-15 14:31:24 -08003074 for(i = 0; i < DLM_LVB_LEN; i++)
3075 seq_printf(m, "0x%x\t", lvb[i]);
3076
Sunil Mushran8ddb7b02008-05-13 13:45:15 -07003077#ifdef CONFIG_OCFS2_FS_STATS
Sunil Mushran5bc970e2010-12-28 23:26:03 -08003078# define lock_num_prmode(_l) ((_l)->l_lock_prmode.ls_gets)
3079# define lock_num_exmode(_l) ((_l)->l_lock_exmode.ls_gets)
3080# define lock_num_prmode_failed(_l) ((_l)->l_lock_prmode.ls_fail)
3081# define lock_num_exmode_failed(_l) ((_l)->l_lock_exmode.ls_fail)
3082# define lock_total_prmode(_l) ((_l)->l_lock_prmode.ls_total)
3083# define lock_total_exmode(_l) ((_l)->l_lock_exmode.ls_total)
3084# define lock_max_prmode(_l) ((_l)->l_lock_prmode.ls_max)
3085# define lock_max_exmode(_l) ((_l)->l_lock_exmode.ls_max)
3086# define lock_refresh(_l) ((_l)->l_lock_refresh)
Sunil Mushran8ddb7b02008-05-13 13:45:15 -07003087#else
Sunil Mushran5bc970e2010-12-28 23:26:03 -08003088# define lock_num_prmode(_l) (0)
3089# define lock_num_exmode(_l) (0)
Sunil Mushran8ddb7b02008-05-13 13:45:15 -07003090# define lock_num_prmode_failed(_l) (0)
3091# define lock_num_exmode_failed(_l) (0)
Randy Dunlapdd25e552008-05-28 14:41:00 -07003092# define lock_total_prmode(_l) (0ULL)
3093# define lock_total_exmode(_l) (0ULL)
Sunil Mushran8ddb7b02008-05-13 13:45:15 -07003094# define lock_max_prmode(_l) (0)
3095# define lock_max_exmode(_l) (0)
3096# define lock_refresh(_l) (0)
3097#endif
3098 /* The following seq_print was added in version 2 of this output */
Sunil Mushran5bc970e2010-12-28 23:26:03 -08003099 seq_printf(m, "%u\t"
3100 "%u\t"
Sunil Mushran8ddb7b02008-05-13 13:45:15 -07003101 "%u\t"
3102 "%u\t"
3103 "%llu\t"
3104 "%llu\t"
3105 "%u\t"
3106 "%u\t"
3107 "%u\t",
3108 lock_num_prmode(lockres),
3109 lock_num_exmode(lockres),
3110 lock_num_prmode_failed(lockres),
3111 lock_num_exmode_failed(lockres),
3112 lock_total_prmode(lockres),
3113 lock_total_exmode(lockres),
3114 lock_max_prmode(lockres),
3115 lock_max_exmode(lockres),
3116 lock_refresh(lockres));
3117
Mark Fashehccd979b2005-12-15 14:31:24 -08003118 /* End the line */
3119 seq_printf(m, "\n");
3120 return 0;
3121}
3122
Jan Engelhardt90d99772008-01-22 20:52:20 +01003123static const struct seq_operations ocfs2_dlm_seq_ops = {
Mark Fashehccd979b2005-12-15 14:31:24 -08003124 .start = ocfs2_dlm_seq_start,
3125 .stop = ocfs2_dlm_seq_stop,
3126 .next = ocfs2_dlm_seq_next,
3127 .show = ocfs2_dlm_seq_show,
3128};
3129
3130static int ocfs2_dlm_debug_release(struct inode *inode, struct file *file)
3131{
Joe Perches33fa1d92010-07-12 13:50:19 -07003132 struct seq_file *seq = file->private_data;
Mark Fashehccd979b2005-12-15 14:31:24 -08003133 struct ocfs2_dlm_seq_priv *priv = seq->private;
3134 struct ocfs2_lock_res *res = &priv->p_iter_res;
3135
3136 ocfs2_remove_lockres_tracking(res);
3137 ocfs2_put_dlm_debug(priv->p_dlm_debug);
3138 return seq_release_private(inode, file);
3139}
3140
3141static int ocfs2_dlm_debug_open(struct inode *inode, struct file *file)
3142{
Mark Fashehccd979b2005-12-15 14:31:24 -08003143 struct ocfs2_dlm_seq_priv *priv;
Mark Fashehccd979b2005-12-15 14:31:24 -08003144 struct ocfs2_super *osb;
3145
Rob Jones1848cb552014-10-09 15:25:09 -07003146 priv = __seq_open_private(file, &ocfs2_dlm_seq_ops, sizeof(*priv));
Mark Fashehccd979b2005-12-15 14:31:24 -08003147 if (!priv) {
Rob Jones1848cb552014-10-09 15:25:09 -07003148 mlog_errno(-ENOMEM);
3149 return -ENOMEM;
Mark Fashehccd979b2005-12-15 14:31:24 -08003150 }
Rob Jones1848cb552014-10-09 15:25:09 -07003151
Theodore Ts'o8e18e292006-09-27 01:50:46 -07003152 osb = inode->i_private;
Mark Fashehccd979b2005-12-15 14:31:24 -08003153 ocfs2_get_dlm_debug(osb->osb_dlm_debug);
3154 priv->p_dlm_debug = osb->osb_dlm_debug;
3155 INIT_LIST_HEAD(&priv->p_iter_res.l_debug_list);
3156
Mark Fashehccd979b2005-12-15 14:31:24 -08003157 ocfs2_add_lockres_tracking(&priv->p_iter_res,
3158 priv->p_dlm_debug);
3159
Rob Jones1848cb552014-10-09 15:25:09 -07003160 return 0;
Mark Fashehccd979b2005-12-15 14:31:24 -08003161}
3162
Arjan van de Ven4b6f5d22006-03-28 01:56:42 -08003163static const struct file_operations ocfs2_dlm_debug_fops = {
Mark Fashehccd979b2005-12-15 14:31:24 -08003164 .open = ocfs2_dlm_debug_open,
3165 .release = ocfs2_dlm_debug_release,
3166 .read = seq_read,
3167 .llseek = seq_lseek,
3168};
3169
3170static int ocfs2_dlm_init_debug(struct ocfs2_super *osb)
3171{
3172 int ret = 0;
3173 struct ocfs2_dlm_debug *dlm_debug = osb->osb_dlm_debug;
3174
3175 dlm_debug->d_locking_state = debugfs_create_file("locking_state",
3176 S_IFREG|S_IRUSR,
3177 osb->osb_debug_root,
3178 osb,
3179 &ocfs2_dlm_debug_fops);
Linus Torvalds8f443e22015-04-21 09:17:28 -07003180 if (!dlm_debug->d_locking_state) {
Mark Fashehccd979b2005-12-15 14:31:24 -08003181 ret = -EINVAL;
3182 mlog(ML_ERROR,
3183 "Unable to create locking state debugfs file.\n");
3184 goto out;
3185 }
3186
3187 ocfs2_get_dlm_debug(dlm_debug);
3188out:
3189 return ret;
3190}
3191
3192static void ocfs2_dlm_shutdown_debug(struct ocfs2_super *osb)
3193{
3194 struct ocfs2_dlm_debug *dlm_debug = osb->osb_dlm_debug;
3195
3196 if (dlm_debug) {
3197 debugfs_remove(dlm_debug->d_locking_state);
3198 ocfs2_put_dlm_debug(dlm_debug);
3199 }
3200}
3201
3202int ocfs2_dlm_init(struct ocfs2_super *osb)
3203{
Sunil Mushranc271c5c2006-12-05 17:56:35 -08003204 int status = 0;
Joel Becker4670c462008-02-01 14:39:35 -08003205 struct ocfs2_cluster_connection *conn = NULL;
Mark Fashehccd979b2005-12-15 14:31:24 -08003206
Mark Fasheh0abd6d12008-01-29 16:59:56 -08003207 if (ocfs2_mount_local(osb)) {
3208 osb->node_num = 0;
Sunil Mushranc271c5c2006-12-05 17:56:35 -08003209 goto local;
Mark Fasheh0abd6d12008-01-29 16:59:56 -08003210 }
Sunil Mushranc271c5c2006-12-05 17:56:35 -08003211
Mark Fashehccd979b2005-12-15 14:31:24 -08003212 status = ocfs2_dlm_init_debug(osb);
3213 if (status < 0) {
3214 mlog_errno(status);
3215 goto bail;
3216 }
3217
Mark Fasheh34d024f2007-09-24 15:56:19 -07003218 /* launch downconvert thread */
Joseph Qi5afc44e2015-11-05 18:44:13 -08003219 osb->dc_task = kthread_run(ocfs2_downconvert_thread, osb, "ocfs2dc-%s",
3220 osb->uuid_str);
Mark Fasheh34d024f2007-09-24 15:56:19 -07003221 if (IS_ERR(osb->dc_task)) {
3222 status = PTR_ERR(osb->dc_task);
3223 osb->dc_task = NULL;
Mark Fashehccd979b2005-12-15 14:31:24 -08003224 mlog_errno(status);
3225 goto bail;
3226 }
3227
Mark Fashehccd979b2005-12-15 14:31:24 -08003228 /* for now, uuid == domain */
Joel Becker9c6c8772008-02-01 15:17:30 -08003229 status = ocfs2_cluster_connect(osb->osb_cluster_stack,
Goldwyn Rodriguesc74a3bd2014-01-21 15:48:21 -08003230 osb->osb_cluster_name,
3231 strlen(osb->osb_cluster_name),
Joel Becker9c6c8772008-02-01 15:17:30 -08003232 osb->uuid_str,
Joel Becker4670c462008-02-01 14:39:35 -08003233 strlen(osb->uuid_str),
Joel Becker553b5eb2010-01-29 17:19:06 -08003234 &lproto, ocfs2_do_node_down, osb,
Joel Becker4670c462008-02-01 14:39:35 -08003235 &conn);
3236 if (status) {
Mark Fashehccd979b2005-12-15 14:31:24 -08003237 mlog_errno(status);
3238 goto bail;
3239 }
3240
Goldwyn Rodrigues3e834152014-01-21 15:48:24 -08003241 status = ocfs2_cluster_this_node(conn, &osb->node_num);
Mark Fasheh0abd6d12008-01-29 16:59:56 -08003242 if (status < 0) {
3243 mlog_errno(status);
3244 mlog(ML_ERROR,
3245 "could not find this host's node number\n");
Joel Becker286eaa92008-02-01 15:03:57 -08003246 ocfs2_cluster_disconnect(conn, 0);
Mark Fasheh0abd6d12008-01-29 16:59:56 -08003247 goto bail;
3248 }
3249
Sunil Mushranc271c5c2006-12-05 17:56:35 -08003250local:
Mark Fashehccd979b2005-12-15 14:31:24 -08003251 ocfs2_super_lock_res_init(&osb->osb_super_lockres, osb);
3252 ocfs2_rename_lock_res_init(&osb->osb_rename_lockres, osb);
wengang wang6ca497a2009-03-06 21:29:10 +08003253 ocfs2_nfs_sync_lock_res_init(&osb->osb_nfs_sync_lockres, osb);
Srinivas Eeda83273932009-06-03 17:02:55 -07003254 ocfs2_orphan_scan_lock_res_init(&osb->osb_orphan_scan.os_lockres, osb);
Mark Fashehccd979b2005-12-15 14:31:24 -08003255
Joel Becker4670c462008-02-01 14:39:35 -08003256 osb->cconn = conn;
Mark Fashehccd979b2005-12-15 14:31:24 -08003257bail:
3258 if (status < 0) {
3259 ocfs2_dlm_shutdown_debug(osb);
Mark Fasheh34d024f2007-09-24 15:56:19 -07003260 if (osb->dc_task)
3261 kthread_stop(osb->dc_task);
Mark Fashehccd979b2005-12-15 14:31:24 -08003262 }
3263
Mark Fashehccd979b2005-12-15 14:31:24 -08003264 return status;
3265}
3266
Joel Becker286eaa92008-02-01 15:03:57 -08003267void ocfs2_dlm_shutdown(struct ocfs2_super *osb,
3268 int hangup_pending)
Mark Fashehccd979b2005-12-15 14:31:24 -08003269{
Mark Fashehccd979b2005-12-15 14:31:24 -08003270 ocfs2_drop_osb_locks(osb);
3271
Joel Becker4670c462008-02-01 14:39:35 -08003272 /*
3273 * Now that we have dropped all locks and ocfs2_dismount_volume()
3274 * has disabled recovery, the DLM won't be talking to us. It's
3275 * safe to tear things down before disconnecting the cluster.
3276 */
3277
Mark Fasheh34d024f2007-09-24 15:56:19 -07003278 if (osb->dc_task) {
3279 kthread_stop(osb->dc_task);
3280 osb->dc_task = NULL;
Mark Fashehccd979b2005-12-15 14:31:24 -08003281 }
3282
3283 ocfs2_lock_res_free(&osb->osb_super_lockres);
3284 ocfs2_lock_res_free(&osb->osb_rename_lockres);
wengang wang6ca497a2009-03-06 21:29:10 +08003285 ocfs2_lock_res_free(&osb->osb_nfs_sync_lockres);
Srinivas Eeda83273932009-06-03 17:02:55 -07003286 ocfs2_lock_res_free(&osb->osb_orphan_scan.os_lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08003287
Joel Becker286eaa92008-02-01 15:03:57 -08003288 ocfs2_cluster_disconnect(osb->cconn, hangup_pending);
Joel Becker4670c462008-02-01 14:39:35 -08003289 osb->cconn = NULL;
Mark Fashehccd979b2005-12-15 14:31:24 -08003290
3291 ocfs2_dlm_shutdown_debug(osb);
Mark Fashehccd979b2005-12-15 14:31:24 -08003292}
3293
Mark Fashehccd979b2005-12-15 14:31:24 -08003294static int ocfs2_drop_lock(struct ocfs2_super *osb,
Mark Fasheh0d5dc6c2006-09-14 14:44:51 -07003295 struct ocfs2_lock_res *lockres)
Mark Fashehccd979b2005-12-15 14:31:24 -08003296{
Joel Becker7431cd72008-02-01 12:15:37 -08003297 int ret;
Mark Fashehccd979b2005-12-15 14:31:24 -08003298 unsigned long flags;
Joel Beckerbd3e7612008-02-01 12:14:57 -08003299 u32 lkm_flags = 0;
Mark Fashehccd979b2005-12-15 14:31:24 -08003300
3301 /* We didn't get anywhere near actually using this lockres. */
3302 if (!(lockres->l_flags & OCFS2_LOCK_INITIALIZED))
3303 goto out;
3304
Mark Fashehb80fc012006-09-12 22:08:14 -07003305 if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB)
Joel Beckerbd3e7612008-02-01 12:14:57 -08003306 lkm_flags |= DLM_LKF_VALBLK;
Mark Fashehb80fc012006-09-12 22:08:14 -07003307
Mark Fashehccd979b2005-12-15 14:31:24 -08003308 spin_lock_irqsave(&lockres->l_lock, flags);
3309
3310 mlog_bug_on_msg(!(lockres->l_flags & OCFS2_LOCK_FREEING),
3311 "lockres %s, flags 0x%lx\n",
3312 lockres->l_name, lockres->l_flags);
3313
3314 while (lockres->l_flags & OCFS2_LOCK_BUSY) {
3315 mlog(0, "waiting on busy lock \"%s\": flags = %lx, action = "
3316 "%u, unlock_action = %u\n",
3317 lockres->l_name, lockres->l_flags, lockres->l_action,
3318 lockres->l_unlock_action);
3319
3320 spin_unlock_irqrestore(&lockres->l_lock, flags);
3321
3322 /* XXX: Today we just wait on any busy
3323 * locks... Perhaps we need to cancel converts in the
3324 * future? */
3325 ocfs2_wait_on_busy_lock(lockres);
3326
3327 spin_lock_irqsave(&lockres->l_lock, flags);
3328 }
3329
Mark Fasheh0d5dc6c2006-09-14 14:44:51 -07003330 if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB) {
3331 if (lockres->l_flags & OCFS2_LOCK_ATTACHED &&
Joel Beckerbd3e7612008-02-01 12:14:57 -08003332 lockres->l_level == DLM_LOCK_EX &&
Mark Fasheh0d5dc6c2006-09-14 14:44:51 -07003333 !(lockres->l_flags & OCFS2_LOCK_NEEDS_REFRESH))
3334 lockres->l_ops->set_lvb(lockres);
3335 }
Mark Fashehccd979b2005-12-15 14:31:24 -08003336
3337 if (lockres->l_flags & OCFS2_LOCK_BUSY)
3338 mlog(ML_ERROR, "destroying busy lock: \"%s\"\n",
3339 lockres->l_name);
3340 if (lockres->l_flags & OCFS2_LOCK_BLOCKED)
3341 mlog(0, "destroying blocked lock: \"%s\"\n", lockres->l_name);
3342
3343 if (!(lockres->l_flags & OCFS2_LOCK_ATTACHED)) {
3344 spin_unlock_irqrestore(&lockres->l_lock, flags);
3345 goto out;
3346 }
3347
3348 lockres_clear_flags(lockres, OCFS2_LOCK_ATTACHED);
3349
3350 /* make sure we never get here while waiting for an ast to
3351 * fire. */
3352 BUG_ON(lockres->l_action != OCFS2_AST_INVALID);
3353
3354 /* is this necessary? */
3355 lockres_or_flags(lockres, OCFS2_LOCK_BUSY);
3356 lockres->l_unlock_action = OCFS2_UNLOCK_DROP_LOCK;
3357 spin_unlock_irqrestore(&lockres->l_lock, flags);
3358
3359 mlog(0, "lock %s\n", lockres->l_name);
3360
Joel Beckera796d282010-01-28 19:22:39 -08003361 ret = ocfs2_dlm_unlock(osb->cconn, &lockres->l_lksb, lkm_flags);
Joel Becker7431cd72008-02-01 12:15:37 -08003362 if (ret) {
3363 ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08003364 mlog(ML_ERROR, "lockres flags: %lu\n", lockres->l_flags);
Joel Beckercf0acdc2008-01-29 16:59:55 -08003365 ocfs2_dlm_dump_lksb(&lockres->l_lksb);
Mark Fashehccd979b2005-12-15 14:31:24 -08003366 BUG();
3367 }
Coly Li73ac36e2009-01-07 18:09:16 -08003368 mlog(0, "lock %s, successful return from ocfs2_dlm_unlock\n",
Mark Fashehccd979b2005-12-15 14:31:24 -08003369 lockres->l_name);
3370
3371 ocfs2_wait_on_busy_lock(lockres);
3372out:
Mark Fashehccd979b2005-12-15 14:31:24 -08003373 return 0;
3374}
3375
Jan Kara84d86f82014-04-03 14:46:57 -07003376static void ocfs2_process_blocked_lock(struct ocfs2_super *osb,
3377 struct ocfs2_lock_res *lockres);
3378
Mark Fashehccd979b2005-12-15 14:31:24 -08003379/* Mark the lockres as being dropped. It will no longer be
3380 * queued if blocking, but we still may have to wait on it
Mark Fasheh34d024f2007-09-24 15:56:19 -07003381 * being dequeued from the downconvert thread before we can consider
Sunil Mushran2bd63212010-01-25 16:57:38 -08003382 * it safe to drop.
Mark Fashehccd979b2005-12-15 14:31:24 -08003383 *
3384 * You can *not* attempt to call cluster_lock on this lockres anymore. */
Jan Kara84d86f82014-04-03 14:46:57 -07003385void ocfs2_mark_lockres_freeing(struct ocfs2_super *osb,
3386 struct ocfs2_lock_res *lockres)
Mark Fashehccd979b2005-12-15 14:31:24 -08003387{
3388 int status;
3389 struct ocfs2_mask_waiter mw;
Jan Kara84d86f82014-04-03 14:46:57 -07003390 unsigned long flags, flags2;
Mark Fashehccd979b2005-12-15 14:31:24 -08003391
3392 ocfs2_init_mask_waiter(&mw);
3393
3394 spin_lock_irqsave(&lockres->l_lock, flags);
3395 lockres->l_flags |= OCFS2_LOCK_FREEING;
Jan Kara84d86f82014-04-03 14:46:57 -07003396 if (lockres->l_flags & OCFS2_LOCK_QUEUED && current == osb->dc_task) {
3397 /*
3398 * We know the downconvert is queued but not in progress
3399 * because we are the downconvert thread and processing
3400 * different lock. So we can just remove the lock from the
3401 * queue. This is not only an optimization but also a way
3402 * to avoid the following deadlock:
3403 * ocfs2_dentry_post_unlock()
3404 * ocfs2_dentry_lock_put()
3405 * ocfs2_drop_dentry_lock()
3406 * iput()
3407 * ocfs2_evict_inode()
3408 * ocfs2_clear_inode()
3409 * ocfs2_mark_lockres_freeing()
3410 * ... blocks waiting for OCFS2_LOCK_QUEUED
3411 * since we are the downconvert thread which
3412 * should clear the flag.
3413 */
3414 spin_unlock_irqrestore(&lockres->l_lock, flags);
3415 spin_lock_irqsave(&osb->dc_task_lock, flags2);
3416 list_del_init(&lockres->l_blocked_list);
3417 osb->blocked_lock_count--;
3418 spin_unlock_irqrestore(&osb->dc_task_lock, flags2);
3419 /*
3420 * Warn if we recurse into another post_unlock call. Strictly
3421 * speaking it isn't a problem but we need to be careful if
3422 * that happens (stack overflow, deadlocks, ...) so warn if
3423 * ocfs2 grows a path for which this can happen.
3424 */
3425 WARN_ON_ONCE(lockres->l_ops->post_unlock);
3426 /* Since the lock is freeing we don't do much in the fn below */
3427 ocfs2_process_blocked_lock(osb, lockres);
3428 return;
3429 }
Mark Fashehccd979b2005-12-15 14:31:24 -08003430 while (lockres->l_flags & OCFS2_LOCK_QUEUED) {
3431 lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_QUEUED, 0);
3432 spin_unlock_irqrestore(&lockres->l_lock, flags);
3433
3434 mlog(0, "Waiting on lockres %s\n", lockres->l_name);
3435
3436 status = ocfs2_wait_for_mask(&mw);
3437 if (status)
3438 mlog_errno(status);
3439
3440 spin_lock_irqsave(&lockres->l_lock, flags);
3441 }
3442 spin_unlock_irqrestore(&lockres->l_lock, flags);
3443}
3444
Mark Fashehd680efe2006-09-08 14:14:34 -07003445void ocfs2_simple_drop_lockres(struct ocfs2_super *osb,
3446 struct ocfs2_lock_res *lockres)
3447{
3448 int ret;
3449
Jan Kara84d86f82014-04-03 14:46:57 -07003450 ocfs2_mark_lockres_freeing(osb, lockres);
Mark Fasheh0d5dc6c2006-09-14 14:44:51 -07003451 ret = ocfs2_drop_lock(osb, lockres);
Mark Fashehd680efe2006-09-08 14:14:34 -07003452 if (ret)
3453 mlog_errno(ret);
3454}
3455
Mark Fashehccd979b2005-12-15 14:31:24 -08003456static void ocfs2_drop_osb_locks(struct ocfs2_super *osb)
3457{
Mark Fashehd680efe2006-09-08 14:14:34 -07003458 ocfs2_simple_drop_lockres(osb, &osb->osb_super_lockres);
3459 ocfs2_simple_drop_lockres(osb, &osb->osb_rename_lockres);
wengang wang6ca497a2009-03-06 21:29:10 +08003460 ocfs2_simple_drop_lockres(osb, &osb->osb_nfs_sync_lockres);
Srinivas Eeda83273932009-06-03 17:02:55 -07003461 ocfs2_simple_drop_lockres(osb, &osb->osb_orphan_scan.os_lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08003462}
3463
Mark Fashehccd979b2005-12-15 14:31:24 -08003464int ocfs2_drop_inode_locks(struct inode *inode)
3465{
3466 int status, err;
Mark Fashehccd979b2005-12-15 14:31:24 -08003467
Mark Fashehccd979b2005-12-15 14:31:24 -08003468 /* No need to call ocfs2_mark_lockres_freeing here -
3469 * ocfs2_clear_inode has done it for us. */
3470
3471 err = ocfs2_drop_lock(OCFS2_SB(inode->i_sb),
Tiger Yang50008632007-03-20 16:01:38 -07003472 &OCFS2_I(inode)->ip_open_lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08003473 if (err < 0)
3474 mlog_errno(err);
3475
3476 status = err;
3477
3478 err = ocfs2_drop_lock(OCFS2_SB(inode->i_sb),
Mark Fashehe63aecb62007-10-18 15:30:42 -07003479 &OCFS2_I(inode)->ip_inode_lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08003480 if (err < 0)
3481 mlog_errno(err);
3482 if (err < 0 && !status)
3483 status = err;
3484
3485 err = ocfs2_drop_lock(OCFS2_SB(inode->i_sb),
Mark Fasheh0d5dc6c2006-09-14 14:44:51 -07003486 &OCFS2_I(inode)->ip_rw_lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08003487 if (err < 0)
3488 mlog_errno(err);
3489 if (err < 0 && !status)
3490 status = err;
3491
Mark Fashehccd979b2005-12-15 14:31:24 -08003492 return status;
3493}
3494
Joel Beckerde551242008-02-01 14:45:08 -08003495static unsigned int ocfs2_prepare_downconvert(struct ocfs2_lock_res *lockres,
3496 int new_level)
Mark Fashehccd979b2005-12-15 14:31:24 -08003497{
3498 assert_spin_locked(&lockres->l_lock);
3499
Joel Beckerbd3e7612008-02-01 12:14:57 -08003500 BUG_ON(lockres->l_blocking <= DLM_LOCK_NL);
Mark Fashehccd979b2005-12-15 14:31:24 -08003501
3502 if (lockres->l_level <= new_level) {
Sunil Mushran9b915182010-02-26 19:42:44 -08003503 mlog(ML_ERROR, "lockres %s, lvl %d <= %d, blcklst %d, mask %d, "
3504 "type %d, flags 0x%lx, hold %d %d, act %d %d, req %d, "
3505 "block %d, pgen %d\n", lockres->l_name, lockres->l_level,
3506 new_level, list_empty(&lockres->l_blocked_list),
3507 list_empty(&lockres->l_mask_waiters), lockres->l_type,
3508 lockres->l_flags, lockres->l_ro_holders,
3509 lockres->l_ex_holders, lockres->l_action,
3510 lockres->l_unlock_action, lockres->l_requested,
3511 lockres->l_blocking, lockres->l_pending_gen);
Mark Fashehccd979b2005-12-15 14:31:24 -08003512 BUG();
3513 }
3514
Sunil Mushran9b915182010-02-26 19:42:44 -08003515 mlog(ML_BASTS, "lockres %s, level %d => %d, blocking %d\n",
3516 lockres->l_name, lockres->l_level, new_level, lockres->l_blocking);
Mark Fashehccd979b2005-12-15 14:31:24 -08003517
3518 lockres->l_action = OCFS2_AST_DOWNCONVERT;
3519 lockres->l_requested = new_level;
3520 lockres_or_flags(lockres, OCFS2_LOCK_BUSY);
Joel Beckerde551242008-02-01 14:45:08 -08003521 return lockres_set_pending(lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08003522}
3523
3524static int ocfs2_downconvert_lock(struct ocfs2_super *osb,
3525 struct ocfs2_lock_res *lockres,
3526 int new_level,
Joel Beckerde551242008-02-01 14:45:08 -08003527 int lvb,
3528 unsigned int generation)
Mark Fashehccd979b2005-12-15 14:31:24 -08003529{
Joel Beckerbd3e7612008-02-01 12:14:57 -08003530 int ret;
3531 u32 dlm_flags = DLM_LKF_CONVERT;
Mark Fashehccd979b2005-12-15 14:31:24 -08003532
Sunil Mushran9b915182010-02-26 19:42:44 -08003533 mlog(ML_BASTS, "lockres %s, level %d => %d\n", lockres->l_name,
3534 lockres->l_level, new_level);
3535
Eric Rene7ee2c082017-01-10 16:57:33 -08003536 /*
3537 * On DLM_LKF_VALBLK, fsdlm behaves differently with o2cb. It always
3538 * expects DLM_LKF_VALBLK being set if the LKB has LVB, so that
3539 * we can recover correctly from node failure. Otherwise, we may get
3540 * invalid LVB in LKB, but without DLM_SBF_VALNOTVALID being set.
3541 */
3542 if (!ocfs2_is_o2cb_active() &&
3543 lockres->l_ops->flags & LOCK_TYPE_USES_LVB)
3544 lvb = 1;
3545
Mark Fashehccd979b2005-12-15 14:31:24 -08003546 if (lvb)
Joel Beckerbd3e7612008-02-01 12:14:57 -08003547 dlm_flags |= DLM_LKF_VALBLK;
Mark Fashehccd979b2005-12-15 14:31:24 -08003548
Joel Becker4670c462008-02-01 14:39:35 -08003549 ret = ocfs2_dlm_lock(osb->cconn,
Joel Becker7431cd72008-02-01 12:15:37 -08003550 new_level,
3551 &lockres->l_lksb,
3552 dlm_flags,
3553 lockres->l_name,
Joel Beckera796d282010-01-28 19:22:39 -08003554 OCFS2_LOCK_ID_MAX_LEN - 1);
Joel Beckerde551242008-02-01 14:45:08 -08003555 lockres_clear_pending(lockres, generation, osb);
Joel Becker7431cd72008-02-01 12:15:37 -08003556 if (ret) {
3557 ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08003558 ocfs2_recover_from_dlm_error(lockres, 1);
3559 goto bail;
3560 }
3561
3562 ret = 0;
3563bail:
Mark Fashehccd979b2005-12-15 14:31:24 -08003564 return ret;
3565}
3566
Joel Becker24ef1812008-01-29 17:37:32 -08003567/* returns 1 when the caller should unlock and call ocfs2_dlm_unlock */
Mark Fashehccd979b2005-12-15 14:31:24 -08003568static int ocfs2_prepare_cancel_convert(struct ocfs2_super *osb,
3569 struct ocfs2_lock_res *lockres)
3570{
3571 assert_spin_locked(&lockres->l_lock);
3572
Mark Fashehccd979b2005-12-15 14:31:24 -08003573 if (lockres->l_unlock_action == OCFS2_UNLOCK_CANCEL_CONVERT) {
3574 /* If we're already trying to cancel a lock conversion
3575 * then just drop the spinlock and allow the caller to
3576 * requeue this lock. */
Sunil Mushran9b915182010-02-26 19:42:44 -08003577 mlog(ML_BASTS, "lockres %s, skip convert\n", lockres->l_name);
Mark Fashehccd979b2005-12-15 14:31:24 -08003578 return 0;
3579 }
3580
3581 /* were we in a convert when we got the bast fire? */
3582 BUG_ON(lockres->l_action != OCFS2_AST_CONVERT &&
3583 lockres->l_action != OCFS2_AST_DOWNCONVERT);
3584 /* set things up for the unlockast to know to just
3585 * clear out the ast_action and unset busy, etc. */
3586 lockres->l_unlock_action = OCFS2_UNLOCK_CANCEL_CONVERT;
3587
3588 mlog_bug_on_msg(!(lockres->l_flags & OCFS2_LOCK_BUSY),
3589 "lock %s, invalid flags: 0x%lx\n",
3590 lockres->l_name, lockres->l_flags);
3591
Sunil Mushran9b915182010-02-26 19:42:44 -08003592 mlog(ML_BASTS, "lockres %s\n", lockres->l_name);
3593
Mark Fashehccd979b2005-12-15 14:31:24 -08003594 return 1;
3595}
3596
3597static int ocfs2_cancel_convert(struct ocfs2_super *osb,
3598 struct ocfs2_lock_res *lockres)
3599{
3600 int ret;
Mark Fashehccd979b2005-12-15 14:31:24 -08003601
Joel Becker4670c462008-02-01 14:39:35 -08003602 ret = ocfs2_dlm_unlock(osb->cconn, &lockres->l_lksb,
Joel Beckera796d282010-01-28 19:22:39 -08003603 DLM_LKF_CANCEL);
Joel Becker7431cd72008-02-01 12:15:37 -08003604 if (ret) {
3605 ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08003606 ocfs2_recover_from_dlm_error(lockres, 0);
3607 }
3608
Sunil Mushran9b915182010-02-26 19:42:44 -08003609 mlog(ML_BASTS, "lockres %s\n", lockres->l_name);
Mark Fashehccd979b2005-12-15 14:31:24 -08003610
Mark Fashehccd979b2005-12-15 14:31:24 -08003611 return ret;
3612}
3613
Mark Fashehb5e500e2006-09-13 22:01:16 -07003614static int ocfs2_unblock_lock(struct ocfs2_super *osb,
3615 struct ocfs2_lock_res *lockres,
3616 struct ocfs2_unblock_ctl *ctl)
Mark Fashehccd979b2005-12-15 14:31:24 -08003617{
3618 unsigned long flags;
3619 int blocking;
3620 int new_level;
Sunil Mushran079b8052010-02-03 10:16:54 -08003621 int level;
Mark Fashehccd979b2005-12-15 14:31:24 -08003622 int ret = 0;
Mark Fasheh5ef0d4e2006-09-13 21:21:52 -07003623 int set_lvb = 0;
Joel Beckerde551242008-02-01 14:45:08 -08003624 unsigned int gen;
Mark Fashehccd979b2005-12-15 14:31:24 -08003625
Mark Fashehccd979b2005-12-15 14:31:24 -08003626 spin_lock_irqsave(&lockres->l_lock, flags);
3627
Mark Fashehccd979b2005-12-15 14:31:24 -08003628recheck:
Sunil Mushrandb0f6ce2010-02-01 16:55:50 -08003629 /*
3630 * Is it still blocking? If not, we have no more work to do.
3631 */
3632 if (!(lockres->l_flags & OCFS2_LOCK_BLOCKED)) {
3633 BUG_ON(lockres->l_blocking != DLM_LOCK_NL);
3634 spin_unlock_irqrestore(&lockres->l_lock, flags);
3635 ret = 0;
3636 goto leave;
3637 }
3638
Mark Fashehccd979b2005-12-15 14:31:24 -08003639 if (lockres->l_flags & OCFS2_LOCK_BUSY) {
Joel Beckerde551242008-02-01 14:45:08 -08003640 /* XXX
3641 * This is a *big* race. The OCFS2_LOCK_PENDING flag
3642 * exists entirely for one reason - another thread has set
3643 * OCFS2_LOCK_BUSY, but has *NOT* yet called dlm_lock().
3644 *
3645 * If we do ocfs2_cancel_convert() before the other thread
3646 * calls dlm_lock(), our cancel will do nothing. We will
3647 * get no ast, and we will have no way of knowing the
3648 * cancel failed. Meanwhile, the other thread will call
3649 * into dlm_lock() and wait...forever.
3650 *
3651 * Why forever? Because another node has asked for the
3652 * lock first; that's why we're here in unblock_lock().
3653 *
3654 * The solution is OCFS2_LOCK_PENDING. When PENDING is
3655 * set, we just requeue the unblock. Only when the other
3656 * thread has called dlm_lock() and cleared PENDING will
3657 * we then cancel their request.
3658 *
3659 * All callers of dlm_lock() must set OCFS2_DLM_PENDING
3660 * at the same time they set OCFS2_DLM_BUSY. They must
3661 * clear OCFS2_DLM_PENDING after dlm_lock() returns.
3662 */
Sunil Mushran9b915182010-02-26 19:42:44 -08003663 if (lockres->l_flags & OCFS2_LOCK_PENDING) {
3664 mlog(ML_BASTS, "lockres %s, ReQ: Pending\n",
3665 lockres->l_name);
Joel Beckerde551242008-02-01 14:45:08 -08003666 goto leave_requeue;
Sunil Mushran9b915182010-02-26 19:42:44 -08003667 }
Joel Beckerde551242008-02-01 14:45:08 -08003668
Mark Fashehd680efe2006-09-08 14:14:34 -07003669 ctl->requeue = 1;
Mark Fashehccd979b2005-12-15 14:31:24 -08003670 ret = ocfs2_prepare_cancel_convert(osb, lockres);
3671 spin_unlock_irqrestore(&lockres->l_lock, flags);
3672 if (ret) {
3673 ret = ocfs2_cancel_convert(osb, lockres);
3674 if (ret < 0)
3675 mlog_errno(ret);
3676 }
3677 goto leave;
3678 }
3679
Sunil Mushrana1912822010-01-21 10:50:03 -08003680 /*
3681 * This prevents livelocks. OCFS2_LOCK_UPCONVERT_FINISHING flag is
3682 * set when the ast is received for an upconvert just before the
3683 * OCFS2_LOCK_BUSY flag is cleared. Now if the fs received a bast
3684 * on the heels of the ast, we want to delay the downconvert just
3685 * enough to allow the up requestor to do its task. Because this
3686 * lock is in the blocked queue, the lock will be downconverted
3687 * as soon as the requestor is done with the lock.
3688 */
3689 if (lockres->l_flags & OCFS2_LOCK_UPCONVERT_FINISHING)
3690 goto leave_requeue;
3691
Sunil Mushran0d741252010-01-29 09:44:11 -08003692 /*
3693 * How can we block and yet be at NL? We were trying to upconvert
3694 * from NL and got canceled. The code comes back here, and now
3695 * we notice and clear BLOCKING.
3696 */
3697 if (lockres->l_level == DLM_LOCK_NL) {
3698 BUG_ON(lockres->l_ex_holders || lockres->l_ro_holders);
Sunil Mushran9b915182010-02-26 19:42:44 -08003699 mlog(ML_BASTS, "lockres %s, Aborting dc\n", lockres->l_name);
Sunil Mushran0d741252010-01-29 09:44:11 -08003700 lockres->l_blocking = DLM_LOCK_NL;
3701 lockres_clear_flags(lockres, OCFS2_LOCK_BLOCKED);
3702 spin_unlock_irqrestore(&lockres->l_lock, flags);
3703 goto leave;
3704 }
3705
Mark Fashehccd979b2005-12-15 14:31:24 -08003706 /* if we're blocking an exclusive and we have *any* holders,
3707 * then requeue. */
Joel Beckerbd3e7612008-02-01 12:14:57 -08003708 if ((lockres->l_blocking == DLM_LOCK_EX)
Sunil Mushran9b915182010-02-26 19:42:44 -08003709 && (lockres->l_ex_holders || lockres->l_ro_holders)) {
3710 mlog(ML_BASTS, "lockres %s, ReQ: EX/PR Holders %u,%u\n",
3711 lockres->l_name, lockres->l_ex_holders,
3712 lockres->l_ro_holders);
Mark Fashehf7fbfdd2006-09-13 21:02:29 -07003713 goto leave_requeue;
Sunil Mushran9b915182010-02-26 19:42:44 -08003714 }
Mark Fashehccd979b2005-12-15 14:31:24 -08003715
3716 /* If it's a PR we're blocking, then only
3717 * requeue if we've got any EX holders */
Joel Beckerbd3e7612008-02-01 12:14:57 -08003718 if (lockres->l_blocking == DLM_LOCK_PR &&
Sunil Mushran9b915182010-02-26 19:42:44 -08003719 lockres->l_ex_holders) {
3720 mlog(ML_BASTS, "lockres %s, ReQ: EX Holders %u\n",
3721 lockres->l_name, lockres->l_ex_holders);
Mark Fashehf7fbfdd2006-09-13 21:02:29 -07003722 goto leave_requeue;
Sunil Mushran9b915182010-02-26 19:42:44 -08003723 }
Mark Fashehf7fbfdd2006-09-13 21:02:29 -07003724
3725 /*
3726 * Can we get a lock in this state if the holder counts are
3727 * zero? The meta data unblock code used to check this.
3728 */
3729 if ((lockres->l_ops->flags & LOCK_TYPE_REQUIRES_REFRESH)
Sunil Mushran9b915182010-02-26 19:42:44 -08003730 && (lockres->l_flags & OCFS2_LOCK_REFRESHING)) {
3731 mlog(ML_BASTS, "lockres %s, ReQ: Lock Refreshing\n",
3732 lockres->l_name);
Mark Fashehf7fbfdd2006-09-13 21:02:29 -07003733 goto leave_requeue;
Sunil Mushran9b915182010-02-26 19:42:44 -08003734 }
Mark Fashehccd979b2005-12-15 14:31:24 -08003735
Mark Fasheh16d5b952006-09-13 21:10:12 -07003736 new_level = ocfs2_highest_compat_lock_level(lockres->l_blocking);
3737
3738 if (lockres->l_ops->check_downconvert
Sunil Mushran9b915182010-02-26 19:42:44 -08003739 && !lockres->l_ops->check_downconvert(lockres, new_level)) {
3740 mlog(ML_BASTS, "lockres %s, ReQ: Checkpointing\n",
3741 lockres->l_name);
Mark Fasheh16d5b952006-09-13 21:10:12 -07003742 goto leave_requeue;
Sunil Mushran9b915182010-02-26 19:42:44 -08003743 }
Mark Fasheh16d5b952006-09-13 21:10:12 -07003744
Mark Fashehccd979b2005-12-15 14:31:24 -08003745 /* If we get here, then we know that there are no more
3746 * incompatible holders (and anyone asking for an incompatible
3747 * lock is blocked). We can now downconvert the lock */
Mark Fashehcc567d82006-09-13 21:52:21 -07003748 if (!lockres->l_ops->downconvert_worker)
Mark Fashehccd979b2005-12-15 14:31:24 -08003749 goto downconvert;
3750
3751 /* Some lockres types want to do a bit of work before
3752 * downconverting a lock. Allow that here. The worker function
3753 * may sleep, so we save off a copy of what we're blocking as
3754 * it may change while we're not holding the spin lock. */
3755 blocking = lockres->l_blocking;
Sunil Mushran079b8052010-02-03 10:16:54 -08003756 level = lockres->l_level;
Mark Fashehccd979b2005-12-15 14:31:24 -08003757 spin_unlock_irqrestore(&lockres->l_lock, flags);
3758
Mark Fashehcc567d82006-09-13 21:52:21 -07003759 ctl->unblock_action = lockres->l_ops->downconvert_worker(lockres, blocking);
Mark Fashehd680efe2006-09-08 14:14:34 -07003760
Sunil Mushran9b915182010-02-26 19:42:44 -08003761 if (ctl->unblock_action == UNBLOCK_STOP_POST) {
3762 mlog(ML_BASTS, "lockres %s, UNBLOCK_STOP_POST\n",
3763 lockres->l_name);
Mark Fashehd680efe2006-09-08 14:14:34 -07003764 goto leave;
Sunil Mushran9b915182010-02-26 19:42:44 -08003765 }
Mark Fashehccd979b2005-12-15 14:31:24 -08003766
3767 spin_lock_irqsave(&lockres->l_lock, flags);
Sunil Mushran079b8052010-02-03 10:16:54 -08003768 if ((blocking != lockres->l_blocking) || (level != lockres->l_level)) {
Mark Fashehccd979b2005-12-15 14:31:24 -08003769 /* If this changed underneath us, then we can't drop
3770 * it just yet. */
Sunil Mushran9b915182010-02-26 19:42:44 -08003771 mlog(ML_BASTS, "lockres %s, block=%d:%d, level=%d:%d, "
3772 "Recheck\n", lockres->l_name, blocking,
3773 lockres->l_blocking, level, lockres->l_level);
Mark Fashehccd979b2005-12-15 14:31:24 -08003774 goto recheck;
3775 }
3776
3777downconvert:
Mark Fashehd680efe2006-09-08 14:14:34 -07003778 ctl->requeue = 0;
Mark Fashehccd979b2005-12-15 14:31:24 -08003779
Mark Fasheh5ef0d4e2006-09-13 21:21:52 -07003780 if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB) {
Joel Beckerbd3e7612008-02-01 12:14:57 -08003781 if (lockres->l_level == DLM_LOCK_EX)
Mark Fasheh5ef0d4e2006-09-13 21:21:52 -07003782 set_lvb = 1;
3783
3784 /*
3785 * We only set the lvb if the lock has been fully
3786 * refreshed - otherwise we risk setting stale
3787 * data. Otherwise, there's no need to actually clear
3788 * out the lvb here as it's value is still valid.
3789 */
3790 if (set_lvb && !(lockres->l_flags & OCFS2_LOCK_NEEDS_REFRESH))
3791 lockres->l_ops->set_lvb(lockres);
3792 }
3793
Joel Beckerde551242008-02-01 14:45:08 -08003794 gen = ocfs2_prepare_downconvert(lockres, new_level);
Mark Fashehccd979b2005-12-15 14:31:24 -08003795 spin_unlock_irqrestore(&lockres->l_lock, flags);
Joel Beckerde551242008-02-01 14:45:08 -08003796 ret = ocfs2_downconvert_lock(osb, lockres, new_level, set_lvb,
3797 gen);
3798
Mark Fashehccd979b2005-12-15 14:31:24 -08003799leave:
Tao Mac1e8d352011-03-07 16:43:21 +08003800 if (ret)
3801 mlog_errno(ret);
Mark Fashehccd979b2005-12-15 14:31:24 -08003802 return ret;
Mark Fashehf7fbfdd2006-09-13 21:02:29 -07003803
3804leave_requeue:
3805 spin_unlock_irqrestore(&lockres->l_lock, flags);
3806 ctl->requeue = 1;
3807
Mark Fashehf7fbfdd2006-09-13 21:02:29 -07003808 return 0;
Mark Fashehccd979b2005-12-15 14:31:24 -08003809}
3810
Mark Fashehd680efe2006-09-08 14:14:34 -07003811static int ocfs2_data_convert_worker(struct ocfs2_lock_res *lockres,
3812 int blocking)
Mark Fashehccd979b2005-12-15 14:31:24 -08003813{
3814 struct inode *inode;
3815 struct address_space *mapping;
Goldwyn Rodrigues5e98d492010-06-28 10:04:32 -05003816 struct ocfs2_inode_info *oi;
Mark Fashehccd979b2005-12-15 14:31:24 -08003817
Mark Fashehccd979b2005-12-15 14:31:24 -08003818 inode = ocfs2_lock_res_inode(lockres);
3819 mapping = inode->i_mapping;
3820
Goldwyn Rodrigues5e98d492010-06-28 10:04:32 -05003821 if (S_ISDIR(inode->i_mode)) {
3822 oi = OCFS2_I(inode);
3823 oi->ip_dir_lock_gen++;
3824 mlog(0, "generation: %u\n", oi->ip_dir_lock_gen);
3825 goto out;
3826 }
3827
Mark Fasheh1044e402008-02-28 17:16:03 -08003828 if (!S_ISREG(inode->i_mode))
Mark Fashehf1f54062007-10-18 15:13:59 -07003829 goto out;
3830
Mark Fasheh7f4a2a92006-12-11 11:06:36 -08003831 /*
3832 * We need this before the filemap_fdatawrite() so that it can
3833 * transfer the dirty bit from the PTE to the
3834 * page. Unfortunately this means that even for EX->PR
3835 * downconverts, we'll lose our mappings and have to build
3836 * them up again.
3837 */
3838 unmap_mapping_range(mapping, 0, 0, 0);
3839
Mark Fashehccd979b2005-12-15 14:31:24 -08003840 if (filemap_fdatawrite(mapping)) {
Mark Fashehb06970532006-03-03 10:24:33 -08003841 mlog(ML_ERROR, "Could not sync inode %llu for downconvert!",
3842 (unsigned long long)OCFS2_I(inode)->ip_blkno);
Mark Fashehccd979b2005-12-15 14:31:24 -08003843 }
3844 sync_mapping_buffers(mapping);
Joel Beckerbd3e7612008-02-01 12:14:57 -08003845 if (blocking == DLM_LOCK_EX) {
Mark Fashehccd979b2005-12-15 14:31:24 -08003846 truncate_inode_pages(mapping, 0);
Mark Fashehccd979b2005-12-15 14:31:24 -08003847 } else {
3848 /* We only need to wait on the I/O if we're not also
3849 * truncating pages because truncate_inode_pages waits
3850 * for us above. We don't truncate pages if we're
3851 * blocking anything < EXMODE because we want to keep
3852 * them around in that case. */
3853 filemap_fdatawait(mapping);
3854 }
3855
Andreas Gruenbacherb8a7a3a2016-03-24 14:38:37 +01003856 forget_all_cached_acls(inode);
3857
Mark Fashehf1f54062007-10-18 15:13:59 -07003858out:
Mark Fashehd680efe2006-09-08 14:14:34 -07003859 return UNBLOCK_CONTINUE;
Mark Fashehccd979b2005-12-15 14:31:24 -08003860}
3861
Tao Maa4338482009-08-18 11:19:29 +08003862static int ocfs2_ci_checkpointed(struct ocfs2_caching_info *ci,
3863 struct ocfs2_lock_res *lockres,
3864 int new_level)
Mark Fasheh810d5ae2006-09-13 21:39:52 -07003865{
Tao Maa4338482009-08-18 11:19:29 +08003866 int checkpointed = ocfs2_ci_fully_checkpointed(ci);
Mark Fasheh810d5ae2006-09-13 21:39:52 -07003867
Joel Beckerbd3e7612008-02-01 12:14:57 -08003868 BUG_ON(new_level != DLM_LOCK_NL && new_level != DLM_LOCK_PR);
3869 BUG_ON(lockres->l_level != DLM_LOCK_EX && !checkpointed);
Mark Fasheh810d5ae2006-09-13 21:39:52 -07003870
3871 if (checkpointed)
3872 return 1;
3873
Tao Maa4338482009-08-18 11:19:29 +08003874 ocfs2_start_checkpoint(OCFS2_SB(ocfs2_metadata_cache_get_super(ci)));
Mark Fasheh810d5ae2006-09-13 21:39:52 -07003875 return 0;
3876}
3877
Tao Maa4338482009-08-18 11:19:29 +08003878static int ocfs2_check_meta_downconvert(struct ocfs2_lock_res *lockres,
3879 int new_level)
3880{
3881 struct inode *inode = ocfs2_lock_res_inode(lockres);
3882
3883 return ocfs2_ci_checkpointed(INODE_CACHE(inode), lockres, new_level);
3884}
3885
Mark Fasheh810d5ae2006-09-13 21:39:52 -07003886static void ocfs2_set_meta_lvb(struct ocfs2_lock_res *lockres)
3887{
3888 struct inode *inode = ocfs2_lock_res_inode(lockres);
3889
3890 __ocfs2_stuff_meta_lvb(inode);
3891}
3892
Mark Fashehd680efe2006-09-08 14:14:34 -07003893/*
3894 * Does the final reference drop on our dentry lock. Right now this
Mark Fasheh34d024f2007-09-24 15:56:19 -07003895 * happens in the downconvert thread, but we could choose to simplify the
Mark Fashehd680efe2006-09-08 14:14:34 -07003896 * dlmglue API and push these off to the ocfs2_wq in the future.
3897 */
3898static void ocfs2_dentry_post_unlock(struct ocfs2_super *osb,
3899 struct ocfs2_lock_res *lockres)
3900{
3901 struct ocfs2_dentry_lock *dl = ocfs2_lock_res_dl(lockres);
3902 ocfs2_dentry_lock_put(osb, dl);
3903}
3904
3905/*
3906 * d_delete() matching dentries before the lock downconvert.
3907 *
3908 * At this point, any process waiting to destroy the
3909 * dentry_lock due to last ref count is stopped by the
3910 * OCFS2_LOCK_QUEUED flag.
3911 *
3912 * We have two potential problems
3913 *
3914 * 1) If we do the last reference drop on our dentry_lock (via dput)
3915 * we'll wind up in ocfs2_release_dentry_lock(), waiting on
3916 * the downconvert to finish. Instead we take an elevated
3917 * reference and push the drop until after we've completed our
3918 * unblock processing.
3919 *
3920 * 2) There might be another process with a final reference,
3921 * waiting on us to finish processing. If this is the case, we
3922 * detect it and exit out - there's no more dentries anyway.
3923 */
3924static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res *lockres,
3925 int blocking)
3926{
3927 struct ocfs2_dentry_lock *dl = ocfs2_lock_res_dl(lockres);
3928 struct ocfs2_inode_info *oi = OCFS2_I(dl->dl_inode);
3929 struct dentry *dentry;
3930 unsigned long flags;
3931 int extra_ref = 0;
3932
3933 /*
3934 * This node is blocking another node from getting a read
3935 * lock. This happens when we've renamed within a
3936 * directory. We've forced the other nodes to d_delete(), but
3937 * we never actually dropped our lock because it's still
3938 * valid. The downconvert code will retain a PR for this node,
3939 * so there's no further work to do.
3940 */
Joel Beckerbd3e7612008-02-01 12:14:57 -08003941 if (blocking == DLM_LOCK_PR)
Mark Fashehd680efe2006-09-08 14:14:34 -07003942 return UNBLOCK_CONTINUE;
3943
3944 /*
3945 * Mark this inode as potentially orphaned. The code in
3946 * ocfs2_delete_inode() will figure out whether it actually
3947 * needs to be freed or not.
3948 */
3949 spin_lock(&oi->ip_lock);
3950 oi->ip_flags |= OCFS2_INODE_MAYBE_ORPHANED;
3951 spin_unlock(&oi->ip_lock);
3952
3953 /*
3954 * Yuck. We need to make sure however that the check of
3955 * OCFS2_LOCK_FREEING and the extra reference are atomic with
3956 * respect to a reference decrement or the setting of that
3957 * flag.
3958 */
3959 spin_lock_irqsave(&lockres->l_lock, flags);
3960 spin_lock(&dentry_attach_lock);
3961 if (!(lockres->l_flags & OCFS2_LOCK_FREEING)
3962 && dl->dl_count) {
3963 dl->dl_count++;
3964 extra_ref = 1;
3965 }
3966 spin_unlock(&dentry_attach_lock);
3967 spin_unlock_irqrestore(&lockres->l_lock, flags);
3968
3969 mlog(0, "extra_ref = %d\n", extra_ref);
3970
3971 /*
3972 * We have a process waiting on us in ocfs2_dentry_iput(),
3973 * which means we can't have any more outstanding
3974 * aliases. There's no need to do any more work.
3975 */
3976 if (!extra_ref)
3977 return UNBLOCK_CONTINUE;
3978
3979 spin_lock(&dentry_attach_lock);
3980 while (1) {
3981 dentry = ocfs2_find_local_alias(dl->dl_inode,
3982 dl->dl_parent_blkno, 1);
3983 if (!dentry)
3984 break;
3985 spin_unlock(&dentry_attach_lock);
3986
alex chen10ab8812015-02-10 14:09:23 -08003987 if (S_ISDIR(dl->dl_inode->i_mode))
3988 shrink_dcache_parent(dentry);
3989
Al Viroa4555892014-10-21 20:11:25 -04003990 mlog(0, "d_delete(%pd);\n", dentry);
Mark Fashehd680efe2006-09-08 14:14:34 -07003991
3992 /*
3993 * The following dcache calls may do an
3994 * iput(). Normally we don't want that from the
3995 * downconverting thread, but in this case it's ok
3996 * because the requesting node already has an
3997 * exclusive lock on the inode, so it can't be queued
3998 * for a downconvert.
3999 */
4000 d_delete(dentry);
4001 dput(dentry);
4002
4003 spin_lock(&dentry_attach_lock);
4004 }
4005 spin_unlock(&dentry_attach_lock);
4006
4007 /*
4008 * If we are the last holder of this dentry lock, there is no
4009 * reason to downconvert so skip straight to the unlock.
4010 */
4011 if (dl->dl_count == 1)
4012 return UNBLOCK_STOP_POST;
4013
4014 return UNBLOCK_CONTINUE_POST;
4015}
4016
Tao Ma8dec98e2009-08-18 11:19:58 +08004017static int ocfs2_check_refcount_downconvert(struct ocfs2_lock_res *lockres,
4018 int new_level)
4019{
4020 struct ocfs2_refcount_tree *tree =
4021 ocfs2_lock_res_refcount_tree(lockres);
4022
4023 return ocfs2_ci_checkpointed(&tree->rf_ci, lockres, new_level);
4024}
4025
4026static int ocfs2_refcount_convert_worker(struct ocfs2_lock_res *lockres,
4027 int blocking)
4028{
4029 struct ocfs2_refcount_tree *tree =
4030 ocfs2_lock_res_refcount_tree(lockres);
4031
4032 ocfs2_metadata_cache_purge(&tree->rf_ci);
4033
4034 return UNBLOCK_CONTINUE;
4035}
4036
Jan Kara9e33d692008-08-25 19:56:50 +02004037static void ocfs2_set_qinfo_lvb(struct ocfs2_lock_res *lockres)
4038{
4039 struct ocfs2_qinfo_lvb *lvb;
4040 struct ocfs2_mem_dqinfo *oinfo = ocfs2_lock_res_qinfo(lockres);
4041 struct mem_dqinfo *info = sb_dqinfo(oinfo->dqi_gi.dqi_sb,
4042 oinfo->dqi_gi.dqi_type);
4043
Mark Fasheha641dc22008-12-24 16:03:48 -08004044 lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
Jan Kara9e33d692008-08-25 19:56:50 +02004045 lvb->lvb_version = OCFS2_QINFO_LVB_VERSION;
4046 lvb->lvb_bgrace = cpu_to_be32(info->dqi_bgrace);
4047 lvb->lvb_igrace = cpu_to_be32(info->dqi_igrace);
4048 lvb->lvb_syncms = cpu_to_be32(oinfo->dqi_syncms);
4049 lvb->lvb_blocks = cpu_to_be32(oinfo->dqi_gi.dqi_blocks);
4050 lvb->lvb_free_blk = cpu_to_be32(oinfo->dqi_gi.dqi_free_blk);
4051 lvb->lvb_free_entry = cpu_to_be32(oinfo->dqi_gi.dqi_free_entry);
Jan Kara9e33d692008-08-25 19:56:50 +02004052}
4053
4054void ocfs2_qinfo_unlock(struct ocfs2_mem_dqinfo *oinfo, int ex)
4055{
4056 struct ocfs2_lock_res *lockres = &oinfo->dqi_gqlock;
4057 struct ocfs2_super *osb = OCFS2_SB(oinfo->dqi_gi.dqi_sb);
4058 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
4059
Jan Kara9e33d692008-08-25 19:56:50 +02004060 if (!ocfs2_is_hard_readonly(osb) && !ocfs2_mount_local(osb))
4061 ocfs2_cluster_unlock(osb, lockres, level);
Jan Kara9e33d692008-08-25 19:56:50 +02004062}
4063
4064static int ocfs2_refresh_qinfo(struct ocfs2_mem_dqinfo *oinfo)
4065{
4066 struct mem_dqinfo *info = sb_dqinfo(oinfo->dqi_gi.dqi_sb,
4067 oinfo->dqi_gi.dqi_type);
4068 struct ocfs2_lock_res *lockres = &oinfo->dqi_gqlock;
4069 struct ocfs2_qinfo_lvb *lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
Joel Becker85eb8b72008-11-25 15:31:27 +01004070 struct buffer_head *bh = NULL;
Jan Kara9e33d692008-08-25 19:56:50 +02004071 struct ocfs2_global_disk_dqinfo *gdinfo;
4072 int status = 0;
4073
Joel Becker1c520df2009-06-19 15:14:13 -07004074 if (ocfs2_dlm_lvb_valid(&lockres->l_lksb) &&
4075 lvb->lvb_version == OCFS2_QINFO_LVB_VERSION) {
Jan Kara9e33d692008-08-25 19:56:50 +02004076 info->dqi_bgrace = be32_to_cpu(lvb->lvb_bgrace);
4077 info->dqi_igrace = be32_to_cpu(lvb->lvb_igrace);
4078 oinfo->dqi_syncms = be32_to_cpu(lvb->lvb_syncms);
4079 oinfo->dqi_gi.dqi_blocks = be32_to_cpu(lvb->lvb_blocks);
4080 oinfo->dqi_gi.dqi_free_blk = be32_to_cpu(lvb->lvb_free_blk);
4081 oinfo->dqi_gi.dqi_free_entry =
4082 be32_to_cpu(lvb->lvb_free_entry);
4083 } else {
Jan Karaae4f6ef2010-04-28 19:04:29 +02004084 status = ocfs2_read_quota_phys_block(oinfo->dqi_gqinode,
4085 oinfo->dqi_giblk, &bh);
Joel Becker85eb8b72008-11-25 15:31:27 +01004086 if (status) {
Jan Kara9e33d692008-08-25 19:56:50 +02004087 mlog_errno(status);
4088 goto bail;
4089 }
4090 gdinfo = (struct ocfs2_global_disk_dqinfo *)
4091 (bh->b_data + OCFS2_GLOBAL_INFO_OFF);
4092 info->dqi_bgrace = le32_to_cpu(gdinfo->dqi_bgrace);
4093 info->dqi_igrace = le32_to_cpu(gdinfo->dqi_igrace);
4094 oinfo->dqi_syncms = le32_to_cpu(gdinfo->dqi_syncms);
4095 oinfo->dqi_gi.dqi_blocks = le32_to_cpu(gdinfo->dqi_blocks);
4096 oinfo->dqi_gi.dqi_free_blk = le32_to_cpu(gdinfo->dqi_free_blk);
4097 oinfo->dqi_gi.dqi_free_entry =
4098 le32_to_cpu(gdinfo->dqi_free_entry);
4099 brelse(bh);
4100 ocfs2_track_lock_refresh(lockres);
4101 }
4102
4103bail:
4104 return status;
4105}
4106
4107/* Lock quota info, this function expects at least shared lock on the quota file
4108 * so that we can safely refresh quota info from disk. */
4109int ocfs2_qinfo_lock(struct ocfs2_mem_dqinfo *oinfo, int ex)
4110{
4111 struct ocfs2_lock_res *lockres = &oinfo->dqi_gqlock;
4112 struct ocfs2_super *osb = OCFS2_SB(oinfo->dqi_gi.dqi_sb);
4113 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
4114 int status = 0;
4115
Jan Kara9e33d692008-08-25 19:56:50 +02004116 /* On RO devices, locking really isn't needed... */
4117 if (ocfs2_is_hard_readonly(osb)) {
4118 if (ex)
4119 status = -EROFS;
4120 goto bail;
4121 }
4122 if (ocfs2_mount_local(osb))
4123 goto bail;
4124
4125 status = ocfs2_cluster_lock(osb, lockres, level, 0, 0);
4126 if (status < 0) {
4127 mlog_errno(status);
4128 goto bail;
4129 }
4130 if (!ocfs2_should_refresh_lock_res(lockres))
4131 goto bail;
4132 /* OK, we have the lock but we need to refresh the quota info */
4133 status = ocfs2_refresh_qinfo(oinfo);
4134 if (status)
4135 ocfs2_qinfo_unlock(oinfo, ex);
4136 ocfs2_complete_lock_res_refresh(lockres, status);
4137bail:
Jan Kara9e33d692008-08-25 19:56:50 +02004138 return status;
4139}
4140
Tao Ma8dec98e2009-08-18 11:19:58 +08004141int ocfs2_refcount_lock(struct ocfs2_refcount_tree *ref_tree, int ex)
4142{
4143 int status;
4144 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
4145 struct ocfs2_lock_res *lockres = &ref_tree->rf_lockres;
4146 struct ocfs2_super *osb = lockres->l_priv;
4147
4148
4149 if (ocfs2_is_hard_readonly(osb))
4150 return -EROFS;
4151
4152 if (ocfs2_mount_local(osb))
4153 return 0;
4154
4155 status = ocfs2_cluster_lock(osb, lockres, level, 0, 0);
4156 if (status < 0)
4157 mlog_errno(status);
4158
4159 return status;
4160}
4161
4162void ocfs2_refcount_unlock(struct ocfs2_refcount_tree *ref_tree, int ex)
4163{
4164 int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
4165 struct ocfs2_lock_res *lockres = &ref_tree->rf_lockres;
4166 struct ocfs2_super *osb = lockres->l_priv;
4167
4168 if (!ocfs2_mount_local(osb))
4169 ocfs2_cluster_unlock(osb, lockres, level);
4170}
4171
Adrian Bunk00600052008-01-29 00:11:41 +02004172static void ocfs2_process_blocked_lock(struct ocfs2_super *osb,
4173 struct ocfs2_lock_res *lockres)
Mark Fashehccd979b2005-12-15 14:31:24 -08004174{
4175 int status;
Mark Fashehd680efe2006-09-08 14:14:34 -07004176 struct ocfs2_unblock_ctl ctl = {0, 0,};
Mark Fashehccd979b2005-12-15 14:31:24 -08004177 unsigned long flags;
4178
4179 /* Our reference to the lockres in this function can be
4180 * considered valid until we remove the OCFS2_LOCK_QUEUED
4181 * flag. */
4182
Mark Fashehccd979b2005-12-15 14:31:24 -08004183 BUG_ON(!lockres);
4184 BUG_ON(!lockres->l_ops);
Mark Fashehccd979b2005-12-15 14:31:24 -08004185
Sunil Mushran9b915182010-02-26 19:42:44 -08004186 mlog(ML_BASTS, "lockres %s blocked\n", lockres->l_name);
Mark Fashehccd979b2005-12-15 14:31:24 -08004187
4188 /* Detect whether a lock has been marked as going away while
Mark Fasheh34d024f2007-09-24 15:56:19 -07004189 * the downconvert thread was processing other things. A lock can
Mark Fashehccd979b2005-12-15 14:31:24 -08004190 * still be marked with OCFS2_LOCK_FREEING after this check,
4191 * but short circuiting here will still save us some
4192 * performance. */
4193 spin_lock_irqsave(&lockres->l_lock, flags);
4194 if (lockres->l_flags & OCFS2_LOCK_FREEING)
4195 goto unqueue;
4196 spin_unlock_irqrestore(&lockres->l_lock, flags);
4197
Mark Fashehb5e500e2006-09-13 22:01:16 -07004198 status = ocfs2_unblock_lock(osb, lockres, &ctl);
Mark Fashehccd979b2005-12-15 14:31:24 -08004199 if (status < 0)
4200 mlog_errno(status);
4201
4202 spin_lock_irqsave(&lockres->l_lock, flags);
4203unqueue:
Mark Fashehd680efe2006-09-08 14:14:34 -07004204 if (lockres->l_flags & OCFS2_LOCK_FREEING || !ctl.requeue) {
Mark Fashehccd979b2005-12-15 14:31:24 -08004205 lockres_clear_flags(lockres, OCFS2_LOCK_QUEUED);
4206 } else
4207 ocfs2_schedule_blocked_lock(osb, lockres);
4208
Sunil Mushran9b915182010-02-26 19:42:44 -08004209 mlog(ML_BASTS, "lockres %s, requeue = %s.\n", lockres->l_name,
Mark Fashehd680efe2006-09-08 14:14:34 -07004210 ctl.requeue ? "yes" : "no");
Mark Fashehccd979b2005-12-15 14:31:24 -08004211 spin_unlock_irqrestore(&lockres->l_lock, flags);
4212
Mark Fashehd680efe2006-09-08 14:14:34 -07004213 if (ctl.unblock_action != UNBLOCK_CONTINUE
4214 && lockres->l_ops->post_unlock)
4215 lockres->l_ops->post_unlock(osb, lockres);
Mark Fashehccd979b2005-12-15 14:31:24 -08004216}
4217
4218static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb,
4219 struct ocfs2_lock_res *lockres)
4220{
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004221 unsigned long flags;
4222
Mark Fashehccd979b2005-12-15 14:31:24 -08004223 assert_spin_locked(&lockres->l_lock);
4224
4225 if (lockres->l_flags & OCFS2_LOCK_FREEING) {
4226 /* Do not schedule a lock for downconvert when it's on
4227 * the way to destruction - any nodes wanting access
4228 * to the resource will get it soon. */
Sunil Mushran9b915182010-02-26 19:42:44 -08004229 mlog(ML_BASTS, "lockres %s won't be scheduled: flags 0x%lx\n",
Mark Fashehccd979b2005-12-15 14:31:24 -08004230 lockres->l_name, lockres->l_flags);
4231 return;
4232 }
4233
4234 lockres_or_flags(lockres, OCFS2_LOCK_QUEUED);
4235
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004236 spin_lock_irqsave(&osb->dc_task_lock, flags);
Mark Fashehccd979b2005-12-15 14:31:24 -08004237 if (list_empty(&lockres->l_blocked_list)) {
4238 list_add_tail(&lockres->l_blocked_list,
4239 &osb->blocked_lock_list);
4240 osb->blocked_lock_count++;
4241 }
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004242 spin_unlock_irqrestore(&osb->dc_task_lock, flags);
Mark Fashehccd979b2005-12-15 14:31:24 -08004243}
Mark Fasheh34d024f2007-09-24 15:56:19 -07004244
4245static void ocfs2_downconvert_thread_do_work(struct ocfs2_super *osb)
4246{
4247 unsigned long processed;
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004248 unsigned long flags;
Mark Fasheh34d024f2007-09-24 15:56:19 -07004249 struct ocfs2_lock_res *lockres;
4250
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004251 spin_lock_irqsave(&osb->dc_task_lock, flags);
Mark Fasheh34d024f2007-09-24 15:56:19 -07004252 /* grab this early so we know to try again if a state change and
4253 * wake happens part-way through our work */
4254 osb->dc_work_sequence = osb->dc_wake_sequence;
4255
4256 processed = osb->blocked_lock_count;
Joseph Qi209f7512d02015-08-06 15:46:23 -07004257 /*
4258 * blocked lock processing in this loop might call iput which can
4259 * remove items off osb->blocked_lock_list. Downconvert up to
4260 * 'processed' number of locks, but stop short if we had some
4261 * removed in ocfs2_mark_lockres_freeing when downconverting.
4262 */
4263 while (processed && !list_empty(&osb->blocked_lock_list)) {
Mark Fasheh34d024f2007-09-24 15:56:19 -07004264 lockres = list_entry(osb->blocked_lock_list.next,
4265 struct ocfs2_lock_res, l_blocked_list);
4266 list_del_init(&lockres->l_blocked_list);
4267 osb->blocked_lock_count--;
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004268 spin_unlock_irqrestore(&osb->dc_task_lock, flags);
Mark Fasheh34d024f2007-09-24 15:56:19 -07004269
4270 BUG_ON(!processed);
4271 processed--;
4272
4273 ocfs2_process_blocked_lock(osb, lockres);
4274
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004275 spin_lock_irqsave(&osb->dc_task_lock, flags);
Mark Fasheh34d024f2007-09-24 15:56:19 -07004276 }
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004277 spin_unlock_irqrestore(&osb->dc_task_lock, flags);
Mark Fasheh34d024f2007-09-24 15:56:19 -07004278}
4279
4280static int ocfs2_downconvert_thread_lists_empty(struct ocfs2_super *osb)
4281{
4282 int empty = 0;
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004283 unsigned long flags;
Mark Fasheh34d024f2007-09-24 15:56:19 -07004284
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004285 spin_lock_irqsave(&osb->dc_task_lock, flags);
Mark Fasheh34d024f2007-09-24 15:56:19 -07004286 if (list_empty(&osb->blocked_lock_list))
4287 empty = 1;
4288
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004289 spin_unlock_irqrestore(&osb->dc_task_lock, flags);
Mark Fasheh34d024f2007-09-24 15:56:19 -07004290 return empty;
4291}
4292
4293static int ocfs2_downconvert_thread_should_wake(struct ocfs2_super *osb)
4294{
4295 int should_wake = 0;
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004296 unsigned long flags;
Mark Fasheh34d024f2007-09-24 15:56:19 -07004297
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004298 spin_lock_irqsave(&osb->dc_task_lock, flags);
Mark Fasheh34d024f2007-09-24 15:56:19 -07004299 if (osb->dc_work_sequence != osb->dc_wake_sequence)
4300 should_wake = 1;
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004301 spin_unlock_irqrestore(&osb->dc_task_lock, flags);
Mark Fasheh34d024f2007-09-24 15:56:19 -07004302
4303 return should_wake;
4304}
4305
Adrian Bunk200bfae2008-02-17 10:20:38 +02004306static int ocfs2_downconvert_thread(void *arg)
Mark Fasheh34d024f2007-09-24 15:56:19 -07004307{
4308 int status = 0;
4309 struct ocfs2_super *osb = arg;
4310
4311 /* only quit once we've been asked to stop and there is no more
4312 * work available */
4313 while (!(kthread_should_stop() &&
4314 ocfs2_downconvert_thread_lists_empty(osb))) {
4315
4316 wait_event_interruptible(osb->dc_event,
4317 ocfs2_downconvert_thread_should_wake(osb) ||
4318 kthread_should_stop());
4319
4320 mlog(0, "downconvert_thread: awoken\n");
4321
4322 ocfs2_downconvert_thread_do_work(osb);
4323 }
4324
4325 osb->dc_task = NULL;
4326 return status;
4327}
4328
4329void ocfs2_wake_downconvert_thread(struct ocfs2_super *osb)
4330{
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004331 unsigned long flags;
4332
4333 spin_lock_irqsave(&osb->dc_task_lock, flags);
Mark Fasheh34d024f2007-09-24 15:56:19 -07004334 /* make sure the voting thread gets a swipe at whatever changes
4335 * the caller may have made to the voting state */
4336 osb->dc_wake_sequence++;
Srinivas Eedaa75e9cc2012-01-30 21:51:22 -08004337 spin_unlock_irqrestore(&osb->dc_task_lock, flags);
Mark Fasheh34d024f2007-09-24 15:56:19 -07004338 wake_up(&osb->dc_event);
4339}