glock.c 45.8 KB
Newer Older
David Teigland's avatar
David Teigland committed
1 2
/*
 * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
3
 * Copyright (C) 2004-2008 Red Hat, Inc.  All rights reserved.
David Teigland's avatar
David Teigland committed
4 5 6
 *
 * This copyrighted material is made available to anyone wishing to use,
 * modify, copy, or redistribute it subject to the terms and conditions
7
 * of the GNU General Public License version 2.
David Teigland's avatar
David Teigland committed
8 9 10 11 12 13 14 15 16
 */

#include <linux/sched.h>
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/buffer_head.h>
#include <linux/delay.h>
#include <linux/sort.h>
#include <linux/jhash.h>
17
#include <linux/kallsyms.h>
18
#include <linux/gfs2_ondisk.h>
19
#include <linux/list.h>
20
#include <linux/wait.h>
akpm@linux-foundation.org's avatar
akpm@linux-foundation.org committed
21
#include <linux/module.h>
David Teigland's avatar
David Teigland committed
22
#include <asm/uaccess.h>
23 24
#include <linux/seq_file.h>
#include <linux/debugfs.h>
25 26
#include <linux/kthread.h>
#include <linux/freezer.h>
27 28
#include <linux/workqueue.h>
#include <linux/jiffies.h>
29 30 31
#include <linux/rcupdate.h>
#include <linux/rculist_bl.h>
#include <linux/bit_spinlock.h>
David Teigland's avatar
David Teigland committed
32 33

#include "gfs2.h"
34
#include "incore.h"
David Teigland's avatar
David Teigland committed
35 36 37 38 39 40 41
#include "glock.h"
#include "glops.h"
#include "inode.h"
#include "lops.h"
#include "meta_io.h"
#include "quota.h"
#include "super.h"
42
#include "util.h"
43
#include "bmap.h"
Steven Whitehouse's avatar
Steven Whitehouse committed
44 45
#define CREATE_TRACE_POINTS
#include "trace_gfs2.h"
David Teigland's avatar
David Teigland committed
46

47 48 49 50 51
struct gfs2_glock_iter {
	int hash;			/* hash bucket index         */
	struct gfs2_sbd *sdp;		/* incore superblock         */
	struct gfs2_glock *gl;		/* current glock struct      */
	char string[512];		/* scratch space             */
52 53
};

David Teigland's avatar
David Teigland committed
54 55
typedef void (*glock_examiner) (struct gfs2_glock * gl);

56 57 58
static int __dump_glock(struct seq_file *seq, const struct gfs2_glock *gl);
#define GLOCK_BUG_ON(gl,x) do { if (unlikely(x)) { __dump_glock(NULL, gl); BUG(); } } while(0)
static void do_xmote(struct gfs2_glock *gl, struct gfs2_holder *gh, unsigned int target);
59

60
static struct dentry *gfs2_root;
61
static struct workqueue_struct *glock_workqueue;
62
struct workqueue_struct *gfs2_delete_workqueue;
63 64
static LIST_HEAD(lru_list);
static atomic_t lru_count = ATOMIC_INIT(0);
Julia Lawall's avatar
Julia Lawall committed
65
static DEFINE_SPINLOCK(lru_lock);
66

67
#define GFS2_GL_HASH_SHIFT      15
68 69 70
#define GFS2_GL_HASH_SIZE       (1 << GFS2_GL_HASH_SHIFT)
#define GFS2_GL_HASH_MASK       (GFS2_GL_HASH_SIZE - 1)

71
static struct hlist_bl_head gl_hash_table[GFS2_GL_HASH_SIZE];
72
static struct dentry *gfs2_root;
73

David Teigland's avatar
David Teigland committed
74 75 76 77 78 79 80
/**
 * gl_hash() - Turn glock number into hash bucket number
 * @lock: The glock number
 *
 * Returns: The number of the corresponding hash bucket
 */

81 82
static unsigned int gl_hash(const struct gfs2_sbd *sdp,
			    const struct lm_lockname *name)
David Teigland's avatar
David Teigland committed
83 84 85
{
	unsigned int h;

86
	h = jhash(&name->ln_number, sizeof(u64), 0);
David Teigland's avatar
David Teigland committed
87
	h = jhash(&name->ln_type, sizeof(unsigned int), h);
88
	h = jhash(&sdp, sizeof(struct gfs2_sbd *), h);
David Teigland's avatar
David Teigland committed
89 90 91 92 93
	h &= GFS2_GL_HASH_MASK;

	return h;
}

94 95
static inline void spin_lock_bucket(unsigned int hash)
{
96
	hlist_bl_lock(&gl_hash_table[hash]);
97
}
David Teigland's avatar
David Teigland committed
98

99 100
static inline void spin_unlock_bucket(unsigned int hash)
{
101
	hlist_bl_unlock(&gl_hash_table[hash]);
102
}
David Teigland's avatar
David Teigland committed
103

104
static void gfs2_glock_dealloc(struct rcu_head *rcu)
David Teigland's avatar
David Teigland committed
105
{
106
	struct gfs2_glock *gl = container_of(rcu, struct gfs2_glock, gl_rcu);
David Teigland's avatar
David Teigland committed
107

108 109 110 111
	if (gl->gl_ops->go_flags & GLOF_ASPACE)
		kmem_cache_free(gfs2_glock_aspace_cachep, gl);
	else
		kmem_cache_free(gfs2_glock_cachep, gl);
112 113 114
}

void gfs2_glock_free(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
115 116 117
{
	struct gfs2_sbd *sdp = gl->gl_sbd;

118
	call_rcu(&gl->gl_rcu, gfs2_glock_dealloc);
119 120
	if (atomic_dec_and_test(&sdp->sd_glock_disposal))
		wake_up(&sdp->sd_glock_wait);
David Teigland's avatar
David Teigland committed
121 122 123 124 125 126 127 128
}

/**
 * gfs2_glock_hold() - increment reference count on glock
 * @gl: The glock to hold
 *
 */

129
void gfs2_glock_hold(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
130
{
131
	GLOCK_BUG_ON(gl, atomic_read(&gl->gl_ref) == 0);
132
	atomic_inc(&gl->gl_ref);
David Teigland's avatar
David Teigland committed
133 134
}

135 136 137 138 139 140 141 142 143 144 145 146 147
/**
 * demote_ok - Check to see if it's ok to unlock a glock
 * @gl: the glock
 *
 * Returns: 1 if it's ok
 */

static int demote_ok(const struct gfs2_glock *gl)
{
	const struct gfs2_glock_operations *glops = gl->gl_ops;

	if (gl->gl_state == LM_ST_UNLOCKED)
		return 0;
148
	if (!list_empty(&gl->gl_holders))
149 150 151 152 153 154
		return 0;
	if (glops->go_demote_ok)
		return glops->go_demote_ok(gl);
	return 1;
}

155

156 157 158 159 160 161 162 163 164 165
void gfs2_glock_add_to_lru(struct gfs2_glock *gl)
{
	spin_lock(&lru_lock);

	if (!list_empty(&gl->gl_lru))
		list_del_init(&gl->gl_lru);
	else
		atomic_inc(&lru_count);

	list_add_tail(&gl->gl_lru, &lru_list);
166
	set_bit(GLF_LRU, &gl->gl_flags);
167 168 169
	spin_unlock(&lru_lock);
}

170
static void __gfs2_glock_remove_from_lru(struct gfs2_glock *gl)
171 172 173 174 175 176
{
	if (!list_empty(&gl->gl_lru)) {
		list_del_init(&gl->gl_lru);
		atomic_dec(&lru_count);
		clear_bit(GLF_LRU, &gl->gl_flags);
	}
177 178 179 180 181 182
}

static void gfs2_glock_remove_from_lru(struct gfs2_glock *gl)
{
	spin_lock(&lru_lock);
	__gfs2_glock_remove_from_lru(gl);
183 184 185
	spin_unlock(&lru_lock);
}

186
/**
187
 * __gfs2_glock_schedule_for_reclaim - Add a glock to the reclaim list
188 189
 * @gl: the glock
 *
190 191
 * If the glock is demotable, then we add it (or move it) to the end
 * of the glock LRU list.
192 193
 */

194
static void __gfs2_glock_schedule_for_reclaim(struct gfs2_glock *gl)
195
{
196 197
	if (demote_ok(gl))
		gfs2_glock_add_to_lru(gl);
198 199
}

200 201 202 203 204 205 206 207
/**
 * gfs2_glock_put_nolock() - Decrement reference count on glock
 * @gl: The glock to put
 *
 * This function should only be used if the caller has its own reference
 * to the glock, in addition to the one it is dropping.
 */

208
void gfs2_glock_put_nolock(struct gfs2_glock *gl)
209 210 211 212 213
{
	if (atomic_dec_and_test(&gl->gl_ref))
		GLOCK_BUG_ON(gl, 1);
}

David Teigland's avatar
David Teigland committed
214 215 216 217 218 219
/**
 * gfs2_glock_put() - Decrement reference count on glock
 * @gl: The glock to put
 *
 */

220
void gfs2_glock_put(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
221
{
222 223
	struct gfs2_sbd *sdp = gl->gl_sbd;
	struct address_space *mapping = gfs2_glock2aspace(gl);
David Teigland's avatar
David Teigland committed
224

225 226 227
	if (atomic_dec_and_lock(&gl->gl_ref, &lru_lock)) {
		__gfs2_glock_remove_from_lru(gl);
		spin_unlock(&lru_lock);
228 229 230
		spin_lock_bucket(gl->gl_hash);
		hlist_bl_del_rcu(&gl->gl_list);
		spin_unlock_bucket(gl->gl_hash);
231
		GLOCK_BUG_ON(gl, !list_empty(&gl->gl_holders));
232 233 234
		GLOCK_BUG_ON(gl, mapping && mapping->nrpages);
		trace_gfs2_glock_put(gl);
		sdp->sd_lockstruct.ls_ops->lm_put_lock(gl);
David Teigland's avatar
David Teigland committed
235 236 237 238 239 240 241 242 243 244 245
	}
}

/**
 * search_bucket() - Find struct gfs2_glock by lock number
 * @bucket: the bucket to search
 * @name: The lock name
 *
 * Returns: NULL, or the struct gfs2_glock with the requested number
 */

246
static struct gfs2_glock *search_bucket(unsigned int hash,
247
					const struct gfs2_sbd *sdp,
248
					const struct lm_lockname *name)
David Teigland's avatar
David Teigland committed
249 250
{
	struct gfs2_glock *gl;
251
	struct hlist_bl_node *h;
David Teigland's avatar
David Teigland committed
252

253
	hlist_bl_for_each_entry_rcu(gl, h, &gl_hash_table[hash], gl_list) {
David Teigland's avatar
David Teigland committed
254 255
		if (!lm_name_equal(&gl->gl_name, name))
			continue;
256 257
		if (gl->gl_sbd != sdp)
			continue;
258 259
		if (atomic_inc_not_zero(&gl->gl_ref))
			return gl;
David Teigland's avatar
David Teigland committed
260 261 262 263 264
	}

	return NULL;
}

265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
/**
 * may_grant - check if its ok to grant a new lock
 * @gl: The glock
 * @gh: The lock request which we wish to grant
 *
 * Returns: true if its ok to grant the lock
 */

static inline int may_grant(const struct gfs2_glock *gl, const struct gfs2_holder *gh)
{
	const struct gfs2_holder *gh_head = list_entry(gl->gl_holders.next, const struct gfs2_holder, gh_list);
	if ((gh->gh_state == LM_ST_EXCLUSIVE ||
	     gh_head->gh_state == LM_ST_EXCLUSIVE) && gh != gh_head)
		return 0;
	if (gl->gl_state == gh->gh_state)
		return 1;
	if (gh->gh_flags & GL_EXACT)
		return 0;
283 284 285 286 287 288
	if (gl->gl_state == LM_ST_EXCLUSIVE) {
		if (gh->gh_state == LM_ST_SHARED && gh_head->gh_state == LM_ST_SHARED)
			return 1;
		if (gh->gh_state == LM_ST_DEFERRED && gh_head->gh_state == LM_ST_DEFERRED)
			return 1;
	}
289 290 291 292 293 294 295 296 297 298 299 300
	if (gl->gl_state != LM_ST_UNLOCKED && (gh->gh_flags & LM_FLAG_ANY))
		return 1;
	return 0;
}

static void gfs2_holder_wake(struct gfs2_holder *gh)
{
	clear_bit(HIF_WAIT, &gh->gh_iflags);
	smp_mb__after_clear_bit();
	wake_up_bit(&gh->gh_iflags, HIF_WAIT);
}

301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
/**
 * do_error - Something unexpected has happened during a lock request
 *
 */

static inline void do_error(struct gfs2_glock *gl, const int ret)
{
	struct gfs2_holder *gh, *tmp;

	list_for_each_entry_safe(gh, tmp, &gl->gl_holders, gh_list) {
		if (test_bit(HIF_HOLDER, &gh->gh_iflags))
			continue;
		if (ret & LM_OUT_ERROR)
			gh->gh_error = -EIO;
		else if (gh->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB))
			gh->gh_error = GLR_TRYFAILED;
		else
			continue;
		list_del_init(&gh->gh_list);
		trace_gfs2_glock_queue(gh, 0);
		gfs2_holder_wake(gh);
	}
}

325 326 327 328
/**
 * do_promote - promote as many requests as possible on the current queue
 * @gl: The glock
 * 
329 330
 * Returns: 1 if there is a blocked holder at the head of the list, or 2
 *          if a type specific operation is underway.
331 332 333
 */

static int do_promote(struct gfs2_glock *gl)
334 335
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
{
	const struct gfs2_glock_operations *glops = gl->gl_ops;
	struct gfs2_holder *gh, *tmp;
	int ret;

restart:
	list_for_each_entry_safe(gh, tmp, &gl->gl_holders, gh_list) {
		if (test_bit(HIF_HOLDER, &gh->gh_iflags))
			continue;
		if (may_grant(gl, gh)) {
			if (gh->gh_list.prev == &gl->gl_holders &&
			    glops->go_lock) {
				spin_unlock(&gl->gl_spin);
				/* FIXME: eliminate this eventually */
				ret = glops->go_lock(gh);
				spin_lock(&gl->gl_spin);
				if (ret) {
353 354
					if (ret == 1)
						return 2;
355 356
					gh->gh_error = ret;
					list_del_init(&gh->gh_list);
Steven Whitehouse's avatar
Steven Whitehouse committed
357
					trace_gfs2_glock_queue(gh, 0);
358 359 360 361
					gfs2_holder_wake(gh);
					goto restart;
				}
				set_bit(HIF_HOLDER, &gh->gh_iflags);
Steven Whitehouse's avatar
Steven Whitehouse committed
362
				trace_gfs2_promote(gh, 1);
363 364 365 366
				gfs2_holder_wake(gh);
				goto restart;
			}
			set_bit(HIF_HOLDER, &gh->gh_iflags);
Steven Whitehouse's avatar
Steven Whitehouse committed
367
			trace_gfs2_promote(gh, 0);
368 369 370 371 372
			gfs2_holder_wake(gh);
			continue;
		}
		if (gh->gh_list.prev == &gl->gl_holders)
			return 1;
373
		do_error(gl, 0);
374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
		break;
	}
	return 0;
}

/**
 * find_first_waiter - find the first gh that's waiting for the glock
 * @gl: the glock
 */

static inline struct gfs2_holder *find_first_waiter(const struct gfs2_glock *gl)
{
	struct gfs2_holder *gh;

	list_for_each_entry(gh, &gl->gl_holders, gh_list) {
		if (!test_bit(HIF_HOLDER, &gh->gh_iflags))
			return gh;
	}
	return NULL;
}

/**
 * state_change - record that the glock is now in a different state
 * @gl: the glock
 * @new_state the new state
 *
 */

static void state_change(struct gfs2_glock *gl, unsigned int new_state)
{
	int held1, held2;

	held1 = (gl->gl_state != LM_ST_UNLOCKED);
	held2 = (new_state != LM_ST_UNLOCKED);

	if (held1 != held2) {
		if (held2)
			gfs2_glock_hold(gl);
		else
413
			gfs2_glock_put_nolock(gl);
414
	}
415 416
	if (held1 && held2 && list_empty(&gl->gl_holders))
		clear_bit(GLF_QUEUED, &gl->gl_flags);
417

418 419 420 421
	if (new_state != gl->gl_target)
		/* shorten our minimum hold time */
		gl->gl_hold_time = max(gl->gl_hold_time - GL_GLOCK_HOLD_DECR,
				       GL_GLOCK_MIN_HOLD);
422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445
	gl->gl_state = new_state;
	gl->gl_tchange = jiffies;
}

static void gfs2_demote_wake(struct gfs2_glock *gl)
{
	gl->gl_demote_state = LM_ST_EXCLUSIVE;
	clear_bit(GLF_DEMOTE, &gl->gl_flags);
	smp_mb__after_clear_bit();
	wake_up_bit(&gl->gl_flags, GLF_DEMOTE);
}

/**
 * finish_xmote - The DLM has replied to one of our lock requests
 * @gl: The glock
 * @ret: The status from the DLM
 *
 */

static void finish_xmote(struct gfs2_glock *gl, unsigned int ret)
{
	const struct gfs2_glock_operations *glops = gl->gl_ops;
	struct gfs2_holder *gh;
	unsigned state = ret & LM_OUT_ST_MASK;
446
	int rv;
447 448

	spin_lock(&gl->gl_spin);
Steven Whitehouse's avatar
Steven Whitehouse committed
449
	trace_gfs2_glock_state_change(gl, state);
450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508
	state_change(gl, state);
	gh = find_first_waiter(gl);

	/* Demote to UN request arrived during demote to SH or DF */
	if (test_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags) &&
	    state != LM_ST_UNLOCKED && gl->gl_demote_state == LM_ST_UNLOCKED)
		gl->gl_target = LM_ST_UNLOCKED;

	/* Check for state != intended state */
	if (unlikely(state != gl->gl_target)) {
		if (gh && !test_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags)) {
			/* move to back of queue and try next entry */
			if (ret & LM_OUT_CANCELED) {
				if ((gh->gh_flags & LM_FLAG_PRIORITY) == 0)
					list_move_tail(&gh->gh_list, &gl->gl_holders);
				gh = find_first_waiter(gl);
				gl->gl_target = gh->gh_state;
				goto retry;
			}
			/* Some error or failed "try lock" - report it */
			if ((ret & LM_OUT_ERROR) ||
			    (gh->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB))) {
				gl->gl_target = gl->gl_state;
				do_error(gl, ret);
				goto out;
			}
		}
		switch(state) {
		/* Unlocked due to conversion deadlock, try again */
		case LM_ST_UNLOCKED:
retry:
			do_xmote(gl, gh, gl->gl_target);
			break;
		/* Conversion fails, unlock and try again */
		case LM_ST_SHARED:
		case LM_ST_DEFERRED:
			do_xmote(gl, gh, LM_ST_UNLOCKED);
			break;
		default: /* Everything else */
			printk(KERN_ERR "GFS2: wanted %u got %u\n", gl->gl_target, state);
			GLOCK_BUG_ON(gl, 1);
		}
		spin_unlock(&gl->gl_spin);
		return;
	}

	/* Fast path - we got what we asked for */
	if (test_and_clear_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags))
		gfs2_demote_wake(gl);
	if (state != LM_ST_UNLOCKED) {
		if (glops->go_xmote_bh) {
			spin_unlock(&gl->gl_spin);
			rv = glops->go_xmote_bh(gl, gh);
			spin_lock(&gl->gl_spin);
			if (rv) {
				do_error(gl, rv);
				goto out;
			}
		}
509 510 511
		rv = do_promote(gl);
		if (rv == 2)
			goto out_locked;
512 513 514
	}
out:
	clear_bit(GLF_LOCK, &gl->gl_flags);
515
out_locked:
516 517 518 519 520 521 522 523 524 525 526 527
	spin_unlock(&gl->gl_spin);
}

/**
 * do_xmote - Calls the DLM to change the state of a lock
 * @gl: The lock state
 * @gh: The holder (only for promotes)
 * @target: The target lock state
 *
 */

static void do_xmote(struct gfs2_glock *gl, struct gfs2_holder *gh, unsigned int target)
528 529
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
530 531 532 533 534 535 536 537
{
	const struct gfs2_glock_operations *glops = gl->gl_ops;
	struct gfs2_sbd *sdp = gl->gl_sbd;
	unsigned int lck_flags = gh ? gh->gh_flags : 0;
	int ret;

	lck_flags &= (LM_FLAG_TRY | LM_FLAG_TRY_1CB | LM_FLAG_NOEXP |
		      LM_FLAG_PRIORITY);
538 539
	GLOCK_BUG_ON(gl, gl->gl_state == target);
	GLOCK_BUG_ON(gl, gl->gl_state == gl->gl_target);
540 541 542 543 544
	if ((target == LM_ST_UNLOCKED || target == LM_ST_DEFERRED) &&
	    glops->go_inval) {
		set_bit(GLF_INVALIDATE_IN_PROGRESS, &gl->gl_flags);
		do_error(gl, 0); /* Fail queued try locks */
	}
545
	gl->gl_req = target;
546 547 548 549 550 551 552 553
	spin_unlock(&gl->gl_spin);
	if (glops->go_xmote_th)
		glops->go_xmote_th(gl);
	if (test_bit(GLF_INVALIDATE_IN_PROGRESS, &gl->gl_flags))
		glops->go_inval(gl, target == LM_ST_DEFERRED ? 0 : DIO_METADATA);
	clear_bit(GLF_INVALIDATE_IN_PROGRESS, &gl->gl_flags);

	gfs2_glock_hold(gl);
554 555 556 557 558 559
	if (sdp->sd_lockstruct.ls_ops->lm_lock)	{
		/* lock_dlm */
		ret = sdp->sd_lockstruct.ls_ops->lm_lock(gl, target, lck_flags);
		GLOCK_BUG_ON(gl, ret);
	} else { /* lock_nolock */
		finish_xmote(gl, target);
560 561 562
		if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
			gfs2_glock_put(gl);
	}
563

564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591
	spin_lock(&gl->gl_spin);
}

/**
 * find_first_holder - find the first "holder" gh
 * @gl: the glock
 */

static inline struct gfs2_holder *find_first_holder(const struct gfs2_glock *gl)
{
	struct gfs2_holder *gh;

	if (!list_empty(&gl->gl_holders)) {
		gh = list_entry(gl->gl_holders.next, struct gfs2_holder, gh_list);
		if (test_bit(HIF_HOLDER, &gh->gh_iflags))
			return gh;
	}
	return NULL;
}

/**
 * run_queue - do all outstanding tasks related to a glock
 * @gl: The glock in question
 * @nonblock: True if we must not block in run_queue
 *
 */

static void run_queue(struct gfs2_glock *gl, const int nonblock)
592 593
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
594 595
{
	struct gfs2_holder *gh = NULL;
596
	int ret;
597 598 599 600 601 602 603 604 605

	if (test_and_set_bit(GLF_LOCK, &gl->gl_flags))
		return;

	GLOCK_BUG_ON(gl, test_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags));

	if (test_bit(GLF_DEMOTE, &gl->gl_flags) &&
	    gl->gl_demote_state != gl->gl_state) {
		if (find_first_holder(gl))
606
			goto out_unlock;
607 608 609
		if (nonblock)
			goto out_sched;
		set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags);
610
		GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE);
611 612 613 614
		gl->gl_target = gl->gl_demote_state;
	} else {
		if (test_bit(GLF_DEMOTE, &gl->gl_flags))
			gfs2_demote_wake(gl);
615 616
		ret = do_promote(gl);
		if (ret == 0)
617
			goto out_unlock;
618
		if (ret == 2)
619
			goto out;
620 621 622 623 624 625
		gh = find_first_waiter(gl);
		gl->gl_target = gh->gh_state;
		if (!(gh->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB)))
			do_error(gl, 0); /* Fail queued try locks */
	}
	do_xmote(gl, gh, gl->gl_target);
626
out:
627 628 629
	return;

out_sched:
630 631
	clear_bit(GLF_LOCK, &gl->gl_flags);
	smp_mb__after_clear_bit();
632 633
	gfs2_glock_hold(gl);
	if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
634
		gfs2_glock_put_nolock(gl);
635 636
	return;

637
out_unlock:
638
	clear_bit(GLF_LOCK, &gl->gl_flags);
639 640
	smp_mb__after_clear_bit();
	return;
641 642
}

643 644 645 646
static void delete_work_func(struct work_struct *work)
{
	struct gfs2_glock *gl = container_of(work, struct gfs2_glock, gl_delete);
	struct gfs2_sbd *sdp = gl->gl_sbd;
647
	struct gfs2_inode *ip;
648
	struct inode *inode;
649 650 651 652
	u64 no_addr = gl->gl_name.ln_number;

	ip = gl->gl_object;
	/* Note: Unsafe to dereference ip as we don't hold right refs/locks */
653 654

	if (ip)
655
		inode = gfs2_ilookup(sdp->sd_vfs, no_addr, 1);
656 657 658 659 660
	else
		inode = gfs2_lookup_by_inum(sdp, no_addr, NULL, GFS2_BLKST_UNLINKED);
	if (inode && !IS_ERR(inode)) {
		d_prune_aliases(inode);
		iput(inode);
661 662 663 664
	}
	gfs2_glock_put(gl);
}

665 666
static void glock_work_func(struct work_struct *work)
{
667
	unsigned long delay = 0;
668
	struct gfs2_glock *gl = container_of(work, struct gfs2_glock, gl_work.work);
669
	int drop_ref = 0;
670

671
	if (test_and_clear_bit(GLF_REPLY_PENDING, &gl->gl_flags)) {
672
		finish_xmote(gl, gl->gl_reply);
673 674
		drop_ref = 1;
	}
675
	spin_lock(&gl->gl_spin);
676
	if (test_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) &&
677 678
	    gl->gl_state != LM_ST_UNLOCKED &&
	    gl->gl_demote_state != LM_ST_EXCLUSIVE) {
679
		unsigned long holdtime, now = jiffies;
680

681
		holdtime = gl->gl_tchange + gl->gl_hold_time;
682 683
		if (time_before(now, holdtime))
			delay = holdtime - now;
684 685 686 687 688

		if (!delay) {
			clear_bit(GLF_PENDING_DEMOTE, &gl->gl_flags);
			set_bit(GLF_DEMOTE, &gl->gl_flags);
		}
689 690
	}
	run_queue(gl, 0);
691
	spin_unlock(&gl->gl_spin);
692
	if (!delay)
693
		gfs2_glock_put(gl);
694 695 696 697 698 699
	else {
		if (gl->gl_name.ln_type != LM_TYPE_INODE)
			delay = 0;
		if (queue_delayed_work(glock_workqueue, &gl->gl_work, delay) == 0)
			gfs2_glock_put(gl);
	}
700 701
	if (drop_ref)
		gfs2_glock_put(gl);
702 703
}

David Teigland's avatar
David Teigland committed
704 705 706 707 708 709 710 711 712 713 714 715 716
/**
 * gfs2_glock_get() - Get a glock, or create one if one doesn't exist
 * @sdp: The GFS2 superblock
 * @number: the lock number
 * @glops: The glock_operations to use
 * @create: If 0, don't create the glock if it doesn't exist
 * @glp: the glock is returned here
 *
 * This does not lock a glock, just finds/creates structures for one.
 *
 * Returns: errno
 */

717
int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number,
718
		   const struct gfs2_glock_operations *glops, int create,
David Teigland's avatar
David Teigland committed
719 720
		   struct gfs2_glock **glp)
{
721
	struct super_block *s = sdp->sd_vfs;
722
	struct lm_lockname name = { .ln_number = number, .ln_type = glops->go_type };
David Teigland's avatar
David Teigland committed
723
	struct gfs2_glock *gl, *tmp;
724
	unsigned int hash = gl_hash(sdp, &name);
725
	struct address_space *mapping;
726
	struct kmem_cache *cachep;
David Teigland's avatar
David Teigland committed
727

728
	rcu_read_lock();
729
	gl = search_bucket(hash, sdp, &name);
730
	rcu_read_unlock();
David Teigland's avatar
David Teigland committed
731

732 733
	*glp = gl;
	if (gl)
David Teigland's avatar
David Teigland committed
734
		return 0;
735 736
	if (!create)
		return -ENOENT;
David Teigland's avatar
David Teigland committed
737

738
	if (glops->go_flags & GLOF_ASPACE)
739
		cachep = gfs2_glock_aspace_cachep;
740
	else
741 742
		cachep = gfs2_glock_cachep;
	gl = kmem_cache_alloc(cachep, GFP_KERNEL);
David Teigland's avatar
David Teigland committed
743 744 745
	if (!gl)
		return -ENOMEM;

746
	atomic_inc(&sdp->sd_glock_disposal);
747
	gl->gl_flags = 0;
David Teigland's avatar
David Teigland committed
748
	gl->gl_name = name;
749
	atomic_set(&gl->gl_ref, 1);
David Teigland's avatar
David Teigland committed
750
	gl->gl_state = LM_ST_UNLOCKED;
751
	gl->gl_target = LM_ST_UNLOCKED;
752
	gl->gl_demote_state = LM_ST_EXCLUSIVE;
753
	gl->gl_hash = hash;
David Teigland's avatar
David Teigland committed
754
	gl->gl_ops = glops;
755 756 757
	snprintf(gl->gl_strname, GDLM_STRNAME_BYTES, "%8x%16llx", name.ln_type, (unsigned long long)number);
	memset(&gl->gl_lksb, 0, sizeof(struct dlm_lksb));
	gl->gl_lksb.sb_lvbptr = gl->gl_lvb;
758
	gl->gl_tchange = jiffies;
759
	gl->gl_object = NULL;
David Teigland's avatar
David Teigland committed
760
	gl->gl_sbd = sdp;
761
	gl->gl_hold_time = GL_GLOCK_DFT_HOLD;
762
	INIT_DELAYED_WORK(&gl->gl_work, glock_work_func);
763
	INIT_WORK(&gl->gl_delete, delete_work_func);
David Teigland's avatar
David Teigland committed
764

765 766 767 768 769 770 771 772 773
	mapping = gfs2_glock2aspace(gl);
	if (mapping) {
                mapping->a_ops = &gfs2_meta_aops;
		mapping->host = s->s_bdev->bd_inode;
		mapping->flags = 0;
		mapping_set_gfp_mask(mapping, GFP_NOFS);
		mapping->assoc_mapping = NULL;
		mapping->backing_dev_info = s->s_bdi;
		mapping->writeback_index = 0;
David Teigland's avatar
David Teigland committed
774 775
	}

776
	spin_lock_bucket(hash);
777
	tmp = search_bucket(hash, sdp, &name);
David Teigland's avatar
David Teigland committed
778
	if (tmp) {
779 780
		spin_unlock_bucket(hash);
		kmem_cache_free(cachep, gl);
781
		atomic_dec(&sdp->sd_glock_disposal);
David Teigland's avatar
David Teigland committed
782 783
		gl = tmp;
	} else {
784 785
		hlist_bl_add_head_rcu(&gl->gl_list, &gl_hash_table[hash]);
		spin_unlock_bucket(hash);
David Teigland's avatar
David Teigland committed
786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801
	}

	*glp = gl;

	return 0;
}

/**
 * gfs2_holder_init - initialize a struct gfs2_holder in the default way
 * @gl: the glock
 * @state: the state we're requesting
 * @flags: the modifier flags
 * @gh: the holder structure
 *
 */

802
void gfs2_holder_init(struct gfs2_glock *gl, unsigned int state, unsigned flags,
David Teigland's avatar
David Teigland committed
803 804 805 806
		      struct gfs2_holder *gh)
{
	INIT_LIST_HEAD(&gh->gh_list);
	gh->gh_gl = gl;
807
	gh->gh_ip = (unsigned long)__builtin_return_address(0);
808
	gh->gh_owner_pid = get_pid(task_pid(current));
David Teigland's avatar
David Teigland committed
809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825
	gh->gh_state = state;
	gh->gh_flags = flags;
	gh->gh_error = 0;
	gh->gh_iflags = 0;
	gfs2_glock_hold(gl);
}

/**
 * gfs2_holder_reinit - reinitialize a struct gfs2_holder so we can requeue it
 * @state: the state we're requesting
 * @flags: the modifier flags
 * @gh: the holder structure
 *
 * Don't mess with the glock.
 *
 */

826
void gfs2_holder_reinit(unsigned int state, unsigned flags, struct gfs2_holder *gh)
David Teigland's avatar
David Teigland committed
827 828
{
	gh->gh_state = state;
829
	gh->gh_flags = flags;
830
	gh->gh_iflags = 0;
831
	gh->gh_ip = (unsigned long)__builtin_return_address(0);
Bob Peterson's avatar
Bob Peterson committed
832 833 834
	if (gh->gh_owner_pid)
		put_pid(gh->gh_owner_pid);
	gh->gh_owner_pid = get_pid(task_pid(current));
David Teigland's avatar
David Teigland committed
835 836 837 838 839 840 841 842 843 844
}

/**
 * gfs2_holder_uninit - uninitialize a holder structure (drop glock reference)
 * @gh: the holder structure
 *
 */

void gfs2_holder_uninit(struct gfs2_holder *gh)
{
845
	put_pid(gh->gh_owner_pid);
David Teigland's avatar
David Teigland committed
846 847
	gfs2_glock_put(gh->gh_gl);
	gh->gh_gl = NULL;
848
	gh->gh_ip = 0;
David Teigland's avatar
David Teigland committed
849 850
}

851 852 853 854 855 856 857 858 859 860
/**
 * gfs2_glock_holder_wait
 * @word: unused
 *
 * This function and gfs2_glock_demote_wait both show up in the WCHAN
 * field. Thus I've separated these otherwise identical functions in
 * order to be more informative to the user.
 */

static int gfs2_glock_holder_wait(void *word)
861 862 863 864 865
{
        schedule();
        return 0;
}

866 867 868 869 870 871
static int gfs2_glock_demote_wait(void *word)
{
	schedule();
	return 0;
}

872
static void wait_on_holder(struct gfs2_holder *gh)
873
{
874 875
	unsigned long time1 = jiffies;

876
	might_sleep();
877
	wait_on_bit(&gh->gh_iflags, HIF_WAIT, gfs2_glock_holder_wait, TASK_UNINTERRUPTIBLE);
878 879 880 881 882
	if (time_after(jiffies, time1 + HZ)) /* have we waited > a second? */
		/* Lengthen the minimum hold time. */
		gh->gh_gl->gl_hold_time = min(gh->gh_gl->gl_hold_time +
					      GL_GLOCK_HOLD_INCR,
					      GL_GLOCK_MAX_HOLD);
883 884
}

885
static void wait_on_demote(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
886
{
887
	might_sleep();
888
	wait_on_bit(&gl->gl_flags, GLF_DEMOTE, gfs2_glock_demote_wait, TASK_UNINTERRUPTIBLE);
David Teigland's avatar
David Teigland committed
889 890 891
}

/**
892 893 894
 * handle_callback - process a demote request
 * @gl: the glock
 * @state: the state the caller wants us to change to
David Teigland's avatar
David Teigland committed
895
 *
896 897
 * There are only two requests that we are going to see in actual
 * practise: LM_ST_SHARED and LM_ST_UNLOCKED
David Teigland's avatar
David Teigland committed
898 899
 */

900
static void handle_callback(struct gfs2_glock *gl, unsigned int state,
901
			    unsigned long delay)
David Teigland's avatar
David Teigland committed
902
{
903
	int bit = delay ? GLF_PENDING_DEMOTE : GLF_DEMOTE;
David Teigland's avatar
David Teigland committed
904

905 906 907 908 909 910 911
	set_bit(bit, &gl->gl_flags);
	if (gl->gl_demote_state == LM_ST_EXCLUSIVE) {
		gl->gl_demote_state = state;
		gl->gl_demote_time = jiffies;
	} else if (gl->gl_demote_state != LM_ST_UNLOCKED &&
			gl->gl_demote_state != state) {
		gl->gl_demote_state = LM_ST_UNLOCKED;
David Teigland's avatar
David Teigland committed
912
	}
913 914
	if (gl->gl_ops->go_callback)
		gl->gl_ops->go_callback(gl);
Steven Whitehouse's avatar
Steven Whitehouse committed
915
	trace_gfs2_demote_rq(gl);
David Teigland's avatar
David Teigland committed
916 917 918
}

/**
919
 * gfs2_glock_wait - wait on a glock acquisition
David Teigland's avatar
David Teigland committed
920 921 922 923 924
 * @gh: the glock holder
 *
 * Returns: 0 on success
 */

925
int gfs2_glock_wait(struct gfs2_holder *gh)
David Teigland's avatar
David Teigland committed
926
{
927
	wait_on_holder(gh);
David Teigland's avatar
David Teigland committed
928 929 930
	return gh->gh_error;
}

931
void gfs2_print_dbg(struct seq_file *seq, const char *fmt, ...)
932
{
933
	struct va_format vaf;
934 935 936
	va_list args;

	va_start(args, fmt);
937

938 939
	if (seq) {
		struct gfs2_glock_iter *gi = seq->private;
940
		vsprintf(gi->string, fmt, args);
941 942
		seq_printf(seq, gi->string);
	} else {
943 944 945 946
		vaf.fmt = fmt;
		vaf.va = &args;

		printk(KERN_ERR " %pV", &vaf);
947
	}
948

949 950 951
	va_end(args);
}

David Teigland's avatar
David Teigland committed
952 953 954 955
/**
 * add_to_queue - Add a holder to the wait queue (but look for recursion)
 * @gh: the holder structure to add
 *
956 957 958 959
 * Eventually we should move the recursive locking trap to a
 * debugging option or something like that. This is the fast
 * path and needs to have the minimum number of distractions.
 * 
David Teigland's avatar
David Teigland committed
960 961
 */

962
static inline void add_to_queue(struct gfs2_holder *gh)
963 964
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
David Teigland's avatar
David Teigland committed
965 966
{
	struct gfs2_glock *gl = gh->gh_gl;
967 968 969 970
	struct gfs2_sbd *sdp = gl->gl_sbd;
	struct list_head *insert_pt = NULL;
	struct gfs2_holder *gh2;
	int try_lock = 0;
David Teigland's avatar
David Teigland committed
971

972
	BUG_ON(gh->gh_owner_pid == NULL);
973 974
	if (test_and_set_bit(HIF_WAIT, &gh->gh_iflags))
		BUG();
975

976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993
	if (gh->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB)) {
		if (test_bit(GLF_LOCK, &gl->gl_flags))
			try_lock = 1;
		if (test_bit(GLF_INVALIDATE_IN_PROGRESS, &gl->gl_flags))
			goto fail;
	}

	list_for_each_entry(gh2, &gl->gl_holders, gh_list) {
		if (unlikely(gh2->gh_owner_pid == gh->gh_owner_pid &&
		    (gh->gh_gl->gl_ops->go_type != LM_TYPE_FLOCK)))
			goto trap_recursive;
		if (try_lock &&
		    !(gh2->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB)) &&
		    !may_grant(gl, gh)) {
fail:
			gh->gh_error = GLR_TRYFAILED;
			gfs2_holder_wake(gh);
			return;
994
		}
995 996 997 998 999
		if (test_bit(HIF_HOLDER, &gh2->gh_iflags))
			continue;
		if (unlikely((gh->gh_flags & LM_FLAG_PRIORITY) && !insert_pt))
			insert_pt = &gh2->gh_list;
	}
1000
	set_bit(GLF_QUEUED, &gl->gl_flags);
1001
	trace_gfs2_glock_queue(gh, 1);
1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012
	if (likely(insert_pt == NULL)) {
		list_add_tail(&gh->gh_list, &gl->gl_holders);
		if (unlikely(gh->gh_flags & LM_FLAG_PRIORITY))
			goto do_cancel;
		return;
	}
	list_add_tail(&gh->gh_list, insert_pt);
do_cancel:
	gh = list_entry(gl->gl_holders.next, struct gfs2_holder, gh_list);
	if (!(gh->gh_flags & LM_FLAG_PRIORITY)) {
		spin_unlock(&gl->gl_spin);
Steven Whitehouse's avatar
Steven Whitehouse committed
1013
		if (sdp->sd_lockstruct.ls_ops->lm_cancel)
1014
			sdp->sd_lockstruct.ls_ops->lm_cancel(gl);
1015
		spin_lock(&gl->gl_spin);
David Teigland's avatar
David Teigland committed
1016
	}
1017
	return;
David Teigland's avatar
David Teigland committed
1018

1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029
trap_recursive:
	print_symbol(KERN_ERR "original: %s\n", gh2->gh_ip);
	printk(KERN_ERR "pid: %d\n", pid_nr(gh2->gh_owner_pid));
	printk(KERN_ERR "lock type: %d req lock state : %d\n",
	       gh2->gh_gl->gl_name.ln_type, gh2->gh_state);
	print_symbol(KERN_ERR "new: %s\n", gh->gh_ip);
	printk(KERN_ERR "pid: %d\n", pid_nr(gh->gh_owner_pid));
	printk(KERN_ERR "lock type: %d req lock state : %d\n",
	       gh->gh_gl->gl_name.ln_type, gh->gh_state);
	__dump_glock(NULL, gl);
	BUG();
David Teigland's avatar
David Teigland committed
1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046
}

/**
 * gfs2_glock_nq - enqueue a struct gfs2_holder onto a glock (acquire a glock)
 * @gh: the holder structure
 *
 * if (gh->gh_flags & GL_ASYNC), this never returns an error
 *
 * Returns: 0, GLR_TRYFAILED, or errno on failure
 */

int gfs2_glock_nq(struct gfs2_holder *gh)
{
	struct gfs2_glock *gl = gh->gh_gl;
	struct gfs2_sbd *sdp = gl->gl_sbd;
	int error = 0;

1047
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
David Teigland's avatar
David Teigland committed
1048 1049
		return -EIO;

1050 1051 1052
	if (test_bit(GLF_LRU, &gl->gl_flags))
		gfs2_glock_remove_from_lru(gl);

David Teigland's avatar
David Teigland committed
1053 1054
	spin_lock(&gl->gl_spin);
	add_to_queue(gh);
1055 1056 1057
	if ((LM_FLAG_NOEXP & gh->gh_flags) &&
	    test_and_clear_bit(GLF_FROZEN, &gl->gl_flags))
		set_bit(GLF_REPLY_PENDING, &gl->gl_flags);
1058
	run_queue(gl, 1);
David Teigland's avatar
David Teigland committed
1059 1060
	spin_unlock(&gl->gl_spin);

1061 1062
	if (!(gh->gh_flags & GL_ASYNC))
		error = gfs2_glock_wait(gh);
David Teigland's avatar
David Teigland committed
1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075