glock.c 45 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 96 97 98
static inline void spin_lock_bucket(unsigned int hash)
{
	struct hlist_bl_head *bl = &gl_hash_table[hash];
	bit_spin_lock(0, (unsigned long *)bl);
}
David Teigland's avatar
David Teigland committed
99

100 101 102 103 104
static inline void spin_unlock_bucket(unsigned int hash)
{
	struct hlist_bl_head *bl = &gl_hash_table[hash];
	__bit_spin_unlock(0, (unsigned long *)bl);
}
David Teigland's avatar
David Teigland committed
105

106
static void gfs2_glock_dealloc(struct rcu_head *rcu)
David Teigland's avatar
David Teigland committed
107
{
108
	struct gfs2_glock *gl = container_of(rcu, struct gfs2_glock, gl_rcu);
David Teigland's avatar
David Teigland committed
109

110 111 112 113
	if (gl->gl_ops->go_flags & GLOF_ASPACE)
		kmem_cache_free(gfs2_glock_aspace_cachep, gl);
	else
		kmem_cache_free(gfs2_glock_cachep, gl);
114 115 116
}

void gfs2_glock_free(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
117 118 119
{
	struct gfs2_sbd *sdp = gl->gl_sbd;

120
	call_rcu(&gl->gl_rcu, gfs2_glock_dealloc);
121 122
	if (atomic_dec_and_test(&sdp->sd_glock_disposal))
		wake_up(&sdp->sd_glock_wait);
David Teigland's avatar
David Teigland committed
123 124 125 126 127 128 129 130
}

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

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

148 149
	/* assert_spin_locked(&gl->gl_spin); */

150 151
	if (gl->gl_state == LM_ST_UNLOCKED)
		return 0;
152 153 154 155
	if (test_bit(GLF_LFLUSH, &gl->gl_flags))
		return 0;
	if ((gl->gl_name.ln_type != LM_TYPE_INODE) &&
	    !list_empty(&gl->gl_holders))
156 157 158 159 160 161
		return 0;
	if (glops->go_demote_ok)
		return glops->go_demote_ok(gl);
	return 1;
}

162

163
/**
164
 * __gfs2_glock_schedule_for_reclaim - Add a glock to the reclaim list
165 166
 * @gl: the glock
 *
167 168
 * If the glock is demotable, then we add it (or move it) to the end
 * of the glock LRU list.
169 170
 */

171
static void __gfs2_glock_schedule_for_reclaim(struct gfs2_glock *gl)
172
{
173 174 175 176 177 178 179 180
	if (demote_ok(gl)) {
		spin_lock(&lru_lock);

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

181
		list_add_tail(&gl->gl_lru, &lru_list);
182
		spin_unlock(&lru_lock);
183
	}
184 185 186 187 188 189 190
}

void gfs2_glock_schedule_for_reclaim(struct gfs2_glock *gl)
{
	spin_lock(&gl->gl_spin);
	__gfs2_glock_schedule_for_reclaim(gl);
	spin_unlock(&gl->gl_spin);
191 192
}

193 194 195 196 197 198 199 200
/**
 * 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.
 */

201
void gfs2_glock_put_nolock(struct gfs2_glock *gl)
202 203 204 205 206
{
	if (atomic_dec_and_test(&gl->gl_ref))
		GLOCK_BUG_ON(gl, 1);
}

David Teigland's avatar
David Teigland committed
207 208 209 210 211 212
/**
 * gfs2_glock_put() - Decrement reference count on glock
 * @gl: The glock to put
 *
 */

213
void gfs2_glock_put(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
214
{
215 216
	struct gfs2_sbd *sdp = gl->gl_sbd;
	struct address_space *mapping = gfs2_glock2aspace(gl);
David Teigland's avatar
David Teigland committed
217

218 219 220 221 222
	if (atomic_dec_and_test(&gl->gl_ref)) {
		spin_lock_bucket(gl->gl_hash);
		hlist_bl_del_rcu(&gl->gl_list);
		spin_unlock_bucket(gl->gl_hash);
		spin_lock(&lru_lock);
223 224 225 226 227
		if (!list_empty(&gl->gl_lru)) {
			list_del_init(&gl->gl_lru);
			atomic_dec(&lru_count);
		}
		spin_unlock(&lru_lock);
228
		GLOCK_BUG_ON(gl, !list_empty(&gl->gl_holders));
229 230 231
		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
232 233 234 235 236 237 238 239 240 241 242
	}
}

/**
 * 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
 */

243
static struct gfs2_glock *search_bucket(unsigned int hash,
244
					const struct gfs2_sbd *sdp,
245
					const struct lm_lockname *name)
David Teigland's avatar
David Teigland committed
246 247
{
	struct gfs2_glock *gl;
248
	struct hlist_bl_node *h;
David Teigland's avatar
David Teigland committed
249

250
	hlist_bl_for_each_entry_rcu(gl, h, &gl_hash_table[hash], gl_list) {
David Teigland's avatar
David Teigland committed
251 252
		if (!lm_name_equal(&gl->gl_name, name))
			continue;
253 254
		if (gl->gl_sbd != sdp)
			continue;
255 256
		if (atomic_inc_not_zero(&gl->gl_ref))
			return gl;
David Teigland's avatar
David Teigland committed
257 258 259 260 261
	}

	return NULL;
}

262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279
/**
 * 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;
280 281 282 283 284 285
	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;
	}
286 287 288 289 290 291 292 293 294 295 296 297
	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);
}

298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321
/**
 * 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);
	}
}

322 323 324 325
/**
 * do_promote - promote as many requests as possible on the current queue
 * @gl: The glock
 * 
326 327
 * Returns: 1 if there is a blocked holder at the head of the list, or 2
 *          if a type specific operation is underway.
328 329 330
 */

static int do_promote(struct gfs2_glock *gl)
331 332
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349
{
	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) {
350 351
					if (ret == 1)
						return 2;
352 353
					gh->gh_error = ret;
					list_del_init(&gh->gh_list);
Steven Whitehouse's avatar
Steven Whitehouse committed
354
					trace_gfs2_glock_queue(gh, 0);
355 356 357 358
					gfs2_holder_wake(gh);
					goto restart;
				}
				set_bit(HIF_HOLDER, &gh->gh_iflags);
Steven Whitehouse's avatar
Steven Whitehouse committed
359
				trace_gfs2_promote(gh, 1);
360 361 362 363
				gfs2_holder_wake(gh);
				goto restart;
			}
			set_bit(HIF_HOLDER, &gh->gh_iflags);
Steven Whitehouse's avatar
Steven Whitehouse committed
364
			trace_gfs2_promote(gh, 0);
365 366 367 368 369
			gfs2_holder_wake(gh);
			continue;
		}
		if (gh->gh_list.prev == &gl->gl_holders)
			return 1;
370
		do_error(gl, 0);
371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409
		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
410
			gfs2_glock_put_nolock(gl);
411
	}
412 413
	if (held1 && held2 && list_empty(&gl->gl_holders))
		clear_bit(GLF_QUEUED, &gl->gl_flags);
414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438

	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;
439
	int rv;
440 441

	spin_lock(&gl->gl_spin);
Steven Whitehouse's avatar
Steven Whitehouse committed
442
	trace_gfs2_glock_state_change(gl, state);
443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501
	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;
			}
		}
502 503 504
		rv = do_promote(gl);
		if (rv == 2)
			goto out_locked;
505 506 507
	}
out:
	clear_bit(GLF_LOCK, &gl->gl_flags);
508
out_locked:
509 510 511 512 513 514 515 516 517 518 519 520
	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)
521 522
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
523 524 525 526 527 528 529 530
{
	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);
531 532
	GLOCK_BUG_ON(gl, gl->gl_state == target);
	GLOCK_BUG_ON(gl, gl->gl_state == gl->gl_target);
533 534 535 536 537
	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 */
	}
538
	gl->gl_req = target;
539 540 541 542 543 544 545 546 547 548 549 550 551
	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);
	if (target != LM_ST_UNLOCKED && (gl->gl_state == LM_ST_SHARED ||
	    gl->gl_state == LM_ST_DEFERRED) &&
	    !(lck_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB)))
		lck_flags |= LM_FLAG_TRY_1CB;

552 553 554 555 556 557
	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);
558 559 560
		if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
			gfs2_glock_put(gl);
	}
561

562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589
	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)
590 591
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
592 593
{
	struct gfs2_holder *gh = NULL;
594
	int ret;
595 596 597 598 599 600 601 602 603

	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))
604
			goto out_unlock;
605 606 607
		if (nonblock)
			goto out_sched;
		set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags);
608
		GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE);
609 610 611 612
		gl->gl_target = gl->gl_demote_state;
	} else {
		if (test_bit(GLF_DEMOTE, &gl->gl_flags))
			gfs2_demote_wake(gl);
613 614
		ret = do_promote(gl);
		if (ret == 0)
615
			goto out_unlock;
616
		if (ret == 2)
617
			goto out;
618 619 620 621 622 623
		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);
624
out:
625 626 627
	return;

out_sched:
628 629
	clear_bit(GLF_LOCK, &gl->gl_flags);
	smp_mb__after_clear_bit();
630 631
	gfs2_glock_hold(gl);
	if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
632
		gfs2_glock_put_nolock(gl);
633 634
	return;

635
out_unlock:
636
	clear_bit(GLF_LOCK, &gl->gl_flags);
637 638
	smp_mb__after_clear_bit();
	return;
639 640
}

641 642 643 644
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;
645
	struct gfs2_inode *ip;
646
	struct inode *inode;
647 648 649 650
	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 */
651 652 653

	if (ip)
		inode = gfs2_ilookup(sdp->sd_vfs, no_addr);
654 655 656 657 658
	else
		inode = gfs2_lookup_by_inum(sdp, no_addr, NULL, GFS2_BLKST_UNLINKED);
	if (inode && !IS_ERR(inode)) {
		d_prune_aliases(inode);
		iput(inode);
659 660 661 662
	}
	gfs2_glock_put(gl);
}

663 664
static void glock_work_func(struct work_struct *work)
{
665
	unsigned long delay = 0;
666
	struct gfs2_glock *gl = container_of(work, struct gfs2_glock, gl_work.work);
667
	int drop_ref = 0;
668

669
	if (test_and_clear_bit(GLF_REPLY_PENDING, &gl->gl_flags)) {
670
		finish_xmote(gl, gl->gl_reply);
671 672
		drop_ref = 1;
	}
673
	spin_lock(&gl->gl_spin);
674 675 676
	if (test_and_clear_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) &&
	    gl->gl_state != LM_ST_UNLOCKED &&
	    gl->gl_demote_state != LM_ST_EXCLUSIVE) {
677 678 679 680 681 682 683
		unsigned long holdtime, now = jiffies;
		holdtime = gl->gl_tchange + gl->gl_ops->go_min_hold_time;
		if (time_before(now, holdtime))
			delay = holdtime - now;
		set_bit(delay ? GLF_PENDING_DEMOTE : GLF_DEMOTE, &gl->gl_flags);
	}
	run_queue(gl, 0);
684
	spin_unlock(&gl->gl_spin);
685 686 687
	if (!delay ||
	    queue_delayed_work(glock_workqueue, &gl->gl_work, delay) == 0)
		gfs2_glock_put(gl);
688 689
	if (drop_ref)
		gfs2_glock_put(gl);
690 691
}

David Teigland's avatar
David Teigland committed
692 693 694 695 696 697 698 699 700 701 702 703 704
/**
 * 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
 */

705
int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number,
706
		   const struct gfs2_glock_operations *glops, int create,
David Teigland's avatar
David Teigland committed
707 708
		   struct gfs2_glock **glp)
{
709
	struct super_block *s = sdp->sd_vfs;
710
	struct lm_lockname name = { .ln_number = number, .ln_type = glops->go_type };
David Teigland's avatar
David Teigland committed
711
	struct gfs2_glock *gl, *tmp;
712
	unsigned int hash = gl_hash(sdp, &name);
713
	struct address_space *mapping;
714
	struct kmem_cache *cachep;
David Teigland's avatar
David Teigland committed
715

716
	rcu_read_lock();
717
	gl = search_bucket(hash, sdp, &name);
718
	rcu_read_unlock();
David Teigland's avatar
David Teigland committed
719

720 721
	*glp = gl;
	if (gl)
David Teigland's avatar
David Teigland committed
722
		return 0;
723 724
	if (!create)
		return -ENOENT;
David Teigland's avatar
David Teigland committed
725

726
	if (glops->go_flags & GLOF_ASPACE)
727
		cachep = gfs2_glock_aspace_cachep;
728
	else
729 730
		cachep = gfs2_glock_cachep;
	gl = kmem_cache_alloc(cachep, GFP_KERNEL);
David Teigland's avatar
David Teigland committed
731 732 733
	if (!gl)
		return -ENOMEM;

734
	atomic_inc(&sdp->sd_glock_disposal);
735
	gl->gl_flags = 0;
David Teigland's avatar
David Teigland committed
736
	gl->gl_name = name;
737
	atomic_set(&gl->gl_ref, 1);
David Teigland's avatar
David Teigland committed
738
	gl->gl_state = LM_ST_UNLOCKED;
739
	gl->gl_target = LM_ST_UNLOCKED;
740
	gl->gl_demote_state = LM_ST_EXCLUSIVE;
741
	gl->gl_hash = hash;
David Teigland's avatar
David Teigland committed
742
	gl->gl_ops = glops;
743 744 745
	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;
746
	gl->gl_tchange = jiffies;
747
	gl->gl_object = NULL;
David Teigland's avatar
David Teigland committed
748
	gl->gl_sbd = sdp;
749
	INIT_DELAYED_WORK(&gl->gl_work, glock_work_func);
750
	INIT_WORK(&gl->gl_delete, delete_work_func);
David Teigland's avatar
David Teigland committed
751

752 753 754 755 756 757 758 759 760
	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
761 762
	}

763
	spin_lock_bucket(hash);
764
	tmp = search_bucket(hash, sdp, &name);
David Teigland's avatar
David Teigland committed
765
	if (tmp) {
766 767
		spin_unlock_bucket(hash);
		kmem_cache_free(cachep, gl);
768
		atomic_dec(&sdp->sd_glock_disposal);
David Teigland's avatar
David Teigland committed
769 770
		gl = tmp;
	} else {
771 772
		hlist_bl_add_head_rcu(&gl->gl_list, &gl_hash_table[hash]);
		spin_unlock_bucket(hash);
David Teigland's avatar
David Teigland committed
773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788
	}

	*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
 *
 */

789
void gfs2_holder_init(struct gfs2_glock *gl, unsigned int state, unsigned flags,
David Teigland's avatar
David Teigland committed
790 791 792 793
		      struct gfs2_holder *gh)
{
	INIT_LIST_HEAD(&gh->gh_list);
	gh->gh_gl = gl;
794
	gh->gh_ip = (unsigned long)__builtin_return_address(0);
795
	gh->gh_owner_pid = get_pid(task_pid(current));
David Teigland's avatar
David Teigland committed
796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812
	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.
 *
 */

813
void gfs2_holder_reinit(unsigned int state, unsigned flags, struct gfs2_holder *gh)
David Teigland's avatar
David Teigland committed
814 815
{
	gh->gh_state = state;
816
	gh->gh_flags = flags;
817
	gh->gh_iflags = 0;
818
	gh->gh_ip = (unsigned long)__builtin_return_address(0);
Bob Peterson's avatar
Bob Peterson committed
819 820 821
	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
822 823 824 825 826 827 828 829 830 831
}

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

void gfs2_holder_uninit(struct gfs2_holder *gh)
{
832
	put_pid(gh->gh_owner_pid);
David Teigland's avatar
David Teigland committed
833 834
	gfs2_glock_put(gh->gh_gl);
	gh->gh_gl = NULL;
835
	gh->gh_ip = 0;
David Teigland's avatar
David Teigland committed
836 837
}

838 839 840 841 842 843 844 845 846 847
/**
 * 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)
848 849 850 851 852
{
        schedule();
        return 0;
}

853 854 855 856 857 858
static int gfs2_glock_demote_wait(void *word)
{
	schedule();
	return 0;
}

859
static void wait_on_holder(struct gfs2_holder *gh)
860
{
861
	might_sleep();
862
	wait_on_bit(&gh->gh_iflags, HIF_WAIT, gfs2_glock_holder_wait, TASK_UNINTERRUPTIBLE);
863 864
}

865
static void wait_on_demote(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
866
{
867
	might_sleep();
868
	wait_on_bit(&gl->gl_flags, GLF_DEMOTE, gfs2_glock_demote_wait, TASK_UNINTERRUPTIBLE);
David Teigland's avatar
David Teigland committed
869 870 871
}

/**
872 873 874
 * 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
875
 *
876 877
 * 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
878 879
 */

880
static void handle_callback(struct gfs2_glock *gl, unsigned int state,
881
			    unsigned long delay)
David Teigland's avatar
David Teigland committed
882
{
883
	int bit = delay ? GLF_PENDING_DEMOTE : GLF_DEMOTE;
David Teigland's avatar
David Teigland committed
884

885 886 887 888 889 890 891
	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
892
	}
893 894
	if (gl->gl_ops->go_callback)
		gl->gl_ops->go_callback(gl);
Steven Whitehouse's avatar
Steven Whitehouse committed
895
	trace_gfs2_demote_rq(gl);
David Teigland's avatar
David Teigland committed
896 897 898
}

/**
899
 * gfs2_glock_wait - wait on a glock acquisition
David Teigland's avatar
David Teigland committed
900 901 902 903 904
 * @gh: the glock holder
 *
 * Returns: 0 on success
 */

905
int gfs2_glock_wait(struct gfs2_holder *gh)
David Teigland's avatar
David Teigland committed
906
{
907
	wait_on_holder(gh);
David Teigland's avatar
David Teigland committed
908 909 910
	return gh->gh_error;
}

911
void gfs2_print_dbg(struct seq_file *seq, const char *fmt, ...)
912
{
913
	struct va_format vaf;
914 915 916
	va_list args;

	va_start(args, fmt);
917

918 919
	if (seq) {
		struct gfs2_glock_iter *gi = seq->private;
920
		vsprintf(gi->string, fmt, args);
921 922
		seq_printf(seq, gi->string);
	} else {
923 924 925 926
		vaf.fmt = fmt;
		vaf.va = &args;

		printk(KERN_ERR " %pV", &vaf);
927
	}
928

929 930 931
	va_end(args);
}

David Teigland's avatar
David Teigland committed
932 933 934 935
/**
 * add_to_queue - Add a holder to the wait queue (but look for recursion)
 * @gh: the holder structure to add
 *
936 937 938 939
 * 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
940 941
 */

942
static inline void add_to_queue(struct gfs2_holder *gh)
943 944
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
David Teigland's avatar
David Teigland committed
945 946
{
	struct gfs2_glock *gl = gh->gh_gl;
947 948 949 950
	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
951

952
	BUG_ON(gh->gh_owner_pid == NULL);
953 954
	if (test_and_set_bit(HIF_WAIT, &gh->gh_iflags))
		BUG();
955

956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973
	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;
974
		}
975 976 977 978 979
		if (test_bit(HIF_HOLDER, &gh2->gh_iflags))
			continue;
		if (unlikely((gh->gh_flags & LM_FLAG_PRIORITY) && !insert_pt))
			insert_pt = &gh2->gh_list;
	}
980
	set_bit(GLF_QUEUED, &gl->gl_flags);
981
	trace_gfs2_glock_queue(gh, 1);
982 983 984 985 986 987 988 989 990 991 992
	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
993
		if (sdp->sd_lockstruct.ls_ops->lm_cancel)
994
			sdp->sd_lockstruct.ls_ops->lm_cancel(gl);
995
		spin_lock(&gl->gl_spin);
David Teigland's avatar
David Teigland committed
996
	}
997
	return;
David Teigland's avatar
David Teigland committed
998

999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009
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
1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026
}

/**
 * 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;

1027
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
David Teigland's avatar
David Teigland committed
1028 1029 1030 1031
		return -EIO;

	spin_lock(&gl->gl_spin);
	add_to_queue(gh);
1032 1033 1034
	if ((LM_FLAG_NOEXP & gh->gh_flags) &&
	    test_and_clear_bit(GLF_FROZEN, &gl->gl_flags))
		set_bit(GLF_REPLY_PENDING, &gl->gl_flags);
1035
	run_queue(gl, 1);
David Teigland's avatar
David Teigland committed
1036 1037
	spin_unlock(&gl->gl_spin);

1038 1039
	if (!(gh->gh_flags & GL_ASYNC))
		error = gfs2_glock_wait(gh);
David Teigland's avatar
David Teigland committed
1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052

	return error;
}

/**
 * gfs2_glock_poll - poll to see if an async request has been completed
 * @gh: the holder
 *
 * Returns: 1 if the request is ready to be gfs2_glock_wait()ed on
 */

int gfs2_glock_poll(struct gfs2_holder *gh)
{
1053
	return test_bit(HIF_WAIT, &gh->gh_iflags) ? 0 : 1;
David Teigland's avatar
David Teigland committed
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064
}

/**
 * gfs2_glock_dq - dequeue a struct gfs2_holder from a glock (release a glock)
 * @gh: the glock holder
 *
 */

void gfs2_glock_dq(struct gfs2_holder *gh)
{
	struct gfs2_glock *gl = gh->gh_gl;
1065
	const struct gfs2_glock_operations *glops = gl->gl_ops;
1066
	unsigned delay = 0;
1067
	int fast_path = 0;
David Teigland's avatar
David Teigland committed
1068

1069
	spin_lock(&gl->gl_spin);
David Teigland's avatar
David Teigland committed
1070
	if (gh->gh_flags & GL_NOCACHE)
1071
		handle_callback(gl, LM_ST_UNLOCKED, 0);
David Teigland's avatar
David Teigland committed
1072 1073

	list_del_init(&gh->gh_list);
1074
	if (find_first_holder(gl) == NULL) {
1075
		if (glops->go_unlock) {
1076
			GLOCK_BUG_ON(gl, test_and_set_bit(GLF_LOCK, &gl->gl_flags));
1077
			spin_unlock(&gl->gl_spin);
David Teigland's avatar
David Teigland committed
1078
			glops->go_unlock(gh);
1079
			spin_lock(&gl->gl_spin);
1080
			clear_bit(GLF_LOCK, &gl->gl_flags);
1081
		}