glock.c 47 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>
David Teigland's avatar
David Teigland committed
29 30

#include "gfs2.h"
31
#include "incore.h"
David Teigland's avatar
David Teigland committed
32 33 34 35 36 37 38
#include "glock.h"
#include "glops.h"
#include "inode.h"
#include "lops.h"
#include "meta_io.h"
#include "quota.h"
#include "super.h"
39
#include "util.h"
40
#include "bmap.h"
Steven Whitehouse's avatar
Steven Whitehouse committed
41 42
#define CREATE_TRACE_POINTS
#include "trace_gfs2.h"
David Teigland's avatar
David Teigland committed
43

44
struct gfs2_gl_hash_bucket {
45
        struct hlist_head hb_list;
46 47
};

48 49 50 51 52
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             */
53 54
};

David Teigland's avatar
David Teigland committed
55 56
typedef void (*glock_examiner) (struct gfs2_glock * gl);

57
static int gfs2_dump_lockstate(struct gfs2_sbd *sdp);
58 59 60
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);
61

62
static struct dentry *gfs2_root;
63
static struct workqueue_struct *glock_workqueue;
64
struct workqueue_struct *gfs2_delete_workqueue;
65 66
static LIST_HEAD(lru_list);
static atomic_t lru_count = ATOMIC_INIT(0);
Julia Lawall's avatar
Julia Lawall committed
67
static DEFINE_SPINLOCK(lru_lock);
68

69
#define GFS2_GL_HASH_SHIFT      15
70 71 72
#define GFS2_GL_HASH_SIZE       (1 << GFS2_GL_HASH_SHIFT)
#define GFS2_GL_HASH_MASK       (GFS2_GL_HASH_SIZE - 1)

73
static struct gfs2_gl_hash_bucket gl_hash_table[GFS2_GL_HASH_SIZE];
74
static struct dentry *gfs2_root;
75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114

/*
 * Despite what you might think, the numbers below are not arbitrary :-)
 * They are taken from the ipv4 routing hash code, which is well tested
 * and thus should be nearly optimal. Later on we might tweek the numbers
 * but for now this should be fine.
 *
 * The reason for putting the locks in a separate array from the list heads
 * is that we can have fewer locks than list heads and save memory. We use
 * the same hash function for both, but with a different hash mask.
 */
#if defined(CONFIG_SMP) || defined(CONFIG_DEBUG_SPINLOCK) || \
	defined(CONFIG_PROVE_LOCKING)

#ifdef CONFIG_LOCKDEP
# define GL_HASH_LOCK_SZ        256
#else
# if NR_CPUS >= 32
#  define GL_HASH_LOCK_SZ       4096
# elif NR_CPUS >= 16
#  define GL_HASH_LOCK_SZ       2048
# elif NR_CPUS >= 8
#  define GL_HASH_LOCK_SZ       1024
# elif NR_CPUS >= 4
#  define GL_HASH_LOCK_SZ       512
# else
#  define GL_HASH_LOCK_SZ       256
# endif
#endif

/* We never want more locks than chains */
#if GFS2_GL_HASH_SIZE < GL_HASH_LOCK_SZ
# undef GL_HASH_LOCK_SZ
# define GL_HASH_LOCK_SZ GFS2_GL_HASH_SIZE
#endif

static rwlock_t gl_hash_locks[GL_HASH_LOCK_SZ];

static inline rwlock_t *gl_lock_addr(unsigned int x)
{
115
	return &gl_hash_locks[x & (GL_HASH_LOCK_SZ-1)];
116 117
}
#else /* not SMP, so no spinlocks required */
Randy Dunlap's avatar
Randy Dunlap committed
118
static inline rwlock_t *gl_lock_addr(unsigned int x)
119 120 121 122
{
	return NULL;
}
#endif
123

David Teigland's avatar
David Teigland committed
124 125 126 127 128 129 130
/**
 * gl_hash() - Turn glock number into hash bucket number
 * @lock: The glock number
 *
 * Returns: The number of the corresponding hash bucket
 */

131 132
static unsigned int gl_hash(const struct gfs2_sbd *sdp,
			    const struct lm_lockname *name)
David Teigland's avatar
David Teigland committed
133 134 135
{
	unsigned int h;

136
	h = jhash(&name->ln_number, sizeof(u64), 0);
David Teigland's avatar
David Teigland committed
137
	h = jhash(&name->ln_type, sizeof(unsigned int), h);
138
	h = jhash(&sdp, sizeof(struct gfs2_sbd *), h);
David Teigland's avatar
David Teigland committed
139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154
	h &= GFS2_GL_HASH_MASK;

	return h;
}

/**
 * glock_free() - Perform a few checks and then release struct gfs2_glock
 * @gl: The glock to release
 *
 * Also calls lock module to release its internal structure for this glock.
 *
 */

static void glock_free(struct gfs2_glock *gl)
{
	struct gfs2_sbd *sdp = gl->gl_sbd;
155 156
	struct address_space *mapping = gfs2_glock2aspace(gl);
	struct kmem_cache *cachep = gfs2_glock_cachep;
David Teigland's avatar
David Teigland committed
157

158
	GLOCK_BUG_ON(gl, mapping && mapping->nrpages);
Steven Whitehouse's avatar
Steven Whitehouse committed
159
	trace_gfs2_glock_put(gl);
160 161 162
	if (mapping)
		cachep = gfs2_glock_aspace_cachep;
	sdp->sd_lockstruct.ls_ops->lm_put_lock(cachep, gl);
David Teigland's avatar
David Teigland committed
163 164 165 166 167 168 169 170
}

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

171
void gfs2_glock_hold(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
172
{
173
	GLOCK_BUG_ON(gl, atomic_read(&gl->gl_ref) == 0);
174
	atomic_inc(&gl->gl_ref);
David Teigland's avatar
David Teigland committed
175 176
}

177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196
/**
 * 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;
	if (!list_empty(&gl->gl_holders))
		return 0;
	if (glops->go_demote_ok)
		return glops->go_demote_ok(gl);
	return 1;
}

197 198 199 200 201 202 203 204
/**
 * gfs2_glock_schedule_for_reclaim - Add a glock to the reclaim list
 * @gl: the glock
 *
 */

static void gfs2_glock_schedule_for_reclaim(struct gfs2_glock *gl)
{
205 206 207 208 209
	int may_reclaim;
	may_reclaim = (demote_ok(gl) &&
		       (atomic_read(&gl->gl_ref) == 1 ||
			(gl->gl_name.ln_type == LM_TYPE_INODE &&
			 atomic_read(&gl->gl_ref) <= 2)));
210
	spin_lock(&lru_lock);
211
	if (list_empty(&gl->gl_lru) && may_reclaim) {
212 213 214 215 216 217
		list_add_tail(&gl->gl_lru, &lru_list);
		atomic_inc(&lru_count);
	}
	spin_unlock(&lru_lock);
}

218 219 220 221 222 223 224 225
/**
 * 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.
 */

226
void gfs2_glock_put_nolock(struct gfs2_glock *gl)
227 228 229 230 231 232
{
	if (atomic_dec_and_test(&gl->gl_ref))
		GLOCK_BUG_ON(gl, 1);
	gfs2_glock_schedule_for_reclaim(gl);
}

David Teigland's avatar
David Teigland committed
233 234 235 236 237 238 239 240 241 242
/**
 * gfs2_glock_put() - Decrement reference count on glock
 * @gl: The glock to put
 *
 */

int gfs2_glock_put(struct gfs2_glock *gl)
{
	int rv = 0;

243
	write_lock(gl_lock_addr(gl->gl_hash));
244
	if (atomic_dec_and_lock(&gl->gl_ref, &lru_lock)) {
245
		hlist_del(&gl->gl_list);
246 247 248 249 250
		if (!list_empty(&gl->gl_lru)) {
			list_del_init(&gl->gl_lru);
			atomic_dec(&lru_count);
		}
		spin_unlock(&lru_lock);
251
		write_unlock(gl_lock_addr(gl->gl_hash));
252
		GLOCK_BUG_ON(gl, !list_empty(&gl->gl_holders));
David Teigland's avatar
David Teigland committed
253 254 255 256
		glock_free(gl);
		rv = 1;
		goto out;
	}
257 258 259
	spin_lock(&gl->gl_spin);
	gfs2_glock_schedule_for_reclaim(gl);
	spin_unlock(&gl->gl_spin);
260
	write_unlock(gl_lock_addr(gl->gl_hash));
261
out:
David Teigland's avatar
David Teigland committed
262 263 264 265 266 267 268 269 270 271 272
	return rv;
}

/**
 * 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
 */

273
static struct gfs2_glock *search_bucket(unsigned int hash,
274
					const struct gfs2_sbd *sdp,
275
					const struct lm_lockname *name)
David Teigland's avatar
David Teigland committed
276 277
{
	struct gfs2_glock *gl;
278
	struct hlist_node *h;
David Teigland's avatar
David Teigland committed
279

280
	hlist_for_each_entry(gl, h, &gl_hash_table[hash].hb_list, gl_list) {
David Teigland's avatar
David Teigland committed
281 282
		if (!lm_name_equal(&gl->gl_name, name))
			continue;
283 284
		if (gl->gl_sbd != sdp)
			continue;
David Teigland's avatar
David Teigland committed
285

286
		atomic_inc(&gl->gl_ref);
David Teigland's avatar
David Teigland committed
287 288 289 290 291 292 293

		return gl;
	}

	return NULL;
}

294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311
/**
 * 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;
312 313 314 315 316 317
	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;
	}
318 319 320 321 322 323 324 325 326 327 328 329
	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);
}

330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353
/**
 * 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);
	}
}

354 355 356 357
/**
 * do_promote - promote as many requests as possible on the current queue
 * @gl: The glock
 * 
358 359
 * Returns: 1 if there is a blocked holder at the head of the list, or 2
 *          if a type specific operation is underway.
360 361 362
 */

static int do_promote(struct gfs2_glock *gl)
363 364
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381
{
	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) {
382 383
					if (ret == 1)
						return 2;
384 385
					gh->gh_error = ret;
					list_del_init(&gh->gh_list);
Steven Whitehouse's avatar
Steven Whitehouse committed
386
					trace_gfs2_glock_queue(gh, 0);
387 388 389 390
					gfs2_holder_wake(gh);
					goto restart;
				}
				set_bit(HIF_HOLDER, &gh->gh_iflags);
Steven Whitehouse's avatar
Steven Whitehouse committed
391
				trace_gfs2_promote(gh, 1);
392 393 394 395
				gfs2_holder_wake(gh);
				goto restart;
			}
			set_bit(HIF_HOLDER, &gh->gh_iflags);
Steven Whitehouse's avatar
Steven Whitehouse committed
396
			trace_gfs2_promote(gh, 0);
397 398 399 400 401
			gfs2_holder_wake(gh);
			continue;
		}
		if (gh->gh_list.prev == &gl->gl_holders)
			return 1;
402
		do_error(gl, 0);
403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441
		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
442
			gfs2_glock_put_nolock(gl);
443
	}
444 445
	if (held1 && held2 && list_empty(&gl->gl_holders))
		clear_bit(GLF_QUEUED, &gl->gl_flags);
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

	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;
471
	int rv;
472 473

	spin_lock(&gl->gl_spin);
Steven Whitehouse's avatar
Steven Whitehouse committed
474
	trace_gfs2_glock_state_change(gl, state);
475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533
	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;
			}
		}
534 535 536
		rv = do_promote(gl);
		if (rv == 2)
			goto out_locked;
537 538 539
	}
out:
	clear_bit(GLF_LOCK, &gl->gl_flags);
540
out_locked:
541 542 543 544 545 546 547 548 549 550 551 552
	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)
553 554
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
555 556 557 558 559 560 561 562
{
	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);
563 564
	GLOCK_BUG_ON(gl, gl->gl_state == target);
	GLOCK_BUG_ON(gl, gl->gl_state == gl->gl_target);
565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582
	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 */
	}
	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;

583 584 585 586 587 588
	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);
589 590 591
		if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
			gfs2_glock_put(gl);
	}
592

593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620
	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)
621 622
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
623 624
{
	struct gfs2_holder *gh = NULL;
625
	int ret;
626 627 628 629 630 631 632 633 634

	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))
635
			goto out_unlock;
636 637 638
		if (nonblock)
			goto out_sched;
		set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags);
639
		GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE);
640 641 642 643
		gl->gl_target = gl->gl_demote_state;
	} else {
		if (test_bit(GLF_DEMOTE, &gl->gl_flags))
			gfs2_demote_wake(gl);
644 645
		ret = do_promote(gl);
		if (ret == 0)
646
			goto out_unlock;
647
		if (ret == 2)
648
			goto out;
649 650 651 652 653 654
		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);
655
out:
656 657 658
	return;

out_sched:
659 660
	clear_bit(GLF_LOCK, &gl->gl_flags);
	smp_mb__after_clear_bit();
661 662
	gfs2_glock_hold(gl);
	if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
663
		gfs2_glock_put_nolock(gl);
664 665
	return;

666
out_unlock:
667
	clear_bit(GLF_LOCK, &gl->gl_flags);
668 669
	smp_mb__after_clear_bit();
	return;
670 671
}

672 673 674 675
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;
676
	struct gfs2_inode *ip;
677
	struct inode *inode;
678 679 680 681
	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 */
682 683 684

	if (ip)
		inode = gfs2_ilookup(sdp->sd_vfs, no_addr);
685 686 687 688 689
	else
		inode = gfs2_lookup_by_inum(sdp, no_addr, NULL, GFS2_BLKST_UNLINKED);
	if (inode && !IS_ERR(inode)) {
		d_prune_aliases(inode);
		iput(inode);
690 691 692 693
	}
	gfs2_glock_put(gl);
}

694 695
static void glock_work_func(struct work_struct *work)
{
696
	unsigned long delay = 0;
697
	struct gfs2_glock *gl = container_of(work, struct gfs2_glock, gl_work.work);
698
	int drop_ref = 0;
699

700
	if (test_and_clear_bit(GLF_REPLY_PENDING, &gl->gl_flags)) {
701
		finish_xmote(gl, gl->gl_reply);
702 703
		drop_ref = 1;
	}
704
	spin_lock(&gl->gl_spin);
705 706 707
	if (test_and_clear_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) &&
	    gl->gl_state != LM_ST_UNLOCKED &&
	    gl->gl_demote_state != LM_ST_EXCLUSIVE) {
708 709 710 711 712 713 714
		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);
715
	spin_unlock(&gl->gl_spin);
716 717 718
	if (!delay ||
	    queue_delayed_work(glock_workqueue, &gl->gl_work, delay) == 0)
		gfs2_glock_put(gl);
719 720
	if (drop_ref)
		gfs2_glock_put(gl);
721 722
}

David Teigland's avatar
David Teigland committed
723 724 725 726 727 728 729 730 731 732 733 734 735
/**
 * 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
 */

736
int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number,
737
		   const struct gfs2_glock_operations *glops, int create,
David Teigland's avatar
David Teigland committed
738 739
		   struct gfs2_glock **glp)
{
740
	struct super_block *s = sdp->sd_vfs;
741
	struct lm_lockname name = { .ln_number = number, .ln_type = glops->go_type };
David Teigland's avatar
David Teigland committed
742
	struct gfs2_glock *gl, *tmp;
743
	unsigned int hash = gl_hash(sdp, &name);
744
	struct address_space *mapping;
David Teigland's avatar
David Teigland committed
745

746
	read_lock(gl_lock_addr(hash));
747
	gl = search_bucket(hash, sdp, &name);
748
	read_unlock(gl_lock_addr(hash));
David Teigland's avatar
David Teigland committed
749

750 751
	*glp = gl;
	if (gl)
David Teigland's avatar
David Teigland committed
752
		return 0;
753 754
	if (!create)
		return -ENOENT;
David Teigland's avatar
David Teigland committed
755

756 757 758 759
	if (glops->go_flags & GLOF_ASPACE)
		gl = kmem_cache_alloc(gfs2_glock_aspace_cachep, GFP_KERNEL);
	else
		gl = kmem_cache_alloc(gfs2_glock_cachep, GFP_KERNEL);
David Teigland's avatar
David Teigland committed
760 761 762
	if (!gl)
		return -ENOMEM;

763
	atomic_inc(&sdp->sd_glock_disposal);
764
	gl->gl_flags = 0;
David Teigland's avatar
David Teigland committed
765
	gl->gl_name = name;
766
	atomic_set(&gl->gl_ref, 1);
David Teigland's avatar
David Teigland committed
767
	gl->gl_state = LM_ST_UNLOCKED;
768
	gl->gl_target = LM_ST_UNLOCKED;
769
	gl->gl_demote_state = LM_ST_EXCLUSIVE;
770
	gl->gl_hash = hash;
David Teigland's avatar
David Teigland committed
771
	gl->gl_ops = glops;
772 773 774
	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;
775
	gl->gl_tchange = jiffies;
776
	gl->gl_object = NULL;
David Teigland's avatar
David Teigland committed
777
	gl->gl_sbd = sdp;
778
	INIT_DELAYED_WORK(&gl->gl_work, glock_work_func);
779
	INIT_WORK(&gl->gl_delete, delete_work_func);
David Teigland's avatar
David Teigland committed
780

781 782 783 784 785 786 787 788 789
	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
790 791
	}

792
	write_lock(gl_lock_addr(hash));
793
	tmp = search_bucket(hash, sdp, &name);
David Teigland's avatar
David Teigland committed
794
	if (tmp) {
795
		write_unlock(gl_lock_addr(hash));
David Teigland's avatar
David Teigland committed
796 797 798
		glock_free(gl);
		gl = tmp;
	} else {
799
		hlist_add_head(&gl->gl_list, &gl_hash_table[hash].hb_list);
800
		write_unlock(gl_lock_addr(hash));
David Teigland's avatar
David Teigland committed
801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816
	}

	*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
 *
 */

817
void gfs2_holder_init(struct gfs2_glock *gl, unsigned int state, unsigned flags,
David Teigland's avatar
David Teigland committed
818 819 820 821
		      struct gfs2_holder *gh)
{
	INIT_LIST_HEAD(&gh->gh_list);
	gh->gh_gl = gl;
822
	gh->gh_ip = (unsigned long)__builtin_return_address(0);
823
	gh->gh_owner_pid = get_pid(task_pid(current));
David Teigland's avatar
David Teigland committed
824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840
	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.
 *
 */

841
void gfs2_holder_reinit(unsigned int state, unsigned flags, struct gfs2_holder *gh)
David Teigland's avatar
David Teigland committed
842 843
{
	gh->gh_state = state;
844
	gh->gh_flags = flags;
845
	gh->gh_iflags = 0;
846
	gh->gh_ip = (unsigned long)__builtin_return_address(0);
Bob Peterson's avatar
Bob Peterson committed
847 848 849
	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
850 851 852 853 854 855 856 857 858 859
}

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

void gfs2_holder_uninit(struct gfs2_holder *gh)
{
860
	put_pid(gh->gh_owner_pid);
David Teigland's avatar
David Teigland committed
861 862
	gfs2_glock_put(gh->gh_gl);
	gh->gh_gl = NULL;
863
	gh->gh_ip = 0;
David Teigland's avatar
David Teigland committed
864 865
}

866 867 868 869 870 871 872 873 874 875
/**
 * 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)
876 877 878 879 880
{
        schedule();
        return 0;
}

881 882 883 884 885 886
static int gfs2_glock_demote_wait(void *word)
{
	schedule();
	return 0;
}

887
static void wait_on_holder(struct gfs2_holder *gh)
888
{
889
	might_sleep();
890
	wait_on_bit(&gh->gh_iflags, HIF_WAIT, gfs2_glock_holder_wait, TASK_UNINTERRUPTIBLE);
891 892
}

893
static void wait_on_demote(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
894
{
895
	might_sleep();
896
	wait_on_bit(&gl->gl_flags, GLF_DEMOTE, gfs2_glock_demote_wait, TASK_UNINTERRUPTIBLE);
David Teigland's avatar
David Teigland committed
897 898 899
}

/**
900 901 902
 * 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
903
 *
904 905
 * 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
906 907
 */

908
static void handle_callback(struct gfs2_glock *gl, unsigned int state,
909
			    unsigned long delay)
David Teigland's avatar
David Teigland committed
910
{
911
	int bit = delay ? GLF_PENDING_DEMOTE : GLF_DEMOTE;
David Teigland's avatar
David Teigland committed
912

913 914 915 916 917 918 919
	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
920
	}
921 922
	if (gl->gl_ops->go_callback)
		gl->gl_ops->go_callback(gl);
Steven Whitehouse's avatar
Steven Whitehouse committed
923
	trace_gfs2_demote_rq(gl);
David Teigland's avatar
David Teigland committed
924 925 926
}

/**
927
 * gfs2_glock_wait - wait on a glock acquisition
David Teigland's avatar
David Teigland committed
928 929 930 931 932
 * @gh: the glock holder
 *
 * Returns: 0 on success
 */

933
int gfs2_glock_wait(struct gfs2_holder *gh)
David Teigland's avatar
David Teigland committed
934
{
935
	wait_on_holder(gh);
David Teigland's avatar
David Teigland committed
936 937 938
	return gh->gh_error;
}

939
void gfs2_print_dbg(struct seq_file *seq, const char *fmt, ...)
940
{
941
	struct va_format vaf;
942 943 944
	va_list args;

	va_start(args, fmt);
945

946 947
	if (seq) {
		struct gfs2_glock_iter *gi = seq->private;
948
		vsprintf(gi->string, fmt, args);
949 950
		seq_printf(seq, gi->string);
	} else {
951 952 953 954
		vaf.fmt = fmt;
		vaf.va = &args;

		printk(KERN_ERR " %pV", &vaf);
955
	}
956

957 958 959
	va_end(args);
}

David Teigland's avatar
David Teigland committed
960 961 962 963
/**
 * add_to_queue - Add a holder to the wait queue (but look for recursion)
 * @gh: the holder structure to add
 *
964 965 966 967
 * 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
968 969
 */

970
static inline void add_to_queue(struct gfs2_holder *gh)
971 972
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
David Teigland's avatar
David Teigland committed
973 974
{
	struct gfs2_glock *gl = gh->gh_gl;
975 976 977 978
	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
979

980
	BUG_ON(gh->gh_owner_pid == NULL);
981 982
	if (test_and_set_bit(HIF_WAIT, &gh->gh_iflags))
		BUG();
983

984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001
	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;
1002
		}
1003 1004 1005 1006 1007
		if (test_bit(HIF_HOLDER, &gh2->gh_iflags))
			continue;
		if (unlikely((gh->gh_flags & LM_FLAG_PRIORITY) && !insert_pt))
			insert_pt = &gh2->gh_list;
	}
1008
	set_bit(GLF_QUEUED, &gl->gl_flags);
1009 1010 1011 1012 1013 1014
	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;
	}
Steven Whitehouse's avatar
Steven Whitehouse committed
1015
	trace_gfs2_glock_queue(gh, 1);
1016 1017 1018 1019 1020
	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
1021
		if (sdp->sd_lockstruct.ls_ops->lm_cancel)
1022
			sdp->sd_lockstruct.ls_ops->lm_cancel(gl);
1023
		spin_lock(&gl->gl_spin);
David Teigland's avatar
David Teigland committed
1024
	}
1025
	return;
David Teigland's avatar
David Teigland committed
1026

1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037
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
1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
}

/**
 * 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;

1055
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
David Teigland's avatar
David Teigland committed
1056 1057 1058 1059
		return -EIO;

	spin_lock(&gl->gl_spin);
	add_to_queue(gh);
1060 1061 1062
	if ((LM_FLAG_NOEXP & gh->gh_flags) &&
	    test_and_clear_bit(GLF_FROZEN, &gl->gl_flags))
		set_bit(GLF_REPLY_PENDING, &gl->gl_flags);
1063
	run_queue(gl, 1);
David Teigland's avatar
David Teigland committed
1064 1065
	spin_unlock(&gl->gl_spin);

1066 1067
	if (!(gh->gh_flags & GL_ASYNC))
		error = gfs2_glock_wait(gh);
David Teigland's avatar
David Teigland committed
1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080

	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)
{
1081
	return test_bit(HIF_WAIT, &gh->gh_iflags) ? 0 : 1;
David Teigland's avatar
David Teigland committed
1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092
}

/**
 * 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;
1093
	const struct gfs2_glock_operations *glops = gl->gl_ops;
1094
	unsigned delay = 0;
1095
	int fast_path = 0;
David Teigland's avatar
David Teigland committed
1096

1097
	spin_lock(&gl->gl_spin);
David Teigland's avatar
David Teigland committed
1098
	if (gh->gh_flags & GL_NOCACHE)
1099
		handle_callback(gl, LM_ST_UNLOCKED, 0);
David Teigland's avatar
David Teigland committed
1100 1101

	list_del_init(&gh->gh_list);
1102
	if (find_first_holder(gl) == NULL) {
1103
		if (glops->go_unlock) {
1104
			GLOCK_BUG_ON(gl, test_and_set_bit(GLF_LOCK, &gl->gl_flags));
1105
			spin_unlock(&gl->gl_spin);
David Teigland's avatar
David Teigland committed
1106
			glops->go_unlock(gh);
1107
			spin_lock(&gl->gl_spin);
1108
			clear_bit(GLF_LOCK, &gl->gl_flags);
1109
		}
1110 1111 1112 1113
		if (list_empty(&gl->gl_holders) &&
		    !test_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) &&
		    !test_bit(GLF_DEMOTE, &gl->gl_flags))
			fast_path = 1;
David Teigland's avatar
David Teigland committed
1114
	}
Steven Whitehouse's avatar
Steven Whitehouse committed
1115
	trace_gfs2_glock_queue(gh, 0);
David Teigland's avatar
David Teigland committed
1116
	spin_unlock(&gl->gl_spin);
1117 1118
	if (likely(fast_path))
		return;