glock.c 51.3 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
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

David Teigland's avatar
David Teigland committed
12 13 14 15 16 17 18
#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>
19
#include <linux/kallsyms.h>
20
#include <linux/gfs2_ondisk.h>
21
#include <linux/list.h>
22
#include <linux/wait.h>
akpm@linux-foundation.org's avatar
akpm@linux-foundation.org committed
23
#include <linux/module.h>
David Teigland's avatar
David Teigland committed
24
#include <asm/uaccess.h>
25 26
#include <linux/seq_file.h>
#include <linux/debugfs.h>
27 28
#include <linux/kthread.h>
#include <linux/freezer.h>
29 30
#include <linux/workqueue.h>
#include <linux/jiffies.h>
31 32 33
#include <linux/rcupdate.h>
#include <linux/rculist_bl.h>
#include <linux/bit_spinlock.h>
34
#include <linux/percpu.h>
35
#include <linux/list_sort.h>
36
#include <linux/lockref.h>
David Teigland's avatar
David Teigland committed
37 38

#include "gfs2.h"
39
#include "incore.h"
David Teigland's avatar
David Teigland committed
40 41 42 43 44 45 46
#include "glock.h"
#include "glops.h"
#include "inode.h"
#include "lops.h"
#include "meta_io.h"
#include "quota.h"
#include "super.h"
47
#include "util.h"
48
#include "bmap.h"
Steven Whitehouse's avatar
Steven Whitehouse committed
49 50
#define CREATE_TRACE_POINTS
#include "trace_gfs2.h"
David Teigland's avatar
David Teigland committed
51

52
struct gfs2_glock_iter {
53 54 55 56 57
	int hash;			/* hash bucket index           */
	unsigned nhash;			/* Index within current bucket */
	struct gfs2_sbd *sdp;		/* incore superblock           */
	struct gfs2_glock *gl;		/* current glock struct        */
	loff_t last_pos;		/* last position               */
58 59
};

David Teigland's avatar
David Teigland committed
60 61
typedef void (*glock_examiner) (struct gfs2_glock * gl);

62
static void do_xmote(struct gfs2_glock *gl, struct gfs2_holder *gh, unsigned int target);
63

64
static struct dentry *gfs2_root;
65
static struct workqueue_struct *glock_workqueue;
66
struct workqueue_struct *gfs2_delete_workqueue;
67 68
static LIST_HEAD(lru_list);
static atomic_t lru_count = ATOMIC_INIT(0);
Julia Lawall's avatar
Julia Lawall committed
69
static DEFINE_SPINLOCK(lru_lock);
70

71
#define GFS2_GL_HASH_SHIFT      15
72 73 74
#define GFS2_GL_HASH_SIZE       (1 << GFS2_GL_HASH_SHIFT)
#define GFS2_GL_HASH_MASK       (GFS2_GL_HASH_SIZE - 1)

75
static struct hlist_bl_head gl_hash_table[GFS2_GL_HASH_SIZE];
76
static struct dentry *gfs2_root;
77

David Teigland's avatar
David Teigland committed
78 79 80 81 82 83 84
/**
 * gl_hash() - Turn glock number into hash bucket number
 * @lock: The glock number
 *
 * Returns: The number of the corresponding hash bucket
 */

85 86
static unsigned int gl_hash(const struct gfs2_sbd *sdp,
			    const struct lm_lockname *name)
David Teigland's avatar
David Teigland committed
87 88 89
{
	unsigned int h;

90
	h = jhash(&name->ln_number, sizeof(u64), 0);
David Teigland's avatar
David Teigland committed
91
	h = jhash(&name->ln_type, sizeof(unsigned int), h);
92
	h = jhash(&sdp, sizeof(struct gfs2_sbd *), h);
David Teigland's avatar
David Teigland committed
93 94 95 96 97
	h &= GFS2_GL_HASH_MASK;

	return h;
}

98 99
static inline void spin_lock_bucket(unsigned int hash)
{
100
	hlist_bl_lock(&gl_hash_table[hash]);
101
}
David Teigland's avatar
David Teigland committed
102

103 104
static inline void spin_unlock_bucket(unsigned int hash)
{
105
	hlist_bl_unlock(&gl_hash_table[hash]);
106
}
David Teigland's avatar
David Teigland committed
107

108
static void gfs2_glock_dealloc(struct rcu_head *rcu)
David Teigland's avatar
David Teigland committed
109
{
110
	struct gfs2_glock *gl = container_of(rcu, struct gfs2_glock, gl_rcu);
David Teigland's avatar
David Teigland committed
111

112
	if (gl->gl_ops->go_flags & GLOF_ASPACE) {
113
		kmem_cache_free(gfs2_glock_aspace_cachep, gl);
114
	} else {
115
		kfree(gl->gl_lksb.sb_lvbptr);
116
		kmem_cache_free(gfs2_glock_cachep, gl);
117
	}
118 119 120
}

void gfs2_glock_free(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
121 122 123
{
	struct gfs2_sbd *sdp = gl->gl_sbd;

124
	call_rcu(&gl->gl_rcu, gfs2_glock_dealloc);
125 126
	if (atomic_dec_and_test(&sdp->sd_glock_disposal))
		wake_up(&sdp->sd_glock_wait);
David Teigland's avatar
David Teigland committed
127 128 129 130 131 132 133 134
}

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

135
static void gfs2_glock_hold(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
136
{
137 138
	GLOCK_BUG_ON(gl, __lockref_is_dead(&gl->gl_lockref));
	lockref_get(&gl->gl_lockref);
David Teigland's avatar
David Teigland committed
139 140
}

141 142 143 144 145 146 147 148 149 150 151 152 153
/**
 * 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;
154
	if (!list_empty(&gl->gl_holders))
155 156 157 158 159 160
		return 0;
	if (glops->go_demote_ok)
		return glops->go_demote_ok(gl);
	return 1;
}

161

162 163 164 165 166 167 168 169 170 171
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);
172
	set_bit(GLF_LRU, &gl->gl_flags);
173 174 175
	spin_unlock(&lru_lock);
}

176
static void gfs2_glock_remove_from_lru(struct gfs2_glock *gl)
177
{
178
	spin_lock(&lru_lock);
179 180 181 182 183 184 185 186
	if (!list_empty(&gl->gl_lru)) {
		list_del_init(&gl->gl_lru);
		atomic_dec(&lru_count);
		clear_bit(GLF_LRU, &gl->gl_flags);
	}
	spin_unlock(&lru_lock);
}

David Teigland's avatar
David Teigland committed
187 188 189 190 191 192
/**
 * gfs2_glock_put() - Decrement reference count on glock
 * @gl: The glock to put
 *
 */

193
void gfs2_glock_put(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
194
{
195 196
	struct gfs2_sbd *sdp = gl->gl_sbd;
	struct address_space *mapping = gfs2_glock2aspace(gl);
David Teigland's avatar
David Teigland committed
197

198 199 200 201 202
	if (lockref_put_or_lock(&gl->gl_lockref))
		return;

	lockref_mark_dead(&gl->gl_lockref);

203
	gfs2_glock_remove_from_lru(gl);
204 205 206 207 208 209 210 211
	spin_unlock(&gl->gl_lockref.lock);
	spin_lock_bucket(gl->gl_hash);
	hlist_bl_del_rcu(&gl->gl_list);
	spin_unlock_bucket(gl->gl_hash);
	GLOCK_BUG_ON(gl, !list_empty(&gl->gl_holders));
	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
212 213 214 215 216 217 218 219 220 221
}

/**
 * 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
 */

222
static struct gfs2_glock *search_bucket(unsigned int hash,
223
					const struct gfs2_sbd *sdp,
224
					const struct lm_lockname *name)
David Teigland's avatar
David Teigland committed
225 226
{
	struct gfs2_glock *gl;
227
	struct hlist_bl_node *h;
David Teigland's avatar
David Teigland committed
228

229
	hlist_bl_for_each_entry_rcu(gl, h, &gl_hash_table[hash], gl_list) {
David Teigland's avatar
David Teigland committed
230 231
		if (!lm_name_equal(&gl->gl_name, name))
			continue;
232 233
		if (gl->gl_sbd != sdp)
			continue;
234
		if (lockref_get_not_dead(&gl->gl_lockref))
235
			return gl;
David Teigland's avatar
David Teigland committed
236 237 238 239 240
	}

	return NULL;
}

241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
/**
 * 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;
259 260 261 262 263 264
	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;
	}
265 266 267 268 269 270 271 272
	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);
273
	smp_mb__after_atomic();
274 275 276
	wake_up_bit(&gh->gh_iflags, HIF_WAIT);
}

277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300
/**
 * 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);
	}
}

301 302 303 304
/**
 * do_promote - promote as many requests as possible on the current queue
 * @gl: The glock
 * 
305 306
 * Returns: 1 if there is a blocked holder at the head of the list, or 2
 *          if a type specific operation is underway.
307 308 309
 */

static int do_promote(struct gfs2_glock *gl)
310 311
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
{
	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) {
329 330
					if (ret == 1)
						return 2;
331 332
					gh->gh_error = ret;
					list_del_init(&gh->gh_list);
Steven Whitehouse's avatar
Steven Whitehouse committed
333
					trace_gfs2_glock_queue(gh, 0);
334 335 336 337
					gfs2_holder_wake(gh);
					goto restart;
				}
				set_bit(HIF_HOLDER, &gh->gh_iflags);
Steven Whitehouse's avatar
Steven Whitehouse committed
338
				trace_gfs2_promote(gh, 1);
339 340 341 342
				gfs2_holder_wake(gh);
				goto restart;
			}
			set_bit(HIF_HOLDER, &gh->gh_iflags);
Steven Whitehouse's avatar
Steven Whitehouse committed
343
			trace_gfs2_promote(gh, 0);
344 345 346 347 348
			gfs2_holder_wake(gh);
			continue;
		}
		if (gh->gh_list.prev == &gl->gl_holders)
			return 1;
349
		do_error(gl, 0);
350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385
		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) {
386
		GLOCK_BUG_ON(gl, __lockref_is_dead(&gl->gl_lockref));
387
		if (held2)
388
			gl->gl_lockref.count++;
389
		else
390
			gl->gl_lockref.count--;
391
	}
392 393
	if (held1 && held2 && list_empty(&gl->gl_holders))
		clear_bit(GLF_QUEUED, &gl->gl_flags);
394

395 396 397 398
	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);
399 400 401 402 403 404 405 406
	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);
407
	smp_mb__after_atomic();
408 409 410 411 412 413 414 415 416 417 418 419 420 421 422
	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;
423
	int rv;
424 425

	spin_lock(&gl->gl_spin);
Steven Whitehouse's avatar
Steven Whitehouse committed
426
	trace_gfs2_glock_state_change(gl, state);
427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
	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 */
466
			pr_err("wanted %u got %u\n", gl->gl_target, state);
467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485
			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;
			}
		}
486 487 488
		rv = do_promote(gl);
		if (rv == 2)
			goto out_locked;
489 490 491
	}
out:
	clear_bit(GLF_LOCK, &gl->gl_flags);
492
out_locked:
493 494 495 496 497 498 499 500 501 502 503 504
	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)
505 506
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
507 508 509 510 511 512 513 514
{
	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);
515 516
	GLOCK_BUG_ON(gl, gl->gl_state == target);
	GLOCK_BUG_ON(gl, gl->gl_state == gl->gl_target);
517 518 519 520 521
	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 */
	}
522
	gl->gl_req = target;
523 524 525 526 527
	set_bit(GLF_BLOCKING, &gl->gl_flags);
	if ((gl->gl_req == LM_ST_UNLOCKED) ||
	    (gl->gl_state == LM_ST_EXCLUSIVE) ||
	    (lck_flags & (LM_FLAG_TRY|LM_FLAG_TRY_1CB)))
		clear_bit(GLF_BLOCKING, &gl->gl_flags);
528
	spin_unlock(&gl->gl_spin);
529 530
	if (glops->go_sync)
		glops->go_sync(gl);
531 532 533 534 535
	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);
536 537 538
	if (sdp->sd_lockstruct.ls_ops->lm_lock)	{
		/* lock_dlm */
		ret = sdp->sd_lockstruct.ls_ops->lm_lock(gl, target, lck_flags);
539
		if (ret) {
540
			pr_err("lm_lock ret %d\n", ret);
541 542
			GLOCK_BUG_ON(gl, 1);
		}
543 544
	} else { /* lock_nolock */
		finish_xmote(gl, target);
545 546 547
		if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
			gfs2_glock_put(gl);
	}
548

549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576
	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)
577 578
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
579 580
{
	struct gfs2_holder *gh = NULL;
581
	int ret;
582 583 584 585 586 587 588 589 590

	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))
591
			goto out_unlock;
592 593 594
		if (nonblock)
			goto out_sched;
		set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags);
595
		GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE);
596 597 598 599
		gl->gl_target = gl->gl_demote_state;
	} else {
		if (test_bit(GLF_DEMOTE, &gl->gl_flags))
			gfs2_demote_wake(gl);
600 601
		ret = do_promote(gl);
		if (ret == 0)
602
			goto out_unlock;
603
		if (ret == 2)
604
			goto out;
605 606 607 608 609 610
		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);
611
out:
612 613 614
	return;

out_sched:
615
	clear_bit(GLF_LOCK, &gl->gl_flags);
616
	smp_mb__after_atomic();
617
	gl->gl_lockref.count++;
618
	if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
619
		gl->gl_lockref.count--;
620 621
	return;

622
out_unlock:
623
	clear_bit(GLF_LOCK, &gl->gl_flags);
624
	smp_mb__after_atomic();
625
	return;
626 627
}

628 629 630 631
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;
632
	struct gfs2_inode *ip;
633
	struct inode *inode;
634 635 636 637
	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 */
638 639

	if (ip)
640
		inode = gfs2_ilookup(sdp->sd_vfs, no_addr, 1);
641 642 643 644 645
	else
		inode = gfs2_lookup_by_inum(sdp, no_addr, NULL, GFS2_BLKST_UNLINKED);
	if (inode && !IS_ERR(inode)) {
		d_prune_aliases(inode);
		iput(inode);
646 647 648 649
	}
	gfs2_glock_put(gl);
}

650 651
static void glock_work_func(struct work_struct *work)
{
652
	unsigned long delay = 0;
653
	struct gfs2_glock *gl = container_of(work, struct gfs2_glock, gl_work.work);
654
	int drop_ref = 0;
655

656
	if (test_and_clear_bit(GLF_REPLY_PENDING, &gl->gl_flags)) {
657
		finish_xmote(gl, gl->gl_reply);
658 659
		drop_ref = 1;
	}
660
	spin_lock(&gl->gl_spin);
661
	if (test_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) &&
662 663
	    gl->gl_state != LM_ST_UNLOCKED &&
	    gl->gl_demote_state != LM_ST_EXCLUSIVE) {
664
		unsigned long holdtime, now = jiffies;
665

666
		holdtime = gl->gl_tchange + gl->gl_hold_time;
667 668
		if (time_before(now, holdtime))
			delay = holdtime - now;
669 670 671 672 673

		if (!delay) {
			clear_bit(GLF_PENDING_DEMOTE, &gl->gl_flags);
			set_bit(GLF_DEMOTE, &gl->gl_flags);
		}
674 675
	}
	run_queue(gl, 0);
676
	spin_unlock(&gl->gl_spin);
677
	if (!delay)
678
		gfs2_glock_put(gl);
679 680 681 682 683 684
	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);
	}
685 686
	if (drop_ref)
		gfs2_glock_put(gl);
687 688
}

David Teigland's avatar
David Teigland committed
689 690 691 692 693 694 695 696 697 698 699 700 701
/**
 * 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
 */

702
int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number,
703
		   const struct gfs2_glock_operations *glops, int create,
David Teigland's avatar
David Teigland committed
704 705
		   struct gfs2_glock **glp)
{
706
	struct super_block *s = sdp->sd_vfs;
707
	struct lm_lockname name = { .ln_number = number, .ln_type = glops->go_type };
David Teigland's avatar
David Teigland committed
708
	struct gfs2_glock *gl, *tmp;
709
	unsigned int hash = gl_hash(sdp, &name);
710
	struct address_space *mapping;
711
	struct kmem_cache *cachep;
David Teigland's avatar
David Teigland committed
712

713
	rcu_read_lock();
714
	gl = search_bucket(hash, sdp, &name);
715
	rcu_read_unlock();
David Teigland's avatar
David Teigland committed
716

717 718
	*glp = gl;
	if (gl)
David Teigland's avatar
David Teigland committed
719
		return 0;
720 721
	if (!create)
		return -ENOENT;
David Teigland's avatar
David Teigland committed
722

723
	if (glops->go_flags & GLOF_ASPACE)
724
		cachep = gfs2_glock_aspace_cachep;
725
	else
726
		cachep = gfs2_glock_cachep;
727
	gl = kmem_cache_alloc(cachep, GFP_NOFS);
David Teigland's avatar
David Teigland committed
728 729 730
	if (!gl)
		return -ENOMEM;

731 732 733
	memset(&gl->gl_lksb, 0, sizeof(struct dlm_lksb));

	if (glops->go_flags & GLOF_LVB) {
734
		gl->gl_lksb.sb_lvbptr = kzalloc(GFS2_MIN_LVB_SIZE, GFP_NOFS);
735
		if (!gl->gl_lksb.sb_lvbptr) {
736 737 738 739 740
			kmem_cache_free(cachep, gl);
			return -ENOMEM;
		}
	}

741
	atomic_inc(&sdp->sd_glock_disposal);
742
	gl->gl_sbd = sdp;
743
	gl->gl_flags = 0;
David Teigland's avatar
David Teigland committed
744
	gl->gl_name = name;
745
	gl->gl_lockref.count = 1;
David Teigland's avatar
David Teigland committed
746
	gl->gl_state = LM_ST_UNLOCKED;
747
	gl->gl_target = LM_ST_UNLOCKED;
748
	gl->gl_demote_state = LM_ST_EXCLUSIVE;
749
	gl->gl_hash = hash;
David Teigland's avatar
David Teigland committed
750
	gl->gl_ops = glops;
751 752 753 754 755 756 757
	gl->gl_dstamp = ktime_set(0, 0);
	preempt_disable();
	/* We use the global stats to estimate the initial per-glock stats */
	gl->gl_stats = this_cpu_ptr(sdp->sd_lkstats)->lkstats[glops->go_type];
	preempt_enable();
	gl->gl_stats.stats[GFS2_LKS_DCOUNT] = 0;
	gl->gl_stats.stats[GFS2_LKS_QCOUNT] = 0;
758
	gl->gl_tchange = jiffies;
759
	gl->gl_object = NULL;
760
	gl->gl_hold_time = GL_GLOCK_DFT_HOLD;
761
	INIT_DELAYED_WORK(&gl->gl_work, glock_work_func);
762
	INIT_WORK(&gl->gl_delete, delete_work_func);
David Teigland's avatar
David Teigland committed
763

764 765 766 767 768 769
	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);
770
		mapping->private_data = NULL;
771
		mapping->writeback_index = 0;
David Teigland's avatar
David Teigland committed
772 773
	}

774
	spin_lock_bucket(hash);
775
	tmp = search_bucket(hash, sdp, &name);
David Teigland's avatar
David Teigland committed
776
	if (tmp) {
777
		spin_unlock_bucket(hash);
778
		kfree(gl->gl_lksb.sb_lvbptr);
779
		kmem_cache_free(cachep, gl);
780
		atomic_dec(&sdp->sd_glock_disposal);
David Teigland's avatar
David Teigland committed
781 782
		gl = tmp;
	} else {
783 784
		hlist_bl_add_head_rcu(&gl->gl_list, &gl_hash_table[hash]);
		spin_unlock_bucket(hash);
David Teigland's avatar
David Teigland committed
785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800
	}

	*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
 *
 */

801
void gfs2_holder_init(struct gfs2_glock *gl, unsigned int state, unsigned flags,
David Teigland's avatar
David Teigland committed
802 803 804 805
		      struct gfs2_holder *gh)
{
	INIT_LIST_HEAD(&gh->gh_list);
	gh->gh_gl = gl;
806
	gh->gh_ip = _RET_IP_;
807
	gh->gh_owner_pid = get_pid(task_pid(current));
David Teigland's avatar
David Teigland committed
808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824
	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.
 *
 */

825
void gfs2_holder_reinit(unsigned int state, unsigned flags, struct gfs2_holder *gh)
David Teigland's avatar
David Teigland committed
826 827
{
	gh->gh_state = state;
828
	gh->gh_flags = flags;
829
	gh->gh_iflags = 0;
830
	gh->gh_ip = _RET_IP_;
831
	put_pid(gh->gh_owner_pid);
Bob Peterson's avatar
Bob Peterson committed
832
	gh->gh_owner_pid = get_pid(task_pid(current));
David Teigland's avatar
David Teigland committed
833 834 835 836 837 838 839 840 841 842
}

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

void gfs2_holder_uninit(struct gfs2_holder *gh)
{
843
	put_pid(gh->gh_owner_pid);
David Teigland's avatar
David Teigland committed
844 845
	gfs2_glock_put(gh->gh_gl);
	gh->gh_gl = NULL;
846
	gh->gh_ip = 0;
David Teigland's avatar
David Teigland committed
847 848
}

849 850 851 852 853 854 855 856
/**
 * gfs2_glock_wait - wait on a glock acquisition
 * @gh: the glock holder
 *
 * Returns: 0 on success
 */

int gfs2_glock_wait(struct gfs2_holder *gh)
857
{
858 859
	unsigned long time1 = jiffies;

860
	might_sleep();
861
	wait_on_bit(&gh->gh_iflags, HIF_WAIT, TASK_UNINTERRUPTIBLE);
862 863 864 865 866
	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);
867
	return gh->gh_error;
868 869
}

David Teigland's avatar
David Teigland committed
870
/**
871 872 873
 * 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
874
 *
875 876
 * 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
877 878
 */

879
static void handle_callback(struct gfs2_glock *gl, unsigned int state,
880
			    unsigned long delay, bool remote)
David Teigland's avatar
David Teigland committed
881
{
882
	int bit = delay ? GLF_PENDING_DEMOTE : GLF_DEMOTE;
David Teigland's avatar
David Teigland committed
883

884 885 886 887 888 889 890
	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
891
	}
892
	if (gl->gl_ops->go_callback)
893
		gl->gl_ops->go_callback(gl, remote);
894
	trace_gfs2_demote_rq(gl, remote);
David Teigland's avatar
David Teigland committed
895 896
}

897
void gfs2_print_dbg(struct seq_file *seq, const char *fmt, ...)
898
{
899
	struct va_format vaf;
900 901 902
	va_list args;

	va_start(args, fmt);
903

904
	if (seq) {
905
		seq_vprintf(seq, fmt, args);
906
	} else {
907 908 909
		vaf.fmt = fmt;
		vaf.va = &args;

910
		pr_err("%pV", &vaf);
911
	}
912

913 914 915
	va_end(args);
}

David Teigland's avatar
David Teigland committed
916 917 918 919
/**
 * add_to_queue - Add a holder to the wait queue (but look for recursion)
 * @gh: the holder structure to add
 *
920 921 922 923
 * 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
924 925
 */

926
static inline void add_to_queue(struct gfs2_holder *gh)
927 928
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
David Teigland's avatar
David Teigland committed
929 930
{
	struct gfs2_glock *gl = gh->gh_gl;
931 932 933
	struct gfs2_sbd *sdp = gl->gl_sbd;
	struct list_head *insert_pt = NULL;
	struct gfs2_holder *gh2;
934
	int try_futile = 0;
David Teigland's avatar
David Teigland committed
935

936
	BUG_ON(gh->gh_owner_pid == NULL);
937 938
	if (test_and_set_bit(HIF_WAIT, &gh->gh_iflags))
		BUG();
939

940 941
	if (gh->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB)) {
		if (test_bit(GLF_LOCK, &gl->gl_flags))
942
			try_futile = !may_grant(gl, gh);
943 944 945 946 947 948 949 950
		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;
951 952
		if (try_futile &&
		    !(gh2->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB))) {
953 954 955 956
fail:
			gh->gh_error = GLR_TRYFAILED;
			gfs2_holder_wake(gh);
			return;
957
		}
958 959 960 961 962
		if (test_bit(HIF_HOLDER, &gh2->gh_iflags))
			continue;
		if (unlikely((gh->gh_flags & LM_FLAG_PRIORITY) && !insert_pt))
			insert_pt = &gh2->gh_list;
	}
963
	set_bit(GLF_QUEUED, &gl->gl_flags);
964
	trace_gfs2_glock_queue(gh, 1);
965 966
	gfs2_glstats_inc(gl, GFS2_LKS_QCOUNT);
	gfs2_sbstats_inc(gl, GFS2_LKS_QCOUNT);
967 968 969 970 971 972 973 974 975 976 977
	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
978
		if (sdp->sd_lockstruct.ls_ops->lm_cancel)
979
			sdp->sd_lockstruct.ls_ops->lm_cancel(gl);
980
		spin_lock(&gl->gl_spin);
David Teigland's avatar
David Teigland committed
981
	}
982
	return;
David Teigland's avatar
David Teigland committed
983

984
trap_recursive:
985 986 987
	pr_err("original: %pSR\n", (void *)gh2->gh_ip);
	pr_err("pid: %d\n", pid_nr(gh2->gh_owner_pid));
	pr_err("lock type: %d req lock state : %d\n",
988
	       gh2->gh_gl->gl_name.ln_type, gh2->gh_state);
989 990 991
	pr_err("new: %pSR\n", (void *)gh->gh_ip);
	pr_err("pid: %d\n", pid_nr(gh->gh_owner_pid));
	pr_err("lock type: %d req lock state : %d\n",
992
	       gh->gh_gl->gl_name.ln_type, gh->gh_state);
993
	gfs2_dump_glock(NULL, gl);
Steven Whitehouse's avatar