glock.c 52.5 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>
32
#include <linux/percpu.h>
33
#include <linux/list_sort.h>
David Teigland's avatar
David Teigland committed
34 35

#include "gfs2.h"
36
#include "incore.h"
David Teigland's avatar
David Teigland committed
37 38 39 40 41 42 43
#include "glock.h"
#include "glops.h"
#include "inode.h"
#include "lops.h"
#include "meta_io.h"
#include "quota.h"
#include "super.h"
44
#include "util.h"
45
#include "bmap.h"
Steven Whitehouse's avatar
Steven Whitehouse committed
46 47
#define CREATE_TRACE_POINTS
#include "trace_gfs2.h"
David Teigland's avatar
David Teigland committed
48

49
struct gfs2_glock_iter {
50 51 52 53 54
	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               */
55 56
};

David Teigland's avatar
David Teigland committed
57 58
typedef void (*glock_examiner) (struct gfs2_glock * gl);

59
static void do_xmote(struct gfs2_glock *gl, struct gfs2_holder *gh, unsigned int target);
60

61
static struct dentry *gfs2_root;
62
static struct workqueue_struct *glock_workqueue;
63
struct workqueue_struct *gfs2_delete_workqueue;
64 65
static LIST_HEAD(lru_list);
static atomic_t lru_count = ATOMIC_INIT(0);
Julia Lawall's avatar
Julia Lawall committed
66
static DEFINE_SPINLOCK(lru_lock);
67

68
#define GFS2_GL_HASH_SHIFT      15
69 70 71
#define GFS2_GL_HASH_SIZE       (1 << GFS2_GL_HASH_SHIFT)
#define GFS2_GL_HASH_MASK       (GFS2_GL_HASH_SIZE - 1)

72
static struct hlist_bl_head gl_hash_table[GFS2_GL_HASH_SIZE];
73
static struct dentry *gfs2_root;
74

David Teigland's avatar
David Teigland committed
75 76 77 78 79 80 81
/**
 * gl_hash() - Turn glock number into hash bucket number
 * @lock: The glock number
 *
 * Returns: The number of the corresponding hash bucket
 */

82 83
static unsigned int gl_hash(const struct gfs2_sbd *sdp,
			    const struct lm_lockname *name)
David Teigland's avatar
David Teigland committed
84 85 86
{
	unsigned int h;

87
	h = jhash(&name->ln_number, sizeof(u64), 0);
David Teigland's avatar
David Teigland committed
88
	h = jhash(&name->ln_type, sizeof(unsigned int), h);
89
	h = jhash(&sdp, sizeof(struct gfs2_sbd *), h);
David Teigland's avatar
David Teigland committed
90 91 92 93 94
	h &= GFS2_GL_HASH_MASK;

	return h;
}

95 96
static inline void spin_lock_bucket(unsigned int hash)
{
97
	hlist_bl_lock(&gl_hash_table[hash]);
98
}
David Teigland's avatar
David Teigland committed
99

100 101
static inline void spin_unlock_bucket(unsigned int hash)
{
102
	hlist_bl_unlock(&gl_hash_table[hash]);
103
}
David Teigland's avatar
David Teigland committed
104

105
static void gfs2_glock_dealloc(struct rcu_head *rcu)
David Teigland's avatar
David Teigland committed
106
{
107
	struct gfs2_glock *gl = container_of(rcu, struct gfs2_glock, gl_rcu);
David Teigland's avatar
David Teigland committed
108

109
	if (gl->gl_ops->go_flags & GLOF_ASPACE) {
110
		kmem_cache_free(gfs2_glock_aspace_cachep, gl);
111
	} else {
112
		kfree(gl->gl_lksb.sb_lvbptr);
113
		kmem_cache_free(gfs2_glock_cachep, gl);
114
	}
115 116 117
}

void gfs2_glock_free(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
118 119 120
{
	struct gfs2_sbd *sdp = gl->gl_sbd;

121
	call_rcu(&gl->gl_rcu, gfs2_glock_dealloc);
122 123
	if (atomic_dec_and_test(&sdp->sd_glock_disposal))
		wake_up(&sdp->sd_glock_wait);
David Teigland's avatar
David Teigland committed
124 125 126 127 128 129 130 131
}

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

132
void gfs2_glock_hold(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
133
{
134
	GLOCK_BUG_ON(gl, atomic_read(&gl->gl_ref) == 0);
135
	atomic_inc(&gl->gl_ref);
David Teigland's avatar
David Teigland committed
136 137
}

138 139 140 141 142 143 144 145 146 147 148 149 150
/**
 * 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;
151
	if (!list_empty(&gl->gl_holders))
152 153 154 155 156 157
		return 0;
	if (glops->go_demote_ok)
		return glops->go_demote_ok(gl);
	return 1;
}

158

159 160 161 162 163 164 165 166 167 168
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);
169
	set_bit(GLF_LRU, &gl->gl_flags);
170 171 172
	spin_unlock(&lru_lock);
}

173
static void __gfs2_glock_remove_from_lru(struct gfs2_glock *gl)
174 175 176 177 178 179
{
	if (!list_empty(&gl->gl_lru)) {
		list_del_init(&gl->gl_lru);
		atomic_dec(&lru_count);
		clear_bit(GLF_LRU, &gl->gl_flags);
	}
180 181 182 183 184 185
}

static void gfs2_glock_remove_from_lru(struct gfs2_glock *gl)
{
	spin_lock(&lru_lock);
	__gfs2_glock_remove_from_lru(gl);
186 187 188
	spin_unlock(&lru_lock);
}

189 190 191 192 193 194 195 196
/**
 * 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.
 */

197
void gfs2_glock_put_nolock(struct gfs2_glock *gl)
198 199 200 201 202
{
	if (atomic_dec_and_test(&gl->gl_ref))
		GLOCK_BUG_ON(gl, 1);
}

David Teigland's avatar
David Teigland committed
203 204 205 206 207 208
/**
 * gfs2_glock_put() - Decrement reference count on glock
 * @gl: The glock to put
 *
 */

209
void gfs2_glock_put(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
210
{
211 212
	struct gfs2_sbd *sdp = gl->gl_sbd;
	struct address_space *mapping = gfs2_glock2aspace(gl);
David Teigland's avatar
David Teigland committed
213

214 215 216
	if (atomic_dec_and_lock(&gl->gl_ref, &lru_lock)) {
		__gfs2_glock_remove_from_lru(gl);
		spin_unlock(&lru_lock);
217 218 219
		spin_lock_bucket(gl->gl_hash);
		hlist_bl_del_rcu(&gl->gl_list);
		spin_unlock_bucket(gl->gl_hash);
220
		GLOCK_BUG_ON(gl, !list_empty(&gl->gl_holders));
221 222 223
		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
224 225 226 227 228 229 230 231 232 233 234
	}
}

/**
 * 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
 */

235
static struct gfs2_glock *search_bucket(unsigned int hash,
236
					const struct gfs2_sbd *sdp,
237
					const struct lm_lockname *name)
David Teigland's avatar
David Teigland committed
238 239
{
	struct gfs2_glock *gl;
240
	struct hlist_bl_node *h;
David Teigland's avatar
David Teigland committed
241

242
	hlist_bl_for_each_entry_rcu(gl, h, &gl_hash_table[hash], gl_list) {
David Teigland's avatar
David Teigland committed
243 244
		if (!lm_name_equal(&gl->gl_name, name))
			continue;
245 246
		if (gl->gl_sbd != sdp)
			continue;
247 248
		if (atomic_inc_not_zero(&gl->gl_ref))
			return gl;
David Teigland's avatar
David Teigland committed
249 250 251 252 253
	}

	return NULL;
}

254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271
/**
 * 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;
272 273 274 275 276 277
	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;
	}
278 279 280 281 282 283 284 285 286 287 288 289
	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);
}

290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313
/**
 * 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);
	}
}

314 315 316 317
/**
 * do_promote - promote as many requests as possible on the current queue
 * @gl: The glock
 * 
318 319
 * Returns: 1 if there is a blocked holder at the head of the list, or 2
 *          if a type specific operation is underway.
320 321 322
 */

static int do_promote(struct gfs2_glock *gl)
323 324
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341
{
	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) {
342 343
					if (ret == 1)
						return 2;
344 345
					gh->gh_error = ret;
					list_del_init(&gh->gh_list);
Steven Whitehouse's avatar
Steven Whitehouse committed
346
					trace_gfs2_glock_queue(gh, 0);
347 348 349 350
					gfs2_holder_wake(gh);
					goto restart;
				}
				set_bit(HIF_HOLDER, &gh->gh_iflags);
Steven Whitehouse's avatar
Steven Whitehouse committed
351
				trace_gfs2_promote(gh, 1);
352 353 354 355
				gfs2_holder_wake(gh);
				goto restart;
			}
			set_bit(HIF_HOLDER, &gh->gh_iflags);
Steven Whitehouse's avatar
Steven Whitehouse committed
356
			trace_gfs2_promote(gh, 0);
357 358 359 360 361
			gfs2_holder_wake(gh);
			continue;
		}
		if (gh->gh_list.prev == &gl->gl_holders)
			return 1;
362
		do_error(gl, 0);
363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401
		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
402
			gfs2_glock_put_nolock(gl);
403
	}
404 405
	if (held1 && held2 && list_empty(&gl->gl_holders))
		clear_bit(GLF_QUEUED, &gl->gl_flags);
406

407 408 409 410
	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);
411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434
	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;
435
	int rv;
436 437

	spin_lock(&gl->gl_spin);
Steven Whitehouse's avatar
Steven Whitehouse committed
438
	trace_gfs2_glock_state_change(gl, state);
439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
	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;
			}
		}
498 499 500
		rv = do_promote(gl);
		if (rv == 2)
			goto out_locked;
501 502 503
	}
out:
	clear_bit(GLF_LOCK, &gl->gl_flags);
504
out_locked:
505 506 507 508 509 510 511 512 513 514 515 516
	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)
517 518
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
519 520 521 522 523 524 525 526
{
	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);
527 528
	GLOCK_BUG_ON(gl, gl->gl_state == target);
	GLOCK_BUG_ON(gl, gl->gl_state == gl->gl_target);
529 530 531 532 533
	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 */
	}
534
	gl->gl_req = target;
535 536 537 538 539
	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);
540
	spin_unlock(&gl->gl_spin);
541 542
	if (glops->go_sync)
		glops->go_sync(gl);
543 544 545 546 547
	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);
548 549 550
	if (sdp->sd_lockstruct.ls_ops->lm_lock)	{
		/* lock_dlm */
		ret = sdp->sd_lockstruct.ls_ops->lm_lock(gl, target, lck_flags);
551 552 553 554
		if (ret) {
			printk(KERN_ERR "GFS2: lm_lock ret %d\n", ret);
			GLOCK_BUG_ON(gl, 1);
		}
555 556
	} else { /* lock_nolock */
		finish_xmote(gl, target);
557 558 559
		if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
			gfs2_glock_put(gl);
	}
560

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
	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)
589 590
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
591 592
{
	struct gfs2_holder *gh = NULL;
593
	int ret;
594 595 596 597 598 599 600 601 602

	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))
603
			goto out_unlock;
604 605 606
		if (nonblock)
			goto out_sched;
		set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags);
607
		GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE);
608 609 610 611
		gl->gl_target = gl->gl_demote_state;
	} else {
		if (test_bit(GLF_DEMOTE, &gl->gl_flags))
			gfs2_demote_wake(gl);
612 613
		ret = do_promote(gl);
		if (ret == 0)
614
			goto out_unlock;
615
		if (ret == 2)
616
			goto out;
617 618 619 620 621 622
		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);
623
out:
624 625 626
	return;

out_sched:
627 628
	clear_bit(GLF_LOCK, &gl->gl_flags);
	smp_mb__after_clear_bit();
629 630
	gfs2_glock_hold(gl);
	if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
631
		gfs2_glock_put_nolock(gl);
632 633
	return;

634
out_unlock:
635
	clear_bit(GLF_LOCK, &gl->gl_flags);
636 637
	smp_mb__after_clear_bit();
	return;
638 639
}

640 641 642 643
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;
644
	struct gfs2_inode *ip;
645
	struct inode *inode;
646 647 648 649
	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 */
650 651

	if (ip)
652
		inode = gfs2_ilookup(sdp->sd_vfs, no_addr, 1);
653 654 655 656 657
	else
		inode = gfs2_lookup_by_inum(sdp, no_addr, NULL, GFS2_BLKST_UNLINKED);
	if (inode && !IS_ERR(inode)) {
		d_prune_aliases(inode);
		iput(inode);
658 659 660 661
	}
	gfs2_glock_put(gl);
}

662 663
static void glock_work_func(struct work_struct *work)
{
664
	unsigned long delay = 0;
665
	struct gfs2_glock *gl = container_of(work, struct gfs2_glock, gl_work.work);
666
	int drop_ref = 0;
667

668
	if (test_and_clear_bit(GLF_REPLY_PENDING, &gl->gl_flags)) {
669
		finish_xmote(gl, gl->gl_reply);
670 671
		drop_ref = 1;
	}
672
	spin_lock(&gl->gl_spin);
673
	if (test_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) &&
674 675
	    gl->gl_state != LM_ST_UNLOCKED &&
	    gl->gl_demote_state != LM_ST_EXCLUSIVE) {
676
		unsigned long holdtime, now = jiffies;
677

678
		holdtime = gl->gl_tchange + gl->gl_hold_time;
679 680
		if (time_before(now, holdtime))
			delay = holdtime - now;
681 682 683 684 685

		if (!delay) {
			clear_bit(GLF_PENDING_DEMOTE, &gl->gl_flags);
			set_bit(GLF_DEMOTE, &gl->gl_flags);
		}
686 687
	}
	run_queue(gl, 0);
688
	spin_unlock(&gl->gl_spin);
689
	if (!delay)
690
		gfs2_glock_put(gl);
691 692 693 694 695 696
	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);
	}
697 698
	if (drop_ref)
		gfs2_glock_put(gl);
699 700
}

David Teigland's avatar
David Teigland committed
701 702 703 704 705 706 707 708 709 710 711 712 713
/**
 * 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
 */

714
int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number,
715
		   const struct gfs2_glock_operations *glops, int create,
David Teigland's avatar
David Teigland committed
716 717
		   struct gfs2_glock **glp)
{
718
	struct super_block *s = sdp->sd_vfs;
719
	struct lm_lockname name = { .ln_number = number, .ln_type = glops->go_type };
David Teigland's avatar
David Teigland committed
720
	struct gfs2_glock *gl, *tmp;
721
	unsigned int hash = gl_hash(sdp, &name);
722
	struct address_space *mapping;
723
	struct kmem_cache *cachep;
David Teigland's avatar
David Teigland committed
724

725
	rcu_read_lock();
726
	gl = search_bucket(hash, sdp, &name);
727
	rcu_read_unlock();
David Teigland's avatar
David Teigland committed
728

729 730
	*glp = gl;
	if (gl)
David Teigland's avatar
David Teigland committed
731
		return 0;
732 733
	if (!create)
		return -ENOENT;
David Teigland's avatar
David Teigland committed
734

735
	if (glops->go_flags & GLOF_ASPACE)
736
		cachep = gfs2_glock_aspace_cachep;
737
	else
738 739
		cachep = gfs2_glock_cachep;
	gl = kmem_cache_alloc(cachep, GFP_KERNEL);
David Teigland's avatar
David Teigland committed
740 741 742
	if (!gl)
		return -ENOMEM;

743 744 745
	memset(&gl->gl_lksb, 0, sizeof(struct dlm_lksb));

	if (glops->go_flags & GLOF_LVB) {
746 747
		gl->gl_lksb.sb_lvbptr = kzalloc(GFS2_MIN_LVB_SIZE, GFP_KERNEL);
		if (!gl->gl_lksb.sb_lvbptr) {
748 749 750 751 752
			kmem_cache_free(cachep, gl);
			return -ENOMEM;
		}
	}

753
	atomic_inc(&sdp->sd_glock_disposal);
754
	gl->gl_sbd = sdp;
755
	gl->gl_flags = 0;
David Teigland's avatar
David Teigland committed
756
	gl->gl_name = name;
757
	atomic_set(&gl->gl_ref, 1);
David Teigland's avatar
David Teigland committed
758
	gl->gl_state = LM_ST_UNLOCKED;
759
	gl->gl_target = LM_ST_UNLOCKED;
760
	gl->gl_demote_state = LM_ST_EXCLUSIVE;
761
	gl->gl_hash = hash;
David Teigland's avatar
David Teigland committed
762
	gl->gl_ops = glops;
763 764 765 766 767 768 769
	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;
770
	gl->gl_tchange = jiffies;
771
	gl->gl_object = NULL;
772
	gl->gl_hold_time = GL_GLOCK_DFT_HOLD;
773
	INIT_DELAYED_WORK(&gl->gl_work, glock_work_func);
774
	INIT_WORK(&gl->gl_delete, delete_work_func);
David Teigland's avatar
David Teigland committed
775

776 777 778 779 780 781
	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);
782
		mapping->private_data = NULL;
783 784
		mapping->backing_dev_info = s->s_bdi;
		mapping->writeback_index = 0;
David Teigland's avatar
David Teigland committed
785 786
	}

787
	spin_lock_bucket(hash);
788
	tmp = search_bucket(hash, sdp, &name);
David Teigland's avatar
David Teigland committed
789
	if (tmp) {
790
		spin_unlock_bucket(hash);
791
		kfree(gl->gl_lksb.sb_lvbptr);
792
		kmem_cache_free(cachep, gl);
793
		atomic_dec(&sdp->sd_glock_disposal);
David Teigland's avatar
David Teigland committed
794 795
		gl = tmp;
	} else {
796 797
		hlist_bl_add_head_rcu(&gl->gl_list, &gl_hash_table[hash]);
		spin_unlock_bucket(hash);
David Teigland's avatar
David Teigland committed
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813
	}

	*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
 *
 */

814
void gfs2_holder_init(struct gfs2_glock *gl, unsigned int state, unsigned flags,
David Teigland's avatar
David Teigland committed
815 816 817 818
		      struct gfs2_holder *gh)
{
	INIT_LIST_HEAD(&gh->gh_list);
	gh->gh_gl = gl;
819
	gh->gh_ip = (unsigned long)__builtin_return_address(0);
820
	gh->gh_owner_pid = get_pid(task_pid(current));
David Teigland's avatar
David Teigland committed
821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837
	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.
 *
 */

838
void gfs2_holder_reinit(unsigned int state, unsigned flags, struct gfs2_holder *gh)
David Teigland's avatar
David Teigland committed
839 840
{
	gh->gh_state = state;
841
	gh->gh_flags = flags;
842
	gh->gh_iflags = 0;
843
	gh->gh_ip = (unsigned long)__builtin_return_address(0);
Bob Peterson's avatar
Bob Peterson committed
844 845 846
	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
847 848 849 850 851 852 853 854 855 856
}

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

void gfs2_holder_uninit(struct gfs2_holder *gh)
{
857
	put_pid(gh->gh_owner_pid);
David Teigland's avatar
David Teigland committed
858 859
	gfs2_glock_put(gh->gh_gl);
	gh->gh_gl = NULL;
860
	gh->gh_ip = 0;
David Teigland's avatar
David Teigland committed
861 862
}

863 864 865 866 867 868 869 870 871 872
/**
 * 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)
873 874 875 876 877
{
        schedule();
        return 0;
}

878 879 880 881 882 883
static int gfs2_glock_demote_wait(void *word)
{
	schedule();
	return 0;
}

884 885 886 887 888 889 890 891
/**
 * gfs2_glock_wait - wait on a glock acquisition
 * @gh: the glock holder
 *
 * Returns: 0 on success
 */

int gfs2_glock_wait(struct gfs2_holder *gh)
892
{
893 894
	unsigned long time1 = jiffies;

895
	might_sleep();
896
	wait_on_bit(&gh->gh_iflags, HIF_WAIT, gfs2_glock_holder_wait, TASK_UNINTERRUPTIBLE);
897 898 899 900 901
	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);
902
	return gh->gh_error;
903 904
}

David Teigland's avatar
David Teigland committed
905
/**
906 907 908
 * 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
909
 *
910 911
 * 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
912 913
 */

914
static void handle_callback(struct gfs2_glock *gl, unsigned int state,
915
			    unsigned long delay)
David Teigland's avatar
David Teigland committed
916
{
917
	int bit = delay ? GLF_PENDING_DEMOTE : GLF_DEMOTE;
David Teigland's avatar
David Teigland committed
918

919 920 921 922 923 924 925
	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
926
	}
927 928
	if (gl->gl_ops->go_callback)
		gl->gl_ops->go_callback(gl);
Steven Whitehouse's avatar
Steven Whitehouse committed
929
	trace_gfs2_demote_rq(gl);
David Teigland's avatar
David Teigland committed
930 931
}

932
void gfs2_print_dbg(struct seq_file *seq, const char *fmt, ...)
933
{
934
	struct va_format vaf;
935 936 937
	va_list args;

	va_start(args, fmt);
938

939
	if (seq) {
940
		seq_vprintf(seq, fmt, args);
941
	} else {
942 943 944 945
		vaf.fmt = fmt;
		vaf.va = &args;

		printk(KERN_ERR " %pV", &vaf);
946
	}
947

948 949 950
	va_end(args);
}

David Teigland's avatar
David Teigland committed
951 952 953 954
/**
 * add_to_queue - Add a holder to the wait queue (but look for recursion)
 * @gh: the holder structure to add
 *
955 956 957 958
 * 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
959 960
 */

961
static inline void add_to_queue(struct gfs2_holder *gh)
962 963
__releases(&gl->gl_spin)
__acquires(&gl->gl_spin)
David Teigland's avatar
David Teigland committed
964 965
{
	struct gfs2_glock *gl = gh->gh_gl;
966 967 968
	struct gfs2_sbd *sdp = gl->gl_sbd;
	struct list_head *insert_pt = NULL;
	struct gfs2_holder *gh2;
969
	int try_futile = 0;
David Teigland's avatar
David Teigland committed
970

971
	BUG_ON(gh->gh_owner_pid == NULL);
972 973
	if (test_and_set_bit(HIF_WAIT, &gh->gh_iflags))
		BUG();
974

975 976
	if (gh->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB)) {
		if (test_bit(GLF_LOCK, &gl->gl_flags))
977
			try_futile = !may_grant(gl, gh);
978 979 980 981 982 983 984 985
		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;
986 987
		if (try_futile &&
		    !(gh2->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB))) {
988 989 990 991
fail:
			gh->gh_error = GLR_TRYFAILED;
			gfs2_holder_wake(gh);
			return;
992
		}
993 994 995 996 997
		if (test_bit(HIF_HOLDER, &gh2->gh_iflags))
			continue;
		if (unlikely((gh->gh_flags & LM_FLAG_PRIORITY) && !insert_pt))
			insert_pt = &gh2->gh_list;
	}
998
	set_bit(GLF_QUEUED, &gl->gl_flags);
999
	trace_gfs2_glock_queue(gh, 1);
1000 1001
	gfs2_glstats_inc(gl, GFS2_LKS_QCOUNT);
	gfs2_sbstats_inc(gl, GFS2_LKS_QCOUNT);
1002 1003 1004 1005 1006 1007 1008 1009 1010 1011