glock.c 51 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>
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>
37
#include <linux/rhashtable.h>
David Teigland's avatar
David Teigland committed
38 39

#include "gfs2.h"
40
#include "incore.h"
David Teigland's avatar
David Teigland committed
41 42 43 44 45 46 47
#include "glock.h"
#include "glops.h"
#include "inode.h"
#include "lops.h"
#include "meta_io.h"
#include "quota.h"
#include "super.h"
48
#include "util.h"
49
#include "bmap.h"
50 51
#define CREATE_TRACE_POINTS
#include "trace_gfs2.h"
David Teigland's avatar
David Teigland committed
52

53
struct gfs2_glock_iter {
54
	struct gfs2_sbd *sdp;		/* incore superblock           */
55
	struct rhashtable_iter hti;	/* rhashtable iterator         */
56 57
	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
#define GFS2_GL_HASH_SIZE       (1 << GFS2_GL_HASH_SHIFT)

74 75 76 77 78 79
static struct rhashtable_params ht_parms = {
	.nelem_hint = GFS2_GL_HASH_SIZE * 3 / 4,
	.key_len = sizeof(struct lm_lockname),
	.key_offset = offsetof(struct gfs2_glock, gl_name),
	.head_offset = offsetof(struct gfs2_glock, gl_node),
};
David Teigland's avatar
David Teigland committed
80

81
static struct rhashtable gl_hash_table;
David Teigland's avatar
David Teigland committed
82

83
void gfs2_glock_free(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
84
{
85
	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
David Teigland's avatar
David Teigland committed
86

87
	if (gl->gl_ops->go_flags & GLOF_ASPACE) {
88
		kmem_cache_free(gfs2_glock_aspace_cachep, gl);
89
	} else {
90
		kfree(gl->gl_lksb.sb_lvbptr);
91
		kmem_cache_free(gfs2_glock_cachep, gl);
92
	}
93 94
	if (atomic_dec_and_test(&sdp->sd_glock_disposal))
		wake_up(&sdp->sd_glock_wait);
David Teigland's avatar
David Teigland committed
95 96 97 98 99 100 101 102
}

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

103
static void gfs2_glock_hold(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
104
{
105 106
	GLOCK_BUG_ON(gl, __lockref_is_dead(&gl->gl_lockref));
	lockref_get(&gl->gl_lockref);
David Teigland's avatar
David Teigland committed
107 108
}

109 110 111 112 113 114 115 116 117 118 119 120 121
/**
 * 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;
122
	if (!list_empty(&gl->gl_holders))
123 124 125 126 127 128
		return 0;
	if (glops->go_demote_ok)
		return glops->go_demote_ok(gl);
	return 1;
}

129

130 131 132 133 134 135 136 137 138 139
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);
140
	set_bit(GLF_LRU, &gl->gl_flags);
141 142 143
	spin_unlock(&lru_lock);
}

144
static void gfs2_glock_remove_from_lru(struct gfs2_glock *gl)
145
{
146
	spin_lock(&lru_lock);
147 148 149 150 151 152 153 154
	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
155 156 157 158 159 160
/**
 * gfs2_glock_put() - Decrement reference count on glock
 * @gl: The glock to put
 *
 */

161
void gfs2_glock_put(struct gfs2_glock *gl)
David Teigland's avatar
David Teigland committed
162
{
163
	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
164
	struct address_space *mapping = gfs2_glock2aspace(gl);
David Teigland's avatar
David Teigland committed
165

166 167 168 169 170
	if (lockref_put_or_lock(&gl->gl_lockref))
		return;

	lockref_mark_dead(&gl->gl_lockref);

171
	gfs2_glock_remove_from_lru(gl);
172
	spin_unlock(&gl->gl_lockref.lock);
173
	rhashtable_remove_fast(&gl_hash_table, &gl->gl_node, ht_parms);
174 175 176 177
	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
178 179
}

180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197
/**
 * 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;
198 199 200 201 202 203
	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;
	}
204 205 206 207 208 209 210 211
	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);
212
	smp_mb__after_atomic();
213 214 215
	wake_up_bit(&gh->gh_iflags, HIF_WAIT);
}

216 217 218 219 220
/**
 * do_error - Something unexpected has happened during a lock request
 *
 */

221
static void do_error(struct gfs2_glock *gl, const int ret)
222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239
{
	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);
	}
}

240 241 242 243
/**
 * do_promote - promote as many requests as possible on the current queue
 * @gl: The glock
 * 
244 245
 * Returns: 1 if there is a blocked holder at the head of the list, or 2
 *          if a type specific operation is underway.
246 247 248
 */

static int do_promote(struct gfs2_glock *gl)
249 250
__releases(&gl->gl_lockref.lock)
__acquires(&gl->gl_lockref.lock)
251 252 253 254 255 256 257 258 259 260 261 262
{
	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) {
263
				spin_unlock(&gl->gl_lockref.lock);
264 265
				/* FIXME: eliminate this eventually */
				ret = glops->go_lock(gh);
266
				spin_lock(&gl->gl_lockref.lock);
267
				if (ret) {
268 269
					if (ret == 1)
						return 2;
270 271
					gh->gh_error = ret;
					list_del_init(&gh->gh_list);
272
					trace_gfs2_glock_queue(gh, 0);
273 274 275 276
					gfs2_holder_wake(gh);
					goto restart;
				}
				set_bit(HIF_HOLDER, &gh->gh_iflags);
277
				trace_gfs2_promote(gh, 1);
278 279 280 281
				gfs2_holder_wake(gh);
				goto restart;
			}
			set_bit(HIF_HOLDER, &gh->gh_iflags);
282
			trace_gfs2_promote(gh, 0);
283 284 285 286 287
			gfs2_holder_wake(gh);
			continue;
		}
		if (gh->gh_list.prev == &gl->gl_holders)
			return 1;
288
		do_error(gl, 0);
289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
		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) {
325
		GLOCK_BUG_ON(gl, __lockref_is_dead(&gl->gl_lockref));
326
		if (held2)
327
			gl->gl_lockref.count++;
328
		else
329
			gl->gl_lockref.count--;
330
	}
331 332
	if (held1 && held2 && list_empty(&gl->gl_holders))
		clear_bit(GLF_QUEUED, &gl->gl_flags);
333

334 335 336 337
	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);
338 339 340 341 342 343 344 345
	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);
346
	smp_mb__after_atomic();
347 348 349 350 351 352 353 354 355 356 357 358 359 360 361
	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;
362
	int rv;
363

364
	spin_lock(&gl->gl_lockref.lock);
365
	trace_gfs2_glock_state_change(gl, state);
366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404
	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 */
405
			pr_err("wanted %u got %u\n", gl->gl_target, state);
406 407
			GLOCK_BUG_ON(gl, 1);
		}
408
		spin_unlock(&gl->gl_lockref.lock);
409 410 411 412 413 414 415 416
		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) {
417
			spin_unlock(&gl->gl_lockref.lock);
418
			rv = glops->go_xmote_bh(gl, gh);
419
			spin_lock(&gl->gl_lockref.lock);
420 421 422 423 424
			if (rv) {
				do_error(gl, rv);
				goto out;
			}
		}
425 426 427
		rv = do_promote(gl);
		if (rv == 2)
			goto out_locked;
428 429 430
	}
out:
	clear_bit(GLF_LOCK, &gl->gl_flags);
431
out_locked:
432
	spin_unlock(&gl->gl_lockref.lock);
433 434 435 436 437 438 439 440 441 442 443
}

/**
 * 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)
444 445
__releases(&gl->gl_lockref.lock)
__acquires(&gl->gl_lockref.lock)
446 447
{
	const struct gfs2_glock_operations *glops = gl->gl_ops;
448
	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
449
	unsigned int lck_flags = (unsigned int)(gh ? gh->gh_flags : 0);
450 451 452 453
	int ret;

	lck_flags &= (LM_FLAG_TRY | LM_FLAG_TRY_1CB | LM_FLAG_NOEXP |
		      LM_FLAG_PRIORITY);
454 455
	GLOCK_BUG_ON(gl, gl->gl_state == target);
	GLOCK_BUG_ON(gl, gl->gl_state == gl->gl_target);
456 457 458 459 460
	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 */
	}
461
	gl->gl_req = target;
462 463 464 465 466
	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);
467
	spin_unlock(&gl->gl_lockref.lock);
468 469
	if (glops->go_sync)
		glops->go_sync(gl);
470 471 472 473 474
	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);
475 476 477
	if (sdp->sd_lockstruct.ls_ops->lm_lock)	{
		/* lock_dlm */
		ret = sdp->sd_lockstruct.ls_ops->lm_lock(gl, target, lck_flags);
478 479 480 481 482 483 484 485
		if (ret == -EINVAL && gl->gl_target == LM_ST_UNLOCKED &&
		    target == LM_ST_UNLOCKED &&
		    test_bit(SDF_SKIP_DLM_UNLOCK, &sdp->sd_flags)) {
			finish_xmote(gl, target);
			if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
				gfs2_glock_put(gl);
		}
		else if (ret) {
486
			pr_err("lm_lock ret %d\n", ret);
487 488
			GLOCK_BUG_ON(gl, 1);
		}
489 490
	} else { /* lock_nolock */
		finish_xmote(gl, target);
491 492 493
		if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
			gfs2_glock_put(gl);
	}
494

495
	spin_lock(&gl->gl_lockref.lock);
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
}

/**
 * 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)
523 524
__releases(&gl->gl_lockref.lock)
__acquires(&gl->gl_lockref.lock)
525 526
{
	struct gfs2_holder *gh = NULL;
527
	int ret;
528 529 530 531 532 533 534 535 536

	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))
537
			goto out_unlock;
538 539 540
		if (nonblock)
			goto out_sched;
		set_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags);
541
		GLOCK_BUG_ON(gl, gl->gl_demote_state == LM_ST_EXCLUSIVE);
542 543 544 545
		gl->gl_target = gl->gl_demote_state;
	} else {
		if (test_bit(GLF_DEMOTE, &gl->gl_flags))
			gfs2_demote_wake(gl);
546 547
		ret = do_promote(gl);
		if (ret == 0)
548
			goto out_unlock;
549
		if (ret == 2)
550
			goto out;
551 552 553 554 555 556
		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);
557
out:
558 559 560
	return;

out_sched:
561
	clear_bit(GLF_LOCK, &gl->gl_flags);
562
	smp_mb__after_atomic();
563
	gl->gl_lockref.count++;
564
	if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
565
		gl->gl_lockref.count--;
566 567
	return;

568
out_unlock:
569
	clear_bit(GLF_LOCK, &gl->gl_flags);
570
	smp_mb__after_atomic();
571
	return;
572 573
}

574 575 576
static void delete_work_func(struct work_struct *work)
{
	struct gfs2_glock *gl = container_of(work, struct gfs2_glock, gl_delete);
577
	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
578
	struct inode *inode;
579 580
	u64 no_addr = gl->gl_name.ln_number;

581 582 583 584 585 586
	/* If someone's using this glock to create a new dinode, the block must
	   have been freed by another node, then re-used, in which case our
	   iopen callback is too late after the fact. Ignore it. */
	if (test_bit(GLF_INODE_CREATING, &gl->gl_flags))
		goto out;

587
	inode = gfs2_lookup_by_inum(sdp, no_addr, NULL, GFS2_BLKST_UNLINKED);
588 589 590
	if (inode && !IS_ERR(inode)) {
		d_prune_aliases(inode);
		iput(inode);
591
	}
592
out:
593 594 595
	gfs2_glock_put(gl);
}

596 597
static void glock_work_func(struct work_struct *work)
{
598
	unsigned long delay = 0;
599
	struct gfs2_glock *gl = container_of(work, struct gfs2_glock, gl_work.work);
600
	int drop_ref = 0;
601

602
	if (test_and_clear_bit(GLF_REPLY_PENDING, &gl->gl_flags)) {
603
		finish_xmote(gl, gl->gl_reply);
604 605
		drop_ref = 1;
	}
606
	spin_lock(&gl->gl_lockref.lock);
607
	if (test_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) &&
608 609
	    gl->gl_state != LM_ST_UNLOCKED &&
	    gl->gl_demote_state != LM_ST_EXCLUSIVE) {
610
		unsigned long holdtime, now = jiffies;
611

612
		holdtime = gl->gl_tchange + gl->gl_hold_time;
613 614
		if (time_before(now, holdtime))
			delay = holdtime - now;
615 616 617 618 619

		if (!delay) {
			clear_bit(GLF_PENDING_DEMOTE, &gl->gl_flags);
			set_bit(GLF_DEMOTE, &gl->gl_flags);
		}
620 621
	}
	run_queue(gl, 0);
622
	spin_unlock(&gl->gl_lockref.lock);
623
	if (!delay)
624
		gfs2_glock_put(gl);
625 626 627 628 629 630
	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);
	}
631 632
	if (drop_ref)
		gfs2_glock_put(gl);
633 634
}

David Teigland's avatar
David Teigland committed
635 636 637 638 639 640 641 642 643 644 645 646 647
/**
 * 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
 */

648
int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number,
649
		   const struct gfs2_glock_operations *glops, int create,
David Teigland's avatar
David Teigland committed
650 651
		   struct gfs2_glock **glp)
{
652
	struct super_block *s = sdp->sd_vfs;
653 654 655
	struct lm_lockname name = { .ln_number = number,
				    .ln_type = glops->go_type,
				    .ln_sbd = sdp };
656
	struct gfs2_glock *gl, *tmp = NULL;
657
	struct address_space *mapping;
658
	struct kmem_cache *cachep;
659
	int ret, tries = 0;
David Teigland's avatar
David Teigland committed
660

661 662 663
	gl = rhashtable_lookup_fast(&gl_hash_table, &name, ht_parms);
	if (gl && !lockref_get_not_dead(&gl->gl_lockref))
		gl = NULL;
David Teigland's avatar
David Teigland committed
664

665 666
	*glp = gl;
	if (gl)
David Teigland's avatar
David Teigland committed
667
		return 0;
668 669
	if (!create)
		return -ENOENT;
David Teigland's avatar
David Teigland committed
670

671
	if (glops->go_flags & GLOF_ASPACE)
672
		cachep = gfs2_glock_aspace_cachep;
673
	else
674
		cachep = gfs2_glock_cachep;
675
	gl = kmem_cache_alloc(cachep, GFP_NOFS);
David Teigland's avatar
David Teigland committed
676 677 678
	if (!gl)
		return -ENOMEM;

679 680 681
	memset(&gl->gl_lksb, 0, sizeof(struct dlm_lksb));

	if (glops->go_flags & GLOF_LVB) {
682
		gl->gl_lksb.sb_lvbptr = kzalloc(GFS2_MIN_LVB_SIZE, GFP_NOFS);
683
		if (!gl->gl_lksb.sb_lvbptr) {
684 685 686 687 688
			kmem_cache_free(cachep, gl);
			return -ENOMEM;
		}
	}

689
	atomic_inc(&sdp->sd_glock_disposal);
690
	gl->gl_node.next = NULL;
691
	gl->gl_flags = 0;
David Teigland's avatar
David Teigland committed
692
	gl->gl_name = name;
693
	gl->gl_lockref.count = 1;
David Teigland's avatar
David Teigland committed
694
	gl->gl_state = LM_ST_UNLOCKED;
695
	gl->gl_target = LM_ST_UNLOCKED;
696
	gl->gl_demote_state = LM_ST_EXCLUSIVE;
David Teigland's avatar
David Teigland committed
697
	gl->gl_ops = glops;
698 699 700 701 702 703 704
	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;
705
	gl->gl_tchange = jiffies;
706
	gl->gl_object = NULL;
707
	gl->gl_hold_time = GL_GLOCK_DFT_HOLD;
708
	INIT_DELAYED_WORK(&gl->gl_work, glock_work_func);
709
	INIT_WORK(&gl->gl_delete, delete_work_func);
David Teigland's avatar
David Teigland committed
710

711 712 713 714 715 716
	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);
717
		mapping->private_data = NULL;
718
		mapping->writeback_index = 0;
David Teigland's avatar
David Teigland committed
719 720
	}

721 722 723 724 725 726
again:
	ret = rhashtable_lookup_insert_fast(&gl_hash_table, &gl->gl_node,
					    ht_parms);
	if (ret == 0) {
		*glp = gl;
		return 0;
David Teigland's avatar
David Teigland committed
727 728
	}

729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746
	if (ret == -EEXIST) {
		ret = 0;
		tmp = rhashtable_lookup_fast(&gl_hash_table, &name, ht_parms);
		if (tmp == NULL || !lockref_get_not_dead(&tmp->gl_lockref)) {
			if (++tries < 100) {
				cond_resched();
				goto again;
			}
			tmp = NULL;
			ret = -ENOMEM;
		}
	} else {
		WARN_ON_ONCE(ret);
	}
	kfree(gl->gl_lksb.sb_lvbptr);
	kmem_cache_free(cachep, gl);
	atomic_dec(&sdp->sd_glock_disposal);
	*glp = tmp;
David Teigland's avatar
David Teigland committed
747

748
	return ret;
David Teigland's avatar
David Teigland committed
749 750 751 752 753 754 755 756 757 758 759
}

/**
 * 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
 *
 */

760
void gfs2_holder_init(struct gfs2_glock *gl, unsigned int state, u16 flags,
David Teigland's avatar
David Teigland committed
761 762 763 764
		      struct gfs2_holder *gh)
{
	INIT_LIST_HEAD(&gh->gh_list);
	gh->gh_gl = gl;
765
	gh->gh_ip = _RET_IP_;
766
	gh->gh_owner_pid = get_pid(task_pid(current));
David Teigland's avatar
David Teigland committed
767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783
	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.
 *
 */

784
void gfs2_holder_reinit(unsigned int state, u16 flags, struct gfs2_holder *gh)
David Teigland's avatar
David Teigland committed
785 786
{
	gh->gh_state = state;
787
	gh->gh_flags = flags;
788
	gh->gh_iflags = 0;
789
	gh->gh_ip = _RET_IP_;
790
	put_pid(gh->gh_owner_pid);
Bob Peterson's avatar
Bob Peterson committed
791
	gh->gh_owner_pid = get_pid(task_pid(current));
David Teigland's avatar
David Teigland committed
792 793 794 795 796 797 798 799 800 801
}

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

void gfs2_holder_uninit(struct gfs2_holder *gh)
{
802
	put_pid(gh->gh_owner_pid);
David Teigland's avatar
David Teigland committed
803
	gfs2_glock_put(gh->gh_gl);
804
	gfs2_holder_mark_uninitialized(gh);
805
	gh->gh_ip = 0;
David Teigland's avatar
David Teigland committed
806 807
}

808 809 810 811 812 813 814 815
/**
 * gfs2_glock_wait - wait on a glock acquisition
 * @gh: the glock holder
 *
 * Returns: 0 on success
 */

int gfs2_glock_wait(struct gfs2_holder *gh)
816
{
817 818
	unsigned long time1 = jiffies;

819
	might_sleep();
820
	wait_on_bit(&gh->gh_iflags, HIF_WAIT, TASK_UNINTERRUPTIBLE);
821 822 823 824 825
	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);
826
	return gh->gh_error;
827 828
}

David Teigland's avatar
David Teigland committed
829
/**
830 831 832
 * 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
833
 *
834 835
 * 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
836 837
 */

838
static void handle_callback(struct gfs2_glock *gl, unsigned int state,
839
			    unsigned long delay, bool remote)
David Teigland's avatar
David Teigland committed
840
{
841
	int bit = delay ? GLF_PENDING_DEMOTE : GLF_DEMOTE;
David Teigland's avatar
David Teigland committed
842

843 844 845 846 847 848 849
	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
850
	}
851
	if (gl->gl_ops->go_callback)
852
		gl->gl_ops->go_callback(gl, remote);
853
	trace_gfs2_demote_rq(gl, remote);
David Teigland's avatar
David Teigland committed
854 855
}

856
void gfs2_print_dbg(struct seq_file *seq, const char *fmt, ...)
857
{
858
	struct va_format vaf;
859 860 861
	va_list args;

	va_start(args, fmt);
862

863
	if (seq) {
864
		seq_vprintf(seq, fmt, args);
865
	} else {
866 867 868
		vaf.fmt = fmt;
		vaf.va = &args;

869
		pr_err("%pV", &vaf);
870
	}
871

872 873 874
	va_end(args);
}

David Teigland's avatar
David Teigland committed
875 876 877 878
/**
 * add_to_queue - Add a holder to the wait queue (but look for recursion)
 * @gh: the holder structure to add
 *
879 880 881 882
 * 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
883 884
 */

885
static inline void add_to_queue(struct gfs2_holder *gh)
886 887
__releases(&gl->gl_lockref.lock)
__acquires(&gl->gl_lockref.lock)
David Teigland's avatar
David Teigland committed
888 889
{
	struct gfs2_glock *gl = gh->gh_gl;
890
	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
891 892
	struct list_head *insert_pt = NULL;
	struct gfs2_holder *gh2;
893
	int try_futile = 0;
David Teigland's avatar
David Teigland committed
894

895
	BUG_ON(gh->gh_owner_pid == NULL);
896 897
	if (test_and_set_bit(HIF_WAIT, &gh->gh_iflags))
		BUG();
898

899 900
	if (gh->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB)) {
		if (test_bit(GLF_LOCK, &gl->gl_flags))
901
			try_futile = !may_grant(gl, gh);
902 903 904 905 906 907 908 909
		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;
910 911
		if (try_futile &&
		    !(gh2->gh_flags & (LM_FLAG_TRY | LM_FLAG_TRY_1CB))) {
912 913 914 915
fail:
			gh->gh_error = GLR_TRYFAILED;
			gfs2_holder_wake(gh);
			return;
916
		}
917 918 919 920 921
		if (test_bit(HIF_HOLDER, &gh2->gh_iflags))
			continue;
		if (unlikely((gh->gh_flags & LM_FLAG_PRIORITY) && !insert_pt))
			insert_pt = &gh2->gh_list;
	}
922
	set_bit(GLF_QUEUED, &gl->gl_flags);
923
	trace_gfs2_glock_queue(gh, 1);
924 925
	gfs2_glstats_inc(gl, GFS2_LKS_QCOUNT);
	gfs2_sbstats_inc(gl, GFS2_LKS_QCOUNT);
926 927 928 929 930 931 932 933 934 935
	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)) {
936
		spin_unlock(&gl->gl_lockref.lock);
937
		if (sdp->sd_lockstruct.ls_ops->lm_cancel)
938
			sdp->sd_lockstruct.ls_ops->lm_cancel(gl);
939
		spin_lock(&gl->gl_lockref.lock);
David Teigland's avatar
David Teigland committed
940
	}
941
	return;
David Teigland's avatar
David Teigland committed
942

943
trap_recursive:
944 945 946
	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",
947
	       gh2->gh_gl->gl_name.ln_type, gh2->gh_state);
948 949 950
	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",
951
	       gh->gh_gl->gl_name.ln_type, gh->gh_state);
952
	gfs2_dump_glock(NULL, gl);
953
	BUG();
David Teigland's avatar
David Teigland committed
954 955 956 957 958 959 960 961 962 963 964 965 966 967
}

/**
 * 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;
968
	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
David Teigland's avatar
David Teigland committed
969 970
	int error = 0;

971
	if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags)))
David Teigland's avatar
David Teigland committed
972 973
		return -EIO;

974 975 976
	if (test_bit(GLF_LRU, &gl->gl_flags))
		gfs2_glock_remove_from_lru(gl);

977
	spin_lock(&gl->gl_lockref.lock);
David Teigland's avatar
David Teigland committed
978
	add_to_queue(gh);
979 980
	if (unlikely((LM_FLAG_NOEXP & gh->gh_flags) &&
		     test_and_clear_bit(GLF_FROZEN, &gl->gl_flags))) {
981
		set_bit(GLF_REPLY_PENDING, &gl->gl_flags);
982 983 984 985
		gl->gl_lockref.count++;
		if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
			gl->gl_lockref.count--;
	}
986
	run_queue(gl, 1);
987
	spin_unlock(&gl->gl_lockref.lock);
David Teigland's avatar
David Teigland committed
988

989 990
	if (!(gh->gh_flags & GL_ASYNC))
		error = gfs2_glock_wait(gh);
David Teigland's avatar
David Teigland committed
991 992 993 994 995 996 997 998 999 1000 1001 1002 1003

	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)
{
1004
	return test_bit(HIF_WAIT, &gh->gh_iflags) ? 0 : 1;
David Teigland's avatar
David Teigland committed
1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015
}

/**
 * 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;
1016
	const struct gfs2_glock_operations *glops = gl->gl_ops;
1017
	unsigned delay = 0;
1018
	int fast_path = 0;
David Teigland's avatar
David Teigland committed
1019

1020
	spin_lock(&gl->gl_lockref.lock);
David Teigland's avatar
David Teigland committed
1021
	if (gh->gh_flags & GL_NOCACHE)
1022
		handle_callback(gl, LM_ST_UNLOCKED, 0, false);
David Teigland's avatar
David Teigland committed
1023 1024

	list_del_init(&gh->gh_list);
1025
	clear_bit(HIF_HOLDER, &gh->gh_iflags);
1026
	if (find_first_holder(gl) == NULL) {
1027
		if (glops->go_unlock) {
1028
			GLOCK_BUG_ON(gl, test_and_set_bit(GLF_LOCK, &gl->gl_flags));
1029
			spin_unlock(&gl->gl_lockref.lock);
David Teigland's avatar
David Teigland committed
1030
			glops->go_unlock(gh);
1031
			spin_lock(&gl->gl_lockref.lock);
1032
			clear_bit(GLF_LOCK, &gl->gl_flags);
1033
		}
1034 1035 1036 1037
		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
1038
	}
1039 1040
	if (!test_bit(GLF_LFLUSH, &gl->gl_flags) && demote_ok(gl) &&
	    (glops->go_flags & GLOF_LRU))
1041 1042
		gfs2_glock_add_to_lru(gl);

1043
	trace_gfs2_glock_queue(gh, 0);
1044
	spin_unlock(&gl->gl_lockref.lock);
1045 1046
	if (likely(fast_path))
		return;
1047 1048 1049

	gfs2_glock_hold(gl);
	if (test_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) &&
1050 1051 1052
	    !test_bit(GLF_DEMOTE, &gl->gl_flags) &&
	    gl->gl_name.ln_type == LM_TYPE_INODE)
		delay = gl->gl_hold_time;
1053 1054
	if (queue_delayed_work(glock_workqueue, &gl->gl_work, delay) == 0)
		gfs2_glock_put(gl);
David Teigland's avatar
David Teigland committed
1055 1056
}

1057 1058 1059 1060
void gfs2_glock_dq_wait(struct gfs2_holder *gh)
{
	struct gfs2_glock *gl = gh->gh_gl;
	gfs2_glock_dq(gh);
1061
	might_sleep();
1062
	wait_on_bit(&gl->gl_flags, GLF_DEMOTE, TASK_UNINTERRUPTIBLE);
1063 1064
}

David Teigland's avatar
David Teigland committed
1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082
/**
 * gfs2_glock_dq_uninit - dequeue a holder from a glock and initialize it
 * @gh: the holder structure
 *
 */

void gfs2_glock_dq_uninit(struct gfs2_holder *gh)
{
	gfs2_glock_dq(gh);
	gfs2_holder_uninit(gh);
}

/**
 * gfs2_glock_nq_num - acquire a glock based on lock number
 * @sdp: the filesystem
 * @number: the lock number
 * @glops: the glock operations for the type of glock
 * @state: the state to acquire the glock in
1083
 * @flags: modifier flags for the acquisition
David Teigland's avatar
David Teigland committed
1084 1085 1086 1087 1088
 * @gh: the struct gfs2_holder
 *
 * Returns: errno
 */

1089
int gfs2_glock_nq_num(struct gfs2_sbd *sdp, u64 number,
1090
		      const struct gfs2_glock_operations *glops,
1091
		      unsigned int state, u16 flags, struct gfs2_holder *gh)
David Teigland's avatar
David Teigland committed
1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113
{
	struct gfs2_glock *gl;
	int error;

	error = gfs2_glock_get(sdp, number, glops, CREATE, &gl);
	if (!error) {
		error = gfs2_glock_nq_init(gl, state, flags, gh);
		gfs2_glock_put(gl);
	}

	return error;
}

/**
 * glock_compare - Compare two struct gfs2_glock structures for sorting
 * @arg_a: the first structure
 * @arg_b: the second structure
 *
 */

static int glock_compare(const void *arg_a, const void *arg_b)
{
1114 1115 1116 1117
	const struct gfs2_holder *gh_a = *(const struct gfs2_holder **)arg_a;
	const struct gfs2_holder *gh_b = *(const struct gfs2_holder **)arg_b;
	const struct lm_lockname *a = &gh_a->gh_gl->gl_name;
	const struct lm_lockname *b = &gh_b->gh_gl->gl_name;
David Teigland's avatar
David Teigland committed
1118 1119

	if (a->ln_number > b->ln_number)
1120 1121 1122
		return 1;
	if (a->ln_number < b->ln_number)
		return -1;
1123
	BUG_ON(gh_a->gh_gl->gl_ops->go_type == gh_b->gh_gl->gl_ops->go_type);
1124
	return 0;
David Teigland's avatar
David Teigland committed
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172
}

/**
 * nq_m_sync - synchonously acquire more than one glock in deadlock free order
 * @num_gh: the number of structures
 * @ghs: an array of struct gfs2_holder structures
 *
 * Returns: 0 on success (all glocks acquired),
 *          errno on failure (no glocks acquired)
 */

static int nq_m_sync(unsigned int num_gh, struct gfs2_holder *ghs,
		     struct gfs2_holder **p)
{
	unsigned int x;
	int error = 0;

	for (x = 0; x < num_gh; x++)
		p[x] = &ghs[x];

	sort(p, num_gh, sizeof(struct gfs2_holder *), glock_compare, NULL);

	for (x = 0; x < num_gh; x++) {
		p[x]->gh_flags &= ~(LM_FLAG_TRY | GL_ASYNC);

		error = gfs2_glock_nq(p[x]);
		if (error) {
			while (x--)
				gfs2_glock_dq(p[x]);
			break;
		}
	}

	return error;
}

/**
 * gfs2_glock_nq_m - acquire multiple glocks
 * @num_gh: the number of structures
 * @ghs: an array of struct gfs2_holder structures
 *
 *
 * Returns: 0 on success (all glocks acquired),
 *          errno on failure (no glocks acquired)
 */

int gfs2_glock_nq_m(unsigned int num_gh, struct gfs2_holder *ghs)
{
1173 1174
	struct gfs2_holder *tmp[4];
	struct gfs2_holder **pph = tmp;
David Teigland's avatar
David Teigland committed
1175 1176
	int error = 0;

1177 1178
	switch(num_gh) {
	case 0:
David Teigland's avatar
David Teigland committed
1179
		return 0;
1180
	case 1:
David Teigland's avatar
David Teigland committed
1181 1182
		ghs->gh_flags &= ~(LM_FLAG_TRY | GL_ASYNC);
		return gfs2_glock_nq(ghs);
1183 1184
	default:
		if (num_gh <= 4)
David Teigland's avatar
David Teigland committed
1185
			break;
1186 1187 1188
		pph = kmalloc(num_gh * sizeof(struct gfs2_holder *), GFP_NOFS);
		if (!pph)
			return -ENOMEM;
David Teigland's avatar
David Teigland committed
1189 1190
	}

1191
	error = nq_m_sync(num_gh, ghs, pph);
David Teigland's avatar
David Teigland committed
1192

1193 1194
	if (pph != tmp)
		kfree(pph);
David Teigland's avatar
David Teigland committed
1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207

	return error;
}

/**
 * gfs2_glock_dq_m - release multiple glocks
 * @num_gh: the number of structures
 * @ghs: an array of struct gfs2_holder structures
 *
 */

void gfs2_glock_dq_m(unsigned int num_gh, struct gfs2_holder *ghs)
{
1208 1209
	while (num_gh--)
		gfs2_glock_dq(&ghs[num_gh]);
David Teigland's avatar
David Teigland committed
1210 1211
}

1212
void gfs2_glock_cb(struct gfs2_glock *gl, unsigned int state)
1213
{
1214 1215 1216
	unsigned long delay = 0;
	unsigned long holdtime;
	unsigned long now = jiffies;
David Teigland's avatar
David Teigland committed
1217

1218
	gfs2_glock_hold(gl);
1219 1220 1221
	holdtime = gl->gl_tchange + gl->gl_hold_time;
	if (test_bit(GLF_QUEUED, &gl->gl_flags) &&
	    gl->gl_name.ln_type == LM_TYPE_INODE) {
1222 1223 1224
		if (time_before(now, holdtime))
			delay = holdtime - now;
		if (test_bit(GLF_REPLY_PENDING, &gl->gl_flags))
1225
			delay = gl->gl_hold_time;
1226
	}
David Teigland's avatar
David Teigland committed
1227

1228
	spin_lock(&gl->gl_lockref.lock);
1229
	handle_callback(gl, state, delay, true);
1230
	spin_unlock(&gl->gl_lockref.lock);
1231 1232
	if (queue_delayed_work(glock_workqueue, &gl->gl_work, delay) == 0)
		gfs2_glock_put(gl);
David Teigland's avatar
David Teigland committed
1233 1234
}

1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264
/**
 * gfs2_should_freeze - Figure out if glock should be frozen
 * @gl: The glock in question
 *
 * Glocks are not frozen if (a) the result of the dlm operation is
 * an error, (b) the locking operation was an unlock operation or
 * (c) if there is a "noexp" flagged request anywhere in the queue
 *
 * Returns: 1 if freezing should occur, 0 otherwise
 */

static int gfs2_should_freeze(const struct gfs2_glock *gl)
{
	const struct gfs2_holder *gh;

	if (gl->gl_reply & ~LM_OUT_ST_MASK)
		return 0;
	if (gl->gl_target == LM_ST_UNLOCKED)
		return 0;

	list_for_each_entry(gh, &gl->gl_holders, gh_list) {
		if (test_bit(HIF_HOLDER, &gh->gh_iflags))
			continue;
		if (LM_FLAG_NOEXP & gh->gh_flags)
			return 0;
	}

	return 1;
}

David Teigland's avatar
David Teigland committed
1265
/**
1266 1267 1268
 * gfs2_glock_complete - Callback used by locking
 * @gl: Pointer to the glock
 * @ret: The return value from the dlm
David Teigland's avatar
David Teigland committed
1269
 *
1270
 * The gl_reply field is under the gl_lockref.lock lock so that it is ok
1271
 * to use a bitfield shared with other glock state fields.
David Teigland's avatar
David Teigland committed
1272 1273
 */

1274
void gfs2_glock_complete(struct gfs2_glock *gl, int ret)
David Teigland's avatar
David Teigland committed
1275
{
1276
	struct lm_lockstruct *ls = &gl->gl_name.ln_sbd->sd_lockstruct;
1277

1278
	spin_lock(&gl->gl_lockref.lock);
1279
	gl->gl_reply = ret;
1280

1281
	if (unlikely(test_bit(DFL_BLOCK_LOCKS, &ls->ls_recover_flags))) {
1282
		if (gfs2_should_freeze(gl)) {
1283
			set_bit(GLF_FROZEN, &gl->gl_flags);
1284
			spin_unlock(&gl->gl_lockref.lock);
David Teigland's avatar
David Teigland committed
1285
			return;
1286
		}
David Teigland's avatar
David Teigland committed
1287
	}
1288

1289
	gl->gl_lockref.count++;
1290
	set_bit(GLF_REPLY_PENDING, &gl->gl_flags);
1291
	spin_unlock(&gl->gl_lockref.lock);
1292

1293 1294
	if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
		gfs2_glock_put(gl);
David Teigland's avatar
David Teigland committed
1295 1296
}

1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336
static int glock_cmp(void *priv, struct list_head *a, struct list_head *b)
{
	struct gfs2_glock *gla, *glb;

	gla = list_entry(a, struct gfs2_glock, gl_lru);
	glb = list_entry(b, struct gfs2_glock, gl_lru);

	if (gla->gl_name.ln_number > glb->gl_name.ln_number)
		return 1;
	if (gla->gl_name.ln_number < glb->gl_name.ln_number)
		return -1;

	return 0;
}

/**
 * gfs2_dispose_glock_lru - Demote a list of glocks
 * @list: The list to dispose of
 *
 * Disposing of glocks may involve disk accesses, so that here we sort
 * the glocks by number (i.e. disk location of the inodes) so that if
 * there are any such accesses, they'll be sent in order (mostly).
 *
 * Must be called under the lru_lock, but may drop and retake this
 * lock. While the lru_lock is dropped, entries may vanish from the
 * list, but no new entries will appear on the list (since it is
 * private)
 */

static void gfs2_dispose_glock_lru(struct list_head *list)
__releases(&lru_lock)
__acquires(&lru_lock)
{
	struct gfs2_glock *gl;

	list_sort(NULL, list, glock_cmp);

	while(!list_empty(list)) {
		gl = list_entry(list->next, struct gfs2_glock, gl_lru);
		list_del_init(&gl->gl_lru);
1337
		if (!spin_trylock(&gl->gl_lockref.lock)) {
1338
add_back_to_lru:
1339 1340 1341 1342
			list_add(&gl->gl_lru, &lru_list);
			atomic_inc(&lru_count);
			continue;
		}
1343
		if (test_and_set_bit(GLF_LOCK, &gl->gl_flags)) {
1344
			spin_unlock(&gl->gl_lockref.lock);
1345 1346