xfs_extfree_item.c 15.5 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2 3
 * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc.
 * All Rights Reserved.
L
Linus Torvalds 已提交
4
 *
5 6
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License as
L
Linus Torvalds 已提交
7 8
 * published by the Free Software Foundation.
 *
9 10 11 12
 * This program is distributed in the hope that it would be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
L
Linus Torvalds 已提交
13
 *
14 15 16
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write the Free Software Foundation,
 * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
L
Linus Torvalds 已提交
17 18
 */
#include "xfs.h"
19
#include "xfs_fs.h"
L
Linus Torvalds 已提交
20 21
#include "xfs_types.h"
#include "xfs_log.h"
22
#include "xfs_inum.h"
L
Linus Torvalds 已提交
23 24 25
#include "xfs_trans.h"
#include "xfs_buf_item.h"
#include "xfs_sb.h"
26
#include "xfs_ag.h"
L
Linus Torvalds 已提交
27 28 29 30 31 32 33 34 35 36 37
#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_trans_priv.h"
#include "xfs_extfree_item.h"


kmem_zone_t	*xfs_efi_zone;
kmem_zone_t	*xfs_efd_zone;

STATIC void	xfs_efi_item_unlock(xfs_efi_log_item_t *);

38 39 40 41 42 43
void
xfs_efi_item_free(xfs_efi_log_item_t *efip)
{
	int nexts = efip->efi_format.efi_nextents;

	if (nexts > XFS_EFI_MAX_FAST_EXTENTS) {
44
		kmem_free(efip);
45 46 47 48
	} else {
		kmem_zone_free(xfs_efi_zone, efip);
	}
}
L
Linus Torvalds 已提交
49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84

/*
 * This returns the number of iovecs needed to log the given efi item.
 * We only need 1 iovec for an efi item.  It just logs the efi_log_format
 * structure.
 */
/*ARGSUSED*/
STATIC uint
xfs_efi_item_size(xfs_efi_log_item_t *efip)
{
	return 1;
}

/*
 * This is called to fill in the vector of log iovecs for the
 * given efi log item. We use only 1 iovec, and we point that
 * at the efi_log_format structure embedded in the efi item.
 * It is at this point that we assert that all of the extent
 * slots in the efi item have been filled.
 */
STATIC void
xfs_efi_item_format(xfs_efi_log_item_t	*efip,
		    xfs_log_iovec_t	*log_vector)
{
	uint	size;

	ASSERT(efip->efi_next_extent == efip->efi_format.efi_nextents);

	efip->efi_format.efi_type = XFS_LI_EFI;

	size = sizeof(xfs_efi_log_format_t);
	size += (efip->efi_format.efi_nextents - 1) * sizeof(xfs_extent_t);
	efip->efi_format.efi_size = 1;

	log_vector->i_addr = (xfs_caddr_t)&(efip->efi_format);
	log_vector->i_len = size;
85
	XLOG_VEC_SET_TYPE(log_vector, XLOG_REG_TYPE_EFI_FORMAT);
L
Linus Torvalds 已提交
86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
	ASSERT(size >= sizeof(xfs_efi_log_format_t));
}


/*
 * Pinning has no meaning for an efi item, so just return.
 */
/*ARGSUSED*/
STATIC void
xfs_efi_item_pin(xfs_efi_log_item_t *efip)
{
	return;
}


/*
 * While EFIs cannot really be pinned, the unpin operation is the
 * last place at which the EFI is manipulated during a transaction.
 * Here we coordinate with xfs_efi_cancel() to determine who gets to
 * free the EFI.
 */
/*ARGSUSED*/
STATIC void
xfs_efi_item_unpin(xfs_efi_log_item_t *efip, int stale)
{
	xfs_mount_t	*mp;

	mp = efip->efi_item.li_mountp;
114
	spin_lock(&mp->m_ail->xa_lock);
L
Linus Torvalds 已提交
115 116 117 118
	if (efip->efi_flags & XFS_EFI_CANCELED) {
		/*
		 * xfs_trans_delete_ail() drops the AIL lock.
		 */
D
Donald Douwsma 已提交
119
		xfs_trans_delete_ail(mp, (xfs_log_item_t *)efip);
120
		xfs_efi_item_free(efip);
L
Linus Torvalds 已提交
121 122
	} else {
		efip->efi_flags |= XFS_EFI_COMMITTED;
123
		spin_unlock(&mp->m_ail->xa_lock);
L
Linus Torvalds 已提交
124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140
	}
}

/*
 * like unpin only we have to also clear the xaction descriptor
 * pointing the log item if we free the item.  This routine duplicates
 * unpin because efi_flags is protected by the AIL lock.  Freeing
 * the descriptor and then calling unpin would force us to drop the AIL
 * lock which would open up a race condition.
 */
STATIC void
xfs_efi_item_unpin_remove(xfs_efi_log_item_t *efip, xfs_trans_t *tp)
{
	xfs_mount_t	*mp;
	xfs_log_item_desc_t	*lidp;

	mp = efip->efi_item.li_mountp;
141
	spin_lock(&mp->m_ail->xa_lock);
L
Linus Torvalds 已提交
142 143 144 145 146 147 148 149 150 151
	if (efip->efi_flags & XFS_EFI_CANCELED) {
		/*
		 * free the xaction descriptor pointing to this item
		 */
		lidp = xfs_trans_find_item(tp, (xfs_log_item_t *) efip);
		xfs_trans_free_item(tp, lidp);
		/*
		 * pull the item off the AIL.
		 * xfs_trans_delete_ail() drops the AIL lock.
		 */
D
Donald Douwsma 已提交
152
		xfs_trans_delete_ail(mp, (xfs_log_item_t *)efip);
153
		xfs_efi_item_free(efip);
L
Linus Torvalds 已提交
154 155
	} else {
		efip->efi_flags |= XFS_EFI_COMMITTED;
156
		spin_unlock(&mp->m_ail->xa_lock);
L
Linus Torvalds 已提交
157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181
	}
}

/*
 * Efi items have no locking or pushing.  However, since EFIs are
 * pulled from the AIL when their corresponding EFDs are committed
 * to disk, their situation is very similar to being pinned.  Return
 * XFS_ITEM_PINNED so that the caller will eventually flush the log.
 * This should help in getting the EFI out of the AIL.
 */
/*ARGSUSED*/
STATIC uint
xfs_efi_item_trylock(xfs_efi_log_item_t *efip)
{
	return XFS_ITEM_PINNED;
}

/*
 * Efi items have no locking, so just return.
 */
/*ARGSUSED*/
STATIC void
xfs_efi_item_unlock(xfs_efi_log_item_t *efip)
{
	if (efip->efi_item.li_flags & XFS_LI_ABORTED)
182
		xfs_efi_item_free(efip);
L
Linus Torvalds 已提交
183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227
	return;
}

/*
 * The EFI is logged only once and cannot be moved in the log, so
 * simply return the lsn at which it's been logged.  The canceled
 * flag is not paid any attention here.  Checking for that is delayed
 * until the EFI is unpinned.
 */
/*ARGSUSED*/
STATIC xfs_lsn_t
xfs_efi_item_committed(xfs_efi_log_item_t *efip, xfs_lsn_t lsn)
{
	return lsn;
}

/*
 * There isn't much you can do to push on an efi item.  It is simply
 * stuck waiting for all of its corresponding efd items to be
 * committed to disk.
 */
/*ARGSUSED*/
STATIC void
xfs_efi_item_push(xfs_efi_log_item_t *efip)
{
	return;
}

/*
 * The EFI dependency tracking op doesn't do squat.  It can't because
 * it doesn't know where the free extent is coming from.  The dependency
 * tracking has to be handled by the "enclosing" metadata object.  For
 * example, for inodes, the inode is locked throughout the extent freeing
 * so the dependency should be recorded there.
 */
/*ARGSUSED*/
STATIC void
xfs_efi_item_committing(xfs_efi_log_item_t *efip, xfs_lsn_t lsn)
{
	return;
}

/*
 * This is the ops vector shared by all efi log items.
 */
228
static struct xfs_item_ops xfs_efi_item_ops = {
L
Linus Torvalds 已提交
229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276
	.iop_size	= (uint(*)(xfs_log_item_t*))xfs_efi_item_size,
	.iop_format	= (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
					xfs_efi_item_format,
	.iop_pin	= (void(*)(xfs_log_item_t*))xfs_efi_item_pin,
	.iop_unpin	= (void(*)(xfs_log_item_t*, int))xfs_efi_item_unpin,
	.iop_unpin_remove = (void(*)(xfs_log_item_t*, xfs_trans_t *))
					xfs_efi_item_unpin_remove,
	.iop_trylock	= (uint(*)(xfs_log_item_t*))xfs_efi_item_trylock,
	.iop_unlock	= (void(*)(xfs_log_item_t*))xfs_efi_item_unlock,
	.iop_committed	= (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
					xfs_efi_item_committed,
	.iop_push	= (void(*)(xfs_log_item_t*))xfs_efi_item_push,
	.iop_pushbuf	= NULL,
	.iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
					xfs_efi_item_committing
};


/*
 * Allocate and initialize an efi item with the given number of extents.
 */
xfs_efi_log_item_t *
xfs_efi_init(xfs_mount_t	*mp,
	     uint		nextents)

{
	xfs_efi_log_item_t	*efip;
	uint			size;

	ASSERT(nextents > 0);
	if (nextents > XFS_EFI_MAX_FAST_EXTENTS) {
		size = (uint)(sizeof(xfs_efi_log_item_t) +
			((nextents - 1) * sizeof(xfs_extent_t)));
		efip = (xfs_efi_log_item_t*)kmem_zalloc(size, KM_SLEEP);
	} else {
		efip = (xfs_efi_log_item_t*)kmem_zone_zalloc(xfs_efi_zone,
							     KM_SLEEP);
	}

	efip->efi_item.li_type = XFS_LI_EFI;
	efip->efi_item.li_ops = &xfs_efi_item_ops;
	efip->efi_item.li_mountp = mp;
	efip->efi_format.efi_nextents = nextents;
	efip->efi_format.efi_id = (__psint_t)(void*)efip;

	return (efip);
}

277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332
/*
 * Copy an EFI format buffer from the given buf, and into the destination
 * EFI format structure.
 * The given buffer can be in 32 bit or 64 bit form (which has different padding),
 * one of which will be the native format for this kernel.
 * It will handle the conversion of formats if necessary.
 */
int
xfs_efi_copy_format(xfs_log_iovec_t *buf, xfs_efi_log_format_t *dst_efi_fmt)
{
	xfs_efi_log_format_t *src_efi_fmt = (xfs_efi_log_format_t *)buf->i_addr;
	uint i;
	uint len = sizeof(xfs_efi_log_format_t) + 
		(src_efi_fmt->efi_nextents - 1) * sizeof(xfs_extent_t);  
	uint len32 = sizeof(xfs_efi_log_format_32_t) + 
		(src_efi_fmt->efi_nextents - 1) * sizeof(xfs_extent_32_t);  
	uint len64 = sizeof(xfs_efi_log_format_64_t) + 
		(src_efi_fmt->efi_nextents - 1) * sizeof(xfs_extent_64_t);  

	if (buf->i_len == len) {
		memcpy((char *)dst_efi_fmt, (char*)src_efi_fmt, len);
		return 0;
	} else if (buf->i_len == len32) {
		xfs_efi_log_format_32_t *src_efi_fmt_32 =
			(xfs_efi_log_format_32_t *)buf->i_addr;

		dst_efi_fmt->efi_type     = src_efi_fmt_32->efi_type;
		dst_efi_fmt->efi_size     = src_efi_fmt_32->efi_size;
		dst_efi_fmt->efi_nextents = src_efi_fmt_32->efi_nextents;
		dst_efi_fmt->efi_id       = src_efi_fmt_32->efi_id;
		for (i = 0; i < dst_efi_fmt->efi_nextents; i++) {
			dst_efi_fmt->efi_extents[i].ext_start =
				src_efi_fmt_32->efi_extents[i].ext_start;
			dst_efi_fmt->efi_extents[i].ext_len =
				src_efi_fmt_32->efi_extents[i].ext_len;
		}
		return 0;
	} else if (buf->i_len == len64) {
		xfs_efi_log_format_64_t *src_efi_fmt_64 =
			(xfs_efi_log_format_64_t *)buf->i_addr;

		dst_efi_fmt->efi_type     = src_efi_fmt_64->efi_type;
		dst_efi_fmt->efi_size     = src_efi_fmt_64->efi_size;
		dst_efi_fmt->efi_nextents = src_efi_fmt_64->efi_nextents;
		dst_efi_fmt->efi_id       = src_efi_fmt_64->efi_id;
		for (i = 0; i < dst_efi_fmt->efi_nextents; i++) {
			dst_efi_fmt->efi_extents[i].ext_start =
				src_efi_fmt_64->efi_extents[i].ext_start;
			dst_efi_fmt->efi_extents[i].ext_len =
				src_efi_fmt_64->efi_extents[i].ext_len;
		}
		return 0;
	}
	return EFSCORRUPTED;
}

L
Linus Torvalds 已提交
333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
/*
 * This is called by the efd item code below to release references to
 * the given efi item.  Each efd calls this with the number of
 * extents that it has logged, and when the sum of these reaches
 * the total number of extents logged by this efi item we can free
 * the efi item.
 *
 * Freeing the efi item requires that we remove it from the AIL.
 * We'll use the AIL lock to protect our counters as well as
 * the removal from the AIL.
 */
void
xfs_efi_release(xfs_efi_log_item_t	*efip,
		uint			nextents)
{
	xfs_mount_t	*mp;
	int		extents_left;

	mp = efip->efi_item.li_mountp;
	ASSERT(efip->efi_next_extent > 0);
	ASSERT(efip->efi_flags & XFS_EFI_COMMITTED);

355
	spin_lock(&mp->m_ail->xa_lock);
L
Linus Torvalds 已提交
356 357 358 359 360 361 362
	ASSERT(efip->efi_next_extent >= nextents);
	efip->efi_next_extent -= nextents;
	extents_left = efip->efi_next_extent;
	if (extents_left == 0) {
		/*
		 * xfs_trans_delete_ail() drops the AIL lock.
		 */
D
Donald Douwsma 已提交
363
		xfs_trans_delete_ail(mp, (xfs_log_item_t *)efip);
364
		xfs_efi_item_free(efip);
L
Linus Torvalds 已提交
365
	} else {
366
		spin_unlock(&mp->m_ail->xa_lock);
L
Linus Torvalds 已提交
367 368 369
	}
}

370 371 372 373
STATIC void
xfs_efd_item_free(xfs_efd_log_item_t *efdp)
{
	int nexts = efdp->efd_format.efd_nextents;
L
Linus Torvalds 已提交
374

375
	if (nexts > XFS_EFD_MAX_FAST_EXTENTS) {
376
		kmem_free(efdp);
377 378 379 380
	} else {
		kmem_zone_free(xfs_efd_zone, efdp);
	}
}
L
Linus Torvalds 已提交
381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416

/*
 * This returns the number of iovecs needed to log the given efd item.
 * We only need 1 iovec for an efd item.  It just logs the efd_log_format
 * structure.
 */
/*ARGSUSED*/
STATIC uint
xfs_efd_item_size(xfs_efd_log_item_t *efdp)
{
	return 1;
}

/*
 * This is called to fill in the vector of log iovecs for the
 * given efd log item. We use only 1 iovec, and we point that
 * at the efd_log_format structure embedded in the efd item.
 * It is at this point that we assert that all of the extent
 * slots in the efd item have been filled.
 */
STATIC void
xfs_efd_item_format(xfs_efd_log_item_t	*efdp,
		    xfs_log_iovec_t	*log_vector)
{
	uint	size;

	ASSERT(efdp->efd_next_extent == efdp->efd_format.efd_nextents);

	efdp->efd_format.efd_type = XFS_LI_EFD;

	size = sizeof(xfs_efd_log_format_t);
	size += (efdp->efd_format.efd_nextents - 1) * sizeof(xfs_extent_t);
	efdp->efd_format.efd_size = 1;

	log_vector->i_addr = (xfs_caddr_t)&(efdp->efd_format);
	log_vector->i_len = size;
417
	XLOG_VEC_SET_TYPE(log_vector, XLOG_REG_TYPE_EFD_FORMAT);
L
Linus Torvalds 已提交
418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469
	ASSERT(size >= sizeof(xfs_efd_log_format_t));
}


/*
 * Pinning has no meaning for an efd item, so just return.
 */
/*ARGSUSED*/
STATIC void
xfs_efd_item_pin(xfs_efd_log_item_t *efdp)
{
	return;
}


/*
 * Since pinning has no meaning for an efd item, unpinning does
 * not either.
 */
/*ARGSUSED*/
STATIC void
xfs_efd_item_unpin(xfs_efd_log_item_t *efdp, int stale)
{
	return;
}

/*ARGSUSED*/
STATIC void
xfs_efd_item_unpin_remove(xfs_efd_log_item_t *efdp, xfs_trans_t *tp)
{
	return;
}

/*
 * Efd items have no locking, so just return success.
 */
/*ARGSUSED*/
STATIC uint
xfs_efd_item_trylock(xfs_efd_log_item_t *efdp)
{
	return XFS_ITEM_LOCKED;
}

/*
 * Efd items have no locking or pushing, so return failure
 * so that the caller doesn't bother with us.
 */
/*ARGSUSED*/
STATIC void
xfs_efd_item_unlock(xfs_efd_log_item_t *efdp)
{
	if (efdp->efd_item.li_flags & XFS_LI_ABORTED)
470
		xfs_efd_item_free(efdp);
L
Linus Torvalds 已提交
471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491
	return;
}

/*
 * When the efd item is committed to disk, all we need to do
 * is delete our reference to our partner efi item and then
 * free ourselves.  Since we're freeing ourselves we must
 * return -1 to keep the transaction code from further referencing
 * this item.
 */
/*ARGSUSED*/
STATIC xfs_lsn_t
xfs_efd_item_committed(xfs_efd_log_item_t *efdp, xfs_lsn_t lsn)
{
	/*
	 * If we got a log I/O error, it's always the case that the LR with the
	 * EFI got unpinned and freed before the EFD got aborted.
	 */
	if ((efdp->efd_item.li_flags & XFS_LI_ABORTED) == 0)
		xfs_efi_release(efdp->efd_efip, efdp->efd_format.efd_nextents);

492
	xfs_efd_item_free(efdp);
L
Linus Torvalds 已提交
493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523
	return (xfs_lsn_t)-1;
}

/*
 * There isn't much you can do to push on an efd item.  It is simply
 * stuck waiting for the log to be flushed to disk.
 */
/*ARGSUSED*/
STATIC void
xfs_efd_item_push(xfs_efd_log_item_t *efdp)
{
	return;
}

/*
 * The EFD dependency tracking op doesn't do squat.  It can't because
 * it doesn't know where the free extent is coming from.  The dependency
 * tracking has to be handled by the "enclosing" metadata object.  For
 * example, for inodes, the inode is locked throughout the extent freeing
 * so the dependency should be recorded there.
 */
/*ARGSUSED*/
STATIC void
xfs_efd_item_committing(xfs_efd_log_item_t *efip, xfs_lsn_t lsn)
{
	return;
}

/*
 * This is the ops vector shared by all efd log items.
 */
524
static struct xfs_item_ops xfs_efd_item_ops = {
L
Linus Torvalds 已提交
525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573
	.iop_size	= (uint(*)(xfs_log_item_t*))xfs_efd_item_size,
	.iop_format	= (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
					xfs_efd_item_format,
	.iop_pin	= (void(*)(xfs_log_item_t*))xfs_efd_item_pin,
	.iop_unpin	= (void(*)(xfs_log_item_t*, int))xfs_efd_item_unpin,
	.iop_unpin_remove = (void(*)(xfs_log_item_t*, xfs_trans_t*))
					xfs_efd_item_unpin_remove,
	.iop_trylock	= (uint(*)(xfs_log_item_t*))xfs_efd_item_trylock,
	.iop_unlock	= (void(*)(xfs_log_item_t*))xfs_efd_item_unlock,
	.iop_committed	= (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
					xfs_efd_item_committed,
	.iop_push	= (void(*)(xfs_log_item_t*))xfs_efd_item_push,
	.iop_pushbuf	= NULL,
	.iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
					xfs_efd_item_committing
};


/*
 * Allocate and initialize an efd item with the given number of extents.
 */
xfs_efd_log_item_t *
xfs_efd_init(xfs_mount_t	*mp,
	     xfs_efi_log_item_t	*efip,
	     uint		nextents)

{
	xfs_efd_log_item_t	*efdp;
	uint			size;

	ASSERT(nextents > 0);
	if (nextents > XFS_EFD_MAX_FAST_EXTENTS) {
		size = (uint)(sizeof(xfs_efd_log_item_t) +
			((nextents - 1) * sizeof(xfs_extent_t)));
		efdp = (xfs_efd_log_item_t*)kmem_zalloc(size, KM_SLEEP);
	} else {
		efdp = (xfs_efd_log_item_t*)kmem_zone_zalloc(xfs_efd_zone,
							     KM_SLEEP);
	}

	efdp->efd_item.li_type = XFS_LI_EFD;
	efdp->efd_item.li_ops = &xfs_efd_item_ops;
	efdp->efd_item.li_mountp = mp;
	efdp->efd_efip = efip;
	efdp->efd_format.efd_nextents = nextents;
	efdp->efd_format.efd_efi_id = efip->efi_format.efi_id;

	return (efdp);
}