xfs_inode.c 91.4 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1
/*
2
 * Copyright (c) 2000-2006 Silicon Graphics, Inc.
3
 * All Rights Reserved.
Linus Torvalds's avatar
Linus Torvalds committed
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
Linus Torvalds's avatar
Linus Torvalds committed
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.
Linus Torvalds's avatar
Linus Torvalds committed
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
Linus Torvalds's avatar
Linus Torvalds committed
17
 */
18 19
#include <linux/log2.h>

Linus Torvalds's avatar
Linus Torvalds committed
20
#include "xfs.h"
21
#include "xfs_fs.h"
22
#include "xfs_shared.h"
23 24 25
#include "xfs_format.h"
#include "xfs_log_format.h"
#include "xfs_trans_resv.h"
26
#include "xfs_inum.h"
Linus Torvalds's avatar
Linus Torvalds committed
27 28 29
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_mount.h"
30
#include "xfs_inode.h"
31
#include "xfs_da_format.h"
Dave Chinner's avatar
Dave Chinner committed
32 33
#include "xfs_da_btree.h"
#include "xfs_dir2.h"
34
#include "xfs_attr_sf.h"
Dave Chinner's avatar
Dave Chinner committed
35
#include "xfs_attr.h"
36 37
#include "xfs_trans_space.h"
#include "xfs_trans.h"
Linus Torvalds's avatar
Linus Torvalds committed
38
#include "xfs_buf_item.h"
39 40 41
#include "xfs_inode_item.h"
#include "xfs_ialloc.h"
#include "xfs_bmap.h"
42
#include "xfs_bmap_util.h"
Linus Torvalds's avatar
Linus Torvalds committed
43 44
#include "xfs_error.h"
#include "xfs_quota.h"
45
#include "xfs_filestream.h"
46
#include "xfs_cksum.h"
47
#include "xfs_trace.h"
Dave Chinner's avatar
Dave Chinner committed
48
#include "xfs_icache.h"
Dave Chinner's avatar
Dave Chinner committed
49
#include "xfs_symlink.h"
50 51
#include "xfs_trans_priv.h"
#include "xfs_log.h"
52
#include "xfs_bmap_btree.h"
Linus Torvalds's avatar
Linus Torvalds committed
53 54 55 56

kmem_zone_t *xfs_inode_zone;

/*
57
 * Used in xfs_itruncate_extents().  This is the maximum number of extents
Linus Torvalds's avatar
Linus Torvalds committed
58 59 60 61 62 63
 * freed from a file in a single transaction.
 */
#define	XFS_ITRUNC_MAX_EXTENTS	2

STATIC int xfs_iflush_int(xfs_inode_t *, xfs_buf_t *);

64 65
STATIC int xfs_iunlink_remove(xfs_trans_t *, xfs_inode_t *);

66 67 68 69 70 71 72 73 74 75 76 77 78 79
/*
 * helper function to extract extent size hint from inode
 */
xfs_extlen_t
xfs_get_extsz_hint(
	struct xfs_inode	*ip)
{
	if ((ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE) && ip->i_d.di_extsize)
		return ip->i_d.di_extsize;
	if (XFS_IS_REALTIME_INODE(ip))
		return ip->i_mount->m_sb.sb_rextsize;
	return 0;
}

80
/*
81 82 83 84 85 86 87 88 89 90
 * These two are wrapper routines around the xfs_ilock() routine used to
 * centralize some grungy code.  They are used in places that wish to lock the
 * inode solely for reading the extents.  The reason these places can't just
 * call xfs_ilock(ip, XFS_ILOCK_SHARED) is that the inode lock also guards to
 * bringing in of the extents from disk for a file in b-tree format.  If the
 * inode is in b-tree format, then we need to lock the inode exclusively until
 * the extents are read in.  Locking it exclusively all the time would limit
 * our parallelism unnecessarily, though.  What we do instead is check to see
 * if the extents have been read in yet, and only lock the inode exclusively
 * if they have not.
91
 *
92
 * The functions return a value which should be given to the corresponding
93
 * xfs_iunlock() call.
94 95
 */
uint
96 97
xfs_ilock_data_map_shared(
	struct xfs_inode	*ip)
98
{
99
	uint			lock_mode = XFS_ILOCK_SHARED;
100

101 102
	if (ip->i_d.di_format == XFS_DINODE_FMT_BTREE &&
	    (ip->i_df.if_flags & XFS_IFEXTENTS) == 0)
103 104 105 106 107
		lock_mode = XFS_ILOCK_EXCL;
	xfs_ilock(ip, lock_mode);
	return lock_mode;
}

108 109 110
uint
xfs_ilock_attr_map_shared(
	struct xfs_inode	*ip)
111
{
112 113 114 115 116 117 118
	uint			lock_mode = XFS_ILOCK_SHARED;

	if (ip->i_d.di_aformat == XFS_DINODE_FMT_BTREE &&
	    (ip->i_afp->if_flags & XFS_IFEXTENTS) == 0)
		lock_mode = XFS_ILOCK_EXCL;
	xfs_ilock(ip, lock_mode);
	return lock_mode;
119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286
}

/*
 * The xfs inode contains 2 locks: a multi-reader lock called the
 * i_iolock and a multi-reader lock called the i_lock.  This routine
 * allows either or both of the locks to be obtained.
 *
 * The 2 locks should always be ordered so that the IO lock is
 * obtained first in order to prevent deadlock.
 *
 * ip -- the inode being locked
 * lock_flags -- this parameter indicates the inode's locks
 *       to be locked.  It can be:
 *		XFS_IOLOCK_SHARED,
 *		XFS_IOLOCK_EXCL,
 *		XFS_ILOCK_SHARED,
 *		XFS_ILOCK_EXCL,
 *		XFS_IOLOCK_SHARED | XFS_ILOCK_SHARED,
 *		XFS_IOLOCK_SHARED | XFS_ILOCK_EXCL,
 *		XFS_IOLOCK_EXCL | XFS_ILOCK_SHARED,
 *		XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL
 */
void
xfs_ilock(
	xfs_inode_t		*ip,
	uint			lock_flags)
{
	trace_xfs_ilock(ip, lock_flags, _RET_IP_);

	/*
	 * You can't set both SHARED and EXCL for the same lock,
	 * and only XFS_IOLOCK_SHARED, XFS_IOLOCK_EXCL, XFS_ILOCK_SHARED,
	 * and XFS_ILOCK_EXCL are valid values to set in lock_flags.
	 */
	ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) !=
	       (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL));
	ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) !=
	       (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
	ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_DEP_MASK)) == 0);

	if (lock_flags & XFS_IOLOCK_EXCL)
		mrupdate_nested(&ip->i_iolock, XFS_IOLOCK_DEP(lock_flags));
	else if (lock_flags & XFS_IOLOCK_SHARED)
		mraccess_nested(&ip->i_iolock, XFS_IOLOCK_DEP(lock_flags));

	if (lock_flags & XFS_ILOCK_EXCL)
		mrupdate_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags));
	else if (lock_flags & XFS_ILOCK_SHARED)
		mraccess_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags));
}

/*
 * This is just like xfs_ilock(), except that the caller
 * is guaranteed not to sleep.  It returns 1 if it gets
 * the requested locks and 0 otherwise.  If the IO lock is
 * obtained but the inode lock cannot be, then the IO lock
 * is dropped before returning.
 *
 * ip -- the inode being locked
 * lock_flags -- this parameter indicates the inode's locks to be
 *       to be locked.  See the comment for xfs_ilock() for a list
 *	 of valid values.
 */
int
xfs_ilock_nowait(
	xfs_inode_t		*ip,
	uint			lock_flags)
{
	trace_xfs_ilock_nowait(ip, lock_flags, _RET_IP_);

	/*
	 * You can't set both SHARED and EXCL for the same lock,
	 * and only XFS_IOLOCK_SHARED, XFS_IOLOCK_EXCL, XFS_ILOCK_SHARED,
	 * and XFS_ILOCK_EXCL are valid values to set in lock_flags.
	 */
	ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) !=
	       (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL));
	ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) !=
	       (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
	ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_DEP_MASK)) == 0);

	if (lock_flags & XFS_IOLOCK_EXCL) {
		if (!mrtryupdate(&ip->i_iolock))
			goto out;
	} else if (lock_flags & XFS_IOLOCK_SHARED) {
		if (!mrtryaccess(&ip->i_iolock))
			goto out;
	}
	if (lock_flags & XFS_ILOCK_EXCL) {
		if (!mrtryupdate(&ip->i_lock))
			goto out_undo_iolock;
	} else if (lock_flags & XFS_ILOCK_SHARED) {
		if (!mrtryaccess(&ip->i_lock))
			goto out_undo_iolock;
	}
	return 1;

 out_undo_iolock:
	if (lock_flags & XFS_IOLOCK_EXCL)
		mrunlock_excl(&ip->i_iolock);
	else if (lock_flags & XFS_IOLOCK_SHARED)
		mrunlock_shared(&ip->i_iolock);
 out:
	return 0;
}

/*
 * xfs_iunlock() is used to drop the inode locks acquired with
 * xfs_ilock() and xfs_ilock_nowait().  The caller must pass
 * in the flags given to xfs_ilock() or xfs_ilock_nowait() so
 * that we know which locks to drop.
 *
 * ip -- the inode being unlocked
 * lock_flags -- this parameter indicates the inode's locks to be
 *       to be unlocked.  See the comment for xfs_ilock() for a list
 *	 of valid values for this parameter.
 *
 */
void
xfs_iunlock(
	xfs_inode_t		*ip,
	uint			lock_flags)
{
	/*
	 * You can't set both SHARED and EXCL for the same lock,
	 * and only XFS_IOLOCK_SHARED, XFS_IOLOCK_EXCL, XFS_ILOCK_SHARED,
	 * and XFS_ILOCK_EXCL are valid values to set in lock_flags.
	 */
	ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) !=
	       (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL));
	ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) !=
	       (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
	ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_DEP_MASK)) == 0);
	ASSERT(lock_flags != 0);

	if (lock_flags & XFS_IOLOCK_EXCL)
		mrunlock_excl(&ip->i_iolock);
	else if (lock_flags & XFS_IOLOCK_SHARED)
		mrunlock_shared(&ip->i_iolock);

	if (lock_flags & XFS_ILOCK_EXCL)
		mrunlock_excl(&ip->i_lock);
	else if (lock_flags & XFS_ILOCK_SHARED)
		mrunlock_shared(&ip->i_lock);

	trace_xfs_iunlock(ip, lock_flags, _RET_IP_);
}

/*
 * give up write locks.  the i/o lock cannot be held nested
 * if it is being demoted.
 */
void
xfs_ilock_demote(
	xfs_inode_t		*ip,
	uint			lock_flags)
{
	ASSERT(lock_flags & (XFS_IOLOCK_EXCL|XFS_ILOCK_EXCL));
	ASSERT((lock_flags & ~(XFS_IOLOCK_EXCL|XFS_ILOCK_EXCL)) == 0);

	if (lock_flags & XFS_ILOCK_EXCL)
		mrdemote(&ip->i_lock);
	if (lock_flags & XFS_IOLOCK_EXCL)
		mrdemote(&ip->i_iolock);

	trace_xfs_ilock_demote(ip, lock_flags, _RET_IP_);
}

287
#if defined(DEBUG) || defined(XFS_WARN)
288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309
int
xfs_isilocked(
	xfs_inode_t		*ip,
	uint			lock_flags)
{
	if (lock_flags & (XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)) {
		if (!(lock_flags & XFS_ILOCK_SHARED))
			return !!ip->i_lock.mr_writer;
		return rwsem_is_locked(&ip->i_lock.mr_lock);
	}

	if (lock_flags & (XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED)) {
		if (!(lock_flags & XFS_IOLOCK_SHARED))
			return !!ip->i_iolock.mr_writer;
		return rwsem_is_locked(&ip->i_iolock.mr_lock);
	}

	ASSERT(0);
	return 0;
}
#endif

Dave Chinner's avatar
Dave Chinner committed
310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491
#ifdef DEBUG
int xfs_locked_n;
int xfs_small_retries;
int xfs_middle_retries;
int xfs_lots_retries;
int xfs_lock_delays;
#endif

/*
 * Bump the subclass so xfs_lock_inodes() acquires each lock with
 * a different value
 */
static inline int
xfs_lock_inumorder(int lock_mode, int subclass)
{
	if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL))
		lock_mode |= (subclass + XFS_LOCK_INUMORDER) << XFS_IOLOCK_SHIFT;
	if (lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL))
		lock_mode |= (subclass + XFS_LOCK_INUMORDER) << XFS_ILOCK_SHIFT;

	return lock_mode;
}

/*
 * The following routine will lock n inodes in exclusive mode.
 * We assume the caller calls us with the inodes in i_ino order.
 *
 * We need to detect deadlock where an inode that we lock
 * is in the AIL and we start waiting for another inode that is locked
 * by a thread in a long running transaction (such as truncate). This can
 * result in deadlock since the long running trans might need to wait
 * for the inode we just locked in order to push the tail and free space
 * in the log.
 */
void
xfs_lock_inodes(
	xfs_inode_t	**ips,
	int		inodes,
	uint		lock_mode)
{
	int		attempts = 0, i, j, try_lock;
	xfs_log_item_t	*lp;

	ASSERT(ips && (inodes >= 2)); /* we need at least two */

	try_lock = 0;
	i = 0;

again:
	for (; i < inodes; i++) {
		ASSERT(ips[i]);

		if (i && (ips[i] == ips[i-1]))	/* Already locked */
			continue;

		/*
		 * If try_lock is not set yet, make sure all locked inodes
		 * are not in the AIL.
		 * If any are, set try_lock to be used later.
		 */

		if (!try_lock) {
			for (j = (i - 1); j >= 0 && !try_lock; j--) {
				lp = (xfs_log_item_t *)ips[j]->i_itemp;
				if (lp && (lp->li_flags & XFS_LI_IN_AIL)) {
					try_lock++;
				}
			}
		}

		/*
		 * If any of the previous locks we have locked is in the AIL,
		 * we must TRY to get the second and subsequent locks. If
		 * we can't get any, we must release all we have
		 * and try again.
		 */

		if (try_lock) {
			/* try_lock must be 0 if i is 0. */
			/*
			 * try_lock means we have an inode locked
			 * that is in the AIL.
			 */
			ASSERT(i != 0);
			if (!xfs_ilock_nowait(ips[i], xfs_lock_inumorder(lock_mode, i))) {
				attempts++;

				/*
				 * Unlock all previous guys and try again.
				 * xfs_iunlock will try to push the tail
				 * if the inode is in the AIL.
				 */

				for(j = i - 1; j >= 0; j--) {

					/*
					 * Check to see if we've already
					 * unlocked this one.
					 * Not the first one going back,
					 * and the inode ptr is the same.
					 */
					if ((j != (i - 1)) && ips[j] ==
								ips[j+1])
						continue;

					xfs_iunlock(ips[j], lock_mode);
				}

				if ((attempts % 5) == 0) {
					delay(1); /* Don't just spin the CPU */
#ifdef DEBUG
					xfs_lock_delays++;
#endif
				}
				i = 0;
				try_lock = 0;
				goto again;
			}
		} else {
			xfs_ilock(ips[i], xfs_lock_inumorder(lock_mode, i));
		}
	}

#ifdef DEBUG
	if (attempts) {
		if (attempts < 5) xfs_small_retries++;
		else if (attempts < 100) xfs_middle_retries++;
		else xfs_lots_retries++;
	} else {
		xfs_locked_n++;
	}
#endif
}

/*
 * xfs_lock_two_inodes() can only be used to lock one type of lock
 * at a time - the iolock or the ilock, but not both at once. If
 * we lock both at once, lockdep will report false positives saying
 * we have violated locking orders.
 */
void
xfs_lock_two_inodes(
	xfs_inode_t		*ip0,
	xfs_inode_t		*ip1,
	uint			lock_mode)
{
	xfs_inode_t		*temp;
	int			attempts = 0;
	xfs_log_item_t		*lp;

	if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL))
		ASSERT((lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)) == 0);
	ASSERT(ip0->i_ino != ip1->i_ino);

	if (ip0->i_ino > ip1->i_ino) {
		temp = ip0;
		ip0 = ip1;
		ip1 = temp;
	}

 again:
	xfs_ilock(ip0, xfs_lock_inumorder(lock_mode, 0));

	/*
	 * If the first lock we have locked is in the AIL, we must TRY to get
	 * the second lock. If we can't get it, we must release the first one
	 * and try again.
	 */
	lp = (xfs_log_item_t *)ip0->i_itemp;
	if (lp && (lp->li_flags & XFS_LI_IN_AIL)) {
		if (!xfs_ilock_nowait(ip1, xfs_lock_inumorder(lock_mode, 1))) {
			xfs_iunlock(ip0, lock_mode);
			if ((++attempts % 5) == 0)
				delay(1); /* Don't just spin the CPU */
			goto again;
		}
	} else {
		xfs_ilock(ip1, xfs_lock_inumorder(lock_mode, 1));
	}
}


492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507
void
__xfs_iflock(
	struct xfs_inode	*ip)
{
	wait_queue_head_t *wq = bit_waitqueue(&ip->i_flags, __XFS_IFLOCK_BIT);
	DEFINE_WAIT_BIT(wait, &ip->i_flags, __XFS_IFLOCK_BIT);

	do {
		prepare_to_wait_exclusive(wq, &wait.wait, TASK_UNINTERRUPTIBLE);
		if (xfs_isiflocked(ip))
			io_schedule();
	} while (!xfs_iflock_nowait(ip));

	finish_wait(wq, &wait.wait);
}

Linus Torvalds's avatar
Linus Torvalds committed
508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534
STATIC uint
_xfs_dic2xflags(
	__uint16_t		di_flags)
{
	uint			flags = 0;

	if (di_flags & XFS_DIFLAG_ANY) {
		if (di_flags & XFS_DIFLAG_REALTIME)
			flags |= XFS_XFLAG_REALTIME;
		if (di_flags & XFS_DIFLAG_PREALLOC)
			flags |= XFS_XFLAG_PREALLOC;
		if (di_flags & XFS_DIFLAG_IMMUTABLE)
			flags |= XFS_XFLAG_IMMUTABLE;
		if (di_flags & XFS_DIFLAG_APPEND)
			flags |= XFS_XFLAG_APPEND;
		if (di_flags & XFS_DIFLAG_SYNC)
			flags |= XFS_XFLAG_SYNC;
		if (di_flags & XFS_DIFLAG_NOATIME)
			flags |= XFS_XFLAG_NOATIME;
		if (di_flags & XFS_DIFLAG_NODUMP)
			flags |= XFS_XFLAG_NODUMP;
		if (di_flags & XFS_DIFLAG_RTINHERIT)
			flags |= XFS_XFLAG_RTINHERIT;
		if (di_flags & XFS_DIFLAG_PROJINHERIT)
			flags |= XFS_XFLAG_PROJINHERIT;
		if (di_flags & XFS_DIFLAG_NOSYMLINKS)
			flags |= XFS_XFLAG_NOSYMLINKS;
535 536 537 538
		if (di_flags & XFS_DIFLAG_EXTSIZE)
			flags |= XFS_XFLAG_EXTSIZE;
		if (di_flags & XFS_DIFLAG_EXTSZINHERIT)
			flags |= XFS_XFLAG_EXTSZINHERIT;
539 540
		if (di_flags & XFS_DIFLAG_NODEFRAG)
			flags |= XFS_XFLAG_NODEFRAG;
541 542
		if (di_flags & XFS_DIFLAG_FILESTREAM)
			flags |= XFS_XFLAG_FILESTREAM;
Linus Torvalds's avatar
Linus Torvalds committed
543 544 545 546 547 548 549 550 551
	}

	return flags;
}

uint
xfs_ip2xflags(
	xfs_inode_t		*ip)
{
552
	xfs_icdinode_t		*dic = &ip->i_d;
Linus Torvalds's avatar
Linus Torvalds committed
553

554
	return _xfs_dic2xflags(dic->di_flags) |
555
				(XFS_IFORK_Q(ip) ? XFS_XFLAG_HASATTR : 0);
Linus Torvalds's avatar
Linus Torvalds committed
556 557 558 559
}

uint
xfs_dic2xflags(
560
	xfs_dinode_t		*dip)
Linus Torvalds's avatar
Linus Torvalds committed
561
{
562
	return _xfs_dic2xflags(be16_to_cpu(dip->di_flags)) |
563
				(XFS_DFORK_Q(dip) ? XFS_XFLAG_HASATTR : 0);
Linus Torvalds's avatar
Linus Torvalds committed
564 565
}

Dave Chinner's avatar
Dave Chinner committed
566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587
/*
 * Lookups up an inode from "name". If ci_name is not NULL, then a CI match
 * is allowed, otherwise it has to be an exact match. If a CI match is found,
 * ci_name->name will point to a the actual name (caller must free) or
 * will be set to NULL if an exact match is found.
 */
int
xfs_lookup(
	xfs_inode_t		*dp,
	struct xfs_name		*name,
	xfs_inode_t		**ipp,
	struct xfs_name		*ci_name)
{
	xfs_ino_t		inum;
	int			error;
	uint			lock_mode;

	trace_xfs_lookup(dp, name);

	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
		return XFS_ERROR(EIO);

588
	lock_mode = xfs_ilock_data_map_shared(dp);
Dave Chinner's avatar
Dave Chinner committed
589
	error = xfs_dir_lookup(NULL, dp, name, &inum, ci_name);
590
	xfs_iunlock(dp, lock_mode);
Dave Chinner's avatar
Dave Chinner committed
591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608

	if (error)
		goto out;

	error = xfs_iget(dp->i_mount, NULL, inum, 0, 0, ipp);
	if (error)
		goto out_free_name;

	return 0;

out_free_name:
	if (ci_name)
		kmem_free(ci_name->name);
out:
	*ipp = NULL;
	return error;
}

Linus Torvalds's avatar
Linus Torvalds committed
609 610 611 612 613 614 615
/*
 * Allocate an inode on disk and return a copy of its in-core version.
 * The in-core inode is locked exclusively.  Set mode, nlink, and rdev
 * appropriately within the inode.  The uid and gid for the inode are
 * set according to the contents of the given cred structure.
 *
 * Use xfs_dialloc() to allocate the on-disk inode. If xfs_dialloc()
616 617 618 619
 * has a free inode available, call xfs_iget() to obtain the in-core
 * version of the allocated inode.  Finally, fill in the inode and
 * log its initial contents.  In this case, ialloc_context would be
 * set to NULL.
Linus Torvalds's avatar
Linus Torvalds committed
620
 *
621 622 623 624 625
 * If xfs_dialloc() does not have an available inode, it will replenish
 * its supply by doing an allocation. Since we can only do one
 * allocation within a transaction without deadlocks, we must commit
 * the current transaction before returning the inode itself.
 * In this case, therefore, we will set ialloc_context and return.
Linus Torvalds's avatar
Linus Torvalds committed
626 627 628 629 630 631 632 633
 * The caller should then commit the current transaction, start a new
 * transaction, and call xfs_ialloc() again to actually get the inode.
 *
 * To ensure that some other process does not grab the inode that
 * was allocated during the first call to xfs_ialloc(), this routine
 * also returns the [locked] bp pointing to the head of the freelist
 * as ialloc_context.  The caller should hold this buffer across
 * the commit and pass it back into this routine on the second call.
634 635 636 637 638
 *
 * If we are allocating quota inodes, we do not have a parent inode
 * to attach to or associate with (i.e. pip == NULL) because they
 * are not linked into the directory structure - they are attached
 * directly to the superblock - and so have no parent.
Linus Torvalds's avatar
Linus Torvalds committed
639 640 641 642 643
 */
int
xfs_ialloc(
	xfs_trans_t	*tp,
	xfs_inode_t	*pip,
Al Viro's avatar
Al Viro committed
644
	umode_t		mode,
645
	xfs_nlink_t	nlink,
Linus Torvalds's avatar
Linus Torvalds committed
646
	xfs_dev_t	rdev,
647
	prid_t		prid,
Linus Torvalds's avatar
Linus Torvalds committed
648 649 650 651
	int		okalloc,
	xfs_buf_t	**ialloc_context,
	xfs_inode_t	**ipp)
{
652
	struct xfs_mount *mp = tp->t_mountp;
Linus Torvalds's avatar
Linus Torvalds committed
653 654 655 656
	xfs_ino_t	ino;
	xfs_inode_t	*ip;
	uint		flags;
	int		error;
657
	timespec_t	tv;
Linus Torvalds's avatar
Linus Torvalds committed
658 659 660 661 662

	/*
	 * Call the space management code to pick
	 * the on-disk inode to be allocated.
	 */
663
	error = xfs_dialloc(tp, pip ? pip->i_ino : 0, mode, okalloc,
664
			    ialloc_context, &ino);
665
	if (error)
Linus Torvalds's avatar
Linus Torvalds committed
666
		return error;
667
	if (*ialloc_context || ino == NULLFSINO) {
Linus Torvalds's avatar
Linus Torvalds committed
668 669 670 671 672 673 674 675 676 677
		*ipp = NULL;
		return 0;
	}
	ASSERT(*ialloc_context == NULL);

	/*
	 * Get the in-core inode with the lock held exclusively.
	 * This is because we're setting fields here we need
	 * to prevent others from looking at until we're done.
	 */
678
	error = xfs_iget(mp, tp, ino, XFS_IGET_CREATE,
679
			 XFS_ILOCK_EXCL, &ip);
680
	if (error)
Linus Torvalds's avatar
Linus Torvalds committed
681 682 683
		return error;
	ASSERT(ip != NULL);

Al Viro's avatar
Al Viro committed
684
	ip->i_d.di_mode = mode;
Linus Torvalds's avatar
Linus Torvalds committed
685 686 687
	ip->i_d.di_onlink = 0;
	ip->i_d.di_nlink = nlink;
	ASSERT(ip->i_d.di_nlink == nlink);
688 689
	ip->i_d.di_uid = xfs_kuid_to_uid(current_fsuid());
	ip->i_d.di_gid = xfs_kgid_to_gid(current_fsgid());
690
	xfs_set_projid(ip, prid);
Linus Torvalds's avatar
Linus Torvalds committed
691 692 693 694 695 696 697 698
	memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));

	/*
	 * If the superblock version is up to where we support new format
	 * inodes and this is currently an old format inode, then change
	 * the inode version number now.  This way we only do the conversion
	 * here rather than here and in the flush/logging code.
	 */
699
	if (xfs_sb_version_hasnlink(&mp->m_sb) &&
700 701
	    ip->i_d.di_version == 1) {
		ip->i_d.di_version = 2;
Linus Torvalds's avatar
Linus Torvalds committed
702 703 704 705 706 707 708 709 710
		/*
		 * We've already zeroed the old link count, the projid field,
		 * and the pad field.
		 */
	}

	/*
	 * Project ids won't be stored on disk if we are using a version 1 inode.
	 */
711
	if ((prid != 0) && (ip->i_d.di_version == 1))
Linus Torvalds's avatar
Linus Torvalds committed
712 713
		xfs_bump_ino_vers2(tp, ip);

714
	if (pip && XFS_INHERIT_GID(pip)) {
Linus Torvalds's avatar
Linus Torvalds committed
715
		ip->i_d.di_gid = pip->i_d.di_gid;
716
		if ((pip->i_d.di_mode & S_ISGID) && S_ISDIR(mode)) {
Linus Torvalds's avatar
Linus Torvalds committed
717 718 719 720 721 722 723 724 725 726 727
			ip->i_d.di_mode |= S_ISGID;
		}
	}

	/*
	 * If the group ID of the new file does not match the effective group
	 * ID or one of the supplementary group IDs, the S_ISGID bit is cleared
	 * (and only if the irix_sgid_inherit compatibility variable is set).
	 */
	if ((irix_sgid_inherit) &&
	    (ip->i_d.di_mode & S_ISGID) &&
728
	    (!in_group_p(xfs_gid_to_kgid(ip->i_d.di_gid)))) {
Linus Torvalds's avatar
Linus Torvalds committed
729 730 731 732 733 734
		ip->i_d.di_mode &= ~S_ISGID;
	}

	ip->i_d.di_size = 0;
	ip->i_d.di_nextents = 0;
	ASSERT(ip->i_d.di_nblocks == 0);
735 736 737 738 739 740 741

	nanotime(&tv);
	ip->i_d.di_mtime.t_sec = (__int32_t)tv.tv_sec;
	ip->i_d.di_mtime.t_nsec = (__int32_t)tv.tv_nsec;
	ip->i_d.di_atime = ip->i_d.di_mtime;
	ip->i_d.di_ctime = ip->i_d.di_mtime;

Linus Torvalds's avatar
Linus Torvalds committed
742 743 744 745 746 747 748
	/*
	 * di_gen will have been taken care of in xfs_iread.
	 */
	ip->i_d.di_extsize = 0;
	ip->i_d.di_dmevmask = 0;
	ip->i_d.di_dmstate = 0;
	ip->i_d.di_flags = 0;
749 750 751 752 753 754 755 756 757 758 759 760 761

	if (ip->i_d.di_version == 3) {
		ASSERT(ip->i_d.di_ino == ino);
		ASSERT(uuid_equal(&ip->i_d.di_uuid, &mp->m_sb.sb_uuid));
		ip->i_d.di_crc = 0;
		ip->i_d.di_changecount = 1;
		ip->i_d.di_lsn = 0;
		ip->i_d.di_flags2 = 0;
		memset(&(ip->i_d.di_pad2[0]), 0, sizeof(ip->i_d.di_pad2));
		ip->i_d.di_crtime = ip->i_d.di_mtime;
	}


Linus Torvalds's avatar
Linus Torvalds committed
762 763 764 765 766 767 768 769 770 771 772 773 774
	flags = XFS_ILOG_CORE;
	switch (mode & S_IFMT) {
	case S_IFIFO:
	case S_IFCHR:
	case S_IFBLK:
	case S_IFSOCK:
		ip->i_d.di_format = XFS_DINODE_FMT_DEV;
		ip->i_df.if_u2.if_rdev = rdev;
		ip->i_df.if_flags = 0;
		flags |= XFS_ILOG_DEV;
		break;
	case S_IFREG:
	case S_IFDIR:
775
		if (pip && (pip->i_d.di_flags & XFS_DIFLAG_ANY)) {
776 777
			uint	di_flags = 0;

778
			if (S_ISDIR(mode)) {
779 780
				if (pip->i_d.di_flags & XFS_DIFLAG_RTINHERIT)
					di_flags |= XFS_DIFLAG_RTINHERIT;
781 782 783 784
				if (pip->i_d.di_flags & XFS_DIFLAG_EXTSZINHERIT) {
					di_flags |= XFS_DIFLAG_EXTSZINHERIT;
					ip->i_d.di_extsize = pip->i_d.di_extsize;
				}
785
			} else if (S_ISREG(mode)) {
Christoph Hellwig's avatar
Christoph Hellwig committed
786
				if (pip->i_d.di_flags & XFS_DIFLAG_RTINHERIT)
787
					di_flags |= XFS_DIFLAG_REALTIME;
788 789 790 791
				if (pip->i_d.di_flags & XFS_DIFLAG_EXTSZINHERIT) {
					di_flags |= XFS_DIFLAG_EXTSIZE;
					ip->i_d.di_extsize = pip->i_d.di_extsize;
				}
Linus Torvalds's avatar
Linus Torvalds committed
792 793 794
			}
			if ((pip->i_d.di_flags & XFS_DIFLAG_NOATIME) &&
			    xfs_inherit_noatime)
795
				di_flags |= XFS_DIFLAG_NOATIME;
Linus Torvalds's avatar
Linus Torvalds committed
796 797
			if ((pip->i_d.di_flags & XFS_DIFLAG_NODUMP) &&
			    xfs_inherit_nodump)
798
				di_flags |= XFS_DIFLAG_NODUMP;
Linus Torvalds's avatar
Linus Torvalds committed
799 800
			if ((pip->i_d.di_flags & XFS_DIFLAG_SYNC) &&
			    xfs_inherit_sync)
801
				di_flags |= XFS_DIFLAG_SYNC;
Linus Torvalds's avatar
Linus Torvalds committed
802 803
			if ((pip->i_d.di_flags & XFS_DIFLAG_NOSYMLINKS) &&
			    xfs_inherit_nosymlinks)
804 805 806
				di_flags |= XFS_DIFLAG_NOSYMLINKS;
			if (pip->i_d.di_flags & XFS_DIFLAG_PROJINHERIT)
				di_flags |= XFS_DIFLAG_PROJINHERIT;
807 808 809
			if ((pip->i_d.di_flags & XFS_DIFLAG_NODEFRAG) &&
			    xfs_inherit_nodefrag)
				di_flags |= XFS_DIFLAG_NODEFRAG;
810 811
			if (pip->i_d.di_flags & XFS_DIFLAG_FILESTREAM)
				di_flags |= XFS_DIFLAG_FILESTREAM;
812
			ip->i_d.di_flags |= di_flags;
Linus Torvalds's avatar
Linus Torvalds committed
813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832
		}
		/* FALLTHROUGH */
	case S_IFLNK:
		ip->i_d.di_format = XFS_DINODE_FMT_EXTENTS;
		ip->i_df.if_flags = XFS_IFEXTENTS;
		ip->i_df.if_bytes = ip->i_df.if_real_bytes = 0;
		ip->i_df.if_u1.if_extents = NULL;
		break;
	default:
		ASSERT(0);
	}
	/*
	 * Attribute fork settings for new inode.
	 */
	ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
	ip->i_d.di_anextents = 0;

	/*
	 * Log the new values stuffed into the inode.
	 */
833
	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
Linus Torvalds's avatar
Linus Torvalds committed
834 835
	xfs_trans_log_inode(tp, ip, flags);

836
	/* now that we have an i_mode we can setup inode ops and unlock */
837
	xfs_setup_inode(ip);
Linus Torvalds's avatar
Linus Torvalds committed
838 839 840 841 842

	*ipp = ip;
	return 0;
}

Dave Chinner's avatar
Dave Chinner committed
843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918
/*
 * Allocates a new inode from disk and return a pointer to the
 * incore copy. This routine will internally commit the current
 * transaction and allocate a new one if the Space Manager needed
 * to do an allocation to replenish the inode free-list.
 *
 * This routine is designed to be called from xfs_create and
 * xfs_create_dir.
 *
 */
int
xfs_dir_ialloc(
	xfs_trans_t	**tpp,		/* input: current transaction;
					   output: may be a new transaction. */
	xfs_inode_t	*dp,		/* directory within whose allocate
					   the inode. */
	umode_t		mode,
	xfs_nlink_t	nlink,
	xfs_dev_t	rdev,
	prid_t		prid,		/* project id */
	int		okalloc,	/* ok to allocate new space */
	xfs_inode_t	**ipp,		/* pointer to inode; it will be
					   locked. */
	int		*committed)

{
	xfs_trans_t	*tp;
	xfs_trans_t	*ntp;
	xfs_inode_t	*ip;
	xfs_buf_t	*ialloc_context = NULL;
	int		code;
	void		*dqinfo;
	uint		tflags;

	tp = *tpp;
	ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES);

	/*
	 * xfs_ialloc will return a pointer to an incore inode if
	 * the Space Manager has an available inode on the free
	 * list. Otherwise, it will do an allocation and replenish
	 * the freelist.  Since we can only do one allocation per
	 * transaction without deadlocks, we will need to commit the
	 * current transaction and start a new one.  We will then
	 * need to call xfs_ialloc again to get the inode.
	 *
	 * If xfs_ialloc did an allocation to replenish the freelist,
	 * it returns the bp containing the head of the freelist as
	 * ialloc_context. We will hold a lock on it across the
	 * transaction commit so that no other process can steal
	 * the inode(s) that we've just allocated.
	 */
	code = xfs_ialloc(tp, dp, mode, nlink, rdev, prid, okalloc,
			  &ialloc_context, &ip);

	/*
	 * Return an error if we were unable to allocate a new inode.
	 * This should only happen if we run out of space on disk or
	 * encounter a disk error.
	 */
	if (code) {
		*ipp = NULL;
		return code;
	}
	if (!ialloc_context && !ip) {
		*ipp = NULL;
		return XFS_ERROR(ENOSPC);
	}

	/*
	 * If the AGI buffer is non-NULL, then we were unable to get an
	 * inode in one operation.  We need to commit the current
	 * transaction and call xfs_ialloc() again.  It is guaranteed
	 * to succeed the second time.
	 */
	if (ialloc_context) {
919 920
		struct xfs_trans_res tres;

Dave Chinner's avatar
Dave Chinner committed
921 922 923 924 925 926 927 928 929 930 931 932
		/*
		 * Normally, xfs_trans_commit releases all the locks.
		 * We call bhold to hang on to the ialloc_context across
		 * the commit.  Holding this buffer prevents any other
		 * processes from doing any allocations in this
		 * allocation group.
		 */
		xfs_trans_bhold(tp, ialloc_context);
		/*
		 * Save the log reservation so we can use
		 * them in the next transaction.
		 */
933 934
		tres.tr_logres = xfs_trans_get_log_res(tp);
		tres.tr_logcount = xfs_trans_get_log_count(tp);
Dave Chinner's avatar
Dave Chinner committed
935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976

		/*
		 * We want the quota changes to be associated with the next
		 * transaction, NOT this one. So, detach the dqinfo from this
		 * and attach it to the next transaction.
		 */
		dqinfo = NULL;
		tflags = 0;
		if (tp->t_dqinfo) {
			dqinfo = (void *)tp->t_dqinfo;
			tp->t_dqinfo = NULL;
			tflags = tp->t_flags & XFS_TRANS_DQ_DIRTY;
			tp->t_flags &= ~(XFS_TRANS_DQ_DIRTY);
		}

		ntp = xfs_trans_dup(tp);
		code = xfs_trans_commit(tp, 0);
		tp = ntp;
		if (committed != NULL) {
			*committed = 1;
		}
		/*
		 * If we get an error during the commit processing,
		 * release the buffer that is still held and return
		 * to the caller.
		 */
		if (code) {
			xfs_buf_relse(ialloc_context);
			if (dqinfo) {
				tp->t_dqinfo = dqinfo;
				xfs_trans_free_dqinfo(tp);
			}
			*tpp = ntp;
			*ipp = NULL;
			return code;
		}

		/*
		 * transaction commit worked ok so we can drop the extra ticket
		 * reference that we gained in xfs_trans_dup()
		 */
		xfs_log_ticket_put(tp->t_ticket);
977 978 979
		tres.tr_logflags = XFS_TRANS_PERM_LOG_RES;
		code = xfs_trans_reserve(tp, &tres, 0, 0);

Dave Chinner's avatar
Dave Chinner committed
980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101
		/*
		 * Re-attach the quota info that we detached from prev trx.
		 */
		if (dqinfo) {
			tp->t_dqinfo = dqinfo;
			tp->t_flags |= tflags;
		}

		if (code) {
			xfs_buf_relse(ialloc_context);
			*tpp = ntp;
			*ipp = NULL;
			return code;
		}
		xfs_trans_bjoin(tp, ialloc_context);

		/*
		 * Call ialloc again. Since we've locked out all
		 * other allocations in this allocation group,
		 * this call should always succeed.
		 */
		code = xfs_ialloc(tp, dp, mode, nlink, rdev, prid,
				  okalloc, &ialloc_context, &ip);

		/*
		 * If we get an error at this point, return to the caller
		 * so that the current transaction can be aborted.
		 */
		if (code) {
			*tpp = tp;
			*ipp = NULL;
			return code;
		}
		ASSERT(!ialloc_context && ip);

	} else {
		if (committed != NULL)
			*committed = 0;
	}

	*ipp = ip;
	*tpp = tp;

	return 0;
}

/*
 * Decrement the link count on an inode & log the change.
 * If this causes the link count to go to zero, initiate the
 * logging activity required to truncate a file.
 */
int				/* error */
xfs_droplink(
	xfs_trans_t *tp,
	xfs_inode_t *ip)
{
	int	error;

	xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);

	ASSERT (ip->i_d.di_nlink > 0);
	ip->i_d.di_nlink--;
	drop_nlink(VFS_I(ip));
	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);

	error = 0;
	if (ip->i_d.di_nlink == 0) {
		/*
		 * We're dropping the last link to this file.
		 * Move the on-disk inode to the AGI unlinked list.
		 * From xfs_inactive() we will pull the inode from
		 * the list and free it.
		 */
		error = xfs_iunlink(tp, ip);
	}
	return error;
}

/*
 * This gets called when the inode's version needs to be changed from 1 to 2.
 * Currently this happens when the nlink field overflows the old 16-bit value
 * or when chproj is called to change the project for the first time.
 * As a side effect the superblock version will also get rev'd
 * to contain the NLINK bit.
 */
void
xfs_bump_ino_vers2(
	xfs_trans_t	*tp,
	xfs_inode_t	*ip)
{
	xfs_mount_t	*mp;

	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
	ASSERT(ip->i_d.di_version == 1);

	ip->i_d.di_version = 2;
	ip->i_d.di_onlink = 0;
	memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
	mp = tp->t_mountp;
	if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
		spin_lock(&mp->m_sb_lock);
		if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
			xfs_sb_version_addnlink(&mp->m_sb);
			spin_unlock(&mp->m_sb_lock);
			xfs_mod_sb(tp, XFS_SB_VERSIONNUM);
		} else {
			spin_unlock(&mp->m_sb_lock);
		}
	}
	/* Caller must log the inode */
}

/*
 * Increment the link count on an inode & log the change.
 */
int
xfs_bumplink(
	xfs_trans_t *tp,
	xfs_inode_t *ip)
{
	xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);

1102
	ASSERT(ip->i_d.di_nlink > 0 || (VFS_I(ip)->i_state & I_LINKABLE));
Dave Chinner's avatar
Dave Chinner committed
1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121
	ip->i_d.di_nlink++;
	inc_nlink(VFS_I(ip));
	if ((ip->i_d.di_version == 1) &&
	    (ip->i_d.di_nlink > XFS_MAXLINK_1)) {
		/*
		 * The inode has increased its number of links beyond
		 * what can fit in an old format inode.  It now needs
		 * to be converted to a version 2 inode with a 32 bit
		 * link count.  If this is the first inode in the file
		 * system to do this, then we need to bump the superblock
		 * version number as well.
		 */
		xfs_bump_ino_vers2(tp, ip);
	}

	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
	return 0;
}

Dave Chinner's avatar
Dave Chinner committed
1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143
int
xfs_create(
	xfs_inode_t		*dp,
	struct xfs_name		*name,
	umode_t			mode,
	xfs_dev_t		rdev,
	xfs_inode_t		**ipp)
{
	int			is_dir = S_ISDIR(mode);
	struct xfs_mount	*mp = dp->i_mount;
	struct xfs_inode	*ip = NULL;
	struct xfs_trans	*tp = NULL;
	int			error;
	xfs_bmap_free_t		free_list;
	xfs_fsblock_t		first_block;
	bool                    unlock_dp_on_error = false;
	uint			cancel_flags;
	int			committed;
	prid_t			prid;
	struct xfs_dquot	*udqp = NULL;
	struct xfs_dquot	*gdqp = NULL;
	struct xfs_dquot	*pdqp = NULL;
1144
	struct xfs_trans_res	tres;
Dave Chinner's avatar
Dave Chinner committed
1145 1146 1147 1148 1149 1150 1151
	uint			resblks;

	trace_xfs_create(dp, name);

	if (XFS_FORCED_SHUTDOWN(mp))
		return XFS_ERROR(EIO);

1152
	prid = xfs_get_initial_prid(dp);
Dave Chinner's avatar
Dave Chinner committed
1153 1154 1155 1156

	/*
	 * Make sure that we have allocated dquot(s) on disk.
	 */
1157 1158
	error = xfs_qm_vop_dqalloc(dp, xfs_kuid_to_uid(current_fsuid()),
					xfs_kgid_to_gid(current_fsgid()), prid,
Dave Chinner's avatar
Dave Chinner committed
1159 1160 1161 1162 1163 1164 1165 1166
					XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT,
					&udqp, &gdqp, &pdqp);
	if (error)
		return error;

	if (is_dir) {
		rdev = 0;
		resblks = XFS_MKDIR_SPACE_RES(mp, name->len);
1167 1168
		tres.tr_logres = M_RES(mp)->tr_mkdir.tr_logres;
		tres.tr_logcount = XFS_MKDIR_LOG_COUNT;
Dave Chinner's avatar
Dave Chinner committed
1169 1170 1171
		tp = xfs_trans_alloc(mp, XFS_TRANS_MKDIR);
	} else {
		resblks = XFS_CREATE_SPACE_RES(mp, name->len);
1172 1173
		tres.tr_logres = M_RES(mp)->tr_create.tr_logres;
		tres.tr_logcount = XFS_CREATE_LOG_COUNT;
Dave Chinner's avatar
Dave Chinner committed
1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184
		tp = xfs_trans_alloc(mp, XFS_TRANS_CREATE);
	}

	cancel_flags = XFS_TRANS_RELEASE_LOG_RES;

	/*
	 * Initially assume that the file does not exist and
	 * reserve the resources for that case.  If that is not
	 * the case we'll drop the one we have and get a more
	 * appropriate transaction later.
	 */
1185 1186
	tres.tr_logflags = XFS_TRANS_PERM_LOG_RES;
	error = xfs_trans_reserve(tp, &tres, resblks, 0);
Dave Chinner's avatar
Dave Chinner committed
1187 1188 1189
	if (error == ENOSPC) {
		/* flush outstanding delalloc blocks and retry */
		xfs_flush_inodes(mp);
1190
		error = xfs_trans_reserve(tp, &tres, resblks, 0);
Dave Chinner's avatar
Dave Chinner committed
1191 1192 1193 1194
	}
	if (error == ENOSPC) {
		/* No space at all so try a "no-allocation" reservation */
		resblks = 0;
1195
		error = xfs_trans_reserve(tp, &tres, 0, 0);
Dave Chinner's avatar
Dave Chinner committed
1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 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 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315
	}
	if (error) {
		cancel_flags = 0;
		goto out_trans_cancel;
	}

	xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
	unlock_dp_on_error = true;

	xfs_bmap_init(&free_list, &first_block);

	/*
	 * Reserve disk quota and the inode.
	 */
	error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp,
						pdqp, resblks, 1, 0);
	if (error)
		goto out_trans_cancel;

	error = xfs_dir_canenter(tp, dp, name, resblks);
	if (error)
		goto out_trans_cancel;

	/*
	 * A newly created regular or special file just has one directory
	 * entry pointing to them, but a directory also the "." entry
	 * pointing to itself.
	 */
	error = xfs_dir_ialloc(&tp, dp, mode, is_dir ? 2 : 1, rdev,
			       prid, resblks > 0, &ip, &committed);
	if (error) {
		if (error == ENOSPC)
			goto out_trans_cancel;
		goto out_trans_abort;
	}

	/*
	 * Now we join the directory inode to the transaction.  We do not do it
	 * earlier because xfs_dir_ialloc might commit the previous transaction
	 * (and release all the locks).  An error from here on will result in
	 * the transaction cancel unlocking dp so don't do it explicitly in the
	 * error path.
	 */
	xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
	unlock_dp_on_error = false;

	error = xfs_dir_createname(tp, dp, name, ip->i_ino,
					&first_block, &free_list, resblks ?
					resblks - XFS_IALLOC_SPACE_RES(mp) : 0);
	if (error) {
		ASSERT(error != ENOSPC);
		goto out_trans_abort;
	}
	xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
	xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);

	if (is_dir) {
		error = xfs_dir_init(tp, ip, dp);
		if (error)
			goto out_bmap_cancel;

		error = xfs_bumplink(tp, dp);
		if (error)
			goto out_bmap_cancel;
	}

	/*
	 * If this is a synchronous mount, make sure that the
	 * create transaction goes to disk before returning to
	 * the user.
	 */
	if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC))
		xfs_trans_set_sync(tp);

	/*
	 * Attach the dquot(s) to the inodes and modify them incore.
	 * These ids of the inode couldn't have changed since the new
	 * inode has been locked ever since it was created.
	 */
	xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp);

	error = xfs_bmap_finish(&tp, &free_list, &committed);
	if (error)
		goto out_bmap_cancel;

	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
	if (error)
		goto out_release_inode;

	xfs_qm_dqrele(udqp);
	xfs_qm_dqrele(gdqp);
	xfs_qm_dqrele(pdqp);

	*ipp = ip;
	return 0;

 out_bmap_cancel:
	xfs_bmap_cancel(&free_list);
 out_trans_abort:
	cancel_flags |= XFS_TRANS_ABORT;
 out_trans_cancel:
	xfs_trans_cancel(tp, cancel_flags);
 out_release_inode:
	/*
	 * Wait until after the current transaction is aborted to
	 * release the inode.  This prevents recursive transactions
	 * and deadlocks from xfs_inactive.
	 */
	if (ip)
		IRELE(ip);

	xfs_qm_dqrele(udqp);
	xfs_qm_dqrele(gdqp);
	xfs_qm_dqrele(pdqp);

	if (unlock_dp_on_error)
		xfs_iunlock(dp, XFS_ILOCK_EXCL);
	return error;
}

Zhi Yong Wu's avatar
Zhi Yong Wu committed
1316 1317 1318 1319
int
xfs_create_tmpfile(
	struct xfs_inode	*dp,
	struct dentry		*dentry,
1320 1321
	umode_t			mode,
	struct xfs_inode	**ipp)
Zhi Yong Wu's avatar
Zhi Yong Wu committed
1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400
{
	struct xfs_mount	*mp = dp->i_mount;
	struct xfs_inode	*ip = NULL;
	struct xfs_trans	*tp = NULL;
	int			error;
	uint			cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
	prid_t                  prid;
	struct xfs_dquot	*udqp = NULL;
	struct xfs_dquot	*gdqp = NULL;
	struct xfs_dquot	*pdqp = NULL;
	struct xfs_trans_res	*tres;
	uint			resblks;

	if (XFS_FORCED_SHUTDOWN(mp))
		return XFS_ERROR(EIO);

	prid = xfs_get_initial_prid(dp);

	/*
	 * Make sure that we have allocated dquot(s) on disk.
	 */
	error = xfs_qm_vop_dqalloc(dp, xfs_kuid_to_uid(current_fsuid()),
				xfs_kgid_to_gid(current_fsgid()), prid,
				XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT,
				&udqp, &gdqp, &pdqp);
	if (error)
		return error;

	resblks = XFS_IALLOC_SPACE_RES(mp);
	tp = xfs_trans_alloc(mp, XFS_TRANS_CREATE_TMPFILE);

	tres = &M_RES(mp)->tr_create_tmpfile;
	error = xfs_trans_reserve(tp, tres, resblks, 0);
	if (error == ENOSPC) {
		/* No space at all so try a "no-allocation" reservation */
		resblks = 0;
		error = xfs_trans_reserve(tp, tres, 0, 0);
	}
	if (error) {
		cancel_flags = 0;
		goto out_trans_cancel;
	}

	error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp,
						pdqp, resblks, 1, 0);
	if (error)
		goto out_trans_cancel;

	error = xfs_dir_ialloc(&tp, dp, mode, 1, 0,
				prid, resblks > 0, &ip, NULL);
	if (error) {
		if (error == ENOSPC)
			goto out_trans_cancel;
		goto out_trans_abort;
	}

	if (mp->m_flags & XFS_MOUNT_WSYNC)
		xfs_trans_set_sync(tp);

	/*
	 * Attach the dquot(s) to the inodes and modify them incore.
	 * These ids of the inode couldn't have changed since the new
	 * inode has been locked ever since it was created.
	 */
	xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp);

	ip->i_d.di_nlink--;
	error = xfs_iunlink(tp, ip);
	if (error)
		goto out_trans_abort;

	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
	if (error)
		goto out_release_inode;

	xfs_qm_dqrele(udqp);
	xfs_qm_dqrele(gdqp);
	xfs_qm_dqrele(pdqp);

1401
	*ipp = ip;
Zhi Yong Wu's avatar
Zhi Yong Wu committed
1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
	return 0;

 out_trans_abort:
	cancel_flags |= XFS_TRANS_ABORT;
 out_trans_cancel:
	xfs_trans_cancel(tp, cancel_flags);
 out_release_inode:
	/*
	 * Wait until after the current transaction is aborted to
	 * release the inode.  This prevents recursive transactions
	 * and deadlocks from xfs_inactive.
	 */
	if (ip)
		IRELE(ip);

	xfs_qm_dqrele(udqp);
	xfs_qm_dqrele(gdqp);
	xfs_qm_dqrele(pdqp);

	return error;
}

Dave Chinner's avatar
Dave Chinner committed
1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456
int
xfs_link(
	xfs_inode_t		*tdp,
	xfs_inode_t		*sip,
	struct xfs_name		*target_name)
{
	xfs_mount_t		*mp = tdp->i_mount;
	xfs_trans_t		*tp;
	int			error;
	xfs_bmap_free_t         free_list;
	xfs_fsblock_t           first_block;
	int			cancel_flags;
	int			committed;
	int			resblks;

	trace_xfs_link(tdp, target_name);

	ASSERT(!S_ISDIR(sip->i_d.di_mode));

	if (XFS_FORCED_SHUTDOWN(mp))
		return XFS_ERROR(EIO);

	error = xfs_qm_dqattach(sip, 0);
	if (error)
		goto std_return;

	error = xfs_qm_dqattach(tdp, 0);
	if (error)
		goto std_return;

	tp = xfs_trans_alloc(mp, XFS_TRANS_LINK);
	cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
	resblks = XFS_LINK_SPACE_RES(mp, target_name->len);
1457
	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_link, resblks, 0);
Dave Chinner's avatar
Dave Chinner committed
1458 1459
	if (error == ENOSPC) {
		resblks = 0;
1460
		error = xfs_trans_reserve(tp, &M_RES(mp)->tr_link, 0, 0);
Dave Chinner's avatar
Dave Chinner committed
1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488
	}
	if (error) {
		cancel_flags = 0;
		goto error_return;
	}

	xfs_lock_two_inodes(sip, tdp, XFS_ILOCK_EXCL);

	xfs_trans_ijoin(tp, sip, XFS_ILOCK_EXCL);
	xfs_trans_ijoin(tp, tdp, XFS_ILOCK_EXCL);

	/*
	 * If we are using project inheritance, we only allow hard link
	 * creation in our tree when the project IDs are the same; else
	 * the tree quota mechanism could be circumvented.
	 */
	if (unlikely((tdp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) &&
		     (xfs_get_projid(tdp) != xfs_get_projid(sip)))) {
		error = XFS_ERROR(EXDEV);
		goto error_return;
	}

	error = xfs_dir_canenter(tp, tdp, target_name, resblks);
	if (error)
		goto error_return;

	xfs_bmap_init(&free_list, &first_block);

1489 1490 1491 1492 1493 1494
	if (sip->i_d.di_nlink == 0) {
		error = xfs_iunlink_remove(tp, sip);
		if (error)
			goto abort_return;
	}

Dave Chinner's avatar
Dave Chinner committed
1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530
	error = xfs_dir_createname(tp, tdp, target_name, sip->i_ino,
					&first_block, &free_list, resblks);
	if (error)
		goto abort_return;
	xfs_trans_ichgtime(tp, tdp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
	xfs_trans_log_inode(tp, tdp, XFS_ILOG_CORE);

	error = xfs_bumplink(tp, sip);
	if (error)
		goto abort_return;

	/*
	 * If this is a synchronous mount, make sure that the
	 * link transaction goes to disk before returning to
	 * the user.
	 */
	if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) {
		xfs_trans_set_sync(tp);
	}

	error = xfs_bmap_finish (&tp, &free_list, &committed);
	if (error) {
		xfs_bmap_cancel(&free_list);
		goto abort_return;
	}

	return xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);

 abort_return:
	cancel_flags |= XFS_TRANS_ABORT;
 error_return:
	xfs_trans_cancel(tp, cancel_flags);
 std_return:
	return error;
}

Linus Torvalds's avatar
Linus Torvalds committed
1531
/*
1532 1533 1534
 * Free up the underlying blocks past new_size.  The new size must be smaller
 * than the current size.  This routine can be used both for the attribute and
 * data fork, and does not modify the inode size, which is left to the caller.
Linus Torvalds's avatar
Linus Torvalds committed
1535
 *
1536 1537 1538 1539 1540 1541 1542 1543 1544
 * The transaction passed to this routine must have made a permanent log
 * reservation of at least XFS_ITRUNCATE_LOG_RES.  This routine may commit the
 * given transaction and start new ones, so make sure everything involved in
 * the transaction is tidy before calling here.  Some transaction will be
 * returned to the caller to be committed.  The incoming transaction must
 * already include the inode, and both inode locks must be held exclusively.
 * The inode must also be "held" within the transaction.  On return the inode
 * will be "held" within the returned transaction.  This routine does NOT
 * require any disk space to be reserved for it within the transaction.
Linus Torvalds's avatar
Linus Torvalds committed
1545
 *
1546 1547 1548 1549 1550
 * If we get an error, we must return with the inode locked and linked into the
 * current transaction. This keeps things simple for the higher level code,
 * because it always knows that the inode is locked and held in the transaction
 * that returns to it whether errors occur or not.  We don't mark the inode
 * dirty on error so that transactions can be easily aborted if possible.
Linus Torvalds's avatar
Linus Torvalds committed
1551 1552
 */
int
1553 1554 1555 1556 1557
xfs_itruncate_extents(
	struct xfs_trans	**tpp,
	struct xfs_inode	*ip,
	int			whichfork,
	xfs_fsize_t		new_size)
Linus Torvalds's avatar
Linus Torvalds committed
1558
{
1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569
	struct xfs_mount	*mp = ip->i_mount;
	struct xfs_trans	*tp = *tpp;
	struct xfs_trans	*ntp;
	xfs_bmap_free_t		free_list;
	xfs_fsblock_t		first_block;
	xfs_fileoff_t		first_unmap_block;
	xfs_fileoff_t		last_block;
	xfs_filblks_t		unmap_len;
	int			committed;
	int			error = 0;
	int			done = 0;
Linus Torvalds's avatar
Linus Torvalds committed
1570

1571 1572 1573
	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
	ASSERT(!atomic_read(&VFS_I(ip)->i_count) ||
	       xfs_isilocked(ip, XFS_IOLOCK_EXCL));
1574
	ASSERT(new_size <= XFS_ISIZE(ip));
1575
	ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES);
Linus Torvalds's avatar
Linus Torvalds committed
1576
	ASSERT(ip->i_itemp != NULL);
1577
	ASSERT(ip->i_itemp->ili_lock_flags == 0);
1578
	ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
Linus Torvalds's avatar
Linus Torvalds committed
1579

1580 1581
	trace_xfs_itruncate_extents_start(ip, new_size);

Linus Torvalds's avatar
Linus Torvalds committed
1582 1583 1584 1585 1586 1587 1588 1589 1590
	/*
	 * Since it is possible for space to become allocated beyond
	 * the end of the file (in a crash where the space is allocated
	 * but the inode size is not yet updated), simply remove any
	 * blocks which show up between the new EOF and the maximum
	 * possible file size.  If the first block to be removed is
	 * beyond the maximum file size (ie it is the same as last_block),
	 * then there is nothing to do.
	 */
1591
	first_unmap_block = XFS_B_TO_FSB(mp, (xfs_ufsize_t)new_size);
1592
	last_block = XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes);
1593 1594 1595 1596 1597
	if (first_unmap_block == last_block)
		return 0;

	ASSERT(first_unmap_block < last_block);
	unmap_len = last_block - first_unmap_block + 1;
Linus Torvalds's avatar
Linus Torvalds committed
1598
	while (!done) {
1599
		xfs_bmap_init(&free_list, &first_block);
1600
		error = xfs_bunmapi(tp, ip,
1601
				    first_unmap_block, unmap_len,
1602
				    xfs_bmapi_aflag(whichfork),
Linus Torvalds's avatar
Linus Torvalds committed
1603
				    XFS_ITRUNC_MAX_EXTENTS,
1604
				    &first_block, &free_list,
1605
				    &done);
1606 1607
		if (error)
			goto out_bmap_cancel;
Linus Torvalds's avatar
Linus Torvalds committed
1608 1609 1610 1611 1612

		/*
		 * Duplicate the transaction that has the permanent
		 * reservation and commit the old transaction.
		 */
1613
		error = xfs_bmap_finish(&tp, &free_list, &committed);
1614
		if (committed)
1615
			xfs_trans_ijoin(tp, ip, 0);
1616 1617
		if (error)
			goto out_bmap_cancel;
Linus Torvalds's avatar
Linus Torvalds committed
1618 1619 1620

		if (committed) {
			/*
1621
			 * Mark the inode dirty so it will be logged and
1622
			 * moved forward in the log as part of every commit.
Linus Torvalds's avatar
Linus Torvalds committed
1623
			 */
1624
			xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
Linus Torvalds's avatar
Linus Torvalds committed
1625
		}
1626

1627 1628 1629
		ntp = xfs_trans_dup(tp);
		error = xfs_trans_commit(tp, 0);
		tp = ntp;
1630

1631
		xfs_trans_ijoin(tp, ip, 0);
1632

1633
		if (error)
1634 1635
			goto out;

1636
		/*
1637
		 * Transaction commit worked ok so we can drop the extra ticket
1638 1639
		 * reference that we gained in xfs_trans_dup()
		 */
1640
		xfs_log_ticket_put(tp->t_ticket);
1641
		error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0);
1642
		if (error)
1643
			goto out;
Linus Torvalds's avatar
Linus Torvalds committed
1644
	}
1645

1646 1647 1648 1649 1650 1651 1652 1653
	/*
	 * Always re-log the inode so that our permanent transaction can keep
	 * on rolling it forward in the log.
	 */
	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);

	trace_xfs_itruncate_extents_end(ip, new_size);

1654 1655 1656 1657
out:
	*tpp = tp;
	return error;
out_bmap_cancel:
Linus Torvalds's avatar
Linus Torvalds committed
1658
	/*
1659 1660 1661
	 * If the bunmapi call encounters an error, return to the caller where
	 * the transaction can be properly aborted.  We just need to make sure
	 * we're not holding any resources that we were not when we came in.
Linus Torvalds's avatar
Linus Torvalds committed
1662
	 */
1663 1664 1665 1666
	xfs_bmap_cancel(&free_list);
	goto out;
}

Dave Chinner's avatar
Dave Chinner committed
1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744
int
xfs_release(
	xfs_inode_t	*ip)
{
	xfs_mount_t	*mp = ip->i_mount;
	int		error;

	if (!S_ISREG(ip->i_d.di_mode) || (ip->i_d.di_mode == 0))
		return 0;

	/* If this is a read-only mount, don't do this (would generate I/O) */
	if (mp->m_flags & XFS_MOUNT_RDONLY)
		return 0;

	if (!XFS_FORCED_SHUTDOWN(mp)) {
		int truncated;

		/*
		 * If we previously truncated this file and removed old data
		 * in the process, we want to initiate "early" writeout on
		 * the last close.  This is an attempt to combat the notorious
		 * NULL files problem which is particularly noticeable from a
		 * truncate down, buffered (re-)write (delalloc), followed by
		 * a crash.  What we are effectively doing here is
		 * significantly reducing the time window where we'd otherwise
		 * be exposed to that problem.
		 */
		truncated = xfs_iflags_test_and_clear(ip, XFS_ITRUNCATED);
		if (truncated) {
			xfs_iflags_clear(ip, XFS_IDIRTY_RELEASE);
			if (VN_DIRTY(VFS_I(ip)) && ip->i_delayed_blks > 0) {
				error = -filemap_flush(VFS_I(ip)->i_mapping);
				if (error)
					return error;
			}
		}
	}

	if (ip->i_d.di_nlink == 0)
		return 0;

	if (xfs_can_free_eofblocks(ip, false)) {

		/*
		 * If we can't get the iolock just skip truncating the blocks
		 * past EOF because we could deadlock with the mmap_sem
		 * otherwise.  We'll get another chance to drop them once the
		 * last reference to the inode is dropped, so we'll never leak
		 * blocks permanently.
		 *
		 * Further, check if the inode is being opened, written and
		 * closed frequently and we have delayed allocation blocks
		 * outstanding (e.g. streaming writes from the NFS server),
		 * truncating the blocks past EOF will cause fragmentation to
		 * occur.
		 *
		 * In this case don't do the truncation, either, but we have to
		 * be careful how we detect this case. Blocks beyond EOF show
		 * up as i_delayed_blks even when the inode is clean, so we
		 * need to truncate them away first before checking for a dirty
		 * release. Hence on the first dirty close we will still remove
		 * the speculative allocation, but after that we will leave it
		 * in place.
		 */
		if (xfs_iflags_test(ip, XFS_IDIRTY_RELEASE))
			return 0;

		error = xfs_free_eofblocks(mp, ip, true);
		if (error && error != EAGAIN)
			return error;

		/* delalloc blocks after truncation means it really is dirty */
		if (ip->i_delayed_blks)
			xfs_iflags_set(ip, XFS_IDIRTY_RELEASE);
	}
	return 0;
}

1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796
/*
 * xfs_inactive_truncate
 *
 * Called to perform a truncate when an inode becomes unlinked.
 */
STATIC int
xfs_inactive_truncate(
	struct xfs_inode *ip)
{
	struct xfs_mount	*mp = ip->i_mount;
	struct xfs_trans	*tp;
	int			error;

	tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0);
	if (error) {
		ASSERT(XFS_FORCED_SHUTDOWN(mp));
		xfs_trans_cancel(tp, 0);
		return error;
	}

	xfs_ilock(ip, XFS_ILOCK_EXCL);
	xfs_trans_ijoin(tp, ip, 0);

	/*
	 * Log the inode size first to prevent stale data exposure in the event
	 * of a system crash before the truncate completes. See the related
	 * comment in xfs_setattr_size() for details.
	 */
	ip->i_d.di_size = 0;
	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);

	error = xfs_itruncate_extents(&tp, ip, XFS_DATA_FORK, 0);
	if (error)
		goto error_trans_cancel;

	ASSERT(ip->i_d.di_nextents == 0);

	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
	if (error)
		goto error_unlock;

	xfs_iunlock(ip, XFS_ILOCK_EXCL);
	return 0;

error_trans_cancel:
	xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
error_unlock:
	xfs_iunlock(ip, XFS_ILOCK_EXCL);
	return error;
}

1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813
/*
 * xfs_inactive_ifree()
 *
 * Perform the inode free when an inode is unlinked.
 */
STATIC int
xfs_inactive_ifree(
	struct xfs_inode *ip)
{
	xfs_bmap_free_t		free_list;
	xfs_fsblock_t		first_block;
	int			committed;
	struct xfs_mount	*mp = ip->i_mount;
	struct xfs_trans	*tp;
	int			error;

	tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832

	/*
	 * The ifree transaction might need to allocate blocks for record
	 * insertion to the finobt. We don't want to fail here at ENOSPC, so
	 * allow ifree to dip into the reserved block pool if necessary.
	 *
	 * Freeing large sets of inodes generally means freeing inode chunks,
	 * directory and file data blocks, so this should be relatively safe.
	 * Only under severe circumstances should it be possible to free enough
	 * inodes to exhaust the reserve block pool via finobt expansion while
	 * at the same time not creating free space in the filesystem.
	 *
	 * Send a warning if the reservation does happen to fail, as the inode
	 * now remains allocated and sits on the unlinked list until the fs is
	 * repaired.
	 */
	tp->t_flags |= XFS_TRANS_RESERVE;
	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ifree,
				  XFS_IFREE_SPACE_RES(mp), 0);
1833
	if (error) {
1834 1835 1836 1837 1838 1839 1840
		if (error == ENOSPC) {
			xfs_warn_ratelimited(mp,
			"Failed to remove inode(s) from unlinked list. "
			"Please free space, unmount and run xfs_repair.");
		} else {
			ASSERT(XFS_FORCED_SHUTDOWN(mp));
		}
1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888
		xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES);
		return error;
	}

	xfs_ilock(ip, XFS_ILOCK_EXCL);
	xfs_trans_ijoin(tp, ip, 0);

	xfs_bmap_init(&free_list, &first_block);
	error = xfs_ifree(tp, ip, &free_list);
	if (error) {
		/*
		 * If we fail to free the inode, shut down.  The cancel
		 * might do that, we need to make sure.  Otherwise the
		 * inode might be lost for a long time or forever.
		 */
		if (!XFS_FORCED_SHUTDOWN(mp)) {
			xfs_notice(mp, "%s: xfs_ifree returned error %d",
				__func__, error);
			xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
		}
		xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
		xfs_iunlock(ip, XFS_ILOCK_EXCL);
		return error;
	}

	/*
	 * Credit the quota account(s). The inode is gone.
	 */
	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_ICOUNT, -1);

	/*
	 * Just ignore errors at this point.  There is nothing we can
	 * do except to try to keep going. Make sure it's not a silent
	 * error.
	 */
	error = xfs_bmap_finish(&tp,  &free_list, &committed);
	if (error)
		xfs_notice(mp, "%s: xfs_bmap_finish returned error %d",
			__func__, error);
	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
	if (error)
		xfs_notice(mp, "%s: xfs_trans_commit returned error %d",
			__func__, error);

	xfs_iunlock(ip, XFS_ILOCK_EXCL);
	return 0;
}

Dave Chinner's avatar
Dave Chinner committed
1889 1890 1891 1892 1893 1894 1895 1896
/*
 * xfs_inactive
 *
 * This is called when the vnode reference count for the vnode
 * goes to zero.  If the file has been unlinked, then it must
 * now be truncated.  Also, we clear all of the read-ahead state
 * kept for the inode here since the file is now closed.
 */
1897
void
Dave Chinner's avatar
Dave Chinner committed
1898 1899 1900
xfs_inactive(
	xfs_inode_t	*ip)
{
1901 1902 1903
	struct xfs_mount	*mp;
	int			error;
	int			truncate = 0;
Dave Chinner's avatar
Dave Chinner committed
1904 1905 1906 1907 1908

	/*
	 * If the inode is already free, then there can be nothing
	 * to clean up here.
	 */
1909
	if (ip->i_d.di_mode == 0) {
Dave Chinner's avatar
Dave Chinner committed
1910 1911
		ASSERT(ip->i_df.if_real_bytes == 0);
		ASSERT(ip->i_df.if_broot_bytes == 0);
1912
		return;
Dave Chinner's avatar
Dave Chinner committed
1913 1914 1915 1916 1917 1918
	}

	mp = ip->i_mount;

	/* If this is a read-only mount, don't do this (would generate I/O) */
	if (mp->m_flags & XFS_MOUNT_RDONLY)
1919
		return;
Dave Chinner's avatar
Dave Chinner committed
1920 1921 1922 1923 1924 1925 1926

	if (ip->i_d.di_nlink != 0) {
		/*
		 * force is true because we are evicting an inode from the
		 * cache. Post-eof blocks must be freed, lest we end up with
		 * broken free space accounting.
		 */
1927 1928 1929 1930
		if (xfs_can_free_eofblocks(ip, true))
			xfs_free_eofblocks(mp, ip, false);

		return;
Dave Chinner's avatar
Dave Chinner committed
1931 1932 1933 1934 1935 1936 1937 1938 1939
	}

	if (S_ISREG(ip->i_d.di_mode) &&
	    (ip->i_d.di_size != 0 || XFS_ISIZE(ip) != 0 ||
	     ip->i_d.di_nextents > 0 || ip->i_delayed_blks > 0))
		truncate = 1;

	error = xfs_qm_dqattach(ip, 0);
	if (error)
1940
		return;
Dave Chinner's avatar
Dave Chinner committed
1941

1942
	if (S_ISLNK(ip->i_d.di_mode))
1943
		error = xfs_inactive_symlink(ip);
1944 1945 1946
	else if (truncate)
		error = xfs_inactive_truncate(ip);
	if (error)
1947
		return;
Dave Chinner's avatar
Dave Chinner committed
1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959

	/*
	 * If there are attributes associated with the file then blow them away
	 * now.  The code calls a routine that recursively deconstructs the
	 * attribute fork.  We need to just commit the current transaction
	 * because we can't use it for xfs_attr_inactive().
	 */
	if (ip->i_d.di_anextents > 0) {
		ASSERT(ip->i_d.di_forkoff != 0);

		error = xfs_attr_inactive(ip);
		if (error)
1960
			return;
Dave Chinner's avatar
Dave Chinner committed
1961 1962 1963 1964 1965 1966 1967 1968 1969 1970
	}

	if (ip->i_afp)
		xfs_idestroy_fork(ip, XFS_ATTR_FORK);

	ASSERT(ip->i_d.di_anextents == 0);

	/*
	 * Free the inode.
	 */
1971 1972
	error = xfs_inactive_ifree(ip);
	if (error)
1973
		return;
Dave Chinner's avatar
Dave Chinner committed
1974 1975 1976 1977 1978 1979 1980

	/*
	 * Release the dquots held by inode, if any.
	 */
	xfs_qm_dqdetach(ip);
}

Linus Torvalds's avatar
Linus Torvalds committed
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
/*
 * This is called when the inode's link count goes to 0.
 * We place the on-disk inode on a list in the AGI.  It
 * will be pulled from this list when the inode is freed.
 */
int
xfs_iunlink(
	xfs_trans_t	*tp,
	xfs_inode_t	*ip)
{
	xfs_mount_t	*mp;
	xfs_agi_t	*agi;
	xfs_dinode_t	*dip;
	xfs_buf_t	*agibp;
	xfs_buf_t	*ibp;
	xfs_agino_t	agino;
	short		bucket_index;
	int		offset;
	int		error;

	ASSERT(ip->i_d.di_nlink == 0);
	ASSERT(ip->i_d.di_mode != 0);

	mp = tp->t_mountp;

	/*
	 * Get the agi buffer first.  It ensures lock ordering
	 * on the list.
	 */
2010
	error = xfs_read_agi(mp, tp, XFS_INO_TO_AGNO(mp, ip->i_ino), &agibp);
2011
	if (error)
Linus Torvalds's avatar
Linus Torvalds committed
2012 2013
		return error;
	agi = XFS_BUF_TO_AGI(agibp);
2014

Linus Torvalds's avatar
Linus Torvalds committed
2015 2016 2017 2018 2019 2020 2021 2022
	/*
	 * Get the index into the agi hash table for the
	 * list this inode will go on.
	 */
	agino = XFS_INO_TO_AGINO(mp, ip->i_ino);
	ASSERT(agino != 0);
	bucket_index = agino % XFS_AGI_UNLINKED_BUCKETS;
	ASSERT(agi->agi_unlinked[bucket_index]);
2023
	ASSERT(be32_to_cpu(agi->agi_unlinked[bucket_index]) != agino);
Linus Torvalds's avatar
Linus Torvalds committed
2024

2025
	if (agi->agi_unlinked[bucket_index] != cpu_to_be32(NULLAGINO)) {
Linus Torvalds's avatar
Linus Torvalds committed
2026 2027 2028 2029 2030 2031
		/*
		 * There is already another inode in the bucket we need
		 * to add ourselves to.  Add us at the front of the list.
		 * Here we put the head pointer into our next pointer,
		 * and then we fall through to point the head at us.
		 */
2032 2033
		error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &dip, &ibp,
				       0, 0);
2034 2035 2036
		if (error)
			return error;

2037
		ASSERT(dip->di_next_unlinked == cpu_to_be32(NULLAGINO));
Linus Torvalds's avatar
Linus Torvalds committed
2038
		dip->di_next_unlinked = agi->agi_unlinked[bucket_index];
2039
		offset = ip->i_imap.im_boffset +
Linus Torvalds's avatar
Linus Torvalds committed
2040
			offsetof(xfs_dinode_t, di_next_unlinked);
2041 2042 2043 2044

		/* need to recalc the inode CRC if appropriate */
		xfs_dinode_calc_crc(mp, dip);

Linus Torvalds's avatar
Linus Torvalds committed
2045 2046 2047 2048 2049 2050 2051 2052 2053 2054
		xfs_trans_inode_buf(tp, ibp);
		xfs_trans_log_buf(tp, ibp, offset,
				  (offset + sizeof(xfs_agino_t) - 1));
		xfs_inobp_check(mp, ibp);
	}

	/*
	 * Point the bucket head pointer at the inode being inserted.
	 */
	ASSERT(agino != 0);
2055
	agi->agi_unlinked[bucket_index] = cpu_to_be32(agino);
Linus Torvalds's avatar
Linus Torvalds committed
2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080
	offset = offsetof(xfs_agi_t, agi_unlinked) +
		(sizeof(xfs_agino_t) * bucket_index);
	xfs_trans_log_buf(tp, agibp, offset,
			  (offset + sizeof(xfs_agino_t) - 1));
	return 0;
}

/*
 * Pull the on-disk inode from the AGI unlinked list.
 */
STATIC int
xfs_iunlink_remove(
	xfs_trans_t	*tp,
	xfs_inode_t	*ip)
{
	xfs_ino_t	next_ino;
	xfs_mount_t	*mp;
	xfs_agi_t	*agi;
	xfs_dinode_t	*dip;
	xfs_buf_t	*agibp;
	xfs_buf_t	*ibp;
	xfs_agnumber_t	agno;
	xfs_agino_t	agino;
	xfs_agino_t	next_agino;
	xfs_buf_t	*last_ibp;
2081
	xfs_dinode_t	*last_dip = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
2082
	short		bucket_index;
2083
	int		offset, last_offset = 0;
Linus Torvalds's avatar
Linus Torvalds committed
2084 2085 2086 2087 2088 2089 2090 2091 2092
	int		error;

	mp = tp->t_mountp;
	agno = XFS_INO_TO_AGNO(mp, ip->i_ino);

	/*
	 * Get the agi buffer first.  It ensures lock ordering
	 * on the list.
	 */
2093 2094
	error = xfs_read_agi(mp, tp, agno, &agibp);
	if (error)
Linus Torvalds's avatar
Linus Torvalds committed
2095
		return error;
2096

Linus Torvalds's avatar
Linus Torvalds committed
2097
	agi = XFS_BUF_TO_AGI(agibp);
2098

Linus Torvalds's avatar
Linus Torvalds committed
2099 2100 2101 2102 2103 2104 2105
	/*
	 * Get the index into the agi hash table for the
	 * list this inode will go on.
	 */
	agino = XFS_INO_TO_AGINO(mp, ip->i_ino);
	ASSERT(agino != 0);
	bucket_index = agino % XFS_AGI_UNLINKED_BUCKETS;
2106
	ASSERT(agi->agi_unlinked[bucket_index] != cpu_to_be32(NULLAGINO));
Linus Torvalds's avatar
Linus Torvalds committed
2107 2108
	ASSERT(agi->agi_unlinked[bucket_index]);

2109
	if (be32_to_cpu(agi->agi_unlinked[bucket_index]) == agino) {
Linus Torvalds's avatar
Linus Torvalds committed
2110
		/*
2111 2112 2113 2114 2115
		 * We're at the head of the list.  Get the inode's on-disk
		 * buffer to see if there is anyone after us on the list.
		 * Only modify our next pointer if it is not already NULLAGINO.
		 * This saves us the overhead of dealing with the buffer when
		 * there is no need to change it.
Linus Torvalds's avatar
Linus Torvalds committed
2116
		 */
2117 2118
		error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &dip, &ibp,
				       0, 0);
Linus Torvalds's avatar
Linus Torvalds committed
2119
		if (error) {
2120
			xfs_warn(mp, "%s: xfs_imap_to_bp returned error %d.",
2121
				__func__, error);
Linus Torvalds's avatar
Linus Torvalds committed
2122 2123
			return error;
		}
2124
		next_agino = be32_to_cpu(dip->di_next_unlinked);
Linus Torvalds's avatar
Linus Torvalds committed
2125 2126
		ASSERT(next_agino != 0);
		if (next_agino != NULLAGINO) {
2127
			dip->di_next_unlinked = cpu_to_be32(NULLAGINO);
2128
			offset = ip->i_imap.im_boffset +
Linus Torvalds's avatar
Linus Torvalds committed
2129
				offsetof(xfs_dinode_t, di_next_unlinked);
2130 2131 2132 2133

			/* need to recalc the inode CRC if appropriate */
			xfs_dinode_calc_crc(mp, dip);

Linus Torvalds's avatar
Linus Torvalds committed
2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145
			xfs_trans_inode_buf(tp, ibp);
			xfs_trans_log_buf(tp, ibp, offset,
					  (offset + sizeof(xfs_agino_t) - 1));
			xfs_inobp_check(mp, ibp);
		} else {
			xfs_trans_brelse(tp, ibp);
		}
		/*
		 * Point the bucket head pointer at the next inode.
		 */
		ASSERT(next_agino != 0);
		ASSERT(next_agino != agino);
2146
		agi->agi_unlinked[bucket_index] = cpu_to_be32(next_agino);
Linus Torvalds's avatar
Linus Torvalds committed
2147 2148 2149 2150 2151 2152 2153 2154
		offset = offsetof(xfs_agi_t, agi_unlinked) +
			(sizeof(xfs_agino_t) * bucket_index);
		xfs_trans_log_buf(tp, agibp, offset,
				  (offset + sizeof(xfs_agino_t) - 1));
	} else {
		/*
		 * We need to search the list for the inode being freed.
		 */
2155
		next_agino = be32_to_cpu(agi->agi_unlinked[bucket_index]);
Linus Torvalds's avatar
Linus Torvalds committed
2156 2157
		last_ibp = NULL;
		while (next_agino != agino) {
Christoph Hellwig's avatar
Christoph Hellwig committed
2158 2159 2160
			struct xfs_imap	imap;

			if (last_ibp)
Linus Torvalds's avatar
Linus Torvalds committed
2161
				xfs_trans_brelse(tp, last_ibp);
Christoph Hellwig's avatar
Christoph Hellwig committed
2162 2163

			imap.im_blkno = 0;
Linus Torvalds's avatar
Linus Torvalds committed
2164
			next_ino = XFS_AGINO_TO_INO(mp, agno, next_agino);
Christoph Hellwig's avatar
Christoph Hellwig committed
2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175

			error = xfs_imap(mp, tp, next_ino, &imap, 0);
			if (error) {
				xfs_warn(mp,
	"%s: xfs_imap returned error %d.",
					 __func__, error);
				return error;
			}

			error = xfs_imap_to_bp(mp, tp, &imap, &last_dip,
					       &last_ibp, 0, 0);
Linus Torvalds's avatar
Linus Torvalds committed
2176
			if (error) {
2177
				xfs_warn(mp,
Christoph Hellwig's avatar
Christoph Hellwig committed
2178
	"%s: xfs_imap_to_bp returned error %d.",
2179
					__func__, error);
Linus Torvalds's avatar
Linus Torvalds committed
2180 2181
				return error;
			}
Christoph Hellwig's avatar
Christoph Hellwig committed
2182 2183

			last_offset = imap.im_boffset;
2184
			next_agino = be32_to_cpu(last_dip->di_next_unlinked);
Linus Torvalds's avatar
Linus Torvalds committed
2185 2186 2187
			ASSERT(next_agino != NULLAGINO);
			ASSERT(next_agino != 0);
		}
2188

Linus Torvalds's avatar
Linus Torvalds committed
2189
		/*
2190 2191
		 * Now last_ibp points to the buffer previous to us on the
		 * unlinked list.  Pull us from the list.
Linus Torvalds's avatar
Linus Torvalds committed
2192
		 */
2193 2194
		error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &dip, &ibp,
				       0, 0);
Linus Torvalds's avatar
Linus Torvalds committed
2195
		if (error) {
2196
			xfs_warn(mp, "%s: xfs_imap_to_bp(2) returned error %d.",
2197
				__func__, error);
Linus Torvalds's avatar
Linus Torvalds committed
2198 2199
			return error;
		}
2200
		next_agino = be32_to_cpu(dip->di_next_unlinked);
Linus Torvalds's avatar
Linus Torvalds committed
2201 2202 2203
		ASSERT(next_agino != 0);
		ASSERT(next_agino != agino);
		if (next_agino != NULLAGINO) {
2204
			dip->di_next_unlinked = cpu_to_be32(NULLAGINO);
2205
			offset = ip->i_imap.im_boffset +
Linus Torvalds's avatar
Linus Torvalds committed
2206
				offsetof(xfs_dinode_t, di_next_unlinked);
2207 2208 2209 2210

			/* need to recalc the inode CRC if appropriate */
			xfs_dinode_calc_crc(mp, dip);

Linus Torvalds's avatar
Linus Torvalds committed
2211 2212 2213 2214 2215 2216 2217 2218 2219 2220
			xfs_trans_inode_buf(tp, ibp);
			xfs_trans_log_buf(tp, ibp, offset,
					  (offset + sizeof(xfs_agino_t) - 1));
			xfs_inobp_check(mp, ibp);
		} else {
			xfs_trans_brelse(tp, ibp);
		}
		/*
		 * Point the previous inode on the list to the next inode.
		 */
2221
		last_dip->di_next_unlinked = cpu_to_be32(next_agino);
Linus Torvalds's avatar
Linus Torvalds committed
2222 2223
		ASSERT(next_agino != 0);
		offset = last_offset + offsetof(xfs_dinode_t, di_next_unlinked);
2224 2225 2226 2227

		/* need to recalc the inode CRC if appropriate */
		xfs_dinode_calc_crc(mp, last_dip);

Linus Torvalds's avatar
Linus Torvalds committed
2228 2229 2230 2231 2232 2233 2234 2235
		xfs_trans_inode_buf(tp, last_ibp);
		xfs_trans_log_buf(tp, last_ibp, offset,
				  (offset + sizeof(xfs_agino_t) - 1));
		xfs_inobp_check(mp, last_ibp);
	}
	return 0;
}

2236
/*
2237
 * A big issue when freeing the inode cluster is that we _cannot_ skip any
2238 2239 2240
 * inodes that are in memory - they all must be marked stale and attached to
 * the cluster buffer.
 */
2241
STATIC int
Linus Torvalds's avatar
Linus Torvalds committed
2242 2243 2244 2245 2246 2247 2248
xfs_ifree_cluster(
	xfs_inode_t	*free_ip,
	xfs_trans_t	*tp,
	xfs_ino_t	inum)
{
	xfs_mount_t		*mp = free_ip->i_mount;
	int			blks_per_cluster;
2249
	int			inodes_per_cluster;
Linus Torvalds's avatar
Linus Torvalds committed
2250
	int			nbufs;
2251
	int			i, j;
Linus Torvalds's avatar
Linus Torvalds committed
2252 2253
	xfs_daddr_t		blkno;
	xfs_buf_t		*bp;
2254
	xfs_inode_t		*ip;
Linus Torvalds's avatar
Linus Torvalds committed
2255 2256
	xfs_inode_log_item_t	*iip;
	xfs_log_item_t		*lip;
Dave Chinner's avatar
Dave Chinner committed
2257
	struct xfs_perag	*pag;
Linus Torvalds's avatar
Linus Torvalds committed
2258

Dave Chinner's avatar
Dave Chinner committed
2259
	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, inum));
2260 2261 2262
	blks_per_cluster = xfs_icluster_size_fsb(mp);
	inodes_per_cluster = blks_per_cluster << mp->m_sb.sb_inopblog;
	nbufs = mp->m_ialloc_blks / blks_per_cluster;
Linus Torvalds's avatar
Linus Torvalds committed
2263

2264
	for (j = 0; j < nbufs; j++, inum += inodes_per_cluster) {
Linus Torvalds's avatar
Linus Torvalds committed
2265 2266 2267
		blkno = XFS_AGB_TO_DADDR(mp, XFS_INO_TO_AGNO(mp, inum),
					 XFS_INO_TO_AGBNO(mp, inum));

2268 2269 2270 2271 2272 2273 2274 2275 2276
		/*
		 * We obtain and lock the backing buffer first in the process
		 * here, as we have to ensure that any dirty inode that we
		 * can't get the flush lock on is attached to the buffer.
		 * If we scan the in-memory inodes first, then buffer IO can
		 * complete before we get a lock on it, and hence we may fail
		 * to mark all the active inodes on the buffer stale.
		 */
		bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, blkno,
2277 2278
					mp->m_bsize * blks_per_cluster,
					XBF_UNMAPPED);
2279

2280 2281
		if (!bp)
			return ENOMEM;
2282 2283 2284 2285 2286 2287 2288 2289 2290 2291

		/*
		 * This buffer may not have been correctly initialised as we
		 * didn't read it from disk. That's not important because we are
		 * only using to mark the buffer as stale in the log, and to
		 * attach stale cached inodes on it. That means it will never be
		 * dispatched for IO. If it is, we want to know about it, and we
		 * want it to fail. We can acheive this by adding a write
		 * verifier to the buffer.
		 */
2292
		 bp->b_ops = &xfs_inode_buf_ops;
2293

2294 2295 2296
		/*
		 * Walk the inodes already attached to the buffer and mark them
		 * stale. These will all have the flush locks held, so an
2297 2298 2299
		 * in-memory inode walk can't lock them. By marking them all
		 * stale first, we will not attempt to lock them in the loop
		 * below as the XFS_ISTALE flag will be set.
2300
		 */
2301
		lip = bp->b_fspriv;
2302 2303 2304 2305
		while (lip) {
			if (lip->li_type == XFS_LI_INODE) {
				iip = (xfs_inode_log_item_t *)lip;
				ASSERT(iip->ili_logged == 1);
2306
				lip->li_cb = xfs_istale_done;
2307 2308 2309 2310 2311 2312 2313
				xfs_trans_ail_copy_lsn(mp->m_ail,
							&iip->ili_flush_lsn,
							&iip->ili_item.li_lsn);
				xfs_iflags_set(iip->ili_inode, XFS_ISTALE);
			}
			lip = lip->li_bio_list;
		}
Linus Torvalds's avatar
Linus Torvalds committed
2314

2315

Linus Torvalds's avatar
Linus Torvalds committed
2316
		/*
2317 2318 2319 2320
		 * For each inode in memory attempt to add it to the inode
		 * buffer and set it up for being staled on buffer IO
		 * completion.  This is safe as we've locked out tail pushing
		 * and flushing by locking the buffer.
Linus Torvalds's avatar
Linus Torvalds committed
2321
		 *
2322 2323 2324
		 * We have already marked every inode that was part of a
		 * transaction stale above, which means there is no point in
		 * even trying to lock them.
Linus Torvalds's avatar
Linus Torvalds committed
2325
		 */
2326
		for (i = 0; i < inodes_per_cluster; i++) {
2327
retry:
2328
			rcu_read_lock();
2329 2330
			ip = radix_tree_lookup(&pag->pag_ici_root,
					XFS_INO_TO_AGINO(mp, (inum + i)));
Linus Torvalds's avatar
Linus Torvalds committed
2331

2332 2333 2334
			/* Inode not in memory, nothing to do */
			if (!ip) {
				rcu_read_unlock();
Linus Torvalds's avatar
Linus Torvalds committed
2335 2336 2337
				continue;
			}

2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353
			/*
			 * because this is an RCU protected lookup, we could
			 * find a recently freed or even reallocated inode
			 * during the lookup. We need to check under the
			 * i_flags_lock for a valid inode here. Skip it if it
			 * is not valid, the wrong inode or stale.
			 */
			spin_lock(&ip->i_flags_lock);
			if (ip->i_ino != inum + i ||
			    __xfs_iflags_test(ip, XFS_ISTALE)) {
				spin_unlock(&ip->i_flags_lock);
				rcu_read_unlock();
				continue;
			}
			spin_unlock(&ip->i_flags_lock);

2354 2355 2356 2357 2358 2359 2360
			/*
			 * Don't try to lock/unlock the current inode, but we
			 * _cannot_ skip the other inodes that we did not find
			 * in the list attached to the buffer and are not
			 * already marked stale. If we can't lock it, back off
			 * and retry.
			 */
2361 2362
			if (ip != free_ip &&
			    !xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) {
2363
				rcu_read_unlock();
2364 2365
				delay(1);
				goto retry;
Linus Torvalds's avatar
Linus Torvalds committed
2366
			}
2367
			rcu_read_unlock();
Linus Torvalds's avatar
Linus Torvalds committed
2368

2369
			xfs_iflock(ip);
2370
			xfs_iflags_set(ip, XFS_ISTALE);
Linus Torvalds's avatar
Linus Torvalds committed
2371

2372 2373 2374 2375
			/*
			 * we don't need to attach clean inodes or those only
			 * with unlogged changes (which we throw away, anyway).
			 */
Linus Torvalds's avatar
Linus Torvalds committed
2376
			iip = ip->i_itemp;
2377
			if (!iip || xfs_inode_clean(ip)) {
2378
				ASSERT(ip != free_ip);
Linus Torvalds's avatar
Linus Torvalds committed
2379 2380 2381 2382 2383
				xfs_ifunlock(ip);
				xfs_iunlock(ip, XFS_ILOCK_EXCL);
				continue;
			}

2384 2385
			iip->ili_last_fields = iip->ili_fields;
			iip->ili_fields = 0;
Linus Torvalds's avatar
Linus Torvalds committed
2386
			iip->ili_logged = 1;
2387 2388
			xfs_trans_ail_copy_lsn(mp->m_ail, &iip->ili_flush_lsn,
						&iip->ili_item.li_lsn);
Linus Torvalds's avatar
Linus Torvalds committed
2389

2390 2391
			xfs_buf_attach_iodone(bp, xfs_istale_done,
						  &iip->ili_item);
2392 2393

			if (ip != free_ip)
Linus Torvalds's avatar
Linus Torvalds committed
2394 2395 2396
				xfs_iunlock(ip, XFS_ILOCK_EXCL);
		}

2397
		xfs_trans_stale_inode_buf(tp, bp);
Linus Torvalds's avatar
Linus Torvalds committed
2398 2399 2400
		xfs_trans_binval(tp, bp);
	}

Dave Chinner's avatar
Dave Chinner committed
2401
	xfs_perag_put(pag);
2402
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424
}

/*
 * This is called to return an inode to the inode free list.
 * The inode should already be truncated to 0 length and have
 * no pages associated with it.  This routine also assumes that
 * the inode is already a part of the transaction.
 *
 * The on-disk copy of the inode will have been added to the list
 * of unlinked inodes in the AGI. We need to remove the inode from
 * that list atomically with respect to freeing it here.
 */
int
xfs_ifree(
	xfs_trans_t	*tp,
	xfs_inode_t	*ip,
	xfs_bmap_free_t	*flist)
{
	int			error;
	int			delete;
	xfs_ino_t		first_ino;

Christoph Hellwig's avatar
Christoph Hellwig committed
2425
	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
Linus Torvalds's avatar
Linus Torvalds committed
2426 2427 2428
	ASSERT(ip->i_d.di_nlink == 0);
	ASSERT(ip->i_d.di_nextents == 0);
	ASSERT(ip->i_d.di_anextents == 0);
2429
	ASSERT(ip->i_d.di_size == 0 || !S_ISREG(ip->i_d.di_mode));
Linus Torvalds's avatar
Linus Torvalds committed
2430 2431 2432 2433 2434 2435
	ASSERT(ip->i_d.di_nblocks == 0);

	/*
	 * Pull the on-disk inode from the AGI unlinked list.
	 */
	error = xfs_iunlink_remove(tp, ip);
2436
	if (error)
Linus Torvalds's avatar
Linus Torvalds committed
2437 2438 2439
		return error;

	error = xfs_difree(tp, ip->i_ino, flist, &delete, &first_ino);
2440
	if (error)
Linus Torvalds's avatar
Linus Torvalds committed
2441
		return error;
2442

Linus Torvalds's avatar
Linus Torvalds committed
2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455
	ip->i_d.di_mode = 0;		/* mark incore inode as free */
	ip->i_d.di_flags = 0;
	ip->i_d.di_dmevmask = 0;
	ip->i_d.di_forkoff = 0;		/* mark the attr fork not in use */
	ip->i_d.di_format = XFS_DINODE_FMT_EXTENTS;
	ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
	/*
	 * Bump the generation count so no one will be confused
	 * by reincarnations of this inode.
	 */
	ip->i_d.di_gen++;
	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);

2456
	if (delete)
2457
		error = xfs_ifree_cluster(ip, tp, first_ino);
Linus Torvalds's avatar
Linus Torvalds committed
2458

2459
	return error;
Linus Torvalds's avatar
Linus Torvalds committed
2460 2461 2462
}

/*
2463 2464 2465
 * This is called to unpin an inode.  The caller must have the inode locked
 * in at least shared mode so that the buffer cannot be subsequently pinned
 * once someone is waiting for it to be unpinned.
Linus Torvalds's avatar
Linus Torvalds committed
2466
 */
2467
static void
2468
xfs_iunpin(
2469
	struct xfs_inode	*ip)
Linus Torvalds's avatar
Linus Torvalds committed
2470
{
Christoph Hellwig's avatar
Christoph Hellwig committed
2471
	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
Linus Torvalds's avatar
Linus Torvalds committed
2472

2473 2474
	trace_xfs_inode_unpin_nowait(ip, _RET_IP_);

2475
	/* Give the log a push to start the unpinning I/O */
2476
	xfs_log_force_lsn(ip->i_mount, ip->i_itemp->ili_last_lsn, 0);
2477

2478
}
Linus Torvalds's avatar
Linus Torvalds committed
2479

2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496
static void
__xfs_iunpin_wait(
	struct xfs_inode	*ip)
{
	wait_queue_head_t *wq = bit_waitqueue(&ip->i_flags, __XFS_IPINNED_BIT);
	DEFINE_WAIT_BIT(wait, &ip->i_flags, __XFS_IPINNED_BIT);

	xfs_iunpin(ip);

	do {
		prepare_to_wait(wq, &wait.wait, TASK_UNINTERRUPTIBLE);
		if (xfs_ipincount(ip))
			io_schedule();
	} while (xfs_ipincount(ip));
	finish_wait(wq, &wait.wait);
}

2497
void
2498
xfs_iunpin_wait(
2499
	struct xfs_inode	*ip)
2500
{
2501 2502
	if (xfs_ipincount(ip))
		__xfs_iunpin_wait(ip);
Linus Torvalds's avatar
Linus Torvalds committed
2503 2504
}

2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531
/*
 * Removing an inode from the namespace involves removing the directory entry
 * and dropping the link count on the inode. Removing the directory entry can
 * result in locking an AGF (directory blocks were freed) and removing a link
 * count can result in placing the inode on an unlinked list which results in
 * locking an AGI.
 *
 * The big problem here is that we have an ordering constraint on AGF and AGI
 * locking - inode allocation locks the AGI, then can allocate a new extent for
 * new inodes, locking the AGF after the AGI. Similarly, freeing the inode
 * removes the inode from the unlinked list, requiring that we lock the AGI
 * first, and then freeing the inode can result in an inode chunk being freed
 * and hence freeing disk space requiring that we lock an AGF.
 *
 * Hence the ordering that is imposed by other parts of the code is AGI before
 * AGF. This means we cannot remove the directory entry before we drop the inode
 * reference count and put it on the unlinked list as this results in a lock
 * order of AGF then AGI, and this can deadlock against inode allocation and
 * freeing. Therefore we must drop the link counts before we remove the
 * directory entry.
 *
 * This is still safe from a transactional point of view - it is not until we
 * get to xfs_bmap_finish() that we have the possibility of multiple
 * transactions in this operation. Hence as long as we remove the directory
 * entry and drop the link count in the first transaction of the remove
 * operation, there are no transactional constraints on the ordering here.
 */
Dave Chinner's avatar
Dave Chinner committed
2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581
int
xfs_remove(
	xfs_inode_t             *dp,
	struct xfs_name		*name,
	xfs_inode_t		*ip)
{
	xfs_mount_t		*mp = dp->i_mount;
	xfs_trans_t             *tp = NULL;
	int			is_dir = S_ISDIR(ip->i_d.di_mode);
	int                     error = 0;
	xfs_bmap_free_t         free_list;
	xfs_fsblock_t           first_block;
	int			cancel_flags;
	int			committed;
	int			link_zero;
	uint			resblks;
	uint			log_count;

	trace_xfs_remove(dp, name);

	if (XFS_FORCED_SHUTDOWN(mp))
		return XFS_ERROR(EIO);

	error = xfs_qm_dqattach(dp, 0);
	if (error)
		goto std_return;

	error = xfs_qm_dqattach(ip, 0);
	if (error)
		goto std_return;

	if (is_dir) {
		tp = xfs_trans_alloc(mp, XFS_TRANS_RMDIR);
		log_count = XFS_DEFAULT_LOG_COUNT;
	} else {
		tp = xfs_trans_alloc(mp, XFS_TRANS_REMOVE);
		log_count = XFS_REMOVE_LOG_COUNT;
	}
	cancel_flags = XFS_TRANS_RELEASE_LOG_RES;

	/*
	 * We try to get the real space reservation first,
	 * allowing for directory btree deletion(s) implying
	 * possible bmap insert(s).  If we can't get the space
	 * reservation then we use 0 instead, and avoid the bmap
	 * btree insert(s) in the directory code by, if the bmap
	 * insert tries to happen, instead trimming the LAST
	 * block from the directory.
	 */
	resblks = XFS_REMOVE_SPACE_RES(mp);
2582
	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_remove, resblks, 0);
Dave Chinner's avatar
Dave Chinner committed
2583 2584
	if (error == ENOSPC) {
		resblks = 0;
2585
		error = xfs_trans_reserve(tp, &M_RES(mp)->tr_remove, 0, 0);
Dave Chinner's avatar
Dave Chinner committed
2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600
	}
	if (error) {
		ASSERT(error != ENOSPC);
		cancel_flags = 0;
		goto out_trans_cancel;
	}

	xfs_lock_two_inodes(dp, ip, XFS_ILOCK_EXCL);

	xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);

	/*
	 * If we're removing a directory perform some additional validation.
	 */
2601
	cancel_flags |= XFS_TRANS_ABORT;
Dave Chinner's avatar
Dave Chinner committed
2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612
	if (is_dir) {
		ASSERT(ip->i_d.di_nlink >= 2);
		if (ip->i_d.di_nlink != 2) {
			error = XFS_ERROR(ENOTEMPTY);
			goto out_trans_cancel;
		}
		if (!xfs_dir_isempty(ip)) {
			error = XFS_ERROR(ENOTEMPTY);
			goto out_trans_cancel;
		}

2613
		/* Drop the link from ip's "..".  */
Dave Chinner's avatar
Dave Chinner committed
2614 2615
		error = xfs_droplink(tp, dp);
		if (error)
2616
			goto out_trans_cancel;
Dave Chinner's avatar
Dave Chinner committed
2617

2618
		/* Drop the "." link from ip to self.  */
Dave Chinner's avatar
Dave Chinner committed
2619 2620
		error = xfs_droplink(tp, ip);
		if (error)
2621
			goto out_trans_cancel;
Dave Chinner's avatar
Dave Chinner committed
2622 2623 2624 2625 2626 2627 2628 2629
	} else {
		/*
		 * When removing a non-directory we need to log the parent
		 * inode here.  For a directory this is done implicitly
		 * by the xfs_droplink call for the ".." entry.
		 */
		xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
	}
2630
	xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
Dave Chinner's avatar
Dave Chinner committed
2631

2632
	/* Drop the link from dp to ip. */
Dave Chinner's avatar
Dave Chinner committed
2633 2634
	error = xfs_droplink(tp, ip);
	if (error)
2635
		goto out_trans_cancel;
Dave Chinner's avatar
Dave Chinner committed
2636

2637
	/* Determine if this is the last link while the inode is locked */
Dave Chinner's avatar
Dave Chinner committed
2638 2639
	link_zero = (ip->i_d.di_nlink == 0);

2640 2641 2642 2643 2644 2645 2646 2647
	xfs_bmap_init(&free_list, &first_block);
	error = xfs_dir_removename(tp, dp, name, ip->i_ino,
					&first_block, &free_list, resblks);
	if (error) {
		ASSERT(error != ENOENT);
		goto out_bmap_cancel;
	}

Dave Chinner's avatar
Dave Chinner committed
2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663
	/*
	 * If this is a synchronous mount, make sure that the
	 * remove transaction goes to disk before returning to
	 * the user.
	 */
	if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC))
		xfs_trans_set_sync(tp);

	error = xfs_bmap_finish(&tp, &free_list, &committed);
	if (error)
		goto out_bmap_cancel;

	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
	if (error)
		goto std_return;

2664
	if (is_dir && xfs_inode_is_filestream(ip))
Dave Chinner's avatar
Dave Chinner committed
2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676
		xfs_filestream_deassociate(ip);

	return 0;

 out_bmap_cancel:
	xfs_bmap_cancel(&free_list);
 out_trans_cancel:
	xfs_trans_cancel(tp, cancel_flags);
 std_return:
	return error;
}

Dave Chinner's avatar
Dave Chinner committed
2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762
/*
 * Enter all inodes for a rename transaction into a sorted array.
 */
STATIC void
xfs_sort_for_rename(
	xfs_inode_t	*dp1,	/* in: old (source) directory inode */
	xfs_inode_t	*dp2,	/* in: new (target) directory inode */
	xfs_inode_t	*ip1,	/* in: inode of old entry */
	xfs_inode_t	*ip2,	/* in: inode of new entry, if it
				   already exists, NULL otherwise. */
	xfs_inode_t	**i_tab,/* out: array of inode returned, sorted */
	int		*num_inodes)  /* out: number of inodes in array */
{
	xfs_inode_t		*temp;
	int			i, j;

	/*
	 * i_tab contains a list of pointers to inodes.  We initialize
	 * the table here & we'll sort it.  We will then use it to
	 * order the acquisition of the inode locks.
	 *
	 * Note that the table may contain duplicates.  e.g., dp1 == dp2.
	 */
	i_tab[0] = dp1;
	i_tab[1] = dp2;
	i_tab[2] = ip1;
	if (ip2) {
		*num_inodes = 4;
		i_tab[3] = ip2;
	} else {
		*num_inodes = 3;
		i_tab[3] = NULL;
	}

	/*
	 * Sort the elements via bubble sort.  (Remember, there are at
	 * most 4 elements to sort, so this is adequate.)
	 */
	for (i = 0; i < *num_inodes; i++) {
		for (j = 1; j < *num_inodes; j++) {
			if (i_tab[j]->i_ino < i_tab[j-1]->i_ino) {
				temp = i_tab[j];
				i_tab[j] = i_tab[j-1];
				i_tab[j-1] = temp;
			}
		}
	}
}

/*
 * xfs_rename
 */
int
xfs_rename(
	xfs_inode_t	*src_dp,
	struct xfs_name	*src_name,
	xfs_inode_t	*src_ip,
	xfs_inode_t	*target_dp,
	struct xfs_name	*target_name,
	xfs_inode_t	*target_ip)
{
	xfs_trans_t	*tp = NULL;
	xfs_mount_t	*mp = src_dp->i_mount;
	int		new_parent;		/* moving to a new dir */
	int		src_is_directory;	/* src_name is a directory */
	int		error;
	xfs_bmap_free_t free_list;
	xfs_fsblock_t   first_block;
	int		cancel_flags;
	int		committed;
	xfs_inode_t	*inodes[4];
	int		spaceres;
	int		num_inodes;

	trace_xfs_rename(src_dp, target_dp, src_name, target_name);

	new_parent = (src_dp != target_dp);
	src_is_directory = S_ISDIR(src_ip->i_d.di_mode);

	xfs_sort_for_rename(src_dp, target_dp, src_ip, target_ip,
				inodes, &num_inodes);

	xfs_bmap_init(&free_list, &first_block);
	tp = xfs_trans_alloc(mp, XFS_TRANS_RENAME);
	cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
	spaceres = XFS_RENAME_SPACE_RES(mp, target_name->len);
2763
	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_rename, spaceres, 0);
Dave Chinner's avatar
Dave Chinner committed
2764 2765
	if (error == ENOSPC) {
		spaceres = 0;
2766
		error = xfs_trans_reserve(tp, &M_RES(mp)->tr_rename, 0, 0);
Dave Chinner's avatar
Dave Chinner committed
2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981
	}
	if (error) {
		xfs_trans_cancel(tp, 0);
		goto std_return;
	}

	/*
	 * Attach the dquots to the inodes
	 */
	error = xfs_qm_vop_rename_dqattach(inodes);
	if (error) {
		xfs_trans_cancel(tp, cancel_flags);
		goto std_return;
	}

	/*
	 * Lock all the participating inodes. Depending upon whether
	 * the target_name exists in the target directory, and
	 * whether the target directory is the same as the source
	 * directory, we can lock from 2 to 4 inodes.
	 */
	xfs_lock_inodes(inodes, num_inodes, XFS_ILOCK_EXCL);

	/*
	 * Join all the inodes to the transaction. From this point on,
	 * we can rely on either trans_commit or trans_cancel to unlock
	 * them.
	 */
	xfs_trans_ijoin(tp, src_dp, XFS_ILOCK_EXCL);
	if (new_parent)
		xfs_trans_ijoin(tp, target_dp, XFS_ILOCK_EXCL);
	xfs_trans_ijoin(tp, src_ip, XFS_ILOCK_EXCL);
	if (target_ip)
		xfs_trans_ijoin(tp, target_ip, XFS_ILOCK_EXCL);

	/*
	 * If we are using project inheritance, we only allow renames
	 * into our tree when the project IDs are the same; else the
	 * tree quota mechanism would be circumvented.
	 */
	if (unlikely((target_dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) &&
		     (xfs_get_projid(target_dp) != xfs_get_projid(src_ip)))) {
		error = XFS_ERROR(EXDEV);
		goto error_return;
	}

	/*
	 * Set up the target.
	 */
	if (target_ip == NULL) {
		/*
		 * If there's no space reservation, check the entry will
		 * fit before actually inserting it.
		 */
		error = xfs_dir_canenter(tp, target_dp, target_name, spaceres);
		if (error)
			goto error_return;
		/*
		 * If target does not exist and the rename crosses
		 * directories, adjust the target directory link count
		 * to account for the ".." reference from the new entry.
		 */
		error = xfs_dir_createname(tp, target_dp, target_name,
						src_ip->i_ino, &first_block,
						&free_list, spaceres);
		if (error == ENOSPC)
			goto error_return;
		if (error)
			goto abort_return;

		xfs_trans_ichgtime(tp, target_dp,
					XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);

		if (new_parent && src_is_directory) {
			error = xfs_bumplink(tp, target_dp);
			if (error)
				goto abort_return;
		}
	} else { /* target_ip != NULL */
		/*
		 * If target exists and it's a directory, check that both
		 * target and source are directories and that target can be
		 * destroyed, or that neither is a directory.
		 */
		if (S_ISDIR(target_ip->i_d.di_mode)) {
			/*
			 * Make sure target dir is empty.
			 */
			if (!(xfs_dir_isempty(target_ip)) ||
			    (target_ip->i_d.di_nlink > 2)) {
				error = XFS_ERROR(EEXIST);
				goto error_return;
			}
		}

		/*
		 * Link the source inode under the target name.
		 * If the source inode is a directory and we are moving
		 * it across directories, its ".." entry will be
		 * inconsistent until we replace that down below.
		 *
		 * In case there is already an entry with the same
		 * name at the destination directory, remove it first.
		 */
		error = xfs_dir_replace(tp, target_dp, target_name,
					src_ip->i_ino,
					&first_block, &free_list, spaceres);
		if (error)
			goto abort_return;

		xfs_trans_ichgtime(tp, target_dp,
					XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);

		/*
		 * Decrement the link count on the target since the target
		 * dir no longer points to it.
		 */
		error = xfs_droplink(tp, target_ip);
		if (error)
			goto abort_return;

		if (src_is_directory) {
			/*
			 * Drop the link from the old "." entry.
			 */
			error = xfs_droplink(tp, target_ip);
			if (error)
				goto abort_return;
		}
	} /* target_ip != NULL */

	/*
	 * Remove the source.
	 */
	if (new_parent && src_is_directory) {
		/*
		 * Rewrite the ".." entry to point to the new
		 * directory.
		 */
		error = xfs_dir_replace(tp, src_ip, &xfs_name_dotdot,
					target_dp->i_ino,
					&first_block, &free_list, spaceres);
		ASSERT(error != EEXIST);
		if (error)
			goto abort_return;
	}

	/*
	 * We always want to hit the ctime on the source inode.
	 *
	 * This isn't strictly required by the standards since the source
	 * inode isn't really being changed, but old unix file systems did
	 * it and some incremental backup programs won't work without it.
	 */
	xfs_trans_ichgtime(tp, src_ip, XFS_ICHGTIME_CHG);
	xfs_trans_log_inode(tp, src_ip, XFS_ILOG_CORE);

	/*
	 * Adjust the link count on src_dp.  This is necessary when
	 * renaming a directory, either within one parent when
	 * the target existed, or across two parent directories.
	 */
	if (src_is_directory && (new_parent || target_ip != NULL)) {

		/*
		 * Decrement link count on src_directory since the
		 * entry that's moved no longer points to it.
		 */
		error = xfs_droplink(tp, src_dp);
		if (error)
			goto abort_return;
	}

	error = xfs_dir_removename(tp, src_dp, src_name, src_ip->i_ino,
					&first_block, &free_list, spaceres);
	if (error)
		goto abort_return;

	xfs_trans_ichgtime(tp, src_dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
	xfs_trans_log_inode(tp, src_dp, XFS_ILOG_CORE);
	if (new_parent)
		xfs_trans_log_inode(tp, target_dp, XFS_ILOG_CORE);

	/*
	 * If this is a synchronous mount, make sure that the
	 * rename transaction goes to disk before returning to
	 * the user.
	 */
	if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) {
		xfs_trans_set_sync(tp);
	}

	error = xfs_bmap_finish(&tp, &free_list, &committed);
	if (error) {
		xfs_bmap_cancel(&free_list);
		xfs_trans_cancel(tp, (XFS_TRANS_RELEASE_LOG_RES |
				 XFS_TRANS_ABORT));
		goto std_return;
	}

	/*
	 * trans_commit will unlock src_ip, target_ip & decrement
	 * the vnode references.
	 */
	return xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);

 abort_return:
	cancel_flags |= XFS_TRANS_ABORT;
 error_return:
	xfs_bmap_cancel(&free_list);
	xfs_trans_cancel(tp, cancel_flags);
 std_return:
	return error;
}

2982 2983 2984 2985
STATIC int
xfs_iflush_cluster(
	xfs_inode_t	*ip,
	xfs_buf_t	*bp)
Linus Torvalds's avatar
Linus Torvalds committed
2986
{
2987 2988 2989 2990 2991 2992 2993 2994 2995 2996
	xfs_mount_t		*mp = ip->i_mount;
	struct xfs_perag	*pag;
	unsigned long		first_index, mask;
	unsigned long		inodes_per_cluster;
	int			ilist_size;
	xfs_inode_t		**ilist;
	xfs_inode_t		*iq;
	int			nr_found;
	int			clcount = 0;
	int			bufwasdelwri;
Linus Torvalds's avatar
Linus Torvalds committed
2997 2998
	int			i;

2999
	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino));
Linus Torvalds's avatar
Linus Torvalds committed
3000

3001
	inodes_per_cluster = mp->m_inode_cluster_size >> mp->m_sb.sb_inodelog;
3002 3003 3004 3005
	ilist_size = inodes_per_cluster * sizeof(xfs_inode_t *);
	ilist = kmem_alloc(ilist_size, KM_MAYFAIL|KM_NOFS);
	if (!ilist)
		goto out_put;
Linus Torvalds's avatar
Linus Torvalds committed
3006

3007
	mask = ~(((mp->m_inode_cluster_size >> mp->m_sb.sb_inodelog)) - 1);
3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018
	first_index = XFS_INO_TO_AGINO(mp, ip->i_ino) & mask;
	rcu_read_lock();
	/* really need a gang lookup range call here */
	nr_found = radix_tree_gang_lookup(&pag->pag_ici_root, (void**)ilist,
					first_index, inodes_per_cluster);
	if (nr_found == 0)
		goto out_free;

	for (i = 0; i < nr_found; i++) {
		iq = ilist[i];
		if (iq == ip)
3019
			continue;
3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034

		/*
		 * because this is an RCU protected lookup, we could find a
		 * recently freed or even reallocated inode during the lookup.
		 * We need to check under the i_flags_lock for a valid inode
		 * here. Skip it if it is not valid or the wrong inode.
		 */
		spin_lock(&ip->i_flags_lock);
		if (!ip->i_ino ||
		    (XFS_INO_TO_AGINO(mp, iq->i_ino) & mask) != first_index) {
			spin_unlock(&ip->i_flags_lock);
			continue;
		}
		spin_unlock(&ip->i_flags_lock);

3035 3036 3037 3038 3039
		/*
		 * Do an un-protected check to see if the inode is dirty and
		 * is a candidate for flushing.  These checks will be repeated
		 * later after the appropriate locks are acquired.
		 */
3040
		if (xfs_inode_clean(iq) && xfs_ipincount(iq) == 0)
3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063
			continue;

		/*
		 * Try to get locks.  If any are unavailable or it is pinned,
		 * then this inode cannot be flushed and is skipped.
		 */

		if (!xfs_ilock_nowait(iq, XFS_ILOCK_SHARED))
			continue;
		if (!xfs_iflock_nowait(iq)) {
			xfs_iunlock(iq, XFS_ILOCK_SHARED);
			continue;
		}
		if (xfs_ipincount(iq)) {
			xfs_ifunlock(iq);
			xfs_iunlock(iq, XFS_ILOCK_SHARED);
			continue;
		}

		/*
		 * arriving here means that this inode can be flushed.  First
		 * re-check that it's dirty before flushing.
		 */
3064 3065
		if (!xfs_inode_clean(iq)) {
			int	error;
3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083
			error = xfs_iflush_int(iq, bp);
			if (error) {
				xfs_iunlock(iq, XFS_ILOCK_SHARED);
				goto cluster_corrupt_out;
			}
			clcount++;
		} else {
			xfs_ifunlock(iq);
		}
		xfs_iunlock(iq, XFS_ILOCK_SHARED);
	}

	if (clcount) {
		XFS_STATS_INC(xs_icluster_flushcnt);
		XFS_STATS_ADD(xs_icluster_flushinode, clcount);
	}

out_free:
3084
	rcu_read_unlock();
3085
	kmem_free(ilist);
3086 3087
out_put:
	xfs_perag_put(pag);
3088 3089 3090 3091 3092 3093 3094 3095
	return 0;


cluster_corrupt_out:
	/*
	 * Corruption detected in the clustering loop.  Invalidate the
	 * inode buffer and shut down the filesystem.
	 */
3096
	rcu_read_unlock();
3097
	/*
3098
	 * Clean up the buffer.  If it was delwri, just release it --
3099 3100 3101
	 * brelse can handle it with no problems.  If not, shut down the
	 * filesystem before releasing the buffer.
	 */
3102
	bufwasdelwri = (bp->b_flags & _XBF_DELWRI_Q);
3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113
	if (bufwasdelwri)
		xfs_buf_relse(bp);

	xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);

	if (!bufwasdelwri) {
		/*
		 * Just like incore_relse: if we have b_iodone functions,
		 * mark the buffer as an error and call them.  Otherwise
		 * mark it as stale and brelse.
		 */
3114
		if (bp->b_iodone) {
3115
			XFS_BUF_UNDONE(bp);
3116
			xfs_buf_stale(bp);
3117
			xfs_buf_ioerror(bp, EIO);
3118
			xfs_buf_ioend(bp, 0);
3119
		} else {
3120
			xfs_buf_stale(bp);
3121 3122 3123 3124 3125 3126 3127
			xfs_buf_relse(bp);
		}
	}

	/*
	 * Unlocks the flush lock
	 */
3128
	xfs_iflush_abort(iq, false);
3129
	kmem_free(ilist);
3130
	xfs_perag_put(pag);
3131 3132 3133
	return XFS_ERROR(EFSCORRUPTED);
}

Linus Torvalds's avatar
Linus Torvalds committed
3134
/*
3135 3136 3137 3138 3139 3140 3141
 * Flush dirty inode metadata into the backing buffer.
 *
 * The caller must have the inode lock and the inode flush lock held.  The
 * inode lock will still be held upon return to the caller, and the inode
 * flush lock will be released after the inode has reached the disk.
 *
 * The caller must write out the buffer returned in *bpp and release it.
Linus Torvalds's avatar
Linus Torvalds committed
3142 3143 3144
 */
int
xfs_iflush(
3145 3146
	struct xfs_inode	*ip,
	struct xfs_buf		**bpp)
Linus Torvalds's avatar
Linus Torvalds committed
3147
{
3148 3149 3150
	struct xfs_mount	*mp = ip->i_mount;
	struct xfs_buf		*bp;
	struct xfs_dinode	*dip;
Linus Torvalds's avatar
Linus Torvalds committed
3151 3152 3153 3154
	int			error;

	XFS_STATS_INC(xs_iflush_count);

Christoph Hellwig's avatar
Christoph Hellwig committed
3155
	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
3156
	ASSERT(xfs_isiflocked(ip));
Linus Torvalds's avatar
Linus Torvalds committed
3157
	ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE ||
3158
	       ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK));
Linus Torvalds's avatar
Linus Torvalds committed
3159

3160
	*bpp = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
3161 3162 3163

	xfs_iunpin_wait(ip);

3164 3165 3166
	/*
	 * For stale inodes we cannot rely on the backing buffer remaining
	 * stale in cache for the remaining life of the stale inode and so
3167
	 * xfs_imap_to_bp() below may give us a buffer that no longer contains
3168 3169 3170 3171 3172 3173 3174 3175 3176
	 * inodes below. We have to check this after ensuring the inode is
	 * unpinned so that it is safe to reclaim the stale inode after the
	 * flush call.
	 */
	if (xfs_iflags_test(ip, XFS_ISTALE)) {
		xfs_ifunlock(ip);
		return 0;
	}

Linus Torvalds's avatar
Linus Torvalds committed
3177 3178 3179
	/*
	 * This may have been unpinned because the filesystem is shutting
	 * down forcibly. If that's the case we must not write this inode
3180 3181 3182 3183
	 * to disk, because the log record didn't make it to disk.
	 *
	 * We also have to remove the log item from the AIL in this case,
	 * as we wait for an empty AIL as part of the unmount process.
Linus Torvalds's avatar
Linus Torvalds committed
3184 3185
	 */
	if (XFS_FORCED_SHUTDOWN(mp)) {
3186 3187
		error = XFS_ERROR(EIO);
		goto abort_out;
Linus Torvalds's avatar
Linus Torvalds committed
3188 3189
	}

3190 3191 3192
	/*
	 * Get the buffer containing the on-disk inode.
	 */
3193 3194
	error = xfs_imap_to_bp(mp, NULL, &ip->i_imap, &dip, &bp, XBF_TRYLOCK,
			       0);
3195 3196 3197 3198 3199
	if (error || !bp) {
		xfs_ifunlock(ip);
		return error;
	}

Linus Torvalds's avatar
Linus Torvalds committed
3200 3201 3202 3203
	/*
	 * First flush out the inode that xfs_iflush was called with.
	 */
	error = xfs_iflush_int(ip, bp);
3204
	if (error)
Linus Torvalds's avatar
Linus Torvalds committed
3205 3206
		goto corrupt_out;

3207 3208 3209 3210
	/*
	 * If the buffer is pinned then push on the log now so we won't
	 * get stuck waiting in the write for too long.
	 */
3211
	if (xfs_buf_ispinned(bp))
3212
		xfs_log_force(mp, 0);
3213

Linus Torvalds's avatar
Linus Torvalds committed
3214 3215 3216 3217
	/*
	 * inode clustering:
	 * see if other inodes can be gathered into this write
	 */
3218 3219 3220
	error = xfs_iflush_cluster(ip, bp);
	if (error)
		goto cluster_corrupt_out;
Linus Torvalds's avatar
Linus Torvalds committed
3221

3222 3223
	*bpp = bp;
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
3224 3225 3226

corrupt_out:
	xfs_buf_relse(bp);
3227
	xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
Linus Torvalds's avatar
Linus Torvalds committed
3228
cluster_corrupt_out:
3229 3230
	error = XFS_ERROR(EFSCORRUPTED);
abort_out:
Linus Torvalds's avatar
Linus Torvalds committed
3231 3232 3233
	/*
	 * Unlocks the flush lock
	 */
3234
	xfs_iflush_abort(ip, false);
3235
	return error;
Linus Torvalds's avatar
Linus Torvalds committed
3236 3237 3238 3239
}

STATIC int
xfs_iflush_int(
3240 3241
	struct xfs_inode	*ip,
	struct xfs_buf		*bp)
Linus Torvalds's avatar
Linus Torvalds committed
3242
{
3243 3244 3245
	struct xfs_inode_log_item *iip = ip->i_itemp;
	struct xfs_dinode	*dip;
	struct xfs_mount	*mp = ip->i_mount;
Linus Torvalds's avatar
Linus Torvalds committed
3246

Christoph Hellwig's avatar
Christoph Hellwig committed
3247
	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
3248
	ASSERT(xfs_isiflocked(ip));
Linus Torvalds's avatar
Linus Torvalds committed
3249
	ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE ||
3250
	       ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK));
3251
	ASSERT(iip != NULL && iip->ili_fields != 0);
Linus Torvalds's avatar
Linus Torvalds committed
3252 3253

	/* set *dip = inode's place in the buffer */
3254
	dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset);
Linus Torvalds's avatar
Linus Torvalds committed
3255

3256
	if (XFS_TEST_ERROR(dip->di_magic != cpu_to_be16(XFS_DINODE_MAGIC),
Linus Torvalds's avatar
Linus Torvalds committed
3257
			       mp, XFS_ERRTAG_IFLUSH_1, XFS_RANDOM_IFLUSH_1)) {
3258 3259 3260
		xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
			"%s: Bad inode %Lu magic number 0x%x, ptr 0x%p",
			__func__, ip->i_ino, be16_to_cpu(dip->di_magic), dip);
Linus Torvalds's avatar
Linus Torvalds committed
3261 3262 3263 3264
		goto corrupt_out;
	}
	if (XFS_TEST_ERROR(ip->i_d.di_magic != XFS_DINODE_MAGIC,
				mp, XFS_ERRTAG_IFLUSH_2, XFS_RANDOM_IFLUSH_2)) {
3265 3266 3267
		xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
			"%s: Bad inode %Lu, ptr 0x%p, magic number 0x%x",
			__func__, ip->i_ino, ip, ip->i_d.di_magic);
Linus Torvalds's avatar
Linus Torvalds committed
3268 3269
		goto corrupt_out;
	}
3270
	if (S_ISREG(ip->i_d.di_mode)) {
Linus Torvalds's avatar
Linus Torvalds committed
3271 3272 3273 3274
		if (XFS_TEST_ERROR(
		    (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS) &&
		    (ip->i_d.di_format != XFS_DINODE_FMT_BTREE),
		    mp, XFS_ERRTAG_IFLUSH_3, XFS_RANDOM_IFLUSH_3)) {
3275 3276 3277
			xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
				"%s: Bad regular inode %Lu, ptr 0x%p",
				__func__, ip->i_ino, ip);
Linus Torvalds's avatar
Linus Torvalds committed
3278 3279
			goto corrupt_out;
		}
3280
	} else if (S_ISDIR(ip->i_d.di_mode)) {
Linus Torvalds's avatar
Linus Torvalds committed
3281 3282 3283 3284 3285
		if (XFS_TEST_ERROR(
		    (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS) &&
		    (ip->i_d.di_format != XFS_DINODE_FMT_BTREE) &&
		    (ip->i_d.di_format != XFS_DINODE_FMT_LOCAL),
		    mp, XFS_ERRTAG_IFLUSH_4, XFS_RANDOM_IFLUSH_4)) {
3286 3287 3288
			xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
				"%s: Bad directory inode %Lu, ptr 0x%p",
				__func__, ip->i_ino, ip);
Linus Torvalds's avatar
Linus Torvalds committed
3289 3290 3291 3292 3293 3294
			goto corrupt_out;
		}
	}
	if (XFS_TEST_ERROR(ip->i_d.di_nextents + ip->i_d.di_anextents >
				ip->i_d.di_nblocks, mp, XFS_ERRTAG_IFLUSH_5,
				XFS_RANDOM_IFLUSH_5)) {
3295 3296 3297 3298
		xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
			"%s: detected corrupt incore inode %Lu, "
			"total extents = %d, nblocks = %Ld, ptr 0x%p",
			__func__, ip->i_ino,
Linus Torvalds's avatar
Linus Torvalds committed
3299
			ip->i_d.di_nextents + ip->i_d.di_anextents,
3300
			ip->i_d.di_nblocks, ip);
Linus Torvalds's avatar
Linus Torvalds committed
3301 3302 3303 3304
		goto corrupt_out;
	}
	if (XFS_TEST_ERROR(ip->i_d.di_forkoff > mp->m_sb.sb_inodesize,
				mp, XFS_ERRTAG_IFLUSH_6, XFS_RANDOM_IFLUSH_6)) {
3305 3306 3307
		xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
			"%s: bad inode %Lu, forkoff 0x%x, ptr 0x%p",
			__func__, ip->i_ino, ip->i_d.di_forkoff, ip);
Linus Torvalds's avatar
Linus Torvalds committed
3308 3309
		goto corrupt_out;
	}
3310

Linus Torvalds's avatar
Linus Torvalds committed
3311
	/*
3312 3313 3314 3315 3316 3317 3318
	 * Inode item log recovery for v1/v2 inodes are dependent on the
	 * di_flushiter count for correct sequencing. We bump the flush
	 * iteration count so we can detect flushes which postdate a log record
	 * during recovery. This is redundant as we now log every change and
	 * hence this can't happen but we need to still do it to ensure
	 * backwards compatibility with old kernels that predate logging all
	 * inode changes.
Linus Torvalds's avatar
Linus Torvalds committed
3319
	 */
3320 3321
	if (ip->i_d.di_version < 3)
		ip->i_d.di_flushiter++;
Linus Torvalds's avatar
Linus Torvalds committed
3322 3323 3324 3325 3326 3327 3328

	/*
	 * Copy the dirty parts of the inode into the on-disk
	 * inode.  We always copy out the core of the inode,
	 * because if the inode is dirty at all the core must
	 * be.
	 */
3329
	xfs_dinode_to_disk(dip, &ip->i_d);
Linus Torvalds's avatar
Linus Torvalds committed
3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340

	/* Wrap, we never let the log put out DI_MAX_FLUSH */
	if (ip->i_d.di_flushiter == DI_MAX_FLUSH)
		ip->i_d.di_flushiter = 0;

	/*
	 * If this is really an old format inode and the superblock version
	 * has not been updated to support only new format inodes, then
	 * convert back to the old inode format.  If the superblock version
	 * has been updated, then make the conversion permanent.
	 */
3341 3342
	ASSERT(ip->i_d.di_version == 1 || xfs_sb_version_hasnlink(&mp->m_sb));
	if (ip->i_d.di_version == 1) {
3343
		if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
Linus Torvalds's avatar
Linus Torvalds committed
3344 3345 3346 3347
			/*
			 * Convert it back.
			 */
			ASSERT(ip->i_d.di_nlink <= XFS_MAXLINK_1);
3348
			dip->di_onlink = cpu_to_be16(ip->i_d.di_nlink);
Linus Torvalds's avatar
Linus Torvalds committed
3349 3350 3351 3352 3353 3354
		} else {
			/*
			 * The superblock version has already been bumped,
			 * so just make the conversion to the new inode
			 * format permanent.
			 */
3355 3356
			ip->i_d.di_version = 2;
			dip->di_version = 2;
Linus Torvalds's avatar
Linus Torvalds committed
3357
			ip->i_d.di_onlink = 0;
3358
			dip->di_onlink = 0;
Linus Torvalds's avatar
Linus Torvalds committed
3359
			memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
3360 3361
			memset(&(dip->di_pad[0]), 0,
			      sizeof(dip->di_pad));
3362
			ASSERT(xfs_get_projid(ip) == 0);
Linus Torvalds's avatar
Linus Torvalds committed
3363 3364 3365
		}
	}

3366
	xfs_iflush_fork(ip, dip, iip, XFS_DATA_FORK);
3367
	if (XFS_IFORK_Q(ip))
3368
		xfs_iflush_fork(ip, dip, iip, XFS_ATTR_FORK);
Linus Torvalds's avatar
Linus Torvalds committed
3369 3370 3371
	xfs_inobp_check(mp, bp);

	/*
3372 3373 3374 3375 3376 3377 3378
	 * We've recorded everything logged in the inode, so we'd like to clear
	 * the ili_fields bits so we don't log and flush things unnecessarily.
	 * However, we can't stop logging all this information until the data
	 * we've copied into the disk buffer is written to disk.  If we did we
	 * might overwrite the copy of the inode in the log with all the data
	 * after re-logging only part of it, and in the face of a crash we
	 * wouldn't have all the data we need to recover.
Linus Torvalds's avatar
Linus Torvalds committed
3379
	 *
3380 3381 3382 3383 3384 3385
	 * What we do is move the bits to the ili_last_fields field.  When
	 * logging the inode, these bits are moved back to the ili_fields field.
	 * In the xfs_iflush_done() routine we clear ili_last_fields, since we
	 * know that the information those bits represent is permanently on
	 * disk.  As long as the flush completes before the inode is logged
	 * again, then both ili_fields and ili_last_fields will be cleared.
Linus Torvalds's avatar
Linus Torvalds committed
3386
	 *
3387 3388 3389 3390 3391 3392 3393 3394
	 * We can play with the ili_fields bits here, because the inode lock
	 * must be held exclusively in order to set bits there and the flush
	 * lock protects the ili_last_fields bits.  Set ili_logged so the flush
	 * done routine can tell whether or not to look in the AIL.  Also, store
	 * the current LSN of the inode so that we can tell whether the item has
	 * moved in the AIL from xfs_iflush_done().  In order to read the lsn we
	 * need the AIL lock, because it is a 64 bit value that cannot be read
	 * atomically.
Linus Torvalds's avatar
Linus Torvalds committed
3395
	 */
3396 3397 3398
	iip->ili_last_fields = iip->ili_fields;
	iip->ili_fields = 0;
	iip->ili_logged = 1;
Linus Torvalds's avatar
Linus Torvalds committed
3399

3400 3401
	xfs_trans_ail_copy_lsn(mp->m_ail, &iip->ili_flush_lsn,
				&iip->ili_item.li_lsn);
Linus Torvalds's avatar
Linus Torvalds committed
3402

3403 3404 3405 3406 3407 3408 3409
	/*
	 * Attach the function xfs_iflush_done to the inode's
	 * buffer.  This will remove the inode from the AIL
	 * and unlock the inode's flush lock when the inode is
	 * completely written to disk.
	 */
	xfs_buf_attach_iodone(bp, xfs_iflush_done, &iip->ili_item);
Linus Torvalds's avatar
Linus Torvalds committed
3410

3411 3412 3413 3414 3415 3416
	/* update the lsn in the on disk inode if required */
	if (ip->i_d.di_version == 3)
		dip->di_lsn = cpu_to_be64(iip->ili_item.li_lsn);

	/* generate the checksum. */
	xfs_dinode_calc_crc(mp, dip);
Linus Torvalds's avatar
Linus Torvalds committed
3417

3418 3419
	ASSERT(bp->b_fspriv != NULL);
	ASSERT(bp->b_iodone != NULL);
Linus Torvalds's avatar
Linus Torvalds committed
3420 3421 3422 3423 3424
	return 0;

corrupt_out:
	return XFS_ERROR(EFSCORRUPTED);
}