xfs_attr.c 43.0 KB
Newer Older
D
Dave Chinner 已提交
1
// SPDX-License-Identifier: GPL-2.0
L
Linus Torvalds 已提交
2
/*
3 4
 * Copyright (c) 2000-2005 Silicon Graphics, Inc.
 * All Rights Reserved.
L
Linus Torvalds 已提交
5 6
 */
#include "xfs.h"
7
#include "xfs_fs.h"
8
#include "xfs_shared.h"
9 10 11
#include "xfs_format.h"
#include "xfs_log_format.h"
#include "xfs_trans_resv.h"
L
Linus Torvalds 已提交
12
#include "xfs_mount.h"
13
#include "xfs_defer.h"
14
#include "xfs_da_format.h"
15 16
#include "xfs_da_btree.h"
#include "xfs_attr_sf.h"
L
Linus Torvalds 已提交
17
#include "xfs_inode.h"
18
#include "xfs_trans.h"
L
Linus Torvalds 已提交
19
#include "xfs_bmap.h"
20
#include "xfs_bmap_btree.h"
L
Linus Torvalds 已提交
21 22
#include "xfs_attr.h"
#include "xfs_attr_leaf.h"
23
#include "xfs_attr_remote.h"
L
Linus Torvalds 已提交
24 25
#include "xfs_quota.h"
#include "xfs_trans_space.h"
C
Christoph Hellwig 已提交
26
#include "xfs_trace.h"
27
#include "xfs_attr_item.h"
28
#include "xfs_log.h"
29 30 31

struct kmem_cache		*xfs_attri_cache;
struct kmem_cache		*xfs_attrd_cache;
L
Linus Torvalds 已提交
32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

/*
 * xfs_attr.c
 *
 * Provide the external interfaces to manage attribute lists.
 */

/*========================================================================
 * Function prototypes for the kernel.
 *========================================================================*/

/*
 * Internal routines when attribute list fits inside the inode.
 */
STATIC int xfs_attr_shortform_addname(xfs_da_args_t *args);

/*
 * Internal routines when attribute list is one block.
 */
51
STATIC int xfs_attr_leaf_get(xfs_da_args_t *args);
L
Linus Torvalds 已提交
52
STATIC int xfs_attr_leaf_removename(xfs_da_args_t *args);
53
STATIC int xfs_attr_leaf_hasname(struct xfs_da_args *args, struct xfs_buf **bp);
54
STATIC int xfs_attr_leaf_try_add(struct xfs_da_args *args, struct xfs_buf *bp);
L
Linus Torvalds 已提交
55 56 57 58

/*
 * Internal routines when attribute list is more than one block.
 */
59
STATIC int xfs_attr_node_get(xfs_da_args_t *args);
60
STATIC void xfs_attr_restore_rmt_blk(struct xfs_da_args *args);
61
static int xfs_attr_node_try_addname(struct xfs_attr_item *attr);
62
STATIC int xfs_attr_node_addname_find_attr(struct xfs_attr_item *attr);
63
STATIC int xfs_attr_node_remove_attr(struct xfs_attr_item *attr);
64 65
STATIC int xfs_attr_node_hasname(xfs_da_args_t *args,
				 struct xfs_da_state **state);
L
Linus Torvalds 已提交
66 67
STATIC int xfs_attr_fillstate(xfs_da_state_t *state);
STATIC int xfs_attr_refillstate(xfs_da_state_t *state);
68 69
STATIC int xfs_attr_node_removename(struct xfs_da_args *args,
				    struct xfs_da_state *state);
L
Linus Torvalds 已提交
70

71
int
72 73 74 75
xfs_inode_hasattr(
	struct xfs_inode	*ip)
{
	if (!XFS_IFORK_Q(ip) ||
76
	    (ip->i_afp->if_format == XFS_DINODE_FMT_EXTENTS &&
77
	     ip->i_afp->if_nextents == 0))
78 79 80 81
		return 0;
	return 1;
}

82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101
/*
 * Returns true if the there is exactly only block in the attr fork, in which
 * case the attribute fork consists of a single leaf block entry.
 */
bool
xfs_attr_is_leaf(
	struct xfs_inode	*ip)
{
	struct xfs_ifork	*ifp = ip->i_afp;
	struct xfs_iext_cursor	icur;
	struct xfs_bmbt_irec	imap;

	if (ifp->if_nextents != 1 || ifp->if_format != XFS_DINODE_FMT_EXTENTS)
		return false;

	xfs_iext_first(ifp, &icur);
	xfs_iext_get_extent(ifp, &icur, &imap);
	return imap.br_startoff == 0 && imap.br_blockcount == 1;
}

L
Linus Torvalds 已提交
102 103 104 105
/*========================================================================
 * Overall external interface routines.
 *========================================================================*/

106 107 108 109
/*
 * Retrieve an extended attribute and its value.  Must have ilock.
 * Returns 0 on successful retrieval, otherwise an error.
 */
110 111 112 113
int
xfs_attr_get_ilocked(
	struct xfs_da_args	*args)
{
114
	ASSERT(xfs_isilocked(args->dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
115

116
	if (!xfs_inode_hasattr(args->dp))
117
		return -ENOATTR;
118

119
	if (args->dp->i_afp->if_format == XFS_DINODE_FMT_LOCAL)
120
		return xfs_attr_shortform_getvalue(args);
121
	if (xfs_attr_is_leaf(args->dp))
122
		return xfs_attr_leaf_get(args);
123
	return xfs_attr_node_get(args);
124 125
}

126 127 128
/*
 * Retrieve an extended attribute by name, and its value if requested.
 *
C
Christoph Hellwig 已提交
129 130 131
 * If args->valuelen is zero, then the caller does not want the value, just an
 * indication whether the attribute exists and the size of the value if it
 * exists. The size is returned in args.valuelen.
132
 *
133 134 135 136 137
 * If args->value is NULL but args->valuelen is non-zero, allocate the buffer
 * for the value after existence of the attribute has been determined. The
 * caller always has to free args->value if it is set, no matter if this
 * function was successful or not.
 *
138
 * If the attribute is found, but exceeds the size limit set by the caller in
139 140
 * args->valuelen, return -ERANGE with the size of the attribute that was found
 * in args->valuelen.
141
 */
142 143
int
xfs_attr_get(
144
	struct xfs_da_args	*args)
L
Linus Torvalds 已提交
145
{
146 147 148
	uint			lock_mode;
	int			error;

149
	XFS_STATS_INC(args->dp->i_mount, xs_attr_get);
150

151
	if (xfs_is_shutdown(args->dp->i_mount))
D
Dave Chinner 已提交
152
		return -EIO;
L
Linus Torvalds 已提交
153

154 155 156
	args->geo = args->dp->i_mount->m_attr_geo;
	args->whichfork = XFS_ATTR_FORK;
	args->hashval = xfs_da_hashname(args->name, args->namelen);
157

158
	/* Entirely possible to look up a name which doesn't exist */
159
	args->op_flags = XFS_DA_OP_OKNOENT;
L
Linus Torvalds 已提交
160

161
	lock_mode = xfs_ilock_attr_map_shared(args->dp);
162
	error = xfs_attr_get_ilocked(args);
163
	xfs_iunlock(args->dp, lock_mode);
164

165
	return error;
L
Linus Torvalds 已提交
166 167
}

168 169 170
/*
 * Calculate how many blocks we need for the new attribute,
 */
171
int
172
xfs_attr_calc_size(
173
	struct xfs_da_args	*args,
174 175
	int			*local)
{
176
	struct xfs_mount	*mp = args->dp->i_mount;
177 178 179 180 181 182 183
	int			size;
	int			nblks;

	/*
	 * Determine space new attribute will use, and if it would be
	 * "local" or "remote" (note: local != inline).
	 */
184
	size = xfs_attr_leaf_newentsize(args, local);
185 186
	nblks = XFS_DAENTER_SPACE_RES(mp, XFS_ATTR_FORK);
	if (*local) {
187
		if (size > (args->geo->blksize / 2)) {
188 189 190 191 192 193 194 195
			/* Double split possible */
			nblks *= 2;
		}
	} else {
		/*
		 * Out of line attribute, cannot double split, but
		 * make room for the attribute value itself.
		 */
196
		uint	dblocks = xfs_attr3_rmt_blocks(mp, args->valuelen);
197 198 199 200 201 202 203
		nblks += dblocks;
		nblks += XFS_NEXTENTADD_SPACE_RES(mp, dblocks, XFS_ATTR_FORK);
	}

	return nblks;
}

204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225
/* Initialize transaction reservation for attr operations */
void
xfs_init_attr_trans(
	struct xfs_da_args	*args,
	struct xfs_trans_res	*tres,
	unsigned int		*total)
{
	struct xfs_mount	*mp = args->dp->i_mount;

	if (args->value) {
		tres->tr_logres = M_RES(mp)->tr_attrsetm.tr_logres +
				 M_RES(mp)->tr_attrsetrt.tr_logres *
				 args->total;
		tres->tr_logcount = XFS_ATTRSET_LOG_COUNT;
		tres->tr_logflags = XFS_TRANS_PERM_LOG_RES;
		*total = args->total;
	} else {
		*tres = M_RES(mp)->tr_attrrm;
		*total = XFS_ATTRRM_SPACE_RES(mp);
	}
}

226 227 228 229 230
/*
 * Add an attr to a shortform fork. If there is no space,
 * xfs_attr_shortform_addname() will convert to leaf format and return -ENOSPC.
 * to use.
 */
231 232 233 234 235 236
STATIC int
xfs_attr_try_sf_addname(
	struct xfs_inode	*dp,
	struct xfs_da_args	*args)
{

237 238 239 240 241 242 243
	int			error;

	/*
	 * Build initial attribute list (if required).
	 */
	if (dp->i_afp->if_format == XFS_DINODE_FMT_EXTENTS)
		xfs_attr_shortform_create(args);
244 245 246 247 248 249 250 251 252

	error = xfs_attr_shortform_addname(args);
	if (error == -ENOSPC)
		return error;

	/*
	 * Commit the shortform mods, and we're done.
	 * NOTE: this is also the error path (EEXIST, etc).
	 */
253
	if (!error && !(args->op_flags & XFS_DA_OP_NOTIME))
254 255
		xfs_trans_ichgtime(args->trans, dp, XFS_ICHGTIME_CHG);

256
	if (xfs_has_wsync(dp->i_mount))
257 258
		xfs_trans_set_sync(args->trans);

259
	return error;
260 261
}

262
static int
263
xfs_attr_sf_addname(
264
	struct xfs_attr_item		*attr)
265
{
266
	struct xfs_da_args		*args = attr->xattri_da_args;
267 268
	struct xfs_inode		*dp = args->dp;
	int				error = 0;
269 270

	error = xfs_attr_try_sf_addname(dp, args);
271 272 273 274 275
	if (error != -ENOSPC) {
		ASSERT(!error || error == -EEXIST);
		attr->xattri_dela_state = XFS_DAS_DONE;
		goto out;
	}
276

277 278 279 280
	/*
	 * It won't fit in the shortform, transform to a leaf block.  GROT:
	 * another possible req'mt for a double-split btree op.
	 */
281
	error = xfs_attr_shortform_to_leaf(args, &attr->xattri_leaf_bp);
282 283 284 285 286 287
	if (error)
		return error;

	/*
	 * Prevent the leaf buffer from being unlocked so that a concurrent AIL
	 * push cannot grab the half-baked leaf buffer and run into problems
288
	 * with the write verifier.
289
	 */
290
	xfs_trans_bhold(args->trans, attr->xattri_leaf_bp);
291 292 293 294 295
	attr->xattri_dela_state = XFS_DAS_LEAF_ADD;
	error = -EAGAIN;
out:
	trace_xfs_attr_sf_addname_return(attr->xattri_dela_state, args->dp);
	return error;
296 297
}

298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317
/*
 * When we bump the state to REPLACE, we may actually need to skip over the
 * state. When LARP mode is enabled, we don't need to run the atomic flags flip,
 * so we skip straight over the REPLACE state and go on to REMOVE_OLD.
 */
static void
xfs_attr_dela_state_set_replace(
	struct xfs_attr_item	*attr,
	enum xfs_delattr_state	replace)
{
	struct xfs_da_args	*args = attr->xattri_da_args;

	ASSERT(replace == XFS_DAS_LEAF_REPLACE ||
			replace == XFS_DAS_NODE_REPLACE);

	attr->xattri_dela_state = replace;
	if (xfs_has_larp(args->dp->i_mount))
		attr->xattri_dela_state++;
}

318
static int
319 320 321 322 323 324
xfs_attr_leaf_addname(
	struct xfs_attr_item	*attr)
{
	struct xfs_da_args	*args = attr->xattri_da_args;
	int			error;

325
	ASSERT(xfs_attr_is_leaf(args->dp));
326

327 328 329 330 331 332 333
	/*
	 * Use the leaf buffer we may already hold locked as a result of
	 * a sf-to-leaf conversion. The held buffer is no longer valid
	 * after this call, regardless of the result.
	 */
	error = xfs_attr_leaf_try_add(args, attr->xattri_leaf_bp);
	attr->xattri_leaf_bp = NULL;
334

335 336
	if (error == -ENOSPC) {
		error = xfs_attr3_leaf_to_node(args);
337 338 339
		if (error)
			return error;

340 341 342 343 344 345 346
		/*
		 * We're not in leaf format anymore, so roll the transaction and
		 * retry the add to the newly allocated node block.
		 */
		attr->xattri_dela_state = XFS_DAS_NODE_ADD;
		error = -EAGAIN;
		goto out;
347 348 349
	}
	if (error)
		return error;
350

351 352 353 354 355
	/*
	 * We need to commit and roll if we need to allocate remote xattr blocks
	 * or perform more xattr manipulations. Otherwise there is nothing more
	 * to do and we can return success.
	 */
356 357 358 359
	if (args->rmtblkno) {
		attr->xattri_dela_state = XFS_DAS_LEAF_SET_RMT;
		error = -EAGAIN;
	} else if (args->op_flags & XFS_DA_OP_RENAME) {
360
		xfs_attr_dela_state_set_replace(attr, XFS_DAS_LEAF_REPLACE);
361
		error = -EAGAIN;
362 363
	} else {
		attr->xattri_dela_state = XFS_DAS_DONE;
364
	}
365
out:
366
	trace_xfs_attr_leaf_addname_return(attr->xattri_dela_state, args->dp);
367
	return error;
368 369
}

370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386
static int
xfs_attr_node_addname(
	struct xfs_attr_item	*attr)
{
	struct xfs_da_args	*args = attr->xattri_da_args;
	int			error;

	ASSERT(!attr->xattri_leaf_bp);

	error = xfs_attr_node_addname_find_attr(attr);
	if (error)
		return error;

	error = xfs_attr_node_try_addname(attr);
	if (error)
		return error;

387 388 389 390
	if (args->rmtblkno) {
		attr->xattri_dela_state = XFS_DAS_NODE_SET_RMT;
		error = -EAGAIN;
	} else if (args->op_flags & XFS_DA_OP_RENAME) {
391
		xfs_attr_dela_state_set_replace(attr, XFS_DAS_NODE_REPLACE);
392 393 394 395 396 397 398 399 400
		error = -EAGAIN;
	} else {
		attr->xattri_dela_state = XFS_DAS_DONE;
	}

	trace_xfs_attr_node_addname_return(attr->xattri_dela_state, args->dp);
	return error;
}

401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417
static int
xfs_attr_rmtval_alloc(
	struct xfs_attr_item		*attr)
{
	struct xfs_da_args              *args = attr->xattri_da_args;
	int				error = 0;

	/*
	 * If there was an out-of-line value, allocate the blocks we
	 * identified for its storage and copy the value.  This is done
	 * after we create the attribute so that we don't overflow the
	 * maximum size of a transaction and/or hit a deadlock.
	 */
	if (attr->xattri_blkcnt > 0) {
		error = xfs_attr_rmtval_set_blk(attr);
		if (error)
			return error;
418 419 420 421 422
		/* Roll the transaction only if there is more to allocate. */
		if (attr->xattri_blkcnt > 0) {
			error = -EAGAIN;
			goto out;
		}
423 424 425 426 427 428 429 430 431 432
	}

	error = xfs_attr_rmtval_set_value(args);
	if (error)
		return error;

	/* If this is not a rename, clear the incomplete flag and we're done. */
	if (!(args->op_flags & XFS_DA_OP_RENAME)) {
		error = xfs_attr3_leaf_clearflag(args);
		attr->xattri_dela_state = XFS_DAS_DONE;
433 434 435 436 437 438 439
	} else {
		/*
		 * We are running a REPLACE operation, so we need to bump the
		 * state to the step in that operation.
		 */
		attr->xattri_dela_state++;
		xfs_attr_dela_state_set_replace(attr, attr->xattri_dela_state);
440 441 442 443 444
	}
out:
	trace_xfs_attr_rmtval_alloc(attr->xattri_dela_state, args->dp);
	return error;
}
445

446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516
/*
 * Remove the original attr we have just replaced. This is dependent on the
 * original lookup and insert placing the old attr in args->blkno/args->index
 * and the new attr in args->blkno2/args->index2.
 */
static int
xfs_attr_leaf_remove_attr(
	struct xfs_attr_item		*attr)
{
	struct xfs_da_args              *args = attr->xattri_da_args;
	struct xfs_inode		*dp = args->dp;
	struct xfs_buf			*bp = NULL;
	int				forkoff;
	int				error;

	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno,
				   &bp);
	if (error)
		return error;

	xfs_attr3_leaf_remove(bp, args);

	forkoff = xfs_attr_shortform_allfit(bp, dp);
	if (forkoff)
		error = xfs_attr3_leaf_to_shortform(bp, args, forkoff);
		/* bp is gone due to xfs_da_shrink_inode */

	return error;
}

/*
 * Shrink an attribute from leaf to shortform. Used by the node format remove
 * path when the node format collapses to a single block and so we have to check
 * if it can be collapsed further.
 */
static int
xfs_attr_leaf_shrink(
	struct xfs_da_args	*args,
	struct xfs_da_state     *state)
{
	struct xfs_inode	*dp = args->dp;
	int			error, forkoff;
	struct xfs_buf		*bp;

	if (!xfs_attr_is_leaf(dp))
		return 0;

	/*
	 * Have to get rid of the copy of this dabuf in the state.
	 */
	if (state) {
		ASSERT(state->path.active == 1);
		ASSERT(state->path.blk[0].bp);
		state->path.blk[0].bp = NULL;
	}

	error = xfs_attr3_leaf_read(args->trans, args->dp, 0, &bp);
	if (error)
		return error;

	forkoff = xfs_attr_shortform_allfit(bp, dp);
	if (forkoff) {
		error = xfs_attr3_leaf_to_shortform(bp, args, forkoff);
		/* bp is gone due to xfs_da_shrink_inode */
	} else {
		xfs_trans_brelse(args->trans, bp);
	}

	return error;
}

517 518
/*
 * Set the attribute specified in @args.
519 520 521 522
 * This routine is meant to function as a delayed operation, and may return
 * -EAGAIN when the transaction needs to be rolled.  Calling functions will need
 * to handle this, and recall the function until a successful error code is
 * returned.
523 524
 */
int
525
xfs_attr_set_iter(
526
	struct xfs_attr_item		*attr)
527
{
528
	struct xfs_da_args              *args = attr->xattri_da_args;
529
	int				error = 0;
530 531

	/* State machine switch */
532
next_state:
533
	switch (attr->xattri_dela_state) {
534
	case XFS_DAS_UNINIT:
535 536 537 538 539
		ASSERT(0);
		return -EFSCORRUPTED;
	case XFS_DAS_SF_ADD:
		return xfs_attr_sf_addname(attr);
	case XFS_DAS_LEAF_ADD:
540
		return xfs_attr_leaf_addname(attr);
541 542
	case XFS_DAS_NODE_ADD:
		return xfs_attr_node_addname(attr);
543

544 545 546 547 548
	case XFS_DAS_LEAF_SET_RMT:
	case XFS_DAS_NODE_SET_RMT:
		error = xfs_attr_rmtval_find_space(attr);
		if (error)
			return error;
549
		attr->xattri_dela_state++;
550
		fallthrough;
551

552
	case XFS_DAS_LEAF_ALLOC_RMT:
553
	case XFS_DAS_NODE_ALLOC_RMT:
554 555
		error = xfs_attr_rmtval_alloc(attr);
		if (error)
556
			return error;
557 558
		if (attr->xattri_dela_state == XFS_DAS_DONE)
			break;
559
		goto next_state;
560

561 562
	case XFS_DAS_LEAF_REPLACE:
	case XFS_DAS_NODE_REPLACE:
563
		/*
564 565
		 * We must "flip" the incomplete flags on the "new" and "old"
		 * attribute/value pairs so that one disappears and one appears
566
		 * atomically.
567
		 */
568 569 570 571
		error = xfs_attr3_leaf_flipflags(args);
		if (error)
			return error;
		/*
572 573
		 * We must commit the flag value change now to make it atomic
		 * and then we can start the next trans in series at REMOVE_OLD.
574 575
		 */
		error = -EAGAIN;
576
		attr->xattri_dela_state++;
577 578 579 580
		break;

	case XFS_DAS_LEAF_REMOVE_OLD:
	case XFS_DAS_NODE_REMOVE_OLD:
581
		/*
582 583 584 585 586
		 * If we have a remote attr, start the process of removing it
		 * by invalidating any cached buffers.
		 *
		 * If we don't have a remote attr, we skip the remote block
		 * removal state altogether with a second state increment.
587 588 589
		 */
		xfs_attr_restore_rmt_blk(args);
		if (args->rmtblkno) {
590
			error = xfs_attr_rmtval_invalidate(args);
591 592
			if (error)
				return error;
593 594
		} else {
			attr->xattri_dela_state++;
595 596
		}

597 598 599 600 601 602 603 604 605 606 607
		attr->xattri_dela_state++;
		goto next_state;

	case XFS_DAS_LEAF_REMOVE_RMT:
	case XFS_DAS_NODE_REMOVE_RMT:
		error = xfs_attr_rmtval_remove(attr);
		if (error == -EAGAIN)
			break;
		if (error)
			return error;

608
		/*
609 610 611 612 613 614
		 * We've finished removing the remote attr blocks, so commit the
		 * transaction and move on to removing the attr name from the
		 * leaf/node block. Removing the attr might require a full
		 * transaction reservation for btree block freeing, so we
		 * can't do that in the same transaction where we removed the
		 * remote attr blocks.
615
		 */
616
		error = -EAGAIN;
617
		attr->xattri_dela_state++;
618
		break;
619

620 621 622 623
	case XFS_DAS_LEAF_REMOVE_ATTR:
		error = xfs_attr_leaf_remove_attr(attr);
		attr->xattri_dela_state = XFS_DAS_DONE;
		break;
624

625 626 627 628 629
	case XFS_DAS_NODE_REMOVE_ATTR:
		error = xfs_attr_node_remove_attr(attr);
		if (!error)
			error = xfs_attr_leaf_shrink(args, NULL);
		attr->xattri_dela_state = XFS_DAS_DONE;
630 631
		break;
	default:
632
		ASSERT(0);
633 634
		break;
	}
635 636

	trace_xfs_attr_set_iter_return(attr->xattri_dela_state, args->dp);
637
	return error;
638 639
}

640

641 642 643
/*
 * Return EEXIST if attr is found, or ENOATTR if not
 */
D
Dave Chinner 已提交
644 645
static int
xfs_attr_lookup(
646 647 648 649 650 651 652 653 654
	struct xfs_da_args	*args)
{
	struct xfs_inode	*dp = args->dp;
	struct xfs_buf		*bp = NULL;
	int			error;

	if (!xfs_inode_hasattr(dp))
		return -ENOATTR;

C
Christoph Hellwig 已提交
655
	if (dp->i_afp->if_format == XFS_DINODE_FMT_LOCAL)
656 657
		return xfs_attr_sf_findname(args, NULL, NULL);

658
	if (xfs_attr_is_leaf(dp)) {
659 660 661 662 663 664 665 666 667 668 669
		error = xfs_attr_leaf_hasname(args, &bp);

		if (bp)
			xfs_trans_brelse(args->trans, bp);

		return error;
	}

	return xfs_attr_node_hasname(args, NULL);
}

670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698
static int
xfs_attr_item_init(
	struct xfs_da_args	*args,
	unsigned int		op_flags,	/* op flag (set or remove) */
	struct xfs_attr_item	**attr)		/* new xfs_attr_item */
{

	struct xfs_attr_item	*new;

	new = kmem_zalloc(sizeof(struct xfs_attr_item), KM_NOFS);
	new->xattri_op_flags = op_flags;
	new->xattri_da_args = args;

	*attr = new;
	return 0;
}

/* Sets an attribute for an inode as a deferred operation */
static int
xfs_attr_defer_add(
	struct xfs_da_args	*args)
{
	struct xfs_attr_item	*new;
	int			error = 0;

	error = xfs_attr_item_init(args, XFS_ATTR_OP_FLAGS_SET, &new);
	if (error)
		return error;

699
	new->xattri_dela_state = xfs_attr_init_add_state(args);
700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717
	xfs_defer_add(args->trans, XFS_DEFER_OPS_TYPE_ATTR, &new->xattri_list);
	trace_xfs_attr_defer_add(new->xattri_dela_state, args->dp);

	return 0;
}

/* Sets an attribute for an inode as a deferred operation */
static int
xfs_attr_defer_replace(
	struct xfs_da_args	*args)
{
	struct xfs_attr_item	*new;
	int			error = 0;

	error = xfs_attr_item_init(args, XFS_ATTR_OP_FLAGS_REPLACE, &new);
	if (error)
		return error;

718
	new->xattri_dela_state = xfs_attr_init_replace_state(args);
719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744
	xfs_defer_add(args->trans, XFS_DEFER_OPS_TYPE_ATTR, &new->xattri_list);
	trace_xfs_attr_defer_replace(new->xattri_dela_state, args->dp);

	return 0;
}

/* Removes an attribute for an inode as a deferred operation */
static int
xfs_attr_defer_remove(
	struct xfs_da_args	*args)
{

	struct xfs_attr_item	*new;
	int			error;

	error  = xfs_attr_item_init(args, XFS_ATTR_OP_FLAGS_REMOVE, &new);
	if (error)
		return error;

	new->xattri_dela_state = XFS_DAS_UNINIT;
	xfs_defer_add(args->trans, XFS_DEFER_OPS_TYPE_ATTR, &new->xattri_list);
	trace_xfs_attr_defer_remove(new->xattri_dela_state, args->dp);

	return 0;
}

745
/*
746
 * Note: If args->value is NULL the attribute will be removed, just like the
747 748
 * Linux ->setattr API.
 */
749 750
int
xfs_attr_set(
751
	struct xfs_da_args	*args)
L
Linus Torvalds 已提交
752
{
753
	struct xfs_inode	*dp = args->dp;
754 755
	struct xfs_mount	*mp = dp->i_mount;
	struct xfs_trans_res	tres;
756
	bool			rsvd = (args->attr_filter & XFS_ATTR_ROOT);
757
	int			error, local;
758
	int			rmt_blks = 0;
759
	unsigned int		total;
760
	int			delayed = xfs_has_larp(mp);
761

762
	if (xfs_is_shutdown(dp->i_mount))
D
Dave Chinner 已提交
763
		return -EIO;
764

765 766 767 768
	error = xfs_qm_dqattach(dp);
	if (error)
		return error;

769 770 771
	args->geo = mp->m_attr_geo;
	args->whichfork = XFS_ATTR_FORK;
	args->hashval = xfs_da_hashname(args->name, args->namelen);
L
Linus Torvalds 已提交
772 773

	/*
774 775 776
	 * We have no control over the attribute names that userspace passes us
	 * to remove, so we have to allow the name lookup prior to attribute
	 * removal to fail as well.
L
Linus Torvalds 已提交
777
	 */
778
	args->op_flags = XFS_DA_OP_OKNOENT;
779

780
	if (args->value) {
781 782
		XFS_STATS_INC(mp, xs_attr_set);

783 784
		args->op_flags |= XFS_DA_OP_ADDNAME;
		args->total = xfs_attr_calc_size(args, &local);
785 786 787 788 789 790 791

		/*
		 * If the inode doesn't have an attribute fork, add one.
		 * (inode must not be locked when we call this routine)
		 */
		if (XFS_IFORK_Q(dp) == 0) {
			int sf_size = sizeof(struct xfs_attr_sf_hdr) +
792
				xfs_attr_sf_entsize_byname(args->namelen,
793
						args->valuelen);
794 795 796 797 798 799

			error = xfs_bmap_add_attrfork(dp, sf_size, rsvd);
			if (error)
				return error;
		}

800 801
		if (!local)
			rmt_blks = xfs_attr3_rmt_blocks(mp, args->valuelen);
802 803
	} else {
		XFS_STATS_INC(mp, xs_attr_remove);
804
		rmt_blks = xfs_attr3_rmt_blocks(mp, XFS_XATTR_SIZE_MAX);
805
	}
L
Linus Torvalds 已提交
806

807 808 809 810 811 812
	if (delayed) {
		error = xfs_attr_use_log_assist(mp);
		if (error)
			return error;
	}

L
Linus Torvalds 已提交
813 814 815 816
	/*
	 * Root fork attributes can use reserved data blocks for this
	 * operation if necessary
	 */
817
	xfs_init_attr_trans(args, &tres, &total);
818
	error = xfs_trans_alloc_inode(dp, &tres, total, 0, rsvd, &args->trans);
819
	if (error)
820
		goto drop_incompat;
L
Linus Torvalds 已提交
821

822 823 824
	if (args->value || xfs_inode_hasattr(dp)) {
		error = xfs_iext_count_may_overflow(dp, XFS_ATTR_FORK,
				XFS_IEXT_ATTR_MANIP_CNT(rmt_blks));
825 826 827
		if (error == -EFBIG)
			error = xfs_iext_count_upgrade(args->trans, dp,
					XFS_IEXT_ATTR_MANIP_CNT(rmt_blks));
828 829 830 831
		if (error)
			goto out_trans_cancel;
	}

D
Dave Chinner 已提交
832
	error = xfs_attr_lookup(args);
833 834 835 836 837 838 839 840 841
	switch (error) {
	case -EEXIST:
		/* if no value, we are performing a remove operation */
		if (!args->value) {
			error = xfs_attr_defer_remove(args);
			break;
		}
		/* Pure create fails if the attr already exists */
		if (args->attr_flags & XATTR_CREATE)
842 843
			goto out_trans_cancel;

844 845 846 847 848
		error = xfs_attr_defer_replace(args);
		break;
	case -ENOATTR:
		/* Can't remove what isn't there. */
		if (!args->value)
849
			goto out_trans_cancel;
850

851 852
		/* Pure replace fails if no existing attr to replace. */
		if (args->attr_flags & XATTR_REPLACE)
853 854
			goto out_trans_cancel;

855 856 857 858
		error = xfs_attr_defer_add(args);
		break;
	default:
		goto out_trans_cancel;
859
	}
860 861
	if (error)
		goto out_trans_cancel;
L
Linus Torvalds 已提交
862 863 864 865 866

	/*
	 * If this is a synchronous mount, make sure that the
	 * transaction goes to disk before returning to the user.
	 */
867
	if (xfs_has_wsync(mp))
868
		xfs_trans_set_sync(args->trans);
L
Linus Torvalds 已提交
869

870
	if (!(args->op_flags & XFS_DA_OP_NOTIME))
871
		xfs_trans_ichgtime(args->trans, dp, XFS_ICHGTIME_CHG);
872

L
Linus Torvalds 已提交
873 874 875
	/*
	 * Commit the last in the sequence of transactions.
	 */
876 877
	xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE);
	error = xfs_trans_commit(args->trans);
878
out_unlock:
L
Linus Torvalds 已提交
879
	xfs_iunlock(dp, XFS_ILOCK_EXCL);
880 881 882
drop_incompat:
	if (delayed)
		xlog_drop_incompat_feat(mp->m_log);
883
	return error;
L
Linus Torvalds 已提交
884

885
out_trans_cancel:
886 887
	if (args->trans)
		xfs_trans_cancel(args->trans);
888
	goto out_unlock;
L
Linus Torvalds 已提交
889 890
}

891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924
int __init
xfs_attri_init_cache(void)
{
	xfs_attri_cache = kmem_cache_create("xfs_attri",
					    sizeof(struct xfs_attri_log_item),
					    0, 0, NULL);

	return xfs_attri_cache != NULL ? 0 : -ENOMEM;
}

void
xfs_attri_destroy_cache(void)
{
	kmem_cache_destroy(xfs_attri_cache);
	xfs_attri_cache = NULL;
}

int __init
xfs_attrd_init_cache(void)
{
	xfs_attrd_cache = kmem_cache_create("xfs_attrd",
					    sizeof(struct xfs_attrd_log_item),
					    0, 0, NULL);

	return xfs_attrd_cache != NULL ? 0 : -ENOMEM;
}

void
xfs_attrd_destroy_cache(void)
{
	kmem_cache_destroy(xfs_attrd_cache);
	xfs_attrd_cache = NULL;
}

L
Linus Torvalds 已提交
925 926 927 928
/*========================================================================
 * External routines when attribute list is inside the inode
 *========================================================================*/

929 930 931 932 933 934 935 936
static inline int xfs_attr_sf_totsize(struct xfs_inode *dp)
{
	struct xfs_attr_shortform *sf;

	sf = (struct xfs_attr_shortform *)dp->i_afp->if_u1.if_data;
	return be16_to_cpu(sf->hdr.totsize);
}

L
Linus Torvalds 已提交
937 938 939 940 941 942 943
/*
 * Add a name to the shortform attribute list structure
 * This is the external routine.
 */
STATIC int
xfs_attr_shortform_addname(xfs_da_args_t *args)
{
944
	int newsize, forkoff, retval;
L
Linus Torvalds 已提交
945

946 947
	trace_xfs_attr_sf_addname(args);

L
Linus Torvalds 已提交
948
	retval = xfs_attr_shortform_lookup(args);
949
	if (retval == -ENOATTR && (args->attr_flags & XATTR_REPLACE))
E
Eric Sandeen 已提交
950
		return retval;
951
	if (retval == -EEXIST) {
952
		if (args->attr_flags & XATTR_CREATE)
E
Eric Sandeen 已提交
953
			return retval;
954
		retval = xfs_attr_sf_removename(args);
955 956 957 958 959 960 961
		if (retval)
			return retval;
		/*
		 * Since we have removed the old attr, clear ATTR_REPLACE so
		 * that the leaf format add routine won't trip over the attr
		 * not being around.
		 */
962
		args->attr_flags &= ~XATTR_REPLACE;
L
Linus Torvalds 已提交
963 964
	}

965 966
	if (args->namelen >= XFS_ATTR_SF_ENTSIZE_MAX ||
	    args->valuelen >= XFS_ATTR_SF_ENTSIZE_MAX)
D
Dave Chinner 已提交
967
		return -ENOSPC;
968

969 970
	newsize = xfs_attr_sf_totsize(args->dp);
	newsize += xfs_attr_sf_entsize_byname(args->namelen, args->valuelen);
971 972 973

	forkoff = xfs_attr_shortform_bytesfit(args->dp, newsize);
	if (!forkoff)
D
Dave Chinner 已提交
974
		return -ENOSPC;
975 976

	xfs_attr_shortform_add(args, forkoff);
E
Eric Sandeen 已提交
977
	return 0;
L
Linus Torvalds 已提交
978 979 980 981 982 983 984
}


/*========================================================================
 * External routines when attribute list is one block
 *========================================================================*/

985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008
/* Store info about a remote block */
STATIC void
xfs_attr_save_rmt_blk(
	struct xfs_da_args	*args)
{
	args->blkno2 = args->blkno;
	args->index2 = args->index;
	args->rmtblkno2 = args->rmtblkno;
	args->rmtblkcnt2 = args->rmtblkcnt;
	args->rmtvaluelen2 = args->rmtvaluelen;
}

/* Set stored info about a remote block */
STATIC void
xfs_attr_restore_rmt_blk(
	struct xfs_da_args	*args)
{
	args->blkno = args->blkno2;
	args->index = args->index2;
	args->rmtblkno = args->rmtblkno2;
	args->rmtblkcnt = args->rmtblkcnt2;
	args->rmtvaluelen = args->rmtvaluelen2;
}

L
Linus Torvalds 已提交
1009
/*
1010
 * Tries to add an attribute to an inode in leaf form
L
Linus Torvalds 已提交
1011
 *
1012 1013 1014 1015 1016 1017
 * This function is meant to execute as part of a delayed operation and leaves
 * the transaction handling to the caller.  On success the attribute is added
 * and the inode and transaction are left dirty.  If there is not enough space,
 * the attr data is converted to node format and -ENOSPC is returned. Caller is
 * responsible for handling the dirty inode and transaction or adding the attr
 * in node format.
L
Linus Torvalds 已提交
1018
 */
D
David Chinner 已提交
1019
STATIC int
1020 1021 1022
xfs_attr_leaf_try_add(
	struct xfs_da_args	*args,
	struct xfs_buf		*bp)
L
Linus Torvalds 已提交
1023
{
1024
	int			error;
1025

L
Linus Torvalds 已提交
1026
	/*
1027 1028 1029 1030
	 * If the caller provided a buffer to us, it is locked and held in
	 * the transaction because it just did a shortform to leaf conversion.
	 * Hence we don't need to read it again. Otherwise read in the leaf
	 * buffer.
L
Linus Torvalds 已提交
1031
	 */
1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044
	if (bp) {
		xfs_trans_bhold_release(args->trans, bp);
	} else {
		error = xfs_attr3_leaf_read(args->trans, args->dp, 0, &bp);
		if (error)
			return error;
	}

	/*
	 * Look up the xattr name to set the insertion point for the new xattr.
	 */
	error = xfs_attr3_leaf_lookup_int(bp, args);
	if (error != -ENOATTR && error != -EEXIST)
1045
		goto out_brelse;
1046 1047 1048
	if (error == -ENOATTR && (args->attr_flags & XATTR_REPLACE))
		goto out_brelse;
	if (error == -EEXIST) {
1049
		if (args->attr_flags & XATTR_CREATE)
1050
			goto out_brelse;
1051 1052 1053

		trace_xfs_attr_leaf_replace(args);

1054
		/* save the attribute state for later removal*/
1055
		args->op_flags |= XFS_DA_OP_RENAME;	/* an atomic rename */
1056
		xfs_attr_save_rmt_blk(args);
1057 1058 1059 1060 1061 1062 1063 1064 1065

		/*
		 * clear the remote attr state now that it is saved so that the
		 * values reflect the state of the attribute we are about to
		 * add, not the attribute we just found and will remove later.
		 */
		args->rmtblkno = 0;
		args->rmtblkcnt = 0;
		args->rmtvaluelen = 0;
L
Linus Torvalds 已提交
1066 1067
	}

1068 1069
	return xfs_attr3_leaf_add(bp, args);

1070 1071
out_brelse:
	xfs_trans_brelse(args->trans, bp);
1072
	return error;
1073
}
L
Linus Torvalds 已提交
1074

1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095
/*
 * Return EEXIST if attr is found, or ENOATTR if not
 */
STATIC int
xfs_attr_leaf_hasname(
	struct xfs_da_args	*args,
	struct xfs_buf		**bp)
{
	int                     error = 0;

	error = xfs_attr3_leaf_read(args->trans, args->dp, 0, bp);
	if (error)
		return error;

	error = xfs_attr3_leaf_lookup_int(*bp, args);
	if (error != -ENOATTR && error != -EEXIST)
		xfs_trans_brelse(args->trans, *bp);

	return error;
}

L
Linus Torvalds 已提交
1096 1097 1098 1099 1100 1101 1102
/*
 * Remove a name from the leaf attribute list structure
 *
 * This leaf block cannot have a "remote" value, we only call this routine
 * if bmap_one_block() says there is only one block (ie: no remote blks).
 */
STATIC int
1103 1104
xfs_attr_leaf_removename(
	struct xfs_da_args	*args)
L
Linus Torvalds 已提交
1105
{
1106 1107 1108
	struct xfs_inode	*dp;
	struct xfs_buf		*bp;
	int			error, forkoff;
L
Linus Torvalds 已提交
1109

1110 1111
	trace_xfs_attr_leaf_removename(args);

L
Linus Torvalds 已提交
1112 1113 1114 1115 1116
	/*
	 * Remove the attribute.
	 */
	dp = args->dp;

1117 1118
	error = xfs_attr_leaf_hasname(args, &bp);

D
Dave Chinner 已提交
1119
	if (error == -ENOATTR) {
1120
		xfs_trans_brelse(args->trans, bp);
D
Dave Chinner 已提交
1121
		return error;
1122 1123
	} else if (error != -EEXIST)
		return error;
L
Linus Torvalds 已提交
1124

D
Dave Chinner 已提交
1125
	xfs_attr3_leaf_remove(bp, args);
L
Linus Torvalds 已提交
1126 1127 1128 1129

	/*
	 * If the result is small enough, shrink it all into the inode.
	 */
1130 1131 1132
	forkoff = xfs_attr_shortform_allfit(bp, dp);
	if (forkoff)
		return xfs_attr3_leaf_to_shortform(bp, args, forkoff);
L
Linus Torvalds 已提交
1133
		/* bp is gone due to xfs_da_shrink_inode */
1134

D
Dave Chinner 已提交
1135
	return 0;
L
Linus Torvalds 已提交
1136 1137 1138 1139 1140 1141 1142
}

/*
 * Look up a name in a leaf attribute list structure.
 *
 * This leaf block cannot have a "remote" value, we only call this routine
 * if bmap_one_block() says there is only one block (ie: no remote blks).
1143 1144
 *
 * Returns 0 on successful retrieval, otherwise an error.
L
Linus Torvalds 已提交
1145
 */
1146
STATIC int
L
Linus Torvalds 已提交
1147 1148
xfs_attr_leaf_get(xfs_da_args_t *args)
{
1149
	struct xfs_buf *bp;
L
Linus Torvalds 已提交
1150 1151
	int error;

1152 1153
	trace_xfs_attr_leaf_get(args);

1154
	error = xfs_attr_leaf_hasname(args, &bp);
L
Linus Torvalds 已提交
1155

1156
	if (error == -ENOATTR)  {
1157
		xfs_trans_brelse(args->trans, bp);
D
Dave Chinner 已提交
1158
		return error;
1159 1160 1161 1162
	} else if (error != -EEXIST)
		return error;


D
Dave Chinner 已提交
1163
	error = xfs_attr3_leaf_getvalue(bp, args);
1164
	xfs_trans_brelse(args->trans, bp);
1165
	return error;
L
Linus Torvalds 已提交
1166 1167
}

1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182
/*
 * Return EEXIST if attr is found, or ENOATTR if not
 * statep: If not null is set to point at the found state.  Caller will
 *         be responsible for freeing the state in this case.
 */
STATIC int
xfs_attr_node_hasname(
	struct xfs_da_args	*args,
	struct xfs_da_state	**statep)
{
	struct xfs_da_state	*state;
	int			retval, error;

	state = xfs_da_state_alloc(args);
	if (statep != NULL)
1183
		*statep = state;
1184 1185 1186 1187 1188

	/*
	 * Search to see if name exists, and get back a pointer to it.
	 */
	error = xfs_da3_node_lookup_int(state, &retval);
1189 1190
	if (error)
		retval = error;
1191

1192
	if (!statep)
1193
		xfs_da_state_free(state);
1194

1195 1196 1197
	return retval;
}

L
Linus Torvalds 已提交
1198
/*========================================================================
1199
 * External routines when attribute list size > geo->blksize
L
Linus Torvalds 已提交
1200 1201 1202
 *========================================================================*/

STATIC int
1203
xfs_attr_node_addname_find_attr(
1204
	 struct xfs_attr_item		*attr)
L
Linus Torvalds 已提交
1205
{
1206
	struct xfs_da_args		*args = attr->xattri_da_args;
1207
	int				retval;
1208

L
Linus Torvalds 已提交
1209 1210 1211 1212
	/*
	 * Search to see if name already exists, and get back a pointer
	 * to where it should go.
	 */
1213
	retval = xfs_attr_node_hasname(args, &attr->xattri_da_state);
1214
	if (retval != -ENOATTR && retval != -EEXIST)
1215
		goto error;
1216

1217
	if (retval == -ENOATTR && (args->attr_flags & XATTR_REPLACE))
1218
		goto error;
1219
	if (retval == -EEXIST) {
1220
		if (args->attr_flags & XATTR_CREATE)
1221
			goto error;
1222 1223 1224

		trace_xfs_attr_node_replace(args);

1225
		/* save the attribute state for later removal*/
1226
		args->op_flags |= XFS_DA_OP_RENAME;	/* atomic rename op */
1227
		xfs_attr_save_rmt_blk(args);
1228 1229 1230 1231 1232 1233

		/*
		 * clear the remote attr state now that it is saved so that the
		 * values reflect the state of the attribute we are about to
		 * add, not the attribute we just found and will remove later.
		 */
L
Linus Torvalds 已提交
1234 1235
		args->rmtblkno = 0;
		args->rmtblkcnt = 0;
1236
		args->rmtvaluelen = 0;
L
Linus Torvalds 已提交
1237 1238
	}

1239 1240
	return 0;
error:
1241 1242
	if (attr->xattri_da_state)
		xfs_da_state_free(attr->xattri_da_state);
1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254
	return retval;
}

/*
 * Add a name to a Btree-format attribute list.
 *
 * This will involve walking down the Btree, and may involve splitting
 * leaf nodes and even splitting intermediate nodes up to and including
 * the root node (a special case of an intermediate node).
 *
 * "Remote" attribute values confuse the issue and atomic rename operations
 * add a whole extra layer of confusion on top of that.
1255 1256 1257 1258 1259
 *
 * This routine is meant to function as a delayed operation, and may return
 * -EAGAIN when the transaction needs to be rolled.  Calling functions will need
 * to handle this, and recall the function until a successful error code is
 *returned.
1260
 */
1261 1262
static int
xfs_attr_node_try_addname(
1263
	struct xfs_attr_item		*attr)
1264
{
1265 1266
	struct xfs_da_args		*args = attr->xattri_da_args;
	struct xfs_da_state		*state = attr->xattri_da_state;
1267 1268
	struct xfs_da_state_blk		*blk;
	int				error;
1269 1270 1271 1272 1273 1274

	trace_xfs_attr_node_addname(args);

	blk = &state->path.blk[state->path.active-1];
	ASSERT(blk->magic == XFS_ATTR_LEAF_MAGIC);

1275 1276
	error = xfs_attr3_leaf_add(blk->bp, state->args);
	if (error == -ENOSPC) {
L
Linus Torvalds 已提交
1277 1278 1279 1280 1281 1282 1283
		if (state->path.active == 1) {
			/*
			 * Its really a single leaf node, but it had
			 * out-of-line values so it looked like it *might*
			 * have been a b-tree.
			 */
			xfs_da_state_free(state);
1284
			state = NULL;
D
Dave Chinner 已提交
1285
			error = xfs_attr3_leaf_to_node(args);
1286
			if (error)
1287
				goto out;
L
Linus Torvalds 已提交
1288 1289

			/*
1290 1291 1292 1293 1294
			 * Now that we have converted the leaf to a node, we can
			 * roll the transaction, and try xfs_attr3_leaf_add
			 * again on re-entry.  No need to set dela_state to do
			 * this. dela_state is still unset by this function at
			 * this point.
L
Linus Torvalds 已提交
1295
			 */
1296
			trace_xfs_attr_node_addname_return(
1297
					attr->xattri_dela_state, args->dp);
1298
			return -EAGAIN;
L
Linus Torvalds 已提交
1299 1300 1301 1302 1303 1304 1305 1306
		}

		/*
		 * Split as many Btree elements as required.
		 * This code tracks the new and old attr's location
		 * in the index/blkno/rmtblkno/rmtblkcnt fields and
		 * in the index2/blkno2/rmtblkno2/rmtblkcnt2 fields.
		 */
1307
		error = xfs_da3_split(state);
1308
		if (error)
1309
			goto out;
L
Linus Torvalds 已提交
1310 1311 1312 1313
	} else {
		/*
		 * Addition succeeded, update Btree hashvals.
		 */
1314
		xfs_da3_fixhashpath(state, &state->path);
L
Linus Torvalds 已提交
1315 1316
	}

1317 1318 1319
out:
	if (state)
		xfs_da_state_free(state);
1320
	return error;
1321 1322 1323
}


1324 1325
static int
xfs_attr_node_remove_attr(
1326
	struct xfs_attr_item		*attr)
1327
{
1328
	struct xfs_da_args		*args = attr->xattri_da_args;
1329
	struct xfs_da_state		*state = NULL;
1330
	struct xfs_mount		*mp = args->dp->i_mount;
1331 1332 1333
	int				retval = 0;
	int				error = 0;

1334 1335 1336 1337
	/*
	 * Re-find the "old" attribute entry after any split ops. The INCOMPLETE
	 * flag means that we will find the "old" attr, not the "new" one.
	 */
1338 1339
	if (!xfs_has_larp(mp))
		args->attr_filter |= XFS_ATTR_INCOMPLETE;
1340 1341 1342 1343 1344
	state = xfs_da_state_alloc(args);
	state->inleaf = 0;
	error = xfs_da3_node_lookup_int(state, &retval);
	if (error)
		goto out;
L
Linus Torvalds 已提交
1345

1346
	error = xfs_attr_node_removename(args, state);
L
Linus Torvalds 已提交
1347

1348 1349 1350 1351 1352
	/*
	 * Check to see if the tree needs to be collapsed.
	 */
	if (retval && (state->path.active > 1)) {
		error = xfs_da3_join(state);
L
Linus Torvalds 已提交
1353 1354 1355 1356 1357 1358 1359 1360 1361
		if (error)
			goto out;
	}
	retval = error = 0;

out:
	if (state)
		xfs_da_state_free(state);
	if (error)
E
Eric Sandeen 已提交
1362 1363
		return error;
	return retval;
L
Linus Torvalds 已提交
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
/*
 * Mark an attribute entry INCOMPLETE and save pointers to the relevant buffers
 * for later deletion of the entry.
 */
STATIC int
xfs_attr_leaf_mark_incomplete(
	struct xfs_da_args	*args,
	struct xfs_da_state	*state)
{
	int			error;

	/*
	 * Fill in disk block numbers in the state structure
	 * so that we can get the buffers back after we commit
	 * several transactions in the following calls.
	 */
	error = xfs_attr_fillstate(state);
	if (error)
		return error;

	/*
	 * Mark the attribute as INCOMPLETE
	 */
	return xfs_attr3_leaf_setflag(args);
}

1393 1394 1395 1396 1397 1398 1399
/*
 * Initial setup for xfs_attr_node_removename.  Make sure the attr is there and
 * the blocks are valid.  Attr keys with remote blocks will be marked
 * incomplete.
 */
STATIC
int xfs_attr_node_removename_setup(
1400
	struct xfs_attr_item		*attr)
1401
{
1402 1403
	struct xfs_da_args		*args = attr->xattri_da_args;
	struct xfs_da_state		**state = &attr->xattri_da_state;
1404
	int				error;
1405 1406 1407

	error = xfs_attr_node_hasname(args, state);
	if (error != -EEXIST)
1408
		goto out;
1409
	error = 0;
1410 1411 1412 1413 1414 1415 1416 1417

	ASSERT((*state)->path.blk[(*state)->path.active - 1].bp != NULL);
	ASSERT((*state)->path.blk[(*state)->path.active - 1].magic ==
		XFS_ATTR_LEAF_MAGIC);

	if (args->rmtblkno > 0) {
		error = xfs_attr_leaf_mark_incomplete(args, *state);
		if (error)
1418
			goto out;
1419

1420
		error = xfs_attr_rmtval_invalidate(args);
1421
	}
1422 1423 1424
out:
	if (error)
		xfs_da_state_free(*state);
1425

1426
	return error;
1427 1428
}

1429
STATIC int
1430
xfs_attr_node_removename(
1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447
	struct xfs_da_args	*args,
	struct xfs_da_state	*state)
{
	struct xfs_da_state_blk	*blk;
	int			retval;

	/*
	 * Remove the name and update the hashvals in the tree.
	 */
	blk = &state->path.blk[state->path.active-1];
	ASSERT(blk->magic == XFS_ATTR_LEAF_MAGIC);
	retval = xfs_attr3_leaf_remove(blk->bp, args);
	xfs_da3_fixhashpath(state, &state->path);

	return retval;
}

L
Linus Torvalds 已提交
1448
/*
1449
 * Remove the attribute specified in @args.
L
Linus Torvalds 已提交
1450 1451 1452 1453
 *
 * This will involve walking down the Btree, and may involve joining
 * leaf nodes and even joining intermediate nodes up to and including
 * the root node (a special case of an intermediate node).
1454 1455 1456 1457 1458
 *
 * This routine is meant to function as either an in-line or delayed operation,
 * and may return -EAGAIN when the transaction needs to be rolled.  Calling
 * functions will need to handle this, and call the function until a
 * successful error code is returned.
L
Linus Torvalds 已提交
1459
 */
1460 1461
int
xfs_attr_remove_iter(
1462
	struct xfs_attr_item		*attr)
L
Linus Torvalds 已提交
1463
{
1464 1465
	struct xfs_da_args		*args = attr->xattri_da_args;
	struct xfs_da_state		*state = attr->xattri_da_state;
1466
	int				retval, error = 0;
1467
	struct xfs_inode		*dp = args->dp;
L
Linus Torvalds 已提交
1468

1469 1470
	trace_xfs_attr_node_removename(args);

1471
	switch (attr->xattri_dela_state) {
1472 1473 1474
	case XFS_DAS_UNINIT:
		if (!xfs_inode_hasattr(dp))
			return -ENOATTR;
L
Linus Torvalds 已提交
1475

1476 1477 1478 1479 1480
		/*
		 * Shortform or leaf formats don't require transaction rolls and
		 * thus state transitions. Call the right helper and return.
		 */
		if (dp->i_afp->if_format == XFS_DINODE_FMT_LOCAL)
1481
			return xfs_attr_sf_removename(args);
1482 1483 1484

		if (xfs_attr_is_leaf(dp))
			return xfs_attr_leaf_removename(args);
1485 1486

		/*
1487 1488
		 * Node format may require transaction rolls. Set up the
		 * state context and fall into the state machine.
1489
		 */
1490 1491
		if (!attr->xattri_da_state) {
			error = xfs_attr_node_removename_setup(attr);
1492 1493
			if (error)
				return error;
1494
			state = attr->xattri_da_state;
1495 1496
		}

1497
		fallthrough;
1498
	case XFS_DAS_RMTBLK:
1499
		attr->xattri_dela_state = XFS_DAS_RMTBLK;
L
Linus Torvalds 已提交
1500 1501

		/*
1502 1503 1504 1505
		 * If there is an out-of-line value, de-allocate the blocks.
		 * This is done before we remove the attribute so that we don't
		 * overflow the maximum size of a transaction and/or hit a
		 * deadlock.
L
Linus Torvalds 已提交
1506
		 */
1507 1508 1509 1510 1511
		if (args->rmtblkno > 0) {
			/*
			 * May return -EAGAIN. Roll and repeat until all remote
			 * blocks are removed.
			 */
1512
			error = xfs_attr_rmtval_remove(attr);
1513 1514
			if (error == -EAGAIN) {
				trace_xfs_attr_remove_iter_return(
1515
					attr->xattri_dela_state, args->dp);
1516
				return error;
1517
			} else if (error) {
1518
				goto out;
1519
			}
L
Linus Torvalds 已提交
1520

1521 1522 1523 1524 1525 1526 1527 1528 1529
			/*
			 * Refill the state structure with buffers (the prior
			 * calls released our buffers) and close out this
			 * transaction before proceeding.
			 */
			ASSERT(args->rmtblkno == 0);
			error = xfs_attr_refillstate(state);
			if (error)
				goto out;
1530 1531 1532 1533

			attr->xattri_dela_state = XFS_DAS_RM_NAME;
			trace_xfs_attr_remove_iter_return(
					attr->xattri_dela_state, args->dp);
1534 1535 1536
			return -EAGAIN;
		}

1537
		fallthrough;
1538 1539 1540 1541 1542
	case XFS_DAS_RM_NAME:
		/*
		 * If we came here fresh from a transaction roll, reattach all
		 * the buffers to the current transaction.
		 */
1543
		if (attr->xattri_dela_state == XFS_DAS_RM_NAME) {
1544 1545 1546 1547
			error = xfs_attr_refillstate(state);
			if (error)
				goto out;
		}
L
Linus Torvalds 已提交
1548

1549
		retval = xfs_attr_node_removename(args, state);
1550 1551 1552 1553 1554 1555 1556 1557 1558 1559

		/*
		 * Check to see if the tree needs to be collapsed. If so, roll
		 * the transacton and fall into the shrink state.
		 */
		if (retval && (state->path.active > 1)) {
			error = xfs_da3_join(state);
			if (error)
				goto out;

1560
			attr->xattri_dela_state = XFS_DAS_RM_SHRINK;
1561
			trace_xfs_attr_remove_iter_return(
1562
					attr->xattri_dela_state, args->dp);
1563 1564 1565
			return -EAGAIN;
		}

1566
		fallthrough;
1567 1568 1569 1570 1571 1572 1573
	case XFS_DAS_RM_SHRINK:
		/*
		 * If the result is small enough, push it all into the inode.
		 * This is our final state so it's safe to return a dirty
		 * transaction.
		 */
		if (xfs_attr_is_leaf(dp))
1574
			error = xfs_attr_leaf_shrink(args, state);
1575 1576 1577 1578 1579 1580 1581
		ASSERT(error != -EAGAIN);
		break;
	default:
		ASSERT(0);
		error = -EINVAL;
		goto out;
	}
L
Linus Torvalds 已提交
1582
out:
1583 1584
	if (state)
		xfs_da_state_free(state);
E
Eric Sandeen 已提交
1585
	return error;
L
Linus Torvalds 已提交
1586 1587 1588 1589 1590 1591
}

/*
 * Fill in the disk block numbers in the state structure for the buffers
 * that are attached to the state structure.
 * This is done so that we can quickly reattach ourselves to those buffers
1592
 * after some set of transaction commits have released these buffers.
L
Linus Torvalds 已提交
1593 1594 1595 1596 1597 1598 1599 1600
 */
STATIC int
xfs_attr_fillstate(xfs_da_state_t *state)
{
	xfs_da_state_path_t *path;
	xfs_da_state_blk_t *blk;
	int level;

1601 1602
	trace_xfs_attr_fillstate(state->args);

L
Linus Torvalds 已提交
1603 1604 1605 1606 1607 1608 1609 1610
	/*
	 * Roll down the "path" in the state structure, storing the on-disk
	 * block number for those buffers in the "path".
	 */
	path = &state->path;
	ASSERT((path->active >= 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
	for (blk = path->blk, level = 0; level < path->active; blk++, level++) {
		if (blk->bp) {
D
Dave Chinner 已提交
1611
			blk->disk_blkno = xfs_buf_daddr(blk->bp);
L
Linus Torvalds 已提交
1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625
			blk->bp = NULL;
		} else {
			blk->disk_blkno = 0;
		}
	}

	/*
	 * Roll down the "altpath" in the state structure, storing the on-disk
	 * block number for those buffers in the "altpath".
	 */
	path = &state->altpath;
	ASSERT((path->active >= 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
	for (blk = path->blk, level = 0; level < path->active; blk++, level++) {
		if (blk->bp) {
D
Dave Chinner 已提交
1626
			blk->disk_blkno = xfs_buf_daddr(blk->bp);
L
Linus Torvalds 已提交
1627 1628 1629 1630 1631 1632
			blk->bp = NULL;
		} else {
			blk->disk_blkno = 0;
		}
	}

E
Eric Sandeen 已提交
1633
	return 0;
L
Linus Torvalds 已提交
1634 1635 1636 1637 1638
}

/*
 * Reattach the buffers to the state structure based on the disk block
 * numbers stored in the state structure.
1639
 * This is done after some set of transaction commits have released those
L
Linus Torvalds 已提交
1640 1641 1642 1643 1644 1645 1646 1647 1648
 * buffers from our grip.
 */
STATIC int
xfs_attr_refillstate(xfs_da_state_t *state)
{
	xfs_da_state_path_t *path;
	xfs_da_state_blk_t *blk;
	int level, error;

1649 1650
	trace_xfs_attr_refillstate(state->args);

L
Linus Torvalds 已提交
1651 1652 1653 1654 1655 1656 1657 1658
	/*
	 * Roll down the "path" in the state structure, storing the on-disk
	 * block number for those buffers in the "path".
	 */
	path = &state->path;
	ASSERT((path->active >= 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
	for (blk = path->blk, level = 0; level < path->active; blk++, level++) {
		if (blk->disk_blkno) {
C
Christoph Hellwig 已提交
1659 1660 1661
			error = xfs_da3_node_read_mapped(state->args->trans,
					state->args->dp, blk->disk_blkno,
					&blk->bp, XFS_ATTR_FORK);
L
Linus Torvalds 已提交
1662
			if (error)
E
Eric Sandeen 已提交
1663
				return error;
L
Linus Torvalds 已提交
1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676
		} else {
			blk->bp = NULL;
		}
	}

	/*
	 * Roll down the "altpath" in the state structure, storing the on-disk
	 * block number for those buffers in the "altpath".
	 */
	path = &state->altpath;
	ASSERT((path->active >= 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
	for (blk = path->blk, level = 0; level < path->active; blk++, level++) {
		if (blk->disk_blkno) {
C
Christoph Hellwig 已提交
1677 1678 1679
			error = xfs_da3_node_read_mapped(state->args->trans,
					state->args->dp, blk->disk_blkno,
					&blk->bp, XFS_ATTR_FORK);
L
Linus Torvalds 已提交
1680
			if (error)
E
Eric Sandeen 已提交
1681
				return error;
L
Linus Torvalds 已提交
1682 1683 1684 1685 1686
		} else {
			blk->bp = NULL;
		}
	}

E
Eric Sandeen 已提交
1687
	return 0;
L
Linus Torvalds 已提交
1688 1689 1690
}

/*
1691
 * Retrieve the attribute data from a node attribute list.
L
Linus Torvalds 已提交
1692 1693 1694 1695
 *
 * This routine gets called for any attribute fork that has more than one
 * block, ie: both true Btree attr lists and for single-leaf-blocks with
 * "remote" values taking up more blocks.
1696 1697
 *
 * Returns 0 on successful retrieval, otherwise an error.
L
Linus Torvalds 已提交
1698
 */
1699
STATIC int
1700 1701
xfs_attr_node_get(
	struct xfs_da_args	*args)
L
Linus Torvalds 已提交
1702
{
1703 1704 1705 1706
	struct xfs_da_state	*state;
	struct xfs_da_state_blk	*blk;
	int			i;
	int			error;
L
Linus Torvalds 已提交
1707

1708 1709
	trace_xfs_attr_node_get(args);

L
Linus Torvalds 已提交
1710 1711 1712
	/*
	 * Search to see if name exists, and get back a pointer to it.
	 */
1713 1714
	error = xfs_attr_node_hasname(args, &state);
	if (error != -EEXIST)
1715 1716 1717 1718 1719 1720
		goto out_release;

	/*
	 * Get the value, local or "remote"
	 */
	blk = &state->path.blk[state->path.active - 1];
1721
	error = xfs_attr3_leaf_getvalue(blk->bp, args);
L
Linus Torvalds 已提交
1722 1723 1724 1725

	/*
	 * If not in a transaction, we have to release all the buffers.
	 */
1726
out_release:
1727
	for (i = 0; state != NULL && i < state->path.active; i++) {
1728
		xfs_trans_brelse(args->trans, state->path.blk[i].bp);
L
Linus Torvalds 已提交
1729 1730 1731
		state->path.blk[i].bp = NULL;
	}

1732 1733 1734
	if (state)
		xfs_da_state_free(state);
	return error;
L
Linus Torvalds 已提交
1735
}
1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752

/* Returns true if the attribute entry name is valid. */
bool
xfs_attr_namecheck(
	const void	*name,
	size_t		length)
{
	/*
	 * MAXNAMELEN includes the trailing null, but (name/length) leave it
	 * out, so use >= for the length check.
	 */
	if (length >= MAXNAMELEN)
		return false;

	/* There shouldn't be any nulls here */
	return !memchr(name, 0, length);
}