send.c 152.7 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
/*
 * Copyright (C) 2012 Alexander Block.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will 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.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

#include <linux/bsearch.h>
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/sort.h>
#include <linux/mount.h>
#include <linux/xattr.h>
#include <linux/posix_acl_xattr.h>
#include <linux/radix-tree.h>
27
#include <linux/vmalloc.h>
A
Andy Shevchenko 已提交
28
#include <linux/string.h>
29 30 31

#include "send.h"
#include "backref.h"
32
#include "hash.h"
33 34 35 36
#include "locking.h"
#include "disk-io.h"
#include "btrfs_inode.h"
#include "transaction.h"
37
#include "compression.h"
38 39 40 41 42 43 44 45 46 47 48 49 50 51 52

/*
 * A fs_path is a helper to dynamically build path names with unknown size.
 * It reallocates the internal buffer on demand.
 * It allows fast adding of path elements on the right side (normal path) and
 * fast adding to the left side (reversed path). A reversed path can also be
 * unreversed if needed.
 */
struct fs_path {
	union {
		struct {
			char *start;
			char *end;

			char *buf;
53 54
			unsigned short buf_len:15;
			unsigned short reversed:1;
55 56
			char inline_buf[];
		};
57 58 59 60 61 62
		/*
		 * Average path length does not exceed 200 bytes, we'll have
		 * better packing in the slab and higher chance to satisfy
		 * a allocation later during send.
		 */
		char pad[256];
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88
	};
};
#define FS_PATH_INLINE_SIZE \
	(sizeof(struct fs_path) - offsetof(struct fs_path, inline_buf))


/* reused for each extent */
struct clone_root {
	struct btrfs_root *root;
	u64 ino;
	u64 offset;

	u64 found_refs;
};

#define SEND_CTX_MAX_NAME_CACHE_SIZE 128
#define SEND_CTX_NAME_CACHE_CLEAN_SIZE (SEND_CTX_MAX_NAME_CACHE_SIZE * 2)

struct send_ctx {
	struct file *send_filp;
	loff_t send_off;
	char *send_buf;
	u32 send_size;
	u32 send_max_size;
	u64 total_send_size;
	u64 cmd_send_size[BTRFS_SEND_C_MAX + 1];
89
	u64 flags;	/* 'flags' member of btrfs_ioctl_send_args is u64 */
90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111

	struct btrfs_root *send_root;
	struct btrfs_root *parent_root;
	struct clone_root *clone_roots;
	int clone_roots_cnt;

	/* current state of the compare_tree call */
	struct btrfs_path *left_path;
	struct btrfs_path *right_path;
	struct btrfs_key *cmp_key;

	/*
	 * infos of the currently processed inode. In case of deleted inodes,
	 * these are the values from the deleted inode.
	 */
	u64 cur_ino;
	u64 cur_inode_gen;
	int cur_inode_new;
	int cur_inode_new_gen;
	int cur_inode_deleted;
	u64 cur_inode_size;
	u64 cur_inode_mode;
L
Liu Bo 已提交
112
	u64 cur_inode_rdev;
113
	u64 cur_inode_last_extent;
114 115 116 117 118 119 120 121 122 123

	u64 send_progress;

	struct list_head new_refs;
	struct list_head deleted_refs;

	struct radix_tree_root name_cache;
	struct list_head name_cache_list;
	int name_cache_size;

L
Liu Bo 已提交
124 125
	struct file_ra_state ra;

126
	char *read_buf;
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

	/*
	 * We process inodes by their increasing order, so if before an
	 * incremental send we reverse the parent/child relationship of
	 * directories such that a directory with a lower inode number was
	 * the parent of a directory with a higher inode number, and the one
	 * becoming the new parent got renamed too, we can't rename/move the
	 * directory with lower inode number when we finish processing it - we
	 * must process the directory with higher inode number first, then
	 * rename/move it and then rename/move the directory with lower inode
	 * number. Example follows.
	 *
	 * Tree state when the first send was performed:
	 *
	 * .
	 * |-- a                   (ino 257)
	 *     |-- b               (ino 258)
	 *         |
	 *         |
	 *         |-- c           (ino 259)
	 *         |   |-- d       (ino 260)
	 *         |
	 *         |-- c2          (ino 261)
	 *
	 * Tree state when the second (incremental) send is performed:
	 *
	 * .
	 * |-- a                   (ino 257)
	 *     |-- b               (ino 258)
	 *         |-- c2          (ino 261)
	 *             |-- d2      (ino 260)
	 *                 |-- cc  (ino 259)
	 *
	 * The sequence of steps that lead to the second state was:
	 *
	 * mv /a/b/c/d /a/b/c2/d2
	 * mv /a/b/c /a/b/c2/d2/cc
	 *
	 * "c" has lower inode number, but we can't move it (2nd mv operation)
	 * before we move "d", which has higher inode number.
	 *
	 * So we just memorize which move/rename operations must be performed
	 * later when their respective parent is processed and moved/renamed.
	 */

	/* Indexed by parent directory inode number. */
	struct rb_root pending_dir_moves;

	/*
	 * Reverse index, indexed by the inode number of a directory that
	 * is waiting for the move/rename of its immediate parent before its
	 * own move/rename can be performed.
	 */
	struct rb_root waiting_dir_moves;
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

	/*
	 * A directory that is going to be rm'ed might have a child directory
	 * which is in the pending directory moves index above. In this case,
	 * the directory can only be removed after the move/rename of its child
	 * is performed. Example:
	 *
	 * Parent snapshot:
	 *
	 * .                        (ino 256)
	 * |-- a/                   (ino 257)
	 *     |-- b/               (ino 258)
	 *         |-- c/           (ino 259)
	 *         |   |-- x/       (ino 260)
	 *         |
	 *         |-- y/           (ino 261)
	 *
	 * Send snapshot:
	 *
	 * .                        (ino 256)
	 * |-- a/                   (ino 257)
	 *     |-- b/               (ino 258)
	 *         |-- YY/          (ino 261)
	 *              |-- x/      (ino 260)
	 *
	 * Sequence of steps that lead to the send snapshot:
	 * rm -f /a/b/c/foo.txt
	 * mv /a/b/y /a/b/YY
	 * mv /a/b/c/x /a/b/YY
	 * rmdir /a/b/c
	 *
	 * When the child is processed, its move/rename is delayed until its
	 * parent is processed (as explained above), but all other operations
	 * like update utimes, chown, chgrp, etc, are performed and the paths
	 * that it uses for those operations must use the orphanized name of
	 * its parent (the directory we're going to rm later), so we need to
	 * memorize that name.
	 *
	 * Indexed by the inode number of the directory to be deleted.
	 */
	struct rb_root orphan_dirs;
222 223 224 225 226 227 228 229 230 231 232 233 234 235
};

struct pending_dir_move {
	struct rb_node node;
	struct list_head list;
	u64 parent_ino;
	u64 ino;
	u64 gen;
	struct list_head update_refs;
};

struct waiting_dir_move {
	struct rb_node node;
	u64 ino;
236 237 238 239 240 241
	/*
	 * There might be some directory that could not be removed because it
	 * was waiting for this directory inode to be moved first. Therefore
	 * after this directory is moved, we can try to rmdir the ino rmdir_ino.
	 */
	u64 rmdir_ino;
242
	bool orphanized;
243 244 245 246 247 248
};

struct orphan_dir_info {
	struct rb_node node;
	u64 ino;
	u64 gen;
249 250 251 252
};

struct name_cache_entry {
	struct list_head list;
253 254 255 256 257 258 259 260 261
	/*
	 * radix_tree has only 32bit entries but we need to handle 64bit inums.
	 * We use the lower 32bit of the 64bit inum to store it in the tree. If
	 * more then one inum would fall into the same entry, we use radix_list
	 * to store the additional entries. radix_list is also used to store
	 * entries where two entries have the same inum but different
	 * generations.
	 */
	struct list_head radix_list;
262 263 264 265 266 267 268 269 270 271
	u64 ino;
	u64 gen;
	u64 parent_ino;
	u64 parent_gen;
	int ret;
	int need_later_update;
	int name_len;
	char name[];
};

272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304
static void inconsistent_snapshot_error(struct send_ctx *sctx,
					enum btrfs_compare_tree_result result,
					const char *what)
{
	const char *result_string;

	switch (result) {
	case BTRFS_COMPARE_TREE_NEW:
		result_string = "new";
		break;
	case BTRFS_COMPARE_TREE_DELETED:
		result_string = "deleted";
		break;
	case BTRFS_COMPARE_TREE_CHANGED:
		result_string = "updated";
		break;
	case BTRFS_COMPARE_TREE_SAME:
		ASSERT(0);
		result_string = "unchanged";
		break;
	default:
		ASSERT(0);
		result_string = "unexpected";
	}

	btrfs_err(sctx->send_root->fs_info,
		  "Send: inconsistent snapshot, found %s %s for inode %llu without updated inode item, send root is %llu, parent root is %llu",
		  result_string, what, sctx->cmp_key->objectid,
		  sctx->send_root->root_key.objectid,
		  (sctx->parent_root ?
		   sctx->parent_root->root_key.objectid : 0));
}

305 306
static int is_waiting_for_move(struct send_ctx *sctx, u64 ino);

307 308 309 310 311
static struct waiting_dir_move *
get_waiting_dir_move(struct send_ctx *sctx, u64 ino);

static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino);

312 313 314 315 316 317 318
static int need_send_hole(struct send_ctx *sctx)
{
	return (sctx->parent_root && !sctx->cur_inode_new &&
		!sctx->cur_inode_new_gen && !sctx->cur_inode_deleted &&
		S_ISREG(sctx->cur_inode_mode));
}

319 320 321 322 323 324 325 326 327 328 329 330 331
static void fs_path_reset(struct fs_path *p)
{
	if (p->reversed) {
		p->start = p->buf + p->buf_len - 1;
		p->end = p->start;
		*p->start = 0;
	} else {
		p->start = p->buf;
		p->end = p->start;
		*p->start = 0;
	}
}

332
static struct fs_path *fs_path_alloc(void)
333 334 335
{
	struct fs_path *p;

336
	p = kmalloc(sizeof(*p), GFP_KERNEL);
337 338 339 340 341 342 343 344 345
	if (!p)
		return NULL;
	p->reversed = 0;
	p->buf = p->inline_buf;
	p->buf_len = FS_PATH_INLINE_SIZE;
	fs_path_reset(p);
	return p;
}

346
static struct fs_path *fs_path_alloc_reversed(void)
347 348 349
{
	struct fs_path *p;

350
	p = fs_path_alloc();
351 352 353 354 355 356 357
	if (!p)
		return NULL;
	p->reversed = 1;
	fs_path_reset(p);
	return p;
}

358
static void fs_path_free(struct fs_path *p)
359 360 361
{
	if (!p)
		return;
362 363
	if (p->buf != p->inline_buf)
		kfree(p->buf);
364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382
	kfree(p);
}

static int fs_path_len(struct fs_path *p)
{
	return p->end - p->start;
}

static int fs_path_ensure_buf(struct fs_path *p, int len)
{
	char *tmp_buf;
	int path_len;
	int old_buf_len;

	len++;

	if (p->buf_len >= len)
		return 0;

383 384 385 386 387
	if (len > PATH_MAX) {
		WARN_ON(1);
		return -ENOMEM;
	}

388 389 390
	path_len = p->end - p->start;
	old_buf_len = p->buf_len;

391 392 393
	/*
	 * First time the inline_buf does not suffice
	 */
394
	if (p->buf == p->inline_buf) {
395
		tmp_buf = kmalloc(len, GFP_KERNEL);
396 397 398
		if (tmp_buf)
			memcpy(tmp_buf, p->buf, old_buf_len);
	} else {
399
		tmp_buf = krealloc(p->buf, len, GFP_KERNEL);
400
	}
401 402 403 404 405 406 407 408
	if (!tmp_buf)
		return -ENOMEM;
	p->buf = tmp_buf;
	/*
	 * The real size of the buffer is bigger, this will let the fast path
	 * happen most of the time
	 */
	p->buf_len = ksize(p->buf);
409

410 411 412 413 414 415 416 417 418 419 420 421
	if (p->reversed) {
		tmp_buf = p->buf + old_buf_len - path_len - 1;
		p->end = p->buf + p->buf_len - 1;
		p->start = p->end - path_len;
		memmove(p->start, tmp_buf, path_len + 1);
	} else {
		p->start = p->buf;
		p->end = p->start + path_len;
	}
	return 0;
}

422 423
static int fs_path_prepare_for_add(struct fs_path *p, int name_len,
				   char **prepared)
424 425 426 427 428 429 430 431 432 433 434 435 436 437 438
{
	int ret;
	int new_len;

	new_len = p->end - p->start + name_len;
	if (p->start != p->end)
		new_len++;
	ret = fs_path_ensure_buf(p, new_len);
	if (ret < 0)
		goto out;

	if (p->reversed) {
		if (p->start != p->end)
			*--p->start = '/';
		p->start -= name_len;
439
		*prepared = p->start;
440 441 442
	} else {
		if (p->start != p->end)
			*p->end++ = '/';
443
		*prepared = p->end;
444 445 446 447 448 449 450 451 452 453 454
		p->end += name_len;
		*p->end = 0;
	}

out:
	return ret;
}

static int fs_path_add(struct fs_path *p, const char *name, int name_len)
{
	int ret;
455
	char *prepared;
456

457
	ret = fs_path_prepare_for_add(p, name_len, &prepared);
458 459
	if (ret < 0)
		goto out;
460
	memcpy(prepared, name, name_len);
461 462 463 464 465 466 467 468

out:
	return ret;
}

static int fs_path_add_path(struct fs_path *p, struct fs_path *p2)
{
	int ret;
469
	char *prepared;
470

471
	ret = fs_path_prepare_for_add(p, p2->end - p2->start, &prepared);
472 473
	if (ret < 0)
		goto out;
474
	memcpy(prepared, p2->start, p2->end - p2->start);
475 476 477 478 479 480 481 482 483 484

out:
	return ret;
}

static int fs_path_add_from_extent_buffer(struct fs_path *p,
					  struct extent_buffer *eb,
					  unsigned long off, int len)
{
	int ret;
485
	char *prepared;
486

487
	ret = fs_path_prepare_for_add(p, len, &prepared);
488 489 490
	if (ret < 0)
		goto out;

491
	read_extent_buffer(eb, prepared, off, len);
492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534

out:
	return ret;
}

static int fs_path_copy(struct fs_path *p, struct fs_path *from)
{
	int ret;

	p->reversed = from->reversed;
	fs_path_reset(p);

	ret = fs_path_add_path(p, from);

	return ret;
}


static void fs_path_unreverse(struct fs_path *p)
{
	char *tmp;
	int len;

	if (!p->reversed)
		return;

	tmp = p->start;
	len = p->end - p->start;
	p->start = p->buf;
	p->end = p->start + len;
	memmove(p->start, tmp, len + 1);
	p->reversed = 0;
}

static struct btrfs_path *alloc_path_for_send(void)
{
	struct btrfs_path *path;

	path = btrfs_alloc_path();
	if (!path)
		return NULL;
	path->search_commit_root = 1;
	path->skip_locking = 1;
535
	path->need_commit_sem = 1;
536 537 538
	return path;
}

539
static int write_buf(struct file *filp, const void *buf, u32 len, loff_t *off)
540 541 542 543 544 545 546 547 548
{
	int ret;
	mm_segment_t old_fs;
	u32 pos = 0;

	old_fs = get_fs();
	set_fs(KERNEL_DS);

	while (pos < len) {
F
Fabian Frederick 已提交
549 550
		ret = vfs_write(filp, (__force const char __user *)buf + pos,
				len - pos, off);
551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588
		/* TODO handle that correctly */
		/*if (ret == -ERESTARTSYS) {
			continue;
		}*/
		if (ret < 0)
			goto out;
		if (ret == 0) {
			ret = -EIO;
			goto out;
		}
		pos += ret;
	}

	ret = 0;

out:
	set_fs(old_fs);
	return ret;
}

static int tlv_put(struct send_ctx *sctx, u16 attr, const void *data, int len)
{
	struct btrfs_tlv_header *hdr;
	int total_len = sizeof(*hdr) + len;
	int left = sctx->send_max_size - sctx->send_size;

	if (unlikely(left < total_len))
		return -EOVERFLOW;

	hdr = (struct btrfs_tlv_header *) (sctx->send_buf + sctx->send_size);
	hdr->tlv_type = cpu_to_le16(attr);
	hdr->tlv_len = cpu_to_le16(len);
	memcpy(hdr + 1, data, len);
	sctx->send_size += total_len;

	return 0;
}

D
David Sterba 已提交
589 590 591 592 593 594 595
#define TLV_PUT_DEFINE_INT(bits) \
	static int tlv_put_u##bits(struct send_ctx *sctx,	 	\
			u##bits attr, u##bits value)			\
	{								\
		__le##bits __tmp = cpu_to_le##bits(value);		\
		return tlv_put(sctx, attr, &__tmp, sizeof(__tmp));	\
	}
596

D
David Sterba 已提交
597
TLV_PUT_DEFINE_INT(64)
598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673

static int tlv_put_string(struct send_ctx *sctx, u16 attr,
			  const char *str, int len)
{
	if (len == -1)
		len = strlen(str);
	return tlv_put(sctx, attr, str, len);
}

static int tlv_put_uuid(struct send_ctx *sctx, u16 attr,
			const u8 *uuid)
{
	return tlv_put(sctx, attr, uuid, BTRFS_UUID_SIZE);
}

static int tlv_put_btrfs_timespec(struct send_ctx *sctx, u16 attr,
				  struct extent_buffer *eb,
				  struct btrfs_timespec *ts)
{
	struct btrfs_timespec bts;
	read_extent_buffer(eb, &bts, (unsigned long)ts, sizeof(bts));
	return tlv_put(sctx, attr, &bts, sizeof(bts));
}


#define TLV_PUT(sctx, attrtype, attrlen, data) \
	do { \
		ret = tlv_put(sctx, attrtype, attrlen, data); \
		if (ret < 0) \
			goto tlv_put_failure; \
	} while (0)

#define TLV_PUT_INT(sctx, attrtype, bits, value) \
	do { \
		ret = tlv_put_u##bits(sctx, attrtype, value); \
		if (ret < 0) \
			goto tlv_put_failure; \
	} while (0)

#define TLV_PUT_U8(sctx, attrtype, data) TLV_PUT_INT(sctx, attrtype, 8, data)
#define TLV_PUT_U16(sctx, attrtype, data) TLV_PUT_INT(sctx, attrtype, 16, data)
#define TLV_PUT_U32(sctx, attrtype, data) TLV_PUT_INT(sctx, attrtype, 32, data)
#define TLV_PUT_U64(sctx, attrtype, data) TLV_PUT_INT(sctx, attrtype, 64, data)
#define TLV_PUT_STRING(sctx, attrtype, str, len) \
	do { \
		ret = tlv_put_string(sctx, attrtype, str, len); \
		if (ret < 0) \
			goto tlv_put_failure; \
	} while (0)
#define TLV_PUT_PATH(sctx, attrtype, p) \
	do { \
		ret = tlv_put_string(sctx, attrtype, p->start, \
			p->end - p->start); \
		if (ret < 0) \
			goto tlv_put_failure; \
	} while(0)
#define TLV_PUT_UUID(sctx, attrtype, uuid) \
	do { \
		ret = tlv_put_uuid(sctx, attrtype, uuid); \
		if (ret < 0) \
			goto tlv_put_failure; \
	} while (0)
#define TLV_PUT_BTRFS_TIMESPEC(sctx, attrtype, eb, ts) \
	do { \
		ret = tlv_put_btrfs_timespec(sctx, attrtype, eb, ts); \
		if (ret < 0) \
			goto tlv_put_failure; \
	} while (0)

static int send_header(struct send_ctx *sctx)
{
	struct btrfs_stream_header hdr;

	strcpy(hdr.magic, BTRFS_SEND_STREAM_MAGIC);
	hdr.version = cpu_to_le32(BTRFS_SEND_STREAM_VERSION);

674 675
	return write_buf(sctx->send_filp, &hdr, sizeof(hdr),
					&sctx->send_off);
676 677 678 679 680 681 682 683 684
}

/*
 * For each command/item we want to send to userspace, we call this function.
 */
static int begin_cmd(struct send_ctx *sctx, int cmd)
{
	struct btrfs_cmd_header *hdr;

685
	if (WARN_ON(!sctx->send_buf))
686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706
		return -EINVAL;

	BUG_ON(sctx->send_size);

	sctx->send_size += sizeof(*hdr);
	hdr = (struct btrfs_cmd_header *)sctx->send_buf;
	hdr->cmd = cpu_to_le16(cmd);

	return 0;
}

static int send_cmd(struct send_ctx *sctx)
{
	int ret;
	struct btrfs_cmd_header *hdr;
	u32 crc;

	hdr = (struct btrfs_cmd_header *)sctx->send_buf;
	hdr->len = cpu_to_le32(sctx->send_size - sizeof(*hdr));
	hdr->crc = 0;

707
	crc = btrfs_crc32c(0, (unsigned char *)sctx->send_buf, sctx->send_size);
708 709
	hdr->crc = cpu_to_le32(crc);

710 711
	ret = write_buf(sctx->send_filp, sctx->send_buf, sctx->send_size,
					&sctx->send_off);
712 713 714 715 716 717 718 719 720 721 722 723 724 725

	sctx->total_send_size += sctx->send_size;
	sctx->cmd_send_size[le16_to_cpu(hdr->cmd)] += sctx->send_size;
	sctx->send_size = 0;

	return ret;
}

/*
 * Sends a move instruction to user space
 */
static int send_rename(struct send_ctx *sctx,
		     struct fs_path *from, struct fs_path *to)
{
726
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
727 728
	int ret;

729
	btrfs_debug(fs_info, "send_rename %s -> %s", from->start, to->start);
730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750

	ret = begin_cmd(sctx, BTRFS_SEND_C_RENAME);
	if (ret < 0)
		goto out;

	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, from);
	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH_TO, to);

	ret = send_cmd(sctx);

tlv_put_failure:
out:
	return ret;
}

/*
 * Sends a link instruction to user space
 */
static int send_link(struct send_ctx *sctx,
		     struct fs_path *path, struct fs_path *lnk)
{
751
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
752 753
	int ret;

754
	btrfs_debug(fs_info, "send_link %s -> %s", path->start, lnk->start);
755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774

	ret = begin_cmd(sctx, BTRFS_SEND_C_LINK);
	if (ret < 0)
		goto out;

	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path);
	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH_LINK, lnk);

	ret = send_cmd(sctx);

tlv_put_failure:
out:
	return ret;
}

/*
 * Sends an unlink instruction to user space
 */
static int send_unlink(struct send_ctx *sctx, struct fs_path *path)
{
775
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
776 777
	int ret;

778
	btrfs_debug(fs_info, "send_unlink %s", path->start);
779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797

	ret = begin_cmd(sctx, BTRFS_SEND_C_UNLINK);
	if (ret < 0)
		goto out;

	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path);

	ret = send_cmd(sctx);

tlv_put_failure:
out:
	return ret;
}

/*
 * Sends a rmdir instruction to user space
 */
static int send_rmdir(struct send_ctx *sctx, struct fs_path *path)
{
798
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
799 800
	int ret;

801
	btrfs_debug(fs_info, "send_rmdir %s", path->start);
802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818

	ret = begin_cmd(sctx, BTRFS_SEND_C_RMDIR);
	if (ret < 0)
		goto out;

	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path);

	ret = send_cmd(sctx);

tlv_put_failure:
out:
	return ret;
}

/*
 * Helper function to retrieve some fields from an inode item.
 */
819 820 821
static int __get_inode_info(struct btrfs_root *root, struct btrfs_path *path,
			  u64 ino, u64 *size, u64 *gen, u64 *mode, u64 *uid,
			  u64 *gid, u64 *rdev)
822 823 824 825 826 827 828 829 830 831
{
	int ret;
	struct btrfs_inode_item *ii;
	struct btrfs_key key;

	key.objectid = ino;
	key.type = BTRFS_INODE_ITEM_KEY;
	key.offset = 0;
	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret) {
832 833 834
		if (ret > 0)
			ret = -ENOENT;
		return ret;
835 836 837 838 839 840 841 842 843 844 845 846 847 848
	}

	ii = btrfs_item_ptr(path->nodes[0], path->slots[0],
			struct btrfs_inode_item);
	if (size)
		*size = btrfs_inode_size(path->nodes[0], ii);
	if (gen)
		*gen = btrfs_inode_generation(path->nodes[0], ii);
	if (mode)
		*mode = btrfs_inode_mode(path->nodes[0], ii);
	if (uid)
		*uid = btrfs_inode_uid(path->nodes[0], ii);
	if (gid)
		*gid = btrfs_inode_gid(path->nodes[0], ii);
849 850
	if (rdev)
		*rdev = btrfs_inode_rdev(path->nodes[0], ii);
851

852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867
	return ret;
}

static int get_inode_info(struct btrfs_root *root,
			  u64 ino, u64 *size, u64 *gen,
			  u64 *mode, u64 *uid, u64 *gid,
			  u64 *rdev)
{
	struct btrfs_path *path;
	int ret;

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;
	ret = __get_inode_info(root, path, ino, size, gen, mode, uid, gid,
			       rdev);
868 869 870 871 872 873 874 875 876
	btrfs_free_path(path);
	return ret;
}

typedef int (*iterate_inode_ref_t)(int num, u64 dir, int index,
				   struct fs_path *p,
				   void *ctx);

/*
877 878
 * Helper function to iterate the entries in ONE btrfs_inode_ref or
 * btrfs_inode_extref.
879 880 881
 * The iterate callback may return a non zero value to stop iteration. This can
 * be a negative value for error codes or 1 to simply stop it.
 *
882
 * path must point to the INODE_REF or INODE_EXTREF when called.
883
 */
884
static int iterate_inode_ref(struct btrfs_root *root, struct btrfs_path *path,
885 886 887
			     struct btrfs_key *found_key, int resolve,
			     iterate_inode_ref_t iterate, void *ctx)
{
888
	struct extent_buffer *eb = path->nodes[0];
889 890
	struct btrfs_item *item;
	struct btrfs_inode_ref *iref;
891
	struct btrfs_inode_extref *extref;
892 893
	struct btrfs_path *tmp_path;
	struct fs_path *p;
894
	u32 cur = 0;
895
	u32 total;
896
	int slot = path->slots[0];
897 898 899
	u32 name_len;
	char *start;
	int ret = 0;
900
	int num = 0;
901
	int index;
902 903 904 905
	u64 dir;
	unsigned long name_off;
	unsigned long elem_size;
	unsigned long ptr;
906

907
	p = fs_path_alloc_reversed();
908 909 910 911 912
	if (!p)
		return -ENOMEM;

	tmp_path = alloc_path_for_send();
	if (!tmp_path) {
913
		fs_path_free(p);
914 915 916 917
		return -ENOMEM;
	}


918 919 920
	if (found_key->type == BTRFS_INODE_REF_KEY) {
		ptr = (unsigned long)btrfs_item_ptr(eb, slot,
						    struct btrfs_inode_ref);
921
		item = btrfs_item_nr(slot);
922 923 924 925 926 927 928 929
		total = btrfs_item_size(eb, item);
		elem_size = sizeof(*iref);
	} else {
		ptr = btrfs_item_ptr_offset(eb, slot);
		total = btrfs_item_size_nr(eb, slot);
		elem_size = sizeof(*extref);
	}

930 931 932
	while (cur < total) {
		fs_path_reset(p);

933 934 935 936 937 938 939 940 941 942 943 944 945 946
		if (found_key->type == BTRFS_INODE_REF_KEY) {
			iref = (struct btrfs_inode_ref *)(ptr + cur);
			name_len = btrfs_inode_ref_name_len(eb, iref);
			name_off = (unsigned long)(iref + 1);
			index = btrfs_inode_ref_index(eb, iref);
			dir = found_key->offset;
		} else {
			extref = (struct btrfs_inode_extref *)(ptr + cur);
			name_len = btrfs_inode_extref_name_len(eb, extref);
			name_off = (unsigned long)&extref->name;
			index = btrfs_inode_extref_index(eb, extref);
			dir = btrfs_inode_extref_parent(eb, extref);
		}

947
		if (resolve) {
948 949 950
			start = btrfs_ref_to_path(root, tmp_path, name_len,
						  name_off, eb, dir,
						  p->buf, p->buf_len);
951 952 953 954 955 956 957 958 959 960
			if (IS_ERR(start)) {
				ret = PTR_ERR(start);
				goto out;
			}
			if (start < p->buf) {
				/* overflow , try again with larger buffer */
				ret = fs_path_ensure_buf(p,
						p->buf_len + p->buf - start);
				if (ret < 0)
					goto out;
961 962 963 964
				start = btrfs_ref_to_path(root, tmp_path,
							  name_len, name_off,
							  eb, dir,
							  p->buf, p->buf_len);
965 966 967 968 969 970 971 972
				if (IS_ERR(start)) {
					ret = PTR_ERR(start);
					goto out;
				}
				BUG_ON(start < p->buf);
			}
			p->start = start;
		} else {
973 974
			ret = fs_path_add_from_extent_buffer(p, eb, name_off,
							     name_len);
975 976 977 978
			if (ret < 0)
				goto out;
		}

979 980
		cur += elem_size + name_len;
		ret = iterate(num, dir, index, p, ctx);
981 982 983 984 985 986 987
		if (ret)
			goto out;
		num++;
	}

out:
	btrfs_free_path(tmp_path);
988
	fs_path_free(p);
989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003
	return ret;
}

typedef int (*iterate_dir_item_t)(int num, struct btrfs_key *di_key,
				  const char *name, int name_len,
				  const char *data, int data_len,
				  u8 type, void *ctx);

/*
 * Helper function to iterate the entries in ONE btrfs_dir_item.
 * The iterate callback may return a non zero value to stop iteration. This can
 * be a negative value for error codes or 1 to simply stop it.
 *
 * path must point to the dir item when called.
 */
1004
static int iterate_dir_item(struct btrfs_root *root, struct btrfs_path *path,
1005 1006 1007 1008 1009 1010 1011 1012 1013
			    struct btrfs_key *found_key,
			    iterate_dir_item_t iterate, void *ctx)
{
	int ret = 0;
	struct extent_buffer *eb;
	struct btrfs_item *item;
	struct btrfs_dir_item *di;
	struct btrfs_key di_key;
	char *buf = NULL;
1014
	int buf_len;
1015 1016 1017 1018 1019 1020 1021 1022 1023
	u32 name_len;
	u32 data_len;
	u32 cur;
	u32 len;
	u32 total;
	int slot;
	int num;
	u8 type;

1024 1025 1026 1027 1028 1029 1030
	/*
	 * Start with a small buffer (1 page). If later we end up needing more
	 * space, which can happen for xattrs on a fs with a leaf size greater
	 * then the page size, attempt to increase the buffer. Typically xattr
	 * values are small.
	 */
	buf_len = PATH_MAX;
1031
	buf = kmalloc(buf_len, GFP_KERNEL);
1032 1033 1034 1035 1036 1037 1038
	if (!buf) {
		ret = -ENOMEM;
		goto out;
	}

	eb = path->nodes[0];
	slot = path->slots[0];
1039
	item = btrfs_item_nr(slot);
1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051
	di = btrfs_item_ptr(eb, slot, struct btrfs_dir_item);
	cur = 0;
	len = 0;
	total = btrfs_item_size(eb, item);

	num = 0;
	while (cur < total) {
		name_len = btrfs_dir_name_len(eb, di);
		data_len = btrfs_dir_data_len(eb, di);
		type = btrfs_dir_type(eb, di);
		btrfs_dir_item_key_to_cpu(eb, di, &di_key);

1052 1053 1054 1055 1056
		if (type == BTRFS_FT_XATTR) {
			if (name_len > XATTR_NAME_MAX) {
				ret = -ENAMETOOLONG;
				goto out;
			}
1057 1058
			if (name_len + data_len >
					BTRFS_MAX_XATTR_SIZE(root->fs_info)) {
1059 1060 1061 1062 1063 1064 1065
				ret = -E2BIG;
				goto out;
			}
		} else {
			/*
			 * Path too long
			 */
1066
			if (name_len + data_len > PATH_MAX) {
1067 1068 1069
				ret = -ENAMETOOLONG;
				goto out;
			}
1070 1071
		}

1072 1073 1074 1075 1076 1077 1078
		if (name_len + data_len > buf_len) {
			buf_len = name_len + data_len;
			if (is_vmalloc_addr(buf)) {
				vfree(buf);
				buf = NULL;
			} else {
				char *tmp = krealloc(buf, buf_len,
1079
						GFP_KERNEL | __GFP_NOWARN);
1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093

				if (!tmp)
					kfree(buf);
				buf = tmp;
			}
			if (!buf) {
				buf = vmalloc(buf_len);
				if (!buf) {
					ret = -ENOMEM;
					goto out;
				}
			}
		}

1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113
		read_extent_buffer(eb, buf, (unsigned long)(di + 1),
				name_len + data_len);

		len = sizeof(*di) + name_len + data_len;
		di = (struct btrfs_dir_item *)((char *)di + len);
		cur += len;

		ret = iterate(num, &di_key, buf, name_len, buf + name_len,
				data_len, type, ctx);
		if (ret < 0)
			goto out;
		if (ret) {
			ret = 0;
			goto out;
		}

		num++;
	}

out:
1114
	kvfree(buf);
1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135
	return ret;
}

static int __copy_first_ref(int num, u64 dir, int index,
			    struct fs_path *p, void *ctx)
{
	int ret;
	struct fs_path *pt = ctx;

	ret = fs_path_copy(pt, p);
	if (ret < 0)
		return ret;

	/* we want the first only */
	return 1;
}

/*
 * Retrieve the first path of an inode. If an inode has more then one
 * ref/hardlink, this is ignored.
 */
1136
static int get_inode_path(struct btrfs_root *root,
1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161
			  u64 ino, struct fs_path *path)
{
	int ret;
	struct btrfs_key key, found_key;
	struct btrfs_path *p;

	p = alloc_path_for_send();
	if (!p)
		return -ENOMEM;

	fs_path_reset(path);

	key.objectid = ino;
	key.type = BTRFS_INODE_REF_KEY;
	key.offset = 0;

	ret = btrfs_search_slot_for_read(root, &key, p, 1, 0);
	if (ret < 0)
		goto out;
	if (ret) {
		ret = 1;
		goto out;
	}
	btrfs_item_key_to_cpu(p->nodes[0], &found_key, p->slots[0]);
	if (found_key.objectid != ino ||
1162 1163
	    (found_key.type != BTRFS_INODE_REF_KEY &&
	     found_key.type != BTRFS_INODE_EXTREF_KEY)) {
1164 1165 1166 1167
		ret = -ENOENT;
		goto out;
	}

1168 1169
	ret = iterate_inode_ref(root, p, &found_key, 1,
				__copy_first_ref, path);
1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181
	if (ret < 0)
		goto out;
	ret = 0;

out:
	btrfs_free_path(p);
	return ret;
}

struct backref_ctx {
	struct send_ctx *sctx;

1182
	struct btrfs_path *path;
1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195
	/* number of total found references */
	u64 found;

	/*
	 * used for clones found in send_root. clones found behind cur_objectid
	 * and cur_offset are not considered as allowed clones.
	 */
	u64 cur_objectid;
	u64 cur_offset;

	/* may be truncated in case it's the last extent in a file */
	u64 extent_len;

1196 1197 1198
	/* data offset in the file extent item */
	u64 data_offset;

1199
	/* Just to check for bugs in backref resolving */
1200
	int found_itself;
1201 1202 1203 1204
};

static int __clone_root_cmp_bsearch(const void *key, const void *elt)
{
1205
	u64 root = (u64)(uintptr_t)key;
1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228
	struct clone_root *cr = (struct clone_root *)elt;

	if (root < cr->root->objectid)
		return -1;
	if (root > cr->root->objectid)
		return 1;
	return 0;
}

static int __clone_root_cmp_sort(const void *e1, const void *e2)
{
	struct clone_root *cr1 = (struct clone_root *)e1;
	struct clone_root *cr2 = (struct clone_root *)e2;

	if (cr1->root->objectid < cr2->root->objectid)
		return -1;
	if (cr1->root->objectid > cr2->root->objectid)
		return 1;
	return 0;
}

/*
 * Called for every backref that is found for the current extent.
1229
 * Results are collected in sctx->clone_roots->ino/offset/found_refs
1230 1231 1232 1233 1234 1235 1236 1237 1238
 */
static int __iterate_backrefs(u64 ino, u64 offset, u64 root, void *ctx_)
{
	struct backref_ctx *bctx = ctx_;
	struct clone_root *found;
	int ret;
	u64 i_size;

	/* First check if the root is in the list of accepted clone sources */
1239
	found = bsearch((void *)(uintptr_t)root, bctx->sctx->clone_roots,
1240 1241 1242 1243 1244 1245 1246 1247 1248
			bctx->sctx->clone_roots_cnt,
			sizeof(struct clone_root),
			__clone_root_cmp_bsearch);
	if (!found)
		return 0;

	if (found->root == bctx->sctx->send_root &&
	    ino == bctx->cur_objectid &&
	    offset == bctx->cur_offset) {
1249
		bctx->found_itself = 1;
1250 1251 1252
	}

	/*
1253
	 * There are inodes that have extents that lie behind its i_size. Don't
1254 1255
	 * accept clones from these extents.
	 */
1256 1257 1258
	ret = __get_inode_info(found->root, bctx->path, ino, &i_size, NULL, NULL,
			       NULL, NULL, NULL);
	btrfs_release_path(bctx->path);
1259 1260 1261
	if (ret < 0)
		return ret;

1262
	if (offset + bctx->data_offset + bctx->extent_len > i_size)
1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277
		return 0;

	/*
	 * Make sure we don't consider clones from send_root that are
	 * behind the current inode/offset.
	 */
	if (found->root == bctx->sctx->send_root) {
		/*
		 * TODO for the moment we don't accept clones from the inode
		 * that is currently send. We may change this when
		 * BTRFS_IOC_CLONE_RANGE supports cloning from and to the same
		 * file.
		 */
		if (ino >= bctx->cur_objectid)
			return 0;
1278 1279 1280 1281
#if 0
		if (ino > bctx->cur_objectid)
			return 0;
		if (offset + bctx->extent_len > bctx->cur_offset)
1282
			return 0;
1283
#endif
1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302
	}

	bctx->found++;
	found->found_refs++;
	if (ino < found->ino) {
		found->ino = ino;
		found->offset = offset;
	} else if (found->ino == ino) {
		/*
		 * same extent found more then once in the same file.
		 */
		if (found->offset > offset + bctx->extent_len)
			found->offset = offset;
	}

	return 0;
}

/*
1303 1304 1305 1306 1307 1308
 * Given an inode, offset and extent item, it finds a good clone for a clone
 * instruction. Returns -ENOENT when none could be found. The function makes
 * sure that the returned clone is usable at the point where sending is at the
 * moment. This means, that no clones are accepted which lie behind the current
 * inode+offset.
 *
1309 1310 1311 1312 1313 1314 1315 1316
 * path must point to the extent item when called.
 */
static int find_extent_clone(struct send_ctx *sctx,
			     struct btrfs_path *path,
			     u64 ino, u64 data_offset,
			     u64 ino_size,
			     struct clone_root **found)
{
1317
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
1318 1319 1320
	int ret;
	int extent_type;
	u64 logical;
1321
	u64 disk_byte;
1322 1323
	u64 num_bytes;
	u64 extent_item_pos;
1324
	u64 flags = 0;
1325 1326
	struct btrfs_file_extent_item *fi;
	struct extent_buffer *eb = path->nodes[0];
1327
	struct backref_ctx *backref_ctx = NULL;
1328 1329 1330
	struct clone_root *cur_clone_root;
	struct btrfs_key found_key;
	struct btrfs_path *tmp_path;
1331
	int compressed;
1332 1333 1334 1335 1336 1337
	u32 i;

	tmp_path = alloc_path_for_send();
	if (!tmp_path)
		return -ENOMEM;

1338 1339 1340
	/* We only use this path under the commit sem */
	tmp_path->need_commit_sem = 0;

1341
	backref_ctx = kmalloc(sizeof(*backref_ctx), GFP_KERNEL);
1342 1343 1344 1345 1346
	if (!backref_ctx) {
		ret = -ENOMEM;
		goto out;
	}

1347 1348
	backref_ctx->path = tmp_path;

1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365
	if (data_offset >= ino_size) {
		/*
		 * There may be extents that lie behind the file's size.
		 * I at least had this in combination with snapshotting while
		 * writing large files.
		 */
		ret = 0;
		goto out;
	}

	fi = btrfs_item_ptr(eb, path->slots[0],
			struct btrfs_file_extent_item);
	extent_type = btrfs_file_extent_type(eb, fi);
	if (extent_type == BTRFS_FILE_EXTENT_INLINE) {
		ret = -ENOENT;
		goto out;
	}
1366
	compressed = btrfs_file_extent_compression(eb, fi);
1367 1368

	num_bytes = btrfs_file_extent_num_bytes(eb, fi);
1369 1370
	disk_byte = btrfs_file_extent_disk_bytenr(eb, fi);
	if (disk_byte == 0) {
1371 1372 1373
		ret = -ENOENT;
		goto out;
	}
1374
	logical = disk_byte + btrfs_file_extent_offset(eb, fi);
1375

1376 1377
	down_read(&fs_info->commit_root_sem);
	ret = extent_from_logical(fs_info, disk_byte, tmp_path,
1378
				  &found_key, &flags);
1379
	up_read(&fs_info->commit_root_sem);
1380 1381 1382 1383
	btrfs_release_path(tmp_path);

	if (ret < 0)
		goto out;
1384
	if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398
		ret = -EIO;
		goto out;
	}

	/*
	 * Setup the clone roots.
	 */
	for (i = 0; i < sctx->clone_roots_cnt; i++) {
		cur_clone_root = sctx->clone_roots + i;
		cur_clone_root->ino = (u64)-1;
		cur_clone_root->offset = 0;
		cur_clone_root->found_refs = 0;
	}

1399 1400 1401 1402 1403 1404
	backref_ctx->sctx = sctx;
	backref_ctx->found = 0;
	backref_ctx->cur_objectid = ino;
	backref_ctx->cur_offset = data_offset;
	backref_ctx->found_itself = 0;
	backref_ctx->extent_len = num_bytes;
1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417
	/*
	 * For non-compressed extents iterate_extent_inodes() gives us extent
	 * offsets that already take into account the data offset, but not for
	 * compressed extents, since the offset is logical and not relative to
	 * the physical extent locations. We must take this into account to
	 * avoid sending clone offsets that go beyond the source file's size,
	 * which would result in the clone ioctl failing with -EINVAL on the
	 * receiving end.
	 */
	if (compressed == BTRFS_COMPRESS_NONE)
		backref_ctx->data_offset = 0;
	else
		backref_ctx->data_offset = btrfs_file_extent_offset(eb, fi);
1418 1419 1420 1421 1422 1423 1424

	/*
	 * The last extent of a file may be too large due to page alignment.
	 * We need to adjust extent_len in this case so that the checks in
	 * __iterate_backrefs work.
	 */
	if (data_offset + num_bytes >= ino_size)
1425
		backref_ctx->extent_len = ino_size - data_offset;
1426 1427 1428 1429

	/*
	 * Now collect all backrefs.
	 */
1430 1431 1432 1433
	if (compressed == BTRFS_COMPRESS_NONE)
		extent_item_pos = logical - found_key.objectid;
	else
		extent_item_pos = 0;
1434 1435 1436
	ret = iterate_extent_inodes(fs_info, found_key.objectid,
				    extent_item_pos, 1, __iterate_backrefs,
				    backref_ctx);
1437

1438 1439 1440
	if (ret < 0)
		goto out;

1441
	if (!backref_ctx->found_itself) {
1442 1443
		/* found a bug in backref code? */
		ret = -EIO;
1444
		btrfs_err(fs_info,
J
Jeff Mahoney 已提交
1445
			  "did not find backref in send_root. inode=%llu, offset=%llu, disk_byte=%llu found extent=%llu",
1446
			  ino, data_offset, disk_byte, found_key.objectid);
1447 1448 1449
		goto out;
	}

1450 1451 1452
	btrfs_debug(fs_info,
		    "find_extent_clone: data_offset=%llu, ino=%llu, num_bytes=%llu, logical=%llu",
		    data_offset, ino, num_bytes, logical);
1453

1454
	if (!backref_ctx->found)
1455
		btrfs_debug(fs_info, "no clones found");
1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477

	cur_clone_root = NULL;
	for (i = 0; i < sctx->clone_roots_cnt; i++) {
		if (sctx->clone_roots[i].found_refs) {
			if (!cur_clone_root)
				cur_clone_root = sctx->clone_roots + i;
			else if (sctx->clone_roots[i].root == sctx->send_root)
				/* prefer clones from send_root over others */
				cur_clone_root = sctx->clone_roots + i;
		}

	}

	if (cur_clone_root) {
		*found = cur_clone_root;
		ret = 0;
	} else {
		ret = -ENOENT;
	}

out:
	btrfs_free_path(tmp_path);
1478
	kfree(backref_ctx);
1479 1480 1481
	return ret;
}

1482
static int read_symlink(struct btrfs_root *root,
1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504
			u64 ino,
			struct fs_path *dest)
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct btrfs_file_extent_item *ei;
	u8 type;
	u8 compression;
	unsigned long off;
	int len;

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	key.objectid = ino;
	key.type = BTRFS_EXTENT_DATA_KEY;
	key.offset = 0;
	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto out;
1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519
	if (ret) {
		/*
		 * An empty symlink inode. Can happen in rare error paths when
		 * creating a symlink (transaction committed before the inode
		 * eviction handler removed the symlink inode items and a crash
		 * happened in between or the subvol was snapshoted in between).
		 * Print an informative message to dmesg/syslog so that the user
		 * can delete the symlink.
		 */
		btrfs_err(root->fs_info,
			  "Found empty symlink inode %llu at root %llu",
			  ino, root->root_key.objectid);
		ret = -EIO;
		goto out;
	}
1520 1521 1522 1523 1524 1525 1526 1527 1528

	ei = btrfs_item_ptr(path->nodes[0], path->slots[0],
			struct btrfs_file_extent_item);
	type = btrfs_file_extent_type(path->nodes[0], ei);
	compression = btrfs_file_extent_compression(path->nodes[0], ei);
	BUG_ON(type != BTRFS_FILE_EXTENT_INLINE);
	BUG_ON(compression);

	off = btrfs_file_extent_inline_start(ei);
1529
	len = btrfs_file_extent_inline_len(path->nodes[0], path->slots[0], ei);
1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557

	ret = fs_path_add_from_extent_buffer(dest, path->nodes[0], off, len);

out:
	btrfs_free_path(path);
	return ret;
}

/*
 * Helper function to generate a file name that is unique in the root of
 * send_root and parent_root. This is used to generate names for orphan inodes.
 */
static int gen_unique_name(struct send_ctx *sctx,
			   u64 ino, u64 gen,
			   struct fs_path *dest)
{
	int ret = 0;
	struct btrfs_path *path;
	struct btrfs_dir_item *di;
	char tmp[64];
	int len;
	u64 idx = 0;

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	while (1) {
1558
		len = snprintf(tmp, sizeof(tmp), "o%llu-%llu-%llu",
1559
				ino, gen, idx);
1560
		ASSERT(len < sizeof(tmp));
1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622

		di = btrfs_lookup_dir_item(NULL, sctx->send_root,
				path, BTRFS_FIRST_FREE_OBJECTID,
				tmp, strlen(tmp), 0);
		btrfs_release_path(path);
		if (IS_ERR(di)) {
			ret = PTR_ERR(di);
			goto out;
		}
		if (di) {
			/* not unique, try again */
			idx++;
			continue;
		}

		if (!sctx->parent_root) {
			/* unique */
			ret = 0;
			break;
		}

		di = btrfs_lookup_dir_item(NULL, sctx->parent_root,
				path, BTRFS_FIRST_FREE_OBJECTID,
				tmp, strlen(tmp), 0);
		btrfs_release_path(path);
		if (IS_ERR(di)) {
			ret = PTR_ERR(di);
			goto out;
		}
		if (di) {
			/* not unique, try again */
			idx++;
			continue;
		}
		/* unique */
		break;
	}

	ret = fs_path_add(dest, tmp, strlen(tmp));

out:
	btrfs_free_path(path);
	return ret;
}

enum inode_state {
	inode_state_no_change,
	inode_state_will_create,
	inode_state_did_create,
	inode_state_will_delete,
	inode_state_did_delete,
};

static int get_cur_inode_state(struct send_ctx *sctx, u64 ino, u64 gen)
{
	int ret;
	int left_ret;
	int right_ret;
	u64 left_gen;
	u64 right_gen;

	ret = get_inode_info(sctx->send_root, ino, NULL, &left_gen, NULL, NULL,
1623
			NULL, NULL);
1624 1625 1626 1627 1628 1629 1630 1631
	if (ret < 0 && ret != -ENOENT)
		goto out;
	left_ret = ret;

	if (!sctx->parent_root) {
		right_ret = -ENOENT;
	} else {
		ret = get_inode_info(sctx->parent_root, ino, NULL, &right_gen,
1632
				NULL, NULL, NULL, NULL);
1633 1634 1635 1636 1637 1638
		if (ret < 0 && ret != -ENOENT)
			goto out;
		right_ret = ret;
	}

	if (!left_ret && !right_ret) {
1639
		if (left_gen == gen && right_gen == gen) {
1640
			ret = inode_state_no_change;
1641
		} else if (left_gen == gen) {
1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 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
			if (ino < sctx->send_progress)
				ret = inode_state_did_create;
			else
				ret = inode_state_will_create;
		} else if (right_gen == gen) {
			if (ino < sctx->send_progress)
				ret = inode_state_did_delete;
			else
				ret = inode_state_will_delete;
		} else  {
			ret = -ENOENT;
		}
	} else if (!left_ret) {
		if (left_gen == gen) {
			if (ino < sctx->send_progress)
				ret = inode_state_did_create;
			else
				ret = inode_state_will_create;
		} else {
			ret = -ENOENT;
		}
	} else if (!right_ret) {
		if (right_gen == gen) {
			if (ino < sctx->send_progress)
				ret = inode_state_did_delete;
			else
				ret = inode_state_will_delete;
		} else {
			ret = -ENOENT;
		}
	} else {
		ret = -ENOENT;
	}

out:
	return ret;
}

static int is_inode_existent(struct send_ctx *sctx, u64 ino, u64 gen)
{
	int ret;

	ret = get_cur_inode_state(sctx, ino, gen);
	if (ret < 0)
		goto out;

	if (ret == inode_state_no_change ||
	    ret == inode_state_did_create ||
	    ret == inode_state_will_delete)
		ret = 1;
	else
		ret = 0;

out:
	return ret;
}

/*
 * Helper function to lookup a dir item in a dir.
 */
static int lookup_dir_item_inode(struct btrfs_root *root,
				 u64 dir, const char *name, int name_len,
				 u64 *found_inode,
				 u8 *found_type)
{
	int ret = 0;
	struct btrfs_dir_item *di;
	struct btrfs_key key;
	struct btrfs_path *path;

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	di = btrfs_lookup_dir_item(NULL, root, path,
			dir, name, name_len, 0);
	if (!di) {
		ret = -ENOENT;
		goto out;
	}
	if (IS_ERR(di)) {
		ret = PTR_ERR(di);
		goto out;
	}
	btrfs_dir_item_key_to_cpu(path->nodes[0], di, &key);
1727 1728 1729 1730
	if (key.type == BTRFS_ROOT_ITEM_KEY) {
		ret = -ENOENT;
		goto out;
	}
1731 1732 1733 1734 1735 1736 1737 1738
	*found_inode = key.objectid;
	*found_type = btrfs_dir_type(path->nodes[0], di);

out:
	btrfs_free_path(path);
	return ret;
}

1739 1740 1741 1742
/*
 * Looks up the first btrfs_inode_ref of a given ino. It returns the parent dir,
 * generation of the parent dir and the name of the dir entry.
 */
1743
static int get_first_ref(struct btrfs_root *root, u64 ino,
1744 1745 1746 1747 1748 1749 1750
			 u64 *dir, u64 *dir_gen, struct fs_path *name)
{
	int ret;
	struct btrfs_key key;
	struct btrfs_key found_key;
	struct btrfs_path *path;
	int len;
1751
	u64 parent_dir;
1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	key.objectid = ino;
	key.type = BTRFS_INODE_REF_KEY;
	key.offset = 0;

	ret = btrfs_search_slot_for_read(root, &key, path, 1, 0);
	if (ret < 0)
		goto out;
	if (!ret)
		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
				path->slots[0]);
1767 1768 1769
	if (ret || found_key.objectid != ino ||
	    (found_key.type != BTRFS_INODE_REF_KEY &&
	     found_key.type != BTRFS_INODE_EXTREF_KEY)) {
1770 1771 1772 1773
		ret = -ENOENT;
		goto out;
	}

1774
	if (found_key.type == BTRFS_INODE_REF_KEY) {
1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
		struct btrfs_inode_ref *iref;
		iref = btrfs_item_ptr(path->nodes[0], path->slots[0],
				      struct btrfs_inode_ref);
		len = btrfs_inode_ref_name_len(path->nodes[0], iref);
		ret = fs_path_add_from_extent_buffer(name, path->nodes[0],
						     (unsigned long)(iref + 1),
						     len);
		parent_dir = found_key.offset;
	} else {
		struct btrfs_inode_extref *extref;
		extref = btrfs_item_ptr(path->nodes[0], path->slots[0],
					struct btrfs_inode_extref);
		len = btrfs_inode_extref_name_len(path->nodes[0], extref);
		ret = fs_path_add_from_extent_buffer(name, path->nodes[0],
					(unsigned long)&extref->name, len);
		parent_dir = btrfs_inode_extref_parent(path->nodes[0], extref);
	}
1792 1793 1794 1795
	if (ret < 0)
		goto out;
	btrfs_release_path(path);

1796 1797 1798 1799 1800 1801
	if (dir_gen) {
		ret = get_inode_info(root, parent_dir, NULL, dir_gen, NULL,
				     NULL, NULL, NULL);
		if (ret < 0)
			goto out;
	}
1802

1803
	*dir = parent_dir;
1804 1805 1806 1807 1808 1809

out:
	btrfs_free_path(path);
	return ret;
}

1810
static int is_first_ref(struct btrfs_root *root,
1811 1812 1813 1814 1815 1816 1817
			u64 ino, u64 dir,
			const char *name, int name_len)
{
	int ret;
	struct fs_path *tmp_name;
	u64 tmp_dir;

1818
	tmp_name = fs_path_alloc();
1819 1820 1821
	if (!tmp_name)
		return -ENOMEM;

1822
	ret = get_first_ref(root, ino, &tmp_dir, NULL, tmp_name);
1823 1824 1825
	if (ret < 0)
		goto out;

1826
	if (dir != tmp_dir || name_len != fs_path_len(tmp_name)) {
1827 1828 1829 1830
		ret = 0;
		goto out;
	}

1831
	ret = !memcmp(tmp_name->start, name, name_len);
1832 1833

out:
1834
	fs_path_free(tmp_name);
1835 1836 1837
	return ret;
}

1838 1839 1840 1841 1842 1843 1844 1845 1846 1847
/*
 * Used by process_recorded_refs to determine if a new ref would overwrite an
 * already existing ref. In case it detects an overwrite, it returns the
 * inode/gen in who_ino/who_gen.
 * When an overwrite is detected, process_recorded_refs does proper orphanizing
 * to make sure later references to the overwritten inode are possible.
 * Orphanizing is however only required for the first ref of an inode.
 * process_recorded_refs does an additional is_first_ref check to see if
 * orphanizing is really required.
 */
1848 1849 1850 1851 1852
static int will_overwrite_ref(struct send_ctx *sctx, u64 dir, u64 dir_gen,
			      const char *name, int name_len,
			      u64 *who_ino, u64 *who_gen)
{
	int ret = 0;
1853
	u64 gen;
1854 1855 1856 1857 1858 1859 1860 1861 1862 1863
	u64 other_inode = 0;
	u8 other_type = 0;

	if (!sctx->parent_root)
		goto out;

	ret = is_inode_existent(sctx, dir, dir_gen);
	if (ret <= 0)
		goto out;

1864 1865
	/*
	 * If we have a parent root we need to verify that the parent dir was
1866
	 * not deleted and then re-created, if it was then we have no overwrite
1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881
	 * and we can just unlink this entry.
	 */
	if (sctx->parent_root) {
		ret = get_inode_info(sctx->parent_root, dir, NULL, &gen, NULL,
				     NULL, NULL, NULL);
		if (ret < 0 && ret != -ENOENT)
			goto out;
		if (ret) {
			ret = 0;
			goto out;
		}
		if (gen != dir_gen)
			goto out;
	}

1882 1883 1884 1885 1886 1887 1888 1889 1890
	ret = lookup_dir_item_inode(sctx->parent_root, dir, name, name_len,
			&other_inode, &other_type);
	if (ret < 0 && ret != -ENOENT)
		goto out;
	if (ret) {
		ret = 0;
		goto out;
	}

1891 1892 1893 1894 1895
	/*
	 * Check if the overwritten ref was already processed. If yes, the ref
	 * was already unlinked/moved, so we can safely assume that we will not
	 * overwrite anything at this point in time.
	 */
1896 1897
	if (other_inode > sctx->send_progress ||
	    is_waiting_for_move(sctx, other_inode)) {
1898
		ret = get_inode_info(sctx->parent_root, other_inode, NULL,
1899
				who_gen, NULL, NULL, NULL, NULL);
1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912
		if (ret < 0)
			goto out;

		ret = 1;
		*who_ino = other_inode;
	} else {
		ret = 0;
	}

out:
	return ret;
}

1913 1914 1915 1916 1917 1918 1919
/*
 * Checks if the ref was overwritten by an already processed inode. This is
 * used by __get_cur_name_and_parent to find out if the ref was orphanized and
 * thus the orphan name needs be used.
 * process_recorded_refs also uses it to avoid unlinking of refs that were
 * overwritten.
 */
1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948
static int did_overwrite_ref(struct send_ctx *sctx,
			    u64 dir, u64 dir_gen,
			    u64 ino, u64 ino_gen,
			    const char *name, int name_len)
{
	int ret = 0;
	u64 gen;
	u64 ow_inode;
	u8 other_type;

	if (!sctx->parent_root)
		goto out;

	ret = is_inode_existent(sctx, dir, dir_gen);
	if (ret <= 0)
		goto out;

	/* check if the ref was overwritten by another ref */
	ret = lookup_dir_item_inode(sctx->send_root, dir, name, name_len,
			&ow_inode, &other_type);
	if (ret < 0 && ret != -ENOENT)
		goto out;
	if (ret) {
		/* was never and will never be overwritten */
		ret = 0;
		goto out;
	}

	ret = get_inode_info(sctx->send_root, ow_inode, NULL, &gen, NULL, NULL,
1949
			NULL, NULL);
1950 1951 1952 1953 1954 1955 1956 1957
	if (ret < 0)
		goto out;

	if (ow_inode == ino && gen == ino_gen) {
		ret = 0;
		goto out;
	}

1958 1959 1960
	/*
	 * We know that it is or will be overwritten. Check this now.
	 * The current inode being processed might have been the one that caused
1961 1962
	 * inode 'ino' to be orphanized, therefore check if ow_inode matches
	 * the current inode being processed.
1963
	 */
1964 1965 1966
	if ((ow_inode < sctx->send_progress) ||
	    (ino != sctx->cur_ino && ow_inode == sctx->cur_ino &&
	     gen == sctx->cur_inode_gen))
1967 1968 1969 1970 1971 1972 1973 1974
		ret = 1;
	else
		ret = 0;

out:
	return ret;
}

1975 1976 1977 1978 1979
/*
 * Same as did_overwrite_ref, but also checks if it is the first ref of an inode
 * that got overwritten. This is used by process_recorded_refs to determine
 * if it has to use the path as returned by get_cur_path or the orphan name.
 */
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989
static int did_overwrite_first_ref(struct send_ctx *sctx, u64 ino, u64 gen)
{
	int ret = 0;
	struct fs_path *name = NULL;
	u64 dir;
	u64 dir_gen;

	if (!sctx->parent_root)
		goto out;

1990
	name = fs_path_alloc();
1991 1992 1993
	if (!name)
		return -ENOMEM;

1994
	ret = get_first_ref(sctx->parent_root, ino, &dir, &dir_gen, name);
1995 1996 1997 1998 1999 2000 2001
	if (ret < 0)
		goto out;

	ret = did_overwrite_ref(sctx, dir, dir_gen, ino, gen,
			name->start, fs_path_len(name));

out:
2002
	fs_path_free(name);
2003 2004 2005
	return ret;
}

2006 2007 2008 2009
/*
 * Insert a name cache entry. On 32bit kernels the radix tree index is 32bit,
 * so we need to do some special handling in case we have clashes. This function
 * takes care of this with the help of name_cache_entry::radix_list.
2010
 * In case of error, nce is kfreed.
2011
 */
2012 2013 2014 2015
static int name_cache_insert(struct send_ctx *sctx,
			     struct name_cache_entry *nce)
{
	int ret = 0;
2016 2017 2018 2019 2020
	struct list_head *nce_head;

	nce_head = radix_tree_lookup(&sctx->name_cache,
			(unsigned long)nce->ino);
	if (!nce_head) {
2021
		nce_head = kmalloc(sizeof(*nce_head), GFP_KERNEL);
2022 2023
		if (!nce_head) {
			kfree(nce);
2024
			return -ENOMEM;
2025
		}
2026
		INIT_LIST_HEAD(nce_head);
2027

2028
		ret = radix_tree_insert(&sctx->name_cache, nce->ino, nce_head);
2029 2030 2031
		if (ret < 0) {
			kfree(nce_head);
			kfree(nce);
2032
			return ret;
2033
		}
2034
	}
2035
	list_add_tail(&nce->radix_list, nce_head);
2036 2037 2038 2039 2040 2041 2042 2043 2044
	list_add_tail(&nce->list, &sctx->name_cache_list);
	sctx->name_cache_size++;

	return ret;
}

static void name_cache_delete(struct send_ctx *sctx,
			      struct name_cache_entry *nce)
{
2045
	struct list_head *nce_head;
2046

2047 2048
	nce_head = radix_tree_lookup(&sctx->name_cache,
			(unsigned long)nce->ino);
2049 2050 2051 2052 2053
	if (!nce_head) {
		btrfs_err(sctx->send_root->fs_info,
	      "name_cache_delete lookup failed ino %llu cache size %d, leaking memory",
			nce->ino, sctx->name_cache_size);
	}
2054

2055
	list_del(&nce->radix_list);
2056 2057
	list_del(&nce->list);
	sctx->name_cache_size--;
2058

2059 2060 2061 2062
	/*
	 * We may not get to the final release of nce_head if the lookup fails
	 */
	if (nce_head && list_empty(nce_head)) {
2063 2064 2065
		radix_tree_delete(&sctx->name_cache, (unsigned long)nce->ino);
		kfree(nce_head);
	}
2066 2067 2068 2069 2070
}

static struct name_cache_entry *name_cache_search(struct send_ctx *sctx,
						    u64 ino, u64 gen)
{
2071 2072
	struct list_head *nce_head;
	struct name_cache_entry *cur;
2073

2074 2075
	nce_head = radix_tree_lookup(&sctx->name_cache, (unsigned long)ino);
	if (!nce_head)
2076 2077
		return NULL;

2078 2079 2080 2081
	list_for_each_entry(cur, nce_head, radix_list) {
		if (cur->ino == ino && cur->gen == gen)
			return cur;
	}
2082 2083 2084
	return NULL;
}

2085 2086 2087 2088
/*
 * Removes the entry from the list and adds it back to the end. This marks the
 * entry as recently used so that name_cache_clean_unused does not remove it.
 */
2089 2090 2091 2092 2093 2094
static void name_cache_used(struct send_ctx *sctx, struct name_cache_entry *nce)
{
	list_del(&nce->list);
	list_add_tail(&nce->list, &sctx->name_cache_list);
}

2095 2096 2097
/*
 * Remove some entries from the beginning of name_cache_list.
 */
2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116
static void name_cache_clean_unused(struct send_ctx *sctx)
{
	struct name_cache_entry *nce;

	if (sctx->name_cache_size < SEND_CTX_NAME_CACHE_CLEAN_SIZE)
		return;

	while (sctx->name_cache_size > SEND_CTX_MAX_NAME_CACHE_SIZE) {
		nce = list_entry(sctx->name_cache_list.next,
				struct name_cache_entry, list);
		name_cache_delete(sctx, nce);
		kfree(nce);
	}
}

static void name_cache_free(struct send_ctx *sctx)
{
	struct name_cache_entry *nce;

2117 2118 2119
	while (!list_empty(&sctx->name_cache_list)) {
		nce = list_entry(sctx->name_cache_list.next,
				struct name_cache_entry, list);
2120
		name_cache_delete(sctx, nce);
2121
		kfree(nce);
2122 2123 2124
	}
}

2125 2126 2127 2128 2129 2130 2131 2132
/*
 * Used by get_cur_path for each ref up to the root.
 * Returns 0 if it succeeded.
 * Returns 1 if the inode is not existent or got overwritten. In that case, the
 * name is an orphan name. This instructs get_cur_path to stop iterating. If 1
 * is returned, parent_ino/parent_gen are not guaranteed to be valid.
 * Returns <0 in case of error.
 */
2133 2134 2135 2136 2137 2138 2139 2140 2141 2142
static int __get_cur_name_and_parent(struct send_ctx *sctx,
				     u64 ino, u64 gen,
				     u64 *parent_ino,
				     u64 *parent_gen,
				     struct fs_path *dest)
{
	int ret;
	int nce_ret;
	struct name_cache_entry *nce = NULL;

2143 2144 2145 2146 2147
	/*
	 * First check if we already did a call to this function with the same
	 * ino/gen. If yes, check if the cache entry is still up-to-date. If yes
	 * return the cached result.
	 */
2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165
	nce = name_cache_search(sctx, ino, gen);
	if (nce) {
		if (ino < sctx->send_progress && nce->need_later_update) {
			name_cache_delete(sctx, nce);
			kfree(nce);
			nce = NULL;
		} else {
			name_cache_used(sctx, nce);
			*parent_ino = nce->parent_ino;
			*parent_gen = nce->parent_gen;
			ret = fs_path_add(dest, nce->name, nce->name_len);
			if (ret < 0)
				goto out;
			ret = nce->ret;
			goto out;
		}
	}

2166 2167 2168 2169 2170
	/*
	 * If the inode is not existent yet, add the orphan name and return 1.
	 * This should only happen for the parent dir that we determine in
	 * __record_new_ref
	 */
2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182
	ret = is_inode_existent(sctx, ino, gen);
	if (ret < 0)
		goto out;

	if (!ret) {
		ret = gen_unique_name(sctx, ino, gen, dest);
		if (ret < 0)
			goto out;
		ret = 1;
		goto out_cache;
	}

2183 2184 2185 2186
	/*
	 * Depending on whether the inode was already processed or not, use
	 * send_root or parent_root for ref lookup.
	 */
2187
	if (ino < sctx->send_progress)
2188 2189
		ret = get_first_ref(sctx->send_root, ino,
				    parent_ino, parent_gen, dest);
2190
	else
2191 2192
		ret = get_first_ref(sctx->parent_root, ino,
				    parent_ino, parent_gen, dest);
2193 2194 2195
	if (ret < 0)
		goto out;

2196 2197 2198 2199
	/*
	 * Check if the ref was overwritten by an inode's ref that was processed
	 * earlier. If yes, treat as orphan and return 1.
	 */
2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212
	ret = did_overwrite_ref(sctx, *parent_ino, *parent_gen, ino, gen,
			dest->start, dest->end - dest->start);
	if (ret < 0)
		goto out;
	if (ret) {
		fs_path_reset(dest);
		ret = gen_unique_name(sctx, ino, gen, dest);
		if (ret < 0)
			goto out;
		ret = 1;
	}

out_cache:
2213 2214 2215
	/*
	 * Store the result of the lookup in the name cache.
	 */
2216
	nce = kmalloc(sizeof(*nce) + fs_path_len(dest) + 1, GFP_KERNEL);
2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277
	if (!nce) {
		ret = -ENOMEM;
		goto out;
	}

	nce->ino = ino;
	nce->gen = gen;
	nce->parent_ino = *parent_ino;
	nce->parent_gen = *parent_gen;
	nce->name_len = fs_path_len(dest);
	nce->ret = ret;
	strcpy(nce->name, dest->start);

	if (ino < sctx->send_progress)
		nce->need_later_update = 0;
	else
		nce->need_later_update = 1;

	nce_ret = name_cache_insert(sctx, nce);
	if (nce_ret < 0)
		ret = nce_ret;
	name_cache_clean_unused(sctx);

out:
	return ret;
}

/*
 * Magic happens here. This function returns the first ref to an inode as it
 * would look like while receiving the stream at this point in time.
 * We walk the path up to the root. For every inode in between, we check if it
 * was already processed/sent. If yes, we continue with the parent as found
 * in send_root. If not, we continue with the parent as found in parent_root.
 * If we encounter an inode that was deleted at this point in time, we use the
 * inodes "orphan" name instead of the real name and stop. Same with new inodes
 * that were not created yet and overwritten inodes/refs.
 *
 * When do we have have orphan inodes:
 * 1. When an inode is freshly created and thus no valid refs are available yet
 * 2. When a directory lost all it's refs (deleted) but still has dir items
 *    inside which were not processed yet (pending for move/delete). If anyone
 *    tried to get the path to the dir items, it would get a path inside that
 *    orphan directory.
 * 3. When an inode is moved around or gets new links, it may overwrite the ref
 *    of an unprocessed inode. If in that case the first ref would be
 *    overwritten, the overwritten inode gets "orphanized". Later when we
 *    process this overwritten inode, it is restored at a new place by moving
 *    the orphan inode.
 *
 * sctx->send_progress tells this function at which point in time receiving
 * would be.
 */
static int get_cur_path(struct send_ctx *sctx, u64 ino, u64 gen,
			struct fs_path *dest)
{
	int ret = 0;
	struct fs_path *name = NULL;
	u64 parent_inode = 0;
	u64 parent_gen = 0;
	int stop = 0;

2278
	name = fs_path_alloc();
2279 2280 2281 2282 2283 2284 2285 2286 2287
	if (!name) {
		ret = -ENOMEM;
		goto out;
	}

	dest->reversed = 1;
	fs_path_reset(dest);

	while (!stop && ino != BTRFS_FIRST_FREE_OBJECTID) {
2288 2289
		struct waiting_dir_move *wdm;

2290 2291
		fs_path_reset(name);

2292 2293 2294 2295 2296 2297 2298 2299
		if (is_waiting_for_rm(sctx, ino)) {
			ret = gen_unique_name(sctx, ino, gen, name);
			if (ret < 0)
				goto out;
			ret = fs_path_add_path(dest, name);
			break;
		}

2300 2301 2302 2303 2304
		wdm = get_waiting_dir_move(sctx, ino);
		if (wdm && wdm->orphanized) {
			ret = gen_unique_name(sctx, ino, gen, name);
			stop = 1;
		} else if (wdm) {
2305 2306 2307 2308 2309 2310 2311 2312 2313 2314
			ret = get_first_ref(sctx->parent_root, ino,
					    &parent_inode, &parent_gen, name);
		} else {
			ret = __get_cur_name_and_parent(sctx, ino, gen,
							&parent_inode,
							&parent_gen, name);
			if (ret)
				stop = 1;
		}

2315 2316
		if (ret < 0)
			goto out;
2317

2318 2319 2320 2321 2322 2323 2324 2325 2326
		ret = fs_path_add_path(dest, name);
		if (ret < 0)
			goto out;

		ino = parent_inode;
		gen = parent_gen;
	}

out:
2327
	fs_path_free(name);
2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347
	if (!ret)
		fs_path_unreverse(dest);
	return ret;
}

/*
 * Sends a BTRFS_SEND_C_SUBVOL command/item to userspace
 */
static int send_subvol_begin(struct send_ctx *sctx)
{
	int ret;
	struct btrfs_root *send_root = sctx->send_root;
	struct btrfs_root *parent_root = sctx->parent_root;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct btrfs_root_ref *ref;
	struct extent_buffer *leaf;
	char *name = NULL;
	int namelen;

2348
	path = btrfs_alloc_path();
2349 2350 2351
	if (!path)
		return -ENOMEM;

2352
	name = kmalloc(BTRFS_PATH_NAME_MAX, GFP_KERNEL);
2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393
	if (!name) {
		btrfs_free_path(path);
		return -ENOMEM;
	}

	key.objectid = send_root->objectid;
	key.type = BTRFS_ROOT_BACKREF_KEY;
	key.offset = 0;

	ret = btrfs_search_slot_for_read(send_root->fs_info->tree_root,
				&key, path, 1, 0);
	if (ret < 0)
		goto out;
	if (ret) {
		ret = -ENOENT;
		goto out;
	}

	leaf = path->nodes[0];
	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
	if (key.type != BTRFS_ROOT_BACKREF_KEY ||
	    key.objectid != send_root->objectid) {
		ret = -ENOENT;
		goto out;
	}
	ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref);
	namelen = btrfs_root_ref_name_len(leaf, ref);
	read_extent_buffer(leaf, name, (unsigned long)(ref + 1), namelen);
	btrfs_release_path(path);

	if (parent_root) {
		ret = begin_cmd(sctx, BTRFS_SEND_C_SNAPSHOT);
		if (ret < 0)
			goto out;
	} else {
		ret = begin_cmd(sctx, BTRFS_SEND_C_SUBVOL);
		if (ret < 0)
			goto out;
	}

	TLV_PUT_STRING(sctx, BTRFS_SEND_A_PATH, name, namelen);
2394 2395 2396 2397 2398 2399 2400 2401

	if (!btrfs_is_empty_uuid(sctx->send_root->root_item.received_uuid))
		TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
			    sctx->send_root->root_item.received_uuid);
	else
		TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
			    sctx->send_root->root_item.uuid);

2402
	TLV_PUT_U64(sctx, BTRFS_SEND_A_CTRANSID,
2403
		    le64_to_cpu(sctx->send_root->root_item.ctransid));
2404
	if (parent_root) {
2405 2406 2407 2408 2409 2410
		if (!btrfs_is_empty_uuid(parent_root->root_item.received_uuid))
			TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID,
				     parent_root->root_item.received_uuid);
		else
			TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID,
				     parent_root->root_item.uuid);
2411
		TLV_PUT_U64(sctx, BTRFS_SEND_A_CLONE_CTRANSID,
2412
			    le64_to_cpu(sctx->parent_root->root_item.ctransid));
2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425
	}

	ret = send_cmd(sctx);

tlv_put_failure:
out:
	btrfs_free_path(path);
	kfree(name);
	return ret;
}

static int send_truncate(struct send_ctx *sctx, u64 ino, u64 gen, u64 size)
{
2426
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
2427 2428 2429
	int ret = 0;
	struct fs_path *p;

2430
	btrfs_debug(fs_info, "send_truncate %llu size=%llu", ino, size);
2431

2432
	p = fs_path_alloc();
2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449
	if (!p)
		return -ENOMEM;

	ret = begin_cmd(sctx, BTRFS_SEND_C_TRUNCATE);
	if (ret < 0)
		goto out;

	ret = get_cur_path(sctx, ino, gen, p);
	if (ret < 0)
		goto out;
	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
	TLV_PUT_U64(sctx, BTRFS_SEND_A_SIZE, size);

	ret = send_cmd(sctx);

tlv_put_failure:
out:
2450
	fs_path_free(p);
2451 2452 2453 2454 2455
	return ret;
}

static int send_chmod(struct send_ctx *sctx, u64 ino, u64 gen, u64 mode)
{
2456
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
2457 2458 2459
	int ret = 0;
	struct fs_path *p;

2460
	btrfs_debug(fs_info, "send_chmod %llu mode=%llu", ino, mode);
2461

2462
	p = fs_path_alloc();
2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479
	if (!p)
		return -ENOMEM;

	ret = begin_cmd(sctx, BTRFS_SEND_C_CHMOD);
	if (ret < 0)
		goto out;

	ret = get_cur_path(sctx, ino, gen, p);
	if (ret < 0)
		goto out;
	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
	TLV_PUT_U64(sctx, BTRFS_SEND_A_MODE, mode & 07777);

	ret = send_cmd(sctx);

tlv_put_failure:
out:
2480
	fs_path_free(p);
2481 2482 2483 2484 2485
	return ret;
}

static int send_chown(struct send_ctx *sctx, u64 ino, u64 gen, u64 uid, u64 gid)
{
2486
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
2487 2488 2489
	int ret = 0;
	struct fs_path *p;

2490 2491
	btrfs_debug(fs_info, "send_chown %llu uid=%llu, gid=%llu",
		    ino, uid, gid);
2492

2493
	p = fs_path_alloc();
2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511
	if (!p)
		return -ENOMEM;

	ret = begin_cmd(sctx, BTRFS_SEND_C_CHOWN);
	if (ret < 0)
		goto out;

	ret = get_cur_path(sctx, ino, gen, p);
	if (ret < 0)
		goto out;
	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
	TLV_PUT_U64(sctx, BTRFS_SEND_A_UID, uid);
	TLV_PUT_U64(sctx, BTRFS_SEND_A_GID, gid);

	ret = send_cmd(sctx);

tlv_put_failure:
out:
2512
	fs_path_free(p);
2513 2514 2515 2516 2517
	return ret;
}

static int send_utimes(struct send_ctx *sctx, u64 ino, u64 gen)
{
2518
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
2519 2520 2521 2522 2523 2524 2525 2526
	int ret = 0;
	struct fs_path *p = NULL;
	struct btrfs_inode_item *ii;
	struct btrfs_path *path = NULL;
	struct extent_buffer *eb;
	struct btrfs_key key;
	int slot;

2527
	btrfs_debug(fs_info, "send_utimes %llu", ino);
2528

2529
	p = fs_path_alloc();
2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542
	if (!p)
		return -ENOMEM;

	path = alloc_path_for_send();
	if (!path) {
		ret = -ENOMEM;
		goto out;
	}

	key.objectid = ino;
	key.type = BTRFS_INODE_ITEM_KEY;
	key.offset = 0;
	ret = btrfs_search_slot(NULL, sctx->send_root, &key, path, 0, 0);
2543 2544
	if (ret > 0)
		ret = -ENOENT;
2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559
	if (ret < 0)
		goto out;

	eb = path->nodes[0];
	slot = path->slots[0];
	ii = btrfs_item_ptr(eb, slot, struct btrfs_inode_item);

	ret = begin_cmd(sctx, BTRFS_SEND_C_UTIMES);
	if (ret < 0)
		goto out;

	ret = get_cur_path(sctx, ino, gen, p);
	if (ret < 0)
		goto out;
	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
2560 2561 2562
	TLV_PUT_BTRFS_TIMESPEC(sctx, BTRFS_SEND_A_ATIME, eb, &ii->atime);
	TLV_PUT_BTRFS_TIMESPEC(sctx, BTRFS_SEND_A_MTIME, eb, &ii->mtime);
	TLV_PUT_BTRFS_TIMESPEC(sctx, BTRFS_SEND_A_CTIME, eb, &ii->ctime);
2563
	/* TODO Add otime support when the otime patches get into upstream */
2564 2565 2566 2567 2568

	ret = send_cmd(sctx);

tlv_put_failure:
out:
2569
	fs_path_free(p);
2570 2571 2572 2573 2574 2575 2576 2577 2578
	btrfs_free_path(path);
	return ret;
}

/*
 * Sends a BTRFS_SEND_C_MKXXX or SYMLINK command to user space. We don't have
 * a valid path yet because we did not process the refs yet. So, the inode
 * is created as orphan.
 */
2579
static int send_create_inode(struct send_ctx *sctx, u64 ino)
2580
{
2581
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
2582 2583 2584
	int ret = 0;
	struct fs_path *p;
	int cmd;
2585
	u64 gen;
2586
	u64 mode;
2587
	u64 rdev;
2588

2589
	btrfs_debug(fs_info, "send_create_inode %llu", ino);
2590

2591
	p = fs_path_alloc();
2592 2593 2594
	if (!p)
		return -ENOMEM;

L
Liu Bo 已提交
2595 2596 2597 2598 2599 2600 2601 2602 2603 2604
	if (ino != sctx->cur_ino) {
		ret = get_inode_info(sctx->send_root, ino, NULL, &gen, &mode,
				     NULL, NULL, &rdev);
		if (ret < 0)
			goto out;
	} else {
		gen = sctx->cur_inode_gen;
		mode = sctx->cur_inode_mode;
		rdev = sctx->cur_inode_rdev;
	}
2605

2606
	if (S_ISREG(mode)) {
2607
		cmd = BTRFS_SEND_C_MKFILE;
2608
	} else if (S_ISDIR(mode)) {
2609
		cmd = BTRFS_SEND_C_MKDIR;
2610
	} else if (S_ISLNK(mode)) {
2611
		cmd = BTRFS_SEND_C_SYMLINK;
2612
	} else if (S_ISCHR(mode) || S_ISBLK(mode)) {
2613
		cmd = BTRFS_SEND_C_MKNOD;
2614
	} else if (S_ISFIFO(mode)) {
2615
		cmd = BTRFS_SEND_C_MKFIFO;
2616
	} else if (S_ISSOCK(mode)) {
2617
		cmd = BTRFS_SEND_C_MKSOCK;
2618
	} else {
2619
		btrfs_warn(sctx->send_root->fs_info, "unexpected inode type %o",
2620 2621 2622 2623 2624 2625 2626 2627 2628
				(int)(mode & S_IFMT));
		ret = -ENOTSUPP;
		goto out;
	}

	ret = begin_cmd(sctx, cmd);
	if (ret < 0)
		goto out;

2629
	ret = gen_unique_name(sctx, ino, gen, p);
2630 2631 2632 2633
	if (ret < 0)
		goto out;

	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
2634
	TLV_PUT_U64(sctx, BTRFS_SEND_A_INO, ino);
2635 2636 2637

	if (S_ISLNK(mode)) {
		fs_path_reset(p);
2638
		ret = read_symlink(sctx->send_root, ino, p);
2639 2640 2641 2642 2643
		if (ret < 0)
			goto out;
		TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH_LINK, p);
	} else if (S_ISCHR(mode) || S_ISBLK(mode) ||
		   S_ISFIFO(mode) || S_ISSOCK(mode)) {
2644 2645
		TLV_PUT_U64(sctx, BTRFS_SEND_A_RDEV, new_encode_dev(rdev));
		TLV_PUT_U64(sctx, BTRFS_SEND_A_MODE, mode);
2646 2647 2648 2649 2650 2651 2652 2653 2654
	}

	ret = send_cmd(sctx);
	if (ret < 0)
		goto out;


tlv_put_failure:
out:
2655
	fs_path_free(p);
2656 2657 2658
	return ret;
}

2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683
/*
 * We need some special handling for inodes that get processed before the parent
 * directory got created. See process_recorded_refs for details.
 * This function does the check if we already created the dir out of order.
 */
static int did_create_dir(struct send_ctx *sctx, u64 dir)
{
	int ret = 0;
	struct btrfs_path *path = NULL;
	struct btrfs_key key;
	struct btrfs_key found_key;
	struct btrfs_key di_key;
	struct extent_buffer *eb;
	struct btrfs_dir_item *di;
	int slot;

	path = alloc_path_for_send();
	if (!path) {
		ret = -ENOMEM;
		goto out;
	}

	key.objectid = dir;
	key.type = BTRFS_DIR_INDEX_KEY;
	key.offset = 0;
2684 2685 2686 2687
	ret = btrfs_search_slot(NULL, sctx->send_root, &key, path, 0, 0);
	if (ret < 0)
		goto out;

2688
	while (1) {
2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699
		eb = path->nodes[0];
		slot = path->slots[0];
		if (slot >= btrfs_header_nritems(eb)) {
			ret = btrfs_next_leaf(sctx->send_root, path);
			if (ret < 0) {
				goto out;
			} else if (ret > 0) {
				ret = 0;
				break;
			}
			continue;
2700
		}
2701 2702 2703

		btrfs_item_key_to_cpu(eb, &found_key, slot);
		if (found_key.objectid != key.objectid ||
2704 2705 2706 2707 2708 2709 2710 2711
		    found_key.type != key.type) {
			ret = 0;
			goto out;
		}

		di = btrfs_item_ptr(eb, slot, struct btrfs_dir_item);
		btrfs_dir_item_key_to_cpu(eb, di, &di_key);

2712 2713
		if (di_key.type != BTRFS_ROOT_ITEM_KEY &&
		    di_key.objectid < sctx->send_progress) {
2714 2715 2716 2717
			ret = 1;
			goto out;
		}

2718
		path->slots[0]++;
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
	}

out:
	btrfs_free_path(path);
	return ret;
}

/*
 * Only creates the inode if it is:
 * 1. Not a directory
 * 2. Or a directory which was not created already due to out of order
 *    directories. See did_create_dir and process_recorded_refs for details.
 */
static int send_create_inode_if_needed(struct send_ctx *sctx)
{
	int ret;

	if (S_ISDIR(sctx->cur_inode_mode)) {
		ret = did_create_dir(sctx, sctx->cur_ino);
		if (ret < 0)
			goto out;
		if (ret) {
			ret = 0;
			goto out;
		}
	}

	ret = send_create_inode(sctx, sctx->cur_ino);
	if (ret < 0)
		goto out;

out:
	return ret;
}

2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769
struct recorded_ref {
	struct list_head list;
	char *dir_path;
	char *name;
	struct fs_path *full_path;
	u64 dir;
	u64 dir_gen;
	int dir_path_len;
	int name_len;
};

/*
 * We need to process new refs before deleted refs, but compare_tree gives us
 * everything mixed. So we first record all refs and later process them.
 * This function is a helper to record one ref.
 */
2770
static int __record_ref(struct list_head *head, u64 dir,
2771 2772 2773 2774
		      u64 dir_gen, struct fs_path *path)
{
	struct recorded_ref *ref;

2775
	ref = kmalloc(sizeof(*ref), GFP_KERNEL);
2776 2777 2778 2779 2780 2781 2782
	if (!ref)
		return -ENOMEM;

	ref->dir = dir;
	ref->dir_gen = dir_gen;
	ref->full_path = path;

A
Andy Shevchenko 已提交
2783 2784 2785 2786
	ref->name = (char *)kbasename(ref->full_path->start);
	ref->name_len = ref->full_path->end - ref->name;
	ref->dir_path = ref->full_path->start;
	if (ref->name == ref->full_path->start)
2787
		ref->dir_path_len = 0;
A
Andy Shevchenko 已提交
2788
	else
2789 2790 2791 2792 2793 2794 2795
		ref->dir_path_len = ref->full_path->end -
				ref->full_path->start - 1 - ref->name_len;

	list_add_tail(&ref->list, head);
	return 0;
}

2796 2797 2798 2799
static int dup_ref(struct recorded_ref *ref, struct list_head *list)
{
	struct recorded_ref *new;

2800
	new = kmalloc(sizeof(*ref), GFP_KERNEL);
2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811
	if (!new)
		return -ENOMEM;

	new->dir = ref->dir;
	new->dir_gen = ref->dir_gen;
	new->full_path = NULL;
	INIT_LIST_HEAD(&new->list);
	list_add_tail(&new->list, list);
	return 0;
}

2812
static void __free_recorded_refs(struct list_head *head)
2813 2814 2815
{
	struct recorded_ref *cur;

2816 2817
	while (!list_empty(head)) {
		cur = list_entry(head->next, struct recorded_ref, list);
2818
		fs_path_free(cur->full_path);
2819
		list_del(&cur->list);
2820 2821 2822 2823 2824 2825
		kfree(cur);
	}
}

static void free_recorded_refs(struct send_ctx *sctx)
{
2826 2827
	__free_recorded_refs(&sctx->new_refs);
	__free_recorded_refs(&sctx->deleted_refs);
2828 2829 2830
}

/*
2831
 * Renames/moves a file/dir to its orphan name. Used when the first
2832 2833 2834 2835 2836 2837 2838 2839 2840
 * ref of an unprocessed inode gets overwritten and for all non empty
 * directories.
 */
static int orphanize_inode(struct send_ctx *sctx, u64 ino, u64 gen,
			  struct fs_path *path)
{
	int ret;
	struct fs_path *orphan;

2841
	orphan = fs_path_alloc();
2842 2843 2844 2845 2846 2847 2848 2849 2850 2851
	if (!orphan)
		return -ENOMEM;

	ret = gen_unique_name(sctx, ino, gen, orphan);
	if (ret < 0)
		goto out;

	ret = send_rename(sctx, path, orphan);

out:
2852
	fs_path_free(orphan);
2853 2854 2855
	return ret;
}

2856 2857 2858 2859 2860 2861 2862
static struct orphan_dir_info *
add_orphan_dir_info(struct send_ctx *sctx, u64 dir_ino)
{
	struct rb_node **p = &sctx->orphan_dirs.rb_node;
	struct rb_node *parent = NULL;
	struct orphan_dir_info *entry, *odi;

2863
	odi = kmalloc(sizeof(*odi), GFP_KERNEL);
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
	if (!odi)
		return ERR_PTR(-ENOMEM);
	odi->ino = dir_ino;
	odi->gen = 0;

	while (*p) {
		parent = *p;
		entry = rb_entry(parent, struct orphan_dir_info, node);
		if (dir_ino < entry->ino) {
			p = &(*p)->rb_left;
		} else if (dir_ino > entry->ino) {
			p = &(*p)->rb_right;
		} else {
			kfree(odi);
			return entry;
		}
	}

	rb_link_node(&odi->node, parent, p);
	rb_insert_color(&odi->node, &sctx->orphan_dirs);
	return odi;
}

static struct orphan_dir_info *
get_orphan_dir_info(struct send_ctx *sctx, u64 dir_ino)
{
	struct rb_node *n = sctx->orphan_dirs.rb_node;
	struct orphan_dir_info *entry;

	while (n) {
		entry = rb_entry(n, struct orphan_dir_info, node);
		if (dir_ino < entry->ino)
			n = n->rb_left;
		else if (dir_ino > entry->ino)
			n = n->rb_right;
		else
			return entry;
	}
	return NULL;
}

static int is_waiting_for_rm(struct send_ctx *sctx, u64 dir_ino)
{
	struct orphan_dir_info *odi = get_orphan_dir_info(sctx, dir_ino);

	return odi != NULL;
}

static void free_orphan_dir_info(struct send_ctx *sctx,
				 struct orphan_dir_info *odi)
{
	if (!odi)
		return;
	rb_erase(&odi->node, &sctx->orphan_dirs);
	kfree(odi);
}

2921 2922 2923 2924 2925
/*
 * Returns 1 if a directory can be removed at this point in time.
 * We check this by iterating all dir items and checking if the inode behind
 * the dir item was already processed.
 */
2926 2927
static int can_rmdir(struct send_ctx *sctx, u64 dir, u64 dir_gen,
		     u64 send_progress)
2928 2929 2930 2931 2932 2933 2934 2935 2936
{
	int ret = 0;
	struct btrfs_root *root = sctx->parent_root;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct btrfs_key found_key;
	struct btrfs_key loc;
	struct btrfs_dir_item *di;

2937 2938 2939 2940 2941 2942
	/*
	 * Don't try to rmdir the top/root subvolume dir.
	 */
	if (dir == BTRFS_FIRST_FREE_OBJECTID)
		return 0;

2943 2944 2945 2946 2947 2948 2949
	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	key.objectid = dir;
	key.type = BTRFS_DIR_INDEX_KEY;
	key.offset = 0;
2950 2951 2952
	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto out;
2953 2954

	while (1) {
2955 2956
		struct waiting_dir_move *dm;

2957 2958 2959 2960 2961 2962 2963
		if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0)
				goto out;
			else if (ret > 0)
				break;
			continue;
2964
		}
2965 2966 2967 2968
		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
				      path->slots[0]);
		if (found_key.objectid != key.objectid ||
		    found_key.type != key.type)
2969 2970 2971 2972 2973 2974
			break;

		di = btrfs_item_ptr(path->nodes[0], path->slots[0],
				struct btrfs_dir_item);
		btrfs_dir_item_key_to_cpu(path->nodes[0], di, &loc);

2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989
		dm = get_waiting_dir_move(sctx, loc.objectid);
		if (dm) {
			struct orphan_dir_info *odi;

			odi = add_orphan_dir_info(sctx, dir);
			if (IS_ERR(odi)) {
				ret = PTR_ERR(odi);
				goto out;
			}
			odi->gen = dir_gen;
			dm->rmdir_ino = dir;
			ret = 0;
			goto out;
		}

2990
		if (loc.objectid > send_progress) {
2991 2992 2993 2994
			struct orphan_dir_info *odi;

			odi = get_orphan_dir_info(sctx, dir);
			free_orphan_dir_info(sctx, odi);
2995 2996 2997 2998
			ret = 0;
			goto out;
		}

2999
		path->slots[0]++;
3000 3001 3002 3003 3004 3005 3006 3007 3008
	}

	ret = 1;

out:
	btrfs_free_path(path);
	return ret;
}

3009 3010
static int is_waiting_for_move(struct send_ctx *sctx, u64 ino)
{
3011
	struct waiting_dir_move *entry = get_waiting_dir_move(sctx, ino);
3012

3013
	return entry != NULL;
3014 3015
}

3016
static int add_waiting_dir_move(struct send_ctx *sctx, u64 ino, bool orphanized)
3017 3018 3019 3020 3021
{
	struct rb_node **p = &sctx->waiting_dir_moves.rb_node;
	struct rb_node *parent = NULL;
	struct waiting_dir_move *entry, *dm;

3022
	dm = kmalloc(sizeof(*dm), GFP_KERNEL);
3023 3024 3025
	if (!dm)
		return -ENOMEM;
	dm->ino = ino;
3026
	dm->rmdir_ino = 0;
3027
	dm->orphanized = orphanized;
3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046

	while (*p) {
		parent = *p;
		entry = rb_entry(parent, struct waiting_dir_move, node);
		if (ino < entry->ino) {
			p = &(*p)->rb_left;
		} else if (ino > entry->ino) {
			p = &(*p)->rb_right;
		} else {
			kfree(dm);
			return -EEXIST;
		}
	}

	rb_link_node(&dm->node, parent, p);
	rb_insert_color(&dm->node, &sctx->waiting_dir_moves);
	return 0;
}

3047 3048
static struct waiting_dir_move *
get_waiting_dir_move(struct send_ctx *sctx, u64 ino)
3049 3050 3051 3052 3053 3054
{
	struct rb_node *n = sctx->waiting_dir_moves.rb_node;
	struct waiting_dir_move *entry;

	while (n) {
		entry = rb_entry(n, struct waiting_dir_move, node);
3055
		if (ino < entry->ino)
3056
			n = n->rb_left;
3057
		else if (ino > entry->ino)
3058
			n = n->rb_right;
3059 3060
		else
			return entry;
3061
	}
3062 3063 3064 3065 3066 3067 3068 3069 3070 3071
	return NULL;
}

static void free_waiting_dir_move(struct send_ctx *sctx,
				  struct waiting_dir_move *dm)
{
	if (!dm)
		return;
	rb_erase(&dm->node, &sctx->waiting_dir_moves);
	kfree(dm);
3072 3073
}

3074 3075 3076
static int add_pending_dir_move(struct send_ctx *sctx,
				u64 ino,
				u64 ino_gen,
3077 3078
				u64 parent_ino,
				struct list_head *new_refs,
3079 3080
				struct list_head *deleted_refs,
				const bool is_orphan)
3081 3082 3083
{
	struct rb_node **p = &sctx->pending_dir_moves.rb_node;
	struct rb_node *parent = NULL;
C
Chris Mason 已提交
3084
	struct pending_dir_move *entry = NULL, *pm;
3085 3086 3087 3088
	struct recorded_ref *cur;
	int exists = 0;
	int ret;

3089
	pm = kmalloc(sizeof(*pm), GFP_KERNEL);
3090 3091 3092
	if (!pm)
		return -ENOMEM;
	pm->parent_ino = parent_ino;
3093 3094
	pm->ino = ino;
	pm->gen = ino_gen;
3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111
	INIT_LIST_HEAD(&pm->list);
	INIT_LIST_HEAD(&pm->update_refs);
	RB_CLEAR_NODE(&pm->node);

	while (*p) {
		parent = *p;
		entry = rb_entry(parent, struct pending_dir_move, node);
		if (parent_ino < entry->parent_ino) {
			p = &(*p)->rb_left;
		} else if (parent_ino > entry->parent_ino) {
			p = &(*p)->rb_right;
		} else {
			exists = 1;
			break;
		}
	}

3112
	list_for_each_entry(cur, deleted_refs, list) {
3113 3114 3115 3116
		ret = dup_ref(cur, &pm->update_refs);
		if (ret < 0)
			goto out;
	}
3117
	list_for_each_entry(cur, new_refs, list) {
3118 3119 3120 3121 3122
		ret = dup_ref(cur, &pm->update_refs);
		if (ret < 0)
			goto out;
	}

3123
	ret = add_waiting_dir_move(sctx, pm->ino, is_orphan);
3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159
	if (ret)
		goto out;

	if (exists) {
		list_add_tail(&pm->list, &entry->list);
	} else {
		rb_link_node(&pm->node, parent, p);
		rb_insert_color(&pm->node, &sctx->pending_dir_moves);
	}
	ret = 0;
out:
	if (ret) {
		__free_recorded_refs(&pm->update_refs);
		kfree(pm);
	}
	return ret;
}

static struct pending_dir_move *get_pending_dir_moves(struct send_ctx *sctx,
						      u64 parent_ino)
{
	struct rb_node *n = sctx->pending_dir_moves.rb_node;
	struct pending_dir_move *entry;

	while (n) {
		entry = rb_entry(n, struct pending_dir_move, node);
		if (parent_ino < entry->parent_ino)
			n = n->rb_left;
		else if (parent_ino > entry->parent_ino)
			n = n->rb_right;
		else
			return entry;
	}
	return NULL;
}

3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201
static int path_loop(struct send_ctx *sctx, struct fs_path *name,
		     u64 ino, u64 gen, u64 *ancestor_ino)
{
	int ret = 0;
	u64 parent_inode = 0;
	u64 parent_gen = 0;
	u64 start_ino = ino;

	*ancestor_ino = 0;
	while (ino != BTRFS_FIRST_FREE_OBJECTID) {
		fs_path_reset(name);

		if (is_waiting_for_rm(sctx, ino))
			break;
		if (is_waiting_for_move(sctx, ino)) {
			if (*ancestor_ino == 0)
				*ancestor_ino = ino;
			ret = get_first_ref(sctx->parent_root, ino,
					    &parent_inode, &parent_gen, name);
		} else {
			ret = __get_cur_name_and_parent(sctx, ino, gen,
							&parent_inode,
							&parent_gen, name);
			if (ret > 0) {
				ret = 0;
				break;
			}
		}
		if (ret < 0)
			break;
		if (parent_inode == start_ino) {
			ret = 1;
			if (*ancestor_ino == 0)
				*ancestor_ino = ino;
			break;
		}
		ino = parent_inode;
		gen = parent_gen;
	}
	return ret;
}

3202 3203 3204 3205
static int apply_dir_move(struct send_ctx *sctx, struct pending_dir_move *pm)
{
	struct fs_path *from_path = NULL;
	struct fs_path *to_path = NULL;
3206
	struct fs_path *name = NULL;
3207 3208
	u64 orig_progress = sctx->send_progress;
	struct recorded_ref *cur;
3209
	u64 parent_ino, parent_gen;
3210 3211
	struct waiting_dir_move *dm = NULL;
	u64 rmdir_ino = 0;
3212 3213
	u64 ancestor;
	bool is_orphan;
3214 3215
	int ret;

3216
	name = fs_path_alloc();
3217
	from_path = fs_path_alloc();
3218 3219 3220 3221
	if (!name || !from_path) {
		ret = -ENOMEM;
		goto out;
	}
3222

3223 3224 3225
	dm = get_waiting_dir_move(sctx, pm->ino);
	ASSERT(dm);
	rmdir_ino = dm->rmdir_ino;
3226
	is_orphan = dm->orphanized;
3227
	free_waiting_dir_move(sctx, dm);
3228

3229
	if (is_orphan) {
3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242
		ret = gen_unique_name(sctx, pm->ino,
				      pm->gen, from_path);
	} else {
		ret = get_first_ref(sctx->parent_root, pm->ino,
				    &parent_ino, &parent_gen, name);
		if (ret < 0)
			goto out;
		ret = get_cur_path(sctx, parent_ino, parent_gen,
				   from_path);
		if (ret < 0)
			goto out;
		ret = fs_path_add_path(from_path, name);
	}
3243 3244
	if (ret < 0)
		goto out;
3245

3246
	sctx->send_progress = sctx->cur_ino + 1;
3247
	ret = path_loop(sctx, name, pm->ino, pm->gen, &ancestor);
3248 3249
	if (ret < 0)
		goto out;
3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264
	if (ret) {
		LIST_HEAD(deleted_refs);
		ASSERT(ancestor > BTRFS_FIRST_FREE_OBJECTID);
		ret = add_pending_dir_move(sctx, pm->ino, pm->gen, ancestor,
					   &pm->update_refs, &deleted_refs,
					   is_orphan);
		if (ret < 0)
			goto out;
		if (rmdir_ino) {
			dm = get_waiting_dir_move(sctx, pm->ino);
			ASSERT(dm);
			dm->rmdir_ino = rmdir_ino;
		}
		goto out;
	}
3265 3266
	fs_path_reset(name);
	to_path = name;
3267
	name = NULL;
3268 3269 3270 3271 3272 3273 3274 3275
	ret = get_cur_path(sctx, pm->ino, pm->gen, to_path);
	if (ret < 0)
		goto out;

	ret = send_rename(sctx, from_path, to_path);
	if (ret < 0)
		goto out;

3276 3277 3278 3279 3280 3281 3282 3283
	if (rmdir_ino) {
		struct orphan_dir_info *odi;

		odi = get_orphan_dir_info(sctx, rmdir_ino);
		if (!odi) {
			/* already deleted */
			goto finish;
		}
3284
		ret = can_rmdir(sctx, rmdir_ino, odi->gen, sctx->cur_ino);
3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304
		if (ret < 0)
			goto out;
		if (!ret)
			goto finish;

		name = fs_path_alloc();
		if (!name) {
			ret = -ENOMEM;
			goto out;
		}
		ret = get_cur_path(sctx, rmdir_ino, odi->gen, name);
		if (ret < 0)
			goto out;
		ret = send_rmdir(sctx, name);
		if (ret < 0)
			goto out;
		free_orphan_dir_info(sctx, odi);
	}

finish:
3305 3306 3307 3308 3309 3310 3311 3312 3313
	ret = send_utimes(sctx, pm->ino, pm->gen);
	if (ret < 0)
		goto out;

	/*
	 * After rename/move, need to update the utimes of both new parent(s)
	 * and old parent(s).
	 */
	list_for_each_entry(cur, &pm->update_refs, list) {
3314 3315 3316 3317 3318 3319 3320
		/*
		 * The parent inode might have been deleted in the send snapshot
		 */
		ret = get_inode_info(sctx->send_root, cur->dir, NULL,
				     NULL, NULL, NULL, NULL, NULL);
		if (ret == -ENOENT) {
			ret = 0;
3321
			continue;
3322 3323 3324 3325
		}
		if (ret < 0)
			goto out;

3326 3327 3328 3329 3330 3331
		ret = send_utimes(sctx, cur->dir, cur->dir_gen);
		if (ret < 0)
			goto out;
	}

out:
3332
	fs_path_free(name);
3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397
	fs_path_free(from_path);
	fs_path_free(to_path);
	sctx->send_progress = orig_progress;

	return ret;
}

static void free_pending_move(struct send_ctx *sctx, struct pending_dir_move *m)
{
	if (!list_empty(&m->list))
		list_del(&m->list);
	if (!RB_EMPTY_NODE(&m->node))
		rb_erase(&m->node, &sctx->pending_dir_moves);
	__free_recorded_refs(&m->update_refs);
	kfree(m);
}

static void tail_append_pending_moves(struct pending_dir_move *moves,
				      struct list_head *stack)
{
	if (list_empty(&moves->list)) {
		list_add_tail(&moves->list, stack);
	} else {
		LIST_HEAD(list);
		list_splice_init(&moves->list, &list);
		list_add_tail(&moves->list, stack);
		list_splice_tail(&list, stack);
	}
}

static int apply_children_dir_moves(struct send_ctx *sctx)
{
	struct pending_dir_move *pm;
	struct list_head stack;
	u64 parent_ino = sctx->cur_ino;
	int ret = 0;

	pm = get_pending_dir_moves(sctx, parent_ino);
	if (!pm)
		return 0;

	INIT_LIST_HEAD(&stack);
	tail_append_pending_moves(pm, &stack);

	while (!list_empty(&stack)) {
		pm = list_first_entry(&stack, struct pending_dir_move, list);
		parent_ino = pm->ino;
		ret = apply_dir_move(sctx, pm);
		free_pending_move(sctx, pm);
		if (ret)
			goto out;
		pm = get_pending_dir_moves(sctx, parent_ino);
		if (pm)
			tail_append_pending_moves(pm, &stack);
	}
	return 0;

out:
	while (!list_empty(&stack)) {
		pm = list_first_entry(&stack, struct pending_dir_move, list);
		free_pending_move(sctx, pm);
	}
	return ret;
}

3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437
/*
 * We might need to delay a directory rename even when no ancestor directory
 * (in the send root) with a higher inode number than ours (sctx->cur_ino) was
 * renamed. This happens when we rename a directory to the old name (the name
 * in the parent root) of some other unrelated directory that got its rename
 * delayed due to some ancestor with higher number that got renamed.
 *
 * Example:
 *
 * Parent snapshot:
 * .                                       (ino 256)
 * |---- a/                                (ino 257)
 * |     |---- file                        (ino 260)
 * |
 * |---- b/                                (ino 258)
 * |---- c/                                (ino 259)
 *
 * Send snapshot:
 * .                                       (ino 256)
 * |---- a/                                (ino 258)
 * |---- x/                                (ino 259)
 *       |---- y/                          (ino 257)
 *             |----- file                 (ino 260)
 *
 * Here we can not rename 258 from 'b' to 'a' without the rename of inode 257
 * from 'a' to 'x/y' happening first, which in turn depends on the rename of
 * inode 259 from 'c' to 'x'. So the order of rename commands the send stream
 * must issue is:
 *
 * 1 - rename 259 from 'c' to 'x'
 * 2 - rename 257 from 'a' to 'x/y'
 * 3 - rename 258 from 'b' to 'a'
 *
 * Returns 1 if the rename of sctx->cur_ino needs to be delayed, 0 if it can
 * be done right away and < 0 on error.
 */
static int wait_for_dest_dir_move(struct send_ctx *sctx,
				  struct recorded_ref *parent_ref,
				  const bool is_orphan)
{
3438
	struct btrfs_fs_info *fs_info = sctx->parent_root->fs_info;
3439 3440 3441 3442 3443 3444 3445
	struct btrfs_path *path;
	struct btrfs_key key;
	struct btrfs_key di_key;
	struct btrfs_dir_item *di;
	u64 left_gen;
	u64 right_gen;
	int ret = 0;
3446
	struct waiting_dir_move *wdm;
3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466

	if (RB_EMPTY_ROOT(&sctx->waiting_dir_moves))
		return 0;

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	key.objectid = parent_ref->dir;
	key.type = BTRFS_DIR_ITEM_KEY;
	key.offset = btrfs_name_hash(parent_ref->name, parent_ref->name_len);

	ret = btrfs_search_slot(NULL, sctx->parent_root, &key, path, 0, 0);
	if (ret < 0) {
		goto out;
	} else if (ret > 0) {
		ret = 0;
		goto out;
	}

3467 3468
	di = btrfs_match_dir_item_name(fs_info, path, parent_ref->name,
				       parent_ref->name_len);
3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504
	if (!di) {
		ret = 0;
		goto out;
	}
	/*
	 * di_key.objectid has the number of the inode that has a dentry in the
	 * parent directory with the same name that sctx->cur_ino is being
	 * renamed to. We need to check if that inode is in the send root as
	 * well and if it is currently marked as an inode with a pending rename,
	 * if it is, we need to delay the rename of sctx->cur_ino as well, so
	 * that it happens after that other inode is renamed.
	 */
	btrfs_dir_item_key_to_cpu(path->nodes[0], di, &di_key);
	if (di_key.type != BTRFS_INODE_ITEM_KEY) {
		ret = 0;
		goto out;
	}

	ret = get_inode_info(sctx->parent_root, di_key.objectid, NULL,
			     &left_gen, NULL, NULL, NULL, NULL);
	if (ret < 0)
		goto out;
	ret = get_inode_info(sctx->send_root, di_key.objectid, NULL,
			     &right_gen, NULL, NULL, NULL, NULL);
	if (ret < 0) {
		if (ret == -ENOENT)
			ret = 0;
		goto out;
	}

	/* Different inode, no need to delay the rename of sctx->cur_ino */
	if (right_gen != left_gen) {
		ret = 0;
		goto out;
	}

3505 3506
	wdm = get_waiting_dir_move(sctx, di_key.objectid);
	if (wdm && !wdm->orphanized) {
3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521
		ret = add_pending_dir_move(sctx,
					   sctx->cur_ino,
					   sctx->cur_inode_gen,
					   di_key.objectid,
					   &sctx->new_refs,
					   &sctx->deleted_refs,
					   is_orphan);
		if (!ret)
			ret = 1;
	}
out:
	btrfs_free_path(path);
	return ret;
}

3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552
/*
 * Check if ino ino1 is an ancestor of inode ino2 in the given root.
 * Return 1 if true, 0 if false and < 0 on error.
 */
static int is_ancestor(struct btrfs_root *root,
		       const u64 ino1,
		       const u64 ino1_gen,
		       const u64 ino2,
		       struct fs_path *fs_path)
{
	u64 ino = ino2;

	while (ino > BTRFS_FIRST_FREE_OBJECTID) {
		int ret;
		u64 parent;
		u64 parent_gen;

		fs_path_reset(fs_path);
		ret = get_first_ref(root, ino, &parent, &parent_gen, fs_path);
		if (ret < 0) {
			if (ret == -ENOENT && ino == ino2)
				ret = 0;
			return ret;
		}
		if (parent == ino1)
			return parent_gen == ino1_gen ? 1 : 0;
		ino = parent;
	}
	return 0;
}

3553
static int wait_for_parent_move(struct send_ctx *sctx,
3554 3555
				struct recorded_ref *parent_ref,
				const bool is_orphan)
3556
{
3557
	int ret = 0;
3558 3559 3560 3561 3562 3563 3564
	u64 ino = parent_ref->dir;
	u64 parent_ino_before, parent_ino_after;
	struct fs_path *path_before = NULL;
	struct fs_path *path_after = NULL;
	int len1, len2;

	path_after = fs_path_alloc();
3565 3566
	path_before = fs_path_alloc();
	if (!path_after || !path_before) {
3567 3568 3569 3570
		ret = -ENOMEM;
		goto out;
	}

3571
	/*
3572 3573 3574
	 * Our current directory inode may not yet be renamed/moved because some
	 * ancestor (immediate or not) has to be renamed/moved first. So find if
	 * such ancestor exists and make sure our own rename/move happens after
3575 3576
	 * that ancestor is processed to avoid path build infinite loops (done
	 * at get_cur_path()).
3577
	 */
3578 3579
	while (ino > BTRFS_FIRST_FREE_OBJECTID) {
		if (is_waiting_for_move(sctx, ino)) {
3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592
			/*
			 * If the current inode is an ancestor of ino in the
			 * parent root, we need to delay the rename of the
			 * current inode, otherwise don't delayed the rename
			 * because we can end up with a circular dependency
			 * of renames, resulting in some directories never
			 * getting the respective rename operations issued in
			 * the send stream or getting into infinite path build
			 * loops.
			 */
			ret = is_ancestor(sctx->parent_root,
					  sctx->cur_ino, sctx->cur_inode_gen,
					  ino, path_before);
3593 3594
			if (ret)
				break;
3595
		}
3596 3597 3598 3599 3600

		fs_path_reset(path_before);
		fs_path_reset(path_after);

		ret = get_first_ref(sctx->send_root, ino, &parent_ino_after,
3601
				    NULL, path_after);
3602 3603 3604 3605
		if (ret < 0)
			goto out;
		ret = get_first_ref(sctx->parent_root, ino, &parent_ino_before,
				    NULL, path_before);
3606
		if (ret < 0 && ret != -ENOENT) {
3607
			goto out;
3608
		} else if (ret == -ENOENT) {
3609
			ret = 0;
3610
			break;
3611 3612 3613 3614
		}

		len1 = fs_path_len(path_before);
		len2 = fs_path_len(path_after);
3615 3616 3617
		if (ino > sctx->cur_ino &&
		    (parent_ino_before != parent_ino_after || len1 != len2 ||
		     memcmp(path_before->start, path_after->start, len1))) {
3618
			ret = 1;
3619
			break;
3620 3621 3622 3623
		}
		ino = parent_ino_after;
	}

3624 3625 3626 3627
out:
	fs_path_free(path_before);
	fs_path_free(path_after);

3628 3629 3630 3631 3632 3633
	if (ret == 1) {
		ret = add_pending_dir_move(sctx,
					   sctx->cur_ino,
					   sctx->cur_inode_gen,
					   ino,
					   &sctx->new_refs,
3634
					   &sctx->deleted_refs,
3635
					   is_orphan);
3636 3637 3638 3639
		if (!ret)
			ret = 1;
	}

3640 3641 3642
	return ret;
}

3643 3644 3645
/*
 * This does all the move/link/unlink/rmdir magic.
 */
3646
static int process_recorded_refs(struct send_ctx *sctx, int *pending_move)
3647
{
3648
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
3649 3650
	int ret = 0;
	struct recorded_ref *cur;
3651
	struct recorded_ref *cur2;
3652
	struct list_head check_dirs;
3653
	struct fs_path *valid_path = NULL;
3654
	u64 ow_inode = 0;
3655 3656 3657
	u64 ow_gen;
	int did_overwrite = 0;
	int is_orphan = 0;
3658
	u64 last_dir_ino_rm = 0;
3659
	bool can_rename = true;
3660

3661
	btrfs_debug(fs_info, "process_recorded_refs %llu", sctx->cur_ino);
3662

3663 3664 3665 3666 3667
	/*
	 * This should never happen as the root dir always has the same ref
	 * which is always '..'
	 */
	BUG_ON(sctx->cur_ino <= BTRFS_FIRST_FREE_OBJECTID);
3668
	INIT_LIST_HEAD(&check_dirs);
3669

3670
	valid_path = fs_path_alloc();
3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708
	if (!valid_path) {
		ret = -ENOMEM;
		goto out;
	}

	/*
	 * First, check if the first ref of the current inode was overwritten
	 * before. If yes, we know that the current inode was already orphanized
	 * and thus use the orphan name. If not, we can use get_cur_path to
	 * get the path of the first ref as it would like while receiving at
	 * this point in time.
	 * New inodes are always orphan at the beginning, so force to use the
	 * orphan name in this case.
	 * The first ref is stored in valid_path and will be updated if it
	 * gets moved around.
	 */
	if (!sctx->cur_inode_new) {
		ret = did_overwrite_first_ref(sctx, sctx->cur_ino,
				sctx->cur_inode_gen);
		if (ret < 0)
			goto out;
		if (ret)
			did_overwrite = 1;
	}
	if (sctx->cur_inode_new || did_overwrite) {
		ret = gen_unique_name(sctx, sctx->cur_ino,
				sctx->cur_inode_gen, valid_path);
		if (ret < 0)
			goto out;
		is_orphan = 1;
	} else {
		ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen,
				valid_path);
		if (ret < 0)
			goto out;
	}

	list_for_each_entry(cur, &sctx->new_refs, list) {
3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748
		/*
		 * We may have refs where the parent directory does not exist
		 * yet. This happens if the parent directories inum is higher
		 * the the current inum. To handle this case, we create the
		 * parent directory out of order. But we need to check if this
		 * did already happen before due to other refs in the same dir.
		 */
		ret = get_cur_inode_state(sctx, cur->dir, cur->dir_gen);
		if (ret < 0)
			goto out;
		if (ret == inode_state_will_create) {
			ret = 0;
			/*
			 * First check if any of the current inodes refs did
			 * already create the dir.
			 */
			list_for_each_entry(cur2, &sctx->new_refs, list) {
				if (cur == cur2)
					break;
				if (cur2->dir == cur->dir) {
					ret = 1;
					break;
				}
			}

			/*
			 * If that did not happen, check if a previous inode
			 * did already create the dir.
			 */
			if (!ret)
				ret = did_create_dir(sctx, cur->dir);
			if (ret < 0)
				goto out;
			if (!ret) {
				ret = send_create_inode(sctx, cur->dir);
				if (ret < 0)
					goto out;
			}
		}

3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760
		/*
		 * Check if this new ref would overwrite the first ref of
		 * another unprocessed inode. If yes, orphanize the
		 * overwritten inode. If we find an overwritten ref that is
		 * not the first ref, simply unlink it.
		 */
		ret = will_overwrite_ref(sctx, cur->dir, cur->dir_gen,
				cur->name, cur->name_len,
				&ow_inode, &ow_gen);
		if (ret < 0)
			goto out;
		if (ret) {
3761 3762 3763
			ret = is_first_ref(sctx->parent_root,
					   ow_inode, cur->dir, cur->name,
					   cur->name_len);
3764 3765 3766
			if (ret < 0)
				goto out;
			if (ret) {
3767
				struct name_cache_entry *nce;
3768
				struct waiting_dir_move *wdm;
3769

3770 3771 3772 3773
				ret = orphanize_inode(sctx, ow_inode, ow_gen,
						cur->full_path);
				if (ret < 0)
					goto out;
3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787

				/*
				 * If ow_inode has its rename operation delayed
				 * make sure that its orphanized name is used in
				 * the source path when performing its rename
				 * operation.
				 */
				if (is_waiting_for_move(sctx, ow_inode)) {
					wdm = get_waiting_dir_move(sctx,
								   ow_inode);
					ASSERT(wdm);
					wdm->orphanized = true;
				}

3788 3789 3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802
				/*
				 * Make sure we clear our orphanized inode's
				 * name from the name cache. This is because the
				 * inode ow_inode might be an ancestor of some
				 * other inode that will be orphanized as well
				 * later and has an inode number greater than
				 * sctx->send_progress. We need to prevent
				 * future name lookups from using the old name
				 * and get instead the orphan name.
				 */
				nce = name_cache_search(sctx, ow_inode, ow_gen);
				if (nce) {
					name_cache_delete(sctx, nce);
					kfree(nce);
				}
3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815

				/*
				 * ow_inode might currently be an ancestor of
				 * cur_ino, therefore compute valid_path (the
				 * current path of cur_ino) again because it
				 * might contain the pre-orphanization name of
				 * ow_inode, which is no longer valid.
				 */
				fs_path_reset(valid_path);
				ret = get_cur_path(sctx, sctx->cur_ino,
					   sctx->cur_inode_gen, valid_path);
				if (ret < 0)
					goto out;
3816 3817 3818 3819 3820 3821 3822
			} else {
				ret = send_unlink(sctx, cur->full_path);
				if (ret < 0)
					goto out;
			}
		}

3823 3824 3825 3826 3827 3828 3829 3830 3831 3832
		if (S_ISDIR(sctx->cur_inode_mode) && sctx->parent_root) {
			ret = wait_for_dest_dir_move(sctx, cur, is_orphan);
			if (ret < 0)
				goto out;
			if (ret == 1) {
				can_rename = false;
				*pending_move = 1;
			}
		}

3833 3834 3835 3836 3837 3838 3839 3840 3841 3842 3843
		if (S_ISDIR(sctx->cur_inode_mode) && sctx->parent_root &&
		    can_rename) {
			ret = wait_for_parent_move(sctx, cur, is_orphan);
			if (ret < 0)
				goto out;
			if (ret == 1) {
				can_rename = false;
				*pending_move = 1;
			}
		}

3844 3845 3846 3847 3848
		/*
		 * link/move the ref to the new place. If we have an orphan
		 * inode, move it and update valid_path. If not, link or move
		 * it depending on the inode mode.
		 */
3849
		if (is_orphan && can_rename) {
3850 3851 3852 3853 3854 3855 3856
			ret = send_rename(sctx, valid_path, cur->full_path);
			if (ret < 0)
				goto out;
			is_orphan = 0;
			ret = fs_path_copy(valid_path, cur->full_path);
			if (ret < 0)
				goto out;
3857
		} else if (can_rename) {
3858 3859 3860 3861 3862 3863
			if (S_ISDIR(sctx->cur_inode_mode)) {
				/*
				 * Dirs can't be linked, so move it. For moved
				 * dirs, we always have one new and one deleted
				 * ref. The deleted ref is ignored later.
				 */
3864 3865 3866 3867 3868
				ret = send_rename(sctx, valid_path,
						  cur->full_path);
				if (!ret)
					ret = fs_path_copy(valid_path,
							   cur->full_path);
3869 3870 3871 3872 3873 3874 3875 3876 3877
				if (ret < 0)
					goto out;
			} else {
				ret = send_link(sctx, cur->full_path,
						valid_path);
				if (ret < 0)
					goto out;
			}
		}
3878
		ret = dup_ref(cur, &check_dirs);
3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889
		if (ret < 0)
			goto out;
	}

	if (S_ISDIR(sctx->cur_inode_mode) && sctx->cur_inode_deleted) {
		/*
		 * Check if we can already rmdir the directory. If not,
		 * orphanize it. For every dir item inside that gets deleted
		 * later, we do this check again and rmdir it then if possible.
		 * See the use of check_dirs for more details.
		 */
3890 3891
		ret = can_rmdir(sctx, sctx->cur_ino, sctx->cur_inode_gen,
				sctx->cur_ino);
3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906
		if (ret < 0)
			goto out;
		if (ret) {
			ret = send_rmdir(sctx, valid_path);
			if (ret < 0)
				goto out;
		} else if (!is_orphan) {
			ret = orphanize_inode(sctx, sctx->cur_ino,
					sctx->cur_inode_gen, valid_path);
			if (ret < 0)
				goto out;
			is_orphan = 1;
		}

		list_for_each_entry(cur, &sctx->deleted_refs, list) {
3907
			ret = dup_ref(cur, &check_dirs);
3908 3909 3910
			if (ret < 0)
				goto out;
		}
3911 3912 3913 3914 3915 3916 3917
	} else if (S_ISDIR(sctx->cur_inode_mode) &&
		   !list_empty(&sctx->deleted_refs)) {
		/*
		 * We have a moved dir. Add the old parent to check_dirs
		 */
		cur = list_entry(sctx->deleted_refs.next, struct recorded_ref,
				list);
3918
		ret = dup_ref(cur, &check_dirs);
3919 3920
		if (ret < 0)
			goto out;
3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933
	} else if (!S_ISDIR(sctx->cur_inode_mode)) {
		/*
		 * We have a non dir inode. Go through all deleted refs and
		 * unlink them if they were not already overwritten by other
		 * inodes.
		 */
		list_for_each_entry(cur, &sctx->deleted_refs, list) {
			ret = did_overwrite_ref(sctx, cur->dir, cur->dir_gen,
					sctx->cur_ino, sctx->cur_inode_gen,
					cur->name, cur->name_len);
			if (ret < 0)
				goto out;
			if (!ret) {
3934 3935 3936
				ret = send_unlink(sctx, cur->full_path);
				if (ret < 0)
					goto out;
3937
			}
3938
			ret = dup_ref(cur, &check_dirs);
3939 3940 3941 3942 3943 3944 3945
			if (ret < 0)
				goto out;
		}
		/*
		 * If the inode is still orphan, unlink the orphan. This may
		 * happen when a previous inode did overwrite the first ref
		 * of this inode and no new refs were added for the current
3946 3947 3948
		 * inode. Unlinking does not mean that the inode is deleted in
		 * all cases. There may still be links to this inode in other
		 * places.
3949
		 */
3950
		if (is_orphan) {
3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962
			ret = send_unlink(sctx, valid_path);
			if (ret < 0)
				goto out;
		}
	}

	/*
	 * We did collect all parent dirs where cur_inode was once located. We
	 * now go through all these dirs and check if they are pending for
	 * deletion and if it's finally possible to perform the rmdir now.
	 * We also update the inode stats of the parent dirs here.
	 */
3963
	list_for_each_entry(cur, &check_dirs, list) {
3964 3965 3966 3967 3968
		/*
		 * In case we had refs into dirs that were not processed yet,
		 * we don't need to do the utime and rmdir logic for these dirs.
		 * The dir will be processed later.
		 */
3969
		if (cur->dir > sctx->cur_ino)
3970 3971
			continue;

3972
		ret = get_cur_inode_state(sctx, cur->dir, cur->dir_gen);
3973 3974 3975 3976 3977 3978
		if (ret < 0)
			goto out;

		if (ret == inode_state_did_create ||
		    ret == inode_state_no_change) {
			/* TODO delayed utimes */
3979
			ret = send_utimes(sctx, cur->dir, cur->dir_gen);
3980 3981
			if (ret < 0)
				goto out;
3982 3983
		} else if (ret == inode_state_did_delete &&
			   cur->dir != last_dir_ino_rm) {
3984 3985
			ret = can_rmdir(sctx, cur->dir, cur->dir_gen,
					sctx->cur_ino);
3986 3987 3988
			if (ret < 0)
				goto out;
			if (ret) {
3989 3990
				ret = get_cur_path(sctx, cur->dir,
						   cur->dir_gen, valid_path);
3991 3992 3993 3994 3995
				if (ret < 0)
					goto out;
				ret = send_rmdir(sctx, valid_path);
				if (ret < 0)
					goto out;
3996
				last_dir_ino_rm = cur->dir;
3997 3998 3999 4000 4001 4002 4003
			}
		}
	}

	ret = 0;

out:
4004
	__free_recorded_refs(&check_dirs);
4005
	free_recorded_refs(sctx);
4006
	fs_path_free(valid_path);
4007 4008 4009
	return ret;
}

4010 4011
static int record_ref(struct btrfs_root *root, int num, u64 dir, int index,
		      struct fs_path *name, void *ctx, struct list_head *refs)
4012 4013 4014 4015 4016 4017
{
	int ret = 0;
	struct send_ctx *sctx = ctx;
	struct fs_path *p;
	u64 gen;

4018
	p = fs_path_alloc();
4019 4020 4021
	if (!p)
		return -ENOMEM;

4022
	ret = get_inode_info(root, dir, NULL, &gen, NULL, NULL,
4023
			NULL, NULL);
4024 4025 4026 4027 4028 4029 4030 4031 4032 4033
	if (ret < 0)
		goto out;

	ret = get_cur_path(sctx, dir, gen, p);
	if (ret < 0)
		goto out;
	ret = fs_path_add_path(p, name);
	if (ret < 0)
		goto out;

4034
	ret = __record_ref(refs, dir, gen, p);
4035 4036 4037

out:
	if (ret)
4038
		fs_path_free(p);
4039 4040 4041
	return ret;
}

4042 4043 4044 4045 4046 4047 4048 4049 4050 4051
static int __record_new_ref(int num, u64 dir, int index,
			    struct fs_path *name,
			    void *ctx)
{
	struct send_ctx *sctx = ctx;
	return record_ref(sctx->send_root, num, dir, index, name,
			  ctx, &sctx->new_refs);
}


4052 4053 4054 4055 4056
static int __record_deleted_ref(int num, u64 dir, int index,
				struct fs_path *name,
				void *ctx)
{
	struct send_ctx *sctx = ctx;
4057 4058
	return record_ref(sctx->parent_root, num, dir, index, name,
			  ctx, &sctx->deleted_refs);
4059 4060 4061 4062 4063 4064
}

static int record_new_ref(struct send_ctx *sctx)
{
	int ret;

4065 4066
	ret = iterate_inode_ref(sctx->send_root, sctx->left_path,
				sctx->cmp_key, 0, __record_new_ref, sctx);
4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078
	if (ret < 0)
		goto out;
	ret = 0;

out:
	return ret;
}

static int record_deleted_ref(struct send_ctx *sctx)
{
	int ret;

4079 4080
	ret = iterate_inode_ref(sctx->parent_root, sctx->right_path,
				sctx->cmp_key, 0, __record_deleted_ref, sctx);
4081 4082 4083 4084 4085 4086 4087 4088 4089 4090
	if (ret < 0)
		goto out;
	ret = 0;

out:
	return ret;
}

struct find_ref_ctx {
	u64 dir;
4091 4092
	u64 dir_gen;
	struct btrfs_root *root;
4093 4094 4095 4096 4097 4098 4099 4100 4101
	struct fs_path *name;
	int found_idx;
};

static int __find_iref(int num, u64 dir, int index,
		       struct fs_path *name,
		       void *ctx_)
{
	struct find_ref_ctx *ctx = ctx_;
4102 4103
	u64 dir_gen;
	int ret;
4104 4105 4106

	if (dir == ctx->dir && fs_path_len(name) == fs_path_len(ctx->name) &&
	    strncmp(name->start, ctx->name->start, fs_path_len(name)) == 0) {
4107 4108 4109 4110 4111 4112 4113 4114 4115 4116
		/*
		 * To avoid doing extra lookups we'll only do this if everything
		 * else matches.
		 */
		ret = get_inode_info(ctx->root, dir, NULL, &dir_gen, NULL,
				     NULL, NULL, NULL);
		if (ret)
			return ret;
		if (dir_gen != ctx->dir_gen)
			return 0;
4117 4118 4119 4120 4121 4122
		ctx->found_idx = num;
		return 1;
	}
	return 0;
}

4123
static int find_iref(struct btrfs_root *root,
4124 4125
		     struct btrfs_path *path,
		     struct btrfs_key *key,
4126
		     u64 dir, u64 dir_gen, struct fs_path *name)
4127 4128 4129 4130 4131 4132
{
	int ret;
	struct find_ref_ctx ctx;

	ctx.dir = dir;
	ctx.name = name;
4133
	ctx.dir_gen = dir_gen;
4134
	ctx.found_idx = -1;
4135
	ctx.root = root;
4136

4137
	ret = iterate_inode_ref(root, path, key, 0, __find_iref, &ctx);
4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150
	if (ret < 0)
		return ret;

	if (ctx.found_idx == -1)
		return -ENOENT;

	return ctx.found_idx;
}

static int __record_changed_new_ref(int num, u64 dir, int index,
				    struct fs_path *name,
				    void *ctx)
{
4151
	u64 dir_gen;
4152 4153 4154
	int ret;
	struct send_ctx *sctx = ctx;

4155 4156 4157 4158 4159
	ret = get_inode_info(sctx->send_root, dir, NULL, &dir_gen, NULL,
			     NULL, NULL, NULL);
	if (ret)
		return ret;

4160
	ret = find_iref(sctx->parent_root, sctx->right_path,
4161
			sctx->cmp_key, dir, dir_gen, name);
4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173
	if (ret == -ENOENT)
		ret = __record_new_ref(num, dir, index, name, sctx);
	else if (ret > 0)
		ret = 0;

	return ret;
}

static int __record_changed_deleted_ref(int num, u64 dir, int index,
					struct fs_path *name,
					void *ctx)
{
4174
	u64 dir_gen;
4175 4176 4177
	int ret;
	struct send_ctx *sctx = ctx;

4178 4179 4180 4181 4182
	ret = get_inode_info(sctx->parent_root, dir, NULL, &dir_gen, NULL,
			     NULL, NULL, NULL);
	if (ret)
		return ret;

4183
	ret = find_iref(sctx->send_root, sctx->left_path, sctx->cmp_key,
4184
			dir, dir_gen, name);
4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196
	if (ret == -ENOENT)
		ret = __record_deleted_ref(num, dir, index, name, sctx);
	else if (ret > 0)
		ret = 0;

	return ret;
}

static int record_changed_ref(struct send_ctx *sctx)
{
	int ret = 0;

4197
	ret = iterate_inode_ref(sctx->send_root, sctx->left_path,
4198 4199 4200
			sctx->cmp_key, 0, __record_changed_new_ref, sctx);
	if (ret < 0)
		goto out;
4201
	ret = iterate_inode_ref(sctx->parent_root, sctx->right_path,
4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223 4224 4225
			sctx->cmp_key, 0, __record_changed_deleted_ref, sctx);
	if (ret < 0)
		goto out;
	ret = 0;

out:
	return ret;
}

/*
 * Record and process all refs at once. Needed when an inode changes the
 * generation number, which means that it was deleted and recreated.
 */
static int process_all_refs(struct send_ctx *sctx,
			    enum btrfs_compare_tree_result cmd)
{
	int ret;
	struct btrfs_root *root;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct btrfs_key found_key;
	struct extent_buffer *eb;
	int slot;
	iterate_inode_ref_t cb;
4226
	int pending_move = 0;
4227 4228 4229 4230 4231 4232 4233 4234 4235 4236 4237 4238

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	if (cmd == BTRFS_COMPARE_TREE_NEW) {
		root = sctx->send_root;
		cb = __record_new_ref;
	} else if (cmd == BTRFS_COMPARE_TREE_DELETED) {
		root = sctx->parent_root;
		cb = __record_deleted_ref;
	} else {
4239 4240 4241 4242
		btrfs_err(sctx->send_root->fs_info,
				"Wrong command %d in process_all_refs", cmd);
		ret = -EINVAL;
		goto out;
4243 4244 4245 4246 4247
	}

	key.objectid = sctx->cmp_key->objectid;
	key.type = BTRFS_INODE_REF_KEY;
	key.offset = 0;
4248 4249 4250
	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto out;
4251

4252
	while (1) {
4253 4254
		eb = path->nodes[0];
		slot = path->slots[0];
4255 4256 4257 4258 4259 4260 4261 4262 4263
		if (slot >= btrfs_header_nritems(eb)) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0)
				goto out;
			else if (ret > 0)
				break;
			continue;
		}

4264 4265 4266
		btrfs_item_key_to_cpu(eb, &found_key, slot);

		if (found_key.objectid != key.objectid ||
4267 4268
		    (found_key.type != BTRFS_INODE_REF_KEY &&
		     found_key.type != BTRFS_INODE_EXTREF_KEY))
4269 4270
			break;

4271
		ret = iterate_inode_ref(root, path, &found_key, 0, cb, sctx);
4272 4273 4274
		if (ret < 0)
			goto out;

4275
		path->slots[0]++;
4276
	}
4277
	btrfs_release_path(path);
4278

4279 4280 4281 4282 4283
	/*
	 * We don't actually care about pending_move as we are simply
	 * re-creating this inode and will be rename'ing it into place once we
	 * rename the parent directory.
	 */
4284
	ret = process_recorded_refs(sctx, &pending_move);
4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 4335 4336 4337 4338 4339
out:
	btrfs_free_path(path);
	return ret;
}

static int send_set_xattr(struct send_ctx *sctx,
			  struct fs_path *path,
			  const char *name, int name_len,
			  const char *data, int data_len)
{
	int ret = 0;

	ret = begin_cmd(sctx, BTRFS_SEND_C_SET_XATTR);
	if (ret < 0)
		goto out;

	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path);
	TLV_PUT_STRING(sctx, BTRFS_SEND_A_XATTR_NAME, name, name_len);
	TLV_PUT(sctx, BTRFS_SEND_A_XATTR_DATA, data, data_len);

	ret = send_cmd(sctx);

tlv_put_failure:
out:
	return ret;
}

static int send_remove_xattr(struct send_ctx *sctx,
			  struct fs_path *path,
			  const char *name, int name_len)
{
	int ret = 0;

	ret = begin_cmd(sctx, BTRFS_SEND_C_REMOVE_XATTR);
	if (ret < 0)
		goto out;

	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path);
	TLV_PUT_STRING(sctx, BTRFS_SEND_A_XATTR_NAME, name, name_len);

	ret = send_cmd(sctx);

tlv_put_failure:
out:
	return ret;
}

static int __process_new_xattr(int num, struct btrfs_key *di_key,
			       const char *name, int name_len,
			       const char *data, int data_len,
			       u8 type, void *ctx)
{
	int ret;
	struct send_ctx *sctx = ctx;
	struct fs_path *p;
4340
	struct posix_acl_xattr_header dummy_acl;
4341

4342
	p = fs_path_alloc();
4343 4344 4345 4346
	if (!p)
		return -ENOMEM;

	/*
4347
	 * This hack is needed because empty acls are stored as zero byte
4348
	 * data in xattrs. Problem with that is, that receiving these zero byte
4349
	 * acls will fail later. To fix this, we send a dummy acl list that
4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368
	 * only contains the version number and no entries.
	 */
	if (!strncmp(name, XATTR_NAME_POSIX_ACL_ACCESS, name_len) ||
	    !strncmp(name, XATTR_NAME_POSIX_ACL_DEFAULT, name_len)) {
		if (data_len == 0) {
			dummy_acl.a_version =
					cpu_to_le32(POSIX_ACL_XATTR_VERSION);
			data = (char *)&dummy_acl;
			data_len = sizeof(dummy_acl);
		}
	}

	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
	if (ret < 0)
		goto out;

	ret = send_set_xattr(sctx, p, name, name_len, data, data_len);

out:
4369
	fs_path_free(p);
4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381
	return ret;
}

static int __process_deleted_xattr(int num, struct btrfs_key *di_key,
				   const char *name, int name_len,
				   const char *data, int data_len,
				   u8 type, void *ctx)
{
	int ret;
	struct send_ctx *sctx = ctx;
	struct fs_path *p;

4382
	p = fs_path_alloc();
4383 4384 4385 4386 4387 4388 4389 4390 4391 4392
	if (!p)
		return -ENOMEM;

	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
	if (ret < 0)
		goto out;

	ret = send_remove_xattr(sctx, p, name, name_len);

out:
4393
	fs_path_free(p);
4394 4395 4396 4397 4398 4399 4400
	return ret;
}

static int process_new_xattr(struct send_ctx *sctx)
{
	int ret = 0;

4401 4402
	ret = iterate_dir_item(sctx->send_root, sctx->left_path,
			       sctx->cmp_key, __process_new_xattr, sctx);
4403 4404 4405 4406 4407 4408

	return ret;
}

static int process_deleted_xattr(struct send_ctx *sctx)
{
4409 4410
	return iterate_dir_item(sctx->parent_root, sctx->right_path,
				sctx->cmp_key, __process_deleted_xattr, sctx);
4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 4429 4430 4431
}

struct find_xattr_ctx {
	const char *name;
	int name_len;
	int found_idx;
	char *found_data;
	int found_data_len;
};

static int __find_xattr(int num, struct btrfs_key *di_key,
			const char *name, int name_len,
			const char *data, int data_len,
			u8 type, void *vctx)
{
	struct find_xattr_ctx *ctx = vctx;

	if (name_len == ctx->name_len &&
	    strncmp(name, ctx->name, name_len) == 0) {
		ctx->found_idx = num;
		ctx->found_data_len = data_len;
4432
		ctx->found_data = kmemdup(data, data_len, GFP_KERNEL);
4433 4434 4435 4436 4437 4438 4439
		if (!ctx->found_data)
			return -ENOMEM;
		return 1;
	}
	return 0;
}

4440
static int find_xattr(struct btrfs_root *root,
4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454
		      struct btrfs_path *path,
		      struct btrfs_key *key,
		      const char *name, int name_len,
		      char **data, int *data_len)
{
	int ret;
	struct find_xattr_ctx ctx;

	ctx.name = name;
	ctx.name_len = name_len;
	ctx.found_idx = -1;
	ctx.found_data = NULL;
	ctx.found_data_len = 0;

4455
	ret = iterate_dir_item(root, path, key, __find_xattr, &ctx);
4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 4476 4477 4478 4479 4480
	if (ret < 0)
		return ret;

	if (ctx.found_idx == -1)
		return -ENOENT;
	if (data) {
		*data = ctx.found_data;
		*data_len = ctx.found_data_len;
	} else {
		kfree(ctx.found_data);
	}
	return ctx.found_idx;
}


static int __process_changed_new_xattr(int num, struct btrfs_key *di_key,
				       const char *name, int name_len,
				       const char *data, int data_len,
				       u8 type, void *ctx)
{
	int ret;
	struct send_ctx *sctx = ctx;
	char *found_data = NULL;
	int found_data_len  = 0;

4481 4482 4483
	ret = find_xattr(sctx->parent_root, sctx->right_path,
			 sctx->cmp_key, name, name_len, &found_data,
			 &found_data_len);
4484 4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508
	if (ret == -ENOENT) {
		ret = __process_new_xattr(num, di_key, name, name_len, data,
				data_len, type, ctx);
	} else if (ret >= 0) {
		if (data_len != found_data_len ||
		    memcmp(data, found_data, data_len)) {
			ret = __process_new_xattr(num, di_key, name, name_len,
					data, data_len, type, ctx);
		} else {
			ret = 0;
		}
	}

	kfree(found_data);
	return ret;
}

static int __process_changed_deleted_xattr(int num, struct btrfs_key *di_key,
					   const char *name, int name_len,
					   const char *data, int data_len,
					   u8 type, void *ctx)
{
	int ret;
	struct send_ctx *sctx = ctx;

4509 4510
	ret = find_xattr(sctx->send_root, sctx->left_path, sctx->cmp_key,
			 name, name_len, NULL, NULL);
4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522 4523
	if (ret == -ENOENT)
		ret = __process_deleted_xattr(num, di_key, name, name_len, data,
				data_len, type, ctx);
	else if (ret >= 0)
		ret = 0;

	return ret;
}

static int process_changed_xattr(struct send_ctx *sctx)
{
	int ret = 0;

4524
	ret = iterate_dir_item(sctx->send_root, sctx->left_path,
4525 4526 4527
			sctx->cmp_key, __process_changed_new_xattr, sctx);
	if (ret < 0)
		goto out;
4528
	ret = iterate_dir_item(sctx->parent_root, sctx->right_path,
4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553
			sctx->cmp_key, __process_changed_deleted_xattr, sctx);

out:
	return ret;
}

static int process_all_new_xattrs(struct send_ctx *sctx)
{
	int ret;
	struct btrfs_root *root;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct btrfs_key found_key;
	struct extent_buffer *eb;
	int slot;

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	root = sctx->send_root;

	key.objectid = sctx->cmp_key->objectid;
	key.type = BTRFS_XATTR_ITEM_KEY;
	key.offset = 0;
4554 4555 4556
	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto out;
4557

4558
	while (1) {
4559 4560
		eb = path->nodes[0];
		slot = path->slots[0];
4561 4562 4563 4564 4565 4566 4567 4568 4569 4570
		if (slot >= btrfs_header_nritems(eb)) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0) {
				goto out;
			} else if (ret > 0) {
				ret = 0;
				break;
			}
			continue;
		}
4571

4572
		btrfs_item_key_to_cpu(eb, &found_key, slot);
4573 4574 4575 4576 4577 4578
		if (found_key.objectid != key.objectid ||
		    found_key.type != key.type) {
			ret = 0;
			goto out;
		}

4579 4580
		ret = iterate_dir_item(root, path, &found_key,
				       __process_new_xattr, sctx);
4581 4582 4583
		if (ret < 0)
			goto out;

4584
		path->slots[0]++;
4585 4586 4587 4588 4589 4590 4591
	}

out:
	btrfs_free_path(path);
	return ret;
}

J
Josef Bacik 已提交
4592 4593 4594 4595 4596 4597 4598 4599
static ssize_t fill_read_buf(struct send_ctx *sctx, u64 offset, u32 len)
{
	struct btrfs_root *root = sctx->send_root;
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct inode *inode;
	struct page *page;
	char *addr;
	struct btrfs_key key;
4600
	pgoff_t index = offset >> PAGE_SHIFT;
J
Josef Bacik 已提交
4601
	pgoff_t last_index;
4602
	unsigned pg_offset = offset & ~PAGE_MASK;
J
Josef Bacik 已提交
4603 4604 4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 4619 4620 4621
	ssize_t ret = 0;

	key.objectid = sctx->cur_ino;
	key.type = BTRFS_INODE_ITEM_KEY;
	key.offset = 0;

	inode = btrfs_iget(fs_info->sb, &key, root, NULL);
	if (IS_ERR(inode))
		return PTR_ERR(inode);

	if (offset + len > i_size_read(inode)) {
		if (offset > i_size_read(inode))
			len = 0;
		else
			len = offset - i_size_read(inode);
	}
	if (len == 0)
		goto out;

4622
	last_index = (offset + len - 1) >> PAGE_SHIFT;
L
Liu Bo 已提交
4623 4624 4625 4626 4627 4628 4629

	/* initial readahead */
	memset(&sctx->ra, 0, sizeof(struct file_ra_state));
	file_ra_state_init(&sctx->ra, inode->i_mapping);
	btrfs_force_ra(inode->i_mapping, &sctx->ra, NULL, index,
		       last_index - index + 1);

J
Josef Bacik 已提交
4630 4631
	while (index <= last_index) {
		unsigned cur_len = min_t(unsigned, len,
4632
					 PAGE_SIZE - pg_offset);
4633
		page = find_or_create_page(inode->i_mapping, index, GFP_KERNEL);
J
Josef Bacik 已提交
4634 4635 4636 4637 4638 4639 4640 4641 4642 4643
		if (!page) {
			ret = -ENOMEM;
			break;
		}

		if (!PageUptodate(page)) {
			btrfs_readpage(NULL, page);
			lock_page(page);
			if (!PageUptodate(page)) {
				unlock_page(page);
4644
				put_page(page);
J
Josef Bacik 已提交
4645 4646 4647 4648 4649 4650 4651 4652 4653
				ret = -EIO;
				break;
			}
		}

		addr = kmap(page);
		memcpy(sctx->read_buf + ret, addr + pg_offset, cur_len);
		kunmap(page);
		unlock_page(page);
4654
		put_page(page);
J
Josef Bacik 已提交
4655 4656 4657 4658 4659 4660 4661 4662 4663 4664
		index++;
		pg_offset = 0;
		len -= cur_len;
		ret += cur_len;
	}
out:
	iput(inode);
	return ret;
}

4665 4666 4667 4668 4669 4670
/*
 * Read some bytes from the current inode/file and send a write command to
 * user space.
 */
static int send_write(struct send_ctx *sctx, u64 offset, u32 len)
{
4671
	struct btrfs_fs_info *fs_info = sctx->send_root->fs_info;
4672 4673
	int ret = 0;
	struct fs_path *p;
J
Josef Bacik 已提交
4674
	ssize_t num_read = 0;
4675

4676
	p = fs_path_alloc();
4677 4678 4679
	if (!p)
		return -ENOMEM;

4680
	btrfs_debug(fs_info, "send_write offset=%llu, len=%d", offset, len);
4681

J
Josef Bacik 已提交
4682 4683 4684 4685
	num_read = fill_read_buf(sctx, offset, len);
	if (num_read <= 0) {
		if (num_read < 0)
			ret = num_read;
4686
		goto out;
J
Josef Bacik 已提交
4687
	}
4688 4689 4690 4691 4692 4693 4694 4695 4696 4697 4698

	ret = begin_cmd(sctx, BTRFS_SEND_C_WRITE);
	if (ret < 0)
		goto out;

	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
	if (ret < 0)
		goto out;

	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
	TLV_PUT_U64(sctx, BTRFS_SEND_A_FILE_OFFSET, offset);
4699
	TLV_PUT(sctx, BTRFS_SEND_A_DATA, sctx->read_buf, num_read);
4700 4701 4702 4703 4704

	ret = send_cmd(sctx);

tlv_put_failure:
out:
4705
	fs_path_free(p);
4706 4707
	if (ret < 0)
		return ret;
4708
	return num_read;
4709 4710 4711 4712 4713 4714 4715 4716 4717 4718 4719 4720 4721
}

/*
 * Send a clone command to user space.
 */
static int send_clone(struct send_ctx *sctx,
		      u64 offset, u32 len,
		      struct clone_root *clone_root)
{
	int ret = 0;
	struct fs_path *p;
	u64 gen;

4722 4723 4724 4725
	btrfs_debug(sctx->send_root->fs_info,
		    "send_clone offset=%llu, len=%d, clone_root=%llu, clone_inode=%llu, clone_offset=%llu",
		    offset, len, clone_root->root->objectid, clone_root->ino,
		    clone_root->offset);
4726

4727
	p = fs_path_alloc();
4728 4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742
	if (!p)
		return -ENOMEM;

	ret = begin_cmd(sctx, BTRFS_SEND_C_CLONE);
	if (ret < 0)
		goto out;

	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
	if (ret < 0)
		goto out;

	TLV_PUT_U64(sctx, BTRFS_SEND_A_FILE_OFFSET, offset);
	TLV_PUT_U64(sctx, BTRFS_SEND_A_CLONE_LEN, len);
	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);

4743
	if (clone_root->root == sctx->send_root) {
4744
		ret = get_inode_info(sctx->send_root, clone_root->ino, NULL,
4745
				&gen, NULL, NULL, NULL, NULL);
4746 4747 4748 4749
		if (ret < 0)
			goto out;
		ret = get_cur_path(sctx, clone_root->ino, gen, p);
	} else {
4750
		ret = get_inode_path(clone_root->root, clone_root->ino, p);
4751 4752 4753 4754
	}
	if (ret < 0)
		goto out;

4755 4756 4757 4758 4759 4760 4761 4762 4763 4764 4765 4766 4767 4768 4769
	/*
	 * If the parent we're using has a received_uuid set then use that as
	 * our clone source as that is what we will look for when doing a
	 * receive.
	 *
	 * This covers the case that we create a snapshot off of a received
	 * subvolume and then use that as the parent and try to receive on a
	 * different host.
	 */
	if (!btrfs_is_empty_uuid(clone_root->root->root_item.received_uuid))
		TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID,
			     clone_root->root->root_item.received_uuid);
	else
		TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID,
			     clone_root->root->root_item.uuid);
4770
	TLV_PUT_U64(sctx, BTRFS_SEND_A_CLONE_CTRANSID,
4771
		    le64_to_cpu(clone_root->root->root_item.ctransid));
4772 4773 4774 4775 4776 4777 4778 4779
	TLV_PUT_PATH(sctx, BTRFS_SEND_A_CLONE_PATH, p);
	TLV_PUT_U64(sctx, BTRFS_SEND_A_CLONE_OFFSET,
			clone_root->offset);

	ret = send_cmd(sctx);

tlv_put_failure:
out:
4780
	fs_path_free(p);
4781 4782 4783
	return ret;
}

4784 4785 4786 4787 4788 4789 4790 4791 4792
/*
 * Send an update extent command to user space.
 */
static int send_update_extent(struct send_ctx *sctx,
			      u64 offset, u32 len)
{
	int ret = 0;
	struct fs_path *p;

4793
	p = fs_path_alloc();
4794 4795 4796 4797 4798 4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812
	if (!p)
		return -ENOMEM;

	ret = begin_cmd(sctx, BTRFS_SEND_C_UPDATE_EXTENT);
	if (ret < 0)
		goto out;

	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
	if (ret < 0)
		goto out;

	TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
	TLV_PUT_U64(sctx, BTRFS_SEND_A_FILE_OFFSET, offset);
	TLV_PUT_U64(sctx, BTRFS_SEND_A_SIZE, len);

	ret = send_cmd(sctx);

tlv_put_failure:
out:
4813
	fs_path_free(p);
4814 4815 4816
	return ret;
}

4817 4818 4819 4820 4821 4822 4823 4824 4825 4826
static int send_hole(struct send_ctx *sctx, u64 end)
{
	struct fs_path *p = NULL;
	u64 offset = sctx->cur_inode_last_extent;
	u64 len;
	int ret = 0;

	p = fs_path_alloc();
	if (!p)
		return -ENOMEM;
4827 4828 4829
	ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p);
	if (ret < 0)
		goto tlv_put_failure;
4830 4831 4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849
	memset(sctx->read_buf, 0, BTRFS_SEND_READ_SIZE);
	while (offset < end) {
		len = min_t(u64, end - offset, BTRFS_SEND_READ_SIZE);

		ret = begin_cmd(sctx, BTRFS_SEND_C_WRITE);
		if (ret < 0)
			break;
		TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p);
		TLV_PUT_U64(sctx, BTRFS_SEND_A_FILE_OFFSET, offset);
		TLV_PUT(sctx, BTRFS_SEND_A_DATA, sctx->read_buf, len);
		ret = send_cmd(sctx);
		if (ret < 0)
			break;
		offset += len;
	}
tlv_put_failure:
	fs_path_free(p);
	return ret;
}

4850 4851 4852 4853 4854 4855 4856 4857 4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947 4948 4949 4950 4951 4952 4953 4954 4955
static int send_extent_data(struct send_ctx *sctx,
			    const u64 offset,
			    const u64 len)
{
	u64 sent = 0;

	if (sctx->flags & BTRFS_SEND_FLAG_NO_FILE_DATA)
		return send_update_extent(sctx, offset, len);

	while (sent < len) {
		u64 size = len - sent;
		int ret;

		if (size > BTRFS_SEND_READ_SIZE)
			size = BTRFS_SEND_READ_SIZE;
		ret = send_write(sctx, offset + sent, size);
		if (ret < 0)
			return ret;
		if (!ret)
			break;
		sent += ret;
	}
	return 0;
}

static int clone_range(struct send_ctx *sctx,
		       struct clone_root *clone_root,
		       const u64 disk_byte,
		       u64 data_offset,
		       u64 offset,
		       u64 len)
{
	struct btrfs_path *path;
	struct btrfs_key key;
	int ret;

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	/*
	 * We can't send a clone operation for the entire range if we find
	 * extent items in the respective range in the source file that
	 * refer to different extents or if we find holes.
	 * So check for that and do a mix of clone and regular write/copy
	 * operations if needed.
	 *
	 * Example:
	 *
	 * mkfs.btrfs -f /dev/sda
	 * mount /dev/sda /mnt
	 * xfs_io -f -c "pwrite -S 0xaa 0K 100K" /mnt/foo
	 * cp --reflink=always /mnt/foo /mnt/bar
	 * xfs_io -c "pwrite -S 0xbb 50K 50K" /mnt/foo
	 * btrfs subvolume snapshot -r /mnt /mnt/snap
	 *
	 * If when we send the snapshot and we are processing file bar (which
	 * has a higher inode number than foo) we blindly send a clone operation
	 * for the [0, 100K[ range from foo to bar, the receiver ends up getting
	 * a file bar that matches the content of file foo - iow, doesn't match
	 * the content from bar in the original filesystem.
	 */
	key.objectid = clone_root->ino;
	key.type = BTRFS_EXTENT_DATA_KEY;
	key.offset = clone_root->offset;
	ret = btrfs_search_slot(NULL, clone_root->root, &key, path, 0, 0);
	if (ret < 0)
		goto out;
	if (ret > 0 && path->slots[0] > 0) {
		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0] - 1);
		if (key.objectid == clone_root->ino &&
		    key.type == BTRFS_EXTENT_DATA_KEY)
			path->slots[0]--;
	}

	while (true) {
		struct extent_buffer *leaf = path->nodes[0];
		int slot = path->slots[0];
		struct btrfs_file_extent_item *ei;
		u8 type;
		u64 ext_len;
		u64 clone_len;

		if (slot >= btrfs_header_nritems(leaf)) {
			ret = btrfs_next_leaf(clone_root->root, path);
			if (ret < 0)
				goto out;
			else if (ret > 0)
				break;
			continue;
		}

		btrfs_item_key_to_cpu(leaf, &key, slot);

		/*
		 * We might have an implicit trailing hole (NO_HOLES feature
		 * enabled). We deal with it after leaving this loop.
		 */
		if (key.objectid != clone_root->ino ||
		    key.type != BTRFS_EXTENT_DATA_KEY)
			break;

		ei = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item);
		type = btrfs_file_extent_type(leaf, ei);
		if (type == BTRFS_FILE_EXTENT_INLINE) {
			ext_len = btrfs_file_extent_inline_len(leaf, slot, ei);
4956
			ext_len = PAGE_ALIGN(ext_len);
4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988 4989 4990 4991 4992 4993 4994 4995 4996 4997 4998 4999 5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011 5012 5013 5014
		} else {
			ext_len = btrfs_file_extent_num_bytes(leaf, ei);
		}

		if (key.offset + ext_len <= clone_root->offset)
			goto next;

		if (key.offset > clone_root->offset) {
			/* Implicit hole, NO_HOLES feature enabled. */
			u64 hole_len = key.offset - clone_root->offset;

			if (hole_len > len)
				hole_len = len;
			ret = send_extent_data(sctx, offset, hole_len);
			if (ret < 0)
				goto out;

			len -= hole_len;
			if (len == 0)
				break;
			offset += hole_len;
			clone_root->offset += hole_len;
			data_offset += hole_len;
		}

		if (key.offset >= clone_root->offset + len)
			break;

		clone_len = min_t(u64, ext_len, len);

		if (btrfs_file_extent_disk_bytenr(leaf, ei) == disk_byte &&
		    btrfs_file_extent_offset(leaf, ei) == data_offset)
			ret = send_clone(sctx, offset, clone_len, clone_root);
		else
			ret = send_extent_data(sctx, offset, clone_len);

		if (ret < 0)
			goto out;

		len -= clone_len;
		if (len == 0)
			break;
		offset += clone_len;
		clone_root->offset += clone_len;
		data_offset += clone_len;
next:
		path->slots[0]++;
	}

	if (len > 0)
		ret = send_extent_data(sctx, offset, len);
	else
		ret = 0;
out:
	btrfs_free_path(path);
	return ret;
}

5015 5016 5017 5018 5019 5020 5021 5022 5023 5024
static int send_write_or_clone(struct send_ctx *sctx,
			       struct btrfs_path *path,
			       struct btrfs_key *key,
			       struct clone_root *clone_root)
{
	int ret = 0;
	struct btrfs_file_extent_item *ei;
	u64 offset = key->offset;
	u64 len;
	u8 type;
5025
	u64 bs = sctx->send_root->fs_info->sb->s_blocksize;
5026 5027 5028 5029

	ei = btrfs_item_ptr(path->nodes[0], path->slots[0],
			struct btrfs_file_extent_item);
	type = btrfs_file_extent_type(path->nodes[0], ei);
5030
	if (type == BTRFS_FILE_EXTENT_INLINE) {
5031 5032
		len = btrfs_file_extent_inline_len(path->nodes[0],
						   path->slots[0], ei);
5033 5034 5035 5036 5037
		/*
		 * it is possible the inline item won't cover the whole page,
		 * but there may be items after this page.  Make
		 * sure to send the whole thing
		 */
5038
		len = PAGE_ALIGN(len);
5039
	} else {
5040
		len = btrfs_file_extent_num_bytes(path->nodes[0], ei);
5041
	}
5042 5043 5044 5045 5046 5047 5048 5049

	if (offset + len > sctx->cur_inode_size)
		len = sctx->cur_inode_size - offset;
	if (len == 0) {
		ret = 0;
		goto out;
	}

5050
	if (clone_root && IS_ALIGNED(offset + len, bs)) {
5051 5052 5053 5054 5055 5056 5057
		u64 disk_byte;
		u64 data_offset;

		disk_byte = btrfs_file_extent_disk_bytenr(path->nodes[0], ei);
		data_offset = btrfs_file_extent_offset(path->nodes[0], ei);
		ret = clone_range(sctx, clone_root, disk_byte, data_offset,
				  offset, len);
5058
	} else {
5059
		ret = send_extent_data(sctx, offset, len);
5060 5061 5062 5063 5064 5065 5066 5067 5068 5069 5070 5071 5072 5073 5074 5075 5076 5077 5078 5079 5080 5081 5082
	}
out:
	return ret;
}

static int is_extent_unchanged(struct send_ctx *sctx,
			       struct btrfs_path *left_path,
			       struct btrfs_key *ekey)
{
	int ret = 0;
	struct btrfs_key key;
	struct btrfs_path *path = NULL;
	struct extent_buffer *eb;
	int slot;
	struct btrfs_key found_key;
	struct btrfs_file_extent_item *ei;
	u64 left_disknr;
	u64 right_disknr;
	u64 left_offset;
	u64 right_offset;
	u64 left_offset_fixed;
	u64 left_len;
	u64 right_len;
5083 5084
	u64 left_gen;
	u64 right_gen;
5085 5086 5087 5088 5089 5090 5091 5092 5093 5094 5095 5096 5097 5098 5099 5100
	u8 left_type;
	u8 right_type;

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	eb = left_path->nodes[0];
	slot = left_path->slots[0];
	ei = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
	left_type = btrfs_file_extent_type(eb, ei);

	if (left_type != BTRFS_FILE_EXTENT_REG) {
		ret = 0;
		goto out;
	}
5101 5102 5103 5104
	left_disknr = btrfs_file_extent_disk_bytenr(eb, ei);
	left_len = btrfs_file_extent_num_bytes(eb, ei);
	left_offset = btrfs_file_extent_offset(eb, ei);
	left_gen = btrfs_file_extent_generation(eb, ei);
5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145

	/*
	 * Following comments will refer to these graphics. L is the left
	 * extents which we are checking at the moment. 1-8 are the right
	 * extents that we iterate.
	 *
	 *       |-----L-----|
	 * |-1-|-2a-|-3-|-4-|-5-|-6-|
	 *
	 *       |-----L-----|
	 * |--1--|-2b-|...(same as above)
	 *
	 * Alternative situation. Happens on files where extents got split.
	 *       |-----L-----|
	 * |-----------7-----------|-6-|
	 *
	 * Alternative situation. Happens on files which got larger.
	 *       |-----L-----|
	 * |-8-|
	 * Nothing follows after 8.
	 */

	key.objectid = ekey->objectid;
	key.type = BTRFS_EXTENT_DATA_KEY;
	key.offset = ekey->offset;
	ret = btrfs_search_slot_for_read(sctx->parent_root, &key, path, 0, 0);
	if (ret < 0)
		goto out;
	if (ret) {
		ret = 0;
		goto out;
	}

	/*
	 * Handle special case where the right side has no extents at all.
	 */
	eb = path->nodes[0];
	slot = path->slots[0];
	btrfs_item_key_to_cpu(eb, &found_key, slot);
	if (found_key.objectid != key.objectid ||
	    found_key.type != key.type) {
5146 5147
		/* If we're a hole then just pretend nothing changed */
		ret = (left_disknr) ? 0 : 1;
5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162
		goto out;
	}

	/*
	 * We're now on 2a, 2b or 7.
	 */
	key = found_key;
	while (key.offset < ekey->offset + left_len) {
		ei = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
		right_type = btrfs_file_extent_type(eb, ei);
		if (right_type != BTRFS_FILE_EXTENT_REG) {
			ret = 0;
			goto out;
		}

5163 5164 5165 5166 5167
		right_disknr = btrfs_file_extent_disk_bytenr(eb, ei);
		right_len = btrfs_file_extent_num_bytes(eb, ei);
		right_offset = btrfs_file_extent_offset(eb, ei);
		right_gen = btrfs_file_extent_generation(eb, ei);

5168 5169 5170 5171
		/*
		 * Are we at extent 8? If yes, we know the extent is changed.
		 * This may only happen on the first iteration.
		 */
5172
		if (found_key.offset + right_len <= ekey->offset) {
5173 5174
			/* If we're a hole just pretend nothing changed */
			ret = (left_disknr) ? 0 : 1;
5175 5176 5177 5178 5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189
			goto out;
		}

		left_offset_fixed = left_offset;
		if (key.offset < ekey->offset) {
			/* Fix the right offset for 2a and 7. */
			right_offset += ekey->offset - key.offset;
		} else {
			/* Fix the left offset for all behind 2a and 2b */
			left_offset_fixed += key.offset - ekey->offset;
		}

		/*
		 * Check if we have the same extent.
		 */
5190
		if (left_disknr != right_disknr ||
5191 5192
		    left_offset_fixed != right_offset ||
		    left_gen != right_gen) {
5193 5194 5195 5196 5197 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207 5208 5209 5210 5211
			ret = 0;
			goto out;
		}

		/*
		 * Go to the next extent.
		 */
		ret = btrfs_next_item(sctx->parent_root, path);
		if (ret < 0)
			goto out;
		if (!ret) {
			eb = path->nodes[0];
			slot = path->slots[0];
			btrfs_item_key_to_cpu(eb, &found_key, slot);
		}
		if (ret || found_key.objectid != key.objectid ||
		    found_key.type != key.type) {
			key.offset += right_len;
			break;
5212 5213 5214 5215
		}
		if (found_key.offset != key.offset + right_len) {
			ret = 0;
			goto out;
5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 5230 5231 5232 5233 5234
		}
		key = found_key;
	}

	/*
	 * We're now behind the left extent (treat as unchanged) or at the end
	 * of the right side (treat as changed).
	 */
	if (key.offset >= ekey->offset + left_len)
		ret = 1;
	else
		ret = 0;


out:
	btrfs_free_path(path);
	return ret;
}

5235 5236 5237 5238 5239 5240 5241 5242 5243 5244 5245 5246 5247 5248 5249 5250 5251 5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264 5265
static int get_last_extent(struct send_ctx *sctx, u64 offset)
{
	struct btrfs_path *path;
	struct btrfs_root *root = sctx->send_root;
	struct btrfs_file_extent_item *fi;
	struct btrfs_key key;
	u64 extent_end;
	u8 type;
	int ret;

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	sctx->cur_inode_last_extent = 0;

	key.objectid = sctx->cur_ino;
	key.type = BTRFS_EXTENT_DATA_KEY;
	key.offset = offset;
	ret = btrfs_search_slot_for_read(root, &key, path, 0, 1);
	if (ret < 0)
		goto out;
	ret = 0;
	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
	if (key.objectid != sctx->cur_ino || key.type != BTRFS_EXTENT_DATA_KEY)
		goto out;

	fi = btrfs_item_ptr(path->nodes[0], path->slots[0],
			    struct btrfs_file_extent_item);
	type = btrfs_file_extent_type(path->nodes[0], fi);
	if (type == BTRFS_FILE_EXTENT_INLINE) {
5266 5267
		u64 size = btrfs_file_extent_inline_len(path->nodes[0],
							path->slots[0], fi);
5268
		extent_end = ALIGN(key.offset + size,
5269
				   sctx->send_root->fs_info->sectorsize);
5270 5271 5272 5273 5274 5275 5276 5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293 5294 5295 5296 5297 5298 5299 5300
	} else {
		extent_end = key.offset +
			btrfs_file_extent_num_bytes(path->nodes[0], fi);
	}
	sctx->cur_inode_last_extent = extent_end;
out:
	btrfs_free_path(path);
	return ret;
}

static int maybe_send_hole(struct send_ctx *sctx, struct btrfs_path *path,
			   struct btrfs_key *key)
{
	struct btrfs_file_extent_item *fi;
	u64 extent_end;
	u8 type;
	int ret = 0;

	if (sctx->cur_ino != key->objectid || !need_send_hole(sctx))
		return 0;

	if (sctx->cur_inode_last_extent == (u64)-1) {
		ret = get_last_extent(sctx, key->offset - 1);
		if (ret)
			return ret;
	}

	fi = btrfs_item_ptr(path->nodes[0], path->slots[0],
			    struct btrfs_file_extent_item);
	type = btrfs_file_extent_type(path->nodes[0], fi);
	if (type == BTRFS_FILE_EXTENT_INLINE) {
5301 5302
		u64 size = btrfs_file_extent_inline_len(path->nodes[0],
							path->slots[0], fi);
5303
		extent_end = ALIGN(key->offset + size,
5304
				   sctx->send_root->fs_info->sectorsize);
5305 5306 5307 5308
	} else {
		extent_end = key->offset +
			btrfs_file_extent_num_bytes(path->nodes[0], fi);
	}
5309 5310 5311 5312 5313 5314 5315 5316 5317 5318 5319 5320 5321 5322 5323

	if (path->slots[0] == 0 &&
	    sctx->cur_inode_last_extent < key->offset) {
		/*
		 * We might have skipped entire leafs that contained only
		 * file extent items for our current inode. These leafs have
		 * a generation number smaller (older) than the one in the
		 * current leaf and the leaf our last extent came from, and
		 * are located between these 2 leafs.
		 */
		ret = get_last_extent(sctx, key->offset - 1);
		if (ret)
			return ret;
	}

5324 5325 5326 5327 5328 5329
	if (sctx->cur_inode_last_extent < key->offset)
		ret = send_hole(sctx, key->offset);
	sctx->cur_inode_last_extent = extent_end;
	return ret;
}

5330 5331 5332 5333 5334
static int process_extent(struct send_ctx *sctx,
			  struct btrfs_path *path,
			  struct btrfs_key *key)
{
	struct clone_root *found_clone = NULL;
5335
	int ret = 0;
5336 5337 5338 5339 5340 5341 5342 5343 5344 5345

	if (S_ISLNK(sctx->cur_inode_mode))
		return 0;

	if (sctx->parent_root && !sctx->cur_inode_new) {
		ret = is_extent_unchanged(sctx, path, key);
		if (ret < 0)
			goto out;
		if (ret) {
			ret = 0;
5346
			goto out_hole;
5347
		}
5348 5349 5350 5351 5352 5353 5354 5355 5356 5357 5358 5359 5360 5361 5362 5363 5364 5365 5366 5367 5368 5369 5370 5371 5372 5373
	} else {
		struct btrfs_file_extent_item *ei;
		u8 type;

		ei = btrfs_item_ptr(path->nodes[0], path->slots[0],
				    struct btrfs_file_extent_item);
		type = btrfs_file_extent_type(path->nodes[0], ei);
		if (type == BTRFS_FILE_EXTENT_PREALLOC ||
		    type == BTRFS_FILE_EXTENT_REG) {
			/*
			 * The send spec does not have a prealloc command yet,
			 * so just leave a hole for prealloc'ed extents until
			 * we have enough commands queued up to justify rev'ing
			 * the send spec.
			 */
			if (type == BTRFS_FILE_EXTENT_PREALLOC) {
				ret = 0;
				goto out;
			}

			/* Have a hole, just skip it. */
			if (btrfs_file_extent_disk_bytenr(path->nodes[0], ei) == 0) {
				ret = 0;
				goto out;
			}
		}
5374 5375 5376 5377 5378 5379 5380 5381
	}

	ret = find_extent_clone(sctx, path, key->objectid, key->offset,
			sctx->cur_inode_size, &found_clone);
	if (ret != -ENOENT && ret < 0)
		goto out;

	ret = send_write_or_clone(sctx, path, key, found_clone);
5382 5383 5384 5385
	if (ret)
		goto out;
out_hole:
	ret = maybe_send_hole(sctx, path, key);
5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400 5401 5402 5403 5404 5405 5406 5407
out:
	return ret;
}

static int process_all_extents(struct send_ctx *sctx)
{
	int ret;
	struct btrfs_root *root;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct btrfs_key found_key;
	struct extent_buffer *eb;
	int slot;

	root = sctx->send_root;
	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	key.objectid = sctx->cmp_key->objectid;
	key.type = BTRFS_EXTENT_DATA_KEY;
	key.offset = 0;
5408 5409 5410
	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto out;
5411

5412
	while (1) {
5413 5414
		eb = path->nodes[0];
		slot = path->slots[0];
5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426

		if (slot >= btrfs_header_nritems(eb)) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0) {
				goto out;
			} else if (ret > 0) {
				ret = 0;
				break;
			}
			continue;
		}

5427 5428 5429 5430 5431 5432 5433 5434 5435 5436 5437 5438
		btrfs_item_key_to_cpu(eb, &found_key, slot);

		if (found_key.objectid != key.objectid ||
		    found_key.type != key.type) {
			ret = 0;
			goto out;
		}

		ret = process_extent(sctx, path, &found_key);
		if (ret < 0)
			goto out;

5439
		path->slots[0]++;
5440 5441 5442 5443 5444 5445 5446
	}

out:
	btrfs_free_path(path);
	return ret;
}

5447 5448 5449
static int process_recorded_refs_if_needed(struct send_ctx *sctx, int at_end,
					   int *pending_move,
					   int *refs_processed)
5450 5451 5452 5453 5454 5455
{
	int ret = 0;

	if (sctx->cur_ino == 0)
		goto out;
	if (!at_end && sctx->cur_ino == sctx->cmp_key->objectid &&
5456
	    sctx->cmp_key->type <= BTRFS_INODE_EXTREF_KEY)
5457 5458 5459 5460
		goto out;
	if (list_empty(&sctx->new_refs) && list_empty(&sctx->deleted_refs))
		goto out;

5461
	ret = process_recorded_refs(sctx, pending_move);
5462 5463 5464
	if (ret < 0)
		goto out;

5465
	*refs_processed = 1;
5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480
out:
	return ret;
}

static int finish_inode_if_needed(struct send_ctx *sctx, int at_end)
{
	int ret = 0;
	u64 left_mode;
	u64 left_uid;
	u64 left_gid;
	u64 right_mode;
	u64 right_uid;
	u64 right_gid;
	int need_chmod = 0;
	int need_chown = 0;
5481 5482
	int pending_move = 0;
	int refs_processed = 0;
5483

5484 5485
	ret = process_recorded_refs_if_needed(sctx, at_end, &pending_move,
					      &refs_processed);
5486 5487 5488
	if (ret < 0)
		goto out;

5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503
	/*
	 * We have processed the refs and thus need to advance send_progress.
	 * Now, calls to get_cur_xxx will take the updated refs of the current
	 * inode into account.
	 *
	 * On the other hand, if our current inode is a directory and couldn't
	 * be moved/renamed because its parent was renamed/moved too and it has
	 * a higher inode number, we can only move/rename our current inode
	 * after we moved/renamed its parent. Therefore in this case operate on
	 * the old path (pre move/rename) of our current inode, and the
	 * move/rename will be performed later.
	 */
	if (refs_processed && !pending_move)
		sctx->send_progress = sctx->cur_ino + 1;

5504 5505 5506 5507 5508 5509
	if (sctx->cur_ino == 0 || sctx->cur_inode_deleted)
		goto out;
	if (!at_end && sctx->cmp_key->objectid == sctx->cur_ino)
		goto out;

	ret = get_inode_info(sctx->send_root, sctx->cur_ino, NULL, NULL,
5510
			&left_mode, &left_uid, &left_gid, NULL);
5511 5512 5513
	if (ret < 0)
		goto out;

5514 5515 5516
	if (!sctx->parent_root || sctx->cur_inode_new) {
		need_chown = 1;
		if (!S_ISLNK(sctx->cur_inode_mode))
5517
			need_chmod = 1;
5518 5519 5520 5521 5522 5523
	} else {
		ret = get_inode_info(sctx->parent_root, sctx->cur_ino,
				NULL, NULL, &right_mode, &right_uid,
				&right_gid, NULL);
		if (ret < 0)
			goto out;
5524

5525 5526 5527 5528
		if (left_uid != right_uid || left_gid != right_gid)
			need_chown = 1;
		if (!S_ISLNK(sctx->cur_inode_mode) && left_mode != right_mode)
			need_chmod = 1;
5529 5530 5531
	}

	if (S_ISREG(sctx->cur_inode_mode)) {
5532
		if (need_send_hole(sctx)) {
5533 5534 5535
			if (sctx->cur_inode_last_extent == (u64)-1 ||
			    sctx->cur_inode_last_extent <
			    sctx->cur_inode_size) {
5536 5537 5538 5539 5540 5541 5542 5543 5544 5545 5546
				ret = get_last_extent(sctx, (u64)-1);
				if (ret)
					goto out;
			}
			if (sctx->cur_inode_last_extent <
			    sctx->cur_inode_size) {
				ret = send_hole(sctx, sctx->cur_inode_size);
				if (ret)
					goto out;
			}
		}
5547 5548 5549 5550 5551 5552 5553 5554 5555 5556 5557 5558 5559 5560 5561 5562 5563 5564 5565 5566
		ret = send_truncate(sctx, sctx->cur_ino, sctx->cur_inode_gen,
				sctx->cur_inode_size);
		if (ret < 0)
			goto out;
	}

	if (need_chown) {
		ret = send_chown(sctx, sctx->cur_ino, sctx->cur_inode_gen,
				left_uid, left_gid);
		if (ret < 0)
			goto out;
	}
	if (need_chmod) {
		ret = send_chmod(sctx, sctx->cur_ino, sctx->cur_inode_gen,
				left_mode);
		if (ret < 0)
			goto out;
	}

	/*
5567 5568
	 * If other directory inodes depended on our current directory
	 * inode's move/rename, now do their move/rename operations.
5569
	 */
5570 5571 5572 5573
	if (!is_waiting_for_move(sctx, sctx->cur_ino)) {
		ret = apply_children_dir_moves(sctx);
		if (ret)
			goto out;
5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584
		/*
		 * Need to send that every time, no matter if it actually
		 * changed between the two trees as we have done changes to
		 * the inode before. If our inode is a directory and it's
		 * waiting to be moved/renamed, we will send its utimes when
		 * it's moved/renamed, therefore we don't need to do it here.
		 */
		sctx->send_progress = sctx->cur_ino + 1;
		ret = send_utimes(sctx, sctx->cur_ino, sctx->cur_inode_gen);
		if (ret < 0)
			goto out;
5585 5586
	}

5587 5588 5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601 5602
out:
	return ret;
}

static int changed_inode(struct send_ctx *sctx,
			 enum btrfs_compare_tree_result result)
{
	int ret = 0;
	struct btrfs_key *key = sctx->cmp_key;
	struct btrfs_inode_item *left_ii = NULL;
	struct btrfs_inode_item *right_ii = NULL;
	u64 left_gen = 0;
	u64 right_gen = 0;

	sctx->cur_ino = key->objectid;
	sctx->cur_inode_new_gen = 0;
5603
	sctx->cur_inode_last_extent = (u64)-1;
5604 5605 5606 5607 5608 5609

	/*
	 * Set send_progress to current inode. This will tell all get_cur_xxx
	 * functions that the current inode's refs are not updated yet. Later,
	 * when process_recorded_refs is finished, it is set to cur_ino + 1.
	 */
5610 5611 5612 5613 5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624 5625 5626 5627 5628 5629 5630 5631 5632
	sctx->send_progress = sctx->cur_ino;

	if (result == BTRFS_COMPARE_TREE_NEW ||
	    result == BTRFS_COMPARE_TREE_CHANGED) {
		left_ii = btrfs_item_ptr(sctx->left_path->nodes[0],
				sctx->left_path->slots[0],
				struct btrfs_inode_item);
		left_gen = btrfs_inode_generation(sctx->left_path->nodes[0],
				left_ii);
	} else {
		right_ii = btrfs_item_ptr(sctx->right_path->nodes[0],
				sctx->right_path->slots[0],
				struct btrfs_inode_item);
		right_gen = btrfs_inode_generation(sctx->right_path->nodes[0],
				right_ii);
	}
	if (result == BTRFS_COMPARE_TREE_CHANGED) {
		right_ii = btrfs_item_ptr(sctx->right_path->nodes[0],
				sctx->right_path->slots[0],
				struct btrfs_inode_item);

		right_gen = btrfs_inode_generation(sctx->right_path->nodes[0],
				right_ii);
5633 5634 5635 5636 5637 5638 5639 5640

		/*
		 * The cur_ino = root dir case is special here. We can't treat
		 * the inode as deleted+reused because it would generate a
		 * stream that tries to delete/mkdir the root dir.
		 */
		if (left_gen != right_gen &&
		    sctx->cur_ino != BTRFS_FIRST_FREE_OBJECTID)
5641 5642 5643 5644 5645 5646 5647 5648 5649 5650 5651
			sctx->cur_inode_new_gen = 1;
	}

	if (result == BTRFS_COMPARE_TREE_NEW) {
		sctx->cur_inode_gen = left_gen;
		sctx->cur_inode_new = 1;
		sctx->cur_inode_deleted = 0;
		sctx->cur_inode_size = btrfs_inode_size(
				sctx->left_path->nodes[0], left_ii);
		sctx->cur_inode_mode = btrfs_inode_mode(
				sctx->left_path->nodes[0], left_ii);
L
Liu Bo 已提交
5652 5653
		sctx->cur_inode_rdev = btrfs_inode_rdev(
				sctx->left_path->nodes[0], left_ii);
5654
		if (sctx->cur_ino != BTRFS_FIRST_FREE_OBJECTID)
5655
			ret = send_create_inode_if_needed(sctx);
5656 5657 5658 5659 5660 5661 5662 5663 5664
	} else if (result == BTRFS_COMPARE_TREE_DELETED) {
		sctx->cur_inode_gen = right_gen;
		sctx->cur_inode_new = 0;
		sctx->cur_inode_deleted = 1;
		sctx->cur_inode_size = btrfs_inode_size(
				sctx->right_path->nodes[0], right_ii);
		sctx->cur_inode_mode = btrfs_inode_mode(
				sctx->right_path->nodes[0], right_ii);
	} else if (result == BTRFS_COMPARE_TREE_CHANGED) {
5665 5666 5667 5668 5669 5670 5671
		/*
		 * We need to do some special handling in case the inode was
		 * reported as changed with a changed generation number. This
		 * means that the original inode was deleted and new inode
		 * reused the same inum. So we have to treat the old inode as
		 * deleted and the new one as new.
		 */
5672
		if (sctx->cur_inode_new_gen) {
5673 5674 5675
			/*
			 * First, process the inode as if it was deleted.
			 */
5676 5677 5678 5679 5680 5681 5682 5683 5684 5685 5686 5687
			sctx->cur_inode_gen = right_gen;
			sctx->cur_inode_new = 0;
			sctx->cur_inode_deleted = 1;
			sctx->cur_inode_size = btrfs_inode_size(
					sctx->right_path->nodes[0], right_ii);
			sctx->cur_inode_mode = btrfs_inode_mode(
					sctx->right_path->nodes[0], right_ii);
			ret = process_all_refs(sctx,
					BTRFS_COMPARE_TREE_DELETED);
			if (ret < 0)
				goto out;

5688 5689 5690
			/*
			 * Now process the inode as if it was new.
			 */
5691 5692 5693 5694 5695 5696 5697
			sctx->cur_inode_gen = left_gen;
			sctx->cur_inode_new = 1;
			sctx->cur_inode_deleted = 0;
			sctx->cur_inode_size = btrfs_inode_size(
					sctx->left_path->nodes[0], left_ii);
			sctx->cur_inode_mode = btrfs_inode_mode(
					sctx->left_path->nodes[0], left_ii);
L
Liu Bo 已提交
5698 5699
			sctx->cur_inode_rdev = btrfs_inode_rdev(
					sctx->left_path->nodes[0], left_ii);
5700
			ret = send_create_inode_if_needed(sctx);
5701 5702 5703 5704 5705 5706
			if (ret < 0)
				goto out;

			ret = process_all_refs(sctx, BTRFS_COMPARE_TREE_NEW);
			if (ret < 0)
				goto out;
5707 5708 5709 5710 5711
			/*
			 * Advance send_progress now as we did not get into
			 * process_recorded_refs_if_needed in the new_gen case.
			 */
			sctx->send_progress = sctx->cur_ino + 1;
5712 5713 5714 5715 5716

			/*
			 * Now process all extents and xattrs of the inode as if
			 * they were all new.
			 */
5717 5718 5719 5720 5721 5722 5723 5724 5725 5726 5727 5728 5729 5730 5731 5732 5733 5734 5735 5736 5737 5738
			ret = process_all_extents(sctx);
			if (ret < 0)
				goto out;
			ret = process_all_new_xattrs(sctx);
			if (ret < 0)
				goto out;
		} else {
			sctx->cur_inode_gen = left_gen;
			sctx->cur_inode_new = 0;
			sctx->cur_inode_new_gen = 0;
			sctx->cur_inode_deleted = 0;
			sctx->cur_inode_size = btrfs_inode_size(
					sctx->left_path->nodes[0], left_ii);
			sctx->cur_inode_mode = btrfs_inode_mode(
					sctx->left_path->nodes[0], left_ii);
		}
	}

out:
	return ret;
}

5739 5740 5741 5742 5743 5744 5745 5746 5747 5748
/*
 * We have to process new refs before deleted refs, but compare_trees gives us
 * the new and deleted refs mixed. To fix this, we record the new/deleted refs
 * first and later process them in process_recorded_refs.
 * For the cur_inode_new_gen case, we skip recording completely because
 * changed_inode did already initiate processing of refs. The reason for this is
 * that in this case, compare_tree actually compares the refs of 2 different
 * inodes. To fix this, process_all_refs is used in changed_inode to handle all
 * refs of the right tree as deleted and all refs of the left tree as new.
 */
5749 5750 5751 5752 5753
static int changed_ref(struct send_ctx *sctx,
		       enum btrfs_compare_tree_result result)
{
	int ret = 0;

5754 5755 5756 5757
	if (sctx->cur_ino != sctx->cmp_key->objectid) {
		inconsistent_snapshot_error(sctx, result, "reference");
		return -EIO;
	}
5758 5759 5760 5761 5762 5763 5764 5765 5766 5767 5768 5769 5770 5771

	if (!sctx->cur_inode_new_gen &&
	    sctx->cur_ino != BTRFS_FIRST_FREE_OBJECTID) {
		if (result == BTRFS_COMPARE_TREE_NEW)
			ret = record_new_ref(sctx);
		else if (result == BTRFS_COMPARE_TREE_DELETED)
			ret = record_deleted_ref(sctx);
		else if (result == BTRFS_COMPARE_TREE_CHANGED)
			ret = record_changed_ref(sctx);
	}

	return ret;
}

5772 5773 5774 5775 5776
/*
 * Process new/deleted/changed xattrs. We skip processing in the
 * cur_inode_new_gen case because changed_inode did already initiate processing
 * of xattrs. The reason is the same as in changed_ref
 */
5777 5778 5779 5780 5781
static int changed_xattr(struct send_ctx *sctx,
			 enum btrfs_compare_tree_result result)
{
	int ret = 0;

5782 5783 5784 5785
	if (sctx->cur_ino != sctx->cmp_key->objectid) {
		inconsistent_snapshot_error(sctx, result, "xattr");
		return -EIO;
	}
5786 5787 5788 5789 5790 5791 5792 5793 5794 5795 5796 5797 5798

	if (!sctx->cur_inode_new_gen && !sctx->cur_inode_deleted) {
		if (result == BTRFS_COMPARE_TREE_NEW)
			ret = process_new_xattr(sctx);
		else if (result == BTRFS_COMPARE_TREE_DELETED)
			ret = process_deleted_xattr(sctx);
		else if (result == BTRFS_COMPARE_TREE_CHANGED)
			ret = process_changed_xattr(sctx);
	}

	return ret;
}

5799 5800 5801 5802 5803
/*
 * Process new/deleted/changed extents. We skip processing in the
 * cur_inode_new_gen case because changed_inode did already initiate processing
 * of extents. The reason is the same as in changed_ref
 */
5804 5805 5806 5807 5808
static int changed_extent(struct send_ctx *sctx,
			  enum btrfs_compare_tree_result result)
{
	int ret = 0;

5809
	if (sctx->cur_ino != sctx->cmp_key->objectid) {
5810 5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824 5825 5826 5827 5828 5829 5830 5831 5832 5833 5834 5835 5836 5837 5838 5839 5840 5841 5842 5843 5844 5845 5846 5847 5848 5849 5850 5851 5852 5853 5854 5855 5856 5857 5858 5859 5860 5861 5862 5863 5864 5865 5866 5867

		if (result == BTRFS_COMPARE_TREE_CHANGED) {
			struct extent_buffer *leaf_l;
			struct extent_buffer *leaf_r;
			struct btrfs_file_extent_item *ei_l;
			struct btrfs_file_extent_item *ei_r;

			leaf_l = sctx->left_path->nodes[0];
			leaf_r = sctx->right_path->nodes[0];
			ei_l = btrfs_item_ptr(leaf_l,
					      sctx->left_path->slots[0],
					      struct btrfs_file_extent_item);
			ei_r = btrfs_item_ptr(leaf_r,
					      sctx->right_path->slots[0],
					      struct btrfs_file_extent_item);

			/*
			 * We may have found an extent item that has changed
			 * only its disk_bytenr field and the corresponding
			 * inode item was not updated. This case happens due to
			 * very specific timings during relocation when a leaf
			 * that contains file extent items is COWed while
			 * relocation is ongoing and its in the stage where it
			 * updates data pointers. So when this happens we can
			 * safely ignore it since we know it's the same extent,
			 * but just at different logical and physical locations
			 * (when an extent is fully replaced with a new one, we
			 * know the generation number must have changed too,
			 * since snapshot creation implies committing the current
			 * transaction, and the inode item must have been updated
			 * as well).
			 * This replacement of the disk_bytenr happens at
			 * relocation.c:replace_file_extents() through
			 * relocation.c:btrfs_reloc_cow_block().
			 */
			if (btrfs_file_extent_generation(leaf_l, ei_l) ==
			    btrfs_file_extent_generation(leaf_r, ei_r) &&
			    btrfs_file_extent_ram_bytes(leaf_l, ei_l) ==
			    btrfs_file_extent_ram_bytes(leaf_r, ei_r) &&
			    btrfs_file_extent_compression(leaf_l, ei_l) ==
			    btrfs_file_extent_compression(leaf_r, ei_r) &&
			    btrfs_file_extent_encryption(leaf_l, ei_l) ==
			    btrfs_file_extent_encryption(leaf_r, ei_r) &&
			    btrfs_file_extent_other_encoding(leaf_l, ei_l) ==
			    btrfs_file_extent_other_encoding(leaf_r, ei_r) &&
			    btrfs_file_extent_type(leaf_l, ei_l) ==
			    btrfs_file_extent_type(leaf_r, ei_r) &&
			    btrfs_file_extent_disk_bytenr(leaf_l, ei_l) !=
			    btrfs_file_extent_disk_bytenr(leaf_r, ei_r) &&
			    btrfs_file_extent_disk_num_bytes(leaf_l, ei_l) ==
			    btrfs_file_extent_disk_num_bytes(leaf_r, ei_r) &&
			    btrfs_file_extent_offset(leaf_l, ei_l) ==
			    btrfs_file_extent_offset(leaf_r, ei_r) &&
			    btrfs_file_extent_num_bytes(leaf_l, ei_l) ==
			    btrfs_file_extent_num_bytes(leaf_r, ei_r))
				return 0;
		}

5868 5869 5870
		inconsistent_snapshot_error(sctx, result, "extent");
		return -EIO;
	}
5871 5872 5873 5874 5875 5876 5877 5878 5879 5880

	if (!sctx->cur_inode_new_gen && !sctx->cur_inode_deleted) {
		if (result != BTRFS_COMPARE_TREE_DELETED)
			ret = process_extent(sctx, sctx->left_path,
					sctx->cmp_key);
	}

	return ret;
}

5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892 5893 5894 5895 5896 5897 5898 5899 5900 5901 5902 5903 5904 5905 5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931 5932 5933 5934 5935 5936 5937 5938
static int dir_changed(struct send_ctx *sctx, u64 dir)
{
	u64 orig_gen, new_gen;
	int ret;

	ret = get_inode_info(sctx->send_root, dir, NULL, &new_gen, NULL, NULL,
			     NULL, NULL);
	if (ret)
		return ret;

	ret = get_inode_info(sctx->parent_root, dir, NULL, &orig_gen, NULL,
			     NULL, NULL, NULL);
	if (ret)
		return ret;

	return (orig_gen != new_gen) ? 1 : 0;
}

static int compare_refs(struct send_ctx *sctx, struct btrfs_path *path,
			struct btrfs_key *key)
{
	struct btrfs_inode_extref *extref;
	struct extent_buffer *leaf;
	u64 dirid = 0, last_dirid = 0;
	unsigned long ptr;
	u32 item_size;
	u32 cur_offset = 0;
	int ref_name_len;
	int ret = 0;

	/* Easy case, just check this one dirid */
	if (key->type == BTRFS_INODE_REF_KEY) {
		dirid = key->offset;

		ret = dir_changed(sctx, dirid);
		goto out;
	}

	leaf = path->nodes[0];
	item_size = btrfs_item_size_nr(leaf, path->slots[0]);
	ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
	while (cur_offset < item_size) {
		extref = (struct btrfs_inode_extref *)(ptr +
						       cur_offset);
		dirid = btrfs_inode_extref_parent(leaf, extref);
		ref_name_len = btrfs_inode_extref_name_len(leaf, extref);
		cur_offset += ref_name_len + sizeof(*extref);
		if (dirid == last_dirid)
			continue;
		ret = dir_changed(sctx, dirid);
		if (ret)
			break;
		last_dirid = dirid;
	}
out:
	return ret;
}

5939 5940 5941 5942
/*
 * Updates compare related fields in sctx and simply forwards to the actual
 * changed_xxx functions.
 */
5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953
static int changed_cb(struct btrfs_root *left_root,
		      struct btrfs_root *right_root,
		      struct btrfs_path *left_path,
		      struct btrfs_path *right_path,
		      struct btrfs_key *key,
		      enum btrfs_compare_tree_result result,
		      void *ctx)
{
	int ret = 0;
	struct send_ctx *sctx = ctx;

5954
	if (result == BTRFS_COMPARE_TREE_SAME) {
5955 5956 5957 5958 5959 5960 5961 5962 5963 5964
		if (key->type == BTRFS_INODE_REF_KEY ||
		    key->type == BTRFS_INODE_EXTREF_KEY) {
			ret = compare_refs(sctx, left_path, key);
			if (!ret)
				return 0;
			if (ret < 0)
				return ret;
		} else if (key->type == BTRFS_EXTENT_DATA_KEY) {
			return maybe_send_hole(sctx, left_path, key);
		} else {
5965
			return 0;
5966
		}
5967 5968 5969 5970
		result = BTRFS_COMPARE_TREE_CHANGED;
		ret = 0;
	}

5971 5972 5973 5974 5975 5976 5977 5978
	sctx->left_path = left_path;
	sctx->right_path = right_path;
	sctx->cmp_key = key;

	ret = finish_inode_if_needed(sctx, 0);
	if (ret < 0)
		goto out;

5979 5980 5981 5982 5983
	/* Ignore non-FS objects */
	if (key->objectid == BTRFS_FREE_INO_OBJECTID ||
	    key->objectid == BTRFS_FREE_SPACE_OBJECTID)
		goto out;

5984 5985
	if (key->type == BTRFS_INODE_ITEM_KEY)
		ret = changed_inode(sctx, result);
5986 5987
	else if (key->type == BTRFS_INODE_REF_KEY ||
		 key->type == BTRFS_INODE_EXTREF_KEY)
5988 5989 5990 5991 5992 5993 5994 5995 5996 5997 5998 5999 6000 6001 6002 6003 6004 6005 6006 6007 6008 6009 6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021 6022 6023 6024 6025 6026 6027 6028 6029 6030 6031 6032 6033 6034 6035 6036 6037 6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051 6052 6053 6054 6055 6056
		ret = changed_ref(sctx, result);
	else if (key->type == BTRFS_XATTR_ITEM_KEY)
		ret = changed_xattr(sctx, result);
	else if (key->type == BTRFS_EXTENT_DATA_KEY)
		ret = changed_extent(sctx, result);

out:
	return ret;
}

static int full_send_tree(struct send_ctx *sctx)
{
	int ret;
	struct btrfs_root *send_root = sctx->send_root;
	struct btrfs_key key;
	struct btrfs_key found_key;
	struct btrfs_path *path;
	struct extent_buffer *eb;
	int slot;

	path = alloc_path_for_send();
	if (!path)
		return -ENOMEM;

	key.objectid = BTRFS_FIRST_FREE_OBJECTID;
	key.type = BTRFS_INODE_ITEM_KEY;
	key.offset = 0;

	ret = btrfs_search_slot_for_read(send_root, &key, path, 1, 0);
	if (ret < 0)
		goto out;
	if (ret)
		goto out_finish;

	while (1) {
		eb = path->nodes[0];
		slot = path->slots[0];
		btrfs_item_key_to_cpu(eb, &found_key, slot);

		ret = changed_cb(send_root, NULL, path, NULL,
				&found_key, BTRFS_COMPARE_TREE_NEW, sctx);
		if (ret < 0)
			goto out;

		key.objectid = found_key.objectid;
		key.type = found_key.type;
		key.offset = found_key.offset + 1;

		ret = btrfs_next_item(send_root, path);
		if (ret < 0)
			goto out;
		if (ret) {
			ret  = 0;
			break;
		}
	}

out_finish:
	ret = finish_inode_if_needed(sctx, 1);

out:
	btrfs_free_path(path);
	return ret;
}

static int send_subvol(struct send_ctx *sctx)
{
	int ret;

6057 6058 6059 6060 6061
	if (!(sctx->flags & BTRFS_SEND_FLAG_OMIT_STREAM_HEADER)) {
		ret = send_header(sctx);
		if (ret < 0)
			goto out;
	}
6062 6063 6064 6065 6066 6067 6068 6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082 6083 6084 6085

	ret = send_subvol_begin(sctx);
	if (ret < 0)
		goto out;

	if (sctx->parent_root) {
		ret = btrfs_compare_trees(sctx->send_root, sctx->parent_root,
				changed_cb, sctx);
		if (ret < 0)
			goto out;
		ret = finish_inode_if_needed(sctx, 1);
		if (ret < 0)
			goto out;
	} else {
		ret = full_send_tree(sctx);
		if (ret < 0)
			goto out;
	}

out:
	free_recorded_refs(sctx);
	return ret;
}

6086 6087 6088 6089 6090 6091 6092 6093 6094 6095 6096 6097 6098 6099 6100 6101 6102 6103 6104 6105 6106 6107 6108 6109 6110 6111 6112 6113 6114 6115 6116 6117 6118 6119 6120 6121 6122 6123 6124 6125 6126 6127 6128 6129 6130
/*
 * If orphan cleanup did remove any orphans from a root, it means the tree
 * was modified and therefore the commit root is not the same as the current
 * root anymore. This is a problem, because send uses the commit root and
 * therefore can see inode items that don't exist in the current root anymore,
 * and for example make calls to btrfs_iget, which will do tree lookups based
 * on the current root and not on the commit root. Those lookups will fail,
 * returning a -ESTALE error, and making send fail with that error. So make
 * sure a send does not see any orphans we have just removed, and that it will
 * see the same inodes regardless of whether a transaction commit happened
 * before it started (meaning that the commit root will be the same as the
 * current root) or not.
 */
static int ensure_commit_roots_uptodate(struct send_ctx *sctx)
{
	int i;
	struct btrfs_trans_handle *trans = NULL;

again:
	if (sctx->parent_root &&
	    sctx->parent_root->node != sctx->parent_root->commit_root)
		goto commit_trans;

	for (i = 0; i < sctx->clone_roots_cnt; i++)
		if (sctx->clone_roots[i].root->node !=
		    sctx->clone_roots[i].root->commit_root)
			goto commit_trans;

	if (trans)
		return btrfs_end_transaction(trans, sctx->send_root);

	return 0;

commit_trans:
	/* Use any root, all fs roots will get their commit roots updated. */
	if (!trans) {
		trans = btrfs_join_transaction(sctx->send_root);
		if (IS_ERR(trans))
			return PTR_ERR(trans);
		goto again;
	}

	return btrfs_commit_transaction(trans, sctx->send_root);
}

6131 6132 6133 6134 6135 6136 6137 6138 6139 6140
static void btrfs_root_dec_send_in_progress(struct btrfs_root* root)
{
	spin_lock(&root->root_item_lock);
	root->send_in_progress--;
	/*
	 * Not much left to do, we don't know why it's unbalanced and
	 * can't blindly reset it to 0.
	 */
	if (root->send_in_progress < 0)
		btrfs_err(root->fs_info,
6141 6142
			  "send_in_progres unbalanced %d root %llu",
			  root->send_in_progress, root->root_key.objectid);
6143 6144 6145
	spin_unlock(&root->root_item_lock);
}

6146 6147 6148
long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_)
{
	int ret = 0;
6149 6150
	struct btrfs_root *send_root = BTRFS_I(file_inode(mnt_file))->root;
	struct btrfs_fs_info *fs_info = send_root->fs_info;
6151 6152 6153 6154 6155 6156
	struct btrfs_root *clone_root;
	struct btrfs_ioctl_send_args *arg = NULL;
	struct btrfs_key key;
	struct send_ctx *sctx = NULL;
	u32 i;
	u64 *clone_sources_tmp = NULL;
6157
	int clone_sources_to_rollback = 0;
6158
	unsigned alloc_size;
6159
	int sort_clone_roots = 0;
6160
	int index;
6161 6162 6163 6164

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

6165 6166
	/*
	 * The subvolume must remain read-only during send, protect against
6167
	 * making it RW. This also protects against deletion.
6168 6169 6170 6171 6172
	 */
	spin_lock(&send_root->root_item_lock);
	send_root->send_in_progress++;
	spin_unlock(&send_root->root_item_lock);

J
Josef Bacik 已提交
6173 6174 6175 6176 6177 6178
	/*
	 * This is done when we lookup the root, it should already be complete
	 * by the time we get here.
	 */
	WARN_ON(send_root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE);

6179 6180 6181 6182 6183 6184 6185 6186 6187
	/*
	 * Userspace tools do the checks and warn the user if it's
	 * not RO.
	 */
	if (!btrfs_root_readonly(send_root)) {
		ret = -EPERM;
		goto out;
	}

6188 6189 6190 6191 6192 6193 6194
	arg = memdup_user(arg_, sizeof(*arg));
	if (IS_ERR(arg)) {
		ret = PTR_ERR(arg);
		arg = NULL;
		goto out;
	}

6195 6196 6197 6198 6199 6200
	if (arg->clone_sources_count >
	    ULLONG_MAX / sizeof(*arg->clone_sources)) {
		ret = -EINVAL;
		goto out;
	}

6201
	if (!access_ok(VERIFY_READ, arg->clone_sources,
6202 6203
			sizeof(*arg->clone_sources) *
			arg->clone_sources_count)) {
6204 6205 6206 6207
		ret = -EFAULT;
		goto out;
	}

6208
	if (arg->flags & ~BTRFS_SEND_FLAG_MASK) {
6209 6210 6211 6212
		ret = -EINVAL;
		goto out;
	}

6213
	sctx = kzalloc(sizeof(struct send_ctx), GFP_KERNEL);
6214 6215 6216 6217 6218 6219 6220
	if (!sctx) {
		ret = -ENOMEM;
		goto out;
	}

	INIT_LIST_HEAD(&sctx->new_refs);
	INIT_LIST_HEAD(&sctx->deleted_refs);
6221
	INIT_RADIX_TREE(&sctx->name_cache, GFP_KERNEL);
6222 6223
	INIT_LIST_HEAD(&sctx->name_cache_list);

6224 6225
	sctx->flags = arg->flags;

6226
	sctx->send_filp = fget(arg->send_fd);
6227 6228
	if (!sctx->send_filp) {
		ret = -EBADF;
6229 6230 6231 6232
		goto out;
	}

	sctx->send_root = send_root;
6233 6234 6235 6236 6237 6238 6239 6240 6241
	/*
	 * Unlikely but possible, if the subvolume is marked for deletion but
	 * is slow to remove the directory entry, send can still be started
	 */
	if (btrfs_root_dead(sctx->send_root)) {
		ret = -EPERM;
		goto out;
	}

6242 6243 6244
	sctx->clone_roots_cnt = arg->clone_sources_count;

	sctx->send_max_size = BTRFS_SEND_BUF_SIZE;
6245
	sctx->send_buf = kmalloc(sctx->send_max_size, GFP_KERNEL | __GFP_NOWARN);
6246
	if (!sctx->send_buf) {
6247 6248 6249 6250 6251
		sctx->send_buf = vmalloc(sctx->send_max_size);
		if (!sctx->send_buf) {
			ret = -ENOMEM;
			goto out;
		}
6252 6253
	}

6254
	sctx->read_buf = kmalloc(BTRFS_SEND_READ_SIZE, GFP_KERNEL | __GFP_NOWARN);
6255
	if (!sctx->read_buf) {
6256 6257 6258 6259 6260
		sctx->read_buf = vmalloc(BTRFS_SEND_READ_SIZE);
		if (!sctx->read_buf) {
			ret = -ENOMEM;
			goto out;
		}
6261 6262
	}

6263 6264
	sctx->pending_dir_moves = RB_ROOT;
	sctx->waiting_dir_moves = RB_ROOT;
6265
	sctx->orphan_dirs = RB_ROOT;
6266

6267 6268
	alloc_size = sizeof(struct clone_root) * (arg->clone_sources_count + 1);

6269
	sctx->clone_roots = kzalloc(alloc_size, GFP_KERNEL | __GFP_NOWARN);
6270
	if (!sctx->clone_roots) {
6271 6272 6273 6274 6275
		sctx->clone_roots = vzalloc(alloc_size);
		if (!sctx->clone_roots) {
			ret = -ENOMEM;
			goto out;
		}
6276 6277
	}

6278 6279
	alloc_size = arg->clone_sources_count * sizeof(*arg->clone_sources);

6280
	if (arg->clone_sources_count) {
6281
		clone_sources_tmp = kmalloc(alloc_size, GFP_KERNEL | __GFP_NOWARN);
6282
		if (!clone_sources_tmp) {
6283 6284 6285 6286 6287
			clone_sources_tmp = vmalloc(alloc_size);
			if (!clone_sources_tmp) {
				ret = -ENOMEM;
				goto out;
			}
6288 6289 6290
		}

		ret = copy_from_user(clone_sources_tmp, arg->clone_sources,
6291
				alloc_size);
6292 6293 6294 6295 6296 6297 6298 6299 6300
		if (ret) {
			ret = -EFAULT;
			goto out;
		}

		for (i = 0; i < arg->clone_sources_count; i++) {
			key.objectid = clone_sources_tmp[i];
			key.type = BTRFS_ROOT_ITEM_KEY;
			key.offset = (u64)-1;
6301 6302 6303

			index = srcu_read_lock(&fs_info->subvol_srcu);

6304 6305
			clone_root = btrfs_read_fs_root_no_name(fs_info, &key);
			if (IS_ERR(clone_root)) {
6306
				srcu_read_unlock(&fs_info->subvol_srcu, index);
6307 6308 6309
				ret = PTR_ERR(clone_root);
				goto out;
			}
6310
			spin_lock(&clone_root->root_item_lock);
6311 6312
			if (!btrfs_root_readonly(clone_root) ||
			    btrfs_root_dead(clone_root)) {
6313
				spin_unlock(&clone_root->root_item_lock);
6314
				srcu_read_unlock(&fs_info->subvol_srcu, index);
6315 6316 6317
				ret = -EPERM;
				goto out;
			}
6318
			clone_root->send_in_progress++;
6319
			spin_unlock(&clone_root->root_item_lock);
6320 6321
			srcu_read_unlock(&fs_info->subvol_srcu, index);

6322
			sctx->clone_roots[i].root = clone_root;
6323
			clone_sources_to_rollback = i + 1;
6324
		}
6325
		kvfree(clone_sources_tmp);
6326 6327 6328 6329 6330 6331 6332
		clone_sources_tmp = NULL;
	}

	if (arg->parent_root) {
		key.objectid = arg->parent_root;
		key.type = BTRFS_ROOT_ITEM_KEY;
		key.offset = (u64)-1;
6333 6334 6335

		index = srcu_read_lock(&fs_info->subvol_srcu);

6336
		sctx->parent_root = btrfs_read_fs_root_no_name(fs_info, &key);
6337
		if (IS_ERR(sctx->parent_root)) {
6338
			srcu_read_unlock(&fs_info->subvol_srcu, index);
6339
			ret = PTR_ERR(sctx->parent_root);
6340 6341
			goto out;
		}
6342

6343 6344
		spin_lock(&sctx->parent_root->root_item_lock);
		sctx->parent_root->send_in_progress++;
6345 6346
		if (!btrfs_root_readonly(sctx->parent_root) ||
				btrfs_root_dead(sctx->parent_root)) {
6347
			spin_unlock(&sctx->parent_root->root_item_lock);
6348
			srcu_read_unlock(&fs_info->subvol_srcu, index);
6349 6350 6351 6352
			ret = -EPERM;
			goto out;
		}
		spin_unlock(&sctx->parent_root->root_item_lock);
6353 6354

		srcu_read_unlock(&fs_info->subvol_srcu, index);
6355 6356 6357 6358 6359 6360 6361 6362 6363 6364 6365 6366 6367
	}

	/*
	 * Clones from send_root are allowed, but only if the clone source
	 * is behind the current send position. This is checked while searching
	 * for possible clone sources.
	 */
	sctx->clone_roots[sctx->clone_roots_cnt++].root = sctx->send_root;

	/* We do a bsearch later */
	sort(sctx->clone_roots, sctx->clone_roots_cnt,
			sizeof(*sctx->clone_roots), __clone_root_cmp_sort,
			NULL);
6368
	sort_clone_roots = 1;
6369

6370 6371 6372 6373
	ret = ensure_commit_roots_uptodate(sctx);
	if (ret)
		goto out;

6374
	current->journal_info = BTRFS_SEND_TRANS_STUB;
6375
	ret = send_subvol(sctx);
6376
	current->journal_info = NULL;
6377 6378 6379
	if (ret < 0)
		goto out;

6380 6381 6382 6383 6384 6385 6386 6387
	if (!(sctx->flags & BTRFS_SEND_FLAG_OMIT_END_CMD)) {
		ret = begin_cmd(sctx, BTRFS_SEND_C_END);
		if (ret < 0)
			goto out;
		ret = send_cmd(sctx);
		if (ret < 0)
			goto out;
	}
6388 6389

out:
6390 6391 6392 6393 6394 6395 6396 6397 6398 6399 6400 6401 6402 6403 6404 6405 6406 6407 6408 6409 6410 6411 6412 6413 6414 6415 6416 6417
	WARN_ON(sctx && !ret && !RB_EMPTY_ROOT(&sctx->pending_dir_moves));
	while (sctx && !RB_EMPTY_ROOT(&sctx->pending_dir_moves)) {
		struct rb_node *n;
		struct pending_dir_move *pm;

		n = rb_first(&sctx->pending_dir_moves);
		pm = rb_entry(n, struct pending_dir_move, node);
		while (!list_empty(&pm->list)) {
			struct pending_dir_move *pm2;

			pm2 = list_first_entry(&pm->list,
					       struct pending_dir_move, list);
			free_pending_move(sctx, pm2);
		}
		free_pending_move(sctx, pm);
	}

	WARN_ON(sctx && !ret && !RB_EMPTY_ROOT(&sctx->waiting_dir_moves));
	while (sctx && !RB_EMPTY_ROOT(&sctx->waiting_dir_moves)) {
		struct rb_node *n;
		struct waiting_dir_move *dm;

		n = rb_first(&sctx->waiting_dir_moves);
		dm = rb_entry(n, struct waiting_dir_move, node);
		rb_erase(&dm->node, &sctx->waiting_dir_moves);
		kfree(dm);
	}

6418 6419 6420 6421 6422 6423 6424 6425 6426 6427
	WARN_ON(sctx && !ret && !RB_EMPTY_ROOT(&sctx->orphan_dirs));
	while (sctx && !RB_EMPTY_ROOT(&sctx->orphan_dirs)) {
		struct rb_node *n;
		struct orphan_dir_info *odi;

		n = rb_first(&sctx->orphan_dirs);
		odi = rb_entry(n, struct orphan_dir_info, node);
		free_orphan_dir_info(sctx, odi);
	}

6428 6429 6430 6431 6432 6433 6434 6435 6436 6437 6438
	if (sort_clone_roots) {
		for (i = 0; i < sctx->clone_roots_cnt; i++)
			btrfs_root_dec_send_in_progress(
					sctx->clone_roots[i].root);
	} else {
		for (i = 0; sctx && i < clone_sources_to_rollback; i++)
			btrfs_root_dec_send_in_progress(
					sctx->clone_roots[i].root);

		btrfs_root_dec_send_in_progress(send_root);
	}
6439 6440
	if (sctx && !IS_ERR_OR_NULL(sctx->parent_root))
		btrfs_root_dec_send_in_progress(sctx->parent_root);
6441

6442
	kfree(arg);
6443
	kvfree(clone_sources_tmp);
6444 6445 6446 6447 6448

	if (sctx) {
		if (sctx->send_filp)
			fput(sctx->send_filp);

6449
		kvfree(sctx->clone_roots);
6450
		kvfree(sctx->send_buf);
6451
		kvfree(sctx->read_buf);
6452 6453 6454 6455 6456 6457 6458 6459

		name_cache_free(sctx);

		kfree(sctx);
	}

	return ret;
}