1 // SPDX-License-Identifier: GPL-2.0
3 * Copyright (C) 2007 Oracle. All rights reserved.
7 #include <linux/slab.h>
8 #include <linux/pagemap.h>
9 #include <linux/highmem.h>
10 #include <linux/sched/mm.h>
11 #include <crypto/hash.h>
15 #include "transaction.h"
17 #include "print-tree.h"
18 #include "compression.h"
20 #define __MAX_CSUM_ITEMS(r, size) ((unsigned long)(((BTRFS_LEAF_DATA_SIZE(r) - \
21 sizeof(struct btrfs_item) * 2) / \
24 #define MAX_CSUM_ITEMS(r, size) (min_t(u32, __MAX_CSUM_ITEMS(r, size), \
28 * Set inode's size according to filesystem options
30 * @inode: inode we want to update the disk_i_size for
31 * @new_i_size: i_size we want to set to, 0 if we use i_size
33 * With NO_HOLES set this simply sets the disk_is_size to whatever i_size_read()
34 * returns as it is perfectly fine with a file that has holes without hole file
37 * However without NO_HOLES we need to only return the area that is contiguous
38 * from the 0 offset of the file. Otherwise we could end up adjust i_size up
39 * to an extent that has a gap in between.
41 * Finally new_i_size should only be set in the case of truncate where we're not
42 * ready to use i_size_read() as the limiter yet.
44 void btrfs_inode_safe_disk_i_size_write(struct btrfs_inode *inode, u64 new_i_size)
46 struct btrfs_fs_info *fs_info = inode->root->fs_info;
47 u64 start, end, i_size;
50 i_size = new_i_size ?: i_size_read(&inode->vfs_inode);
51 if (btrfs_fs_incompat(fs_info, NO_HOLES)) {
52 inode->disk_i_size = i_size;
56 spin_lock(&inode->lock);
57 ret = find_contiguous_extent_bit(&inode->file_extent_tree, 0, &start,
59 if (!ret && start == 0)
60 i_size = min(i_size, end + 1);
63 inode->disk_i_size = i_size;
64 spin_unlock(&inode->lock);
68 * Mark range within a file as having a new extent inserted
70 * @inode: inode being modified
71 * @start: start file offset of the file extent we've inserted
72 * @len: logical length of the file extent item
74 * Call when we are inserting a new file extent where there was none before.
75 * Does not need to call this in the case where we're replacing an existing file
76 * extent, however if not sure it's fine to call this multiple times.
78 * The start and len must match the file extent item, so thus must be sectorsize
81 int btrfs_inode_set_file_extent_range(struct btrfs_inode *inode, u64 start,
87 ASSERT(IS_ALIGNED(start + len, inode->root->fs_info->sectorsize));
89 if (btrfs_fs_incompat(inode->root->fs_info, NO_HOLES))
91 return set_extent_bits(&inode->file_extent_tree, start, start + len - 1,
96 * Marks an inode range as not having a backing extent
98 * @inode: inode being modified
99 * @start: start file offset of the file extent we've inserted
100 * @len: logical length of the file extent item
102 * Called when we drop a file extent, for example when we truncate. Doesn't
103 * need to be called for cases where we're replacing a file extent, like when
104 * we've COWed a file extent.
106 * The start and len must match the file extent item, so thus must be sectorsize
109 int btrfs_inode_clear_file_extent_range(struct btrfs_inode *inode, u64 start,
115 ASSERT(IS_ALIGNED(start + len, inode->root->fs_info->sectorsize) ||
118 if (btrfs_fs_incompat(inode->root->fs_info, NO_HOLES))
120 return clear_extent_bit(&inode->file_extent_tree, start,
121 start + len - 1, EXTENT_DIRTY, 0, 0, NULL);
124 static inline u32 max_ordered_sum_bytes(struct btrfs_fs_info *fs_info,
127 u32 ncsums = (PAGE_SIZE - sizeof(struct btrfs_ordered_sum)) / csum_size;
129 return ncsums * fs_info->sectorsize;
132 int btrfs_insert_file_extent(struct btrfs_trans_handle *trans,
133 struct btrfs_root *root,
134 u64 objectid, u64 pos,
135 u64 disk_offset, u64 disk_num_bytes,
136 u64 num_bytes, u64 offset, u64 ram_bytes,
137 u8 compression, u8 encryption, u16 other_encoding)
140 struct btrfs_file_extent_item *item;
141 struct btrfs_key file_key;
142 struct btrfs_path *path;
143 struct extent_buffer *leaf;
145 path = btrfs_alloc_path();
148 file_key.objectid = objectid;
149 file_key.offset = pos;
150 file_key.type = BTRFS_EXTENT_DATA_KEY;
152 ret = btrfs_insert_empty_item(trans, root, path, &file_key,
156 BUG_ON(ret); /* Can't happen */
157 leaf = path->nodes[0];
158 item = btrfs_item_ptr(leaf, path->slots[0],
159 struct btrfs_file_extent_item);
160 btrfs_set_file_extent_disk_bytenr(leaf, item, disk_offset);
161 btrfs_set_file_extent_disk_num_bytes(leaf, item, disk_num_bytes);
162 btrfs_set_file_extent_offset(leaf, item, offset);
163 btrfs_set_file_extent_num_bytes(leaf, item, num_bytes);
164 btrfs_set_file_extent_ram_bytes(leaf, item, ram_bytes);
165 btrfs_set_file_extent_generation(leaf, item, trans->transid);
166 btrfs_set_file_extent_type(leaf, item, BTRFS_FILE_EXTENT_REG);
167 btrfs_set_file_extent_compression(leaf, item, compression);
168 btrfs_set_file_extent_encryption(leaf, item, encryption);
169 btrfs_set_file_extent_other_encoding(leaf, item, other_encoding);
171 btrfs_mark_buffer_dirty(leaf);
173 btrfs_free_path(path);
177 static struct btrfs_csum_item *
178 btrfs_lookup_csum(struct btrfs_trans_handle *trans,
179 struct btrfs_root *root,
180 struct btrfs_path *path,
183 struct btrfs_fs_info *fs_info = root->fs_info;
185 struct btrfs_key file_key;
186 struct btrfs_key found_key;
187 struct btrfs_csum_item *item;
188 struct extent_buffer *leaf;
190 const u32 csum_size = fs_info->csum_size;
193 file_key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
194 file_key.offset = bytenr;
195 file_key.type = BTRFS_EXTENT_CSUM_KEY;
196 ret = btrfs_search_slot(trans, root, &file_key, path, 0, cow);
199 leaf = path->nodes[0];
202 if (path->slots[0] == 0)
205 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
206 if (found_key.type != BTRFS_EXTENT_CSUM_KEY)
209 csum_offset = (bytenr - found_key.offset) >>
210 fs_info->sectorsize_bits;
211 csums_in_item = btrfs_item_size(leaf, path->slots[0]);
212 csums_in_item /= csum_size;
214 if (csum_offset == csums_in_item) {
217 } else if (csum_offset > csums_in_item) {
221 item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_csum_item);
222 item = (struct btrfs_csum_item *)((unsigned char *)item +
223 csum_offset * csum_size);
231 int btrfs_lookup_file_extent(struct btrfs_trans_handle *trans,
232 struct btrfs_root *root,
233 struct btrfs_path *path, u64 objectid,
236 struct btrfs_key file_key;
237 int ins_len = mod < 0 ? -1 : 0;
240 file_key.objectid = objectid;
241 file_key.offset = offset;
242 file_key.type = BTRFS_EXTENT_DATA_KEY;
244 return btrfs_search_slot(trans, root, &file_key, path, ins_len, cow);
248 * Find checksums for logical bytenr range [disk_bytenr, disk_bytenr + len) and
249 * estore the result to @dst.
251 * Return >0 for the number of sectors we found.
252 * Return 0 for the range [disk_bytenr, disk_bytenr + sectorsize) has no csum
253 * for it. Caller may want to try next sector until one range is hit.
254 * Return <0 for fatal error.
256 static int search_csum_tree(struct btrfs_fs_info *fs_info,
257 struct btrfs_path *path, u64 disk_bytenr,
260 struct btrfs_root *csum_root;
261 struct btrfs_csum_item *item = NULL;
262 struct btrfs_key key;
263 const u32 sectorsize = fs_info->sectorsize;
264 const u32 csum_size = fs_info->csum_size;
270 ASSERT(IS_ALIGNED(disk_bytenr, sectorsize) &&
271 IS_ALIGNED(len, sectorsize));
273 /* Check if the current csum item covers disk_bytenr */
274 if (path->nodes[0]) {
275 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
276 struct btrfs_csum_item);
277 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
278 itemsize = btrfs_item_size(path->nodes[0], path->slots[0]);
280 csum_start = key.offset;
281 csum_len = (itemsize / csum_size) * sectorsize;
283 if (in_range(disk_bytenr, csum_start, csum_len))
287 /* Current item doesn't contain the desired range, search again */
288 btrfs_release_path(path);
289 csum_root = btrfs_csum_root(fs_info, disk_bytenr);
290 item = btrfs_lookup_csum(NULL, csum_root, path, disk_bytenr, 0);
295 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
296 itemsize = btrfs_item_size(path->nodes[0], path->slots[0]);
298 csum_start = key.offset;
299 csum_len = (itemsize / csum_size) * sectorsize;
300 ASSERT(in_range(disk_bytenr, csum_start, csum_len));
303 ret = (min(csum_start + csum_len, disk_bytenr + len) -
304 disk_bytenr) >> fs_info->sectorsize_bits;
305 read_extent_buffer(path->nodes[0], dst, (unsigned long)item,
308 if (ret == -ENOENT || ret == -EFBIG)
314 * Locate the file_offset of @cur_disk_bytenr of a @bio.
316 * Bio of btrfs represents read range of
317 * [bi_sector << 9, bi_sector << 9 + bi_size).
318 * Knowing this, we can iterate through each bvec to locate the page belong to
319 * @cur_disk_bytenr and get the file offset.
321 * @inode is used to determine if the bvec page really belongs to @inode.
323 * Return 0 if we can't find the file offset
324 * Return >0 if we find the file offset and restore it to @file_offset_ret
326 static int search_file_offset_in_bio(struct bio *bio, struct inode *inode,
327 u64 disk_bytenr, u64 *file_offset_ret)
329 struct bvec_iter iter;
331 u64 cur = bio->bi_iter.bi_sector << SECTOR_SHIFT;
334 bio_for_each_segment(bvec, bio, iter) {
335 struct page *page = bvec.bv_page;
337 if (cur > disk_bytenr)
339 if (cur + bvec.bv_len <= disk_bytenr) {
343 ASSERT(in_range(disk_bytenr, cur, bvec.bv_len));
344 if (page->mapping && page->mapping->host &&
345 page->mapping->host == inode) {
347 *file_offset_ret = page_offset(page) + bvec.bv_offset +
356 * Lookup the checksum for the read bio in csum tree.
358 * @inode: inode that the bio is for.
359 * @bio: bio to look up.
360 * @dst: Buffer of size nblocks * btrfs_super_csum_size() used to return
361 * checksum (nblocks = bio->bi_iter.bi_size / fs_info->sectorsize). If
362 * NULL, the checksum buffer is allocated and returned in
363 * btrfs_bio(bio)->csum instead.
365 * Return: BLK_STS_RESOURCE if allocating memory fails, BLK_STS_OK otherwise.
367 blk_status_t btrfs_lookup_bio_sums(struct inode *inode, struct bio *bio, u8 *dst)
369 struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
370 struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
371 struct btrfs_bio *bbio = NULL;
372 struct btrfs_path *path;
373 const u32 sectorsize = fs_info->sectorsize;
374 const u32 csum_size = fs_info->csum_size;
375 u32 orig_len = bio->bi_iter.bi_size;
376 u64 orig_disk_bytenr = bio->bi_iter.bi_sector << SECTOR_SHIFT;
379 const unsigned int nblocks = orig_len >> fs_info->sectorsize_bits;
381 blk_status_t ret = BLK_STS_OK;
383 if ((BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM) ||
384 test_bit(BTRFS_FS_STATE_NO_CSUMS, &fs_info->fs_state))
388 * This function is only called for read bio.
390 * This means two things:
391 * - All our csums should only be in csum tree
392 * No ordered extents csums, as ordered extents are only for write
394 * - No need to bother any other info from bvec
395 * Since we're looking up csums, the only important info is the
396 * disk_bytenr and the length, which can be extracted from bi_iter
399 ASSERT(bio_op(bio) == REQ_OP_READ);
400 path = btrfs_alloc_path();
402 return BLK_STS_RESOURCE;
405 bbio = btrfs_bio(bio);
407 if (nblocks * csum_size > BTRFS_BIO_INLINE_CSUM_SIZE) {
408 bbio->csum = kmalloc_array(nblocks, csum_size, GFP_NOFS);
410 btrfs_free_path(path);
411 return BLK_STS_RESOURCE;
414 bbio->csum = bbio->csum_inline;
422 * If requested number of sectors is larger than one leaf can contain,
423 * kick the readahead for csum tree.
425 if (nblocks > fs_info->csums_per_leaf)
426 path->reada = READA_FORWARD;
429 * the free space stuff is only read when it hasn't been
430 * updated in the current transaction. So, we can safely
431 * read from the commit root and sidestep a nasty deadlock
432 * between reading the free space cache and updating the csum tree.
434 if (btrfs_is_free_space_inode(BTRFS_I(inode))) {
435 path->search_commit_root = 1;
436 path->skip_locking = 1;
439 for (cur_disk_bytenr = orig_disk_bytenr;
440 cur_disk_bytenr < orig_disk_bytenr + orig_len;
441 cur_disk_bytenr += (count * sectorsize)) {
442 u64 search_len = orig_disk_bytenr + orig_len - cur_disk_bytenr;
443 unsigned int sector_offset;
447 * Although both cur_disk_bytenr and orig_disk_bytenr is u64,
448 * we're calculating the offset to the bio start.
450 * Bio size is limited to UINT_MAX, thus unsigned int is large
451 * enough to contain the raw result, not to mention the right
454 ASSERT(cur_disk_bytenr - orig_disk_bytenr < UINT_MAX);
455 sector_offset = (cur_disk_bytenr - orig_disk_bytenr) >>
456 fs_info->sectorsize_bits;
457 csum_dst = csum + sector_offset * csum_size;
459 count = search_csum_tree(fs_info, path, cur_disk_bytenr,
460 search_len, csum_dst);
462 ret = errno_to_blk_status(count);
464 btrfs_bio_free_csum(bbio);
469 * We didn't find a csum for this range. We need to make sure
470 * we complain loudly about this, because we are not NODATASUM.
472 * However for the DATA_RELOC inode we could potentially be
473 * relocating data extents for a NODATASUM inode, so the inode
474 * itself won't be marked with NODATASUM, but the extent we're
475 * copying is in fact NODATASUM. If we don't find a csum we
476 * assume this is the case.
479 memset(csum_dst, 0, csum_size);
482 if (BTRFS_I(inode)->root->root_key.objectid ==
483 BTRFS_DATA_RELOC_TREE_OBJECTID) {
487 ret = search_file_offset_in_bio(bio, inode,
488 cur_disk_bytenr, &file_offset);
490 set_extent_bits(io_tree, file_offset,
491 file_offset + sectorsize - 1,
494 btrfs_warn_rl(fs_info,
495 "csum hole found for disk bytenr range [%llu, %llu)",
496 cur_disk_bytenr, cur_disk_bytenr + sectorsize);
501 btrfs_free_path(path);
505 int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
506 struct list_head *list, int search_commit)
508 struct btrfs_fs_info *fs_info = root->fs_info;
509 struct btrfs_key key;
510 struct btrfs_path *path;
511 struct extent_buffer *leaf;
512 struct btrfs_ordered_sum *sums;
513 struct btrfs_csum_item *item;
515 unsigned long offset;
519 const u32 csum_size = fs_info->csum_size;
521 ASSERT(IS_ALIGNED(start, fs_info->sectorsize) &&
522 IS_ALIGNED(end + 1, fs_info->sectorsize));
524 path = btrfs_alloc_path();
529 path->skip_locking = 1;
530 path->reada = READA_FORWARD;
531 path->search_commit_root = 1;
534 key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
536 key.type = BTRFS_EXTENT_CSUM_KEY;
538 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
541 if (ret > 0 && path->slots[0] > 0) {
542 leaf = path->nodes[0];
543 btrfs_item_key_to_cpu(leaf, &key, path->slots[0] - 1);
544 if (key.objectid == BTRFS_EXTENT_CSUM_OBJECTID &&
545 key.type == BTRFS_EXTENT_CSUM_KEY) {
546 offset = (start - key.offset) >> fs_info->sectorsize_bits;
547 if (offset * csum_size <
548 btrfs_item_size(leaf, path->slots[0] - 1))
553 while (start <= end) {
554 leaf = path->nodes[0];
555 if (path->slots[0] >= btrfs_header_nritems(leaf)) {
556 ret = btrfs_next_leaf(root, path);
561 leaf = path->nodes[0];
564 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
565 if (key.objectid != BTRFS_EXTENT_CSUM_OBJECTID ||
566 key.type != BTRFS_EXTENT_CSUM_KEY ||
570 if (key.offset > start)
573 size = btrfs_item_size(leaf, path->slots[0]);
574 csum_end = key.offset + (size / csum_size) * fs_info->sectorsize;
575 if (csum_end <= start) {
580 csum_end = min(csum_end, end + 1);
581 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
582 struct btrfs_csum_item);
583 while (start < csum_end) {
584 size = min_t(size_t, csum_end - start,
585 max_ordered_sum_bytes(fs_info, csum_size));
586 sums = kzalloc(btrfs_ordered_sum_size(fs_info, size),
593 sums->bytenr = start;
594 sums->len = (int)size;
596 offset = (start - key.offset) >> fs_info->sectorsize_bits;
598 size >>= fs_info->sectorsize_bits;
600 read_extent_buffer(path->nodes[0],
602 ((unsigned long)item) + offset,
605 start += fs_info->sectorsize * size;
606 list_add_tail(&sums->list, &tmplist);
612 while (ret < 0 && !list_empty(&tmplist)) {
613 sums = list_entry(tmplist.next, struct btrfs_ordered_sum, list);
614 list_del(&sums->list);
617 list_splice_tail(&tmplist, list);
619 btrfs_free_path(path);
624 * Calculate checksums of the data contained inside a bio
626 * @inode: Owner of the data inside the bio
627 * @bio: Contains the data to be checksummed
628 * @offset: If (u64)-1, @bio may contain discontiguous bio vecs, so the
629 * file offsets are determined from the page offsets in the bio.
630 * Otherwise, this is the starting file offset of the bio vecs in
631 * @bio, which must be contiguous.
632 * @one_ordered: If true, @bio only refers to one ordered extent.
634 blk_status_t btrfs_csum_one_bio(struct btrfs_inode *inode, struct bio *bio,
635 u64 offset, bool one_ordered)
637 struct btrfs_fs_info *fs_info = inode->root->fs_info;
638 SHASH_DESC_ON_STACK(shash, fs_info->csum_shash);
639 struct btrfs_ordered_sum *sums;
640 struct btrfs_ordered_extent *ordered = NULL;
641 const bool use_page_offsets = (offset == (u64)-1);
643 struct bvec_iter iter;
646 unsigned int blockcount;
647 unsigned long total_bytes = 0;
648 unsigned long this_sum_bytes = 0;
652 nofs_flag = memalloc_nofs_save();
653 sums = kvzalloc(btrfs_ordered_sum_size(fs_info, bio->bi_iter.bi_size),
655 memalloc_nofs_restore(nofs_flag);
658 return BLK_STS_RESOURCE;
660 sums->len = bio->bi_iter.bi_size;
661 INIT_LIST_HEAD(&sums->list);
663 sums->bytenr = bio->bi_iter.bi_sector << 9;
666 shash->tfm = fs_info->csum_shash;
668 bio_for_each_segment(bvec, bio, iter) {
669 if (use_page_offsets)
670 offset = page_offset(bvec.bv_page) + bvec.bv_offset;
673 ordered = btrfs_lookup_ordered_extent(inode, offset);
675 * The bio range is not covered by any ordered extent,
676 * must be a code logic error.
678 if (unlikely(!ordered)) {
680 "no ordered extent for root %llu ino %llu offset %llu\n",
681 inode->root->root_key.objectid,
682 btrfs_ino(inode), offset);
684 return BLK_STS_IOERR;
688 blockcount = BTRFS_BYTES_TO_BLKS(fs_info,
689 bvec.bv_len + fs_info->sectorsize
692 for (i = 0; i < blockcount; i++) {
694 !in_range(offset, ordered->file_offset,
695 ordered->num_bytes)) {
696 unsigned long bytes_left;
698 sums->len = this_sum_bytes;
700 btrfs_add_ordered_sum(ordered, sums);
701 btrfs_put_ordered_extent(ordered);
703 bytes_left = bio->bi_iter.bi_size - total_bytes;
705 nofs_flag = memalloc_nofs_save();
706 sums = kvzalloc(btrfs_ordered_sum_size(fs_info,
707 bytes_left), GFP_KERNEL);
708 memalloc_nofs_restore(nofs_flag);
709 BUG_ON(!sums); /* -ENOMEM */
710 sums->len = bytes_left;
711 ordered = btrfs_lookup_ordered_extent(inode,
713 ASSERT(ordered); /* Logic error */
714 sums->bytenr = (bio->bi_iter.bi_sector << 9)
719 data = bvec_kmap_local(&bvec);
720 crypto_shash_digest(shash,
721 data + (i * fs_info->sectorsize),
725 index += fs_info->csum_size;
726 offset += fs_info->sectorsize;
727 this_sum_bytes += fs_info->sectorsize;
728 total_bytes += fs_info->sectorsize;
733 btrfs_add_ordered_sum(ordered, sums);
734 btrfs_put_ordered_extent(ordered);
739 * helper function for csum removal, this expects the
740 * key to describe the csum pointed to by the path, and it expects
741 * the csum to overlap the range [bytenr, len]
743 * The csum should not be entirely contained in the range and the
744 * range should not be entirely contained in the csum.
746 * This calls btrfs_truncate_item with the correct args based on the
747 * overlap, and fixes up the key as required.
749 static noinline void truncate_one_csum(struct btrfs_fs_info *fs_info,
750 struct btrfs_path *path,
751 struct btrfs_key *key,
754 struct extent_buffer *leaf;
755 const u32 csum_size = fs_info->csum_size;
757 u64 end_byte = bytenr + len;
758 u32 blocksize_bits = fs_info->sectorsize_bits;
760 leaf = path->nodes[0];
761 csum_end = btrfs_item_size(leaf, path->slots[0]) / csum_size;
762 csum_end <<= blocksize_bits;
763 csum_end += key->offset;
765 if (key->offset < bytenr && csum_end <= end_byte) {
770 * A simple truncate off the end of the item
772 u32 new_size = (bytenr - key->offset) >> blocksize_bits;
773 new_size *= csum_size;
774 btrfs_truncate_item(path, new_size, 1);
775 } else if (key->offset >= bytenr && csum_end > end_byte &&
776 end_byte > key->offset) {
781 * we need to truncate from the beginning of the csum
783 u32 new_size = (csum_end - end_byte) >> blocksize_bits;
784 new_size *= csum_size;
786 btrfs_truncate_item(path, new_size, 0);
788 key->offset = end_byte;
789 btrfs_set_item_key_safe(fs_info, path, key);
796 * deletes the csum items from the csum tree for a given
799 int btrfs_del_csums(struct btrfs_trans_handle *trans,
800 struct btrfs_root *root, u64 bytenr, u64 len)
802 struct btrfs_fs_info *fs_info = trans->fs_info;
803 struct btrfs_path *path;
804 struct btrfs_key key;
805 u64 end_byte = bytenr + len;
807 struct extent_buffer *leaf;
809 const u32 csum_size = fs_info->csum_size;
810 u32 blocksize_bits = fs_info->sectorsize_bits;
812 ASSERT(root->root_key.objectid == BTRFS_CSUM_TREE_OBJECTID ||
813 root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID);
815 path = btrfs_alloc_path();
820 key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
821 key.offset = end_byte - 1;
822 key.type = BTRFS_EXTENT_CSUM_KEY;
824 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
827 if (path->slots[0] == 0)
830 } else if (ret < 0) {
834 leaf = path->nodes[0];
835 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
837 if (key.objectid != BTRFS_EXTENT_CSUM_OBJECTID ||
838 key.type != BTRFS_EXTENT_CSUM_KEY) {
842 if (key.offset >= end_byte)
845 csum_end = btrfs_item_size(leaf, path->slots[0]) / csum_size;
846 csum_end <<= blocksize_bits;
847 csum_end += key.offset;
849 /* this csum ends before we start, we're done */
850 if (csum_end <= bytenr)
853 /* delete the entire item, it is inside our range */
854 if (key.offset >= bytenr && csum_end <= end_byte) {
858 * Check how many csum items preceding this one in this
859 * leaf correspond to our range and then delete them all
862 if (key.offset > bytenr && path->slots[0] > 0) {
863 int slot = path->slots[0] - 1;
868 btrfs_item_key_to_cpu(leaf, &pk, slot);
869 if (pk.offset < bytenr ||
870 pk.type != BTRFS_EXTENT_CSUM_KEY ||
872 BTRFS_EXTENT_CSUM_OBJECTID)
874 path->slots[0] = slot;
876 key.offset = pk.offset;
880 ret = btrfs_del_items(trans, root, path,
881 path->slots[0], del_nr);
884 if (key.offset == bytenr)
886 } else if (key.offset < bytenr && csum_end > end_byte) {
887 unsigned long offset;
888 unsigned long shift_len;
889 unsigned long item_offset;
894 * Our bytes are in the middle of the csum,
895 * we need to split this item and insert a new one.
897 * But we can't drop the path because the
898 * csum could change, get removed, extended etc.
900 * The trick here is the max size of a csum item leaves
901 * enough room in the tree block for a single
902 * item header. So, we split the item in place,
903 * adding a new header pointing to the existing
904 * bytes. Then we loop around again and we have
905 * a nicely formed csum item that we can neatly
908 offset = (bytenr - key.offset) >> blocksize_bits;
911 shift_len = (len >> blocksize_bits) * csum_size;
913 item_offset = btrfs_item_ptr_offset(leaf,
916 memzero_extent_buffer(leaf, item_offset + offset,
921 * btrfs_split_item returns -EAGAIN when the
922 * item changed size or key
924 ret = btrfs_split_item(trans, root, path, &key, offset);
925 if (ret && ret != -EAGAIN) {
926 btrfs_abort_transaction(trans, ret);
931 key.offset = end_byte - 1;
933 truncate_one_csum(fs_info, path, &key, bytenr, len);
934 if (key.offset < bytenr)
937 btrfs_release_path(path);
939 btrfs_free_path(path);
943 static int find_next_csum_offset(struct btrfs_root *root,
944 struct btrfs_path *path,
947 const u32 nritems = btrfs_header_nritems(path->nodes[0]);
948 struct btrfs_key found_key;
949 int slot = path->slots[0] + 1;
952 if (nritems == 0 || slot >= nritems) {
953 ret = btrfs_next_leaf(root, path);
956 } else if (ret > 0) {
957 *next_offset = (u64)-1;
960 slot = path->slots[0];
963 btrfs_item_key_to_cpu(path->nodes[0], &found_key, slot);
965 if (found_key.objectid != BTRFS_EXTENT_CSUM_OBJECTID ||
966 found_key.type != BTRFS_EXTENT_CSUM_KEY)
967 *next_offset = (u64)-1;
969 *next_offset = found_key.offset;
974 int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans,
975 struct btrfs_root *root,
976 struct btrfs_ordered_sum *sums)
978 struct btrfs_fs_info *fs_info = root->fs_info;
979 struct btrfs_key file_key;
980 struct btrfs_key found_key;
981 struct btrfs_path *path;
982 struct btrfs_csum_item *item;
983 struct btrfs_csum_item *item_end;
984 struct extent_buffer *leaf = NULL;
993 const u32 csum_size = fs_info->csum_size;
995 path = btrfs_alloc_path();
999 next_offset = (u64)-1;
1001 bytenr = sums->bytenr + total_bytes;
1002 file_key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
1003 file_key.offset = bytenr;
1004 file_key.type = BTRFS_EXTENT_CSUM_KEY;
1006 item = btrfs_lookup_csum(trans, root, path, bytenr, 1);
1007 if (!IS_ERR(item)) {
1009 leaf = path->nodes[0];
1010 item_end = btrfs_item_ptr(leaf, path->slots[0],
1011 struct btrfs_csum_item);
1012 item_end = (struct btrfs_csum_item *)((char *)item_end +
1013 btrfs_item_size(leaf, path->slots[0]));
1016 ret = PTR_ERR(item);
1017 if (ret != -EFBIG && ret != -ENOENT)
1020 if (ret == -EFBIG) {
1022 /* we found one, but it isn't big enough yet */
1023 leaf = path->nodes[0];
1024 item_size = btrfs_item_size(leaf, path->slots[0]);
1025 if ((item_size / csum_size) >=
1026 MAX_CSUM_ITEMS(fs_info, csum_size)) {
1027 /* already at max size, make a new one */
1031 /* We didn't find a csum item, insert one. */
1032 ret = find_next_csum_offset(root, path, &next_offset);
1040 * At this point, we know the tree has a checksum item that ends at an
1041 * offset matching the start of the checksum range we want to insert.
1042 * We try to extend that item as much as possible and then add as many
1043 * checksums to it as they fit.
1045 * First check if the leaf has enough free space for at least one
1046 * checksum. If it has go directly to the item extension code, otherwise
1047 * release the path and do a search for insertion before the extension.
1049 if (btrfs_leaf_free_space(leaf) >= csum_size) {
1050 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
1051 csum_offset = (bytenr - found_key.offset) >>
1052 fs_info->sectorsize_bits;
1056 btrfs_release_path(path);
1057 path->search_for_extension = 1;
1058 ret = btrfs_search_slot(trans, root, &file_key, path,
1060 path->search_for_extension = 0;
1065 if (path->slots[0] == 0)
1070 leaf = path->nodes[0];
1071 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
1072 csum_offset = (bytenr - found_key.offset) >> fs_info->sectorsize_bits;
1074 if (found_key.type != BTRFS_EXTENT_CSUM_KEY ||
1075 found_key.objectid != BTRFS_EXTENT_CSUM_OBJECTID ||
1076 csum_offset >= MAX_CSUM_ITEMS(fs_info, csum_size)) {
1081 if (csum_offset == btrfs_item_size(leaf, path->slots[0]) /
1087 tmp = sums->len - total_bytes;
1088 tmp >>= fs_info->sectorsize_bits;
1090 extend_nr = max_t(int, 1, tmp);
1093 * A log tree can already have checksum items with a subset of
1094 * the checksums we are trying to log. This can happen after
1095 * doing a sequence of partial writes into prealloc extents and
1096 * fsyncs in between, with a full fsync logging a larger subrange
1097 * of an extent for which a previous fast fsync logged a smaller
1098 * subrange. And this happens in particular due to merging file
1099 * extent items when we complete an ordered extent for a range
1100 * covered by a prealloc extent - this is done at
1101 * btrfs_mark_extent_written().
1103 * So if we try to extend the previous checksum item, which has
1104 * a range that ends at the start of the range we want to insert,
1105 * make sure we don't extend beyond the start offset of the next
1106 * checksum item. If we are at the last item in the leaf, then
1107 * forget the optimization of extending and add a new checksum
1108 * item - it is not worth the complexity of releasing the path,
1109 * getting the first key for the next leaf, repeat the btree
1110 * search, etc, because log trees are temporary anyway and it
1111 * would only save a few bytes of leaf space.
1113 if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
1114 if (path->slots[0] + 1 >=
1115 btrfs_header_nritems(path->nodes[0])) {
1116 ret = find_next_csum_offset(root, path, &next_offset);
1123 ret = find_next_csum_offset(root, path, &next_offset);
1127 tmp = (next_offset - bytenr) >> fs_info->sectorsize_bits;
1129 extend_nr = min_t(int, extend_nr, tmp);
1132 diff = (csum_offset + extend_nr) * csum_size;
1134 MAX_CSUM_ITEMS(fs_info, csum_size) * csum_size);
1136 diff = diff - btrfs_item_size(leaf, path->slots[0]);
1137 diff = min_t(u32, btrfs_leaf_free_space(leaf), diff);
1141 btrfs_extend_item(path, diff);
1147 btrfs_release_path(path);
1152 tmp = sums->len - total_bytes;
1153 tmp >>= fs_info->sectorsize_bits;
1154 tmp = min(tmp, (next_offset - file_key.offset) >>
1155 fs_info->sectorsize_bits);
1157 tmp = max_t(u64, 1, tmp);
1158 tmp = min_t(u64, tmp, MAX_CSUM_ITEMS(fs_info, csum_size));
1159 ins_size = csum_size * tmp;
1161 ins_size = csum_size;
1163 ret = btrfs_insert_empty_item(trans, root, path, &file_key,
1167 if (WARN_ON(ret != 0))
1169 leaf = path->nodes[0];
1171 item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_csum_item);
1172 item_end = (struct btrfs_csum_item *)((unsigned char *)item +
1173 btrfs_item_size(leaf, path->slots[0]));
1174 item = (struct btrfs_csum_item *)((unsigned char *)item +
1175 csum_offset * csum_size);
1177 ins_size = (u32)(sums->len - total_bytes) >> fs_info->sectorsize_bits;
1178 ins_size *= csum_size;
1179 ins_size = min_t(u32, (unsigned long)item_end - (unsigned long)item,
1181 write_extent_buffer(leaf, sums->sums + index, (unsigned long)item,
1185 ins_size /= csum_size;
1186 total_bytes += ins_size * fs_info->sectorsize;
1188 btrfs_mark_buffer_dirty(path->nodes[0]);
1189 if (total_bytes < sums->len) {
1190 btrfs_release_path(path);
1195 btrfs_free_path(path);
1199 void btrfs_extent_item_to_extent_map(struct btrfs_inode *inode,
1200 const struct btrfs_path *path,
1201 struct btrfs_file_extent_item *fi,
1202 const bool new_inline,
1203 struct extent_map *em)
1205 struct btrfs_fs_info *fs_info = inode->root->fs_info;
1206 struct btrfs_root *root = inode->root;
1207 struct extent_buffer *leaf = path->nodes[0];
1208 const int slot = path->slots[0];
1209 struct btrfs_key key;
1210 u64 extent_start, extent_end;
1212 u8 type = btrfs_file_extent_type(leaf, fi);
1213 int compress_type = btrfs_file_extent_compression(leaf, fi);
1215 btrfs_item_key_to_cpu(leaf, &key, slot);
1216 extent_start = key.offset;
1217 extent_end = btrfs_file_extent_end(path);
1218 em->ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi);
1219 em->generation = btrfs_file_extent_generation(leaf, fi);
1220 if (type == BTRFS_FILE_EXTENT_REG ||
1221 type == BTRFS_FILE_EXTENT_PREALLOC) {
1222 em->start = extent_start;
1223 em->len = extent_end - extent_start;
1224 em->orig_start = extent_start -
1225 btrfs_file_extent_offset(leaf, fi);
1226 em->orig_block_len = btrfs_file_extent_disk_num_bytes(leaf, fi);
1227 bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
1229 em->block_start = EXTENT_MAP_HOLE;
1232 if (compress_type != BTRFS_COMPRESS_NONE) {
1233 set_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
1234 em->compress_type = compress_type;
1235 em->block_start = bytenr;
1236 em->block_len = em->orig_block_len;
1238 bytenr += btrfs_file_extent_offset(leaf, fi);
1239 em->block_start = bytenr;
1240 em->block_len = em->len;
1241 if (type == BTRFS_FILE_EXTENT_PREALLOC)
1242 set_bit(EXTENT_FLAG_PREALLOC, &em->flags);
1244 } else if (type == BTRFS_FILE_EXTENT_INLINE) {
1245 em->block_start = EXTENT_MAP_INLINE;
1246 em->start = extent_start;
1247 em->len = extent_end - extent_start;
1249 * Initialize orig_start and block_len with the same values
1250 * as in inode.c:btrfs_get_extent().
1252 em->orig_start = EXTENT_MAP_HOLE;
1253 em->block_len = (u64)-1;
1254 if (!new_inline && compress_type != BTRFS_COMPRESS_NONE) {
1255 set_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
1256 em->compress_type = compress_type;
1260 "unknown file extent item type %d, inode %llu, offset %llu, "
1261 "root %llu", type, btrfs_ino(inode), extent_start,
1262 root->root_key.objectid);
1267 * Returns the end offset (non inclusive) of the file extent item the given path
1268 * points to. If it points to an inline extent, the returned offset is rounded
1269 * up to the sector size.
1271 u64 btrfs_file_extent_end(const struct btrfs_path *path)
1273 const struct extent_buffer *leaf = path->nodes[0];
1274 const int slot = path->slots[0];
1275 struct btrfs_file_extent_item *fi;
1276 struct btrfs_key key;
1279 btrfs_item_key_to_cpu(leaf, &key, slot);
1280 ASSERT(key.type == BTRFS_EXTENT_DATA_KEY);
1281 fi = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item);
1283 if (btrfs_file_extent_type(leaf, fi) == BTRFS_FILE_EXTENT_INLINE) {
1284 end = btrfs_file_extent_ram_bytes(leaf, fi);
1285 end = ALIGN(key.offset + end, leaf->fs_info->sectorsize);
1287 end = key.offset + btrfs_file_extent_num_bytes(leaf, fi);