^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) #include <linux/blkdev.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) #include <linux/iversion.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) #include "compression.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) #include "ctree.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) #include "delalloc-space.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) #include "reflink.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) #include "transaction.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #define BTRFS_MAX_DEDUPE_LEN SZ_16M
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) static int clone_finish_inode_update(struct btrfs_trans_handle *trans,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) struct inode *inode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) u64 endoff,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) const u64 destoff,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) const u64 olen,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) int no_time_update)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) struct btrfs_root *root = BTRFS_I(inode)->root;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) inode_inc_iversion(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) if (!no_time_update)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) inode->i_mtime = inode->i_ctime = current_time(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) * We round up to the block size at eof when determining which
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) * extents to clone above, but shouldn't round up the file size.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) if (endoff > destoff + olen)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) endoff = destoff + olen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) if (endoff > inode->i_size) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) i_size_write(inode, endoff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) btrfs_inode_safe_disk_i_size_write(inode, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) ret = btrfs_update_inode(trans, root, inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) if (ret) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) btrfs_abort_transaction(trans, ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) btrfs_end_transaction(trans);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) ret = btrfs_end_transaction(trans);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) static int copy_inline_to_page(struct btrfs_inode *inode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) const u64 file_offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) char *inline_data,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) const u64 size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) const u64 datal,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) const u8 comp_type)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) const u64 block_size = btrfs_inode_sectorsize(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) const u64 range_end = file_offset + block_size - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) const size_t inline_size = size - btrfs_file_extent_calc_inline_size(0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) char *data_start = inline_data + btrfs_file_extent_calc_inline_size(0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) struct extent_changeset *data_reserved = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) struct page *page = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) struct address_space *mapping = inode->vfs_inode.i_mapping;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) ASSERT(IS_ALIGNED(file_offset, block_size));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) * We have flushed and locked the ranges of the source and destination
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) * inodes, we also have locked the inodes, so we are safe to do a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) * reservation here. Also we must not do the reservation while holding
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) * a transaction open, otherwise we would deadlock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) ret = btrfs_delalloc_reserve_space(inode, &data_reserved, file_offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) block_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) page = find_or_create_page(mapping, file_offset >> PAGE_SHIFT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) btrfs_alloc_write_mask(mapping));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) if (!page) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) ret = -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) set_page_extent_mapped(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) clear_extent_bit(&inode->io_tree, file_offset, range_end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) 0, 0, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) ret = btrfs_set_extent_delalloc(inode, file_offset, range_end, 0, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) * After dirtying the page our caller will need to start a transaction,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) * and if we are low on metadata free space, that can cause flushing of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) * delalloc for all inodes in order to get metadata space released.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) * However we are holding the range locked for the whole duration of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) * the clone/dedupe operation, so we may deadlock if that happens and no
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) * other task releases enough space. So mark this inode as not being
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) * possible to flush to avoid such deadlock. We will clear that flag
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) * when we finish cloning all extents, since a transaction is started
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) * after finding each extent to clone.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) set_bit(BTRFS_INODE_NO_DELALLOC_FLUSH, &inode->runtime_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) if (comp_type == BTRFS_COMPRESS_NONE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) char *map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) map = kmap(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) memcpy(map, data_start, datal);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) flush_dcache_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) kunmap(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) ret = btrfs_decompress(comp_type, data_start, page, 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) inline_size, datal);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) flush_dcache_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) * If our inline data is smaller then the block/page size, then the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) * remaining of the block/page is equivalent to zeroes. We had something
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) * like the following done:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) * $ xfs_io -f -c "pwrite -S 0xab 0 500" file
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) * $ sync # (or fsync)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) * $ xfs_io -c "falloc 0 4K" file
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) * $ xfs_io -c "pwrite -S 0xcd 4K 4K"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) * So what's in the range [500, 4095] corresponds to zeroes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) if (datal < block_size) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) char *map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) map = kmap(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) memset(map + datal, 0, block_size - datal);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) flush_dcache_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) kunmap(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) SetPageUptodate(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) ClearPageChecked(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) set_page_dirty(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) out_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) if (page) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) unlock_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) put_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) btrfs_delalloc_release_space(inode, data_reserved, file_offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) block_size, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) btrfs_delalloc_release_extents(inode, block_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) extent_changeset_free(data_reserved);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) * Deal with cloning of inline extents. We try to copy the inline extent from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) * the source inode to destination inode when possible. When not possible we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) * copy the inline extent's data into the respective page of the inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) static int clone_copy_inline_extent(struct inode *dst,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) struct btrfs_path *path,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) struct btrfs_key *new_key,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) const u64 drop_start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) const u64 datal,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) const u64 size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) const u8 comp_type,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) char *inline_data,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) struct btrfs_trans_handle **trans_out)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) struct btrfs_fs_info *fs_info = btrfs_sb(dst->i_sb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) struct btrfs_root *root = BTRFS_I(dst)->root;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) const u64 aligned_end = ALIGN(new_key->offset + datal,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) fs_info->sectorsize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) struct btrfs_trans_handle *trans = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) struct btrfs_key key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) if (new_key->offset > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) ret = copy_inline_to_page(BTRFS_I(dst), new_key->offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) inline_data, size, datal, comp_type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) key.objectid = btrfs_ino(BTRFS_I(dst));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) key.type = BTRFS_EXTENT_DATA_KEY;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) key.offset = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) if (ret < 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) } else if (ret > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) ret = btrfs_next_leaf(root, path);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) else if (ret > 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) goto copy_inline_extent;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) if (key.objectid == btrfs_ino(BTRFS_I(dst)) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) key.type == BTRFS_EXTENT_DATA_KEY) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) * There's an implicit hole at file offset 0, copy the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) * inline extent's data to the page.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) ASSERT(key.offset > 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) goto copy_to_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) } else if (i_size_read(dst) <= datal) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) struct btrfs_file_extent_item *ei;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) ei = btrfs_item_ptr(path->nodes[0], path->slots[0],
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) struct btrfs_file_extent_item);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) * If it's an inline extent replace it with the source inline
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) * extent, otherwise copy the source inline extent data into
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) * the respective page at the destination inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) if (btrfs_file_extent_type(path->nodes[0], ei) ==
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) BTRFS_FILE_EXTENT_INLINE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) goto copy_inline_extent;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) goto copy_to_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) copy_inline_extent:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) * We have no extent items, or we have an extent at offset 0 which may
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) * or may not be inlined. All these cases are dealt the same way.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) if (i_size_read(dst) > datal) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) * At the destination offset 0 we have either a hole, a regular
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) * extent or an inline extent larger then the one we want to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) * clone. Deal with all these cases by copying the inline extent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) * data into the respective page at the destination inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) goto copy_to_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) * Release path before starting a new transaction so we don't hold locks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) * that would confuse lockdep.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) btrfs_release_path(path);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) * If we end up here it means were copy the inline extent into a leaf
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) * of the destination inode. We know we will drop or adjust at most one
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) * extent item in the destination root.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) * 1 unit - adjusting old extent (we may have to split it)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) * 1 unit - add new extent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) * 1 unit - inode update
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) trans = btrfs_start_transaction(root, 3);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) if (IS_ERR(trans)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) ret = PTR_ERR(trans);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) trans = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) ret = btrfs_drop_extents(trans, root, dst, drop_start, aligned_end, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) ret = btrfs_insert_empty_item(trans, root, path, new_key, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) write_extent_buffer(path->nodes[0], inline_data,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) btrfs_item_ptr_offset(path->nodes[0],
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) path->slots[0]),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) inode_add_bytes(dst, datal);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &BTRFS_I(dst)->runtime_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) ret = btrfs_inode_set_file_extent_range(BTRFS_I(dst), 0, aligned_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) if (!ret && !trans) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) * No transaction here means we copied the inline extent into a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) * page of the destination inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) * 1 unit to update inode item
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) trans = btrfs_start_transaction(root, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) if (IS_ERR(trans)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) ret = PTR_ERR(trans);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) trans = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) if (ret && trans) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) btrfs_abort_transaction(trans, ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) btrfs_end_transaction(trans);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) if (!ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) *trans_out = trans;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) copy_to_page:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) * Release our path because we don't need it anymore and also because
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) * copy_inline_to_page() needs to reserve data and metadata, which may
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) * need to flush delalloc when we are low on available space and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) * therefore cause a deadlock if writeback of an inline extent needs to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) * write to the same leaf or an ordered extent completion needs to write
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) * to the same leaf.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) btrfs_release_path(path);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) ret = copy_inline_to_page(BTRFS_I(dst), new_key->offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) inline_data, size, datal, comp_type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) * btrfs_clone() - clone a range from inode file to another
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) * @src: Inode to clone from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) * @inode: Inode to clone to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) * @off: Offset within source to start clone from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) * @olen: Original length, passed by user, of range to clone
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) * @olen_aligned: Block-aligned value of olen
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) * @destoff: Offset within @inode to start clone
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) * @no_time_update: Whether to update mtime/ctime on the target inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) static int btrfs_clone(struct inode *src, struct inode *inode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) const u64 off, const u64 olen, const u64 olen_aligned,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) const u64 destoff, int no_time_update)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) struct btrfs_path *path = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) struct extent_buffer *leaf;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) struct btrfs_trans_handle *trans;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) char *buf = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) struct btrfs_key key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) u32 nritems;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) int slot;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) const u64 len = olen_aligned;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) u64 last_dest_end = destoff;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) ret = -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) buf = kvmalloc(fs_info->nodesize, GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) if (!buf)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) path = btrfs_alloc_path();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) if (!path) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) kvfree(buf);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) path->reada = READA_FORWARD;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) /* Clone data */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) key.objectid = btrfs_ino(BTRFS_I(src));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) key.type = BTRFS_EXTENT_DATA_KEY;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) key.offset = off;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) while (1) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) u64 next_key_min_offset = key.offset + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) struct btrfs_file_extent_item *extent;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) u64 extent_gen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) int type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) u32 size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) struct btrfs_key new_key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) u64 disko = 0, diskl = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) u64 datao = 0, datal = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) u8 comp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) u64 drop_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) /* Note the key will change type as we walk through the tree */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) path->leave_spinning = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) ret = btrfs_search_slot(NULL, BTRFS_I(src)->root, &key, path,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) 0, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) * First search, if no extent item that starts at offset off was
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) * found but the previous item is an extent item, it's possible
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) * it might overlap our target range, therefore process it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) if (key.offset == off && ret > 0 && path->slots[0] > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) btrfs_item_key_to_cpu(path->nodes[0], &key,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) path->slots[0] - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) if (key.type == BTRFS_EXTENT_DATA_KEY)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) path->slots[0]--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) nritems = btrfs_header_nritems(path->nodes[0]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) process_slot:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) if (path->slots[0] >= nritems) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) ret = btrfs_next_leaf(BTRFS_I(src)->root, path);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397) if (ret > 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) nritems = btrfs_header_nritems(path->nodes[0]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) leaf = path->nodes[0];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) slot = path->slots[0];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) btrfs_item_key_to_cpu(leaf, &key, slot);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) if (key.type > BTRFS_EXTENT_DATA_KEY ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) key.objectid != btrfs_ino(BTRFS_I(src)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) ASSERT(key.type == BTRFS_EXTENT_DATA_KEY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) extent = btrfs_item_ptr(leaf, slot,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) struct btrfs_file_extent_item);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) extent_gen = btrfs_file_extent_generation(leaf, extent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) comp = btrfs_file_extent_compression(leaf, extent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) type = btrfs_file_extent_type(leaf, extent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) if (type == BTRFS_FILE_EXTENT_REG ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) type == BTRFS_FILE_EXTENT_PREALLOC) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) disko = btrfs_file_extent_disk_bytenr(leaf, extent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) diskl = btrfs_file_extent_disk_num_bytes(leaf, extent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) datao = btrfs_file_extent_offset(leaf, extent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) datal = btrfs_file_extent_num_bytes(leaf, extent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) } else if (type == BTRFS_FILE_EXTENT_INLINE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) /* Take upper bound, may be compressed */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) datal = btrfs_file_extent_ram_bytes(leaf, extent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) * The first search might have left us at an extent item that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429) * ends before our target range's start, can happen if we have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) * holes and NO_HOLES feature enabled.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) if (key.offset + datal <= off) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) path->slots[0]++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) goto process_slot;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) } else if (key.offset >= off + len) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) next_key_min_offset = key.offset + datal;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) size = btrfs_item_size_nr(leaf, slot);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) read_extent_buffer(leaf, buf, btrfs_item_ptr_offset(leaf, slot),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) btrfs_release_path(path);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) path->leave_spinning = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) memcpy(&new_key, &key, sizeof(new_key));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) new_key.objectid = btrfs_ino(BTRFS_I(inode));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) if (off <= key.offset)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) new_key.offset = key.offset + destoff - off;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) new_key.offset = destoff;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) * Deal with a hole that doesn't have an extent item that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) * represents it (NO_HOLES feature enabled).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) * This hole is either in the middle of the cloning range or at
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) * the beginning (fully overlaps it or partially overlaps it).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) if (new_key.offset != last_dest_end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) drop_start = last_dest_end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) drop_start = new_key.offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) if (type == BTRFS_FILE_EXTENT_REG ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) type == BTRFS_FILE_EXTENT_PREALLOC) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) struct btrfs_replace_extent_info clone_info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) * a | --- range to clone ---| b
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) * | ------------- extent ------------- |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473) /* Subtract range b */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) if (key.offset + datal > off + len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) datal = off + len - key.offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) /* Subtract range a */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478) if (off > key.offset) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) datao += off - key.offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) datal -= off - key.offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) clone_info.disk_offset = disko;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) clone_info.disk_len = diskl;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) clone_info.data_offset = datao;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) clone_info.data_len = datal;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487) clone_info.file_offset = new_key.offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) clone_info.extent_buf = buf;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) clone_info.is_new_extent = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) ret = btrfs_replace_file_extents(inode, path, drop_start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) new_key.offset + datal - 1, &clone_info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) &trans);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) } else if (type == BTRFS_FILE_EXTENT_INLINE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) * Inline extents always have to start at file offset 0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) * and can never be bigger then the sector size. We can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) * never clone only parts of an inline extent, since all
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) * reflink operations must start at a sector size aligned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) * offset, and the length must be aligned too or end at
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) * the i_size (which implies the whole inlined data).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) ASSERT(key.offset == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) ASSERT(datal <= fs_info->sectorsize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) if (WARN_ON(key.offset != 0) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) WARN_ON(datal > fs_info->sectorsize)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) ret = -EUCLEAN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) ret = clone_copy_inline_extent(inode, path, &new_key,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) drop_start, datal, size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) comp, buf, &trans);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) btrfs_release_path(path);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) * If this is a new extent update the last_reflink_trans of both
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) * inodes. This is used by fsync to make sure it does not log
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) * multiple checksum items with overlapping ranges. For older
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) * extents we don't need to do it since inode logging skips the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526) * checksums for older extents. Also ignore holes and inline
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) * extents because they don't have checksums in the csum tree.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529) if (extent_gen == trans->transid && disko > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) BTRFS_I(src)->last_reflink_trans = trans->transid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) BTRFS_I(inode)->last_reflink_trans = trans->transid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) last_dest_end = ALIGN(new_key.offset + datal,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) fs_info->sectorsize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) ret = clone_finish_inode_update(trans, inode, last_dest_end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) destoff, olen, no_time_update);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) if (new_key.offset + datal >= destoff + len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) btrfs_release_path(path);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) key.offset = next_key_min_offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) if (fatal_signal_pending(current)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) ret = -EINTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) if (last_dest_end < destoff + len) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) * We have an implicit hole that fully or partially overlaps our
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) * cloning range at its end. This means that we either have the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) * NO_HOLES feature enabled or the implicit hole happened due to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) * mixing buffered and direct IO writes against this file.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) btrfs_release_path(path);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563) path->leave_spinning = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) * When using NO_HOLES and we are cloning a range that covers
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) * only a hole (no extents) into a range beyond the current
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) * i_size, punching a hole in the target range will not create
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569) * an extent map defining a hole, because the range starts at or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) * beyond current i_size. If the file previously had an i_size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) * greater than the new i_size set by this clone operation, we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) * need to make sure the next fsync is a full fsync, so that it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) * detects and logs a hole covering a range from the current
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) * i_size to the new i_size. If the clone range covers extents,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) * besides a hole, then we know the full sync flag was already
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) * set by previous calls to btrfs_replace_file_extents() that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) * replaced file extent items.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579) if (last_dest_end >= i_size_read(inode))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) &BTRFS_I(inode)->runtime_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) ret = btrfs_replace_file_extents(inode, path, last_dest_end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) destoff + len - 1, NULL, &trans);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) ret = clone_finish_inode_update(trans, inode, destoff + len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589) destoff, olen, no_time_update);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) btrfs_free_path(path);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) kvfree(buf);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) clear_bit(BTRFS_INODE_NO_DELALLOC_FLUSH, &BTRFS_I(inode)->runtime_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) static void btrfs_double_extent_unlock(struct inode *inode1, u64 loff1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) struct inode *inode2, u64 loff2, u64 len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603) unlock_extent(&BTRFS_I(inode1)->io_tree, loff1, loff1 + len - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) unlock_extent(&BTRFS_I(inode2)->io_tree, loff2, loff2 + len - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) static void btrfs_double_extent_lock(struct inode *inode1, u64 loff1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608) struct inode *inode2, u64 loff2, u64 len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610) if (inode1 < inode2) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) swap(inode1, inode2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) swap(loff1, loff2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) } else if (inode1 == inode2 && loff2 < loff1) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) swap(loff1, loff2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616) lock_extent(&BTRFS_I(inode1)->io_tree, loff1, loff1 + len - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617) lock_extent(&BTRFS_I(inode2)->io_tree, loff2, loff2 + len - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) static int btrfs_extent_same_range(struct inode *src, u64 loff, u64 len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621) struct inode *dst, u64 dst_loff)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) const u64 bs = BTRFS_I(src)->root->fs_info->sb->s_blocksize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) * Lock destination range to serialize with concurrent readpages() and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628) * source range to serialize with relocation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630) btrfs_double_extent_lock(src, loff, dst, dst_loff, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631) ret = btrfs_clone(src, dst, loff, len, ALIGN(len, bs), dst_loff, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) btrfs_double_extent_unlock(src, loff, dst, dst_loff, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637) static int btrfs_extent_same(struct inode *src, u64 loff, u64 olen,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638) struct inode *dst, u64 dst_loff)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640) int ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641) u64 i, tail_len, chunk_count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) struct btrfs_root *root_dst = BTRFS_I(dst)->root;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644) spin_lock(&root_dst->root_item_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645) if (root_dst->send_in_progress) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 646) btrfs_warn_rl(root_dst->fs_info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 647) "cannot deduplicate to root %llu while send operations are using it (%d in progress)",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 648) root_dst->root_key.objectid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 649) root_dst->send_in_progress);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 650) spin_unlock(&root_dst->root_item_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 651) return -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 652) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 653) root_dst->dedupe_in_progress++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 654) spin_unlock(&root_dst->root_item_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 655)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 656) tail_len = olen % BTRFS_MAX_DEDUPE_LEN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 657) chunk_count = div_u64(olen, BTRFS_MAX_DEDUPE_LEN);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 658)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 659) for (i = 0; i < chunk_count; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 660) ret = btrfs_extent_same_range(src, loff, BTRFS_MAX_DEDUPE_LEN,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 661) dst, dst_loff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 662) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 663) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 664)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 665) loff += BTRFS_MAX_DEDUPE_LEN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 666) dst_loff += BTRFS_MAX_DEDUPE_LEN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 667) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 668)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 669) if (tail_len > 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 670) ret = btrfs_extent_same_range(src, loff, tail_len, dst, dst_loff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 671) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 672) spin_lock(&root_dst->root_item_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 673) root_dst->dedupe_in_progress--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 674) spin_unlock(&root_dst->root_item_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 675)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 676) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 677) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 678)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 679) static noinline int btrfs_clone_files(struct file *file, struct file *file_src,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 680) u64 off, u64 olen, u64 destoff)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 681) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 682) struct inode *inode = file_inode(file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 683) struct inode *src = file_inode(file_src);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 684) struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 685) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 686) int wb_ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 687) u64 len = olen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 688) u64 bs = fs_info->sb->s_blocksize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 689)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 690) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 691) * VFS's generic_remap_file_range_prep() protects us from cloning the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 692) * eof block into the middle of a file, which would result in corruption
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 693) * if the file size is not blocksize aligned. So we don't need to check
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 694) * for that case here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 695) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 696) if (off + len == src->i_size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 697) len = ALIGN(src->i_size, bs) - off;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 698)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 699) if (destoff > inode->i_size) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 700) const u64 wb_start = ALIGN_DOWN(inode->i_size, bs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 701)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 702) ret = btrfs_cont_expand(inode, inode->i_size, destoff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 703) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 704) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 705) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 706) * We may have truncated the last block if the inode's size is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 707) * not sector size aligned, so we need to wait for writeback to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 708) * complete before proceeding further, otherwise we can race
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 709) * with cloning and attempt to increment a reference to an
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 710) * extent that no longer exists (writeback completed right after
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 711) * we found the previous extent covering eof and before we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 712) * attempted to increment its reference count).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 713) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 714) ret = btrfs_wait_ordered_range(inode, wb_start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 715) destoff - wb_start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 716) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 717) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 718) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 719)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 720) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 721) * Lock destination range to serialize with concurrent readpages() and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 722) * source range to serialize with relocation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 723) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 724) btrfs_double_extent_lock(src, off, inode, destoff, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 725) ret = btrfs_clone(src, inode, off, olen, len, destoff, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 726) btrfs_double_extent_unlock(src, off, inode, destoff, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 727)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 728) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 729) * We may have copied an inline extent into a page of the destination
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 730) * range, so wait for writeback to complete before truncating pages
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 731) * from the page cache. This is a rare case.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 732) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 733) wb_ret = btrfs_wait_ordered_range(inode, destoff, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 734) ret = ret ? ret : wb_ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 735) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 736) * Truncate page cache pages so that future reads will see the cloned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 737) * data immediately and not the previous data.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 738) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 739) truncate_inode_pages_range(&inode->i_data,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 740) round_down(destoff, PAGE_SIZE),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 741) round_up(destoff + len, PAGE_SIZE) - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 742)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 743) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 744) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 745)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 746) static int btrfs_remap_file_range_prep(struct file *file_in, loff_t pos_in,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 747) struct file *file_out, loff_t pos_out,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 748) loff_t *len, unsigned int remap_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 749) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 750) struct inode *inode_in = file_inode(file_in);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 751) struct inode *inode_out = file_inode(file_out);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 752) u64 bs = BTRFS_I(inode_out)->root->fs_info->sb->s_blocksize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 753) bool same_inode = inode_out == inode_in;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 754) u64 wb_len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 755) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 756)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 757) if (!(remap_flags & REMAP_FILE_DEDUP)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 758) struct btrfs_root *root_out = BTRFS_I(inode_out)->root;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 759)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 760) if (btrfs_root_readonly(root_out))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 761) return -EROFS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 762)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 763) if (file_in->f_path.mnt != file_out->f_path.mnt ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 764) inode_in->i_sb != inode_out->i_sb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 765) return -EXDEV;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 766) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 767)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 768) /* Don't make the dst file partly checksummed */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 769) if ((BTRFS_I(inode_in)->flags & BTRFS_INODE_NODATASUM) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 770) (BTRFS_I(inode_out)->flags & BTRFS_INODE_NODATASUM)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 771) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 772) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 773)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 774) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 775) * Now that the inodes are locked, we need to start writeback ourselves
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 776) * and can not rely on the writeback from the VFS's generic helper
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 777) * generic_remap_file_range_prep() because:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 778) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 779) * 1) For compression we must call filemap_fdatawrite_range() range
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 780) * twice (btrfs_fdatawrite_range() does it for us), and the generic
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 781) * helper only calls it once;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 782) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 783) * 2) filemap_fdatawrite_range(), called by the generic helper only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 784) * waits for the writeback to complete, i.e. for IO to be done, and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 785) * not for the ordered extents to complete. We need to wait for them
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 786) * to complete so that new file extent items are in the fs tree.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 787) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 788) if (*len == 0 && !(remap_flags & REMAP_FILE_DEDUP))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 789) wb_len = ALIGN(inode_in->i_size, bs) - ALIGN_DOWN(pos_in, bs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 790) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 791) wb_len = ALIGN(*len, bs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 792)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 793) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 794) * Since we don't lock ranges, wait for ongoing lockless dio writes (as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 795) * any in progress could create its ordered extents after we wait for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 796) * existing ordered extents below).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 797) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 798) inode_dio_wait(inode_in);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 799) if (!same_inode)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 800) inode_dio_wait(inode_out);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 801)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 802) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 803) * Workaround to make sure NOCOW buffered write reach disk as NOCOW.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 804) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 805) * Btrfs' back references do not have a block level granularity, they
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 806) * work at the whole extent level.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 807) * NOCOW buffered write without data space reserved may not be able
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 808) * to fall back to CoW due to lack of data space, thus could cause
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 809) * data loss.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 810) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 811) * Here we take a shortcut by flushing the whole inode, so that all
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 812) * nocow write should reach disk as nocow before we increase the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 813) * reference of the extent. We could do better by only flushing NOCOW
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 814) * data, but that needs extra accounting.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 815) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 816) * Also we don't need to check ASYNC_EXTENT, as async extent will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 817) * CoWed anyway, not affecting nocow part.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 818) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 819) ret = filemap_flush(inode_in->i_mapping);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 820) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 821) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 822)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 823) ret = btrfs_wait_ordered_range(inode_in, ALIGN_DOWN(pos_in, bs),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 824) wb_len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 825) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 826) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 827) ret = btrfs_wait_ordered_range(inode_out, ALIGN_DOWN(pos_out, bs),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 828) wb_len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 829) if (ret < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 830) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 831)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 832) return generic_remap_file_range_prep(file_in, pos_in, file_out, pos_out,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 833) len, remap_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 834) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 835)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 836) loff_t btrfs_remap_file_range(struct file *src_file, loff_t off,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 837) struct file *dst_file, loff_t destoff, loff_t len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 838) unsigned int remap_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 839) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 840) struct inode *src_inode = file_inode(src_file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 841) struct inode *dst_inode = file_inode(dst_file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 842) bool same_inode = dst_inode == src_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 843) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 844)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 845) if (remap_flags & ~(REMAP_FILE_DEDUP | REMAP_FILE_ADVISORY))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 846) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 847)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 848) if (same_inode)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 849) inode_lock(src_inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 850) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 851) lock_two_nondirectories(src_inode, dst_inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 852)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 853) ret = btrfs_remap_file_range_prep(src_file, off, dst_file, destoff,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 854) &len, remap_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 855) if (ret < 0 || len == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 856) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 857)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 858) if (remap_flags & REMAP_FILE_DEDUP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 859) ret = btrfs_extent_same(src_inode, off, len, dst_inode, destoff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 860) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 861) ret = btrfs_clone_files(dst_file, src_file, off, len, destoff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 862)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 863) out_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 864) if (same_inode)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 865) inode_unlock(src_inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 866) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 867) unlock_two_nondirectories(src_inode, dst_inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 868)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 869) return ret < 0 ? ret : len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 870) }
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags