2017-01-27 15:56:37 +00:00
|
|
|
/*
|
|
|
|
* This program is free software; you can redistribute it and/or
|
|
|
|
* modify it under the terms of the GNU General Public
|
|
|
|
* License v2 as published by the Free Software Foundation.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
* General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public
|
|
|
|
* License along with this program; if not, write to the
|
|
|
|
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
|
|
|
|
* Boston, MA 021110-1307, USA.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#if BTRFSCONVERT_EXT2
|
|
|
|
|
|
|
|
#include "kerncompat.h"
|
2022-09-15 11:53:43 +00:00
|
|
|
#include <sys/stat.h>
|
2017-03-12 17:46:52 +00:00
|
|
|
#include <linux/limits.h>
|
2022-09-15 11:53:43 +00:00
|
|
|
#include <errno.h>
|
2017-07-12 20:05:38 +00:00
|
|
|
#include <pthread.h>
|
2022-09-15 11:53:43 +00:00
|
|
|
#include <limits.h>
|
|
|
|
#include <stdio.h>
|
|
|
|
#include <stdlib.h>
|
|
|
|
#include <string.h>
|
|
|
|
#include "kernel-lib/sizes.h"
|
2020-08-18 13:56:04 +00:00
|
|
|
#include "kernel-shared/transaction.h"
|
2023-04-19 21:17:17 +00:00
|
|
|
#include "kernel-shared/file-item.h"
|
2022-09-15 11:53:43 +00:00
|
|
|
#include "common/extent-cache.h"
|
|
|
|
#include "common/messages.h"
|
2017-01-27 15:56:37 +00:00
|
|
|
#include "convert/common.h"
|
2022-09-15 11:53:43 +00:00
|
|
|
#include "convert/source-fs.h"
|
2017-01-27 15:56:37 +00:00
|
|
|
#include "convert/source-ext2.h"
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Open Ext2fs in readonly mode, read block allocation bitmap and
|
|
|
|
* inode bitmap into memory.
|
|
|
|
*/
|
|
|
|
static int ext2_open_fs(struct btrfs_convert_context *cctx, const char *name)
|
|
|
|
{
|
|
|
|
errcode_t ret;
|
|
|
|
ext2_filsys ext2_fs;
|
|
|
|
ext2_ino_t ino;
|
|
|
|
u32 ro_feature;
|
2017-06-03 09:57:45 +00:00
|
|
|
int open_flag = EXT2_FLAG_SOFTSUPP_FEATURES | EXT2_FLAG_64BITS;
|
2017-01-27 15:56:37 +00:00
|
|
|
|
2017-06-03 09:57:45 +00:00
|
|
|
ret = ext2fs_open(name, open_flag, 0, 0, unix_io_manager, &ext2_fs);
|
2017-01-27 15:56:37 +00:00
|
|
|
if (ret) {
|
2017-08-29 17:42:10 +00:00
|
|
|
if (ret != EXT2_ET_BAD_MAGIC)
|
2022-09-19 10:17:19 +00:00
|
|
|
error("ext2fs_open: %s", error_message(ret));
|
2017-01-27 15:56:37 +00:00
|
|
|
return -1;
|
|
|
|
}
|
2021-03-10 15:16:25 +00:00
|
|
|
|
|
|
|
if (ext2_fs->super->s_feature_incompat & EXT3_FEATURE_INCOMPAT_RECOVER) {
|
|
|
|
error("source filesystem requires recovery, run e2fsck first");
|
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
|
2017-01-27 15:56:37 +00:00
|
|
|
/*
|
|
|
|
* We need to know exactly the used space, some RO compat flags like
|
|
|
|
* BIGALLOC will affect how used space is present.
|
|
|
|
* So we need manuall check any unsupported RO compat flags
|
|
|
|
*/
|
|
|
|
ro_feature = ext2_fs->super->s_feature_ro_compat;
|
|
|
|
if (ro_feature & ~EXT2_LIB_FEATURE_RO_COMPAT_SUPP) {
|
|
|
|
error(
|
|
|
|
"unsupported RO features detected: %x, abort convert to avoid possible corruption",
|
|
|
|
ro_feature & ~EXT2_LIB_FEATURE_COMPAT_SUPP);
|
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
ret = ext2fs_read_inode_bitmap(ext2_fs);
|
|
|
|
if (ret) {
|
2022-09-19 10:17:19 +00:00
|
|
|
error("ext2fs_read_inode_bitmap: %s", error_message(ret));
|
2017-01-27 15:56:37 +00:00
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
ret = ext2fs_read_block_bitmap(ext2_fs);
|
|
|
|
if (ret) {
|
2022-09-19 10:17:19 +00:00
|
|
|
error("ext2fs_read_block_bitmap: %s", error_message(ret));
|
2017-01-27 15:56:37 +00:00
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* search each block group for a free inode. this set up
|
|
|
|
* uninit block/inode bitmaps appropriately.
|
|
|
|
*/
|
|
|
|
ino = 1;
|
|
|
|
while (ino <= ext2_fs->super->s_inodes_count) {
|
|
|
|
ext2_ino_t foo;
|
|
|
|
ext2fs_new_inode(ext2_fs, ino, 0, NULL, &foo);
|
|
|
|
ino += EXT2_INODES_PER_GROUP(ext2_fs->super);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!(ext2_fs->super->s_feature_incompat &
|
|
|
|
EXT2_FEATURE_INCOMPAT_FILETYPE)) {
|
|
|
|
error("filetype feature is missing");
|
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
|
|
|
|
cctx->fs_data = ext2_fs;
|
|
|
|
cctx->blocksize = ext2_fs->blocksize;
|
2022-06-07 10:24:21 +00:00
|
|
|
cctx->block_count = ext2fs_blocks_count(ext2_fs->super);
|
|
|
|
cctx->total_bytes = cctx->block_count * cctx->blocksize;
|
2021-08-17 20:39:33 +00:00
|
|
|
cctx->label = strndup((char *)ext2_fs->super->s_volume_name, 16);
|
2017-01-27 15:56:37 +00:00
|
|
|
cctx->first_data_block = ext2_fs->super->s_first_data_block;
|
|
|
|
cctx->inodes_count = ext2_fs->super->s_inodes_count;
|
|
|
|
cctx->free_inodes_count = ext2_fs->super->s_free_inodes_count;
|
2021-08-17 14:47:18 +00:00
|
|
|
memcpy(cctx->fs_uuid, ext2_fs->super->s_uuid, SOURCE_FS_UUID_SIZE);
|
2017-01-27 15:56:37 +00:00
|
|
|
return 0;
|
|
|
|
fail:
|
|
|
|
ext2fs_close(ext2_fs);
|
2018-02-14 08:09:56 +00:00
|
|
|
ext2fs_free(ext2_fs);
|
2017-01-27 15:56:37 +00:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __ext2_add_one_block(ext2_filsys fs, char *bitmap,
|
|
|
|
unsigned long group_nr, struct cache_tree *used)
|
|
|
|
{
|
|
|
|
unsigned long offset;
|
|
|
|
unsigned i;
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
offset = fs->super->s_first_data_block;
|
|
|
|
offset /= EXT2FS_CLUSTER_RATIO(fs);
|
|
|
|
offset += group_nr * EXT2_CLUSTERS_PER_GROUP(fs->super);
|
|
|
|
for (i = 0; i < EXT2_CLUSTERS_PER_GROUP(fs->super); i++) {
|
|
|
|
if ((i + offset) >= ext2fs_blocks_count(fs->super))
|
|
|
|
break;
|
|
|
|
|
|
|
|
if (ext2fs_test_bit(i, bitmap)) {
|
|
|
|
u64 start;
|
|
|
|
|
|
|
|
start = (i + offset) * EXT2FS_CLUSTER_RATIO(fs);
|
|
|
|
start *= fs->blocksize;
|
|
|
|
ret = add_merge_cache_extent(used, start,
|
|
|
|
fs->blocksize);
|
|
|
|
if (ret < 0)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Read all used ext2 space into cctx->used cache tree
|
|
|
|
*/
|
|
|
|
static int ext2_read_used_space(struct btrfs_convert_context *cctx)
|
|
|
|
{
|
|
|
|
ext2_filsys fs = (ext2_filsys)cctx->fs_data;
|
|
|
|
blk64_t blk_itr = EXT2FS_B2C(fs, fs->super->s_first_data_block);
|
2017-02-01 14:06:04 +00:00
|
|
|
struct cache_tree *used_tree = &cctx->used_space;
|
2017-01-27 15:56:37 +00:00
|
|
|
char *block_bitmap = NULL;
|
|
|
|
unsigned long i;
|
|
|
|
int block_nbytes;
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
block_nbytes = EXT2_CLUSTERS_PER_GROUP(fs->super) / 8;
|
2017-01-30 15:58:57 +00:00
|
|
|
if (!block_nbytes) {
|
|
|
|
error("EXT2_CLUSTERS_PER_GROUP too small: %llu",
|
|
|
|
(unsigned long long)(EXT2_CLUSTERS_PER_GROUP(fs->super)));
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
2017-01-27 15:56:37 +00:00
|
|
|
|
|
|
|
block_bitmap = malloc(block_nbytes);
|
|
|
|
if (!block_bitmap)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
for (i = 0; i < fs->group_desc_count; i++) {
|
2017-06-03 09:57:45 +00:00
|
|
|
ret = ext2fs_get_block_bitmap_range2(fs->block_map, blk_itr,
|
2017-01-27 15:56:37 +00:00
|
|
|
block_nbytes * 8, block_bitmap);
|
|
|
|
if (ret) {
|
|
|
|
error("fail to get bitmap from ext2, %s",
|
2017-09-12 13:23:05 +00:00
|
|
|
error_message(ret));
|
|
|
|
ret = -EINVAL;
|
2017-01-27 15:56:37 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
ret = __ext2_add_one_block(fs, block_bitmap, i, used_tree);
|
|
|
|
if (ret < 0) {
|
2018-10-25 12:10:54 +00:00
|
|
|
errno = -ret;
|
|
|
|
error("fail to build used space tree, %m");
|
2017-01-27 15:56:37 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
blk_itr += EXT2_CLUSTERS_PER_GROUP(fs->super);
|
|
|
|
}
|
|
|
|
|
|
|
|
free(block_bitmap);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void ext2_close_fs(struct btrfs_convert_context *cctx)
|
|
|
|
{
|
2021-08-17 20:39:33 +00:00
|
|
|
if (cctx->label) {
|
|
|
|
free(cctx->label);
|
|
|
|
cctx->label = NULL;
|
2017-01-27 15:56:37 +00:00
|
|
|
}
|
|
|
|
ext2fs_close(cctx->fs_data);
|
2018-02-14 08:09:56 +00:00
|
|
|
ext2fs_free(cctx->fs_data);
|
2017-01-27 15:56:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static u8 ext2_filetype_conversion_table[EXT2_FT_MAX] = {
|
|
|
|
[EXT2_FT_UNKNOWN] = BTRFS_FT_UNKNOWN,
|
|
|
|
[EXT2_FT_REG_FILE] = BTRFS_FT_REG_FILE,
|
|
|
|
[EXT2_FT_DIR] = BTRFS_FT_DIR,
|
|
|
|
[EXT2_FT_CHRDEV] = BTRFS_FT_CHRDEV,
|
|
|
|
[EXT2_FT_BLKDEV] = BTRFS_FT_BLKDEV,
|
|
|
|
[EXT2_FT_FIFO] = BTRFS_FT_FIFO,
|
|
|
|
[EXT2_FT_SOCK] = BTRFS_FT_SOCK,
|
|
|
|
[EXT2_FT_SYMLINK] = BTRFS_FT_SYMLINK,
|
|
|
|
};
|
|
|
|
|
|
|
|
static int ext2_dir_iterate_proc(ext2_ino_t dir, int entry,
|
|
|
|
struct ext2_dir_entry *dirent,
|
|
|
|
int offset, int blocksize,
|
|
|
|
char *buf,void *priv_data)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
int file_type;
|
|
|
|
u64 objectid;
|
|
|
|
char dotdot[] = "..";
|
|
|
|
struct dir_iterate_data *idata = (struct dir_iterate_data *)priv_data;
|
|
|
|
int name_len;
|
|
|
|
|
|
|
|
name_len = dirent->name_len & 0xFF;
|
|
|
|
|
|
|
|
objectid = dirent->inode + INO_OFFSET;
|
|
|
|
if (!strncmp(dirent->name, dotdot, name_len)) {
|
|
|
|
if (name_len == 2) {
|
2022-09-30 12:44:38 +00:00
|
|
|
if (idata->parent != 0) {
|
|
|
|
error("dotdot entry parent not zero: %llu",
|
|
|
|
idata->parent);
|
|
|
|
return BLOCK_ABORT;
|
|
|
|
}
|
2017-01-27 15:56:37 +00:00
|
|
|
idata->parent = objectid;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
if (dirent->inode < EXT2_GOOD_OLD_FIRST_INO)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
file_type = dirent->name_len >> 8;
|
2022-09-30 12:44:38 +00:00
|
|
|
if (file_type >= EXT2_FT_MAX) {
|
|
|
|
error("invalid file type %d for %*s", file_type, name_len, dirent->name);
|
|
|
|
return BLOCK_ABORT;
|
|
|
|
}
|
2017-01-27 15:56:37 +00:00
|
|
|
|
|
|
|
ret = convert_insert_dirent(idata->trans, idata->root, dirent->name,
|
|
|
|
name_len, idata->objectid, objectid,
|
|
|
|
ext2_filetype_conversion_table[file_type],
|
|
|
|
idata->index_cnt, idata->inode);
|
|
|
|
if (ret < 0) {
|
|
|
|
idata->errcode = ret;
|
|
|
|
return BLOCK_ABORT;
|
|
|
|
}
|
|
|
|
|
|
|
|
idata->index_cnt++;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ext2_create_dir_entries(struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_root *root, u64 objectid,
|
|
|
|
struct btrfs_inode_item *btrfs_inode,
|
|
|
|
ext2_filsys ext2_fs, ext2_ino_t ext2_ino)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
errcode_t err;
|
|
|
|
struct dir_iterate_data data = {
|
|
|
|
.trans = trans,
|
|
|
|
.root = root,
|
|
|
|
.inode = btrfs_inode,
|
|
|
|
.objectid = objectid,
|
|
|
|
.index_cnt = 2,
|
|
|
|
.parent = 0,
|
|
|
|
.errcode = 0,
|
|
|
|
};
|
|
|
|
|
|
|
|
err = ext2fs_dir_iterate2(ext2_fs, ext2_ino, 0, NULL,
|
|
|
|
ext2_dir_iterate_proc, &data);
|
|
|
|
if (err)
|
|
|
|
goto error;
|
|
|
|
ret = data.errcode;
|
|
|
|
if (ret == 0 && data.parent == objectid) {
|
|
|
|
ret = btrfs_insert_inode_ref(trans, root, "..", 2,
|
|
|
|
objectid, objectid, 0);
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
error:
|
2022-09-19 10:17:19 +00:00
|
|
|
error("ext2fs_dir_iterate2: %s", error_message(err));
|
2017-01-27 15:56:37 +00:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ext2_block_iterate_proc(ext2_filsys fs, blk_t *blocknr,
|
|
|
|
e2_blkcnt_t blockcnt, blk_t ref_block,
|
|
|
|
int ref_offset, void *priv_data)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
struct blk_iterate_data *idata;
|
|
|
|
idata = (struct blk_iterate_data *)priv_data;
|
|
|
|
ret = block_iterate_proc(*blocknr, blockcnt, idata);
|
|
|
|
if (ret) {
|
|
|
|
idata->errcode = ret;
|
|
|
|
return BLOCK_ABORT;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* traverse file's data blocks, record these data blocks as file extents.
|
|
|
|
*/
|
|
|
|
static int ext2_create_file_extents(struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_root *root, u64 objectid,
|
|
|
|
struct btrfs_inode_item *btrfs_inode,
|
|
|
|
ext2_filsys ext2_fs, ext2_ino_t ext2_ino,
|
2017-01-30 16:52:15 +00:00
|
|
|
u32 convert_flags)
|
2017-01-27 15:56:37 +00:00
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
char *buffer = NULL;
|
|
|
|
errcode_t err;
|
|
|
|
u32 last_block;
|
2017-05-18 02:14:19 +00:00
|
|
|
u32 sectorsize = root->fs_info->sectorsize;
|
2017-01-27 15:56:37 +00:00
|
|
|
u64 inode_size = btrfs_stack_inode_size(btrfs_inode);
|
|
|
|
struct blk_iterate_data data;
|
|
|
|
|
|
|
|
init_blk_iterate_data(&data, trans, root, btrfs_inode, objectid,
|
2017-01-30 16:52:15 +00:00
|
|
|
convert_flags & CONVERT_FLAG_DATACSUM);
|
2017-01-27 15:56:37 +00:00
|
|
|
|
|
|
|
err = ext2fs_block_iterate2(ext2_fs, ext2_ino, BLOCK_FLAG_DATA_ONLY,
|
|
|
|
NULL, ext2_block_iterate_proc, &data);
|
|
|
|
if (err)
|
|
|
|
goto error;
|
|
|
|
ret = data.errcode;
|
|
|
|
if (ret)
|
|
|
|
goto fail;
|
2017-01-30 16:52:15 +00:00
|
|
|
if ((convert_flags & CONVERT_FLAG_INLINE_DATA) && data.first_block == 0
|
2018-03-20 06:42:24 +00:00
|
|
|
&& data.num_blocks > 0 && inode_size < sectorsize
|
2018-01-26 07:26:02 +00:00
|
|
|
&& inode_size <= BTRFS_MAX_INLINE_DATA_SIZE(root->fs_info)) {
|
2017-01-27 15:56:37 +00:00
|
|
|
u64 num_bytes = data.num_blocks * sectorsize;
|
|
|
|
u64 disk_bytenr = data.disk_block * sectorsize;
|
|
|
|
u64 nbytes;
|
|
|
|
|
|
|
|
buffer = malloc(num_bytes);
|
|
|
|
if (!buffer)
|
|
|
|
return -ENOMEM;
|
|
|
|
ret = read_disk_extent(root, disk_bytenr, num_bytes, buffer);
|
|
|
|
if (ret)
|
|
|
|
goto fail;
|
|
|
|
if (num_bytes > inode_size)
|
|
|
|
num_bytes = inode_size;
|
|
|
|
ret = btrfs_insert_inline_extent(trans, root, objectid,
|
|
|
|
0, buffer, num_bytes);
|
|
|
|
if (ret)
|
|
|
|
goto fail;
|
|
|
|
nbytes = btrfs_stack_inode_nbytes(btrfs_inode) + num_bytes;
|
|
|
|
btrfs_set_stack_inode_nbytes(btrfs_inode, nbytes);
|
|
|
|
} else if (data.num_blocks > 0) {
|
|
|
|
ret = record_file_blocks(&data, data.first_block,
|
|
|
|
data.disk_block, data.num_blocks);
|
|
|
|
if (ret)
|
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
data.first_block += data.num_blocks;
|
|
|
|
last_block = (inode_size + sectorsize - 1) / sectorsize;
|
|
|
|
if (last_block > data.first_block) {
|
|
|
|
ret = record_file_blocks(&data, data.first_block, 0,
|
|
|
|
last_block - data.first_block);
|
|
|
|
}
|
|
|
|
fail:
|
|
|
|
free(buffer);
|
|
|
|
return ret;
|
|
|
|
error:
|
2022-09-19 10:17:19 +00:00
|
|
|
error("ext2fs_block_iterate2: %s", error_message(err));
|
2017-01-27 15:56:37 +00:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
2017-01-31 10:08:09 +00:00
|
|
|
static int ext2_create_symlink(struct btrfs_trans_handle *trans,
|
2017-01-27 15:56:37 +00:00
|
|
|
struct btrfs_root *root, u64 objectid,
|
|
|
|
struct btrfs_inode_item *btrfs_inode,
|
|
|
|
ext2_filsys ext2_fs, ext2_ino_t ext2_ino,
|
|
|
|
struct ext2_inode *ext2_inode)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
char *pathname;
|
|
|
|
u64 inode_size = btrfs_stack_inode_size(btrfs_inode);
|
2017-06-03 09:57:45 +00:00
|
|
|
if (ext2fs_inode_data_blocks2(ext2_fs, ext2_inode)) {
|
2017-01-27 15:56:37 +00:00
|
|
|
btrfs_set_stack_inode_size(btrfs_inode, inode_size + 1);
|
|
|
|
ret = ext2_create_file_extents(trans, root, objectid,
|
2017-01-30 16:52:15 +00:00
|
|
|
btrfs_inode, ext2_fs, ext2_ino,
|
|
|
|
CONVERT_FLAG_DATACSUM |
|
|
|
|
CONVERT_FLAG_INLINE_DATA);
|
2017-01-27 15:56:37 +00:00
|
|
|
btrfs_set_stack_inode_size(btrfs_inode, inode_size);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
pathname = (char *)&(ext2_inode->i_block[0]);
|
|
|
|
BUG_ON(pathname[inode_size] != 0);
|
|
|
|
ret = btrfs_insert_inline_extent(trans, root, objectid, 0,
|
|
|
|
pathname, inode_size + 1);
|
|
|
|
btrfs_set_stack_inode_nbytes(btrfs_inode, inode_size + 1);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Following xattr/acl related codes are based on codes in
|
|
|
|
* fs/ext3/xattr.c and fs/ext3/acl.c
|
|
|
|
*/
|
|
|
|
#define EXT2_XATTR_BHDR(ptr) ((struct ext2_ext_attr_header *)(ptr))
|
|
|
|
#define EXT2_XATTR_BFIRST(ptr) \
|
|
|
|
((struct ext2_ext_attr_entry *)(EXT2_XATTR_BHDR(ptr) + 1))
|
|
|
|
#define EXT2_XATTR_IHDR(inode) \
|
|
|
|
((struct ext2_ext_attr_header *) ((void *)(inode) + \
|
|
|
|
EXT2_GOOD_OLD_INODE_SIZE + (inode)->i_extra_isize))
|
|
|
|
#define EXT2_XATTR_IFIRST(inode) \
|
|
|
|
((struct ext2_ext_attr_entry *) ((void *)EXT2_XATTR_IHDR(inode) + \
|
|
|
|
sizeof(EXT2_XATTR_IHDR(inode)->h_magic)))
|
|
|
|
|
|
|
|
static int ext2_xattr_check_names(struct ext2_ext_attr_entry *entry,
|
|
|
|
const void *end)
|
|
|
|
{
|
|
|
|
struct ext2_ext_attr_entry *next;
|
|
|
|
|
|
|
|
while (!EXT2_EXT_IS_LAST_ENTRY(entry)) {
|
|
|
|
next = EXT2_EXT_ATTR_NEXT(entry);
|
|
|
|
if ((void *)next >= end)
|
|
|
|
return -EIO;
|
|
|
|
entry = next;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ext2_xattr_check_block(const char *buf, size_t size)
|
|
|
|
{
|
|
|
|
int error;
|
|
|
|
struct ext2_ext_attr_header *header = EXT2_XATTR_BHDR(buf);
|
|
|
|
|
|
|
|
if (header->h_magic != EXT2_EXT_ATTR_MAGIC ||
|
|
|
|
header->h_blocks != 1)
|
|
|
|
return -EIO;
|
|
|
|
error = ext2_xattr_check_names(EXT2_XATTR_BFIRST(buf), buf + size);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ext2_xattr_check_entry(struct ext2_ext_attr_entry *entry,
|
|
|
|
size_t size)
|
|
|
|
{
|
|
|
|
size_t value_size = entry->e_value_size;
|
|
|
|
|
2018-03-14 00:56:57 +00:00
|
|
|
if (value_size > size || entry->e_value_offs + value_size > size)
|
2017-01-27 15:56:37 +00:00
|
|
|
return -EIO;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ext2_acl_to_xattr(void *dst, const void *src,
|
|
|
|
size_t dst_size, size_t src_size)
|
|
|
|
{
|
|
|
|
int i, count;
|
|
|
|
const void *end = src + src_size;
|
|
|
|
acl_ea_header *ext_acl = (acl_ea_header *)dst;
|
|
|
|
acl_ea_entry *dst_entry = ext_acl->a_entries;
|
|
|
|
ext2_acl_entry *src_entry;
|
|
|
|
|
|
|
|
if (src_size < sizeof(ext2_acl_header))
|
|
|
|
goto fail;
|
|
|
|
if (((ext2_acl_header *)src)->a_version !=
|
|
|
|
cpu_to_le32(EXT2_ACL_VERSION))
|
|
|
|
goto fail;
|
|
|
|
src += sizeof(ext2_acl_header);
|
|
|
|
count = ext2_acl_count(src_size);
|
|
|
|
if (count <= 0)
|
|
|
|
goto fail;
|
|
|
|
|
2022-09-30 12:44:38 +00:00
|
|
|
if (dst_size < acl_ea_size(count)) {
|
|
|
|
error("not enough space to store ACLs");
|
|
|
|
goto fail;
|
|
|
|
}
|
2017-01-27 15:56:37 +00:00
|
|
|
ext_acl->a_version = cpu_to_le32(ACL_EA_VERSION);
|
|
|
|
for (i = 0; i < count; i++, dst_entry++) {
|
|
|
|
src_entry = (ext2_acl_entry *)src;
|
|
|
|
if (src + sizeof(ext2_acl_entry_short) > end)
|
|
|
|
goto fail;
|
|
|
|
dst_entry->e_tag = src_entry->e_tag;
|
|
|
|
dst_entry->e_perm = src_entry->e_perm;
|
|
|
|
switch (le16_to_cpu(src_entry->e_tag)) {
|
|
|
|
case ACL_USER_OBJ:
|
|
|
|
case ACL_GROUP_OBJ:
|
|
|
|
case ACL_MASK:
|
|
|
|
case ACL_OTHER:
|
|
|
|
src += sizeof(ext2_acl_entry_short);
|
|
|
|
dst_entry->e_id = cpu_to_le32(ACL_UNDEFINED_ID);
|
|
|
|
break;
|
|
|
|
case ACL_USER:
|
|
|
|
case ACL_GROUP:
|
|
|
|
src += sizeof(ext2_acl_entry);
|
|
|
|
if (src > end)
|
|
|
|
goto fail;
|
|
|
|
dst_entry->e_id = src_entry->e_id;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
goto fail;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (src != end)
|
|
|
|
goto fail;
|
|
|
|
return 0;
|
|
|
|
fail:
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static char *xattr_prefix_table[] = {
|
|
|
|
[1] = "user.",
|
|
|
|
[2] = "system.posix_acl_access",
|
|
|
|
[3] = "system.posix_acl_default",
|
|
|
|
[4] = "trusted.",
|
|
|
|
[6] = "security.",
|
|
|
|
};
|
|
|
|
|
|
|
|
static int ext2_copy_single_xattr(struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_root *root, u64 objectid,
|
|
|
|
struct ext2_ext_attr_entry *entry,
|
|
|
|
const void *data, u32 datalen)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
int name_len;
|
|
|
|
int name_index;
|
|
|
|
void *databuf = NULL;
|
|
|
|
char namebuf[XATTR_NAME_MAX + 1];
|
|
|
|
|
|
|
|
name_index = entry->e_name_index;
|
|
|
|
if (name_index >= ARRAY_SIZE(xattr_prefix_table) ||
|
|
|
|
xattr_prefix_table[name_index] == NULL)
|
|
|
|
return -EOPNOTSUPP;
|
|
|
|
name_len = strlen(xattr_prefix_table[name_index]) +
|
|
|
|
entry->e_name_len;
|
|
|
|
if (name_len >= sizeof(namebuf))
|
|
|
|
return -ERANGE;
|
|
|
|
|
|
|
|
if (name_index == 2 || name_index == 3) {
|
|
|
|
size_t bufsize = acl_ea_size(ext2_acl_count(datalen));
|
|
|
|
databuf = malloc(bufsize);
|
|
|
|
if (!databuf)
|
|
|
|
return -ENOMEM;
|
|
|
|
ret = ext2_acl_to_xattr(databuf, data, bufsize, datalen);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
data = databuf;
|
|
|
|
datalen = bufsize;
|
|
|
|
}
|
|
|
|
strncpy(namebuf, xattr_prefix_table[name_index], XATTR_NAME_MAX);
|
|
|
|
strncat(namebuf, EXT2_EXT_ATTR_NAME(entry), entry->e_name_len);
|
2018-01-26 07:26:00 +00:00
|
|
|
if (name_len + datalen > BTRFS_LEAF_DATA_SIZE(root->fs_info) -
|
2017-01-27 15:56:37 +00:00
|
|
|
sizeof(struct btrfs_item) - sizeof(struct btrfs_dir_item)) {
|
2022-09-19 10:17:19 +00:00
|
|
|
error("skip large xattr on inode %llu name %.*s",
|
2017-01-27 15:56:37 +00:00
|
|
|
objectid - INO_OFFSET, name_len, namebuf);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
ret = btrfs_insert_xattr_item(trans, root, namebuf, name_len,
|
|
|
|
data, datalen, objectid);
|
|
|
|
out:
|
|
|
|
free(databuf);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ext2_copy_extended_attrs(struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_root *root, u64 objectid,
|
|
|
|
struct btrfs_inode_item *btrfs_inode,
|
|
|
|
ext2_filsys ext2_fs, ext2_ino_t ext2_ino)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
int inline_ea = 0;
|
|
|
|
errcode_t err;
|
|
|
|
u32 datalen;
|
|
|
|
u32 block_size = ext2_fs->blocksize;
|
|
|
|
u32 inode_size = EXT2_INODE_SIZE(ext2_fs->super);
|
|
|
|
struct ext2_inode_large *ext2_inode;
|
|
|
|
struct ext2_ext_attr_entry *entry;
|
|
|
|
void *data;
|
|
|
|
char *buffer = NULL;
|
|
|
|
char inode_buf[EXT2_GOOD_OLD_INODE_SIZE];
|
|
|
|
|
|
|
|
if (inode_size <= EXT2_GOOD_OLD_INODE_SIZE) {
|
|
|
|
ext2_inode = (struct ext2_inode_large *)inode_buf;
|
|
|
|
} else {
|
|
|
|
ext2_inode = (struct ext2_inode_large *)malloc(inode_size);
|
|
|
|
if (!ext2_inode)
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
err = ext2fs_read_inode_full(ext2_fs, ext2_ino, (void *)ext2_inode,
|
|
|
|
inode_size);
|
|
|
|
if (err) {
|
2022-09-19 10:17:19 +00:00
|
|
|
error("ext2fs_read_inode_full: %s", error_message(err));
|
2017-01-27 15:56:37 +00:00
|
|
|
ret = -1;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (ext2_ino > ext2_fs->super->s_first_ino &&
|
|
|
|
inode_size > EXT2_GOOD_OLD_INODE_SIZE) {
|
|
|
|
if (EXT2_GOOD_OLD_INODE_SIZE +
|
|
|
|
ext2_inode->i_extra_isize > inode_size) {
|
|
|
|
ret = -EIO;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
if (ext2_inode->i_extra_isize != 0 &&
|
|
|
|
EXT2_XATTR_IHDR(ext2_inode)->h_magic ==
|
|
|
|
EXT2_EXT_ATTR_MAGIC) {
|
|
|
|
inline_ea = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (inline_ea) {
|
|
|
|
int total;
|
|
|
|
void *end = (void *)ext2_inode + inode_size;
|
|
|
|
entry = EXT2_XATTR_IFIRST(ext2_inode);
|
|
|
|
total = end - (void *)entry;
|
|
|
|
ret = ext2_xattr_check_names(entry, end);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
while (!EXT2_EXT_IS_LAST_ENTRY(entry)) {
|
|
|
|
ret = ext2_xattr_check_entry(entry, total);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
data = (void *)EXT2_XATTR_IFIRST(ext2_inode) +
|
|
|
|
entry->e_value_offs;
|
|
|
|
datalen = entry->e_value_size;
|
|
|
|
ret = ext2_copy_single_xattr(trans, root, objectid,
|
|
|
|
entry, data, datalen);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
entry = EXT2_EXT_ATTR_NEXT(entry);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (ext2_inode->i_file_acl == 0)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
buffer = malloc(block_size);
|
|
|
|
if (!buffer) {
|
|
|
|
ret = -ENOMEM;
|
|
|
|
goto out;
|
|
|
|
}
|
2017-06-03 09:57:45 +00:00
|
|
|
err = ext2fs_read_ext_attr2(ext2_fs, ext2_inode->i_file_acl, buffer);
|
2017-01-27 15:56:37 +00:00
|
|
|
if (err) {
|
2022-09-19 10:17:19 +00:00
|
|
|
error("ext2fs_read_ext_attr2: %s", error_message(err));
|
2017-01-27 15:56:37 +00:00
|
|
|
ret = -1;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
ret = ext2_xattr_check_block(buffer, block_size);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
entry = EXT2_XATTR_BFIRST(buffer);
|
|
|
|
while (!EXT2_EXT_IS_LAST_ENTRY(entry)) {
|
|
|
|
ret = ext2_xattr_check_entry(entry, block_size);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
data = buffer + entry->e_value_offs;
|
|
|
|
datalen = entry->e_value_size;
|
|
|
|
ret = ext2_copy_single_xattr(trans, root, objectid,
|
|
|
|
entry, data, datalen);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
entry = EXT2_EXT_ATTR_NEXT(entry);
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
free(buffer);
|
|
|
|
if ((void *)ext2_inode != inode_buf)
|
|
|
|
free(ext2_inode);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline dev_t old_decode_dev(u16 val)
|
|
|
|
{
|
|
|
|
return MKDEV((val >> 8) & 255, val & 255);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void ext2_copy_inode_item(struct btrfs_inode_item *dst,
|
|
|
|
struct ext2_inode *src, u32 blocksize)
|
|
|
|
{
|
|
|
|
btrfs_set_stack_inode_generation(dst, 1);
|
|
|
|
btrfs_set_stack_inode_sequence(dst, 0);
|
|
|
|
btrfs_set_stack_inode_transid(dst, 1);
|
|
|
|
btrfs_set_stack_inode_size(dst, src->i_size);
|
|
|
|
btrfs_set_stack_inode_nbytes(dst, 0);
|
|
|
|
btrfs_set_stack_inode_block_group(dst, 0);
|
|
|
|
btrfs_set_stack_inode_nlink(dst, src->i_links_count);
|
|
|
|
btrfs_set_stack_inode_uid(dst, src->i_uid | (src->i_uid_high << 16));
|
|
|
|
btrfs_set_stack_inode_gid(dst, src->i_gid | (src->i_gid_high << 16));
|
|
|
|
btrfs_set_stack_inode_mode(dst, src->i_mode);
|
|
|
|
btrfs_set_stack_inode_rdev(dst, 0);
|
|
|
|
btrfs_set_stack_inode_flags(dst, 0);
|
|
|
|
btrfs_set_stack_timespec_sec(&dst->atime, src->i_atime);
|
|
|
|
btrfs_set_stack_timespec_nsec(&dst->atime, 0);
|
|
|
|
btrfs_set_stack_timespec_sec(&dst->ctime, src->i_ctime);
|
|
|
|
btrfs_set_stack_timespec_nsec(&dst->ctime, 0);
|
|
|
|
btrfs_set_stack_timespec_sec(&dst->mtime, src->i_mtime);
|
|
|
|
btrfs_set_stack_timespec_nsec(&dst->mtime, 0);
|
|
|
|
btrfs_set_stack_timespec_sec(&dst->otime, 0);
|
|
|
|
btrfs_set_stack_timespec_nsec(&dst->otime, 0);
|
|
|
|
|
|
|
|
if (S_ISDIR(src->i_mode)) {
|
|
|
|
btrfs_set_stack_inode_size(dst, 0);
|
|
|
|
btrfs_set_stack_inode_nlink(dst, 1);
|
|
|
|
}
|
|
|
|
if (S_ISREG(src->i_mode)) {
|
|
|
|
btrfs_set_stack_inode_size(dst, (u64)src->i_size_high << 32 |
|
|
|
|
(u64)src->i_size);
|
|
|
|
}
|
|
|
|
if (!S_ISREG(src->i_mode) && !S_ISDIR(src->i_mode) &&
|
|
|
|
!S_ISLNK(src->i_mode)) {
|
|
|
|
if (src->i_block[0]) {
|
|
|
|
btrfs_set_stack_inode_rdev(dst,
|
|
|
|
old_decode_dev(src->i_block[0]));
|
|
|
|
} else {
|
|
|
|
btrfs_set_stack_inode_rdev(dst,
|
2017-08-22 18:43:49 +00:00
|
|
|
decode_dev(src->i_block[1]));
|
2017-01-27 15:56:37 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
memset(&dst->reserved, 0, sizeof(dst->reserved));
|
|
|
|
}
|
2020-12-25 07:45:38 +00:00
|
|
|
|
2021-03-14 18:49:13 +00:00
|
|
|
#if HAVE_EXT4_EPOCH_MASK_DEFINE
|
2021-02-24 15:31:11 +00:00
|
|
|
|
2020-12-25 07:45:38 +00:00
|
|
|
/*
|
|
|
|
* Copied and modified from fs/ext4/ext4.h
|
|
|
|
*/
|
|
|
|
static inline void ext4_decode_extra_time(__le32 * tv_sec, __le32 * tv_nsec,
|
|
|
|
__le32 extra)
|
|
|
|
{
|
|
|
|
if (extra & cpu_to_le32(EXT4_EPOCH_MASK))
|
|
|
|
*tv_sec += (u64)(le32_to_cpu(extra) & EXT4_EPOCH_MASK) << 32;
|
|
|
|
*tv_nsec = (le32_to_cpu(extra) & EXT4_NSEC_MASK) >> EXT4_EPOCH_BITS;
|
|
|
|
}
|
|
|
|
|
|
|
|
#define EXT4_COPY_XTIME(xtime, dst, tv_sec, tv_nsec) \
|
|
|
|
do { \
|
|
|
|
tv_sec = src->i_ ## xtime ; \
|
|
|
|
if (inode_includes(inode_size, i_ ## xtime ## _extra)) { \
|
|
|
|
tv_sec = src->i_ ## xtime ; \
|
|
|
|
ext4_decode_extra_time(&tv_sec, &tv_nsec, src->i_ ## xtime ## _extra); \
|
|
|
|
btrfs_set_stack_timespec_sec(&dst->xtime , tv_sec); \
|
|
|
|
btrfs_set_stack_timespec_nsec(&dst->xtime , tv_nsec); \
|
|
|
|
} else { \
|
|
|
|
btrfs_set_stack_timespec_sec(&dst->xtime , tv_sec); \
|
|
|
|
btrfs_set_stack_timespec_nsec(&dst->xtime , 0); \
|
|
|
|
} \
|
|
|
|
} while (0);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Decode and copy i_[cma]time_extra and i_crtime{,_extra} field
|
|
|
|
*/
|
|
|
|
static int ext4_copy_inode_timespec_extra(struct btrfs_inode_item *dst,
|
|
|
|
ext2_ino_t ext2_ino, u32 s_inode_size,
|
|
|
|
ext2_filsys ext2_fs)
|
|
|
|
{
|
|
|
|
struct ext2_inode_large *src;
|
|
|
|
u32 inode_size, tv_sec, tv_nsec;
|
|
|
|
int ret, err;
|
|
|
|
ret = 0;
|
|
|
|
|
|
|
|
src = (struct ext2_inode_large *)malloc(s_inode_size);
|
|
|
|
if (!src)
|
|
|
|
return -ENOMEM;
|
|
|
|
err = ext2fs_read_inode_full(ext2_fs, ext2_ino, (void *)src,
|
|
|
|
s_inode_size);
|
|
|
|
if (err) {
|
2022-09-19 10:17:19 +00:00
|
|
|
error("ext2fs_read_inode_full: %s", error_message(err));
|
2020-12-25 07:45:38 +00:00
|
|
|
ret = -1;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
inode_size = EXT2_GOOD_OLD_INODE_SIZE + src->i_extra_isize;
|
|
|
|
|
|
|
|
EXT4_COPY_XTIME(atime, dst, tv_sec, tv_nsec);
|
|
|
|
EXT4_COPY_XTIME(mtime, dst, tv_sec, tv_nsec);
|
|
|
|
EXT4_COPY_XTIME(ctime, dst, tv_sec, tv_nsec);
|
|
|
|
|
|
|
|
tv_sec = src->i_crtime;
|
|
|
|
if (inode_includes(inode_size, i_crtime_extra)) {
|
|
|
|
tv_sec = src->i_crtime;
|
|
|
|
ext4_decode_extra_time(&tv_sec, &tv_nsec, src->i_crtime_extra);
|
|
|
|
btrfs_set_stack_timespec_sec(&dst->otime, tv_sec);
|
|
|
|
btrfs_set_stack_timespec_nsec(&dst->otime, tv_nsec);
|
|
|
|
} else {
|
|
|
|
btrfs_set_stack_timespec_sec(&dst->otime, tv_sec);
|
|
|
|
btrfs_set_stack_timespec_nsec(&dst->otime, 0);
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
free(src);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2021-03-14 18:49:13 +00:00
|
|
|
#else /* HAVE_EXT4_EPOCH_MASK_DEFINE */
|
2021-02-24 15:31:11 +00:00
|
|
|
|
|
|
|
static int ext4_copy_inode_timespec_extra(struct btrfs_inode_item *dst,
|
|
|
|
ext2_ino_t ext2_ino, u32 s_inode_size,
|
|
|
|
ext2_filsys ext2_fs)
|
|
|
|
{
|
|
|
|
static int warn = 0;
|
|
|
|
|
|
|
|
if (!warn) {
|
|
|
|
warning(
|
|
|
|
"extended inode (size %u) found but e2fsprogs don't support reading extra timespec",
|
|
|
|
s_inode_size);
|
|
|
|
warn = 1;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2021-03-14 18:49:13 +00:00
|
|
|
#endif /* !HAVE_EXT4_EPOCH_MASK_DEFINE */
|
2021-02-24 15:31:11 +00:00
|
|
|
|
2017-01-27 15:56:37 +00:00
|
|
|
static int ext2_check_state(struct btrfs_convert_context *cctx)
|
|
|
|
{
|
|
|
|
ext2_filsys fs = cctx->fs_data;
|
|
|
|
|
|
|
|
if (!(fs->super->s_state & EXT2_VALID_FS))
|
|
|
|
return 1;
|
|
|
|
else if (fs->super->s_state & EXT2_ERROR_FS)
|
|
|
|
return 1;
|
|
|
|
else
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* EXT2_*_FL to BTRFS_INODE_FLAG_* stringification helper */
|
|
|
|
#define COPY_ONE_EXT2_FLAG(flags, ext2_inode, name) ({ \
|
|
|
|
if (ext2_inode->i_flags & EXT2_##name##_FL) \
|
|
|
|
flags |= BTRFS_INODE_##name; \
|
|
|
|
})
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Convert EXT2_*_FL to corresponding BTRFS_INODE_* flags
|
|
|
|
*
|
|
|
|
* Only a subset of EXT_*_FL is supported in btrfs.
|
|
|
|
*/
|
|
|
|
static void ext2_convert_inode_flags(struct btrfs_inode_item *dst,
|
|
|
|
struct ext2_inode *src)
|
|
|
|
{
|
2017-06-28 21:29:49 +00:00
|
|
|
u64 flags = btrfs_stack_inode_flags(dst);
|
2017-01-27 15:56:37 +00:00
|
|
|
|
|
|
|
COPY_ONE_EXT2_FLAG(flags, src, APPEND);
|
|
|
|
COPY_ONE_EXT2_FLAG(flags, src, SYNC);
|
|
|
|
COPY_ONE_EXT2_FLAG(flags, src, IMMUTABLE);
|
|
|
|
COPY_ONE_EXT2_FLAG(flags, src, NODUMP);
|
|
|
|
COPY_ONE_EXT2_FLAG(flags, src, NOATIME);
|
|
|
|
COPY_ONE_EXT2_FLAG(flags, src, DIRSYNC);
|
|
|
|
btrfs_set_stack_inode_flags(dst, flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* copy a single inode. do all the required works, such as cloning
|
|
|
|
* inode item, creating file extents and creating directory entries.
|
|
|
|
*/
|
|
|
|
static int ext2_copy_single_inode(struct btrfs_trans_handle *trans,
|
|
|
|
struct btrfs_root *root, u64 objectid,
|
|
|
|
ext2_filsys ext2_fs, ext2_ino_t ext2_ino,
|
|
|
|
struct ext2_inode *ext2_inode,
|
2017-01-30 16:52:15 +00:00
|
|
|
u32 convert_flags)
|
2017-01-27 15:56:37 +00:00
|
|
|
{
|
|
|
|
int ret;
|
2020-12-25 07:45:38 +00:00
|
|
|
int s_inode_size;
|
2017-01-27 15:56:37 +00:00
|
|
|
struct btrfs_inode_item btrfs_inode;
|
|
|
|
|
|
|
|
if (ext2_inode->i_links_count == 0)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
ext2_copy_inode_item(&btrfs_inode, ext2_inode, ext2_fs->blocksize);
|
2020-12-25 07:45:38 +00:00
|
|
|
s_inode_size = EXT2_INODE_SIZE(ext2_fs->super);
|
|
|
|
if (s_inode_size > EXT2_GOOD_OLD_INODE_SIZE) {
|
|
|
|
ret = ext4_copy_inode_timespec_extra(&btrfs_inode, ext2_ino,
|
|
|
|
s_inode_size, ext2_fs);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2017-01-30 16:52:15 +00:00
|
|
|
if (!(convert_flags & CONVERT_FLAG_DATACSUM)
|
|
|
|
&& S_ISREG(ext2_inode->i_mode)) {
|
2017-01-27 15:56:37 +00:00
|
|
|
u32 flags = btrfs_stack_inode_flags(&btrfs_inode) |
|
|
|
|
BTRFS_INODE_NODATASUM;
|
|
|
|
btrfs_set_stack_inode_flags(&btrfs_inode, flags);
|
|
|
|
}
|
|
|
|
ext2_convert_inode_flags(&btrfs_inode, ext2_inode);
|
|
|
|
|
|
|
|
switch (ext2_inode->i_mode & S_IFMT) {
|
|
|
|
case S_IFREG:
|
|
|
|
ret = ext2_create_file_extents(trans, root, objectid,
|
2017-01-30 16:52:15 +00:00
|
|
|
&btrfs_inode, ext2_fs, ext2_ino, convert_flags);
|
2017-01-27 15:56:37 +00:00
|
|
|
break;
|
|
|
|
case S_IFDIR:
|
|
|
|
ret = ext2_create_dir_entries(trans, root, objectid,
|
|
|
|
&btrfs_inode, ext2_fs, ext2_ino);
|
|
|
|
break;
|
|
|
|
case S_IFLNK:
|
2017-01-31 10:08:09 +00:00
|
|
|
ret = ext2_create_symlink(trans, root, objectid,
|
2017-01-27 15:56:37 +00:00
|
|
|
&btrfs_inode, ext2_fs, ext2_ino, ext2_inode);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
ret = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2017-01-30 16:52:15 +00:00
|
|
|
if (convert_flags & CONVERT_FLAG_XATTR) {
|
2017-01-27 15:56:37 +00:00
|
|
|
ret = ext2_copy_extended_attrs(trans, root, objectid,
|
|
|
|
&btrfs_inode, ext2_fs, ext2_ino);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
return btrfs_insert_inode(trans, root, objectid, &btrfs_inode);
|
|
|
|
}
|
|
|
|
|
btrfs-progs: convert: handle ext4 orphan file feature properly
[BUG]
Since e2fsprog 1.47, even with a newly created empty ext4 filesystem,
btrfs-convert would result an fs that btrfs-check would complain:
# mkfs.ext4 -F test.img
# btrfs-convert test.img
# btrfs-check test.img
Opening filesystem to check...
Checking filesystem on test.img
UUID: e45da158-8967-4e4d-9c9f-66b0d127dbce
[1/7] checking root items
[2/7] checking extents
[3/7] checking free space cache
[4/7] checking fs roots
root 5 inode 266 errors 2000, link count wrong
ERROR: errors found in fs roots
found 26333184 bytes used, error(s) found <<<
total csum bytes: 25540
total tree bytes: 180224
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 145423
file data blocks allocated: 33947648
referenced 26284032
[CAUSE]
Ext4 has a new compat feature, COMPAT_ORPHAN_FILE, as a better way to
track all the orphan inodes.
This new feature would create a new special inode for this purpose, and
such orphan file inode would not be reachable from any other inode, but
only from super block.
Unfortunately btrfs-convert only skip ext2 known special inodes, not the
newer one.
[FIX]
According to the kernel document, we can locate the orphan file inode
using ext2 super block s_orphan_file_inum, and skip it for
btrfs-convert.
And such skip would only happen if we have the definition of
EXT4_FEATURE_COMPAT_ORPHAN_FILE, to be compatible with older e2fsprogs.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
2023-03-23 02:33:59 +00:00
|
|
|
static bool ext2_is_special_inode(ext2_filsys ext2_fs, ext2_ino_t ino)
|
2017-01-30 16:14:03 +00:00
|
|
|
{
|
|
|
|
if (ino < EXT2_GOOD_OLD_FIRST_INO && ino != EXT2_ROOT_INO)
|
|
|
|
return 1;
|
btrfs-progs: convert: handle ext4 orphan file feature properly
[BUG]
Since e2fsprog 1.47, even with a newly created empty ext4 filesystem,
btrfs-convert would result an fs that btrfs-check would complain:
# mkfs.ext4 -F test.img
# btrfs-convert test.img
# btrfs-check test.img
Opening filesystem to check...
Checking filesystem on test.img
UUID: e45da158-8967-4e4d-9c9f-66b0d127dbce
[1/7] checking root items
[2/7] checking extents
[3/7] checking free space cache
[4/7] checking fs roots
root 5 inode 266 errors 2000, link count wrong
ERROR: errors found in fs roots
found 26333184 bytes used, error(s) found <<<
total csum bytes: 25540
total tree bytes: 180224
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 145423
file data blocks allocated: 33947648
referenced 26284032
[CAUSE]
Ext4 has a new compat feature, COMPAT_ORPHAN_FILE, as a better way to
track all the orphan inodes.
This new feature would create a new special inode for this purpose, and
such orphan file inode would not be reachable from any other inode, but
only from super block.
Unfortunately btrfs-convert only skip ext2 known special inodes, not the
newer one.
[FIX]
According to the kernel document, we can locate the orphan file inode
using ext2 super block s_orphan_file_inum, and skip it for
btrfs-convert.
And such skip would only happen if we have the definition of
EXT4_FEATURE_COMPAT_ORPHAN_FILE, to be compatible with older e2fsprogs.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
2023-03-23 02:33:59 +00:00
|
|
|
#ifdef EXT4_FEATURE_COMPAT_ORPHAN_FILE
|
|
|
|
/*
|
|
|
|
* If we have COMPAT_ORPHAN_FILE feature, we have a special inode
|
|
|
|
* recording all the orphan files. We need to skip such special inode.
|
|
|
|
*/
|
|
|
|
if (ext2_fs->super->s_feature_compat & EXT4_FEATURE_COMPAT_ORPHAN_FILE &&
|
|
|
|
ino == ext2_fs->super->s_orphan_file_inum)
|
|
|
|
return 1;
|
|
|
|
#endif
|
2017-01-30 16:14:03 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2017-01-27 15:56:37 +00:00
|
|
|
/*
|
|
|
|
* scan ext2's inode bitmap and copy all used inodes.
|
|
|
|
*/
|
|
|
|
static int ext2_copy_inodes(struct btrfs_convert_context *cctx,
|
|
|
|
struct btrfs_root *root,
|
2017-01-30 16:52:15 +00:00
|
|
|
u32 convert_flags, struct task_ctx *p)
|
2017-01-27 15:56:37 +00:00
|
|
|
{
|
|
|
|
ext2_filsys ext2_fs = cctx->fs_data;
|
2020-07-29 08:40:36 +00:00
|
|
|
int ret = 0;
|
2017-01-27 15:56:37 +00:00
|
|
|
errcode_t err;
|
|
|
|
ext2_inode_scan ext2_scan;
|
|
|
|
struct ext2_inode ext2_inode;
|
|
|
|
ext2_ino_t ext2_ino;
|
|
|
|
u64 objectid;
|
|
|
|
struct btrfs_trans_handle *trans;
|
|
|
|
|
|
|
|
trans = btrfs_start_transaction(root, 1);
|
2017-08-28 14:48:16 +00:00
|
|
|
if (IS_ERR(trans))
|
|
|
|
return PTR_ERR(trans);
|
2017-01-27 15:56:37 +00:00
|
|
|
err = ext2fs_open_inode_scan(ext2_fs, 0, &ext2_scan);
|
|
|
|
if (err) {
|
2020-07-29 08:40:36 +00:00
|
|
|
error("ext2fs_open_inode_scan failed: %s", error_message(err));
|
|
|
|
btrfs_commit_transaction(trans, root);
|
|
|
|
return -EIO;
|
2017-01-27 15:56:37 +00:00
|
|
|
}
|
|
|
|
while (!(err = ext2fs_get_next_inode(ext2_scan, &ext2_ino,
|
|
|
|
&ext2_inode))) {
|
|
|
|
/* no more inodes */
|
|
|
|
if (ext2_ino == 0)
|
|
|
|
break;
|
btrfs-progs: convert: handle ext4 orphan file feature properly
[BUG]
Since e2fsprog 1.47, even with a newly created empty ext4 filesystem,
btrfs-convert would result an fs that btrfs-check would complain:
# mkfs.ext4 -F test.img
# btrfs-convert test.img
# btrfs-check test.img
Opening filesystem to check...
Checking filesystem on test.img
UUID: e45da158-8967-4e4d-9c9f-66b0d127dbce
[1/7] checking root items
[2/7] checking extents
[3/7] checking free space cache
[4/7] checking fs roots
root 5 inode 266 errors 2000, link count wrong
ERROR: errors found in fs roots
found 26333184 bytes used, error(s) found <<<
total csum bytes: 25540
total tree bytes: 180224
total fs tree bytes: 49152
total extent tree bytes: 16384
btree space waste bytes: 145423
file data blocks allocated: 33947648
referenced 26284032
[CAUSE]
Ext4 has a new compat feature, COMPAT_ORPHAN_FILE, as a better way to
track all the orphan inodes.
This new feature would create a new special inode for this purpose, and
such orphan file inode would not be reachable from any other inode, but
only from super block.
Unfortunately btrfs-convert only skip ext2 known special inodes, not the
newer one.
[FIX]
According to the kernel document, we can locate the orphan file inode
using ext2 super block s_orphan_file_inum, and skip it for
btrfs-convert.
And such skip would only happen if we have the definition of
EXT4_FEATURE_COMPAT_ORPHAN_FILE, to be compatible with older e2fsprogs.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
2023-03-23 02:33:59 +00:00
|
|
|
if (ext2_is_special_inode(ext2_fs, ext2_ino))
|
2017-01-27 15:56:37 +00:00
|
|
|
continue;
|
|
|
|
objectid = ext2_ino + INO_OFFSET;
|
|
|
|
ret = ext2_copy_single_inode(trans, root,
|
|
|
|
objectid, ext2_fs, ext2_ino,
|
2017-01-30 16:52:15 +00:00
|
|
|
&ext2_inode, convert_flags);
|
2017-07-12 20:05:38 +00:00
|
|
|
pthread_mutex_lock(&p->mutex);
|
2017-01-27 15:56:37 +00:00
|
|
|
p->cur_copy_inodes++;
|
2017-07-12 20:05:38 +00:00
|
|
|
pthread_mutex_unlock(&p->mutex);
|
2020-07-29 08:40:36 +00:00
|
|
|
if (ret) {
|
|
|
|
error("failed to copy ext2 inode %llu: %d",
|
|
|
|
(unsigned long long)ext2_ino, ret);
|
|
|
|
goto out;
|
|
|
|
}
|
btrfs-progs: convert: Workaround delayed ref bug by limiting the size of a transaction
In convert we use trans->block_reserved >= 4096 as a threshold to commit
transaction, where block_reserved is the number of new tree blocks
allocated inside a transaction.
The problem is, we still have a hidden bug in delayed ref implementation
in btrfs-progs, when we have a large enough transaction, delayed ref may
failed to find certain tree blocks in extent tree and cause transaction
abort.
This fix will workaround it by committing transaction at a much lower
threshold.
The old 4096 means 4096 new tree blocks, when using default (16K)
nodesize, it's 64M, which can contain over 12k inlined data extent or
csum for around 60G, or over 800K file extents.
The new threshold will limit the size of new tree blocks to 2M, aligning
with the chunk preallocator threshold, and reducing the possibility to
hit that delayed ref bug.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
2019-05-27 05:10:54 +00:00
|
|
|
/*
|
|
|
|
* blocks_used is the number of new tree blocks allocated in
|
|
|
|
* current transaction.
|
|
|
|
* Use a small amount of it to workaround a bug where delayed
|
|
|
|
* ref may fail to locate tree blocks in extent tree.
|
|
|
|
*
|
|
|
|
* 2M is the threshold to kick chunk preallocator into work,
|
|
|
|
* For default (16K) nodesize it will be 128 tree blocks,
|
|
|
|
* large enough to contain over 300 inlined files or
|
|
|
|
* around 26k file extents. Which should be good enough.
|
|
|
|
*/
|
|
|
|
if (trans->blocks_used >= SZ_2M / root->fs_info->nodesize) {
|
2017-01-27 15:56:37 +00:00
|
|
|
ret = btrfs_commit_transaction(trans, root);
|
2020-07-29 08:40:36 +00:00
|
|
|
if (ret < 0) {
|
2022-09-30 07:12:06 +00:00
|
|
|
errno = -ret;
|
|
|
|
error_msg(ERROR_MSG_COMMIT_TRANS, "%m");
|
2020-07-29 08:40:36 +00:00
|
|
|
goto out;
|
|
|
|
}
|
2017-01-27 15:56:37 +00:00
|
|
|
trans = btrfs_start_transaction(root, 1);
|
2020-07-29 08:40:36 +00:00
|
|
|
if (IS_ERR(trans)) {
|
|
|
|
ret = PTR_ERR(trans);
|
2022-09-30 07:12:06 +00:00
|
|
|
errno = -ret;
|
|
|
|
error_msg(ERROR_MSG_START_TRANS, "%m");
|
2020-07-29 08:40:36 +00:00
|
|
|
trans = NULL;
|
|
|
|
goto out;
|
|
|
|
}
|
2017-01-27 15:56:37 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
if (err) {
|
2020-07-29 08:40:36 +00:00
|
|
|
error("ext2fs_get_next_inode failed: %s", error_message(err));
|
|
|
|
ret = -EIO;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
if (ret < 0) {
|
|
|
|
if (trans)
|
|
|
|
btrfs_abort_transaction(trans, ret);
|
|
|
|
} else {
|
|
|
|
ret = btrfs_commit_transaction(trans, root);
|
2022-09-30 07:12:06 +00:00
|
|
|
if (ret < 0) {
|
|
|
|
errno = -ret;
|
|
|
|
error_msg(ERROR_MSG_COMMIT_TRANS, "%m");
|
|
|
|
}
|
2017-01-27 15:56:37 +00:00
|
|
|
}
|
|
|
|
ext2fs_close_inode_scan(ext2_scan);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
const struct btrfs_convert_operations ext2_convert_ops = {
|
|
|
|
.name = "ext2",
|
|
|
|
.open_fs = ext2_open_fs,
|
|
|
|
.read_used_space = ext2_read_used_space,
|
|
|
|
.copy_inodes = ext2_copy_inodes,
|
|
|
|
.close_fs = ext2_close_fs,
|
|
|
|
.check_state = ext2_check_state,
|
|
|
|
};
|
|
|
|
|
|
|
|
#endif /* BTRFSCONVERT_EXT2 */
|