X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=fs%2Fbtrfs%2Fioctl.c;h=bd88f25889f7c5daf94bf0aec8645041bfb2ce60;hb=405f55712dfe464b3240d7816cc4fe4174831be2;hp=3d85f18bbba6f2890de0eac73944c8f278bac056;hpb=cb8e70901d36f32017614f16d2cf7cc089544574;p=safe%2Fjmp%2Flinux-2.6 diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 3d85f18..bd88f25 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -27,7 +27,6 @@ #include #include #include -#include #include #include #include @@ -38,9 +37,9 @@ #include #include #include -#include #include #include +#include "compat.h" #include "ctree.h" #include "disk-io.h" #include "transaction.h" @@ -50,7 +49,177 @@ #include "volumes.h" #include "locking.h" +/* Mask out flags that are inappropriate for the given type of inode. */ +static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags) +{ + if (S_ISDIR(mode)) + return flags; + else if (S_ISREG(mode)) + return flags & ~FS_DIRSYNC_FL; + else + return flags & (FS_NODUMP_FL | FS_NOATIME_FL); +} + +/* + * Export inode flags to the format expected by the FS_IOC_GETFLAGS ioctl. + */ +static unsigned int btrfs_flags_to_ioctl(unsigned int flags) +{ + unsigned int iflags = 0; + + if (flags & BTRFS_INODE_SYNC) + iflags |= FS_SYNC_FL; + if (flags & BTRFS_INODE_IMMUTABLE) + iflags |= FS_IMMUTABLE_FL; + if (flags & BTRFS_INODE_APPEND) + iflags |= FS_APPEND_FL; + if (flags & BTRFS_INODE_NODUMP) + iflags |= FS_NODUMP_FL; + if (flags & BTRFS_INODE_NOATIME) + iflags |= FS_NOATIME_FL; + if (flags & BTRFS_INODE_DIRSYNC) + iflags |= FS_DIRSYNC_FL; + + return iflags; +} + +/* + * Update inode->i_flags based on the btrfs internal flags. + */ +void btrfs_update_iflags(struct inode *inode) +{ + struct btrfs_inode *ip = BTRFS_I(inode); + + inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC); + + if (ip->flags & BTRFS_INODE_SYNC) + inode->i_flags |= S_SYNC; + if (ip->flags & BTRFS_INODE_IMMUTABLE) + inode->i_flags |= S_IMMUTABLE; + if (ip->flags & BTRFS_INODE_APPEND) + inode->i_flags |= S_APPEND; + if (ip->flags & BTRFS_INODE_NOATIME) + inode->i_flags |= S_NOATIME; + if (ip->flags & BTRFS_INODE_DIRSYNC) + inode->i_flags |= S_DIRSYNC; +} + +/* + * Inherit flags from the parent inode. + * + * Unlike extN we don't have any flags we don't want to inherit currently. + */ +void btrfs_inherit_iflags(struct inode *inode, struct inode *dir) +{ + unsigned int flags; + + if (!dir) + return; + + flags = BTRFS_I(dir)->flags; + + if (S_ISREG(inode->i_mode)) + flags &= ~BTRFS_INODE_DIRSYNC; + else if (!S_ISDIR(inode->i_mode)) + flags &= (BTRFS_INODE_NODUMP | BTRFS_INODE_NOATIME); + + BTRFS_I(inode)->flags = flags; + btrfs_update_iflags(inode); +} + +static int btrfs_ioctl_getflags(struct file *file, void __user *arg) +{ + struct btrfs_inode *ip = BTRFS_I(file->f_path.dentry->d_inode); + unsigned int flags = btrfs_flags_to_ioctl(ip->flags); + + if (copy_to_user(arg, &flags, sizeof(flags))) + return -EFAULT; + return 0; +} + +static int btrfs_ioctl_setflags(struct file *file, void __user *arg) +{ + struct inode *inode = file->f_path.dentry->d_inode; + struct btrfs_inode *ip = BTRFS_I(inode); + struct btrfs_root *root = ip->root; + struct btrfs_trans_handle *trans; + unsigned int flags, oldflags; + int ret; + + if (copy_from_user(&flags, arg, sizeof(flags))) + return -EFAULT; + + if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \ + FS_NOATIME_FL | FS_NODUMP_FL | \ + FS_SYNC_FL | FS_DIRSYNC_FL)) + return -EOPNOTSUPP; + + if (!is_owner_or_cap(inode)) + return -EACCES; + + mutex_lock(&inode->i_mutex); + + flags = btrfs_mask_flags(inode->i_mode, flags); + oldflags = btrfs_flags_to_ioctl(ip->flags); + if ((flags ^ oldflags) & (FS_APPEND_FL | FS_IMMUTABLE_FL)) { + if (!capable(CAP_LINUX_IMMUTABLE)) { + ret = -EPERM; + goto out_unlock; + } + } + + ret = mnt_want_write(file->f_path.mnt); + if (ret) + goto out_unlock; + + if (flags & FS_SYNC_FL) + ip->flags |= BTRFS_INODE_SYNC; + else + ip->flags &= ~BTRFS_INODE_SYNC; + if (flags & FS_IMMUTABLE_FL) + ip->flags |= BTRFS_INODE_IMMUTABLE; + else + ip->flags &= ~BTRFS_INODE_IMMUTABLE; + if (flags & FS_APPEND_FL) + ip->flags |= BTRFS_INODE_APPEND; + else + ip->flags &= ~BTRFS_INODE_APPEND; + if (flags & FS_NODUMP_FL) + ip->flags |= BTRFS_INODE_NODUMP; + else + ip->flags &= ~BTRFS_INODE_NODUMP; + if (flags & FS_NOATIME_FL) + ip->flags |= BTRFS_INODE_NOATIME; + else + ip->flags &= ~BTRFS_INODE_NOATIME; + if (flags & FS_DIRSYNC_FL) + ip->flags |= BTRFS_INODE_DIRSYNC; + else + ip->flags &= ~BTRFS_INODE_DIRSYNC; + + trans = btrfs_join_transaction(root, 1); + BUG_ON(!trans); + + ret = btrfs_update_inode(trans, root, inode); + BUG_ON(ret); + + btrfs_update_iflags(inode); + inode->i_ctime = CURRENT_TIME; + btrfs_end_transaction(trans, root); + + mnt_drop_write(file->f_path.mnt); + out_unlock: + mutex_unlock(&inode->i_mutex); + return 0; +} + +static int btrfs_ioctl_getversion(struct file *file, int __user *arg) +{ + struct inode *inode = file->f_path.dentry->d_inode; + + return put_user(inode->i_generation, arg); +} static noinline int create_subvol(struct btrfs_root *root, struct dentry *dentry, @@ -67,9 +236,10 @@ static noinline int create_subvol(struct btrfs_root *root, int err; u64 objectid; u64 new_dirid = BTRFS_FIRST_FREE_OBJECTID; + u64 index = 0; unsigned long nr = 1; - ret = btrfs_check_free_space(root, 1, 0); + ret = btrfs_check_metadata_free_space(root); if (ret) goto fail_commit; @@ -81,22 +251,25 @@ static noinline int create_subvol(struct btrfs_root *root, if (ret) goto fail; - leaf = btrfs_alloc_free_block(trans, root, root->leafsize, 0, - objectid, trans->transid, 0, 0, 0); + leaf = btrfs_alloc_free_block(trans, root, root->leafsize, + 0, objectid, NULL, 0, 0, 0); if (IS_ERR(leaf)) { ret = PTR_ERR(leaf); goto fail; } - btrfs_set_header_nritems(leaf, 0); - btrfs_set_header_level(leaf, 0); + memset_extent_buffer(leaf, 0, 0, sizeof(struct btrfs_header)); btrfs_set_header_bytenr(leaf, leaf->start); btrfs_set_header_generation(leaf, trans->transid); + btrfs_set_header_backref_rev(leaf, BTRFS_MIXED_BACKREF_REV); btrfs_set_header_owner(leaf, objectid); write_extent_buffer(leaf, root->fs_info->fsid, (unsigned long)btrfs_header_fsid(leaf), BTRFS_FSID_SIZE); + write_extent_buffer(leaf, root->fs_info->chunk_tree_uuid, + (unsigned long)btrfs_header_chunk_tree_uuid(leaf), + BTRFS_UUID_SIZE); btrfs_mark_buffer_dirty(leaf); inode_item = &root_item.inode; @@ -108,9 +281,11 @@ static noinline int create_subvol(struct btrfs_root *root, inode_item->mode = cpu_to_le32(S_IFDIR | 0755); btrfs_set_root_bytenr(&root_item, leaf->start); + btrfs_set_root_generation(&root_item, trans->transid); btrfs_set_root_level(&root_item, 0); btrfs_set_root_refs(&root_item, 1); btrfs_set_root_used(&root_item, 0); + btrfs_set_root_last_snapshot(&root_item, 0); memset(&root_item.drop_progress, 0, sizeof(root_item.drop_progress)); root_item.drop_level = 0; @@ -122,7 +297,7 @@ static noinline int create_subvol(struct btrfs_root *root, btrfs_set_root_dirid(&root_item, new_dirid); key.objectid = objectid; - key.offset = 1; + key.offset = 0; btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY); ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key, &root_item); @@ -133,30 +308,47 @@ static noinline int create_subvol(struct btrfs_root *root, * insert the directory item */ key.offset = (u64)-1; - dir = root->fs_info->sb->s_root->d_inode; - ret = btrfs_insert_dir_item(trans, root->fs_info->tree_root, + dir = dentry->d_parent->d_inode; + ret = btrfs_set_inode_index(dir, &index); + BUG_ON(ret); + + ret = btrfs_insert_dir_item(trans, root, name, namelen, dir->i_ino, &key, - BTRFS_FT_DIR, 0); + BTRFS_FT_DIR, index); if (ret) goto fail; - ret = btrfs_insert_inode_ref(trans, root->fs_info->tree_root, - name, namelen, objectid, - root->fs_info->sb->s_root->d_inode->i_ino, 0); - if (ret) - goto fail; + btrfs_i_size_write(dir, dir->i_size + namelen * 2); + ret = btrfs_update_inode(trans, root, dir); + BUG_ON(ret); + + /* add the backref first */ + ret = btrfs_add_root_ref(trans, root->fs_info->tree_root, + objectid, BTRFS_ROOT_BACKREF_KEY, + root->root_key.objectid, + dir->i_ino, index, name, namelen); + + BUG_ON(ret); + + /* now add the forward ref */ + ret = btrfs_add_root_ref(trans, root->fs_info->tree_root, + root->root_key.objectid, BTRFS_ROOT_REF_KEY, + objectid, + dir->i_ino, index, name, namelen); + + BUG_ON(ret); ret = btrfs_commit_transaction(trans, root); if (ret) goto fail_commit; - new_root = btrfs_read_fs_root(root->fs_info, &key, name, namelen); + new_root = btrfs_read_fs_root_no_name(root->fs_info, &key); BUG_ON(!new_root); trans = btrfs_start_transaction(new_root, 1); BUG_ON(!trans); - ret = btrfs_create_subvol_root(new_root, dentry, trans, new_dirid, + ret = btrfs_create_subvol_root(trans, new_root, dentry, new_dirid, BTRFS_I(dir)->block_group); if (ret) goto fail; @@ -171,22 +363,23 @@ fail_commit: return ret; } -static int create_snapshot(struct btrfs_root *root, char *name, int namelen) +static int create_snapshot(struct btrfs_root *root, struct dentry *dentry, + char *name, int namelen) { struct btrfs_pending_snapshot *pending_snapshot; struct btrfs_trans_handle *trans; - int ret; + int ret = 0; int err; unsigned long nr = 0; if (!root->ref_cows) return -EINVAL; - ret = btrfs_check_free_space(root, 1, 0); + ret = btrfs_check_metadata_free_space(root); if (ret) goto fail_unlock; - pending_snapshot = kmalloc(sizeof(*pending_snapshot), GFP_NOFS); + pending_snapshot = kzalloc(sizeof(*pending_snapshot), GFP_NOFS); if (!pending_snapshot) { ret = -ENOMEM; goto fail_unlock; @@ -199,12 +392,12 @@ static int create_snapshot(struct btrfs_root *root, char *name, int namelen) } memcpy(pending_snapshot->name, name, namelen); pending_snapshot->name[namelen] = '\0'; + pending_snapshot->dentry = dentry; trans = btrfs_start_transaction(root, 1); BUG_ON(!trans); pending_snapshot->root = root; list_add(&pending_snapshot->list, &trans->transaction->pending_snapshots); - ret = btrfs_update_inode(trans, root, root->inode); err = btrfs_commit_transaction(trans, root); fail_unlock: @@ -228,7 +421,8 @@ static inline int btrfs_may_create(struct inode *dir, struct dentry *child) * inside this filesystem so it's quite a bit simpler. */ static noinline int btrfs_mksubvol(struct path *parent, char *name, - int mode, int namelen) + int mode, int namelen, + struct btrfs_root *snap_src) { struct dentry *dentry; int error; @@ -245,7 +439,8 @@ static noinline int btrfs_mksubvol(struct path *parent, char *name, goto out_dput; if (!IS_POSIXACL(parent->dentry->d_inode)) - mode &= ~current->fs->umask; + mode &= ~current_umask(); + error = mnt_want_write(parent->mnt); if (error) goto out_dput; @@ -254,11 +449,6 @@ static noinline int btrfs_mksubvol(struct path *parent, char *name, if (error) goto out_drop_write; - mode &= (S_IRWXUGO|S_ISVTX); - error = security_inode_mkdir(parent->dentry->d_inode, dentry, mode); - if (error) - goto out_drop_write; - /* * Actually perform the low-level subvolume creation after all * this VFS fuzz. @@ -269,8 +459,64 @@ static noinline int btrfs_mksubvol(struct path *parent, char *name, * Also we should pass on the mode eventually to allow creating new * subvolume with specific mode bits. */ - error = create_subvol(BTRFS_I(parent->dentry->d_inode)->root, dentry, - name, namelen); + if (snap_src) { + struct dentry *dir = dentry->d_parent; + struct dentry *test = dir->d_parent; + struct btrfs_path *path = btrfs_alloc_path(); + int ret; + u64 test_oid; + u64 parent_oid = BTRFS_I(dir->d_inode)->root->root_key.objectid; + + test_oid = snap_src->root_key.objectid; + + ret = btrfs_find_root_ref(snap_src->fs_info->tree_root, + path, parent_oid, test_oid); + if (ret == 0) + goto create; + btrfs_release_path(snap_src->fs_info->tree_root, path); + + /* we need to make sure we aren't creating a directory loop + * by taking a snapshot of something that has our current + * subvol in its directory tree. So, this loops through + * the dentries and checks the forward refs for each subvolume + * to see if is references the subvolume where we are + * placing this new snapshot. + */ + while (1) { + if (!test || + dir == snap_src->fs_info->sb->s_root || + test == snap_src->fs_info->sb->s_root || + test->d_inode->i_sb != snap_src->fs_info->sb) { + break; + } + if (S_ISLNK(test->d_inode->i_mode)) { + printk(KERN_INFO "Btrfs symlink in snapshot " + "path, failed\n"); + error = -EMLINK; + btrfs_free_path(path); + goto out_drop_write; + } + test_oid = + BTRFS_I(test->d_inode)->root->root_key.objectid; + ret = btrfs_find_root_ref(snap_src->fs_info->tree_root, + path, test_oid, parent_oid); + if (ret == 0) { + printk(KERN_INFO "Btrfs snapshot creation " + "failed, looping\n"); + error = -EMLINK; + btrfs_free_path(path); + goto out_drop_write; + } + btrfs_release_path(snap_src->fs_info->tree_root, path); + test = test->d_parent; + } +create: + btrfs_free_path(path); + error = create_snapshot(snap_src, dentry, name, namelen); + } else { + error = create_subvol(BTRFS_I(parent->dentry->d_inode)->root, + dentry, name, namelen); + } if (error) goto out_drop_write; @@ -285,7 +531,7 @@ out_unlock: } -int btrfs_defrag_file(struct file *file) +static int btrfs_defrag_file(struct file *file) { struct inode *inode = fdentry(file)->d_inode; struct btrfs_root *root = BTRFS_I(inode)->root; @@ -300,7 +546,7 @@ int btrfs_defrag_file(struct file *file) unsigned long i; int ret; - ret = btrfs_check_free_space(root, inode->i_size, 0); + ret = btrfs_check_data_free_space(root, inode, inode->i_size); if (ret) return -ENOSPC; @@ -363,10 +609,6 @@ out_unlock: return 0; } -/* - * Called inside transaction, so use GFP_NOFS - */ - static int btrfs_ioctl_resize(struct btrfs_root *root, void __user *arg) { u64 new_size; @@ -381,15 +623,15 @@ static int btrfs_ioctl_resize(struct btrfs_root *root, void __user *arg) int namelen; int mod = 0; - vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS); + if (root->fs_info->sb->s_flags & MS_RDONLY) + return -EROFS; - if (!vol_args) - return -ENOMEM; + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; - if (copy_from_user(vol_args, arg, sizeof(*vol_args))) { - ret = -EFAULT; - goto out; - } + vol_args = memdup_user(arg, sizeof(*vol_args)); + if (IS_ERR(vol_args)) + return PTR_ERR(vol_args); vol_args->name[BTRFS_PATH_NAME_MAX] = '\0'; namelen = strlen(vol_args->name); @@ -403,11 +645,13 @@ static int btrfs_ioctl_resize(struct btrfs_root *root, void __user *arg) *devstr = '\0'; devstr = vol_args->name; devid = simple_strtoull(devstr, &end, 10); - printk(KERN_INFO "resizing devid %llu\n", devid); + printk(KERN_INFO "resizing devid %llu\n", + (unsigned long long)devid); } - device = btrfs_find_device(root, devid, NULL); + device = btrfs_find_device(root, devid, NULL, NULL); if (!device) { - printk(KERN_INFO "resizer unable to find device %llu\n", devid); + printk(KERN_INFO "resizer unable to find device %llu\n", + (unsigned long long)devid); ret = -EINVAL; goto out_unlock; } @@ -465,31 +709,28 @@ static int btrfs_ioctl_resize(struct btrfs_root *root, void __user *arg) out_unlock: mutex_unlock(&root->fs_info->volume_mutex); -out: kfree(vol_args); return ret; } static noinline int btrfs_ioctl_snap_create(struct file *file, - void __user *arg) + void __user *arg, int subvol) { struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root; struct btrfs_ioctl_vol_args *vol_args; struct btrfs_dir_item *di; struct btrfs_path *path; + struct file *src_file; u64 root_dirid; int namelen; - int ret; - - vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS); + int ret = 0; - if (!vol_args) - return -ENOMEM; + if (root->fs_info->sb->s_flags & MS_RDONLY) + return -EROFS; - if (copy_from_user(vol_args, arg, sizeof(*vol_args))) { - ret = -EFAULT; - goto out; - } + vol_args = memdup_user(arg, sizeof(*vol_args)); + if (IS_ERR(vol_args)) + return PTR_ERR(vol_args); vol_args->name[BTRFS_PATH_NAME_MAX] = '\0'; namelen = strlen(vol_args->name); @@ -520,12 +761,30 @@ static noinline int btrfs_ioctl_snap_create(struct file *file, goto out; } - if (root == root->fs_info->tree_root) { + if (subvol) { ret = btrfs_mksubvol(&file->f_path, vol_args->name, file->f_path.dentry->d_inode->i_mode, - namelen); + namelen, NULL); } else { - ret = create_snapshot(root, vol_args->name, namelen); + struct inode *src_inode; + src_file = fget(vol_args->fd); + if (!src_file) { + ret = -EINVAL; + goto out; + } + + src_inode = src_file->f_path.dentry->d_inode; + if (src_inode->i_sb != file->f_path.dentry->d_inode->i_sb) { + printk(KERN_INFO "btrfs: Snapshot src from " + "another FS\n"); + ret = -EINVAL; + fput(src_file); + goto out; + } + ret = btrfs_mksubvol(&file->f_path, vol_args->name, + file->f_path.dentry->d_inode->i_mode, + namelen, BTRFS_I(src_inode)->root); + fput(src_file); } out: @@ -537,65 +796,77 @@ static int btrfs_ioctl_defrag(struct file *file) { struct inode *inode = fdentry(file)->d_inode; struct btrfs_root *root = BTRFS_I(inode)->root; + int ret; + + ret = mnt_want_write(file->f_path.mnt); + if (ret) + return ret; switch (inode->i_mode & S_IFMT) { case S_IFDIR: + if (!capable(CAP_SYS_ADMIN)) { + ret = -EPERM; + goto out; + } btrfs_defrag_root(root, 0); btrfs_defrag_root(root->fs_info->extent_root, 0); break; case S_IFREG: + if (!(file->f_mode & FMODE_WRITE)) { + ret = -EINVAL; + goto out; + } btrfs_defrag_file(file); break; } - - return 0; +out: + mnt_drop_write(file->f_path.mnt); + return ret; } -long btrfs_ioctl_add_dev(struct btrfs_root *root, void __user *arg) +static long btrfs_ioctl_add_dev(struct btrfs_root *root, void __user *arg) { struct btrfs_ioctl_vol_args *vol_args; int ret; - vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS); + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; - if (!vol_args) - return -ENOMEM; + vol_args = memdup_user(arg, sizeof(*vol_args)); + if (IS_ERR(vol_args)) + return PTR_ERR(vol_args); - if (copy_from_user(vol_args, arg, sizeof(*vol_args))) { - ret = -EFAULT; - goto out; - } vol_args->name[BTRFS_PATH_NAME_MAX] = '\0'; ret = btrfs_init_new_device(root, vol_args->name); -out: kfree(vol_args); return ret; } -long btrfs_ioctl_rm_dev(struct btrfs_root *root, void __user *arg) +static long btrfs_ioctl_rm_dev(struct btrfs_root *root, void __user *arg) { struct btrfs_ioctl_vol_args *vol_args; int ret; - vol_args = kmalloc(sizeof(*vol_args), GFP_NOFS); + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; - if (!vol_args) - return -ENOMEM; + if (root->fs_info->sb->s_flags & MS_RDONLY) + return -EROFS; + + vol_args = memdup_user(arg, sizeof(*vol_args)); + if (IS_ERR(vol_args)) + return PTR_ERR(vol_args); - if (copy_from_user(vol_args, arg, sizeof(*vol_args))) { - ret = -EFAULT; - goto out; - } vol_args->name[BTRFS_PATH_NAME_MAX] = '\0'; ret = btrfs_rm_device(root, vol_args->name); -out: kfree(vol_args); return ret; } -long btrfs_ioctl_clone(struct file *file, unsigned long src_fd) +static long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, + u64 off, u64 olen, u64 destoff) { struct inode *inode = fdentry(file)->d_inode; struct btrfs_root *root = BTRFS_I(inode)->root; @@ -609,12 +880,39 @@ long btrfs_ioctl_clone(struct file *file, unsigned long src_fd) u32 nritems; int slot; int ret; + u64 len = olen; + u64 bs = root->fs_info->sb->s_blocksize; + u64 hint_byte; + + /* + * TODO: + * - split compressed inline extents. annoying: we need to + * decompress into destination's address_space (the file offset + * may change, so source mapping won't do), then recompress (or + * otherwise reinsert) a subrange. + * - allow ranges within the same file to be cloned (provided + * they don't overlap)? + */ + + /* the destination must be opened for writing */ + if (!(file->f_mode & FMODE_WRITE)) + return -EINVAL; + + ret = mnt_want_write(file->f_path.mnt); + if (ret) + return ret; - src_file = fget(src_fd); - if (!src_file) - return -EBADF; + src_file = fget(srcfd); + if (!src_file) { + ret = -EBADF; + goto out_drop_write; + } src = src_file->f_dentry->d_inode; + ret = -EINVAL; + if (src == inode) + goto out_fput; + ret = -EISDIR; if (S_ISDIR(src->i_mode) || S_ISDIR(inode->i_mode)) goto out_fput; @@ -643,27 +941,44 @@ long btrfs_ioctl_clone(struct file *file, unsigned long src_fd) mutex_lock(&inode->i_mutex); } - ret = -ENOTEMPTY; - if (inode->i_size) + /* determine range to clone */ + ret = -EINVAL; + if (off >= src->i_size || off + len > src->i_size) + goto out_unlock; + if (len == 0) + olen = len = src->i_size - off; + /* if we extend to eof, continue to block boundary */ + if (off + len == src->i_size) + len = ((src->i_size + bs-1) & ~(bs-1)) + - off; + + /* verify the end result is block aligned */ + if ((off & (bs-1)) || + ((off + len) & (bs-1))) goto out_unlock; /* do any pending delalloc/csum calc on src, one way or another, and lock file content */ while (1) { struct btrfs_ordered_extent *ordered; - lock_extent(&BTRFS_I(src)->io_tree, 0, (u64)-1, GFP_NOFS); - ordered = btrfs_lookup_first_ordered_extent(inode, (u64)-1); + lock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS); + ordered = btrfs_lookup_first_ordered_extent(inode, off+len); if (BTRFS_I(src)->delalloc_bytes == 0 && !ordered) break; - unlock_extent(&BTRFS_I(src)->io_tree, 0, (u64)-1, GFP_NOFS); + unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS); if (ordered) btrfs_put_ordered_extent(ordered); - btrfs_wait_ordered_range(src, 0, (u64)-1); + btrfs_wait_ordered_range(src, off, off+len); } trans = btrfs_start_transaction(root, 1); BUG_ON(!trans); + /* punch hole in destination first */ + btrfs_drop_extents(trans, root, inode, off, off + len, + off + len, 0, &hint_byte); + + /* clone data */ key.objectid = src->i_ino; key.type = BTRFS_EXTENT_DATA_KEY; key.offset = 0; @@ -690,59 +1005,133 @@ long btrfs_ioctl_clone(struct file *file, unsigned long src_fd) slot = path->slots[0]; btrfs_item_key_to_cpu(leaf, &key, slot); - if (btrfs_key_type(&key) > BTRFS_CSUM_ITEM_KEY || + if (btrfs_key_type(&key) > BTRFS_EXTENT_DATA_KEY || key.objectid != src->i_ino) break; - if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY || - btrfs_key_type(&key) == BTRFS_CSUM_ITEM_KEY) { + if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY) { + struct btrfs_file_extent_item *extent; + int type; u32 size; struct btrfs_key new_key; + u64 disko = 0, diskl = 0; + u64 datao = 0, datal = 0; + u8 comp; size = btrfs_item_size_nr(leaf, slot); read_extent_buffer(leaf, buf, btrfs_item_ptr_offset(leaf, slot), size); + + extent = btrfs_item_ptr(leaf, slot, + struct btrfs_file_extent_item); + comp = btrfs_file_extent_compression(leaf, extent); + type = btrfs_file_extent_type(leaf, extent); + if (type == BTRFS_FILE_EXTENT_REG || + type == BTRFS_FILE_EXTENT_PREALLOC) { + disko = btrfs_file_extent_disk_bytenr(leaf, + extent); + diskl = btrfs_file_extent_disk_num_bytes(leaf, + extent); + datao = btrfs_file_extent_offset(leaf, extent); + datal = btrfs_file_extent_num_bytes(leaf, + extent); + } else if (type == BTRFS_FILE_EXTENT_INLINE) { + /* take upper bound, may be compressed */ + datal = btrfs_file_extent_ram_bytes(leaf, + extent); + } btrfs_release_path(root, path); + if (key.offset + datal < off || + key.offset >= off+len) + goto next; + memcpy(&new_key, &key, sizeof(new_key)); new_key.objectid = inode->i_ino; - ret = btrfs_insert_empty_item(trans, root, path, - &new_key, size); - if (ret) - goto out; - - leaf = path->nodes[0]; - slot = path->slots[0]; - write_extent_buffer(leaf, buf, + new_key.offset = key.offset + destoff - off; + + if (type == BTRFS_FILE_EXTENT_REG || + type == BTRFS_FILE_EXTENT_PREALLOC) { + ret = btrfs_insert_empty_item(trans, root, path, + &new_key, size); + if (ret) + goto out; + + leaf = path->nodes[0]; + slot = path->slots[0]; + write_extent_buffer(leaf, buf, btrfs_item_ptr_offset(leaf, slot), size); - btrfs_mark_buffer_dirty(leaf); - } - if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY) { - struct btrfs_file_extent_item *extent; - int found_type; - - extent = btrfs_item_ptr(leaf, slot, + extent = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item); - found_type = btrfs_file_extent_type(leaf, extent); - if (found_type == BTRFS_FILE_EXTENT_REG) { - u64 ds = btrfs_file_extent_disk_bytenr(leaf, - extent); - u64 dl = btrfs_file_extent_disk_num_bytes(leaf, - extent); - /* ds == 0 means there's a hole */ - if (ds != 0) { + + if (off > key.offset) { + datao += off - key.offset; + datal -= off - key.offset; + } + if (key.offset + datao + datal + key.offset > + off + len) + datal = off + len - key.offset - datao; + /* disko == 0 means it's a hole */ + if (!disko) + datao = 0; + + btrfs_set_file_extent_offset(leaf, extent, + datao); + btrfs_set_file_extent_num_bytes(leaf, extent, + datal); + if (disko) { + inode_add_bytes(inode, datal); ret = btrfs_inc_extent_ref(trans, root, - ds, dl, leaf->start, - root->root_key.objectid, - trans->transid, - inode->i_ino); + disko, diskl, 0, + root->root_key.objectid, + inode->i_ino, + new_key.offset - datao); BUG_ON(ret); } + } else if (type == BTRFS_FILE_EXTENT_INLINE) { + u64 skip = 0; + u64 trim = 0; + if (off > key.offset) { + skip = off - key.offset; + new_key.offset += skip; + } + + if (key.offset + datal > off+len) + trim = key.offset + datal - (off+len); + + if (comp && (skip || trim)) { + ret = -EINVAL; + goto out; + } + size -= skip + trim; + datal -= skip + trim; + ret = btrfs_insert_empty_item(trans, root, path, + &new_key, size); + if (ret) + goto out; + + if (skip) { + u32 start = + btrfs_file_extent_calc_inline_size(0); + memmove(buf+start, buf+start+skip, + datal); + } + + leaf = path->nodes[0]; + slot = path->slots[0]; + write_extent_buffer(leaf, buf, + btrfs_item_ptr_offset(leaf, slot), + size); + inode_add_bytes(inode, datal); } + + btrfs_mark_buffer_dirty(leaf); } + +next: btrfs_release_path(root, path); key.offset++; } @@ -751,13 +1140,13 @@ out: btrfs_release_path(root, path); if (ret == 0) { inode->i_mtime = inode->i_ctime = CURRENT_TIME; - inode_set_bytes(inode, inode_get_bytes(src)); - btrfs_i_size_write(inode, src->i_size); + if (destoff + olen > inode->i_size) + btrfs_i_size_write(inode, destoff + olen); BTRFS_I(inode)->flags = BTRFS_I(src)->flags; ret = btrfs_update_inode(trans, root, inode); } btrfs_end_transaction(trans, root); - unlock_extent(&BTRFS_I(src)->io_tree, 0, (u64)-1, GFP_NOFS); + unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS); if (ret) vmtruncate(inode, 0); out_unlock: @@ -767,16 +1156,28 @@ out_unlock: btrfs_free_path(path); out_fput: fput(src_file); +out_drop_write: + mnt_drop_write(file->f_path.mnt); return ret; } +static long btrfs_ioctl_clone_range(struct file *file, void __user *argp) +{ + struct btrfs_ioctl_clone_range_args args; + + if (copy_from_user(&args, argp, sizeof(args))) + return -EFAULT; + return btrfs_ioctl_clone(file, args.src_fd, args.src_offset, + args.src_length, args.dest_offset); +} + /* * there are many ways the trans_start and trans_end ioctls can lead * to deadlocks. They should only be used by applications that * basically own the machine, and have a very in depth understanding * of all the possible deadlocks and enospc problems. */ -long btrfs_ioctl_trans_start(struct file *file) +static long btrfs_ioctl_trans_start(struct file *file) { struct inode *inode = fdentry(file)->d_inode; struct btrfs_root *root = BTRFS_I(inode)->root; @@ -791,6 +1192,10 @@ long btrfs_ioctl_trans_start(struct file *file) goto out; } + ret = mnt_want_write(file->f_path.mnt); + if (ret) + goto out; + mutex_lock(&root->fs_info->trans_mutex); root->fs_info->open_ioctl_trans++; mutex_unlock(&root->fs_info->trans_mutex); @@ -830,6 +1235,8 @@ long btrfs_ioctl_trans_end(struct file *file) root->fs_info->open_ioctl_trans--; mutex_unlock(&root->fs_info->trans_mutex); + mnt_drop_write(file->f_path.mnt); + out: return ret; } @@ -838,28 +1245,38 @@ long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg) { struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root; + void __user *argp = (void __user *)arg; switch (cmd) { + case FS_IOC_GETFLAGS: + return btrfs_ioctl_getflags(file, argp); + case FS_IOC_SETFLAGS: + return btrfs_ioctl_setflags(file, argp); + case FS_IOC_GETVERSION: + return btrfs_ioctl_getversion(file, argp); case BTRFS_IOC_SNAP_CREATE: - return btrfs_ioctl_snap_create(file, (void __user *)arg); + return btrfs_ioctl_snap_create(file, argp, 0); + case BTRFS_IOC_SUBVOL_CREATE: + return btrfs_ioctl_snap_create(file, argp, 1); case BTRFS_IOC_DEFRAG: return btrfs_ioctl_defrag(file); case BTRFS_IOC_RESIZE: - return btrfs_ioctl_resize(root, (void __user *)arg); + return btrfs_ioctl_resize(root, argp); case BTRFS_IOC_ADD_DEV: - return btrfs_ioctl_add_dev(root, (void __user *)arg); + return btrfs_ioctl_add_dev(root, argp); case BTRFS_IOC_RM_DEV: - return btrfs_ioctl_rm_dev(root, (void __user *)arg); + return btrfs_ioctl_rm_dev(root, argp); case BTRFS_IOC_BALANCE: return btrfs_balance(root->fs_info->dev_root); case BTRFS_IOC_CLONE: - return btrfs_ioctl_clone(file, arg); + return btrfs_ioctl_clone(file, arg, 0, 0, 0); + case BTRFS_IOC_CLONE_RANGE: + return btrfs_ioctl_clone_range(file, argp); case BTRFS_IOC_TRANS_START: return btrfs_ioctl_trans_start(file); case BTRFS_IOC_TRANS_END: return btrfs_ioctl_trans_end(file); case BTRFS_IOC_SYNC: - btrfs_start_delalloc_inodes(root); btrfs_sync_fs(file->f_dentry->d_sb, 1); return 0; }