return -EINVAL;
atomic_inc(&root->will_be_snapshoted);
- smp_mb__after_atomic_inc();
+ smp_mb__after_atomic();
btrfs_wait_nocow_write(root);
ret = btrfs_start_delalloc_inodes(root, 0);
if (ret)
goto fail;
+ /*
+ * If orphan cleanup did remove any orphans, it means the tree was
+ * modified and therefore the commit root is not the same as the
+ * current root anymore. This is a problem, because send uses the
+ * commit root and therefore can see inode items that don't exist
+ * in the current root anymore, and for example make calls to
+ * btrfs_iget, which will do tree lookups based on the current root
+ * and not on the commit root. Those lookups will fail, returning a
+ * -ESTALE error, and making send fail with that error. So make sure
+ * a send does not see any orphans we have just removed, and that it
+ * will see the same inodes regardless of whether a transaction
+ * commit happened before it started (meaning that the commit root
+ * will be the same as the current root) or not.
+ */
+ if (readonly && pending_snapshot->snap->node !=
+ pending_snapshot->snap->commit_root) {
+ trans = btrfs_join_transaction(pending_snapshot->snap);
+ if (IS_ERR(trans) && PTR_ERR(trans) != -ENOENT) {
+ ret = PTR_ERR(trans);
+ goto fail;
+ }
+ if (!IS_ERR(trans)) {
+ ret = btrfs_commit_transaction(trans,
+ pending_snapshot->snap);
+ if (ret)
+ goto fail;
+ }
+ }
+
inode = btrfs_lookup_dentry(dentry->d_parent->d_inode, dentry);
if (IS_ERR(inode)) {
ret = PTR_ERR(inode);
sizestr = vol_args->name;
devstr = strchr(sizestr, ':');
if (devstr) {
- char *end;
sizestr = devstr + 1;
*devstr = '\0';
devstr = vol_args->name;
- devid = simple_strtoull(devstr, &end, 10);
+ ret = kstrtoull(devstr, 10, &devid);
+ if (ret)
+ goto out_free;
if (!devid) {
ret = -EINVAL;
goto out_free;
new_size = old_size - new_size;
} else if (mod > 0) {
if (new_size > ULLONG_MAX - old_size) {
- ret = -EINVAL;
+ ret = -ERANGE;
goto out_free;
}
new_size = old_size + new_size;
struct btrfs_path *path,
struct btrfs_key *key,
struct btrfs_ioctl_search_key *sk,
- char *buf,
+ size_t *buf_size,
+ char __user *ubuf,
unsigned long *sk_offset,
int *num_found)
{
if (!key_in_sk(key, sk))
continue;
- if (sizeof(sh) + item_len > BTRFS_SEARCH_ARGS_BUFSIZE)
+ if (sizeof(sh) + item_len > *buf_size) {
+ if (*num_found) {
+ ret = 1;
+ goto out;
+ }
+
+ /*
+ * return one empty item back for v1, which does not
+ * handle -EOVERFLOW
+ */
+
+ *buf_size = sizeof(sh) + item_len;
item_len = 0;
+ ret = -EOVERFLOW;
+ }
- if (sizeof(sh) + item_len + *sk_offset >
- BTRFS_SEARCH_ARGS_BUFSIZE) {
+ if (sizeof(sh) + item_len + *sk_offset > *buf_size) {
ret = 1;
- goto overflow;
+ goto out;
}
sh.objectid = key->objectid;
sh.transid = found_transid;
/* copy search result header */
- memcpy(buf + *sk_offset, &sh, sizeof(sh));
+ if (copy_to_user(ubuf + *sk_offset, &sh, sizeof(sh))) {
+ ret = -EFAULT;
+ goto out;
+ }
+
*sk_offset += sizeof(sh);
if (item_len) {
- char *p = buf + *sk_offset;
+ char __user *up = ubuf + *sk_offset;
/* copy the item */
- read_extent_buffer(leaf, p,
- item_off, item_len);
+ if (read_extent_buffer_to_user(leaf, up,
+ item_off, item_len)) {
+ ret = -EFAULT;
+ goto out;
+ }
+
*sk_offset += item_len;
}
(*num_found)++;
- if (*num_found >= sk->nr_items)
- break;
+ if (ret) /* -EOVERFLOW from above */
+ goto out;
+
+ if (*num_found >= sk->nr_items) {
+ ret = 1;
+ goto out;
+ }
}
advance_key:
ret = 0;
key->objectid++;
} else
ret = 1;
-overflow:
+out:
+ /*
+ * 0: all items from this leaf copied, continue with next
+ * 1: * more items can be copied, but unused buffer is too small
+ * * all items were found
+ * Either way, it will stops the loop which iterates to the next
+ * leaf
+ * -EOVERFLOW: item was to large for buffer
+ * -EFAULT: could not copy extent buffer back to userspace
+ */
return ret;
}
static noinline int search_ioctl(struct inode *inode,
- struct btrfs_ioctl_search_args *args)
+ struct btrfs_ioctl_search_key *sk,
+ size_t *buf_size,
+ char __user *ubuf)
{
struct btrfs_root *root;
struct btrfs_key key;
struct btrfs_path *path;
- struct btrfs_ioctl_search_key *sk = &args->key;
struct btrfs_fs_info *info = BTRFS_I(inode)->root->fs_info;
int ret;
int num_found = 0;
unsigned long sk_offset = 0;
+ if (*buf_size < sizeof(struct btrfs_ioctl_search_header)) {
+ *buf_size = sizeof(struct btrfs_ioctl_search_header);
+ return -EOVERFLOW;
+ }
+
path = btrfs_alloc_path();
if (!path)
return -ENOMEM;
ret = 0;
goto err;
}
- ret = copy_to_sk(root, path, &key, sk, args->buf,
+ ret = copy_to_sk(root, path, &key, sk, buf_size, ubuf,
&sk_offset, &num_found);
btrfs_release_path(path);
- if (ret || num_found >= sk->nr_items)
+ if (ret)
break;
}
- ret = 0;
+ if (ret > 0)
+ ret = 0;
err:
sk->nr_items = num_found;
btrfs_free_path(path);
static noinline int btrfs_ioctl_tree_search(struct file *file,
void __user *argp)
{
- struct btrfs_ioctl_search_args *args;
- struct inode *inode;
- int ret;
+ struct btrfs_ioctl_search_args __user *uargs;
+ struct btrfs_ioctl_search_key sk;
+ struct inode *inode;
+ int ret;
+ size_t buf_size;
if (!capable(CAP_SYS_ADMIN))
return -EPERM;
- args = memdup_user(argp, sizeof(*args));
- if (IS_ERR(args))
- return PTR_ERR(args);
+ uargs = (struct btrfs_ioctl_search_args __user *)argp;
+
+ if (copy_from_user(&sk, &uargs->key, sizeof(sk)))
+ return -EFAULT;
+
+ buf_size = sizeof(uargs->buf);
inode = file_inode(file);
- ret = search_ioctl(inode, args);
- if (ret == 0 && copy_to_user(argp, args, sizeof(*args)))
+ ret = search_ioctl(inode, &sk, &buf_size, uargs->buf);
+
+ /*
+ * In the origin implementation an overflow is handled by returning a
+ * search header with a len of zero, so reset ret.
+ */
+ if (ret == -EOVERFLOW)
+ ret = 0;
+
+ if (ret == 0 && copy_to_user(&uargs->key, &sk, sizeof(sk)))
ret = -EFAULT;
- kfree(args);
+ return ret;
+}
+
+static noinline int btrfs_ioctl_tree_search_v2(struct file *file,
+ void __user *argp)
+{
+ struct btrfs_ioctl_search_args_v2 __user *uarg;
+ struct btrfs_ioctl_search_args_v2 args;
+ struct inode *inode;
+ int ret;
+ size_t buf_size;
+ const size_t buf_limit = 16 * 1024 * 1024;
+
+ if (!capable(CAP_SYS_ADMIN))
+ return -EPERM;
+
+ /* copy search header and buffer size */
+ uarg = (struct btrfs_ioctl_search_args_v2 __user *)argp;
+ if (copy_from_user(&args, uarg, sizeof(args)))
+ return -EFAULT;
+
+ buf_size = args.buf_size;
+
+ if (buf_size < sizeof(struct btrfs_ioctl_search_header))
+ return -EOVERFLOW;
+
+ /* limit result size to 16MB */
+ if (buf_size > buf_limit)
+ buf_size = buf_limit;
+
+ inode = file_inode(file);
+ ret = search_ioctl(inode, &args.key, &buf_size,
+ (char *)(&uarg->buf[0]));
+ if (ret == 0 && copy_to_user(&uarg->key, &args.key, sizeof(args.key)))
+ ret = -EFAULT;
+ else if (ret == -EOVERFLOW &&
+ copy_to_user(&uarg->buf_size, &buf_size, sizeof(buf_size)))
+ ret = -EFAULT;
+
return ret;
}
* again is not run concurrently.
*/
spin_lock(&dest->root_item_lock);
- root_flags = btrfs_root_flags(&root->root_item);
- if (root->send_in_progress == 0) {
- btrfs_set_root_flags(&root->root_item,
+ root_flags = btrfs_root_flags(&dest->root_item);
+ if (dest->send_in_progress == 0) {
+ btrfs_set_root_flags(&dest->root_item,
root_flags | BTRFS_ROOT_SUBVOL_DEAD);
spin_unlock(&dest->root_item_lock);
} else {
spin_unlock(&dest->root_item_lock);
btrfs_warn(root->fs_info,
"Attempt to delete subvolume %llu during send",
- root->root_key.objectid);
+ dest->root_key.objectid);
err = -EPERM;
goto out_dput;
}
out_unlock:
if (err) {
spin_lock(&dest->root_item_lock);
- root_flags = btrfs_root_flags(&root->root_item);
- btrfs_set_root_flags(&root->root_item,
+ root_flags = btrfs_root_flags(&dest->root_item);
+ btrfs_set_root_flags(&dest->root_item,
root_flags & ~BTRFS_ROOT_SUBVOL_DEAD);
spin_unlock(&dest->root_item_lock);
}
lock_extent(&BTRFS_I(inode)->io_tree, off, off + len - 1);
ordered = btrfs_lookup_first_ordered_extent(inode,
off + len - 1);
- if (!ordered &&
+ if ((!ordered ||
+ ordered->file_offset + ordered->len <= off ||
+ ordered->file_offset >= off + len) &&
!test_range_bit(&BTRFS_I(inode)->io_tree, off,
- off + len - 1, EXTENT_DELALLOC, 0, NULL))
+ off + len - 1, EXTENT_DELALLOC, 0, NULL)) {
+ if (ordered)
+ btrfs_put_ordered_extent(ordered);
break;
+ }
unlock_extent(&BTRFS_I(inode)->io_tree, off, off + len - 1);
if (ordered)
btrfs_put_ordered_extent(ordered);
return ret;
}
+static int clone_finish_inode_update(struct btrfs_trans_handle *trans,
+ struct inode *inode,
+ u64 endoff,
+ const u64 destoff,
+ const u64 olen)
+{
+ struct btrfs_root *root = BTRFS_I(inode)->root;
+ int ret;
+
+ inode_inc_iversion(inode);
+ inode->i_mtime = inode->i_ctime = CURRENT_TIME;
+ /*
+ * We round up to the block size at eof when determining which
+ * extents to clone above, but shouldn't round up the file size.
+ */
+ if (endoff > destoff + olen)
+ endoff = destoff + olen;
+ if (endoff > inode->i_size)
+ btrfs_i_size_write(inode, endoff);
+
+ ret = btrfs_update_inode(trans, root, inode);
+ if (ret) {
+ btrfs_abort_transaction(trans, root, ret);
+ btrfs_end_transaction(trans, root);
+ goto out;
+ }
+ ret = btrfs_end_transaction(trans, root);
+out:
+ return ret;
+}
+
+static void clone_update_extent_map(struct inode *inode,
+ const struct btrfs_trans_handle *trans,
+ const struct btrfs_path *path,
+ struct btrfs_file_extent_item *fi,
+ const u64 hole_offset,
+ const u64 hole_len)
+{
+ struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+ struct extent_map *em;
+ int ret;
+
+ em = alloc_extent_map();
+ if (!em) {
+ set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
+ &BTRFS_I(inode)->runtime_flags);
+ return;
+ }
+
+ if (fi) {
+ btrfs_extent_item_to_extent_map(inode, path, fi, false, em);
+ em->generation = -1;
+ if (btrfs_file_extent_type(path->nodes[0], fi) ==
+ BTRFS_FILE_EXTENT_INLINE)
+ set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
+ &BTRFS_I(inode)->runtime_flags);
+ } else {
+ em->start = hole_offset;
+ em->len = hole_len;
+ em->ram_bytes = em->len;
+ em->orig_start = hole_offset;
+ em->block_start = EXTENT_MAP_HOLE;
+ em->block_len = 0;
+ em->orig_block_len = 0;
+ em->compress_type = BTRFS_COMPRESS_NONE;
+ em->generation = trans->transid;
+ }
+
+ while (1) {
+ write_lock(&em_tree->lock);
+ ret = add_extent_mapping(em_tree, em, 1);
+ write_unlock(&em_tree->lock);
+ if (ret != -EEXIST) {
+ free_extent_map(em);
+ break;
+ }
+ btrfs_drop_extent_cache(inode, em->start,
+ em->start + em->len - 1, 0);
+ }
+
+ if (unlikely(ret))
+ set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
+ &BTRFS_I(inode)->runtime_flags);
+}
+
/**
* btrfs_clone() - clone a range from inode file to another
*
* @destoff: Offset within @inode to start clone
*/
static int btrfs_clone(struct inode *src, struct inode *inode,
- u64 off, u64 olen, u64 olen_aligned, u64 destoff)
+ const u64 off, const u64 olen, const u64 olen_aligned,
+ const u64 destoff)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_path *path = NULL;
int slot;
int ret;
int no_quota;
- u64 len = olen_aligned;
+ const u64 len = olen_aligned;
u64 last_disko = 0;
+ u64 last_dest_end = destoff;
ret = -ENOMEM;
buf = vmalloc(btrfs_level_size(root, 0));
/* clone data */
key.objectid = btrfs_ino(src);
key.type = BTRFS_EXTENT_DATA_KEY;
- key.offset = 0;
+ key.offset = off;
while (1) {
/*
0, 0);
if (ret < 0)
goto out;
+ /*
+ * First search, if no extent item that starts at offset off was
+ * found but the previous item is an extent item, it's possible
+ * it might overlap our target range, therefore process it.
+ */
+ if (key.offset == off && ret > 0 && path->slots[0] > 0) {
+ btrfs_item_key_to_cpu(path->nodes[0], &key,
+ path->slots[0] - 1);
+ if (key.type == BTRFS_EXTENT_DATA_KEY)
+ path->slots[0]--;
+ }
nritems = btrfs_header_nritems(path->nodes[0]);
process_slot:
u64 disko = 0, diskl = 0;
u64 datao = 0, datal = 0;
u8 comp;
- u64 endoff;
+ u64 drop_start;
extent = btrfs_item_ptr(leaf, slot,
struct btrfs_file_extent_item);
extent);
}
- if (key.offset + datal <= off ||
- key.offset >= off + len - 1) {
+ /*
+ * The first search might have left us at an extent
+ * item that ends before our target range's start, can
+ * happen if we have holes and NO_HOLES feature enabled.
+ */
+ if (key.offset + datal <= off) {
path->slots[0]++;
goto process_slot;
+ } else if (key.offset >= off + len) {
+ break;
}
size = btrfs_item_size_nr(leaf, slot);
else
new_key.offset = destoff;
+ /*
+ * Deal with a hole that doesn't have an extent item
+ * that represents it (NO_HOLES feature enabled).
+ * This hole is either in the middle of the cloning
+ * range or at the beginning (fully overlaps it or
+ * partially overlaps it).
+ */
+ if (new_key.offset != last_dest_end)
+ drop_start = last_dest_end;
+ else
+ drop_start = new_key.offset;
+
/*
* 1 - adjusting old extent (we may have to split it)
* 1 - add new extent
* | ------------- extent ------------- |
*/
- /* substract range b */
+ /* subtract range b */
if (key.offset + datal > off + len)
datal = off + len - key.offset;
- /* substract range a */
+ /* subtract range a */
if (off > key.offset) {
datao += off - key.offset;
datal -= off - key.offset;
}
ret = btrfs_drop_extents(trans, root, inode,
- new_key.offset,
+ drop_start,
new_key.offset + datal,
1);
if (ret) {
aligned_end = ALIGN(new_key.offset + datal,
root->sectorsize);
ret = btrfs_drop_extents(trans, root, inode,
- new_key.offset,
+ drop_start,
aligned_end,
1);
if (ret) {
btrfs_item_ptr_offset(leaf, slot),
size);
inode_add_bytes(inode, datal);
+ extent = btrfs_item_ptr(leaf, slot,
+ struct btrfs_file_extent_item);
}
+ /* If we have an implicit hole (NO_HOLES feature). */
+ if (drop_start < new_key.offset)
+ clone_update_extent_map(inode, trans,
+ path, NULL, drop_start,
+ new_key.offset - drop_start);
+
+ clone_update_extent_map(inode, trans, path,
+ extent, 0, 0);
+
btrfs_mark_buffer_dirty(leaf);
btrfs_release_path(path);
- inode_inc_iversion(inode);
- inode->i_mtime = inode->i_ctime = CURRENT_TIME;
-
- /*
- * we round up to the block size at eof when
- * determining which extents to clone above,
- * but shouldn't round up the file size
- */
- endoff = new_key.offset + datal;
- if (endoff > destoff+olen)
- endoff = destoff+olen;
- if (endoff > inode->i_size)
- btrfs_i_size_write(inode, endoff);
-
- ret = btrfs_update_inode(trans, root, inode);
- if (ret) {
- btrfs_abort_transaction(trans, root, ret);
- btrfs_end_transaction(trans, root);
+ last_dest_end = new_key.offset + datal;
+ ret = clone_finish_inode_update(trans, inode,
+ last_dest_end,
+ destoff, olen);
+ if (ret)
goto out;
- }
- ret = btrfs_end_transaction(trans, root);
+ if (new_key.offset + datal >= destoff + len)
+ break;
}
btrfs_release_path(path);
key.offset++;
}
ret = 0;
+ if (last_dest_end < destoff + len) {
+ /*
+ * We have an implicit hole (NO_HOLES feature is enabled) that
+ * fully or partially overlaps our cloning range at its end.
+ */
+ btrfs_release_path(path);
+
+ /*
+ * 1 - remove extent(s)
+ * 1 - inode update
+ */
+ trans = btrfs_start_transaction(root, 2);
+ if (IS_ERR(trans)) {
+ ret = PTR_ERR(trans);
+ goto out;
+ }
+ ret = btrfs_drop_extents(trans, root, inode,
+ last_dest_end, destoff + len, 1);
+ if (ret) {
+ if (ret != -EOPNOTSUPP)
+ btrfs_abort_transaction(trans, root, ret);
+ btrfs_end_transaction(trans, root);
+ goto out;
+ }
+ ret = clone_finish_inode_update(trans, inode, destoff + len,
+ destoff, olen);
+ if (ret)
+ goto out;
+ clone_update_extent_map(inode, trans, path, NULL, last_dest_end,
+ destoff + len - last_dest_end);
+ }
+
out:
- btrfs_release_path(path);
btrfs_free_path(path);
vfree(buf);
return ret;
goto out_unlock;
}
- /* truncate page cache pages from target inode range */
- truncate_inode_pages_range(&inode->i_data, destoff,
- PAGE_CACHE_ALIGN(destoff + len) - 1);
+ /*
+ * Lock the target range too. Right after we replace the file extent
+ * items in the fs tree (which now point to the cloned data), we might
+ * have a worker replace them with extent items relative to a write
+ * operation that was issued before this clone operation (i.e. confront
+ * with inode.c:btrfs_finish_ordered_io).
+ */
+ if (same_inode) {
+ u64 lock_start = min_t(u64, off, destoff);
+ u64 lock_len = max_t(u64, off, destoff) + len - lock_start;
- lock_extent_range(src, off, len);
+ lock_extent_range(src, lock_start, lock_len);
+ } else {
+ lock_extent_range(src, off, len);
+ lock_extent_range(inode, destoff, len);
+ }
ret = btrfs_clone(src, inode, off, olen, len, destoff);
- unlock_extent(&BTRFS_I(src)->io_tree, off, off + len - 1);
+ if (same_inode) {
+ u64 lock_start = min_t(u64, off, destoff);
+ u64 lock_end = max_t(u64, off, destoff) + len - 1;
+
+ unlock_extent(&BTRFS_I(src)->io_tree, lock_start, lock_end);
+ } else {
+ unlock_extent(&BTRFS_I(src)->io_tree, off, off + len - 1);
+ unlock_extent(&BTRFS_I(inode)->io_tree, destoff,
+ destoff + len - 1);
+ }
+ /*
+ * Truncate page cache pages so that future reads will see the cloned
+ * data immediately and not the previous data.
+ */
+ truncate_inode_pages_range(&inode->i_data, destoff,
+ PAGE_CACHE_ALIGN(destoff + len) - 1);
out_unlock:
if (!same_inode) {
if (inode < src) {
return btrfs_ioctl_trans_end(file);
case BTRFS_IOC_TREE_SEARCH:
return btrfs_ioctl_tree_search(file, argp);
+ case BTRFS_IOC_TREE_SEARCH_V2:
+ return btrfs_ioctl_tree_search_v2(file, argp);
case BTRFS_IOC_INO_LOOKUP:
return btrfs_ioctl_ino_lookup(file, argp);
case BTRFS_IOC_INO_PATHS: