btrfs: Change the hole range to a more accurate value.

Commit 3ac0d7b fixed the btrfs expanding
write problem but the hole punched is sometimes too large for some
iovec, which has unmapped data ranges.
This patch will change to hole range to a more accurate value using the
counts checked by the write check routines.

Reported-by: Al Viro <[email protected]>
Signed-off-by: Qu Wenruo <[email protected]>
Signed-off-by: Chris Mason <[email protected]>

Author: Qu Wenruo

fs/btrfs/file.c

@@ -1783,7 +1783,7 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb,
 	start_pos = round_down(pos, root->sectorsize);
 	if (start_pos > i_size_read(inode)) {
 		/* Expand hole size to cover write data, preventing empty gap */
-		end_pos = round_up(pos + iov->iov_len, root->sectorsize);
+		end_pos = round_up(pos + count, root->sectorsize);
 		err = btrfs_cont_expand(inode, i_size_read(inode), end_pos);
 		if (err) {
 			mutex_unlock(&inode->i_mutex);