Re: [PATCH] Btrfs: add missing inode item update in fallocate()

[Liu Bo <bo.li.liu@oracle.com>]

On Thu, Mar 12, 2015 at 03:36:58PM +0000, Filipe Manana wrote:
> If we fallocate(), without the keep size flag, into an area already covered
> by an extent previously fallocated, we were updating the inode's i_size but
> we weren't updating the inode item in the fs/subvol tree. A following umount
> + mount would result in a loss of the inode's size (and an fsync would miss
> too the fact that the inode changed).
> 
> Reproducer:
> 
>   $ mkfs.btrfs -f /dev/sdd
>   $ mount /dev/sdd /mnt
>   $ fallocate -n -l 1M /mnt/foobar
>   $ fallocate -l 512K /mnt/foobar
>   $ umount /mnt
>   $ mount /dev/sdd /mnt
>   $ od -t x1 /mnt/foobar
>   0000000
> 
> The expected result is:
> 
>   $ od -t x1 /mnt/foobar
>   0000000 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>   *
>   2000000
> 
> A test case for fstests follows soon.
> 
> Signed-off-by: Filipe Manana <fdmanana@suse.com>
> ---
>  fs/btrfs/file.c | 21 ++++++++++++++++-----
>  1 file changed, 16 insertions(+), 5 deletions(-)
> 
> diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
> index 8f256b1..fb4bd79 100644
> --- a/fs/btrfs/file.c
> +++ b/fs/btrfs/file.c
> @@ -2677,13 +2677,10 @@ static long btrfs_fallocate(struct file *file, int mode,
>  							1 << inode->i_blkbits,
>  							offset + len,
>  							&alloc_hint);
> -
> -			if (ret < 0) {
> -				free_extent_map(em);
> -				break;
> -			}
>  		} else if (actual_end > inode->i_size &&
>  			   !(mode & FALLOC_FL_KEEP_SIZE)) {
> +			struct btrfs_trans_handle *trans;
> +
>  			/*
>  			 * We didn't need to allocate any more space, but we
>  			 * still extended the size of the file so we need to
> @@ -2692,8 +2689,22 @@ static long btrfs_fallocate(struct file *file, int mode,
>  			inode->i_ctime = CURRENT_TIME;
>  			i_size_write(inode, actual_end);
>  			btrfs_ordered_update_i_size(inode, actual_end, NULL);
> +			/* 1 unit for inode item */
> +			trans = btrfs_start_transaction(root, 1);
> +			if (IS_ERR(trans)) {
> +				ret = PTR_ERR(trans);

I prefer putting this earlier, before i_size updates, to
protect us from another isize inconsistence(disk_i_size, isize vs isize in inode_item).

Basically I mean,
			...
			trans = btrfs_start_transaction(root, 1);
			if (IS_ERR(trans)) {
				ret = PTR_ERR(trans);
			} else {
	  			inode->i_ctime = CURRENT_TIME;
				...
			}

> +			} else {
> +				ret = btrfs_update_inode(trans, root, inode);
> +				if (ret)
> +					btrfs_end_transaction(trans, root);
> +				else
> +					ret = btrfs_end_transaction(trans,
> +								    root);

calling same end_transaction() for two times seems kind of weird for me,
what about this?

ret2 = btrfs_end_transaction(trans, root);
if (!ret)
	ret = ret2;

But anyway, I'm okay with both.

Thanks,

-liubo

>  		}
>  		free_extent_map(em);
> +		if (ret < 0)
> +			break;
>  
>  		cur_offset = last_byte;
>  		if (cur_offset >= alloc_end) {
> -- 
> 2.1.3
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html