Re: [PATCH 1/1 linux-next] ext4: atomically read inode size

[Theodore Ts'o]

On Mon, Jan 23, 2017 at 06:56:09PM +0100, Fabian Frederick wrote:
> See i_size_read() comments in include/linux/fs.h
> 
> Signed-off-by: Fabian Frederick <fabf@xxxxxxxxx>
> ---
>  fs/ext4/super.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/ext4/super.c b/fs/ext4/super.c
> index 9d15a62..3746f77 100644
> --- a/fs/ext4/super.c
> +++ b/fs/ext4/super.c
> @@ -3798,7 +3798,8 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
>  		goto cantfind_ext4;
>  
>  	/* check blocks count against device size */
> -	blocks_count = sb->s_bdev->bd_inode->i_size >> sb->s_blocksize_bits;
> +	blocks_count = i_size_read(sb->s_bdev->bd_inode) >>
> +		       sb->s_blocksize_bits;
>  	if (blocks_count && ext4_blocks_count(es) > blocks_count) {
>  		ext4_msg(sb, KERN_WARNING, "bad geometry: block count %llu "
>  		       "exceeds size of device (%llu blocks)",

Um, this is at mount time, and we're talking about the i_size of the
block device --- which typically isn't moving around a huge amount.

      	     	       		       	      - Ted