Re: ENOSPC error returned too late

[Theodore Tso <tytso@mit.edu>]

On Mon, Aug 18, 2008 at 04:21:40PM +0530, Aneesh Kumar K.V wrote:
> --- a/fs/ext4/balloc.c
> +++ b/fs/ext4/balloc.c
> @@ -1628,6 +1628,9 @@ ext4_fsblk_t ext4_has_free_blocks(struct ext4_sb_info *sbi,
>  		free_blocks =
>  			percpu_counter_sum_and_set(&sbi->s_freeblocks_counter);
>  #endif
> +	if (free_blocks <= root_blocks)
> +		/* we don't have free space */
> +		return 0;
>  	if (free_blocks - root_blocks < nblocks)
>  		return free_blocks - root_blocks;
>  	return nblocks;

Thanks, I'll try it out.  Given that ext4_fsblk_t is an unsigned long
(BTW, this is one of the reasons why the common coding style guide
dislikes typedefs; it hides this kind of errors), it's obviously a
good patch to add regardless of whether it solves are problem, since
if freeblocks < root_blocks, the following condition is going to do
the Wrong Thing.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>

						- Ted