On Mon, 27 Oct 2008 21:58:54 +0900 Takashi Sato <t-sato@yk.j...> wrote:

> -void thaw_bdev(struct block_device *bdev, struct super_block *sb)
> +int thaw_bdev(struct block_device *bdev, struct super_block *sb)
>  {
> + int error = 0;
> +
> + mutex_lock(&bdev->bd_fsfreeze_mutex);
> + if (!bdev->bd_fsfreeze_count) {
> + mutex_unlock(&bdev->bd_fsfreeze_mutex);
> + return -EINVAL;

This would be a programming error, yes?

If so, a WARN_ON is more appropriate than a silent runtime error.

> + }