[RFC / PATCH] mtd/ubi: flush wl before clearing update marker

[Artem Bityutskiy]

On Sun, 2009-11-29 at 19:46 +0100, Sebastian Andrzej Siewior wrote:
> ubiupdatevol -t does the following:
> - ubi_start_update()
>   - set_update_marker()
>   - for all LEBs ubi_eba_unmap_leb()
>   - clear_update_marker()
>   - ubi_wl_flush()
> 
> ubi_wl_flush() physically erases all PEB, once it returns all PEBs are
> empty. clear_update_marker() has the update marker written after return.
> If there is a power cut between the last two functions then the UBI
> volume has no longer the "update" marker set and may have some valid
> LEBs while some of them may be gone.
> If that volume in question happens to be a UBIFS volume, then mount
> will fail with

Very good catch, thanks. Pushed to the ubi-2.6.git tree, although I've

[snip]

>  	if (bytes == 0) {
> +
Removed this junk newline
> +		err = ubi_wl_flush(ubi);
> +		if (err)
> +			return err;

[snip]

>  	vol->upd_buf = vmalloc(ubi->leb_size);
> @@ -362,16 +365,17 @@ int ubi_more_update_data(struct ubi_device *ubi, struct ubi_volume *vol,
>  
>  	ubi_assert(vol->upd_received <= vol->upd_bytes);
>  	if (vol->upd_received == vol->upd_bytes) {
> +
And this.
> +		err = ubi_wl_flush(ubi);
> +		if (err)
> +			return err;

[snip]

-- 
Best Regards,
Artem Bityutskiy (Артём Битюцкий)