Re: [PATCH 5/6] UBI: Split __wl_get_peb()

[Guido Martínez]

Hi Richard,

On Mon, Nov 24, 2014 at 02:20:35PM +0100, Richard Weinberger wrote:
> Make it two functions, wl_get_wle() and wl_get_peb().
> wl_get_peb() works exactly like __wl_get_peb() but wl_get_wle()
> does not call produce_free_peb().
> While refilling the fastmap user pool we cannot release ubi->wl_lock
> as produce_free_peb() does.
> Hence the fastmap logic uses now wl_get_wle().
> 
> Signed-off-by: Richard Weinberger <richard@xxxxxx>
> ---
>  drivers/mtd/ubi/wl.c | 61 ++++++++++++++++++++++++++++++++--------------------
>  1 file changed, 38 insertions(+), 23 deletions(-)
> 
> diff --git a/drivers/mtd/ubi/wl.c b/drivers/mtd/ubi/wl.c
> index 7730b97..f028b68 100644
> --- a/drivers/mtd/ubi/wl.c
> +++ b/drivers/mtd/ubi/wl.c
> @@ -499,13 +499,46 @@ out:
>  #endif
>  
>  /**
> - * __wl_get_peb - get a physical eraseblock.
> + * wl_get_wle - get a mean wl entry to be used by wl_get_peb() or
> + * refill_wl_user_pool().
> + * @ubi: UBI device description object
> + *
> + * This function returns a a wear leveling entry in case of success and
> + * NULL in case of failure.
> + */
> +static struct ubi_wl_entry *wl_get_wle(struct ubi_device *ubi)
> +{
> +	struct ubi_wl_entry *e;
> +
> +	e = find_mean_wl_entry(ubi, &ubi->free);
> +	if (!e) {
> +		ubi_err(ubi, "no free eraseblocks");
> +		return NULL;
> +	}
> +
> +	self_check_in_wl_tree(ubi, e, &ubi->free);
> +
> +	/*
> +	 * Move the physical eraseblock to the protection queue where it will
> +	 * be protected from being moved for some time.
> +	 */
> +	rb_erase(&e->u.rb, &ubi->free);
> +	ubi->free_count--;
> +	dbg_wl("PEB %d EC %d", e->pnum, e->ec);
> +
> +	return e;
> +}
> +
> +/**
> + * wl_get_peb - get a physical eraseblock.
>   * @ubi: UBI device description object
>   *
>   * This function returns a physical eraseblock in case of success and a
>   * negative error code in case of failure.
> + * It is the low level component of ubi_wl_get_peb() in the non-fastmap
> + * case.
>   */
> -static int __wl_get_peb(struct ubi_device *ubi)
> +static int wl_get_peb(struct ubi_device *ubi)
This bit breaks the build since (at this point) the usage of
__wl_get_peb on refill_wl_user_pool is not changed. There's also a
comment mentioning __wl_get_peb, which should be updated.

I can see both changes in the next patch, but it's way better if we
don't break the build :).

Besides from these two and the comments by Tanya, the patch looks good.

>  {
>  	int err;
>  	struct ubi_wl_entry *e;
> @@ -524,27 +557,9 @@ retry:
>  		goto retry;
>  	}
>  
> -	e = find_mean_wl_entry(ubi, &ubi->free);
> -	if (!e) {
> -		ubi_err(ubi, "no free eraseblocks");
> -		return -ENOSPC;
> -	}
> -
> -	self_check_in_wl_tree(ubi, e, &ubi->free);
> -
> -	/*
> -	 * Move the physical eraseblock to the protection queue where it will
> -	 * be protected from being moved for some time.
> -	 */
> -	rb_erase(&e->u.rb, &ubi->free);
> -	ubi->free_count--;
> -	dbg_wl("PEB %d EC %d", e->pnum, e->ec);
> -#ifndef CONFIG_MTD_UBI_FASTMAP
> -	/* We have to enqueue e only if fastmap is disabled,
> -	 * is fastmap enabled prot_queue_add() will be called by
> -	 * ubi_wl_get_peb() after removing e from the pool. */
> +	e = wl_get_wle(ubi);
>  	prot_queue_add(ubi, e);
> -#endif
> +
>  	return e->pnum;
>  }
>  
> @@ -704,7 +719,7 @@ int ubi_wl_get_peb(struct ubi_device *ubi)
>  	int peb, err;
>  
>  	spin_lock(&ubi->wl_lock);
> -	peb = __wl_get_peb(ubi);
> +	peb = wl_get_peb(ubi);
>  	spin_unlock(&ubi->wl_lock);
>  
>  	if (peb < 0)

-- 
Guido Martínez, VanguardiaSur
www.vanguardiasur.com.ar
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/