[PATCH 1/2] fix: imsm: size must be in K for rounding to chunk

[PATCH 1/2] fix: imsm: size must be in K for rounding to chunk

[Czarnowska, Anna]

From 77d66de281e818f93e97b3150dc0e610634463de Mon Sep 17 00:00:00 2001
From: Anna Czarnowska <anna.czarnowska@intel.com>
Date: Tue, 8 Feb 2011 11:13:14 +0100
Subject: [PATCH 1/2] fix: imsm: size must be in K for rounding to chunk
Cc: linux-raid@vger.kernel.org, Williams, Dan J <dan.j.williams@intel.com>, Ciechanowski, Ed <ed.ciechanowski@intel.com>

chunk is in K so size must be converted to K before it is rounded.
Otherwise we may get wrong freesize returned
resulting in creation failure.

Signed-off-by: Anna Czarnowska <anna.czarnowska@intel.com>
---
 super-intel.c |    7 ++++---
 1 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/super-intel.c b/super-intel.c
index 6a21b26..5c16137 100644
--- a/super-intel.c
+++ b/super-intel.c
@@ -4379,7 +4379,8 @@ static int reserve_space(struct supertype *st, int raiddisks,
 	maxsize = merge_extents(super, extent_cnt);
 	minsize = size;
 	if (size == 0)
-		minsize = chunk;
+		/* chunk is in K */
+		minsize = chunk * 2;
 
 	if (cnt < raiddisks ||
 	    (super->orom && used && used != raiddisks) ||
@@ -4392,8 +4393,8 @@ static int reserve_space(struct supertype *st, int raiddisks,
 	if (size == 0) {
 		size = maxsize;
 		if (chunk) {
-			size /= chunk;
-			size *= chunk;
+			size /= 2 * chunk;
+			size *= 2 * chunk;
 		}
 	}
 
-- 
1.7.1

Re: [PATCH 1/2] fix: imsm: size must be in K for rounding to chunk

[NeilBrown]

On Thu, 10 Feb 2011 08:33:33 +0000 "Czarnowska, Anna"
<anna.czarnowska@intel.com> wrote:

> >From 77d66de281e818f93e97b3150dc0e610634463de Mon Sep 17 00:00:00 2001
> From: Anna Czarnowska <anna.czarnowska@intel.com>
> Date: Tue, 8 Feb 2011 11:13:14 +0100
> Subject: [PATCH 1/2] fix: imsm: size must be in K for rounding to chunk
> Cc: linux-raid@vger.kernel.org, Williams, Dan J <dan.j.williams@intel.com>, Ciechanowski, Ed <ed.ciechanowski@intel.com>
> 
> chunk is in K so size must be converted to K before it is rounded.
> Otherwise we may get wrong freesize returned
> resulting in creation failure.
> 
> Signed-off-by: Anna Czarnowska <anna.czarnowska@intel.com>

Thanks.  I've applied both patches in this series.

NeilBrown


> ---
>  super-intel.c |    7 ++++---
>  1 files changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/super-intel.c b/super-intel.c
> index 6a21b26..5c16137 100644
> --- a/super-intel.c
> +++ b/super-intel.c
> @@ -4379,7 +4379,8 @@ static int reserve_space(struct supertype *st, int raiddisks,
>  	maxsize = merge_extents(super, extent_cnt);
>  	minsize = size;
>  	if (size == 0)
> -		minsize = chunk;
> +		/* chunk is in K */
> +		minsize = chunk * 2;
>  
>  	if (cnt < raiddisks ||
>  	    (super->orom && used && used != raiddisks) ||
> @@ -4392,8 +4393,8 @@ static int reserve_space(struct supertype *st, int raiddisks,
>  	if (size == 0) {
>  		size = maxsize;
>  		if (chunk) {
> -			size /= chunk;
> -			size *= chunk;
> +			size /= 2 * chunk;
> +			size *= 2 * chunk;
>  		}
>  	}
>