From mboxrd@z Thu Jan  1 00:00:00 1970
From: Steven Whitehouse <swhiteho@redhat.com>
Date: Thu, 07 Jul 2011 14:57:30 +0100
Subject: [Cluster-devel] [PATCH 1/2] libgfs2: Added new RGSIZE macros
In-Reply-To: <1310046550-30496-1-git-send-email-cmaiolino@redhat.com>
References: <1310046550-30496-1-git-send-email-cmaiolino@redhat.com>
Message-ID: <1310047050.2730.13.camel@menhir>
List-Id: <cluster-devel.redhat.com>
To: cluster-devel.redhat.com
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit

Hi,

On Thu, 2011-07-07 at 10:49 -0300, Carlos Maiolino wrote:
> This patch add new macros related to
> default rgsizes which makes the purpose
> of the values easier to understand
> ---
>  gfs2/libgfs2/libgfs2.h |    5 +++++
>  1 files changed, 5 insertions(+), 0 deletions(-)
> 
> diff --git a/gfs2/libgfs2/libgfs2.h b/gfs2/libgfs2/libgfs2.h
> index be5bdfa..a0c3eea 100644
> --- a/gfs2/libgfs2/libgfs2.h
> +++ b/gfs2/libgfs2/libgfs2.h
> @@ -258,6 +258,11 @@ struct metapath {
>  #define GFS2_MIN_GROW_SIZE          (10)
>  #define GFS2_EXCESSIVE_RGS          (10000)
>  
> +#define GFS2_EXP_MIN_RGSIZE         (1)
> +#define GFS2_DEFAULT_MIN_RGSIZE     (32)
> +/* Look at this!  Why can't we go bigger than 2GB? */
> +#define GFS2_MAX_RGSIZE             (2048)
> +
>  #define DATA (1)
>  #define META (2)
>  #define DINODE (3)

The limit is due to 32 bit offsets used within the rgrp's data fields,
and there shouldn't be a need to go to a greater size than that really
since the intent is to divide the disk into a reasonably large number of
rgrps to reduce the likelihood of rgrp contention.

Anyway, looks good modulo my comment on the other patch,

Steve.