From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <andrew@lunn.ch>
Date: Thu, 13 Feb 2014 10:36:50 +0100
From: Andrew Lunn <andrew@lunn.ch>
Message-ID: <20140213093650.GA7193@lunn.ch>
References: <1392122903-805-1-git-send-email-antonio@meshcoding.com>
 <1392122903-805-13-git-send-email-antonio@meshcoding.com>
 <20140212084022.GE30814@lunn.ch> <52FB6701.5000406@meshcoding.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <52FB6701.5000406@meshcoding.com>
Subject: Re: [B.A.T.M.A.N.] [RFC 12/23] batman-adv: add base throughput
 attribute
Reply-To: The list for a Better Approach To Mobile Ad-hoc Networking
 <b.a.t.m.a.n@lists.open-mesh.org>
List-Id: The list for a Better Approach To Mobile Ad-hoc Networking
 <b.a.t.m.a.n.lists.open-mesh.org>
List-Unsubscribe: <https://lists.open-mesh.org/mm/options/b.a.t.m.a.n>,
 <mailto:b.a.t.m.a.n-request@lists.open-mesh.org?subject=unsubscribe>
List-Archive: <http://lists.open-mesh.org/pipermail/b.a.t.m.a.n/>
List-Post: <mailto:b.a.t.m.a.n@lists.open-mesh.org>
List-Help: <mailto:b.a.t.m.a.n-request@lists.open-mesh.org?subject=help>
List-Subscribe: <https://lists.open-mesh.org/mm/listinfo/b.a.t.m.a.n>,
 <mailto:b.a.t.m.a.n-request@lists.open-mesh.org?subject=subscribe>
To: The list for a Better Approach To Mobile Ad-hoc Networking <b.a.t.m.a.n@lists.open-mesh.org>

On Wed, Feb 12, 2014 at 01:20:17PM +0100, Antonio Quartulli wrote:
> On 12/02/14 09:40, Andrew Lunn wrote:
> > On Tue, Feb 11, 2014 at 01:48:12PM +0100, Antonio Quartulli wrote:
> >> From: Antonio Quartulli <antonio@open-mesh.com>
> >>
> >> when batman-adv is asked to estimate/compute the throughput
> >> of an interface, but it fails for whatever reason, then the
> >> value in this attribute is used
> >>
> >> Signed-off-by: Antonio Quartulli <antonio@open-mesh.com>
> >> ---
> >>  bat_v.c | 3 +++
> >>  main.h  | 7 +++++++
> >>  sysfs.c | 7 +++++++
> >>  types.h | 1 +
> >>  4 files changed, 18 insertions(+)
> >>
> >> diff --git a/bat_v.c b/bat_v.c
> >> index d5457b3..3aa42ec 100644
> >> --- a/bat_v.c
> >> +++ b/bat_v.c
> >> @@ -75,6 +75,9 @@ static struct batadv_algo_ops batadv_batman_v __read_mostly = {
> >>  
> >>  int batadv_v_mesh_init(struct batadv_priv *bat_priv)
> >>  {
> >> +	atomic_set(&bat_priv->bat_v.base_throughput,
> >> +		   BATADV_DEFAULT_BASE_THROUGHPUT);
> >> +
> >>  	return batadv_v_ogm_init(bat_priv);
> >>  }
> >>  
> >> diff --git a/main.h b/main.h
> >> index 47fb9ed..5699c9f 100644
> >> --- a/main.h
> >> +++ b/main.h
> >> @@ -33,6 +33,13 @@
> >>  #define BATADV_MAX_METRIC 0xFFFFFFFF
> >>  #define BATADV_JITTER 20
> >>  
> >> +/**
> >> + * BATADV_DEFAULT_BASE_THROUGHPUT - default value used as throughput for
> >> + * hard_ifaces for which it is not possible to measure/estimate the real one.
> >> + * Value is expressed in Mbps/10
> > 
> > Mbps/10 is a rather odd unit. How about just Kbps?
> 
> This is the unit used in cfg/mac80211 because most of the rates are
> either in the form of XX.0Mbps of XX.YMbps (so just one decimal number
> needs to be saved).
> Moreover I thought that having an higher precision is not very useful
> because the routing decision cannot be taken on such a fine grained value.
> Actually the originator table prints the metric in Mbps (XX.YMbps).

It is not precision which i think is important, it is having a unit a
user expects. Networking people work in Kbps, or MBps. If it is
printed in the originator table as XX.YMBps, then i think this sysfs
file also needs to be XX.YMBps. That does however make your parsing
more complex, since it is no longer an int. I suppose there is the use
case of an ADSL link, with bandwidth of say 2Mbps down, 256Kbps up,
and then you do need the decimal point in XX.YMBps.

    Andrew