Hi Bishal,

On Thu, Dec 15, 2011 at 8:02 AM, Bishal Thapa <[email protected]> wrote:
> Hi Yeoh,
>  A revelation for me. I was quite sure before that each mesh interface had 
> its own mesh table. I had two interfaces "mesh" and "nesh" off of "wlan0" and 
> "wlan1" and frankly, I thought I saw different values populating "two" tables 
> when I did:
> 1. iw dev mesh mpath dump
> 2. iw dev nesh mpath dump

The two statements are not conflicting with each other.  1. There is
only path table per host (see mesh_paths in mesh_pathtbl.c).
Each mpath entry (see mesh_path in mesh.h) contains a
ieee80211_sub_if_data pointer.  And 2. when you are requesting an
mpath dump on an interface, you only get those entries whose
ieee80211_sub_if_data match your interface (see ieee80211_dump_mpath()
in cfg.c, and path_lookup() in mesh_pathtbl.c).

> Thank you for clarifying that Javier. To answer your question Yeoh, the code 
> as it stands, will create two interfaces with different IFACE names, and have 
> two MAC addresses. The "multi-radio" is happening at the Layer 2, not at 
> Layer-1 (like in 802.11n or any multi-antenna case). As far as I understand, 
> only Layer 3 and above will think there is only one "mesh node". At layer-2 
> there will be "two or more mesh nodes" (with two or more distinct MAC 
> addresses). Am I right, Javier?

I think you could implement multi-radio support by reusing the same
MAC, but I believe this would not work when/if multiple interfaces
were on the same channel.  My preference would be to keep one MAC
address per PHY.

> The fields that already exists when we do mpath dump
>
> DEST ADDR | NEXT HOP | IFACE | ....
>
> So, I believe the IFACE would say "mesh" or "nesh" based on what interface 
> you use for mpath dump.

It's just faster to try it rather than to share your doubts with the
list, don't you think so? :P

$ for i in {0..2};do iw mesh$i mpath dump; done
DEST ADDR         NEXT HOP          IFACE       SN      METRIC  QLEN    
EXPTIMEDTIM     DRET    FLAGS
42:00:00:00:02:00 42:00:00:00:01:00 mesh0       3       16386   0       
3243046752      100     0       0x4
42:00:00:00:01:00 42:00:00:00:01:00 mesh0       3       8193    0       
3243046752      200     1       0x4
DEST ADDR         NEXT HOP          IFACE       SN      METRIC  QLEN    
EXPTIMEDTIM     DRET    FLAGS
42:00:00:00:02:00 42:00:00:00:02:00 mesh1       3       8193    0       
3243046752      0       0       0x14
42:00:00:00:00:00 42:00:00:00:00:00 mesh1       6       8193    0       
3243046752      100     0       0x14
DEST ADDR         NEXT HOP          IFACE       SN      METRIC  QLEN    
EXPTIMEDTIM     DRET    FLAGS
42:00:00:00:01:00 42:00:00:00:01:00 mesh2       2       8193    0       
3243046752      0       0       0x14
42:00:00:00:00:00 42:00:00:00:01:00 mesh2       6       16386   0       
3243046752      100     0       0x4

Anyway, for multi-interface forwarding you would need an extra IFACE,
for instance:

DEST ADDR         IFACE NEXT HOP          IFACE SN      METRIC  ...
42:00:00:00:02:00 mesh1 42:00:00:00:01:00 mesh0 3       16386   ...

> Finally, here is my attempt to draw a layering diagram (Cheated off of 
> Javier's diagram from 802.11s paper)
>
>
> Layer 3
> --------
> 802.1d Bridge (if used)
> --------
> (mac80211) [Path Table, Neighbor Table, Mesh Kernel Services]
> (mac80211) [Multi-radio Mesh Connection : Mesh-1 Mesh-2 ...]
> --------
> PHY [Radio-1 interface, Radio-2 interface, ...]
>
> Comments?

Looks correct to me.

> Stefano,
>  I am certainly interested. Though we will have to coordinate pretty well and 
> pretty soon given (end of semester and) holidays are coming in no time. Lets 
> email each other off-of-the-list, and cc Javier and Yeoh.

If I'm to be copied, I would rather include the list.

Thanks!

Javier


> ----- Original Message -----
> From: "Yeoh Chun-Yeow" <[email protected]>
> To: "Javier Cardona" <[email protected]>
> Cc: [email protected], "Bishal Thapa" <[email protected]>
> Sent: Thursday, December 15, 2011 5:07:25 AM GMT -05:00 US/Canada Eastern
> Subject: Re: Question about joining meshes together
>
> Hi, Bishal and Javier,
>
>
>
> In terms of implementation, I just wish to know how are we going to 
> differentiate the MAC addresses for a particular Mesh Node with multiple 
> radio support. If let say Mesh Node A has two radio cards creating two mesh 
> node interfaces separately, it will have two MAC addresses. We should assume 
> only one MAC address occurs in our mesh forwarding table, am I right?
>
>
> A quick check on indicates batman node ( 
> http://www.open-mesh.org/wiki/batman-adv/Understand-your-batman-adv-network ) 
> only announcing one MAC address which is the batX interface.
>
>
> Please advice. Thanks
>
>
> Regards,
> Chun-Yeow
> _______________________________________________
> Devel mailing list
> [email protected]
> http://open80211s.com/mailman/listinfo/devel



-- 
Javier Cardona
cozybit Inc.
http://www.cozybit.com
_______________________________________________
Devel mailing list
[email protected]
http://open80211s.com/mailman/listinfo/devel

Reply via email to