[PATCH] Implement APuP Access Point Micro Peering

[Gio]

Hi!

I am happy to read from you, there have been progress about this patch 
indeed, I haven't reported here because I interpreted not receiving 
comments except for space/tab indentation as a lack of interest, but you 
bringing again this thread to life of course changed my mind

continuing in line


On 2024-12-30 10:48, Jouni Malinen wrote:
> On Fri, May 10, 2024 at 06:31:16PM +0200, gio at eigenlab.org wrote:
>> Access Point Micro Peering is a simpler and hopefully more useful successor to
>> Ad Hoc, Wireless Distribution System, 802.11s mesh mode, Multi-AP and EasyMesh.
>> When enabled almost plain APs communicate between them via 4-address mode,
>> like in WDS but all of them are AP, so they can eventually communicate also with
>> plain stations and more AP nodes in sight, without more trickery.
>> APuP has low hardware requirements, just AP mode support + 4-address mode, and
>> no more unnecessary complications, like hardcoded bridging or routing algorithm
>> in WiFi stack.
>> For each AP in sight an interface is created, and then it can be used as
>> convenient in each case, bridging, routing etc.
>> Those interfaces could be simply bridged in a trivial topology (which happens
>> automatically if wds_bridge is not an empty string), or feeded to a
>> routing daemon.
> What's the current state of this effort? This patch is clearly not ready
> to be included since it breaks existing functionality (e.g., hostapd
> crashing due to NULL pointer dereferencing in i802_set_wds_sta() due to
> ifname_wds == NULL with ap_wds_sta test case) and has TODO comments
> implying that this is not really complete.


As said before there have been progress, one of them fixing the crash 
you noticed too, you can take a look of current status here


https://gitlab.com/g10h4ck/hostap/-/commit/6bb15f81e6857989c0b722fc1a49275492114148 



>
> Is this mechanism defined somewhere? This seems to be adding new WDS
> STAs based on received Beacon frames

Yeah it is basically that, there is no formal specification yet, but a 
wall of text explaining what it is and how this idea came out


https://github.com/G10h4ck/lime-packages/tree/lime_curtigghio/packages/lime-curtigghio#readme


> without any kind of authentication
> or security which seems like a completely unrealistic deployment model
> due to how open it would be against various attacks.

Yeah no proper authentication or encryption implemented yet, not an 
immediate problem in my use-case.

We do use plain open AP which is included in hostapd since eons ;-)

But sure I am interested and have been participating in discussions on 
implementing authentication and encryption that make sense in this kind 
of setup, if you think implementing this might convince you to merge 
this into hostapd, I can give priority to this too. What kind of 
mechanisms (preferably reusing already existent code) would you suggest 
to explore first?

For now I was more interested into getting informed opinions from people 
(like you) that have much more experience in hostapd code then me.


Cheers

G10h4ck