[Commotion-dev] other potential approaches to adhoc on Android 4.x

Thu Feb 28 20:21:39 UTC 2013

At this point, they know we are ready, willing, and able.  They just need to
figure out if and how they can use outside devs.

.hc

On 02/28/2013 02:11 PM, Josh King wrote:
> Hey Hans,
> 
> Thanks for forwarding that along, that's very interesting and way more
> positive than I expected. Do you know what the best avenue is to get
> involved with the effort to 'bring engineering muscle to the problem'?
> 
> On 02/28/2013 02:03 PM, Hans-Christoph Steiner wrote:
>>
>> I was checking in with the Google crew about official Adhoc support in
>> Android, and got this trove of interesting information, which Lucas said I
>> could forward:
>>
>> -------- Original Message --------
>> Subject: Re: Android Ad-Hoc
>> Date: Thu, 28 Feb 2013 09:24:30 -0800
>> From: Lucas Dixon <ldixon at google.com>
>>
>> Hi!
>> Sorry for the delay in replying. We've had some success at getting adhoc
>> wifi to work on Android 4.x. here's my notes...
>>
>> Note that there are several Motorola devices that do ship with working ad-
>> hoc, and their GUI includes some UX for ad hoc. e.g. Motorola Droid 4. In
>> the list ofWiFI networks, it lists ad hoc networks with a separate icon.
>> When clicking "Add Network", there is a tick box that states it is a "Peer
>> to peer (Ad-hoc) network.". In advanced settings there is an option to
>> specify the wifi channel for ad hoc networks. We're exploring proposing
>> this as UX for having ad-hoc wifi more generally supported.
>>
>> We've ben wondering about virtual interfaces and if they could be used to
>> allow Ad-Hoc WiFi at the same time as normal wifi-connections. That way a
>> WiFi hotspot can both act as a P2P node and as an uplink, both on WiFi.
>> Cool.
>>
>> Bringing IBSS support to Android (the long standing Issue
>> 82<https://code.google.com/p/android/issues/detail?id=82> and
>> some old hacks to work around it<http://szym.net/2010/12/adhoc-wifi-in-android/>
>> ):
>>
>> No changes need to the kernel are necessary for 4.X. It's just a matter of
>> how the Wifi service talks to wpa_supplicant. Szymon knows all the details.
>>
>> Some changes to drivers are sometimes needed, e.g.
>> https://android.googlesource.com/kernel/omap.git/+/96034c20069fe841f4be741755f94671a328702e
>> This is a change to bcmdhd in cfg80211 for our Galaxy Nexus devices, and
>> some other devices. Most Nexus devices use bcmdhd (but Nexus 4 is using
>> something else).
>>
>> So the change to Android framework is as basic as:
>> https://android-review.googlesource.com/#/c/43070/
>> (I think it will be necessary to add |frequency| parameter, because some
>> drivers refuse to form/join IBSS without that.)
>>
>> Some drivers explicitly disable IBSS, but that also can be often fixed with
>> a one line change in the driver module code.
>>
>> Szymon has been writing some benchmarks for AdHoc, WiFi Direct, and
>> Bluetooth. Hoping to have some more stats on that in the not too distant
>> future.
>>
>> I'll have a chat with the interested people and we'll try and make a plan
>> to leverage your ability to bring engineering muscle to the problem. :)
>>
>> Thanks!
>> lucas
>>
>>
>> On Thu, Feb 21, 2013 at 2:59 PM, Hans-Christoph Steiner <
>> hans at guardianproject.info> wrote:
>>
>>>
>>> Hey All,
>>>
>>> I guess wanted to check in on this issue, since it seems that the situation
>>> has gotten worse for adhoc in Android 4.x.  If dev resources are an issue,
>>> I'm
>>> sure some of us can find time to work on this.
>>>
>>> I still believe there is a ton of potential for mesh networking on
>>> Android, so
>>> I'd love to help make this happen!
>>>
>>> .hc
>>>
>>> On 08/02/2012 08:14 PM, Hans-Christoph Steiner wrote:
>>>>
>>>> Also, I just realized, it might work without a static IP if there is a
>>>> working DHCP server on the adhoc network.  For the work I'm doing now,
>>>> it requires a static IP, so I've only tested the adhoc with a static IP.
>>>>
>>>> As for how I did this, its part of the Commotion MeshTether app I'm
>>>> working on.  Starting at line 361 is the adhoc config stuff.  Szymon,
>>>> this code should look somewhat familiar to you since this app started as
>>>> a fork of Barnacle :)
>>>>
>>>>
>>> https://code.commotionwireless.net/projects/commotion-android/repository/revisions/master/entry/src/net/commotionwireless/meshtether/MeshService.java
>>>>
>>>> .hc
>>>>
>>>> On 08/02/2012 06:59 PM, Szymon Jakubczak wrote:
>>>>> Hi Hans-Christoph,
>>>>>
>>>>> That's some good news! Could you explain how you tested this? Does it
>>>>> seem to work on all phones with a particular version of Android?
>>>>>
>>>>> I've managed to find a Nexus S and build a working image. I'll try to
>>>>> combine the patch from havlenapetr and see if I can make it create
>>>>> ad-hoc networks.
>>>>>
>>>>> --Szymon
>>>>>
>>>>> On Thu, Aug 2, 2012 at 11:29 AM, Hans-Christoph Steiner
>>>>> <hans at guardianproject.info> wrote:
>>>>>>
>>>>>> Hey Lally,
>>>>>>
>>>>>> Glad to here of any progress! :)  Sounds good to me.  I've recently
>>>>>> discovered that Android already supports connecting to Adhoc networks
>>>>>> with certain conditions:
>>>>>>
>>>>>> 1) some other device has created the network
>>>>>> 2) you know the SSID and/or BSSID of the network
>>>>>> 3) you set a static IP
>>>>>>
>>>>>> Using WifiManager and System.Settings, I can associate to an existing
>>>>>> adhoc network and use it fine.  This is good to know, since it means
>>>>>> that we can do some mesh stuff with existing Android installs, but the
>>>>>> lack of ability to create the adhoc network definitely limits the
>>>>>> possibilities.
>>>>>>
>>>>>> .hc
>>>>>>
>>>>>> On 07/31/2012 04:29 PM, Lally Singh wrote:
>>>>>>> Hi everyone.
>>>>>>>
>>>>>>> Sorry for the huge delay.  So, at
>>>>>>> https://github.com/lally/adhoc-frameworks_base I've got a copy of
>>>>>>> frameworks_base, ready for us to hack on.  Additionally, I've got a
>>> PandaBoard
>>>>>>> ES <http://pandaboard.org/content/platform> to test on.  So, if it's
>>>>>>> acceptable for everyone, I'm happy to be a guinea pig/tester/worker
>>> elf for
>>>>>>> us to get the patches tested, running, cleaned up, and on their way to
>>>>>>> submission into android proper.
>>>>>>>
>>>>>>> So, while I figure out how to get the pandaboard configured over its
>>> serial
>>>>>>> link (srsly), shall we start putting the appropriate patches together?
>>>>>>>
>>>>>>> Also, if anyone objects to this way of doing things, or has a better
>>> idea,
>>>>>>> or suggestions, or comments, by all means, speak up!  I'm just trying
>>> to
>>>>>>> get the ball rolling, but I don't want to roll over anyone in the
>>> process.
>>>>>>>
>>>>>>> Cheers,
>>>>>>> $ ls
>>>>>>>
>>>>>>> Lally Singh | Software Engineer |  lally at google.com |  646-707-2716
>>>>>>>
>>>
>>
>>
>>
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 939 bytes
Desc: OpenPGP digital signature
URL: <http://lists.chambana.net/pipermail/commotion-dev/attachments/20130228/bc833326/attachment.sig>