[Commotion-dev] Boilerplate Android code pushed

Hans-Christoph Steiner hans at at.or.at
Tue Feb 7 03:17:50 UTC 2012 

I forgot to mention: neither /proc/calibration nor /data/calibration exist now on this Droid-CM7.1. I don't know if either one was there before, but that file location does seem to be required.

.hc

On Feb 6, 2012, at 10:16 PM, Hans-Christoph Steiner wrote:

> 
> I've tried now wiping the cache and doing a whole system update, and still no luck.  I haven't tried wiping the data yet.
> 
> - after a reboot, tiwlan0_drv is not loaded
> - insmod /system/lib/modules/tiwlan_drv.ko works fine
> - after insmod, "Wireless & network settings" says: Wi-Fi Error
> - /system/bin/wlan_loader -f /system/etc/wifi/fw_wlan1271.bin -e /proc/calibration -i /system/etc/wifi/tiwlan.ini gives me:
>   Set property wlan.driver.status = ok - Fail
> 
> If you want, I can hook this phone up to a computer where I can give you a shell, and you can hunt around. Its even got a Debian chroot on it, you can get a full Debian shell by running /data/local/bin/debian, thanks to LilDebi.apk.  :)
> 
> .hc
> 
> On Feb 2, 2012, at 8:26 PM, Jeremy Lakeman wrote:
> 
>> We've got a lot of cheap Huewai IDEOS phones, CPU and flash is quite
>> slow and they have ridiculously tiny screens.
>> 
>> Samsung galaxy's work quite well, at least with the stock firmware.
>> I've also seen them work with CM, so I assume those roms have the same
>> driver files in the same locations so we aren't hitting the same
>> issue.
>> 
>> If we can disable our "generic" profile for CM phones, we could work
>> towards supporting the built in modified wpa supplicant. We shouldn't
>> actually need to load the wifi driver ourselves since Cyanogen
>> supports joining adhoc networks anyway. Not sure if we can initiate a
>> network though.
>> 
>> I wonder if your phone is having trouble loading the (at a guess)
>> tiwlan_drv.ko module.
>> 
>> On stock Droids we do something like;
>> insmod /system/lib/modules/tiwlan_drv.ko
>> /system/bin/wlan_loader -f /system/etc/wifi/fw_wlan1271.bin -e
>> /proc/calibration -i [path]/tiwlan.ini
>> Then use ifconfig / iwconfig to get the network running
>> 
>> If you have firmware binaries in different places we wouldn't use this
>> script and would fall back to trying "generic"
>> 
>> On Fri, Feb 3, 2012 at 11:25 AM, Hans-Christoph Steiner
>> <hans at guardianproject.info> wrote:
>>> 
>>> I poked around a little bit, my /data/misc/wifi/wpa_supplicant.conf file
>>> seems intact.  "netcfg tiwlan0 up" and down no longer work, it tells me:
>>> 
>>> # /system/bin/netcfg tiwlan0 up
>>> action 'up' failed (No such device)
>>> 
>>> I rebooted and cleared the cache, no luck there. I'll try flashing CM7.1
>>> again.  Do you have a phone you recommend for wifi hacking with Android?
>>> We have a good collection, so we probably have a better one than a
>>> Droid/Milestone running CM7.1.
>>> 
>>> .hc
>>> 
>>> 
>>> On 02/02/2012 04:58 PM, Jeremy Lakeman wrote:
>>>> Ah, we've had some strange reports like that from CM7 users.
>>>> No we don't write over any files in the system folder. At least not directly.
>>>> We'll probably end up trying to use our "generic" chipset profile
>>>> which will try to call /system/lib/libhardware_legacy.so
>>>> wifi_load_driver() / wifi_unload_driver().
>>>> CM7 seems to include this library, but using it might be the cause of
>>>> this issue. We haven't flashed CM7 onto any of our phones yet to test
>>>> what the root cause is.
>>>> You may need to boot into recovery mode and clear the cache, or worst
>>>> case do a factory reset.
>>>> 
>>>> On Fri, Feb 3, 2012 at 3:48 AM, Hans-Christoph Steiner
>>>> <hans at guardianproject.info> wrote:
>>>> 
>>>> I got it running on my Droid with CM7.1, but then I couldn't make it
>>>> stop running ;)  Every time I turned on my wifi since setting up
>>>> Serval, the Serval app would then show up and turn it back off.  I
>>>> first clicked the "Suspend" button, and it turned to "Start".  I then
>>>> clicked "Reset".  I then force closed it, and cleared the data.
>>>> 
>>>> Now I've removed it entirely, and my wifi isn't working any more.  It
>>>> turns on, but never scans for any networks and it seems to have
>>>> cleared all the saved networks in my phone.  I cannot manually add
>>>> networks either.
>>>> 
>>>> Does this app mess with the wpa_supplicant.conf that comes with the
>>>> phone, or any of the standard Android files?
>>>> 
>>>> .hc
>>>> 
>>>> On 01/10/2012 10:08 PM, Jeremy Lakeman wrote:
>>>>>>> If that process completely locks up (and it can) you may need to
>>>>>>> force close the application or reboot. After that it will disable
>>>>>>> running the same process for turning on adhoc mode.
>>>>>>> 
>>>>>>> Let me know if you manage to get past that step, or you are still
>>>>>>> having issues.
>>>>>>> 
>>>>>>> On Wed, Jan 11, 2012 at 1:14 PM, Hans-Christoph Steiner
>>>>>>> <hans at at.or.at> wrote:
>>>>>>>> 
>>>>>>>> A little bug report, in case you're interested:
>>>>>>>> 
>>>>>>>> I just downloaded Serval Mesh from the Marketplace on my Droid
>>>>>>>> running CM7.1.  It gets thru "Unpacking the stuff I need",
>>>>>>>> "Checking for root access", and "I think I know how to control
>>>>>>>> your wifi", then says "Skipped check for experimental support
>>>>>>>> since we already support your handset.", and gets stuck on "Test
>>>>>>>> if WiFi control works".  It just seems to try again and again.
>>>>>>>> 
>>>>>>>> .hc
>>>>>>>> 
>>>>>>>> On Jan 10, 2012, at 6:54 PM, Jeremy Lakeman wrote:
>>>>>>>> 
>>>>>>>>> Serval's software is available here
>>>>>>>>> https://github.com/servalproject/batphone The specific pieces
>>>>>>>>> you might want to look at are;
>>>>>>>>> 
>>>>>>>>> data/conf/wifichipsets/
>>>>>>>>> src/org/servalproject/system/ChipsetDetection.java
>>>>>>>>> src/org/servalproject/system/WiFiRadio.java
>>>>>>>>> 
>>>>>>>>> and the edify scripts are interpreted by (including the call
>>>>>>>>> to libhardware_legacy wifi_load_driver / wifi_unload_driver);
>>>>>>>>> native/adhoc-edify/
>>>>>>>>> 
>>>>>>>>> (The data directory is zipped in the build process, and
>>>>>>>>> unzipped by our first run process into
>>>>>>>>> /data/data/org.servalproject)
>>>>>>>>> 
>>>>>>>>> And the apk is available on the market;
>>>>>>>>> https://market.android.com/details?id=org.servalproject
>>>>>>>>> 
>>>>>>>>> This approach works with a fair number of stock (but rooted)
>>>>>>>>> handsets. We haven't gotten around to trying to using wifi
>>>>>>>>> supplicants that already support adhoc mode.
>>>>>>>>> 
>>>>>>>>> On Wed, Jan 11, 2012 at 9:45 AM, Hans-Christoph Steiner
>>>>>>>>> <hans at at.or.at> wrote:
>>>>>>>>>> 
>>>>>>>>>> Wow, that sounds quite elaborate, I'd like to see it in
>>>>>>>>>> action :)  Can you point us towards the code that is doing
>>>>>>>>>> all this?
>>>>>>>>>> 
>>>>>>>>>> My guess is that the level of detail will be needed to handle
>>>>>>>>>> a lot of cases, but I think we have a good chance of
>>>>>>>>>> simplifying things in the long run by including our own
>>>>>>>>>> binaries in the app for things that are not kernel mods.
>>>>>>>>>> 
>>>>>>>>>> I'd love to find a way to include kernel mods that work
>>>>>>>>>> across kernel versions, but I have not found a way to do
>>>>>>>>>> that.  It must be possible, since people like nvidia are
>>>>>>>>>> shipping binary kernel mods that work across a number of
>>>>>>>>>> kernel versions.
>>>>>>>>>> 
>>>>>>>>>> .hc
>>>>>>>>>> 
>>>>>>>>>> On Jan 10, 2012, at 5:28 PM, Jeremy Lakeman wrote:
>>>>>>>>>> 
>>>>>>>>>>> Getting Adhoc and olsrd running is something the Serval
>>>>>>>>>>> Mesh software is already trying to do. We default to BATMAN
>>>>>>>>>>> but we have a setting to run olsr instead.
>>>>>>>>>>> 
>>>>>>>>>>> We started with wifi tether's adhoc support and have tried
>>>>>>>>>>> to improve on it. We've written a generic approach to
>>>>>>>>>>> discovering how to control the wifi driver on each handset.
>>>>>>>>>>> For the handsets that need something different, we've got a
>>>>>>>>>>> bunch of text files describing file names we should be
>>>>>>>>>>> looking for, and a matching control script.
>>>>>>>>>>> 
>>>>>>>>>>> But if we don't have a specific script that matches this
>>>>>>>>>>> handset, we look through the filesystem looking for *.ko.
>>>>>>>>>>> Then read through every short script in /system looking for
>>>>>>>>>>> insmod commands and any arguments that we need to pass
>>>>>>>>>>> through. And we've got a fallback method that tries a load
>>>>>>>>>>> driver library call that may or may not work.
>>>>>>>>>>> 
>>>>>>>>>>> Hopefully one of these approaches will succeed.
>>>>>>>>>>> 
>>>>>>>>>>> 
>>>>>>>>>>> 
>>>>>>>>>>> 
>>>>>>>>>>> On Wed, Jan 11, 2012 at 7:11 AM, Hans-Christoph Steiner
>>>>>>>>>>> <hans at at.or.at> wrote:
>>>>>>>>>>>> 
>>>>>>>>>>>> Hey Brian,
>>>>>>>>>>>> 
>>>>>>>>>>>> I briefly checked out your code, I'll dig in more soon.
>>>>>>>>>>>> Did you get ad-hoc mode running after all?  I assume so
>>>>>>>>>>>> since you said you got olsrd running.  Apparently, on
>>>>>>>>>>>> some devices, you have to replace the wpa_supplicant
>>>>>>>>>>>> binary in order to get ad-hoc mode working:
>>>>>>>>>>>> http://forum.xda-developers.com/showthread.php?t=929865
>>>>>>>>>>>> 
>>>>>>>>>>>> Cyanogenmod seems to include everything needed to run
>>>>>>>>>>>> ad-hoc though.
>>>>>>>>>>>> 
>>>>>>>>>>>> I just found this, which looks like it should provide us
>>>>>>>>>>>> with some useful code:
>>>>>>>>>>>> http://code.google.com/p/adhoc-on-android/
>>>>>>>>>>>> 
>>>>>>>>>>>> On approach that might work for us is to include our own
>>>>>>>>>>>> custom wpa_supplicant in the app.
>>>>>>>>>>>> 
>>>>>>>>>>>> .hc
>>>>>>>>>>>> 
>>>>>>>>>>>> On Jan 10, 2012, at 2:47 PM, Brian Duggan wrote:
>>>>>>>>>>>> 
>>>>>>> Hi Commoters,
>>>>>>> 
>>>>>>> I've been doing some preliminary work on getting olsrd to work in
>>>>>>> a user-friendly way on Android. The first challenge is creating an
>>>>>>> ad-hoc interface. Google doesn't support this, and doesn't look
>>>>>>> like they will support ad-hoc because they feel WiFi Direct is
>>>>>>> functions as a replacement. See bug 82:
>>>>>>> 
>>>>>>> http://code.google.com/p/android/issues/detail?id=82
>>>>>>> 
>>>>>>> I pushed some code to the redux branch of the commotion-android
>>>>>>> repository. It's half code, half comments describing what will need
>>>>>>> to be done to automate the entire process of creating an ad-hoc
>>>>>>> interface the Android Way.
>>>>>>> 
>>>>>>> It doesn't address current olsrd incompatibility with some Android
>>>>>>> wifi drivers, namely the dhd driver. To date, I've only been able
>>>>>>> to get olsrd to run stably on devices using the bcm4325 driver.
>>>>>>> 
>>>>>>> Check it out, if you're interested:
>>>>>>> 
>>>>>>> git clone -b redux
>>>>>>> git://git.chambana.net/commotion/commotion-android
>>>>>>> 
>>>>>>> The master branch is still an Android Wifi Tether overlay, but will
>>>>>>> get olsrd up and running on bcm4325 (and other non-dhd?) devices.
>>>>>>> Some testing would be appreciated here. It would be great to start
>>>>>>> probing which devices olsrd works on and which don't.
>>>>>>> 
>>>>>>> Thanks! Brian
>>>>>>> 
>>>>>>>>>>>>> _______________________________________________
>>>>>>>>>>>>> Commotion-dev mailing list
>>>>>>>>>>>>> Commotion-dev at lists.chambana.net
>>>>>>>>>>>>> http://lists.chambana.net/mailman/listinfo/commotion-dev
>>>>>>>>>>>> 
>>>>>>>>>>>> 
>>>>>>>>>>>> 
>>>>>>>>>>>> 
>>>>>>>>>>>>> 
>>>> -
>>>> ----------------------------------------------------------------------------
>>>>>>>>>>>> 
>>>>>>>>>>>> Man has survived hitherto because he was too ignorant to
>>>>>>>>>>>> know how to realize his wishes.  Now that he can realize
>>>>>>>>>>>> them, he must either change them, or perish.    -William
>>>>>>>>>>>> Carlos Williams
>>>>>>>>>>>> 
>>>>>>>>>>>> 
>>>>>>>>>>>> _______________________________________________
>>>>>>>>>>>> Commotion-dev mailing list
>>>>>>>>>>>> Commotion-dev at lists.chambana.net
>>>>>>>>>>>> http://lists.chambana.net/mailman/listinfo/commotion-dev
>>>>>>>>>>> _______________________________________________
>>>>>>>>>>> Commotion-dev mailing list
>>>>>>>>>>> Commotion-dev at lists.chambana.net
>>>>>>>>>>> http://lists.chambana.net/mailman/listinfo/commotion-dev
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>> ----------------------------------------------------------------------------
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>>> 
>>>> There is no way to peace, peace is the way.       -A.J. Muste
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>> _______________________________________________ Commotion-dev
>>>>>>>>> mailing list Commotion-dev at lists.chambana.net
>>>>>>>>> http://lists.chambana.net/mailman/listinfo/commotion-dev
>>>>>>>> 
>>>>>>>> 
>>>>>>>> 
>>>>>>>> 
>>>>>>>> ----------------------------------------------------------------------------
>>>>>>>> 
>>>>>>>> 
>>>>>>>> 
>>>> I have always wished for my computer to be as easy to use as my
>>>> telephone; my wish has come true because I can no longer figure out how
>>>> to use my telephone."  --Bjarne Stroustrup (creator of C++)
>>>>>>>> 
> 
> 
> 
> ----------------------------------------------------------------------------
> 
> I spent 33 years and four months in active military service and during that period I spent most of my time as a high class muscle man for Big Business, for Wall Street and the bankers.      - General Smedley Butler
> 



----------------------------------------------------------------------------

Man has survived hitherto because he was too ignorant to know how to realize his wishes.  Now that he can realize them, he must either change them, or perish.    -William Carlos Williams

More information about the Commotion-dev mailing list