[Commotion-dev] OpenWRT AP not functioning

Josh King jking at chambana.net
Fri Jul 26 12:29:43 UTC 2013


Yes, we should definitely continue to go with wpad assuming we can 
identify a version that works reliably. Presumably, the space savings 
is why OpenWRT compiles hostapd and wpa_supplicant into the wpad 
multi-call binary in the first place. Will was building images with a 
new version pinning as well, as he suspected that the problem might be 
a mismatch between the nl80211 headers included in wpad vs. the version 
made available with the kernel. He'll send updates to the list.

As an aside, something we should look into doing at some point is 
rolling our own version of wpad-mini that has the IBSS-RSN compile 
flags added. The addition of those flags are the only reason we 
switched from wpad-mini to wpad, but wpad also contains the whole 
WPA-Enterprise stack and we could potentially save up to 400-600KB by 
stripping that out, from what I recall.

On Fri 26 Jul 2013 07:43:34 AM EDT, Dan Staples wrote:
> On Thu 25 Jul 2013 11:31:20 PM EDT, Ben West wrote:
>> I did notice approx 1.5 months ago that AA r36682+ did seem to cause
>> IBSS-RSN to stop working entirely, at least with nodes that had
>> packages hostapd and wpa_supplicant (but NOT wpad) installed.
>>
>> Right now, I have a collection of nodes running AA r36681 with JUST
>> wpad, any they appear to be doing IBSS-RSN adhoc just fine.
>>
>> Stability issues notwithstanding, is there a reason to go with the
>> full hostapd / wpa_supplicant packages?  I noticed you save
>> 150-200kBytes in firmware size using wpad instead.
>>
>> On Thu, Jul 25, 2013 at 8:59 PM, Dan Staples
>> <danstaples at opentechinstitute.org
>> <mailto:danstaples at opentechinstitute.org>> wrote:
>>
>>
>>     Update: turns out building w/ wpa_supplicant and hostapd as separate
>>     packages (instead of wpad which combines them into one binary) solves
>>     the problem. A newer version of wpad/hostapd/wpa_supplicant is
>>     available
>>     since we pinned AA to revisions from April, so Will and Jordan
>>     experimented with using the new version.
>>
>>     I just built with OpenWRT pinned @36682 and the packages feed @36537,
>>     and can confirm it fixes the problem, while including the newer
>>     version
>>     of wpad/hostapd/wpa_supplicant. IBSS-RSN works as well.
>>
>>     Thanks Will and Jordan for doing a lot of hard work investigating this
>>     today!
>>
>>     --
>>     Dan Staples
>>
>>     Open Technology Institute
>>     https://commotionwireless.net
>>     OpenPGP key: http://disman.tl/pgp.asc
>>     Fingerprint: 2480 095D 4B16 436F 35AB 7305 F670 74ED BD86 43A9
>>     _______________________________________________
>>     Commotion-dev mailing list
>>     Commotion-dev at lists.chambana.net
>>     <mailto:Commotion-dev at lists.chambana.net>
>>     https://lists.chambana.net/mailman/listinfo/commotion-dev
>>
>>
>>
>>
>> --
>> Ben West
>> http://gowasabi.net
>> ben at gowasabi.net <mailto:ben at gowasabi.net>
>> 314-246-9434
>
> 150-200k seems like a good enough reason to go with wpad instead of
> separate hostapd/wpa_supplicant, especially since we are tight on
> space. And since AP mode and IBSS-RSN both work w/ wpad at the
> revisions I tested, we might as well move our pinning up to that, don't
> you think?
>
> --
> Dan Staples
>
> Open Technology Institute
> https://commotionwireless.net
> OpenPGP key: http://disman.tl/pgp.asc
> Fingerprint: 2480 095D 4B16 436F 35AB 7305 F670 74ED BD86 43A9
> _______________________________________________
> Commotion-dev mailing list
> Commotion-dev at lists.chambana.net
> https://lists.chambana.net/mailman/listinfo/commotion-dev
>


More information about the Commotion-dev mailing list