problem with iPhone and Samsung mobiles

Tamas Selmeci tselmeci
Wed Feb 18 22:47:53 PST 2009

Hi all!

Jouni Malinen wrote:

> iPod touch worked fine in my tests and so did Nokia N95, both of which
> are using power save mode.

Nokia N95 worked fine at us too, no problem experienced. As far as I 
remember we have also an iPod Touch here (therefore I don't understand 
why it haven't already been tested), I'm going to test it myself with ath9k.

The interesting thing is that by turning WPA/WPA2 encryption suite on 
(but strictly with only CCMP since TKIP is not supported by 802.11n) our 
other Samsung mobile (a Samsung Omnia), it has been able to work 
perfectly, without any problem. I would have thought that i780 and Omnia 
has somewhat similar wifi subsystem/chip/anything, but now the situation 
doesn't seem to be that. Perhaps i780 is more similar to Adam C?cile's 
Samsung device.

>> All the hostapd logging facilities are on (I'm watching the syslog). 
>> Have been tried with compat-ath9k-20090125 and compat-ath9k-20090211, 
>> nothing changed. Similarly, Atheros 5416 and 9160 behaved the same.
> Would you be able to test wireless-testing Git tree instead of
> compat-wireless? compat-wireless is known (well, by some, I did not
> remember this limitation until pointed out again) to not support power
> save buffering properly with kernels older than 2.6.29 (and yes, I know
> it has not been released yet). So your best bets might be to either take
> the latest 2.6.29-rc5 version of the Linux tree or test with
> wireless-testing.git as far as anything requiring power save buffering
> is concerned. The limitation in compat-wireless is not specific to
> ath9k; it applies to any driver using mac80211.

That can be a problem and thank you very much for the idea. 
Unfortunately at this time I'm too overloaded to make a try with 
2.6.29-rc5, since it would involve several changes in our environment 
(build system, patches, tuning all the components etc.), and now I'm 
under a rather high pressure to get our product into an as final state 
as possible to the Cebit exhibition. Luckily, I have an other colleague 
who could be asked to perform such tests, or at least prepare the 
environment to me thus saving a bunch of time.

I'm going to report the results of these tests - it can take up 2-3 weeks...

Best regards,
Tamas Selmeci
R&D Engineer

DENSION Audio Systems Ltd.
H-1116 Budapest, Sztregova u. 1.

More information about the Hostap mailing list