[ath9k-devel] AP Limitation because of ATH9K_HTC_MAX_STA

[David Jardin <ich@djardin.de>]

Hi Adrian,

thanks for the response. That's bad news.

I thought about using the Pi's as they have much more RAM and CPU power
than normal consumer access points have - which seems to be the limiting
factor of clients on those devices. And as we are a non profit
organization we simply can't afford business class access points (Cisco
etc.) as they are way to expensive.

David

Am 08.07.13 22:07, schrieb Adrian Chadd:
> The ath9k_htc stuff isn't ready to do what you want. And even if it
> was, the performance would be sub-par compared to the SoC designs.
> There's just not a lot of RAM on these things; even if we do
> eventually push almost everything into the driver instead of firmware,
> there's still only a limited amount of RAM for packet buffers. So
> you're never going to reach the same level of performance as
> direct-attach.
>
> I suggest you spend the money on some slightly more useful access
> points. You'll likely be able to serve at least 2x the clients on a
> single 2x2 tplink or dlink access point for about $50-$70 each.
>
> Good luck,
>
>
> -adrian
> (Why does everyone see "Raspberry pi" as the solution to everything
> these days? Sheesh..)
>
>
> On 8 July 2013 01:35, David Jardin <ich@djardin.de> wrote:
>> Hi,
>>
>> I'm organizing all the technical stuff for several conferences about
>> Joomla (opensource content managment system) with 200-300 participants
>> and therefore I was looking for a low-budget wifi-solution that is able
>> to handle this number of clients. I decided to use a few raspberry pi's
>> with atheros-based usb wifi dongles which worked great in my little test
>> environment but crashed in the real conference setup.
>>
>> I took a look at the logfiles and it seems that the issue is related to
>> the number of clients that can be handled by the driver/firmware, which
>> is 8 (ATH9K_HTC_MAX_STA). Is there any chance to increase this limit to
>> make the dongles usable in large-scale applications?
>>
>> David
>> _______________________________________________
>> ath9k-devel mailing list
>> ath9k-devel at lists.ath9k.org
>> https://lists.ath9k.org/mailman/listinfo/ath9k-devel