[ath9k-devel] Does ath9k support beam forming?

[ath9k-devel] Does ath9k support beam forming?

[Ben Greear]

On 08/28/2012 10:05 PM, Mohammed Shafi wrote:
> Hi Ben,
>
> On Wed, Aug 29, 2012 at 12:29 AM, Ben Greear <greearb@candelatech.com> wrote:
>> I saw some RFC patches about beam-forming from 2010, but grepping current
>> code doesn't find anything that indicates these patches ever made it
>> upstream.
>> Of course, I could be looking for the wrong thing...
>
> i think even though the patches are ready for submission,
> I think the testing them proved to be very challenging, where there
> does not seems to be consistent improvement(if not decrease)
> in throughput.  I can be wrong.
>
>>
>> Does the current ath9k code support beam-forming?
>
> no. hopefully we can support them(yet we need some considerable
> amount of time to test  i suppose).

Can you post the latest patch series for this?

And maybe we can make the functionality configurable by debugfs
so that we can test it out easily?

Are there any particular test cases that would help verify the
patches?

Thanks,
Ben


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

[ath9k-devel] Does ath9k support beam forming?

[Ben Greear]

On 08/30/2012 10:28 PM, Mohammed Shafi wrote:
> Hi Ben,
>
> sorry for the delayed reply was bit involved in some other tasks.
>
>>
>>
>> Can you post the latest patch series for this?
>
> there are some commitments which i have to finish it off and i can start
> doing this only out of my free time.  Further it would take some good
> amount of time for testing.
> Unless its not pushed officially validating requires even more effort.
> i hope we can make this work into ath9k.
>
>>
>> And maybe we can make the functionality configurable by debugfs
>> so that we can test it out easily?
>
> yes sure.
>
>>
>> Are there any particular test cases that would help verify the
>> patches?
>
> throughput test cases with variation attenuation and antenna orientation ?
> definitely it should not cause any throughput regressions :)

I'll soon have the ability to script throughput tests v/s various
attenuations (using programmable attenuator that I'm building).

I can do over-the-air tests as well.  Probably 2-4 weeks until I
get everything automated, but when it is, I'll run some baseline
tests and then if you can post patches (RFC is fine if you are
not sure it's ready for upstream) then I can do some tests with
that as well and compare the results...

Thanks,
Ben


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com