Subject: Re: Using ath5k/ath9k radio for constant-tx noise source.
To: Zefir Kurtisi <zefir.kurtisi@neratec.com>,
        "linux-wireless@vger.kernel.org" <linux-wireless@vger.kernel.org>
References: <55D4D406.5000603@candelatech.com> <55D5EE0E.3020408@neratec.com>
 <6d50965d-af35-dad9-23fe-162678a0f02f@candelatech.com>
 <27058308-3d36-b415-900f-3e3b65b11dab@neratec.com>
Cc: Christian Lamparter <chunkeey@googlemail.com>
From: Ben Greear <greearb@candelatech.com>
Message-ID: <126bb617-0de3-9366-513d-d69b0cbe5238@candelatech.com> (sfid-20160915_175928_388674_1661FC06)
Date: Thu, 15 Sep 2016 08:59:15 -0700
MIME-Version: 1.0
In-Reply-To: <27058308-3d36-b415-900f-3e3b65b11dab@neratec.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Sender: linux-wireless-owner@vger.kernel.org

On 09/15/2016 08:28 AM, Zefir Kurtisi wrote:
> Hi Ben,
>
> On 09/15/2016 02:22 AM, Ben Greear wrote:
>> On 08/20/2015 08:11 AM, Zefir Kurtisi wrote:
>>> On 08/19/2015 09:07 PM, Ben Greear wrote:
>>>> I have a commercial AP that is using a CM9 ath5k radio (evidently, I could be
>>>> wrong)
>>>> and it has the ability to do a constant transmit of raw noise (RF probe shows
>>>> noise, but a monitor-port sniffer does not see any frames from the CM9).
>>>>
>>>> I don't know the low-level details of how it is doing this, but I suspect
>>>> it is using something like madwifi for a driver.
>>>>
>>>> Does anyone know how this can be done with modern software and
>>>> ath5k or ath9k NICs?
>>>>
>>>> Thanks,
>>>> Ben
>>>>
>>>
>>> Maybe slightly related: some years ago when DFS became a topic and it was hard to
>>> get hands on radar pattern generators, Christian Lamparter wrote a variant of the
>>> carl9170 fw [1] which can generate radar pulses to test ath9k and other DFS radar
>>> detectors. Pulses are generated by enabling txout at defined sampling intervals.
>>>
>>> It should be doable to mimic what you are looking for by generating a _very_ long
>>> pulse.
>>
>> Sorry to revive such an old thread..but I'm back poking at this.
>>
> Whew, that year went by so incredibly fast ;)
>
>> I've used the modified carl9170 firmware to generate pulses, with
>> the control being 'pulse-width' and 'pulse-interval'.
>>
>> This sort of works, and sometimes our ath10k in an isolation chamber reports
>> a radar event.
>>
>> But, after some reading, I am thinking I need more control to better mimic
>> a radar.
>>
>> If I understand things properly, I need something like this:
>>
>> A pulse event being:  pulse width, pulse period:  For instance 1us, 200us
>> Then, I need to configure an amount of pulse events, maybe 10-30 consecutive pulse
>> events.
>> Then, I need a quiet period to mimic the radar sweeping full circle (15 seconds
>> perhaps)
>>
>> From what I can tell, the carl9170 modified firmware is missing the features
>> to do this, though it should not be too difficult to add.
>>
> Yes, that's essentially it - the last step is even not needed if your goal is to
> estimate DFS detection probabilities, since in the certification lab they usually
> just repeatedly fire the radar pattern and count detection events.
>
> When I played with the modified carl9170 FW, I estimated that developing an solid
> and reliable radar pattern pulse scheduler would take me 2-4 weeks, so being in a
> hurry I ended up using an SDR (Ettus USRP N200, see [1]). I developed a pulse
> scheduler to feed arbitrary patterns (covering those for DFS testing), which is
> available in [2]. It has not been maintained ever since, but might help you as
> starting point if you decide to go that route.

Out of curiosity, about how much did that kit cost?

>
>> If someone has an idea whether the control above is appropriate, I'd
>> appreciate feedback before I start hacking...
>>
>> This document seemed useful, for instance:
>>
>> https://dl.cdn-anritsu.com/en-en/test-measurement/files/Product-Introductions/Product-Introduction/mx370073a-el1200.pdf
>>
> We use the R&S SMBV100A [3], which we know is also used in some certification
> labs. Unfortunately it is not exactly cheap - if you are not going to prepare your
> product for certification, the SDR approach is affordable and good enough.

Thanks for the info.  We previously used the carl9170 to do some const-tx signal generation
to reproduce some interesting crashes in ath10k, and now a user wants a cheap radar emulator.

We could also use it for some basic testing in our own labs.

I'll work on allowing the specific number of pulses as well as a quiet period in
the carl9170.  Hopefully just a day or two of work on top of what I already have.

Thanks much for the suggestions.

--Ben
>
>>
>> Thanks,
>> Ben
>>
>
> Good luck,
> Zefir
>
> [1] https://www.ettus.com/product/details/UN200-KIT
> [2] https://github.com/zefir-kurtisi/USRP-Radar-Relay
> [3]
> https://www.rohde-schwarz.com/us/product/smbv100a-productstartpage_63493-10220.html
>


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