Re: [ath5k-devel] Rookie needs helps with ath5k basics

Re: [ath5k-devel] Rookie needs helps with ath5k basics

[Sergey Ryazanov]

Cc linux-wireless since Rostislav Lisovy just working on adding
802.11p to the stack.

2014-11-07 16:49 GMT+03:00 Hernán Maximiliano González Calderón
<hernan.gonzalez.calderon@gmail.com>:
> Hello everyone,
>
> I am still working to adapt the ath5k module to transmit in the
> 5850..5925GHz range, in order to comply with IEEE 802.11p requirements. Our
> plan is to liberate the code to the community as soon as we develop it.
>
> I have already compiled a new regdomains database with wireless-regdb and
> crda, and we are using the module in ATH5K_TEST_CHANNELS mode. The database
> is now defined as follows:
>
> (2402 - 2472 @ 40), (3, 27)
> (5170 - 5250 @ 40), (3, 17)
> (5250 - 5330 @ 40), (3, 20)
> (5490 - 5600 @ 40), (3, 20)
> (5650 - 5710 @ 40), (3, 20)
> (5735 - 5835 @ 40), (3, 30)
> (5835 - 5925 @ 10), (3, 30)
>
> However, when I execute "iw wlan1 ibss join TFG 5850" it returns the -22
> error number, indicating that we are using a frequency not defined.
>
> 2014-02-19 17:22 GMT+01:00 Hernán Maximiliano González Calderón
> <hernan.gonzalez.calderon@gmail.com>:
>>
>> Thanks for the quick reply and sorry for not giving an answer until now,
>> but first I had to talk with my project advisor. The reason we chose ath5k
>> was that the cards we bought used it and all information we gather about
>> this kind of projects were related to that driver.
>>
>> I also have talked with my advisor and whatever we accomplish will come
>> back to the community.
>>
>> I am just starting with the project and I am needing some guides, the tips
>> and info you all gave me will be very helpful. I will keep on working and
>> will tell you if I get something done.
>>
>> Thanks a lot,
>> Hernán M. G. C.
>>
>>
>> 2014-02-18 2:03 GMT+01:00 Adrian Chadd <adrian@freebsd.org>:
>>>
>>> ... because some of the 802.11p NICs are actually ath5k NICs that have
>>> the relevant bandpass filters for 5.9GHz and high output amplifiers.
>>>
>>>
>>>
>>> -a
>>>
>>>
>>> On 17 February 2014 01:27, Holger Schurig <holgerschurig@gmail.com>
>>> wrote:
>>> > Okay, I admit that I cannot help you, I have no clue on the driver
>>> > level.
>>> >
>>> > But maybe I can help with the methodology. :-)
>>> >
>>> > You mention 802.11p (car-to-car-communication). Is there any specific
>>> > reason you base it on ath5k and not on ath9k?  If you look at the
>>> > number of commits, then you should see that ath9k is much more lively.
>>> > People are actively working with that code and might be able to be
>>> > answer more specific questions.
>>> > Another thing that I noted: I have seen over the years many requests
>>> > of information from uni projects in this mailing list. But I'm quite
>>> > unsure if ever something came back into the Linux kernel. How do you
>>> > plan to tackle that?  I have the feeling that people are more likely
>>> > to cooperate if the work doesn't end up in yet another black hole ...
>>> >
>>> > And one tip: ask specific questions, not broad ones. For example, look
>>> > at what features you need to implement 802.11p. Now look at what OSI
>>> > level this has to be done, e.g. at hardware level (frequency,
>>> > bandwidth), driver level, or protocoll layer (mac80211, user-space
>>> > layer (e.g. wpa_supplicant). That would allow you to ask questions not
>>> > like "Tell me everything", but "Oh, I need to do XYZ, where can I do
>>> > it?". It might even help you in finding your way, e.g. by looking into
>>> > git commits inside the ath/ath9k subdirectories that might have
>>> > something to do with what you need.
>>> > _______________________________________________
>>> > ath5k-devel mailing list
>>> > ath5k-devel@lists.ath5k.org
>>> > https://lists.ath5k.org/mailman/listinfo/ath5k-devel

-- 
BR,
Sergey

Re: [ath5k-devel] Rookie needs helps with ath5k basics

[Bruno Randolf]

Hi,

I don't remember all the details but there are various places which can
limit the available channels in ath5k. Check

ath5k_is_standard_channel()
ath5k_setup_channels()
ath5k_setup_bands()

in drivers/net/wireless/ath/ath5k/base.c

Good luck!

bruno

On 11/07/2014 02:27 PM, Sergey Ryazanov wrote:
> Cc linux-wireless since Rostislav Lisovy just working on adding
> 802.11p to the stack.
> 
> 2014-11-07 16:49 GMT+03:00 Hernán Maximiliano González Calderón
> <hernan.gonzalez.calderon@gmail.com>:
>> Hello everyone,
>>
>> I am still working to adapt the ath5k module to transmit in the
>> 5850..5925GHz range, in order to comply with IEEE 802.11p requirements. Our
>> plan is to liberate the code to the community as soon as we develop it.
>>
>> I have already compiled a new regdomains database with wireless-regdb and
>> crda, and we are using the module in ATH5K_TEST_CHANNELS mode. The database
>> is now defined as follows:
>>
>> (2402 - 2472 @ 40), (3, 27)
>> (5170 - 5250 @ 40), (3, 17)
>> (5250 - 5330 @ 40), (3, 20)
>> (5490 - 5600 @ 40), (3, 20)
>> (5650 - 5710 @ 40), (3, 20)
>> (5735 - 5835 @ 40), (3, 30)
>> (5835 - 5925 @ 10), (3, 30)
>>
>> However, when I execute "iw wlan1 ibss join TFG 5850" it returns the -22
>> error number, indicating that we are using a frequency not defined.
>>
>> 2014-02-19 17:22 GMT+01:00 Hernán Maximiliano González Calderón
>> <hernan.gonzalez.calderon@gmail.com>:
>>>
>>> Thanks for the quick reply and sorry for not giving an answer until now,
>>> but first I had to talk with my project advisor. The reason we chose ath5k
>>> was that the cards we bought used it and all information we gather about
>>> this kind of projects were related to that driver.
>>>
>>> I also have talked with my advisor and whatever we accomplish will come
>>> back to the community.
>>>
>>> I am just starting with the project and I am needing some guides, the tips
>>> and info you all gave me will be very helpful. I will keep on working and
>>> will tell you if I get something done.
>>>
>>> Thanks a lot,
>>> Hernán M. G. C.
>>>
>>>
>>> 2014-02-18 2:03 GMT+01:00 Adrian Chadd <adrian@freebsd.org>:
>>>>
>>>> ... because some of the 802.11p NICs are actually ath5k NICs that have
>>>> the relevant bandpass filters for 5.9GHz and high output amplifiers.
>>>>
>>>>
>>>>
>>>> -a
>>>>
>>>>
>>>> On 17 February 2014 01:27, Holger Schurig <holgerschurig@gmail.com>
>>>> wrote:
>>>>> Okay, I admit that I cannot help you, I have no clue on the driver
>>>>> level.
>>>>>
>>>>> But maybe I can help with the methodology. :-)
>>>>>
>>>>> You mention 802.11p (car-to-car-communication). Is there any specific
>>>>> reason you base it on ath5k and not on ath9k?  If you look at the
>>>>> number of commits, then you should see that ath9k is much more lively.
>>>>> People are actively working with that code and might be able to be
>>>>> answer more specific questions.
>>>>> Another thing that I noted: I have seen over the years many requests
>>>>> of information from uni projects in this mailing list. But I'm quite
>>>>> unsure if ever something came back into the Linux kernel. How do you
>>>>> plan to tackle that?  I have the feeling that people are more likely
>>>>> to cooperate if the work doesn't end up in yet another black hole ...
>>>>>
>>>>> And one tip: ask specific questions, not broad ones. For example, look
>>>>> at what features you need to implement 802.11p. Now look at what OSI
>>>>> level this has to be done, e.g. at hardware level (frequency,
>>>>> bandwidth), driver level, or protocoll layer (mac80211, user-space
>>>>> layer (e.g. wpa_supplicant). That would allow you to ask questions not
>>>>> like "Tell me everything", but "Oh, I need to do XYZ, where can I do
>>>>> it?". It might even help you in finding your way, e.g. by looking into
>>>>> git commits inside the ath/ath9k subdirectories that might have
>>>>> something to do with what you need.
>>>>> _______________________________________________
>>>>> ath5k-devel mailing list
>>>>> ath5k-devel@lists.ath5k.org
>>>>> https://lists.ath5k.org/mailman/listinfo/ath5k-devel
> 

Re: [ath5k-devel] Rookie needs helps with ath5k basics

[Adrian Chadd]

Hi,

right. The default NIC EEPROM setup doesn't include those channels
because they haven't tested them out.

If you have an 802.11p regulatory compliant NIC then it should have
the 11p frequencies show up in the EEPROM channel range.

So you should definitely first check that the NIC EEPROM has those as
available channels. :)



-adrian


On 7 November 2014 07:04, Bruno Randolf <br1@einfach.org> wrote:
> Hi,
>
> I don't remember all the details but there are various places which can
> limit the available channels in ath5k. Check
>
> ath5k_is_standard_channel()
> ath5k_setup_channels()
> ath5k_setup_bands()
>
> in drivers/net/wireless/ath/ath5k/base.c
>
> Good luck!
>
> bruno
>
> On 11/07/2014 02:27 PM, Sergey Ryazanov wrote:
>> Cc linux-wireless since Rostislav Lisovy just working on adding
>> 802.11p to the stack.
>>
>> 2014-11-07 16:49 GMT+03:00 Hernán Maximiliano González Calderón
>> <hernan.gonzalez.calderon@gmail.com>:
>>> Hello everyone,
>>>
>>> I am still working to adapt the ath5k module to transmit in the
>>> 5850..5925GHz range, in order to comply with IEEE 802.11p requirements. Our
>>> plan is to liberate the code to the community as soon as we develop it.
>>>
>>> I have already compiled a new regdomains database with wireless-regdb and
>>> crda, and we are using the module in ATH5K_TEST_CHANNELS mode. The database
>>> is now defined as follows:
>>>
>>> (2402 - 2472 @ 40), (3, 27)
>>> (5170 - 5250 @ 40), (3, 17)
>>> (5250 - 5330 @ 40), (3, 20)
>>> (5490 - 5600 @ 40), (3, 20)
>>> (5650 - 5710 @ 40), (3, 20)
>>> (5735 - 5835 @ 40), (3, 30)
>>> (5835 - 5925 @ 10), (3, 30)
>>>
>>> However, when I execute "iw wlan1 ibss join TFG 5850" it returns the -22
>>> error number, indicating that we are using a frequency not defined.
>>>
>>> 2014-02-19 17:22 GMT+01:00 Hernán Maximiliano González Calderón
>>> <hernan.gonzalez.calderon@gmail.com>:
>>>>
>>>> Thanks for the quick reply and sorry for not giving an answer until now,
>>>> but first I had to talk with my project advisor. The reason we chose ath5k
>>>> was that the cards we bought used it and all information we gather about
>>>> this kind of projects were related to that driver.
>>>>
>>>> I also have talked with my advisor and whatever we accomplish will come
>>>> back to the community.
>>>>
>>>> I am just starting with the project and I am needing some guides, the tips
>>>> and info you all gave me will be very helpful. I will keep on working and
>>>> will tell you if I get something done.
>>>>
>>>> Thanks a lot,
>>>> Hernán M. G. C.
>>>>
>>>>
>>>> 2014-02-18 2:03 GMT+01:00 Adrian Chadd <adrian@freebsd.org>:
>>>>>
>>>>> ... because some of the 802.11p NICs are actually ath5k NICs that have
>>>>> the relevant bandpass filters for 5.9GHz and high output amplifiers.
>>>>>
>>>>>
>>>>>
>>>>> -a
>>>>>
>>>>>
>>>>> On 17 February 2014 01:27, Holger Schurig <holgerschurig@gmail.com>
>>>>> wrote:
>>>>>> Okay, I admit that I cannot help you, I have no clue on the driver
>>>>>> level.
>>>>>>
>>>>>> But maybe I can help with the methodology. :-)
>>>>>>
>>>>>> You mention 802.11p (car-to-car-communication). Is there any specific
>>>>>> reason you base it on ath5k and not on ath9k?  If you look at the
>>>>>> number of commits, then you should see that ath9k is much more lively.
>>>>>> People are actively working with that code and might be able to be
>>>>>> answer more specific questions.
>>>>>> Another thing that I noted: I have seen over the years many requests
>>>>>> of information from uni projects in this mailing list. But I'm quite
>>>>>> unsure if ever something came back into the Linux kernel. How do you
>>>>>> plan to tackle that?  I have the feeling that people are more likely
>>>>>> to cooperate if the work doesn't end up in yet another black hole ...
>>>>>>
>>>>>> And one tip: ask specific questions, not broad ones. For example, look
>>>>>> at what features you need to implement 802.11p. Now look at what OSI
>>>>>> level this has to be done, e.g. at hardware level (frequency,
>>>>>> bandwidth), driver level, or protocoll layer (mac80211, user-space
>>>>>> layer (e.g. wpa_supplicant). That would allow you to ask questions not
>>>>>> like "Tell me everything", but "Oh, I need to do XYZ, where can I do
>>>>>> it?". It might even help you in finding your way, e.g. by looking into
>>>>>> git commits inside the ath/ath9k subdirectories that might have
>>>>>> something to do with what you need.
>>>>>> _______________________________________________
>>>>>> ath5k-devel mailing list
>>>>>> ath5k-devel@lists.ath5k.org
>>>>>> https://lists.ath5k.org/mailman/listinfo/ath5k-devel
>>
>

Re: [ath5k-devel] Rookie needs helps with ath5k basics

[Rostislav Lisovy]

Hello Hernan and all the others;

Sergey Ryazanov <ryazanov.s.a@...> writes:
> 
> Cc linux-wireless since Rostislav Lisovy just working on adding
> 802.11p to the stack.
> 
> 2014-11-07 16:49 GMT+03:00 Hernán Maximiliano González Calderón
> <hernan.gonzalez.calderon@...>:
> > Hello everyone,
> >
> > I am still working to adapt the ath5k module to transmit in the
> > 5850..5925GHz range, in order to comply with IEEE 802.11p requirements.

802.11p implementation is not just about the "WNIC driver" modification -- 
some work has to be done in the cfg80211/mac80211 subsystem as well.
All the necessary changes are already in the mac80211-next git. To make
this work, you will need to modify the WNIC driver (as you are trying
to accomplish) + user-space configuration tool.
Some quick and dirty ath9k modification to support 5.9GHz is at
https://github.com/CTU-IIG/802.11p-linux/commits/its-g5_v2
(I read the other email pointing out that the EEPROM should contain
the information about the 5.9GHz channels so I am not sure how dangerous
is what I am doing). The 'iw' modification is at
https://github.com/CTU-IIG/802.11p-iw/tree/its-g5_v2
A bit obsolete but still helpful tutorial on how to run the thing is
at https://gist.github.com/lisovy/80dde5a792e774a706a9 (obsolete in
the sense that the ath9k driver modification should be rebased to
the newest mac80211-next).

The questions that are still unanswered (and will definitely require
more effort to answer) are the 802.11p standard interpretation
regarding the OCB mode / 5.9GHz usage limitation. More specific --
Is the 5.9GHz limited for ITS (Intelligent transportation system)
usage only. Or for any usage as long as it uses the OCB mode? If one
of this limitation is valid, who does actually define the restriction --
regulatory rules of particular countries?
Another question is if the OCB mode may be used outside the 5.9GHz range?
Do we need to define the OCB_ONLY flag for 5.9GHz channels in
CRDA/wireless-regdb?

Best regards;
Rostislav

Re: [ath5k-devel] Rookie needs helps with ath5k basics

[Rostislav Lisovy]

Rostislav Lisovy <lisovy@...> writes:
> 
> The questions that are still unanswered (and will definitely require
> more effort to answer) are the 802.11p standard interpretation
> regarding the OCB mode / 5.9GHz usage limitation. More specific --
> Is the 5.9GHz limited for ITS (Intelligent transportation system)
> usage only. Or for any usage as long as it uses the OCB mode? If one
> of this limitation is valid, who does actually define the restriction -

I don't know why but the message got truncated. The last paragraph
should be:

The questions that are still unanswered (and will definitely require
more effort to answer) are the 802.11p standard interpretation
regarding the OCB mode / 5.9GHz usage limitation. More specific --
Is the 5.9GHz limited for ITS (Intelligent transportation system) usage
only. Or for any usage as long as it uses the OCB mode? If one of this
limitation is valid, who does actually define the restriction --
regulatory rules of particular countries?
Another question is if the OCB mode may be used outside the 5.9GHz
range? Do we need to define the OCB_ONLY flag for 5.9GHz channels in
CRDA/wireless-regdb?

Best regards;
Rostislav