Re: [ath9k-devel] 'Superchannel'?
.. since someone already posted, I may as well follow up. If people wish to reverse engineer how all of this works (read: I'd likely be doing the same if I were involved) and collaborate on how to do this on the ath9k mailing list, I don't think anyone here will actively try to stop you. But please keep in mind the ramifications of doing so. You may not personally get in trouble by making their kit work on non-standard frequencies (and at non-regulation TX power), but this isn't to say that this won't be causing issues for Qualcomm in the future. You may even be in a country where this is legal (and boy have I heard _that_ line be dragged out time and time again!) but again, the repercussions extend outside of your country and likely to tens (if not hundreds) of thousands of open source users. The last thing any of us want to see is their newer stuff go back to being 100% binary blob drivers. So with that in mind: * Everything you likely need to get this show off the road is in ath9k; * I ask you keep in mind is to look at how TX power calibration is done. (Hint: there's code to generate the relevant TX power curve based on the channel frequency when using closed-loop TX power control; and there's similar code to select the relevant parameters for open loop TX power control.) You're going to _have_ to do this if you're going to even remotely stay within legal (and non-equipment-damaging) levels! * Please keep in mind that whilst your NIC may work fine with it, others may not have in any way been tested for frequencies outside of the standard. Who knows what kind of crazy harmonic distortion is going to occur in the TX path when you start running the NIC at outside frequencies. And finally: * You keep in mind that you're taking responsibility for others who may come along and use it. I've already had people contact me privately, wanting further information so they can run their home networks outside of congested 2.4ghz bands. (Which I always reply then buy 5ghz kit.) This can and will occur. So if you're going to do this AND release it publicly, I suggest you only do so _IF_ you are able to verify the kit is running at something even remotely resembling correct. Ie, you say card X from vendor Y works correct from these frequencies to these frequencies and we've tested thirty of them, so it is highly unlikely that it's a fluke that one card performs better. This is the kind of testing I bet NIC vendors are doing before they get their cards certified and sent to you - but they're testing for certified frequencies, not other things. :-) If you're going to do this, document on your website(s) exactly what kind of things can and will go wrong. If you don't have a spectrum analyser so you can do spectral curve tests, I highly suggest against even attempting this project. No, a Ubiquiti AP or Wispy unit doesn't count - I mean a $20k + spectrum analyser from Agilent. As always, look both sides before you cross the road, and set a good example for others. Adrian ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
On Sunday 28 Aug 2011, Adrian Chadd wrote: Guys (and girls, and fuzzy others..) I'll do some poking of my atheros contacts and see what can be found. Alex/Alex/Jerry, please follow up with me privately. Adrian ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel Could I make a request. One of the things that would be useful as a result of doing this work would a standard way for people with licences to extend the range as defined by CRDA. In the case discussed here it is a HAM licence for frequencies below 2.4GHz, in my case the area of interest is what is known as Band C here in the UK at 5.8GHz. You do need a licence (£1 per station per year to a minimum of £50 per year) from Ofcom (our FCC) but it is easy to obtain. The 5.8 band users are also allowed more power than you would be in the normal 802.11a band. Once that standard way (preferably without rebuilding the kernel) has been defined it should be made public with lots of caviats written around it saying that this can only be used if you posess a licence and that you are personally liable for its misuse. David ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
On Sun, 2011-08-28 at 11:04 +0100, David Goodenough wrote: On Sunday 28 Aug 2011, Adrian Chadd wrote: Guys (and girls, and fuzzy others..) I'll do some poking of my atheros contacts and see what can be found. Alex/Alex/Jerry, please follow up with me privately. Adrian ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel Could I make a request. One of the things that would be useful as a result of doing this work would a standard way for people with licences to extend the range as defined by CRDA. In the case discussed here it is a HAM licence for frequencies below 2.4GHz, in my case the area of interest is what is known as Band C here in the UK at 5.8GHz. You do need a licence (£1 per station per year to a minimum of £50 per year) from Ofcom (our FCC) but it is easy to obtain. The 5.8 band users are also allowed more power than you would be in the normal 802.11a band. Once that standard way (preferably without rebuilding the kernel) has been defined it should be made public with lots of caviats written around it saying that this can only be used if you posess a licence and that you are personally liable for its misuse. David ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel David, A standard would be nice but can it be done without opening the Pandora box for misuse. US Amateur also have privileges in the 5cm band which overlaps the the 5 Ghz ISM band. We are secondary users of this band with little power restriction other than not causing interference to other user of this spectrum. It would be nice to see a open source solution without having to buy more expensive commercial gear. Jerry ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
On Fri, Aug 26, 2011 at 09:18:47PM -0400, Jerald A DeLong wrote: I would also be very interested in this discussion. Jerry, KD4YAL Hi Jerald, Obviously Alex shows bizarre behavior - hi got a full support from important persons (do not including myself) and disappear. So please tell us about your project and I try to renew the correspondence. 73! Alex Hacker. ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
On 28 August 2011 01:10, Alex Hacker hac...@epn.ru wrote: On Fri, Aug 26, 2011 at 09:18:47PM -0400, Jerald A DeLong wrote: I would also be very interested in this discussion. Jerry, KD4YAL Hi Jerald, Obviously Alex shows bizarre behavior - hi got a full support from important persons (do not including myself) and disappear. So please tell us about your project and I try to renew the correspondence. It's something to take off-list. It's also not easy to do correctly, for all the reasons we've discussed off-list. I think with a bit of persistence, those with the relevant licences could get some further help. But you're going to have to go code diving to understand what's involved in the TX path. :) (The RX path looks mostly easy.) Adrian ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
Tnx guys for all the replies. Now I see the actual situation with regulatory and FCC and the ath9k development a little bit clearer. Best regards JoeSemler Am 25.08.2011 um 10:15 schrieb Adrian Chadd adr...@freebsd.org: On 25 August 2011 15:43, Joe Semler josef.sem...@gmail.com wrote: Hy guys, I'm following this discussion according regulatory, frequency and power limitation now for a while in this forum. Oc it's a good policy to have a clear regulatory for ath9k and for our openWRT. But is it not a little bit to much regulatory? My opinion is, that it's in the responsibility of the operator to fulfill the law. [snip] I'm going to stay out of that discussion, because it's rather .. well, delicate. :) Wold be really great when we could find to such a regulatory. It would help a lot of radio amateurs to use openWRT instead of airOS for HAMNET. For the minority of users that are licenced to operate at different frequencies and power restrictions, I think the best bet is to try to build some relations with the vendor(s) in question (eg Atheros) and talk directly with some of the developers there. It's annoying, but do you really want to see a proliferation of people rolling out drivers which let users select frequencies outside the regulatory limits? Then the next revision of hardware suddenly will likely stop you from doing it. Then everyone loses. Adrian ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
Obviously Alex shows bizarre behavior - hi got a full support from important persons (do not including myself) and disappear. I did not mean to disappear. I took it off list so people that shouldn't be using non part15 frequencies don't get ahold of the info. If the powers that be on the mailing list are OK with me pursuing this on the list, then I will be happy to share what I have so far with the group. Thanks! Alex ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
On Sun, 2011-08-28 at 01:17 +0800, Adrian Chadd wrote: On 28 August 2011 01:10, Alex Hacker hac...@epn.ru wrote: On Fri, Aug 26, 2011 at 09:18:47PM -0400, Jerald A DeLong wrote: I would also be very interested in this discussion. Jerry, KD4YAL Hi Jerald, Obviously Alex shows bizarre behavior - hi got a full support from important persons (do not including myself) and disappear. So please tell us about your project and I try to renew the correspondence. It's something to take off-list. It's also not easy to do correctly, for all the reasons we've discussed off-list. I think with a bit of persistence, those with the relevant licences could get some further help. But you're going to have to go code diving to understand what's involved in the TX path. :) (The RX path looks mostly easy.) Adrian Adrian, I am fine with taking it off the list and any help would be appreciated. Jerry, KD4YAL ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
I would also be very interested in this discussion. Jerry, KD4YAL ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
On 25 August 2011 16:07, Alex Hacker hac...@epn.ru wrote: I think the best thing to do is engage some of the Atheros developers directly, rather than on the mailing list. Since you have a licence to tinker with this kind of stuff, you're allowed to, but this may give others (who don't have an amateur licence) the same idea. :) Adrian Hello everybody, Yeah, some people (like U.. teens) sell this idea on the open market. :) Hah. Why doesn't that surprise me. Actually you need only a half of hour to find that synth VCO can QSY 2272..3000 (11g) and 3500..6400 (11a). This is a 'Secret de Polichinelle'... 73! I've heard stories too of people blowing things up when trying to set the synth frequencies too far out of whack. Anyway, OP: contact me privately and I'll start a thread with some other developers to see if we can help you out. Adrian ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
On 25 August 2011 15:43, Joe Semler josef.sem...@gmail.com wrote: Hy guys, I'm following this discussion according regulatory, frequency and power limitation now for a while in this forum. Oc it's a good policy to have a clear regulatory for ath9k and for our openWRT. But is it not a little bit to much regulatory? My opinion is, that it's in the responsibility of the operator to fulfill the law. [snip] I'm going to stay out of that discussion, because it's rather .. well, delicate. :) Wold be really great when we could find to such a regulatory. It would help a lot of radio amateurs to use openWRT instead of airOS for HAMNET. For the minority of users that are licenced to operate at different frequencies and power restrictions, I think the best bet is to try to build some relations with the vendor(s) in question (eg Atheros) and talk directly with some of the developers there. It's annoying, but do you really want to see a proliferation of people rolling out drivers which let users select frequencies outside the regulatory limits? Then the next revision of hardware suddenly will likely stop you from doing it. Then everyone loses. Adrian ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
On Thu, Aug 25, 2011 at 1:15 AM, Adrian Chadd adr...@freebsd.org wrote: On 25 August 2011 15:43, Joe Semler josef.sem...@gmail.com wrote: Hy guys, I'm following this discussion according regulatory, frequency and power limitation now for a while in this forum. Oc it's a good policy to have a clear regulatory for ath9k and for our openWRT. But is it not a little bit to much regulatory? My opinion is, that it's in the responsibility of the operator to fulfill the law. [snip] I'm going to stay out of that discussion, because it's rather .. well, delicate. :) Its very simple -- if you know what you are doing, you can enable whatever frequency you want and sign your own regulatory database. For the clueless user we provide defaults that do abide by regulatory rules. You have the freedom to make a change if you know what you are doing. This was by design! Luis ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
* changing regulatory not advisable This may not be applicable if the amateur licence he has permits him to transmit on these frequencies. *second number is hw specific value for the channel, we use this index to do some HAL configuration like ANI, calibration or any other operation in the driver code using this index for struct ath9k_channel ... and the hardware doesn't have calibration data for those lower frequencies, so you may find you have to do all kinds of crazy hacks in the EEPROM code for things to work. Adrian ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
On Wed, Aug 24, 2011 at 12:32 PM, Adrian Chadd adr...@freebsd.org wrote: * changing regulatory not advisable This may not be applicable if the amateur licence he has permits him to transmit on these frequencies. oh ok, i don't know much about regulatory *second number is hw specific value for the channel, we use this index to do some HAL configuration like ANI, calibration or any other operation in the driver code using this index for struct ath9k_channel ... and the hardware doesn't have calibration data for those lower frequencies, so you may find you have to do all kinds of crazy hacks in the EEPROM code for things to work. hmmm ok, going to be lots of hardware code reading :) Adrian ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
This may not be applicable if the amateur licence he has permits him to transmit on these frequencies. Exactly. According to the FCC it's OK for us to modify stuff to work on the amateur bands. Which in this case there is a band that starts below the frequency of channel 1 and is shared with unlicensed devices up to the frequency of channel 4 or so. *second number is hw specific value for the channel, we use this index to do some HAL configuration like ANI, calibration or any other operation in the driver code using this index for struct ath9k_channel Is there a chart somewhere that maps the index number to frequency? ... and the hardware doesn't have calibration data for those lower frequencies, so you may find you have to do all kinds of crazy hacks in the EEPROM code for things to work. How does Ubiquiti do it then? Their M series products have Atheros 802.11n chipsets and a Compliance Test regdomain that allows channels 0-255 (basically 2.3-2.7 GHz) on the 2.4GHz band. *you should have hit BUILD_BUG_ON(ARRAY_SIZE(ath9k_2ghz_chantable) + ARRAY_SIZE(ath9k_5ghz_chantable) != ATH9K_NUM_CHANNELS); I did. But changing the value of ATH9K_NUM_CHANNELS fixed that ;) Thanks for the replies! -Alex ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
Re: [ath9k-devel] 'Superchannel'?
On Wed, Aug 24, 2011 at 1:35 AM, Alex S. al_91...@hotmail.com wrote: Hi, I am working on an amateur radio project, so I'd like to get access to the 2.4GHz channels below 2412. I'm using OpenWRT and have modified my CRDA regulatory.bin file to allow those frequencies, but I still only have access to channels 1-11. Poking around in the source for ath9k, I see ath9k_2ghz_chantable in init.c. Will adding values here open up extra channels? If so, what exactly is the format of CHAN2G(2467, 11)? The 2467 is obviously the frequency but what is the significance of the second number? I added the channels from 2312-2407 and after recompiling it doesn't seem to have changed anything. * changing regulatory not advisable *second number is hw specific value for the channel, we use this index to do some HAL configuration like ANI, calibration or any other operation in the driver code using this index for struct ath9k_channel *you should have hit BUILD_BUG_ON(ARRAY_SIZE(ath9k_2ghz_chantable) + ARRAY_SIZE(ath9k_5ghz_chantable) != ATH9K_NUM_CHANNELS); Thanks for any help, Alex ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel ___ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel