You can use [strike]text[/strike] to strike-out text.

You can see the full bbcode tags used here. (It's a link in the "posting rules" box on the bottom left corner.)

Hi,

@bassmadrigal thank you very much

--
Best regards,
Andrzej Telszewski

You're completely missing multiple crucial pieces of information.

Each government has its own rules of which channels and power you should be able to use, and they want you to be absolutely unable to circumvent those rules.

Your "can't understand why" ideas are a recipe for circumvention. How does a device know for sure where it is? It can't, not even if it has GPS onboard. Networks do not provide location information, let alone a way that can't be spoofed. This is never going to be good enough when the law says you must prevent people from using illegal channels and illegal power settings.

There was recently a huge controversy about proposed FCC type approval standards which said, explicitly, "explain how your device cannot be reflashed with DD-WRT to circumvent our rules". And if the FCC in the States adds new restrictions, those restrictions end up affecting everybody, because manufacturers don't want to make a hundred different country-specific devices.

So: imagine you're a Chinese manufacturer that just wants to turn an honest profit around the world. You are forced, through type approval, to bake into the hardware a default location that provides just the minimum common channels and powers that are valid everywhere -- maybe by using a weird default location, like Venezuela -- and to make it difficult, in random obnoxious and lazy ways, for people to change that location.

So the Linux kernel has to discover how random devices on multiple buses have been set up by manufacturers from the shallow end of the gene pool, and has drivers from multiple developers who all have different ideas about how to prevent you from using your hardware to break the law. Those opinions range all the way from "I work for the manufacturer and I don't want to go to jail" to "I'm a hacker and it's my duty to enable other hackers to hack".

If you want to know what a massive pain all this can be, just google how to enable channels 12 and 13 on OpenWRT with some of the Atheros chips. The standard Linux regulatory database is cryptographically signed...

... Or you could just put the exact command you started this thread with into rc.local and get on with your life.

55020: (Sorry, you didn't put any name up!) Yes, I'm well aware of government rules. I used to design RF equipment many years ago, and know only too well what a pain they are! My "other" hobby is RC models, and you should see what havoc the European Union is creating on the 2.4 GHz band at present - and yes, this will affect WiFi as well!

As I understand it, the standard Linux regulatory database contains all the necessary information - after all, that's what I'm using when I tell it to adopt the "GB" domain rules. All that needs to be added is some means of discovering where in the world you are. I don't know about the US, but when I log on to a network pretty much anywhere in Europe, the laptop - like my 'phone - seems to immediately know which country I am in. It seems to me that its just a case of introducing the left hand to the right hand!

I'm sure the regulators would be far happier having all this happen hidden from sight of the end user, rather than having to force people to tinker about to make things compliant. As an example: One of the major brands of RC equipment designed and sold in the US is illegal in Europe! The EU versions have to have the power output restricted. And the reason that Japan was so late entering the market was because it was ALL illegal over there, until common sense prevailed!

Now as far as I am aware, here in the UK, we have a relatively liberal use of 2.4 GHz for WiFi and other uses. Certainly - as you've implied - we can use more channels than the US. But this is a very congested little island! We NEED more channels to alleviate congestion, and having everything default to the US standard is a right, royal, pain in the posterior!

I accept that it *may* be impractical to implement an auto standard selection. But that would seem to me to be a preferable approach - even from a regulatory point of view - than forcing people to tinker to get the service they need and for which they have paid! After all, If I can force it to work to UK standards, somebody in the US will be able to do the same. And as we have already discovered over here, trying to identify non-compliant equipment at the point of entry into the country is totally impractical, when it is as easy to order from Hong Kong as it is to get it from the shop down the road!

Rules - like laws - have to be enforceable if they are to be practicable!

--
Pete

Well, first, in the walled garden of 3G there have long been huge complaints from people near borders and on coastlines that their phones keep roaming to adjacent countries and running up huge bills, and the only practical fix for that so far has been to promise to abolish roaming charges, maybe, eventually. The base station's location is known and trustworthy, but not the phone's.

And secondly, your Windows lappie or Android tablet is getting its location by proprietary heuristics from a vast amount of abusive snooping, including sniffing wifi SSIDs and comparing them to laboriously acquired databases, and your Linux system is getting its location from looking your IP up in old versions of the Maxmind GeoIP database, which randomly puts me anywhere from London to "somewhere in Europe" (especially on IPv6). I'd love to know how a wifi device could have that onboard, or get it from the network before it has worked out how to connect to the network...

Your personal import comes off the same production line as the millions of other devices that are imported in bulk, and type approval of bulk imports most certainly is enforced, and so the manufacturer just manufactures devices that meet worldwide type approval standards, and so your personal import also meets worldwide type approval standards. You might think you're fighting the law, but the law's already won.

-Dave S.

(btw, there's enough in my LQ posts to give you my name, photo, bio, employer, a recording of my voice last Tuesday afternoon, and my home location within a 71m radius, but of course I could be spoofing any or all of that)

Hi Dave,

Yes, I could have looked all that up, but unlike our Home Secretary, I've no wish to snoop!

Phone's running up large bills close to borders is just laziness on the part of the network operators and phone manufacturers. A mobile phone can be tracked - without GPS - to within a hundred meters or so of its location by the network operators. But if they are making money from it, I doubt they will change, unless forced!

And as regards worldwide standards - there ain't no such thing! There are regional standards that may be approved world-wide, but that's another matter entirely. If I had a mind to (and I don't!), I can order US standard RC equipment from the Far East and it would probably arrive here nearly as quickly - and probably cheaper - than if I'd ordered EU compliant stuff locally. It would not be checked for compliance at the point of entry, and its pot luck if it would even attract import duty!

Of course, if I were caught using it, I could be fined, but that possibility is so remote as to be not worth considering.

But this is drifting way off the topic!

We live in an imperfect world, but even with that imperfection, it would be nice if there was a standard *and documented* way of ensuring that programmable devices, like WiFi cards, were working to the right standard for the region in which they are used.

--
Pete

One problem has emerged with my ath9k.conf script - based on Chris' solution earlier in this thread - but only when using the slackware "generic" kernel! If you boot into single user mode (1), the script causes cfg80211 to continually loop round, re-configuring the regulatory domain to "00". This makes the terminal almost unusable! It does the same in level 3 as well, but stops looping at the login prompt.

However, it works fine in my "home-rolled" kernel.

I only noticed this after doing the recent big upgrade to eudev, when I temporarily reverted to the generic kernel to check for issues, as requested by the developers. Not a problem for me, but worth noting, in case it causes anyone else an issue.....

--
Pete

Very strange - I'm using the stock generic kernel but I can't replicate the problem. I was going to suggest that, since you'd been trying different things to make it work, there may be some old script lying around causing trouble i.e. multiple influences on the module loading. However I don't see how that effect could be manifest only for one particular kernel.

chris

I know! It doesn't make sense to me either! More details in this thread:
http://www.linuxquestions.org/questi...368/page6.html

As you can see, I've compared the kernels, and can't find any significant differences. Have you done the eudev upgrade? I think its related to that.....

--
Pete

I first tried an upgrade but had lots of problems with nvidia module in run level 4; also some problems running containers - to the point that eventually I just rebuilt the system and everything has been fine since then.

chris

So it could be OK with a fresh install, but something going awry in an upgrade? It wouldn't surprise me, bearing in mind the scale of that upgrade! Its a bit like trying to replace the foundations of a house, without damaging the superstructure! Something is bound to break!

Ah, well, it all works fine under my own kernel, so I'm not going to worry any further about it. But it was probably worth raising in case anyone else runs into it....

Cheers,

--
Pete