Turning off Radar Detection on DFS channels?

You're recommending an "expert" check box in the webUI, where when checked, you could override the DFS ?

I'd love to live in your bubble/reality ...

3 Likes

How many nW do you think the weather radar receiver gets back from it’s own high power output from a single 1mm particle flying in the air lets say 300km away compared to your 25mW transmitting power with beamforming right to the radar?

You have to make the difference between working two way radio and only receiving radio waves.

5 Likes

and the DFS documents reference specific requirements for anything within 35km (21 miles) of Doppler Radar (TDWR) locations. Radar systems protected by the DFS standard come in many forms, including radar on boats, radars for TV weather stations, and military radars.

~ From: https://www.networkcomputing.com/wireless-infrastructure/dynamic-frequency-selection-why-its-critical-80211ac

My answer is yes.

4 Likes

speaking of radar I use a small wire antenna with adsb to detect aircraft's up to 230NM. which is a very large diameter. many European countries could fit in that diameter.

LAW exist for a reason and those shouldn't be crossed just to get better 5Ghz coverage in your house. use other channels add another AP instead of breaking the law.

5 Likes

It is sad to see just how right this might be. Regarding wireless, people seem to be not respectful enough to use a DFS function controlled by them responsibly.
It gets especially sad because the laws around DFS marked the end of such central devices as routers being operable by fully free software. As per the laws, the drivers have to be non-free.
Because I am trying to avoid non-free software, I bought a few ancient 5GHz capable APs, which came out before the regulations and work without non-free drivers (they can be used with ath9k) and flashed LibreCMC, a fully free fork of OpenWRT on them. I am operating them on the channels 36-48, on which indoor use is allowed without DFS. I had to carefully select the right channels to not pick DFS-only channels because somehow it would let me choose them on this specific non-DFS device. DFS is a great idea, but as long as it requires me to give up control over my router to some company I will not use DFS channels. Who knows which security vulnerabilities or backdoors might be hiding in software for which only a very small amount of people have access to the source code and which may not be modified?
Sadly I can only use ancient Wi-Fi standards (up to Wi-Fi 4) with these ancient routers because there are no new routers made for only the DFS-free channels working without a binary blob for DFS. However, if it is true that Wi-Fi 6 has no such requirements there may be a light at the end of the tunnel.

Yes there is, just a different one than you are proposing:

This should work quite well if neighbouring APs are your problem. Neither you want your neighbour's wifi signals nor does your neighbour want yours. So just make sure all signals stay where they belong. And without the wifis getting in each other's way all the time, both of you can enjoy lightning-fast wifi.
It surprises to see how uncommon his is. It seems to be more common among people who fear that little bit of non-ionizing electromagnetic waves wifis use to communicate as if it was nuclear radiation. Please do not fall for that.
This definitely deserves its own thread. Although it is simple in theory, it needs to be thought through and materials need to be tested. It would be mostly physics. Bear in mind that for most wifi shielding solutions (which might sometimes be offered at insane price points) the product is not something scientific that works, instead it is often emotions. Also, people who fear wifi radiation and do not understand how their wifi router works might do things as pointless as putting the material around their APs antenna, so some products might be deliberately designed to not make a huge difference. Anyway, with the right research it should be simple as it is something every microwave (oven) does - and thankfully very, very well I might add - to avoid heating up (more like burning) the humans on the outside in the same way the meal on the inside should be heated.

1 Like

Can you provide a reference for that claim?

The wireless reg db is is open.

Or do you mean the actual WiFi chip behavior itself?

If so, how does that differ from an other code in the chip?

It boils down to that when the manufacturers have to stop the users from changing the software (because they could change it in a way so that DFS does not work any more).

DFS is described in Recommendation ITU-R M. 1652-1 to ensure compatible operation with radiodetermination systems.

The DFS mechanism shall ensure that the probability of selecting a given channel is the same for all available channels within the 5 250 -5 350 MHz and 5 470 -5 725 MHz bands. The DFS mechanism shall also ensure, on average, a near-uniform spread of the loading of the spectrum.

WAS/RLAN shall implement a dynamic frequency selection providing a mitigation against interference to radar at least as efficient as DFS described in ETSI Standard EN 301 893 V2.1.1. Settings (hardware and/or software) of WAS/RLAN related to DFS shall not be accessible to the user if changing those settings results in the WAS/RLANs no longer being compliant with the DFS requirements. This includes (a) not allowing the user to change the country of operation and/or the operating frequency band if that results in the equipment no longer being compliant with the DFS requirements and (b) not accepting software and/or firmware which results in the equipment no longer being compliant with the DFS requirements.

https://eur-lex.europa.eu/legal-content/en/TXT/?uri=CELEX%3A32022D0179

We all know that you can install alternative software like OpenWRT on a router. However, if you try to load something else than the (in some way digitally signed) binary blob from the manufacturer on the wifi chipset the chipset will refuse to work.

Result: Only non-free software allowed!

Free Software refers to software coming with a licence that respects the four essential freedoms of free software.

I see nothing in the Wikipedia link explaining: "laws around DFS marked the end of such central devices as routers being operable by fully free software. As per the laws, the drivers have to be non-free."

In fact, your link seem to imply (as I already knew) LibreCMC is a libre distro.

I was hoping you provide a link with information supporting your statement about laws requiring DFS code to be non-free - or to explain what you're referencing as non-free.

Again:

  • the WiFi chip comes from the manufacturer (nothing to do with a the libre nature of any distro), are you referring to that?
  • The wireless reg db is open, so I assume you're referring to something else?

It's already been made clear that it cannot be legally altered. That's been made clear, but I do not understand the non-free comment.

If the device does not accept alternative software because alternative software could be used to circumvent DFS, then this leaves the user no choice but to use the non-free software from the manufacturer.

Free Software may be used, studied, shared and improved by its users. This is impossible when the device only accepts software signed by the manufacturer.

https://www.fsf.org/blogs/licensing/from-tpp-to-saving-wifi-the-fsf-fights-for-you

https://libreplanet.org/wiki/Save_WiFi

Thanks.

  • What alternative software?
  • Exactly what "non-free software from the manufacturer" are you referring to?

Again, are you referring to the code burnt into the WiFi chip (but if so, why are you discussing something that never changes - regardless if the device runs the manufacturer's OS or an alternative OS)?

Are you referring to a specific manufacturer and their non-free driver blobs (e.g. Broadcom) - but if so, what does this have to do with DFS specifically?

You linked LibreCMC - and this is an OpenWrt forum. Both are "alternative software" for a "device" - so hopefully you understand now why I'm seeking clarity on what non-free/free reference you're making.

Or are you implying something else?

2 Likes

Maybe you did not see my edit above yet.

By the way, OpenWRT is not fully free any more. It includes those blobs because modern wifi chipsets refuse to work without them.

https://www.fsf.org/blogs/licensing/from-tpp-to-saving-wifi-the-fsf-fights-for-you

https://libreplanet.org/wiki/Save_WiFi

I don't understand the edit part either.

What are you loading?

Where are you loading it?

What is the "something else" you're referring to?

Oh!

  • :bulb: If one can install OpenWrt (or LibreCMC) on such a [post-2015] device, they've already gotten around that (i.e. installing alternative software by using serial access by soldering the connectors, or "finding" the obscurity that prevents firmware from being uploaded to the web GUI)
  • Linux is compliant (again, see: https://wireless.wiki.kernel.org/en/developers/regulatory/wireless-regdb)
  • That leads me to this - :spiral_notepad: To be clear (I don't want to explain the implications as it relates to FCC-land); but this is a requirement of the manufacturer, not the end consumer :wink:
  • This was addressed here in 2015
  • I'm not sure how this relates to a WiFi chip specifically - your references seem to be focused on the OS running on the embedded board, you then go on to say yourself that it can be changed

There's quite in-depth discussions in various communities about this years ago. I really don't understand how the links are related to today's' discussion specific to DFS.

No, they have not. OpenWRT is a huge collection of software (an operating system) and includes the non-free binary blobs from the manufacturer. Upon booting up the wifi is initialised by loading those binary blobs, which then run on the wifi chipset. However, before running the blob, the wifi chipset checks if it has a valid signature from the manufacturer. If not it refuses to run it.

The Regulatory Database you are referring to is just a database. To ensure operation according to the regulations, a software that reads the regdb, checks its integrity and operates the wifi chip accordingly is needed as well. That is the software, the driver, the binary blob I am referring to. It comes from the manufacturer of the chipset and is non-free. It cannot be reverse engineered as an alternative cannot be tested. Even if a reverse engineered version existed it could not be used.

This is the part that doesn't make sense. That's why I asked you to provide a reference for this specific claim.

  • Do you have a reference? - or
  • Are you referring to a specific known manufacturer that never openly releases data for WiFi software development (e.g. Broadcom); or only does so under Non Disclosure Agreements (e.g. DD-WRT - https://openwrt.org/meta/infobox/broadcom_wifi)?

If so, I would advise using a device that has for example a MediaTek chip - where there is openly developed software avaliable.

You then seem to skip subjects and reference permanently burnt-in code existing on the WiFi chip itself. I don't understand how this is related to the former claim of an OS having non-free blobs.

What?

Do you have a reference for this too?

I think you are confusing the embedded board itself that boots the OS (i.e. what was locked down in 2015 by ET Docket No. 15-170) and the WiFi chip (that is connected to the board).

Terminology might be part of the confusion here:
Do you mean "closed-source" or something similar, when you say "non-free"?
"Free" is not quite the same as "open source".

To my knowledge the blobs themselves are free (if published by the OEM) and can be legally used, but we do not have the source code for them. E.g. ath10k firmware blob, ath11k firmware blob, mwlwifi firmware blob, mediatek blobs etc. The upper-level drivers have open source, but all those newish wifi chips have a closed-source low-level firmware blob.

(But I do not quite get why you specifically blame DFS functionality for that blob. The blobs handle much else too, as the troubles with mwlwifi show)

3 Likes

The FCC has proposed rules (ET Docket No. 15-170) that will require device makers with WiFi and other Radio Frequency (RF) devices to cryptographically lock down [i.e. digital signature checking] the RF-controlling software [i.e. the software that runs on the wifi chip] on those devices so as to prevent users from installing the software of their choice.

For other references of where such things are done take a look at Microsoft Pluton, the Intel Management Engine, the AMD Platform Security Processor or ARM Trust Zone. All these things are threats to our freedom and Microsoft Pluton is a threat to our society. It would make it next to impossible for whistleblowers like Snowden to provide us with proof for their information.

Exactly, open source is the term huge corporations use and is the laxer one. It is not about user freedom but only one way of developing software that might be lucrative in an economic sense sometimes.
Free Software (as in freedom) is the term coined by the Free Software movement around Richard Stallman, who started the GNU project to reverse engineer UNIX, (originating from the academic hacker culture) before open source even existed and is the origin of all the later ideas like open source. Free Software is about user freedom Open Source is about the cases where some of these ideas increase profit because some software only needs to be developed once.

The Free Software Initiative and the Open Source project are working together although they do not get along too well.

Often, those pieces of software where freedom is most needed are those corporations or others are most unwilling to hand over control to the user. After all, if they where acting in the users' best interest there would be next to no one striving for freedom.

1 Like

Thanks for the clarification - I understand now.

The id est (i.e.) and emphasis in this quote seem to be added by you, and it seems to identify the source of your confusion. That's why I asked for a reference - instead you altered the text contained in your link to match your statements.

Just to be clear, in 2015 manufacturers locked down the ability to install different OSes on devices. I provided 2 examples of this above how in 2023 we install "Alternative Software":

I have no proof otherwise - but I have never heard of evidence of cryptographic lock down of the interaction between the WiFi blobs and the permanent code on the WiFi chip. It seems you're confusing the manufacturer of the embedded device (i.e. the person who needs FCC certification on the finished product) and the wholesale manufacturer of the WiFi chips/WiFi SoCs provided to the former.

It sounds plausible; but to be clear, that's not what is generally understood to have occurred to devices post-ET Docket No. 15-170. Do you have some reference?

1 Like

I did not "alter" the text to match my statements, I added "[i.e. ...]" for clarification. How else would "cryptographically lock[ing] down the RF-controlling software" work?

I can imagine that before the wifi chips supported cryptographically locking down only the wifi firmware, they instead locked down all firmware to comply with the new laws.

Also consider this (although a device needing non-free blobs can either mean there just are no reverse engineered drivers yet or it is partly locked down somehow):

Still before the new laws, some drivers where reverse engineered, now no newer wifi chipset is compatible with LibreCMC.

Thank you for the discussion.

If you choose and at your convenience, feel free to provide a reference that matches your assertion above.

  • I believe this is more likely to be than the crypto and lockdown assertions (BTW there's still no information or reference to this)
  • :warning: Are you saying that the manufacturers have secretly switched out blobs?
  • I don't understand the quote you added - it seems like you keep endorsing LibreCMC over OpenWrt

Modular transmitters need external data and power from the host to operate. Such transmitters, if not approved by the manufacturer, may be sold as sub-assemblies or components to the host device manufacturer. In such instances, the host device manufacturer is responsible for the approval of the final end product including sub-assemblies

I also quoted the current EU law above, highlighting this part. Not accepting certain software is essentially locking down the wifi chipset.

1 Like