wifi command set secrecy - why?

I just privately sent you the full documention for the chipsets you listed.

Thanks but nothing received at this end... (no chance to have been received and remained unnoticed in various "folders", I do email in a really oldfashioned way).

You didn't mention any of the parts you can't find data on. Not sure how anybody will help you out.

You may need to nag the manufacturer for data. Lots of specs are not for download off websites. I vaguely deal with one place that produces off the shelf products, but we get custom software. The interactions are bizarre and there are NDAs from us and them. We get internal infomation on the product, they know how we use it and there's no full docs on how any of it works as implemented. All parties are happy.

I see, yet another nullsayer wisecracking at a message he did not understand. I have been trying more than you think can be done for years, but thanks for enlightening me that the Earth is round.

Did I misunderstand your complete inability to get documentation to use some sort of wifi chipset with your own network stack?

Sucks to be you.

I think they're just embarrassed about how awful the hardware interface is.

Sylvia.

Hi Dimiter, The situation presented is somewhat vague. So I'll consider some fringe possibilities. Could it be, the chipset maker that you are dealing with does not actually have all the information? Are they the true owners of the IP? Could what you are looking at was copied from some other company? And any documentation of such might put them in legal jeopardy? (or be bad for their image at least?) More likely, the reason is mundane. Documentation process & management is costly. Why should they spend the time & money if its not going to make them any more competitive? Are you going to choose another source, one who did document everything, but charges more for their product? How critical would most of the product's customers consider this information? regards, RS

Have a look at the Linux driver if it exists. They usually reverse-engineer the needed specs.

There are various hardware interfaces and most are indeed half baked but this has little to do with the protocol used over them. Some go via SDIO, others via USB, SPI, plain UART (the latter at crazy speeds yet still too slow). But how to talk to a chipset - any chipset on the market - so one can send/receive IP packets like one would via a normal MAC etc. - is buried in the deepest secrecy. So far I have not come across anyone who knows why the secrecy, let alone how to break it (apart from reverse engineering or just implementing the 802.whatever protocols on your own).

These drivers are not open source, not for the part that matters. The reverse engineered part is available IIRC, but what I try to understand is *why* do they (and I) have to reverse engineer what the likes of microsoft and android makers have access to.

The Linux drivers are open source (whether they are documented enough or what you want, is another matter). If the makers do their own closed source driver for Linux, that's their driver, not the Linux driver.

Why they (makers) keep all that documentation secret is a long debate. They claim to have their reasons. Microsoft, Apple, and Android makers probably have NDA signed prior to seeing any spec.

All wifi chipsets I have seen - and I have probably looked at any maker over the years - are quite explicit they come with "drivers for windows, Linux" etc. These drivers are what talks to the firmware of course, which is what the secrecy is about.

I have not seen any credible claim why they keep these data secret, they just won't talk about it.

If it were just about an NDA.... A few months ago I got contacted by a rep of NXP for our region (another guy at NXP who sends me some errata sheets under NDA on a large power architecture processor of theirs had notified him of my interest, which I had expressed when NXP bought Marvel (it was Marvel, right?)). Being under NDA was no help at all. I offered to sign any sort of NDA, be responsible with decapitation or whatever - nothing. He got in touch with some asshole professor somewhere in France (I saw the latter check on our website, like he was equipped to be able to judge what we were...) and that was the end of the story.

====================================================== Dimiter Popoff, TGI

I believe that there is a fair amount of trade secrecy in making WiFi chipsets, and they're trying to protect their advantages from other manufacturers. Broadcom has been a standout performer in the sekrit sauce club.

This is quite likely the case (being competitive), but the firmware command protocols?... I don't think it is possible they don't know each other's protocols, could well be they use the same or very similar ones. If they hide things from each other it will be in the dsp-ing parts and sort of, where they can get a performance advantage.

There are other possibilities...

Price and power consumption are important. Certainly in the wired interface arena a quarter of a century ago there was secret sauce in how you divided MAC and packet level processing between the various processors. Many unfortunate choices were made at that time.

Then there's the possible issue that they don't want to let miscreants easily change RF parameters, since that would enable them to commit all sorts of RF sins. Security through obscurity is better than nothing, although maybe they use stronger techniques.

Also, impeding reverse engineering allows them to have more leverage w.r.t. licencing their technology, especially if drivers are only issued in the form of big blobs of optimised code.

Dimiter_Popoff snipped-for-privacy@tgi-sci.com wrote in news:snrdr4$bvs$1@dont- email.me:

You're mumbling.

Cisco Systems had a modem with hardware that ran under Linux.

They had to release the source code. In that code are numerous settings that do not apply to the gear it gets put on. So a good bit of the "wifi API" must be in there.

There may be many hardware specific hooks involved and it only works on their gear, but one could load it and tweak out settings on their router that were not previously available. Great routers too.

It was called DD-WRT.

There have since been numerous more:

snipped-for-privacy@decadence.org wrote in news:sns494$1s25$1 @gioia.aioe.org:

Supposed to have written "router".

I know there are open source routers. Please read my posts as many times as it takes before entering the discussion.

The first Ethernet chip I used, the "SONIC" from NSC, introduced in the early 90-s (or was it late 80-s), was completely documented, never had to look beyond its datasheet to use it. The Motorola parts with MACs were were also documented as far as I have noticed (never used any of them). You must be talking about some other "world" (PC?) I am not familiar with, but in my world things were documented as usual.

This sounds like a credible excuse but it does not explain why also all the embedded wifi modules are so strict about not allowing you to do IP packets, you *must* go through their tcp/ip stack. Surely you cannot do any RF-evil by doing IP packets and being unable to tinker with the radio.

This can be some motivation for them but it still does not explain the "no IP packets" policy, which is the bizarre part of it all and which is likely driven by what drives the secrecy I am wondering about. And if we all can only speculate about the *why* obviously it is very very serious.

====================================================== Dimiter Popoff, TGI