12 replies to this topic

[Mistrale]

New Open Source Server is Now Runnig

We Vu+ have clearly stated that Vu+ is committed to the philosophy of open source as we understand that it is the very reason why numerous voluntary communities and developers are working hard around the world.

We fully understand the frustration end users felt by the fact that the sole beneficiary of open source communities for long has decided to take a contradictory action of shutting down the access to C++ part of GIT source server.

So we prepared the open GIT server to keep the philosophy of open source.
You can freely access the source code following the link below and we expect your warm attention and support.

http://code.vuplus.com/

Bron: http://www2.vuplus.c...r-is-now-runnig

[Erik Slagter]

Is the source code of the drivers also accessible? Otherwise I don't think this site will have any added value to what's already available.

[Taykun345]

From what i can see in their GIT, they are also working on 3.1 kernel. Great news!

[littlesat]

It seems even to be 3.1.1

[Erik Slagter]

Open source drivers?

[Sjaaky]

They probably can't publish the source of their drivers as lots of it is covered by a NDA with Broadcom.

[Erik Slagter]

That is a common problem with low level drivers, it can be worked around by placing the NDA code in a precompiled library and make the open source parts of the driver call the nda'd parts.

Anyway, I think it's utter nonsense that this kind of software is NDA'd, they tend to defend it by using the argument of protecting "intellectual property" from competitors, but I really can't see how the interface to the hardware would reveal any critical information about the hardware. I think it's actually concealment of bugs in the hardware.

[MiLo]

It's also to prevent some Chinese building the same box for a quarter of the price, with crappy components but with the same logo and almost indistinguishable from the outside. The Chinese box will then be sold as the real thing, giving the real manufacturer a bad name, while stealing most of the market.

[Erik Slagter]

They can do that already, as shown... It appears that even a tpm check in the closed source drivers doesn't help.

[Gennar1]

Another good reason to keep the low-level driver closed source is that usually a manufacturer wants to sell the same chip with some feature disabled at a discounted price (saving some buck for the licence fees).
Typical examples are Macrovision or advanced audio codecs (DTS, AC3+,...).
All modern SoCs have programmable audio decoders, so they can support a wide range of audio formats, but you have to pay a fee for each decoder sold to Dolby, DTS, ... If the driver were open source, you could easily enable new audio formats on low-end chips porting the driver from the more advanced versions.

Edited by Gennar1, 1 December 2011 - 16:35.

[Erik Slagter]

I think that easily can be reverse-engineered. Also, if the company finds out, they can lawsuit the offender.

[MiLo]

..Also, if the company finds out, they can lawsuit the offender.

Only if the offender is a company of some sort. They can't sue you personally. So if one bloke posts how to implement a DTS decoder, anyone with a compiler can do it for himself without paying the fee to the DTS folks.

[Erik Slagter]

What they're doing now is security by obscurity. Sooner or later someone will find out and the security will be null. What they should be practising is either make different issues of a certain IC, but if that appears too expensive (marketing is alien to me....) they should at least use one-time blowable fuses and then nobody will be tempted to try and reverse engineer enabling the various codecs. And then the driver can be open source. Compare to encryption, never trust an encryption algorithm that's not open to the public, it's not necessary for a good encryption algorithm to keep the base a secret.