debian for riscv open source hardware

[hthi]

https://riscv.org/. Isa for open source, no restrictive license hardware. Debian for riscv is being made. https://www.youtube.com/watch?v=NBYbHqPNHGs.

[tomazzi]

https://riscv.org/. Isa for open source, no restrictive license hardware.

Not really.
I've just started to read the specifications and I've found this:

The RISC-V Instruction Set Manual, page 61:

Guaranteed Non-Standard Encoding Space

To support development of proprietary custom extensions, portions of the encoding space are
guaranteed to never be used by standard extensions.

In English:
Just like this is in case of ARM, every chip will have proprietary bootloaders and proprietary extensions, which will be not available for open source solutions, not to mention Free Software.

Regards.

edit: page 61, not 62 ofc.

[Head_on_a_Stick]

How about OpenSPARC?

http://www.oracle.com/technetwork/syste ... index.html

You'll need OpenBSD for that, ofc.

[pylkko]

https://riscv.org/. Isa for open source, no restrictive license hardware.

Not really.
I've just started to read the specifications and I've found this:

The RISC-V Instruction Set Manual, page 61:

Guaranteed Non-Standard Encoding Space

To support development of proprietary custom extensions, portions of the encoding space are
guaranteed to never be used by standard extensions.

In English:
Just like this is in case of ARM, every chip will have proprietary bootloaders and proprietary extensions, which will be not available for open source solutions, not to mention Free Software.

Regards.

edit: page 61, not 62 ofc.

But is this really so? I may not understand some part of your communication, but it seems to me that you are saying that since the specification allows for non-standard extensions, this automatically causes the situation that every chip manufacturer will certainly use non-standard extensions, make them proprietary and also that they will have proprietary bootloaders. Seems a bit non sequitur, unless you mean that you're guessing that that will happen for other reasons or maybe I am not uderstanding something?

[Head_on_a_Stick]

The BeagleBoard (various versions) as well as the PandaBoard use the TI OMAP family of SoCs. These come with free startup software as well as free drivers for the peripherals.

http://www.fsf.org/resources/hw/single-board-computers

EDIT: Somewhat dated link, does anybody have more current information?

This topic is relevant to my interests

[tomazzi]

But is this really so? I may not understand some part of your communication, but it seems to me that you are saying that since the specification allows for non-standard extensions, this automatically causes the situation that every chip manufacturer will certainly use non-standard extensions, make them proprietary and also that they will have proprietary bootloaders.

There's a huge difference between "allowing" and "guaranteeing support" for non-standard, proprietary extensions.

Open ISA specification has completly nothing to do with open hardware (implementation of the ISA).

The ARM ISA is also open, but if You'll take a look at the page linked by Head_on_a_Stick above, then You should quickly realize what's the reality.

Regards.

[pylkko]

There's a huge difference between "allowing" and "guaranteeing support" for non-standard, proprietary extensions.

I'm afraid I don't see your point. It appears to me that there is not a huge difference, and that even if there were your previous statement would not hold... I just don't see what you are saying.

Open ISA specification has completly nothing to do with open hardware (implementation of the ISA).

What do you mean? Has absolutely nothing to do. Certainly is has something to do?

The ARM ISA is also open, but if You'll take a look at the page linked by Head_on_a_Stick above, then You should quickly realize what's the reality.

Regards.

Really? Because to me it looked like as countering your claim that all ARM boards have non-open bootloaders?


Furthermore, correct me if I'm wrong, when you brought up the point that the specification allows for custom proprietary extensions, did you not bring is up while say "not really" to the OP's point that there is a very liberal license? Now my understanding is (and it might not be perfect) that even if RISC-V allows for these things, nevertheless the license is BSD, meaning that - and I cite WIkipedia here:

The RISC-V authors aim to provide several CPU designs freely available under a BSD license. Such licenses allow derivative works, such as RISC-V chip designs, to be either open and free, like RISC-V itself, or closed and proprietary.

and it continues...

By contrast, commercial chip vendors such as ARM Holdings and MIPS Technologies charge substantial license fees for the use of their patents.[8] They also require non-disclosure agreements before releasing documents that describe their designs' advantages and instruction set. Many design advances are completely proprietary, never described even to customers. The secrecy interferes with legitimate public educational use, security auditing, and the development of public, low–cost free and open-source software compilers, and operating systems.

SO, if I am getting this right, even if RISCV does not use a license model that requires derivative works to be only open and free (for example GPL), it nevertheless is very different from ARM and especially in the context that the OP was talking about, the license. Or do you disagree with that passage from wikipedia, because obviously wikipedia is not the last authority on anything.

[Head_on_a_Stick]

As RISC-V is released under a BSD-style licence, perhaps the GPL'd OpenRISC project would be more suitable for these boards:

https://openrisc.github.io/

There is also Novena:

A new open-hardware computing platform, flexible and powerful, designed for use as a desktop, laptop, or standalone board.

https://kosagi.com//w/index.php?title=Novena_Main_Page

[tomazzi]

SO, if I am getting this right, even if RISCV does not use a license model that requires derivative works to be only open and free (for example GPL), it nevertheless is very different from ARM and especially in the context that the OP was talking about, the license.

Apparently You don't get it.
Instruction Set Architecture (ISA) is merely a draft. It does not describe or define the final hardware. For a real-life example You can take the x86 ISA - it is used by both Intel and AMD, but the hardware is completely different and closed (proprietary).
The same example applies to the ARM ISA - f.e. Broadcom and Samsung CPUs are using different hardware implementation.

Both in case of ARM and x86, the ISA specification is open and available free of charge - but the real, final products are containing unofficial, non-documented proprietary extensions and the hardware spesc are closed.

In other words, just because You have an open ISA it does not mean that the hardware is open.

And in case of RISC-V, the ISA specification contains explicit garantee for supporting proprietary solutions...

-------------
Moreover, the project is supported by both Microshit and Google - microshit: no comment needed, Google: they've already demonstrated how to turn GPL-ed project into a cyber-weapon working against the end users: they've made Android.

The problem with Android is that due to the licensing terms, it must stay partially open... and some crazy guys have decided to remove Google's crapware - they've made cyanogen...
The solution for the above "problem" is to hide Google's crapware in the proprietary extensions, bootloaders and firmware blobs in the new CPU family, what means that the open source part of the software will not allow to change anything significant in the device.

But, the project is still in a very early stage of development - I think that they need at least 5 years to create final product.

Regards.

edit:
PS: And to be honest, I can't find anything revolutionary in the RISC-V ISA - for me, it is just yet another set of opcodes...
...but it's just my opinion...

[Head_on_a_Stick]

Instruction Set Architecture (ISA) is merely a draft. It does not describe or define the final hardware. For a real-life example You can take the x86 ISA - it is used by both Intel and AMD, but the hardware is completely different and closed (proprietary).

Whilst it is true that the microarchitecture of the x86 series is closed, the reference "Rocket" implementation of RISC-V does indeed have a fully open microarchitecture:

http://www-inst.eecs.berkeley.edu/~cs25 ... scv.pdf#13

[tomazzi]

Yes, but microarchitecture is part of hardware implementation of the ISA.

The ISA *must* be open, because the opcodes are used by the preprocessor-compiler-assembler toolchain to generate the machine-language representation of the program.

Regards.

[pylkko]

SO, if I am getting this right, even if RISCV does not use a license model that requires derivative works to be only open and free (for example GPL), it nevertheless is very different from ARM and especially in the context that the OP was talking about, the license.

Apparently You don't get it.

So, does that mean that you do recognize - or that you do not recognize - that when you said "Just like this is in case of ARM, every chip will have proprietary bootloaders" that it is not actually the case that all ARM platforms have proprietary bootloaders? As you pointed out, this is something that "I do not get". But maybe you can try to explain.

Are you saying or not that RISC-V is just like ARM when you look at the licenses/rights to make further products and that all the future bootloaders absolutely have to be - just like with ARM - proprietary and non free?? It is indeed really hard to get what you are saying here. It seems like you are using a lot of strong verbiage, but that the actual content of the statements are quite obscure. It seems as if you are dismissing RISC-V as "just like ARM", whereas the statement cited from Wikipedia is claiming that RISC-V is not just like ARM in many interesting ways. Is this not the case?

[tomazzi]

So, does that mean that you do recognize - or that you do not recognize - that when you said "Just like this is in case of ARM, every chip will have proprietary bootloaders" that it is not actually the case that all ARM platforms have proprietary bootloaders?

Every ARM-based chip I know of is using proprietary bootloader, even such "primitive" ones like the NXP/Philips LPCxxxx series.
If You know about some ARM-based chip that does not use bootloader (proprietary initialization code) - please let me know.

Are you saying or not that RISC-V is just like ARM when you look at the licenses/rights to make further products and that all the future bootloaders absolutely have to be - just like with ARM - proprietary and non free?? It is indeed really hard to get what you are saying here. It seems like you are using a lot of strong verbiage, but that the actual content of the statements are quite obscure. It seems as if you are dismissing RISC-V as "just like ARM", whereas the statement cited from Wikipedia is claiming that RISC-V is not just like ARM in many interesting ways. Is this not the case?

First, RISC-V is not a product yet - it works only in the simulators (not counting that reference chip, which can be seen only on a photo).
I'm saying that there's completely no guarantee that the hardware will be open in the final product.
And taking into account that guaranteed support for proprietary extensions, I do claim that RISC-V is just like ARM, which have also started as an educational project.

Regards.

[pylkko]

Every ARM-based chip I know of is using proprietary bootloader, even such "primitive" ones like the NXP/Philips LPCxxxx series.
If You know about some ARM-based chip that does not use bootloader (proprietary initialization code) - please let me know.

Ah, now I see what you want to say. You mean that when HOAS wrote:

The BeagleBoard (various versions) as well as the PandaBoard use the TI OMAP family of SoCs. These come with free startup software as well as free drivers for the peripherals.

That the "free startup software" contains what you call "proprietary initialization code". OK.

First, RISC-V is not a product yet - it works only in the simulators (not counting that reference chip, which can be seen only on a photo).

I believe that no one here has said that RISC-V actually "exists" in that sense.

I'm saying that there's completely no guarantee that the hardware will be open in the final product.

Well, can you see why I it appears to me that earlier you made a stronger claim? Did you not claim that it is guaranteed that RISC-V will not be open in the final product. The way I see it, that claim is completely different.

And taking into account that guaranteed support for proprietary extensions, I do claim that RISC-V is just like ARM, which have also started as an educational project.

Regards.

What about the license fees or mandatory non-disclosure agreements? Surely that is different in such a way that RISC-V is not like ARM??

[hthi]

Riscv must be subject to a critical evaluation too. I have followed and corresponded with lowric. http://www.lowrisc.org/. Lowrisc is attempting to manufacture a fully open source cpu. Not libre software. About the lowrisc cpu everything will be open source. A less restrictive license than fsf's. Bradbury from lowrisc has told me, that they wanted to make the mainboard open source. He expects it will be impossible from start because getting all peripherals open sourced is to big a task.

The riscv isa is open source. No one can make it closed source or demand license fees. A manufacture can or cannot decide to close derivatives. Consumers decide if they want to buy it. Any entity can manufacture a riscv fsf approved device. About arm and intel you cannot make that decision by yourself. Arm or intel has to approve which they do not. According to my information.

The google objection is valid. Google could if they wanted to put big resources into riscv. They have not. Thinkpenguin has said, google wanted to buy arm tablet hardware. One can assume, that google buys big numbers. Google insisted getting all the source code.. They did not get it. It could be google wanted to make their own closed software.

https://www.crowdsupply.com/eoma68/micro-desktop is a free software arm computer. I cannot see through if it became free software because of reverse engineering, illicit actions or agreements. Maybe it was a combination. The computer card is relevant because they are on the market. In general counting on getting free software arm and intel devices is not. The riscv path is better because from step 1 you can get open source or free software devices.

I consider riscv another example of the difficulty for free software or open source software to accumulate resources and produce software and devices.