“…and the restricts essentially come down to no reverse-engineering/disassembly of the binaries…”
if my understanding is right, purism can’t keep going to reverse engineering the intel blobs, so what will be the purism choice? keep using corebot with binary blobs and be less purist, or do they start to move on other architecture?
if my memory is good i remember an answer from @todd-weaver on reddit, writing something about x86 is the present, but other architecture are considered for the future, and i think they are already working on it, i don’t know if they are just planning it or if they will have some work already done
any words on this topic from purism? could you share a bit more on it?
I believe the hope is to move to RISC-V in the future. If the FSP situation does indeed mean Purism is barred from continuing their reverse engineering work, I suspect the initial status will just be that Librem laptops stay where they are freedom-wise. Maybe there would be a switch to AMD or ARM, but that starts the whole reverse engineering process over again, and those companies might also follow Intel and put their firmware under similar licenses.
I think fully-fledged RISC-V CPUs are a few years away at the least, but, I did recently come across this:
I wonder if Purism could partner with them to start getting RISC-V CPUs into Librem laptops
risc-v is really interesting, but now seems not enough powerfull for consumers, i hope they will develop something more powerfull asap
now there is power9 that is also completly open, the only 2 issue are licence fee and power consumption, for the first issue, should be nice if purism could talk with ibm trying to follow the same purpose to try to push power architecture on consumer area too, with a lower licence cost, for the second issue, maybe a larger battery/laptop could solve the problem, maybe will be not thiny like macbook, but who cares? we are there not for shiny jewelery, but for freedom
While this new license is 100 times simpler than the previous license, there is nothing new or changing with our goals and advancements effected by the license change.
We will be pushing on all fronts to advance freeing the hardware; removing the FSP, working with Intel to liberate it, working on moving to RISC-V (when it catches up), etc. It is a multi-faceted approach to solving this problem, since no single way is guaranteed.
what means that you are working with intel to liberate it? they have the source code, and if they want it, they should just release the source
anyway i think is a good move to have different approach to solve the problem
do you have any info about risc-v roadmap? i found nothing about it, they are not really communicative, i also mail them asking some info but i got no answer, i hope they are more friendly with partners
edit: what about power9? seems more ready than risc-v and there is already a debian version for ppc64le
A lot of the FSP (silicon init anyway) is already released as open source with a BSD license (see https://github.com/tianocore/edk2-platforms/tree/devel-MinPlatform/Silicon/Intel/KabylakeSiliconPkg) and from my understanding more parts are getting periodically released. The question is “what is open and what isn’t, and is there anything that is closed that is actually required”. That’s one of the things I’ll hopefully be working on in the near future.
If you know where to look, the code that was voluntarily removed was mirrored in a lot of places. I won’t post it here but it can be found if you ask around or search related Reddit threads.
I’m glad that you can at least privately reverse it, maybe…either way I do miss your blog posts if nothing else because of what I can learn from you.