GNU Linux-libre 4.12 Kernel Released, More Driver Deblobbing

[yoshi314]

i wonder how usable practically is that kernel. how many firmware free configurations are out there?

[starshipeleven]

4G mobile networks prevents any backdoors, WAN IP changes and it is somewhere else than in your site.

[Davidovitch]

Although I can't say I am tempted to use linux-libre for professional/personal use right now due to practical hardware compatibility issues, I do think these efforts are of the utmost importance. So go team!

It would be interesting to learn what hardware exactly is supported though. On the Arch Linux based Parabola distribution there are some clues to be found on what doesn't work (see the Parabola installation wiki). Maybe somewhere there exists a more elaborate list or compatibility matrix?

There a number of efforts out there that are attempting to de-blob the entire software stack (from microcode to firmware and beyond), and I think that especially the ladies and gentlemen of Pursim and their coreboot work deserve applause as well. A fully de-blobbed machine (from microde to kernel drivers), practically usable in a day-to-day context really is getting closer, huraaai :-)

[nadro]

This kernel should remove support for newer Intel and AMD CPUs and a lot of other devices too, because those devices require proprietary firmware. It was always funny for me when peoples say something like this "I don't want proprietary firmwares in my OS, because I want a freedom and full control over my hardware..." and the same peoples use proprietary firmwares integrated with UEFI/BIOS or even hardware There is just one difference between firmware integrated in hardware and in software -> when you have firmware integrated with hardware you can't update it in easy way and thats all. Even firmware integrated in UEFI/BIOS is little problematic, because you need to wait for update from a motherboard manufacturer like in the case with bugged HT on new Intel CPUs.

[Pavel]

Free software community do not have such resources and skills that Asus, Amd and Intel have. I do not want to replace fully working Asus bios with any coreboots or use Stallman made firmware in gpu drivers.

A student from Finland was clever than Stallman with Hurd, you remember.

*Claims free software community doesn't have skills and resources*
*Uses Debian*

*Points out how clever the studend from Finland was*
*Forgets that student never used Debian*

lol you are such a joke.

[starshipeleven]

When you look 4G router firewall logs, nobody attacks. It is easy to monitor network traffic too. If there is backdoor sending data, you would see it.

If you think DHCP requests coming from your own PC in the LAN network are attacks, then yeah. (you showed the logs some time ago and that's what the "attacks" were).

[Zucca]

Okay. They remove all blobs, but have they provided much replacements?
Do they just start coding from zero to replace the blobs or do they first disassemble those blobs and then write open ones based on the observations of the disassembly?

[tuuker]

only way to get firmware source codes is to infiltrate engineering department and take source, no other way.

[reepca]

Maybe somewhere there exists a more elaborate list or compatibility matrix?

h-node.org has a database of which devices are compatible with free software. I can't say much about the reliability of that database beyond what I've tested - contrary to what it says, the R5 230 and radeon 6450 do not work with linux-libre, at least I didn't manage to get them working (one output, no acceleration). Eventually I got an 8400GS, which works, so I can finally use my two screens independently. Note that with incompatible graphics cards you'll get very basic support, usually - my a10-7850k could show one screen with no hardware acceleration, but at least it got the full 1600x1200 resolution.

[RealNC]

steal our freedom

It's statements like this that make the FSF sound like a cult to me :P

*raises hands, slowly walks backwards, saying "ooookey, ooookey"*