Login
You're viewing the functional.cafe public feed.
  • Nov 7, 2026, 8:17 PM

    Sad to see Asahi out there burning bridges for no reason.. relicensing m1n1 and new drivers under GPL2-or-later, seemingly without the consent of all prior contributors, under the guise of "defending against AI bad actors", seems entirely disingenuous and misguided (and also probably illegal, btw). :flan_disappointed:​

    💬 3🔄 1⭐ 0

Replies

  • Nov 7, 2026, 8:47 PM
    Don't we have Corellium to thank for doing the heavy lifting of getting Linux running on Apple Silicon in the first place?

    I've had my own uhh, "challenges" with at least one Asahi attempted contribution (e.g. https://github.com/AsahiLinux/asahi-installer/pull/258#pullrequestreview-1933806739) They don't like my PR? Whatever, their bailiwick, their obsolete with known vulnerabilities version of Python dependency. I guess I am not too surprised they may be steering things in yet another dire direction?
    💬 1🔄 0⭐ 0
  • Nov 7, 2026, 8:51 PM

    @teajaygrey No, Asahi is at the very least responsible for very publicly doing the initial platform bring-up and upstreaming of Linux kernel drivers.

    I don't think anything Corellium did was upstreamed.

    I actually had a relatively positive opinion on the Ashai project and developers before this policy about-face. Oh well. :flan_tired:​

    💬 1🔄 0⭐ 0
  • Nov 7, 2026, 9:02 PM
    Ah, maybe Corellium didn't get anything upstreamed?

    Nonetheless, despite Asahi Linux putting out the call for reverse engineering Apple Silicon in December of 2020 (at least according to https://en.wikipedia.org/wiki/Asahi_Linux#History), I seem to recall Corellium beating them to the punch with headlines in January 2021, e.g. https://www.macrumors.com/2021/01/20/corellium-linux-m1-macs/

    Do Asahi devs claim they didn't derive any sources from https://github.com/corellium/linux-m1 with their upstreaming work as well? Seems, unlikely? Though far be it for me to waste more time doing forensic source reconstruction and retroactive lkml tracking to determine provenance. That sort of thing is what attorneys tend to charge hundreds of dollars and hour for and I prefer to avoid such realms. IIRC, Corellium also faced the brunt of Apple's legal team previously (albeit related to an iOS simulator I guess? e.g. https://appleinsider.com/articles/19/08/15/apple-sues-virtualization-firm-corellium-for-selling-ios-and-itunes-knockoffs); something that Asahi Linux seems to have mostly avoided AFAIK.
    💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Nov 7, 2026, 8:48 PM

    @brynet Would GPL2-or-later even do anything in terms of keeping out the AI fascists, and keeping the work out of their training sets? My impression is that the fascists will do what they please, laws and licenses be damned, sure, but even in a theoretical world where the law applies to everyone, would it do anything?

    💬 1🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Nov 7, 2026, 8:55 PM

    @brynet Yeah, that's what I figured. I was just trying to see if it made sense even from a naive standpoint. Like, if I still believed in the power of licenses to hold the AI fascists at bay, would it even be reasonable for me to believe the GPL would achieve this? I feel like it wouldn't...

    💬 0🔄 0⭐ 0
  • Nov 7, 2026, 9:07 PM

    @brynet This is bad. When fvwm first came into being, its licence predated the GPL, so it was not clear what the over licence of fvwm was. Over the course of its development, it's been widely assumed to be the GPL.

    A few years ago now, I contacted some very old authors in the early code which had non-GPL copyrights on them, so that I could relicence the code.

    It worked -- and managed to dredge up some old memories in some people who did not realise fvwm3 is still alive.

    Heh.

    But you can't just go around changing licences like this, that's insane.

    💬 0🔄 0⭐ 0