2026-05-02 14:00 UTC
Replies (4)
-
@lain@lain.com 2026-05-02 15:05
@phnt@fluffytail.org @icon_of_computational_sin@mstdn.starnix.network @mangeurdenuage@shitposter.world wiat linux is now something i'm not allowed to install unless i want to get punched in the face?
-
@mangeurdenuage@shitposter.world 2026-05-02 14:04
@phnt@fluffytail.org @icon_of_computational_sin@mstdn.starnix.network Yes I'm aware of it. I posted about it previous to this post. >You cannot fully trust Linux nor GNU as vendor, because you cannot fully audit and prove it. Fallacious reasoning. Thank you for the bait. Please go kill yourself.
-
@icon_of_computational_sin@mstdn.starnix.network 2026-05-02 14:18
@phnt@fluffytail.org @mangeurdenuage@shitposter.world no, not really. Here I only refer to the Android bug above. But I would've published this "CopyFail" immediately. Linux licence claims NO WARRANTY and that really does mean NO WARRANTY.
-
@m0xEE@breloma.m0xee.net 2026-05-02 14:41
@phnt@fluffytail.org @icon_of_computational_sin@mstdn.starnix.network @mangeurdenuage@shitposter.world To me CopyFail revealed what a complete and utter clusterfuck modern Linux systems are — and it's not about the fact that vulnerabilities exists or how they are disclosed. Cryptographic routines running in kernel space exported to user space via a socket, what the fuck for? For a 0.1% performance gain?! And kernel module getting loaded when a user — any user at all, creates a socket of a particular type? For real?! This is the shit a ni~ ehm… user gets to deal with now?! And in RHEL AFAIK they went even further and built that module right into the kernel — so there is nothing to unload or prevent from getting loaded. Neither of these things should be possible in a system that is considered even remotely secure! :marseytabletired2: While security folk insist on putting everything in a container — these things exist!