Spoiling Linux Kernel with "sanctioned" code

(printserver.ink)

72 points | by ValdikSS a day ago ago

13 comments

  • flashmozzg 17 hours ago ago

    Is there a CVE for this?

  • mike_hock a day ago ago

    Obvious attack vector for Russia: Submit fixes to severe bugs that can't realistically be fixed any other way.

  • thefounder a day ago ago

    I guess the Russians will have to learn the Chinese way and perhaps the Chinese language as well?

  • 1attice a day ago ago

    I've been thinking lately that what underpinned the FOSS golden age was not actually decentralized VCS and high-quality forges, nor even ZIRP, but rather peacetime.

    After a period of branches and patchsets, full national hard forks are going to become de rigeur, and linux-derived OSes across the world are going to bloom necessarily, as we no longer have the kind of ambient trust required to collaborate across borders.

    Look forward to Euro-linux, Sino-BSD, and I guess probably some sort of GCC-area build as well.

    Patches will be accepted across national boundaries with only the highest scrutiny, which itself will likely be provided by nationalized AI platforms.

    Gods I hate this era

    • V__ 3 minutes ago ago

      OpenSuse is (or will be) "Euro-Linux".

    • gaiagraphia 6 hours ago ago

      This is a great thing for innovation though? Nations/blocs protecting their tech interests will result in more jobs to go round in the industry, more unique ideas, and less centalisation, surely?

      The globalised, hyper-centralised world is a bit boring, tbh.

      • 1attice 3 hours ago ago

        I forecast that you will not be bored, and may have other, stronger feelings. Ask Ukrainians

  • gmerc an hour ago ago

    Perfect usecase for AI, by US legal doctrine, copyright is gone after you feed it through and so should sanctions /s

  • robobully a day ago ago

    This post is apparently not publicly shown on the main page for some reason.

    • ValdikSS a day ago ago

      Why should it be? It has low rating (yet).

  • _user_account 20 hours ago ago

    Yeah, it sucks.

    > This adds ~1ms latency per transfer cycle for rapid bidirectional communication which leads to half the USB 1.1 speed for smaller packets at best.

    Still, I don't think this patch should be applied /for everyone/. Maybe compile out-of-tree and load as a kernel module, if possible?

    • M95D 34 minutes ago ago

      I still have a MB with just a USB 1.1 controller. I would hate it if the USB stopped working after this fix. I think a config option for the delay would be best.

    • ValdikSS 18 hours ago ago

      The patch removes this latency and improves transfer speed, without any drawbacks.