No announcement yet.

Concerns Raised Over The "New" NTFS Linux Driver That Merged Last Year

  • Filter
  • Time
  • Show
Clear All
new posts

  • #21
    Originally posted by birdie View Post

    "ghosting all of us"? The fuck is wrong with you? Paragon has given us a great working driver which needs minor fixes here and there. OK, Konstantin has stepped away from maintaining it, so fucking what? No one in the fucking community can step up? Just "I vote to remove"?
    So what? Basically means he or Paragon released something relatively substantial and then doesn't respond to comment. That isn't a good precedent or normal that should be established for the kernel of the predominant open source operating system. Here y'all go, buh bye ** cricket noises **

    And what other options are there? Either someone else or Company steps up or it gets removed. That's it.

    For years the whole community asked for a native driver, the one was given with no strings attached, oh, god, no one actually cares or wants to continue developing it.
    You act like there's a giant pool of people knowledgeable and capable to work on NTFS on Linux.

    If anything, this just goes to show us the the kernel might need a long term maintainer agreement that committers have to sign if they want to go mainline. Let them keep rocking with DKMS or whatever TF NVIDIA does if they don't want to accept the legally non-enforceable and done-in-good-faith stewardship agreement where you agree to maintain your code with a clear line of communication and agree to either phase out/depreciate code or find a new steward if you're unable to continue maintaining the code.


    • #22
      Originally posted by skeevy420 View Post
      If anything, this just goes to show us the the kernel might need a long term maintainer agreement that committers have to sign if they want to go mainline.
      And Linux mostly has this (for major new subsystems or features(*)). However, what is clearly being reported as an unusual case here is that the company that promised ongoing support is not following through as some expected. Until there is a response from paragon we simply do not know the real why (although that has not stopped the FUDders from posting).

      (*) Which, in practice, is also why some interesting drivers/features/subsystems never get mainlined, as there is no corporate support. Linux has gone corporate, no matter what some might fantasize about it being some community based effort.


      • #23
        Konstantin Komarov is a slavic name. While Paragon's Head Office is situated in Germany, Paragon also has offices in the US, Russia and Japan.

        There could be a very obvious reason that Mr. Komarov isn't responsive right now. This reason can incidentally also be quite orthogonal to the legitimate concerns of the kernel developers.


        • #24
          I do not see the issue, at least the way it is presented here.

          Sure maybe there haven't been updates to the code in quite some time, but are there any inherent issues with the code itself?

          If a software works as intended/does a good job, I will choose it over software that is "maintained" but worse.

          I don't disagree that unmaintained code is not ideal, but also this is open source community code and not corporation shit.

          I do not see a reason for removal, unless there are some hard issues. Everybody can choose wether to compile the feature or exclude it from compilation.

          ​​​​​NTFS3G or whatever it is called is certainly not a better alternative, so not sure how dissmissing such a great and valuable contribution by Paragon would be a good idea.


          • #25
            Originally posted by birdie View Post

            NTFS-3G is known to corrupt volumes when downloading torrents. Maybe it's a very special use case, I dunno.
            I don't think it is, many people use torrents. Has it been reported to the mailing list or in a github issue? I can't see it in the list of open issues on github. It's certainly worth investigating as it sounds quite serious.


            • #26
              Maybe we could associate them as a distant relative of Hans Reiser, then their code could be pulled and never be allowed back in?


              • #27
                Originally posted by microcode View Post
                Gotta say, NTFS-3g is basically fine for most people I think. The performance is good, and I doubt there are many demanding NTFS as a boot volume filesystem on Linux.
                Maybe for read only mounts. But I get corrupted files whenever I do a bunch of writing.

                Winbtrfs work better for me, but it doesn't let you write to a Windows C: drive, and its performance isnt very good (albeit still much better than ntfs-3g).
                Last edited by brucethemoose; 26 April 2022, 11:22 AM.


                • #28
                  Originally posted by CommunityMember View Post

                  So, to follow your logic, others should be able to tell *you* how to spend your revenue (income)?
                  No. But should stop their hypocrite marketing "we love OSS" bullshit. Because they don't.


                  • #29
                    Originally posted by Terrablit View Post

                    Agreed. Other valid complaints:

                    * Still no NTFS utility source despite claims that it's coming. FS utilities are a very important part of being a supported filesystem in the mainline kernel. How are you supposed to check the filesystem without the utilities?
                    * Maintainer (Komarov) doesn't review patches - they either directly apply them or ignore them.
                    * The maintainer doesn't send their own patches for review either - they just directly apply them to their tree. They also don't submit PRs for the work they've done. Everyone who writes code in the kernel needs to get someone else to sign off.
                    * Basically nothing's made it into the kernel for 6+ months despite other people trying to contribute and submit patches because this person is the official maintainer and isn't following protocol or communicating.
                    * Maintainer can't handle things like branching and tagging in git to version and mark stable points.

                    Nobody wants to get tossed into the middle of someone else's work and made responsible for it, and that's what Paragon has done. I mean, props for finally contributing the driver and accommodating the kernel dev requests to get it mainlined. It was a milestone. But the job isn't done, and whoever's de-prioritized this at Paragon needs a reality check. Which is hopefully coming, now that there's some attention on their bad communication skills. Nothing like getting your company name in the news for code littering to clean up your act.
                    Yeah, that Home is why I'm still on ext4: fancy features mean nothing when I can't retrieve my data from something gone wrong.


                    • #30
                      one thing i've learned about open source projects that so many people use it simply as a tool for free labor. either use the ideology behind open source to rile people up to get them to willing to work on something for free while you sit back and soak in the free labor or just release something as open source to wash their hands clean and walk away. leaving it up to other people.