Announcement

Collapse
No announcement yet.

Longene: The Linux Kernel With Windows Support

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Longene: The Linux Kernel With Windows Support

    Phoronix: Longene: The Linux Kernel With Windows Support

    This morning after providing benchmarks of FreeBSD with Linux binary compatibility for gaming, which allows unaltered 32-bit Linux binaries to be executed seamlessly with the FreeBSD kernel (and in a rather fast manner), I was reminded on Twitter about another interesting project: Longene. Longene is a "Linux Unified Kernel" that attempts to implement Microsoft Windows APIs within the Linux kernel. In other words, Windows binary compatibility for the Linux kernel, including for Windows device drivers...

    http://www.phoronix.com/vr.php?view=OTg4NQ

  • #2
    Torvalds once expressed his endorsement of this stuff, as it enhances interoperability. However, the code is probably not remotely up to kernel standards.

    Comment


    • #3
      I'm more interested in how they will resolve things like, making windows video drivers work with Xorg.
      I can think of ways to resolve it.. but they are all kind of nasty.

      Comment


      • #4
        Kernel-level Windows compatibility? I shudder to think of the security risks.

        Comment


        • #5
          Longene is not anything but ReactOS code merged into the Linux kernel...

          Comment


          • #6
            Originally posted by mrugiero View Post
            Longene is not anything but ReactOS code merged into the Linux kernel...
            You got some code snippets to prove it?
            Or to quote wikipedia CITATION NEEDED

            Comment


            • #7
              Originally posted by insanemal View Post
              You got some code snippets to prove it?
              Or to quote wikipedia CITATION NEEDED
              Hmmm... seems like I was wrong, I was sure to remember I read it in Longene's or ReactOS forum, but since I can't find it, I suppose I was wrong, sorry.

              Comment


              • #8
                Well, I've found this:
                http://www.reactos.org/forum/viewtop...longene#p67337

                Although it is not an official source (because it's not said by Longene devs), this might be the cause I thought it.

                Comment


                • #9
                  Honestly, I think this looks like a solution in search of a problem. For Windows compatibility in Linux we have Wine, which, IMO, doesn't suffer from any significant performance problems this is likely to solve. If you want an open-source Windows-like OS, ReactOS is probably a better idea.

                  It seems like this is attempting clumsily to shoehorn together two architectures at a level they are manifestly not designed for.

                  I wish more effort would go into Wine and ReactOS, honestly. This seems like a distraction.

                  Comment


                  • #10
                    Longene is a waste of time.

                    I tried it a year or so ago, at a time when this article probably should have been written, being as there were a few discussions on Wine-dev List about it, articles on the web, and benchmarks. it seems funny, that Phoronix would publish an article randomly about it now, it doesn't seem to me, that development is happening very fast, nor has this project gained any popularity.

                    benchmark article here:

                    http://www.linuxtoday.com/developer/2010060303135OSKNDV

                    Myself, when i tried Longene, i wasn't impressed, at all. I took a look at it, because i had thought it might run a bit quicker, less latency than with normal Wine. (as in theory it should be a bit quicker, latency wise - read the benchmark article, to see what i mean). I personally didn't find any advantages over using Wine, and found many many disadvantages to being stuck with a 'generic' 2.6.34 kernel, that has some ugly hacks and other crap enabled/disabled...

                    from what i remember the Wine-devs didn't have much faith in the project or code either. I don't think the ReactOS people did, either.

                    I don't think Longene will ever become popular, I think focusing on Wine is better.
                    Last edited by ninez; 09-08-2011, 12:12 AM.

                    Comment


                    • #11
                      Originally posted by ninez View Post
                      I don't think the ReactOS people did, either.
                      If anyone cares about, no, they did not, and they don't either now.
                      Source: I check the forums everyday, most members thinks it is a waste of time.

                      Comment


                      • #12
                        Originally posted by insanemal View Post
                        I'm more interested in how they will resolve things like, making windows video drivers work with Xorg.
                        I can think of ways to resolve it.. but they are all kind of nasty.
                        Nasty is an understatement

                        Windows drivers use a far different interface model than Xorg drivers do, and the Windows driver interfaces are not fully documented so it will require a lot of reverse engineering to find out what undocumented Windows driver interface registers and calls that such drivers commonly use.

                        Comment


                        • #13
                          Originally posted by DeepDayze View Post
                          Nasty is an understatement

                          Windows drivers use a far different interface model than Xorg drivers do, and the Windows driver interfaces are not fully documented so it will require a lot of reverse engineering to find out what undocumented Windows driver interface registers and calls that such drivers commonly use.
                          Where did you get that? The device driver interfaces are extensively documented in Windows. They employee people who do nothing but write documentation (66MB compressed).

                          How do you think hardware manufacturers release drivers for every oddball device out there, by reverse-engineering the kernel?

                          Comment


                          • #14
                            Originally posted by DeepDayze View Post
                            Nasty is an understatement

                            Windows drivers use a far different interface model than Xorg drivers do, and the Windows driver interfaces are not fully documented so it will require a lot of reverse engineering to find out what undocumented Windows driver interface registers and calls that such drivers commonly use.

                            You don't implement support for all drivers. Windows has userland drivers with very well documented interfaces for many things so you can implement support for those sorts of thing.

                            example:
                            http://en.wikipedia.org/wiki/Network..._Specification

                            For video acceleration the ideal way from a current perspective would be implement trackers for DirectX and GDI. Then you composite Windows applications into a Wayland desktop. Just like how you are going to do it for your X applications. That way you can get native support for a multitude of APIs without having to support many multiple driver stacks.

                            The alternative being trying to run X graphics acceleration driver + DRI graphics acceleration driver + Windows graphics acceleration driver + OpenCL acceleration driver + video codec acceleration driver etc etc.

                            Comment


                            • #15
                              It's Wine.

                              They are using Wine. Check the screenshots in the application compatibility list on their web site.

                              Comment

                              Working...
                              X