Announcement

Collapse
No announcement yet.

Native Linux Kernel Module Is Out For Microsoft exFAT

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

  • #46
    Originally posted by birdie View Post
    That's funny - people want to destroy me because I see real Linux problems.
    Is this page supposed to be a joke?
    One point even complains that there's adobe flash video tearing... so? Adobe dropped support for Linux, what the hell do you want Linux developers to do to fix that?
    Also,
    No high level, stable, sane (truly forward and backward compatible) and standardized API for developing GUI applications (like core Win32 API - most Windows 95 applications still run fine in Windows 8 - that's 18 years of binary compatibility). Both GTK and Qt (incompatible GTK versions 1, 2, 3 and incompatible Qt versions 2, 3, 4, 5 just for the last decade) don't strive to be backwards compatible.
    yes, more GUIs should strive to be like Win32, known to be really easy to work with due to its awesome backwards compatibility /sarcasm
    The backwards compatibility is why Win32 is garbage and they had to make winforms for .NET.

    I can point to almost anything related to software development and its flaws will almost be entirely with backwards compatibility, you'd think an area that just started existing in the past half-century would be fine with fast progress.
    Last edited by peppercats; 06-30-2013, 11:11 AM.

    Comment


    • #47
      Originally posted by erendorn View Post
      No. The device is performing read and write operations at block level, it is completely agnostic to the filesystem, and both a sd card and a sd card reader are sold without patent license for any filesystem (do you really think all sd cards are sold with a license for each and every file system in existence?). If your camera can read and write exFAT, it needs a license. If your OS can read or write exFAT, it needs a license.
      No. If my OS reads and writes to existing exFAT device, it does not need a license.
      If I need an OS to be able to create and modify exFAT on media of my need, then I need ms license.
      Who is gonna need this FS outside of his camera/usb stick? No one.

      Yes, SD cards ARE sold with license in filesystem they are formatted from factory.

      Comment


      • #48
        Originally posted by RealNC View Post
        FUSE is the way to go for this file system. If something can't be included in mainline, there's no point in having a kernel module. NTFS-3G works very well for a lot of people and it's not tied to a specific kernel version. There's no reason exFAT can't do the same.

        I suspect though that the whole point of this project is about the author simply having fun in implementing it in the kernel. And that's a perfectly valid reason to do this.
        Originally posted by chithanh View Post
        A disadvantage of FUSE is that it results in a high CPU load. This is less of a problem for fast desktop systems, but annoying on smaller embedded CPUs. It was said in this thread that the exFAT kernel driver originally comes from Samsung tablet kernels.
        FUSE change absolutely nothing. FUSE is userspace filesystem, this way the FS loads in userspace, not in kernel. From kernel developer point of view, they don't have to support it as module; from maintainer point of view, they don't have to port it to newer kernel versions. Convenience at exchange of speed.. doesn't change anything else.

        Ofc its much better to have module kernel version. And ofc the thin line HOW you use the FS - on certified device, which is legal; or on any device, which is not legal, does not affect kernel anymore if its FUSEd, but then no difference anyway. In my humble opinion, no one is going to use this FS outside of its own device, so I don't see the legal problem here.

        Comment


        • #49
          Originally posted by brosis View Post
          No. If my OS reads and writes to existing exFAT device, it does not need a license.
          If I need an OS to be able to create and modify exFAT on media of my need, then I need ms license.
          Who is gonna need this FS outside of his camera/usb stick? No one.

          Yes, SD cards ARE sold with license in filesystem they are formatted from factory.
          The sd card license does not give you licenses for other devices.
          That's why your camera has a separate license.
          That's why your PC needs a separate license (if it can use exFAT).
          That's why Samsung paid a license to add exFAT capability to their tablet (using the very module from which comes the code leaked here).
          That's why android manufacturers push for no sd card and MTP (usb raw access= FAT filesystem on the phone, patent fee. usb MTP access= any filesystem on the phone, no patent fee).

          I'll rewrite it once again for you, it's very simple:
          you distribute an exFAT module to or from a country where the related patents are enforceable => you need a license.

          Comment


          • #50
            Microsoft can stuff their patents

            Originally posted by birdie View Post
            This driver is Microsoft's intellectual property, it will be taken offline and you will receive a C&D letter from Microsoft. Good work.

            Legally no one can even use it unless you pay ExFAT tax (which is $2 per device if I'm not mistaken).

            Shame.
            Therefore, this module needs to become available through Pirate Bay as a torrent, and/or mirrored in countries hostile to the US and to software patents. As anyone who ever watched a pirated movie knows, there are workarounds against the legal system. There is even an onion-routed bittorrent network that works like Tor does but is designed for filesharing growing as we speak. I haven't used it mostly because I ignore Hollywood's content almost completely.

            As a person with no assets vulnerable to any lawsuit, I would have exactly NO hesistation about using "illegal" software to extract data from a device enslaved to Microsoft. As a political enemy of Microsoft, I would see that as extra satisfaction. I keep things like libdvdcss on all my systems, whether or not they have optical drives. Since I also keep all my systems encrypted, not even a raid could prove their presence by a court-usable standard anyway. No way in hell I would pay a dime to Microsoft, any more than I would ever pay a Getty extortion note. All they'd ever get from me is a nasty fight, over patents my whole political movement advocates abolition of, not only for software but also for seeds and living things.

            Comment


            • #51
              It has to be GPLv2 for one simple reason

              Samsung are distributing their devices (or specifically tablets) with kernel mode exFAT support - this driver calls into the VFS layer etc. and can be said to derive from the Linux kernel in a meaningful way. Therefore, in order for Samsung to legally ship this binary module pre-installed on their devices, it would have to be licensed as GPLv2.

              It may be covered by patent law in certain countries, but purely from a copyright perspective, whether or not Samsung signed an NDA with Microsoft, they cannot legally distribute this driver on their devices unless they release the source for it under GPLv2 (which it appears they have). This is the reason so many Android closed source driver blobs are user space and use a transition layer - it can then be argued they do not derive from the kernel source and so don't have to be GPLv2.

              The case is clear - copyright wise - this has to be GPLv2 or Samsung are distributing it illegally, regardless of any agreement with Microsoft.

              Comment


              • #52
                Originally posted by Luke View Post
                Therefore, this module needs to become available through Pirate Bay as a torrent, and/or mirrored in countries hostile to the US and to software patents.
                Uh, dude. It is already mirrored several times on github. Whoever has rights to this code seems to have no interest to remove it from there. The Streisand effect needs only kick in when github takes down the repository.
                Originally posted by PhilPotter View Post
                The case is clear - copyright wise - this has to be GPLv2 or Samsung are distributing it illegally, regardless of any agreement with Microsoft.
                Unfortunately this does not make the module license GPLv2 automatically. If Samsung distributes the module in violation of the license, the kernel copyright holders can demand that Samsung stops the distribution and pays damages. But nobody can demand that the license becomes GPL (as some other commenters in this thread seemed to suggest).

                Comment


                • #53
                  Of course, you are correct regarding automatic status - I thought I implied this with my statement but obviously not. Apologies.

                  Comment


                  • #54
                    Some of you may already know this but the license specified in the source code this originally comes from clearly states "Samsung Proprietary". Assuming the code wasn't stolen by an employee, this is likely a case of Samsung mistakenly "open-sourcing" their proprietary code. It has happened at least once or twice before so it seems to be the most likely reason for this code being available.

                    Comment


                    • #55
                      This code was leaked by somebody inside Samsung several months ago, the first instance I'm aware of is here: http://forum.xda-developers.com/show...6#post41330066

                      Considering rxzr's demeanor to my issue on Github, him deleting my reply to his absolutely misinformed and wrong response, and blocking me from the repository; I'll take the freedom to reply here:

                      I've posted this code under the "GPL tag" to let the techy websites spread the word about the driver, rather than treating it as a leak. I repeat: it was publicly available on github, and probably still is.
                      Just because there are a bunch of idiots who do not understand software licensing and still have had the leaked code on their repositories does not mean it is free, you can strip a license from it as you see fit, and throw a GPL tag on it and publish it as some kind of original work. It is nothing less than warez.

                      Because of the linux kernel being licensed under GPL, and Samsung using that kernel in their android devices along with this driver, I probably could've legally gotten this source code by the first request to them.
                      No, GPL applies to linker linked in code at compile time to the kernel. Samsung have distributed their exFat modules on the Galaxy line of phones for over a year and a half, this includes tens of millions of devices. Believe me that you cannot just request it because you have no right to it, the mobile device community has discussed this plenty over the last year.

                      You also could complete the driver written by Ogawa, using this source code, and then suggest including that driver into the kernel tree and get very famous.
                      The idea that this code would be somehow accepted into mainline is of pure na´vety and stupidity. This is not a free-for-all do-whatever-you-want ecosystem.

                      Comment


                      • #56
                        Originally posted by AndreiLux View Post
                        This code was leaked by somebody inside Samsung several months ago, the first instance I'm aware of is here: http://forum.xda-developers.com/show...6#post41330066
                        Here is an earlier reference. It's the earliest one I've found and likely where this code was forked from. It fits the description of a "Samsung tablet kernel".

                        https://github.com/kcrudup/gt-n8000-...535b68ea8b77fe

                        Comment


                        • #57
                          Originally posted by AndreiLux View Post
                          This code was leaked by somebody inside Samsung several months ago, the first instance I'm aware of is here: http://forum.xda-developers.com/show...6#post41330066
                          Thank you.

                          This is exactly what I was talking about in this thread - yet Open Source fanatics laughed at me and said I was wrong - this driver is published legally. Fanaticism turns people into total imbeciles.

                          Comment


                          • #58
                            Originally posted by blunden View Post
                            Here is an earlier reference. It's the earliest one I've found and likely where this code was forked from. It fits the description of a "Samsung tablet kernel".

                            https://github.com/kcrudup/gt-n8000-...535b68ea8b77fe
                            I remember that repository. He seems to update official sources with mega-commits based on official source drops. If Samsung had mistakenly released it, it would be found in one of the source packages; http://opensource.samsung.com/recept...rchValue=n8000 , unless they pulled it. FYI: it's not to be found in any of these packages.

                            Doesn't really confirm the user as the source of the leak though.

                            Comment


                            • #59
                              Since Samsung seems to be relatively friendly to open source and this doesn't hurt them (and the code is already out there), they might as well GPL it because of all this. I hope they do, so it could be pulled into the official kernel.

                              Comment


                              • #60
                                Originally posted by mrugiero View Post
                                Since Samsung seems to be relatively friendly to open source and this doesn't hurt them (and the code is already out there), they might as well GPL it because of all this. I hope they do, so it could be pulled into the official kernel.
                                There's no way this code can be incorporated - it's Microsoft's IP. Only clean room reverse engineering can land in mainline.

                                Comment

                                Working...
                                X