Announcement

Collapse
No announcement yet.

Native Linux Kernel Module Is Out For Microsoft exFAT

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

  • Native Linux Kernel Module Is Out For Microsoft exFAT

    Phoronix: Native Linux Kernel Module Is Out For Microsoft exFAT

    A Linux developer has released a native Linux kernel module for implementing Microsoft exFAT file-system support within the Linux kernel. This is different than earlier exFAT Linux modules that were implemented via FUSE in user-space...

    http://www.phoronix.com/vr.php?view=MTM5NTQ

  • #2
    This module right now will work on the Linux 3.8.11 kernel but the module's author says right now there are issues if using the Linux 3.9 kernel or newer.
    I.e. distros won't use it due to legal issues, enthusiasts won't use it because it's not compatible with mainline.

    I.e. there's a great chance this module will soon be forgotten and forsaken.

    Besides who will verify its source code? What if this module inadvertently messes with the kernel memory layout and allows exploits?

    Comment


    • #3
      Originally posted by birdie View Post
      I.e. distros won't use it due to legal issues, enthusiasts won't use it because it's not compatible with mainline.

      I.e. there's a great chance this module will soon be forgotten and forsaken.

      Besides who will verify its source code? What if this module inadvertently messes with the kernel memory layout and allows exploits?
      LEARN TO FIX PROBLEMS, ALL YOU DO THROUGH YOUR LIFE IS COMPLAINING LIKE A GIRL.

      (sorry for caps to everyone)

      Comment


      • #4
        Originally posted by brosis View Post
        LEARN TO FIX PROBLEMS, ALL YOU DO THROUGH YOUR LIFE IS COMPLAINING LIKE A GIRL.

        (sorry for caps to everyone)
        He does have a point, Brosis. The only distros that will pick this up are the ones without a legal entity behind them that CAN be sued, and the ones that are based out side of the US that CANT be sued for patent infringement on software.

        And until this module is properly fixed for current-mainline and STAYS fixed, even enthusiasts won't pick it up.

        Comment


        • #5
          I can think of a few entities who may find this useful; those entities who currently license exfat for linux (android) cell phones and similar, and yes, MS does have their own module. This probably would provide a slightly greater flexibility for those entities than the MS blob would.

          Edit: Also, the git log appears to be somewhat suspicious;
          https://github.com/rxrz/exfat-nofuse/commits/master
          -- was this entire thing written all today over the last several hours?
          Last edited by droidhacker; 06-25-2013, 03:46 PM.

          Comment


          • #6
            Originally posted by droidhacker View Post
            I can think of a few entities who may find this useful; those entities who currently license exfat for linux (android) cell phones and similar, and yes, MS does have their own module. This probably would provide a slightly greater flexibility for those entities than the MS blob would.

            Edit: Also, the git log appears to be somewhat suspicious;
            https://github.com/rxrz/exfat-nofuse/commits/master
            -- was this entire thing written all today over the last several hours?
            If you'd thought for a second you could have realized that the author could create the code on his own PC and then he simply uploaded the final version to GitHub. ;-)

            Comment


            • #7
              Originally posted by Ericg View Post
              He does have a point, Brosis. The only distros that will pick this up are the ones without a legal entity behind them that CAN be sued, and the ones that are based out side of the US that CANT be sued for patent infringement on software.

              And until this module is properly fixed for current-mainline and STAYS fixed, even enthusiasts won't pick it up.
              Thats completely correct, but its just the same thing as with ntfs. It illegal to use without royalty fee, but thanks to microsoft monopoly (and with exFat - for using their monopoly to push proprietary file system as a standard) people should really give a damn! Proprietary protocols should never ever be elevated into standards.

              Although I completely agree on non-inclusion, because we don't want microsoft tax. A big fat warning for those who look up and another criticism section in already huge article "Criticism of Microsoft" would also be nice. But I applaud to the efforts of this project, because it destroys barriers that MS is so fond of building up; and I will applaud every single bug fixed, because it will destroy birdie aka "Check my Linux sucks list"-guy ego badly.

              Ofc. removing exFat from standard or invalidating its patent claims would be even nicer.

              Comment


              • #8
                That's funny - people want to destroy me because I see real Linux problems.

                Instead of trying to collaborate, trying to work as a whole, trying to change the way Linux is (mis)developed, trying to make Linux stable and usable for a long periods of time (you can work with the same Windows version for great many years before upgrading to a next version) - they blame me.

                Whoa!

                Comment


                • #9
                  We already have Universal Disk Format for sharing read/write data between OS X, Windows and Linux. It's even kind enough to enforce that all file names are valid Unicode, unlike most Linux file systems where they're just a byte string without NUL or slashes.

                  I don't think having fast exFAT support is really going to matter.

                  Comment


                  • #10
                    Legal issues?

                    I'm not convinced the exFAT patents are much of a blocker to mainline inclusion. Last I checked, the mainline kernel includes NTFS and FAT32 drivers and Microsoft has patents on both (and have successfully sued or extracted licenses from Android manufacturers over FAT32).

                    EDIT: Apparently they don't have patents on NTFS, but FAT32 still stands.
                    Last edited by gigaplex; 06-25-2013, 11:33 PM. Reason: Error on NTFS legal status

                    Comment


                    • #11
                      Originally posted by strcat View Post
                      We already have Universal Disk Format for sharing read/write data between OS X, Windows and Linux. It's even kind enough to enforce that all file names are valid Unicode, unlike most Linux file systems where they're just a byte string without NUL or slashes.

                      I don't think having fast exFAT support is really going to matter.
                      It is the standard partition format for large SD cards (more than 32 GB). Devices are not required to, and often don't, support any other format. So if you want to use ordinary, off-the-shelf, large SD cards, this is your only option.

                      Comment


                      • #12
                        Originally posted by TheBlackCat View Post
                        It is the standard partition format for large SD cards (more than 32 GB). Devices are not required to, and often don't, support any other format. So if you want to use ordinary, off-the-shelf, large SD cards, this is your only option.
                        There are really common SD cards that can't be reformatted?

                        Comment


                        • #13
                          Originally posted by birdie View Post
                          That's funny - people want to destroy me because I see real Linux problems.

                          Instead of trying to collaborate, trying to work as a whole, trying to change the way Linux is (mis)developed, trying to make Linux stable and usable for a long periods of time (you can work with the same Windows version for great many years before upgrading to a next version) - they blame me.

                          Whoa!
                          You don't see real problems - your list is crap threefold:
                          1) The attitude of the list itself is to COMPLAIN "why linux is not ready for the desktop", instead of BUGFIX "what prevents linux to shine on desktop"
                          2) The composition of problems is wrong - many problems references as unresolvable, where they have been resolved, like S3TC, omitting S2TC and omitting the fact anyone can use S3TC and intel even turned it on by default; many assumptions are short sighted on purpose, like comparison of closed source from illegal monopoly platform with opensource from 10% platform - its like crying about bad state of tizen vs good state of android, for the sake of crying; many claims are exaggerations, like "X on Linux is slower than X on windows", where its 5-10% slower; trying to show justified and irrelevant complexity like with Kernel+ALSA+PulseAudio+PA_Sink+App vs Windows (why not show how crappy its on Windows with a lot of legacy stuff like this); yelling "ALSA ftw", where ALSA happened due to OSS going crap, and still OSSv4 *is* available.
                          3) The list itself is built in inductive form, multiplied by errors of 1 & especially 2, where it should be built in deductive form.

                          TL; DR Take the list down and rework it from girl's crylist into men's todo list and maybe the devs will even start reading it. But that involves you to change your own mind attitude, which will hardly happen - because you apparently enjoy this state yourself reposting it like a spambot. Which is masochistic IF you are linux user, or saddistic IF you are windows user; but still all-way girlish.

                          Comment


                          • #14
                            Originally posted by birdie View Post
                            That's funny - people want to destroy me because I see real Linux problems.
                            No, because that list is mostly rubbish.

                            Take for example this:

                            strangely some Linux developers oppose to the whole idea of games on Linux
                            With a link to https://mail.gnome.org/archives/wm-s.../msg00003.html

                            Well, he doesn't. You completely made that up as far as I am concerned.

                            Comment


                            • #15
                              Hello.
                              I'm the owner of the github page and of the native exFat linux read-write driver.
                              I'm impressed that the Phoronix community has created a two-page discussion the same day the drivers got released.

                              I would like to reveal a few things about the driver.

                              It's an android exFat driver which I found on github about a month ago.
                              I dislike Microsoft's attempts to bring the reputation of linux and similar open-source projects down, so it was my revenge.

                              I've needed a way to interact with my external 1TB HDD via linux and store large files on it, and there was no non-fuse ntfs driver available for linux either, so I decided to do a good thing and write a native kernel rw exFat driver for linux.
                              There was a ro driver and a fuse implementation before.
                              The first that came up in my mind is to find if it was available on the net already: I've checked android source codes from Samsung's website, but they did not release any parts of the exfat modules there.
                              Then I made a simple search on github and there was a couple of repos of samsung tablet's kernel. There was it.
                              The version of that kernel was 3.0 so I had to put some effort (literally, 4 hours) to make it work on the version 3.8.

                              This driver has not been written by me, I just ported it to the newer kernel version, with minimal adjustments.
                              That is a good thing, because I can state that the driver will work with all 3.x versions of kernel, if only you could spend an hour fixing a couple of lines of code for that version.

                              The reason I did not port it to 3.9 is because I wanted to make a release ASAP and the community to pick this code to do whatever you like: patches/ports/improvements; and release that as branches on github or somewhere else. It is a work of an hour or less for any programmer to make it work with the latest kernel.
                              And it's even a good source to create a driver for other platforms.

                              I used tor all the time and using it right now because I don't want to waste any more time on dealing with microsoft's idiotic policies, that's why I've uploaded the code in 1 hr on github, via the on-site editor.

                              This driver is guaranteed to provide full exFat support on linux, on the same level as it's available on all Samsung tablets with kernel version 3.0.

                              I am not planning on supporting this driver because I believe I've done enough for the community to have the full linux exFat support.
                              Feel free to fork and port as you like.
                              It's not a patch, and can be installed as a separate module by running just one command: `make install`. It's like the vboxdrv, same principle.

                              P.S.
                              I may find more native drivers (ntfs?) from that tablet's source code and release them later.

                              Free software for the free Minds!

                              Comment

                              Working...
                              X