3 ms·

Does anyone know what confluence of events or motives made this land now? I'm grateful for their work--just curious why a commercial software company decided to

Does anyone know what confluence of events or motives made this land now? I'm grateful for their work--just curious why a commercial software company decided to make its main capability available for free.

kasabali 5y ago
Some context: https://arstechnica.com/gadgets/2021/08/paragon-is-working-to-get-its-ntfs3-filesystem-into-the-linux-kernel/ https://arstechnica.com/gadgets/2021/08/paragon-is-working-t... So just a guess; maybe after Samsung's exfat code was merged into mainline, license and support revenue from their own exfat driver took a hit. So they decided it's in their best interest to preemptively get their own ntfs driver in mainline thus they can get more support contracts this way? PS: I noticed this time it's Tuxera who isn't happy with this development. They remembered they maintain ntfs-3g [1], and decided up the version number of ntfs-3g after 4 years [2]. So I guess the theory above makes at least a slight sense. [1] https://lore.kernel.org/lkml/alpine.DEB.2.20.2109030047330.23375@tuxera.com/ https://lore.kernel.org/lkml/alpine.DEB.2.20.2109030047330.2... [2] https://github.com/tuxera/ntfs-3g/wiki/NTFS-3G-Release-History https://github.com/tuxera/ntfs-3g/wiki/NTFS-3G-Release-Histo...
ch_123 5y ago
There’s an official statement here: https://www.paragon-software.com/home/ntfs3-driver-faq/ https://www.paragon-software.com/home/ntfs3-driver-faq/ Reading between the lines, I suspect that the effort of maintaining an out-of-tree file system outweighed any money to be made from selling NTFS support to Linux users, especially when ntfs-3g is probably good enough for many use cases. They have a lot of other products, and it’s possible that some of them leverage this NTFS support as well.
- m4rtink 5y ago
  Anything out of the kernel source tree is pain in the ass for many reasons, ranging from a nuisance to impossible to use due to tainted kernel nullifying OS vendor support guarantees.
- tgsovlerkhgsel 5y ago
  This was my main question: What does this solve that ntfs-3g doesn't? I thought issues with NTFS on linux were a thing of the past.
  - jpgvm 5y ago
    Performance. FUSE filesystems are cool but not really native in the same way an implementation like this is.
    - xoa 5y ago
      And "performance" here isn't some single digit percentage niggling either, it's something like a 4x difference. Enough to genuinely be a pretty big deal for any usage beyond minimal and occasional "get some data off this fs once" type of usage. Of course really the answer is that Microsoft should just adopt Lundman's efforts and add native OpenZFS support to Windows but in the mean time... ;).

kasabali 5y ago

Some context: https://arstechnica.com/gadgets/2021/08/paragon-is-working-to-get-its-ntfs3-filesystem-into-the-linux-kernel/ https://arstechnica.com/gadgets/2021/08/paragon-is-working-t... So just a guess; maybe after Samsung's exfat code was merged into mainline, license and support revenue from their own exfat driver took a hit. So they decided it's in their best interest to preemptively get their own ntfs driver in mainline thus they can get more support contracts this way? PS: I noticed this time it's Tuxera who isn't happy with this development. They remembered they maintain ntfs-3g [1], and decided up the version number of ntfs-3g after 4 years [2]. So I guess the theory above makes at least a slight sense. [1] https://lore.kernel.org/lkml/alpine.DEB.2.20.2109030047330.23375@tuxera.com/ https://lore.kernel.org/lkml/alpine.DEB.2.20.2109030047330.2... [2] https://github.com/tuxera/ntfs-3g/wiki/NTFS-3G-Release-History https://github.com/tuxera/ntfs-3g/wiki/NTFS-3G-Release-Histo...

ch_123 5y ago

There’s an official statement here: https://www.paragon-software.com/home/ntfs3-driver-faq/ https://www.paragon-software.com/home/ntfs3-driver-faq/ Reading between the lines, I suspect that the effort of maintaining an out-of-tree file system outweighed any money to be made from selling NTFS support to Linux users, especially when ntfs-3g is probably good enough for many use cases. They have a lot of other products, and it’s possible that some of them leverage this NTFS support as well.

m4rtink 5y ago
Anything out of the kernel source tree is pain in the ass for many reasons, ranging from a nuisance to impossible to use due to tainted kernel nullifying OS vendor support guarantees.
tgsovlerkhgsel 5y ago
This was my main question: What does this solve that ntfs-3g doesn't? I thought issues with NTFS on linux were a thing of the past.
- jpgvm 5y ago
  Performance. FUSE filesystems are cool but not really native in the same way an implementation like this is.
  - xoa 5y ago
    And "performance" here isn't some single digit percentage niggling either, it's something like a 4x difference. Enough to genuinely be a pretty big deal for any usage beyond minimal and occasional "get some data off this fs once" type of usage. Of course really the answer is that Microsoft should just adopt Lundman's efforts and add native OpenZFS support to Windows but in the mean time... ;).