The New NTFS File-System Driver Has Been Submitted For Linux 5.15

[dragon321]

Ah yes, finally usable kernel driver for NTFS. ntfs-3g is fine and worked good for me but FUSE has some disadvantages. Looking forward to some benchmarks comparing this new NTFS driver to ntfs-3g and Windows.

[discordian]

Great, finally.

Hope someone does backports to the older stable kernels that I can steal.

[uid313]

Now if we could get some Apple APFS support too...

[mroche]

This is pretty exciting stuff. I have a few drives I have to share with Windows (internal and external) that this should produce good improvements for.

Cheers,
Mike

[You-]

For those using this already, are you using a special ntfs-utils or other with it? with it?

[You-]

That's kind of his role. Ultimately he is the only making the call on what to merge.

I think Linux Foundation should be doing this type of outreach and hand-holding. there is a lot of tooling that they could invest in too to make things easier. The proposed kernel patch tracker has no way that I can see of showing revisions or current state and in this case the workflow for inclusion to the kernel was not obvious.

As a non developer, some times commenter, it was also a surprise to learn that whatever is in linux-next isnt automatically pulled in at the next merge window.

[RahulSundaram]

I think Linux Foundation should be doing this type of outreach and hand-holding. there is a lot of tooling that they could invest in too to make things easier. The proposed kernel patch tracker has no way that I can see of showing revisions or current state and in this case the workflow for inclusion to the kernel was not obvious.

They do sponsor a lot of the conferences, documentation works, tooling etc. Linux kernel developers however are very much sticking to the mail based workflow and tracking patches isn't going to be easy in that because metadata isn't standardized and attempts to introduce changes to that isn't always welcomed. As an example:

https://lwn.net/Articles/860607/

There is increasingly more tools used to make the process better however.

[CommunityMember]

Pretty sad Torvalds needed to tell the maintainers what to do.

In this case the issue was that there is no subsystem maintainer group that approves/adds new filesystems (it happens close to never), so while Paragon got lots of feedback from other kernel devs, and the code was apparently in good enough shape, there was no one that felt they had the authority to say "engage!" (send the PR). The result is that for areas for which there is no maintainer, that ends up falling to Linus as the last resort. There are very few areas of the kernel that someone else (or some group) is not the designated coordinator, so such events do not happen often, but not happening often does not mean not happening ever.

[skeevy420]

For those using this already, are you using a special ntfs-utils or other with it? with it?

I like to mount it, mount it
I like to mount it, mount it

Mount It

Seriously though, nothing extra required. Just a simple:

sudo mount -t ntfs3 /dev/sdd2 /mnt

That's literally the command I use to mount my Windows installation for quick access.

[polarathene]

I am using a slightly older version of the patchset for a while now (grabbed from this little gold mine) and I am happy with it. No issues, good job and many thanks to all the devs who contributed to the driver.

Is there more info about the structure there for those less familiar? Not sure why the main branch continues to retain dev versions of prior kernels around? Looking at some like 5.10 and 5.14, I see folders with different versions or suffixes like `-sep`. Is he making updates via new folders instead of updating the original folder for a patch?

The BMQ link needs to be updated to reference PRJC (which it will redirect to now as BMQ has been folded into Alfred's Project C since mid 2020?).

I'm a little confused why 5.14 has BFQ patches, I thought this was already upstreamed? Looking at the two BFQ links from the README, one hasn't been updated since 2016, the other says it's development version of BFQ but it was last updated in Jan 2021 on a branch for 5.6 kernel..?