Re: [PATCH v5 00/14] ntfs filesystem remake

[Namjae Jeon]

On Mon, Jan 19, 2026 at 4:03 PM Christoph Hellwig <hch@xxxxxx> wrote:
>
> On Sun, Jan 18, 2026 at 02:19:51PM +0900, Namjae Jeon wrote:
> > > I'm not sure how many tests are actually run for the ntfs variants
> > > because they lack features needed for many tests, but how many still
> > > fail with this, because with these numbers I suspect there's quite
> > > a few left. Do you have any good grasp why they are failing, i.e.
> > > assumptions in xfsteasts, or missing feature checks?
> > Regarding the xfstests results, many of the 'Not Run' cases are due to
> > fundamental differences in the NTFS architecture. For instance, NTFS
> > does not support certain advanced features like reflink, which causes
> > many tests to be skipped. Also, ntfs does not yet support journaling,
> > leading to failures in tests that assume journal-based consistency.
> > I am currently categorizing these failures to distinguish between
> > NTFS-inherent limitations and areas for future improvement. I will
> > provide a detailed breakdown and analysis of these test results in the
> > cover letter on next version.
>
> Not run is totally fine.  We have plenty of them even for native
> file systems, and having even more for foreign file system support
> is just fine.  What I meant to say is the number of failing tests
> is the much more interesting metric, so maybe you can share that?
Yes, I will share that.
Thanks!
>