On 31 Oct 2001, Georges Heinesch wrote:

> > but NetBSD
> > doesn't appear to make a distinction.
>
> It propably reads/writes NTFS5, but doesn't make use of the additional
> NTFS5 features I mentionned above. This is the same what WinNT makes
> with NTFS5. It is able to handle NTFS5 (in terms of reading and
> writing), but doesn't use the additional functionalities of NTFS5.

That makes sense... A quick grep of the source comes up emtpy for
"encrypt" or "journal".

> > I can read my Windows 2000
> > NTFS partition (presumably "ntfs5") no problem under NetBSD-current.
>
> If Win2000 acccesses this partition, it's for sure NTFS5.
> I guess there's no write support yet for NTFS5?
> Do you know whether NetBSD 1.5 is able to do read NTFS5?

There is ("limited") write support for NTFS4. Wouldn't that mean NetBSD
could also write to NTFS5 partitions but without using any of the new
NTFS5 features?  I haven't tried it personally (my only NTFS partition is
my W2K boot partition, and I don't want to risk hosing it), but I suspect
it will work fine.

There is no difference between the mount_ntfs(8) mangpage on -current and
1.5, so I'd assume the functionality is roughly the same.  There appear to
be atleast a couple bug fixes in -current that aren't on the 1.5 branch,
but a quick scan of CVS Log messages doesn't show any major functionality
differences.

> BTW, is your NTFS partition primary or logical?

Primary.  I avoid extended partitions as much as possible.

> When trying to mount a logical NTFS partition, I get the error ...
>
> dimension# mount -t ntfs /dev/wd1g /mnt/dimension/Data
> mount_ntfs: /dev/wd1g on /mnt/dimension/Data: Argument list too long
> <date> dimension /netbsd: ntfs_readntvattr_plain: POSSIBLE RUN ERROR
> <date> dimension /netbsd: ntfs_readattr_plain: ntfs_readntvattr_plain_
>     failed: o: 0, s: 320048
> <date> dimension /netbsd: ntfs_readattr_plain: attrib: 0 - 78
> [snip]

This appears to be one of the bugs fixed in -current, see PR kern/10119.
You might be able to merge the changes into a 1.5 kernel, or it might be
easier just to run -current.

Rick