Bug#380226: [Parted-maintainers] Bug#380226: NTFS (partition) not recreated correctly after resize: incorrect start sector

[Sven Luther]

On Sat, Dec 16, 2006 at 07:33:38PM +0100, Kurt Roeckx wrote:
> On Sat, Dec 16, 2006 at 05:57:04PM +0100, Sven Luther wrote:
> > On Sat, Dec 16, 2006 at 05:20:02PM +0100, Kurt Roeckx wrote:
> > > 
> > > I just wonder what constraints the end sector should have in that case,
> > > looking at the bug report, the end cylinder also seem to be a multiple
> > > of 2048 - 1, so it seem to me that Microsoft is using 2048 sectors/track
> > > instead of 63.  Vista might not like it that we let the partition stop
> > > on a sector that's a multiple of 63 - 1, and might be unable to use the
> > > last "track".
> > 
> > Is there some kind of version or something for those NTFS partitions, which
> > will enable us to detect them ? 
> 
> Afaik, the version is still 3.1, same as for XP.
> 
> I think they have something in d-i now that detects vista based on some
> files in the fs or something.  Frans probably knows more about this.
> 
> > If i remember the code well, it would be trivial enough to modify the
> > sectors-per-cylinder value if we detect a NTFS partition table which needs it,
> > i think this was already done for the legacy systems, and i know that the
> > amiga filesystems (ffs, sfs, etc), also have some such constraints, and the
> > individual cylinder size of them is hold in the main partition table, so i
> > think i already did something similar.
> 
> That would be one of the resize_contraints (and copy) in libparted/fs/ntfs/ I
> guess, not sure yet how those work.
> 
> But that would still leave the question on how to not change the start
> sector without breaking something else.

You can set constraints on both the start and the end, if i remember well. 

Friendly,

Sven Luther