On Mon, Jun 18, 2001 at 09:11:25PM +0200, Martin Maciaszek wrote:
> On Thu, Jun 14, 2001 at 09:18:57AM +1000, Andrew Gildfind wrote:
> > On Thu, Jun 14, 2001 at 02:25:45AM +1000, Keith Owens wrote:
> > > On Wed, 13 Jun 2001 18:36:38 +0300,
> > > "Tarkan Erimer" <tarkane@xxxxxxxxxxxxx> wrote:
> > > > [question about the implementation of ACLs on XFS and ext2]
> > >
> > > There are at least two groups working on ACLs for Linux and they
> > > disagree about how they should be implemented. SGI has one
> > > implementation of ACL, included in the XFS patch. Andreas Gr|nbacher's
> > > ACLs are different and will not work with XFS.
> > >
> >
> > To clarify that, we don't disagree at all, we're gradually working towards
> > a common implementation. Things have moved slowly, but we have a friendly
> > relationship with the ext2 people, and in due course we want to merge
> > our efforts. Hopefully there will be more visible progress in the next
> > couple of months.
>
> What exactly are the differences in the implementations of ACLs?
> What are the points that you could (so far) not agree on?A
> Where do you already have already agreed on? (Except being POSIX
> compliant :))
>
Primarily the system call interface. There has been a great deal of discussion
on fsdevel/acldevel about this. The current ext2 syscall interface
raised issues for others, particularly its support for NFSv4/NT/other types
of acls. Last time I checked Andreas was in the process of evolving his system
call
interface to address some of those issues.
Andrew
--
Andrew Gildfind - R&D Software Engineer - SGI Melbourne Australia
email: ajag@xxxxxxx - work: +61.3.9834.8200 mobile: 0412.834.183
|