Michael Rothwell writes:
> If we delivered an experimental patch for 2.2.16,
> would people be willing to try it out?
Note that there are _already_ working patches (2.2 and 2.4) for adding EA
functions to the kernel, along with ext2 support for same. In addition,
there is ACL support using these EAs, as well as capability support using
EAs under development. These are "named" extended attributes, so the
API should be usable for HPFS and other EAs I think. I'm not sure if the
Macintosh HFS "extra" data would be considered a named EA, or if it is a
"stream".
The only thing that needs doing is an API implementation for named streams
(one that works well, mind you) which can be used for NTFS, UFS, and HFS
(if it isn't considered an EA). Also, the semantics that appear in the
filesystem/VFS are important.
For EA data, it simply does not appear in the filesystem at all - there is
a special syscall required to access the EA data (consistent with Irix
and possibly Solaris). It returns a fixed amount of data, not a file
handle of any sort.
Info can be found at: <http://acl.bestbits.at/>
Cheers, Andreas
> Jason Venner wrote:
> > What is the minimal interface we can build into the kernel that will
> > allow people to experiment with all the ways they would like to do it,
> > so that we can harvest the resulting knowledge and build a sane/robust
> > system to carry into the future with us?
> >
> > We have" the EA crowd with the data attached to the inode with size
> > limits (This seems to include acls) and if you think about it
> > we have a whole set of attributes we carry around in the
> > inode now (all the stat stuff etc)
> > We have the arbitrary streams crowd.
> > We have hybrids who want some parts of both.
> >
> > There are people arguing about
> > namespace
> > portability across nfs
> > implementations on different underlying file systems
> > Implementations in the vfs layer
> > Implementations in userspace only
> > Some hybrids again.
> > existing tool issues
> > security
> > stability
> > data preservation
> >
> > Can we come up with something that will let us play with all of the
> > above to see what actually works well
--
Andreas Dilger \ "If a man ate a pound of pasta and a pound of antipasto,
\ would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/ -- Dogbert
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Aug 15 2000 - 21:00:36 EST