On Sat, 10 Jun 2000, Alexander Viro wrote:
> Hell knows. If anything, that stuff depends on having revoke() that works
> - otherwise it's asking for trouble big way. If we have that - fine,
> either stacking dentries by hands or using Erez's stuff will do the trick,
> but that's a pretty minor problem compared to revoke(2) done right.
I am glad someone appreciates how non-trivial *correct* implementation of
revoke(2) (well, assuming it is slightly more extended in the natural
direction than the FreeBSD variant) and the somewhat related forced umount
are.
Why do I say they are related? Because my suggestion for revoke(2) would
be to do the same as for open files in forced umount, i.e. allocate an
inode in nullfs and "shift" the VFS part of the real inode into there,
whilst releasing the fs-specific parts. That way app starts getting EIO in
a most natural way (which is what nullfs is for) and the nullfs inode gets
released when app terminates or voluntary closes the file.
Regards,
Tigran
-
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 : Thu Jun 15 2000 - 21:00:22 EST