[clipping CC list to include @freebsd.org, @netbsd.org, and @apple.com]

>>>>> "Chris" == Chris Dillon <cdillon@wolves.k12.mo.us> writes:
    Chris> An origin filesystem (created by and mounted on the same system) which
    Chris> happens to have stuff owned by several different users (this is all
    Chris> pseudo... don't mind sizes, dates, and link counts in this example):

    Chris> drwxr-xr-x   4 root  wheel    512 Aug 18 15:42 ./
    Chris> drwxr-x---   4 harry users    512 Mar 10 10:21 dir1/
    Chris> drwxr-xr-x   2 john  users    512 Jul  1 18:40 dir2/

    Chris> ls -la dir1
    Chris> -rw-r-----   1 harry users    0 Aug 18 15:48 bar
    Chris> -rw-r-----   1 harry users    0 Aug 18 15:48 foo

    Chris> Take this filesystem and mount it as a "foreign" filesystem on another
    Chris> machine.  User 'jake' owns the mountpoint on the machine.

    Chris> drwxr-xr-x   2 jake  users    512 Jan  4  1999 /mnt/

    Chris> mount_foreign /dev/whatever /mnt

    Chris> ls -la /mnt
    Chris> drwxr-xr-x   4 jake  users    512 Aug 18 15:42 ./
    Chris> drwxr-x---   4 jake  users    512 Mar 10 10:21 dir1/
    Chris> drwxr-xr-x   2 jake  users    512 Jul  1 18:40 dir2/

    Chris> ls -la /mnt/dir1/
    Chris> -rw-r-----   1 jake  users    0 Aug 18 15:48 bar
    Chris> -rw-r-----   1 jake  users    0 Aug 18 15:48 foo

    Chris> Note file mode bits were not affected, but everything gained the
    Chris> user/group of the mountpoint.

  I agree up to this point.

    Chris> Now what happens if user 'jake' adds something to the filesystem?

    Chris> touch /mnt/dir1/baz

    Chris> ls -la /mnt/dir1/
    Chris> -rw-r-----   1 jake  users    0 Aug 18 15:48 bar
    Chris> -rw-r-----   1 jake  users    0 Aug 18 15:48 foo
    Chris> -rw-r-----   1 jake  users    0 Aug 18 15:50 baz

    >> From jake's perspective, this happens as if it were an origin
    Chris> filesystem and he owned everything on it.

    Chris> Now, mount the filesystem again on it's origin system.

  First, a question: if the disk was mounted by root on the origin system,
then I'm not sure it is safe to mount it again after it has been in another's
hand. 
  I would suggest that to facilitate cooperation, that the new file should
be made with "jake"'s userid. That may conflict, but I suggest that this
is a higher level issue.

    Chris> 	1) When writing to a foreign filesystem, should file mode bits
    Chris> be inherited from the parent, or be based on the umask of the foreign
    Chris> user writing the file at that time?  Can the umask of the foreign
    Chris> owner be preserved (which isn't necessarily the same thing as
    Chris> inheriting from the parent) and used?

  I'd say you leave things as is for a file system now.

    Chris> 	2) How should chown and chgrp act when attempting to modify
    Chris> credentials on one of these foreign filesystems?  Should it affect
    Chris> only the local credential mapping (temporarily) and not modify the
    Chris> foreign filesystem?  Should you completely ignore the attempts and
    
  I suggest that they fail as non-root can't do chown. If you are root
doing this, then you have no problem, but you don't mount as root.

  chgrp continues to function as normal, which may be wrong if groupids
aren't shared, but I suggest that is too, a higher level problem.
  
    Chris> 	3) What happens if you want to mount the filesystem on a
    Chris> machine besides the origin, but you do NOT want to do credential
    Chris> mapping (i.e. the systems both have the same sets of users)?  This
    Chris> wouldn't matter if you just used a mount option or different
    Chris> filesystem type when mounting, but I'm assuming something automagic is
    Chris> wanted here.

  You have to mount as root.

    Chris> 	4) What happens if you change the credentials of the
    Chris> mountpoint after you have mounted the foreign filesystem?  Should the
    Chris> mappings follow the new credentials, or stay as they were when first
    Chris> mounted?

  It requires some kind of mount/umount operation. It might be as simple as
doing:
	"eject floppy"
  which will fail because the file system is mounted, but it will then
reexamine the mount point.

   :!mcr!:            |  Cow#1: Are you worried about getting Mad Cow Disease?
   Michael Richardson |  Cow#2: No. I'm a duck.
 Home: mcr@sandelman.ottawa.on.ca. PGP key available.