hard links to files on other filesystems disallowed: why?
in [General Linux]
|
Prev: how does user application call NAPI poll method of network device driver.
Next: firefox and google maps
From: John Hasler on 21 Feb 2010 09:08 Marten writes: > Use a 'volser' (volume serial number) to identify partitions as is > done in some mainframe OSen. Linux already has filesystem UUIDs. > I'd much rather deal with 'doofus-root' than with 'hda1' in fstab, for > example. Use volume labels. -- John Hasler jhasler(a)newsguy.com Dancing Horse Hill Elmwood, WI USA
From: despen on 21 Feb 2010 09:31 J G Miller <miller(a)yoyo.ORG> writes: > On Sat, 20 Feb 2010 22:15:25 -0500, despen wrote: > >> There is no requirement that the target of a soft link exist. > > Indeed not, but it does raise the issue of whether or not > dangling links are dangerous and a potential security hole. Disagree. I don't see any danger or a security issue. Soft links do exactly what they are supposed to do. They are easily detectable and only "pretend" to be files if you treat them that way. Ie. the various tools that detect files have variations that detect soft links.
From: Robert Heller on 21 Feb 2010 10:49 At Sun, 21 Feb 2010 15:12:39 +0100 J G Miller <miller(a)yoyo.ORG> wrote: > > On Sat, 20 Feb 2010 22:15:25 -0500, despen wrote: > > > There is no requirement that the target of a soft link exist. > > Indeed not, but it does raise the issue of whether or not > dangling links are dangerous and a potential security hole. It depends on what they are being used for. Firefox for example creates a 'bogus' (dangling) symlink as a lock file. The "dangling" link contains the host and pid of the running firefox process. > -- Robert Heller -- 978-544-6933 Deepwoods Software -- Download the Model Railroad System http://www.deepsoft.com/ -- Binaries for Linux and MS-Windows heller(a)deepsoft.com -- http://www.deepsoft.com/ModelRailroadSystem/
From: unruh on 21 Feb 2010 11:54 On 2010-02-21, John Hasler <jhasler(a)newsguy.com> wrote: > Bill Unruh writes: >> I can change the uuid of a partition at any time I want. > > This would have to be a new, special filesystem UUID generated when the > filesystem was created and stored as part of the filesystem metadata. > You'd use a null UUID in directory entries for files on the same > filesystem as the directory so that the filesystem driver would bypass > all the UUID stuff when dealing with "local" files. > >> DO you really want all data on all disks to be lost if you replace one >> disk? > > Why would that happen? > >> If you made them hard as well, any change would loose you all your >> data. > > See above. Losing the filesystem UUID would merely cause all links into > the filesystem to dangle. No. hard links have a dual set of pointers-- from the file names to the inodes and from the inodes to the directory names ( well really only the number) so if you delete A you do not delete the inode. Now imagine that the disk became unavailable. Do assume that that link still exists, so if that disk had a bullet put through it, you could never delete that inode, since the count could never go to zero? Do you have each rm search for each directory entry to see if it exists to keep track of the number of directory entries pointing to that inode? If the disk containing the inode disappears, that hard link is useless. Ie keeping track of directory entries/inodes becomes a nightmare. Soft links can dangle. hard links are not supposed to dangle. You could get a situation in which the inode resides on disk A, but no directory entry on disk A actually references that inode. If you insisit that at least one entry on disk A must exist which points to the inode, then I see no difference from soft links. If not, you have full inodes ( eg a 10GB file) which have nothing associated with it, but which cannot be erased, because there are 5 hard links from non-existant disks supposedly pointing at it.
From: AZ Nomad on 21 Feb 2010 12:14
On Sat, 20 Feb 2010 18:14:08 +0000 (UTC), Rahul <nospam(a)invalid.invalid> wrote: >Linux only allows hard-links to be made to files on the same filesystem. Is >there a structural reason for this or just legacy of the "founding >fathers"? Just curious. Hardlinks are directory structures. If you tried to point from one file system to another, the other file system wouldn't be under the control of the file system where the link resides. The other file system is totally foreign to the first file system. The might not even be the same file system. For example, the "other filesystem" might be NFS or XFS while the first file system is EXT3. How the hell would you expect EXT3 to understand the internal structures within NFS or XFS? A soft pointer, on the otherhand, is just a filename. |