Re: Names vs. Inodes

From: Crispin Cowan (crispinat_private)
Date: Wed Jul 18 2001 - 18:58:18 PDT

  • Next message: James Morris: "Re: TODO list"

    Greg KH wrote:
    
    > On Wed, Jul 18, 2001 at 03:32:48PM -0700, Crispin Cowan wrote:
    > >
    > > If someone has a practical, detailed way to construct the name, please
    > > speak up.  Note that it is not sufficient to be able to construct *some*
    > > name that *could* have been used to open the file: we need the actual name
    > > that was used to make the request.
    > >
    > > On the other hand, if what you have is a heart-felt conviction that there
    > > must be some way to do it ... save it, we've heard that :-)  We need the
    > > details, not the encouragement.
    >
    > I've said in private email to you and your group that there is a way to
    > do it, and outlined it.  I can do it again here if you really want me
    > to.
    >
    > In short, you have to stop thinking of path names being the end all, and
    > have to start worrying about inodes _if_ you allow multiple mounts.
    > Just knowing the original path that was taken to get to a file isn't
    > going to help you out, you _need_ to know them all.
    
    Your private e-mail seemed to mis-understand the SubDomain security model.
    Yes, the absolute path name is the be-al and end-all that we need.  What you
    outlined is an appropriate model for an ACL (access control list) model, but
    SubDomain is the dual of that.
    
    
    > If you don't allow multiple mounts, then you do not have a problem.  A
    > inode has a dentry list which will contain only one dentry.  Hence the
    > pathname to the file.
    
    As Seth alludes to, hard links are the other way a file can be aliased, and
    according to Chris, the inode (which is what we're presently hooking) provides
    a list of dentries, reflecting the hard link alias issue.
    
    If you're going to the advanced PLUG meeting tonight, we can chat about it.
    
    
    > I can write some specific code, based on the last release of SubDomain
    > to do this if you want proof, but I really don't want to do that :)
    
    We're not just being stubborn; Steve, Chris, and Seth spent two hours off-site
    today, trying to brainstorm a solution, and they keep hitting road blocks.
    This is a tough problem, and putting the name hook back is starting to look
    attractive.
    
    Crispin
    
    --
    Crispin Cowan, Ph.D.
    Chief Scientist, WireX Communications, Inc. http://wirex.com
    Security Hardened Linux Distribution:       http://immunix.org
    Available for purchase: http://wirex.com/Products/Immunix/purchase.html
    
    
    _______________________________________________
    linux-security-module mailing list
    linux-security-moduleat_private
    http://mail.wirex.com/mailman/listinfo/linux-security-module
    



    This archive was generated by hypermail 2b30 : Wed Jul 18 2001 - 19:00:19 PDT