Re: Names vs. Inodes

From: David Wagner (dawat_private)
Date: Sun Jul 22 2001 - 18:58:22 PDT

  • Next message: David Wagner: "Re: Names vs. Inodes"

    Crispin Cowan  wrote:
    >It's a subtle distinction, but basically I am not concerned about rogue
    >root shells, because
    >they should not exist. I don't *want* to have to be concerned about
    >rogue nobody shells,
    >because it is very difficult to ensure that they never exist.
    
    My philosophy is a bit different.  I always assume that if someone can
    get access to a nobody shell on my machine, they can get root access.
    This assumption has borne out extremely well in practice: The number
    of local exploits is too huge to keep track of.  Therefore, in my view,
    you should be confining those nobody shells, if you think the attacker
    can gain control of them.
    
    If you then make the assumption that all malicious processes are confined,
    I don't think you have to worry so much about the million-symlink attack.
    The truly troublesome concern (IMHO) would be if some non-malicious
    local process with good intentions created a symlink that let a confined
    application violate the security policy, but as far as I know this isn't
    an issue for SubDomain.
    
    _______________________________________________
    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 : Sun Jul 22 2001 - 20:23:31 PDT