Re: A Comment from User Space

From: Sandy Harris (sandyat_private)
Date: Sun Apr 22 2001 - 13:15:54 PDT

  • Next message: jmjonesat_private: "Re: Benchmarks (was Re: Hooking into Linux using the LTT)"

    Crispin Cowan wrote:
    > 
    > jmjonesat_private wrote:
    > 
    > > Who's working on the Code to English (documentation) translation here?
    > > I think (conceit), I could perhaps volunteer to help.
    > 
    > Thanks!  That's certainly something that will need doing.  At the moment,
    > the code base is rather fluid, so the code is the documentation.  When it
    > firms up, we'll need a document that explains how to create a security
    > module, and documents the API for each of the hooks.
    
    Arguably, we want a reasonably clear spec before people start coding. This
    is particularly true for security code, since we want people to analyse it
    looking for holes. Ideally, you want to go beyond just a specification, all
    the way to a formal model with provable security properties.
    
    Of course, it's no use specifying something that cannot be coded or that is
    going to be hopelessly inefficient. So we are probably looking at some form
    of iterative process.
    
    Can we specify what we want to hook to in terms of data structures? We
    have structs for a process, file, socket. Is it enough to say a security
    module gets to:
    
    	add fields to those structs,
    	add things to the i-node to initialise file structs
    	intercept a list of procedure calls, some of whose args are
    	  pointers to those structs
    
    Or are there other structures (RPC?, shared memory?, IPSEC tunnels?, ...)
    that they should be able to manipulate?
    
    _______________________________________________
    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 Apr 22 2001 - 13:38:28 PDT