Re: FreeBSD hooks

From: Stephen Smalley (sdsat_private)
Date: Tue Aug 21 2001 - 07:31:28 PDT

  • Next message: James Morris: "Re: Capability tests in netlink and oom_kill"

    On Tue, 21 Aug 2001, Stephen Smalley wrote:
    
    > Linus originally mentioned the possibility of per-process hooks in his
    > message that led to the creation of the LSM project.  However, per-process
    > hooks are problematic for a number of reasons, e.g.:
    > 
    > 1) How do you deal with operations that occur outside of process 
    > context, such as network input operations?
    > 
    > 2) How do you deal with operations between processes, such as
    > signal delivery, where you may have two different sets of hooks
    > for each process?
    
    I should note even though LSM doesn't provide per-process hooks, you
    can always implement per-process hooks in the security module itself.
    LSM does provide a void *security field in the task_struct, so the 
    module could always use that security field to store per-process
    hooks.  The global hooks defined by the module could then call
    the per-process hooks, albeit with the overhead of the extra
    function call.  But this approach also allows the module to deal
    with the situations above (no process context or multiple process
    contexts) however it desires, without putting that complexity
    in LSM itself.
    
    --
    Stephen D. Smalley, NAI Labs
    ssmalleyat_private
    
    
    
    
    
    _______________________________________________
    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 : Tue Aug 21 2001 - 07:34:02 PDT