Re: Security through Permissiveness: A Zen Riddle? (Crispin Cowan)

From: Crispin Cowan (crispinat_private)
Date: Fri Jul 13 2001 - 16:50:19 PDT

  • Next message: Chris Wright: "Re: Security through Permissiveness: A Zen Riddle? (Crispin Cowan)"

    Matt Block wrote:
    
    > From: crispinat_private [mailto:crispinat_private]
    > > IMHO, the priority sequence for LSM is:
    > >  1. Finish the current rendition of LSM and get it into the 2.5
    > kernel(as Greg said)
    > >  2. Audit
    > >  3. Permissive hooks
    >
    > Does your humble opinion represent, by any chance, something that can be
    > viewed
    > as a guideline?
    
    Sort-of.  WireX started this project in response to a perceived community
    need, a perceived opportunity that Linus created when he said he'd be
    interested in such a thing, and (naturally) because it is in our own
    interests, because we have our own security enhancements that we wish could
    be modules.
    
    I am the PI (Principle Investigator) on a DARPA contract (Autonomix
    http://immunix.org/autonomix/ but this web page is a bit out of date) that
    is, in part, paying for WireX's development efforts on LSM.  This means
    that in principle I direct the guys coming from the wirex.com domain
    (Chris, Seth, and Steve) but in practice they know more about kernel code
    than I do, so I mostly provide guidance and let them do what they're good
    at.
    
    LSM is a community project.  It will have no purpose at all if the
    community does not buy into it.  "Community" means both the security
    community (which is famous for squabbling) and the Linux community (which
    is relatively new at the squabbling business :-)  So while I speak with a
    fairly well-informed voice with a sack of money tied to it, I don't really
    hold special powers over LSM.
    
    
    > If you are saying that permissive hooks are low priority, but that they
    > _are_ a recognized goal of the project, then I think I agree; there may
    > be no need to get them in now, but perhaps something _could_ be put in at
    > relatively low cost so as to indicate more clearly that this is a
    > direction in which we think we might want to go.  Even just a very low
    > priority entry on the TODO list would be a great start.
    
    It permissive hooks is a complex question.  Personally, I don't like them:
    I prefer the "simple assurance" property of a restrictive-only interface.
    I recognize the legitimacy of what you can do with permissive hooks, but I
    don't personally think they're worth the cost.  So if/when it comes up
    again, I will oppose permissive hooks, but not veto them.
    
    Within the security community, I know that David Wagner agrees with me on
    this.  Stephen Smalley appears to be on the fence.  In the Linux community,
    we suspect that there will be more resistance to permissive hooks than to
    restrictive hooks.  So even if we were all agreed that permissive hooks
    were desirable, it would still be a "later, and then only if Linus agrees"
    item.
    
    
    > Is there a clear list of projects on which help is needed?
    
    No, but that's a good idea:
    
       * I like Seth's idea (just posted) of implementing BSD-style securelevel
         with an LSM module.
       * I'm not sure if JM Jones wants help with the LSM Test Suite.
       * Greg pointed out at the start of today's "TODO" thread that socket
         mediation needs work.
    
    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 : Fri Jul 13 2001 - 16:51:29 PDT