Re: Buffer overflow prevention

From: Peter Busser (peterat_private)
Date: Fri Aug 15 2003 - 02:54:54 PDT

  • Next message: Florian Weimer: "Re: PointGuard: It's not the Size of the Buffer, it's the Address of the Pointer"

    Hi!
    
    > > AFAIK all those combined do bring real security against generic exploits.
    >   "Real security" is not the word.
    
    Even though PaX is better than W^X, it is far from being perfect.
    
    >   PaX / Propolice / W^X / non-exec stacks don't solve bugs. What they do is
    > to _abort_ execution of a process when it behaves abnormally.
    >   So instead of giving attackers the opportunity to run arbitrary code, you
    > only give them the ability to cause a denial of service.
    
    You could say they trade availability for integrity.
    
    >   This kind of protection should be coupled with tools that automatically
    > restart daemons when they crash (ex: daemontools and monit) to actually keep
    > the service running when under attack. Still, all of this is a couple of
    > unreliable band-aids.
    
    A better way to deal with would be to automatically warn someone with enough
    information to easily find and fix the problem. Restarting the daemon makes the
    problem managable, but it won't solve the bug.
    
    Groetjes,
    Peter Busser
    -- 
    The Adamantix Project
    Taking trustworthy software out of the labs, and into the real world
    http://www.adamantix.org/
    



    This archive was generated by hypermail 2b30 : Fri Aug 15 2003 - 10:13:18 PDT