Fw: Priority Increasing

From: Ashwin Chandra (ashcs_at_ucla.edu)
Date: 02/28/05

  • Next message: Julian Elischer: "Re: intercepting RSVP packets with netgraph" 
    To: <freebsd-hackers@freebsd.org>
Date: Sun, 27 Feb 2005 20:56:03 -0800

    The forkbomb program I wrote is just one parent that forks 750 or so
     children that each malloc around 40 MB's of memory and do a mem traversal
     through it. The children do not fork. I see the overhead of forking could
    be
     causing this, but shouldn't there be some difference in the load of the
     system when each forkbomb process is set to the lowest priority? To fork
    750
     processes would incur overhead until those processes are created (Which
     shouldnt take much of real time) and once they are running, if they other
     processes that have already been created are running "nicely", I don't see
     why there is so much overhead too.

     Do you recommend anotoher way to solve this forkbomb problem and keep the
     system DoS free for others?
     Ash

    > ----- Original Message -----
    > From: "Coleman Kane" <zombyfork@gmail.com>
    > To: "Ashwin Chandra" <ashcs@ucla.edu>
    > Cc: <freebsd-hackers@freebsd.org>
    > Sent: Sunday, February 27, 2005 7:28 PM
    > Subject: Re: Priority Increasing
    >
    >
    >> Well, since the program is running a forkbomb, it is gonna stress out
    >> the kernel. The kernel is constantly creating new process spaces, as
    >> well as filling in the queue.
    >>
    >> Are we talking a O(2^n) forkbomb here (where the forks also fork)?
    >> Remember, there is overhead associated with forking off new processes,
    >> and if your program is doing it continuously, nicing them is not going
    >> to fix the problem. I suggest you fix the program so that it doesn't
    >> forkbomb.
    >>
    >> You can also rlimit it, and force the number of processes to a
    >> specific ceiling. This will result in crashing the program everytime
    >> you hit that limit, however. Try looking into djb's daemontools if you
    >> want to duct-tape it (ports/sysutils/daemontools)
    >>
    >> --
    >> coleman
    >>
    >> On Sun, 27 Feb 2005 18:11:57 -0800, Ashwin Chandra <ashcs@ucla.edu>
    >> wrote:
    >>> Hi all,
    >>> Ive been trying to counter the malicious effects of a forkbomb by
    >>> setting the forkbomb parent and children to a PRI_MAX priority, although
    >>> this is not having any effect on the system load.
    >>>
    >>> Basically in my code when I know which process is acting maliciously
    >>> (forkbomb), I run the following simple code:
    >>>
    >>> FOREACH_KSEGRP_IN_PROC(p, kg)
    >>> {
    >>> printf("old prio:%d", kg->kg_user_pri);
    >>> kg->kg_user_pri = PRI_MAX;
    >>> printf(" new prio:%d", kg->kg_user_pri);
    >>> }
    >>>
    >>> When it prints out, the old prio was 180 and the new gets set to 255
    >>> although there is help to the system...the system is still under stress.
    >>> Do you guys know any good ways of hacking the scheduler to make a
    >>> process that is bad run MUCH MUCH less as to not overload the system?
    >>>
    >>> Ash
    >>> _______________________________________________
    >>> freebsd-hackers@freebsd.org mailing list
    >>> http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
    >>> To unsubscribe, send any mail to
    >>> "freebsd-hackers-unsubscribe@freebsd.org"
    >>>
    >

    _______________________________________________
    freebsd-hackers@freebsd.org mailing list
    http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
    To unsubscribe, send any mail to "freebsd-hackers-unsubscribe@freebsd.org"

  • Next message: Julian Elischer: "Re: intercepting RSVP packets with netgraph"

    Relevant Pages

    • Re: Fw: Priority Increasing
      ... > The forkbomb program I wrote is just one parent that forks 750 or so ... The children do not fork. ... You are talking about malloc'ing almost 30GB of memory and having 750 ...
      (freebsd-hackers)
    • Re: Priority Increasing
      ... > by setting the forkbomb parent and children to a PRI_MAX priority, ... What it does at present is whenever you try to fork and you've hit your ... process limit ), it puts your process to sleep for .5 ... not just do the same thing, perhaps with a shorter sleep. ...
      (freebsd-hackers)