Changes to 'work2'

[David Teigland]

New branch 'work2' available with the following commits:
commit 3cc457d215713035cf9aa2024841184913c350d7
Author: Federico Simoncelli <fsimonce(a)redhat.com&gt;
Date:   Thu Sep 27 12:06:03 2012 -0400

    sanlock: use signal handler for all terminating signals
    
    Signed-off-by: Federico Simoncelli <fsimonce(a)redhat.com&gt;

commit b08e8726c7e854921ec6d8d590c77d4e4d18c4b0
Author: Federico Simoncelli <fsimonce(a)redhat.com&gt;
Date:   Thu Sep 27 12:06:02 2012 -0400

    python: fix the tuple parsing for py_add_lockspace
    
    Signed-off-by: Federico Simoncelli <fsimonce(a)redhat.com&gt;

commit abaf80bbf21b9c771684def60f73f3f1a287bc20
Author: Federico Simoncelli <fsimonce(a)redhat.com&gt;
Date:   Thu Sep 27 15:46:59 2012 -0400

    sanlock: configuring the core dump output
    
    All the setrlimit calls have been moved to a new setup_limits function
    that is called regardless the presence of the -U and -G options.
    The new RLIMIT_CORE limit is set to allow the core dump to be written
    in case of a crash (for more details: man core).
    
    Signed-off-by: Federico Simoncelli <fsimonce(a)redhat.com&gt;

commit aff1f7713ca8f3836e1b76611b3222628f9c2b84
Author: David Teigland <teigland(a)redhat.com&gt;
Date:   Wed Sep 26 14:02:00 2012 -0500

    sanlock: configurable mlockall level
    
    -l 0 -> no mlockall
    -l 1 -> mlockall(CURRENT)
    -l 2 -> mlockall(CURRENT|FUTURE)
    
    The default is now set to 1 (was 2).
    
    CURRENT|FUTURE results in each pthread_create using
    8MB of RSS for the thread stack.  CURRENT alone does
    not.
    
    Signed-off-by: David Teigland <teigland(a)redhat.com&gt;