namecache stage-2 patch for testing

[Matthew Dillon]

::Testing of NAMEI2.DIFF with compares of without.
::All of the testing without used kernel as of Sep 24.
::The testing was performed in this order on both occasions.
::
::-DR
:
:    oops.  don't bother doing any performance comparisons, I blew 
:    it somewhere.  My buildworld is geting slower and slower and sloowwwer
:    :-).
:
:    I'll take this up tomorrow and produce a new patch.
:
:						-Matt

    Update: I am almost ready with another version of the patch, one that
    hopefully fixes the degenerate case that caused the slowdown.  I have
    also tracked down a bug in the slab allocator that could lead to memory
    corruption and a panics (typically a dirty page on clean or cache
    list VM panic), which I will commit today (not quite yet).

    I also (whew!) am testing a patch that fixes the MP lock panic that
    people have been reporting on SMP boxes.

    I am going to sit on this stuff until I can successfully complete a
    buildworld or two or three on my SMP box before I commit the slab fix
    and post the new namecache patch, but it will definitely happen today
    sometime.

					-Matt
					Matthew Dillon 
					<dillon at xxxxxxxxxxxxx>