I think we are basically at the point where the nathanw_sa branch
can be merged down to the mainline.  I would like to do that on
Friday (US/Pacific time).  Here is what I'm planning to do:

	1. Make an announcement that the merge is beginning.  I would
	   really appreciate it if people refrained from checking code
	   into the kernel while this happens.

	2. Merge the kernel portion branch onto the trunk, compile and
	   boot kernels for platforms I can test directly:

		* alpha
		* arm
		* i386
		* sparc
		* sparc64
		* powerpc
		* m68k (if I'm feeling ambitious)
		* mips (if I'm feeling really ambitious)

	3. Commit kernel portion of merge, and bump the kernel verison.

	4. Merge the userland part of the branch.

	5. Build userland for platforms I can test easily:

		* alpha
		* arm
		* sparc64
		* powerpc
		* m68k

	   (common thread here -- diskless systems :-)

	6. Commit userland portion of merge, and ensure all the requisite
	   library version bumps have been performed.

	7. Announce merge is complete, and that tree is once again open.

	8. File critical/high bug reports against platforms which are
	   known to not have the necessary modifications to even compile:

	   Kernel / libc:

		* sh5/evbsh5
		* hppa/hp700
		* pc532
		* x86_64

	   libpthread / libpthread_dbg:

		* sh3
		* vax

		* sh5
		* hppa
		* ns32k
		* x86_64

	   Note that sh3 and vax will be the only platforms which will
	   fail in the autobuild.  sh5, hppa, ns32k, and x86_64 are not
	   currently built by the autobuild.

	   YES, THIS MEANS THAT PLATFORMS WHICH ARE MISSING LIBPTHREAD
	   SUPPORT WILL FAIL TO BUILD.  We hope this is an effective
	   motivational tool.

	9. File serious/high bug reports against platforms which do not
	   have gdb thread support in their "NetBSD native" support files:

		* alpha
		* arm
		* m68k
		* mips
		* sparc
		* sparc64

		* sh3
		* sh5
		* hppa
		* ns32k
		* vax
		* x86_64

	   I think these problems will be easy to fix up, but I want an
	   audit-trail anyhow.

	10. Announce to the world, and beg people to TEST TEST TEST :-)

The merging of this branch represents the light at the end of a long tunnel,
quite a lot of work, almost all of it by Nathan Williams -- I think we all
owe him a big round of applause (and a few rounds of beers at the next Usenix
probably wouldn't hurt :-)

-- 
        -- Jason R. Thorpe <thorpej@wasabisystems.com>