[GIT PULL?] Create and populate toplevel tests/ for kernel tests

[Sam Ravnborg <sam@ravnborg.org>]

Hi Linus.

Will you consider such a primary code-movement for -rc1
or shall we wait until next merge window?

Had we hit -rc2 I would not have sent this pull req and
feel free to flame me anyway.

The rationale to get it merged is obviously to avoid
merge conflicts and the only reason I ask is that I consider
it a low risk patch.

I have not included 8/8 since it was questioned and it
will wait until next merge window. But the first 7 was
straightforward.

You can pull from:
ssh://master.kernel.org/pub/scm/linux/kernel/git/sam/tests.git

diffstat and shortlog below.
I also included mail last with a few of the merge related comments.

	Sam

 Makefile                                        |    1 +
 drivers/misc/Makefile                           |    1 -
 kernel/Makefile                                 |    4 -
 lib/Kconfig.debug                               |   71 +--------------------
 lib/Makefile                                    |    1 -
 tests/Kconfig                                   |   79 +++++++++++++++++++++++
 tests/Makefile                                  |   10 +++
 {kernel => tests}/backtracetest.c               |    0 
 {drivers/misc => tests}/lkdtm.c                 |   12 ++--
 {lib => tests}/locking-selftest-hardirq.h       |    0 
 {lib => tests}/locking-selftest-mutex.h         |    0 
 {lib => tests}/locking-selftest-rlock-hardirq.h |    0 
 {lib => tests}/locking-selftest-rlock-softirq.h |    0 
 {lib => tests}/locking-selftest-rlock.h         |    0 
 {lib => tests}/locking-selftest-rsem.h          |    0 
 {lib => tests}/locking-selftest-softirq.h       |    0 
 {lib => tests}/locking-selftest-spin-hardirq.h  |    0 
 {lib => tests}/locking-selftest-spin-softirq.h  |    0 
 {lib => tests}/locking-selftest-spin.h          |    0 
 {lib => tests}/locking-selftest-wlock-hardirq.h |    0 
 {lib => tests}/locking-selftest-wlock-softirq.h |    0 
 {lib => tests}/locking-selftest-wlock.h         |    0 
 {lib => tests}/locking-selftest-wsem.h          |    0 
 {lib => tests}/locking-selftest.c               |    0 
 {kernel => tests}/rcutorture.c                  |    0 
 {kernel => tests}/rtmutex-tester.c              |    2 +-
 {kernel => tests}/test_kprobes.c                |    0 
 27 files changed, 99 insertions(+), 82 deletions(-)

Ananth N Mavinakayanahalli (7):
      Create tests/ directory
      Move locking selftests to tests/
      Move rcutorture to tests/
      Move rtmutex-tests to tests/
      Move lkdtm to tests/
      Move kprobes smoke tests to tests/
      Move backtrace tests to tests/



On Tue, Feb 12, 2008 at 01:22:46PM -0800, Andrew Morton wrote:
> On Tue, 12 Feb 2008 11:44:52 -0500
> Christoph Hellwig <hch@infradead.org> wrote:
> 
> > On Mon, Feb 11, 2008 at 04:14:52PM +0530, Ananth N Mavinakayanahalli wrote:
> > > The following series of patches create and populate the toplevel tests/
> > > directory. This will henceforth be the place where all in-kernel tests
> > > live.
> > > 
> > > All patches against 2.6.25-rc1 and are just code movement without any
> > > change in functionality.
> > 
> > ACK to patches 1-7, and I agree with Ingo that the x86-specific test
> > should stay under arch/x86.
> 
> OK.  But now is basically the worst time for me (or anyone else) to merge
> large code-motion changes like this, because they need to be carried for
> two months or more.
> 
> And even though git can track renames, putting them into a git tree (say,
> git-kbuild) won't help, because if some other git tree tries to modify a
> file in its original place, I get to fix up the fallout.
> 
> Which I _could_ do, and would do if the patches were particularly risky or
> added/changed functionality or whatever.  But they don't do that, and there
> is little advantage in maintaining them for the >2 months.
> 
> So.  Please redo and resend the patches when we hit 2.6.25-rc6 or so?
> 
> Thanks.
> 
> (linux-next will largely fix all this: git will take care of the renames
> and I'll just base the -mm queue on the consolidated linux-next.  But we
> aren't there yet).
>