Re: [Xen-devel] [PATCH 03/10] blktap: Move from drivers/xen to drivers/block

[Ian Campbell <Ian.Campbell@xxxxxxxxxxxxx>]

On Wed, 2011-03-09 at 23:57 +0000, Daniel Stodden wrote:
> On Wed, 2011-03-09 at 17:26 -0500, Daniel Stodden wrote:
> > On Wed, 2011-03-09 at 16:50 -0500, Daniel Stodden wrote:
> > > On Wed, 2011-03-09 at 05:23 -0500, Ian Campbell wrote:
> > > > On Wed, 2011-03-09 at 00:42 +0000, Daniel Stodden wrote:
> > > > > From: Daniel Stodden <dns@xxxxxxxxxxxx>
> > > > > 
> > > > > Signed-off-by: Daniel Stodden <daniel.stodden@xxxxxxxxxx>
> > > > > ---
> > > > >  drivers/block/Kconfig          |    9 +
> > > > >  drivers/block/Makefile         |    1 +
> > > > >  drivers/block/blktap/Makefile  |    3 +
> > > > >  drivers/block/blktap/blktap.h  |  161 +++++++++++
> > > > >  drivers/block/blktap/control.c |  315 +++++++++++++++++++++
> > > > >  drivers/block/blktap/device.c  |  551 
> > > > > +++++++++++++++++++++++++++++++++++++
> > > > >  drivers/block/blktap/request.c |  418 ++++++++++++++++++++++++++++
> > > > >  drivers/block/blktap/ring.c    |  595 
> > > > > ++++++++++++++++++++++++++++++++++++++++
> > > > >  drivers/block/blktap/sysfs.c   |  288 +++++++++++++++++++
> > > > >  drivers/xen/Kconfig            |   11 -
> > > > >  drivers/xen/Makefile           |    1 -
> > > > >  drivers/xen/blktap/Makefile    |    3 -
> > > > >  drivers/xen/blktap/blktap.h    |  161 -----------
> > > > >  drivers/xen/blktap/control.c   |  315 ---------------------
> > > > >  drivers/xen/blktap/device.c    |  551 
> > > > > -------------------------------------
> > > > >  drivers/xen/blktap/request.c   |  418 ----------------------------
> > > > >  drivers/xen/blktap/ring.c      |  595 
> > > > > ----------------------------------------
> > > > >  drivers/xen/blktap/sysfs.c     |  288 -------------------
> > > > >  18 files changed, 2341 insertions(+), 2343 deletions(-)
> > > > >  create mode 100644 drivers/block/blktap/Makefile
> > > > >  create mode 100644 drivers/block/blktap/blktap.h
> > > > >  create mode 100644 drivers/block/blktap/control.c
> > > > >  create mode 100644 drivers/block/blktap/device.c
> > > > >  create mode 100644 drivers/block/blktap/request.c
> > > > >  create mode 100644 drivers/block/blktap/ring.c
> > > > >  create mode 100644 drivers/block/blktap/sysfs.c
> > > > >  delete mode 100644 drivers/xen/blktap/Makefile
> > > > >  delete mode 100644 drivers/xen/blktap/blktap.h
> > > > >  delete mode 100644 drivers/xen/blktap/control.c
> > > > >  delete mode 100644 drivers/xen/blktap/device.c
> > > > >  delete mode 100644 drivers/xen/blktap/request.c
> > > > >  delete mode 100644 drivers/xen/blktap/ring.c
> > > > >  delete mode 100644 drivers/xen/blktap/sysfs.c
> > > > 
> > > > Given the right options "git format-patch" should display this as:
> > > >  drivers/block/Kconfig                   |    9 +++++++++
> > > >  drivers/block/Makefile                  |    1 +
> > > >  drivers/{xen => block}/blktap/Makefile  |    2 +-
> > > >  drivers/{xen => block}/blktap/blktap.h  |    0 
> > > >  drivers/{xen => block}/blktap/control.c |    0 
> > > >  drivers/{xen => block}/blktap/device.c  |    0 
> > > >  drivers/{xen => block}/blktap/request.c |    0 
> > > >  drivers/{xen => block}/blktap/ring.c    |    0 
> > > >  drivers/{xen => block}/blktap/sysfs.c   |    0 
> > > >  drivers/xen/Kconfig                     |   11 -----------
> > > >  drivers/xen/Makefile                    |    1 -
> > > >  11 files changed, 11 insertions(+), 13 deletions(-)
> > > > [...]
> > > > diff --git a/drivers/xen/blktap/blktap.h b/drivers/block/blktap/blktap.h
> > > > similarity index 100%
> > > > rename from drivers/xen/blktap/blktap.h
> > > > rename to drivers/block/blktap/blktap.h
> > > > [...]
> > > > which is far easier to review.
> > > > 
> > > > I think it's -M you want.
> > > 
> > > Hmm, sounds true. 
> > > 
> > > Back then when I created that change, I'm pretty sure I performed a
> > > number of git mv, applied some fixup and committed. 
> > > 
> > > I rebased that tree a couple times, though. Do you mean to say I must
> > > have broken that commit somehere.
> > > 
> > > Okay, I'll try it again and see if that improves the output.
> > 
> > Thinking of it: Iirc this diff is not only a plain rename, there were
> > some additional changes made in source to have everything compile again.
> > Could that be the reason? 
> > 
> > Want those split out?
> 
> Oooh, scratch that, that's not even the case.
> 
> I seem to finally be getting what you're really talking about. This is
> not really a broken commit, it's solely about send-email not enabling
> rename *detection*. Interesting, I always assumed the renaming case must
> be a property of the commit object, not a diff feature.

I don't know if it is a feature of the commit metadata (which FWIW I'd
expect to be retained over a rebase etc) or something which is computed
at diff generation time. In any case -M is required to get the git patch
syntax, the default is to be compatible with old-school patch utils.

> That correct now?

<1299715326-26236-1-git-send-email-daniel.stodden@xxxxxxxxxx> looks
good.
 
> 
> Daniel
> 
> 
> 
> 
> 
> 



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel