Re: [GIT PULL] VFIO updates for v5.19-rc1

[Alex Williamson]

On Wed, 1 Jun 2022 13:57:52 -0700
Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:

> On Wed, Jun 1, 2022 at 10:11 AM Alex Williamson
> <alex.williamson@xxxxxxxxxx> wrote:
> >
> > I'm not sure where git pull-request is getting the diffstat below, the
> > diff of the actual merge of this against mainline looks far less scary.
> > If I've botched something, please let me know.  
> 
> It's all normal, and due to you having merges in your tree and
> multiple merge bases.
> 
> See
> 
>     Documentation/maintainer/messy-diffstat.rst
> 
> for details (yay, Jonathan scrounged together docs so that I don't end
> up having to write a long email explanation any more, and there are
> links to some of my previous explanations on lore).
> 
> That also has a suggested remedy, ie just do a temporary merge and use
> the diffstat from that one instead.

Ok, so I should have gone the one step further than I did and replaced
the diffstat from pull-request with the one from my local test merge,
which would have looked like this:

 Documentation/driver-api/vfio-mediated-device.rst |   4 +-
 drivers/gpu/drm/i915/gvt/gtt.c                    |   4 +-
 drivers/gpu/drm/i915/gvt/gvt.h                    |   8 +-
 drivers/gpu/drm/i915/gvt/kvmgt.c                  | 115 +---
 drivers/net/ethernet/mellanox/mlx5/core/sriov.c   |  65 +-
 drivers/s390/cio/vfio_ccw_cp.c                    |  47 +-
 drivers/s390/cio/vfio_ccw_cp.h                    |   4 +-
 drivers/s390/cio/vfio_ccw_fsm.c                   |   3 +-
 drivers/s390/cio/vfio_ccw_ops.c                   |   7 +-
 drivers/s390/crypto/vfio_ap_ops.c                 |  50 +-
 drivers/s390/crypto/vfio_ap_private.h             |   3 -
 drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c    |  16 +-
 drivers/vfio/pci/mlx5/cmd.c                       | 236 +++++--
 drivers/vfio/pci/mlx5/cmd.h                       |  52 +-
 drivers/vfio/pci/mlx5/main.c                      | 136 ++--
 drivers/vfio/pci/vfio_pci.c                       |   6 +-
 drivers/vfio/pci/vfio_pci_config.c                |  56 +-
 drivers/vfio/pci/vfio_pci_core.c                  | 254 ++++---
 drivers/vfio/vfio.c                               | 781 ++++++++--------------
 include/linux/mlx5/driver.h                       |  12 +
 include/linux/vfio.h                              |  44 +-
 include/linux/vfio_pci_core.h                     |   3 +-
 include/uapi/linux/vfio.h                         |   4 +-
 virt/kvm/vfio.c                                   | 329 ++++-----
 24 files changed, 1095 insertions(+), 1144 deletions(-)

> But I can also re-create that messy diffstat (and thus verify that
> what you sent me matches what I got) here locally too.
> 
> So while the diffstat is messy and not very useful for a "this is what
> changed" angle (because it has a lot of other changes mixed in), even
> that messy diffstat is actually useful for my secondary reason, namely
> as a verification that yes, I got what you were trying to send and
> just didn't document very clearly because of those multiple merge
> bases.
> 
> I can (and do) also check the shortlog, since the actual log doesn't
> have any issues with merges, it's only "diff" that needs a single
> well-defined <start,end> tuple.

Thanks, Linus.  Replacing the pull-request diffstat didn't feel quite
right, but I'll do that next time I get a crazy listing.  Thanks,

Alex