[GIT PULL] Xen updates for 3.17-rc0

[GIT PULL] Xen updates for 3.17-rc0

[David Vrabel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Linus,

Please git pull the following tag:

 git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip.git stable/for-linus-3.17-rc0-tag

- - Remove unused V2 grant table support.
- - Note that Konrad is xen-blkkback/front maintainer.
- - Add 'xen_nopv' option to disable PV extentions for x86 HVM guests.
- - Misc. minor cleanups.

Note there will be some minor conflicts because "x86/xen: safely map
and unmap grant frames when in atomic context" was cherry-picked from
this branch for 3.16.  It's not clear why git didn't do the right
thing automatically.

Thanks.

David

 Documentation/ABI/testing/sysfs-driver-pciback |   13 +
 Documentation/kernel-parameters.txt            |    4 +
 MAINTAINERS                                    |    7 +
 arch/arm/xen/enlighten.c                       |    6 +-
 arch/arm/xen/grant-table.c                     |    6 +-
 arch/x86/xen/enlighten.c                       |   13 +
 arch/x86/xen/grant-table.c                     |   58 +----
 arch/x86/xen/p2m.c                             |    5 +-
 drivers/xen/events/events_base.c               |    2 +-
 drivers/xen/events/events_fifo.c               |   55 +++--
 drivers/xen/grant-table.c                      |  309 +-----------------------
 drivers/xen/xen-pciback/xenbus.c               |    1 +
 include/xen/grant_table.h                      |   30 +--
 13 files changed, 87 insertions(+), 422 deletions(-)

Andrey Utkin (1):
      xen/events: drop negativity check of unsigned parameter

David Vrabel (4):
      x86/xen: safely map and unmap grant frames when in atomic context
      xen/grant-table: remove support for V2 tables
      xen/events/fifo: reset control block and local HEADs on resume
      xen/events/fifo: ensure all bitops are properly aligned even on x86

Frediano Ziglio (1):
      xen/events/fifo: remove a unecessary use of BM()

Himangi Saraogi (1):
      xen/arm: use BUG_ON

Konrad Rzeszutek Wilk (3):
      xen: Introduce 'xen_nopv' to disable PV extensions for HVM guests.
      MAINTAINERS: Make me the Xen block subsystem (front and back) maintainer
      xen-pciback: Document the 'quirks' sysfs file

Matt Rushton (1):
      xen/setup: Remove Identity Map Debug Message

Wei Yongjun (1):
      xen/pciback: Fix error return code in xen_pcibk_attach()
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Icedove - http://www.enigmail.net/

iQEcBAEBAgAGBQJT4hwvAAoJEFxbo/MsZsTRSw0H/3WRveQfOu75W/DIkLCgs/+0
LK1EoeFCLIGEciUHbdl4g3C6IQ1jYFI8Vc+WUCbfBRHeytvUzni0M3sKt0nlofg6
2q8C9A5OBBxCAfdXek6CrAzSMbIh5ThAgbUADM1JlOxwERi7PDQalGiSKQDRztDb
qf3JbfqWUU+WOw3rSu1+wPRyRfJCOW38eBdENGjftwkjLoyn+zEeptORsIhuillT
oTgSi5YkMb04MZfKJq+asOyozjjkgnb9chby28UnmkrL3nsk3aN6xrFc1JaTIgzI
bokyOl4mHcKvjZYU6b58YVt2X8Oy9HKrfUS3HQocoAs9yvVyWhQz14ZpsfOx+8M=
=vAd6
-----END PGP SIGNATURE-----

Re: [GIT PULL] Xen updates for 3.17-rc0

[Linus Torvalds]

On Wed, Aug 6, 2014 at 2:14 AM, David Vrabel <dvrabel@cantab.net> wrote:
>
> Note there will be some minor conflicts because "x86/xen: safely map
> and unmap grant frames when in atomic context" was cherry-picked from
> this branch for 3.16.  It's not clear why git didn't do the right
> thing automatically.

The reason is that while both branches have the same patch, one of
them has additional changes on top of that patch that means that the
two branches do not agree on what the end result should be.

There are broken systems that think that "same patch" means that it
should be a common base that should be merged on top of (resulting in
the merge being the superset patch). That's BS, and git doesn't do
that, even if that's _often_ the right thing to do (but the patch on
top might have become irrelevant due to other patches in the other
tree). So Git only cares about "branch A resulted in X, branch B
resulted in Y, the two changed the same file differently".

Anyway, now you know. Cherrypicking is *not* the answer to "no merge
conflicts". Not if the two branches then disagree in other details wrt
the cherry-picked code, like in this case.

           Linus