[GIT PULL] KVM changes for 3.16-rc4

[GIT PULL] KVM changes for 3.16-rc4

[Paolo Bonzini]

Linus,

The following changes since commit 7171511eaec5bf23fb06078f59784a3a0626b38f:

  Linux 3.16-rc1 (2014-06-15 17:45:28 -1000)

are available in the git repository at:

  git://git.kernel.org/pub/scm/virt/kvm/kvm.git tags/for-linus

for you to fetch changes up to 9a630d15f16dbe4fec7ef5a4bc570cd46774a968:

  Merge commit '33b458d276bb' into kvm-master (2014-06-30 16:45:40 +0200)

----------------------------------------------------------------

A bunch of one-liners (except the s390 one).

The two more serious bugs ("KVM: SVM: Fix CPL export via SS.DPL" and
"KVM: s390: add sie.h uapi header file to Kbuild and remove header
dependency") were introduced in the 3.16 merge window.

----------------------------------------------------------------
Deng-Cheng Zhu (1):
      MIPS: KVM: Fix memory leak on VCPU

Heiko Carstens (1):
      KVM: s390: add sie.h uapi header file to Kbuild and remove header dependency

Jan Kiszka (1):
      KVM: SVM: Fix CPL export via SS.DPL

Nadav Amit (1):
      KVM: x86: Increase the number of fixed MTRR regs to 10

Paolo Bonzini (3):
      KVM: x86: preserve the high 32-bits of the PAT register
      Merge tag 'kvms390-20140626' of git://git.kernel.org/.../kvms390/linux into kvm-master
      Merge commit '33b458d276bb' into kvm-master

Xiaoming Gao (1):
      kvm: fix wrong address when writing Hyper-V tsc page

 arch/mips/kvm/kvm_mips.c          |  1 +
 arch/s390/include/uapi/asm/Kbuild |  1 +
 arch/s390/include/uapi/asm/sie.h  | 26 ++++++++++++--------------
 arch/x86/include/asm/kvm_host.h   |  4 ++--
 arch/x86/kvm/svm.c                |  1 +
 arch/x86/kvm/x86.c                |  2 +-
 6 files changed, 18 insertions(+), 17 deletions(-)

Re: [GIT PULL] KVM changes for 3.16-rc4

[Linus Torvalds]

On Tue, Jul 1, 2014 at 3:44 AM, Paolo Bonzini <pbonzini@redhat.com> wrote:
>
>   git://git.kernel.org/pub/scm/virt/kvm/kvm.git tags/for-linus

I've pulled this, but I was pretty close to saying "screw this shit".

Look at commit 9a630d15f16d, and pray tell me why those kinds of
commit logs are excusable?

That commit message is totally worthless noise. It doesn't explain
what it merges, much less give any explanation of *why* it is a merge.

Merges need explanations too. Tell what the branch you are merging
does, and why you are doing the merge. Yeah, in this case the "branch"
contains a single commit, but that *still* doesn't excuse not telling
what the merge is, and why it exists at all.

Seriously. Do "git log --merges" on current git, and look at the
discrepancy between merge commit messages. That commit 9a630d15f16d is
pure garbage.

It's not the only crappy one, but it really does stand out. There are
other one-liners in there, but even then they tend to have at least
*some* semblance of actual information in them, ie

    Merge branch 'for-v3.16/ti-clk-drv' of
github.com:t-kristo/linux-pm into clk-next

at least shows that there's a topic branch with a reasonable name, and
where it comes from. I'd really prefer it to talk about what it merges
and why, but it's still *much* better than your completely
information-free merge message.

                Linus

Re: [GIT PULL] KVM changes for 3.16-rc4

[Paolo Bonzini]

Il 01/07/2014 19:47, Linus Torvalds ha scritto:
> Merges need explanations too. Tell what the branch you are merging
> does, and why you are doing the merge. Yeah, in this case the "branch"
> contains a single commit, but that *still* doesn't excuse not telling
> what the merge is, and why it exists at all.
>
> Seriously. Do "git log --merges" on current git, and look at the
> discrepancy between merge commit messages. That commit 9a630d15f16d is
> pure garbage.
>
> It's not the only crappy one, but it really does stand out. There are
> other one-liners in there, but even then they tend to have at least
> *some* semblance of actual information in them, ie
>
>     Merge branch 'for-v3.16/ti-clk-drv' of
> github.com:t-kristo/linux-pm into clk-next
>
> at least shows that there's a topic branch with a reasonable name, and
> where it comes from. I'd really prefer it to talk about what it merges
> and why, but it's still *much* better than your completely
> information-free merge message.

You're definitely right about the poor commit message.

I made this a merge because I really wanted this commit in the 3.17 
development branch too.  So I applied the patch on top of -rc1 and 
merged it into both branches (kvm-master for 3.16 and kvm-next for 
3.17).  For some reason, I did see a need to justify the merge commit in 
kvm-next:

     commit dc720f95939280f9e69cafe7389be6d0fa6f22dd
     Merge: 27e6fb5dae28 33b458d276bb
     Author: Paolo Bonzini <pbonzini@redhat.com>
     Date:   Mon Jun 30 16:51:07 2014 +0200

         Merge commit '33b458d276bb' into kvm-next

         Fix bad x86 regression introduced during merge window.

Probably still not verbose enough, but a little better.

Am I just over-engineering it and I should have simply cherry-picked it? 
  In this particular case the change is not likely to get other changes 
(and thus conflicts) in kvm-next, but in general it could.

Thanks,

Paolo