[GIT PULL] KVM updates for 2.6.35-rc2

[GIT PULL] KVM updates for 2.6.35-rc2

[Avi Kivity]

Linus, please pull from

   git://git.kernel.org/pub/scm/virt/kvm/kvm.git kvm-updates/2.6.35

to receive a bunch of KVM fixes for 2.6.35-rc2, including a workaround 
for a cpu erratum that can allow a guest to crash the host.

Avi Kivity (1):
       KVM: MMU: Remove user access when allowing kernel access to 
gpte.w=0 page

Huang Ying (1):
       KVM: Avoid killing userspace through guest SRAO MCE on unmapped pages

Jan Kiszka (1):
       KVM: Fix order passed to iommu_unmap

Jean Delvare (1):
       KVM: powerpc: fix init/exit annotation

Joerg Roedel (2):
       KVM: SVM: Handle MCEs early in the vmexit process
       KVM: SVM: Implement workaround for Erratum 383

Julia Lawall (1):
       KVM: ia64: Add missing spin_unlock in kvm_arch_hardware_enable()

Marcelo Tosatti (2):
       KVM: MMU: invalidate and flush on spte small->large page size change
       KVM: read apic->irr with ioapic lock held

  arch/ia64/kvm/kvm-ia64.c         |    1 +
  arch/powerpc/kvm/e500.c          |    2 +-
  arch/x86/include/asm/msr-index.h |    1 +
  arch/x86/kvm/mmu.c               |   39 ++++++++++++---
  arch/x86/kvm/paging_tmpl.h       |    7 +--
  arch/x86/kvm/svm.c               |   96 
+++++++++++++++++++++++++++++++++++++-
  include/linux/kvm_host.h         |    1 +
  include/linux/mm.h               |    8 +++
  mm/memory-failure.c              |   30 ++++++++++++
  virt/kvm/ioapic.c                |    3 +-
  virt/kvm/iommu.c                 |    2 +-
  virt/kvm/kvm_main.c              |   30 +++++++++++-
  12 files changed, 201 insertions(+), 19 deletions(-)

-- 
error compiling committee.c: too many arguments to function

Re: [GIT PULL] KVM updates for 2.6.35-rc2

[Linus Torvalds]

On Thu, Jun 10, 2010 at 2:31 AM, Avi Kivity <avi@redhat.com> wrote:
> Linus, please pull from
>
>  git://git.kernel.org/pub/scm/virt/kvm/kvm.git kvm-updates/2.6.35
>
> to receive a bunch of KVM fixes for 2.6.35-rc2, including a workaround for a
> cpu erratum that can allow a guest to crash the host.

Hmm. This looks bigger than I'd wish for at this point. I'll be
offline starting Saturday, so I want -rc3 (on Friday evening) to be in
as well-known a state as humanly possible. Which is why I've been
pushing back a lot on anything that isn't absolutely essential.

                    Linus

Re: [GIT PULL] KVM updates for 2.6.35-rc2

[Avi Kivity]

On 06/10/2010 07:33 PM, Linus Torvalds wrote:
> On Thu, Jun 10, 2010 at 2:31 AM, Avi Kivity<avi@redhat.com>  wrote:
>    
>> Linus, please pull from
>>
>>   git://git.kernel.org/pub/scm/virt/kvm/kvm.git kvm-updates/2.6.35
>>
>> to receive a bunch of KVM fixes for 2.6.35-rc2, including a workaround for a
>> cpu erratum that can allow a guest to crash the host.
>>      
> Hmm. This looks bigger than I'd wish for at this point. I'll be
> offline starting Saturday, so I want -rc3 (on Friday evening) to be in
> as well-known a state as humanly possible. Which is why I've been
> pushing back a lot on anything that isn't absolutely essential.
>    

Ok.  I've dropped the SRAO patch; this still leaves ~100 changed lines, 
but they are almost all due to the erratum 383 workaround, which can 
result in a guest killing the host.  The rest of the patches are less 
critical, but they are one liners and I am comfortable with them in 
-rc3.  The new diffstat also shows no changes to core code:

  arch/ia64/kvm/kvm-ia64.c         |    1 +
  arch/powerpc/kvm/e500.c          |    2 +-
  arch/x86/include/asm/msr-index.h |    1 +
  arch/x86/kvm/mmu.c               |    5 ++
  arch/x86/kvm/svm.c               |   96 
+++++++++++++++++++++++++++++++++++++-
  virt/kvm/ioapic.c                |    3 +-
  virt/kvm/iommu.c                 |    2 +-
  7 files changed, 106 insertions(+), 4 deletions(-)

-- 
I have a truly marvellous patch that fixes the bug which this
signature is too narrow to contain.

Re: [GIT PULL] KVM updates for 2.6.35-rc2

[Linus Torvalds]

On Thu, Jun 10, 2010 at 10:34 AM, Avi Kivity <avi@redhat.com> wrote:
>
> Ok.  I've dropped the SRAO patch; this still leaves ~100 changed lines, but
> they are almost all due to the erratum 383 workaround, which can result in a
> guest killing the host.  The rest of the patches are less critical, but they
> are one liners and I am comfortable with them in -rc3.  The new diffstat
> also shows no changes to core code:

Ok. Pulled the updated version. Thanks,

                  Linus