[PATCH 0/11] Nested SVM fixes v2

[Joerg Roedel <joerg.roedel@amd.com>]

Hi,

this is the second version of the KVM fixes queued up while testing
nested svm with Windows 7 64 bit. Changes to the previous version:

	* Fixed kunmap bug
	* Introduced nested_svm_intercept insteat of
	  nested_svm_exit_handled_atomic()
	* Added missing patch description
	* Added wbinvd/invd skip functions because the instruction
	  emulator fails on these instructions (found during Windows 7
	  XP mode bringup)

Please review and comment (or apply ;-) these patches.

Thanks,

	Joerg

Shortlog:

Joerg Roedel (11):
      KVM: SVM: Don't use kmap_atomic in nested_svm_map
      KVM: SVM: Fix wrong interrupt injection in enable_irq_windows
      KVM: SVM: Fix schedule-while-atomic on nested exception handling
      KVM: SVM: Sync all control registers on nested vmexit
      KVM: SVM: Annotate nested_svm_map with might_sleep()
      KVM: SVM: Fix nested msr intercept handling
      KVM: SVM: Don't sync nested cr8 to lapic and back
      KVM: SVM: Activate nested state only when guest state is complete
      KVM: SVM: Make lazy FPU switching work with nested svm
      KVM: SVM: Remove newlines from nested trace points
      KVM: SVM: Don't call instruction emulator for invd and wbinvd

Diffstat:

 arch/x86/kvm/emulate.c |    4 +
 arch/x86/kvm/svm.c     |  218 ++++++++++++++++++++++++++++++++---------------
 arch/x86/kvm/trace.h   |   12 ++--
 3 files changed, 158 insertions(+), 76 deletions(-)