[v3,00/28] KVM: nSVM: event fixes and migration support
mbox series

Message ID 20200529153934.11694-1-pbonzini@redhat.com
Headers show
Series
  • KVM: nSVM: event fixes and migration support
Related show

Message

Paolo Bonzini May 29, 2020, 3:39 p.m. UTC
This is basically the same as v2 except that it has a small fix to
"KVM: x86: enable event window in inject_pending_event", where
a second pending interrupt or NMI was not enabling the window-open
vmexit (caught by apic.flat).  In addition I've renamed
inject_pending_event to handle_processor_events.

The series now passes kvm-unit-tests and various nested hypervisor tests
so now it's *really* ready for review!  (Thanks Krish for looking at
it so far).

I'm quite pleased with the overall look of the code, though the
INT_CTL arbitration is a bit ugly.  I have plans to implement nested
vGIF and vLS, and then I will probably clean it up.

Paolo

Paolo Bonzini (28):
  KVM: x86: track manually whether an event has been injected
  KVM: x86: enable event window in inject_pending_event
  KVM: nSVM: inject exceptions via svm_check_nested_events
  KVM: nSVM: remove exit_required
  KVM: nSVM: correctly inject INIT vmexits
  KVM: SVM: always update CR3 in VMCB
  KVM: nVMX: always update CR3 in VMCS
  KVM: nSVM: move map argument out of enter_svm_guest_mode
  KVM: nSVM: extract load_nested_vmcb_control
  KVM: nSVM: extract preparation of VMCB for nested run
  KVM: nSVM: move MMU setup to nested_prepare_vmcb_control
  KVM: nSVM: clean up tsc_offset update
  KVM: nSVM: pass vmcb_control_area to copy_vmcb_control_area
  KVM: nSVM: remove trailing padding for struct vmcb_control_area
  KVM: nSVM: save all control fields in svm->nested
  KVM: nSVM: restore clobbered INT_CTL fields after clearing VINTR
  KVM: nSVM: synchronize VMCB controls updated by the processor on every
    vmexit
  KVM: nSVM: remove unnecessary if
  KVM: nSVM: extract svm_set_gif
  KVM: SVM: preserve VGIF across VMCB switch
  KVM: nSVM: synthesize correct EXITINTINFO on vmexit
  KVM: nSVM: remove HF_VINTR_MASK
  KVM: nSVM: remove HF_HIF_MASK
  KVM: nSVM: split nested_vmcb_check_controls
  KVM: nSVM: leave guest mode when clearing EFER.SVME
  KVM: MMU: pass arbitrary CR0/CR4/EFER to kvm_init_shadow_mmu
  selftests: kvm: add a SVM version of state-test
  KVM: nSVM: implement KVM_GET_NESTED_STATE and KVM_SET_NESTED_STATE

Vitaly Kuznetsov (2):
  selftests: kvm: introduce cpu_has_svm() check
  selftests: kvm: fix smm test on SVM

 arch/x86/include/asm/kvm_host.h               |  12 +-
 arch/x86/include/asm/svm.h                    |   9 +-
 arch/x86/include/uapi/asm/kvm.h               |  17 +-
 arch/x86/kvm/cpuid.h                          |   5 +
 arch/x86/kvm/irq.c                            |   1 +
 arch/x86/kvm/mmu.h                            |   2 +-
 arch/x86/kvm/mmu/mmu.c                        |  14 +-
 arch/x86/kvm/svm/nested.c                     | 624 ++++++++++++------
 arch/x86/kvm/svm/svm.c                        | 154 ++---
 arch/x86/kvm/svm/svm.h                        |  33 +-
 arch/x86/kvm/vmx/nested.c                     |   5 -
 arch/x86/kvm/vmx/vmx.c                        |  25 +-
 arch/x86/kvm/x86.c                            | 146 ++--
 .../selftests/kvm/include/x86_64/svm_util.h   |  10 +
 tools/testing/selftests/kvm/x86_64/smm_test.c |  19 +-
 .../testing/selftests/kvm/x86_64/state_test.c |  62 +-
 16 files changed, 708 insertions(+), 430 deletions(-)

Comments

Sean Christopherson May 29, 2020, 5:59 p.m. UTC | #1
> [PATCH v3 00/28] KVM: nSVM: event fixes and migration support

You've got something funky going on with the way you generate cover letters,
looks like it doesn't count patches authored by someone else.  The 'v3' is
also missing from the patches, though I suppose some heathens do that on
purpose.

> Paolo Bonzini (28):
> 
> Vitaly Kuznetsov (2):
Paolo Bonzini May 29, 2020, 7:07 p.m. UTC | #2
On 29/05/20 19:59, Sean Christopherson wrote:
>> [PATCH v3 00/28] KVM: nSVM: event fixes and migration support
> You've got something funky going on with the way you generate cover letters,
> looks like it doesn't count patches authored by someone else.  The 'v3' is
> also missing from the patches, though I suppose some heathens do that on
> purpose.

No, I simply had a draft of the cover letter ready, and when I decided
to include Vitaly's patches I didn't update the subject.

Paolo