On Tue, Mar 26, 2019 at 02:07:46PM +0100, Vitaly Kuznetsov wrote:
> Commit 5bea5123cbf0 ("KVM: VMX: check nested state and CR4.VMXE against
> SMM") introduced a check to vmx_set_cr4() forbidding to set VMXE from SMM.
> The check is correct, however, there is a special case when RSM is called
> to leave SMM: rsm_enter_protected_mode() is called with HF_SMM_MASK still
> set and in case VMXE was set before entering SMM we're failing to return.
>
> Resolve the issue by temporary dropping HF_SMM_MASK around set_cr4() calls
> when ops->set_cr() is called from RSM.
>
> Reported-by: Jon Doron <arilou@xxxxxxxxx>
> Suggested-by: Liran Alon <liran.alon@xxxxxxxxxx>
> Fixes: 5bea5123cbf0 ("KVM: VMX: check nested state and CR4.VMXE against SMM")
> Signed-off-by: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>
Plumbing from_rsm all the way to set_cr() is pretty heinous. What about
going with the idea Jim alluded to, i.e. manually save/restore VMXE during
transitions to/from SMM? It's a little more tedious than it aught to be
due to the placement and naming of the x86_ops hooks for SMM, but IMO the
end result is cleaner even after adding the necessary callbacks.
The following patches are compile tested only.
Sean Christopherson (2):
KVM: x86: Rename pre_{enter,leave}_smm() ops to reference SMM state
save
KVM: x86: Add kvm_x86_ops callback to allow VMX to stash away CR4.VMXE
arch/x86/include/asm/kvm_emulate.h | 3 ++-
arch/x86/include/asm/kvm_host.h | 6 ++++--
arch/x86/kvm/emulate.c | 10 ++++++----
arch/x86/kvm/svm.c | 20 ++++++++++++++++----
arch/x86/kvm/vmx/vmx.c | 30 ++++++++++++++++++++++++++----
arch/x86/kvm/vmx/vmx.h | 2 ++
arch/x86/kvm/x86.c | 23 ++++++++++++++++-------
7 files changed, 72 insertions(+), 22 deletions(-)
--
2.21.0
