When vCPU is in VMX operation and enters SMM mode,
it temporarily exits VMX operation but KVM maintained nested-state
still stores the VMXON region physical address, i.e. even when the
vCPU is in SMM mode then (nested_state->hdr.vmx.vmxon_pa != -1ull).
Therefore, there is no need to explicitly check for
KVM_STATE_NESTED_SMM_VMXON to determine if it is necessary
to save nested-state as part of migration stream.
In addition, destination must enable eVMCS if it is enabled on
source as specified by the KVM_STATE_NESTED_EVMCS flag, even if
the VMXON region is not set. Thus, change the code to require saving
nested-state as part of migration stream in case it is set.
Reviewed-by: Karl Heubaum <karl.heubaum@xxxxxxxxxx>
Signed-off-by: Liran Alon <liran.alon@xxxxxxxxxx>
---
target/i386/machine.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/target/i386/machine.c b/target/i386/machine.c
index 851b249d1a39..e7d72faf9e24 100644
--- a/target/i386/machine.c
+++ b/target/i386/machine.c
@@ -999,7 +999,7 @@ static bool vmx_nested_state_needed(void *opaque)
return ((nested_state->format == KVM_STATE_NESTED_FORMAT_VMX) &&
((nested_state->hdr.vmx.vmxon_pa != -1ull) ||
- (nested_state->hdr.vmx.smm.flags & KVM_STATE_NESTED_SMM_VMXON)));
+ (nested_state->flags & KVM_STATE_NESTED_EVMCS)));
}
static const VMStateDescription vmstate_vmx_nested_state = {
--
2.20.1
