"KVM: VMX" for the scope, or maybe even "KVM: nVMX:" despite this not touching vmx/nested.c. Either way (but not "kvm/vmx:"). On Fri, Feb 18, 2022, Peng Hao wrote: > Make sure nested_vmx_hardware_setup/unsetup are called in pairs under > the same conditions. Probably worth adding a sentence to clarify that the existing code isn't buggy only because free_page() plays nice with getting passed '0' and vmx_bitmap is initialized to zero. Something like: Make sure nested_vmx_hardware_setup/unsetup() are called in pairs under the same conditions. Calling nested_vmx_hardware_unsetup() when nested is false "works" right now because it only calls free_page() on zero- initialized pointers, but it's possible that more code will be added to nested_vmx_hardware_unsetup() in the future. Reviewed-by: Sean Christopherson <seanjc@xxxxxxxxxx> > > Signed-off-by: Peng Hao <flyingpeng@xxxxxxxxxxx> > --- > arch/x86/kvm/vmx/vmx.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c > index 0ffcfe54eea5..5392def71093 100644 > --- a/arch/x86/kvm/vmx/vmx.c > +++ b/arch/x86/kvm/vmx/vmx.c > @@ -7852,7 +7852,7 @@ static __init int hardware_setup(void) > vmx_set_cpu_caps(); > > r = alloc_kvm_area(); > - if (r) > + if (r && nested) > nested_vmx_hardware_unsetup(); > > kvm_set_posted_intr_wakeup_handler(pi_wakeup_handler); > -- > 2.27.0 >
