Re: [PATCH 1/3] KVM: SVM: Refactor EFER.SVME switching logic out of svm_set_efer()

[Sean Christopherson <seanjc@google.com>]

On Fri, Jan 30, 2026, Yosry Ahmed wrote:
> Move the logic of switching EFER.SVME in the guest outside of
> svm_set_efer(). This makes it possible to easily check the skip
> conditions separately (and add more) and reduce indentation level.
> 
> No functional change intended.
> 
> Signed-off-by: Yosry Ahmed <yosry.ahmed@linux.dev>
> ---
>  arch/x86/kvm/svm/svm.c | 72 ++++++++++++++++++++++++------------------
>  1 file changed, 42 insertions(+), 30 deletions(-)
> 
> diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
> index 5f0136dbdde6b..4575a6a7d6c4e 100644
> --- a/arch/x86/kvm/svm/svm.c
> +++ b/arch/x86/kvm/svm/svm.c
> @@ -200,11 +200,49 @@ static int get_npt_level(void)
>  #endif
>  }
>  
> +static int svm_set_efer_svme(struct kvm_vcpu *vcpu, u64 old_efer, u64 new_efer)

Code looks good, but I think we need a better name.  This helper doesn't actually
write vcpu->arch.efer, and the name can also be misconstrued as "set EFER.SVME=1".

How about svm_post_set_efer(), to align with kvm_post_set_cr{0,3,4}()?  It's not
perfect, but I can't come up with something that's more accurate without being
stupidly verbose.