Re: Re: [PATCH v3 2/2] x86/svm: Use the virtual NMI when available

[Abdelkareem Abdelsaamad <abdelkareem.abdelsaamad@xxxxxxxxxx>]

On 16/02/2026 10:15 am, Andrew Cooper wrote:
>So what happens to the NMI we're trying to inject when NMIs are
>currently blocked?

Ah, I see this...
The handling flow will eventually land on the function svm_enable_intr_window
with HVM_INTR_SHADOW_NMI. The logic inside the function svm_enable_intr_window
will drive it to create a vintr and set up the VMCB to intercept that created
vintr. I see that the code logic in svm_enable_intr_window should be extended
to skip creating such vintr and just return the same as the handling with
Event injected NMI when the GENERAL1_INTERCEPT_IRET is already set.

>~Andrew

--Abdelkareem