On Tue, 2026-11-03 at 22:55 -0800, Sriram Nambakam wrote:
> Hi David,
> 
> Thanks for posting this. I've been trying the series and the QEMU VSM
> implementation in my local integration trees. With the fixes below,
> Windows now reports "Virtualization-based security: Running" in msinfo32.

Great, thanks for testing!

> This was on a v7.3-based kernel tree with existing local VM-planes/LVBS
> changes, rather than an otherwise unmodified copy of your v7.1-rc7-based
> tree. I don't want to imply that all of the integration changes were
> defects in the posted series.
>
> The main runtime issue I found was descriptor-table interception while
> OVMF was executing in SMM. hv_divert_dt_write() was diverting firmware
> descriptor-table writes through the VTL register-intercept path. The
> failure trace showed a reflected #GP delivered in SMM with IDT base and
> limit both zero, followed by a triple fault/reset. Adding an is_smm()
> exclusion to the helper fixed that boot failure. I'll reply to patch 23
> with the change.

Thanks. I'll add OVMF with Secure Boot (hence SMM) to the test matrix,
and pull in your fixes. Could I trouble you for a Signed-off-by: for
each of those please? 

> On the QEMU side, I also changed vsm_glue_reset() to make plane 0 runnable
> when stopping plane 1. Otherwise a reset following a VTL switch can leave
> the normal-world vCPU userspace-stopped. I'll send that change separately
> in this thread, since the QEMU glue isn't part of the kernel posting.
> 
> Two build/header items from the newer kernel integration:
> 
>   - In patch 49, I used cpu_feature_enabled(X86_FEATURE_DECODEASSISTS)
>     instead of static_cpu_has(), matching the current SVM code.
>   - I synchronized the tools UAPI header with the PLANE_PROT/INSN_VALID
>     flags and their field-validity comments, and the capability-version
>     requirement for KVM_HV_HCALL_RESULT_RAISE_UD.

I'm playing with the HV_HCALL stuff to make the capability into a
bitmask, like the Xen one, which says which calls are handled in-
kernel. That means the basic support can be submitted first, before
anything else.

> I also encountered an issue in the underlying planes dependency:
> "kvm: Make KVM_SIGNAL_MSI per plane" guards the VM-fd and plane-fd ioctl
> dispatch with CONFIG_HAVE_KVM_MSI. Those guards need to use __KVM_HAVE_MSI;
> otherwise KVM_SIGNAL_MSI is compiled out and returns ENOTTY. The existing
> test_plane_msi case exposed this in my tree. This isn't introduced by one
> of the 55 patches here.

Yeah, I have a few more fixes for the underlying Planes series, which
I'll try to split out for Jörg too.

> Finally, my older LVBS tree had raised the default plane count to two.
> I reverted that local change to respect the KVM_CAP_PLANES opt-in contract
> and added early opt-in for my LVBS userspace path. Your QEMU VSM path
> already has its own opt-in, so I am not reporting that as a missing VSM
> change.

There are going to be a lot of intersecting details like that as the
various trees come together and the dust settles. I'm trying to get
things in some semblance of a mergeable order and form a plan for what
could be posted immediately (the basic hv hypercall exit to userspace),
what could follow on as soon as Planes/memattrs land, and then what
needs more thought.

I'd like to get the *minimum* kernel functionality landed, and then
follow on with "optimisations" like the in-kernel VTL switching and
extra features like VINA etc.

Attachment: smime.p7s
Description: S/MIME cryptographic signature

Reply via email to