Kernel KVM virtualization development
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Paolo Bonzini" <pbonzini@redhat.com>, kvm@vger.kernel.org
Subject: Re: [PATCH 1/2] KVM: nVMX: Reset MMU back to the root MMU if loading L2 state fails for !VM-Entry
Date: Thu, 15 Oct 2026 09:02:20 +0000	[thread overview]
Message-ID: <sashiko-outbox-172536@kernel.org> (raw)
In-Reply-To: <20261015083123.790030-2-pbonzini@redhat.com>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] kvm: vmx: nested state corruption (CR0/CR4/EFER) on KVM_SET_NESTED_STATE failure
--

commit 55d55918900f4eae4bf7636855cdc83ba4433cf5
Author: Amirmohammad Eftekhar <amirmohammadeftekhar@gmail.com>

KVM: nVMX: Reset MMU back to the root MMU if loading L2 state fails for !VM-Entry

This patch resets the vCPU's MMU pointer back to the root MMU if loading L2
state fails when not coming from VM-Entry. This prevents KVM from hitting
a NULL pointer dereference when attempting to run L1 with a nested MMU.

Link: https://patch.msgid.link/20261001214356.3825100-2-seanjc@google.com

> diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c
> index b252168627408..5165275128fce 100644
> --- a/arch/x86/kvm/vmx/nested.c
> +++ b/arch/x86/kvm/vmx/nested.c
> @@ -3770,8 +3770,10 @@ enum nvmx_vmentry_status nested_vmx_enter_non_root_mode(struct kvm_vcpu *vcpu,
>  vmentry_fail_vmexit:
>  	vmx_switch_vmcs(vcpu, &vmx->vmcs01);
>  
> -	if (!from_vmentry)
> +	if (!from_vmentry) {
> +		nested_ept_uninit_mmu_context(vcpu);

[Severity: High]
When userspace invokes KVM_SET_NESTED_STATE, KVM executes
vmx_set_nested_state() which calls nested_vmx_enter_non_root_mode() with
from_vmentry set to false.

This unconditionally calls prepare_vmcs02(), which modifies the vCPU's
software state (such as cr0, cr4, efer, and tsc_offset) to match the
nested L2 state.

If prepare_vmcs02() subsequently fails, for example due to an invalid L2
CR3, it returns an error and jumps to vmentry_fail_vmexit_guest_mode.

While this patch fixes the MMU pointer, since from_vmentry is false, does
this early return bypass load_vmcs12_host_state() and leave the vCPU's
software state permanently corrupted with L2's values?

>  		return NVMX_VMENTRY_VMEXIT;
> +	}
>  
>  	nested_put_vmcs12_pages(vcpu);
>

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261015083123.790030-1-pbonzini@redhat.com?part=1

  reply	other threads:[~2026-10-15  9:02 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-15  8:31 [PATCH 0/2] KVM: x86: Reset MMU back to the root MMU when leaving guest mode Paolo Bonzini
2026-10-15  8:31 ` [PATCH 1/2] KVM: nVMX: Reset MMU back to the root MMU if loading L2 state fails for !VM-Entry Paolo Bonzini
2026-10-15  9:02   ` sashiko-bot [this message]
2026-10-15  9:16     ` Paolo Bonzini
2026-10-15  8:31 ` [PATCH 2/2] KVM: x86: Reset MMU back to the root MMU when leaving guest mode Paolo Bonzini

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=sashiko-outbox-172536@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=kvm@vger.kernel.org \
    --cc=pbonzini@redhat.com \
    --cc=sashiko-reviews@lists.linux.dev \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox