Kernel KVM virtualization development
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Chang S. Bae" <chang.seok.bae@intel.com>
Cc: kvm@vger.kernel.org
Subject: Re: [PATCH v7 02/20] KVM: VMX: Save guest EGPRs in VCPU cache
Date: Wed, 02 Sep 2026 03:26:18 +0000	[thread overview]
Message-ID: <20260902032618.A76FC1F000E9@smtp.kernel.org> (raw)
In-Reply-To: <20260902021804.718676-3-chang.seok.bae@intel.com>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Host kernel data in Extended General Purpose Registers (EGPRs) is leaked to a malicious guest via speculative execution when the guest operates with APX disabled.
--

commit 70442dd137e1b61dbdd5b674afc21cfb7948b2b6
Author: Chang S. Bae <chang.seok.bae@intel.com>
KVM: VMX: Save guest EGPRs in VCPU cache

This commit saves and restores the guest EGPRs on VM exit/entry if the
system supports Advanced Performance Extensions (APX). It skips loading the
EGPR state for guests that do not set XCR0[APX], and places the register
saving path after SPEC_CTRL is restored.

Link: https://lore.kernel.org/adPRA4ZhnvbaXSn0@google.com

> diff --git a/arch/x86/kvm/vmx/vmenter.S b/arch/x86/kvm/vmx/vmenter.S
> index 00d807505fc88..02308fb36f0f5 100644
> --- a/arch/x86/kvm/vmx/vmenter.S
> +++ b/arch/x86/kvm/vmx/vmenter.S

[ ... ]

> @@ -46,6 +47,16 @@ SYM_FUNC_START(__vmx_vcpu_run)
>  	/* Reload @vmx, _ASM_ARG1 may be modified by vmx_update_host_rsp().  */
>  	mov WORD_SIZE(%_ASM_SP), %_ASM_DI
>  
> +#ifdef CONFIG_KVM_APX
> +	ALTERNATIVE "jmp .Lload_egprs_done", "", X86_FEATURE_APX
> +	testl	$KVM_ENTER_EGPR_SWITCH, (%_ASM_SP)
> +	jz	.Lload_egprs_done
> +	LOAD_REGS %_ASM_DI, VMX_vcpu_arch_regs, \
> +		  %r16, %r17, %r18, %r19, %r20, %r21, %r22, %r23, \
> +		  %r24, %r25, %r26, %r27, %r28, %r29, %r30, %r31
> +.Lload_egprs_done:
> +#endif

[Severity: High]
In __vmx_vcpu_run(), when a guest runs with APX disabled,
KVM_ENTER_EGPR_SWITCH is clear and this branch skips loading the guest's
EGPRs. Are the host's physical %r16 through %r31 register values left
intact upon entering the guest when this happens?

If so, could a malicious guest with APX disabled speculatively execute APX
instructions before the architectural #UD exception retires, allowing it to
read the host's %r16 through %r31 values and leak them via a cache
side-channel?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260902021804.718676-1-chang.seok.bae@intel.com?part=2

  reply	other threads:[~2026-09-02  3:26 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-02  2:17 [PATCH v7 00/20] KVM: x86: Enable APX for guests Chang S. Bae
2026-09-02  2:17 ` [PATCH v7 01/20] KVM: x86: Extend VCPU registers for EGPRs Chang S. Bae
2026-09-02  2:17 ` [PATCH v7 02/20] KVM: VMX: Save guest EGPRs in VCPU cache Chang S. Bae
2026-09-02  3:26   ` sashiko-bot [this message]
2026-09-02  2:17 ` [PATCH v7 03/20] KVM: x86: Support APX state for XSAVE ABI Chang S. Bae
2026-09-02  3:08   ` sashiko-bot
2026-09-02  2:17 ` [PATCH v7 04/20] KVM: VMX: Refactor VMX instruction information access Chang S. Bae
2026-09-02  2:17 ` [PATCH v7 05/20] KVM: VMX: Refactor instruction information decoding Chang S. Bae
2026-09-02  2:17 ` [PATCH v7 06/20] KVM: VMX: Remove unused control-register access defines Chang S. Bae
2026-09-02  2:17 ` [PATCH v7 07/20] KVM: VMX: Refactor register index retrieval from exit qualification Chang S. Bae
2026-09-02  2:17 ` [PATCH v7 08/20] KVM: VMX: Support instruction information extension Chang S. Bae
2026-09-02  2:17 ` [PATCH v7 09/20] KVM: nVMX: Propagate extended instruction information Chang S. Bae
2026-09-02  3:07   ` sashiko-bot
2026-09-02  2:17 ` [PATCH v7 10/20] KVM: x86: Support EGPR accessing and tracking for emulator Chang S. Bae
2026-09-02  2:17 ` [PATCH v7 11/20] KVM: x86: Handle EGPR index and REX2-incompatible opcodes Chang S. Bae
2026-09-02  3:06   ` sashiko-bot
2026-09-02  2:17 ` [PATCH v7 12/20] KVM: x86: Support REX2-prefixed opcode decode Chang S. Bae
2026-09-02  2:17 ` [PATCH v7 13/20] KVM: x86: Reject EVEX-prefixed instructions Chang S. Bae
2026-09-02  2:59   ` sashiko-bot
2026-09-02  2:17 ` [PATCH v7 14/20] KVM: x86: Move KVM_SUPPORTED_{XCR0,XSS} into kvm_x86_vendor_init() Chang S. Bae
2026-09-02  2:17 ` [PATCH v7 15/20] KVM: x86: Guard valid XCR0.APX settings Chang S. Bae
2026-09-02  2:18 ` [PATCH v7 16/20] KVM: x86: Add APX to supported XCR0 Chang S. Bae
2026-09-02  3:07   ` sashiko-bot
2026-09-02  2:18 ` [PATCH v7 17/20] KVM: x86: Expose APX foundation feature to userspace Chang S. Bae
2026-09-02  2:18 ` [PATCH v7 18/20] KVM: x86: Expose APX sub-features " Chang S. Bae
2026-09-02  3:00   ` sashiko-bot
2026-09-02  2:18 ` [PATCH v7 19/20] KVM: x86: selftests: Add APX state and ABI test Chang S. Bae
2026-09-02  3:07   ` sashiko-bot
2026-09-02  2:18 ` [PATCH v7 20/20] KVM: x86: selftests: Add APX state handling and XCR0 sanity checks Chang S. Bae

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=20260902032618.A76FC1F000E9@smtp.kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=chang.seok.bae@intel.com \
    --cc=kvm@vger.kernel.org \
    --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