Kernel KVM virtualization development
 help / color / mirror / Atom feed
From: Sean Christopherson <seanjc@google.com>
To: Jim Mattson <jmattson@google.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>,
	Thomas Gleixner <tglx@kernel.org>, Ingo Molnar <mingo@redhat.com>,
	 Borislav Petkov <bp@alien8.de>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	x86@kernel.org,  "H. Peter Anvin" <hpa@zytor.com>,
	Shuah Khan <shuah@kernel.org>,
	kvm@vger.kernel.org,  linux-kernel@vger.kernel.org,
	linux-kselftest@vger.kernel.org,
	 Clayton Pence <ctpence@google.com>
Subject: Re: [PATCH v2 2/3] KVM: x86: Virtualize AMD CPUID faulting
Date: Tue, 12 May 2026 14:33:55 -0700	[thread overview]
Message-ID: <agOcw9ZIXALKrs_a@google.com> (raw)
In-Reply-To: <20260508170714.489136-3-jmattson@google.com>

On Fri, May 08, 2026, Jim Mattson wrote:
>  static __always_inline void kvm_cpu_cap_clear(unsigned int x86_feature)
> diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
> index c8c6cc0406d6..127d031d3310 100644
> --- a/arch/x86/kvm/emulate.c
> +++ b/arch/x86/kvm/emulate.c
> @@ -3593,13 +3593,15 @@ static int em_sti(struct x86_emulate_ctxt *ctxt)
>  static int em_cpuid(struct x86_emulate_ctxt *ctxt)
>  {
>  	u32 eax, ebx, ecx, edx;
> -	u64 msr = 0;
> +	u64 msr[2] = {};
>  
> -	ctxt->ops->get_msr(ctxt, MSR_MISC_FEATURES_ENABLES, &msr);
> -	if (!ctxt->ops->is_smm(ctxt) &&
> -	    (msr & MSR_MISC_FEATURES_ENABLES_CPUID_FAULT) &&
> -	    ctxt->ops->cpl(ctxt))
> -		return emulate_gp(ctxt, 0);
> +	if (!ctxt->ops->is_smm(ctxt) && ctxt->ops->cpl(ctxt)) {
> +		ctxt->ops->get_msr(ctxt, MSR_MISC_FEATURES_ENABLES, &msr[0]);
> +		ctxt->ops->get_msr(ctxt, MSR_K7_HWCR, &msr[1]);
> +		if ((msr[0] & MSR_MISC_FEATURES_ENABLES_CPUID_FAULT) ||
> +		    (msr[1] & BIT_ULL(MSR_K7_HWCR_CPUID_USER_DIS_BIT)))

Hrm, this is all rather ugly (definitely not your fault).  What do you think about
adding x86_emulate_ops.cpuid_fault_enabled() to at least get deduplicate the MSR
lookups?

Gah, we should do more than that.  I was going to say, "If it weren't for
kvm_emulate_cpuid() using kvm_require_cpl(), we could roll *all* of the checks
into cpuid_fault_enabled()".  But that's a rather stupid reason to duplicate the
code, because kvm_emulate_cpuid() is literally the only user of kvm_require_cpl().

Pulling smm.h into cpuid.h isn't great, but the code is much nicer, especially
the next patch to add AMD CPUID faulting.

Compile tested only...

---
From: Sean Christopherson <seanjc@google.com>
Date: Tue, 12 May 2026 14:20:47 -0700
Subject: [PATCH 1/3] KVM: x86: Consolidate CPUID fault handling for emulator
 and interception logic

Extract the logic for emulating CPUID faulting (where CPUID #GPs at CPL>0
outside of SMM) into a dedicated helper and use the helper for both the
full emulator and the intercepted-CPUID paths.

Opportunistically drop kvm_require_cpl(), as kvm_require_cpl() was the one
and only user.

No functional change intended.

Signed-off-by: Sean Christopherson <seanjc@google.com>
---
 arch/x86/include/asm/kvm_host.h |  1 -
 arch/x86/kvm/cpuid.c            |  5 +++--
 arch/x86/kvm/cpuid.h            |  8 ++++++++
 arch/x86/kvm/emulate.c          |  6 +-----
 arch/x86/kvm/kvm_emulate.h      |  1 +
 arch/x86/kvm/x86.c              | 18 ++++++------------
 6 files changed, 19 insertions(+), 20 deletions(-)

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index c470e40a00aa..a9005c61619b 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -2285,7 +2285,6 @@ void kvm_requeue_exception(struct kvm_vcpu *vcpu, unsigned int nr,
 void kvm_inject_page_fault(struct kvm_vcpu *vcpu, struct x86_exception *fault);
 void kvm_inject_emulated_page_fault(struct kvm_vcpu *vcpu,
 				    struct x86_exception *fault);
-bool kvm_require_cpl(struct kvm_vcpu *vcpu, int required_cpl);
 bool kvm_require_dr(struct kvm_vcpu *vcpu, int dr);
 
 static inline int __kvm_irq_line_state(unsigned long *irq_state,
diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index e69156b54cff..1c95d1fa3ead 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -2161,9 +2161,10 @@ int kvm_emulate_cpuid(struct kvm_vcpu *vcpu)
 {
 	u32 eax, ebx, ecx, edx;
 
-	if (!is_smm(vcpu) && cpuid_fault_enabled(vcpu) &&
-	    !kvm_require_cpl(vcpu, 0))
+	if (!kvm_is_cpuid_allowed(vcpu)) {
+		kvm_queue_exception_e(vcpu, GP_VECTOR, 0);
 		return 1;
+	}
 
 	eax = kvm_rax_read(vcpu);
 	ecx = kvm_rcx_read(vcpu);
diff --git a/arch/x86/kvm/cpuid.h b/arch/x86/kvm/cpuid.h
index 8b64d863e19c..95d09ccbf951 100644
--- a/arch/x86/kvm/cpuid.h
+++ b/arch/x86/kvm/cpuid.h
@@ -7,6 +7,8 @@
 #include <asm/processor.h>
 #include <uapi/asm/kvm_para.h>
 
+#include "smm.h"
+
 extern u32 kvm_cpu_caps[NR_KVM_CPU_CAPS] __read_mostly;
 extern bool kvm_is_configuring_cpu_caps __read_mostly;
 
@@ -187,6 +189,12 @@ static inline bool cpuid_fault_enabled(struct kvm_vcpu *vcpu)
 		  MSR_MISC_FEATURES_ENABLES_CPUID_FAULT;
 }
 
+static inline bool kvm_is_cpuid_allowed(struct kvm_vcpu *vcpu)
+{
+	return !cpuid_fault_enabled(vcpu) || is_smm(vcpu) ||
+	       !kvm_x86_call(get_cpl)(vcpu);
+}
+
 static __always_inline void kvm_cpu_cap_clear(unsigned int x86_feature)
 {
 	unsigned int x86_leaf = __feature_leaf(x86_feature);
diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
index c8c6cc0406d6..3ba09093b5ab 100644
--- a/arch/x86/kvm/emulate.c
+++ b/arch/x86/kvm/emulate.c
@@ -3593,12 +3593,8 @@ static int em_sti(struct x86_emulate_ctxt *ctxt)
 static int em_cpuid(struct x86_emulate_ctxt *ctxt)
 {
 	u32 eax, ebx, ecx, edx;
-	u64 msr = 0;
 
-	ctxt->ops->get_msr(ctxt, MSR_MISC_FEATURES_ENABLES, &msr);
-	if (!ctxt->ops->is_smm(ctxt) &&
-	    (msr & MSR_MISC_FEATURES_ENABLES_CPUID_FAULT) &&
-	    ctxt->ops->cpl(ctxt))
+	if (!ctxt->ops->is_cpuid_allowed(ctxt))
 		return emulate_gp(ctxt, 0);
 
 	eax = reg_read(ctxt, VCPU_REGS_RAX);
diff --git a/arch/x86/kvm/kvm_emulate.h b/arch/x86/kvm/kvm_emulate.h
index 0abff36d0994..45d4a03b202e 100644
--- a/arch/x86/kvm/kvm_emulate.h
+++ b/arch/x86/kvm/kvm_emulate.h
@@ -225,6 +225,7 @@ struct x86_emulate_ops {
 			 struct x86_instruction_info *info,
 			 enum x86_intercept_stage stage);
 
+	bool (*is_cpuid_allowed)(struct x86_emulate_ctxt *ctxt);
 	bool (*get_cpuid)(struct x86_emulate_ctxt *ctxt, u32 *eax, u32 *ebx,
 			  u32 *ecx, u32 *edx, bool exact_only);
 	bool (*guest_has_movbe)(struct x86_emulate_ctxt *ctxt);
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 7d5b66287593..c60773349f35 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -1021,18 +1021,6 @@ void kvm_queue_exception_e(struct kvm_vcpu *vcpu, unsigned nr, u32 error_code)
 }
 EXPORT_SYMBOL_FOR_KVM_INTERNAL(kvm_queue_exception_e);
 
-/*
- * Checks if cpl <= required_cpl; if true, return true.  Otherwise queue
- * a #GP and return false.
- */
-bool kvm_require_cpl(struct kvm_vcpu *vcpu, int required_cpl)
-{
-	if (kvm_x86_call(get_cpl)(vcpu) <= required_cpl)
-		return true;
-	kvm_queue_exception_e(vcpu, GP_VECTOR, 0);
-	return false;
-}
-
 bool kvm_require_dr(struct kvm_vcpu *vcpu, int dr)
 {
 	if ((dr != 4 && dr != 5) || !kvm_is_cr4_bit_set(vcpu, X86_CR4_DE))
@@ -8819,6 +8807,11 @@ static int emulator_intercept(struct x86_emulate_ctxt *ctxt,
 					     &ctxt->exception);
 }
 
+static bool emulator_is_cpuid_allowed(struct x86_emulate_ctxt *ctxt)
+{
+	return kvm_is_cpuid_allowed(emul_to_vcpu(ctxt));
+}
+
 static bool emulator_get_cpuid(struct x86_emulate_ctxt *ctxt,
 			      u32 *eax, u32 *ebx, u32 *ecx, u32 *edx,
 			      bool exact_only)
@@ -8955,6 +8948,7 @@ static const struct x86_emulate_ops emulate_ops = {
 	.wbinvd              = emulator_wbinvd,
 	.fix_hypercall       = emulator_fix_hypercall,
 	.intercept           = emulator_intercept,
+	.is_cpuid_allowed    = emulator_is_cpuid_allowed,
 	.get_cpuid           = emulator_get_cpuid,
 	.guest_has_movbe     = emulator_guest_has_movbe,
 	.guest_has_fxsr      = emulator_guest_has_fxsr,

base-commit: 41a2602f2947b91e237cd1ce85afcd162274470a
--

  reply	other threads:[~2026-05-12 21:33 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-08 17:07 [PATCH v2 0/3] KVM: x86: Virtualize AMD CPUID faulting Jim Mattson
2026-05-08 17:07 ` [PATCH v2 1/3] KVM: x86: Remove supports_cpuid_fault() helper Jim Mattson
2026-05-08 17:07 ` [PATCH v2 2/3] KVM: x86: Virtualize AMD CPUID faulting Jim Mattson
2026-05-12 21:33   ` Sean Christopherson [this message]
2026-05-13  0:01     ` Jim Mattson
2026-05-13  0:04       ` Sean Christopherson
2026-05-08 17:07 ` [PATCH v2 3/3] KVM: selftests: Update hwcr_msr_test for CPUID faulting bit Jim Mattson

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=agOcw9ZIXALKrs_a@google.com \
    --to=seanjc@google.com \
    --cc=bp@alien8.de \
    --cc=ctpence@google.com \
    --cc=dave.hansen@linux.intel.com \
    --cc=hpa@zytor.com \
    --cc=jmattson@google.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-kselftest@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=shuah@kernel.org \
    --cc=tglx@kernel.org \
    --cc=x86@kernel.org \
    /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