[PATCH v2 0/2] Use host cpuid facilities

[PATCH v2 0/2] Use host cpuid facilities

[Avi Kivity]

Use the host's cpuid facilities to see if features are supporte.  This
allows the clearcpuid boot parameter to work, and allows Linux to disable
nx if it is not enabled on all cpus.

Avi Kivity (2):
  KVM: SVM: Replace svm_has() by standard Linux cpuid accessors
  KVM: Mask KVM_GET_SUPPORTED_CPUID data with Linux cpuid info
    v2: mask the correct cpuid returned register

 arch/x86/kvm/svm.c |   15 +++++----------
 arch/x86/kvm/x86.c |    9 +++++++++
 2 files changed, 14 insertions(+), 10 deletions(-)

-- 
1.7.3.1

[PATCH v2 1/2] KVM: SVM: Replace svm_has() by standard Linux cpuid accessors

[Avi Kivity]

Instead of querying cpuid directly, use the Linux accessors (boot_cpu_has,
etc.).  This allows the things like the clearcpuid kernel command line to
work (when it's fixed wrt scattered cpuid bits).

Signed-off-by: Avi Kivity <avi@redhat.com>
---
 arch/x86/kvm/svm.c |   15 +++++----------
 1 files changed, 5 insertions(+), 10 deletions(-)

diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index 1a5757a..c6a7798 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -271,11 +271,6 @@ static u32 svm_msrpm_offset(u32 msr)
 
 #define MAX_INST_SIZE 15
 
-static inline u32 svm_has(u32 feat)
-{
-	return svm_features & feat;
-}
-
 static inline void clgi(void)
 {
 	asm volatile (__ex(SVM_CLGI));
@@ -381,7 +376,7 @@ static void svm_queue_exception(struct kvm_vcpu *vcpu, unsigned nr,
 	    nested_svm_check_exception(svm, nr, has_error_code, error_code))
 		return;
 
-	if (nr == BP_VECTOR && !svm_has(SVM_FEATURE_NRIP)) {
+	if (nr == BP_VECTOR && !static_cpu_has(X86_FEATURE_NRIPS)) {
 		unsigned long rip, old_rip = kvm_rip_read(&svm->vcpu);
 
 		/*
@@ -677,7 +672,7 @@ static __init int svm_hardware_setup(void)
 
 	svm_features = cpuid_edx(SVM_CPUID_FUNC);
 
-	if (!svm_has(SVM_FEATURE_NPT))
+	if (!boot_cpu_has(X86_FEATURE_NPT))
 		npt_enabled = false;
 
 	if (npt_enabled && !npt) {
@@ -876,7 +871,7 @@ static void init_vmcb(struct vcpu_svm *svm)
 	svm->nested.vmcb = 0;
 	svm->vcpu.arch.hflags = 0;
 
-	if (svm_has(SVM_FEATURE_PAUSE_FILTER)) {
+	if (boot_cpu_has(X86_FEATURE_PAUSEFILTER)) {
 		control->pause_filter_count = 3000;
 		control->intercept |= (1ULL << INTERCEPT_PAUSE);
 	}
@@ -2743,7 +2738,7 @@ static int svm_set_msr(struct kvm_vcpu *vcpu, unsigned ecx, u64 data)
 		svm->vmcb->save.sysenter_esp = data;
 		break;
 	case MSR_IA32_DEBUGCTLMSR:
-		if (!svm_has(SVM_FEATURE_LBRV)) {
+		if (!boot_cpu_has(X86_FEATURE_LBRV)) {
 			pr_unimpl(vcpu, "%s: MSR_IA32_DEBUGCTL 0x%llx, nop\n",
 					__func__, data);
 			break;
@@ -3527,7 +3522,7 @@ static void svm_set_supported_cpuid(u32 func, struct kvm_cpuid_entry2 *entry)
 				   additional features */
 
 		/* Support next_rip if host supports it */
-		if (svm_has(SVM_FEATURE_NRIP))
+		if (boot_cpu_has(X86_FEATURE_NRIPS))
 			entry->edx |= SVM_FEATURE_NRIP;
 
 		/* Support NPT for the guest if enabled */
-- 
1.7.3.1

[PATCH v2 2/2] KVM: Mask KVM_GET_SUPPORTED_CPUID data with Linux cpuid info

[Avi Kivity]

This allows Linux to mask cpuid bits if, for example, nx is enabled on only
some cpus.

Signed-off-by: Avi Kivity <avi@redhat.com>
---
 arch/x86/kvm/x86.c |    9 +++++++++
 1 files changed, 9 insertions(+), 0 deletions(-)

diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 2044302..788f064 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -2229,6 +2229,11 @@ out:
 	return r;
 }
 
+static void cpuid_mask(u32 *word, int wordnum)
+{
+	*word &= boot_cpu_data.x86_capability[wordnum];
+}
+
 static void do_cpuid_1_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 			   u32 index)
 {
@@ -2303,7 +2308,9 @@ static void do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 		break;
 	case 1:
 		entry->edx &= kvm_supported_word0_x86_features;
+		cpuid_mask(&entry->edx, 0);
 		entry->ecx &= kvm_supported_word4_x86_features;
+		cpuid_mask(&entry->ecx, 4);
 		/* we support x2apic emulation even if host does not support
 		 * it since we emulate x2apic in software */
 		entry->ecx |= F(X2APIC);
@@ -2394,7 +2401,9 @@ static void do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 		break;
 	case 0x80000001:
 		entry->edx &= kvm_supported_word1_x86_features;
+		cpuid_mask(&entry->edx, 1);
 		entry->ecx &= kvm_supported_word6_x86_features;
+		cpuid_mask(&entry->ecx, 6);
 		break;
 	}
 
-- 
1.7.3.1

Re: [PATCH v2 1/2] KVM: SVM: Replace svm_has() by standard Linux cpuid accessors

[Joerg Roedel]

On Tue, Nov 09, 2010 at 04:15:42PM +0200, Avi Kivity wrote:
> -	if (nr == BP_VECTOR && !svm_has(SVM_FEATURE_NRIP)) {
> +	if (nr == BP_VECTOR && !static_cpu_has(X86_FEATURE_NRIPS)) {

What is static_cpu_has and why you use it only here and boot_cpu_has
in all other places?

	Joerg

Re: [PATCH v2 1/2] KVM: SVM: Replace svm_has() by standard Linux cpuid accessors

[Avi Kivity]

On 11/11/2010 04:46 PM, Joerg Roedel wrote:
> On Tue, Nov 09, 2010 at 04:15:42PM +0200, Avi Kivity wrote:
> >  -	if (nr == BP_VECTOR&&  !svm_has(SVM_FEATURE_NRIP)) {
> >  +	if (nr == BP_VECTOR&&  !static_cpu_has(X86_FEATURE_NRIPS)) {
>
> What is static_cpu_has

It's like boot_cpu_has, only it works by patching instead of a dynamic test.

> and why you use it only here and boot_cpu_has
> in all other places?

A nano optimization, this is a more commonly used path.

-- 
error compiling committee.c: too many arguments to function

Re: [PATCH v2 1/2] KVM: SVM: Replace svm_has() by standard Linux cpuid accessors

[Joerg Roedel]

On Thu, Nov 11, 2010 at 04:50:10PM +0200, Avi Kivity wrote:
> On 11/11/2010 04:46 PM, Joerg Roedel wrote:
>> On Tue, Nov 09, 2010 at 04:15:42PM +0200, Avi Kivity wrote:
>> >  -	if (nr == BP_VECTOR&&  !svm_has(SVM_FEATURE_NRIP)) {
>> >  +	if (nr == BP_VECTOR&&  !static_cpu_has(X86_FEATURE_NRIPS)) {
>>
>> What is static_cpu_has
>
> It's like boot_cpu_has, only it works by patching instead of a dynamic test.
>
>> and why you use it only here and boot_cpu_has
>> in all other places?
>
> A nano optimization, this is a more commonly used path.

Ok, I was just curious because I couldn't find the static_cpu_has by a
quick grep. Thanks for the explanation.

Acked-by: Joerg Roedel <joerg.roedel@amd.com>

Re: [PATCH v2 0/2] Use host cpuid facilities

[Marcelo Tosatti]

On Tue, Nov 09, 2010 at 04:15:41PM +0200, Avi Kivity wrote:
> Use the host's cpuid facilities to see if features are supporte.  This
> allows the clearcpuid boot parameter to work, and allows Linux to disable
> nx if it is not enabled on all cpus.
> 
> Avi Kivity (2):
>   KVM: SVM: Replace svm_has() by standard Linux cpuid accessors
>   KVM: Mask KVM_GET_SUPPORTED_CPUID data with Linux cpuid info
>     v2: mask the correct cpuid returned register
> 
>  arch/x86/kvm/svm.c |   15 +++++----------
>  arch/x86/kvm/x86.c |    9 +++++++++
>  2 files changed, 14 insertions(+), 10 deletions(-)

Applied, thanks.