kvm.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v2 1/2] kvm: x86: revert mask out xsaves
@ 2014-12-02  6:14 Wanpeng Li
  2014-12-02  6:14 ` [PATCH v2 2/2] kvm: vmx: enable intel xsaves for guest Wanpeng Li
  2014-12-02  8:55 ` [PATCH v2 1/2] kvm: x86: revert mask out xsaves Paolo Bonzini
  0 siblings, 2 replies; 7+ messages in thread
From: Wanpeng Li @ 2014-12-02  6:14 UTC (permalink / raw)
  To: Paolo Bonzini; +Cc: kvm, linux-kernel, Wanpeng Li

xsaves will be exported to guest in the next patch, so revert the
mask out xsaves patch.

Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com>
---
 arch/x86/kvm/cpuid.c | 10 +---------
 1 file changed, 1 insertion(+), 9 deletions(-)

diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index a4f5ac4..7af07571 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -320,10 +320,6 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 		F(ADX) | F(SMAP) | F(AVX512F) | F(AVX512PF) | F(AVX512ER) |
 		F(AVX512CD);
 
-	/* cpuid 0xD.1.eax */
-	const u32 kvm_supported_word10_x86_features =
-		F(XSAVEOPT) | F(XSAVEC) | F(XGETBV1);
-
 	/* all calls to cpuid_count() should be made on the same cpu */
 	get_cpu();
 
@@ -460,8 +456,6 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 		entry->eax &= supported;
 		entry->edx &= supported >> 32;
 		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
-		if (!supported)
-			break;
 
 		for (idx = 1, i = 1; idx < 64; ++idx) {
 			u64 mask = ((u64)1 << idx);
@@ -469,9 +463,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 				goto out;
 
 			do_cpuid_1_ent(&entry[i], function, idx);
-			if (idx == 1)
-				entry[i].eax &= kvm_supported_word10_x86_features;
-			else if (entry[i].eax == 0 || !(supported & mask))
+			if (entry[i].eax == 0 || !(supported & mask))
 				continue;
 			entry[i].flags |=
 			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
-- 
1.9.1

^ permalink raw reply related	[flat|nested] 7+ messages in thread

* [PATCH v2 2/2] kvm: vmx: enable intel xsaves for guest
  2014-12-02  6:14 [PATCH v2 1/2] kvm: x86: revert mask out xsaves Wanpeng Li
@ 2014-12-02  6:14 ` Wanpeng Li
  2014-12-02  8:51   ` Paolo Bonzini
  2014-12-02  8:55 ` [PATCH v2 1/2] kvm: x86: revert mask out xsaves Paolo Bonzini
  1 sibling, 1 reply; 7+ messages in thread
From: Wanpeng Li @ 2014-12-02  6:14 UTC (permalink / raw)
  To: Paolo Bonzini; +Cc: kvm, linux-kernel, Wanpeng Li

Expose intel xsaves feature to guest.

Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com>
---
v1 -> v2:
 *auto switch msr ia32_xss if this msr is present

 arch/x86/include/asm/kvm_host.h |  1 +
 arch/x86/include/asm/vmx.h      |  3 +++
 arch/x86/include/uapi/asm/vmx.h |  6 +++++-
 arch/x86/kvm/vmx.c              | 35 ++++++++++++++++++++++++++++++++++-
 4 files changed, 43 insertions(+), 2 deletions(-)

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 2896dbc..95dde42 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -362,6 +362,7 @@ struct kvm_vcpu_arch {
 	int mp_state;
 	u64 ia32_misc_enable_msr;
 	bool tpr_access_reporting;
+	u64 ia32_xss;
 
 	/*
 	 * Paging state of the vcpu
diff --git a/arch/x86/include/asm/vmx.h b/arch/x86/include/asm/vmx.h
index bcbfade..bdb79ef 100644
--- a/arch/x86/include/asm/vmx.h
+++ b/arch/x86/include/asm/vmx.h
@@ -69,6 +69,7 @@
 #define SECONDARY_EXEC_PAUSE_LOOP_EXITING	0x00000400
 #define SECONDARY_EXEC_ENABLE_INVPCID		0x00001000
 #define SECONDARY_EXEC_SHADOW_VMCS              0x00004000
+#define SECONDARY_EXEC_XSAVES			0x00100000
 
 
 #define PIN_BASED_EXT_INTR_MASK                 0x00000001
@@ -159,6 +160,8 @@ enum vmcs_field {
 	EOI_EXIT_BITMAP3_HIGH           = 0x00002023,
 	VMREAD_BITMAP                   = 0x00002026,
 	VMWRITE_BITMAP                  = 0x00002028,
+	XSS_EXIT_BIMTAP                 = 0x0000202C,
+	XSS_EXIT_BIMTAP_HIGH            = 0x0000202D,
 	GUEST_PHYSICAL_ADDRESS          = 0x00002400,
 	GUEST_PHYSICAL_ADDRESS_HIGH     = 0x00002401,
 	VMCS_LINK_POINTER               = 0x00002800,
diff --git a/arch/x86/include/uapi/asm/vmx.h b/arch/x86/include/uapi/asm/vmx.h
index 990a2fe..b813bf9 100644
--- a/arch/x86/include/uapi/asm/vmx.h
+++ b/arch/x86/include/uapi/asm/vmx.h
@@ -72,6 +72,8 @@
 #define EXIT_REASON_XSETBV              55
 #define EXIT_REASON_APIC_WRITE          56
 #define EXIT_REASON_INVPCID             58
+#define EXIT_REASON_XSAVES              63
+#define EXIT_REASON_XRSTORS             64
 
 #define VMX_EXIT_REASONS \
 	{ EXIT_REASON_EXCEPTION_NMI,         "EXCEPTION_NMI" }, \
@@ -116,6 +118,8 @@
 	{ EXIT_REASON_INVALID_STATE,         "INVALID_STATE" }, \
 	{ EXIT_REASON_INVD,                  "INVD" }, \
 	{ EXIT_REASON_INVVPID,               "INVVPID" }, \
-	{ EXIT_REASON_INVPCID,               "INVPCID" }
+	{ EXIT_REASON_INVPCID,               "INVPCID" }, \
+	{ EXIT_REASON_XSAVES,                "XSAVES" }, \
+	{ EXIT_REASON_XRSTORS,               "XRSTORS" }
 
 #endif /* _UAPIVMX_H */
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 6a951d8..b87b5b8 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -1045,6 +1045,12 @@ static inline bool cpu_has_vmx_invpcid(void)
 		SECONDARY_EXEC_ENABLE_INVPCID;
 }
 
+static inline bool cpu_has_xss_exit_bitmap(void)
+{
+	return vmcs_config.cpu_based_2nd_exec_ctrl &
+		SECONDARY_EXEC_XSAVES;
+}
+
 static inline bool cpu_has_virtual_nmis(void)
 {
 	return vmcs_config.pin_based_exec_ctrl & PIN_BASED_VIRTUAL_NMIS;
@@ -1773,6 +1779,14 @@ static void vmx_save_host_state(struct kvm_vcpu *vcpu)
 		kvm_set_shared_msr(vmx->guest_msrs[i].index,
 				   vmx->guest_msrs[i].data,
 				   vmx->guest_msrs[i].mask);
+
+	if (cpu_has_xsaves) {
+		u64 host_xss;
+
+		rdmsrl(MSR_IA32_XSS, host_xss);
+		add_atomic_switch_msr(vmx, MSR_IA32_XSS,
+				vcpu->arch.ia32_xss, host_xss);
+	}
 }
 
 static void __vmx_load_host_state(struct vcpu_vmx *vmx)
@@ -2895,7 +2909,8 @@ static __init int setup_vmcs_config(struct vmcs_config *vmcs_conf)
 			SECONDARY_EXEC_ENABLE_INVPCID |
 			SECONDARY_EXEC_APIC_REGISTER_VIRT |
 			SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY |
-			SECONDARY_EXEC_SHADOW_VMCS;
+			SECONDARY_EXEC_SHADOW_VMCS |
+			SECONDARY_EXEC_XSAVES;
 		if (adjust_vmx_controls(min2, opt2,
 					MSR_IA32_VMX_PROCBASED_CTLS2,
 					&_cpu_based_2nd_exec_control) < 0)
@@ -4346,6 +4361,7 @@ static int vmx_vcpu_setup(struct vcpu_vmx *vmx)
 	unsigned long a;
 #endif
 	int i;
+	u64 xss = 0;
 
 	/* I/O */
 	vmcs_write64(IO_BITMAP_A, __pa(vmx_io_bitmap_a));
@@ -4446,6 +4462,9 @@ static int vmx_vcpu_setup(struct vcpu_vmx *vmx)
 	vmcs_writel(CR0_GUEST_HOST_MASK, ~0UL);
 	set_cr4_guest_host_mask(vmx);
 
+	if (cpu_has_xss_exit_bitmap())
+		vmcs_write64(XSS_EXIT_BIMTAP, xss);
+
 	return 0;
 }
 
@@ -5334,6 +5353,18 @@ static int handle_xsetbv(struct kvm_vcpu *vcpu)
 	return 1;
 }
 
+static int handle_xsaves(struct kvm_vcpu *vcpu)
+{
+	skip_emulated_instruction(vcpu);
+	return 1;
+}
+
+static int handle_xrstors(struct kvm_vcpu *vcpu)
+{
+	skip_emulated_instruction(vcpu);
+	return 1;
+}
+
 static int handle_apic_access(struct kvm_vcpu *vcpu)
 {
 	if (likely(fasteoi)) {
@@ -6951,6 +6982,8 @@ static int (*const kvm_vmx_exit_handlers[])(struct kvm_vcpu *vcpu) = {
 	[EXIT_REASON_MONITOR_INSTRUCTION]     = handle_monitor,
 	[EXIT_REASON_INVEPT]                  = handle_invept,
 	[EXIT_REASON_INVVPID]                 = handle_invvpid,
+	[EXIT_REASON_XSAVES]                  = handle_xsaves,
+	[EXIT_REASON_XRSTORS]                 = handle_xrstors,
 };
 
 static const int kvm_vmx_max_exit_handlers =
-- 
1.9.1

^ permalink raw reply related	[flat|nested] 7+ messages in thread

* Re: [PATCH v2 2/2] kvm: vmx: enable intel xsaves for guest
  2014-12-02  8:51   ` Paolo Bonzini
@ 2014-12-02  8:38     ` Wanpeng Li
  0 siblings, 0 replies; 7+ messages in thread
From: Wanpeng Li @ 2014-12-02  8:38 UTC (permalink / raw)
  To: Paolo Bonzini; +Cc: kvm, linux-kernel

Hi Paolo,
On Tue, Dec 02, 2014 at 09:51:22AM +0100, Paolo Bonzini wrote:
>
>
>On 02/12/2014 07:14, Wanpeng Li wrote:
>> Expose intel xsaves feature to guest.
>> 
>> Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com>
>> ---
>> v1 -> v2:
>>  *auto switch msr ia32_xss if this msr is present
>> 
>>  arch/x86/include/asm/kvm_host.h |  1 +
>>  arch/x86/include/asm/vmx.h      |  3 +++
>>  arch/x86/include/uapi/asm/vmx.h |  6 +++++-
>>  arch/x86/kvm/vmx.c              | 35 ++++++++++++++++++++++++++++++++++-
>>  4 files changed, 43 insertions(+), 2 deletions(-)
>> 
>> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
>> index 2896dbc..95dde42 100644
>> --- a/arch/x86/include/asm/kvm_host.h
>> +++ b/arch/x86/include/asm/kvm_host.h
>> @@ -362,6 +362,7 @@ struct kvm_vcpu_arch {
>>  	int mp_state;
>>  	u64 ia32_misc_enable_msr;
>>  	bool tpr_access_reporting;
>> +	u64 ia32_xss;
>
>The patch is not getting/setting ia32_xss when the guest does
>RDMSR/WRMSR.  You also need a QEMU patch to migrate XSS.

Will do.

>
>>  	/*
>>  	 * Paging state of the vcpu
>> diff --git a/arch/x86/include/asm/vmx.h b/arch/x86/include/asm/vmx.h
>> index bcbfade..bdb79ef 100644
>> --- a/arch/x86/include/asm/vmx.h
>> +++ b/arch/x86/include/asm/vmx.h
>> @@ -69,6 +69,7 @@
>>  #define SECONDARY_EXEC_PAUSE_LOOP_EXITING	0x00000400
>>  #define SECONDARY_EXEC_ENABLE_INVPCID		0x00001000
>>  #define SECONDARY_EXEC_SHADOW_VMCS              0x00004000
>> +#define SECONDARY_EXEC_XSAVES			0x00100000
>>  
>>  
>>  #define PIN_BASED_EXT_INTR_MASK                 0x00000001
>> @@ -159,6 +160,8 @@ enum vmcs_field {
>>  	EOI_EXIT_BITMAP3_HIGH           = 0x00002023,
>>  	VMREAD_BITMAP                   = 0x00002026,
>>  	VMWRITE_BITMAP                  = 0x00002028,
>> +	XSS_EXIT_BIMTAP                 = 0x0000202C,
>> +	XSS_EXIT_BIMTAP_HIGH            = 0x0000202D,
>
>s/BIMTAP/BITMAP/

Ok.

>
>>  	GUEST_PHYSICAL_ADDRESS          = 0x00002400,
>>  	GUEST_PHYSICAL_ADDRESS_HIGH     = 0x00002401,
>>  	VMCS_LINK_POINTER               = 0x00002800,
>> diff --git a/arch/x86/include/uapi/asm/vmx.h b/arch/x86/include/uapi/asm/vmx.h
>> index 990a2fe..b813bf9 100644
>> --- a/arch/x86/include/uapi/asm/vmx.h
>> +++ b/arch/x86/include/uapi/asm/vmx.h
>> @@ -72,6 +72,8 @@
>>  #define EXIT_REASON_XSETBV              55
>>  #define EXIT_REASON_APIC_WRITE          56
>>  #define EXIT_REASON_INVPCID             58
>> +#define EXIT_REASON_XSAVES              63
>> +#define EXIT_REASON_XRSTORS             64
>>  
>>  #define VMX_EXIT_REASONS \
>>  	{ EXIT_REASON_EXCEPTION_NMI,         "EXCEPTION_NMI" }, \
>> @@ -116,6 +118,8 @@
>>  	{ EXIT_REASON_INVALID_STATE,         "INVALID_STATE" }, \
>>  	{ EXIT_REASON_INVD,                  "INVD" }, \
>>  	{ EXIT_REASON_INVVPID,               "INVVPID" }, \
>> -	{ EXIT_REASON_INVPCID,               "INVPCID" }
>> +	{ EXIT_REASON_INVPCID,               "INVPCID" }, \
>> +	{ EXIT_REASON_XSAVES,                "XSAVES" }, \
>> +	{ EXIT_REASON_XRSTORS,               "XRSTORS" }
>>  
>>  #endif /* _UAPIVMX_H */
>> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
>> index 6a951d8..b87b5b8 100644
>> --- a/arch/x86/kvm/vmx.c
>> +++ b/arch/x86/kvm/vmx.c
>> @@ -1045,6 +1045,12 @@ static inline bool cpu_has_vmx_invpcid(void)
>>  		SECONDARY_EXEC_ENABLE_INVPCID;
>>  }
>>  
>> +static inline bool cpu_has_xss_exit_bitmap(void)
>> +{
>> +	return vmcs_config.cpu_based_2nd_exec_ctrl &
>> +		SECONDARY_EXEC_XSAVES;
>> +}
>> +
>>  static inline bool cpu_has_virtual_nmis(void)
>>  {
>>  	return vmcs_config.pin_based_exec_ctrl & PIN_BASED_VIRTUAL_NMIS;
>> @@ -1773,6 +1779,14 @@ static void vmx_save_host_state(struct kvm_vcpu *vcpu)
>>  		kvm_set_shared_msr(vmx->guest_msrs[i].index,
>>  				   vmx->guest_msrs[i].data,
>>  				   vmx->guest_msrs[i].mask);
>> +
>> +	if (cpu_has_xsaves) {
>> +		u64 host_xss;
>> +
>> +		rdmsrl(MSR_IA32_XSS, host_xss);
>
>Is this host value fixed?  If so, please load it just once in
>setup_vmcs_config.

Will do.

>
>> +		add_atomic_switch_msr(vmx, MSR_IA32_XSS,
>> +				vcpu->arch.ia32_xss, host_xss);
>
>Also, if host_xss is fixed you can do this add_atomic_switch_msr at
>WRMSR time rather than here, and only if vcpu->arch.ia32_xss !=
>host_xss.  If the two XSS values match, do clear_atomic_switch_msr instead.

Agreed.

>
>> +	}
>>  }
>>  
>>  static void __vmx_load_host_state(struct vcpu_vmx *vmx)
>> @@ -2895,7 +2909,8 @@ static __init int setup_vmcs_config(struct vmcs_config *vmcs_conf)
>>  			SECONDARY_EXEC_ENABLE_INVPCID |
>>  			SECONDARY_EXEC_APIC_REGISTER_VIRT |
>>  			SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY |
>> -			SECONDARY_EXEC_SHADOW_VMCS;
>> +			SECONDARY_EXEC_SHADOW_VMCS |
>> +			SECONDARY_EXEC_XSAVES;
>>  		if (adjust_vmx_controls(min2, opt2,
>>  					MSR_IA32_VMX_PROCBASED_CTLS2,
>>  					&_cpu_based_2nd_exec_control) < 0)
>> @@ -4346,6 +4361,7 @@ static int vmx_vcpu_setup(struct vcpu_vmx *vmx)
>>  	unsigned long a;
>>  #endif
>>  	int i;
>> +	u64 xss = 0;
>
>#define VMX_XSS_EXIT_BITMAP	0
>

Ok.

>>  
>>  	/* I/O */
>>  	vmcs_write64(IO_BITMAP_A, __pa(vmx_io_bitmap_a));
>> @@ -4446,6 +4462,9 @@ static int vmx_vcpu_setup(struct vcpu_vmx *vmx)
>>  	vmcs_writel(CR0_GUEST_HOST_MASK, ~0UL);
>>  	set_cr4_guest_host_mask(vmx);
>>  
>> +	if (cpu_has_xss_exit_bitmap())
>> +		vmcs_write64(XSS_EXIT_BIMTAP, xss);
>> +
>>  	return 0;
>>  }
>>  
>> @@ -5334,6 +5353,18 @@ static int handle_xsetbv(struct kvm_vcpu *vcpu)
>>  	return 1;
>>  }
>>  
>> +static int handle_xsaves(struct kvm_vcpu *vcpu)
>> +{
>> +	skip_emulated_instruction(vcpu);
>> +	return 1;
>
>Please WARN(), this should never happen.

Ok.

>
>> +}
>> +
>> +static int handle_xrstors(struct kvm_vcpu *vcpu)
>> +{
>> +	skip_emulated_instruction(vcpu);
>> +	return 1;
>
>Same here.
>
>> +}
>> +
>>  static int handle_apic_access(struct kvm_vcpu *vcpu)
>>  {
>>  	if (likely(fasteoi)) {
>> @@ -6951,6 +6982,8 @@ static int (*const kvm_vmx_exit_handlers[])(struct kvm_vcpu *vcpu) = {
>>  	[EXIT_REASON_MONITOR_INSTRUCTION]     = handle_monitor,
>>  	[EXIT_REASON_INVEPT]                  = handle_invept,
>>  	[EXIT_REASON_INVVPID]                 = handle_invvpid,
>> +	[EXIT_REASON_XSAVES]                  = handle_xsaves,
>> +	[EXIT_REASON_XRSTORS]                 = handle_xrstors,
>>  };
>>  
>>  static const int kvm_vmx_max_exit_handlers =
>> 

Thanks for your review, I fix these issues in next version.

Regards,
Wanpeng Li 

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v2 1/2] kvm: x86: revert mask out xsaves
  2014-12-02  8:55 ` [PATCH v2 1/2] kvm: x86: revert mask out xsaves Paolo Bonzini
@ 2014-12-02  8:41   ` Wanpeng Li
  2014-12-03  0:50   ` Wanpeng Li
  1 sibling, 0 replies; 7+ messages in thread
From: Wanpeng Li @ 2014-12-02  8:41 UTC (permalink / raw)
  To: Paolo Bonzini; +Cc: kvm, linux-kernel

On Tue, Dec 02, 2014 at 09:55:09AM +0100, Paolo Bonzini wrote:
>
>
>On 02/12/2014 07:14, Wanpeng Li wrote:
>> xsaves will be exported to guest in the next patch, so revert the
>> mask out xsaves patch.
>> 
>> Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com>
>> ---
>>  arch/x86/kvm/cpuid.c | 10 +---------
>>  1 file changed, 1 insertion(+), 9 deletions(-)
>> 
>> diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
>> index a4f5ac4..7af07571 100644
>> --- a/arch/x86/kvm/cpuid.c
>> +++ b/arch/x86/kvm/cpuid.c
>> @@ -320,10 +320,6 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
>>  		F(ADX) | F(SMAP) | F(AVX512F) | F(AVX512PF) | F(AVX512ER) |
>>  		F(AVX512CD);
>>  
>> -	/* cpuid 0xD.1.eax */
>> -	const u32 kvm_supported_word10_x86_features =
>> -		F(XSAVEOPT) | F(XSAVEC) | F(XGETBV1);
>
>Every single bit needs to be whitelisted independently of the others, so
>a full revert is not possible.  We need to be careful about post-Skylake
>processors introducing new bits in this cpuid leaf.
>
>Also, you cannot just add F(XSAVES) here; you must only do it if XSAVES
>is actually supported.  This makes sure that nested virtualization will
>_not_ present XSAVES to the guests until specific support is introduced
>for XSAVES and XRSTORS exits (including the XSS exit bitmap).
>
>In order to do this, you have to introduce a new member in kvm_x86_ops,
>modeling what was done for MPX.  The second patch can then implement
>this new member.

Will do, thanks for your review.

Regards,
Wanpeng Li 

>
>Thanks,
>
>Paolo
>
>>  	/* all calls to cpuid_count() should be made on the same cpu */
>>  	get_cpu();
>>  
>> @@ -460,8 +456,6 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
>>  		entry->eax &= supported;
>>  		entry->edx &= supported >> 32;
>>  		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
>> -		if (!supported)
>> -			break;
>>  
>>  		for (idx = 1, i = 1; idx < 64; ++idx) {
>>  			u64 mask = ((u64)1 << idx);
>> @@ -469,9 +463,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
>>  				goto out;
>>  
>>  			do_cpuid_1_ent(&entry[i], function, idx);
>> -			if (idx == 1)
>> -				entry[i].eax &= kvm_supported_word10_x86_features;
>> -			else if (entry[i].eax == 0 || !(supported & mask))
>> +			if (entry[i].eax == 0 || !(supported & mask))
>>  				continue;
>>  			entry[i].flags |=
>>  			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
>> 

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v2 2/2] kvm: vmx: enable intel xsaves for guest
  2014-12-02  6:14 ` [PATCH v2 2/2] kvm: vmx: enable intel xsaves for guest Wanpeng Li
@ 2014-12-02  8:51   ` Paolo Bonzini
  2014-12-02  8:38     ` Wanpeng Li
  0 siblings, 1 reply; 7+ messages in thread
From: Paolo Bonzini @ 2014-12-02  8:51 UTC (permalink / raw)
  To: Wanpeng Li; +Cc: kvm, linux-kernel



On 02/12/2014 07:14, Wanpeng Li wrote:
> Expose intel xsaves feature to guest.
> 
> Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com>
> ---
> v1 -> v2:
>  *auto switch msr ia32_xss if this msr is present
> 
>  arch/x86/include/asm/kvm_host.h |  1 +
>  arch/x86/include/asm/vmx.h      |  3 +++
>  arch/x86/include/uapi/asm/vmx.h |  6 +++++-
>  arch/x86/kvm/vmx.c              | 35 ++++++++++++++++++++++++++++++++++-
>  4 files changed, 43 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> index 2896dbc..95dde42 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -362,6 +362,7 @@ struct kvm_vcpu_arch {
>  	int mp_state;
>  	u64 ia32_misc_enable_msr;
>  	bool tpr_access_reporting;
> +	u64 ia32_xss;

The patch is not getting/setting ia32_xss when the guest does
RDMSR/WRMSR.  You also need a QEMU patch to migrate XSS.

>  	/*
>  	 * Paging state of the vcpu
> diff --git a/arch/x86/include/asm/vmx.h b/arch/x86/include/asm/vmx.h
> index bcbfade..bdb79ef 100644
> --- a/arch/x86/include/asm/vmx.h
> +++ b/arch/x86/include/asm/vmx.h
> @@ -69,6 +69,7 @@
>  #define SECONDARY_EXEC_PAUSE_LOOP_EXITING	0x00000400
>  #define SECONDARY_EXEC_ENABLE_INVPCID		0x00001000
>  #define SECONDARY_EXEC_SHADOW_VMCS              0x00004000
> +#define SECONDARY_EXEC_XSAVES			0x00100000
>  
>  
>  #define PIN_BASED_EXT_INTR_MASK                 0x00000001
> @@ -159,6 +160,8 @@ enum vmcs_field {
>  	EOI_EXIT_BITMAP3_HIGH           = 0x00002023,
>  	VMREAD_BITMAP                   = 0x00002026,
>  	VMWRITE_BITMAP                  = 0x00002028,
> +	XSS_EXIT_BIMTAP                 = 0x0000202C,
> +	XSS_EXIT_BIMTAP_HIGH            = 0x0000202D,

s/BIMTAP/BITMAP/

>  	GUEST_PHYSICAL_ADDRESS          = 0x00002400,
>  	GUEST_PHYSICAL_ADDRESS_HIGH     = 0x00002401,
>  	VMCS_LINK_POINTER               = 0x00002800,
> diff --git a/arch/x86/include/uapi/asm/vmx.h b/arch/x86/include/uapi/asm/vmx.h
> index 990a2fe..b813bf9 100644
> --- a/arch/x86/include/uapi/asm/vmx.h
> +++ b/arch/x86/include/uapi/asm/vmx.h
> @@ -72,6 +72,8 @@
>  #define EXIT_REASON_XSETBV              55
>  #define EXIT_REASON_APIC_WRITE          56
>  #define EXIT_REASON_INVPCID             58
> +#define EXIT_REASON_XSAVES              63
> +#define EXIT_REASON_XRSTORS             64
>  
>  #define VMX_EXIT_REASONS \
>  	{ EXIT_REASON_EXCEPTION_NMI,         "EXCEPTION_NMI" }, \
> @@ -116,6 +118,8 @@
>  	{ EXIT_REASON_INVALID_STATE,         "INVALID_STATE" }, \
>  	{ EXIT_REASON_INVD,                  "INVD" }, \
>  	{ EXIT_REASON_INVVPID,               "INVVPID" }, \
> -	{ EXIT_REASON_INVPCID,               "INVPCID" }
> +	{ EXIT_REASON_INVPCID,               "INVPCID" }, \
> +	{ EXIT_REASON_XSAVES,                "XSAVES" }, \
> +	{ EXIT_REASON_XRSTORS,               "XRSTORS" }
>  
>  #endif /* _UAPIVMX_H */
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index 6a951d8..b87b5b8 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -1045,6 +1045,12 @@ static inline bool cpu_has_vmx_invpcid(void)
>  		SECONDARY_EXEC_ENABLE_INVPCID;
>  }
>  
> +static inline bool cpu_has_xss_exit_bitmap(void)
> +{
> +	return vmcs_config.cpu_based_2nd_exec_ctrl &
> +		SECONDARY_EXEC_XSAVES;
> +}
> +
>  static inline bool cpu_has_virtual_nmis(void)
>  {
>  	return vmcs_config.pin_based_exec_ctrl & PIN_BASED_VIRTUAL_NMIS;
> @@ -1773,6 +1779,14 @@ static void vmx_save_host_state(struct kvm_vcpu *vcpu)
>  		kvm_set_shared_msr(vmx->guest_msrs[i].index,
>  				   vmx->guest_msrs[i].data,
>  				   vmx->guest_msrs[i].mask);
> +
> +	if (cpu_has_xsaves) {
> +		u64 host_xss;
> +
> +		rdmsrl(MSR_IA32_XSS, host_xss);

Is this host value fixed?  If so, please load it just once in
setup_vmcs_config.

> +		add_atomic_switch_msr(vmx, MSR_IA32_XSS,
> +				vcpu->arch.ia32_xss, host_xss);

Also, if host_xss is fixed you can do this add_atomic_switch_msr at
WRMSR time rather than here, and only if vcpu->arch.ia32_xss !=
host_xss.  If the two XSS values match, do clear_atomic_switch_msr instead.

> +	}
>  }
>  
>  static void __vmx_load_host_state(struct vcpu_vmx *vmx)
> @@ -2895,7 +2909,8 @@ static __init int setup_vmcs_config(struct vmcs_config *vmcs_conf)
>  			SECONDARY_EXEC_ENABLE_INVPCID |
>  			SECONDARY_EXEC_APIC_REGISTER_VIRT |
>  			SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY |
> -			SECONDARY_EXEC_SHADOW_VMCS;
> +			SECONDARY_EXEC_SHADOW_VMCS |
> +			SECONDARY_EXEC_XSAVES;
>  		if (adjust_vmx_controls(min2, opt2,
>  					MSR_IA32_VMX_PROCBASED_CTLS2,
>  					&_cpu_based_2nd_exec_control) < 0)
> @@ -4346,6 +4361,7 @@ static int vmx_vcpu_setup(struct vcpu_vmx *vmx)
>  	unsigned long a;
>  #endif
>  	int i;
> +	u64 xss = 0;

#define VMX_XSS_EXIT_BITMAP	0

>  
>  	/* I/O */
>  	vmcs_write64(IO_BITMAP_A, __pa(vmx_io_bitmap_a));
> @@ -4446,6 +4462,9 @@ static int vmx_vcpu_setup(struct vcpu_vmx *vmx)
>  	vmcs_writel(CR0_GUEST_HOST_MASK, ~0UL);
>  	set_cr4_guest_host_mask(vmx);
>  
> +	if (cpu_has_xss_exit_bitmap())
> +		vmcs_write64(XSS_EXIT_BIMTAP, xss);
> +
>  	return 0;
>  }
>  
> @@ -5334,6 +5353,18 @@ static int handle_xsetbv(struct kvm_vcpu *vcpu)
>  	return 1;
>  }
>  
> +static int handle_xsaves(struct kvm_vcpu *vcpu)
> +{
> +	skip_emulated_instruction(vcpu);
> +	return 1;

Please WARN(), this should never happen.

> +}
> +
> +static int handle_xrstors(struct kvm_vcpu *vcpu)
> +{
> +	skip_emulated_instruction(vcpu);
> +	return 1;

Same here.

> +}
> +
>  static int handle_apic_access(struct kvm_vcpu *vcpu)
>  {
>  	if (likely(fasteoi)) {
> @@ -6951,6 +6982,8 @@ static int (*const kvm_vmx_exit_handlers[])(struct kvm_vcpu *vcpu) = {
>  	[EXIT_REASON_MONITOR_INSTRUCTION]     = handle_monitor,
>  	[EXIT_REASON_INVEPT]                  = handle_invept,
>  	[EXIT_REASON_INVVPID]                 = handle_invvpid,
> +	[EXIT_REASON_XSAVES]                  = handle_xsaves,
> +	[EXIT_REASON_XRSTORS]                 = handle_xrstors,
>  };
>  
>  static const int kvm_vmx_max_exit_handlers =
> 

Thanks,

Paolo

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v2 1/2] kvm: x86: revert mask out xsaves
  2014-12-02  6:14 [PATCH v2 1/2] kvm: x86: revert mask out xsaves Wanpeng Li
  2014-12-02  6:14 ` [PATCH v2 2/2] kvm: vmx: enable intel xsaves for guest Wanpeng Li
@ 2014-12-02  8:55 ` Paolo Bonzini
  2014-12-02  8:41   ` Wanpeng Li
  2014-12-03  0:50   ` Wanpeng Li
  1 sibling, 2 replies; 7+ messages in thread
From: Paolo Bonzini @ 2014-12-02  8:55 UTC (permalink / raw)
  To: Wanpeng Li; +Cc: kvm, linux-kernel



On 02/12/2014 07:14, Wanpeng Li wrote:
> xsaves will be exported to guest in the next patch, so revert the
> mask out xsaves patch.
> 
> Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com>
> ---
>  arch/x86/kvm/cpuid.c | 10 +---------
>  1 file changed, 1 insertion(+), 9 deletions(-)
> 
> diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
> index a4f5ac4..7af07571 100644
> --- a/arch/x86/kvm/cpuid.c
> +++ b/arch/x86/kvm/cpuid.c
> @@ -320,10 +320,6 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
>  		F(ADX) | F(SMAP) | F(AVX512F) | F(AVX512PF) | F(AVX512ER) |
>  		F(AVX512CD);
>  
> -	/* cpuid 0xD.1.eax */
> -	const u32 kvm_supported_word10_x86_features =
> -		F(XSAVEOPT) | F(XSAVEC) | F(XGETBV1);

Every single bit needs to be whitelisted independently of the others, so
a full revert is not possible.  We need to be careful about post-Skylake
processors introducing new bits in this cpuid leaf.

Also, you cannot just add F(XSAVES) here; you must only do it if XSAVES
is actually supported.  This makes sure that nested virtualization will
_not_ present XSAVES to the guests until specific support is introduced
for XSAVES and XRSTORS exits (including the XSS exit bitmap).

In order to do this, you have to introduce a new member in kvm_x86_ops,
modeling what was done for MPX.  The second patch can then implement
this new member.

Thanks,

Paolo

>  	/* all calls to cpuid_count() should be made on the same cpu */
>  	get_cpu();
>  
> @@ -460,8 +456,6 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
>  		entry->eax &= supported;
>  		entry->edx &= supported >> 32;
>  		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
> -		if (!supported)
> -			break;
>  
>  		for (idx = 1, i = 1; idx < 64; ++idx) {
>  			u64 mask = ((u64)1 << idx);
> @@ -469,9 +463,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
>  				goto out;
>  
>  			do_cpuid_1_ent(&entry[i], function, idx);
> -			if (idx == 1)
> -				entry[i].eax &= kvm_supported_word10_x86_features;
> -			else if (entry[i].eax == 0 || !(supported & mask))
> +			if (entry[i].eax == 0 || !(supported & mask))
>  				continue;
>  			entry[i].flags |=
>  			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
> 

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v2 1/2] kvm: x86: revert mask out xsaves
  2014-12-02  8:55 ` [PATCH v2 1/2] kvm: x86: revert mask out xsaves Paolo Bonzini
  2014-12-02  8:41   ` Wanpeng Li
@ 2014-12-03  0:50   ` Wanpeng Li
  1 sibling, 0 replies; 7+ messages in thread
From: Wanpeng Li @ 2014-12-03  0:50 UTC (permalink / raw)
  To: Paolo Bonzini; +Cc: kvm, linux-kernel

On Tue, Dec 02, 2014 at 09:55:09AM +0100, Paolo Bonzini wrote:
>
>
>On 02/12/2014 07:14, Wanpeng Li wrote:
>> xsaves will be exported to guest in the next patch, so revert the
>> mask out xsaves patch.
>> 
>> Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com>
>> ---
>>  arch/x86/kvm/cpuid.c | 10 +---------
>>  1 file changed, 1 insertion(+), 9 deletions(-)
>> 
>> diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
>> index a4f5ac4..7af07571 100644
>> --- a/arch/x86/kvm/cpuid.c
>> +++ b/arch/x86/kvm/cpuid.c
>> @@ -320,10 +320,6 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
>>  		F(ADX) | F(SMAP) | F(AVX512F) | F(AVX512PF) | F(AVX512ER) |
>>  		F(AVX512CD);
>>  
>> -	/* cpuid 0xD.1.eax */
>> -	const u32 kvm_supported_word10_x86_features =
>> -		F(XSAVEOPT) | F(XSAVEC) | F(XGETBV1);
>
>Every single bit needs to be whitelisted independently of the others, so
>a full revert is not possible.  We need to be careful about post-Skylake
>processors introducing new bits in this cpuid leaf.
>
>Also, you cannot just add F(XSAVES) here; you must only do it if XSAVES
>is actually supported.  This makes sure that nested virtualization will
>_not_ present XSAVES to the guests until specific support is introduced
>for XSAVES and XRSTORS exits (including the XSS exit bitmap).

I will try to enable nested xsaves.

Regards,
Wanpeng Li 

>
>In order to do this, you have to introduce a new member in kvm_x86_ops,
>modeling what was done for MPX.  The second patch can then implement
>this new member.
>
>Thanks,
>
>Paolo
>
>>  	/* all calls to cpuid_count() should be made on the same cpu */
>>  	get_cpu();
>>  
>> @@ -460,8 +456,6 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
>>  		entry->eax &= supported;
>>  		entry->edx &= supported >> 32;
>>  		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
>> -		if (!supported)
>> -			break;
>>  
>>  		for (idx = 1, i = 1; idx < 64; ++idx) {
>>  			u64 mask = ((u64)1 << idx);
>> @@ -469,9 +463,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
>>  				goto out;
>>  
>>  			do_cpuid_1_ent(&entry[i], function, idx);
>> -			if (idx == 1)
>> -				entry[i].eax &= kvm_supported_word10_x86_features;
>> -			else if (entry[i].eax == 0 || !(supported & mask))
>> +			if (entry[i].eax == 0 || !(supported & mask))
>>  				continue;
>>  			entry[i].flags |=
>>  			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
>> 

^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2014-12-03  0:50 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-12-02  6:14 [PATCH v2 1/2] kvm: x86: revert mask out xsaves Wanpeng Li
2014-12-02  6:14 ` [PATCH v2 2/2] kvm: vmx: enable intel xsaves for guest Wanpeng Li
2014-12-02  8:51   ` Paolo Bonzini
2014-12-02  8:38     ` Wanpeng Li
2014-12-02  8:55 ` [PATCH v2 1/2] kvm: x86: revert mask out xsaves Paolo Bonzini
2014-12-02  8:41   ` Wanpeng Li
2014-12-03  0:50   ` Wanpeng Li

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).