Re: [PATCH AUTOSEL for 4.14 12/67] KVM: x86: add support for emulating UMIP

2018-03-19 Thread Sasha Levin 
On Thu, Mar 08, 2018 at 01:46:39AM -0500, Paolo Bonzini wrote:
>
>
>- Original Message -
>> From: "Sasha Levin" 
>> To: linux-kernel@vger.kernel.org, sta...@vger.kernel.org
>> Cc: "Paolo Bonzini" , "Sasha Levin" 
>> 
>> Sent: Thursday, March 8, 2018 5:57:36 AM
>> Subject: [PATCH AUTOSEL for 4.14 12/67] KVM: x86: add support for emulating 
>> UMIP
>>
>> From: Paolo Bonzini 
>>
>> [ Upstream commit 66336cab3531d3325ebde36a047250c42cb5 ]
>>
>> The User-Mode Instruction Prevention feature present in recent Intel
>> processor prevents a group of instructions (sgdt, sidt, sldt, smsw, and
>> str) from being executed with CPL > 0. Otherwise, a general protection
>> fault is issued.
>>
>> UMIP instructions in general are also able to trigger vmexits, so we can
>> actually emulate UMIP on older processors.  This commit sets up the
>> infrastructure so that kvm-intel.ko and kvm-amd.ko can set the UMIP
>> feature bit for CPUID even if the feature is not actually available
>> in hardware.
>>
>> Reviewed-by: Wanpeng Li 
>> Signed-off-by: Paolo Bonzini 
>> Signed-off-by: Sasha Levin 
>
>This is not enough (there were a couple bugs, plus you have not
>added the changes needed to the handle SLDT, STR, etc. in the
>emulator).  But since this is a feature, why is it being backported
>to 4.14, especially without any testing??

Hi Paolo,

They were proposed for stable as part of the automatic patch selection
process. I'll drop this patch.

-- 

Thanks,
Sasha

Re: [PATCH AUTOSEL for 4.14 12/67] KVM: x86: add support for emulating UMIP

2018-03-19 Thread Sasha Levin 
On Thu, Mar 08, 2018 at 01:46:39AM -0500, Paolo Bonzini wrote:
>
>
>- Original Message -
>> From: "Sasha Levin" 
>> To: linux-kernel@vger.kernel.org, sta...@vger.kernel.org
>> Cc: "Paolo Bonzini" , "Sasha Levin" 
>> 
>> Sent: Thursday, March 8, 2018 5:57:36 AM
>> Subject: [PATCH AUTOSEL for 4.14 12/67] KVM: x86: add support for emulating 
>> UMIP
>>
>> From: Paolo Bonzini 
>>
>> [ Upstream commit 66336cab3531d3325ebde36a047250c42cb5 ]
>>
>> The User-Mode Instruction Prevention feature present in recent Intel
>> processor prevents a group of instructions (sgdt, sidt, sldt, smsw, and
>> str) from being executed with CPL > 0. Otherwise, a general protection
>> fault is issued.
>>
>> UMIP instructions in general are also able to trigger vmexits, so we can
>> actually emulate UMIP on older processors.  This commit sets up the
>> infrastructure so that kvm-intel.ko and kvm-amd.ko can set the UMIP
>> feature bit for CPUID even if the feature is not actually available
>> in hardware.
>>
>> Reviewed-by: Wanpeng Li 
>> Signed-off-by: Paolo Bonzini 
>> Signed-off-by: Sasha Levin 
>
>This is not enough (there were a couple bugs, plus you have not
>added the changes needed to the handle SLDT, STR, etc. in the
>emulator).  But since this is a feature, why is it being backported
>to 4.14, especially without any testing??

Hi Paolo,

They were proposed for stable as part of the automatic patch selection
process. I'll drop this patch.

-- 

Thanks,
Sasha

Re: [PATCH AUTOSEL for 4.14 12/67] KVM: x86: add support for emulating UMIP

2018-03-07 Thread Paolo Bonzini 


- Original Message -
> From: "Sasha Levin" 
> To: linux-kernel@vger.kernel.org, sta...@vger.kernel.org
> Cc: "Paolo Bonzini" , "Sasha Levin" 
> 
> Sent: Thursday, March 8, 2018 5:57:36 AM
> Subject: [PATCH AUTOSEL for 4.14 12/67] KVM: x86: add support for emulating 
> UMIP
> 
> From: Paolo Bonzini 
> 
> [ Upstream commit 66336cab3531d3325ebde36a047250c42cb5 ]
> 
> The User-Mode Instruction Prevention feature present in recent Intel
> processor prevents a group of instructions (sgdt, sidt, sldt, smsw, and
> str) from being executed with CPL > 0. Otherwise, a general protection
> fault is issued.
> 
> UMIP instructions in general are also able to trigger vmexits, so we can
> actually emulate UMIP on older processors.  This commit sets up the
> infrastructure so that kvm-intel.ko and kvm-amd.ko can set the UMIP
> feature bit for CPUID even if the feature is not actually available
> in hardware.
> 
> Reviewed-by: Wanpeng Li 
> Signed-off-by: Paolo Bonzini 
> Signed-off-by: Sasha Levin 

This is not enough (there were a couple bugs, plus you have not
added the changes needed to the handle SLDT, STR, etc. in the
emulator).  But since this is a feature, why is it being backported
to 4.14, especially without any testing??

Thanks,

Paolo

> ---
>  arch/x86/include/asm/kvm_host.h | 1 +
>  arch/x86/kvm/cpuid.c| 2 ++
>  arch/x86/kvm/svm.c  | 6 ++
>  arch/x86/kvm/vmx.c  | 6 ++
>  4 files changed, 15 insertions(+)
> 
> diff --git a/arch/x86/include/asm/kvm_host.h
> b/arch/x86/include/asm/kvm_host.h
> index 4f8b80199672..52ecf9b2f61e 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -1004,6 +1004,7 @@ struct kvm_x86_ops {
>   void (*handle_external_intr)(struct kvm_vcpu *vcpu);
>   bool (*mpx_supported)(void);
>   bool (*xsaves_supported)(void);
> + bool (*umip_emulated)(void);
>  
>   int (*check_nested_events)(struct kvm_vcpu *vcpu, bool external_intr);
>  
> diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
> index 13f5d4217e4f..f3fc225f5ebb 100644
> --- a/arch/x86/kvm/cpuid.c
> +++ b/arch/x86/kvm/cpuid.c
> @@ -325,6 +325,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2
> *entry, u32 function,
>   unsigned f_invpcid = kvm_x86_ops->invpcid_supported() ? F(INVPCID) : 0;
>   unsigned f_mpx = kvm_mpx_supported() ? F(MPX) : 0;
>   unsigned f_xsaves = kvm_x86_ops->xsaves_supported() ? F(XSAVES) : 0;
> + unsigned f_umip = kvm_x86_ops->umip_emulated() ? F(UMIP) : 0;
>  
>   /* cpuid 1.edx */
>   const u32 kvm_cpuid_1_edx_x86_features =
> @@ -476,6 +477,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2
> *entry, u32 function,
>   entry->ebx |= F(TSC_ADJUST);
>   entry->ecx &= kvm_cpuid_7_0_ecx_x86_features;
>   cpuid_mask(>ecx, CPUID_7_ECX);
> + entry->ecx |= f_umip;
>   /* PKU is not yet implemented for shadow paging. */
>   if (!tdp_enabled || !boot_cpu_has(X86_FEATURE_OSPKE))
>   entry->ecx &= ~F(PKU);
> diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
> index e0bc3ad0f6cd..8ea19bf09202 100644
> --- a/arch/x86/kvm/svm.c
> +++ b/arch/x86/kvm/svm.c
> @@ -5322,6 +5322,11 @@ static bool svm_xsaves_supported(void)
>   return false;
>  }
>  
> +static bool svm_umip_emulated(void)
> +{
> + return false;
> +}
> +
>  static bool svm_has_wbinvd_exit(void)
>  {
>   return true;
> @@ -5633,6 +5638,7 @@ static struct kvm_x86_ops svm_x86_ops __ro_after_init =
> {
>   .invpcid_supported = svm_invpcid_supported,
>   .mpx_supported = svm_mpx_supported,
>   .xsaves_supported = svm_xsaves_supported,
> + .umip_emulated = svm_umip_emulated,
>  
>   .set_supported_cpuid = svm_set_supported_cpuid,
>  
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index 5ffde16253cb..924d88d5ca35 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -9212,6 +9212,11 @@ static bool vmx_xsaves_supported(void)
>   SECONDARY_EXEC_XSAVES;
>  }
>  
> +static bool vmx_umip_emulated(void)
> +{
> + return false;
> +}
> +
>  static void vmx_recover_nmi_blocking(struct vcpu_vmx *vmx)
>  {
>   u32 exit_intr_info;
> @@ -12252,6 +12257,7 @@ static struct kvm_x86_ops vmx_x86_ops __ro_after_init
> = {
>   .handle_external_intr = vmx_handle_external_intr,
>   .mpx_supported = vmx_mpx_supported,
>   .xsaves_supported = vmx_xsaves_supported,
> + .umip_emulated = vmx_umip_emulated,
>  
>   .check_nested_events = vmx_check_nested_events,
>  
> --
> 2.14.1
>

Re: [PATCH AUTOSEL for 4.14 12/67] KVM: x86: add support for emulating UMIP

2018-03-07 Thread Paolo Bonzini 


- Original Message -
> From: "Sasha Levin" 
> To: linux-kernel@vger.kernel.org, sta...@vger.kernel.org
> Cc: "Paolo Bonzini" , "Sasha Levin" 
> 
> Sent: Thursday, March 8, 2018 5:57:36 AM
> Subject: [PATCH AUTOSEL for 4.14 12/67] KVM: x86: add support for emulating 
> UMIP
> 
> From: Paolo Bonzini 
> 
> [ Upstream commit 66336cab3531d3325ebde36a047250c42cb5 ]
> 
> The User-Mode Instruction Prevention feature present in recent Intel
> processor prevents a group of instructions (sgdt, sidt, sldt, smsw, and
> str) from being executed with CPL > 0. Otherwise, a general protection
> fault is issued.
> 
> UMIP instructions in general are also able to trigger vmexits, so we can
> actually emulate UMIP on older processors.  This commit sets up the
> infrastructure so that kvm-intel.ko and kvm-amd.ko can set the UMIP
> feature bit for CPUID even if the feature is not actually available
> in hardware.
> 
> Reviewed-by: Wanpeng Li 
> Signed-off-by: Paolo Bonzini 
> Signed-off-by: Sasha Levin 

This is not enough (there were a couple bugs, plus you have not
added the changes needed to the handle SLDT, STR, etc. in the
emulator).  But since this is a feature, why is it being backported
to 4.14, especially without any testing??

Thanks,

Paolo

> ---
>  arch/x86/include/asm/kvm_host.h | 1 +
>  arch/x86/kvm/cpuid.c| 2 ++
>  arch/x86/kvm/svm.c  | 6 ++
>  arch/x86/kvm/vmx.c  | 6 ++
>  4 files changed, 15 insertions(+)
> 
> diff --git a/arch/x86/include/asm/kvm_host.h
> b/arch/x86/include/asm/kvm_host.h
> index 4f8b80199672..52ecf9b2f61e 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -1004,6 +1004,7 @@ struct kvm_x86_ops {
>   void (*handle_external_intr)(struct kvm_vcpu *vcpu);
>   bool (*mpx_supported)(void);
>   bool (*xsaves_supported)(void);
> + bool (*umip_emulated)(void);
>  
>   int (*check_nested_events)(struct kvm_vcpu *vcpu, bool external_intr);
>  
> diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
> index 13f5d4217e4f..f3fc225f5ebb 100644
> --- a/arch/x86/kvm/cpuid.c
> +++ b/arch/x86/kvm/cpuid.c
> @@ -325,6 +325,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2
> *entry, u32 function,
>   unsigned f_invpcid = kvm_x86_ops->invpcid_supported() ? F(INVPCID) : 0;
>   unsigned f_mpx = kvm_mpx_supported() ? F(MPX) : 0;
>   unsigned f_xsaves = kvm_x86_ops->xsaves_supported() ? F(XSAVES) : 0;
> + unsigned f_umip = kvm_x86_ops->umip_emulated() ? F(UMIP) : 0;
>  
>   /* cpuid 1.edx */
>   const u32 kvm_cpuid_1_edx_x86_features =
> @@ -476,6 +477,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2
> *entry, u32 function,
>   entry->ebx |= F(TSC_ADJUST);
>   entry->ecx &= kvm_cpuid_7_0_ecx_x86_features;
>   cpuid_mask(>ecx, CPUID_7_ECX);
> + entry->ecx |= f_umip;
>   /* PKU is not yet implemented for shadow paging. */
>   if (!tdp_enabled || !boot_cpu_has(X86_FEATURE_OSPKE))
>   entry->ecx &= ~F(PKU);
> diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
> index e0bc3ad0f6cd..8ea19bf09202 100644
> --- a/arch/x86/kvm/svm.c
> +++ b/arch/x86/kvm/svm.c
> @@ -5322,6 +5322,11 @@ static bool svm_xsaves_supported(void)
>   return false;
>  }
>  
> +static bool svm_umip_emulated(void)
> +{
> + return false;
> +}
> +
>  static bool svm_has_wbinvd_exit(void)
>  {
>   return true;
> @@ -5633,6 +5638,7 @@ static struct kvm_x86_ops svm_x86_ops __ro_after_init =
> {
>   .invpcid_supported = svm_invpcid_supported,
>   .mpx_supported = svm_mpx_supported,
>   .xsaves_supported = svm_xsaves_supported,
> + .umip_emulated = svm_umip_emulated,
>  
>   .set_supported_cpuid = svm_set_supported_cpuid,
>  
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index 5ffde16253cb..924d88d5ca35 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -9212,6 +9212,11 @@ static bool vmx_xsaves_supported(void)
>   SECONDARY_EXEC_XSAVES;
>  }
>  
> +static bool vmx_umip_emulated(void)
> +{
> + return false;
> +}
> +
>  static void vmx_recover_nmi_blocking(struct vcpu_vmx *vmx)
>  {
>   u32 exit_intr_info;
> @@ -12252,6 +12257,7 @@ static struct kvm_x86_ops vmx_x86_ops __ro_after_init
> = {
>   .handle_external_intr = vmx_handle_external_intr,
>   .mpx_supported = vmx_mpx_supported,
>   .xsaves_supported = vmx_xsaves_supported,
> + .umip_emulated = vmx_umip_emulated,
>  
>   .check_nested_events = vmx_check_nested_events,
>  
> --
> 2.14.1
>