[PATCH] x86: guest: rely on leaf 0x40000001 to detect Hyper-V

[PATCH] x86: guest: rely on leaf 0x40000001 to detect Hyper-V

[Paolo Bonzini]

The specification says that "Microsoft Hv" is actually a vendor ID field
that is only used for reporting and diagnostic purposes.  The actual
field that you need to check is the interface ID that you get in eax
when querying the HYPERV_CPUID_INTERFACE.

Change ms_hyperv_platform to actually do what the specification suggests.
This roughy matches what Windows looks for, though Windows actually
ignores HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS completely.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 arch/x86/kernel/cpu/mshyperv.c | 14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)

diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
index 381c8b9b3a33..7910e7fd705b 100644
--- a/arch/x86/kernel/cpu/mshyperv.c
+++ b/arch/x86/kernel/cpu/mshyperv.c
@@ -116,18 +116,16 @@ static void hv_machine_crash_shutdown(struct pt_regs *regs)
 static uint32_t  __init ms_hyperv_platform(void)
 {
 	u32 eax;
-	u32 hyp_signature[3];
 
 	if (!boot_cpu_has(X86_FEATURE_HYPERVISOR))
 		return 0;
 
-	cpuid(HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS,
-	      &eax, &hyp_signature[0], &hyp_signature[1], &hyp_signature[2]);
-
-	if (eax >= HYPERV_CPUID_MIN &&
-	    eax <= HYPERV_CPUID_MAX &&
-	    !memcmp("Microsoft Hv", hyp_signature, 12))
-		return HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
+	eax = cpuid_eax(HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS);
+	if (eax >= HYPERV_CPUID_MIN && eax <= HYPERV_CPUID_MAX) {
+		eax = cpuid_eax(HYPERV_CPUID_INTERFACE);
+		if (!memcmp(&eax, "Hv#1", sizeof(eax)))
+			return HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
+	}
 
 	return 0;
 }
-- 
2.5.0

Re: [PATCH] x86: guest: rely on leaf 0x40000001 to detect Hyper-V

[Vitaly Kuznetsov]

Paolo Bonzini <pbonzini@redhat.com> writes:

> The specification says that "Microsoft Hv" is actually a vendor ID field
> that is only used for reporting and diagnostic purposes.  The actual
> field that you need to check is the interface ID that you get in eax
> when querying the HYPERV_CPUID_INTERFACE.
>
> Change ms_hyperv_platform to actually do what the specification suggests.
> This roughy matches what Windows looks for, though Windows actually
> ignores HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS completely.
>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>

(and it seems K. Y. is missing on the CC: list, fixed).

> ---
>  arch/x86/kernel/cpu/mshyperv.c | 14 ++++++--------
>  1 file changed, 6 insertions(+), 8 deletions(-)
>
> diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
> index 381c8b9b3a33..7910e7fd705b 100644
> --- a/arch/x86/kernel/cpu/mshyperv.c
> +++ b/arch/x86/kernel/cpu/mshyperv.c
> @@ -116,18 +116,16 @@ static void hv_machine_crash_shutdown(struct pt_regs *regs)
>  static uint32_t  __init ms_hyperv_platform(void)
>  {
>  	u32 eax;
> -	u32 hyp_signature[3];
>
>  	if (!boot_cpu_has(X86_FEATURE_HYPERVISOR))
>  		return 0;
>
> -	cpuid(HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS,
> -	      &eax, &hyp_signature[0], &hyp_signature[1], &hyp_signature[2]);
> -
> -	if (eax >= HYPERV_CPUID_MIN &&
> -	    eax <= HYPERV_CPUID_MAX &&
> -	    !memcmp("Microsoft Hv", hyp_signature, 12))
> -		return HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
> +	eax = cpuid_eax(HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS);
> +	if (eax >= HYPERV_CPUID_MIN && eax <= HYPERV_CPUID_MAX) {
> +		eax = cpuid_eax(HYPERV_CPUID_INTERFACE);
> +		if (!memcmp(&eax, "Hv#1", sizeof(eax)))
> +			return HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
> +	}
>
>  	return 0;
>  }

-- 
  Vitaly

RE: [PATCH] x86: guest: rely on leaf 0x40000001 to detect Hyper-V

[KY Srinivasan]

> -----Original Message-----
> From: Paolo Bonzini [mailto:paolo.bonzini@gmail.com] On Behalf Of Paolo
> Bonzini
> Sent: Friday, October 2, 2015 2:27 AM
> To: linux-kernel@vger.kernel.org
> Cc: KY Srinivasan <kys@microsoft.com>; Haiyang Zhang
> <haiyangz@microsoft.com>; x86@kernel.org; devel@linuxdriverproject.org;
> alex.williamson@redhat.com
> Subject: [PATCH] x86: guest: rely on leaf 0x40000001 to detect Hyper-V
> 
> The specification says that "Microsoft Hv" is actually a vendor ID field
> that is only used for reporting and diagnostic purposes.  The actual
> field that you need to check is the interface ID that you get in eax
> when querying the HYPERV_CPUID_INTERFACE.
> 
> Change ms_hyperv_platform to actually do what the specification suggests.
> This roughy matches what Windows looks for, though Windows actually
> ignores HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS completely.
> 
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Thanks Paolo.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
> ---
>  arch/x86/kernel/cpu/mshyperv.c | 14 ++++++--------
>  1 file changed, 6 insertions(+), 8 deletions(-)
> 
> diff --git a/arch/x86/kernel/cpu/mshyperv.c
> b/arch/x86/kernel/cpu/mshyperv.c
> index 381c8b9b3a33..7910e7fd705b 100644
> --- a/arch/x86/kernel/cpu/mshyperv.c
> +++ b/arch/x86/kernel/cpu/mshyperv.c
> @@ -116,18 +116,16 @@ static void hv_machine_crash_shutdown(struct
> pt_regs *regs)
>  static uint32_t  __init ms_hyperv_platform(void)
>  {
>  	u32 eax;
> -	u32 hyp_signature[3];
> 
>  	if (!boot_cpu_has(X86_FEATURE_HYPERVISOR))
>  		return 0;
> 
> -	cpuid(HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS,
> -	      &eax, &hyp_signature[0], &hyp_signature[1],
> &hyp_signature[2]);
> -
> -	if (eax >= HYPERV_CPUID_MIN &&
> -	    eax <= HYPERV_CPUID_MAX &&
> -	    !memcmp("Microsoft Hv", hyp_signature, 12))
> -		return HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
> +	eax =
> cpuid_eax(HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS);
> +	if (eax >= HYPERV_CPUID_MIN && eax <= HYPERV_CPUID_MAX) {
> +		eax = cpuid_eax(HYPERV_CPUID_INTERFACE);
> +		if (!memcmp(&eax, "Hv#1", sizeof(eax)))
> +			return
> HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS;
> +	}
> 
>  	return 0;
>  }
> --
> 2.5.0

RE: [PATCH] x86: guest: rely on leaf 0x40000001 to detect Hyper-V

[Thomas Gleixner]

On Fri, 2 Oct 2015, KY Srinivasan wrote:
> > Change ms_hyperv_platform to actually do what the specification suggests.
> > This roughy matches what Windows looks for, though Windows actually
> > ignores HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS completely.
> > 
> > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> 
> Thanks Paolo.
> 
> Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>

Does that mean Acked-by or Reviewed-by? SOB certainly does not apply
here.

Thanks,

	tglx

RE: [PATCH] x86: guest: rely on leaf 0x40000001 to detect Hyper-V

[KY Srinivasan]

> -----Original Message-----
> From: Thomas Gleixner [mailto:tglx@linutronix.de]
> Sent: Friday, October 2, 2015 1:07 PM
> To: KY Srinivasan <kys@microsoft.com>
> Cc: Paolo Bonzini <pbonzini@redhat.com>; linux-kernel@vger.kernel.org;
> Haiyang Zhang <haiyangz@microsoft.com>; x86@kernel.org;
> devel@linuxdriverproject.org; alex.williamson@redhat.com
> Subject: RE: [PATCH] x86: guest: rely on leaf 0x40000001 to detect Hyper-V
> 
> On Fri, 2 Oct 2015, KY Srinivasan wrote:
> > > Change ms_hyperv_platform to actually do what the specification
> suggests.
> > > This roughy matches what Windows looks for, though Windows actually
> > > ignores HYPERV_CPUID_VENDOR_AND_MAX_FUNCTIONS completely.
> > >
> > > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> >
> > Thanks Paolo.
> >
> > Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
> 
> Does that mean Acked-by or Reviewed-by? SOB certainly does not apply
> here.
Thomas,

I have reviewed the patch and am acking this change.

Regards,

K. Y
> 
> Thanks,
> 
> 	tglx