In x86 virtualization environments, including TDX, RDTSC instruction is
handled without causing a VM exit, resulting in minimal overhead and
jitters. On the other hand, other clock sources (such as HPET, ACPI
timer, APIC, etc.) necessitate VM exits to implement, resulting in more
fluctuating measurements compared to TSC. Thus, those clock sources are
not effective for calibrating TSC.

In TD guests, TSC is virtualized by the TDX module, which ensures:

  - Virtual TSC values are consistent among all the TD’s VCPUs;
  - Monotonously incrementing for any single VCPU;
  - The frequency is determined by TD configuration. The host TSC is
    invariant on platforms where TDX is available.

Reliable TSC is architectural guarantee for the TDX platform and it must
work for any sane TDX implementation.

Use TSC as the only reliable clock source in TD guests, bypassing
unstable calibration.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 v2:
  - Slightly updated commit message;
---
 arch/x86/coco/tdx/tdx.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/arch/x86/coco/tdx/tdx.c b/arch/x86/coco/tdx/tdx.c
index 1d6b863c42b0..1583ec64d92e 100644
--- a/arch/x86/coco/tdx/tdx.c
+++ b/arch/x86/coco/tdx/tdx.c
@@ -769,6 +769,9 @@ void __init tdx_early_init(void)
 
 	setup_force_cpu_cap(X86_FEATURE_TDX_GUEST);
 
+	/* TSC is the only reliable clock in TDX guest */
+	setup_force_cpu_cap(X86_FEATURE_TSC_RELIABLE);
+
 	cc_vendor = CC_VENDOR_INTEL;
 	tdx_parse_tdinfo(&cc_mask);
 	cc_set_mask(cc_mask);
-- 
2.41.0

On Wed, Aug 23, 2023 at 02:18:23AM +0300,
"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> wrote:

> In x86 virtualization environments, including TDX, RDTSC instruction is
> handled without causing a VM exit, resulting in minimal overhead and
> jitters. On the other hand, other clock sources (such as HPET, ACPI
> timer, APIC, etc.) necessitate VM exits to implement, resulting in more
> fluctuating measurements compared to TSC. Thus, those clock sources are
> not effective for calibrating TSC.
> 
> In TD guests, TSC is virtualized by the TDX module, which ensures:
> 
>   - Virtual TSC values are consistent among all the TD’s VCPUs;
>   - Monotonously incrementing for any single VCPU;
>   - The frequency is determined by TD configuration. The host TSC is
>     invariant on platforms where TDX is available.
> 
> Reliable TSC is architectural guarantee for the TDX platform and it must
> work for any sane TDX implementation.
> 
> Use TSC as the only reliable clock source in TD guests, bypassing
> unstable calibration.
> 
> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> ---
>  v2:
>   - Slightly updated commit message;
> ---
>  arch/x86/coco/tdx/tdx.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/arch/x86/coco/tdx/tdx.c b/arch/x86/coco/tdx/tdx.c
> index 1d6b863c42b0..1583ec64d92e 100644
> --- a/arch/x86/coco/tdx/tdx.c
> +++ b/arch/x86/coco/tdx/tdx.c
> @@ -769,6 +769,9 @@ void __init tdx_early_init(void)
>  
>  	setup_force_cpu_cap(X86_FEATURE_TDX_GUEST);
>  
> +	/* TSC is the only reliable clock in TDX guest */
> +	setup_force_cpu_cap(X86_FEATURE_TSC_RELIABLE);
> +
>  	cc_vendor = CC_VENDOR_INTEL;
>  	tdx_parse_tdinfo(&cc_mask);
>  	cc_set_mask(cc_mask);
> -- 
> 2.41.0
> 
> 

Reviewed-by: Isaku Yamahata <isaku.yamahata@intel.com>
-- 
Isaku Yamahata <isaku.yamahata@linux.intel.com>

On Wed, 2023-08-23 at 02:18 +0300, Kirill A. Shutemov wrote:
> In x86 virtualization environments, including TDX, RDTSC instruction is
> handled without causing a VM exit, resulting in minimal overhead and
> jitters. On the other hand, other clock sources (such as HPET, ACPI
> timer, APIC, etc.) necessitate VM exits to implement, resulting in more
> fluctuating measurements compared to TSC. Thus, those clock sources are
> not effective for calibrating TSC.
> 
> In TD guests, TSC is virtualized by the TDX module, which ensures:
> 
>   - Virtual TSC values are consistent among all the TD’s VCPUs;
>   - Monotonously incrementing for any single VCPU;

Nit:

    - Virtual TSC is monotonously incrementing ...

Otherwise here "incrementing" looks like a noun, which means you need to use
"monotonous" rather than "monotonously".


>   - The frequency is determined by TD configuration. The host TSC is
>     invariant on platforms where TDX is available.
> 
> Reliable TSC is architectural guarantee for the TDX platform and it must
> work for any sane TDX implementation.
> 
> Use TSC as the only reliable clock source in TD guests, bypassing
> unstable calibration.
> 
> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>

Acked-by: Kai Huang <kai.huang@intel.com>

On Wed, Aug 23, 2023 at 04:47:29AM +0000, Huang, Kai wrote:
> On Wed, 2023-08-23 at 02:18 +0300, Kirill A. Shutemov wrote:
> > In x86 virtualization environments, including TDX, RDTSC instruction is
> > handled without causing a VM exit, resulting in minimal overhead and
> > jitters. On the other hand, other clock sources (such as HPET, ACPI
> > timer, APIC, etc.) necessitate VM exits to implement, resulting in more
> > fluctuating measurements compared to TSC. Thus, those clock sources are
> > not effective for calibrating TSC.
> > 
> > In TD guests, TSC is virtualized by the TDX module, which ensures:
> > 
> >   - Virtual TSC values are consistent among all the TD’s VCPUs;
> >   - Monotonously incrementing for any single VCPU;
> 
> Nit:
> 
>     - Virtual TSC is monotonously incrementing ...
> 
> Otherwise here "incrementing" looks like a noun, which means you need to use
> "monotonous" rather than "monotonously".

Okay, will update.

> >   - The frequency is determined by TD configuration. The host TSC is
> >     invariant on platforms where TDX is available.
> > 
> > Reliable TSC is architectural guarantee for the TDX platform and it must
> > work for any sane TDX implementation.
> > 
> > Use TSC as the only reliable clock source in TD guests, bypassing
> > unstable calibration.
> > 
> > Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> 
> Acked-by: Kai Huang <kai.huang@intel.com>

Thanks.

-- 
  Kiryl Shutsemau / Kirill A. Shutemov

On 8/22/2023 4:18 PM, Kirill A. Shutemov wrote:
> In x86 virtualization environments, including TDX, RDTSC instruction is
> handled without causing a VM exit, resulting in minimal overhead and
> jitters. On the other hand, other clock sources (such as HPET, ACPI
> timer, APIC, etc.) necessitate VM exits to implement, resulting in more
> fluctuating measurements compared to TSC. Thus, those clock sources are
> not effective for calibrating TSC.
> 
> In TD guests, TSC is virtualized by the TDX module, which ensures:
> 
>   - Virtual TSC values are consistent among all the TD’s VCPUs;
>   - Monotonously incrementing for any single VCPU;
>   - The frequency is determined by TD configuration. The host TSC is
>     invariant on platforms where TDX is available.
> 
> Reliable TSC is architectural guarantee for the TDX platform and it must
> work for any sane TDX implementation.
> 
> Use TSC as the only reliable clock source in TD guests, bypassing
> unstable calibration.
> 
> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> ---
>  v2:
>   - Slightly updated commit message;
> ---

Looks good to me.

Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>

>  arch/x86/coco/tdx/tdx.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/arch/x86/coco/tdx/tdx.c b/arch/x86/coco/tdx/tdx.c
> index 1d6b863c42b0..1583ec64d92e 100644
> --- a/arch/x86/coco/tdx/tdx.c
> +++ b/arch/x86/coco/tdx/tdx.c
> @@ -769,6 +769,9 @@ void __init tdx_early_init(void)
>  
>  	setup_force_cpu_cap(X86_FEATURE_TDX_GUEST);
>  
> +	/* TSC is the only reliable clock in TDX guest */
> +	setup_force_cpu_cap(X86_FEATURE_TSC_RELIABLE);
> +
>  	cc_vendor = CC_VENDOR_INTEL;
>  	tdx_parse_tdinfo(&cc_mask);
>  	cc_set_mask(cc_mask);

-- 
Sathyanarayanan Kuppuswamy
Linux Kernel Developer