Re: [PATCH v19 108/130] KVM: TDX: Handle TDX PV HLT hypercall

[Chao Gao]

On Mon, Feb 26, 2024 at 12:26:50AM -0800, isaku.yamahata@xxxxxxxxx wrote:
>From: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
>
>Wire up TDX PV HLT hypercall to the KVM backend function.
>
>Signed-off-by: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
>---
>v19:
>- move tdvps_state_non_arch_check() to this patch
>
>v18:
>- drop buggy_hlt_workaround and use TDH.VP.RD(TD_VCPU_STATE_DETAILS)
>
>Signed-off-by: Isaku Yamahata <isaku.yamahata@xxxxxxxxx>
>---
> arch/x86/kvm/vmx/tdx.c | 26 +++++++++++++++++++++++++-
> arch/x86/kvm/vmx/tdx.h |  4 ++++
> 2 files changed, 29 insertions(+), 1 deletion(-)
>
>diff --git a/arch/x86/kvm/vmx/tdx.c b/arch/x86/kvm/vmx/tdx.c
>index eb68d6c148b6..a2caf2ae838c 100644
>--- a/arch/x86/kvm/vmx/tdx.c
>+++ b/arch/x86/kvm/vmx/tdx.c
>@@ -688,7 +688,18 @@ void tdx_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
> 
> bool tdx_protected_apic_has_interrupt(struct kvm_vcpu *vcpu)
> {
>-	return pi_has_pending_interrupt(vcpu);
>+	bool ret = pi_has_pending_interrupt(vcpu);

Maybe
	bool has_pending_interrupt = pi_has_pending_interrupt(vcpu);

"ret" isn't a good name. or even call pi_has_pending_interrupt() directly in
the if statement below.

>+	union tdx_vcpu_state_details details;
>+	struct vcpu_tdx *tdx = to_tdx(vcpu);
>+
>+	if (ret || vcpu->arch.mp_state != KVM_MP_STATE_HALTED)
>+		return true;

Question: why mp_state matters here?

>+
>+	if (tdx->interrupt_disabled_hlt)
>+		return false;

Shouldn't we move this into vt_interrupt_allowed()? VMX calls the function to
check if interrupt is disabled. KVM can clear tdx->interrupt_disabled_hlt on
every TD-enter and set it only on TD-exit due to the guest making a
TDVMCALL(hlt) w/ interrupt disabled.

>+
>+	details.full = td_state_non_arch_read64(tdx, TD_VCPU_STATE_DETAILS_NON_ARCH);
>+	return !!details.vmxip;
> }
> 
> void tdx_prepare_switch_to_guest(struct kvm_vcpu *vcpu)
>@@ -1130,6 +1141,17 @@ static int tdx_emulate_cpuid(struct kvm_vcpu *vcpu)
> 	return 1;
> }
> 
>+static int tdx_emulate_hlt(struct kvm_vcpu *vcpu)
>+{
>+	struct vcpu_tdx *tdx = to_tdx(vcpu);
>+
>+	/* See tdx_protected_apic_has_interrupt() to avoid heavy seamcall */
>+	tdx->interrupt_disabled_hlt = tdvmcall_a0_read(vcpu);
>+
>+	tdvmcall_set_return_code(vcpu, TDVMCALL_SUCCESS);
>+	return kvm_emulate_halt_noskip(vcpu);
>+}
>+
> static int handle_tdvmcall(struct kvm_vcpu *vcpu)
> {
> 	if (tdvmcall_exit_type(vcpu))
>@@ -1138,6 +1160,8 @@ static int handle_tdvmcall(struct kvm_vcpu *vcpu)
> 	switch (tdvmcall_leaf(vcpu)) {
> 	case EXIT_REASON_CPUID:
> 		return tdx_emulate_cpuid(vcpu);
>+	case EXIT_REASON_HLT:
>+		return tdx_emulate_hlt(vcpu);
> 	default:
> 		break;
> 	}
>diff --git a/arch/x86/kvm/vmx/tdx.h b/arch/x86/kvm/vmx/tdx.h
>index 4399d474764f..11c74c34555f 100644
>--- a/arch/x86/kvm/vmx/tdx.h
>+++ b/arch/x86/kvm/vmx/tdx.h
>@@ -104,6 +104,8 @@ struct vcpu_tdx {
> 	bool host_state_need_restore;
> 	u64 msr_host_kernel_gs_base;
> 
>+	bool interrupt_disabled_hlt;
>+
> 	/*
> 	 * Dummy to make pmu_intel not corrupt memory.
> 	 * TODO: Support PMU for TDX.  Future work.
>@@ -166,6 +168,7 @@ static __always_inline void tdvps_vmcs_check(u32 field, u8 bits)
> }
> 
> static __always_inline void tdvps_management_check(u64 field, u8 bits) {}
>+static __always_inline void tdvps_state_non_arch_check(u64 field, u8 bits) {}
> 
> #define TDX_BUILD_TDVPS_ACCESSORS(bits, uclass, lclass)				\
> static __always_inline u##bits td_##lclass##_read##bits(struct vcpu_tdx *tdx,	\
>@@ -226,6 +229,7 @@ TDX_BUILD_TDVPS_ACCESSORS(32, VMCS, vmcs);
> TDX_BUILD_TDVPS_ACCESSORS(64, VMCS, vmcs);
> 
> TDX_BUILD_TDVPS_ACCESSORS(8, MANAGEMENT, management);
>+TDX_BUILD_TDVPS_ACCESSORS(64, STATE_NON_ARCH, state_non_arch);
> 
> static __always_inline u64 td_tdcs_exec_read64(struct kvm_tdx *kvm_tdx, u32 field)
> {
>-- 
>2.25.1
>
>