xref: /qemu/docs/system/i386/tdx.rst (revision 98721058d6d50ef218e0c26e4f67c8ef96965859)
1Intel Trusted Domain eXtension (TDX)
2====================================
3
4Intel Trusted Domain eXtensions (TDX) refers to an Intel technology that extends
5Virtual Machine Extensions (VMX) and Multi-Key Total Memory Encryption (MKTME)
6with a new kind of virtual machine guest called a Trust Domain (TD). A TD runs
7in a CPU mode that is designed to protect the confidentiality of its memory
8contents and its CPU state from any other software, including the hosting
9Virtual Machine Monitor (VMM), unless explicitly shared by the TD itself.
10
11Prerequisites
12-------------
13
14To run TD, the physical machine needs to have TDX module loaded and initialized
15while KVM hypervisor has TDX support and has TDX enabled. If those requirements
16are met, the ``KVM_CAP_VM_TYPES`` will report the support of ``KVM_X86_TDX_VM``.
17
18Trust Domain Virtual Firmware (TDVF)
19~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
20
21Trust Domain Virtual Firmware (TDVF) is required to provide TD services to boot
22TD Guest OS. TDVF needs to be copied to guest private memory and measured before
23the TD boots.
24
25KVM vcpu ioctl ``KVM_TDX_INIT_MEM_REGION`` can be used to populate the TDVF
26content into its private memory.
27
28Since TDX doesn't support readonly memslot, TDVF cannot be mapped as pflash
29device and it actually works as RAM. "-bios" option is chosen to load TDVF.
30
31OVMF is the opensource firmware that implements the TDVF support. Thus the
32command line to specify and load TDVF is ``-bios OVMF.fd``
33
34Feature Configuration
35---------------------
36
37Unlike non-TDX VM, the CPU features (enumerated by CPU or MSR) of a TD are not
38under full control of VMM. VMM can only configure part of features of a TD on
39``KVM_TDX_INIT_VM`` command of VM scope ``MEMORY_ENCRYPT_OP`` ioctl.
40
41The configurable features have three types:
42
43- Attributes:
44  - PKS (bit 30) controls whether Supervisor Protection Keys is exposed to TD,
45  which determines related CPUID bit and CR4 bit;
46  - PERFMON (bit 63) controls whether PMU is exposed to TD.
47
48- XSAVE related features (XFAM):
49  XFAM is a 64b mask, which has the same format as XCR0 or IA32_XSS MSR. It
50  determines the set of extended features available for use by the guest TD.
51
52- CPUID features:
53  Only some bits of some CPUID leaves are directly configurable by VMM.
54
55What features can be configured is reported via TDX capabilities.
56
57TDX capabilities
58~~~~~~~~~~~~~~~~
59
60The VM scope ``MEMORY_ENCRYPT_OP`` ioctl provides command ``KVM_TDX_CAPABILITIES``
61to get the TDX capabilities from KVM. It returns a data structure of
62``struct kvm_tdx_capabilities``, which tells the supported configuration of
63attributes, XFAM and CPUIDs.
64
65TD attributes
66~~~~~~~~~~~~~
67
68QEMU supports configuring raw 64-bit TD attributes directly via "attributes"
69property of "tdx-guest" object. Note, it's users' responsibility to provide a
70valid value because some bits may not supported by current QEMU or KVM yet.
71
72QEMU also supports the configuration of individual attribute bits that are
73supported by it, via properties of "tdx-guest" object.
74E.g., "sept-ve-disable" (bit 28).
75
76MSR based features
77~~~~~~~~~~~~~~~~~~
78
79Current KVM doesn't support MSR based feature (e.g., MSR_IA32_ARCH_CAPABILITIES)
80configuration for TDX, and it's a future work to enable it in QEMU when KVM adds
81support of it.
82
83Feature check
84~~~~~~~~~~~~~
85
86QEMU checks if the final (CPU) features, determined by given cpu model and
87explicit feature adjustment of "+featureA/-featureB", can be supported or not.
88It can produce feature not supported warning like
89
90  "warning: host doesn't support requested feature: CPUID.07H:EBX.intel-pt [bit 25]"
91
92It can also produce warning like
93
94  "warning: TDX forcibly sets the feature: CPUID.80000007H:EDX.invtsc [bit 8]"
95
96if the fixed-1 feature is requested to be disabled explicitly. This is newly
97added to QEMU for TDX because TDX has fixed-1 features that are forcibly enabled
98by TDX module and VMM cannot disable them.
99
100Launching a TD (TDX VM)
101-----------------------
102
103To launch a TD, the necessary command line options are tdx-guest object and
104split kernel-irqchip, as below:
105
106.. parsed-literal::
107
108    |qemu_system_x86| \\
109        -accel kvm \\
110        -cpu host \\
111        -object tdx-guest,id=tdx0 \\
112        -machine ...,confidential-guest-support=tdx0 \\
113        -bios OVMF.fd \\
114
115Restrictions
116------------
117
118 - kernel-irqchip must be split;
119
120   This is set by default for TDX guest if kernel-irqchip is left on its default
121   'auto' setting.
122
123 - No readonly support for private memory;
124
125 - No SMM support: SMM support requires manipulating the guest register states
126   which is not allowed;
127
128Debugging
129---------
130
131Bit 0 of TD attributes, is DEBUG bit, which decides if the TD runs in off-TD
132debug mode. When in off-TD debug mode, TD's VCPU state and private memory are
133accessible via given SEAMCALLs. This requires KVM to expose APIs to invoke those
134SEAMCALLs and corresonponding QEMU change.
135
136It's targeted as future work.
137
138TD attestation
139--------------
140
141In TD guest, the attestation process is used to verify the TDX guest
142trustworthiness to other entities before provisioning secrets to the guest.
143
144TD attestation is initiated first by calling TDG.MR.REPORT inside TD to get the
145REPORT. Then the REPORT data needs to be converted into a remotely verifiable
146Quote by SGX Quoting Enclave (QE).
147
148It's a future work in QEMU to add support of TD attestation since it lacks
149support in current KVM.
150
151Live Migration
152--------------
153
154Future work.
155
156References
157----------
158
159- `TDX Homepage <https://www.intel.com/content/www/us/en/developer/articles/technical/intel-trust-domain-extensions.html>`__
160
161- `SGX QE <https://github.com/intel/SGXDataCenterAttestationPrimitives/tree/master/QuoteGeneration>`__
162