[PATCH v5 0/1] SEV-SNP: Add KVM support for SNP certificate fetching

[PATCH v5 0/1] SEV-SNP: Add KVM support for SNP certificate fetching

[Michael Roth]

This patchset is also available at:

  https://github.com/amdese/linux/commits/snp-certs-v5

and is based on top of kvm/next (d3d0b8dfe060)

v4 of these patches were previously submitted under:

  [PATCH v4 0/1] SEV-SNP: Add KVM support for SNP certificate fetching
  https://lore.kernel.org/kvm/20250120215818.522175-1-huibo.wang@amd.com/


Overview
--------

The GHCB 2.0 specification defines 2 GHCB request types to allow SNP guests
to send encrypted messages/requests to firmware: SNP Guest Requests and SNP
Extended Guest Requests. These encrypted messages are used for things like
servicing attestation requests issued by the guest. Implementing support for
these is required to be fully GHCB-compliant.

For the most part, KVM only needs to handle forwarding these requests to
firmware (to be issued via the SNP_GUEST_REQUEST firmware command defined
in the SEV-SNP Firmware ABI), and then forwarding the encrypted response to
the guest.

However, in the case of SNP Extended Guest Requests, the host is also
able to provide the certificate data corresponding to the endorsement key
used by firmware to sign attestation report requests. This certificate data
is provided by userspace because:

  1) It allows for different keys/key types to be used for each particular
     guest with requiring any sort of KVM API to configure the certificate
     table in advance on a per-guest basis.

  2) It provides additional flexibility with how attestation requests might
     be handled during live migration where the certificate data for
     source/dest might be different.

  3) It allows all synchronization between certificates and firmware/signing
     key updates to be handled purely by userspace rather than requiring
     some in-kernel mechanism to facilitate it. [1]

To support fetching certificate data from userspace, a new KVM
KVM_EXIT_SNP_REQ_CERTS exit type is used to fetch the data similarly to
KVM_EXIT_MMIO/etc, with an associate KVM capability to detect/enable the
exits depending on whether userspace has been configured to provide
certificate data.

[1] https://lore.kernel.org/kvm/ZS614OSoritrE1d2@google.com/


Testing
-------

For testing this via QEMU, use the following tree:

  https://github.com/amdese/qemu/commits/snp-certs-rfc1-wip4

A basic command-line invocation for SNP with certificate data supplied
would be:

 qemu-system-x86_64 -smp 32,maxcpus=255 -cpu EPYC-Milan-v2
  -machine q35,confidential-guest-support=sev0,memory-backend=ram1
  -object memory-backend-memfd,id=ram1,size=4G,share=true,reserve=false
  -object sev-snp-guest,id=sev0,cbitpos=51,reduced-phys-bits=1,id-auth=,certs-path=/home/mroth/cert.blob
  -bios OVMF.fd

Something like the following simple example can be used to simulate an
exclusive lock being held on the certificate by management tools performing an
update:

  #include <stdlib.h>
  #include <stdio.h>
  #define __USE_GNU
  #include <fcntl.h>
  #include <unistd.h>
  #include <errno.h>
  #include <stdbool.h>
  #include <sys/types.h>
  #include <sys/stat.h>
  
  int main(int argc, void **argv)
  {
      int ret, fd, i = 0;
      char *path = argv[1];
  
      struct flock fl = {
          .l_whence = SEEK_SET,
          .l_start = 0,
          .l_len = 0,
          .l_type = F_WRLCK
      };
  
      fd = open(path, O_RDWR);
      ret = fcntl(fd, F_OFD_SETLK, &fl);
      if (ret) {
          printf("error locking file, ret %d errno %d\n", ret, errno);
          return ret;
      }
  
      while (true) {
          i++;
          printf("now holding lock (%d seconds elapsed)...\n", i);
          usleep(1000 * 1000);
      }
  
      return 0;
  }

The format of the certificate blob is defined in the GHCB 2.0 specification,
but if it's not being parsed on the guest-side then random data will suffice
for testing the KVM bits.

Any feedback/review is appreciated.

Thanks!

-Mike

Changes since v4:

 * Minor documentation updates to make the implementation notes less
   specific to QEMU.
 * Collected Reviewed-by/Tested-by from v3 since there have been no
   functional changes since then and only minor documentation updates.
 * Rebased/re-tested on top of latest kvm/next (d3d0b8dfe060)

Changes since v3:

 * This version updates the documentation scheme about how file locking is
   expected to happen.

Changes since v2:

 * As per discussion during PUCK, drop all the KVM_EXIT_COCO infrastructure
   since there are enough differences with TDX's quote generation to make
   unifying the 2 exits over-complicated for userspace, and the code-sharing
   we stand to gain from placing everything under the KVM_EXIT_COCO_*
   umbrella are of questionable benefit.
 * Update/simplify documentation as per the above.
 * Rebase/re-test on top of latest kvm-coco-queue

Changes since v1:

 * Drop subtype-specific error codes. Instead use standard error codes like
   ENOSPC/etc. and let KVM determine whether a particular error requires
   special handling for a particular KVM_EXIT_COCO subtype. (Sean)
 * Introduce special handling for EAGAIN for KVM_EXIT_COCO_REQ_CERTS such
   that the guest can be instructed to retry if userspace is temporarily unable
   to immediately lock/provide the certificate data. (Sean)
 * Move the 'ret' field of struct kvm_exit_coco to the top-level so all
   sub-types can propagate error codes the same way.
 * Add more clarifying details in KVM documentation about the suggested
   file-locking scheme to avoid races between certificate requests and updates
   to SNP firmware that might modify the endorsement key corresponding to the
   certificate data.

Changes since splitting this off from v15 SNP KVM patchset:

 * Address clang-reported warnings regarding uninitialized variables 
 * Address a memory leak of the request/response buffer pages, and refactor
   the code based on Sean's suggestions:
   https://lore.kernel.org/kvm/ZktbBRLXeOp9X6aH@google.com/
 * Fix SNP Extended Guest Request handling to only attempt to fetch
   certificates if handling MSG_REQ_REPORT (attestation) message types
 * Drop KVM_EXIT_VMGEXIT and introduce KVM_EXIT_COCO events instead
 * Refactor patch layout for easier handling/review

----------------------------------------------------------------
Michael Roth (1):
      KVM: Introduce KVM_EXIT_SNP_REQ_CERTS for SNP certificate-fetching

 Documentation/virt/kvm/api.rst  | 100 ++++++++++++++++++++++++++++++++++++++++
 arch/x86/include/asm/kvm_host.h |   1 +
 arch/x86/kvm/svm/sev.c          |  43 ++++++++++++++---
 arch/x86/kvm/x86.c              |  11 +++++
 include/uapi/linux/kvm.h        |  10 ++++
 include/uapi/linux/sev-guest.h  |   8 ++++
 6 files changed, 167 insertions(+), 6 deletions(-)

[PATCH v5 1/1] KVM: Introduce KVM_EXIT_SNP_REQ_CERTS for SNP certificate-fetching

[Michael Roth]

For SEV-SNP, the host can optionally provide a certificate table to the
guest when it issues an attestation request to firmware (see GHCB 2.0
specification regarding "SNP Extended Guest Requests"). This certificate
table can then be used to verify the endorsement key used by firmware to
sign the attestation report.

While it is possible for guests to obtain the certificates through other
means, handling it via the host provides more flexibility in being able
to keep the certificate data in sync with the endorsement key throughout
host-side operations that might resulting in the endorsement key
changing.

In the case of KVM, userspace will be responsible for fetching the
certificate table and keeping it in sync with any modifications to the
endorsement key by other userspace management tools. Define a new
KVM_EXIT_SNP_REQ_CERTS event where userspace is provided with the GPA of
the buffer the guest has provided as part of the attestation request so
that userspace can write the certificate data into it while relying on
filesystem-based locking to keep the certificates up-to-date relative to
the endorsement keys installed/utilized by firmware at the time the
certificates are fetched.

Also introduce a KVM_CAP_EXIT_SNP_REQ_CERTS capability to enable/disable
the exit for cases where userspace does not support
certificate-fetching, in which case KVM will fall back to returning an
empty certificate table if the guest provides a buffer for it.

  [Melody: Update the documentation scheme about how file locking is
  expected to happen.]

Reviewed-by: Liam Merwick <liam.merwick@oracle.com>
Tested-by: Liam Merwick <liam.merwick@oracle.com>
Tested-by: Dionna Glaze <dionnaglaze@google.com>
Signed-off-by: Michael Roth <michael.roth@amd.com>
Signed-off-by: Melody Wang <huibo.wang@amd.com>
---
 Documentation/virt/kvm/api.rst  | 100 ++++++++++++++++++++++++++++++++
 arch/x86/include/asm/kvm_host.h |   1 +
 arch/x86/kvm/svm/sev.c          |  43 ++++++++++++--
 arch/x86/kvm/x86.c              |  11 ++++
 include/uapi/linux/kvm.h        |  10 ++++
 include/uapi/linux/sev-guest.h  |   8 +++
 6 files changed, 167 insertions(+), 6 deletions(-)

diff --git a/Documentation/virt/kvm/api.rst b/Documentation/virt/kvm/api.rst
index 2b52eb77e29c..4d8aa274b65f 100644
--- a/Documentation/virt/kvm/api.rst
+++ b/Documentation/virt/kvm/api.rst
@@ -7179,6 +7179,89 @@ Please note that the kernel is allowed to use the kvm_run structure as the
 primary storage for certain register types. Therefore, the kernel may use the
 values in kvm_run even if the corresponding bit in kvm_dirty_regs is not set.
 
+::
+
+		/* KVM_EXIT_SNP_REQ_CERTS */
+		struct kvm_exit_snp_req_certs {
+			__u64 gfn;
+			__u32 npages;
+			__u32 ret;
+		};
+
+This event provides a way to request certificate data from userspace and
+have it written into guest memory. This is intended to handle attestation
+requests made by SEV-SNP guests (using the Extended Guest Requests GHCB
+command as defined by the GHCB 2.0 specification for SEV-SNP guests),
+where additional certificate data corresponding to the endorsement key
+used by firmware to sign an attestation report can be optionally provided
+by userspace to pass along to the guest together with the
+firmware-provided attestation report.
+
+KVM will supply in `gfn` the non-private guest page that userspace should
+use to write the contents of certificate data. The format of this
+certificate data is defined in the GHCB 2.0 specification (see section
+"SNP Extended Guest Request"). KVM will also supply in `npages` the
+number of contiguous pages available for writing the certificate data
+into.
+
+  - If the supplied number of pages is sufficient, userspace must write
+    the certificate table blob (in the format defined by the GHCB spec)
+    into the address corresponding to `gfn` and set `ret` to 0 to indicate
+    success. If no certificate data is available, then userspace can
+    either write an empty certificate table into the address corresponding
+    to `gfn`, or it can disable ``KVM_EXIT_SNP_REQ_CERTS`` (via
+    ``KVM_CAP_EXIT_SNP_REQ_CERTS``), in which case KVM will handle
+    returning an empty certificate table to the guest.
+
+  - If the number of pages supplied is not sufficient, userspace must set
+    the required number of pages in `npages` and then set `ret` to
+    ``ENOSPC``.
+
+  - If the certificate cannot be immediately provided, userspace should set
+    `ret` to ``EAGAIN``, which will inform the guest to retry the request
+    later. One scenario where this would be useful is if the certificate
+    is in the process of being updated and cannot be fetched until the
+    update completes (see the NOTE below regarding how file-locking can
+    be used to orchestrate such updates between management/guests).
+
+  - If some other error occurred, userspace must set `ret` to ``EIO``.
+    (This is to reserve special meaning for unused error codes in the
+    future.)
+
+NOTE: The endorsement key used by firmware may change as a result of
+management activities like updating SEV-SNP firmware or loading new
+endorsement keys, so some care should be taken to keep the returned
+certificate data in sync with the actual endorsement key in use by
+firmware at the time the attestation request is sent to SNP firmware. The
+recommended scheme to do this is to use file locking (e.g. via fcntl()'s
+F_OFD_SETLK) in the following manner:
+
+  - The VMM should obtain a shared/read or exclusive/write lock on the
+  certificate blob file before reading it and returning it to KVM, and
+  continue to hold the lock until the attestation request is actually
+  sent to firmware. To facilitate this, the VMM can set the
+  ``immediate_exit`` flag of kvm_run just after supplying the
+  certificate data, and just before and resuming the vCPU. This will
+  ensure the vCPU will exit again to userspace with ``-EINTR`` after
+  it finishes fetching the attestation request from firmware, at which
+  point the VMM can safely drop the file lock.
+
+  - Tools/libraries that perform updates to SNP firmware TCB values or
+    endorsement keys (e.g. via /dev/sev interfaces such as ``SNP_COMMIT``,
+    ``SNP_SET_CONFIG``, or ``SNP_VLEK_LOAD``, see
+    Documentation/virt/coco/sev-guest.rst for more details) in such a way
+    that the certificate blob needs to be updated, should similarly take an
+    exclusive lock on the certificate blob for the duration of any updates
+    to endorsement keys or the certificate blob contents to ensure that
+    VMMs using the above scheme will not return certificate blob data that
+    is out of sync with the endorsement key used by firmware.
+
+This scheme is recommended so that tools could naturally opt to use
+it rather than every service provider coming up with a different solution
+that they will need to work into some custom QEMU/VMM (or other related
+management tools that handle firmware updates/configuration) to solve the
+same problem, but userspace will need to ensure all such tools are using
+this scheme in order for it to work as intended.
 
 .. _cap_enable:
 
@@ -9024,6 +9107,23 @@ Do not use KVM_X86_SW_PROTECTED_VM for "real" VMs, and especially not in
 production.  The behavior and effective ABI for software-protected VMs is
 unstable.
 
+8.42 KVM_CAP_EXIT_SNP_REQ_CERTS
+-------------------------------
+
+:Capability: KVM_CAP_EXIT_SNP_REQ_CERTS
+:Architectures: x86
+:Type: vm
+
+This capability, if enabled, will cause KVM to exit to userspace with
+KVM_EXIT_SNP_REQ_CERTS exit reason to allow for fetching SNP attestation
+certificates from userspace.
+
+Calling KVM_CHECK_EXTENSION for this capability will return a non-zero
+value to indicate KVM support for KVM_EXIT_SNP_REQ_CERTS.
+
+The 1st argument to KVM_ENABLE_CAP should be 1 to indicate userspace support
+for handling this event.
+
 9. Known KVM API problems
 =========================
 
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 0b7af5902ff7..8b11d1a64378 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -1459,6 +1459,7 @@ struct kvm_arch {
 	struct kvm_x86_msr_filter __rcu *msr_filter;
 
 	u32 hypercall_exit_enabled;
+	bool snp_certs_enabled;
 
 	/* Guest can access the SGX PROVISIONKEY. */
 	bool sgx_provisioning_allowed;
diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c
index 0dbb25442ec1..a18e8eed533b 100644
--- a/arch/x86/kvm/svm/sev.c
+++ b/arch/x86/kvm/svm/sev.c
@@ -4088,6 +4088,30 @@ static int snp_handle_guest_req(struct vcpu_svm *svm, gpa_t req_gpa, gpa_t resp_
 	return ret;
 }
 
+static int snp_complete_req_certs(struct kvm_vcpu *vcpu)
+{
+	struct vcpu_svm *svm = to_svm(vcpu);
+	struct vmcb_control_area *control = &svm->vmcb->control;
+
+	if (vcpu->run->snp_req_certs.ret) {
+		if (vcpu->run->snp_req_certs.ret == ENOSPC) {
+			vcpu->arch.regs[VCPU_REGS_RBX] = vcpu->run->snp_req_certs.npages;
+			ghcb_set_sw_exit_info_2(svm->sev_es.ghcb,
+						SNP_GUEST_ERR(SNP_GUEST_VMM_ERR_INVALID_LEN, 0));
+		} else if (vcpu->run->snp_req_certs.ret == EAGAIN) {
+			ghcb_set_sw_exit_info_2(svm->sev_es.ghcb,
+						SNP_GUEST_ERR(SNP_GUEST_VMM_ERR_BUSY, 0));
+		} else {
+			ghcb_set_sw_exit_info_2(svm->sev_es.ghcb,
+						SNP_GUEST_ERR(SNP_GUEST_VMM_ERR_GENERIC, 0));
+		}
+
+		return 1; /* resume guest */
+	}
+
+	return snp_handle_guest_req(svm, control->exit_info_1, control->exit_info_2);
+}
+
 static int snp_handle_ext_guest_req(struct vcpu_svm *svm, gpa_t req_gpa, gpa_t resp_gpa)
 {
 	struct kvm *kvm = svm->vcpu.kvm;
@@ -4103,12 +4127,10 @@ static int snp_handle_ext_guest_req(struct vcpu_svm *svm, gpa_t req_gpa, gpa_t r
 	/*
 	 * As per GHCB spec, requests of type MSG_REPORT_REQ also allow for
 	 * additional certificate data to be provided alongside the attestation
-	 * report via the guest-provided data pages indicated by RAX/RBX. The
-	 * certificate data is optional and requires additional KVM enablement
-	 * to provide an interface for userspace to provide it, but KVM still
-	 * needs to be able to handle extended guest requests either way. So
-	 * provide a stub implementation that will always return an empty
-	 * certificate table in the guest-provided data pages.
+	 * report via the guest-provided data pages indicated by RAX/RBX. If
+	 * userspace enables KVM_EXIT_SNP_REQ_CERTS, then exit to userspace
+	 * to fetch the certificate data. Otherwise, return an empty certificate
+	 * table in the guest-provided data pages.
 	 */
 	if (msg_type == SNP_MSG_REPORT_REQ) {
 		struct kvm_vcpu *vcpu = &svm->vcpu;
@@ -4124,6 +4146,15 @@ static int snp_handle_ext_guest_req(struct vcpu_svm *svm, gpa_t req_gpa, gpa_t r
 		if (!PAGE_ALIGNED(data_gpa))
 			goto request_invalid;
 
+		if (vcpu->kvm->arch.snp_certs_enabled) {
+			vcpu->run->exit_reason = KVM_EXIT_SNP_REQ_CERTS;
+			vcpu->run->snp_req_certs.gfn = gpa_to_gfn(data_gpa);
+			vcpu->run->snp_req_certs.npages = data_npages;
+			vcpu->run->snp_req_certs.ret = 0;
+			vcpu->arch.complete_userspace_io = snp_complete_req_certs;
+			return 0; /* fetch certs from userspace */
+		}
+
 		/*
 		 * As per GHCB spec (see "SNP Extended Guest Request"), the
 		 * certificate table is terminated by 24-bytes of zeroes.
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 02159c967d29..67ff4a89ac81 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -4774,6 +4774,9 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
 	case KVM_CAP_READONLY_MEM:
 		r = kvm ? kvm_arch_has_readonly_mem(kvm) : 1;
 		break;
+	case KVM_CAP_EXIT_SNP_REQ_CERTS:
+		r = 1;
+		break;
 	default:
 		break;
 	}
@@ -6734,6 +6737,14 @@ int kvm_vm_ioctl_enable_cap(struct kvm *kvm,
 		mutex_unlock(&kvm->lock);
 		break;
 	}
+	case KVM_CAP_EXIT_SNP_REQ_CERTS:
+		if (cap->args[0] != 1) {
+			r = -EINVAL;
+			break;
+		}
+		kvm->arch.snp_certs_enabled = true;
+		r = 0;
+		break;
 	default:
 		r = -EINVAL;
 		break;
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index 45e6d8fca9b9..83c4e6929df7 100644
--- a/include/uapi/linux/kvm.h
+++ b/include/uapi/linux/kvm.h
@@ -135,6 +135,12 @@ struct kvm_xen_exit {
 	} u;
 };
 
+struct kvm_exit_snp_req_certs {
+	__u64 gfn;
+	__u32 npages;
+	__u32 ret;
+};
+
 #define KVM_S390_GET_SKEYS_NONE   1
 #define KVM_S390_SKEYS_MAX        1048576
 
@@ -178,6 +184,7 @@ struct kvm_xen_exit {
 #define KVM_EXIT_NOTIFY           37
 #define KVM_EXIT_LOONGARCH_IOCSR  38
 #define KVM_EXIT_MEMORY_FAULT     39
+#define KVM_EXIT_SNP_REQ_CERTS    40
 
 /* For KVM_EXIT_INTERNAL_ERROR */
 /* Emulate instruction failed. */
@@ -446,6 +453,8 @@ struct kvm_run {
 			__u64 gpa;
 			__u64 size;
 		} memory_fault;
+		/* KVM_EXIT_SNP_REQ_CERTS */
+		struct kvm_exit_snp_req_certs snp_req_certs;
 		/* Fix the size of the union. */
 		char padding[256];
 	};
@@ -929,6 +938,7 @@ struct kvm_enable_cap {
 #define KVM_CAP_PRE_FAULT_MEMORY 236
 #define KVM_CAP_X86_APIC_BUS_CYCLES_NS 237
 #define KVM_CAP_X86_GUEST_MODE 238
+#define KVM_CAP_EXIT_SNP_REQ_CERTS 239
 
 struct kvm_irq_routing_irqchip {
 	__u32 irqchip;
diff --git a/include/uapi/linux/sev-guest.h b/include/uapi/linux/sev-guest.h
index fcdfea767fca..4c4ed8bc71d7 100644
--- a/include/uapi/linux/sev-guest.h
+++ b/include/uapi/linux/sev-guest.h
@@ -95,5 +95,13 @@ struct snp_ext_report_req {
 
 #define SNP_GUEST_VMM_ERR_INVALID_LEN	1
 #define SNP_GUEST_VMM_ERR_BUSY		2
+/*
+ * The GHCB spec essentially states that all non-zero error codes other than
+ * those explicitly defined above should be treated as an error by the guest.
+ * Define a generic error to cover that case, and choose a value that is not
+ * likely to overlap with new explicit error codes should more be added to
+ * the GHCB spec later.
+ */
+#define SNP_GUEST_VMM_ERR_GENERIC       ((u32)~0U)
 
 #endif /* __UAPI_LINUX_SEV_GUEST_H_ */
-- 
2.25.1

Re: [PATCH v5 1/1] KVM: Introduce KVM_EXIT_SNP_REQ_CERTS for SNP certificate-fetching

[Joerg Roedel]

Hi Michael,

On Wed, Feb 19, 2025 at 09:15:05AM -0600, Michael Roth wrote:
> +  - If some other error occurred, userspace must set `ret` to ``EIO``.
> +    (This is to reserve special meaning for unused error codes in the
> +    future.)

[...]

> +static int snp_complete_req_certs(struct kvm_vcpu *vcpu)
> +{
> +	struct vcpu_svm *svm = to_svm(vcpu);
> +	struct vmcb_control_area *control = &svm->vmcb->control;
> +
> +	if (vcpu->run->snp_req_certs.ret) {
> +		if (vcpu->run->snp_req_certs.ret == ENOSPC) {
> +			vcpu->arch.regs[VCPU_REGS_RBX] = vcpu->run->snp_req_certs.npages;
> +			ghcb_set_sw_exit_info_2(svm->sev_es.ghcb,
> +						SNP_GUEST_ERR(SNP_GUEST_VMM_ERR_INVALID_LEN, 0));
> +		} else if (vcpu->run->snp_req_certs.ret == EAGAIN) {
> +			ghcb_set_sw_exit_info_2(svm->sev_es.ghcb,
> +						SNP_GUEST_ERR(SNP_GUEST_VMM_ERR_BUSY, 0));
> +		} else {
> +			ghcb_set_sw_exit_info_2(svm->sev_es.ghcb,
> +						SNP_GUEST_ERR(SNP_GUEST_VMM_ERR_GENERIC, 0));
> +		}

According to the documentation above, there should be a block checking
for EIO which injects SNP_GUEST_VMM_ERR_GENERIC and the else block
should return with EINVAL to user-space, no?

Regards,

-- 
Jörg Rödel
jroedel@suse.de

SUSE Software Solutions Germany GmbH
Frankenstraße 146
90461 Nürnberg
Germany
https://www.suse.com/

Geschäftsführer: Ivo Totev, Andrew McDonald, Werner Knoblich
(HRB 36809, AG Nürnberg)

Re: [PATCH v5 1/1] KVM: Introduce KVM_EXIT_SNP_REQ_CERTS for SNP certificate-fetching

[Sean Christopherson]

On Mon, Feb 24, 2025, Joerg Roedel wrote:
> Hi Michael,
> 
> On Wed, Feb 19, 2025 at 09:15:05AM -0600, Michael Roth wrote:
> > +  - If some other error occurred, userspace must set `ret` to ``EIO``.
> > +    (This is to reserve special meaning for unused error codes in the
> > +    future.)
> 
> [...]
> 
> > +static int snp_complete_req_certs(struct kvm_vcpu *vcpu)
> > +{
> > +	struct vcpu_svm *svm = to_svm(vcpu);
> > +	struct vmcb_control_area *control = &svm->vmcb->control;
> > +
> > +	if (vcpu->run->snp_req_certs.ret) {
> > +		if (vcpu->run->snp_req_certs.ret == ENOSPC) {
> > +			vcpu->arch.regs[VCPU_REGS_RBX] = vcpu->run->snp_req_certs.npages;
> > +			ghcb_set_sw_exit_info_2(svm->sev_es.ghcb,
> > +						SNP_GUEST_ERR(SNP_GUEST_VMM_ERR_INVALID_LEN, 0));
> > +		} else if (vcpu->run->snp_req_certs.ret == EAGAIN) {
> > +			ghcb_set_sw_exit_info_2(svm->sev_es.ghcb,
> > +						SNP_GUEST_ERR(SNP_GUEST_VMM_ERR_BUSY, 0));
> > +		} else {
> > +			ghcb_set_sw_exit_info_2(svm->sev_es.ghcb,
> > +						SNP_GUEST_ERR(SNP_GUEST_VMM_ERR_GENERIC, 0));
> > +		}
> 
> According to the documentation above, there should be a block checking
> for EIO which injects SNP_GUEST_VMM_ERR_GENERIC and the else block
> should return with EINVAL to user-space, no?

Yeah.  It feels a bit ridiculous, but it would be quite unfortunate to go through
the extra effort of decoupling KVM's error handling from the GHCB error code, only
for it to all fall apart due to not enforcing the "return" value.

Re: [PATCH v5 1/1] KVM: Introduce KVM_EXIT_SNP_REQ_CERTS for SNP certificate-fetching

[Sean Christopherson]

On Wed, Feb 19, 2025, Michael Roth wrote:
> +  - If the supplied number of pages is sufficient, userspace must write
> +    the certificate table blob (in the format defined by the GHCB spec)
> +    into the address corresponding to `gfn` and set `ret` to 0 to indicate
> +    success. If no certificate data is available, then userspace can
> +    either write an empty certificate table into the address corresponding
> +    to `gfn`, or it can disable ``KVM_EXIT_SNP_REQ_CERTS`` (via
> +    ``KVM_CAP_EXIT_SNP_REQ_CERTS``), in which case KVM will handle

This doesn't match the code.  As is, KVM only allows *enabling* cert requests.
Userspace can only "disable" cert requests by doing nothing.

> +    returning an empty certificate table to the guest.
> +
> +  - If the number of pages supplied is not sufficient, userspace must set
> +    the required number of pages in `npages` and then set `ret` to
> +    ``ENOSPC``.
> +
> +  - If the certificate cannot be immediately provided, userspace should set
> +    `ret` to ``EAGAIN``, which will inform the guest to retry the request
> +    later. One scenario where this would be useful is if the certificate
> +    is in the process of being updated and cannot be fetched until the
> +    update completes (see the NOTE below regarding how file-locking can
> +    be used to orchestrate such updates between management/guests).
> +
> +  - If some other error occurred, userspace must set `ret` to ``EIO``.
> +    (This is to reserve special meaning for unused error codes in the
> +    future.)

I would phrase this differently, and say that KVM responds with -1 (or whatever
the generic KVM-defined error value is) to the guest if ret is set to EIO, and
then state that all other "return" values are reserved.  Because EIO isn't a
placeholder, it's full on ABI once this lands.

> +8.42 KVM_CAP_EXIT_SNP_REQ_CERTS
> +-------------------------------
> +
> +:Capability: KVM_CAP_EXIT_SNP_REQ_CERTS
> +:Architectures: x86
> +:Type: vm
> +
> +This capability, if enabled, will cause KVM to exit to userspace with
> +KVM_EXIT_SNP_REQ_CERTS exit reason to allow for fetching SNP attestation
> +certificates from userspace.
> +
> +Calling KVM_CHECK_EXTENSION for this capability will return a non-zero
> +value to indicate KVM support for KVM_EXIT_SNP_REQ_CERTS.
> +
> +The 1st argument to KVM_ENABLE_CAP should be 1 to indicate userspace support
> +for handling this event.
> +
>  9. Known KVM API problems
>  =========================
>  
> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> index 0b7af5902ff7..8b11d1a64378 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -1459,6 +1459,7 @@ struct kvm_arch {
>  	struct kvm_x86_msr_filter __rcu *msr_filter;
>  
>  	u32 hypercall_exit_enabled;
> +	bool snp_certs_enabled;

Given that this is obviously specific to SNP, I think it belongs in kvm_sev_info.
And trying to enable the capability on a non-SNP VM should fail.

Actually, rather than a generic capability, what about using the SEV "attributes"
to enumerate support, and then use KVM_MEMORY_ENCRYPT_OP to enable cert requests.
That way we don't need to add yet more plumbing to KVM, and it fits with the
direction Paolo went for KVM_SEV_INIT2.

E.g. I think the KVM side of things ends up like:

diff --git a/arch/x86/include/uapi/asm/kvm.h b/arch/x86/include/uapi/asm/kvm.h
index 9e75da97bce0..675795a83c44 100644
--- a/arch/x86/include/uapi/asm/kvm.h
+++ b/arch/x86/include/uapi/asm/kvm.h
@@ -467,6 +467,7 @@ struct kvm_sync_regs {
 /* vendor-specific groups and attributes for system fd */
 #define KVM_X86_GRP_SEV                        1
 #  define KVM_X86_SEV_VMSA_FEATURES    0
+#  define KVM_X86_SNP_SNP_REQ_CERTS    1
 
 struct kvm_vmx_nested_state_data {
        __u8 vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE];
@@ -704,6 +705,8 @@ enum sev_cmd_id {
        KVM_SEV_SNP_LAUNCH_START = 100,
        KVM_SEV_SNP_LAUNCH_UPDATE,
        KVM_SEV_SNP_LAUNCH_FINISH,
+       KVM_SEV_SNP_ENABLE_REQ_CERTS,
+
 
        KVM_SEV_NR_MAX,
 };
diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c
index 74525651770a..ddadd14551fc 100644
--- a/arch/x86/kvm/svm/sev.c
+++ b/arch/x86/kvm/svm/sev.c
@@ -2123,7 +2123,9 @@ int sev_dev_get_attr(u32 group, u64 attr, u64 *val)
        case KVM_X86_SEV_VMSA_FEATURES:
                *val = sev_supported_vmsa_features;
                return 0;
-
+       case KVM_X86_SNP_SNP_REQ_CERTS
+               *val = 1;
+               return 0;
        default:
                return -ENXIO;
        }
@@ -2535,6 +2537,15 @@ static int snp_launch_finish(struct kvm *kvm, struct kvm_sev_cmd *argp)
        return ret;
 }
 
+static int snp_enable_certs(struct kvm *kvm)
+{
+       if (kvm->created_vcpus || !sev_snp_guest(kvm))
+               return -EINVAL;
+
+       to_kvm_sev_info(kvm)->snp_certs_enabled = true;
+       return 0;
+}
+
 int sev_mem_enc_ioctl(struct kvm *kvm, void __user *argp)
 {
        struct kvm_sev_cmd sev_cmd;
@@ -2640,6 +2651,9 @@ int sev_mem_enc_ioctl(struct kvm *kvm, void __user *argp)
        case KVM_SEV_SNP_LAUNCH_FINISH:
                r = snp_launch_finish(kvm, &sev_cmd);
                break;
+       case KVM_SEV_SNP_ENABLE_REQ_CERTS:
+               r = snp_enable_certs(kvm);
+               break;
        default:
                r = -EINVAL;
                goto out;


>  	/* Guest can access the SGX PROVISIONKEY. */
>  	bool sgx_provisioning_allowed;
> diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c
> index 0dbb25442ec1..a18e8eed533b 100644
> --- a/arch/x86/kvm/svm/sev.c
> +++ b/arch/x86/kvm/svm/sev.c
> @@ -4088,6 +4088,30 @@ static int snp_handle_guest_req(struct vcpu_svm *svm, gpa_t req_gpa, gpa_t resp_
>  	return ret;
>  }
>  
> +static int snp_complete_req_certs(struct kvm_vcpu *vcpu)
> +{
> +	struct vcpu_svm *svm = to_svm(vcpu);
> +	struct vmcb_control_area *control = &svm->vmcb->control;
> +
> +	if (vcpu->run->snp_req_certs.ret) {

This should be READ_ONCE() to avoid weirdness if userspace is scribbling kvm_run.

Aha!  Idea.  To reduce indentation and dedup some code, add a helper to fill the
error field.  And then you can use a switch statement too.  E.g.

	switch (ret) {
	case 0:
		return snp_handle_guest_req(svm, control->exit_info_1,
					    control->exit_info_2);
	case ENOSPC:
		vcpu->arch.regs[VCPU_REGS_RBX] = vcpu->run->snp_req_certs.npages;
		return snp_req_certs_err(svm, SNP_GUEST_VMM_ERR_INVALID_LEN);
	case EAGAIN:
		return snp_req_certs_err(svm, SNP_GUEST_VMM_ERR_BUSY);
	case EIO:
		return snp_req_certs_err(svm, SNP_GUEST_VMM_ERR_GENERIC);
	default:
		break;
	}

	return -EINVAL;

> @@ -4124,6 +4146,15 @@ static int snp_handle_ext_guest_req(struct vcpu_svm *svm, gpa_t req_gpa, gpa_t r
>  		if (!PAGE_ALIGNED(data_gpa))
>  			goto request_invalid;
>  
> +		if (vcpu->kvm->arch.snp_certs_enabled) {
> +			vcpu->run->exit_reason = KVM_EXIT_SNP_REQ_CERTS;
> +			vcpu->run->snp_req_certs.gfn = gpa_to_gfn(data_gpa);
> +			vcpu->run->snp_req_certs.npages = data_npages;
> +			vcpu->run->snp_req_certs.ret = 0;
> +			vcpu->arch.complete_userspace_io = snp_complete_req_certs;
> +			return 0; /* fetch certs from userspace */
> +		}
> +
>  		/*
>  		 * As per GHCB spec (see "SNP Extended Guest Request"), the
>  		 * certificate table is terminated by 24-bytes of zeroes.
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 02159c967d29..67ff4a89ac81 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -4774,6 +4774,9 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
>  	case KVM_CAP_READONLY_MEM:
>  		r = kvm ? kvm_arch_has_readonly_mem(kvm) : 1;
>  		break;
> +	case KVM_CAP_EXIT_SNP_REQ_CERTS:
> +		r = 1;

This should be conditional on KVM supporting SNP.  Or if we don't want the support
to be dynamic based on hardware capabilities and module params, at least conditional
on KVM supporting SEV in the Kconfig.

> +		break;
>  	default:
>  		break;
>  	}
> @@ -6734,6 +6737,14 @@ int kvm_vm_ioctl_enable_cap(struct kvm *kvm,
>  		mutex_unlock(&kvm->lock);
>  		break;
>  	}
> +	case KVM_CAP_EXIT_SNP_REQ_CERTS:
> +		if (cap->args[0] != 1) {

Definitely prefer a proper #define for the flag.  Or as above, put "ENABLE" in
the name and ignore the args entirely (probably my vote).

> +			r = -EINVAL;
> +			break;
> +		}
> +		kvm->arch.snp_certs_enabled = true;

This should be guarded with a check that vCPUs haven't been created, so that KVM
doesn't have to think about what can go wrong if userspace toggle the capability
while vCPUs are running.

> +		r = 0;
> +		break;
>  	default:
>  		r = -EINVAL;
>  		break;
> diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
> index 45e6d8fca9b9..83c4e6929df7 100644
> --- a/include/uapi/linux/kvm.h
> +++ b/include/uapi/linux/kvm.h
> @@ -135,6 +135,12 @@ struct kvm_xen_exit {
>  	} u;
>  };
>  
> +struct kvm_exit_snp_req_certs {
> +	__u64 gfn;
> +	__u32 npages;

Space is not at a premium.  Given that data_npages is a u64, and KVM typically
tracks the number of guest pages with gfn_t, I think it makes sense to have this
be a u64.

> +	__u32 ret;

Same here, use a u64.  No reason to pack this tightly.

> +};
> +
>  #define KVM_S390_GET_SKEYS_NONE   1
>  #define KVM_S390_SKEYS_MAX        1048576
>  
> @@ -178,6 +184,7 @@ struct kvm_xen_exit {
>  #define KVM_EXIT_NOTIFY           37
>  #define KVM_EXIT_LOONGARCH_IOCSR  38
>  #define KVM_EXIT_MEMORY_FAULT     39
> +#define KVM_EXIT_SNP_REQ_CERTS    40
>  
>  /* For KVM_EXIT_INTERNAL_ERROR */
>  /* Emulate instruction failed. */
> @@ -446,6 +453,8 @@ struct kvm_run {
>  			__u64 gpa;
>  			__u64 size;
>  		} memory_fault;
> +		/* KVM_EXIT_SNP_REQ_CERTS */
> +		struct kvm_exit_snp_req_certs snp_req_certs;
>  		/* Fix the size of the union. */
>  		char padding[256];
>  	};
> @@ -929,6 +938,7 @@ struct kvm_enable_cap {
>  #define KVM_CAP_PRE_FAULT_MEMORY 236
>  #define KVM_CAP_X86_APIC_BUS_CYCLES_NS 237
>  #define KVM_CAP_X86_GUEST_MODE 238
> +#define KVM_CAP_EXIT_SNP_REQ_CERTS 239
>  
>  struct kvm_irq_routing_irqchip {
>  	__u32 irqchip;
> diff --git a/include/uapi/linux/sev-guest.h b/include/uapi/linux/sev-guest.h
> index fcdfea767fca..4c4ed8bc71d7 100644
> --- a/include/uapi/linux/sev-guest.h
> +++ b/include/uapi/linux/sev-guest.h
> @@ -95,5 +95,13 @@ struct snp_ext_report_req {
>  
>  #define SNP_GUEST_VMM_ERR_INVALID_LEN	1
>  #define SNP_GUEST_VMM_ERR_BUSY		2
> +/*
> + * The GHCB spec essentially states that all non-zero error codes other than
> + * those explicitly defined above should be treated as an error by the guest.
> + * Define a generic error to cover that case, and choose a value that is not
> + * likely to overlap with new explicit error codes should more be added to
> + * the GHCB spec later.
> + */
> +#define SNP_GUEST_VMM_ERR_GENERIC       ((u32)~0U)

Why cast to a u32?  The "U" should take care of things.  If you need to cast, it
should be __u32.

>  
>  #endif /* __UAPI_LINUX_SEV_GUEST_H_ */
> -- 
> 2.25.1
>