summaryrefslogtreecommitdiffstats
path: root/include/uapi
diff options
context:
space:
mode:
authorBrijesh Singh <brijesh.singh@amd.com>2024-07-02 00:31:46 +0200
committerPaolo Bonzini <pbonzini@redhat.com>2024-07-16 17:44:00 +0200
commit88caf544c9305313e1c48ac1a437faa5df8fff06 (patch)
tree2b9ec5a58cdf1e6fbe9217358cde518282582775 /include/uapi
parentKVM: SVM: Remove the need to trigger an UNBLOCK event on AP creation (diff)
downloadlinux-88caf544c9305313e1c48ac1a437faa5df8fff06.tar.xz
linux-88caf544c9305313e1c48ac1a437faa5df8fff06.zip
KVM: SEV: Provide support for SNP_GUEST_REQUEST NAE event
Version 2 of GHCB specification added support for the SNP Guest Request Message NAE event. The event allows for an SEV-SNP guest to make requests to the SEV-SNP firmware through the hypervisor using the SNP_GUEST_REQUEST API defined in the SEV-SNP firmware specification. This is used by guests primarily to request attestation reports from firmware. There are other request types are available as well, but the specifics of what guest requests are being made generally does not affect how they are handled by the hypervisor, which only serves as a proxy for the guest requests and firmware responses. Implement handling for these events. When an SNP Guest Request is issued, the guest will provide its own request/response pages, which could in theory be passed along directly to firmware. However, these pages would need special care: - Both pages are from shared guest memory, so they need to be protected from migration/etc. occurring while firmware reads/writes to them. At a minimum, this requires elevating the ref counts and potentially needing an explicit pinning of the memory. This places additional restrictions on what type of memory backends userspace can use for shared guest memory since there would be some reliance on using refcounted pages. - The response page needs to be switched to Firmware-owned state before the firmware can write to it, which can lead to potential host RMP #PFs if the guest is misbehaved and hands the host a guest page that KVM is writing to for other reasons (e.g. virtio buffers). Both of these issues can be avoided completely by using separately-allocated bounce pages for both the request/response pages and passing those to firmware instead. So that's the approach taken here. Signed-off-by: Brijesh Singh <brijesh.singh@amd.com> Co-developed-by: Alexey Kardashevskiy <aik@amd.com> Signed-off-by: Alexey Kardashevskiy <aik@amd.com> Co-developed-by: Ashish Kalra <ashish.kalra@amd.com> Signed-off-by: Ashish Kalra <ashish.kalra@amd.com> Reviewed-by: Tom Lendacky <thomas.lendacky@amd.com> Reviewed-by: Liam Merwick <liam.merwick@oracle.com> [mdr: ensure FW command failures are indicated to guest, drop extended request handling to be re-written as separate patch, massage commit] Signed-off-by: Michael Roth <michael.roth@amd.com> Message-ID: <20240701223148.3798365-2-michael.roth@amd.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Diffstat (limited to 'include/uapi')
-rw-r--r--include/uapi/linux/sev-guest.h3
1 files changed, 3 insertions, 0 deletions
diff --git a/include/uapi/linux/sev-guest.h b/include/uapi/linux/sev-guest.h
index 154a87a1eca9..fcdfea767fca 100644
--- a/include/uapi/linux/sev-guest.h
+++ b/include/uapi/linux/sev-guest.h
@@ -89,6 +89,9 @@ struct snp_ext_report_req {
#define SNP_GUEST_FW_ERR_MASK GENMASK_ULL(31, 0)
#define SNP_GUEST_VMM_ERR_SHIFT 32
#define SNP_GUEST_VMM_ERR(x) (((u64)x) << SNP_GUEST_VMM_ERR_SHIFT)
+#define SNP_GUEST_FW_ERR(x) ((x) & SNP_GUEST_FW_ERR_MASK)
+#define SNP_GUEST_ERR(vmm_err, fw_err) (SNP_GUEST_VMM_ERR(vmm_err) | \
+ SNP_GUEST_FW_ERR(fw_err))
#define SNP_GUEST_VMM_ERR_INVALID_LEN 1
#define SNP_GUEST_VMM_ERR_BUSY 2