]> www.infradead.org Git - nvme.git/commitdiff
KVM: SEV: Automatically switch reclaimed pages to shared
authorMichael Roth <michael.roth@amd.com>
Tue, 28 May 2024 20:58:09 +0000 (15:58 -0500)
committerPaolo Bonzini <pbonzini@redhat.com>
Mon, 3 Jun 2024 16:36:52 +0000 (12:36 -0400)
Currently there's a consistent pattern of always calling
host_rmp_make_shared() immediately after snp_page_reclaim(), so go ahead
and handle it automatically as part of snp_page_reclaim(). Also rename
it to kvm_rmp_make_shared() to more easily distinguish it as a
KVM-specific variant of the more generic rmp_make_shared() helper.

Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Michael Roth <michael.roth@amd.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
arch/x86/kvm/svm/sev.c

index 176ba117413af9c69de9bbe26f66b81ef58edace..39f8eeea5659f11be593f54f9653ba53df0e9a4a 100644 (file)
@@ -262,6 +262,21 @@ static void sev_decommission(unsigned int handle)
        sev_guest_decommission(&decommission, NULL);
 }
 
+/*
+ * Transition a page to hypervisor-owned/shared state in the RMP table. This
+ * should not fail under normal conditions, but leak the page should that
+ * happen since it will no longer be usable by the host due to RMP protections.
+ */
+static int kvm_rmp_make_shared(struct kvm *kvm, u64 pfn, enum pg_level level)
+{
+       if (KVM_BUG_ON(rmp_make_shared(pfn, level), kvm)) {
+               snp_leak_pages(pfn, page_level_size(level) >> PAGE_SHIFT);
+               return -EIO;
+       }
+
+       return 0;
+}
+
 /*
  * Certain page-states, such as Pre-Guest and Firmware pages (as documented
  * in Chapter 5 of the SEV-SNP Firmware ABI under "Page States") cannot be
@@ -271,32 +286,25 @@ static void sev_decommission(unsigned int handle)
  * Until they are reclaimed and subsequently transitioned via RMPUPDATE, they
  * might not be usable by the host due to being set as immutable or still
  * being associated with a guest ASID.
+ *
+ * Bug the VM and leak the page if reclaim fails, or if the RMP entry can't be
+ * converted back to shared, as the page is no longer usable due to RMP
+ * protections, and it's infeasible for the guest to continue on.
  */
-static int snp_page_reclaim(u64 pfn)
+static int snp_page_reclaim(struct kvm *kvm, u64 pfn)
 {
        struct sev_data_snp_page_reclaim data = {0};
-       int err, rc;
+       int fw_err, rc;
 
        data.paddr = __sme_set(pfn << PAGE_SHIFT);
-       rc = sev_do_cmd(SEV_CMD_SNP_PAGE_RECLAIM, &data, &err);
-       if (WARN_ONCE(rc, "Failed to reclaim PFN %llx", pfn))
+       rc = sev_do_cmd(SEV_CMD_SNP_PAGE_RECLAIM, &data, &fw_err);
+       if (KVM_BUG(rc, kvm, "Failed to reclaim PFN %llx, rc %d fw_err %d", pfn, rc, fw_err)) {
                snp_leak_pages(pfn, 1);
+               return -EIO;
+       }
 
-       return rc;
-}
-
-/*
- * Transition a page to hypervisor-owned/shared state in the RMP table. This
- * should not fail under normal conditions, but leak the page should that
- * happen since it will no longer be usable by the host due to RMP protections.
- */
-static int host_rmp_make_shared(u64 pfn, enum pg_level level)
-{
-       int rc;
-
-       rc = rmp_make_shared(pfn, level);
-       if (WARN_ON_ONCE(rc))
-               snp_leak_pages(pfn, page_level_size(level) >> PAGE_SHIFT);
+       if (kvm_rmp_make_shared(kvm, pfn, PG_LEVEL_4K))
+               return -EIO;
 
        return rc;
 }
@@ -2244,7 +2252,7 @@ fw_err:
         * information to provide information on which CPUID leaves/fields
         * failed CPUID validation.
         */
-       if (!snp_page_reclaim(pfn + i) && !host_rmp_make_shared(pfn + i, PG_LEVEL_4K) &&
+       if (!snp_page_reclaim(kvm, pfn + i) &&
            sev_populate_args->type == KVM_SEV_SNP_PAGE_TYPE_CPUID &&
            sev_populate_args->fw_error == SEV_RET_INVALID_PARAM) {
                void *vaddr = kmap_local_pfn(pfn + i);
@@ -2262,7 +2270,7 @@ err:
        pr_debug("%s: exiting with error ret %d (fw_error %d), restoring %d gmem PFNs to shared.\n",
                 __func__, ret, sev_populate_args->fw_error, n_private);
        for (i = 0; i < n_private; i++)
-               host_rmp_make_shared(pfn + i, PG_LEVEL_4K);
+               kvm_rmp_make_shared(kvm, pfn + i, PG_LEVEL_4K);
 
        return ret;
 }
@@ -2380,8 +2388,7 @@ static int snp_launch_update_vmsa(struct kvm *kvm, struct kvm_sev_cmd *argp)
                ret = __sev_issue_cmd(argp->sev_fd, SEV_CMD_SNP_LAUNCH_UPDATE,
                                      &data, &argp->error);
                if (ret) {
-                       if (!snp_page_reclaim(pfn))
-                               host_rmp_make_shared(pfn, PG_LEVEL_4K);
+                       snp_page_reclaim(kvm, pfn);
 
                        return ret;
                }
@@ -3069,7 +3076,7 @@ void sev_free_vcpu(struct kvm_vcpu *vcpu)
        if (sev_snp_guest(vcpu->kvm)) {
                u64 pfn = __pa(svm->sev_es.vmsa) >> PAGE_SHIFT;
 
-               if (host_rmp_make_shared(pfn, PG_LEVEL_4K))
+               if (kvm_rmp_make_shared(vcpu->kvm, pfn, PG_LEVEL_4K))
                        goto skip_vmsa_free;
        }