linux-kernel - [Patch v2 3/4] KVM:SVM: Pin sev_launch_update_data() pages via sev_get

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20200819151742.7892-4-Eric.VanTassell@amd.com>
Date:   Wed, 19 Aug 2020 10:17:41 -0500
From:   eric van tassell <Eric.VanTassell@....com>
To:     kvm@...r.kernel.org
Cc:     linux-kernel@...r.kernel.org, bp@...en8.de, hpa@...or.com,
        mingo@...hat.com, jmattson@...gle.com, joro@...tes.org,
        pbonzini@...hat.com, sean.j.christopherson@...el.com,
        tglx@...utronix.de, vkuznets@...hat.com, wanpengli@...cent.com,
        x86@...nel.org, rientjes@...gle.com, junaids@...gle.com,
        evantass@....com
Subject: [Patch v2 3/4] KVM:SVM: Pin sev_launch_update_data() pages via sev_get_page()

Add 2 small infrastructure functions here which to enable pinning the SEV
guest pages used for sev_launch_update_data() using sev_get_page().

Pin the memory for the data being passed to launch_update_data() because it
gets encrypted before the guest is first run and must not be moved which
would corrupt it.

Co-developed-by: Brijesh Singh <brijesh.singh@....com>
Signed-off-by: eric van tassell <Eric.VanTassell@....com>
---
 arch/x86/kvm/svm/sev.c | 57 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 57 insertions(+)

diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c
index 8d56d1afb33e..4a0157254fef 100644
--- a/arch/x86/kvm/svm/sev.c
+++ b/arch/x86/kvm/svm/sev.c
@@ -454,6 +454,37 @@ static int sev_get_page(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn)
 	return 0;
 }
 
+static struct kvm_memory_slot *hva_to_memslot(struct kvm *kvm,
+					      unsigned long hva)
+{
+	struct kvm_memslots *slots = kvm_memslots(kvm);
+	struct kvm_memory_slot *memslot;
+
+	kvm_for_each_memslot(memslot, slots) {
+		if (hva >= memslot->userspace_addr &&
+		    hva < memslot->userspace_addr +
+			      (memslot->npages << PAGE_SHIFT))
+			return memslot;
+	}
+
+	return NULL;
+}
+
+static bool hva_to_gfn(struct kvm *kvm, unsigned long hva, gfn_t *gfn)
+{
+	struct kvm_memory_slot *memslot;
+	gpa_t gpa_offset;
+
+	memslot = hva_to_memslot(kvm, hva);
+	if (!memslot)
+		return false;
+
+	gpa_offset = hva - memslot->userspace_addr;
+	*gfn = ((memslot->base_gfn << PAGE_SHIFT) + gpa_offset) >> PAGE_SHIFT;
+
+	return true;
+}
+
 static int sev_launch_update_data(struct kvm *kvm, struct kvm_sev_cmd *argp)
 {
 	unsigned long vaddr, vaddr_end, next_vaddr, npages, pages, size, i;
@@ -462,6 +493,7 @@ static int sev_launch_update_data(struct kvm *kvm, struct kvm_sev_cmd *argp)
 	struct sev_data_launch_update_data *data;
 	struct page **inpages;
 	int ret;
+	int srcu_idx;
 
 	if (!sev_guest(kvm))
 		return -ENOTTY;
@@ -484,6 +516,31 @@ static int sev_launch_update_data(struct kvm *kvm, struct kvm_sev_cmd *argp)
 		goto e_free;
 	}
 
+	/*
+	 * Increment the page ref count so that the pages do not get migrated or
+	 * moved after we are done from the LAUNCH_UPDATE_DATA.
+	 */
+
+	/* ensure hva_to_gfn translations remain valid */
+	srcu_idx = srcu_read_lock(&kvm->srcu);
+
+	for (i = 0; i < npages; i++) {
+		gfn_t gfn;
+
+		if (!hva_to_gfn(kvm, (vaddr + (i * PAGE_SIZE)) & PAGE_MASK, &gfn)) {
+			ret = -EFAULT;
+			break;
+		}
+
+		ret = sev_get_page(kvm, gfn, page_to_pfn(inpages[i]));
+		if (ret)
+			break;
+	}
+
+	srcu_read_unlock(&kvm->srcu, srcu_idx);
+	if (ret)
+		goto e_unpin;
+
 	/*
 	 * The LAUNCH_UPDATE command will perform in-place encryption of the
 	 * memory content (i.e it will write the same memory region with C=1).
-- 
2.17.1