linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: linux-kernel@vger.kernel.org, kvm@vger.kernel.org
Cc: pgonda@google.com, Sean Christopherson <seanjc@google.com>
Subject: [PATCH 09/12] KVM: SEV: Do COPY_ENC_CONTEXT_FROM with both VMs locked
Date: Mon, 22 Nov 2021 19:50:33 -0500	[thread overview]
Message-ID: <20211123005036.2954379-10-pbonzini@redhat.com> (raw)
In-Reply-To: <20211123005036.2954379-1-pbonzini@redhat.com>

Now that we have a facility to lock two VMs with deadlock
protection, use it for the creation of mirror VMs as well.  One of
COPY_ENC_CONTEXT_FROM(dst, src) and COPY_ENC_CONTEXT_FROM(src, dst)
would always fail, so the combination is nonsensical and it is okay to
return -EBUSY if it is attempted.

This sidesteps the question of what happens if a VM is
MOVE_ENC_CONTEXT_FROM'd at the same time as it is
COPY_ENC_CONTEXT_FROM'd: the locking prevents that from
happening.

Cc: Peter Gonda <pgonda@google.com>
Cc: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 arch/x86/kvm/svm/sev.c | 69 +++++++++++++++++-------------------------
 1 file changed, 27 insertions(+), 42 deletions(-)

diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c
index c1eb1c83401d..025d9731b66c 100644
--- a/arch/x86/kvm/svm/sev.c
+++ b/arch/x86/kvm/svm/sev.c
@@ -1548,6 +1548,9 @@ static int sev_lock_two_vms(struct kvm *dst_kvm, struct kvm *src_kvm)
 	struct kvm_sev_info *dst_sev = &to_kvm_svm(dst_kvm)->sev_info;
 	struct kvm_sev_info *src_sev = &to_kvm_svm(src_kvm)->sev_info;
 
+	if (dst_kvm == src_kvm)
+		return -EINVAL;
+
 	/*
 	 * Bail if these VMs are already involved in a migration to avoid
 	 * deadlock between two VMs trying to migrate to/from each other.
@@ -1952,77 +1955,59 @@ int svm_vm_copy_asid_from(struct kvm *kvm, unsigned int source_fd)
 {
 	struct file *source_kvm_file;
 	struct kvm *source_kvm;
-	struct kvm_sev_info source_sev, *mirror_sev;
+	struct kvm_sev_info *source_sev, *mirror_sev;
 	int ret;
 
 	source_kvm_file = fget(source_fd);
 	if (!file_is_kvm(source_kvm_file)) {
 		ret = -EBADF;
-		goto e_source_put;
+		goto e_source_fput;
 	}
 
 	source_kvm = source_kvm_file->private_data;
-	mutex_lock(&source_kvm->lock);
-
-	if (!sev_guest(source_kvm)) {
-		ret = -EINVAL;
-		goto e_source_unlock;
-	}
+	ret = sev_lock_two_vms(kvm, source_kvm);
+	if (ret)
+		goto e_source_fput;
 
-	/* Mirrors of mirrors should work, but let's not get silly */
-	if (is_mirroring_enc_context(source_kvm) || source_kvm == kvm) {
+	/*
+	 * Mirrors of mirrors should work, but let's not get silly.  Also
+	 * disallow out-of-band SEV/SEV-ES init if the target is already an
+	 * SEV guest, or if vCPUs have been created.  KVM relies on vCPUs being
+	 * created after SEV/SEV-ES initialization, e.g. to init intercepts.
+	 */
+	if (sev_guest(kvm) || !sev_guest(source_kvm) ||
+	    is_mirroring_enc_context(source_kvm) || kvm->created_vcpus) {
 		ret = -EINVAL;
-		goto e_source_unlock;
+		goto e_unlock;
 	}
 
-	memcpy(&source_sev, &to_kvm_svm(source_kvm)->sev_info,
-	       sizeof(source_sev));
-
 	/*
 	 * The mirror kvm holds an enc_context_owner ref so its asid can't
 	 * disappear until we're done with it
 	 */
+	source_sev = &to_kvm_svm(source_kvm)->sev_info;
 	kvm_get_kvm(source_kvm);
 
-	fput(source_kvm_file);
-	mutex_unlock(&source_kvm->lock);
-	mutex_lock(&kvm->lock);
-
-	/*
-	 * Disallow out-of-band SEV/SEV-ES init if the target is already an
-	 * SEV guest, or if vCPUs have been created.  KVM relies on vCPUs being
-	 * created after SEV/SEV-ES initialization, e.g. to init intercepts.
-	 */
-	if (sev_guest(kvm) || kvm->created_vcpus) {
-		ret = -EINVAL;
-		goto e_mirror_unlock;
-	}
-
 	/* Set enc_context_owner and copy its encryption context over */
 	mirror_sev = &to_kvm_svm(kvm)->sev_info;
 	mirror_sev->enc_context_owner = source_kvm;
 	mirror_sev->active = true;
-	mirror_sev->asid = source_sev.asid;
-	mirror_sev->fd = source_sev.fd;
-	mirror_sev->es_active = source_sev.es_active;
-	mirror_sev->handle = source_sev.handle;
+	mirror_sev->asid = source_sev->asid;
+	mirror_sev->fd = source_sev->fd;
+	mirror_sev->es_active = source_sev->es_active;
+	mirror_sev->handle = source_sev->handle;
 	INIT_LIST_HEAD(&mirror_sev->regions_list);
+	ret = 0;
+
 	/*
 	 * Do not copy ap_jump_table. Since the mirror does not share the same
 	 * KVM contexts as the original, and they may have different
 	 * memory-views.
 	 */
 
-	mutex_unlock(&kvm->lock);
-	return 0;
-
-e_mirror_unlock:
-	mutex_unlock(&kvm->lock);
-	kvm_put_kvm(source_kvm);
-	return ret;
-e_source_unlock:
-	mutex_unlock(&source_kvm->lock);
-e_source_put:
+e_unlock:
+	sev_unlock_two_vms(kvm, source_kvm);
+e_source_fput:
 	if (source_kvm_file)
 		fput(source_kvm_file);
 	return ret;
-- 
2.27.0



  parent reply	other threads:[~2021-11-23  0:51 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-23  0:50 [PATCH 00/12] Fixes for KVM_CAP_VM_MOVE/COPY_ENC_CONTEXT_FROM Paolo Bonzini
2021-11-23  0:50 ` [PATCH 01/12] selftests: fix check for circular KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM Paolo Bonzini
2021-12-01 15:52   ` Peter Gonda
2021-11-23  0:50 ` [PATCH 02/12] selftests: sev_migrate_tests: free all VMs Paolo Bonzini
2021-12-01 15:54   ` Peter Gonda
2021-11-23  0:50 ` [PATCH 03/12] KVM: SEV: expose KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM capability Paolo Bonzini
2021-11-29 22:28   ` Sean Christopherson
2021-12-01 15:55     ` Peter Gonda
2021-11-23  0:50 ` [PATCH 04/12] KVM: SEV: do not use list_replace_init on an empty list Paolo Bonzini
2021-11-29 22:27   ` Sean Christopherson
2021-11-23  0:50 ` [PATCH 05/12] KVM: SEV: cleanup locking for KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM Paolo Bonzini
2021-12-01 16:11   ` Peter Gonda
2021-11-23  0:50 ` [PATCH 06/12] KVM: SEV: initialize regions_list of a mirror VM Paolo Bonzini
2021-11-29 23:00   ` Sean Christopherson
2021-11-23  0:50 ` [PATCH 07/12] KVM: SEV: move mirror status to destination of KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM Paolo Bonzini
2021-11-29 23:02   ` Sean Christopherson
2021-11-23  0:50 ` [PATCH 08/12] selftests: sev_migrate_tests: add tests for KVM_CAP_VM_COPY_ENC_CONTEXT_FROM Paolo Bonzini
2021-12-01 18:09   ` Peter Gonda
2021-12-07 20:11     ` Peter Gonda
2021-11-23  0:50 ` Paolo Bonzini [this message]
2021-11-29 23:08   ` [PATCH 09/12] KVM: SEV: Do COPY_ENC_CONTEXT_FROM with both VMs locked Sean Christopherson
2021-11-23  0:50 ` [PATCH 10/12] KVM: SEV: Prohibit migration of a VM that has mirrors Paolo Bonzini
2021-11-29 22:54   ` Sean Christopherson
2021-12-01 18:17   ` Peter Gonda
2021-12-01 18:21     ` Paolo Bonzini
2021-11-23  0:50 ` [PATCH 11/12] KVM: SEV: do not take kvm->lock when destroying Paolo Bonzini
2021-11-29 22:31   ` Sean Christopherson
2021-11-23  0:50 ` [PATCH 12/12] KVM: SEV: accept signals in sev_lock_two_vms Paolo Bonzini

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20211123005036.2954379-10-pbonzini@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pgonda@google.com \
    --cc=seanjc@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).