[PATCH v3 24/25] KVM: TDX: Guard VM state transitions with "all" the locks

[Sean Christopherson]

Acquire kvm->lock, kvm->slots_lock, and all vcpu->mutex locks when
servicing ioctls that (a) transition the TD to a new state, i.e. when
doing INIT or FINALIZE or (b) are only valid if the TD is in a specific
state, i.e. when initializing a vCPU or memory region.  Acquiring "all"
the locks fixes several KVM_BUG_ON() situations where a SEAMCALL can fail
due to racing actions, e.g. if tdh_vp_create() contends with either
tdh_mr_extend() or tdh_mr_finalize().

For all intents and purposes, the paths in question are fully serialized,
i.e. there's no reason to try and allow anything remotely interesting to
happen.  Smack 'em with a big hammer instead of trying to be "nice".

Acquire kvm->lock to prevent VM-wide things from happening, slots_lock to
prevent kvm_mmu_zap_all_fast(), and _all_ vCPU mutexes to prevent vCPUs
from interefering.  Use the recently-renamed kvm_arch_vcpu_unlocked_ioctl()
to service the vCPU-scoped ioctls to avoid a lock inversion problem, e.g.
due to taking vcpu->mutex outside kvm->lock.

See also commit ecf371f8b02d ("KVM: SVM: Reject SEV{-ES} intra host
migration if vCPU creation is in-flight"), which fixed a similar bug with
SEV intra-host migration where an in-flight vCPU creation could race with
a VM-wide state transition.

Define a fancy new CLASS to handle the lock+check => unlock logic with
guard()-like syntax:

        CLASS(tdx_vm_state_guard, guard)(kvm);
        if (IS_ERR(guard))
                return PTR_ERR(guard);

to simplify juggling the many locks.

Note!  Take kvm->slots_lock *after* all vcpu->mutex locks, as per KVM's
soon-to-be-documented lock ordering rules[1].

Link: https://lore.kernel.org/all/20251016235538.171962-1-seanjc@google.com [1]
Reported-by: Yan Zhao <yan.y.zhao at intel.com>
Closes: https://lore.kernel.org/all/aLFiPq1smdzN3Ary@yzhao56-desk.sh.intel.com
Signed-off-by: Sean Christopherson <seanjc at google.com>
---
 arch/x86/kvm/vmx/tdx.c | 63 +++++++++++++++++++++++++++++++++++-------
 1 file changed, 53 insertions(+), 10 deletions(-)

diff --git a/arch/x86/kvm/vmx/tdx.c b/arch/x86/kvm/vmx/tdx.c
index 84b5fe654c99..d6541b08423f 100644
--- a/arch/x86/kvm/vmx/tdx.c
+++ b/arch/x86/kvm/vmx/tdx.c
@@ -2632,6 +2632,46 @@ static int tdx_read_cpuid(struct kvm_vcpu *vcpu, u32 leaf, u32 sub_leaf,
 	return -EIO;
 }
 
+typedef void *tdx_vm_state_guard_t;
+
+static tdx_vm_state_guard_t tdx_acquire_vm_state_locks(struct kvm *kvm)
+{
+	int r;
+
+	mutex_lock(&kvm->lock);
+
+	if (kvm->created_vcpus != atomic_read(&kvm->online_vcpus)) {
+		r = -EBUSY;
+		goto out_err;
+	}
+
+	r = kvm_lock_all_vcpus(kvm);
+	if (r)
+		goto out_err;
+
+	/*
+	 * Note the unintuitive ordering!  vcpu->mutex must be taken outside
+	 * kvm->slots_lock!
+	 */
+	mutex_lock(&kvm->slots_lock);
+	return kvm;
+
+out_err:
+	mutex_unlock(&kvm->lock);
+	return ERR_PTR(r);
+}
+
+static void tdx_release_vm_state_locks(struct kvm *kvm)
+{
+	mutex_unlock(&kvm->slots_lock);
+	kvm_unlock_all_vcpus(kvm);
+	mutex_unlock(&kvm->lock);
+}
+
+DEFINE_CLASS(tdx_vm_state_guard, tdx_vm_state_guard_t,
+	     if (!IS_ERR(_T)) tdx_release_vm_state_locks(_T),
+	     tdx_acquire_vm_state_locks(kvm), struct kvm *kvm);
+
 static int tdx_td_init(struct kvm *kvm, struct kvm_tdx_cmd *cmd)
 {
 	struct kvm_tdx_init_vm __user *user_data = u64_to_user_ptr(cmd->data);
@@ -2644,6 +2684,10 @@ static int tdx_td_init(struct kvm *kvm, struct kvm_tdx_cmd *cmd)
 	BUILD_BUG_ON(sizeof(*init_vm) != 256 + sizeof_field(struct kvm_tdx_init_vm, cpuid));
 	BUILD_BUG_ON(sizeof(struct td_params) != 1024);
 
+	CLASS(tdx_vm_state_guard, guard)(kvm);
+	if (IS_ERR(guard))
+		return PTR_ERR(guard);
+
 	if (kvm_tdx->state != TD_STATE_UNINITIALIZED)
 		return -EINVAL;
 
@@ -2743,7 +2787,9 @@ static int tdx_td_finalize(struct kvm *kvm, struct kvm_tdx_cmd *cmd)
 {
 	struct kvm_tdx *kvm_tdx = to_kvm_tdx(kvm);
 
-	guard(mutex)(&kvm->slots_lock);
+	CLASS(tdx_vm_state_guard, guard)(kvm);
+	if (IS_ERR(guard))
+		return PTR_ERR(guard);
 
 	if (!is_hkid_assigned(kvm_tdx) || kvm_tdx->state == TD_STATE_RUNNABLE)
 		return -EINVAL;
@@ -2781,8 +2827,6 @@ int tdx_vm_ioctl(struct kvm *kvm, void __user *argp)
 	if (r)
 		return r;
 
-	guard(mutex)(&kvm->lock);
-
 	switch (tdx_cmd.id) {
 	case KVM_TDX_CAPABILITIES:
 		r = tdx_get_capabilities(&tdx_cmd);
@@ -3090,8 +3134,6 @@ static int tdx_vcpu_init_mem_region(struct kvm_vcpu *vcpu, struct kvm_tdx_cmd *c
 	if (tdx->state != VCPU_TD_STATE_INITIALIZED)
 		return -EINVAL;
 
-	guard(mutex)(&kvm->slots_lock);
-
 	/* Once TD is finalized, the initial guest memory is fixed. */
 	if (kvm_tdx->state == TD_STATE_RUNNABLE)
 		return -EINVAL;
@@ -3147,7 +3189,8 @@ static int tdx_vcpu_init_mem_region(struct kvm_vcpu *vcpu, struct kvm_tdx_cmd *c
 
 int tdx_vcpu_unlocked_ioctl(struct kvm_vcpu *vcpu, void __user *argp)
 {
-	struct kvm_tdx *kvm_tdx = to_kvm_tdx(vcpu->kvm);
+	struct kvm *kvm = vcpu->kvm;
+	struct kvm_tdx *kvm_tdx = to_kvm_tdx(kvm);
 	struct kvm_tdx_cmd cmd;
 	int r;
 
@@ -3155,12 +3198,13 @@ int tdx_vcpu_unlocked_ioctl(struct kvm_vcpu *vcpu, void __user *argp)
 	if (r)
 		return r;
 
+	CLASS(tdx_vm_state_guard, guard)(kvm);
+	if (IS_ERR(guard))
+		return PTR_ERR(guard);
+
 	if (!is_hkid_assigned(kvm_tdx) || kvm_tdx->state == TD_STATE_RUNNABLE)
 		return -EINVAL;
 
-	if (mutex_lock_killable(&vcpu->mutex))
-		return -EINTR;
-
 	vcpu_load(vcpu);
 
 	switch (cmd.id) {
@@ -3177,7 +3221,6 @@ int tdx_vcpu_unlocked_ioctl(struct kvm_vcpu *vcpu, void __user *argp)
 
 	vcpu_put(vcpu);
 
-	mutex_unlock(&vcpu->mutex);
 	return r;
 }
 
-- 
2.51.0.858.gf9c4a03a3a-goog