[PATCH v2 04/20] KVM: TDX: Kick off vCPUs when SEAMCALL is busy during TD page removal

[Binbin Wu <binbin.wu@linux.intel.com>]

From: Yan Zhao <yan.y.zhao@intel.com>

Kick off all vCPUs and prevent tdh_vp_enter() from executing whenever
tdh_mem_range_block()/tdh_mem_track()/tdh_mem_page_remove() encounters
contention, since the page removal path does not expect error and is less
sensitive to the performance penalty caused by kicking off vCPUs.

Although KVM has protected SEPT zap-related SEAMCALLs with kvm->mmu_lock,
KVM may still encounter TDX_OPERAND_BUSY due to the contention in the TDX
module.
- tdh_mem_track() may contend with tdh_vp_enter().
- tdh_mem_range_block()/tdh_mem_page_remove() may contend with
  tdh_vp_enter() and TDCALLs.

Resources     SHARED users      EXCLUSIVE users
------------------------------------------------------------
TDCS epoch    tdh_vp_enter      tdh_mem_track
------------------------------------------------------------
SEPT tree  tdh_mem_page_remove  tdh_vp_enter (0-step mitigation)
                                tdh_mem_range_block
------------------------------------------------------------
SEPT entry                      tdh_mem_range_block (Host lock)
                                tdh_mem_page_remove (Host lock)
                                tdg_mem_page_accept (Guest lock)
                                tdg_mem_page_attr_rd (Guest lock)
                                tdg_mem_page_attr_wr (Guest lock)

Use a TDX specific per-VM flag wait_for_sept_zap along with
KVM_REQ_OUTSIDE_GUEST_MODE to kick off vCPUs and prevent them from entering
TD, thereby avoiding the potential contention. Apply the kick-off and no
vCPU entering only after each SEAMCALL busy error to minimize the window of
no TD entry, as the contention due to 0-step mitigation or TDCALLs is
expected to be rare.

Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
---
Rebased to use helper tdx_operand_busy().
---
 arch/x86/kvm/vmx/tdx.c | 63 ++++++++++++++++++++++++++++++++++++------
 arch/x86/kvm/vmx/tdx.h |  7 +++++
 2 files changed, 61 insertions(+), 9 deletions(-)

diff --git a/arch/x86/kvm/vmx/tdx.c b/arch/x86/kvm/vmx/tdx.c
index 6fa6f7e13e15..25ae2c2826d0 100644
--- a/arch/x86/kvm/vmx/tdx.c
+++ b/arch/x86/kvm/vmx/tdx.c
@@ -294,6 +294,26 @@ static void tdx_clear_page(struct page *page)
 	__mb();
 }
 
+static void tdx_no_vcpus_enter_start(struct kvm *kvm)
+{
+	struct kvm_tdx *kvm_tdx = to_kvm_tdx(kvm);
+
+	lockdep_assert_held_write(&kvm->mmu_lock);
+
+	WRITE_ONCE(kvm_tdx->wait_for_sept_zap, true);
+
+	kvm_make_all_cpus_request(kvm, KVM_REQ_OUTSIDE_GUEST_MODE);
+}
+
+static void tdx_no_vcpus_enter_stop(struct kvm *kvm)
+{
+	struct kvm_tdx *kvm_tdx = to_kvm_tdx(kvm);
+
+	lockdep_assert_held_write(&kvm->mmu_lock);
+
+	WRITE_ONCE(kvm_tdx->wait_for_sept_zap, false);
+}
+
 /* TDH.PHYMEM.PAGE.RECLAIM is allowed only when destroying the TD. */
 static int __tdx_reclaim_page(struct page *page)
 {
@@ -953,6 +973,14 @@ fastpath_t tdx_vcpu_run(struct kvm_vcpu *vcpu, bool force_immediate_exit)
 		return EXIT_FASTPATH_NONE;
 	}
 
+	/*
+	 * Wait until retry of SEPT-zap-related SEAMCALL completes before
+	 * allowing vCPU entry to avoid contention with tdh_vp_enter() and
+	 * TDCALLs.
+	 */
+	if (unlikely(READ_ONCE(to_kvm_tdx(vcpu->kvm)->wait_for_sept_zap)))
+		return EXIT_FASTPATH_EXIT_HANDLED;
+
 	trace_kvm_entry(vcpu, force_immediate_exit);
 
 	if (pi_test_on(&vt->pi_desc)) {
@@ -1467,15 +1495,24 @@ static int tdx_sept_drop_private_spte(struct kvm *kvm, gfn_t gfn,
 	if (KVM_BUG_ON(!is_hkid_assigned(kvm_tdx), kvm))
 		return -EINVAL;
 
-	do {
+	/*
+	 * When zapping private page, write lock is held. So no race condition
+	 * with other vcpu sept operation.
+	 * Race with TDH.VP.ENTER due to (0-step mitigation) and Guest TDCALLs.
+	 */
+	err = tdh_mem_page_remove(&kvm_tdx->td, gpa, tdx_level, &entry,
+				  &level_state);
+
+	if (unlikely(tdx_operand_busy(err))) {
 		/*
-		 * When zapping private page, write lock is held. So no race
-		 * condition with other vcpu sept operation.  Race only with
-		 * TDH.VP.ENTER.
+		 * The second retry is expected to succeed after kicking off all
+		 * other vCPUs and prevent them from invoking TDH.VP.ENTER.
 		 */
+		tdx_no_vcpus_enter_start(kvm);
 		err = tdh_mem_page_remove(&kvm_tdx->td, gpa, tdx_level, &entry,
 					  &level_state);
-	} while (unlikely(tdx_operand_busy(err)));
+		tdx_no_vcpus_enter_stop(kvm);
+	}
 
 	if (KVM_BUG_ON(err, kvm)) {
 		pr_tdx_error_2(TDH_MEM_PAGE_REMOVE, err, entry, level_state);
@@ -1559,9 +1596,13 @@ static int tdx_sept_zap_private_spte(struct kvm *kvm, gfn_t gfn,
 	WARN_ON_ONCE(level != PG_LEVEL_4K);
 
 	err = tdh_mem_range_block(&kvm_tdx->td, gpa, tdx_level, &entry, &level_state);
-	if (unlikely(tdx_operand_busy(err)))
-		return -EBUSY;
 
+	if (unlikely(tdx_operand_busy(err))) {
+		/* After no vCPUs enter, the second retry is expected to succeed */
+		tdx_no_vcpus_enter_start(kvm);
+		err = tdh_mem_range_block(&kvm_tdx->td, gpa, tdx_level, &entry, &level_state);
+		tdx_no_vcpus_enter_stop(kvm);
+	}
 	if (tdx_is_sept_zap_err_due_to_premap(kvm_tdx, err, entry, level) &&
 	    !KVM_BUG_ON(!atomic64_read(&kvm_tdx->nr_premapped), kvm)) {
 		atomic64_dec(&kvm_tdx->nr_premapped);
@@ -1611,9 +1652,13 @@ static void tdx_track(struct kvm *kvm)
 
 	lockdep_assert_held_write(&kvm->mmu_lock);
 
-	do {
+	err = tdh_mem_track(&kvm_tdx->td);
+	if (unlikely(tdx_operand_busy(err))) {
+		/* After no vCPUs enter, the second retry is expected to succeed */
+		tdx_no_vcpus_enter_start(kvm);
 		err = tdh_mem_track(&kvm_tdx->td);
-	} while (unlikely(tdx_operand_busy(err)));
+		tdx_no_vcpus_enter_stop(kvm);
+	}
 
 	if (KVM_BUG_ON(err, kvm))
 		pr_tdx_error(TDH_MEM_TRACK, err);
diff --git a/arch/x86/kvm/vmx/tdx.h b/arch/x86/kvm/vmx/tdx.h
index 4e1e71d5d8e8..c6bafac31d4d 100644
--- a/arch/x86/kvm/vmx/tdx.h
+++ b/arch/x86/kvm/vmx/tdx.h
@@ -36,6 +36,13 @@ struct kvm_tdx {
 
 	/* For KVM_TDX_INIT_MEM_REGION. */
 	atomic64_t nr_premapped;
+
+	/*
+	 * Prevent vCPUs from TD entry to ensure SEPT zap related SEAMCALLs do
+	 * not contend with tdh_vp_enter() and TDCALLs.
+	 * Set/unset is protected with kvm->mmu_lock.
+	 */
+	bool wait_for_sept_zap;
 };
 
 /* TDX module vCPU states */
-- 
2.46.0