diff --git a/arch/arm64/kvm/fpsimd.c b/arch/arm64/kvm/fpsimd.c
index f647241..3cbb999 100644
--- a/arch/arm64/kvm/fpsimd.c
+++ b/arch/arm64/kvm/fpsimd.c
@@ -136,36 +136,6 @@ void kvm_arch_vcpu_put_fp(struct kvm_vcpu *vcpu)
 	local_irq_save(flags);
 
 	if (guest_owns_fp_regs()) {
-		if (vcpu_has_sve(vcpu)) {
-			u64 zcr = read_sysreg_el1(SYS_ZCR);
-
-			/*
-			 * If the vCPU is in the hyp context then ZCR_EL1 is
-			 * loaded with its vEL2 counterpart.
-			 */
-			__vcpu_sys_reg(vcpu, vcpu_sve_zcr_elx(vcpu)) = zcr;
-
-			/*
-			 * Restore the VL that was saved when bound to the CPU,
-			 * which is the maximum VL for the guest. Because the
-			 * layout of the data when saving the sve state depends
-			 * on the VL, we need to use a consistent (i.e., the
-			 * maximum) VL.
-			 * Note that this means that at guest exit ZCR_EL1 is
-			 * not necessarily the same as on guest entry.
-			 *
-			 * ZCR_EL2 holds the guest hypervisor's VL when running
-			 * a nested guest, which could be smaller than the
-			 * max for the vCPU. Similar to above, we first need to
-			 * switch to a VL consistent with the layout of the
-			 * vCPU's SVE state. KVM support for NV implies VHE, so
-			 * using the ZCR_EL1 alias is safe.
-			 */
-			if (!has_vhe() || (vcpu_has_nv(vcpu) && !is_hyp_ctxt(vcpu)))
-				sve_cond_update_zcr_vq(vcpu_sve_max_vq(vcpu) - 1,
-						       SYS_ZCR_EL1);
-		}
-
 		/*
 		 * Flush (save and invalidate) the fpsimd/sve state so that if
 		 * the host tries to use fpsimd/sve, it's not using stale data
diff --git a/arch/arm64/kvm/hyp/entry.S b/arch/arm64/kvm/hyp/entry.S
index 4433a23..9f4e8d6 100644
--- a/arch/arm64/kvm/hyp/entry.S
+++ b/arch/arm64/kvm/hyp/entry.S
@@ -44,6 +44,11 @@
 alternative_else_nop_endif
 	mrs	x1, isr_el1
 	cbz	x1,  1f
+
+	// Ensure that __guest_enter() always provides a context
+	// synchronization event so that callers don't need ISBs for anything
+	// that would usually be synchonized by the ERET.
+	isb
 	mov	x0, #ARM_EXCEPTION_IRQ
 	ret
 
diff --git a/arch/arm64/kvm/hyp/include/hyp/switch.h b/arch/arm64/kvm/hyp/include/hyp/switch.h
index c1ab314..cc9cb639 100644
--- a/arch/arm64/kvm/hyp/include/hyp/switch.h
+++ b/arch/arm64/kvm/hyp/include/hyp/switch.h
@@ -344,6 +344,65 @@ static inline void __hyp_sve_save_host(void)
 			 true);
 }
 
+static inline void fpsimd_lazy_switch_to_guest(struct kvm_vcpu *vcpu)
+{
+	u64 zcr_el1, zcr_el2;
+
+	if (!guest_owns_fp_regs())
+		return;
+
+	if (vcpu_has_sve(vcpu)) {
+		/* A guest hypervisor may restrict the effective max VL. */
+		if (vcpu_has_nv(vcpu) && !is_hyp_ctxt(vcpu))
+			zcr_el2 = __vcpu_sys_reg(vcpu, ZCR_EL2);
+		else
+			zcr_el2 = vcpu_sve_max_vq(vcpu) - 1;
+
+		write_sysreg_el2(zcr_el2, SYS_ZCR);
+
+		zcr_el1 = __vcpu_sys_reg(vcpu, vcpu_sve_zcr_elx(vcpu));
+		write_sysreg_el1(zcr_el1, SYS_ZCR);
+	}
+}
+
+static inline void fpsimd_lazy_switch_to_host(struct kvm_vcpu *vcpu)
+{
+	u64 zcr_el1, zcr_el2;
+
+	if (!guest_owns_fp_regs())
+		return;
+
+	/*
+	 * When the guest owns the FP regs, we know that guest+hyp traps for
+	 * any FPSIMD/SVE/SME features exposed to the guest have been disabled
+	 * by either fpsimd_lazy_switch_to_guest() or kvm_hyp_handle_fpsimd()
+	 * prior to __guest_entry(). As __guest_entry() guarantees a context
+	 * synchronization event, we don't need an ISB here to avoid taking
+	 * traps for anything that was exposed to the guest.
+	 */
+	if (vcpu_has_sve(vcpu)) {
+		zcr_el1 = read_sysreg_el1(SYS_ZCR);
+		__vcpu_sys_reg(vcpu, vcpu_sve_zcr_elx(vcpu)) = zcr_el1;
+
+		/*
+		 * The guest's state is always saved using the guest's max VL.
+		 * Ensure that the host has the guest's max VL active such that
+		 * the host can save the guest's state lazily, but don't
+		 * artificially restrict the host to the guest's max VL.
+		 */
+		if (has_vhe()) {
+			zcr_el2 = vcpu_sve_max_vq(vcpu) - 1;
+			write_sysreg_el2(zcr_el2, SYS_ZCR);
+		} else {
+			zcr_el2 = sve_vq_from_vl(kvm_host_sve_max_vl) - 1;
+			write_sysreg_el2(zcr_el2, SYS_ZCR);
+
+			zcr_el1 = vcpu_sve_max_vq(vcpu) - 1;
+			write_sysreg_el1(zcr_el1, SYS_ZCR);
+		}
+	}
+}
+
 static void kvm_hyp_save_fpsimd_host(struct kvm_vcpu *vcpu)
 {
 	/*
diff --git a/arch/arm64/kvm/hyp/nvhe/hyp-main.c b/arch/arm64/kvm/hyp/nvhe/hyp-main.c
index 4e757a7..75f7e38 100644
--- a/arch/arm64/kvm/hyp/nvhe/hyp-main.c
+++ b/arch/arm64/kvm/hyp/nvhe/hyp-main.c
@@ -5,6 +5,7 @@
  */
 
 #include <hyp/adjust_pc.h>
+#include <hyp/switch.h>
 
 #include <asm/pgtable-types.h>
 #include <asm/kvm_asm.h>
@@ -177,7 +178,9 @@ static void handle___kvm_vcpu_run(struct kvm_cpu_context *host_ctxt)
 		pkvm_put_hyp_vcpu(hyp_vcpu);
 	} else {
 		/* The host is fully trusted, run its vCPU directly. */
+		fpsimd_lazy_switch_to_guest(host_vcpu);
 		ret = __kvm_vcpu_run(host_vcpu);
+		fpsimd_lazy_switch_to_host(host_vcpu);
 	}
 
 out:
@@ -486,12 +489,6 @@ void handle_trap(struct kvm_cpu_context *host_ctxt)
 	case ESR_ELx_EC_SMC64:
 		handle_host_smc(host_ctxt);
 		break;
-	case ESR_ELx_EC_SVE:
-		cpacr_clear_set(0, CPACR_ELx_ZEN);
-		isb();
-		sve_cond_update_zcr_vq(sve_vq_from_vl(kvm_host_sve_max_vl) - 1,
-				       SYS_ZCR_EL2);
-		break;
 	case ESR_ELx_EC_IABT_LOW:
 	case ESR_ELx_EC_DABT_LOW:
 		handle_host_mem_abort(host_ctxt);
diff --git a/arch/arm64/kvm/hyp/nvhe/switch.c b/arch/arm64/kvm/hyp/nvhe/switch.c
index ee74006..a1245fa 100644
--- a/arch/arm64/kvm/hyp/nvhe/switch.c
+++ b/arch/arm64/kvm/hyp/nvhe/switch.c
@@ -40,6 +40,9 @@ static void __activate_cptr_traps(struct kvm_vcpu *vcpu)
 {
 	u64 val = CPTR_EL2_TAM;	/* Same bit irrespective of E2H */
 
+	if (!guest_owns_fp_regs())
+		__activate_traps_fpsimd32(vcpu);
+
 	if (has_hvhe()) {
 		val |= CPACR_ELx_TTA;
 
@@ -48,6 +51,8 @@ static void __activate_cptr_traps(struct kvm_vcpu *vcpu)
 			if (vcpu_has_sve(vcpu))
 				val |= CPACR_ELx_ZEN;
 		}
+
+		write_sysreg(val, cpacr_el1);
 	} else {
 		val |= CPTR_EL2_TTA | CPTR_NVHE_EL2_RES1;
 
@@ -62,12 +67,32 @@ static void __activate_cptr_traps(struct kvm_vcpu *vcpu)
 
 		if (!guest_owns_fp_regs())
 			val |= CPTR_EL2_TFP;
+
+		write_sysreg(val, cptr_el2);
 	}
+}
 
-	if (!guest_owns_fp_regs())
-		__activate_traps_fpsimd32(vcpu);
+static void __deactivate_cptr_traps(struct kvm_vcpu *vcpu)
+{
+	if (has_hvhe()) {
+		u64 val = CPACR_ELx_FPEN;
 
-	kvm_write_cptr_el2(val);
+		if (cpus_have_final_cap(ARM64_SVE))
+			val |= CPACR_ELx_ZEN;
+		if (cpus_have_final_cap(ARM64_SME))
+			val |= CPACR_ELx_SMEN;
+
+		write_sysreg(val, cpacr_el1);
+	} else {
+		u64 val = CPTR_NVHE_EL2_RES1;
+
+		if (!cpus_have_final_cap(ARM64_SVE))
+			val |= CPTR_EL2_TZ;
+		if (!cpus_have_final_cap(ARM64_SME))
+			val |= CPTR_EL2_TSM;
+
+		write_sysreg(val, cptr_el2);
+	}
 }
 
 static void __activate_traps(struct kvm_vcpu *vcpu)
@@ -120,7 +145,7 @@ static void __deactivate_traps(struct kvm_vcpu *vcpu)
 
 	write_sysreg(this_cpu_ptr(&kvm_init_params)->hcr_el2, hcr_el2);
 
-	kvm_reset_cptr_el2(vcpu);
+	__deactivate_cptr_traps(vcpu);
 	write_sysreg(__kvm_hyp_host_vector, vbar_el2);
 }
 
diff --git a/arch/arm64/kvm/hyp/vhe/switch.c b/arch/arm64/kvm/hyp/vhe/switch.c
index 46c1f5c..496abfd 100644
--- a/arch/arm64/kvm/hyp/vhe/switch.c
+++ b/arch/arm64/kvm/hyp/vhe/switch.c
@@ -462,6 +462,8 @@ static int __kvm_vcpu_run_vhe(struct kvm_vcpu *vcpu)
 
 	sysreg_save_host_state_vhe(host_ctxt);
 
+	fpsimd_lazy_switch_to_guest(vcpu);
+
 	/*
 	 * Note that ARM erratum 1165522 requires us to configure both stage 1
 	 * and stage 2 translation for the guest context before we clear
@@ -486,6 +488,8 @@ static int __kvm_vcpu_run_vhe(struct kvm_vcpu *vcpu)
 
 	__deactivate_traps(vcpu);
 
+	fpsimd_lazy_switch_to_host(vcpu);
+
 	sysreg_restore_host_state_vhe(host_ctxt);
 
 	if (guest_owns_fp_regs())