Merge tag 'kvm-ppc-next-4.15-2' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc

Second PPC KVM update for 4.15

This merges in my kvm-ppc-fixes branch to resolve the conflicts
between the fixes that have been applied there and the changes
made in my patch series to allow HPT guests to run on a radix
host on POWER9.  It also resolves another conflict in the code
for the KVM_CAP_PPC_HTM capability.

arch/powerpc/kvm/book3s_64_mmu_hv.c

@@ -651,6 +651,16 @@ int kvmppc_book3s_hv_page_fault(struct kvm_run *run, struct kvm_vcpu *vcpu,
 		hnow_v = hpte_new_to_old_v(hnow_v, hnow_r);
 		hnow_r = hpte_new_to_old_r(hnow_r);
 	}
+
+	/*
+	 * If the HPT is being resized, don't update the HPTE,
+	 * instead let the guest retry after the resize operation is complete.
+	 * The synchronization for mmu_ready test vs. set is provided
+	 * by the HPTE lock.
+	 */
+	if (!kvm->arch.mmu_ready)
+		goto out_unlock;
+
 	if ((hnow_v & ~HPTE_V_HVLOCK) != hpte[0] || hnow_r != hpte[1] ||
 	    rev->guest_rpte != hpte[2])
 		/* HPTE has been changed under us; let the guest retry */

arch/powerpc/kvm/book3s_64_vio.c

@@ -478,28 +478,30 @@ long kvmppc_h_put_tce(struct kvm_vcpu *vcpu, unsigned long liobn,
 		return ret;
 
 	dir = iommu_tce_direction(tce);
+
+	idx = srcu_read_lock(&vcpu->kvm->srcu);
+
 	if ((dir != DMA_NONE) && kvmppc_gpa_to_ua(vcpu->kvm,
-			tce & ~(TCE_PCI_READ | TCE_PCI_WRITE), &ua, NULL))
-		return H_PARAMETER;
+			tce & ~(TCE_PCI_READ | TCE_PCI_WRITE), &ua, NULL)) {
+		ret = H_PARAMETER;
+		goto unlock_exit;
+	}
 
 	entry = ioba >> stt->page_shift;
 
 	list_for_each_entry_lockless(stit, &stt->iommu_tables, next) {
-		if (dir == DMA_NONE) {
+		if (dir == DMA_NONE)
 			ret = kvmppc_tce_iommu_unmap(vcpu->kvm,
 					stit->tbl, entry);
-		} else {
-			idx = srcu_read_lock(&vcpu->kvm->srcu);
+		else
 			ret = kvmppc_tce_iommu_map(vcpu->kvm, stit->tbl,
 					entry, ua, dir);
-			srcu_read_unlock(&vcpu->kvm->srcu, idx);
-		}
 
 		if (ret == H_SUCCESS)
 			continue;
 
 		if (ret == H_TOO_HARD)
-			return ret;
+			goto unlock_exit;
 
 		WARN_ON_ONCE(1);
 		kvmppc_clear_tce(stit->tbl, entry);
@@ -507,7 +509,10 @@ long kvmppc_h_put_tce(struct kvm_vcpu *vcpu, unsigned long liobn,
 
 	kvmppc_tce_put(stt, entry, tce);
 
-	return H_SUCCESS;
+unlock_exit:
+	srcu_read_unlock(&vcpu->kvm->srcu, idx);
+
+	return ret;
 }
 EXPORT_SYMBOL_GPL(kvmppc_h_put_tce);
 

arch/powerpc/kvm/book3s_hv.c

@@ -2717,11 +2717,13 @@ static noinline void kvmppc_run_core(struct kvmppc_vcore *vc)
 	 * Hard-disable interrupts, and check resched flag and signals.
 	 * If we need to reschedule or deliver a signal, clean up
 	 * and return without going into the guest(s).
+	 * If the mmu_ready flag has been cleared, don't go into the
+	 * guest because that means a HPT resize operation is in progress.
 	 */
 	local_irq_disable();
 	hard_irq_disable();
 	if (lazy_irq_pending() || need_resched() ||
-	    recheck_signals(&core_info)) {
+	    recheck_signals(&core_info) || !vc->kvm->arch.mmu_ready) {
 		local_irq_enable();
 		vc->vcore_state = VCORE_INACTIVE;
 		/* Unlock all except the primary vcore */
@@ -3118,9 +3120,28 @@ static void kvmppc_vcore_blocked(struct kvmppc_vcore *vc)
 	trace_kvmppc_vcore_wakeup(do_sleep, block_ns);
 }
 
+static int kvmhv_setup_mmu(struct kvm_vcpu *vcpu)
+{
+	int r = 0;
+	struct kvm *kvm = vcpu->kvm;
+
+	mutex_lock(&kvm->lock);
+	if (!kvm->arch.mmu_ready) {
+		if (!kvm_is_radix(kvm))
+			r = kvmppc_hv_setup_htab_rma(vcpu);
+		if (!r) {
+			if (cpu_has_feature(CPU_FTR_ARCH_300))
+				kvmppc_setup_partition_table(kvm);
+			kvm->arch.mmu_ready = 1;
+		}
+	}
+	mutex_unlock(&kvm->lock);
+	return r;
+}
+
 static int kvmppc_run_vcpu(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
 {
-	int n_ceded, i;
+	int n_ceded, i, r;
 	struct kvmppc_vcore *vc;
 	struct kvm_vcpu *v;
 
@@ -3174,6 +3195,20 @@ static int kvmppc_run_vcpu(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
 
 	while (vcpu->arch.state == KVMPPC_VCPU_RUNNABLE &&
 	       !signal_pending(current)) {
+		/* See if the MMU is ready to go */
+		if (!vcpu->kvm->arch.mmu_ready) {
+			spin_unlock(&vc->lock);
+			r = kvmhv_setup_mmu(vcpu);
+			spin_lock(&vc->lock);
+			if (r) {
+				kvm_run->exit_reason = KVM_EXIT_FAIL_ENTRY;
+				kvm_run->fail_entry.
+					hardware_entry_failure_reason = 0;
+				vcpu->arch.ret = r;
+				break;
+			}
+		}
+
 		if (vc->vcore_state == VCORE_PREEMPT && vc->runner == NULL)
 			kvmppc_vcore_end_preempt(vc);
 
@@ -3293,24 +3328,6 @@ static int kvmppc_vcpu_run_hv(struct kvm_run *run, struct kvm_vcpu *vcpu)
 	/* Order vcpus_running vs. mmu_ready, see kvmppc_alloc_reset_hpt */
 	smp_mb();
 
-	/* On the first time here, set up MMU if necessary */
-	if (!vcpu->kvm->arch.mmu_ready) {
-		mutex_lock(&kvm->lock);
-		r = 0;
-		if (!kvm->arch.mmu_ready) {
-			if (!kvm_is_radix(vcpu->kvm))
-				r = kvmppc_hv_setup_htab_rma(vcpu);
-			if (!r) {
-				if (cpu_has_feature(CPU_FTR_ARCH_300))
-					kvmppc_setup_partition_table(kvm);
-				kvm->arch.mmu_ready = 1;
-			}
-		}
-		mutex_unlock(&kvm->lock);
-		if (r)
-			goto out;
-	}
-
 	flush_all_to_thread(current);
 
 	/* Save userspace EBB and other register values */
@@ -3336,10 +3353,10 @@ static int kvmppc_vcpu_run_hv(struct kvm_run *run, struct kvm_vcpu *vcpu)
 			trace_kvm_hcall_exit(vcpu, r);
 			kvmppc_core_prepare_to_enter(vcpu);
 		} else if (r == RESUME_PAGE_FAULT) {
-			srcu_idx = srcu_read_lock(&vcpu->kvm->srcu);
+			srcu_idx = srcu_read_lock(&kvm->srcu);
 			r = kvmppc_book3s_hv_page_fault(run, vcpu,
 				vcpu->arch.fault_dar, vcpu->arch.fault_dsisr);
-			srcu_read_unlock(&vcpu->kvm->srcu, srcu_idx);
+			srcu_read_unlock(&kvm->srcu, srcu_idx);
 		} else if (r == RESUME_PASSTHROUGH) {
 			if (WARN_ON(xive_enabled()))
 				r = H_SUCCESS;
@@ -3358,9 +3375,8 @@ static int kvmppc_vcpu_run_hv(struct kvm_run *run, struct kvm_vcpu *vcpu)
 	}
 	mtspr(SPRN_VRSAVE, user_vrsave);
 
- out:
 	vcpu->arch.state = KVMPPC_VCPU_NOTREADY;
-	atomic_dec(&vcpu->kvm->arch.vcpus_running);
+	atomic_dec(&kvm->arch.vcpus_running);
 	return r;
 }
 

arch/powerpc/kvm/book3s_hv_rmhandlers.S

@@ -1025,13 +1025,14 @@ ALT_FTR_SECTION_END_IFCLR(CPU_FTR_ARCH_300)
 	beq	no_xive
 	ld	r11, VCPU_XIVE_SAVED_STATE(r4)
 	li	r9, TM_QW1_OS
-	stdcix	r11,r9,r10
 	eieio
+	stdcix	r11,r9,r10
 	lwz	r11, VCPU_XIVE_CAM_WORD(r4)
 	li	r9, TM_QW1_OS + TM_WORD2
 	stwcix	r11,r9,r10
 	li	r9, 1
 	stw	r9, VCPU_XIVE_PUSHED(r4)
+	eieio
 no_xive:
 #endif /* CONFIG_KVM_XICS */
 
@@ -1346,6 +1347,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
 	bne	3f
 BEGIN_FTR_SECTION
 	PPC_MSGSYNC
+	lwsync
 END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
 	lbz	r0, HSTATE_HOST_IPI(r13)
 	cmpwi	r0, 0
@@ -1436,8 +1438,8 @@ guest_exit_cont:		/* r9 = vcpu, r12 = trap, r13 = paca */
 	cmpldi	cr0, r10, 0
 	beq	1f
 	/* First load to pull the context, we ignore the value */
-	lwzx	r11, r7, r10
 	eieio
+	lwzx	r11, r7, r10
 	/* Second load to recover the context state (Words 0 and 1) */
 	ldx	r11, r6, r10
 	b	3f
@@ -1445,8 +1447,8 @@ guest_exit_cont:		/* r9 = vcpu, r12 = trap, r13 = paca */
 	cmpldi	cr0, r10, 0
 	beq	1f
 	/* First load to pull the context, we ignore the value */
-	lwzcix	r11, r7, r10
 	eieio
+	lwzcix	r11, r7, r10
 	/* Second load to recover the context state (Words 0 and 1) */
 	ldcix	r11, r6, r10
 3:	std	r11, VCPU_XIVE_SAVED_STATE(r9)
@@ -1456,6 +1458,7 @@ guest_exit_cont:		/* r9 = vcpu, r12 = trap, r13 = paca */
 	stw	r10, VCPU_XIVE_PUSHED(r9)
 	stb	r10, (VCPU_XIVE_SAVED_STATE+3)(r9)
 	stb	r0, (VCPU_XIVE_SAVED_STATE+4)(r9)
+	eieio
 1:
 #endif /* CONFIG_KVM_XICS */
 	/* Save more register state  */
@@ -2838,6 +2841,10 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
 	PPC_MSGCLR(6)
 	/* see if it's a host IPI */
 	li	r3, 1
+BEGIN_FTR_SECTION
+	PPC_MSGSYNC
+	lwsync
+END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
 	lbz	r0, HSTATE_HOST_IPI(r13)
 	cmpwi	r0, 0
 	bnelr

arch/powerpc/kvm/powerpc.c

@@ -643,7 +643,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
 		break;
 #endif
 	case KVM_CAP_PPC_HTM:
-		r = is_kvmppc_hv_enabled(kvm) &&
+		r = hv_enabled &&
 		    (cur_cpu_spec->cpu_user_features2 & PPC_FEATURE2_HTM_COMP);
 		break;
 	default: