Merge tag 'iommu-updates-v3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu

Pull IOMMU updates from Joerg Roedel:
 "This pull-request includes:

   - change in the IOMMU-API to convert the former iommu_domain_capable
     function to just iommu_capable

   - various fixes in handling RMRR ranges for the VT-d driver (one fix
     requires a device driver core change which was acked by Greg KH)

   - the AMD IOMMU driver now assigns and deassigns complete alias
     groups to fix issues with devices using the wrong PCI request-id

   - MMU-401 support for the ARM SMMU driver

   - multi-master IOMMU group support for the ARM SMMU driver

   - various other small fixes all over the place"

* tag 'iommu-updates-v3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu: (41 commits)
  iommu/vt-d: Work around broken RMRR firmware entries
  iommu/vt-d: Store bus information in RMRR PCI device path
  iommu/vt-d: Only remove domain when device is removed
  driver core: Add BUS_NOTIFY_REMOVED_DEVICE event
  iommu/amd: Fix devid mapping for ivrs_ioapic override
  iommu/irq_remapping: Fix the regression of hpet irq remapping
  iommu: Fix bus notifier breakage
  iommu/amd: Split init_iommu_group() from iommu_init_device()
  iommu: Rework iommu_group_get_for_pci_dev()
  iommu: Make of_device_id array const
  amd_iommu: do not dereference a NULL pointer address.
  iommu/omap: Remove omap_iommu unused owner field
  iommu: Remove iommu_domain_has_cap() API function
  IB/usnic: Convert to use new iommu_capable() API function
  vfio: Convert to use new iommu_capable() API function
  kvm: iommu: Convert to use new iommu_capable() API function
  iommu/tegra: Convert to iommu_capable() API function
  iommu/msm: Convert to iommu_capable() API function
  iommu/vt-d: Convert to iommu_capable() API function
  iommu/fsl: Convert to iommu_capable() API function
  ...

25 files changed, 461 insertions, 356 deletions

diff --git a/Documentation/devicetree/bindings/iommu/arm,smmu.txt b/Documentation/devicetree/bindings/iommu/arm,smmu.txt
index 2d0f7cd867ea..06760503a819 100644
--- a/Documentation/devicetree/bindings/iommu/arm,smmu.txt
+++ b/Documentation/devicetree/bindings/iommu/arm,smmu.txt
@@ -14,6 +14,7 @@ conditions.
                         "arm,smmu-v1"
                         "arm,smmu-v2"
                         "arm,mmu-400"
+                        "arm,mmu-401"
                         "arm,mmu-500"
 
                   depending on the particular implementation and/or the
diff --git a/drivers/base/core.c b/drivers/base/core.c
index 28b808c73e8e..14d162952c3b 100644
--- a/drivers/base/core.c
+++ b/drivers/base/core.c
@@ -1211,6 +1211,9 @@ void device_del(struct device *dev)
 	 */
 	if (platform_notify_remove)
 		platform_notify_remove(dev);
+	if (dev->bus)
+		blocking_notifier_call_chain(&dev->bus->p->bus_notifier,
+					     BUS_NOTIFY_REMOVED_DEVICE, dev);
 	kobject_uevent(&dev->kobj, KOBJ_REMOVE);
 	cleanup_device_parent(dev);
 	kobject_del(&dev->kobj);
diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c
index 801a1d6937e4..417de1f32960 100644
--- a/drivers/infiniband/hw/usnic/usnic_uiom.c
+++ b/drivers/infiniband/hw/usnic/usnic_uiom.c
@@ -507,7 +507,7 @@ int usnic_uiom_attach_dev_to_pd(struct usnic_uiom_pd *pd, struct device *dev)
 	if (err)
 		goto out_free_dev;
 
-	if (!iommu_domain_has_cap(pd->domain, IOMMU_CAP_CACHE_COHERENCY)) {
+	if (!iommu_capable(dev->bus, IOMMU_CAP_CACHE_COHERENCY)) {
 		usnic_err("IOMMU of %s does not support cache coherency\n",
 				dev_name(dev));
 		err = -EINVAL;
diff --git a/drivers/iommu/amd_iommu.c b/drivers/iommu/amd_iommu.c
index ecb0109a5360..505a9adac2d5 100644
--- a/drivers/iommu/amd_iommu.c
+++ b/drivers/iommu/amd_iommu.c
@@ -88,6 +88,27 @@ int amd_iommu_max_glx_val = -1;
 static struct dma_map_ops amd_iommu_dma_ops;
 
 /*
+ * This struct contains device specific data for the IOMMU
+ */
+struct iommu_dev_data {
+	struct list_head list;		  /* For domain->dev_list */
+	struct list_head dev_data_list;	  /* For global dev_data_list */
+	struct list_head alias_list;      /* Link alias-groups together */
+	struct iommu_dev_data *alias_data;/* The alias dev_data */
+	struct protection_domain *domain; /* Domain the device is bound to */
+	u16 devid;			  /* PCI Device ID */
+	bool iommu_v2;			  /* Device can make use of IOMMUv2 */
+	bool passthrough;		  /* Default for device is pt_domain */
+	struct {
+		bool enabled;
+		int qdep;
+	} ats;				  /* ATS state */
+	bool pri_tlp;			  /* PASID TLB required for
+					     PPR completions */
+	u32 errata;			  /* Bitmap for errata to apply */
+};
+
+/*
  * general struct to manage commands send to an IOMMU
  */
 struct iommu_cmd {
@@ -114,8 +135,9 @@ static struct iommu_dev_data *alloc_dev_data(u16 devid)
 	if (!dev_data)
 		return NULL;
 
+	INIT_LIST_HEAD(&dev_data->alias_list);
+
 	dev_data->devid = devid;
-	atomic_set(&dev_data->bind, 0);
 
 	spin_lock_irqsave(&dev_data_list_lock, flags);
 	list_add_tail(&dev_data->dev_data_list, &dev_data_list);
@@ -260,17 +282,13 @@ static bool check_device(struct device *dev)
 	return true;
 }
 
-static int init_iommu_group(struct device *dev)
+static void init_iommu_group(struct device *dev)
 {
 	struct iommu_group *group;
 
 	group = iommu_group_get_for_dev(dev);
-
-	if (IS_ERR(group))
-		return PTR_ERR(group);
-
-	iommu_group_put(group);
-	return 0;
+	if (!IS_ERR(group))
+		iommu_group_put(group);
 }
 
 static int __last_alias(struct pci_dev *pdev, u16 alias, void *data)
@@ -340,7 +358,6 @@ static int iommu_init_device(struct device *dev)
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct iommu_dev_data *dev_data;
 	u16 alias;
-	int ret;
 
 	if (dev->archdata.iommu)
 		return 0;
@@ -362,12 +379,9 @@ static int iommu_init_device(struct device *dev)
 			return -ENOTSUPP;
 		}
 		dev_data->alias_data = alias_data;
-	}
 
-	ret = init_iommu_group(dev);
-	if (ret) {
-		free_dev_data(dev_data);
-		return ret;
+		/* Add device to the alias_list */
+		list_add(&dev_data->alias_list, &alias_data->alias_list);
 	}
 
 	if (pci_iommuv2_capable(pdev)) {
@@ -455,6 +469,15 @@ int __init amd_iommu_init_devices(void)
 			goto out_free;
 	}
 
+	/*
+	 * Initialize IOMMU groups only after iommu_init_device() has
+	 * had a chance to populate any IVRS defined aliases.
+	 */
+	for_each_pci_dev(pdev) {
+		if (check_device(&pdev->dev))
+			init_iommu_group(&pdev->dev);
+	}
+
 	return 0;
 
 out_free:
@@ -1368,6 +1391,9 @@ static int iommu_map_page(struct protection_domain *dom,
 	count     = PAGE_SIZE_PTE_COUNT(page_size);
 	pte       = alloc_pte(dom, bus_addr, page_size, NULL, GFP_KERNEL);
 
+	if (!pte)
+		return -ENOMEM;
+
 	for (i = 0; i < count; ++i)
 		if (IOMMU_PTE_PRESENT(pte[i]))
 			return -EBUSY;
@@ -2122,35 +2148,29 @@ static void do_detach(struct iommu_dev_data *dev_data)
 static int __attach_device(struct iommu_dev_data *dev_data,
 			   struct protection_domain *domain)
 {
+	struct iommu_dev_data *head, *entry;
 	int ret;
 
 	/* lock domain */
 	spin_lock(&domain->lock);
 
-	if (dev_data->alias_data != NULL) {
-		struct iommu_dev_data *alias_data = dev_data->alias_data;
+	head = dev_data;
 
-		/* Some sanity checks */
-		ret = -EBUSY;
-		if (alias_data->domain != NULL &&
-				alias_data->domain != domain)
-			goto out_unlock;
+	if (head->alias_data != NULL)
+		head = head->alias_data;
 
-		if (dev_data->domain != NULL &&
-				dev_data->domain != domain)
-			goto out_unlock;
+	/* Now we have the root of the alias group, if any */
 
-		/* Do real assignment */
-		if (alias_data->domain == NULL)
-			do_attach(alias_data, domain);
-
-		atomic_inc(&alias_data->bind);
-	}
+	ret = -EBUSY;
+	if (head->domain != NULL)
+		goto out_unlock;
 
-	if (dev_data->domain == NULL)
-		do_attach(dev_data, domain);
+	/* Attach alias group root */
+	do_attach(head, domain);
 
-	atomic_inc(&dev_data->bind);
+	/* Attach other devices in the alias group */
+	list_for_each_entry(entry, &head->alias_list, alias_list)
+		do_attach(entry, domain);
 
 	ret = 0;
 
@@ -2298,6 +2318,7 @@ static int attach_device(struct device *dev,
  */
 static void __detach_device(struct iommu_dev_data *dev_data)
 {
+	struct iommu_dev_data *head, *entry;
 	struct protection_domain *domain;
 	unsigned long flags;
 
@@ -2307,15 +2328,14 @@ static void __detach_device(struct iommu_dev_data *dev_data)
 
 	spin_lock_irqsave(&domain->lock, flags);
 
-	if (dev_data->alias_data != NULL) {
-		struct iommu_dev_data *alias_data = dev_data->alias_data;
+	head = dev_data;
+	if (head->alias_data != NULL)
+		head = head->alias_data;
 
-		if (atomic_dec_and_test(&alias_data->bind))
-			do_detach(alias_data);
-	}
+	list_for_each_entry(entry, &head->alias_list, alias_list)
+		do_detach(entry);
 
-	if (atomic_dec_and_test(&dev_data->bind))
-		do_detach(dev_data);
+	do_detach(head);
 
 	spin_unlock_irqrestore(&domain->lock, flags);
 
@@ -2415,6 +2435,7 @@ static int device_change_notifier(struct notifier_block *nb,
 	case BUS_NOTIFY_ADD_DEVICE:
 
 		iommu_init_device(dev);
+		init_iommu_group(dev);
 
 		/*
 		 * dev_data is still NULL and
@@ -3158,7 +3179,6 @@ static void cleanup_domain(struct protection_domain *domain)
 		entry = list_first_entry(&domain->dev_list,
 					 struct iommu_dev_data, list);
 		__detach_device(entry);
-		atomic_set(&entry->bind, 0);
 	}
 
 	write_unlock_irqrestore(&amd_iommu_devtable_lock, flags);
@@ -3384,20 +3404,20 @@ static phys_addr_t amd_iommu_iova_to_phys(struct iommu_domain *dom,
 	return paddr;
 }
 
-static int amd_iommu_domain_has_cap(struct iommu_domain *domain,
-				    unsigned long cap)
+static bool amd_iommu_capable(enum iommu_cap cap)
 {
 	switch (cap) {
 	case IOMMU_CAP_CACHE_COHERENCY:
-		return 1;
+		return true;
 	case IOMMU_CAP_INTR_REMAP:
-		return irq_remapping_enabled;
+		return (irq_remapping_enabled == 1);
 	}
 
-	return 0;
+	return false;
 }
 
 static const struct iommu_ops amd_iommu_ops = {
+	.capable = amd_iommu_capable,
 	.domain_init = amd_iommu_domain_init,
 	.domain_destroy = amd_iommu_domain_destroy,
 	.attach_dev = amd_iommu_attach_device,
@@ -3405,7 +3425,6 @@ static const struct iommu_ops amd_iommu_ops = {
 	.map = amd_iommu_map,
 	.unmap = amd_iommu_unmap,
 	.iova_to_phys = amd_iommu_iova_to_phys,
-	.domain_has_cap = amd_iommu_domain_has_cap,
 	.pgsize_bitmap	= AMD_IOMMU_PGSIZES,
 };
 
@@ -4235,7 +4254,7 @@ static int msi_setup_irq(struct pci_dev *pdev, unsigned int irq,
 	return 0;
 }
 
-static int setup_hpet_msi(unsigned int irq, unsigned int id)
+static int alloc_hpet_msi(unsigned int irq, unsigned int id)
 {
 	struct irq_2_irte *irte_info;
 	struct irq_cfg *cfg;
@@ -4274,6 +4293,6 @@ struct irq_remap_ops amd_iommu_irq_ops = {
 	.compose_msi_msg	= compose_msi_msg,
 	.msi_alloc_irq		= msi_alloc_irq,
 	.msi_setup_irq		= msi_setup_irq,
-	.setup_hpet_msi		= setup_hpet_msi,
+	.alloc_hpet_msi		= alloc_hpet_msi,
 };
 #endif
diff --git a/drivers/iommu/amd_iommu_init.c b/drivers/iommu/amd_iommu_init.c
index 3783e0b44df6..b0522f15730f 100644
--- a/drivers/iommu/amd_iommu_init.c
+++ b/drivers/iommu/amd_iommu_init.c
@@ -712,7 +712,7 @@ static void __init set_dev_entry_from_acpi(struct amd_iommu *iommu,
 	set_iommu_for_device(iommu, devid);
 }
 
-static int __init add_special_device(u8 type, u8 id, u16 devid, bool cmd_line)
+static int __init add_special_device(u8 type, u8 id, u16 *devid, bool cmd_line)
 {
 	struct devid_map *entry;
 	struct list_head *list;
@@ -731,6 +731,8 @@ static int __init add_special_device(u8 type, u8 id, u16 devid, bool cmd_line)
 		pr_info("AMD-Vi: Command-line override present for %s id %d - ignoring\n",
 			type == IVHD_SPECIAL_IOAPIC ? "IOAPIC" : "HPET", id);
 
+		*devid = entry->devid;
+
 		return 0;
 	}
 
@@ -739,7 +741,7 @@ static int __init add_special_device(u8 type, u8 id, u16 devid, bool cmd_line)
 		return -ENOMEM;
 
 	entry->id	= id;
-	entry->devid	= devid;
+	entry->devid	= *devid;
 	entry->cmd_line	= cmd_line;
 
 	list_add_tail(&entry->list, list);
@@ -754,7 +756,7 @@ static int __init add_early_maps(void)
 	for (i = 0; i < early_ioapic_map_size; ++i) {
 		ret = add_special_device(IVHD_SPECIAL_IOAPIC,
 					 early_ioapic_map[i].id,
-					 early_ioapic_map[i].devid,
+					 &early_ioapic_map[i].devid,
 					 early_ioapic_map[i].cmd_line);
 		if (ret)
 			return ret;
@@ -763,7 +765,7 @@ static int __init add_early_maps(void)
 	for (i = 0; i < early_hpet_map_size; ++i) {
 		ret = add_special_device(IVHD_SPECIAL_HPET,
 					 early_hpet_map[i].id,
-					 early_hpet_map[i].devid,
+					 &early_hpet_map[i].devid,
 					 early_hpet_map[i].cmd_line);
 		if (ret)
 			return ret;
@@ -978,10 +980,17 @@ static int __init init_iommu_from_acpi(struct amd_iommu *iommu,
 				    PCI_SLOT(devid),
 				    PCI_FUNC(devid));
 
-			set_dev_entry_from_acpi(iommu, devid, e->flags, 0);
-			ret = add_special_device(type, handle, devid, false);
+			ret = add_special_device(type, handle, &devid, false);
 			if (ret)
 				return ret;
+
+			/*
+			 * add_special_device might update the devid in case a
+			 * command-line override is present. So call
+			 * set_dev_entry_from_acpi after add_special_device.
+			 */
+			set_dev_entry_from_acpi(iommu, devid, e->flags, 0);
+
 			break;
 		}
 		default:
diff --git a/drivers/iommu/amd_iommu_types.h b/drivers/iommu/amd_iommu_types.h
index 8e43b7cba133..cec51a8ba844 100644
--- a/drivers/iommu/amd_iommu_types.h
+++ b/drivers/iommu/amd_iommu_types.h
@@ -418,27 +418,6 @@ struct protection_domain {
 };
 
 /*
- * This struct contains device specific data for the IOMMU
- */
-struct iommu_dev_data {
-	struct list_head list;		  /* For domain->dev_list */
-	struct list_head dev_data_list;	  /* For global dev_data_list */
-	struct iommu_dev_data *alias_data;/* The alias dev_data */
-	struct protection_domain *domain; /* Domain the device is bound to */
-	atomic_t bind;			  /* Domain attach reference count */
-	u16 devid;			  /* PCI Device ID */
-	bool iommu_v2;			  /* Device can make use of IOMMUv2 */
-	bool passthrough;		  /* Default for device is pt_domain */
-	struct {
-		bool enabled;
-		int qdep;
-	} ats;				  /* ATS state */
-	bool pri_tlp;			  /* PASID TLB required for
-					     PPR completions */
-	u32 errata;			  /* Bitmap for errata to apply */
-};
-
-/*
  * For dynamic growth the aperture size is split into ranges of 128MB of
  * DMA address space each. This struct represents one such range.
  */
diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c
index a83cc2a2a2ca..60558f794922 100644
--- a/drivers/iommu/arm-smmu.c
+++ b/drivers/iommu/arm-smmu.c
@@ -24,7 +24,7 @@
  *	- v7/v8 long-descriptor format
  *	- Non-secure access to the SMMU
  *	- 4k and 64k pages, with contiguous pte hints.
- *	- Up to 42-bit addressing (dependent on VA_BITS)
+ *	- Up to 48-bit addressing (dependent on VA_BITS)
  *	- Context fault reporting
  */
 
@@ -59,7 +59,7 @@
 
 /* SMMU global address space */
 #define ARM_SMMU_GR0(smmu)		((smmu)->base)
-#define ARM_SMMU_GR1(smmu)		((smmu)->base + (smmu)->pagesize)
+#define ARM_SMMU_GR1(smmu)		((smmu)->base + (1 << (smmu)->pgshift))
 
 /*
  * SMMU global address space with conditional offset to access secure
@@ -224,7 +224,7 @@
 
 /* Translation context bank */
 #define ARM_SMMU_CB_BASE(smmu)		((smmu)->base + ((smmu)->size >> 1))
-#define ARM_SMMU_CB(smmu, n)		((n) * (smmu)->pagesize)
+#define ARM_SMMU_CB(smmu, n)		((n) * (1 << (smmu)->pgshift))
 
 #define ARM_SMMU_CB_SCTLR		0x0
 #define ARM_SMMU_CB_RESUME		0x8
@@ -326,6 +326,16 @@
 
 #define FSYNR0_WNR			(1 << 4)
 
+static int force_stage;
+module_param_named(force_stage, force_stage, int, S_IRUGO | S_IWUSR);
+MODULE_PARM_DESC(force_stage,
+	"Force SMMU mappings to be installed at a particular stage of translation. A value of '1' or '2' forces the corresponding stage. All other values are ignored (i.e. no stage is forced). Note that selecting a specific stage will disable support for nested translation.");
+
+enum arm_smmu_arch_version {
+	ARM_SMMU_V1 = 1,
+	ARM_SMMU_V2,
+};
+
 struct arm_smmu_smr {
 	u8				idx;
 	u16				mask;
@@ -349,7 +359,7 @@ struct arm_smmu_device {
 
 	void __iomem			*base;
 	unsigned long			size;
-	unsigned long			pagesize;
+	unsigned long			pgshift;
 
 #define ARM_SMMU_FEAT_COHERENT_WALK	(1 << 0)
 #define ARM_SMMU_FEAT_STREAM_MATCH	(1 << 1)
@@ -360,7 +370,7 @@ struct arm_smmu_device {
 
 #define ARM_SMMU_OPT_SECURE_CFG_ACCESS (1 << 0)
 	u32				options;
-	int				version;
+	enum arm_smmu_arch_version	version;
 
 	u32				num_context_banks;
 	u32				num_s2_context_banks;
@@ -370,8 +380,9 @@ struct arm_smmu_device {
 	u32				num_mapping_groups;
 	DECLARE_BITMAP(smr_map, ARM_SMMU_MAX_SMRS);
 
-	unsigned long			input_size;
+	unsigned long			s1_input_size;
 	unsigned long			s1_output_size;
+	unsigned long			s2_input_size;
 	unsigned long			s2_output_size;
 
 	u32				num_global_irqs;
@@ -426,17 +437,17 @@ static void parse_driver_options(struct arm_smmu_device *smmu)
 	} while (arm_smmu_options[++i].opt);
 }
 
-static struct device *dev_get_master_dev(struct device *dev)
+static struct device_node *dev_get_dev_node(struct device *dev)
 {
 	if (dev_is_pci(dev)) {
 		struct pci_bus *bus = to_pci_dev(dev)->bus;
 
 		while (!pci_is_root_bus(bus))
 			bus = bus->parent;
-		return bus->bridge->parent;
+		return bus->bridge->parent->of_node;
 	}
 
-	return dev;
+	return dev->of_node;
 }
 
 static struct arm_smmu_master *find_smmu_master(struct arm_smmu_device *smmu,
@@ -461,15 +472,17 @@ static struct arm_smmu_master *find_smmu_master(struct arm_smmu_device *smmu,
 }
 
 static struct arm_smmu_master_cfg *
-find_smmu_master_cfg(struct arm_smmu_device *smmu, struct device *dev)
+find_smmu_master_cfg(struct device *dev)
 {
-	struct arm_smmu_master *master;
+	struct arm_smmu_master_cfg *cfg = NULL;
+	struct iommu_group *group = iommu_group_get(dev);
 
-	if (dev_is_pci(dev))
-		return dev->archdata.iommu;
+	if (group) {
+		cfg = iommu_group_get_iommudata(group);
+		iommu_group_put(group);
+	}
 
-	master = find_smmu_master(smmu, dev->of_node);
-	return master ? &master->cfg : NULL;
+	return cfg;
 }
 
 static int insert_smmu_master(struct arm_smmu_device *smmu,
@@ -545,7 +558,7 @@ static struct arm_smmu_device *find_smmu_for_device(struct device *dev)
 {
 	struct arm_smmu_device *smmu;
 	struct arm_smmu_master *master = NULL;
-	struct device_node *dev_node = dev_get_master_dev(dev)->of_node;
+	struct device_node *dev_node = dev_get_dev_node(dev);
 
 	spin_lock(&arm_smmu_devices_lock);
 	list_for_each_entry(smmu, &arm_smmu_devices, list) {
@@ -729,7 +742,7 @@ static void arm_smmu_init_context_bank(struct arm_smmu_domain *smmu_domain)
 
 	/* CBAR */
 	reg = cfg->cbar;
-	if (smmu->version == 1)
+	if (smmu->version == ARM_SMMU_V1)
 		reg |= cfg->irptndx << CBAR_IRPTNDX_SHIFT;
 
 	/*
@@ -744,7 +757,7 @@ static void arm_smmu_init_context_bank(struct arm_smmu_domain *smmu_domain)
 	}
 	writel_relaxed(reg, gr1_base + ARM_SMMU_GR1_CBAR(cfg->cbndx));
 
-	if (smmu->version > 1) {
+	if (smmu->version > ARM_SMMU_V1) {
 		/* CBA2R */
 #ifdef CONFIG_64BIT
 		reg = CBA2R_RW64_64BIT;
@@ -755,7 +768,7 @@ static void arm_smmu_init_context_bank(struct arm_smmu_domain *smmu_domain)
 			       gr1_base + ARM_SMMU_GR1_CBA2R(cfg->cbndx));
 
 		/* TTBCR2 */
-		switch (smmu->input_size) {
+		switch (smmu->s1_input_size) {
 		case 32:
 			reg = (TTBCR2_ADDR_32 << TTBCR2_SEP_SHIFT);
 			break;
@@ -817,14 +830,14 @@ static void arm_smmu_init_context_bank(struct arm_smmu_domain *smmu_domain)
 	 * TTBCR
 	 * We use long descriptor, with inner-shareable WBWA tables in TTBR0.
 	 */
-	if (smmu->version > 1) {
+	if (smmu->version > ARM_SMMU_V1) {
 		if (PAGE_SIZE == SZ_4K)
 			reg = TTBCR_TG0_4K;
 		else
 			reg = TTBCR_TG0_64K;
 
 		if (!stage1) {
-			reg |= (64 - smmu->s1_output_size) << TTBCR_T0SZ_SHIFT;
+			reg |= (64 - smmu->s2_input_size) << TTBCR_T0SZ_SHIFT;
 
 			switch (smmu->s2_output_size) {
 			case 32:
@@ -847,7 +860,7 @@ static void arm_smmu_init_context_bank(struct arm_smmu_domain *smmu_domain)
 				break;
 			}
 		} else {
-			reg |= (64 - smmu->input_size) << TTBCR_T0SZ_SHIFT;
+			reg |= (64 - smmu->s1_input_size) << TTBCR_T0SZ_SHIFT;
 		}
 	} else {
 		reg = 0;
@@ -914,7 +927,7 @@ static int arm_smmu_init_domain_context(struct iommu_domain *domain,
 		goto out_unlock;
 
 	cfg->cbndx = ret;
-	if (smmu->version == 1) {
+	if (smmu->version == ARM_SMMU_V1) {
 		cfg->irptndx = atomic_inc_return(&smmu->irptndx);
 		cfg->irptndx %= smmu->num_context_irqs;
 	} else {
@@ -1151,9 +1164,10 @@ static int arm_smmu_domain_add_master(struct arm_smmu_domain *smmu_domain,
 	struct arm_smmu_device *smmu = smmu_domain->smmu;
 	void __iomem *gr0_base = ARM_SMMU_GR0(smmu);
 
+	/* Devices in an IOMMU group may already be configured */
 	ret = arm_smmu_master_configure_smrs(smmu, cfg);
 	if (ret)
-		return ret;
+		return ret == -EEXIST ? 0 : ret;
 
 	for (i = 0; i < cfg->num_streamids; ++i) {
 		u32 idx, s2cr;
@@ -1174,6 +1188,10 @@ static void arm_smmu_domain_remove_master(struct arm_smmu_domain *smmu_domain,
 	struct arm_smmu_device *smmu = smmu_domain->smmu;
 	void __iomem *gr0_base = ARM_SMMU_GR0(smmu);
 
+	/* An IOMMU group is torn down by the first device to be removed */
+	if ((smmu->features & ARM_SMMU_FEAT_STREAM_MATCH) && !cfg->smrs)
+		return;
+
 	/*
 	 * We *must* clear the S2CR first, because freeing the SMR means
 	 * that it can be re-allocated immediately.
@@ -1195,12 +1213,17 @@ static int arm_smmu_attach_dev(struct iommu_domain *domain, struct device *dev)
 	struct arm_smmu_device *smmu, *dom_smmu;
 	struct arm_smmu_master_cfg *cfg;
 
-	smmu = dev_get_master_dev(dev)->archdata.iommu;
+	smmu = find_smmu_for_device(dev);
 	if (!smmu) {
 		dev_err(dev, "cannot attach to SMMU, is it on the same bus?\n");
 		return -ENXIO;
 	}
 
+	if (dev->archdata.iommu) {
+		dev_err(dev, "already attached to IOMMU domain\n");
+		return -EEXIST;
+	}
+
 	/*
 	 * Sanity check the domain. We don't support domains across
 	 * different SMMUs.
@@ -1223,11 +1246,14 @@ static int arm_smmu_attach_dev(struct iommu_domain *domain, struct device *dev)
 	}
 
 	/* Looks ok, so add the device to the domain */
-	cfg = find_smmu_master_cfg(smmu_domain->smmu, dev);
+	cfg = find_smmu_master_cfg(dev);
 	if (!cfg)
 		return -ENODEV;
 
-	return arm_smmu_domain_add_master(smmu_domain, cfg);
+	ret = arm_smmu_domain_add_master(smmu_domain, cfg);
+	if (!ret)
+		dev->archdata.iommu = domain;
+	return ret;
 }
 
 static void arm_smmu_detach_dev(struct iommu_domain *domain, struct device *dev)
@@ -1235,9 +1261,12 @@ static void arm_smmu_detach_dev(struct iommu_domain *domain, struct device *dev)
 	struct arm_smmu_domain *smmu_domain = domain->priv;
 	struct arm_smmu_master_cfg *cfg;
 
-	cfg = find_smmu_master_cfg(smmu_domain->smmu, dev);
-	if (cfg)
-		arm_smmu_domain_remove_master(smmu_domain, cfg);
+	cfg = find_smmu_master_cfg(dev);
+	if (!cfg)
+		return;
+
+	dev->archdata.iommu = NULL;
+	arm_smmu_domain_remove_master(smmu_domain, cfg);
 }
 
 static bool arm_smmu_pte_is_contiguous_range(unsigned long addr,
@@ -1379,6 +1408,7 @@ static int arm_smmu_alloc_init_pmd(struct arm_smmu_device *smmu, pud_t *pud,
 		ret = arm_smmu_alloc_init_pte(smmu, pmd, addr, next, pfn,
 					      prot, stage);
 		phys += next - addr;
+		pfn = __phys_to_pfn(phys);
 	} while (pmd++, addr = next, addr < end);
 
 	return ret;
@@ -1431,9 +1461,11 @@ static int arm_smmu_handle_mapping(struct arm_smmu_domain *smmu_domain,
 
 	if (cfg->cbar == CBAR_TYPE_S2_TRANS) {
 		stage = 2;
+		input_mask = (1ULL << smmu->s2_input_size) - 1;
 		output_mask = (1ULL << smmu->s2_output_size) - 1;
 	} else {
 		stage = 1;
+		input_mask = (1ULL << smmu->s1_input_size) - 1;
 		output_mask = (1ULL << smmu->s1_output_size) - 1;
 	}
 
@@ -1443,7 +1475,6 @@ static int arm_smmu_handle_mapping(struct arm_smmu_domain *smmu_domain,
 	if (size & ~PAGE_MASK)
 		return -EINVAL;
 
-	input_mask = (1ULL << smmu->input_size) - 1;
 	if ((phys_addr_t)iova & ~input_mask)
 		return -ERANGE;
 
@@ -1526,20 +1557,19 @@ static phys_addr_t arm_smmu_iova_to_phys(struct iommu_domain *domain,
 	return __pfn_to_phys(pte_pfn(pte)) | (iova & ~PAGE_MASK);
 }
 
-static int arm_smmu_domain_has_cap(struct iommu_domain *domain,
-				   unsigned long cap)
+static bool arm_smmu_capable(enum iommu_cap cap)
 {
-	struct arm_smmu_domain *smmu_domain = domain->priv;
-	struct arm_smmu_device *smmu = smmu_domain->smmu;
-	u32 features = smmu ? smmu->features : 0;
-
 	switch (cap) {
 	case IOMMU_CAP_CACHE_COHERENCY:
-		return features & ARM_SMMU_FEAT_COHERENT_WALK;
+		/*
+		 * Return true here as the SMMU can always send out coherent
+		 * requests.
+		 */
+		return true;
 	case IOMMU_CAP_INTR_REMAP:
-		return 1; /* MSIs are just memory writes */
+		return true; /* MSIs are just memory writes */
 	default:
-		return 0;
+		return false;
 	}
 }
 
@@ -1549,17 +1579,19 @@ static int __arm_smmu_get_pci_sid(struct pci_dev *pdev, u16 alias, void *data)
 	return 0; /* Continue walking */
 }
 
+static void __arm_smmu_release_pci_iommudata(void *data)
+{
+	kfree(data);
+}
+
 static int arm_smmu_add_device(struct device *dev)
 {
 	struct arm_smmu_device *smmu;
+	struct arm_smmu_master_cfg *cfg;
 	struct iommu_group *group;
+	void (*releasefn)(void *) = NULL;
 	int ret;
 
-	if (dev->archdata.iommu) {
-		dev_warn(dev, "IOMMU driver already assigned to device\n");
-		return -EINVAL;
-	}
-
 	smmu = find_smmu_for_device(dev);
 	if (!smmu)
 		return -ENODEV;
@@ -1571,7 +1603,6 @@ static int arm_smmu_add_device(struct device *dev)
 	}
 
 	if (dev_is_pci(dev)) {
-		struct arm_smmu_master_cfg *cfg;
 		struct pci_dev *pdev = to_pci_dev(dev);
 
 		cfg = kzalloc(sizeof(*cfg), GFP_KERNEL);
@@ -1587,11 +1618,20 @@ static int arm_smmu_add_device(struct device *dev)
 		 */
 		pci_for_each_dma_alias(pdev, __arm_smmu_get_pci_sid,
 				       &cfg->streamids[0]);
-		dev->archdata.iommu = cfg;
+		releasefn = __arm_smmu_release_pci_iommudata;
 	} else {
-		dev->archdata.iommu = smmu;
+		struct arm_smmu_master *master;
+
+		master = find_smmu_master(smmu, dev->of_node);
+		if (!master) {
+			ret = -ENODEV;
+			goto out_put_group;
+		}
+
+		cfg = &master->cfg;
 	}
 
+	iommu_group_set_iommudata(group, cfg, releasefn);
 	ret = iommu_group_add_device(group, dev);
 
 out_put_group:
@@ -1601,14 +1641,11 @@ out_put_group:
 
 static void arm_smmu_remove_device(struct device *dev)
 {
-	if (dev_is_pci(dev))
-		kfree(dev->archdata.iommu);
-
-	dev->archdata.iommu = NULL;
 	iommu_group_remove_device(dev);
 }
 
 static const struct iommu_ops arm_smmu_ops = {
+	.capable	= arm_smmu_capable,
 	.domain_init	= arm_smmu_domain_init,
 	.domain_destroy	= arm_smmu_domain_destroy,
 	.attach_dev	= arm_smmu_attach_dev,
@@ -1616,7 +1653,6 @@ static const struct iommu_ops arm_smmu_ops = {
 	.map		= arm_smmu_map,
 	.unmap		= arm_smmu_unmap,
 	.iova_to_phys	= arm_smmu_iova_to_phys,
-	.domain_has_cap	= arm_smmu_domain_has_cap,
 	.add_device	= arm_smmu_add_device,
 	.remove_device	= arm_smmu_remove_device,
 	.pgsize_bitmap	= (SECTION_SIZE |
@@ -1702,10 +1738,6 @@ static int arm_smmu_device_cfg_probe(struct arm_smmu_device *smmu)
 	u32 id;
 
 	dev_notice(smmu->dev, "probing hardware configuration...\n");
-
-	/* Primecell ID */
-	id = readl_relaxed(gr0_base + ARM_SMMU_GR0_PIDR2);
-	smmu->version = ((id >> PIDR2_ARCH_SHIFT) & PIDR2_ARCH_MASK) + 1;
 	dev_notice(smmu->dev, "SMMUv%d with:\n", smmu->version);
 
 	/* ID0 */
@@ -1716,6 +1748,13 @@ static int arm_smmu_device_cfg_probe(struct arm_smmu_device *smmu)
 		return -ENODEV;
 	}
 #endif
+
+	/* Restrict available stages based on module parameter */
+	if (force_stage == 1)
+		id &= ~(ID0_S2TS | ID0_NTS);
+	else if (force_stage == 2)
+		id &= ~(ID0_S1TS | ID0_NTS);
+
 	if (id & ID0_S1TS) {
 		smmu->features |= ARM_SMMU_FEAT_TRANS_S1;
 		dev_notice(smmu->dev, "\tstage 1 translation\n");
@@ -1732,8 +1771,7 @@ static int arm_smmu_device_cfg_probe(struct arm_smmu_device *smmu)
 	}
 
 	if (!(smmu->features &
-		(ARM_SMMU_FEAT_TRANS_S1 | ARM_SMMU_FEAT_TRANS_S2 |
-		 ARM_SMMU_FEAT_TRANS_NESTED))) {
+		(ARM_SMMU_FEAT_TRANS_S1 | ARM_SMMU_FEAT_TRANS_S2))) {
 		dev_err(smmu->dev, "\tno translation support!\n");
 		return -ENODEV;
 	}
@@ -1779,12 +1817,12 @@ static int arm_smmu_device_cfg_probe(struct arm_smmu_device *smmu)
 
 	/* ID1 */
 	id = readl_relaxed(gr0_base + ARM_SMMU_GR0_ID1);
-	smmu->pagesize = (id & ID1_PAGESIZE) ? SZ_64K : SZ_4K;
+	smmu->pgshift = (id & ID1_PAGESIZE) ? 16 : 12;
 
 	/* Check for size mismatch of SMMU address space from mapped region */
 	size = 1 <<
 		(((id >> ID1_NUMPAGENDXB_SHIFT) & ID1_NUMPAGENDXB_MASK) + 1);
-	size *= (smmu->pagesize << 1);
+	size *= 2 << smmu->pgshift;
 	if (smmu->size != size)
 		dev_warn(smmu->dev,
 			"SMMU address space size (0x%lx) differs from mapped region size (0x%lx)!\n",
@@ -1803,28 +1841,21 @@ static int arm_smmu_device_cfg_probe(struct arm_smmu_device *smmu)
 	/* ID2 */
 	id = readl_relaxed(gr0_base + ARM_SMMU_GR0_ID2);
 	size = arm_smmu