1 files changed, 271 insertions, 212 deletions

diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
index c9c370de9d3d..3bfc56df4f78 100644
--- a/drivers/iommu/iommu.c
+++ b/drivers/iommu/iommu.c
@@ -41,6 +41,7 @@
 
 static struct kset *iommu_group_kset;
 static DEFINE_IDA(iommu_group_ida);
+static DEFINE_IDA(iommu_global_pasid_ida);
 
 static unsigned int iommu_def_domain_type __read_mostly;
 static bool iommu_dma_strict __read_mostly = IS_ENABLED(CONFIG_IOMMU_DEFAULT_DMA_STRICT);
@@ -129,9 +130,12 @@ static int iommu_setup_default_domain(struct iommu_group *group,
 				      int target_type);
 static int iommu_create_device_direct_mappings(struct iommu_domain *domain,
 					       struct device *dev);
-static struct iommu_group *iommu_group_get_for_dev(struct device *dev);
 static ssize_t iommu_group_store_type(struct iommu_group *group,
 				      const char *buf, size_t count);
+static struct group_device *iommu_group_alloc_device(struct iommu_group *group,
+						     struct device *dev);
+static void __iommu_group_free_device(struct iommu_group *group,
+				      struct group_device *grp_dev);
 
 #define IOMMU_GROUP_ATTR(_name, _mode, _show, _store)		\
 struct iommu_group_attribute iommu_group_attr_##_name =		\
@@ -377,28 +381,18 @@ static u32 dev_iommu_get_max_pasids(struct device *dev)
 	return min_t(u32, max_pasids, dev->iommu->iommu_dev->max_pasids);
 }
 
-static int __iommu_probe_device(struct device *dev, struct list_head *group_list)
+/*
+ * Init the dev->iommu and dev->iommu_group in the struct device and get the
+ * driver probed
+ */
+static int iommu_init_device(struct device *dev, const struct iommu_ops *ops)
 {
-	const struct iommu_ops *ops = dev->bus->iommu_ops;
 	struct iommu_device *iommu_dev;
 	struct iommu_group *group;
-	static DEFINE_MUTEX(iommu_probe_device_lock);
 	int ret;
 
-	if (!ops)
-		return -ENODEV;
-	/*
-	 * Serialise to avoid races between IOMMU drivers registering in
-	 * parallel and/or the "replay" calls from ACPI/OF code via client
-	 * driver probe. Once the latter have been cleaned up we should
-	 * probably be able to use device_lock() here to minimise the scope,
-	 * but for now enforcing a simple global ordering is fine.
-	 */
-	mutex_lock(&iommu_probe_device_lock);
-	if (!dev_iommu_get(dev)) {
-		ret = -ENOMEM;
-		goto err_unlock;
-	}
+	if (!dev_iommu_get(dev))
+		return -ENOMEM;
 
 	if (!try_module_get(ops->owner)) {
 		ret = -EINVAL;
@@ -408,124 +402,184 @@ static int __iommu_probe_device(struct device *dev, struct list_head *group_list
 	iommu_dev = ops->probe_device(dev);
 	if (IS_ERR(iommu_dev)) {
 		ret = PTR_ERR(iommu_dev);
-		goto out_module_put;
+		goto err_module_put;
 	}
 
-	dev->iommu->iommu_dev = iommu_dev;
-	dev->iommu->max_pasids = dev_iommu_get_max_pasids(dev);
-	if (ops->is_attach_deferred)
-		dev->iommu->attach_deferred = ops->is_attach_deferred(dev);
+	ret = iommu_device_link(iommu_dev, dev);
+	if (ret)
+		goto err_release;
 
-	group = iommu_group_get_for_dev(dev);
+	group = ops->device_group(dev);
+	if (WARN_ON_ONCE(group == NULL))
+		group = ERR_PTR(-EINVAL);
 	if (IS_ERR(group)) {
 		ret = PTR_ERR(group);
-		goto out_release;
+		goto err_unlink;
 	}
+	dev->iommu_group = group;
 
-	mutex_lock(&group->mutex);
-	if (group_list && !group->default_domain && list_empty(&group->entry))
-		list_add_tail(&group->entry, group_list);
-	mutex_unlock(&group->mutex);
-	iommu_group_put(group);
-
-	mutex_unlock(&iommu_probe_device_lock);
-	iommu_device_link(iommu_dev, dev);
-
+	dev->iommu->iommu_dev = iommu_dev;
+	dev->iommu->max_pasids = dev_iommu_get_max_pasids(dev);
+	if (ops->is_attach_deferred)
+		dev->iommu->attach_deferred = ops->is_attach_deferred(dev);
 	return 0;
 
-out_release:
+err_unlink:
+	iommu_device_unlink(iommu_dev, dev);
+err_release:
 	if (ops->release_device)
 		ops->release_device(dev);
-
-out_module_put:
+err_module_put:
 	module_put(ops->owner);
-
 err_free:
 	dev_iommu_free(dev);
+	return ret;
+}
 
-err_unlock:
-	mutex_unlock(&iommu_probe_device_lock);
+static void iommu_deinit_device(struct device *dev)
+{
+	struct iommu_group *group = dev->iommu_group;
+	const struct iommu_ops *ops = dev_iommu_ops(dev);
 
-	return ret;
+	lockdep_assert_held(&group->mutex);
+
+	iommu_device_unlink(dev->iommu->iommu_dev, dev);
+
+	/*
+	 * release_device() must stop using any attached domain on the device.
+	 * If there are still other devices in the group they are not effected
+	 * by this callback.
+	 *
+	 * The IOMMU driver must set the device to either an identity or
+	 * blocking translation and stop using any domain pointer, as it is
+	 * going to be freed.
+	 */
+	if (ops->release_device)
+		ops->release_device(dev);
+
+	/*
+	 * If this is the last driver to use the group then we must free the
+	 * domains before we do the module_put().
+	 */
+	if (list_empty(&group->devices)) {
+		if (group->default_domain) {
+			iommu_domain_free(group->default_domain);
+			group->default_domain = NULL;
+		}
+		if (group->blocking_domain) {
+			iommu_domain_free(group->blocking_domain);
+			group->blocking_domain = NULL;
+		}
+		group->domain = NULL;
+	}
+
+	/* Caller must put iommu_group */
+	dev->iommu_group = NULL;
+	module_put(ops->owner);
+	dev_iommu_free(dev);
 }
 
-int iommu_probe_device(struct device *dev)
+static int __iommu_probe_device(struct device *dev, struct list_head *group_list)
 {
-	const struct iommu_ops *ops;
+	const struct iommu_ops *ops = dev->bus->iommu_ops;
 	struct iommu_group *group;
+	static DEFINE_MUTEX(iommu_probe_device_lock);
+	struct group_device *gdev;
 	int ret;
 
-	ret = __iommu_probe_device(dev, NULL);
-	if (ret)
-		goto err_out;
+	if (!ops)
+		return -ENODEV;
+	/*
+	 * Serialise to avoid races between IOMMU drivers registering in
+	 * parallel and/or the "replay" calls from ACPI/OF code via client
+	 * driver probe. Once the latter have been cleaned up we should
+	 * probably be able to use device_lock() here to minimise the scope,
+	 * but for now enforcing a simple global ordering is fine.
+	 */
+	mutex_lock(&iommu_probe_device_lock);
 
-	group = iommu_group_get(dev);
-	if (!group) {
-		ret = -ENODEV;
-		goto err_release;
+	/* Device is probed already if in a group */
+	if (dev->iommu_group) {
+		ret = 0;
+		goto out_unlock;
 	}
 
+	ret = iommu_init_device(dev, ops);
+	if (ret)
+		goto out_unlock;
+
+	group = dev->iommu_group;
+	gdev = iommu_group_alloc_device(group, dev);
 	mutex_lock(&group->mutex);
+	if (IS_ERR(gdev)) {
+		ret = PTR_ERR(gdev);
+		goto err_put_group;
+	}
 
+	/*
+	 * The gdev must be in the list before calling
+	 * iommu_setup_default_domain()
+	 */
+	list_add_tail(&gdev->list, &group->devices);
+	WARN_ON(group->default_domain && !group->domain);
 	if (group->default_domain)
 		iommu_create_device_direct_mappings(group->default_domain, dev);
-
 	if (group->domain) {
 		ret = __iommu_device_set_domain(group, dev, group->domain, 0);
 		if (ret)
-			goto err_unlock;
-	} else if (!group->default_domain) {
+			goto err_remove_gdev;
+	} else if (!group->default_domain && !group_list) {
 		ret = iommu_setup_default_domain(group, 0);
 		if (ret)
-			goto err_unlock;
+			goto err_remove_gdev;
+	} else if (!group->default_domain) {
+		/*
+		 * With a group_list argument we defer the default_domain setup
+		 * to the caller by providing a de-duplicated list of groups
+		 * that need further setup.
+		 */
+		if (list_empty(&group->entry))
+			list_add_tail(&group->entry, group_list);
 	}
-
 	mutex_unlock(&group->mutex);
-	iommu_group_put(group);
+	mutex_unlock(&iommu_probe_device_lock);
 
-	ops = dev_iommu_ops(dev);
-	if (ops->probe_finalize)
-		ops->probe_finalize(dev);
+	if (dev_is_pci(dev))
+		iommu_dma_set_pci_32bit_workaround(dev);
 
 	return 0;
 
-err_unlock:
+err_remove_gdev:
+	list_del(&gdev->list);
+	__iommu_group_free_device(group, gdev);
+err_put_group:
+	iommu_deinit_device(dev);
 	mutex_unlock(&group->mutex);
 	iommu_group_put(group);
-err_release:
-	iommu_release_device(dev);
+out_unlock:
+	mutex_unlock(&iommu_probe_device_lock);
 
-err_out:
 	return ret;
-
 }
 
-/*
- * Remove a device from a group's device list and return the group device
- * if successful.
- */
-static struct group_device *
-__iommu_group_remove_device(struct iommu_group *group, struct device *dev)
+int iommu_probe_device(struct device *dev)
 {
-	struct group_device *device;
+	const struct iommu_ops *ops;
+	int ret;
 
-	lockdep_assert_held(&group->mutex);
-	for_each_group_device(group, device) {
-		if (device->dev == dev) {
-			list_del(&device->list);
-			return device;
-		}
-	}
+	ret = __iommu_probe_device(dev, NULL);
+	if (ret)
+		return ret;
 
-	return NULL;
+	ops = dev_iommu_ops(dev);
+	if (ops->probe_finalize)
+		ops->probe_finalize(dev);
+
+	return 0;
 }
 
-/*
- * Release a device from its group and decrements the iommu group reference
- * count.
- */
-static void __iommu_group_release_device(struct iommu_group *group,
-					 struct group_device *grp_dev)
+static void __iommu_group_free_device(struct iommu_group *group,
+				      struct group_device *grp_dev)
 {
 	struct device *dev = grp_dev->dev;
 
@@ -534,54 +588,57 @@ static void __iommu_group_release_device(struct iommu_group *group,
 
 	trace_remove_device_from_group(group->id, dev);
 
+	/*
+	 * If the group has become empty then ownership must have been
+	 * released, and the current domain must be set back to NULL or
+	 * the default domain.
+	 */
+	if (list_empty(&group->devices))
+		WARN_ON(group->owner_cnt ||
+			group->domain != group->default_domain);
+
 	kfree(grp_dev->name);
 	kfree(grp_dev);
-	dev->iommu_group = NULL;
-	kobject_put(group->devices_kobj);
 }
 
-static void iommu_release_device(struct device *dev)
+/* Remove the iommu_group from the struct device. */
+static void __iommu_group_remove_device(struct device *dev)
 {
 	struct iommu_group *group = dev->iommu_group;
 	struct group_device *device;
-	const struct iommu_ops *ops;
-
-	if (!dev->iommu || !group)
-		return;
-
-	iommu_device_unlink(dev->iommu->iommu_dev, dev);
 
 	mutex_lock(&group->mutex);
-	device = __iommu_group_remove_device(group, dev);
+	for_each_group_device(group, device) {
+		if (device->dev != dev)
+			continue;
 
-	/*
-	 * If the group has become empty then ownership must have been released,
-	 * and the current domain must be set back to NULL or the default
-	 * domain.
-	 */
-	if (list_empty(&group->devices))
-		WARN_ON(group->owner_cnt ||
-			group->domain != group->default_domain);
+		list_del(&device->list);
+		__iommu_group_free_device(group, device);
+		if (dev->iommu && dev->iommu->iommu_dev)
+			iommu_deinit_device(dev);
+		else
+			dev->iommu_group = NULL;
+		break;
+	}
+	mutex_unlock(&group->mutex);
 
 	/*
-	 * release_device() must stop using any attached domain on the device.
-	 * If there are still other devices in the group they are not effected
-	 * by this callback.
-	 *
-	 * The IOMMU driver must set the device to either an identity or
-	 * blocking translation and stop using any domain pointer, as it is
-	 * going to be freed.
+	 * Pairs with the get in iommu_init_device() or
+	 * iommu_group_add_device()
 	 */
-	ops = dev_iommu_ops(dev);
-	if (ops->release_device)
-		ops->release_device(dev);
-	mutex_unlock(&group->mutex);
+	iommu_group_put(group);
+}
 
-	if (device)
-		__iommu_group_release_device(group, device);
+static void iommu_release_device(struct device *dev)
+{
+	struct iommu_group *group = dev->iommu_group;
 
-	module_put(ops->owner);
-	dev_iommu_free(dev);
+	if (group)
+		__iommu_group_remove_device(dev);
+
+	/* Free any fwspec if no iommu_driver was ever attached */
+	if (dev->iommu)
+		dev_iommu_free(dev);
 }
 
 static int __init iommu_set_def_domain_type(char *str)
@@ -842,10 +899,9 @@ static void iommu_group_release(struct kobject *kobj)
 
 	ida_free(&iommu_group_ida, group->id);
 
-	if (group->default_domain)
-		iommu_domain_free(group->default_domain);
-	if (group->blocking_domain)
-		iommu_domain_free(group->blocking_domain);
+	/* Domains are free'd by iommu_deinit_device() */
+	WARN_ON(group->default_domain);
+	WARN_ON(group->blocking_domain);
 
 	kfree(group->name);
 	kfree(group);
@@ -1003,14 +1059,12 @@ static int iommu_create_device_direct_mappings(struct iommu_domain *domain,
 	unsigned long pg_size;
 	int ret = 0;
 
-	if (!iommu_is_dma_domain(domain))
-		return 0;
-
-	BUG_ON(!domain->pgsize_bitmap);
-
-	pg_size = 1UL << __ffs(domain->pgsize_bitmap);
+	pg_size = domain->pgsize_bitmap ? 1UL << __ffs(domain->pgsize_bitmap) : 0;
 	INIT_LIST_HEAD(&mappings);
 
+	if (WARN_ON_ONCE(iommu_is_dma_domain(domain) && !pg_size))
+		return -EINVAL;
+
 	iommu_get_resv_regions(dev, &mappings);
 
 	/* We need to consider overlapping regions for different devices */
@@ -1018,13 +1072,17 @@ static int iommu_create_device_direct_mappings(struct iommu_domain *domain,
 		dma_addr_t start, end, addr;
 		size_t map_size = 0;
 
-		start = ALIGN(entry->start, pg_size);
-		end   = ALIGN(entry->start + entry->length, pg_size);
+		if (entry->type == IOMMU_RESV_DIRECT)
+			dev->iommu->require_direct = 1;
 
-		if (entry->type != IOMMU_RESV_DIRECT &&
-		    entry->type != IOMMU_RESV_DIRECT_RELAXABLE)
+		if ((entry->type != IOMMU_RESV_DIRECT &&
+		     entry->type != IOMMU_RESV_DIRECT_RELAXABLE) ||
+		    !iommu_is_dma_domain(domain))
 			continue;
 
+		start = ALIGN(entry->start, pg_size);
+		end   = ALIGN(entry->start + entry->length, pg_size);
+
 		for (addr = start; addr <= end; addr += pg_size) {
 			phys_addr_t phys_addr;
 
@@ -1058,22 +1116,16 @@ out:
 	return ret;
 }
 
-/**
- * iommu_group_add_device - add a device to an iommu group
- * @group: the group into which to add the device (reference should be held)
- * @dev: the device
- *
- * This function is called by an iommu driver to add a device into a
- * group.  Adding a device increments the group reference count.
- */
-int iommu_group_add_device(struct iommu_group *group, struct device *dev)
+/* This is undone by __iommu_group_free_device() */
+static struct group_device *iommu_group_alloc_device(struct iommu_group *group,
+						     struct device *dev)
 {
 	int ret, i = 0;
 	struct group_device *device;
 
 	device = kzalloc(sizeof(*device), GFP_KERNEL);
 	if (!device)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 
 	device->dev = dev;
 
@@ -1104,18 +1156,11 @@ rename:
 		goto err_free_name;
 	}
 
-	kobject_get(group->devices_kobj);
-
-	dev->iommu_group = group;
-
-	mutex_lock(&group->mutex);
-	list_add_tail(&device->list, &group->devices);
-	mutex_unlock(&group->mutex);
 	trace_add_device_to_group(group->id, dev);
 
 	dev_info(dev, "Adding to iommu group %d\n", group->id);
 
-	return 0;
+	return device;
 
 err_free_name:
 	kfree(device->name);
@@ -1124,7 +1169,32 @@ err_remove_link:
 err_free_device:
 	kfree(device);
 	dev_err(dev, "Failed to add to iommu group %d: %d\n", group->id, ret);
-	return ret;
+	return ERR_PTR(ret);
+}
+
+/**
+ * iommu_group_add_device - add a device to an iommu group
+ * @group: the group into which to add the device (reference should be held)
+ * @dev: the device
+ *
+ * This function is called by an iommu driver to add a device into a
+ * group.  Adding a device increments the group reference count.
+ */
+int iommu_group_add_device(struct iommu_group *group, struct device *dev)
+{
+	struct group_device *gdev;
+
+	gdev = iommu_group_alloc_device(group, dev);
+	if (IS_ERR(gdev))
+		return PTR_ERR(gdev);
+
+	iommu_group_ref_get(group);
+	dev->iommu_group = group;
+
+	mutex_lock(&group->mutex);
+	list_add_tail(&gdev->list, &group->devices);
+	mutex_unlock(&group->mutex);
+	return 0;
 }
 EXPORT_SYMBOL_GPL(iommu_group_add_device);
 
@@ -1138,19 +1208,13 @@ EXPORT_SYMBOL_GPL(iommu_group_add_device);
 void iommu_group_remove_device(struct device *dev)
 {
 	struct iommu_group *group = dev->iommu_group;
-	struct group_device *device;
 
 	if (!group)
 		return;
 
 	dev_info(dev, "Removing from iommu group %d\n", group->id);
 
-	mutex_lock(&group->mutex);
-	device = __iommu_group_remove_device(group, dev);
-	mutex_unlock(&group->mutex);
-
-	if (device)
-		__iommu_group_release_device(group, device);
+	__iommu_group_remove_device(dev);
 }
 EXPORT_SYMBOL_GPL(iommu_group_remove_device);
 
@@ -1708,45 +1772,6 @@ iommu_group_alloc_default_domain(struct iommu_group *group, int req_type)
 	return dom;
 }
 
-/**
- * iommu_group_get_for_dev - Find or create the IOMMU group for a device
- * @dev: target device
- *
- * This function is intended to be called by IOMMU drivers and extended to
- * support common, bus-defined algorithms when determining or creating the
- * IOMMU group for a device.  On success, the caller will hold a reference
- * to the returned IOMMU group, which will already include the provided
- * device.  The reference should be released with iommu_group_put().
- */
-static struct iommu_group *iommu_group_get_for_dev(struct device *dev)
-{
-	const struct iommu_ops *ops = dev_iommu_ops(dev);
-	struct iommu_group *group;
-	int ret;
-
-	group = iommu_group_get(dev);
-	if (group)
-		return group;
-
-	group = ops->device_group(dev);
-	if (WARN_ON_ONCE(group == NULL))
-		return ERR_PTR(-EINVAL);
-
-	if (IS_ERR(group))
-		return group;
-
-	ret = iommu_group_add_device(group, dev);
-	if (ret)
-		goto out_put_group;
-
-	return group;
-
-out_put_group:
-	iommu_group_put(group);
-
-	return ERR_PTR(ret);
-}
-
 struct iommu_domain *iommu_group_default_domain(struct iommu_group *group)
 {
 	return group->default_domain;
@@ -1755,16 +1780,8 @@ struct iommu_domain *iommu_group_default_domain(struct iommu_group *group)
 static int probe_iommu_group(struct device *dev, void *data)
 {
 	struct list_head *group_list = data;
-	struct iommu_group *group;
 	int ret;
 
-	/* Device is probed already if in a group */
-	group = iommu_group_get(dev);
-	if (group) {
-		iommu_group_put(group);
-		return 0;
-	}
-
 	ret = __iommu_probe_device(dev, group_list);
 	if (ret == -ENODEV)
 		ret = 0;
@@ -1840,11 +1857,6 @@ int bus_iommu_probe(const struct bus_type *bus)
 	LIST_HEAD(group_list);
 	int ret;
 
-	/*
-	 * This code-path does not allocate the default domain when
-	 * creating the iommu group, so do it after the groups are
-	 * created.
-	 */
 	ret = bus_for_each_dev(bus, NULL, &group_list, probe_iommu_group);
 	if (ret)
 		return ret;
@@ -1857,6 +1869,11 @@ int bus_iommu_probe(const struct bus_type *bus)
 		/* Remove item from the list */
 		list_del_init(&group->entry);
 
+		/*
+		 * We go to the trouble of deferred default domain creation so
+		 * that the cross-group default domain type and the setup of the
+		 * IOMMU_RESV_DIRECT will work correctly in non-hotpug scenarios.
+		 */
 		ret = iommu_setup_default_domain(group, 0);
 		if (ret) {
 			mutex_unlock(&group->mutex);
@@ -2191,6 +2208,21 @@ static int __iommu_device_set_domain(struct iommu_group *group,
 {
 	int ret;
 
+	/*
+	 * If the device requires IOMMU_RESV_DIRECT then we cannot allow
+	 * the blocking domain to be attached as it does not contain the
+	 * required 1:1 mapping. This test effectively excludes the device
+	 * being used with iommu_group_claim_dma_owner() which will block
+	 * vfio and iommufd as well.
+	 */
+	if (dev->iommu->require_direct &&
+	    (new_domain->type == IOMMU_DOMAIN_BLOCKED ||
+	     new_domain == group->blocking_domain)) {
+		dev_warn(dev,
+			 "Firmware has requested this device have a 1:1 IOMMU mapping, rejecting configuring the device without a 1:1 mapping. Contact your platform vendor.\n");
+		return -EINVAL;
+	}
+
 	if (dev->iommu->attach_deferred) {
 		if (new_domain == group->default_domain)
 			return 0;
@@ -3282,7 +3314,7 @@ static void __iommu_release_dma_ownership(struct iommu_group *group)
 
 /**
  * iommu_group_release_dma_owner() - Release DMA ownership of a group
- * @dev: The device
+ * @group: The group
  *
  * Release the DMA ownership claimed by iommu_group_claim_dma_owner().
  */
@@ -3296,7 +3328,7 @@ EXPORT_SYMBOL_GPL(iommu_group_release_dma_owner);
 
 /**
  * iommu_device_release_dma_owner() - Release DMA ownership of a device
- * @group: The device.
+ * @dev: The device.
  *
  * Release the DMA ownership claimed by iommu_device_claim_dma_owner().
  */
@@ -3479,3 +3511,30 @@ struct iommu_domain *iommu_sva_domain_alloc(struct device *dev,
 
 	return domain;
 }
+
+ioasid_t iommu_alloc_global_pasid(struct device *dev)
+{
+	int ret;
+
+	/* max_pasids == 0 means that the device does not support PASID */
+	if (!dev->iommu->max_pasids)
+		return IOMMU_PASID_INVALID;
+
+	/*
+	 * max_pasids is set up by vendor driver based on number of PASID bits
+	 * supported but the IDA allocation is inclusive.
+	 */
+	ret = ida_alloc_range(&iommu_global_pasid_ida, IOMMU_FIRST_GLOBAL_PASID,
+			      dev->iommu->max_pasids - 1, GFP_KERNEL);
+	return ret < 0 ? IOMMU_PASID_INVALID : ret;
+}
+EXPORT_SYMBOL_GPL(iommu_alloc_global_pasid);
+
+void iommu_free_global_pasid(ioasid_t pasid)
+{
+	if (WARN_ON(pasid == IOMMU_PASID_INVALID))
+		return;
+
+	ida_free(&iommu_global_pasid_ida, pasid);
+}
+EXPORT_SYMBOL_GPL(iommu_free_global_pasid);